2008-05-01
[17:16:00.0000] <mcarter__>
Hello

[17:17:00.0000] <mcarter>
In the Server-Sent-Events specification, the 301 Moved Permanently response says that the UA must reconnect to the new URI. What URI should be provided to the removeEventSource function to disconnect it? the original URI, or the redirected URI?

[17:19:00.0000] <Hixie>
mcarter: good question. send mail. :-)

[17:20:00.0000] <mcarter>
ok, will do

[17:21:00.0000] <Hixie>
(probably the origin one)

[17:21:01.0000] <mcarter>
Hixie, something else I need to mention on the list is that the terminology is a bit confusing. On the one hand you have the "event-source" dom element, but then you can call addEventSource on an event-source

[17:21:02.0000] <Hixie>
er, original

[17:22:00.0000] <Hixie>
you can call addEventSource on anything

[17:22:01.0000] <mcarter>
so when I see something like "(It doesn't affect other event sources with the same URI unless they also receive 301 responses, and it doesn't affect future sessions, e.g. if the page is reloaded.)" I don't know if "event sources" means the dom element, or one of the event sources attached to that dom element

[17:22:02.0000] <Hixie>
it means the event sources unless it is spelt event-sourcen orange

[17:22:03.0000] <Hixie>
er

[17:22:04.0000] <Hixie>
man i suck at typing today

[17:22:05.0000] <Hixie>
"it means the event sources unless it is spelt event-sourc in orange"

[17:23:00.0000] <mcarter>
Hixie, hmm, then that sentence seems to imply that you can call addEventSource(url) multiple times on the same event-source element with the same url

[17:25:00.0000] <Hixie>
yes

[17:26:00.0000] <Hixie>
<event-source> as an element is the same as <span>

[17:26:01.0000] <Hixie>
except for one thing

[17:26:02.0000] <Hixie>
which is that its src="" attribute causes addEventSource() and removeEventSource() to be called

[17:26:03.0000] <Hixie>
on itself

[17:26:04.0000] <Hixie>
we could in fact remove <event-source> altogether, it's just there for convenience

[17:28:00.0000] <mcarter>
well, what I'm getting at is that the spec seems to imply that a dom element that implements the RemoteEventTarget interface can have its addEventSource function called twice with an identical url. I'm not sure what that would mean -- that you make two connections to the same URI? how would you close a particular one?

[17:30:00.0000] <Hixie>
the spec doesn't provide a way to remove a particular one

[17:30:01.0000] <Hixie>
it'll remove one each time you call removeEventSource(), but you can't control which one

[17:31:00.0000] <mcarter>
ok

[17:49:00.0000] <Hixie>
so much for a non-confidential xhtml2 wg

[17:49:01.0000] <Hixie>
http://lists.w3.org/Archives/Member/w3c-html-wg/2008AprJun/

[17:49:02.0000] <andersca>
ha

[17:50:00.0000] <Hixie>
oh ok it was a mistake

[17:50:01.0000] <Hixie>
nevermind

[17:51:00.0000] <Hixie>
(http://lists.w3.org/Archives/Public/wai-xtech/2008Apr/0404.html)

[18:02:00.0000] <Philip`>
/me wonders if it's intentional that HTML5 almost always uses the passive voice, like in "The values of the data array may be changed", instead of saying something like "Authors may change the values of the data array" that is often less ambiguous

[18:08:00.0000] <Dashiva>
Maybe to avoid overspecifying

[18:08:01.0000] <Dashiva>
Like, is it possible to change the value without being an author?

[18:09:00.0000] <Philip`>
Treat it as an abbreviation of "authors may write documents which cause scripts to execute which change the values of the data array" (and define that abbreviation somewhere global)

[18:10:00.0000] <Philip`>
It's underspecified now, because the UA might decide to arbitrarily change the data array at some random point in time, and argue that that's allowed because the spec says 'the values may be changed'

[18:11:00.0000] <Philip`>
(ignoring the issue that "may" is inappropriate here - pretend I'm talking about a different case that uses "must" or something :-) )

[18:12:00.0000] <Philip`>
(like "ImageData objects must be initialised so that ..." - who must do that? There's only one possible answer given the context, but you shouldn't have to deduce the meaning that's implicit in the context)

[18:13:00.0000] <Philip`>
Anyway, I'm not sure if I'm just being uselessly picky here - it's not going to cause problems in practice, it's just a bit unclear when trying to read the spec in detail, and maybe there are reasons to write it like how it's written now (e.g. 'it looks ugly and hard to read otherwise')

[18:14:00.0000] <Philip`>
...and those reasons could be more significant than the trivial problems

[19:01:00.0000] <Philip`>
Hmm, maybe I shouldn't focus so much on testing the ability to draw solid lime green rectangles - WebKit seemed to be doing really quite well on my ImageData tests, until I accidentally found that it actually totally breaks if you draw some data with alpha!=1 and more than one colour of pixels...

[19:08:00.0000] <Hixie>
Philip`: see if you can track down teh bug that causes teh whatwg issues graph to change colour when you hover over it

[19:08:01.0000] <Hixie>
and the spec using the passive voice is just that my style, it's probably not ideal

[19:09:00.0000] <Hixie>
if there are cases where it is genuinely ambiguous let me know

[19:28:00.0000] <Philip`>
Hixie: I don't see what you mean about it changing colour

[19:28:01.0000] <Hixie>
originally the bars in the background are yellow

[19:28:02.0000] <Hixie>
after the putImageDate(), the turn gray

[19:29:00.0000] <Philip`>
Hixie: I don't think I've seen any cases where it's actually ambiguous, so I'm only moaning about it on IRC for now :-)

[19:29:01.0000] <Hixie>
:-)

[19:29:02.0000] <Philip`>
Hixie: In which browser(s)?

[19:29:03.0000] <Hixie>
webkit trunk

[19:30:00.0000] <Philip`>
Looks like it could be the same issue as I found, if you're putImageDataing things with transparent pixels

[19:30:01.0000] <Philip`>
https://bugs.webkit.org/show_bug.cgi?id=18821

[19:30:02.0000] <Hixie>
k

[19:30:03.0000] <takkaria>
MikeSmith: ping

[19:31:00.0000] <Philip`>
(It seems to copy the colour data from the leftmost pixel, and mixes it with the alpha from the proper pixels)

[19:31:01.0000] <Philip`>
(and if the leftmost pixel is transparent then it'll be black/grey)

[19:32:00.0000] <Philip`>
(I haven't got a clue how they could have implemented such a bug)

[19:33:00.0000] <Hixie>
off by one byte?

[19:34:00.0000] <Philip`>
It's not just one byte - the colour from the leftmost pixel propagates infinitely far rightwards

[19:35:00.0000] <Hixie>
oh

[19:35:01.0000] <Hixie>
wtf

[19:35:02.0000] <Hixie>
actually that's easier

[19:35:03.0000] <Hixie>
probably a misinitialised variable

[19:35:04.0000] <Hixie>
as in, something that is scoped to a different scope than they realise

[19:35:05.0000] <Philip`>
It's also odd that it only happens when alpha != 1

[19:36:00.0000] <Philip`>
since I don't see why there'd be any point in special-casing that case on purpose

[19:37:00.0000] <Philip`>
But I'll bet they fixed it six days ago anyway, and just haven't got a new nightly build out since then :-)

[19:38:00.0000] <Hixie>
heh

[19:55:00.0000] <MikeSmith>
takkaria: wanted to ask you about progress (if any) on your html5lib porting

[19:59:00.0000] <takkaria>
MikeSmith: well, it's not porting, but not started yet, no

[19:59:01.0000] <MikeSmith>
OK

[22:14:00.0000] <MikeSmith>
takkaria: you still around?

[22:16:00.0000] <takkaria>
MikeSmith: yes

[22:17:00.0000] <MikeSmith>
takkaria: wanted to ask if you're on the whatwg implementors mailing list

[22:17:01.0000] <takkaria>
MikeSmith: yup

[22:17:02.0000] <MikeSmith>
ah, OK

[22:17:03.0000] <takkaria>
have been since it started, I think

[22:20:00.0000] <takkaria>
actually I'm waiting on ink cartridges to arrive so I can print out the relevant chunks of the spec and read them through with marker pen

[22:20:01.0000] <takkaria>
then I intend to start coding

[22:21:00.0000] <MikeSmith>
I just asked about the list because I noticed messages from Edward Yang

[01:16:00.0000] <Hixie>
http://juicystudio.com/article/html5-alt-text-authoring-tools.php

[01:16:01.0000] <Hixie>
apparently my attempts at improving accessibility are seen as an attempt to make html5 appear successful

[01:16:02.0000] <Hixie>
and our attempts at doing and working based on real research is seen as misdirection

[01:20:00.0000] <othermaciej>
Hixie: I apologies for turning the icon size discussion into a bikeshed

[01:21:00.0000] <Hixie>
no worries

[01:21:01.0000] <Hixie>
my working model handles bikesheds very easily :-)

[01:27:00.0000] <Hixie>
/me comments on that link

[02:36:00.0000] <jgraham_>
Wow, Gez's article is surprisingly uninteresting given how much publicity it's had (well one link in IRC and 2 emails on public-html)

[02:38:00.0000] <jgraham_>
I was expecting new arguments, instead it was just the same stuff but with the accusation that people who don't agree that alt should be mandatory are being disingenuous 

[02:45:00.0000] <Lachy>
wow, he seems to be arguing that alt should still be omitted when an authoring tool has nothing useful to put in it, but that it should be considered invalid.

[02:47:00.0000] <Philip`>
Why is that a bad idea?

[02:48:00.0000] <jgraham_>
Philip`: Because it encourages people to stuff alt with meaningless values to make their pages valid

[02:49:00.0000] <Philip`>
It encourages people to set alt values to make their pages valid - some of those may be meaningless, but others may be meaningful and worthwhile

[02:50:00.0000] <Philip`>
(assuming "people" means "authors")

[02:50:01.0000] <Philip`>
(If "people" means "authoring tool developers", then they're not capable of meaningful text, in which case it is a problem)

[02:51:00.0000] <othermaciej>
requiring software to generate nonconforming content in some situations does not make sense as a conformace requirement

[02:51:01.0000] <jgraham_>
It's not quite clear which way that one would swing but, with a site like flickr, if there was an option to set alt text, I would expect the values to be harmful more often than they were helpful

[02:52:00.0000] <jgraham_>
(assuming flickr provided the box out of a desire to allow pages to be conforming)

[02:52:01.0000] <othermaciej>
I have been trying to use VoiceOver more lately to play with some of WebKit's new accessibility stuff

[02:53:00.0000] <othermaciej>
perhaps I will get a sense of how annoying repeated or bad alt values are

[02:53:01.0000] <othermaciej>
(by "repeated" I mean repeats something in the real page content)

[02:53:02.0000] <othermaciej>
jgraham_: it probably would get used a fair bit less than the title or description fields

[02:53:03.0000] <othermaciej>
I would guess a minority of photos have both title and description set

[02:54:00.0000] <jgraham_>
othermaciej: I would expect that at best people would just copy the description into the field

[02:54:01.0000] <jgraham_>
s/field/alt field/

[02:55:00.0000] <othermaciej>
well I think on flickr gallery pages, "go to photo page" may be better alt text than a description in any case

[02:55:01.0000] <othermaciej>
open question whether more description would help a substantive portion of the individual photo pages

[02:56:00.0000] <jgraham_>
Yeah, on the gallery page "go to photo $photo_title" might be the best alt text

[02:56:01.0000] <jgraham_>
I was thinking of the individual photo pages

[03:06:00.0000] <jgraham_>
I can't help but think the required alt thinking is tied to the idea that accessibility only happens as a result of people wielding a stick to make it happen. Therefore taking away anything that can be used as a stick is perceived as bad even if that change also has the potential to create an environment where accessibility is better in the default no-stick case.

[03:07:00.0000] <jgraham_>
But I should stop thinking such things until I get around to finishing the data collection tool thing

[03:07:01.0000] <jgraham_>
:)

[03:41:00.0000] <htmlfivedotnet>
Outsiders opinion: Someone who knows what their doing will use the alt tag correctly, many people don't even know its purpose, which is why garbage gets jammed in. No alt tag should be valid html, but not considered accessible... which may cause someone to figure out what that means, and fix it. 

[03:42:00.0000] <htmlfivedotnet>
*missing alt tag, rather.

[06:47:00.0000] <Philip`>
http://www.adobe.com/openscreenproject/developers/ - SWF and FLV specs now with no licensing restrictions

[06:58:00.0000] <othermaciej>
hmm, I can't actually tell how f4v differs from m4v from this

[07:00:00.0000] <roc>
that's cool

[07:03:00.0000] <Philip`>
Sounds mostly like F4V is just a profile of MP4

[07:08:00.0000] <othermaciej>
I think I would need to be an expert on MPEG to understand if what they are describing is different

[07:09:00.0000] <othermaciej>
I think they are redefining parts of the various MP4 container format specs

[07:09:01.0000] <Philip`>
It'd be useful if there was an MPEG experts group that could be asked that kind of thing

[07:09:02.0000] <othermaciej>
but I can't tell if it is actually compatible with real mpeg 4

[07:10:00.0000] <othermaciej>
some sort of Motion Picture Experts Group maybe?

[07:10:01.0000] <Philip`>
Yeah, someone should form one of those

[07:22:00.0000] <Philip`>
http://www.kaourantin.net/2007/10/new-file-extensions-and-mime-types.html seems to indicate that Flash ignores the filename and ignores the ftyp=F4V field, and treats them the same as any other MP4 file

[12:20:00.0000] <Hixie>
htmlfivedotnet: that's what the spec says at the moment :-)

[12:58:00.0000] <Hixie>
jgraham_: btw i'm happy to help with the "choice of sites" problem

[12:59:00.0000] <Hixie>
jgraham_: i can trivially get you a list of urls for pages that satisfy a particular criteria

[12:59:01.0000] <Hixie>
just tell me how many, and what criteria

[13:07:00.0000] <Hixie>
aw man, i fell into the trap

[13:08:00.0000] <Hixie>
and i was being so careful

[13:09:00.0000] <Philip`>
When near a trap, 'being careful' usually means getting as far away as possible, rather than thoughtfully and slowly and delicately poking it until it rips your arm off

[13:10:00.0000] <Hixie>
i had been just not replying to the e-mails

[13:11:00.0000] <maikmerten>
Hixie: do you happen to have a spare minute or two? ;-)

[13:11:01.0000] <Hixie>
sure

[13:11:02.0000] <maikmerten>
ah, thanks

[13:11:03.0000] <maikmerten>
well, I am pretty curious what you meant with your blog comment reading as follows: "The Theora codec isn't unencumbered, despite popular opinion; it just hasn't had anyone claim patents on it publicly yet."

[13:12:00.0000] <maikmerten>
this sounds like patents actually *surfaced*

[13:12:01.0000] <Hixie>
i don't know of any specific patents myself

[13:12:02.0000] <maikmerten>
and that you're just leaning back to see the train wreck ignite ;)

[13:12:03.0000] <maikmerten>
so you were referring to the "usual" submarine threat

[13:12:04.0000] <maikmerten>
(I assume?)

[13:12:05.0000] <Hixie>
but i am told by video experts that the "submarine" threat is not really hypothetical in the case of theora

[13:12:06.0000] <Hixie>
er

[13:12:07.0000] <Hixie>
note that theora

[13:13:00.0000] <Hixie>
note that theora hasn't been designed in a way to avoid patents or anything

[13:13:01.0000] <Hixie>
(unlike vorbis)

[13:13:02.0000] <maikmerten>
that's not quite true, On2 makes a living out of evading patents

[13:13:03.0000] <maikmerten>
it's their business model

[13:13:04.0000] <Hixie>
i can't speak to that

[13:13:05.0000] <Hixie>
i'm just saying that from a patent point of view, theora isn't clean enough

[13:14:00.0000] <Hixie>
not that anything else is any better

[13:14:01.0000] <Hixie>
but theora isn't the silver bullet people make it out to be

[13:14:02.0000] <maikmerten>
well, I wonder why those "video experts" don't just speak out

[13:14:03.0000] <Hixie>
because there are legal implications to claiming knowledge of patents

[13:14:04.0000] <maikmerten>
*if* there are problems they should just go public with it

[13:14:05.0000] <jgraham_>
Hixie: Great. I need to find a little time to work on it...

[13:15:00.0000] <Hixie>
e.g. if you know that a patent exists and then you violate it, your liability triples, as i understand it

[13:15:01.0000] <maikmerten>
nobody benefits from Theora adopters getting hurt *if* there are problems

[13:15:02.0000] <maikmerten>
and it would be interesting to see who those experts are anyway

[13:15:03.0000] <Hixie>
the law around patents are non-intuitive

[13:15:04.0000] <Hixie>
you can get in trouble just for knowing things

[13:15:05.0000] <maikmerten>
note that "audio experts" claimed the same for Vorbis

[13:15:06.0000] <Hixie>
it's pretty ridiculous

[13:15:07.0000] <Hixie>
again, i can't speak to that

[13:16:00.0000] <Hixie>
the long and short of it is that theora is just as bad as everything else, except for being vendor-specific (not a standard) and not quite as technically good as H.264

[13:16:01.0000] <Hixie>
Sun's H.261 work looks interesting though

[13:16:02.0000] <Hixie>
if that goes anywhere

[13:17:00.0000] <maikmerten>
well, if they change anything about H.261 they're facing the same threat

[13:17:01.0000] <maikmerten>
but it's interesting, yes

[13:17:02.0000] <maikmerten>
(what makes them different is that people assume they *have* lawyers)

[13:18:00.0000] <maikmerten>
and as for being vendor-specific: The spec is in the public domain, so how much more libre can it go anyway

[13:18:01.0000] <Philip`>
(They also have liability if their lawyers make the wrong decisions, so there's a significant incentive for them to be cautious)

[13:18:02.0000] <Hixie>
well, it could be an ISO standard :-)

[13:19:00.0000] <maikmerten>
and for not being as good as H.264 - well, that's not the reference to compare to when talking about royality free standards 

[13:19:01.0000] <maikmerten>
just like OOXML? ;-)

[13:19:02.0000] <maikmerten>
well, HTML could also be an ISO standard

[13:19:03.0000] <Hixie>
it is

[13:19:04.0000] <maikmerten>
so, and what does that change ;)

[13:19:05.0000] <Hixie>
ISO/IEC 15445:2000(E)

[13:19:06.0000] <maikmerten>
yup

[13:20:00.0000] <Hixie>
i'm just saying that a public domain spec by one vendor is not the same as an internationally recognised standard

[13:20:01.0000] <Philip`>
The existence of ISO-HTML doesn't suggest that ISO has extremely high standards of detailed interoperable specifications

[13:20:02.0000] <Hixie>
i agree that H.264 has known patent claims made on it

[13:21:00.0000] <Hixie>
though they may expire in the coming years, possibly before html5 is complete, even

[13:21:01.0000] <Hixie>
(at least for 264 baseline)

[13:21:02.0000] <maikmerten>
one could argue that being an ISO standard can in some cases just be a "checkbox feature", without real practical consequences

[13:21:03.0000] <Hixie>
one could

[13:21:04.0000] <Hixie>
but tell me

[13:21:05.0000] <Philip`>
By the time H.264's patents have expired, it'll be as technically worthless as H.261

[13:22:00.0000] <Hixie>
did you prefer <canvas> when it was apple's thing, or when it got added to html5?

[13:22:01.0000] <maikmerten>
ISO doesn't guarantee quality, nor does it imply that everybody can implement it

[13:22:02.0000] <Hixie>
Philip`: most of the important baseline patents may expire within a couple of years

[13:22:03.0000] <maikmerten>
however, there are policies in many organizations calling for ISO standards, so that's why it definately is a "good thing" to use one

[13:23:00.0000] <Philip`>
Hixie: Unfortunately people can still be sued over a single barely-important patent in an optional feature that everyone happens to rely on

[13:23:01.0000] <Hixie>
maikmerten: people will always be suspicious of vendor-controlled "standards"

[13:23:02.0000] <maikmerten>
too bad the ISO video coding standards fall short somewhere else

[13:23:03.0000] <hsivonen>
(work hat off): I'd be interested in pointers to criteria for evaluating standards for goodness in a way that doesn't involve checking for organizational logos

[13:24:00.0000] <Hixie>
Philip`: yeah, the idea would be to use baseline only

[13:24:01.0000] <Philip`>
hsivonen: You could ask some guys on IRC what they think of it

[13:25:00.0000] <hsivonen>
Philip`: so far the best criteria I have seen are Sam's from http://www.intertwingly.net/blog/2007/01/25/Pro-Choice

[13:25:01.0000] <hsivonen>
"A standard is one that has multiple, inter-operable, independent implementations.  An open standard, at least in the software world, is one where at least one of those implementations is open source."

[13:25:02.0000] <Hixie>
that seems like a dodgy definition

[13:26:00.0000] <Hixie>
but ok

[13:26:01.0000] <hsivonen>
Philip`: (I think I should write something that can be given to politicians. and for that use case, telling them to ask on IRC isn't good)

[13:26:02.0000] <maikmerten>
Hixie: well, I just don't see how being public domain qualifies as "vendor-controlled"

[13:26:03.0000] <maikmerten>
however, it may be a case of "not enough control"

[13:27:00.0000] <Hixie>
maikmerten: who decides what the spec says?

[13:27:01.0000] <hsivonen>
Hixie: I think the definition should also require that you can write the next interoperable independent impl. by following the spec

[13:27:02.0000] <Hixie>
maikmerten: public domain means that anyone can copy it, not that the official version is under anyone's control

[13:27:03.0000] <Hixie>
hsivonen: indeed

[13:27:04.0000] <maikmerten>
Hixie: the official Theora spec: Xiph.org (which is a 403 organization)

[13:28:00.0000] <Hixie>
403 Forbidden? :-)

[13:28:01.0000] <hsivonen>
Forbidden? :-)

[13:28:02.0000] <hendry>
heh

[13:28:03.0000] <Hixie>
maikmerten: and who changes it?

[13:28:04.0000] <Philip`>
Hixie: Sounds like that still wouldn't be technically competitive with whatever the state-of-the-art would be, and technical competitiveness is still important for video (since it's low quality and high bandwidth and therefore worth optimising), which is kind of a pain in terms of convincing people it's worth using

[13:28:05.0000] <Hixie>
maikmerten: who decides what an error is, etc?

[13:28:06.0000] <Philip`>
s/that/H.264 baseline/

[13:28:07.0000] <maikmerten>
err... 403... dammit... I meant 501(c)(3)

[13:29:00.0000] <Hixie>
Philip`: yes, i don't think you'll be able to have a patent unencumbered or royalty-free state of the art video codec before international patent reform.

[13:29:01.0000] <maikmerten>
Hixie: "the xiph.org community" - which, however, is as open as it can get

[13:30:00.0000] <maikmerten>
Hixie: however, the point you may be pointing at is "so, how transparent is this?"

[13:30:01.0000] <maikmerten>
and sure, that's where it may look not as nice as ISO, no doubt about that

[13:32:00.0000] <maikmerten>
and, no doubt about that: If ISO would have a standard with fitting licensing we wouldn't have all these problems

[13:32:01.0000] <Philip`>
Before such reform, hopefully browsers will start plugging into their platform's media functionality so I can watch high-quality video (in any browser, in any practical OS, with only a trivial cost added to the purchase of the computer for codec licensing) and the browser vendors won't have to worry about the patents

[13:32:02.0000] <maikmerten>
(and Theora wouldn't even exist)

[13:32:03.0000] <Hixie>
maikmerten: so what happens if microsoft comes along and asks for something to change in the spec?

[13:32:04.0000] <maikmerten>
but apparently there *is* no international standards body really intersted in royality free standards

[13:32:05.0000] <Hixie>
Philip`: yeah that seems to be happening

[13:33:00.0000] <Philip`>
(Sadly my version of Vista doesn't even include DVD codecs, so I guess it'll be a while before H.264 is available as standard)

[13:33:01.0000] <maikmerten>
Hixie: well, what if Microsoft wants a nicely crafted standard to become an ISO standard? ;-)

[13:34:00.0000] <Philip`>
(but that's okay because people can just write a Flash implementation of <video>)

[13:34:01.0000] <Hixie>
maikmerten: believe me, i'm as upset as anyone over the debacle that was ooxml

[13:34:02.0000] <maikmerten>
Hixie: and believe me, if there were a free ISO standard I'd be all for it

[13:34:03.0000] <maikmerten>
Hixie: I can see the benefits of ISO standards

[13:35:00.0000] <maikmerten>
more transparency

[13:35:01.0000] <Hixie>
i dunno about transparency

[13:35:02.0000] <Hixie>
but yeah

[13:35:03.0000] <maikmerten>
less dependency on single entities

[13:35:04.0000] <Hixie>
i'd love to use theora, and indeed the spec used to require it

[13:35:05.0000] <maikmerten>
well, at least for members, I guess

[13:35:06.0000] <Hixie>
it's just not a practical option

[13:35:07.0000] <Hixie>
big vendors won't implement it

[13:35:08.0000] <Hixie>
at the end of the day it doesn't matter what their reason

[13:35:09.0000] <Hixie>
is

[13:36:00.0000] <Hixie>
if the big vendors don't implement it, we don't get interop, and we've failed

[13:36:01.0000] <maikmerten>
that's very true, I guess

[13:36:02.0000] <maikmerten>
well, too bad nobody actually discloses anything

[13:36:03.0000] <Hixie>
as i keep telling people, we spec writers are only powerful so long as we tell the vendors to do what they want to do anyway :-)

[13:36:04.0000] <Hixie>
yeah

[13:36:05.0000] <Hixie>
the video thing is a pain

[13:37:00.0000] <Hixie>
but there is progress being made, albeit very slowly

[13:37:01.0000] <Hixie>
very, very slowly

[13:37:02.0000] <maikmerten>
if somebody actually *knew* flaws in the Theora spec it would help to communicate this

[13:37:03.0000] <Philip`>
Hixie: That's not necessarily failure - the big vendor could become a small vendor, because all their users leave for browsers that do implement the spec, and then it'd be a success

[13:37:04.0000] <maikmerten>
so the spec can be revised and adoption of the flawed iteration can be discouraged

[13:38:00.0000] <Hixie>
Philip`: also, god could exist

[13:38:01.0000] <maikmerten>
all what is visible is information from "video experts" leaking, which could very well just be staged (nah, that doesn't mean you Hixie)

[13:38:02.0000] <Hixie>
Philip`: i'm not praying for either :-)

[13:39:00.0000] <Hixie>
maikmerten: yeah, i wish people were more forthcoming

[13:39:01.0000] <Hixie>
maikmerten: but like i said, there are legal reasons why people can't say what they know

[13:39:02.0000] <hsivonen>
btw, the Finnish public broadcaster is switching from WMV9 to H.264 now that Flash does H.264

[13:39:03.0000] <maikmerten>
Hixie: yeah, and it's sad the whole system is that broken

[13:39:04.0000] <Hixie>
maikmerten: i've been in meetings about stuff like this where everyone knows something but everyone is under different NDAs about it and so everyone pretends to everyone else not to know it

[13:40:00.0000] <maikmerten>
Hixie: anyway, thanks alot for taking the time and elaborating

[13:40:01.0000] <Hixie>
maikmerten: and despite the fact that each person in the room has privately told me that they know of it, they can't speak of it

[13:40:02.0000] <Hixie>
maikmerten: bugs the hell out of me

[13:40:03.0000] <Hixie>
maikmerten: np

[13:40:04.0000] <maikmerten>
I see

[13:40:05.0000] <maikmerten>
scary

[13:40:06.0000] <Hixie>
yeah

[13:40:07.0000] <Lachy>
JohnResig, yt?

[13:40:08.0000] <Philip`>
Hixie: If HTML5 drags on for as long as you say it will, there's no chance that the current big vendors will still be the big vendors by that time, and probably not a great chance that they'll all even exist :-)

[13:41:00.0000] <Hixie>
Philip`: i don't think microsoft, nokia, apple, and other such companies are likely to quit being relevant in the next 20 years

[13:43:00.0000] <Lachy>
JohnResig, re your selectors api feedback, I'm not convinced it's a good idea to change the spec so significantly as you suggest. I designed it the way it is for a reason.

[13:43:01.0000] <Philip`>
Hixie: It's not about being generally relevant, it's about being relevant in the context of developing web browsers, which is a much more fragile thing since it won't survive when an organisation tries to significantly reposition itself in the market

[13:44:00.0000] <Lachy>
JohnResig, however, I'm am convinced we need to define :scope sooner rather than later, which would solve all use cases you presented without sacrificing any features

[13:44:01.0000] <Hixie>
Philip`: the only way e.g. microsoft is getting out of the browser game is if they beat us with one of their own proprietary platforms

[13:44:02.0000] <Hixie>
Philip`: in which case, it doesn't matter what we say

[13:46:00.0000] <Philip`>
Hixie: They got out of the browser game after IE6, and that wasn't after having beaten the web platform

[13:46:01.0000] <Hixie>
they didn't get out of the browser game, they had the entire market

[13:47:00.0000] <Hixie>
indeed, arguably, it _was_ after having beaten the web platform

[13:47:01.0000] <Hixie>
though not in the sense i meant earlier

[13:47:02.0000] <Philip`>
Stopping all development doesn't sound particularly like being in the game

[13:47:03.0000] <Hixie>
it's a funny way to play the game

[13:47:04.0000] <Hixie>
it worked for them for several years

[13:49:00.0000] <Philip`>
If you're playing a game and then press the Pause key for several years, are you still playing it all that time?

[13:49:01.0000] <Philip`>
If so, I've played far too many games while I've been asleep in bed :-(

[13:49:02.0000] <jgraham_>
Philip`: They were still in the game, they just thought they didn't need to play anymore. I guess they regret that now

[13:52:00.0000] <Philip`>
Anyway, I still really very much hope we won't still be using IE and Firefox and Safari and Opera in 2028 :-)

[13:54:00.0000] <Hixie>
i see no evidence to suggest we'll be using anything else

[13:55:00.0000] <Philip`>
You could look at what web browsers we were using in 1988, and assume (in the absence of evidence otherwise) that things will change just as significantly in the future

[13:55:01.0000] <jgraham_>
Philip`: Not a fair comparison. But it's reasonable to say we're not using much of the Netscape 1 codebase anymore

[13:55:02.0000] <Hixie>
IE, Netscape, and Opera?

[13:55:03.0000] <Philip`>
jgraham_: Why is that unfair?

[13:55:04.0000] <Hixie>
there's one new browser

[13:56:00.0000] <gsnedders>
/me tries to remember 1988

[13:56:01.0000] <gsnedders>
/me fails

[13:56:02.0000] <Hixie>
oh, 1988

[13:56:03.0000] <Hixie>
not 1998

[13:56:04.0000] <Philip`>
Comparing to 1998 would be unfair :-)

[13:56:05.0000] <Philip`>
and would destroy my argument

[13:56:06.0000] <jgraham_>
1988 was before the web existed, so the legacy didn't exist at all

[13:56:07.0000] <Hixie>
the first few years of any technology is radical

[13:56:08.0000] <Hixie>
we're in a much more mature stage now

[13:56:09.0000] <Hixie>
compare word processors now to 20 years ago

[13:56:10.0000] <Hixie>
there's not much difference

[13:56:11.0000] <Hixie>
word perfect got replaced by pages

[13:57:00.0000] <Hixie>
that's about it

[13:57:01.0000] <Philip`>
Compare floppy disks now to 20 years ago

[13:57:02.0000] <Lachy>
Hixie, sure there is. MS Word is a lot more bloated now.

[13:57:03.0000] <Hixie>
still big companies, etc

[13:57:04.0000] <Hixie>
Lachy: :-)

[13:57:05.0000] <Hixie>
ok for canvas text my proposal is:

[13:57:06.0000] <Philip`>
Floppies have gone past the mature stage and died out, despite the legacy of billions of disks with software and data on them

[13:57:07.0000] <Philip`>
(where "billions" is an utterly made up number)

[13:57:08.0000] <hsivonen>
including data in WriteNow files

[13:57:09.0000] <Hixie>
drawHString(x, y, maxWidth, textAlign, s); and drawHString(x, y, maxHeight, textAlign, s);

[13:57:10.0000] <Philip`>
(but it's a reasonable extrapolation from how many I have at home)

[13:58:00.0000] <Hixie>
er

[13:58:01.0000] <Hixie>
drawVString(...) for the second one

[13:58:02.0000] <jgraham_>
/me chooses COBOL and mainframes as his legacy technologies

[13:58:03.0000] <jgraham_>
(that still exist as they did 20 years ago)

[13:58:04.0000] <Hixie>
Philip`: we'll see :-)

[13:58:05.0000] <Hixie>
Philip`: i don't think designing html5 with the assumption that the current players will be replaced would work, though (c.f. xhtml2)

[13:59:00.0000] <Hixie>
comment on drawHString() and drawVString() while i go have lunch. :-D

[13:59:01.0000] <Hixie>
bbl

[14:00:00.0000] <Lachy>
what's the difference between them? drawVString for vertical stings where the letters are stacked on top of each other, and not just rotated 90 deg?

[14:01:00.0000] <hsivonen>
/me finds http://www.legacyj.com/lgcyj_perc1.html

[14:01:01.0000] <Philip`>
Hixie: They look complex and hard to use :-p

[14:01:02.0000] <Philip`>
compared to e.g. translate(x,y);drawString(s)

[14:02:00.0000] <Philip`>
Anyway, it'd be useful to know what use cases they're meant to be handling

[14:05:00.0000] <Lachy>
Philip`, labelling graphs would be a typical use case for text, which is currently achieved by positioning elements over the canvas

[14:05:01.0000] <Philip`>
and the interesting bits are in fonts and styles and sizes and measuring

[14:06:00.0000] <Philip`>
which are currently unspecified and therefore hard to comment on :-)

[14:06:01.0000] <Lachy>
the problem it solves is that it makes the text actually part of the image, and so it can be exported and saved without losing the text labels

[14:07:00.0000] <Philip`>
Lachy: Simplicity of implementation sounds like a more relevant concern in that case - I've heard people complaining that they need complicated hacks to do text, but I don't think I've ever heard them complaining that it makes exporting images hard

[14:08:00.0000] <Lachy>
complexity is just another problem it solves

[14:09:00.0000] <Lachy>
(assuming the solution we creation isn't too complex)

[14:09:01.0000] <Lachy>
*create

[14:12:00.0000] <Philip`>
My semi-practical use case is displaying characters' names above their heads in an FPS game, with the name being obscured by closer walls (hence it's not possible to overlay an HTML element, except with horribly evil clipping hacks)

[14:13:00.0000] <Philip`>
What are the use cases for writing large unformatted chunks of wrapped text?

[14:17:00.0000] <Philip`>
(Oh, I suppose I also want text for an FPS game's HUD, mostly just with big numbers for the health and ammo counts in a nice font)

[14:44:00.0000] <JohnResig>
Lachy: you should probably say as much on the mailing list, so we can discuss it there

[14:57:00.0000] <htmlfivedotnet>
Hixie: What about: drawString(s, x, y, (align='horizontal'), (maxSize='null'), (textAlign='left') ) - if that even makes sense...

[15:20:00.0000] <Hixie>
lachy: drawVString() would be for vertical text (e.g. some CJK)

[15:21:00.0000] <Hixie>
htmlfivedotnet: yeah, i considered that, but i didn't want to have the author have to put in a long keyword for the direction

[15:22:00.0000] <Hixie>
also you're not going to be switching between them, really

[15:22:01.0000] <Hixie>
Philip`: the x and y arguments are in line with how the rest of the API takes coordinates

[15:22:02.0000] <Hixie>
Philip`: you don't translate then rect(width, height), e.g.

[15:23:00.0000] <Philip`>
Hixie: Ah, yes, that's a confusing part of the mozDrawText API

[15:23:01.0000] <Philip`>
so I'll modify my proposal to drawString(s, x, y) :-)

[15:24:00.0000] <htmlfivedotnet>
my thought was just that certain properties would be used more frequently, and having defaults would keep it cleaner unless it needed to be more complex. 

[15:24:01.0000] <Philip`>
(plus a couple of properties that I'm not mentioning now)

[15:24:02.0000] <Hixie>
well, there are three things we want to avoid

[15:24:03.0000] <Hixie>
one is lack of support for vertical text :-)

[15:24:04.0000] <htmlfivedotnet>
and i personally hate having 3 functions that do essesntially the same thing... ala PHP

[15:24:05.0000] <Philip`>
One is complexity, given that graph axis labelling is about the only use case

[15:24:06.0000] <Hixie>
another is text being affected by different font metrics and being unviewable

[15:25:00.0000] <Hixie>
and the other is making sure graph axes can be lined up correctly (so we need left/right/center alignment)

[15:25:01.0000] <Hixie>
i don't think  drawHString(x, y, width, 'left', 'text')  is really that complex

[15:26:00.0000] <Philip`>
For graphs you need top/center/bottom too (measured at 'sensible' points like ascent height / median / baseline, not on the bounding box)

[15:26:01.0000] <Hixie>
hmm

[15:26:02.0000] <Hixie>
yes

[15:26:03.0000] <Hixie>
good point

[15:27:00.0000] <Philip`>
I'd like to see a labelled graph using vertical CJK...

[15:27:01.0000] <Hixie>
no idea how to search for that

[15:28:00.0000] <Philip`>
Look at lots of graphs. If you don't find any vertical CJK ones, it's not a "95% of use cases" situation and isn't worth the complexity :-)

[15:29:00.0000] <Hixie>
i have no idea how to avoid sample bias

[15:29:01.0000] <Hixie>
for this case

[15:29:02.0000] <Philip`>
(I'd guess they'd use LTR decimal numbers anyway)

[15:33:00.0000] <Philip`>
How do you suggest setting font family/style/weight/size?

[15:33:01.0000] <htmlfivedotnet>
http://www.honda.co.jp/PARTNER/SP/environment/image/bargraph_image.gif

[15:33:02.0000] <Hixie>
Philip`: .fontStyle = ''

[15:33:03.0000] <Hixie>
htmlfivedotnet: ooo

[15:33:04.0000] <htmlfivedotnet>
lol

[15:33:05.0000] <Philip`>
What happens when text is wider than 'width'?

[15:34:00.0000] <Hixie>
Philip`: ua tries condensing the font, first using actual condensed fonts, then using horizontal scaling, and if a reasonable amount of that doesn't help either, shrinks the font size

[15:35:00.0000] <Philip`>
Can I set width == Infinity when I don't want the browser to mess with things?

[15:35:01.0000] <Hixie>
why would you ever want it to go off the side of the viewport

[15:35:02.0000] <Hixie>
?

[15:36:00.0000] <Hixie>
http://www.osaka-gaidai.ac.jp/houzinnsosiki_kakudai.html has vertical text in a chart

[15:36:01.0000] <Hixie>
http://www.lac.co.jp/img/siodome_j.gif too, i think

[15:37:00.0000] <Hixie>
the rightmost blue block (train stop?) looks like it has vertical glyphs

[15:37:01.0000] <Hixie>
though they are rotated to the point of almost being horizontal :-/

[15:37:02.0000] <Philip`>
Hixie: I'd want that when writing Google Maps in <canvas>, where text is off the side of the screen and the user can scroll or zoom to read it

[15:37:03.0000] <svl>
/join #ubuntu

[15:37:04.0000] <svl>
urgh

[15:38:00.0000] <Hixie>
Philip`: even with google maps i'd wager we always want the text to be within a sane bounding box

[15:39:00.0000] <Philip`>
Hixie: That siodome_j.gif would need both max-width and max-height, else the text might end up overlapping its containing box in an ugly way

[15:39:01.0000] <htmlfivedotnet>
Hixie: so the 'width' property is a maximum, not an absolute?

[15:39:02.0000] <Dashiva>
It's a suggestion

[15:40:00.0000] <Philip`>
Hixie: Google Maps lets me go hundreds of thousands of kilometres sideways, and the text always stays static in relation to the background

[15:40:01.0000] <Hixie>
Philip`: well you set the font size, so you know the likely width (given that they're square and there's no line wrapping)

[15:40:02.0000] <Hixie>
htmlfivedotnet: max width, right

[15:40:03.0000] <Hixie>
Philip`: so?

[15:42:00.0000] <Philip`>
Hixie: Why is it sufficient to know the likely width given the font size in that case, whereas it's not sufficient in the normal drawHString case for authors to know the likely width and choose sizes/positions appropriately themselves (hence requiring the max-width parameter)?

[15:43:00.0000] <htmlfivedotnet>
Phlip: I think he's saying you would still want to set the maxwidth to say.... 100px. if it ends up off the field of view, fine, but if there is a town like "html5landvilleshiretownsylvania", it would take up half the screen if it wasn't made smaller.

[15:44:00.0000] <Hixie>
Philip`: i'm arguing that it's always normal to have a reasonable idea of your glyph extents in the block progression direction but not reasonable to guarentee that you'll fit in the inline progression direction.

[15:45:00.0000] <Hixie>
because given a font-size, your font metrics in the block progression direction are usually around the same as the font size but in the inline progression direction can vary considerably, and will vary in a way that is compounded with every glyph.

[15:45:01.0000] <Hixie>
(your typical iii vs www case for non-monospace text)

[15:46:00.0000] <Philip`>
Okay, sounds reasonable

[15:47:00.0000] <Hixie>
i wonder how to handle your case of wanting to line up text relative to the ascent top rather than the baseline

[15:50:00.0000] <Philip`>
Would it be feasible to use @font-face { src: ... } to define a font that could be used in the canvas?

[15:51:00.0000] <Hixie>
i expect we could define that to work, yes

[15:51:01.0000] <Hixie>
in fact i was planning on hooking straight into the css render model for this

[15:52:00.0000] <Philip`>
Presumably a problem is that it has to wait until after the font has been downloaded

[15:53:00.0000] <Hixie>
font download presumably will be defined to delay onload

[15:53:01.0000] <Hixie>
hmm

[15:58:00.0000] <Philip`>
(By the way, I'd probably want graph labels to be aligned to what http://en.wikipedia.org/wiki/Image:Typography_Line_Terms.svg calls 'cap height' instead of 'ascender height', since fancy cursive fonts shouldn't get placed so far down that there's space for all their large ascenders, since that'd look ugly)

[16:00:00.0000] <Hixie>
there's another problem, similar to the issue you raised about needing to align to the top, which is that to support Devanagari, Gurmukhi, Bengali, and other hanging indic scripts that want to render multiple font sizes we need to be able to tell the drawHText() method to use a hanging baseline instead of the alphabetic baseline

[16:00:01.0000] <Hixie>
yes, i meant the top of the em box actually

[16:00:02.0000] <Hixie>
not the top ascent height

[16:00:03.0000] <Hixie>
my bad

[16:01:00.0000] <Dashiva>
We need a typography5 to make line-height and other font stuff make sense to regular people...

[16:01:01.0000] <Philip`>
Reinventing CSS sounds like a bad idea

[16:02:00.0000] <Dashiva>
Wouldn't need to. Just add enough syntactic sugar around it so it's possible to use :)

[16:02:01.0000] <Philip`>
It seems easier and more powerful to have a drawElement() method (which marks the canvas as origin-unclean if it draws anything dodgy likes images or iframes or visited-vs-unvisited links)

[16:02:02.0000] <Philip`>
and then people don't have to learn lots of new API

[16:03:00.0000] <Hixie>
easier? no.

[16:03:01.0000] <Philip`>
(learn/define/implement)

[16:03:02.0000] <Hixie>
there are so many issues with that that i don't even want to go there.

[16:05:00.0000] <Philip`>
/me wonders what some of the issues are

[16:07:00.0000] <Hixie>
plugins, rendering off-screen, defining what it means to render an element at all (e.g. what about backgrounds of elements that contain it), what happens with floats near the element in the document, what happens if the element isn't in a document, rendering elements that have 'display' of none, table-column, table-row, etc, compact, run-in, etc, working out where the top-left of an element is, etc etc etc

[16:07:01.0000] <Hixie>
not to mention that it relies on css being implemented

[16:10:00.0000] <Philip`>
Define it as deep-cloning the element, moving it into an empty new HTML document, then rendering that document in a viewport of size w*h, then copying the output back. Easy :-)

[16:10:01.0000] <Philip`>
(and you can probably do that already in Mozilla, with iframes and drawWindow)

[16:11:00.0000] <Philip`>
((except for drawWindow not being web-accessible))

[16:11:01.0000] <Hixie>
uh huh

[16:11:02.0000] <Philip`>
Reliance on CSS seems like only a theoretical concern, given where this is going to be implemented

[16:12:00.0000] <Hixie>
yeah, the css issue is a real issue but one we could ignore

[16:12:01.0000] <Philip`>
Well, maybe there's some minor details I omitted :-p

[16:20:00.0000] <Hixie>
anyone here speak japanese, chinese, or korean, even just a little bit?

[16:21:00.0000] <Hixie>
i want to pick a CJK character for an example without offending native speakers...

[16:22:00.0000] <Hixie>
wikipedia to the rescue!

[16:36:00.0000] <Philip`>
/me hopes Hixie accidentally picks a piece of rude CJK wiki-vandalism

[16:44:00.0000] <myakura>
Hixie: i do speak japanese so might help you

[16:45:00.0000] <Dashiva>
Hixie: Just take the 'symbol' symbol :)

[16:50:00.0000] <Hixie>
myakura: i ended up picking the 私 character

[16:50:01.0000] <Hixie>
is that ok?

[16:52:00.0000] <myakura>
yeah, that means "I" the first person in japanese

[16:53:00.0000] <Hixie>
right

[16:53:01.0000] <Hixie>
is there a good glyph i could use that very clearly is lined up on the ideographic baseline?

[16:54:00.0000] <Dashiva>
国 maybe?

[16:55:00.0000] <myakura>
yeah, that's a nice pick

[16:56:00.0000] <Hixie>
ooh that works

[16:56:01.0000] <Hixie>
what does it mean?

[16:57:00.0000] <Hixie>
actually that one seems has a line on the alphabetic baseline, so it is misleading

[16:58:00.0000] <Hixie>
or maybe that's just the font i'm using

[16:58:01.0000] <Dashiva>
Where is the ideographic baseline?

[16:58:02.0000] <Dashiva>
Between the bottom of the king and the bottom of the square?

[16:59:00.0000] <Hixie>
(king?) it's below the alphabetic baseline and above the bottom of the em square

[16:59:01.0000] <Hixie>
just below the alphabetic baseline


2008-05-02
[17:00:00.0000] <Hixie>
匮 is more obviously aligned on the baseline

[17:02:00.0000] <Hixie>
遺 even more

[17:03:00.0000] <myakura>
or 辺?

[17:04:00.0000] <Hixie>
yes

[17:04:01.0000] <Hixie>
is that one good?

[17:04:02.0000] <myakura>
that means "border" or "edge"

[17:05:00.0000] <Hixie>
so "私辺" isn't rude or anything then? :-)

[17:06:00.0000] <Hixie>
meaningless is fine, the alphabetic part of this example at the moment reads "Á ÿ f Ω"

[17:06:01.0000] <Dashiva>
私達

[17:06:02.0000] <myakura>
"私辺" is meaningless, so fine :)

[17:07:00.0000] <myakura>
ah, 私達 ("we" or "us") is nice

[17:08:00.0000] <Hixie>
cool, i'll use that

[17:09:00.0000] <Hixie>
hey Dashiva did i add you to the acknowledgements yet?

[17:09:01.0000] <Dashiva>
Hehe. I don't think so.

[17:10:00.0000] <Hixie>
what's your name?

[17:11:00.0000] <Philip`>
(Dashiva: Make up an amusing joke name!)

[17:11:01.0000] <Dashiva>
(Philip`: Do you think he'll add it out of spite?)

[17:12:00.0000] <Dashiva>
I sent it in a private message so our log watchers won't find out!

[17:12:01.0000] <Philip`>
(Dashiva: Shh, he might hear us)

[17:12:02.0000] <Dashiva>
(Not that it can't be found in 5 minutes via google)

[17:12:03.0000] <Philip`>
The log watches could just look at the HTML5 diff

[17:12:04.0000] <Philip`>
*watchers

[17:14:00.0000] <Dashiva>
I'm gonna have to make an actually pseudonymous pseudonym 

[17:15:00.0000] <Dashiva>
On the other hand, it's convenient to have a recognizable handle too

[17:20:00.0000] <Hixie>
how about someone who speaks Devanagari? :-)

[17:21:00.0000] <Hixie>
/me hopes that the Devanagari? letter "AA" isn't rude

[17:21:01.0000] <BenMillard>
I can't even pronounce "Devanagari"...

[17:21:02.0000] <Dashiva>
Can't help you on that one. My quaternary and quinary languages are German and Latin

[17:21:03.0000] <Philip`>
I can't even read the Devanagari page on Wikipedia - it's just full of boxes

[17:22:00.0000] <Philip`>
This is where you need the i18n expertise of the relevant W3C WGs :-)

[17:22:01.0000] <Hixie>
yeah

[17:23:00.0000] <jgraham_>
http://krijnhoetmer.nl/irc-logs/whatwg/20070428

[17:23:01.0000] <Hixie>
luckily i've spent too many csswg meetings exposed to the i18n people :-P

[17:23:02.0000] <Hixie>
jgraham_: hah

[17:23:03.0000] <Hixie>
must've been the acid3 acks

[17:23:04.0000] <Dashiva>
jgraham_: That was for acid3

[17:24:00.0000] <Dashiva>
I was in the credits in the directory containing the unfinished test :)

[17:32:00.0000] <Hixie>
http://junkyard.damowmow.com/317

[17:34:00.0000] <BenMillard>
I'm off for lunch, bye all.

[17:34:01.0000] <Philip`>
Hixie: The font's too small

[17:34:02.0000] <Philip`>
Er, at least the small label font is - the others are probably large enough

[17:34:03.0000] <Hixie>
hmm

[17:39:00.0000] <Philip`>
(You should make it SVG and then I could zoom in and couldn't complain that it's too hard to read :-p )

[17:40:00.0000] <Hixie>
be my guest

[17:40:01.0000] <Hixie>
the original is pixel-aligned HTML :-D

[17:40:02.0000] <Hixie>
is http://junkyard.damowmow.com/318 ok ?

[17:41:00.0000] <Philip`>
It's easier to read than 317, but it would still make me unhappy if my eyesight was worse

[17:42:00.0000] <Hixie>
sure

[17:43:00.0000] <Philip`>
Could the labels be moved up by one pixel, so they don't merge with the similarly-coloured lines below them?

[17:44:00.0000] <Dashiva>
Hixie: What if you made the image like 5 times bigger and then scaled it down in regular display?

[17:44:01.0000] <Philip`>
(You could ignore me because I'm just being too picky now :-) )

[17:45:00.0000] <Philip`>
Dashiva: The nice sharp straight lines would disappear or go all smudgy

[17:45:01.0000] <Philip`>
and the text would be ugly since it wouldn't have the right hinting

[17:46:00.0000] <Hixie>
Philip`: http://junkyard.damowmow.com/320

[17:47:00.0000] <Dashiva>
Hey, that's practically readable when I zoom in

[17:47:01.0000] <Philip`>
Hixie: That looks better again

[17:47:02.0000] <Philip`>
The important question is, what's a suitable alt text for this image?

[17:48:00.0000] <Hixie>
a very long paragraph

[17:48:01.0000] <Hixie>
which i would include anyway

[17:48:02.0000] <Hixie>
and thus the alt will be alt=""

[17:51:00.0000] <othermaciej>
drawing an element into the canvas would be handy functionality

[17:51:01.0000] <othermaciej>
computing what is "origin-unclean" in there would be complex

[17:52:00.0000] <Hixie>
drawing an element into the canvas would be handy functionality, but i'm not doing it instead of text drawing functions

[17:55:00.0000] <Philip`>
If drawElement isn't impossible for implementors, I'd kind of like to have the simple text-drawing API (single font, single colour, single line, etc) and keep it as simple as is possible for labelling graphs, then have drawElement for full flexibility (mixing various font styles in a single piece of text, line-wrapping, word-spacing, line-spacing, etc)

[17:56:00.0000] <Philip`>
I wouldn't particularly like to have only one, since one is too simple and the other is too complex

[17:57:00.0000] <roc_>
oh, that again

[17:58:00.0000] <Hixie>
Philip`: there are many reasons to have a drawing api, i just don't want to go down that rathole just yet

[17:58:01.0000] <Hixie>
i agree that we shouldn't go overboard with the text drawing api

[17:59:00.0000] <roc>
othermaciej: there are also issues like drawing the contents of file upload controls, which you may not want the page to 'see'

[18:00:00.0000] <Philip`>
roc: You can set the canvas as origin-unclean whenever something potentially unsafe is drawn to it (which, in the crudest approximation, is whenever drawElement is used at all), and then the page can't use toDataURL/getImageData to recover any information from it

[18:01:00.0000] <Philip`>
It'd be nice if using drawElement didn't disable toDataURL/etc, but in any case it's still better than not having drawElement at all

[18:01:01.0000] <roc>
yeah, that's what I did when I proposed canvas.drawWindow as a Web API

[18:01:02.0000] <roc>
one problem with drawElement is that if you pass an element that's not in a document, you have no style data so you can't really draw it

[18:02:00.0000] <roc>
drawWindow avoids that problem

[18:02:01.0000] <roc>
but it's not a biggie

[18:03:00.0000] <Philip`>
Hixie: If it's plausible that an HTML-element drawing API could be added in a future version, it seems like that'd be worth mentioning somewhere, so that people don't get worried by thinking that the simple text-drawing API is all they're ever going to get and start trying to jam all possible features into it

[18:03:01.0000] <Hixie>
since all the features have to go through me, there is little chance of any jamming going on :-)

[18:04:00.0000] <Hixie>
but if i don't include such a note, remind me to add one once i do the checkin

[18:04:01.0000] <Hixie>
i'm still researching at the moment

[18:04:02.0000] <Philip`>
That's why it's only "trying", by sending emails with elaborate proposals, which will get ignored and cause misery to the proposers

[18:04:03.0000] <Hixie>
right now i'm trying to work out how much of the alignment information should be in the method call and how much should be in out-of-band style

[18:05:00.0000] <Philip`>
I think everything (except string and position) should be out-of-band (in attributes), because that's how the rest of the canvas API works

[18:05:01.0000] <Hixie>
makes sense i guess

[18:06:00.0000] <Philip`>
Also the arguments have no natural ordering, so nobody will remember whether it's drawText(..., maxWidth, textAlign, ...) or drawText(..., textAlign, maxWidth, ...)

[18:08:00.0000] <Hixie>
true

[18:08:01.0000] <Philip`>
Also the string argument should come before the position arguments, to be consistent with drawImage and putImageData

[18:09:00.0000] <Hixie>
.baseline = 'top' 'hanging' 'middle' 'alphabetic' 'ideographic' 'bottom'; .textAlign = 'start' 'middle' 'end'; ?

[18:10:00.0000] <Hixie>
i hate start/middle/end

[18:10:01.0000] <Philip`>
.textPosition = '(top|hanging|...) (start|...)'; ? Nobody's going to want to read the values back, and nobody's going to want to set one without setting the other, so they might as well be one attribute

[18:11:00.0000] <Philip`>
Is anything wrong with left/center/right?

[18:11:01.0000] <Hixie>
most people will rarely set .baseline

[18:11:02.0000] <Hixie>
or at least never set it to anything but top or bottom

[18:11:03.0000] <Hixie>
people in other locales will set it once to set their baseline

[18:11:04.0000] <Hixie>
left/center/right doesn't work for drawVText() if you're passing vertical text

[18:12:00.0000] <Hixie>
(i guess it does work if you're passing latin text, since then it renders inline sideways)

[18:12:01.0000] <Philip`>
Ah, okay

[18:12:02.0000] <Philip`>
(about both things)

[18:12:03.0000] <Philip`>
It should be .textBaseline rather than .baseline because otherwise it's not clear that it's related to text drawing

[18:12:04.0000] <Hixie>
yeah

[18:13:00.0000] <Philip`>
Could allow left/start/top as aliases, and center/middle and right/end/bottom

[18:13:01.0000] <Philip`>
though maybe that'd just be horribly confusing

[18:13:02.0000] <Hixie>
i wonder what css3 does

[18:14:00.0000] <Hixie>
"In vertical text, 'left' aligns to the edge of the line box that would be the start edge for left-to-right text."

[18:14:01.0000] <Hixie>
i wonder what the start edge for left-to-right text is

[18:15:00.0000] <Philip`>
Horizontal text seems (based on my reading a Wikipedia article) more common than vertical text, even in CJK (at least in the computer world), so it makes sense to optimise for the common case

[18:15:01.0000] <Hixie>
yes

[18:15:02.0000] <Hixie>
totally

[18:15:03.0000] <othermaciej>
this is starting to sound scary complicated

[18:15:04.0000] <Hixie>
othermaciej: it's actually not that bad

[18:15:05.0000] <othermaciej>
I hope it does not require implementing a third text layout engine

[18:16:00.0000] <Hixie>
othermaciej: i'm explicitly going to make this require the use of the css engine

[18:16:01.0000] <othermaciej>
(in WebKit, although we share as much code as we can, we effectively have two for CSS and SVG)

[18:16:02.0000] <othermaciej>
ok

[18:16:03.0000] <Hixie>
and i'm thinking of not including drawVText() for now

[18:16:04.0000] <Philip`>
Would it be sensible to forbid newlines? (e.g. replace \n with ' ')

[18:16:05.0000] <othermaciej>
as long as all the state translates readily to CSS parameters

[18:16:06.0000] <Hixie>
but i want to make sure it's specced sanely first, then just have it commented out

[18:16:07.0000] <Philip`>
since presumably newlines make layout harder

[18:16:08.0000] <Hixie>
Philip`: newlines will just be stripped

[18:16:09.0000] <Philip`>
Okay

[18:20:00.0000] <Hixie>
hm, seems that we just want different alignment stuff for HText and VText

[18:22:00.0000] <Hixie>
well it seems css handles it by having complicated rules (different complicated rules in each draft) but i guess we can just leverage those

[18:23:00.0000] <Philip`>
/me notes that mozDrawText seems to just not bother with bidi text at all

[18:24:00.0000] <Hixie>
that's ok for a demo experimental api, but the i18n guys, while less aggressive than the accessibility guys, still bite

[18:24:01.0000] <Philip`>
The users are a more important consideration than the i18n guys

[18:25:00.0000] <Hixie>
to you maybe! i don't want to be bitten!

[18:25:01.0000] <Philip`>
but if the i18n guys are aligned with the interests of the users then it doesn't make that much difference in practice, so hopefully that's the case :-)

[18:25:02.0000] <Hixie>
:-)

[18:25:03.0000] <Hixie>
i18n is far more complex than accessibility, but also far less subjective

[18:26:00.0000] <Philip`>
It also affects far more users

[18:26:01.0000] <Hixie>
it affects all of them :-)

[18:28:00.0000] <Philip`>
Text is hard, I want to stay with pictures and colours

[18:30:00.0000] <Philip`>
Could one argue that the text drawing API is intentionally designed to not harm accessibility, by being so limited (to a single line of text with no markup) that it won't encourage people to move any significant text from out of their HTML and into their inaccessible canvas?

[18:31:00.0000] <Philip`>
Maybe it's safer to omit the "intentionally", but otherwise that seems true

[18:32:00.0000] <Philip`>
(e.g. if you have a caption for a graph, you'd write it as HTML below the canvas, where it's easy for everyone to read, because it's too painful to try to print it all directly in the canvas)

[18:33:00.0000] <Hixie>
i'm not even going to go near that

[18:33:01.0000] <othermaciej>
text drawing on canvas is bad for more than just accessibility

[18:33:02.0000] <othermaciej>
you can't select the text, copy it, do text find in it...

[18:34:00.0000] <othermaciej>
but of course people already do images of text which create the same problems

[18:34:01.0000] <Philip`>
Browsers could still offer text-selection features

[18:34:02.0000] <Philip`>
They just need to be a tiny bit cleverer about it

[18:34:03.0000] <Philip`>
(and will break if people use putImageData)

[18:36:00.0000] <othermaciej>
offer text selection features how?

[18:36:01.0000] <Philip`>
Keep a vector approximation of the canvas, to remember the positions of the text boxes, and clip that against every other shape that gets drawn, and you'd end up with a way of mapping from any pixel back to the text that is visible under that point

[18:36:02.0000] <othermaciej>
by making the canvas backing store a scene graph?

[18:37:00.0000] <Philip`>
You only need to store the visible extents of each text box, and don't have to worry about remembering any of the other shapes

[18:37:01.0000] <Hixie>
ocr

[18:37:02.0000] <Hixie>
then it would work for other images too

[18:37:03.0000] <othermaciej>
maybe you could keep an SVG DOM as the <canvas> backing store

[18:37:04.0000] <Hixie>
the spec does suggest that already :-D

[18:38:00.0000] <othermaciej>
that would sort of defeat the point of canvas as an immediate-mode graphics API

[18:38:01.0000] <Hixie>
i didn't say it made sense

[18:38:02.0000] <Philip`>
The point of the immediate-mode API is that it's an immediate-mode API, not that it happens to be implemented on top of a bitmap

[18:39:00.0000] <Hixie>
actually the spec iirc just mentions it in an aside as a hypothetical way to support toDataURL("image/svg+xml")

[18:39:01.0000] <Philip`>
The IE VML canvas emulation is still an immediate-mode API, even though the implementation is totally not

[18:40:00.0000] <Philip`>
Then again, in real life people care about performance, and the performance characteristics of bitmaps and vector scene graphs are very different

[18:40:01.0000] <othermaciej>
and that makes it lack the actual technical advantages of immediate mode (e.g. no RAM explosion when you draw a huge number of shapes)

[18:41:00.0000] <roc>
supporting bidi wouldn't be that hard as long as you don't have to support styled text runs

[18:41:01.0000] <roc>
I mean multiple styles in the same string

[18:41:02.0000] <Hixie>
indeed, none of that in this api

[18:44:00.0000] <othermaciej>
for cases where you want to apply bitmap-level effects to characters or otherwise use glyphs as graphical elements, a straight drawing API makes sense

[18:44:01.0000] <othermaciej>
but for use cases like chars, a convenient way to have a text layer on top of the canvas containing real text would be superior

[18:44:02.0000] <othermaciej>
*charts

[18:45:00.0000] <othermaciej>
because then select/copy/find works on your chart ticks

[18:45:01.0000] <othermaciej>
you can sort of do this now by positioning other elements over the canvas but that is hard to do right

[18:45:02.0000] <othermaciej>
it would be unfortunate if the only easier path led to a worse user experience

[18:47:00.0000] <Hixie>
i'm open to ideas

[18:47:01.0000] <othermaciej>
my "text layer on top" suggestion is not even a quarter baked

[18:47:02.0000] <othermaciej>
but I could try making a more detailed proposal

[18:48:00.0000] <Hixie>
sure

[18:49:00.0000] <Hixie>
i expect most people just want the text in the bitmap though

[18:49:01.0000] <Hixie>
so they can do things like layer crosshairs over the top or whatever

[18:50:00.0000] <othermaciej>
you mean for charts?

[18:50:01.0000] <Hixie>
yeah

[18:50:02.0000] <Hixie>
consider http://www.whatwg.org/issues/data.html

[18:51:00.0000] <othermaciej>
the chart on this page would clearly be better if the text was real: <http://arstechnica.com/news.ars/post/20080317-firefox-3-goes-on-a-diet-eats-less-memory-than-ie-and-opera.html>

[18:51:01.0000] <Hixie>
well for that kind of static thing i'd recommend using svg, not canvas

[18:51:02.0000] <othermaciej>
sure, but you could make similar charts from live data

[18:52:00.0000] <othermaciej>
like a stock price graph

[18:52:01.0000] <Hixie>
yep

[18:52:02.0000] <othermaciej>
or web site traffic stats

[18:53:00.0000] <Hixie>
why does the safari line just stop on that chart

[18:53:01.0000] <Hixie>
and is safari any better now? :-)

[18:53:02.0000] <Philip`>
Use a static SVG graph template, and then dynamically modify one of its <path>s to have d="M $x[0] $y[0] L $x[1] $y[1] ..." 

[18:53:03.0000] <Philip`>
"Safari 3 and Internet Explorer 8 could not be benchmarked because they crashed during the test."

[18:54:00.0000] <othermaciej>
I think Safari for Windows crashed on that benchmark when the Firefox guys ran it

[18:54:01.0000] <othermaciej>
not for us though, on the latest SafariWin

[18:54:02.0000] <othermaciej>
and yes, memory use is better now

[18:55:00.0000] <Hixie>
cool

[18:55:01.0000] <Hixie>
someone should redo that graph

[18:55:02.0000] <othermaciej>
in this case Ars Technica did not rerun the benchmark they just posted Pav's results

[18:55:03.0000] <othermaciej>
which is lazy reporting

[18:55:04.0000] <othermaciej>
I think we'll post some memory use benchmark results eventually

[19:02:00.0000] <Philip`>
/me tries to think of a use case for applying transformations to text

[19:02:01.0000] <Philip`>
If you've played Unreal Tournament, in DM-Morpheus there's that bit where it has a scrolling text display along one wall of a building to say who's currently winning the game

[19:02:02.0000] <Philip`>
I want one of those in my canvas 3D FPS, because they look cool

[19:03:00.0000] <Hixie>
the baseline values make no sense for drawVText

[19:03:01.0000] <Hixie>
for drawVText() we'd just want left, middle, right

[19:03:02.0000] <jruderman>
othermaciej: "eventually" = "when we catch up with gecko"? :P

[19:03:03.0000] <Philip`>
Is vertical text ever drawn left/right aligned, instead of middle-aligned?

[19:04:00.0000] <othermaciej>
jruderman: not necessarily, I'm not philosophically opposed to posting benchmarks we don't win yet

[19:04:01.0000] <Hixie>
Philip`: well

[19:04:02.0000] <Hixie>
Philip`: you always want it middle aligned

[19:04:03.0000] <Philip`>
othermaciej: Are there no marketing people controlling what benchmarks you publish? :-)

[19:05:00.0000] <Hixie>
Philip`: but you might want it anchored using a point on the left or right

[19:05:01.0000] <Philip`>
Hixie: Oh, right

[19:05:02.0000] <Hixie>
i guess you always know the em height, since you set it using 'font'

[19:05:03.0000] <Hixie>
so really this will always just offset your x by +/-0.5em

[19:05:04.0000] <othermaciej>
Apple publishes benchmark results and controls what goes on the apple.com/safari page

[19:06:00.0000] <othermaciej>
the WebKit blog is mainly about technical information and not as influenced by marketing concerns

[19:06:01.0000] <othermaciej>
though obviously we prefer to post when we have good news

[19:20:00.0000] <Hixie>
ooh, first actual casualty of the way canvas doesn't have separate path objects

[19:20:01.0000] <Hixie>
you can't draw text along a path to a path

[19:20:02.0000] <Hixie>
or rather you can't draw two pieces of text along two different paths to the same path

[19:25:00.0000] <othermaciej>
/me is not following

[19:25:01.0000] <othermaciej>
two different paths to the same path?

[19:26:00.0000] <Hixie>
path3.addTextToPath("foo", path1); path3.addTextToPath("bar", path2);

[19:26:01.0000] <Hixie>
no way to do that in <canvas> without adding path objects

[19:26:02.0000] <Philip`>
Can you use text as a path to draw text around?

[19:26:03.0000] <Hixie>
i am considering both drawing text to a path and drawing text on a path

[19:26:04.0000] <Philip`>
You could make a neat fractal pattern like that

[19:26:05.0000] <Hixie>
but i wished to do both

[19:27:00.0000] <Hixie>
see e.g. the demos at the bottom of http://developer.mozilla.org/en/docs/Drawing_text_using_a_canvas in ff3

[19:28:00.0000] <jruderman>
Philip`: ooh, that sounds neat

[19:28:01.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/#text (see the idl block earlier in that section for more details for the api i'm looking at so far)

[19:28:02.0000] <Hixie>
(this is just for drawing text straight to the canvas using 'fill' so far

[19:29:00.0000] <Hixie>
but i might either add to or replace this with an api that does path stuff)

[19:30:00.0000] <othermaciej>
I hope drawing text on a path has SVG-compatible text layout semantics

[19:31:00.0000] <Hixie>
believe me, i have every intention to take every possible shortcut in defining this

[19:31:01.0000] <othermaciej>
it sounds like the Mozilla API does not

[19:31:02.0000] <Hixie>
which means relying heavily on css and svg

[19:33:00.0000] <Philip`>
Hixie: drawHText should be called drawText, because many people will only ever do horizontal text and won't even think about the existence of vertical text, so they won't understand what the H is for. The H just makes things confusing for those people, and harder to type even for people who understand the reasoning, and it doesn't benefit anyone to keep the H in there

[19:33:01.0000] <Hixie>
i guess that makes sense, since the opposite doesn't apply to users of drawVText()

[19:34:00.0000] <Hixie>
othermaciej: dude, you're replying to yourself now on that bikeshed? :-)

[19:34:01.0000] <Philip`>
People who write vertical text seem to be equally familiar with horizontal text, so presumably they can cope fine with the differences between them

[19:34:02.0000] <othermaciej>
Hixie: I had another random idea

[19:34:03.0000] <Hixie>
othermaciej: (btw, rel values are "or"ed, not "and"ed)

[19:35:00.0000] <Hixie>
othermaciej: (so rel="icon 20x20" means it's two links, one for an icon, and one for a 20x20)

[19:35:01.0000] <Philip`>
Also the canvas API, and the whole of HTML5, and the whole web standards community, are strongly biased towards English (because e.g. all the method names and specs are written in English), so it's too late to be politically correct by balanacing drawVText and drawHText :-)

[19:35:02.0000] <othermaciej>
Hixie: "alternate stylesheet" is a counterexample to that (though perhaps a special exception)

[19:35:03.0000] <Philip`>
s/a//

[19:35:04.0000] <othermaciej>
Hixie: I like my sizes idea best, just wanted to jot that one down

[19:35:05.0000] <Philip`>
Oh, also it should be drawVerticalText instead of drawVText, so that people can work out what it means from the name

[19:35:06.0000] <Hixie>
Philip`: i don't buy that, i have no intention of biasing new apis against the majority of web users just because the standard is written in english

[19:36:00.0000] <othermaciej>
Hixie: it would be handy for Safari/WebKit purposes to have a ruling on this btw to spare us having to make something up

[19:36:01.0000] <Hixie>
othermaciej: for the icon thing? i didn't realise it was important for you too

[19:37:00.0000] <othermaciej>
Hixie: that is why I am telling you now :-)

[19:37:01.0000] <Hixie>
othermaciej: if i can ask, is this for iphone use, or for something else?

[19:38:00.0000] <Hixie>
othermaciej: (just trying to get an idea of what the use cases are)

[19:38:01.0000] <othermaciej>
Hixie: well, iPhone already has an iPhone-specific proprietary rel value to specify an iPhone-sized icon for use on the home screen

[19:38:02.0000] <Philip`>
Hixie: In terms of method naming, "drawVText" is already totally biased towards English, because it's using English words and abbreviations, which is unavoidable

[19:38:03.0000] <othermaciej>
it would be preferable for it to adopt a common thing

[19:38:04.0000] <Hixie>
othermaciej: k

[19:39:00.0000] <othermaciej>
I would also like apps like Fluid and Prism to be able to use proper OSX-compliant icons, thus the multisize thing

[19:39:01.0000] <othermaciej>
as for Apple's future plans, I am not at liberty to be specific

[19:39:02.0000] <Hixie>
sure

[19:39:03.0000] <Hixie>
Philip`: sure, but i try to keep that to a minimum

[19:39:04.0000] <Hixie>
Philip`: anyway, in this case it doesn't matter

[19:40:00.0000] <othermaciej>
tagging scalable/anysize icons is more for completeness but they could be a good alternative to multisize in many cases

[19:40:01.0000] <Hixie>
Philip`: as vertical text isn't as common as horizontal anyway

[19:40:02.0000] <othermaciej>
(if you provide vector art, the UA could always render it down to a multisize bitmap if that is what the OS wants)

[19:40:03.0000] <Hixie>
othermaciej: yeah

[19:40:04.0000] <Hixie>
right, dinner time

[19:40:05.0000] <Hixie>
bbl

[19:41:00.0000] <Philip`>
Hixie: I think maxWidth should be optional - it's only useful if it's used correctly, and it's actively harmful if it's used incorrectly (because authors will get confused about why some of their text keeps getting squashed), and it's not trivial to use it correctly, so people should not be forced to supply a value for it

[20:36:00.0000] <Philip`>
Hmm, getting into the IE8 Technical Beta program (which offers the privilege of being able to report bugs) is surprisingly easy

[20:51:00.0000] <Philip`>
/me sleeps

[00:57:00.0000] <hsivonen>
/me thinks Hixie is asking the wrong question when he asks if anyone "speaks" Devanagari

[02:45:00.0000] <Hixie>
hsivonen: i say "speak perl" etc, too

[02:47:00.0000] <hsivonen>
ok.

[02:50:00.0000] <Philip`>
It's slightly painful when you have to speak Perl over the phone to a non-programmer for them to type it in

[02:50:01.0000] <Hixie>
i actually had to do that the other day

[02:50:02.0000] <Hixie>
though it was over video iChat

[02:50:03.0000] <Hixie>
so instead i just asked to use screen sharing and did it manually

[02:51:00.0000] <Hixie>
i wish there was an option in iChat "ask to open ssh connection"

[02:52:00.0000] <hsivonen>
I need to learn some way to shield my mood from the negativity of bikeshedding

[02:54:00.0000] <hsivonen>
/me does some therapeutic tokenizer hacking

[03:00:00.0000] <Hixie>
i recommend blowing shit up in gta4

[03:00:01.0000] <Hixie>
/me fires up his console

[03:18:00.0000] <Philip`>
http://www.w3.org/QA/2008/05/canvas-text-and-cjk.html

[03:23:00.0000] <krijn>
alt="Japanese Typography"

[03:23:01.0000] <krijn>
Yay \o/

[03:31:00.0000] <Hixie>
hm

[03:31:01.0000] <Hixie>
i had good hopes for the doc he links to

[03:31:02.0000] <Hixie>
but it turns out to not say anything especially useful

[03:31:03.0000] <Hixie>
it's mostly abotu layout

[03:31:04.0000] <Hixie>
which doesn't affect us

[03:47:00.0000] <Lachy>
It's interesting though. I didn't realise they mixed verically stacked letters and rotated letters like that, each for different purposes.

[03:50:00.0000] <Hixie>
sure

[03:50:01.0000] <Hixie>
that's what "full width" characters are for, basically, as i understand it

[04:14:00.0000] <takkaria>
http://www.marcozehe.de/2008/05/01/social-bookmarking-feature-added-to-blog/ terrifies me.  people with screenreaders have 45 images to skip over with near-identical alt text

[04:16:00.0000] <Philip`>
Don't screenreaders have a "skip to next paragraph" key, that would let you avoid all those images easily?

[04:16:01.0000] <Lachy>
oh, wow. I had hoped the silly craze of putting subscription links like that would be over by now.

[04:17:00.0000] <takkaria>
Philip`: we can hope

[04:18:00.0000] <hsivonen>
that's sad

[04:18:01.0000] <hsivonen>
all those icons are useless

[04:18:02.0000] <hsivonen>
I have a bookmarklet for the social bookmarking service I actually use

[04:18:03.0000] <Philip`>
When I last used Links (a few days ago), I was looking at a page that happened to have a "skip to content" link at the top before all the navigation menu stuff, but the browser just complained "cannot find #content" when I tried using the link :-(

[04:19:00.0000] <takkaria>
as far as I can tell, it's mainly done by people who aren't very popular but who really want to be and think that by including those links they'll becoming more so

[04:19:01.0000] <krijn>
Philip`: could've been a site I've built :$

[04:20:00.0000] <krijn>
That happens a lot, when making templates for some backend company who just doesn't care

[04:22:00.0000] <roc>
takkaria: Marco is blind, for what it's worth

[04:22:01.0000] <roc>
so if it sucks for screenreaders, he'd probably know

[04:23:00.0000] <hsivonen>
at least it sucks visually

[04:23:01.0000] <roc>
it's his revenge!

[04:24:00.0000] <krijn>
Gheh

[04:24:01.0000] <othermaciej>
roc: is that really a safe assumption? many sighted people design UIs that are unbelievably bad for sighted users

[04:25:00.0000] <takkaria>
roc: oh, ok.  interesting

[04:25:01.0000] <mpt>
It's begging for a parody

[04:26:00.0000] <roc>
othermaciej: true. But still, he cares about this stuff more than most of us

[04:27:00.0000] <takkaria>
mpt: I actually thought it /was/ a parody when I first saw it...

[04:27:01.0000] <mpt>
http://web.archive.org/web/19970205051143/www.stale.com/cmp/options.html#printable

[04:28:00.0000] <roc>
that was such a great site.

[04:30:00.0000] <mpt>
What was that Weblog that Mark Pilgrim did anonymously? Something about platypuses That had huge amounts of administrivia, too

[04:30:01.0000] <mpt>
/me stares at his "." key

[04:32:00.0000] <hsivonen>
www.ragingplatypus.com ?

[04:32:01.0000] <mpt>
oh, that was syndication buttons

[04:32:02.0000] <mpt>
http://web.archive.org/web/20030313032419/http://www.ragingplatypus.com/#metaplat

[04:32:03.0000] <Hixie>
gta4 has internet cafes where you can "browse the web"

[04:32:04.0000] <Hixie>
is it bad that my first thought was to try to go to see how it does on the acid3 test

[04:33:00.0000] <hsivonen>
Hixie: how did it do?

[04:33:01.0000] <Hixie>
"page not found"

[04:35:00.0000] <Hixie>
wow

[04:35:01.0000] <Hixie>
this web is like a total parody of the real one

[04:37:00.0000] <roc>
I assume if you get frustrated with your Web browsing experience you can blast the computer to smithereens

[04:37:01.0000] <Hixie>
and then firebomb the internet cafe, seems likely

[04:38:00.0000] <othermaciej>
are you browsing on the iFruit?

[04:38:01.0000] <Lachy>
wikipedia says "There are over 100 accessible, fictitious websites within the game"

[04:38:02.0000] <Lachy>
http://en.wikipedia.org/wiki/Grand_Theft_Auto_IV

[04:38:03.0000] <roc>
crazy

[04:39:00.0000] <Hixie>
wow, there are possibly more porn sites on the gta4 web than the real one

[04:39:01.0000] <Lachy>
LOL

[04:39:02.0000] <othermaciej>
no way

[04:40:00.0000] <Lachy>
Hixie, I think you may be seriouly underestimating the amount of porn on the real web. :-)

[04:40:01.0000] <roc>
they should just stick a real browser in there with real internet access

[04:41:00.0000] <roc>
also, they should support video chat with the outside world

[04:41:01.0000] <Hixie>
Lachy: well, i mean, relative to the total size, not in absolute numbers, obviously :-)

[04:41:02.0000] <roc>
so you can call your friends from in-game and they'll see your avatar

[04:41:03.0000] <roc>
then they'll watch you get blown up because you're not paying attention

[04:42:00.0000] <Hixie>
this fake web is remarkably true to the real one

[04:42:01.0000] <Lachy>
any well known sites included within it?

[04:43:00.0000] <Hixie>
there's a page antfarmcam.net that claims to have a webcam of an antfarm, except the camera is down

[04:43:01.0000] <Hixie>
and there are three comments on this blog

[04:43:02.0000] <Hixie>
first is "WTF! This lame cam has been down for months! [...]"

[04:43:03.0000] <Hixie>
second is full of swear words

[04:44:00.0000] <Hixie>
and the last is from "5 months ago" and says "I hop ur antz is ded/ LOL Ants stink"

[04:44:01.0000] <Lachy>
what, no "first post" troll?

[04:48:00.0000] <Hixie>
ok let's see if i really can blow up these computers

[04:48:01.0000] <Hixie>
HAHAHAHA

[04:48:02.0000] <Hixie>
oh yeah

[04:48:03.0000] <Hixie>
hm i scared the other customer

[04:49:00.0000] <hsivonen>
/me awaits out-of-context IRC log quotes to show up on mailing lists

[04:50:00.0000] <deltab>
http://arstechnica.com/journals/thumbs.ars/2008/05/01/grand-theft-auto-iv-story-time-with-ben-and-frank

[05:17:00.0000] <Philip`>
Seems like GTA4's developers have registered all their virtual-internet domain names on the real internet, but they haven't even got DNS records, which seems like an odd waste of a marketing opportunity

[05:18:00.0000] <Hixie>
some do

[05:18:01.0000] <Hixie>
the jazz station in gta4 is better than most real jazz stations i've listened to

[05:22:00.0000] <Philip`>
/me has always wondered how much it cost to license all the music that's used in the games

[05:27:00.0000] <hsivonen>
what are XHTML compact forms? http://krijnhoetmer.nl/irc-logs/xhtml/20080430#l-108

[05:29:00.0000] <Philip`>
hsivonen: Probably http://www.w3.org/MarkUp/2008/ED-xhtmlmime-20080423/#C_10

[05:30:00.0000] <Philip`>
s/Probably/Maybe/

[05:30:01.0000] <hsivonen>
Philip`: thanks

[07:19:00.0000] <hsivonen>
How is "foo<!--->bar" supposed to parse in [R]CDATA?

[07:24:00.0000] <hsivonen>
hmm. I suppose the second hyphen turns escape on and > turns it off

[08:02:00.0000] <hsivonen>
the html5lib test cases rock

[08:17:00.0000] <Philip`>
/me wonders if "hash backet" is a common term, or if it's just a spelling mistake in this hash table implementation that nobody has bothered to fix

[08:18:00.0000] <hsivonen>
oh yeah, still only 7925 bytecodes

[08:20:00.0000] <Philip`>
Can you get bytecode optimisers to make it smaller?

[08:20:01.0000] <hsivonen>
Philip`: I don't have one

[08:20:02.0000] <hsivonen>
Philip`: If I need to make it smaller, I'll move string concatenation for error messages into another method

[08:21:00.0000] <Philip`>
Invalid HTML is the most common case, so you should be optimising the generation of error messages :-)

[08:21:01.0000] <hsivonen>
heh

[08:40:00.0000] <hsivonen>
getting rid of the content model flag and the escape flag wasn't as big a win as I expected, but now HEAD beats the latest release even on PPC Client VM

[08:41:00.0000] <Philip`>
How fast does it parse now, in megabits per second?

[08:42:00.0000] <hsivonen>
http://about.validator.nu/htmlparser/perf3.zip

[08:42:01.0000] <hsivonen>
Philip`: I haven't measured it that way

[08:48:00.0000] <Philip`>
Hmm, it can parse a 53KB file about 30K times in one minute, so it's around 200Mbit/s

[08:48:01.0000] <Philip`>
which is faster than the network interface on that machine

[08:48:02.0000] <Philip`>
so it seems unlikely to be the bottleneck in the kinds of thing I'd use it for

[08:49:00.0000] <Lachy>
http://www.bestkungfu.com/archive/date/2008/05/alt-and-the-flickr-defense/

[08:49:01.0000] <Lachy>
... seems to be more of the same.

[08:52:00.0000] <Philip`>
(XML=55301, methods=26961, switch=27660, Java HotSpot(TM) 64-Bit Server VM (build 1.6.0-b105, mixed mode))

[08:52:01.0000] <Philip`>
(running for 1 minute)

[08:53:00.0000] <Philip`>
(since I'm too lazy to leave it for longer)

[09:25:00.0000] <Lachy>
Interesting statistics http://www.456bereastreet.com/archive/200804/validation_statistics_from_nikita_the_spider/

[09:25:01.0000] <Lachy>
I would have though unescapted & would rate higher than missing alt though, but it seems not

[09:27:00.0000] <Lachy>
though it seems the results are limited to pages where authors explicitly validated their page, rather than a general web crawl. So it's possible the results could be a bit biased in some ways.

[09:29:00.0000] <Philip`>
That's total number of occurrences, not number of pages, which is slightly dodgy since a few pages with thousands of images could distort the results

[09:30:00.0000] <Philip`>
They got 55K altless images out of 10K pages, i.e. 5.5 per page; I got 69K out of 8K pages, i.e. 9 per page

[09:31:00.0000] <Philip`>
Out of all pages with <img>, I saw 20% of those pages never using alt

[09:32:00.0000] <Philip`>
Out of all pages in total, I saw 30% with unencoded ampersands

[09:32:01.0000] <Philip`>
(but I don't know how many unencoded ampersands per page)

[13:38:00.0000] <mcarter>
hello

[14:53:00.0000] <Philip`>
Does anyone happen to know if CSS3 Color drafts have changed since the 2003 CR in a way that would affect (to pick a totally arbitrary example) canvas fill/stroke colours?

[14:53:01.0000] <Philip`>
(All I've noticed so far is the dropping of flavor)

[15:10:00.0000] <Hixie>
dbaron: 23:12 <       Philip`> Does anyone happen to know if CSS3 Color drafts have changed since the 2003 CR in a way that would affect (to pick a totally arbitrary example) canvas fill/stroke colours?

[15:11:00.0000] <dbaron>
All the changes are described in http://csswg.inkedblade.net/spec/css3-color

[15:11:01.0000] <Hixie>
Philip`: ^

[15:11:02.0000] <Hixie>
dbaron: thanks

[15:11:03.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/images/baselines.png is the image i was talking about btw

[15:11:04.0000] <dbaron>
I think some of those technically do affect things

[15:11:05.0000] <Hixie>
in case you spotted any problems

[15:12:00.0000] <dbaron>
e.g., handling of out-of-range values for hsl(), hsla(), and rgba() was undefined.

[15:12:01.0000] <Hixie>
was undefined and is now defined, or was un-defined?

[15:12:02.0000] <Hixie>
i assume the former! :-)

[15:12:03.0000] <dbaron>
the former

[15:12:04.0000] <dbaron>
mostly

[15:12:05.0000] <Hixie>
heh

[15:12:06.0000] <dbaron>
it's now mostly but not completely defined

[15:13:00.0000] <Philip`>
dbaron: Aha, thanks

[15:17:00.0000] <Philip`>
Hixie: Would there be any point in producing an SVG version of that image? It would seem nicer than the current PNG with 5-pixel tall characters since you could zoom in to read it, but it'd probably be a waste of time if the spec just wants to stick with PNG for browser compatibility

[15:27:00.0000] <Hixie>
Philip`: i'm certainly happy for you to create an svg image and would be happy to link the png to the svg image

[15:28:00.0000] <Hixie>
Philip`: but i expect i'll stick to png as the main one

[16:11:00.0000] <Hixie>
man i wish anne would provide an obvious link to the editor's copies of his drafts on the TR versions

[16:11:01.0000] <Hixie>
i can never find the latest versions of his specs

[16:14:00.0000] <Hixie>
christ dev.w3.org is such a mess

[16:14:01.0000] <Hixie>
how am i supposed to guess that access-control is a waf spec, and how am i supposed to guess the waf was created in 2006?

[16:18:00.0000] <Hixie>
hrm

[16:18:01.0000] <Hixie>
did we remove the ability to opt-in to certain methods?

[16:19:00.0000] <Hixie>
i guess we did

[16:20:00.0000] <Lachy>
Hixie, have you been following the selectors api thread about :scope vs. other methods?

[16:20:01.0000] <Hixie>
yeah

[16:20:02.0000] <Lachy>
what's your opinion on the issue?

[16:21:00.0000] <Hixie>
i don't understand the issue

[16:21:01.0000] <Hixie>
i mean, i understand what is being requested

[16:21:02.0000] <Hixie>
i don't understand the use case

[16:22:00.0000] <Lachy>
there are several use cases that need to be addressed.

[16:22:01.0000] <Hixie>
i agree that a pseudo-class that matches the context node might be useful, but it seems like a v2 thing

[16:22:02.0000] <Lachy>
yeah, that was the original plan

[16:22:03.0000] <Hixie>
i don't agree with changing selectors to only look at nodes at the context plan or lower

[16:22:04.0000] <Hixie>
that seems silly

[16:23:00.0000] <Lachy>
yeah, that's what I think. But it is the behaviour of current JS library versions

[16:23:01.0000] <Hixie>
so? this doesn't stop them existing

[16:23:02.0000] <Hixie>
not my fault they are all buggy...

[16:23:03.0000] <Lachy>
yeah, I guess.

[16:25:00.0000] <Lachy>
JohnResig, yt?

[16:26:00.0000] <JohnResig>
Lachy: what's up?

[16:28:00.0000] <Lachy>
I don't see what problem is being solved by changing the API, that won't be solved by adding :scope in the near future

[16:28:01.0000] <Hixie>
(:scope is a terrible name btw)

[16:28:02.0000] <Hixie>
(imho!)

[16:28:03.0000] <Lachy>
Hixie, any better suggestion?

[16:28:04.0000] <Hixie>
sadly not

[16:28:05.0000] <Hixie>
so i'm afraid i'm not much help

[16:28:06.0000] <Hixie>
:-)

[16:28:07.0000] <Lachy>
so far the options are :self, :this, or :scope

[16:28:08.0000] <Hixie>
:context-node

[16:28:09.0000] <Hixie>
:current

[16:29:00.0000] <Lachy>
yeah, they could work

[16:29:01.0000] <Hixie>
:query-root

[16:29:02.0000] <Hixie>
i dunno

[16:29:03.0000] <Hixie>
i guess it doesn't matter that much

[16:29:04.0000] <Hixie>
i do like the :query-root name, since this is similar to :root in many ways

[16:29:05.0000] <Hixie>
indeed some people suggested just using :root

[16:29:06.0000] <Hixie>
which i disagree with

[16:29:07.0000] <Hixie>
but i see the logic

[16:30:00.0000] <Lachy>
naming debates are fun :-)

[16:30:01.0000] <Hixie>
the final decision is up to you obviously

[16:30:02.0000] <Lachy>
that's what I thought the last time I was in a naming debate

[16:30:03.0000] <Hixie>
i'm just throwing in my twopence or however you spel that

[16:30:04.0000] <Hixie>
yeah well

[16:30:05.0000] <JohnResig>
Lachy: so you're perfectly happy with Safari, Opera, Firefox, and Internet Explorer all shipping with no way of finding child (or adjancent) elements and returning inconsistent results? and everyone being stuck with using that broken method for years to come?

[16:31:00.0000] <Hixie>
the w3c hasn't quite caught up with the 21st century yet

[16:31:01.0000] <Lachy>
JohnResig, why couldn't we spec and ship with support for :scope, in the same time it would take to change the API?

[16:31:02.0000] <Hixie>
why would they be inconsistent results?

[16:32:00.0000] <Hixie>
and how is there no way to find child or adjacent elements?

[16:32:01.0000] <Hixie>
i guess it would help to see the use cases

[16:32:02.0000] <JohnResig>
Hixie: note the examples that I provided in the thread - finding elements that are globally-rooted rather than contextually-rooted goes against all previous implementations

[16:33:00.0000] <Hixie>
JohnResig: sure, but like i said, it's not my fault all those implementations are buggy

[16:33:01.0000] <JohnResig>
Hixie: define "buggy" - "buggy" seems to be "we have better control over the selectors that we can provide"

[16:33:02.0000] <Hixie>
buggy as in they don't implement selectors correctly

[16:33:03.0000] <Hixie>
selectors aren't rooted at a different tree

[16:33:04.0000] <Lachy>
it's easier to add contextually-rooted support later (or at the same time), than it is to add globally-rooted selectors

[16:34:00.0000] <Hixie>
but anyway that's not a use case

[16:34:01.0000] <Hixie>
that's just past implementations

[16:34:02.0000] <othermaciej>
JohnResig: old Safari versions fall to infinitesimal market share pretty quickly

[16:34:03.0000] <Lachy>
in fact, it would be impossible to do the latter without introducing a whole new method

[16:34:04.0000] <Hixie>
use cases are what i'd like to see

[16:34:05.0000] <othermaciej>
JohnResig: at least as to WebKit, I think your concern about old versions is overblown

[16:34:06.0000] <JohnResig>
othermaciej: I'm not worried about Safari at all - I'm crazy-concerned about IE

[16:35:00.0000] <JohnResig>
if browsers are released that don't implement :scope then, effectively, developers won't be able to use Element.querySelectorAll in them

[16:35:01.0000] <Hixie>
it's trivial to fake :scope

[16:35:02.0000] <JohnResig>
oh, is this the fun "add an ID to the element then remove it later" - yeah, that' BS

[16:35:03.0000] <JohnResig>
*that's

[16:35:04.0000] <Lachy>
yeah, it can be done with an ID selector (even if that ID is temporarily auto-generated by the API

[16:36:00.0000] <Lachy>
I know it's not optimal

[16:36:01.0000] <Hixie>
it's no more BS than the idea that we need it in the first place :-)

[16:36:02.0000] <JohnResig>
except that's something that should be done by the implementation in the first place

[16:36:03.0000] <Hixie>
but like i said, i'd need use cases to work out what the actual discussion point is

[16:36:04.0000] <Hixie>
arguing about an API design in the absence of use cases is meaningless

[16:37:00.0000] <Hixie>
it could be that the best anchor point is the parent's previous sibling for all i know

[16:37:01.0000] <Lachy>
JohnResig, do you have some example pages that use the JS libraries in the way you describe?

[16:37:02.0000] <JohnResig>
Lachy: using contextual selectors? how many do you want?

[16:37:03.0000] <Hixie>
5 would be nice

[16:43:00.0000] <jwalden>
Hixie: 'If the scheme is "file:", then the user agent may return a UA-specific origin.' -- there shouldn't be a colon in "file", methinks

[16:44:00.0000] <Hixie>
you sure?

[16:45:00.0000] <Hixie>
i forget if scheme includes the colon or not

[16:45:01.0000] <Hixie>
ah it does not

[16:45:02.0000] <Hixie>
ok

[16:45:03.0000] <Hixie>
i have removed it

[16:45:04.0000] <Hixie>
thanks!

[16:45:05.0000] <Hixie>
(won't appear in the spec for a while)

[16:46:00.0000] <Philip`>
Are contextual selectors in jQuery done with thing.filter('selector')?

[16:46:01.0000] <JohnResig>
Here's about 200, give or take, http://www.google.com/codesearch?hl=en&lr=&q=jquery+\.find+-%22New+Wave+JavaScript%22+file%3Ajs%24&btnG=Search

[16:47:00.0000] <Lachy>
/me looks...

[16:47:01.0000] <JohnResig>
Philip`: those are filters - contextual selectors are done with .find()

[16:47:02.0000] <Philip`>
JohnResig: Ah, thanks

[16:47:03.0000] <JohnResig>
Lachy: those are only the ones that use .find() - there's a bunch that do jQuery("selector", element)

[16:48:00.0000] <JohnResig>
this is also ignoring the ones that are used in jQuery directly - and ones that are used by other frameworks

[16:49:00.0000] <JohnResig>
some of these queries won't be applicable to querySelectorAll (unless, of course, it's made to work on document fragments)

[16:49:01.0000] <Lachy>
what's the difference between .find() and jQuery("selector", element)?

[16:49:02.0000] <JohnResig>
Lachy: there is none - they're identical

[16:50:00.0000] <JohnResig>
doing jQuery("selector", element) just maps to jQuery(element).find("selector")

[16:50:01.0000] <Philip`>
The first few I find via dmoz.org are ...

[16:50:02.0000] <Philip`>
http://www.dorez.com/ - $(this).find("+ ul")

[16:50:03.0000] <Philip`>
http://www.queens.edu/athletics/wvolleyball/home.asp - $("#nav").superfish(...).find(">li[ul]")

[16:50:04.0000] <Philip`>
http://www.fortdalles.com/ - $(this).find("a")

[16:50:05.0000] <Philip`>
(plus another dozen or so)

[16:50:06.0000] <JohnResig>
the second one is using an old version of jQuery - but the other two are applicable

[16:51:00.0000] <JohnResig>
the second one would be written .find("> li:has(ul)")

[16:51:01.0000] <JohnResig>
Philip`: what's the search that you used?

[16:52:00.0000] <Lachy>
.find("+ ul") one couldn't work even with :scope, or any other proposed solution, since .querySelector only returns descendant nodes

[16:53:00.0000] <JohnResig>
as the spec is written, that is

[16:55:00.0000] <Lachy>
I'm not even sure how it could be rewritten to add the ability, and doing so would add a significant amount of development time for both the spec and implementations

[16:55:01.0000] <Philip`>
Hmm, I also found one duplicate of the superfish script, and two of the find("a") script, and the rest are:

[16:55:02.0000] <Philip`>
http://cafe.daum.net - $(this).find("CONTENT")

[16:55:03.0000] <Philip`>
http://www.abenteuer-reisen.de/wg/ph/ http://www.abenteuer-reisen.de/wg/cv/wg_cv__rf00__index.htm - $(this).find("option:selected")

[16:55:04.0000] <Philip`>
http://www.mat.uni.torun.pl/wosp2001/ - portlet.find('.portlet-content-container')

[16:55:05.0000] <Philip`>
http://www.rodgersinstruments.com/ - $(this).find("a[@class='active']")

[16:55:06.0000] <Philip`>
http://www.dostavka.ru/ - $node.find("div.node_ajax")

[16:55:07.0000] <Lachy>
$(this).find("a") works with the current API

[16:56:00.0000] <Philip`>
(or at least that's what they were two months ago - at least the last of those sites has changed now)

[16:56:01.0000] <Philip`>
JohnResig: I downloaded ~130K pages that were listed on dmoz.org a couple of months ago, then grepped for /\.find\(["']/

[16:57:00.0000] <JohnResig>
Philip`: ah, neat

[16:57:01.0000] <Lachy>
so the proposed changes are only relevant when one wants to acquire either a child node of the context node (":scope>*") or using descendant combinators, such that all parts of the selector only match elements within the context (":scope foo bar")


2008-05-03
[17:00:00.0000] <Lachy>
with the current spec, only the last simple selector in the sequence needs to match an element within context, whereas all others can match any ancestors or previous siblings (depending on the combinators)

[17:00:01.0000] <Philip`>
JohnResig: Just in case it's of any interest: Of those 130K pages, http://philip.html5.org/misc/jquery-pages.txt lists the ones matching /<script[^>]*jquery/

[17:00:02.0000] <JohnResig>
Philip`: neat - thanks!

[17:00:03.0000] <Hixie>
":scope>*" isn't a valid use case for this, that's what the element traversal stuff is for

[17:01:00.0000] <JohnResig>
Hixie: but :scope > div.foo.bar[baz] is totally valid here

[17:01:01.0000] <Hixie>
yeah, what's the page that does that?

[17:02:00.0000] <Hixie>
most seem to just do a single simple selector chain

[17:02:01.0000] <Hixie>
or whatever the term is these days

[17:02:02.0000] <Lachy>
Philip`, I only used * in :scope>* because I was referring to the general case

[17:02:03.0000] <Hixie>
no combinators, i mean

[17:02:04.0000] <Philip`>
Lachy: s/Philip`/Hixie/ ?

[17:02:05.0000] <Hixie>
oh

[17:02:06.0000] <JohnResig>
here's one with at least two: jQuery(element.form || document).find('[@name=' + element.name + ']:checked')

[17:03:00.0000] <Hixie>
that would work fine

[17:03:01.0000] <Hixie>
there's no combinator there

[17:03:02.0000] <Lachy>
Hixie, there are a couple that I have seen using .find(">x")

[17:03:03.0000] <Hixie>
Lachy: interesting syntax

[17:03:04.0000] <Lachy>
yeah, it's the equivalent to ":scope>x"

[17:04:00.0000] <Hixie>
right

[17:04:01.0000] <Lachy>
and it's the one I'm not sure how to solve without :scope

[17:04:02.0000] <Hixie>
seems like most of them don't do this

[17:04:03.0000] <Hixie>
which would suggest it's not a high prioprity

[17:04:04.0000] <JohnResig>
 jQuery("> tbody:first/tr",o).find("> td:eq(" + COLUMN_LAST_INDEX + ")")

[17:04:05.0000] <Hixie>
priority

[17:04:06.0000] <JohnResig>
Hixie: they don't do it because it's slow - the point of this whole new API is to make it fast

[17:05:00.0000] <JohnResig>
if we're not striving to make things faster - that pure JS can't make faster - then there's really no point

[17:05:01.0000] <Hixie>
this data suggests many people use the api, just not with combinators

[17:05:02.0000] <Hixie>
i've no idea what "tbody:first/tr" is supposed to match

[17:05:03.0000] <Lachy>
nor do I.

[17:05:04.0000] <Hixie>
but it certainly won't work with the api anyway, so that again isn't an argument in favour of this

[17:05:05.0000] <Hixie>
i have to go to another building, but i'll be on irc again briefly

[17:06:00.0000] <Hixie>
s/briefly/momentarily/

[17:06:01.0000] <JohnResig>
sure it is - we support some simple XPath expressions - we map them to CSS before parsing

[17:06:02.0000] <Lachy>
ok, so it tbody/tr the same as "tbody tr" or "tbody>tr"?

[17:06:03.0000] <JohnResig>
"tbody>tr"

[17:07:00.0000] <Philip`>
How would you do :eq with the Selectors API?

[17:08:00.0000] <JohnResig>
Philip`: we wouldn't - we'd have to fall back - which is why I've been arguing for good error reporting so that we can actually fallback gracefully

[17:11:00.0000] <Philip`>
Is this almost like Array.forEach(o.querySelector('tbody').querySelectorAll('tr'), function (tr) { var td = tr.querySelectorAll('td')[COLUMN_LAST_INDEX]; foo(td) }) ?

[17:11:01.0000] <Lachy>
JohnResig, without :scope, how would you propose the API work to find child nodes of the context node in the following scenario?

[17:11:02.0000] <Philip`>
(Oh, except that won't work with nested tables)

[17:12:00.0000] <JohnResig>
Philip`: roughly, yeah - it would seem!

[17:12:01.0000] <Lachy>
<div id=foo>A  <div>B  <div>C</div> B</div> A</div>

[17:13:00.0000] <JohnResig>
Lachy: without scope, just look for the front-leading >. I'm not really sure what the purpose of introducing :scope is if there's only one combinator that it'll work on

[17:13:01.0000] <Lachy>
with foo.querySelectorAll("..."); how would you find all child divs (the ones labelled B)

[17:13:02.0000] <Philip`>
After the querySelector('tbody'), it wants some way to find all the <tr>s that are direct children of that particular <tbody> (and not e.g. children of a tbody which is a descendant of that first tbody)

[17:13:03.0000] <Lachy>
because ">div" is an invalid selector

[17:13:04.0000] <JohnResig>
Lachy: unless it's defined to be valid

[17:13:05.0000] <Lachy>
and allowing that would require redefining the selector syntax

[17:14:00.0000] <JohnResig>
Lachy: new syntax is already being defined with :scope - it's not a leap of imagination

[17:14:01.0000] <Lachy>
the difference is :scope doesn't require chaning the grammar

[17:14:02.0000] <Lachy>
it's just another pseudo class

[17:15:00.0000] <jgraham_>
FWIW selectors like ">div" strike me as a bad idea

[17:15:01.0000] <othermaciej>
Lachy, JohnResig: since you're both here...

[17:15:02.0000] <othermaciej>
I was going to suggest on the list that there could be queryScopedSelector()

[17:15:03.0000] <JohnResig>
Lachy: but yeah, with the current spec, it's impossible to reliably find the <div>B one

[17:15:04.0000] <othermaciej>
which implicitly prepends ":scope "

[17:16:00.0000] <othermaciej>
and that might be what JS libraries want to use

[17:16:01.0000] <JohnResig>
othermaciej: I'd contend that it's what users want to use

[17:16:02.0000] <othermaciej>
at least that would clarify that it is using a different notion of a "scoped selector" instead of standard slector grammar and semantics

[17:16:03.0000] <othermaciej>
JohnResig: is that feedback in favor of or against my suggestion?

[17:17:00.0000] <JohnResig>
othermaciej: I think it's fine if it's implied - I just think that querySelectorAll should do that to begin with.

[17:17:01.0000] <Lachy>
othermaciej, are you suggesting that .queryScopedSelector(">div") implicitly becomes ":scope >div"?

[17:18:00.0000] <Philip`>
Does HTML5's <style scoped> have exactly the same issues?

[17:18:01.0000] <othermaciej>
Lachy: yes

[17:18:02.0000] <Lachy>
Philip`, yeah, introducing :scope addresses use cases for that too

[17:18:03.0000] <JohnResig>
othermaciej: I'm also assuming that there'd be .queryScopedSelectorAll and .queryScopedSelector?

[17:18:04.0000] <othermaciej>
Lachy: similarly, queryScopedSelector("div") becomes queryScopedSelector(":scope div")

[17:19:00.0000] <othermaciej>
Lachy: sorry, just querySelector the second time

[17:19:01.0000] <othermaciej>
JohnResig: presumably, yes

[17:20:00.0000] <othermaciej>
JohnResig: the de facto syntax for selectors supported by JS library query APIs does not match standard selectors, and I am not sure changing the core selector syntax is the way to go, but I think this could be a plausible workaround that gives a simple well-defined notion of a "scoped selector"

[17:20:01.0000] <othermaciej>
with grammar and semantics that better match JS libraries

[17:21:00.0000] <othermaciej>
of course there is still the fact that JS libraries support some extensions that CSS3 Selectors expressly forbids, which maybe should be handled by extending Selectors (things like ":not(a > b)")

[17:21:01.0000] <othermaciej>
I'm not sure what all the differences are

[17:23:00.0000] <Lachy>
I guess there's one advantage to introducing a separate method is that is allows libraries to test for it like: if (element.queryScopedSelector) {... }

[17:23:01.0000] <JohnResig>
I'm just completely confused as to why this wasn't done in the first place - and why it's not the default. Alternatively, I'd like to propose that querySelectorAll("selector") be equivalent to .queryGlobalSelector(":scope selector") (changing the primary querySelectorAll method to actually match precedence)

[17:24:00.0000] <othermaciej>
JohnResig: we had lots of input from UA implementors and CSS experts, but not so much from JS library authors or AJAX developers, when drafting the spec

[17:25:00.0000] <JohnResig>
:-|

[17:25:01.0000] <othermaciej>
so I think everyone assumed that normal selector syntax and semantics were what is desired

[17:25:02.0000] <Lachy>
I certainly assumed that

[17:25:03.0000] <Lachy>
while I was aware of the problem, I wasn't aware it was such a priority

[17:25:04.0000] <Lachy>
and assumed that it could be safely left to v2 and solved with :scope

[17:27:00.0000] <othermaciej>
in fact queryScopedSelector could even be defined without Selectors defining ":scope" first, because it could say it acts as if an unspecified simple selector uniquely matching the scope element is prepended

[17:28:00.0000] <othermaciej>
(I guess to each selector in a group of selectors, not just stringwise)

[17:28:01.0000] <othermaciej>
(since you want comma to work as expected)

[17:28:02.0000] <othermaciej>
so UAs could use an arbitrary mechanism to achieve that, and ":scope" can be added to the CSS Selectors spec later

[17:30:00.0000] <Lachy>
ok, I'll think about it.

[17:30:01.0000] <JohnResig>
othermaciej: it seems like it would be better to not overload that method and, instead, provide a primary filtering method .filterSelector("div.class")

[17:30:02.0000] <JohnResig>
othermaciej: of course, that's more useful if the method also exists on DOMNodeLists

[17:31:00.0000] <Lachy>
I'm not sure how .filterSelector would work on anything but NodeLists

[17:31:01.0000] <othermaciej>
for NodeLists you obviously can't support scoped selectors

[17:31:02.0000] <Lachy>
but then, as I wrote in the email, I really need to see the use cases for it

[17:32:00.0000] <othermaciej>
I'm not sure what filterSelector does

[17:32:01.0000] <Lachy>
indeed. .filterSelector() would have to use the global context

[17:32:02.0000] <Lachy>
othermaciej, it takes a NodeList and evaluates the selector against the element in the context of the document

[17:33:00.0000] <Lachy>
and returns all matching elements

[17:33:01.0000] <othermaciej>
that's sorta the same thing as querySelectorAll, as currently defined

[17:33:02.0000] <othermaciej>
isn't it?

[17:34:00.0000] <Lachy>
it's not quite, but the reason I need use cases for it is because I can't see what it can solve that querySelectorAll can't

[17:35:00.0000] <othermaciej>
is it that it can only match immediate children?

[17:36:00.0000] <othermaciej>
I guess thinking about it, querySelectorAll on a NodeList seems kind of weird

[17:36:01.0000] <othermaciej>
when would you want to do selector matching on a list of DOM subtrees?

[17:36:02.0000] <Lachy>
ok, so the advantages for queryScopedSelector/queryScopedSelectorAll() are: 1. Doesn't break the existing API, 2. Can be tested for by scripts, 3. Doesn't depend upon :scope being defined immediately

[17:36:03.0000] <othermaciej>
seems more likely you'd want to do it on a list of actual nodes

[17:37:00.0000] <Lachy>
othermaciej, I think your interpretation of how it would work is different from mine

[17:38:00.0000] <Lachy>
or I'm not sure what you meant by "selector matching on a list of DOM subtrees?"

[17:38:01.0000] <othermaciej>
Lachy: I am not sure how you see it working either

[17:38:02.0000] <Lachy>
ok, consider this:

[17:38:03.0000] <othermaciej>
I am drawing conclusions based on you saying filterSelector and querySelector all would be different

[17:38:04.0000] <othermaciej>
here is what I assumed the difference is:

[17:38:05.0000] <Lachy>
var list = document.getElementsByTagName("div");

[17:38:06.0000] <othermaciej>
filterSelector - can only match items actually in the node list, not their descendants

[17:39:00.0000] <Lachy>
var list2 = list.filterSelector(".foo");

[17:39:01.0000] <othermaciej>
querySelectorAll - can match items in the node list, or any of their descendants

[17:39:02.0000] <Lachy>
list 2 contains all the divs from the first list that had a class of foo

[17:39:03.0000] <Lachy>
basically, the result is the same as querySelectorAll("div.foo")

[17:39:04.0000] <othermaciej>
that seems to match my interpretation of filterSelector

[17:39:05.0000] <Lachy>
which is useless

[17:40:00.0000] <Lachy>
ok, good

[17:40:01.0000] <othermaciej>
Is my understanding of how NodeList.querySelectorAll would differ correct?

[17:41:00.0000] <othermaciej>
I can see how you could do things with filterSelector + querySelectorAll that can't be done with just QSA

[17:41:01.0000] <Lachy>
I assumed they would were just different possible names for the same thing.

[17:41:02.0000] <othermaciej>
for instance:

[17:41:03.0000] <Lachy>
I didn't think there were 2 separate suggestions

[17:42:00.0000] <othermaciej>
var list = document.querySelectorAll("div *")

[17:42:01.0000] <othermaciej>
var list2 = list.filterSelector("section *")

[17:42:02.0000] <othermaciej>
list 2 is all elements contained in both a div and a section, in either order

[17:42:03.0000] <othermaciej>
no, I guess I am wrong

[17:43:00.0000] <othermaciej>
you could instead query for "div section *, section div *"

[17:43:01.0000] <Lachy>
oh yeah, that's right. It's sort of like the :matches() proposal in Selectors

[17:43:02.0000] <othermaciej>
but with many independent conditions you could get a huge explosion of selector length

[17:43:03.0000] <othermaciej>
and maybe some trickier case is not trivially convertible to a group of two selectors

[17:44:00.0000] <Lachy>
yeah, I'm sure there are some rather complicated cases that it could simplify. But are there any such use cases that people are actually attempting?

[17:44:01.0000] <othermaciej>
you might also want to do foo.childNodes.filterSelector(), but queryScopedSelectorAll could do that

[17:44:02.0000] <othermaciej>
I dunno

[17:44:03.0000] <othermaciej>
I am just trying to understand it

[17:54:00.0000] <Lachy>
I think that would mean foo.queryScopedSelectorAll("div").filterSelector("section *"); would be equivalent to foo.querySelector("section div");

[17:55:00.0000] <othermaciej>
Lachy: I think that's right

[17:58:00.0000] <Lachy>
queryScopedSelector() would need to prepend ":scope " (or equivalent magical selector) to evey sequence of simple selectors, so "strong, em" became ":scope strong, :scope em"

[17:59:00.0000] <othermaciej>
right

[18:02:00.0000] <Lachy>
ok, I've got a lot too think about and a lot more to do. I have to find out if IE, Mozilla and Opera would be willing and able to implement queryScopedSelector, then define it, and still evaluate whether or not it really is the best solution to go for

[18:15:00.0000] <hasather>
Hixie: seen this: http://xiphmont.livejournal.com/31209.html

[18:36:00.0000] <Lachy>
queryScopedSelector() won't work as easily as I initially thought. :-(

[18:36:01.0000] <Lachy>
Consider this:

[18:37:00.0000] <Lachy>
foo.queryScopedSelector(">strong, >em");

[18:38:00.0000] <Lachy>
I would basically need to define a special grammar for selectors used in this method in a way that is compatible with normal selectors, and doesn't break forwards compatibility

[18:38:01.0000] <Lachy>
it's basically a complete nightmare

[18:41:00.0000] <Hixie>
hsivonen: that doesn't look like the kind of stuff i'd write

[18:41:01.0000] <Hixie>
i don't even known how to spell antithesis

[18:43:00.0000] <kevogod>
I remember reading something about that months ago.

[18:44:00.0000] <kevogod>
http://64.233.167.104/search?q=cache:VvbdPyHxdxAJ:ventnorsblog.blogspot.com/2008/04/rated-p-for-proprietary.html%3FshowComment%3D1209354480000+The+distinction+between+the+video+format+war+and+the+browser+war&hl=en&ct=clnk&cd=2&gl=us&client=firefox-a

[18:44:01.0000] <Hixie>
yeah i just posted a link to that

[18:44:02.0000] <kevogod>
Oh, sorry.

[18:45:00.0000] <kevogod>
I am curious how "Michael" became Hixie.

[18:46:00.0000] <Lachy>
Hixie, I was wondering why the article implied you no longer had a blog.

[18:58:00.0000] <Philip`>
kevogod: The keys are like right next to each other

[18:58:01.0000] <kevogod>
Philip`, True, I can see how the mistake was made.

[22:28:00.0000] <Facedown>
positioned elements (other than static) have a higher stacking order than floated non-positioned elements (z-index not explicitly set) , dont they?

[23:07:00.0000] <roc>
yes

[23:13:00.0000] <Facedown>
have you run into any IE specific issues where IE does not honor the rightful stacking order, roc?

[23:13:01.0000] <roc>
IE treats z-index:auto as z-index:0 IIRC

[23:13:02.0000] <roc>
not sure about any other stacking order bugs

[23:14:00.0000] <roc>
I hardly ever use IE :-)

[23:14:01.0000] <Facedown>
no self proclaimed web dev would ever personally use it.. but clients..

[23:14:02.0000] <roc>
I work on Firefox. I don't have clients :-)

[23:15:00.0000] <Facedown>
heh

[02:14:00.0000] <hsivonen>
I wish we could magically make CRLF disappear

[02:22:00.0000] <Facedown>
what's CRLF? 

[02:22:01.0000] <Facedown>
\r\n ?

[02:23:00.0000] <Facedown>
carriage return line feed?

[02:30:00.0000] <hsivonen>
Facedown: yes. It's evil.

[03:12:00.0000] <jgraham_>
Sigh. Nice of Daniel to post such a well argued, non-inflammatory addition to the alt thread

[03:32:00.0000] <Lachy>
oh, yeah, cause arguments like "making alt optional in HTML 5 is ridiculous" is exactly the kind of short and to the point arguments we need.

[03:47:00.0000] <jgraham_>
/me ponders the recursive nature of making non-conformance conforming

[04:15:00.0000] <hsivonen>
hmm. perhaps I should just do a preflight on each buffer to get rid of CRLF instead of making the rest of the tokenizer really complex in order to avoid the preflight

[04:16:00.0000] <hsivonen>
I don't usually admit to hating things, so I seriously dislike CRLF

[04:30:00.0000] <jgraham_>
hsivonen: IIRC we strip out CRLF when we buffer the data

[04:31:00.0000] <jgraham_>
(and have a flag to indicate trailing CR in case CRLF is split over two buffers)

[11:40:00.0000] <jgraham_>
So Nvu/KomoZer adds alt="" even when you specify "Don't add alternate text"

[11:42:00.0000] <jgraham_>
And adds alt="image filename" if you drag and drop an image file in

[11:42:01.0000] <jgraham_>
s/KomoZer/KompoZer/

[11:42:02.0000] <jgraham_>
Which clearly sucks

[11:58:00.0000] <Philip`>
Maybe it'd be vaguely informative to correlate alt usage with <meta name=generator> values

[12:27:00.0000] <Facedown>
what's the best ftp client anyone here has ever used? (on windows preferably)? filezilla 3 is a memory hog like firefox.. slow as shit 

[12:32:00.0000] <Philip`>
I use Krusader (on Linux)

[13:51:00.0000] <Hixie>
i wonder why the alt="" people aren't screaming about <object></object>

[13:52:00.0000] <Hixie>
which is even worse since it doesn't distinguish between unknown important image and decorative image

[13:52:01.0000] <Philip`>
Probably because nobody uses <object> decoratively

[13:53:00.0000] <jgraham_>
Probably because nobody uses <object> for images

[13:53:01.0000] <Philip`>
and theoretical problems aren't worth worrying about

[13:53:02.0000] <Hixie>
well then shouldn't they be arguing that <object> must never be empty?

[13:53:03.0000] <Hixie>
what i find interesting is that the longer this argument goes on, the more people are coming out and defending what the spec says

[13:53:04.0000] <Philip`>
jgraham_: I use <object> for images, at least twice in my life :-)

[13:53:05.0000] <Philip`>
although actually one of those was purely decorative

[13:54:00.0000] <jgraham_>
OK "nobody" for large values of "nobody"

[13:55:00.0000] <Philip`>
Hixie: That's because people flock to authority, which the spec provides, rather than considering the technical arguments ;-)

[13:56:00.0000] <Lachy>
I used <object> for an image once, and I put an <img> inside it as fallback :-)

[13:56:01.0000] <Philip`>
Lachy: And an alt on the image as fallback?

[13:56:02.0000] <Lachy>
yes

[13:56:03.0000] <Hixie>
Philip`: if anyone is being an authority here it's the self-proclaimed accessibility experts who announce their opinions as unquestionable

[13:56:04.0000] <jgraham_>
I am quite tempted to say why I think that the Nvu UI encourages bad alt text and required alt encourages Nvu-like UI. But if I do that it will be taken badly...

[13:57:00.0000] <Philip`>
The logo on http://canvex.lazyilluminati.com/ does <h1><object><img alt></...>

[13:57:01.0000] <Philip`>
which is hopefully sane

[13:57:02.0000] <Lachy>
IMHO, Nvu is a really bad editor

[13:57:03.0000] <Hixie>
jgraham_: just ask Philip` for a list of pages that have nvu's signature, and do a study

[13:58:00.0000] <jgraham_>
Hixie: AFAICT Nvu doesn't have a signature

[13:58:01.0000] <jgraham_>
(at least the version that comes with Ubuntu)

[13:58:02.0000] <Hixie>
jgraham_: and compare it to a list of pages that have the signature of tools that make it optional and don't complain and see what they're like

[13:58:03.0000] <Hixie>
actually this would be an interesting study

[13:58:04.0000] <jgraham_>
Hixie: I agree this would be an interesting study

[13:58:05.0000] <Hixie>
i should get us a list of pages with various detailed generator strings

[13:59:00.0000] <Hixie>
someone remind me later today and i'll see about setting that up

[13:59:01.0000] <Hixie>
gotta go now though

[13:59:02.0000] <Philip`>
http://philip.html5.org/data/underline-generators.txt was my last attempt at looking at generators

[13:59:03.0000] <Lachy>
Hixie, at what time do you want to be reminded?

[14:00:00.0000] <Hixie>
Lachy: now + 8 hours

[14:00:01.0000] <Lachy>
OK, I'll be asleep by then. Someone else will have to do it

[14:01:00.0000] <Philip`>
Google Web History says 5am is not my time of greatest awakeness

[14:01:01.0000] <Hixie>
well remind me in 2 hours, i'll be gone then and will see the reminder when i get back :-D

[14:02:00.0000] <Hixie>
bbl :-)

[14:02:01.0000] <Philip`>
Looking for (?i)generator.*nvu I see one <meta name="GENERATOR" content="NVu"> and one <meta name="generator" content="nvu"> and one <meta name="generator" content="Nvu v1.0">

[14:04:00.0000] <Philip`>
I see more pages with <comment title=" ... " xmlns="http://disruptive-innovations.com/zoo/nvu">, which presumably come from some form of interaction with Nvu

[14:05:00.0000] <Philip`>
...where "more" means 8

[14:08:00.0000] <Philip`>
Hixie: The problem with that kind of study is you have no way of knowing whether use of a particular authoring tool causes bad alt text, rather than some other external influence causing both, so you can't conclude anything

[14:09:00.0000] <Philip`>
e.g. maybe people who choose Nvu are more technically competent than those who choose FrontPage, and they'll do better alt text regardless of the editor's UI

[14:10:00.0000] <jgraham_>
Philip`: Yeah there are systematic biases in all surveys like this

[14:10:01.0000] <jgraham_>
You just have to try and understand them

[14:10:02.0000] <jgraham_>
rather than conclude that everything is pointless

[14:10:03.0000] <Philip`>
You should try to eliminate them

[14:10:04.0000] <Philip`>
and it's only pointless if you find you can't do that :-)

[14:11:00.0000] <jgraham_>
Also pasting <comment title=" ... " xmlns="http://disruptive-innovations.com/zoo/nvu">foo</comment> into the source view of KompoZer and then switching back to normal view seems to lock the application up with 100% CPU

[14:11:01.0000] <Philip`>
Just because it's not possible to do a good large-scale study, that doesn't mean you should do a bad one and think its conclusions have any relevance to reality

[14:12:00.0000] <jgraham_>
Philip`: If you have an idea of what the biases are then you can determine whether they are likely to affect the outcome in the way that you observe

[14:13:00.0000] <jgraham_>
e.g. if you think that all Nvu users are more competent than Frontpage users but find that Nvu pages have worse alt text, the effect is most likely real

[14:14:00.0000] <jgraham_>
rather than an artifact of the people using the software

[14:14:01.0000] <Philip`>
If that happened, it probably just means I didn't have a good idea of the biases

[14:14:02.0000] <Philip`>
(If it didn't happen, I still wouldn't think I had a good idea of the biases)

[14:15:00.0000] <Philip`>
Maybe Nvu users are less competent than FrontPage users, or maybe there's one extremely prolific Nvu-using author who writes really bad alt text

[14:15:01.0000] <jgraham_>
Philip`: You don't make any progress by not doing things because you can't do them perfectly. Of course you should report any problems your study might have

[14:17:00.0000] <Philip`>
jgraham_: Indeed, but 'not quite perfect' and 'so far from perfect that the results are meaningless' are not the same

[14:18:00.0000] <jgraham_>
Philip`: Of course :) I don't think this study would necessarily produce meaningless results though

[14:18:01.0000] <Philip`>
So how could we be sure it was not going to produce meaningless results? :-)

[14:19:00.0000] <jgraham_>
We couldn't — at least for a strict definition of "sure"

[14:19:01.0000] <Philip`>
(preferably without first collecting a load of data and then checking whether it corresponds to our preconceived ideas of what the correct results should be :-) )

[14:19:02.0000] <Philip`>
s/sure/reasonably confident/

[14:19:03.0000] <Philip`>
s/reasonably confident/sufficiently confident that we could make language design decisions based on it/

[14:19:04.0000] <Facedown>
Philip` - <h1><object><img> ? hrm

[14:20:00.0000] <jgraham_>
But we could consider all the thinks that might be wrong with it and work out whether any effects should average out over the sample or have a systematic effect

[14:20:01.0000] <Facedown>
I usually dont put my logos in h1 since i don't think they're real headings

[14:20:02.0000] <Facedown>
<div id="logo"><img></div> or just <img id="logo"> usually does it.. 

[14:20:03.0000] <jgraham_>
We then decide if the systematic effect will be big enough to worry about

[14:20:04.0000] <Philip`>
Facedown: A non-visual UA that replaces the image with its alt text ought to use heading formatting for the alt text

[14:21:00.0000] <jgraham_>
If we're not sure, we do the study, collect the results and see if the significant of the result is greater than our estimate of the systematic effects

[14:21:01.0000] <Facedown>
? to say the logo out louder?

[14:21:02.0000] <Facedown>
isnt that what aural styles are for?

[14:21:03.0000] <Facedown>
hehe

[14:22:00.0000] <Philip`>
Facedown: It could be a non-graphical text browser, rather than an aural one

[14:22:01.0000] <Philip`>
Facedown: and I'd rather aural browsers used the default <h1> aural styles, rather than some which I've made up myself and haven't been able to test

[14:22:02.0000] <jgraham_>
(also if we were doing the analysis in a proper Bayesian sense, our preconceived idea of what the result should be is important)

[14:22:03.0000] <Facedown>
My eyes usually go on the main heading in a newspaper rather than look at the logo

[14:23:00.0000] <Philip`>
Facedown: In my case, the logo is the main heading (since it's just fancy text)

[14:25:00.0000] <Philip`>
jgraham: So, how do we quantify the systematic biases in this case?

[14:26:00.0000] <Philip`>
(Oops, I've got to go out and find food)

[14:26:01.0000] <jgraham>
/me suggests eating the pigeons 

[14:27:00.0000] <jgraham>
(since you sound o hunter-gather about it ;)

[14:27:01.0000] <jgraham>
s/o/so/

[14:28:00.0000] <Philip`>
jgraham: I've not seen any pigeons around where I live :-(

[14:28:01.0000] <Philip`>
Quite a few cats wander around our garden, though

[14:30:00.0000] <jgraham>
Apparently carnivores taste quite bad

[14:51:00.0000] <Philip`>
Hmm, I did manage to sneak up on a cat, but I didn't manage to actually see it before it had already decided to run away

[14:51:01.0000] <Philip`>
and my spear arm wasn't quick enough

[16:00:00.0000] <Philip`>
/me thinks "market-leading HTML5 validator" is an odd phrase to use when the entire market consists of about ten people and one validator

[16:08:00.0000] <hsivonen>
Philip`: if there's only one, it's necessarily the top product in its category, isn't it? :-)

[16:12:00.0000] <Philip`>
hsivonen: It's technically a true statement, but the phrase has connotations that are not applicable in this case :-)

[16:21:00.0000] <zokka>
hello

[16:23:00.0000] <zokka>
i think they should change the code tag

[16:24:00.0000] <zokka>
so it displays special characters

[16:26:00.0000] <hsivonen>
zokka: who's "they" and what special characters are you referring to?

[16:26:01.0000] <zokka>
<>

[16:26:02.0000] <hsivonen>
zokka: there's already <xmp> for that

[16:27:00.0000] <hsivonen>
zokka: it doesn't add to the expressiveness of the language and has the problem of not being able to represent an example of itself

[16:27:01.0000] <zokka>
that went over my head

[16:27:02.0000] <zokka>
:D

[16:28:00.0000] <zokka>
xmp is deprecreated though

[16:28:01.0000] <hsivonen>
zokka: you can use <code> and <pre> to show an example of <code> or <pre>

[16:28:02.0000] <hsivonen>
zokka: you can't use <xmp> to show an example of <xmp>

[16:29:00.0000] <hsivonen>
zokka: and for the "expressiveness" part, <xmp> doesn't really help you say anything you couldn't say with <pre>

[16:29:01.0000] <zokka>
i like the auto escaping of special characterst though on it

[16:29:02.0000] <zokka>
it doesnt seem natural to use escape codes

[16:30:00.0000] <zokka>
i see your meaning though markup might look bad when using html

[16:34:00.0000] <jwalden>
*cough* cdata *cough*

[16:35:00.0000] <zokka>
gotta go


2008-05-04
[17:03:00.0000] <Philip`>
jwalden: Like XML's <![CDATA[... stuff? That makes it hard to write an example of XML code that uses CDATA, so it doesn't solve any problems :-p

[17:03:01.0000] <gsnedders>
That's SGML! :P

[17:03:02.0000] <gsnedders>
Perfectly valid (and useless)  HTML 4.01!

[17:03:03.0000] <Philip`>
I prefer to believe that SGML doesn't exist

[17:04:00.0000] <Philip`>
My life is much simpler that way

[17:04:01.0000] <jwalden>
being able to nest is a false goal, in my book

[17:04:02.0000] <Philip`>
There's just HTML and XML

[17:16:00.0000] <gsnedders>
Philip`: And gsneddersML.

[17:17:00.0000] <gsnedders>
With elements such as "lust" and "rant".

[21:19:00.0000] <Hixie>
Philip`: validator.nu is market leading in the sense that it's changing the game, too

[06:04:00.0000] <Hixie>
boy am i glad i don't own XHR anymore

[06:24:00.0000] <Lachy>
Hey Hixie, did you remember to do that study on the markup generator strings?

[07:33:00.0000] <Philip`>
hsivonen: It's really quite confusing how Validator.nu sometimes stops half way through a document

[07:33:01.0000] <Philip`>
e.g. http://validator.nu/?doc=http%3A%2F%2Fwww.millermicro.com%2Fnatick.html&showimagereport=yes&showsource=yes#imagereport made me wonder why it was missing a load of images that were on the page

[08:13:00.0000] <annevk>
I don't understand the opting into cookies proposal...

[08:14:00.0000] <annevk>
Does it mean user agents can decide to do cookies and if they do they include some headers?

[08:15:00.0000] <annevk>
oh wait, the server includes that header, not the user agent

[08:15:01.0000] <annevk>
hmm

[09:00:00.0000] <Lachy>
oh dear, I thought this thread was over :-( http://lists.w3.org/Archives/Public/public-html/2008May/0002.html

[09:09:00.0000] <Dashiva>
Lachy: Various slashes movies come to mind

[09:09:01.0000] <Dashiva>
*slasher

[09:14:00.0000] <Lachy>
Dashiva, which movies?

[09:17:00.0000] <Dashiva>
The sequels :) "They thought he was dead, but they thought wrong!"

[09:26:00.0000] <Philip`>
The only way to really kill those characters is to not watch the movies, since they'll stop making sequels when there's no audience

[09:29:00.0000] <Philip`>
But maybe it's more like the end-of-game bosses in Final Fantasy style games, where you have to defeat them in combat seven times and then use a special sequence of attacks that you found on GameFAQs to finally see them off

[09:38:00.0000] <Philip`>
/me wishes Gmail interacted correctly with Thunderbird's automatic removal of " (was: ...)" from subject lines, instead of thinking the subject change meant it was a whole new thread

[10:39:00.0000] <Philip`>
/me finds yet another Firefox canvas crash

[10:39:01.0000] <Philip`>
They're worryingly common :-(

[10:41:00.0000] <Dashiva>
Maybe they're hydra-based. Every time you find one, two more are created.

[10:47:00.0000] <Philip`>
Hmm, looks like just a null pointer thingy, so not particularly exploitable :-(

[10:47:01.0000] <hsivonen>
Philip`: I need to do something about image report upon fatal error, yes. thanks

[10:54:00.0000] <Philip`>
Oh, it's already reported as https://bugzilla.mozilla.org/show_bug.cgi?id=421715 - I guess nobody cares enough to fix that issue in FF2

[10:55:00.0000] <Philip`>
/me adds UA-sniffing to his canvas test suite, to avoid that crash

[11:05:00.0000] <gavin_>
Philip`: we can take that for a dot release

[11:05:01.0000] <gavin_>
I'll ask for approval

[11:07:00.0000] <Philip`>
gavin_: Ah, thanks, that would make me happy :-)

[11:10:00.0000] <Dashiva>
/me reads "alt and authoring practices"

[11:11:00.0000] <Dashiva>
That's a nice way to make a gallery that never reveals it does, in fact, contain images

[11:31:00.0000] <Philip`>
Hmph - WebKit doesn't support colours rgb(x,x,x) when x >= 2^31, and Firefox 2 doesn't support them when x >= about 10^39, and Opera 9.2 doesn't support them when x >= 2^31 - 64

[11:32:00.0000] <Philip`>
Silly browsers :-(

[11:32:01.0000] <Philip`>
(FF3 and O9.5 seem to get it right, though)

[11:33:00.0000] <Philip`>
/me wonders if he's testing things that don't have great real-world relevance

[11:36:00.0000] <Dashiva>
Well, if you manage to squeeze 10^39 into one byte, that's pretty significant :)

[11:39:00.0000] <Philip`>
Well, I don't expect the browser to make that pixel appear three hundred orders of magnitude brighter than normal bright white, it should just be clamped to 255 :-p

[11:39:01.0000] <Philip`>
Oops, I didn't mean 10^39 for FF2, I meant 10^310

[11:39:02.0000] <gsnedders>
So, both browsers and standards suck then.

[11:40:00.0000] <Philip`>
gsnedders: Monitors too - we wouldn't need this clamping if it wasn't for their physical limitations

[11:40:01.0000] <gsnedders>
Philip`: Then we'd reach the limits of what the human brain can tell apart

[11:42:00.0000] <Philip`>
gsnedders: We're a huge way away from those limits - look at bright white on a computer monitor, then look at the sun with one eye, then look at a class 4 laser with the other eye, and you'll see there's still a long way for monitors to grow

[11:42:01.0000] <gsnedders>
Philip`: Am I meant to look at all three at once, or just the latter two?

[11:43:00.0000] <Philip`>
gsnedders: I recommend doing them separately and in that sequence

[11:43:01.0000] <gsnedders>
Philip`: That's true, it may be a bit hard if you look at the laser first

[11:46:00.0000] <Dashiva>
Philip`: So basically, you're seeing int and double max value limitations?

[11:47:00.0000] <Philip`>
Dashiva: Yes, except for Opera being crazy and doing signed int minus 64

[12:00:00.0000] <Dashiva>
Geez... if we were following the standard set in the early days of htmlwg, Hixie would've filed half a dozen formal complaints by now :)

[12:15:00.0000] <Philip`>
So, Opera's gone from 81% down to 52% on my ImageData tests since I last checked

[12:15:01.0000] <hsivonen>
/me wonders if KompoZer's alt UI is on mpt's radar for Ubuntu usability

[12:16:00.0000] <Dashiva>
Philip`: More tests, changed spec, or worse support?

[12:16:01.0000] <Philip`>
while WebKit has gone from 0% to 79%

[12:16:02.0000] <Philip`>
Dashiva: More tests and changed spec

[12:17:00.0000] <Philip`>
(though only a few of the extra tests are not the result of spec changes, so mostly it's just spec changes)

[12:19:00.0000] <Philip`>
I'm being nasty and e.g. testing that there isn't an ImageData constructor, in order to make Opera fail

[12:21:00.0000] <Dashiva>
Is that a requirement, though?

[12:23:00.0000] <Philip`>
There isn't an explicit statement in the spec saying "There must not be an ImageData constructor"

[12:25:00.0000] <Philip`>
I don't know if the Web IDL spec says that interfaces that aren't explicitly declared to have a constructor mustn't have a constructor

[12:25:01.0000] <Philip`>
but in any case it matters for interoperability, since it's not nice if people write new ImageData() in Opera and find their code breaks for Firefox users, so that's more important than what the specs explicitly say

[12:28:00.0000] <Dashiva>
Hoo boy

[12:48:00.0000] <hsivonen>
alt sure generates a lot of email

[12:51:00.0000] <Philip`>
hsivonen: It's used a hundred times more than abbr/acronym, so it ought to be worth a reasonable amount of discussion

[12:51:01.0000] <Philip`>
but I won't attempt to define what "a reasonable amount" is :-)

[12:58:00.0000] <Dashiva>
Well, it's not like much of it is about alt anymore

[12:59:00.0000] <Dashiva>
(and if it is, we've heard it before)


2008-05-05
[19:13:00.0000] <Hixie>
/me proposes something which would end up making alt="" required for every image

[19:14:00.0000] <Hixie>
i'll be interested to see whether the alt-must-be-required-whether-useful-or-not camp like this proposal or not

[19:17:00.0000] <Philip`>
Sounds kind of similar to the 'noalt' thing that people have suggested

[19:18:00.0000] <Philip`>
except with more graceful degradation in UAs that do ugly things with <img> without alt

[19:18:01.0000] <Philip`>
and sometimes slightly more informative to users than always announcing the image as "Image"

[19:21:00.0000] <Philip`>
I suppose the problem is that people will just never bother adding 'importantimage', because it achieves no visible effect (including in validators) - UAs don't know that images without importantimage are unimportant, so they'll still have to let their users interact with alt-less images in case they're important, so there would be no difference

[19:21:01.0000] <webben_>
Hixie: Hmm. Seems to me that if the use-case includes bulk-upload photo tools, then those tools can't make the statement that the image "cannot be considered equivalent to any text". What they can do is say the alt is not in any sense a full equivalent to the image.

[19:22:00.0000] <webben_>
that would be useful information to tools like Lynx

[19:23:00.0000] <Philip`>
They could say how confident they are in the quality of the alt text, like <img src="cats.jpg" alt="Some photo, probably of a cat; q=0.2">

[19:25:00.0000] <webben_>
I'm not sure that would add much to current use scenarios over just saying it's not a full equivalent.

[19:26:00.0000] <Philip`>
Indeed, it seems a hard enough job just to get people to use a boolean value correctly

[19:26:01.0000] <webben_>
if you've got some hypothetical future gizmo for recognizing images, you're likely to just run it over anything that isn't q=1

[19:26:02.0000] <Philip`>
You should run it on anything that is q=1 too, because people will use q=1 when they've got rubbish alt text

[19:27:00.0000] <webben_>
Philip`: If you've got another hypothetical future gizmo that tells you whether you're new equivalent is better than the original alt, maybe.

[19:27:01.0000] <webben_>
*your new

[19:30:00.0000] <Philip`>
webben_: I've got a concrete present-day human who can decide whether the original alt text is giving them the information they're expecting or hoping for, and if not then they can use the gizmo and decide whether that's any better

[19:31:00.0000] <webben_>
um, yes, although the information they're expecting or hoping for might be rather different to that actually intended to be provided.

[19:31:01.0000] <webben_>
but quering an analysis tool is different to running it automatically

[19:32:00.0000] <webben_>
(or at least to presenting the results automatically)

[19:32:01.0000] <Philip`>
I'm thinking of e.g. looking at a Flickr page and finding a thing called "Photo" or "IMG_05863.JPG" and then (regardless of how reliable the page says that alt text is) choosing to analyse it in more detail

[19:33:00.0000] <webben_>
yep, that would work.

[19:34:00.0000] <Philip`>
I'm not sure how it'd be possible to do the analysis automatically, since you'd want an extremely concise output (like one or two words) in most cases, else it'll significantly disrupt the flow of the page

[19:34:01.0000] <webben_>
depends how clever the gizmo is, it's all a bit speculative at this point

[19:34:02.0000] <Philip`>
At least it could ignore transparent spacer GIFs

[19:35:00.0000] <webben_>
yes. We really ought to be able to improve at to ignore space GIFs now.

[19:35:01.0000] <webben_>
*improve AT

[19:35:02.0000] <Philip`>
Aside: Maybe HTML5 should support <spacer>, since that's a better solution than spacer GIFs because it's easier for UAs to ignore it, but currently it's unfashionable because it makes validators complain

[19:35:03.0000] <webben_>
and Lynx for that matter.

[19:36:00.0000] <Philip`>
webben_: Do most ATs download images?

[19:36:01.0000] <webben_>
I've never run across an instance of a developer needing a spacer GIF, so I know no reason to support either except to deal with existing content.

[19:36:02.0000] <Philip`>
(I assume they would if they're just stuck on top of IE or Firefox or whatever)

[19:37:00.0000] <Philip`>
It'd be kind of a pain to make Lynx download every linked image file solely to decide whether it looks like a spacer or not

[19:37:01.0000] <webben_>
Philip`: Most AT isn't a browser, but acting with a browser. People using AT don't necessarily turn off image downloading, though some do.

[19:37:02.0000] <Hixie>
Philip`: css is a better solution

[19:38:00.0000] <Philip`>
Hixie: Oh, good point

[19:38:01.0000] <webben_>
Philip`: Depends how annoying seeing "SPACER GIF" splattered across the page is to lynx users, I guess

[19:39:00.0000] <webben_>
or IMAGE when it's not one.

[19:40:00.0000] <webben_>
You're probably right that text browsers' commitment to speed is going to make them even less capable than typical AT.

[19:42:00.0000] <Philip`>
/me wonders if someone like Google could download and analyse every image on the internet, then construct something like a Bloom filter so UAs can quickly check whether a given URI is in the pre-computed set of spacer images, and how large the filter data file would be

[19:44:00.0000] <webben_>
tricky, given images are obtained via URIs, and URIs can return things that vary by who requests them (as well as time)

[19:44:01.0000] <webben_>
you could md5 the downloaded image and use that for the query to a webservice, but that would still require Lynx to download the image

[19:45:00.0000] <Philip`>
Static content like images are more commonly at a single static URI, so it might not be too awful

[19:45:01.0000] <webben_>
Oh yes, it would work for common cases.

[19:46:00.0000] <webben_>
having to make a request to a webservice for each image might be almost as bad for performance as downloading the images in the first place

[19:47:00.0000] <Philip`>
Hmm, Google says there's about 10^9 images, and if I wildly guess there's 10^6 spacer images, and an acceptable false positive rate is 1/1000, then it'd only need 14MB of filter data

[19:48:00.0000] <Philip`>
webben_: The idea is that you just download the global data once, and do local queries on that, like how Firefox's phishing filter works by default

[19:48:01.0000] <webben_>
oh I see.

[19:48:02.0000] <Philip`>
Anyway, this a stupid idea so I'll drop it :-p

[19:49:00.0000] <Philip`>
Oops, that's 14Mbits, so it's not even 2MB

[20:02:00.0000] <Philip`>
/me likes writing code that writes code

[20:02:01.0000] <Philip`>
Sadly I haven't yet found an excuse to write code that writes code that writes code

[01:16:00.0000] <annevk>
Sunava doesn't actually set a clear target date :(

[01:16:01.0000] <annevk>
Oh well, I guess that's normal

[01:20:00.0000] <annevk>
Hixie, you around?

[01:21:00.0000] <Hixie>
can be

[01:21:01.0000] <annevk>
Hixie, it's not that important, but I was wondering about your proposal on the appformats list

[01:21:02.0000] <Hixie>
anything specific?

[01:21:03.0000] <annevk>
is  Access-Include-Credentials provided by the UA or the server?

[01:21:04.0000] <annevk>
and what are the header values?

[01:22:00.0000] <Hixie>
server

[01:22:01.0000] <annevk>
(your proposal says UA, but I expect server)

[01:22:02.0000] <Hixie>
no values (or just "yes" or something)

[01:23:00.0000] <annevk>
GET requests would always work?

[01:23:01.0000] <annevk>
or would GET requests need to be preceded by a normal request too?

[01:28:00.0000] <Hixie>
GET requests work like now but without credentials, and if the server replies with an Access-Include-Credentials they are redone with credentials

[01:28:01.0000] <annevk>
:(

[01:29:00.0000] <annevk>
can the user be prompted in the initial request?

[01:29:01.0000] <annevk>
in case of http auth

[01:37:00.0000] <Hixie>
the user can never be prompted can he?

[01:39:00.0000] <annevk>
guess so

[01:39:01.0000] <annevk>
<img src=cross-site> is prompted iirc

[01:40:00.0000] <Hixie>
i don't think we should ever prompt the user for credentials in any request that will be subject to Access-Control management

[01:42:00.0000] <annevk>
so you'd just get a 403 back (or 401?) and be done with it?

[01:42:01.0000] <annevk>
i guess that's ok

[01:44:00.0000] <annevk>
(that is of course, if the server actually provides Access-Control headers with the 401)

[01:47:00.0000] <Hixie>
makes sense to get back the 401

[01:47:01.0000] <Hixie>
that way the third party host knows to tell the user to log in

[01:48:00.0000] <annevk>
the way the spec is written now all status codes work cross-site

[01:48:01.0000] <annevk>
i hope browsers implement that too as I think it's useful

[01:50:00.0000] <Hixie>
include an example whowing it

[01:50:01.0000] <Hixie>
showing

[01:51:00.0000] <othermaciej>
what's the advantage to Access-Include-Credentials?

[01:51:01.0000] <Hixie>
the main advantage is taht it might get us out of this deadlock

[01:52:00.0000] <othermaciej>
I don't see how it adds any security

[01:52:01.0000] <othermaciej>
it wouldn't defend against anything that anyone has identified as a threat model

[01:52:02.0000] <annevk>
I wonder what the reason is sicking isn't commenting at all other than saying the spec isn't ready

[01:53:00.0000] <annevk>
The only threat model I've seen so far is that people might opt in too soon and others would copy and paste and only Firefox users would be vulnerable

[01:54:00.0000] <othermaciej>
I have not followed the appformats list though

[01:54:01.0000] <annevk>
But Access-Include-Credentials doesn't address that issue

[01:54:02.0000] <othermaciej>
I was thinking of the HP guy's "sensitive financial transaction" pseudo-threat

[01:54:03.0000] <othermaciej>
which is also not addressed

[01:54:04.0000] <Hixie>
what's going on is that mozilla has a "security" person internally who isn't making sense

[01:54:05.0000] <Hixie>
and that is deadlocking mozilla

[01:55:00.0000] <othermaciej>
is it Window or someone else?

[01:56:00.0000] <annevk>
Hmm, adding "bloat" because of a single person seems icky :(

[01:56:01.0000] <othermaciej>
I would rather have a face-to-face meeting than add cargo cult security

[01:59:00.0000] <annevk>
btw, have you been reviewing the spec othermaciej?

[02:00:00.0000] <othermaciej>
annevk: not much so far, but I am booking time for it next week, since we might be implementing it soon

[02:00:01.0000] <othermaciej>
I've asked one of my co-workers to look at it as well (he does most of our dom/networking security stuff)

[02:00:02.0000] <annevk>
double awesome :)

[03:18:00.0000] <Philip`>
Why does VoiceOver have such a stupid selection of voices?

[03:18:01.0000] <Philip`>
Particularly "Bells", which sounds somewhat like bells and nothing like speech

[03:19:00.0000] <jgraham_>
Philip`: That one sentence encompassed my entire VO experience

[03:19:01.0000] <Philip`>
Hmm, there's another one that sings everything

[03:20:00.0000] <jgraham_>
(Oh, VoiceOver I should try that. What a crappy voice. Oh, the others are all worse, and all the keyboard shortcuts are buried somewhere in a PDF and impossible to fingeranyway. Nevermind.)

[03:20:01.0000] <othermaciej>
Philip`: it offers all the voices that Mac OS X's regular text-to-speech function supports

[03:20:02.0000] <othermaciej>
and yes, some are silly and would likely be annoying or useless to regular screen reader users

[03:24:00.0000] <jruderman_>
i wish they were grouped somehow (e.g. male, female, silly)

[03:26:00.0000] <jruderman_>
i love the |say| command, btw

[03:27:00.0000] <jruderman_>
'sleep 90 && say time is up' is so much nicer than the timers that come with most board games

[03:28:00.0000] <othermaciej>
that is a cool command, yeah

[03:41:00.0000] <Hixie>
leopard hides the silly ones these days

[03:41:01.0000] <Hixie>
though they are still there

[03:42:00.0000] <Hixie>
and it grounds male and female

[03:42:01.0000] <Hixie>
groups, even

[03:47:00.0000] <Philip`>
Doesn't for me (in 10.5.2) - VoiceOver Utility just has drop-down lists of two dozen alphabetical voices

[03:53:00.0000] <Hixie>
s/leopard/the places in leopard that i've seen so far/ :-)

[04:15:00.0000] <Philip`>
Whoops, I think I strayed off the topic of a specific instance of alt text usage, and started talking about generalities :-(

[04:34:00.0000] <Hixie>
should context.font = '1em sans-serif'; use the font-size of the <canvas> element?

[04:35:00.0000] <Philip`>
I want "12px" to be precisely 12 canvas coordinate space units (because otherwise it's impossible to do even vaguely predictable layouts); apart from that, I don't have any views

[04:36:00.0000] <Hixie>
what else would a px be?

[04:39:00.0000] <Philip`>
It could be canvas coordinate space units multiplied by the user's text size preference, like in Firefox

[04:39:01.0000] <Hixie>
o_O

[04:39:02.0000] <Hixie>
that seems dumb

[04:39:03.0000] <Philip`>
(though maybe FF changed since it added full-page zoom)

[04:39:04.0000] <Philip`>
It makes sense for normal CSS pxs

[04:39:05.0000] <Philip`>
Wait, am I getting confused?

[04:40:00.0000] <Philip`>
Oh, I'm not

[04:40:01.0000] <Philip`>
data:text/html;charset=utf-8,%3Cstyle%3Espan%7Bfont-size%3A16px%7D%3C%2Fstyle%3E%0ATest%20%3Cspan%3ETest%3C%2Fspan%3E

[04:40:02.0000] <Philip`>
in FF2, changing text size changes all the text

[04:40:03.0000] <Philip`>
(which is nice because it makes tiny text readable)

[04:41:00.0000] <Hixie>
CSS doesn't have a concept of font-size zoom

[04:41:01.0000] <Philip`>
(but isn't nice in canvas, since you want it to line up sensibly with all the other drawn shapes)

[04:41:02.0000] <Philip`>
(and if the text in a canvas is too small to read, get a browser with full-page zoom so the whole canvas gets bigger)

[04:44:00.0000] <Philip`>
/me departs

[04:46:00.0000] <Lachy>
In OSX Leopard, the only decent voice I've found is Alex with the pitch lowered to about 40. That's about as close to a normal sounding, Aussie accent I can get from it.

[04:46:01.0000] <Lachy>
the women's voices in it all sound silly

[04:47:00.0000] <Hixie>
Alex is the only voice that was created in this decade, as far as i can tell

[04:47:01.0000] <Hixie>
there are far better ones available from third parties

[04:49:00.0000] <annevk>
makes sense that em uses the <canvas> font size, just like currentColor uses the <canvas> 'color'


2008-05-06
[01:35:00.0000] <annevk2>
Hixie, prose says the default is '10px sans-serif' and IDL says '8px sans-serif'

[01:36:00.0000] <annevk>
ah, logging is back

[01:36:01.0000] <Hixie>
oops

[01:36:02.0000] <Hixie>
logging?

[01:38:00.0000] <annevk>
krijnh joined a fifteen minutes ago

[01:38:01.0000] <Hixie>
aah

[01:39:00.0000] <annevk>
hmm, I wonder if there's an inverse relation between the number of e-mails left and the number of XXX comments in the spec :)

[01:40:00.0000] <Hixie>
quite possible

[01:40:01.0000] <Hixie>
what the hell did little jacob eat that cost a $120 at a burger joint!

[01:40:02.0000] <Hixie>
jeezus

[01:41:00.0000] <annevk>
GTA IV reference?

[01:41:01.0000] <annevk>
/me wants to play that game

[01:41:02.0000] <Hixie>
yes

[01:42:00.0000] <annevk>
he's probably not so little :p

[01:42:01.0000] <jacobolus>
Hixie: I always get the most expensive thing on the menu :)

[01:43:00.0000] <Hixie>
it's a burger joint! the most expensive thing is like $3!

[01:43:01.0000] <jacobolus>
you've never had a caviar burger?

[01:44:00.0000] <Hixie>
i do not believe that would be on the menu of that particular restaurant

[01:46:00.0000] <roc>
you've never eaten in New York City?

[01:47:00.0000] <Hixie>
i have, but $120 seems still pretty damn high

[01:49:00.0000] <BenMillard>
hixie, maybe he just ordered loads of stuff?

[01:49:01.0000] <BenMillard>
btw, I'm at Sight City in Frankfurt this week thanks to Mozilla

[01:50:00.0000] <BenMillard>
assistive technology convention type thing

[01:50:01.0000] <BenMillard>
http://www.sightcity.net/sightcity08-e.html

[01:50:02.0000] <Hixie>
cool

[01:53:00.0000] <annevk>
BenMillard, please invite all of them to join the HTMLWG/WHATWG and especially to e-mail their thoughts :)

[01:53:01.0000] <annevk>
(on html5, in case that wasn't clear...)

[01:54:00.0000] <Hixie>
christ, more people worried about copyrights on acid tests

[01:54:01.0000] <Hixie>
people are insanely paranoid about copyright

[01:54:02.0000] <BenMillard>
annevk, that's a great idea

[01:55:00.0000] <BenMillard>
annevk, I'll also be asking them about their interest in better support HTML's existing accessibility features, fixing bugs in their HTML support, stuff like that

[01:55:01.0000] <BenMillard>
*better supporting

[01:55:02.0000] <BenMillard>
hope to blog it all when I get back in similar detail as the Boston F2F

[01:56:00.0000] <annevk>
cool

[01:57:00.0000] <BenMillard>
I'll be walking around with a notepad and pen, pouncing on AT vendors like a journalist :P

[01:58:00.0000] <BenMillard>
I get back on Friday evening; off for some lunch now. bye all

[02:10:00.0000] <hsivonen>
annevk: Molly had to cancel

[02:20:00.0000] <MikeSmith>
annevk: you there?

[02:20:01.0000] <annevk>
am now

[02:20:02.0000] <annevk>
hsivonen, ok

[02:20:03.0000] <annevk>
too bad

[02:21:00.0000] <MikeSmith>
annevk: wondering where Arve might be

[02:21:01.0000] <annevk>
sidewalk?

[02:21:02.0000] <annevk>
dunno really

[02:25:00.0000] <annevk>
hsivonen, yt?

[02:25:01.0000] <annevk>
/me is going down to get lunch

[02:26:00.0000] <hsivonen>
annevk: I haven't had breakfast yet

[02:26:01.0000] <hsivonen>
It will be a while before I even get to breakfast

[02:34:00.0000] <hsivonen>
hendry: see above. Molly's event is canceled

[02:44:00.0000] <jgraham__>
/me denies having any opinion on canvas text in general and text-on-a-path in particular

[02:54:00.0000] <jgraham__>
Oh wait a second, you were talking about measuring text. Yeah, I did have a (sensible!) opinion on that :)

[03:48:00.0000] <Lachy>
Hixie, did you intentionally omit 'justify' from the allowed values of canvas context .textAlign?

[03:49:00.0000] <Hixie>
yes

[03:49:01.0000] <Hixie>
<string> also

[03:50:00.0000] <Lachy>
isn't it possible to support it at all?

[03:51:00.0000] <Lachy>
or is it just too complex for this first version?

[03:51:01.0000] <othermaciej>
full justification looks shitty on the web usually

[03:52:00.0000] <othermaciej>
basically no site uses it

[03:52:01.0000] <annevk>
justify should be possible, would even be nice if maxWidth was taken as desiredWidth or something

[03:52:02.0000] <othermaciej>
CSS3 Text module has a super complex spec for fancy full justification

[03:52:03.0000] <Lachy>
the default wordpress template uses justify

[03:52:04.0000] <othermaciej>
which IE might even implement some of

[03:53:00.0000] <Lachy>
so you can't say basically no site uses it.

[03:53:01.0000] <Philip`>
You could imagine that 'left' means 'left or justify', since you can only draw a single line of text so it makes no difference

[03:53:02.0000] <othermaciej>
of the 4 random wordpress.com blogs I looked at, none seems to have full justified text

[03:54:00.0000] <annevk>
offtopic: http://shop.lego.com/ByTheme/Product.aspx?p=10179&cn=416&d=322

[03:54:01.0000] <othermaciej>
you could full-justify a single line to a specific width, but that is likely a bad idea

[03:54:02.0000] <Hixie>
full justification would be meaningless since there's only one line, it would be stretch-to-fit

[03:54:03.0000] <othermaciej>
annevk: multiple people at my work have one of those

[03:54:04.0000] <Hixie>
or space-to-fit

[03:54:05.0000] <Hixie>
and i'm not adding that feature just because text-align happens to have it

[03:55:00.0000] <othermaciej>
full justify only ever looks good if you have an algorithm that is smart about adding justification vs. breaking lines, and is willing to either expand or compress both letter-spacing and word-spacing

[03:55:01.0000] <Lachy>
ok, that makes sense

[03:56:00.0000] <Philip`>
Why can't web browsers render everything using LaTeX?

[03:56:01.0000] <othermaciej>
WebKit's algorithm is not very smart (it breaks lines as normal and adds word-spacing)

[03:56:02.0000] <othermaciej>
Philip`: good justification is expensive to do on the fly, since you have to optimize the whole paragraph and can't really work line-at-a-time

[03:56:03.0000] <othermaciej>
(which I believe is what TeX does)

[03:57:00.0000] <Philip`>
Computers are fast - I would have though they could optimise a ten-line paragraph without a huge amount of trouble

[03:57:01.0000] <Philip`>
*thought

[03:58:00.0000] <Hixie>
one paragraph, sire

[03:58:01.0000] <Hixie>
sure

[03:59:00.0000] <Hixie>
html5 is about 5000 paragraphs

[03:59:01.0000] <othermaciej>
full justify is really most useful for very narrow columns as in newspapers

[03:59:02.0000] <othermaciej>
though I guess it looks good in books too

[04:00:00.0000] <annevk>
i think that's why CSS3 covers it in much detail

[04:00:01.0000] <othermaciej>
in the web context it can create "rivers" through the text when done badly

[04:00:02.0000] <annevk>
probably less so for CSS applied on a screen context

[04:01:00.0000] <Philip`>
I assume good justification also requires automatic hyphenation

[04:02:00.0000] <annevk>
that's another topic discussed quite a bit in the CSS WG...

[04:07:00.0000] <Philip`>
/me would be happier if more basic aspects of text rendering worked, like not having some fonts appear with no antialiasing on Linux, or like not having Safari sometimes render '=' with the top line twice the thickness of the bottom line, before worrying about complex typographical layout details

[04:08:00.0000] <hsivonen>
hmm. *I'm* hung up on machine validation of alt??? 

[04:10:00.0000] <annevk>
maybe you should ask which universe version of you they're referring too

[04:17:00.0000] <Philip`>
/me wonders why Hixie's proposal that makes alt mandatory has had so few responses

[04:17:01.0000] <Hixie>
/me bites back some scathing cynicism

[04:18:00.0000] <othermaciej>
did anyone read it?

[04:19:00.0000] <othermaciej>
probably only people who dislike mandatory alt actually read that far

[04:19:01.0000] <Philip`>
I don't dislike mandatory alt but I read it :-)

[04:19:02.0000] <Lachy>
Philip`, when did Hixie propose to make alt mandatory? Did I miss something?

[04:20:00.0000] <Philip`>
http://lists.w3.org/Archives/Public/public-html/2008May/0073.html

[04:22:00.0000] <Philip`>
(I suppose that wasn't actually saying alt should be mandatory in all cases, but I'm not sure what other cases remain)

[04:24:00.0000] <othermaciej>
Hixie's proposal is clearly meant to be a Modest one

[04:24:01.0000] <othermaciej>
but I don't hate it that much other than the sheer length of importantimage="importantimage"

[04:24:02.0000] <Lachy>
I really don't like the name importantimage for the attribute

[04:25:00.0000] <Hixie>
i believe i noted the attribute name was dumb

[04:25:01.0000] <Hixie>
and asked for better ideas

[04:26:00.0000] <Lachy>
how about introducing <img src="..." kind="Photo"> (where the value is a short, descriptive label of somekind)

[04:26:01.0000] <othermaciej>
the thing is, none of the people who would be expected to herald it as a brilliant breakthrough have spoken up

[04:27:00.0000] <Hixie>
shocking

[04:27:01.0000] <Hixie>
in other news

[04:27:02.0000] <Hixie>
i need to make innerHTML convert undefined and null to strings before processing

[04:27:03.0000] <Hixie>
should i just assume that that is what happens?

[04:27:04.0000] <Hixie>
i don't really follow that part of WebIDL

[04:27:05.0000] <Lachy>
that way you're not altering the meaning of alt based on the presence of another attribtue, and it also gives more potentially useful information in conjuncion with alt

[04:28:00.0000] <othermaciej>
Hixie: heycam's explanation of that in email further confused me

[04:29:00.0000] <Hixie>
as i understand it we have the following cases:

[04:29:01.0000] <othermaciej>
Hixie: it sounds like in the current version you have to specify [NoNull] to get that behavior for null but I am not really sure

[04:29:02.0000] <Hixie>
(for setting DOMString attributes and for passing DOMString arguments)

[04:29:03.0000] <Hixie>
1. null -> "", undefined -> ""

[04:29:04.0000] <Hixie>
2. null -> "", undefined -> "undefined"

[04:29:05.0000] <Lachy>
e.g. it distinguishes these two images, even though they have the same alt text:

[04:29:06.0000] <Lachy>
<img src="house.png" alt="The house is white, with green shutters on the windows." kind="Painting">

[04:29:07.0000] <Lachy>
<img src="house.png" alt="The house is white, with green shutters on the windows." kind="Photograph">

[04:29:08.0000] <Hixie>
3. null -> "null", undefined -> "undefined"

[04:30:00.0000] <Hixie>
4. null -> null, undefined -> null

[04:31:00.0000] <annevk>
case 2 actually exists? :(

[04:31:01.0000] <othermaciej>
I don't think Web IDL has anything to specify handling for null

[04:31:02.0000] <othermaciej>
I am also not offhand aware of APIs that do 1 or 4

[04:31:03.0000] <annevk>
Web IDL has [NoNull]

[04:31:04.0000] <Hixie>
annevk: firefox with innerHTML does that

[04:31:05.0000] <othermaciej>
er

[04:31:06.0000] <othermaciej>
specify handling for undefined I mean

[04:31:07.0000] <Hixie>
[NoNull] is a terrible name as it doesn't convey 1-4 of the above

[04:32:00.0000] <othermaciej>
I think 2 is what you get by default, and 3 is what you get with NoNull

[04:32:01.0000] <annevk>
Hixie, I don't think Firefox is the best reference

[04:32:02.0000] <annevk>
I believe WebKit, Opera, and IE are more aligned last I checked

[04:32:03.0000] <Hixie>
i'm not saying it is

[04:33:00.0000] <othermaciej>
in WebKit's IDL extensions, 3 is what you get by default and 2 (or really null -> null, undefined -> "undefined") is what you get with [ConvertNullToNullString]

[04:33:01.0000] <Lachy>
selectors api uses case 1 for return values from nsresolver

[04:33:02.0000] <othermaciej>
I think in DOM APIs, 2 (or really my variant of 2, but most things don't distinguish null string and empty string) is most common

[04:34:00.0000] <othermaciej>
but 3 is the most native-to-JavaScript behavior

[04:35:00.0000] <othermaciej>
Safari does 2 for assignment to innerHTML afaict

[04:36:00.0000] <othermaciej>
I can't find any place where we treat undefined passed for a string argument as empty string instead of "undefined"

[04:36:01.0000] <Hixie>
can you find any cases where you treat undefined as null?

[04:37:00.0000] <othermaciej>
actually we have a function for it so maybe

[04:38:00.0000] <othermaciej>
lemme see

[04:38:01.0000] <Lachy>
othermaciej, it appears that webkit does for document.querySelector(undefined); It throws a syntax err, the same as if "" was passed

[04:38:02.0000] <othermaciej>
createDocumentType, querySelector and toDataURL

[04:39:00.0000] <othermaciej>
createDocumentType is probably the only one of those that can't be changes for compat reasons

[04:39:01.0000] <othermaciej>
sucks to have three different ways of treating DOMString arguments

[04:39:02.0000] <annevk>
createDocumentType is used?!

[04:39:03.0000] <othermaciej>
and that's not even accounting for the weirdshit on attribute setting or return values

[04:40:00.0000] <othermaciej>
annevk: shockingly, yes

[04:40:01.0000] <Hixie>
getElementsByClassName() seems to treat undefined as "" in webkit trunk

[04:42:00.0000] <othermaciej>
I will not vouch for all our string treatment being the best choice compat-wise, mind you, not all of the cases have been compared against other implementations

[04:42:01.0000] <othermaciej>
Document.execCommand also treats undefined as null/empty for the third argument

[04:42:02.0000] <othermaciej>
(presumably to DTRT when it is omitted)

[04:43:00.0000] <othermaciej>
also for the URL and feature arguments of window.open

[04:59:00.0000] <hendry>
/me is in Dublin

[05:17:00.0000] <Philip`>
Lachy: <img src="..." kind="Photo"> would be much worse than <img src="..." alt="Photo" some-metametadata-attribute> for graceful degradation, since existing UAs won't tell the user it's a photo and will do something ugly like read the filename or ignore the image

[05:30:00.0000] <Lachy>
Philip`, so as far as back compat is concerned, it's no worse than what the current spec says

[05:31:00.0000] <Lachy>
but it's still may be an improvement for future UAs

[05:33:00.0000] <Philip`>
Lachy: Would it be more of an improvement for future UAs than <img importantimage>?

[05:34:00.0000] <Hixie>
woot, i caught up with the backlog i got from playing gta4

[05:35:00.0000] <Philip`>
Does that mean you're going back to GTA4 again? :-)

[05:35:01.0000] <Lachy>
I think importantimage will suffer from many of the same problems as noalt=""

[05:35:02.0000] <gsnedders>
Hixie: are you sure you're old enough to play GTA4?

[05:35:03.0000] <gsnedders>
:D

[05:38:00.0000] <Lachy>
oh, I like Hixie's suggested improvements for WebIDL. I hope heycam accepts them, and then I only need to decide on the best behavior to choose for selectors api

[06:34:00.0000] <heycam`>
Hixie / othermaciej_ , thanks for the comments.  i'll get to making some changes to web idl late next week, when my current busy patch is over.

[06:53:00.0000] <Lachy>
JohnResig, othermaciej, I added a very basic testsuite for selectors api. http://dev.w3.org/cvsweb/2006/webapi/selectors-api-testsuite/

[06:54:00.0000] <Lachy>
it doesn't test much yet, I'm still in the process of working out the best way to write the framework.

[06:54:01.0000] <Lachy>
but take a look, see if what I have is a good start and let me know if you have any suggestions for improvements

[06:56:00.0000] <Lachy>
the plan is to run most tests within an iframe, which loads various HTML and XHTML files, and runs queries on them

[06:56:01.0000] <Lachy>
and each test is a function within an array, written in tests.js

[06:58:00.0000] <Philip`>
Would you have a single iframe, and run each test in it in turn?

[06:59:00.0000] <Philip`>
Oh, actually, I think I'm just misunderstanding the whole thing, so ignore me

[06:59:01.0000] <Lachy>
Philip`, yes, that's most likely

[06:59:02.0000] <Philip`>
Are you always testing DOM-created documents, rather than downloaded-over-HTTP documents?

[07:00:00.0000] <Lachy>
no, the plan is to allow each test to either create a document or load a file into the iframe

[07:01:00.0000] <Philip`>
Ah, right

[07:01:01.0000] <Lachy>
not sure exactly how to do that, since such operations would need to be asynchronous

[07:01:02.0000] <othermaciej>
Lachy: I will ask weinig or somebody to tae a look

[07:01:03.0000] <Lachy>
othermaciej, thanks

[07:01:04.0000] <othermaciej>
Lachy: we have our own regression tests for querySelector but I doubt they are very powerful

[07:01:05.0000] <Philip`>
Lachy: Run test 1, set onload to a function that checks the results and then runs test 2? Not sure what the problem is :-)

[07:02:00.0000] <othermaciej>
we specifically didn't want to try to make tests for every possible kind of selector including edge cases, though come to think of it that would be more convenient than the tests with red and green boxes usually used for this kind of thing

[07:02:01.0000] <Lachy>
Philip`, yeah, something like that could work

[07:03:00.0000] <othermaciej>
Lachy: we have a good test framework for tests that purely involve scripting APIs and not layout

[07:03:01.0000] <Lachy>
othermaciej, ok. That might be good if I could build on that, instead of creating my own

[07:04:00.0000] <Philip`>
Mozilla has a framework for that too

[07:04:01.0000] <Philip`>
(http://developer.mozilla.org/en/docs/Mochitest )

[07:04:02.0000] <othermaciej>
basically you just write a series of assertions in a JS file like this:

[07:04:03.0000] <othermaciej>
shouldBe("Math.abs(1)", "1");

[07:04:04.0000] <othermaciej>
shouldBe("Math.abs(-1)", "1");

[07:04:05.0000] <Lachy>
Opera sort of does too, but it's not really something that's usable by anyone easily

[07:04:06.0000] <othermaciej>
and we have some JS files that implement the functions

[07:05:00.0000] <othermaciej>
ours is real easy to use

[07:05:01.0000] <othermaciej>
and gives pretty results

[07:06:00.0000] <othermaciej>
Lachy: I'll mail you an example (unless you have a WebKit checkout handy in which case I could point you to it)

[07:06:01.0000] <Lachy>
othermaciej, email it

[07:06:02.0000] <Philip`>
othermaciej: Is it easy to run individual tests by themselves? (Mozilla's Mochitests run a whole web server and a special browser mode, which is annoyingly complex)

[07:07:00.0000] <othermaciej>
Philip`: we group tests in files, all it takes to "run" such a suite is to open the html file in any browser

[07:07:01.0000] <othermaciej>
for fully automated execution we have a headless tool that loads all the html files in sequence

[07:07:02.0000] <Philip`>
othermaciej: Okay, sounds good :-)

[07:13:00.0000] <othermaciej>
Lachy: sent

[07:13:01.0000] <othermaciej>
Lachy: the three files starting with js- are the framework

[07:14:00.0000] <othermaciej>
elements-not-in-document.js is the substance of the test case

[07:14:01.0000] <othermaciej>
I might also have a script lying around to generate the wrapper HTML files for the script files (they just include the test script plus the CSS file and two helpers)

[07:15:00.0000] <othermaciej>
but it's pretty trivial

[07:16:00.0000] <othermaciej>
makes good looking output though where it is easy to spot the failures and why they fail

[07:17:00.0000] <othermaciej>
in addition to the basic shouldBe tests there's also shouldBeTrue, shouldBeFalse, shouldBeNaN, shouldBeNull, shouldBeUndefined and shouldThrow

[07:27:00.0000] <Lachy>
othermaciej, thanks

[07:33:00.0000] <gsnedders>
BBC News today reporting 4 in 5 15 y/os drink. That sounds far too lower amount.

[07:37:00.0000] <Lachy>
gsnedders, that sounds about right. I was one of the 1 in 5 that didn't drink at that age.

[07:37:01.0000] <gsnedders>
Lachy: Here it is certainly higher than 4 in 5. Even 9 in 10 seems low to me.

[07:38:00.0000] <gsnedders>
(I am that 1, no matter how many that is in :P)

[07:38:01.0000] <Philip`>
Does it mean 4 in 5 drink, or 4 in 5 say they drink when asked on a survey?

[07:39:00.0000] <gsnedders>
Philip`: The claim is the former due to the latter.

[07:39:01.0000] <gsnedders>
If they just admitted it was the latter there would me less questioning by me :P

[07:41:00.0000] <Lachy>
gsnedders, do you believe that ~1 out of 5 lied by saying they didn't, when they actually do?

[07:41:01.0000] <gsnedders>
Lachy: Yes.

[07:41:02.0000] <Lachy>
what if some said they did, but actually don't

[07:42:00.0000] <gsnedders>
I doubt that would happen, though.

[07:42:01.0000] <gsnedders>
I know nobody who claims that

[07:42:02.0000] <Lachy>
maybe the only thing the survey can reliably tell us is that 15 year old kids can't be trusted

[07:42:03.0000] <gsnedders>
I know plenty of people who claim they don't drink, though

[07:43:00.0000] <gsnedders>
(But having seen them drink I can tell you it isn't true)

[07:43:01.0000] <Lachy>
gsnedders, what's your reason for not drinking?

[07:44:00.0000] <gsnedders>
Lachy: Well, I get allergic reactions drinking the slightest amount

[07:44:01.0000] <Lachy>
or were you just one of the liars? :-)

[07:44:02.0000] <Lachy>
oh, that sucks

[07:44:03.0000] <Philip`>
gsnedders: The group of people you know is quite possibly not representative of the entire country, so you can't use your anecdotes to argue against people who used Science and got written about on a website and must be true

[07:44:04.0000] <Lachy>
allergic to just beer, or all alcohol?

[07:44:05.0000] <gsnedders>
(where slightest amount means even a sip, not half a bottle as some of my friends would take it to be true)

[07:44:06.0000] <gsnedders>
Lachy: Beer, wine, whiskey. Never tried anything else.

[07:45:00.0000] <gsnedders>
Philip`: Oh, the parties at the beach include most of the top two years of my school :P

[07:45:01.0000] <Philip`>
(That is, the general concept of Science, not the international weekly journal)

[07:45:02.0000] <Lachy>
try Vodka Cruisers, Biccardi Breezers, or things like that

[07:45:03.0000] <gsnedders>
(that's around 470 people)

[07:45:04.0000] <Philip`>
You could try water instead

[07:45:05.0000] <Lachy>
... but don't do it till you're 18!

[07:45:06.0000] <Lachy>
(or whatever the legal drinking age is in your country)

[07:45:07.0000] <gsnedders>
Lachy: I fear what they'll do to me with their strength if I am allergic

[07:46:00.0000] <gsnedders>
Lachy: (5, with parental consent in private)

[07:46:01.0000] <gsnedders>
Lachy: (14 with food in restaurants)

[07:46:02.0000] <Lachy>
really? 14 is legal?

[07:46:03.0000] <gsnedders>
There's something at 16, too

[07:47:00.0000] <Lachy>
wow. In Australia, it's 18 in any public venue. Not sure if there's an explicit age limit for drinking in private

[07:47:01.0000] <gsnedders>
Hm, it's changed

[07:47:02.0000] <Lachy>
which country are you in?

[07:47:03.0000] <gsnedders>
It's no longer 14, but 16.

[07:47:04.0000] <gsnedders>
Scotland

[07:47:05.0000] <gsnedders>
We need alcohol for our northern European drinking culture (i.e., getting drunk every weekend)

[07:47:06.0000] <gsnedders>
(if not every day)

[07:48:00.0000] <Lachy>
ah, right. I guess, over there, it's rare for anyone to be sober :-)

[07:49:00.0000] <gsnedders>
I really can't think of many people who don't drink.

[07:49:01.0000] <gsnedders>
Maybe five or so out of one hundred, so one in twenty don't.

[07:50:00.0000] <gsnedders>
/me finds a quote from IRC in IM logs

[07:50:01.0000] <gsnedders>
"The editor RECOMMENDS eating pasta while implementing this specification. Lachlan Hunt RECOMMENDS eating hamburgers while implementing this specification."

[07:50:02.0000] <Lachy>
was that from the discussion about beetroot on burgers?

[07:51:00.0000] <gsnedders>
No, RSS5 saying, "Authors MUST NOT use this language. The editor RECOMMENDS usage of The Atom Syndication Format. [RFC4287]", I think

[07:51:01.0000] <gsnedders>
20080115

[07:51:02.0000] <gsnedders>
http://krijnhoetmer.nl/irc-logs/whatwg/20080115#l-633

[07:52:00.0000] <Lachy>
yeah, that's the one.  See elsewhere in the logs where I listed the essential ignredients

[07:52:01.0000] <gsnedders>
There are so many lovely things in IM logs between me and molz

[07:53:00.0000] <gsnedders>
E.g., (I know she won't mind me quoting this): "02:59:09 Molly Holzschlag: you are NOT awake"

[07:53:01.0000] <gsnedders>
How to start a conversation :)

[07:57:00.0000] <gsnedders>
Oh, and now me shouting, "I AM PERFECTLY NORMAL"

[08:00:00.0000] <gsnedders>
Hah!

[08:00:01.0000] <gsnedders>
It's all so lovely.

[08:00:02.0000] <gsnedders>
Who is cuter, me or Hixie?

[08:03:00.0000] <Lachy>
gsnedders, why would you ask such a question?

[08:03:01.0000] <Philip`>
Sounds like a very dangerous question to answer

[08:04:00.0000] <Lachy>
my answer would be neither, since, AFAIK, neither of them are girls and thus not cute at all

[08:05:00.0000] <Philip`>
Are small fluffy male kittens unable to be cute?

[08:05:01.0000] <Lachy>
only if they have a cute caption. Otherwise, I don't like cats

[08:06:00.0000] <gsnedders>
Lachy: Discussion myself and Molly got into for no apparent reason

[08:06:01.0000] <Lachy>
oh, and there's aways exceptions for children

[08:06:02.0000] <gsnedders>
Lachy: Actually, it more a monologue by Molly

[08:06:03.0000] <Philip`>
What about uncaptioned hamsters?

[08:07:00.0000] <Philip`>
They're always cute as long as you keep your fingers away from their mouths

[08:07:01.0000] <Lachy>
Philip`, I don't know. I haven't seen a LOLHamster

[08:07:02.0000] <gsnedders>
Lachy: According to Molly, Hixie has been wrecked by power, so me

[08:09:00.0000] <Lachy>
http://xkcd.com/262/ :-D

[08:10:00.0000] <gsnedders>
Oooo… /me is getting some XKCD t-shirts!

[08:11:00.0000] <annevk>
their alt sucks btw

[08:11:01.0000] <Lachy>
which ones?

[08:11:02.0000] <annevk>
seems they just copied the title

[08:13:00.0000] <Philip`>
http://www.ohnorobot.com/index.pl?comic=56&s=gluing+captions

[08:13:01.0000] <Philip`>
Transcriptions do exist, so presumably it'd be nice to use them as alt text

[10:03:00.0000] <hasather>
Is http://www.opera.com/dragonfly/ done counting for everyone, or is there some TZ issue?

[10:04:00.0000] <hasather>
no, should be live: http://twitter.com/opera/statuses/804779359

[10:04:01.0000] <hasather>
works now

[10:05:00.0000] <Philip`>
hasather: It sends me to http://www.opera.com/products/dragonfly/ which has a download link

[10:05:01.0000] <hasather>
Philip`: yes, didn't work one minute ago for me

[10:08:00.0000] <Philip`>
It's kind of odd that it's "Built using the open web standards you know and love" but starts with <viewport xmlns="http://www.w3.org/1999/xhtml"> which seems not entirely standard

[10:11:00.0000] <MikeSmith>
/me is looking at Dragonfly page now

[10:11:01.0000] <MikeSmith>
wasn't there supposed to be a feature related to remote debugging?

[10:12:00.0000] <MikeSmith>
that is, HTTP debugging proxy like Fiddler?

[10:12:01.0000] <Philip`>
MikeSmith: "Debug pages whether they are on your computer or a supported device, such as a Windows Mobile phone running Opera Mobile 9.5."

[10:12:02.0000] <Lachy>
oh, when was dragon fly publicly announced?

[10:12:03.0000] <MikeSmith>
ah

[10:12:04.0000] <Philip`>
... "You can connect to any device running Presto Core-2.1 or above, and debug using your computer’s screen and keyboard – No need to struggle inputting test data with your phone’s keypad again."

[10:13:00.0000] <Philip`>
/me assumes that won't work for Opera Mini

[10:47:00.0000] <MikeSmith>
hendry: you there?

[11:17:00.0000] <zcorpan_>
hsivonen: dunno if you already know about this but here are some test cases that might be interesting to you: http://www.validome.org/lang/en/errors/ALL

[11:23:00.0000] <zcorpan_>
hsivonen: for xml documents, perhaps you should compare the root element with the mime type and complain about image/svg+xml containing xhtml?

[11:36:00.0000] <zcorpan_>
hsivonen: the message "Quirky doctype." is a bit unhelpful i think -- it's unclear what's wrong and what to do about it

[11:36:01.0000] <zcorpan_>
hsivonen: consider <!DOCTYPE php PUBLIC "-// ...

[11:50:00.0000] <zcorpan_>
hmm? validome seems to think that entities in id='' is not allowed

[11:56:00.0000] <gsnedders>
so, how far are people on GTA4?

[11:57:00.0000] <zcorpan_>
hsivonen: you don't seem to allow whitespace in the legacy encoding declaration (between text/html; and encoding=...)

[12:06:00.0000] <takkaria>
zcorpan_: I thought that the spec mandates that

[12:27:00.0000] <zcorpan_>
takkaria: mandates what?

[12:31:00.0000] <takkaria>
zcorpan_: nm, I checked the spec anyway and I was wrong

[12:41:00.0000] <gsnedders>
annevk: you asked me to expand on mid:E5851D3C-1B2C-49F0-9A23-61A56DD4A4FC⊙gc — do you want me to bring the thread up again now?

[12:56:00.0000] <gsnedders>
Philip`, jgraham__: any thoughts about where to go for (this!) Friday?

[13:23:00.0000] <jgraham__>
gsnedders: Lets decide a place to meet then improvise

[13:23:01.0000] <jgraham__>
/me votes for meeting somewhere closer to the IoA than the Millpond :)

[13:24:00.0000] <gsnedders>
That's further for me though!

[13:24:01.0000] <gsnedders>
And the closer we get, the less well I know where I am (which isn't good when I haven't really been to Cam. much in years)

[13:26:00.0000] <jgraham__>
/me has to go eat now

[13:26:01.0000] <gsnedders>
But you don't know where you're eating on Friday!

[13:35:00.0000] <Philip`>
gsnedders: What kind of time is it likely to be?

[13:35:01.0000] <gsnedders>
Philip`: What suits you two?

[13:35:02.0000] <gsnedders>
I arrive at around 4pm, IIRC

[13:35:03.0000] <Philip`>
Anything that's not too late and not too early would be just right

[13:35:04.0000] <gsnedders>
And I'm not doing anything else all day, apart from travelling

[13:36:00.0000] <Philip`>
I'll be at the Computer Lab during the day, but could leave whenever I wanted to

[13:36:01.0000] <gsnedders>
15:08 I arrive at the station

[13:36:02.0000] <gsnedders>
Or at least, I'm _meant_ to arrive at the station :)

[13:37:00.0000] <gsnedders>
6pm or so at the earliest would be best, IMO

[13:39:00.0000] <gsnedders>
I'm open to suggestions as to where to meet anywhere where I know in Cambridge (which is mostly just the centre)

[13:44:00.0000] <Philip`>
/me notes that he is useless at suggesting anything

[13:45:00.0000] <Philip`>
I don't know much other than the straight line between King's and Sainsbury's

[13:45:01.0000] <gsnedders>
Heh.

[13:46:00.0000] <gsnedders>
Philip`: And you've been in Cam. for how long?

[13:46:01.0000] <Philip`>
Only 3.5 years

[13:46:02.0000] <gsnedders>
You need to get out more! :P

[13:46:03.0000] <Philip`>
I'm happy enough without doing so :-)

[13:47:00.0000] <gsnedders>
/me notes his latest tweet: "Too many vertices, not enough edges, in the graph of my mind."

[13:48:00.0000] <Philip`>
The more important measurement is how many faces you have

[13:49:00.0000] <gsnedders>
No, you've got it wrong :P

[14:08:00.0000] <jgraham__>
We could meet outside Kings

[14:08:01.0000] <jgraham__>
That seems like a happy medium

[14:26:00.0000] <Philip`>
jgraham__: Sounds reasonable to me

[15:47:00.0000] <Hixie>
good lord

[15:50:00.0000] <Philip`>
?

[15:50:01.0000] <Hixie>
i am seriously fed up with people asking me about copyright stuff

[15:50:02.0000] <Hixie>
who

[15:50:03.0000] <Hixie>
the hell

[15:50:04.0000] <Hixie>
cares

[15:51:00.0000] <Hixie>
i hate the RIAA for making everyone so damn paranoid

[15:52:00.0000] <Philip`>
Just put a CC licence on your work, and then people will happily share it and fail to follow the licence's requirements

[15:53:00.0000] <jruderman>
lol

[15:53:01.0000] <Hixie>
in this particular case it was about the copyright on screenshots of the acid3 test

[15:53:02.0000] <Lachy>
ok, that's insane.

[15:53:03.0000] <Hixie>
which i wouldn't own the copyright on even if it COULD be copyrighted which as far as i can tell it can't!

[15:53:04.0000] <Lachy>
doesn't everyone know acid 3 is in the public domain?

[15:53:05.0000] <Hixie>
it's a page for testing interop! as far as i know it can't be copyrighted! anything you do with it is fair use for interop!

[15:54:00.0000] <Hixie>
grrr

[15:54:01.0000] <Philip`>
Publishing screenshots of Acid 3 on Wikipedia isn't for interoperability

[15:54:02.0000] <Lachy>
the W3C would disagree with that, given their recent discussions about the css test suite licence

[15:55:00.0000] <Hixie>
the w3c smokes too much crack

[15:55:01.0000] <Lachy>
LOL

[15:55:02.0000] <Hixie>
and is also pissing me off

[15:56:00.0000] <Hixie>
(regarding the testsuite stuff, and various other things)

[15:56:01.0000] <gsnedders>
No, the people in GTA4 smoke too much crack :)

[15:56:02.0000] <gsnedders>
The W3C ain't got nothing on them.

[15:57:00.0000] <Lachy>
/me wonders if there's crack dealers in the game to make it more realistic?

[15:57:01.0000] <gsnedders>
Yeah. Some get murdered, too. Also realism.

[15:58:00.0000] <Lachy>
cool

[15:58:01.0000] <gsnedders>
/me has actually almost finished the story line

[15:58:02.0000] <Philip`>
gsnedders: There's a slight difference in that the W3C exists in real life, rather than in a satirical video game

[15:58:03.0000] <gsnedders>
A single week for a GTA game? That's a bit too quick.

[15:58:04.0000] <jgraham__>
What's the deal with the CSS testsuite license?

[15:58:05.0000] <gsnedders>
Philip`: Sssh!

[15:59:00.0000] <Philip`>
(unless the W3C exists in the game too)

[15:59:01.0000] <gsnedders>
I dunno. The Web does.

[15:59:02.0000] <Philip`>
(which wouldn't surprise me)

[15:59:03.0000] <gsnedders>
Heck, Web 2.0 exists in the game!

[15:59:04.0000] <Lachy>
when I was younger, I had games on the super nintendo that took me several weeks, if not months to finish. What the hell is wrong with these game companies these days, making such short games?

[15:59:05.0000] <gsnedders>
(and Radio 2.0 too)

[15:59:06.0000] <gsnedders>
Lachy: +1

[15:59:07.0000] <Philip`>
Lachy: The problem is with people spending too much time playing the game, so it's over too quickly :-)

[15:59:08.0000] <Hixie>
there are some pretty hilarious radio spots in gta4 talking about web 2.0

[16:00:00.0000] <gsnedders>
Lachy: I completed the entire NFS: Carbon the weekend before last (as last weekend it was GTA4)

[16:00:01.0000] <gsnedders>
Hixie: Totally.

[16:00:02.0000] <gsnedders>
Hixie: Doesn't one mention [Something] 3.14?

[16:00:03.0000] <jgraham__>
/me totally sucks at computer games

[16:00:04.0000] <gsnedders>
/me totally sucks at computer games with the exception of racing games, in which he rules

[16:00:05.0000] <Philip`>
Oblivion isn't short

[16:01:00.0000] <Philip`>
I think I've spent something like 200 hours on that

[16:01:01.0000] <gsnedders>
Philip`: Yeah, but there are so few long games now :P

[16:01:02.0000] <Lachy>
gsnedders, yep, Web Pi is coming soon

[16:01:03.0000] <gsnedders>
/me must be up to around 40 hours on GTA4 since last Tues

[16:01:04.0000] <gsnedders>
This is ideal timing. Just before my exams :P

[16:01:05.0000] <gsnedders>
Lachy: It's Web Pi? Ah.

[16:02:00.0000] <Lachy>
gsnedders, didn't you get that's what 3.14 refers to?

[16:02:01.0000] <gsnedders>
Lachy: I was busy trying to kill someone in another car when I heard the 3.14 shortly after the mention of Web 2.0

[16:02:02.0000] <Lachy>
I have a shirt that says "Wake me for Web 3.14" on it

[16:03:00.0000] <Lachy>
there's a photo of it on flickr

[16:03:01.0000] <gsnedders>
Lachy: Quite low precision π, then

[16:04:00.0000] <Lachy>
no, mathematicians realised that memorising an infinite number of decimal places was too hard, so they truncated it.

[16:04:01.0000] <gsnedders>
Lachy: memorising an infinite number of decimal places is hard? I never realised.

[16:04:02.0000] <jgraham__>
Maybe we could do HTML2π — circular edition after HTML5

[16:04:03.0000] <gsnedders>
jgraham__: Not HTML2πr?

[16:05:00.0000] <gsnedders>
Then we know the size of our circle.

[16:05:01.0000] <jgraham__>
r=1; it's a unit circle (the best kind)

[16:05:02.0000] <Lachy>
2π ~= 6.38, so that's a good version number to follow 5

[16:05:03.0000] <Philip`>
/me sees 98 hours 37 minutes 55 seconds spent on Advance Wars: DS, which is probably quite high up his list of games-ordered-by-time-wasted-on-them

[16:06:00.0000] <gsnedders>
/me wonders which .+5 spec will be the first to 2π

[16:06:01.0000] <jgraham__>
Lachy: More like 6.28 :)

[16:06:02.0000] <gsnedders>
Gran Turismo is probably the highest for me

[16:06:03.0000] <jgraham__>
(allthough I let you off because of the ~)

[16:06:04.0000] <Lachy>
jgraham__, maybe in your universe :-)

[16:06:05.0000] <Philip`>
/me thinks more games need to provide play timers

[16:06:06.0000] <gsnedders>
/me doesn't want to know

[16:07:00.0000] <Philip`>
Ultima Online is probably the highest for me, since I played that for around a year, which is kind of scary

[16:08:00.0000] <gsnedders>
I played NFS:HP2 for ages, but I doubt that is above GT3/4


2008-05-07
[17:46:00.0000] <annevk>
gsnedders, I've no idea, do you have a URI?

[17:52:00.0000] <Hixie>
jgraham__: yt?

[17:56:00.0000] <Hixie>
Lachy: yt?

[17:56:01.0000] <Lachy>
Hixie, yo

[17:56:02.0000] <Hixie>
in 2005 you wanted some sort of namespacing system for profile=""

[17:56:03.0000] <Hixie>
do you still want me to reply to your feedback on that?

[17:56:04.0000] <Lachy>
I did?

[17:57:00.0000] <Hixie>
quoting:

[17:57:01.0000] <Hixie>
> If a solution can be found which allows for namespaces, but which does    

[17:57:02.0000] <Hixie>
> not require them to be used in most cases, which doesn't introduce even   

[17:57:03.0000] <Hixie>
> more problems, then I think that would probably be the best solution.

[17:57:04.0000] <Lachy>
ah, probably not. profile="" has proven useless. But if it were useful, I'd probably still agree with what I wrote

[17:58:00.0000] <Hixie>
k

[17:58:01.0000] <Lachy>
on second though, I don't like namespaces, so I change my mind

[17:58:02.0000] <Hixie>
/me likes not replying to e-mails where the person has changed their mind, since it avoids the problem of other people jumping on the bandwagon

[17:59:00.0000] <Hixie>
:-)

[18:00:00.0000] <Lachy>
Hixie, do you have a simple command line script that submits a file to the csswg's spec post processor that I could use on my mac? My .bat file I use on windows won't work :-(

[18:01:00.0000] <Hixie>
i just use wget

[18:01:01.0000] <annevk>
curl -u USER:PASS -F file=@Overview.src.html -F group=none -F output=html -F method=file http://cgi.w3.org/member-bin/process.cgi -o Overview.html

[18:02:00.0000] <Lachy>
yeah, I just used curl, but I don't want to have to type all those parameters each time

[18:02:01.0000] <annevk>
my command line remembers it for me...

[18:02:02.0000] <Hixie>
so put them in a shell script

[18:02:03.0000] <Lachy>
how do I do that?

[18:02:04.0000] <Hixie>
i have an .update.sh file that contains the line above

[18:02:05.0000] <Lachy>
/me googles it

[18:02:06.0000] <Hixie>
(or rather, it's equivalent for wget)

[18:02:07.0000] <Hixie>
its

[18:02:08.0000] <Hixie>
its

[18:02:09.0000] <Hixie>
/me takes the apostrophe out back and shoots it

[18:03:00.0000] <Hixie>
shell scripts are just like batch files

[18:03:01.0000] <annevk>
you're showing typical GTA syndroms :p

[18:03:02.0000] <Hixie>
but they work on your shell (e.g. bash) instead of on the dos shell

[18:06:00.0000] <Lachy>
I have written shell scripts a few years ago, but that was in first year of uni 6 years ago and I haven't used linux much since then

[18:06:01.0000] <Lachy>
I've just forgotten all this stuff

[18:06:02.0000] <Hixie>
yeah well i never learnt it

[18:06:03.0000] <Hixie>
i make it up as i go along :-D

[18:06:04.0000] <Lachy>
woah, 7 years ago

[18:06:05.0000] <Lachy>
crap, I'm getting old now...

[18:06:06.0000] <Hixie>
and fall back on perl whenever i want to do something useful

[18:08:00.0000] <Lachy>
ok, is there anything special I need to do to make sure my .update.sh file never inadvertently gets commited to CVS, or is it enough that it's a hidden file?

[18:08:01.0000] <Hixie>
don't cvs add it

[18:08:02.0000] <Hixie>
:-)

[18:08:03.0000] <Lachy>
ok

[18:26:00.0000] <Philip`>
Why ".update.sh" and not just "update.sh"?

[18:27:00.0000] <Hixie>
so that you can't download it on my site

[18:27:01.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/.update.sh

[18:27:02.0000] <Hixie>
i block anything starting with a dot

[18:27:03.0000] <Hixie>
also so it doesn't appear in the middle of directory listings

[18:27:04.0000] <Philip`>
Aha

[18:29:00.0000] <Philip`>
Lachy: You can also create a .cvsignore file in a directory (as in http://docs.freebsd.org/info/cvs/cvs.info.cvsignore.html ) to make cvs automatically ignore particular files

[18:31:00.0000] <Hixie>
/me drives the final nail into profile=""'s coffin

[18:34:00.0000] <Lachy>
/me waits for people to try and pull that nail out

[18:44:00.0000] <Lachy>
ok, looks like that nails in pretty tight. Which section is next?

[18:44:01.0000] <Hixie>
any requests?

[18:45:00.0000] <Lachy>
hmm, let me check the issues list...

[18:46:00.0000] <Facedown>
Hixie - were you the vim user?

[18:46:01.0000] <Hixie>
no

[18:46:02.0000] <Facedown>
oh

[18:46:03.0000] <Hixie>
emacs all the way

[18:46:04.0000] <Facedown>
hehe

[18:46:05.0000] <Hixie>
though i hear someone implemented vi in javascript

[18:46:06.0000] <Facedown>
indeed

[18:46:07.0000] <Facedown>
its slow but its functional

[18:46:08.0000] <Facedown>
pretty amazing

[18:46:09.0000] <Facedown>
just like the mario game

[18:46:10.0000] <Hixie>
heh

[18:47:00.0000] <Facedown>
will i get killed if i ask a ie6/layout question in here?

[18:47:01.0000] <Facedown>
heh

[18:47:02.0000] <Hixie>
no

[18:47:03.0000] <Facedown>
s/killed/stoned

[18:47:04.0000] <Hixie>
though people might not be able to help :-)

[18:47:05.0000] <Facedown>
I'm trying to absolutely position the navigation element on top of #around-content which is floated and has width of 100%. The reason for the float and width is because I'm using that as a wrapper for a negative margin layout. IE6 seems to hide the nav (probably underneath) the floated el if it has a width of 99.6% or higher -> http://pastie.caboo.se/190178

[18:48:00.0000] <Hixie>
i recommend trying another browser

[18:48:01.0000] <Facedown>
heh

[18:48:02.0000] <Hixie>
told you we might not be able to help :-P

[18:50:00.0000] <htmlfivedotnet>
facedown: I think you'll need to set the height

[18:51:00.0000] <Lachy>
I don't have IE6 available to test it, so I can't help

[18:51:01.0000] <htmlfivedotnet>
/me boots virtualbox

[18:51:02.0000] <htmlfivedotnet>
/me recompiles virtual box for 2.6.24-17

[18:51:03.0000] <Lachy>
Hixie, how about finishing off the semantics-phrasing-* folders?

[18:51:04.0000] <Facedown>
htmlfivedotnet - why height? 

[18:51:05.0000] <Facedown>
hmm

[18:52:00.0000] <htmlfivedotnet>
because you're using IE6 lol

[18:52:01.0000] <Facedown>
height triggers layout.. i dont see how it could make it show up

[18:52:02.0000] <Facedown>
position:absolute already triggers layout by default though

[18:53:00.0000] <Hixie>
Lachy: any in particular?

[18:53:01.0000] <htmlfivedotnet>
well try it out, if it doesn't work, you can flame me lol

[18:53:02.0000] <Facedown>
will do

[18:53:03.0000] <htmlfivedotnet>
i'm booting up my winxp to try a few things

[18:53:04.0000] <htmlfivedotnet>
but i have to recompile, because i upgraded my kernel

[18:54:00.0000] <Philip`>
htmlfivedotnet: You wouldn't get all these kernel problems if you just used Windows

[18:54:01.0000] <Lachy>
Hixie, didn't you already do the -abbr ones recently?

[18:54:02.0000] <Lachy>
I thought you removed folders and their messages once they were dealt with

[18:55:00.0000] <htmlfivedotnet>
Philip`: True

[18:57:00.0000] <Hixie>
Lachy: i do

[18:57:01.0000] <Hixie>
Lachy: if there is still a folder, it means there was more feedback

[18:57:02.0000] <Hixie>
there usually is new feedback immediately after i deal with a section

[18:59:00.0000] <Lachy>
you could clear out the datatemplates folder now, since there's only 2 in there and the sectino was dropped.

[19:00:00.0000] <Hixie>
i want to wait until i drop repetition blocks

[19:00:01.0000] <Lachy>
ok

[19:00:02.0000] <Hixie>
otherwise people think one is in and the other is not

[19:01:00.0000] <Lachy>
what about accesskey? Any idea what you're going to do about that yet?

[19:02:00.0000] <Hixie>
i don't know what to do about accesskey

[19:02:01.0000] <Lachy>
it would be nice if the issues list displayed the dates, so I could more easily tell which were new discussions and which are old

[19:02:02.0000] <Hixie>
heh

[19:02:03.0000] <Hixie>
it shouldn't matter :-)

[19:03:00.0000] <Philip`>
Lachy: You could write a script that correlates the message-IDs against the list archives to work out exactly when each message was posted

[19:03:01.0000] <Lachy>
Philip`, I'm too lazy for that

[19:04:00.0000] <Philip`>
You could just add it to http://canvex.lazyilluminati.com/misc/cgi/issues.cgi/source if you wanted :-)

[19:05:00.0000] <Lachy>
I would have to learn perl first

[19:05:01.0000] <Philip`>
It's just like Python but with some symbols added

[19:05:02.0000] <Lachy>
and since the syntax of perl is incomprehensible, that doesn't seem likely any time soon

[19:05:03.0000] <Philip`>
It's not at all incomprehensible :-(

[19:07:00.0000] <htmlfivedotnet>
english is incomprehensible to those who have never tried to learn it. 

[19:07:01.0000] <htmlfivedotnet>
but then again, even those who have tried might still say the same

[19:07:02.0000] <Lachy>
woah, "open my $f, $0 or die" is a bit demanding isn't it? Couldn't you say please instead of threatening?

[19:09:00.0000] <Hixie>
that's better written as   my $f; open($f, '<', $0) or die;

[19:09:01.0000] <Hixie>
but frankly that's pretty obscure perl

[19:09:02.0000] <Hixie>
whoever wrote that clearly doesn't have good perl readability guidelines

[19:09:03.0000] <Hixie>
:-)

[19:09:04.0000] <Philip`>
I disagree with your judgement of "better" :-)

[19:10:00.0000] <htmlfivedotnet>
Philip`: I get a higher "windows experience index" score when running Vista in a virtual machine, then i do when i install it directly. obviously, not in the video, but everything else is about 25% higher

[19:10:01.0000] <Philip`>
There's no point in wasting semicolons and brackets

[19:10:02.0000] <Philip`>
htmlfivedotnet: Maybe the virtual machine is running the virtual clock 25% slower than real-time? :-)

[19:11:00.0000] <htmlfivedotnet>
Philip`: Damn. You got me.

[19:12:00.0000] <Hixie>
Philip`: does opening $0 open the __DATA__ stream or does it open the script itself?

[19:12:01.0000] <Lachy>
htmlfivedotnet, that's probably because the virtual machine software your using is a slow piece of crap. Are you using the Microsoft VirtualMachine?

[19:12:02.0000] <Philip`>
Hixie: $0 is just the script's filename, so it reads it as a normal file and doesn't do anything clever

[19:12:03.0000] <Lachy>
Virtual PC, I mean

[19:13:00.0000] <htmlfivedotnet>
 Lachy: Yeah. I'm running it in Wine lol

[19:13:01.0000] <Philip`>
(Not a particularly robust way of reading the script, since the working directory might be elsewhere, but it works for this CGI thing)

[19:14:00.0000] <Hixie>
Philip`: k

[19:14:01.0000] <Hixie>
so er, why are you reading your own script? :-)

[19:14:02.0000] <Hixie>
that's pretty universally not good practice in _any_ language :-P

[19:14:03.0000] <Lachy>
ah, ok. You should try VMWare, if it's available for Linux. I use VMWare Fusion on Mac, and it's great.

[19:14:04.0000] <Philip`>
Hixie: So that http://canvex.lazyilluminati.com/misc/cgi/issues.cgi/source works

[19:14:05.0000] <Hixie>
oh, i see

[19:14:06.0000] <Hixie>
yeah that seems fair enough

[19:15:00.0000] <othermaciej>
does anyone here know of cool looking web apps done with standards-based technology?

[19:15:01.0000] <Lachy>
htmlfivedotnet, or you could try a virtual machine program written natively for linux

[19:15:02.0000] <othermaciej>
(preferably ones that are actually useful, not just demos)

[19:16:00.0000] <htmlfivedotnet>
Lachy: I was kidding. I use virtualbox ose. mainly because i can apt-get install it

[19:16:01.0000] <Philip`>
I used to use VMware on Linux, but then I upgraded my kernel and couldn't get it to run without crashing my entire machine, so I gave up

[19:17:00.0000] <Lachy>
othermaciej, timwouldlickit.com is funny and useful :-) (though probably not qutie what you're looking for)

[19:18:00.0000] <htmlfivedotnet>
Philip`: it's just a matter of recompiling the module, and remembering where the heck the source files are to do so

[19:18:01.0000] <htmlfivedotnet>
othermaciej: there was one that would generate a random "web 2.0" startup company name, but i can't remember the name of it

[19:19:00.0000] <Lachy>
htmlfivedotnet, there are several of those

[19:19:01.0000] <Hixie>
othermaciej: without being able to use 'window' or XMLHttpRequest, it's hard to make a cool looking web app

[19:20:00.0000] <Hixie>
Lachy: ok, got rid of the phrasing folder

[19:20:01.0000] <Hixie>
not sure what to do about ruby

[19:20:02.0000] <Hixie>
(didn't get rid of hte phrasing subfolders)

[19:21:00.0000] <othermaciej>
Hixie: I would not consider either of those things an obstacle to consider something to be done with "standards-based technology"

[19:21:01.0000] <Hixie>
oh

[19:21:02.0000] <Hixie>
well then

[19:21:03.0000] <othermaciej>
I maily meant to say "not Flash"

[19:21:04.0000] <Hixie>
GMail!

[19:21:05.0000] <othermaciej>
as opposed to "only RECs"

[19:21:06.0000] <Hixie>
Google Maps!

[19:21:07.0000] <Hixie>
Google Calendar!

[19:21:08.0000] <Hixie>
GMail has flash for the IM part, but that's all

[19:22:00.0000] <Hixie>
Google Reader

[19:22:01.0000] <Philip`>
Microsoft Windows Live Local

[19:22:02.0000] <Hixie>
Google Docs (probably has flash for some of the comms stuff)

[19:22:03.0000] <Hixie>
in fact, http://www.google.com/options/

[19:22:04.0000] <Hixie>
:-)

[19:22:05.0000] <Lachy>
oh, I saw a list of the top 100 web apps a few weeks ago. I wonder if I can find that

[19:22:06.0000] <othermaciej>
GMail is useful but not all that cool looking

[19:23:00.0000] <Philip`>
Microsoft Office Live Workspace

[19:23:01.0000] <Philip`>
(Beta)

[19:27:00.0000] <htmlfivedotnet>
facedown: still there?

[19:28:00.0000] <Lachy>
Hixie, how come the phrasing folder is still showing up in the issues list then?

[19:28:01.0000] <Hixie>
the issues list is updated nightly

[19:28:02.0000] <Lachy>
ok

[19:28:03.0000] <Lachy>
I thought it was live

[19:28:04.0000] <Hixie>
no, that would kill my imap server

[19:28:05.0000] <htmlfivedotnet>
facedown: adding a "clear" to the nav element makes it show up. who knows why. it doesn't make any sense, but it worked

[19:29:00.0000] <Lachy>
wasn't it live originally? Did you take away that feature

[19:29:01.0000] <Philip`>
Hixie: Isn't IMAP designed to have clients maintain connections and get immediate notifications of changes?

[19:30:00.0000] <Hixie>
Philip`: probably

[19:30:01.0000] <Hixie>
Lachy: the voting is live

[19:30:02.0000] <Hixie>
Lachy: the list has always been static, though

[19:30:03.0000] <Hixie>
Philip`: but i have enough trouble with just my own imap clients connecting to this server...

[19:31:00.0000] <Philip`>
Hixie: Oh, okay

[19:31:01.0000] <Lachy>
ok, it's good that the voting is live. It's nice to be able to keep up with all the action as it happens there.

[19:31:02.0000] <Hixie>
hah

[19:32:00.0000] <Philip`>
How many issues have ever had multiple votes?

[19:34:00.0000] <Lachy>
there's 2 votes for mail in the microsyntaxes-numbers folder, and that puts it in the lead

[19:35:00.0000] <Hixie>
people seem to think that repeating their arguments is more effective than voting

[19:35:01.0000] <Hixie>
they're wrong

[19:35:02.0000] <Hixie>
it's certainly not more effective

[19:35:03.0000] <Hixie>
and may even be less effective, though the jury is still out on that one

[19:35:04.0000] <Philip`>
If they only post their argument once, they can only vote for it once

[19:36:00.0000] <Hixie>
voting for an argument twice _definitely_ has no more effect than voting for it once

[19:36:01.0000] <Philip`>
By posting multiple times they can vote on each instance, and they will combine into a single powerful super-vote

[19:36:02.0000] <Lachy>
Philip`, the votes don't add credibility to the argument, they only supposedly help with prioritisation

[19:37:00.0000] <Philip`>
Hanging around on IRC seems a more effective way to influence the spec's direction than voting

[19:37:01.0000] <Hixie>
if anything got real votes, it would indeed help prioritisation

[19:37:02.0000] <Lachy>
But, I question that since I voted for wf3-placeholder a long time ago, and it hasn't been touched

[19:37:03.0000] <Hixie>
webforms in general is blocked on anne and maciej

[19:37:04.0000] <Lachy>
yeah, I know

[19:37:05.0000] <othermaciej>
oh sure, blame me

[19:37:06.0000] <Philip`>
Maybe people should just vote on categories, not on individual messages

[19:37:07.0000] <Hixie>
othermaciej: :-P

[19:38:00.0000] <Lachy>
only 2 months till the forms-tf is finished

[19:38:01.0000] <Philip`>
since it's quite unlikely that several people will make the same choice out of three thousand possibilities

[19:38:02.0000] <Lachy>
then we'll get the results of all the amazing work they've been doing

[19:40:00.0000] <Philip`>
Have they decided yet whether HTML needs to support forms or not?

[19:41:00.0000] <Philip`>
I'm not sure native support for forms is worthwhile, since you can just emulate them with contentEditable and CSS and script

[19:41:01.0000] <Philip`>
(and use ARIA so it's accessible)

[19:42:00.0000] <Hixie>
the point of wf2 is to reduce the dependency on scripts

[19:42:01.0000] <Lachy>
what are the use cases that people are trying to solve with forms? Couldn't they be done more easily with an envelope and postage stamp?

[19:42:02.0000] <htmlfivedotnet>
Lachy: ... ? ... I agree.

[19:42:03.0000] <Lachy>
/me must be getting tired; his jokes are getting bad.

[19:43:00.0000] <Hixie>
i disagree with the premise of your argument

[19:43:01.0000] <Hixie>
:-P

[19:44:00.0000] <Philip`>
Hixie: Why is it pointful to reduce the dependency on scripts?

[19:44:01.0000] <Hixie>
it's easier to maintain scriptless markup

[19:44:02.0000] <Hixie>
generally speaking

[19:44:03.0000] <Hixie>
(except when the scriptless markup reaches the point of being script in another disguise, of course, like xpath and xforms actions)

[19:44:04.0000] <Philip`>
I could make a spec whose point is to reduces web apps' dependency on the letter W

[19:45:00.0000] <Philip`>
So the point of WF2 isn't to reduce the dependency on scripts, it's to make markup easier to maintain? :-)

[19:45:01.0000] <htmlfivedotnet>
and write

[19:45:02.0000] <Hixie>
that's just a general language design principle

[19:46:00.0000] <Philip`>
s/reduces/reduce/

[19:46:01.0000] <Lachy>
Philip`, introducing new forms features to cover things that are commonly solved with scripts, such as datetime controls replacing calendar widgets, means that different authors won't have to spend time reimplementing them

[19:47:00.0000] <Philip`>
(I suppose the different is that there isn't an underlying useful point behind the point of reducing the W dependency)

[19:47:01.0000] <Philip`>
s/different/difference/

[19:47:02.0000] <Philip`>
/me thinks he needs to go to sleep

[19:47:03.0000] <Lachy>
me too. I have to get up in 5 hours.

[19:47:04.0000] <Lachy>
good night

[19:48:00.0000] <htmlfivedotnet>
Philip`: I really don't think you're making any sense. I was trying to figure out what you were trying to figure out, but i figured out that you probably don't even know what you're trying to figure out

[19:49:00.0000] <Philip`>
htmlfivedotnet: The figuring-out process isn't useful if I know in advance what I'm going to have figured out by the end :-)

[19:49:01.0000] <Philip`>
s/htmlfivedotnet:/dcostalis:/

[19:49:02.0000] <dcostalis>
lol

[19:50:00.0000] <dcostalis>
Philip`: is it my client, or am i still talking as htmlfivedotnet?

[19:51:00.0000] <Philip`>
dcostalis: My client says you're talking as dcostalis

[19:51:01.0000] <Philip`>
http://krijnhoetmer.nl/irc-logs/whatwg/20080507#l-341 agrees

[19:51:02.0000] <Hixie>
/me dons asbestos underwear

[19:52:00.0000] <Hixie>
ok lets see how many people i pissed off with my latest e-mail

[19:52:01.0000] <Philip`>
/me re-aims his flames for a headshot

[19:52:02.0000] <Hixie>
i didn't tell you where i donned it

[19:53:00.0000] <Philip`>
That is a disturbing image

[19:53:01.0000] <Hixie>
hah

[21:25:00.0000] <Hixie>
woot, crossed the 2500 barrier

[21:49:00.0000] <Facedown>
2500?

[21:56:00.0000] <Hixie>
http://www.whatwg.org/issues/data.html

[22:02:00.0000] <othermaciej>
/me decides to try using the voting interface

[22:03:00.0000] <othermaciej>
I just voted to encourage Hixie to consider his own feedback

[22:03:01.0000] <othermaciej>
for some reason that feels a little surreal

[22:04:00.0000] <Hixie>
which one?

[22:04:01.0000] <othermaciej>
the icon one

[22:04:02.0000] <Hixie>
oh heh

[22:31:00.0000] <Hixie>
sqlite's handling of types is funny

[22:31:01.0000] <Hixie>
the documentation first says taht types are evil

[22:33:00.0000] <Hixie>
and then goes on to list many ways in which it supports types

[22:33:01.0000] <Hixie>
"In SQLite version 3, the type of a value is associated with the value itself, not with the column or variable in which the value is stored. (This is sometimes called manifest typing.) All other SQL databases engines that we are aware of use the more restrictive system of static typing where the type is associated with the container, not the value."

[22:33:02.0000] <Hixie>
yes that's also known as BEING COMPLIANT WITH THE SPEC

[22:34:00.0000] <Hixie>
that has to be the funniest way of saying "we're the only non-compliant implementation we know of" that i've ever seen

[22:36:00.0000] <roc___>
on trunk, both with and without my GM_ADVANCED patch, I get garbage for the line that uses the Courier bitmapped font. Is this something you have a fix for?

[22:37:00.0000] <jruderman>
weird. i thought static typing would be a huge help for perf in databases.

[22:37:01.0000] <roc___>
oo

[22:37:02.0000] <roc___>
ps

[01:05:00.0000] <zcorpan_>
hsivonen: http://www.sitepoint.com/forums/showthread.php?t=547022

[01:18:00.0000] <hsivonen>
zcorpan_: thanks for the Validome link

[01:19:00.0000] <hsivonen>
zcorpan_: Validator.nu already check root namespace and mime concordance, but for some reason I may have forgotten SVG from the list. thank

[01:22:00.0000] <zcorpan_>
hsivonen: quite obviously, people don't understand what schema actually means, and so saying "using the schema for XHTML 1.0" when validating HTML 4.01 is confusing

[01:24:00.0000] <zcorpan_>
hsivonen: perhaps the preset chooser should say "HTML 4.01 Strict/XHTML 1.0 Strict" and the info should say which ever is appropriate when choosing automatically?

[01:25:00.0000] <zcorpan_>
(or the info can say "XHTML 1.0/HTML 4.01" too i guess)

[01:28:00.0000] <hsivonen>
zcorpan_: seems reasonable even if technically not exactly correct

[01:37:00.0000] <Hixie>
i don't think you need to prioritise technical accuracy in the ui, so long as you report th technically correct errors

[01:41:00.0000] <Hixie>
huh, WHATWG is also an acronym for "where have all the women gone"

[01:41:01.0000] <Hixie>
somewhat appropriate

[02:29:00.0000] <annevk>
/me reads http://www.w3.org/mid/Pine.LNX.4.62.0805070153400.15395⊙hdc

[02:29:01.0000] <annevk>
/me expects flame wars

[02:30:00.0000] <Hixie>
nothing except karl's e-mail so far

[02:30:01.0000] <Philip`>
That's what happens when you post flamebait :-p

[02:34:00.0000] <annevk>
'Summary: profile="" doesn't work in practice so we have dropped it.' contradicts the clean-slate mantra

[02:35:00.0000] <Hixie>
no. html5 used to have it

[02:35:01.0000] <Hixie>
i then dropped it

[02:35:02.0000] <Hixie>
karl of course totally misunderstood what i meant by "doesn't work", as usual

[02:35:03.0000] <annevk>
ooh, it was in HTML5? interesting

[02:36:00.0000] <Hixie>
long ago

[02:36:01.0000] <annevk>
oh well, I don't care either way

[02:36:02.0000] <Hixie>
defined in a much more detailed way than html5, too

[02:37:00.0000] <Hixie>
html4 even

[02:40:00.0000] <zcorpan>
http://www.whatwg.org/specs/web-apps/2006-01-01/#profile

[02:41:00.0000] <Hixie>
yup

[02:47:00.0000] <Philip`>
Hixie: "doesn't work" is pretty vague - there are people who use the feature and it does what they want it to do and they are happy, so it does work for them

[02:48:00.0000] <Hixie>
it doesn't work for the reasons given in that e-mail -- namely that people don't use it

[02:50:00.0000] <Lachy_>
I wonder what Karl bases his claim on that "[profile] doesn't lead to any interoperability issues"? Would that be because user agents do absolutely nothing with it?  Since, if they did something with it, it's more than likely they would have bugs.

[02:51:00.0000] <jgraham__>
It's pretty crazy that in 2008 I'm wasting my time reading email where people are trying to apply Latin grammar rules to English

[02:51:01.0000] <Philip`>
There's a non-zero amount of content that uses it, and a non-zero amount of tools that process it, and they interoperate when used together

[02:52:00.0000] <Philip`>
(I don't know if "non-zero" is larger than one, though)

[02:54:00.0000] <Philip`>
...so (some) people do use it

[02:57:00.0000] <Hixie>
i don't believe anyone ever said it leads to interop issues

[03:15:00.0000] <zcorpan>
Hixie: http://forums.whatwg.org/ gives 500

[03:16:00.0000] <Hixie>
really?

[03:16:01.0000] <Hixie>
the machine is basically idle

[03:16:02.0000] <zcorpan>
really

[03:19:00.0000] <Hixie>
configuration error

[03:19:01.0000] <Hixie>
?

[03:20:00.0000] <Hixie>
cos seriously, the machine isn't doing anything right now

[03:20:01.0000] <annevk>
slightly offtopic but funny: http://diveintomark.org/archives/2008/05/07/when-the-fall-is-all-thats-left

[03:24:00.0000] <zcorpan>
Hixie: well, i can't tell, i just know that i get a 500 :)

[03:25:00.0000] <Hixie>
odd

[03:26:00.0000] <annevk>
loads for me

[03:26:01.0000] <Hixie>
wfm

[03:26:02.0000] <annevk>
empty your cache?

[03:27:00.0000] <annevk>
Hixie, should "cache" in http://wiki.whatwg.org/wiki/MetaExtensions be unendorsed?

[03:28:00.0000] <Hixie>
uh yeah

[03:28:01.0000] <Hixie>
my bad

[03:32:00.0000] <annevk>
though i think other wiki pages uses rejected instead

[03:32:01.0000] <annevk>
oh well, fixed

[03:33:00.0000] <zcorpan>
ah, loads fine now

[03:33:01.0000] <zcorpan>
(though i got 500 in different browsers so i think it was unrelated to the cache)

[03:35:00.0000] <annevk>
proxy cache?

[03:57:00.0000] <MikeSmith>
watching Crockford present "Javascript: The Good Parts"

[04:00:00.0000] <MikeSmith>
"the DOM is one of the worst APIs ever invented... Javascript is unfairly blamed for the awfulness of the DOM"

[04:05:00.0000] <Hixie>
well he's right about that

[04:05:01.0000] <annevk>
he says relying on semicolon insertion is bad :(

[04:05:02.0000] <hsivonen>
I use semicolon insertion all the time in JS

[04:06:00.0000] <annevk>
/me too

[04:06:01.0000] <Hixie>
the comments on mark's msn music drm shutdown post are funny

[04:06:02.0000] <annevk>
/me also uses the with operator for <canvas>

[04:06:03.0000] <Hixie>
"you say this now, but what if google ever shuts down a drm service! what would say then!"

[04:06:04.0000] <Hixie>
uh, hello, when we shut down google video's drm service we refunded everyone who'd ever bought anything. twice.

[04:07:00.0000] <Hixie>
let's see microsoft do that.

[04:07:01.0000] <hsivonen>
why not if (value) { ?

[04:10:00.0000] <annevk>
he also discourages blockless statements...

[04:10:01.0000] <tomg>
the YUI videos are an interesting mixture of very interesting and incredibly tedious 

[04:10:02.0000] <annevk>
what happened with the lazy web?

[04:10:03.0000] <hsivonen>
he's right on that one

[04:13:00.0000] <hsivonen>
the V.nu tokenizer is intentionally full of fallthroughs

[04:13:01.0000] <annevk>
(for people reading the logs, some of the remarks here are only relevant if you attend the talk MikeSmith mentioned earlier)

[04:36:00.0000] <annevk>
wow, he does overdramatize it quite a bit

[04:52:00.0000] <hsivonen>
I hadn't noticed Mini 2 didn't have lists

[06:06:00.0000] <zcorpan>
Hixie: it's unclear to me whether strokeText() actually strokes or not

[06:06:01.0000] <zcorpan>
Hixie: does it?

[06:11:00.0000] <annevk>
it does

[06:11:01.0000] <annevk>
it's defined in terms of strokeStyle

[06:11:02.0000] <zcorpan>
"The strokeStyle attribute represents the color or style to use for the lines around shapes"?

[06:12:00.0000] <zcorpan>
what if it's not possible to stroke for a particular font, as is the case apparently on mobiles?

[06:49:00.0000] <hsivonen>
it's taking ages to upload my slides...

[06:50:00.0000] <MikeSmith>
crap network

[07:01:00.0000] <hsivonen>
it's no longer controversial to say that XSD 1.0 sucks

[07:02:00.0000] <hsivonen>
now the audience laughs

[07:02:01.0000] <MikeSmith>
hsivonen: which session?

[07:02:02.0000] <hsivonen>
MikeSmith: Mike Kay on XSD 1.1

[07:03:00.0000] <MikeSmith>
ah

[07:10:00.0000] <annevk>
zcorpan, dunno, fix mobiles? allow some other behavior?

[07:14:00.0000] <hsivonen>
so far it seems to me that people who don't want the PSVI should use RNG and Schematron instead of XSD 1.1

[07:15:00.0000] <hsivonen>
now assertions are going to be a streamability problem for XSD too

[07:18:00.0000] <hsivonen>
give users the rope. let them hang themselves"

[07:23:00.0000] <Lachy>
cool. I've just been asked to do an interview on HTML5 for a podcast on boagworld.com

[07:25:00.0000] <hsivonen>
Lachy: does Opera PR allow you to talk with journalists?

[07:25:01.0000] <Lachy>
I don't know

[07:25:02.0000] <Lachy>
I will have to check

[07:25:03.0000] <Lachy>
I haven't accepted yet

[07:30:00.0000] <hsivonen>
quality of error messages will be a problem and a differentiating factor for impls

[07:31:00.0000] <zcorpan>
http://forums.whatwg.org/profile.php?mode=viewprofile&u=285 - spam?

[07:32:00.0000] <annevk>
looks like it, yes

[07:32:01.0000] <zcorpan>
he's been online on the forum every time i've checked the past few days. interesting way to spam

[07:33:00.0000] <annevk>
that makes him one page away from the final link

[07:34:00.0000] <annevk>
maybe he just keeps a lot of tabs open

[07:34:01.0000] <zcorpan>
indeed, and harder to spot as being a spammer

[07:36:00.0000] <hsivonen>
/me wants electricity over IP for the laptop

[07:43:00.0000] <hsivonen>
David Orchard is talking about extension and versioning

[07:45:00.0000] <annevk>
Is he mentioning CSS and all too?

[07:45:01.0000] <annevk>
/me is in the lobby working on his slides

[07:46:00.0000] <hsivonen>
annevk: not yet but he's talking about forward compat

[07:48:00.0000] <hsivonen>
well-defined processing model mentioned

[07:50:00.0000] <annevk>
(i think the solution to extensions is to have a well-defined processing model and the solution to versioning is to not have any)

[07:50:01.0000] <hsivonen>
HTML5 a great example of doing forward compat

[07:50:02.0000] <hsivonen>
counter example XML 1.0

[07:53:00.0000] <hsivonen>
annevk: so far this is good stuff

[07:54:00.0000] <annevk>
apart from the XML Schema? :)

[07:54:01.0000] <hsivonen>
annevk: not talking about schema yet

[08:00:00.0000] <hsivonen>
it's interesting how Dave is formalizing what people on this channel take intuitively and tacitly into account when designing stuff

[08:01:00.0000] <annevk>
that's prolly good for when we're not around anymore :)

[08:12:00.0000] <hsivonen>
now we got to XSD sucking totally on expressing this

[08:14:00.0000] <hsivonen>
"this is pretty gory"

[08:16:00.0000] <markpilgrim>
been off irc so long, somebody nicked my nick

[08:18:00.0000] <hsivonen>
"we don't wanna hear about SOAP, because that's evil"

[08:19:00.0000] <annevk>
nice

[08:27:00.0000] <MikeSmith>
annevk: this page from ralphm :

[08:27:01.0000] <MikeSmith>
http://ilabra.org/

[08:27:02.0000] <MikeSmith>
see the "Liliana, our new pottery artisan is originally from Argentina."

[08:28:00.0000] <MikeSmith>
running through html5lib results in unexpected DOM

[08:30:00.0000] <MikeSmith>
<a href="href="http://www.earthnskystudio.com/Pottery.html" &gt;Page Title ">

[08:30:01.0000] <Philip`>
http://james.html5.org/cgi-bin/parsetree/parsetree.py?source=%3Ca+href%3D%22href%3D%22http%3A%2F%2Fwww.earthnskystudio.com%2FPottery.html%22+%26gt%3BPage+Title+%22%3ELiliana%2C+our+new+pottery+artisan+is+originally+from+Argentina%3C%2Fa%3E

[08:30:02.0000] <Philip`>
Why is that unexpected?

[08:31:00.0000] <Philip`>
Seems to be pretty much the same as Firefox and Opera

[08:32:00.0000] <markpilgrim>
/me wonders what hsivonen is watching/reading

[08:32:01.0000] <MikeSmith>
it shouldn't result in that mess :)

[08:32:02.0000] <MikeSmith>
it should do something nice

[08:33:00.0000] <Philip`>
IE silently removes the entire link and link text from that page

[08:33:01.0000] <Philip`>
MikeSmith: Why shouldn't it?

[08:34:00.0000] <Philip`>
Garbage in plus processing model highly constrained by backward compatibility concerns results in garbage out :-p

[08:34:01.0000] <MikeSmith>
hmm, well, I guess doing something with it is rather better than removing it entirely

[08:35:00.0000] <ralphm>
I'd had expected to at least have some valid DOM. I mean, " is not really a valid attribute name, is it?

[08:38:00.0000] <zcorpan>
ralphm: <foo "> is a parse error in both text/html and xml, and trying to create an attribute '"' with dom core apis will raise an exception, so yeah, it's probably not a valid attribute name

[08:39:00.0000] <zcorpan>
ralphm: (although it was a valid attribute name for <embed> in html5 a while ago)

[08:43:00.0000] <ralphm>
zcorpan: well, not all DOM implementations actually check their input on object creation. So in my specific case, I happily got an invalid DOM and only discovered the issue when serializing it to XML and sending it over the wire using XMPP. One could claim that that is the DOM implementation's fault, but it would be nice if html5lib wouldn't feed it in the first place, no?

[08:43:01.0000] <annevk>
markpilgrim, XTech 2008, http://2008.xtech.org/public/schedule/detail/530

[08:43:02.0000] <annevk>
/me is not there

[08:43:03.0000] <Philip`>
ralphm: By "valid DOM", do you mean one that can be serialised to well-formed XML?

[08:44:00.0000] <ralphm>
Philip`, yeah

[08:44:01.0000] <ralphm>
Is that a wrong assumption?

[08:44:02.0000] <zcorpan>
ralphm: the serializer needs to do checks

[08:44:03.0000] <zcorpan>
ralphm: the dom allows things that are not serializable as xml

[08:44:04.0000] <zcorpan>
ralphm: e.g. comment with --> in data

[08:45:00.0000] <Philip`>
<html><foo:bar/></html> too

[08:46:00.0000] <ralphm>
Right, so your view is that html5lib is not at fault, and I should file a bug against the DOM implementation(s)?

[08:47:00.0000] <Philip`>
The DOM implementation ought to be able to represent things that aren't XML-well-formed, because that's necessary for compatibility with HTML content

[08:47:01.0000] <annevk>
Well, basically you would need to file a bug against the software that serializes a DOM to well-formed XML

[08:47:02.0000] <annevk>
(but from your comments it doesn't seem you use such a serializer)

[08:47:03.0000] <annevk>
/me isn't sure if there is one

[08:48:00.0000] <Philip`>
(and the DOM-to-XML serialiser should guarantee well-formedness for any arbitrary DOM)

[08:48:01.0000] <zcorpan>
/me might write such a serializer in js just for fun

[08:48:02.0000] <Philip`>
function serialiseDOMToWellFormedXML(dom) { return '<x/>' }

[08:48:03.0000] <annevk>
(yeah, if it throws an exception it's not the software you want)

[08:49:00.0000] <annevk>
(if it does that it doesn't do what it claims to do...)

[08:49:01.0000] <Philip`>
Any such serialiser is going to have to be lossy for at least some inputs

[08:49:02.0000] <Philip`>
Is there a nice way to define what lossiness is acceptable?

[08:49:03.0000] <zcorpan>
dom to xml 1.0 4th ed 

[08:50:00.0000] <zcorpan>
configurable whether it should halt or alter infoset

[08:51:00.0000] <zcorpan>
what does TagSoup do?

[09:02:00.0000] <gsnedders>
annevk: http://w3.org/mid/E5851D3C-1B2C-49F0-9A23-61A56DD4A4FC⊙gc

[09:40:00.0000] <annevk>
room for hsivonen's talk is near empty :(

[09:40:01.0000] <annevk>
it's pretty obvious people don't care about validation

[09:43:00.0000] <othermaciej>
annevk: have you seen the File IO spec that chaals just posted?

[09:43:01.0000] <annevk>
yeah, I wasn't involved in it

[09:43:02.0000] <virtuelv>
othermaciej: I wrote most of it

[09:43:03.0000] <othermaciej>
virtuelv: it looks like a giant security hole

[09:43:04.0000] <othermaciej>
just replied to the list in more detail

[09:44:00.0000] <virtuelv>
othermaciej: how so?

[09:45:00.0000] <othermaciej>
virtuelv: see my email for some examples of security risks

[09:45:01.0000] <annevk>
one of the things that is unclear from chaals' message is that I believe this is not intended for Web applications but for widgets

[09:45:02.0000] <annevk>
so it's not entirely clear how it replaces "file upload"

[09:45:03.0000] <othermaciej>
if it is not meant for Web applications then I guess I am not as worried about security risks, but I also do not see how it is relevant to the group

[09:45:04.0000] <virtuelv>
othermaciej: in the context of web pages served over HTTP, I agree with you

[09:46:00.0000] <virtuelv>
for widgets/gadgets, not so much

[09:46:01.0000] <othermaciej>
virtuelv: ok, nothing in the spec makes clear that it is not meant to be exposed to web apps served over HTTP

[09:46:02.0000] <othermaciej>
"This document describes an interface for an abstract File I/O interface where web applications can interact with a file system, without any prior knowledge about the underlying filesystem."

[09:46:03.0000] <virtuelv>
othermaciej: that needs to be remedied, I agree

[09:47:00.0000] <virtuelv>
btw, all browseFor* methods have UI interaction restrictions in our implementation

[09:50:00.0000] <othermaciej>
virtuelv: when I read the spec it almost gave me a heart attack

[09:50:01.0000] <othermaciej>
because I assumed Web application was meant to refer to Web apps on the Web

[09:50:02.0000] <virtuelv>
heh

[09:52:00.0000] <MikeSmith>
hsivonen now doing validator.nu presentation at XTech

[09:53:00.0000] <othermaciej>
I'm not sure an API that is only for locally installed applications is in scope for Web API WG though

[09:56:00.0000] <virtuelv>
othermaciej: it sort of is, given the merge between appformats and webapi

[09:58:00.0000] <othermaciej>
our charter says: "Although the primary focus will be handling of content deployed over the Web, the deliverables of the Web Applications Working Group should take into consideration uses of Web technologies for other purposes, such as the purpose of building user interfaces on devices; for example, user interfaces in multimedia devices such as digital cameras and in industrial information tools such RFID/barcode scanners and checko

[09:58:01.0000] <othermaciej>
(the new charter)

[09:58:02.0000] <othermaciej>
it also says: "Furthermore, the Web Applications Working Group deliverables must address issues of accessibility, internationalization, mobility, and security."

[09:59:00.0000] <othermaciej>
I guess it could be considered in scope, though I'm not very interested personally in APIs that aren't for the WEb

[10:13:00.0000] <gsnedders>
othermaciej: I may be wrong, but I think there is no such thing as a split infinitive in German

[10:14:00.0000] <gsnedders>
(totally useless comment)

[10:15:00.0000] <virtuelv>
othermaciej: I take it "widgets != web" for you then?

[10:16:00.0000] <Philip`>
German puts verbs in crazy places

[10:16:01.0000] <Philip`>
"Ich kann blah-blah-blah-blah verb" etc

[10:17:00.0000] <Philip`>
It's too easy to forget what you're talking about before the end of the sentence :-(

[10:21:00.0000] <gsnedders>
Philip`: Better than Latin where the verb can be anywhere, though is typically at the end

[10:24:00.0000] <gsnedders>
Philip`, jgraham__: Seeming jgraham__ wanted somewhere nearer the IoA how about outside the front of Selwyn?

[10:25:00.0000] <gsnedders>
/me can't get much closer without going away from where he knows

[10:25:01.0000] <Philip`>
/me wonders where Selwyn is

[10:25:02.0000] <gsnedders>
Oh dear…

[10:26:00.0000] <gsnedders>
/me really does like Philip`'s knowledge of Cambridge after he's been there for 3.5 years

[10:26:01.0000] <Philip`>
Oh, right, I walk past it twice every day

[10:26:02.0000] <gsnedders>
Philip`: Grange Road, between West Road and Sidwick Avenue

[10:26:03.0000] <gsnedders>
Philip`: That's what makes it even worse.

[10:27:00.0000] <gsnedders>
It's between Newham Croft and the faculty of Comp.Sci. :P

[10:27:01.0000] <gsnedders>
/me may be in his late grandmother's flat before we meet up in which case he's more or less there

[10:28:00.0000] <gsnedders>
(the flat is in one of the blocks in the bit opposite Selwyn)

[10:29:00.0000] <gsnedders>
Philip`: You have any thought about time?

[10:29:01.0000] <Philip`>
Selwyn doesn't really seem any closer than King's

[10:29:02.0000] <Philip`>
(Maybe five minutes or so)

[10:29:03.0000] <Philip`>
(unless I'm just misreading the map badly)

[10:29:04.0000] <Philip`>
but anyway, doesn't make much difference to me :-)

[10:30:00.0000] <gsnedders>
I guess from up there it makes little difference

[10:30:01.0000] <Philip`>
gsnedders: No

[10:31:00.0000] <gsnedders>
/me has no idea how hard it'll be to get somewhere to eat on a Friday night at this time of year

[10:32:00.0000] <Philip`>
/me neither

[10:32:01.0000] <gsnedders>
/me doesn't normally go down when it is his term time :P

[10:33:00.0000] <Philip`>
/me normally just eats in college or at home, since those are easy and inexpensive :-)

[10:33:01.0000] <Philip`>
s/normally/always/

[10:33:02.0000] <gsnedders>
Philip`: We can all come to your home if you want :P

[10:34:00.0000] <Philip`>
Probably a bad idea :-p

[10:37:00.0000] <gsnedders>
Philip`: We can just meet by King's if that's easier :P

[10:37:01.0000] <gsnedders>
Philip`: More places to eat around there, too

[10:39:00.0000] <Philip`>
/me shrugs

[12:23:00.0000] <gsnedders>
GTA4 is far too short. Eight days and thirty-five hours for the first run through of the story-line.

[12:24:00.0000] <Philip`>
Eight days and thirty-five hours? Isn't that just nine days and eleven hours?

[12:24:01.0000] <gsnedders>
Philip`: I mean thirty-five hours of actually playing, over eight days.

[12:24:02.0000] <Philip`>
Ah

[12:25:00.0000] <gsnedders>
4.375 hours per day, average (eek)

[12:25:01.0000] <Philip`>
That's not showing much dedication to the game

[12:26:00.0000] <Philip`>
/me has got dragged back into Advance Wars again :-(

[12:26:01.0000] <Philip`>
(I was looking at it yesterday and noticed most of the characters were only at level 2, so I had to get them a bit more experience...)

[12:26:02.0000] <gsnedders>
Oh dear…

[12:27:00.0000] <gsnedders>
/me hopes he doesn't need to drag Philip` to supper :P

[12:28:00.0000] <Philip`>
(The problem is that I'm not very good at strategy, or at tactics, so I just throw units at the enemy for an hour until they either get worn down or defeat me)

[13:22:00.0000] <othermaciej>
does anyone here know what virtualev's real life name is?

[13:27:00.0000] <othermaciej>
any Opera people around?

[13:27:01.0000] <Philip`>
othermaciej: http://virtuelvis.com/

[13:28:00.0000] <othermaciej>
Philip`: thanks!

[13:59:00.0000] <Hixie>
yeah Philip` your comments seem fine

[13:59:01.0000] <Hixie>
though comments sent to the public-html list are more likely to fall through the cracks

[13:59:02.0000] <Hixie>
so if you want a guarenteed reply, you might want to use the whatwg list

[14:00:00.0000] <Hixie>
(the way to get a guarenteed reply with the htmlwg is to use the issues list)

[14:06:00.0000] <Philip`>
Hixie: If my comments get lost or ignored, I believe I'll notice that at some point in the future and can bring them up again; so in the meantime I'd prefer to attempt to contribute to the ratio of technical discussion on public-html, rather than giving up on it yet :-)

[14:11:00.0000] <Dashiva>
Hixie: Do you know anything of the process to return to invited expert status after being part of a Member

[14:11:01.0000] <Dashiva>
+?

[14:13:00.0000] <jgraham__>
gsnedders: I think Kings is better. Selwyn is in the middle of nowhere

[15:00:00.0000] <Hixie>
Dashiva: no idea

[15:00:01.0000] <Hixie>
Dashiva: ask mikesmith

[15:01:00.0000] <Dashiva>
Will do

[15:03:00.0000] <hasather>
Dashiva: Chaals fixed that for me

[15:05:00.0000] <Dashiva>
I figure there's little point in asking to join Google if I'm not going to be doing w3c work :)

[15:07:00.0000] <gsnedders>
jgraham__: ping

[15:08:00.0000] <jgraham__>
gsnedders: Pong

[15:09:00.0000] <gsnedders>
jgraham__: See prior discussion. Short version: It's probably best to meet somewhere like King's.

[15:09:01.0000] <gsnedders>
jgraham__: Wait, that last thing was you.

[15:09:02.0000] <gsnedders>
I'm slow.

[15:09:03.0000] <gsnedders>
/me is half asleep here

[15:10:00.0000] <jgraham__>
Yeah, I suggest Kings at 7, unless that's too late for you. Finding somewhere to eat will be trivial

[15:10:01.0000] <gsnedders>
The front of King's, I assume?

[15:11:00.0000] <Philip`>
Which part of the front?

[15:11:01.0000] <gsnedders>
7 is fine by me.

[15:11:02.0000] <gsnedders>
Philip`: King's Parade?

[15:11:03.0000] <jgraham__>
By the postbox

[15:11:04.0000] <gsnedders>
There's a postbox there? I guess I'll have to see it when I get there.

[15:11:05.0000] <Philip`>
gsnedders: That's a whole street, which is pretty vague :-p

[15:12:00.0000] <Philip`>
It only took me six months to discover the postbox

[15:12:01.0000] <gsnedders>
Philip`: But that's the front! :P

[15:12:02.0000] <gsnedders>
Philip`: Yeah, but your knowledge of Cambridge, well, leaves plenty to be desired :)

[15:12:03.0000] <jgraham__>
http://www.flickr.com/photos/andrewdunn/490994558/

[15:13:00.0000] <gsnedders>
Right hand side, when facing it, then?

[15:13:01.0000] <jgraham__>
(it's quite a distinctive postbox :)

[15:13:02.0000] <Philip`>
Hint: It's tall and red

[15:13:03.0000] <gsnedders>
Philip`: OMFG.

[15:14:00.0000] <jgraham__>
If it helps, I will be the person who looks a bit like I do on the internet

[15:14:01.0000] <gsnedders>
jgraham__: Which is to say, what?

[15:15:00.0000] <jgraham__>
Mainly composed of ascii 

[15:15:01.0000] <gsnedders>
http://flickr.com/photos/gsnedders/2396366112/in/set-72157604424683450/

[15:15:02.0000] <gsnedders>
c'est moi.

[15:15:03.0000] <jgraham__>
(seriously, if you search flickr you should see my photo without much trouble)

[15:16:00.0000] <gsnedders>
/me concludes looking through jgraham__'s photos for him isn't overly successful

[15:16:01.0000] <gsnedders>
Ah, OK. The guy who looks like a physicist who does a lot online.

[15:17:00.0000] <jgraham>
?!

[15:17:01.0000] <gsnedders>
:P

[15:18:00.0000] <jgraham>
/me wonders which photo /that/ came from

[15:19:00.0000] <gsnedders>
All of them in general

[15:23:00.0000] <jgraham>
Fair enough

[15:24:00.0000] <jgraham>
(although I don't think I look like many actual physicists)

[15:37:00.0000] <gsnedders>
/me should be able to write more of HTTP5 on the train on Fri/Sun

[15:39:00.0000] <gsnedders>
markpilgrim: all the Atom autodiscovery tests on ragingplatypus.com are broken

[16:01:00.0000] <Hixie>
/me gets some gold-level valuable feedback from people at Google working on pages that use contentEditable

[16:02:00.0000] <gsnedders>
/me wants platinum-level valuable feedback

[16:02:01.0000] <gsnedders>
Then I can sell it! Mwhahaha!

[16:04:00.0000] <Hixie>
your market would consist of just me

[16:05:00.0000] <gsnedders>
I know, that's slightly problematic.

[16:05:01.0000] <gsnedders>
You don't have much competition pushing the value up.

[16:06:00.0000] <gsnedders>
That's it. I'm writing HTML 2π.

[16:06:01.0000] <tommorris>
You are competing with the XHTML 2 Working Group. ;)

[16:07:00.0000] <Hixie>
i recommend not taking on more work than you can do :-P

[16:07:01.0000] <Hixie>
given you're already doing HTTP5 and the spec preprocessor :-)

[16:07:02.0000] <gsnedders>
Hixie: Already done that :P

[16:08:00.0000] <gsnedders>
GTA4, HTTP5, spec-gen, participating in far too many WGs, writing short stories. Oh, and school.

[16:08:01.0000] <gsnedders>
Or should I say, Secondary5.

[16:08:02.0000] <gsnedders>
:P

[16:08:03.0000] <gsnedders>
Next year of school sucks. Sixth year! 5 > *.

[16:10:00.0000] <gsnedders>
Hixie: Oh, and meaning to help with XML5, but never finding the time to do so.

[16:12:00.0000] <tommorris>
XML 5? For chrissakes, Sun's already done the version number thing. Stop it already. :)

[16:12:01.0000] <gsnedders>
tommorris: And Bible5, too.

[16:13:00.0000] <gsnedders>
http://lists.w3.org/Archives/Public/www-html/2007Jun/0008.html

[16:14:00.0000] <tommorris>
"Happy shall USER AGENT be that dasheth BLOCK LEVEL ELEMENTS against the rocks."

[16:15:00.0000] <gsnedders>
tommorris: that doesn't look like en-gb-x-Hixie to me

[16:18:00.0000] <tommorris>
The Old Testament specification seems quite clear - kill gay people, don't eat shellfish and don't wear cloths of more than two types of fabric. I blame liberal parsing rules.

[16:18:01.0000] <hsivonen>
http://lists.w3.org/Archives/Public/www-validator/2008May/0045.html

[16:19:00.0000] <Hixie>
what am i looking for?

[16:26:00.0000] <gsnedders>
tommorris: No, it's not as simple as gay people. Men who have sex with other men.

[16:26:01.0000] <jgraham>
hsivonen: Fun

[16:26:02.0000] <tommorris>
that's defined in the accompanying normative BNF segment and referenced test suite

[16:27:00.0000] <gsnedders>
tommorris: And besides, Jesus depreciated the old testament

[16:27:01.0000] <gsnedders>
Also, don't sleep with a women and her mother. That's punishable by death too.

[16:27:02.0000] <tommorris>
Jesus: the original quirks mode parser.

[16:28:00.0000] <gsnedders>
Leviticus 20 is always nice. 

[16:47:00.0000] <gsnedders>
Hixie: If not taking on more work than you can do is your advice for today, can I have some more daily advice tomorrow? :P

[16:48:00.0000] <Hixie>
no, i have too much work tomorrow

[16:49:00.0000] <gsnedders>
Then it isn't daily advice :(

[16:49:01.0000] <gsnedders>
Weekly?

[16:52:00.0000] <hsivonen>
should I reply to http://lists.w3.org/Archives/Public/public-html/2008May/0090.html or just let t be?

[16:52:01.0000] <Hixie>
depends whether by replying you can advance your agenda

[16:53:00.0000] <Hixie>
(e.g. whether you can make the web better)

[16:53:01.0000] <Dashiva>
Wouldn't that be 'our' agenda, not just his? :)

[16:53:02.0000] <gsnedders>
Dashiva: No, our agenda is making sure we're awesome. Is it not?

[16:53:03.0000] <Hixie>
i wouldn't pretend to ask him to reply to advance mine

[16:54:00.0000] <gsnedders>
Or to make Hixie rich and famous?

[16:54:01.0000] <Dashiva>
I thought he was both already

[16:54:02.0000] <Dashiva>
And I don't need htmlwg to convince me I'm awesome :P

[16:54:03.0000] <gsnedders>
Dashiva: You're awesome.

[16:54:04.0000] <hsivonen>
Hixie: that's the thing. I don't knwo if replying and explaining the positive/negative/zero-level stuff again helps

[16:54:05.0000] <hsivonen>
though the book argument seems bogus

[16:55:00.0000] <Hixie>
hsivonen: well, notice how i'm basically given up on that thread

[16:55:01.0000] <Hixie>
hsivonen: i read all the comments still, but i don't have anything to say that i haven't already said

[16:55:02.0000] <Hixie>
hsivonen: and nothing that has been said recently has been new in any way

[16:55:03.0000] <Hixie>
hsivonen: so it hasn't affected by conclusions (which were based on the earlier arguments)

[16:56:00.0000] <Dashiva>
I'm tempted to translate that post above as "Nobody cares about WCAG", but objectively I know it's just spite :)


2008-05-08
[17:11:00.0000] <gsnedders>
What does <> mean in SGML?

[17:12:00.0000] <Hixie>
start tag

[17:12:01.0000] <gsnedders>
Yeah, but a start tag of what

[17:13:00.0000] <Philip`>
Maybe it's like </> and can be inferred from context

[17:13:01.0000] <Philip`>
s//sometimes/

[17:13:02.0000] <Philip`>
(and when it can't be inferred, I assume bad things happen)

[17:13:03.0000] <Dashiva>
<ul><> implies <li>?

[17:14:00.0000] <Philip`>
<ul><>Item 1<>Item 2</ul> could make sense

[17:14:01.0000] <Philip`>
I might just be making this up, though

[17:16:00.0000] <tommorris>
/me shudders at the thought of SGML

[17:16:01.0000] <gsnedders>
/me realises he has the SGML handbook beside him

[17:16:02.0000] <Dashiva>
Masochist

[17:17:00.0000] <Philip`>
http://www.google.com/search?q=%3C%3E :-(

[17:17:01.0000] <tommorris>
prefers standards that have an 'X' in the name, if only because X sounds cool

[17:17:02.0000] <Dashiva>
XSGML

[17:17:03.0000] <Hixie>
yes it's the start tag of whatever only tag is allowed at that point

[17:18:00.0000] <gsnedders>
And if more than one tag is allowed behaviour is undefined, I guess?

[17:18:01.0000] <Hixie>
i forget

[17:18:02.0000] <Hixie>
it's been years since i read the spec

[17:19:00.0000] <Philip`>
tommorris: Do you prefer TORX-head screws too?

[17:19:01.0000] <tommorris>
so long as there is a DTD available for the screwdriver

[17:20:00.0000] <Philip`>
Dusty tool drawer?

[19:44:00.0000] <Hixie>
man, merging wf2 is gonna be such a lot of work

[21:31:00.0000] <othermaciej>
Hixie: I love that your post had a "ratholes" section

[22:07:00.0000] <Hixie>
no comment

[22:33:00.0000] <Hixie>
gsnedders: in some feedback you wrote "The ratio subsection of common microsyntaxes does not do what the rest of the numbers do: define what a valid number is"

[22:33:01.0000] <Hixie>
gsnedders: that's because there's no text that uses that concept

[22:33:02.0000] <Hixie>
gsnedders: so defining it would be pointless

[23:04:00.0000] <Hixie>
gsnedders: similarly, the reason it doesn't report errors is that there is nothing to report

[23:52:00.0000] <Hixie>
gsnedders: i am similarly ignoring other e-mails that request editorial changes that i don't think would make the spec better

[23:52:01.0000] <Hixie>
gsnedders: e.g. "Ratios algorithm continues beyond data (part of detailed review of common microsyntaxes)"

[01:13:00.0000] <hsivonen>
I wonder if I could still get a lightning talk slot

[01:16:00.0000] <MikeSmith>
hsivonen: we have about 9 people signed up already, so maybe not

[01:16:01.0000] <hsivonen>
ok

[01:17:00.0000] <MikeSmith>
but if you have time, may be worth preparing some slides anyway

[01:17:01.0000] <MikeSmith>
last year we had a few people cancel

[01:26:00.0000] <hsivonen>
MikeSmith: ok

[02:01:00.0000] <othermaciej>
hi annevk 

[02:02:00.0000] <annevk>
morning

[02:07:00.0000] <MikeSmith>
http://www.microsoft.com/windowsxp/expertzone/chats/transcripts/08_0417_ez_ie8.mspx

[02:07:01.0000] <MikeSmith>
not much interesting there

[02:08:00.0000] <MikeSmith>
except the usual "At the moment, we're not ready to talk about features in upcoming releases, but we appreciate the input." any time anybody asks anything about standards support they currently lack

[02:08:01.0000] <MikeSmith>
plus some misrepresentations

[02:10:00.0000] <MikeSmith>
e.g., Eric Law saying that the Access Control spec "remains under heavy edit"

[02:11:00.0000] <MikeSmith>
and that that was why Mozilla "pulled support for the feature in their latest beta"

[02:11:01.0000] <othermaciej>
well, that was kind of what Mozilla sort of said

[02:13:00.0000] <annevk>
/me is presenting on Access Control today

[02:14:00.0000] <MikeSmith>
nifty rationalization:

[02:14:01.0000] <MikeSmith>
[[

[02:14:02.0000] <MikeSmith>
Cyrar [MSFT] (Expert):

[02:14:03.0000] <MikeSmith>
qwert666: Stop saying "we're not ready to talk about features for upcoming releases" all the time and be more open about what you are considering to do. 

[02:14:04.0000] <MikeSmith>
Answer: Hey Opaque, sorry for the frustration. We generally have a policy about commenting on features that are not built, tested and ready for release. Suppose that we did comment on a feature that was important to you, but then had to cut it because quality or some other blocking factor. If you had made business plans around that feature - I imagine that you would be pretty frustrated. Our policy gives our customers more predictability in 

[02:15:00.0000] <MikeSmith>
their planning

[02:15:01.0000] <MikeSmith>
]]

[02:15:02.0000] <othermaciej_>
MikeSmith: Apple has the same policy sort of - so I can't cast the first stone

[02:17:00.0000] <othermaciej>
MikeSmith: although by WebKit being open source we have managed to push the window for engine features back from "released to customers" to "checked in"

[02:18:00.0000] <MikeSmith>
yeah, exactly

[02:18:01.0000] <Philip`>
So they're saying their customers are stupid and will be confused by being given more information on which to base their business plans? :-)

[02:18:02.0000] <othermaciej>
but people do ask us "when will feature X from the WebKit nightlies make it into a Safari build"

[02:18:03.0000] <MikeSmith>
Philip`: yeah, kind of like the Bush administration justification for why they like to keep so much stuff secret

[02:18:04.0000] <othermaciej>
and we can't really answer

[02:18:05.0000] <Hixie>
microsoft have a policy of not talking about features that are not built, tested, and ready for release?

[02:19:00.0000] <Hixie>
since when?!?!

[02:19:01.0000] <othermaciej>
hmmm good point

[02:19:02.0000] <othermaciej>
Hixie: maybe that policy doesn't apply to whole new products or product lines

[02:20:00.0000] <Hixie>
how about, say, file systems?

[02:20:01.0000] <Hixie>
WinFS has been announced under various names since at least 1995

[02:20:02.0000] <Hixie>
i remember reading about it when i was looking at being on the Win95 beta

[02:21:00.0000] <MikeSmith>
othermaciej_: so IE team could do something just like Webkit vs Safari relationship .. that is, be open about what they have in development in the engine, but no guarantees if or when it actually will make it into IE

[02:21:01.0000] <othermaciej_>
dude that was totally ready

[02:21:02.0000] <othermaciej_>
WinFS was ready for the world

[02:21:03.0000] <othermaciej_>
but the world was not ready for it

[02:22:00.0000] <Hixie>
which time that it was pulled was it not ready?

[02:22:01.0000] <Hixie>
er i mean, was the world not ready?

[02:22:02.0000] <MikeSmith>
/me wonders if anybody knows John Hrvatin

[02:23:00.0000] <MikeSmith>
"...y name is John Hrvatin. I'm the program manager for developer tools, jscript engine integration, and HTML editing support."

[02:23:01.0000] <othermaciej>
Hixie: just wait until you see all the awesome things coming in Windows Mobile 9 and Windows 7

[02:23:02.0000] <othermaciej>
Hixie: then you'll forget all about this

[02:23:03.0000] <Hixie>
mmm

[02:26:00.0000] <Philip`>
Seems like their unwillingness to talk about the future doesn't just apply to technical features - "As for any future cake plans, I'm afraid we consider all bakery related strategies highly confidential."

[02:27:00.0000] <othermaciej>
they sent the firefox guys a cake but they never sent us one at apple

[02:27:01.0000] <Philip`>
Did you ever send them one?

[02:28:00.0000] <othermaciej>
no, but they didn't send us one first

[02:29:00.0000] <Philip`>
Maybe they are waiting for you to make the first move

[02:31:00.0000] <zcorpan_>
/me wonders where a 32768x32768 icon would be used

[02:31:01.0000] <zcorpan_>
or even 8192x8192

[02:32:00.0000] <zcorpan_>
is it normal for .icns to have such sizes?

[02:42:00.0000] <Philip`>
zcorpan_: The largest I see out of about a hundred <link>ed .ico files is 48x48

[02:43:00.0000] <Philip`>
I'm not aware of many desktop systems that expect 8K-wide icons

[02:43:01.0000] <Philip`>
though I thought there were at least some that could use vector (SVG) icons, for which size doesn't make sense

[02:45:00.0000] <Lachy>
Mac OSX icons go up to 512x512. 8192x8192 is bigger than most (if not all) desktop wallpapers! 

[02:45:01.0000] <annevk>
if sizes is not specified does it mean any?

[02:45:02.0000] <annevk>
it also doesn't seem clear what happens when there's multiple icons with the same set of sizes

[02:46:00.0000] <annevk>
should the UA inspect them all, take the first, etc.

[02:46:01.0000] <annevk>
take the first non-broken

[02:46:02.0000] <zcorpan_>
annevk: not specified should mean unknown

[02:47:00.0000] <annevk>
how is that useful?

[02:47:01.0000] <annevk>
nm

[02:51:00.0000] <Hixie>
zcorpan_: it was a joke on the ridiculousness of mac "icons" these days (though apparently it went above everyone's heads :-P)

[02:52:00.0000] <zcorpan_>
Hixie: ah :)

[02:53:00.0000] <Philip`>
/me is surprised that .ico files (i.e. DIBs) aren't limited to 64Kx64K

[03:02:00.0000] <gsnedders>
Hixie: Hey, you want my feedback from last summer to be useful!?

[03:02:01.0000] <Hixie>
some of it is :-)

[03:02:02.0000] <gsnedders>
Some of it finds major issues with the algorithms, actually. So some :)

[03:03:00.0000] <gsnedders>
/me is online as he is at home as there's no point in him even being at school before this essay is finished.

[03:03:01.0000] <gsnedders>
Fun :\

[03:09:00.0000] <gsnedders>
Hixie: Now I need to go and fix my test cases of that section :P

[03:10:00.0000] <Hixie>
can someone with IE send me the results of this test in IE? http://www.hixie.ch/tests/adhoc/html/flow/image-maps/004-demo.html

[03:11:00.0000] <zcorpan_>
Hixie: "'http://www.hixie.ch/tests/adhoc/html/flow/image-maps/004-demo.html' is currently unavailable."

[03:11:01.0000] <Hixie>
o_O

[03:11:02.0000] <zcorpan_>
ah now it works..

[03:12:00.0000] <Hixie>
probably my memory allocation a bit low

[03:12:01.0000] <zcorpan_>
50%...

[03:12:02.0000] <Philip`>
Any particular versions of IE?

[03:12:03.0000] <zcorpan_>
/me tests in ie8b

[03:13:00.0000] <zcorpan_>
acts as separatorU+0, U+20, U+2c, U+3b

[03:13:01.0000] <zcorpan_>
long list... Total for "acts like a decimal point": 17882

[03:14:00.0000] <zcorpan_>
acts as digitU+30 - U+39

[03:14:01.0000] <Hixie>
can you mail me the list?

[03:14:02.0000] <Hixie>
ian⊙hc

[03:14:03.0000] <zcorpan_>
sure

[03:14:04.0000] <Hixie>
thanks dude

[03:14:05.0000] <Philip`>
In IE8b1 I get "aborts parsing" at the end; in IE6 I get "<AREA shape=RECT coords=1,0,2,2>" at the end

[03:14:06.0000] <Philip`>
(and neither version has the other section)

[03:14:07.0000] <Hixie>
ooo

[03:14:08.0000] <Hixie>
mail me the IE6 list too :-)

[03:14:09.0000] <Hixie>
pretty please :-)

[03:14:10.0000] <Philip`>
This is IE6 in Wine so I don't know if that makes things go weird

[03:15:00.0000] <Hixie>
that's fine

[03:15:01.0000] <zcorpan_>
hmm ie8 hanged when i tried to select all :(

[03:15:02.0000] <Hixie>
huh

[03:15:03.0000] <Hixie>
that's what i got

[03:15:04.0000] <Hixie>
i figured it was a VMWare problem

[03:15:05.0000] <Hixie>
wtf

[03:16:00.0000] <gsnedders>
IE8b1 needs a bigger emphasis on the beta part :)

[03:16:01.0000] <zcorpan_>
ie6 doesnt' work for me

[03:16:02.0000] <Hixie>
script error?

[03:16:03.0000] <zcorpan_>
no ie6 never works for me. it doesn't load pages

[03:17:00.0000] <Philip`>
My IE6 doesn't load any UI, but the browser bit works fine

[03:17:01.0000] <hsivonen>
I think I need to ask the annual microformat question in a few minutes

[03:18:00.0000] <zcorpan_>
Hixie: does the page create many text nodes?

[03:18:01.0000] <Hixie>
yes

[03:18:02.0000] <zcorpan_>
Hixie: maybe that's what fucks up selection

[03:18:03.0000] <Hixie>
Philip`: thanks

[03:19:00.0000] <Hixie>
zcorpan_: yeah

[03:19:01.0000] <Hixie>
zcorpan_: quite possible

[03:19:02.0000] <Hixie>
ok i got the list from Philip`, that should do

[03:19:03.0000] <Philip`>
Hixie: Also sent the list from IE7 in real Windows, in case that's relevant

[03:20:00.0000] <Philip`>
I seem to get very different output in every version of IE

[03:20:01.0000] <Hixie>
Philip`: thanks

[03:21:00.0000] <Philip`>
I wonder if IE5.0 still works...

[03:21:01.0000] <Hixie>
wtf

[03:21:02.0000] <Hixie>
why would ie6 and ie7 give different results

[03:21:03.0000] <Philip`>
Alas, script error in IE5 :-(

[03:22:00.0000] <Philip`>
Hixie: Maybe they use system APIs to detect whether characters are decimal points, and those APIs are working differently?

[03:22:01.0000] <Hixie>
craazy

[03:22:02.0000] <zcorpan_>
Hixie: i managed to copy it now

[03:22:03.0000] <hsivonen>
annual question asked

[03:23:00.0000] <hsivonen>
answer: the microformat folks are lazy

[03:23:01.0000] <Hixie>
hsivonen: what's the question? "do you have a spec yet?"

[03:23:02.0000] <hsivonen>
yeah

[03:23:03.0000] <Philip`>
"Why are they called Microformats and not Mozformats?"

[03:23:04.0000] <Philip`>
or Appformats, which sounds like a reasonable name

[03:24:00.0000] <Lachy>
I have an interview about HTML5 next thursday evening for boagworld.com. I get to pick the topics that will be discussed. Any suggestions for what I should talk about?

[03:24:01.0000] <Hixie>
zcorpan_: did you mail the list from IE8?

[03:25:00.0000] <Hixie>
Lachy: you and the interviewer? or you and someone else? or?

[03:25:01.0000] <hsivonen>
Lachy: bringing parsing to non-browsing apps

[03:25:02.0000] <Hixie>
zcorpan_: nm, got them, thanks!

[03:26:00.0000] <Hixie>
ie8 is same as ie7, it seems

[03:26:01.0000] <zcorpan_>
Hixie: http://lists.w3.org/Archives/Public/www-archive/2008May/0015.html

[03:26:02.0000] <Hixie>
yeah i got it

[03:26:03.0000] <Hixie>
thanks

[03:26:04.0000] <Lachy>
Hixie, I don't understand your question.

[03:26:05.0000] <Hixie>
Lachy: are they interviewing you, or is this a panel discussion, or?

[03:26:06.0000] <Hixie>
Lachy: what's the context?

[03:26:07.0000] <Philip`>
Lachy: Do you know who the audience is? (Like do they already know of HTML5?)

[03:27:00.0000] <Lachy>
AFAIK, it's Paul Boag interviewing me.

[03:27:01.0000] <Philip`>
Hixie: They don't seem the same - I get quite different output from IE7 (on Win2K3) and IE8 (on Vista)

[03:27:02.0000] <Lachy>
boagworld.com seems to be a podcast aimed at web developers

[03:27:03.0000] <Hixie>
zcorpan_'s list was the same as yours as far as i could tell

[03:27:04.0000] <Lachy>
so they're likely to at least know what HTML is

[03:27:05.0000] <Hixie>
maybe it's OS-based

[03:28:00.0000] <gsnedders>
Philip`: That should be an em dash, not a hyphen!

[03:29:00.0000] <Hixie>
Lachy: i'd focus on the community we've built up

[03:29:01.0000] <Hixie>
Lachy: i think it's something that hasn't been widely recognised, and is probably the biggest achievement of the effort so far

[03:30:00.0000] <zcorpan_>
Hixie: if you do document.body.innerHTML=document.body.innerHTML then selection works fine

[03:30:01.0000] <Hixie>
zcorpan_: heh

[03:34:00.0000] <gsnedders>
zcorpan_: would not just normalising the Document object have the same affect?

[03:34:01.0000] <zcorpan_>
gsnedders: how do you normalise the Document object?

[03:35:00.0000] <Lachy>
Hixie, ok. Do you mean how things like the blog, wiki, forums, mailing lists, etc. are all open to anyone, and how they're maintained by volunteers from the community?

[03:35:01.0000] <gsnedders>
zcorpan_: Document.normalizeDocument()

[03:35:02.0000] <Hixie>
Lachy: yeah, and how we're gotten so much input from people all over the place, and how people just volunteer to do stuff

[03:35:03.0000] <gsnedders>
actually, in this case, just Document.normalize() should be enough

[03:36:00.0000] <gsnedders>
And people just volunteer to do too much stuff, more than they can do :)

[03:36:01.0000] <zcorpan_>
gsnedders: just gives a js error in ie8

[03:36:02.0000] <gsnedders>
zcorpan_: both of them?

[03:36:03.0000] <zcorpan_>
gsnedders: yeah

[03:36:04.0000] <gsnedders>
Oh well.

[03:36:05.0000] <Lachy>
yeah, I guess that could be more interesting than just the boring technical details

[03:37:00.0000] <Lachy>
hsivonen, what kind of non-browsing apps did you have in mind?

[03:37:01.0000] <Lachy>
other than validators

[03:37:02.0000] <gsnedders>
Lachy: sanitisers?

[03:37:03.0000] <Philip`>
Select-all on that test page in IE8 does work, it just takes somewhere between one and ten minutes

[03:38:00.0000] <Lachy>
CMSs too, I guess

[03:38:01.0000] <gsnedders>
Lachy: look at what WP does, for example

[03:39:00.0000] <gsnedders>
Lachy: Converting HTML to XHTML

[03:39:01.0000] <Hixie>
gsnedders: i fixed the list of integers stuff based on your feedback and on the test case peopel were just playing with

[03:40:00.0000] <gsnedders>
Hixie: Now I need to fix my test case/impl. which is how I found some of the bugs in the first place :)

[03:40:01.0000] <Hixie>
cool

[03:40:02.0000] <gsnedders>
(probably do that on the train tomorrow)

[03:40:03.0000] <gsnedders>
(and there's free wi-fi, so I may be online once or twice)

[03:50:00.0000] <zcorpan_>
Hixie: fwiw, i think the demo would be faster in ie if you build up an array, join(', ') and then append a single text node

[03:50:01.0000] <Hixie>
yeah

[03:50:02.0000] <Hixie>
it was supposed to be a one-off

[03:50:03.0000] <Hixie>
so it just kind of evolved

[03:50:04.0000] <Hixie>
for the record, IE's parsing of coords='' is fricking wacked

[03:51:00.0000] <Hixie>
gsnedders: the reason not to treat newlines the way you suggested is that IE doesn't.

[03:51:01.0000] <gsnedders>
Hixie: remind me what I suggested?

[03:52:00.0000] <Hixie>
that newline, etc, be added to the list of separators

[03:52:01.0000] <Hixie>
along with space

[03:52:02.0000] <Hixie>
same again with your complaint regarding "1. 1" parsing to 1,1 -- that's what IE does.

[03:54:00.0000] <Hixie>
regarding your request that the spec be stylisticly consistent (that the ratios section use the same style as the the other algorithms)... yeah, that'd be nice. but i'm not rewriting the spec, it's too easy to introduce errors.

[03:54:01.0000] <gsnedders>
:P

[03:55:00.0000] <gsnedders>
/me finally finishes his essay, and can head off to school

[03:56:00.0000] <Hixie>
later

[04:02:00.0000] <hsivonen>
Lachy: Pingback, Open ID, Atom feed aggregators, microformat scrapers, templating systems (incl. XSLT)

[04:07:00.0000] <Hixie>
ok the only thing on http://www.whatwg.org/issues/top that isn't on my list of things i'm not doing yet which hasn't been dealt with yet is the ruby stuff

[04:09:00.0000] <Hixie>
i guess i just have to make some semi-arbitrary decision about how to parse ruby

[04:09:01.0000] <Hixie>
since IE's parsing of ruby is a mess and nobody else does it as far as i know

[04:11:00.0000] <MikeSmith>
Hixie: I seem to remember you listing workers somewhere as a work item you're planning on putting time into at some point. Is that still for post HTML 5?

[04:11:01.0000] <Hixie>
http://hixie.ch/specs/dom/messages/0.9 http://hixie.ch/specs/dom/workers/0.9

[04:12:00.0000] <Hixie>
if it gets implementation traction i'll bring it into html5

[04:12:01.0000] <MikeSmith>
Ok, thanks

[04:12:02.0000] <Hixie>
i'd rather implementations work on other html5 things though

[04:12:03.0000] <Hixie>
like <datagrid>

[04:12:04.0000] <MikeSmith>
datagrid would be definitely great to have

[04:13:00.0000] <MikeSmith>
has any implementors shown signs of planning to work on datagrid implementations?

[04:13:01.0000] <MikeSmith>
that might have been a good opportunity for Google SoC

[04:14:00.0000] <othermaciej>
I think we had it as a suggestes GSoC project for WebKit but no one took the bait

[04:14:01.0000] <othermaciej>
we have students planning to do XBL2 and WF2 though (the former might be too big a project)

[04:15:00.0000] <MikeSmith>
othermaciej: yeah, I noted that at the end of the presentation I did yesterday

[04:15:01.0000] <MikeSmith>
which, fwiw, is online here:

[04:15:02.0000] <MikeSmith>
http://www.w3.org/2008/Talks/05-07-smith-xtech/

[04:16:00.0000] <MikeSmith>
and PDF here: http://w3.org/2008/Talks/05-07-smith-xtech/slides.pdf

[04:16:01.0000] <MikeSmith>
also noted the SVG Filters plan

[04:16:02.0000] <Dashiva>
Hey MikeSmith, got time for a chat?

[04:16:03.0000] <MikeSmith>
Dashiva: aye

[04:18:00.0000] <annevk>
othermaciej, cool that WebKit will do WF2!

[04:29:00.0000] <othermaciej>
MikeSmith: pretty good summary of what is new in browser land

[04:38:00.0000] <MikeSmith>
othermaciej: just tried to cover the last 12 months (since the previous XTech in May 2007)

[04:40:00.0000] <MikeSmith>
but putting it together, surprised me how much has gotten new stuff has gotten implemented/refined in that relatively short amount of time

[04:41:00.0000] <virtuelv>
MikeSmith: were you at Bert's talk?

[04:41:01.0000] <MikeSmith>
virtuelv: nope

[04:41:02.0000] <MikeSmith>
why you ask?

[04:41:03.0000] <othermaciej>
there was actually quite a bunch of new stuff in Safari 3.1

[04:41:04.0000] <othermaciej>
engine-wise

[04:41:05.0000] <virtuelv>
because someone asked me for the URL, and I'm incapable of remembering it

[04:41:06.0000] <othermaciej>
I neglected to make a comprehensive blog post at the time but I may try again soon

[04:55:00.0000] <MikeSmith>
othermaciej: a comprehensive blog post would be great :)

[04:55:01.0000] <MikeSmith>
another thing that has changed over the last 12 months

[04:55:02.0000] <MikeSmith>
is that back in May/June of last year,

[04:55:03.0000] <MikeSmith>
you guys were hardly blogging at all at webkit.org

[04:56:00.0000] <MikeSmith>
but since then, you and dhyatt quite prolific

[04:57:00.0000] <othermaciej>
I see 8 blog posts in June 2007

[04:57:01.0000] <MikeSmith>
OK, well, I guess May

[06:05:00.0000] <zcorpan_>
MikeSmith: great presentation

[06:56:00.0000] <MikeSmith>
Hixie: you still awake?

[06:57:00.0000] <MikeSmith>
virtuelv: I asked Bert to find you and let you know the URL for his slides

[07:01:00.0000] <virtuelv>
MikeSmith: I know, I got the URL just now, he's sitting next to me

[07:30:00.0000] <Lachy>
Hixie, the spec is too big. I tried loading it on my new OLPC yesterday, and it seemed to consume all the memory and get unresponsive.

[07:32:00.0000] <Philip`>
There's already the multipage spec for people with rubbish hardware/software :-)

[07:32:01.0000] <Lachy>
or maybe that's just because Firefox is slow and using up too much memory on it

[07:32:02.0000] <Lachy>
yeah, but I don't like the multipage version. It makes it harder to switch between sections

[07:33:00.0000] <Philip`>
Could the multipage version be improved to reduce that problem?

[07:33:01.0000] <Lachy>
But it means I won't be able to replace my MacBook Pro with it

[07:34:00.0000] <Lachy>
Philip`, only by combining it into a single page

[07:34:01.0000] <Philip`>
Lachy: What do you need that isn't handled by inline links and the table of contents?

[07:36:00.0000] <Lachy>
the ability to search the entire spec for what I'm looking for, without having to know exactly which section it's in and finding that section in the TOC first

[07:36:01.0000] <Philip`>
So a 'search' box could solve that problem?

[07:37:00.0000] <Lachy>
it might be useful to have some sort of spec search, but I like to use the find-as-you-type feature in my browser, which is instant in most cases.

[07:39:00.0000] <Philip`>
/me doesn't like his browser's find-as-you-type feature since it freezes for ten seconds trying to highlight every occurrence of the search term's first letter in the spec

[07:39:01.0000] <Lachy>
Philip`, are you using Opera?

[07:39:02.0000] <Philip`>
Yes

[07:39:03.0000] <Philip`>
(9.2)

[07:40:00.0000] <Philip`>
(It's much faster in 9.5)

[07:40:01.0000] <Lachy>
/me tries it in 9.5

[07:43:00.0000] <Lachy>
weird, Opera's find-as-you-type feature seems to be skipping the TOC in the spec

[07:44:00.0000] <Dashiva>
Not for me (9.5)

[07:44:01.0000] <Lachy>
e.g. if I type "the body", in Firefox I get "The body element" link in the TOC highlighted, so I can just press enter to go there. In opera, it takes me to elsewhere in the spec first.

[07:45:00.0000] <Philip`>
"the body" works for me in 9.2, fails in 9.5 (1933)

[07:46:00.0000] <Lachy>
that's could be a known bug, I will check

[07:46:01.0000] <Philip`>
I don't think it's skipping the TOC, it's just skipping the nested elements

[07:47:00.0000] <Dashiva>
The <code>body</code>

[07:51:00.0000] <Lachy>
Dashiva, it seems to be more complcated than that.

[07:52:00.0000] <Lachy>
I'll file a bug and have someone in desktop analyse it.

[07:56:00.0000] <Lachy>
ah, it's to do with whitespace in the source code.

[08:18:00.0000] <MikeSmith>
hsivonen: you got your slides ready?

[08:18:01.0000] <MikeSmith>
deb says we can fit you in

[08:18:02.0000] <MikeSmith>
or if you don't have them ready, please do get them ready

[08:18:03.0000] <MikeSmith>
and send them to xtech-lightning⊙uc

[08:19:00.0000] <MikeSmith>
or at least the URL if you have them online

[08:23:00.0000] <hsivonen>
MikeSmith: yeah, practically ready

[08:27:00.0000] <MikeSmith>
hsivonen: thanks

[08:32:00.0000] <hsivonen>
MikeSmith: slides sent

[08:34:00.0000] <MikeSmith>
hsivonen: thanks

[09:11:00.0000] <zcorpan_>
Hixie: what does "Draft Recommendation" mean?

[09:21:00.0000] <Philip`>
All specs start as Not Working, then go to Working Draft then Draft Recommendation then Recommendation and then they're obsolete and you can start again

[09:21:01.0000] <Philip`>
It's a nice smooth cycle

[09:36:00.0000] <zcorpan_>
Philip`: i'd expect a next Not Working or Working Draft be started during Draft Recommendation stage, if those are the stages

[09:42:00.0000] <hsivonen>
anne is about to speak

[09:42:01.0000] <hsivonen>
Steven Pemberton's talk inspired a lot of commentary from the floor

[09:54:00.0000] <MikeSmith>
/me smiles at Philip` description of spec lifecycle

[09:57:00.0000] <hsivonen>
anne tells us that OPTIONS is now supported everywhere

[09:58:00.0000] <Philip`>
/me wonders what "everywhere" means

[10:06:00.0000] <MikeSmith>
Doug Crockford commenting on Anne's presentation about Access-Control spec: "It should be abandoned"

[10:06:01.0000] <Philip`>
Do I get a prize for guessing what it should be replaced with?

[10:07:00.0000] <MikeSmith>
"cookies have always been bad... I'm not blaming you for cookies"

[10:09:00.0000] <MikeSmith>
"we need to create a new class of service, so those are going to need to have new APIs [for cross-site requests]"

[10:10:00.0000] <MikeSmith>
chaals: You might only care about enabling Ajax apps only, but we have to care about the rest of the Web as well [e.g., XBL, XSLT, etc.]

[10:11:00.0000] <Lachy>
Is Crockford promoting JSONRequest again?

[10:11:01.0000] <MikeSmith>
Lachy: no, actually

[10:11:02.0000] <Lachy>
wow

[10:12:00.0000] <Lachy>
is it JSONRequest2 this time?

[10:12:01.0000] <MikeSmith>
he was just commenting in general about what problems he thinks need to be solved and how they should be solved

[10:12:02.0000] <Lachy>
I wish I was there

[10:12:03.0000] <Lachy>
is there an audio stream available?

[10:12:04.0000] <MikeSmith>
nope

[10:13:00.0000] <zcorpan_>
/me appriciates the scribing.. :)

[10:13:01.0000] <MikeSmith>
he's making some good points here

[10:13:02.0000] <Lachy>
I blame chaals for not sending me

[10:13:03.0000] <MikeSmith>
he says he is encouraged by the cross-document messaging spec

[10:15:00.0000] <MikeSmith>
arve now challenging him on why he says iframe communication is better than cross-site request mechanisms

[10:15:01.0000] <MikeSmith>
crockford said it is absolutely better because it provides better encapsulation

[10:15:02.0000] <Lachy>
does anyone have skype? If so, call me so I can listen to it

[10:16:00.0000] <MikeSmith>
he says the frustration we are all dealing with is that the current state of the art does not permit it to build the kinds of apps we want to build

[10:16:01.0000] <MikeSmith>
Lachy: what's your Skype handle?

[10:16:02.0000] <Lachy>
lachlanhunt

[10:17:00.0000] <gsnedders>
How very imaginative.

[10:17:01.0000] <Lachy>
:-)

[10:17:02.0000] <MikeSmith>
hsivonen talking now

[10:18:00.0000] <MikeSmith>
Lachy: me is sideshowbarker

[10:18:01.0000] <MikeSmith>
tried to call you but got refused

[10:18:02.0000] <MikeSmith>
call me

[10:18:03.0000] <zcorpan_>
can i call too? :)

[10:18:04.0000] <Lachy>
I can't hear anything

[10:19:00.0000] <MikeSmith>
Lachy: we are in a big room

[10:19:01.0000] <MikeSmith>
sorry

[10:19:02.0000] <Lachy>
can you turn up your mic volume?

[10:19:03.0000] <MikeSmith>
Lachy: I can try

[10:19:04.0000] <Lachy>
it just got a bit louder then

[10:20:00.0000] <Lachy>
till you held the call

[10:20:01.0000] <zcorpan_>
/me hears something

[10:20:02.0000] <Lachy>
MikeSmith, can you make it a conf call?

[10:20:03.0000] <MikeSmith>
I could if I knew how

[10:21:00.0000] <Lachy>
you should be able to right click on my name and say add to call

[10:21:01.0000] <Lachy>
or something like that

[10:22:00.0000] <Lachy>
there's a lot of noise

[10:23:00.0000] <Lachy>
ok, I can hear it well now

[10:23:01.0000] <MikeSmith>
that's because I'm sitting right next to anne

[10:23:02.0000] <MikeSmith>
so you're hearing anne

[10:24:00.0000] <MikeSmith>
hsivonen talking now

[10:37:00.0000] <Lachy>
MikeSmith, are there any more talks tonight, or was that the last?

[10:44:00.0000] <annevk>
in Microsoft talk

[10:45:00.0000] <annevk>
silverlight brings the web and desktop together or something

[10:46:00.0000] <MikeSmith>
Lachy: I can Skype you in to Silverlight talk if you want

[10:47:00.0000] <MikeSmith>
at 7pm local time, we do lightnight talks

[10:47:01.0000] <MikeSmith>
lightning

[10:47:02.0000] <MikeSmith>
.NET is a completely new way to do application design and development

[10:47:03.0000] <Philip`>
If a speaker is unpopular, do they get electrocuted?

[10:48:00.0000] <MikeSmith>
Philip`: no, they get the gong

[10:48:01.0000] <MikeSmith>
gong show

[10:49:00.0000] <Lachy>
can you ask them to give the Microsoft speaker the gong? No-one wants to hear about silverlight. That's a dead technology.

[10:50:00.0000] <Lachy>
unless he's talking about why it failed.

[10:53:00.0000] <hsivonen>
Lachy: MS is paying for the privilige to give this pitch

[10:53:01.0000] <annevk>
did he just say? as you can see we have lots of bugs?

[10:53:02.0000] <annevk>
s/say?/say:/

[10:53:03.0000] <annevk>
s/bugs?/bugs./

[10:54:00.0000] <hsivonen>
supporting many browsers costs more than only doing silverlight

[10:54:01.0000] <hsivonen>
allegedly

[10:54:02.0000] <Philip`>
/me can usually support three browsers for quite little cost

[10:54:03.0000] <hsivonen>
now pitching their debugger over Firebug

[10:55:00.0000] <annevk>
ooh wait

[10:55:01.0000] <annevk>
he's saying how complex the web platform is, clearly i should pay more attention

[10:55:02.0000] <hsivonen>
the point seems to be that Debugging in IE.old suck compare to silverlight

[10:56:00.0000] <annevk>
he's now talking "blah blah blah" and hsivonen and i are confused :)

[10:56:01.0000] <hsivonen>
did he just say view source is a problem?

[10:57:00.0000] <annevk>
"it's pretty, look"

[10:59:00.0000] <Philip`>
/me wonders if anyone else in the audience is being more convinced by the presentation

[10:59:01.0000] <hsivonen>
"no portability issue"

[10:59:02.0000] <Lachy>
how long is each lightning talk?

[10:59:03.0000] <annevk>
"easier than javascript, 1/3 development time, more usable, obviously"

[11:00:00.0000] <annevk>
"deployment issues, Flash had that too"

[11:00:01.0000] <hsivonen>
Lachy: 20 times 20 seconds

[11:01:00.0000] <annevk>
"Microsoft is not about taking over Flash, ??, whatever"

[11:01:01.0000] <hsivonen>
"it's not about taking over Flash"

[11:01:02.0000] <annevk>
"It's for when you use .NET already and make desktop applications?" ???

[11:01:03.0000] <annevk>
/me wonders how that relates to the JavaScript pitch

[11:04:00.0000] <MikeSmith>
Lachy: each 20x20 talk is 6 minutes 40 seconds total (maximum)

[11:15:00.0000] <Lachy>
MikeSmith, I couldn't hear it very well at all. So it'd be easier for me to just read whatever you can write in here

[12:26:00.0000] <Lachy>
anne's talk was quite good. I could hear him quite clearly (unlike previous speakers)

[12:27:00.0000] <annevk>
thanks

[12:27:01.0000] <annevk>
it was totally unprepared

[12:27:02.0000] <annevk>
except for the slides :)

[12:29:00.0000] <annevk>
now is RDFa

[12:29:01.0000] <annevk>
next is Henri again on HTML5

[12:52:00.0000] <annevk>
my presentation is here: http://annevankesteren.nl/2008/xtech-html5 (requires support for media=projection and <img src=svg>

[12:52:01.0000] <annevk>
)

[12:55:00.0000] <hsivonen>
http://hsivonen.iki.fi/xtech2008/

[13:39:00.0000] <Hixie>
ISSUE-41 seems to make two assumptions that I believe to be false:

[13:41:00.0000] <Hixie>
1. that making html into a generic syntax is desireable, or that having generic syntaxes available for web authors to extend with custom vocabularies arbitrarily is desireable, or that there will be enough new standard vocabularies that the cost of adding new syntax each time is an important factor in the deployment of a new language

[13:42:00.0000] <Hixie>
2. that it is possible to extend html's parsing model in a generic way without running into compatibility problems with legacy content or with the copy-and-paste mentality of authors while the new syntax is being deployed

[13:47:00.0000] <Hixie>
i guess there's no point trying to raise these points (which I've already raised) before the tag gets to do its thing

[14:48:00.0000] <Hixie>
hahahahaha

[14:48:01.0000] <Hixie>
i just remembered

[14:48:02.0000] <Hixie>
we're in 2008 Q2

[14:48:03.0000] <Hixie>
it's time to publish a LC!

[14:49:00.0000] <gsnedders>
FPWD to LC. Fun!

[15:07:00.0000] <tommorris_>
is my brain operating correctly, or is there a teleconference today?

[15:08:00.0000] <tommorris_>
ah, yes, midnight london time is the cal

[15:08:01.0000] <tommorris_>
*call

[16:29:00.0000] <Hixie>
i was going to deal with http://www.w3.org/mid/tkrat.c998060f76b1e71c⊙gn (which i just got to in my pile of feedback)

[16:29:01.0000] <Hixie>
but i disagree with almost every single point

[16:30:00.0000] <Hixie>
mcarter: did you ever send your feedback on TCPConnection?

[16:35:00.0000] <mcarter>
Hixie, I have an initial draft, though I was trying to get a working demonstration first though

[16:36:00.0000] <Hixie>
k

[16:37:00.0000] <Hixie>
i'll wait til you've sent your feedback before looking at that section, so as to not make mistakes you have already covered in your feedback :-)

[16:37:01.0000] <mcarter>
You should be able to expect javascript+proxy daemon that will speak the current protocol, and a proposed protocol by the end of next week

[16:37:02.0000] <Hixie>
cool

[16:37:03.0000] <Hixie>
no rush

[16:38:00.0000] <mcarter>
I'm pretty excited about this stuff actually

[16:38:01.0000] <mcarter>
Its not so much that I'm rushing, rather all my time is going into it these days

[16:38:02.0000] <Hixie>
:-)

[16:55:00.0000] <aroben>
/me kind of liked DOMStrong

[16:55:01.0000] <Hixie>
he

[16:55:02.0000] <Hixie>
h


2008-05-09
[17:20:00.0000] <Hixie>
hsivonen: not sure what to do about http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2007-July/012131.html

[17:59:00.0000] <Lachy>
Hixie, yt?

[18:00:00.0000] <Lachy>
Typo in the definition for the sizes attribute "If the specified, the attribute..."

[18:05:00.0000] <Hixie>
Lachy: fixed

[18:08:00.0000] <Lachy>
Hixie, why have you included 8192x8192 and 32768x32768 in the example? Aren't they a bit excessive?

[18:09:00.0000] <Hixie>
man, nobody understands my sense of humour :-/

[22:31:00.0000] <MikeSmith>
heh

[22:31:01.0000] <MikeSmith>
comment on Mark Pilgrim's blog:

[22:32:00.0000] <MikeSmith>
commenter: I can’t help but think what an arrogant a$$ you’d have to be

[22:32:01.0000] <MikeSmith>
Mark replies: You must be new here.

[22:32:02.0000] <MikeSmith>
http://diveintomark.org/archives/2008/05/07/when-the-fall-is-all-thats-left#comment-12014

[22:34:00.0000] <Hixie>
heh

[22:41:00.0000] <othermaciej>
I got into a stupid flamewar on reddit over Mark's post

[22:41:01.0000] <othermaciej>
I am embarassed for myself

[22:43:00.0000] <othermaciej>
hey roc

[22:43:01.0000] <roc>
hey

[23:06:00.0000] <Hixie>
othermaciej: uri? :-)

[23:06:01.0000] <othermaciej>
Hixie: I'm too embarassed

[23:08:00.0000] <Hixie>
heh

[23:09:00.0000] <othermaciej>
ok if you really want to see it it's here: http://reddit.com/info/6igdz/comments/ but the thread is downvoted so it's a little buried

[23:20:00.0000] <roc>
ok, now I have to guess which one of these people is maciej. that should be fun

[23:20:01.0000] <othermaciej>
my reddit nick is obvious

[23:20:02.0000] <doublec>
"That's like telling me to go ahead and rationalize away being so smart and good looking.:

[23:20:03.0000] <doublec>
hehe

[23:22:00.0000] <othermaciej>
I'm not entirely embarassed about that part

[23:22:01.0000] <roc>
oh, right. I'm just a reddit noob

[23:23:00.0000] <othermaciej>
just resorting to "do you know who I am?!?!?" at the end

[23:23:01.0000] <doublec>
well, he did ask

[23:24:00.0000] <doublec>
he didn't appear to believe you though - unless he really is the ff team lead. Somehow I doubt it.

[23:29:00.0000] <jwalden>
I, for one, welcome our new ff team overlord

[23:31:00.0000] <Hixie>
wow, that guy really was more of a jerk

[23:32:00.0000] <Hixie>
i love how he didn't reply :-)

[23:35:00.0000] <roc>
our Gecko team isn't that much bigger, actually

[00:16:00.0000] <Hixie>
hsivonen: i don't really understand what i should do to address http://www.w3.org/mid/01575703-3A06-4F51-BE27-86A9EBB44C54⊙if

[00:16:01.0000] <Hixie>
hsivonen: in particular, i don't know what to say about xmlns that isn't either obvious, or out of scope

[00:50:00.0000] <MikeSmith>
Hixie: r1577 typo : "keywods"

[00:50:01.0000] <Hixie>
thanks, in the queue

[00:51:00.0000] <MikeSmith>
and comment: might it better be, "User agents must process the links on a per-*keyword* basis" (instead of "per-link" basis) ?

[00:53:00.0000] <Hixie>
nah, it's per-link, not per keyword

[00:53:01.0000] <Hixie>
e.g. link rel="alternate stylesheet" is one link

[00:53:02.0000] <Hixie>
and <link rel="up up icon"> is two links

[00:54:00.0000] <MikeSmith>
ah, OK

[01:12:00.0000] <Hixie>
MikeSmith: i agree entirely with your mail... so... how do we get a changelog? :-)

[01:14:00.0000] <annevk>
morning

[01:15:00.0000] <Hixie>
heya

[01:15:01.0000] <Hixie>
how's xtech?

[01:15:02.0000] <Hixie>
i haven't seen any controversial blog posts about any hot happenings yet

[01:19:00.0000] <MikeSmith>
Hixie: we get it by somebody taking time to sit down and write it up

[01:20:00.0000] <MikeSmith>
which I can help with (or do alone if I have to) starting this weekend (when I get back home to tokyo from dublin) and first part of next week

[01:20:01.0000] <Hixie>
the working group hasn't had much luck waiting for things to happen that way

[01:20:02.0000] <MikeSmith>
yeah

[01:21:00.0000] <MikeSmith>
which is why I think to make sure it gets done I will probably need to be the one to write it up

[01:23:00.0000] <MikeSmith>
unless some magic shoemaker's elves or something sneak in and do it first

[01:23:01.0000] <annevk>
Hixie, crockford attended my presentation which gave some discussion

[01:23:02.0000] <Hixie>
annevk: ooh, do tell

[01:23:03.0000] <annevk>
the lightning talks were pretty funny

[01:24:00.0000] <annevk>
but there was no real new information :(

[01:24:01.0000] <MikeSmith>
Hixie: about XTech, Doug Crockford was at annevk session on Access Control spec, and we had a 15-minute discussion at the end of that

[01:24:02.0000] <Hixie>
what was there to discuss?

[01:24:03.0000] <MikeSmith>
er, what annevk just said :)

[01:24:04.0000] <MikeSmith>
/me types too slow

[01:24:05.0000] <annevk>
his claims that the spec should be abondoned

[01:24:06.0000] <annevk>
(art explicitly asked for his feedback at the end of the talk)

[01:24:07.0000] <MikeSmith>
yeah, "abandoned" was his exact word

[01:25:00.0000] <MikeSmith>
though he had great things to say about the cross-document messaging spec

[01:25:01.0000] <annevk>
also that we should solve everything with it and that lots of complexity for people like hsivonen is acceptable (hsivonen explained how easy Access Control was to implement for him)

[01:27:00.0000] <annevk>
other things: lightning talk introducing HTML5 concept from me and on HTML5 parsing from hsivonen

[01:27:01.0000] <annevk>
talk on RDFa that showed quite complex examples (the funny bit was that the namespace declarations were not shown which would've made it near impossible to grasp)

[01:28:00.0000] <Hixie>
sounds like the same thing as 2005

[01:28:01.0000] <Hixie>
glad i didn't go

[01:35:00.0000] <annevk>
it was mostly meeting the other people that made it good

[01:36:00.0000] <annevk>
though some of the stuff was interesting too, such as what SilverLight is doing, how OAuth works, that OpenID 2 is incompatible with OpenID 1, etc.

[01:38:00.0000] <Hixie>
what's silverlight doing?

[01:39:00.0000] <annevk>
it was quite confused, but the interface stuff they're doing looks quite fancy

[01:39:01.0000] <annevk>
they're far ahead in ready components and all though libraries will solve that problem for us I guess

[01:40:00.0000] <annevk>
he also advocated against cross-browser scripting and all quite a bit

[01:41:00.0000] <annevk>
quite aggresively, I suppose, though he also said it wasn't about replacing Flash (yeah right)

[02:02:00.0000] <Hixie>
annevk: it's about replacing html, not flash

[02:04:00.0000] <othermaciej>
a little of both I think

[02:05:00.0000] <Hixie>
flash has like 2% of the content market tops

[02:05:01.0000] <Hixie>
they're doing a platform play, to replace their windows monopoly -- that means 98%, not 2%.

[02:06:00.0000] <roc>
I think it's both

[02:06:01.0000] <roc>
flash isn't dominant now, but they're afraid it could be

[02:07:00.0000] <Hixie>
oh i'm sure they're not ignoring it

[02:07:01.0000] <Hixie>
i just mean that that isn't their primary goal

[03:00:00.0000] <Lachy>
/me wonders how many sites actually use silverlight at all?

[03:01:00.0000] <Lachy>
I've only ever seen it used on microsoft.com, and I refused to install the plugin

[03:12:00.0000] <gsnedders>
writing unambiguous ABNF is hard.

[03:12:01.0000] <gsnedders>
(and the damned signal is broken so I'm stuck nowhere on a train)

[03:12:02.0000] <gsnedders>
(and we literally just started moving!)

[03:15:00.0000] <Philip`>
It ought to be possible to get tools that check ABNF for ambiguities

[03:15:01.0000] <Philip`>
No idea if such things exist, though

[03:15:02.0000] <gsnedders>
I don't think so, sadly

[03:16:00.0000] <Philip`>
You could write it in a real parser-generator language, and have that parser-generator check for ambiguity, then manually translate it into ABNF

[03:17:00.0000] <gsnedders>
I guess, but I'm lazy :)

[03:18:00.0000] <gsnedders>
(train stuck again, still not made it to Inverkeithing)

[03:18:01.0000] <gsnedders>
(hopefully will make it to Cam. by 16:00 though)

[03:18:02.0000] <gsnedders>
(certainly before I'm meant to meet you and jgraham_ though)

[03:23:00.0000] <othermaciej>
yacc or bison makes it pretty straightforward to write a checked grammar

[03:23:01.0000] <othermaciej>
(just make all the actions empty)

[03:24:00.0000] <Philip`>
http://ejohn.org/blog/processingjs/

[03:24:01.0000] <gsnedders>
I'm not sure if I can bothered to learn another language, though

[03:24:02.0000] <gsnedders>
yacc is similar to BNF, no?

[03:27:00.0000] <Philip`>
/me sees that olliej (I think) already complained about how it uses ImageData non-conformingly

[03:27:01.0000] <Philip`>
(and also in a way that won't work if there's !=1 device pixel per canvas pixel)

[03:29:00.0000] <gsnedders>
/me wonders if anything actually uses the CRLF allowed in LWS in HTTP/1.1

[03:30:00.0000] <gsnedders>
/me has only ever seen it in test-cases

[03:30:01.0000] <gsnedders>
(i.e., do I actually have to allow LF there?)

[03:50:00.0000] <gsnedders>
http://pastebin.ca/1012564 — header-value is what was causing me issues at making sure it was unambiguous

[03:51:00.0000] <gsnedders>
anybody able to check that?

[03:53:00.0000] <gsnedders>
(no, I don't trust myself)

[03:55:00.0000] <Philip`>
What does "unambiguous" mean?

[03:56:00.0000] <gsnedders>
free of anything that is unclear

[03:56:01.0000] <gsnedders>
i.e., it is completely clear

[03:56:02.0000] <Philip`>
/me hopes it just means there's a single parse tree for any input, else he's confused

[03:57:00.0000] <gsnedders>
well, in the case of ABNF, there's no character which could match more than one production where it is in the string

[03:58:00.0000] <gsnedders>
Philip`: I can give you an English lesson tonight, if you want :P

[03:58:01.0000] <othermaciej>
in the case of grammars, "ambiguous" means something more specific than clear/unclear

[03:59:00.0000] <othermaciej>
it means there's only one way to parse

[03:59:01.0000] <othermaciej>
if you use an ABNF-like LALR(1) parser generator, it will report shift-reduce or reduce-reduce conflicts in such cases

[03:59:02.0000] <othermaciej>
though it is possible the grammar isn't ambiguous but you have just hit a limitation of LALR(1)

[03:59:03.0000] <gsnedders>
/me sucks at defining English

[03:59:04.0000] <othermaciej>
still, it is a pretty good approximation

[04:00:00.0000] <othermaciej>
I do think yacc/bison would make for a good way to test a grammar

[04:00:01.0000] <gsnedders>
othermaciej: I have enough to do without learning even more :)

[04:00:02.0000] <gsnedders>
(languages, that is)

[04:00:03.0000] <othermaciej>
this grammar might be simple enough for lex to do it

[04:01:00.0000] <gsnedders>
/me has an English exam next week, and a French one the week after

[04:02:00.0000] <gsnedders>
othermaciej: Where do you come from, with a name like Maciej, BTW?

[04:02:01.0000] <othermaciej>
gsnedders: straight outta hell

[04:03:00.0000] <gsnedders>
othermaciej: Ah. OK.

[04:03:01.0000] <gsnedders>
(No, seriously?)

[04:04:00.0000] <gsnedders>
Doing it in ABNF gives me a reason to write an ABNF checker, that checks for ambiguities, though

[04:04:01.0000] <gsnedders>
Not that I'll bother, probably :P

[04:04:02.0000] <othermaciej>
yacc is just a tool

[04:05:00.0000] <othermaciej>
learning its (trivial) syntax is almost certainly easier than writing a program to parse a different syntax and do the same checks

[04:05:01.0000] <gsnedders>
yacc is just another tool you have to learn how to use :)

[04:06:00.0000] <Philip`>
Learning how to use a tool is much easier than making your own, and often it's easier than doing your work without the tool

[04:06:01.0000] <Philip`>
Also it's a good opportunity for procrastination

[04:07:00.0000] <Philip`>
/me has spent a lot of time playing with LaTeX when he's meant to be writing reports

[04:10:00.0000] <gsnedders>
That's true. I just end up playing around with type-setting maths in LaTeX :P

[04:13:00.0000] <gsnedders>
/me wonders how to do this without being ambiguous

[04:15:00.0000] <gsnedders>
I need to properly parse headers, while still allowing garbage headers

[04:16:00.0000] <Philip`>
/me doesn't notice any obvious ambiguity in the 'header' thing

[04:17:00.0000] <gsnedders>
Philip`: In what I linked? Good.

[04:17:01.0000] <gsnedders>
Now to make it far more complex and more likely wrong.

[04:18:00.0000] <gsnedders>
I want something like `invalid-header = *( header-content / lws ) [CR] LF`, but that is totally ambiguous

[04:18:01.0000] <gsnedders>
(combined with the valid header)

[04:22:00.0000] <gsnedders>
("If a strict parser is in use, and anything is matched by the invalid-header rule, it is a fatal error.")

[04:22:01.0000] <Philip`>
ABNF is not a good way to do this :-p

[04:22:02.0000] <gsnedders>
It is mostly totally fine.

[04:23:00.0000] <gsnedders>
There are very few quirks like that

[04:23:01.0000] <Philip`>
Mostly, except for all the error handling

[04:23:02.0000] <gsnedders>
There's very little error handling at a parser level :)

[04:23:03.0000] <gsnedders>
It's mostly a case of either it is totally invalid, which is always a fatal error, or has a garbage as a header, which is a fatal error only in strict parsers

[04:24:00.0000] <gsnedders>
Once you start doing header specific stuff, then you get real quirks. But I'm not using ABNF for that.

[04:24:01.0000] <Philip`>
Seems easiest to say something like "If a prefix of the string matches 'header', then process it like a header. Otherwise, fatal error if strict, else skip to the next line"

[04:24:02.0000] <Philip`>
rather than trying to work out how to make a grammar rule that's a complement of another

[04:25:00.0000] <gsnedders>
It's actually really horrible to do using prose. I tried doing parts of it using prose, and it was far harder

[04:25:01.0000] <gsnedders>
I almost ended up using POSIX regex, as that not only has a format specification, but has negation, which ABNF does not

[04:26:00.0000] <Philip`>
Normal regexps only give you a boolean output, which isn't very useful when you want to get an HTTP header out of it

[04:27:00.0000] <gsnedders>
"Taking all matches and submatches of the header rule…"

[04:27:01.0000] <gsnedders>
or the "named header submatch"

[04:27:02.0000] <gsnedders>
It's certainly possible to do

[04:27:03.0000] <Philip`>
Just invent your own string-parsing grammar formalism

[04:28:00.0000] <gsnedders>
It's nicer to use ABNF as then I can just say, "This specification inherits all the rules from RFC3986, which as they are already given in ABNF they are not repeated here."

[04:29:00.0000] <Philip`>
/me goes away for a while

[04:29:01.0000] <gsnedders>
See you at 7, then

[04:29:02.0000] <Philip`>
Or 7:05 if I'm late

[04:30:00.0000] <Philip`>
It should only take me 24 minutes to get there, but I'm never very good at departing on time

[04:33:00.0000] <gsnedders>
Philip`: As long as you recognise me and James :P

[04:34:00.0000] <gsnedders>
Philip`: And I don't have a clue how long it takes me to get there, as I don't know where I'm coming from, and I don't know how long it takes to get around Cambridge

[04:44:00.0000] <gsnedders>
ABNF is too vague. It doesn't specify what to do if multiple alternations match a given input

[06:52:00.0000] <zcorpan>
Hixie: how about "HTML <code>Document</code>" and "XML <code>Document</code>"?

[06:54:00.0000] <zcorpan>
Hixie: that should make it clearer that it's about the dom and not a strem of bytes

[06:54:01.0000] <zcorpan>
s/strem/stream/

[08:59:00.0000] <Hixie>
zcorpan: yeah maybe

[09:05:00.0000] <Hixie>
good to see the accessibility experts dismiss the problems hit by people using ATs...

[09:07:00.0000] <Lachy>
Hixie, are you referring to this? http://www.paciellogroup.com/blog/misc/uc/

[09:07:01.0000] <Lachy>
/me is just reading it now

[09:14:00.0000] <Philip`>
/me assumes it's a reference to http://lists.w3.org/Archives/Public/www-archive/2008May/0016.html

[09:21:00.0000] <Lachy>
In the section Complex Data Images, in their proposal, it seems that they're using alt for text that would be more appropriate for either title="" or as an image caption.

[09:25:00.0000] <Lachy>
and it seems they didn't figure out what the example from the spec "Bubbles traveled everywhere with us" is actually a picture of. But I guess, it's a bit obscure from the caption.

[09:26:00.0000] <Lachy>
http://flickr.com/photos/carinda/1100670787/ Clearly, bubbles is a cat :-)

[09:27:00.0000] <Philip`>
http://lists.w3.org/Archives/Public/public-html/2008May/0079.html sounds like a reasonable argument that empty alt="" (as the spec requires) in those cases is bad, since it removes the distinction between purely decorative images (which you can happily ignore) and informative images (which you might want to see as text by default, but go to some extra effort to see when you decide the image is interesting)

[09:32:00.0000] <Lachy>
For those cases, I still like the kind="" attribute I suggested a few days ago, where for the example in the email, it would be: <img src="rendering-mode-pie-chart.png" kind="Pie Chart" alt="">

[09:33:00.0000] <Lachy>
I don't like the importantimage attribute, since it's likely to suffer from copy and paste errors, and be included where it shouldn't, and it's harder to detect whether or not it should be there.

[09:33:01.0000] <Philip`>
Why is that better than <img alt="Pie Chart" importantimage>? It's worse because it hides the important image entirely from current UAs, and I'm not sure (/can't remember) what makes it better overall

[09:33:02.0000] <Philip`>
Ah

[09:34:00.0000] <Philip`>
importantimage won't suffer from copy and paste errors because nobody will ever bother using it and so it won't be copied :-)

[09:34:01.0000] <Lachy>
with an attribute that says a) what the image is and b) why it's important, errors are more likely to be detected, and it would be easy to analyse with hsivonen's image report tool.

[09:35:00.0000] <Philip`>
With the importantimage thing, the alt attribute says (in combination with its human-understable context) what the image is and why it's important

[09:36:00.0000] <Lachy>
but it also pretends that the text given is its alternative, when it isn't.

[09:36:01.0000] <Lachy>
it's entirely possible to have an important image with a good alt text supplied

[09:36:02.0000] <Philip`>
It doesn't pretend the text is an alternative, if the spec doesn't say that the alt text is necessarily an alternative

[09:37:00.0000] <Lachy>
yes it does

[09:37:01.0000] <Lachy>
that's the whole point of it.

[09:37:02.0000] <Lachy>
also, importantimage is a really bad name

[09:37:03.0000] <Philip`>
The point is to help users, and users might be better helped if alt text isn't always a strict alternative

[09:38:00.0000] <Philip`>
I agree that particular syntax isn't much good

[09:39:00.0000] <Lachy>
importantimage suffers from many of the same problems as noalt

[09:39:01.0000] <Lachy>
it's basically the same thing with a different name

[09:40:00.0000] <Philip`>
importantimage images will still have alt text, so non-hyper-intelligent UAs won't be totally stuck when trying to present the image textually

[09:41:00.0000] <Philip`>
and people won't stick in importantimage to make the validator shut up, because their pages will be equally machine-checkably-valid regardless of whether they have that attribute

[09:41:01.0000] <Lachy>
kind="" also gives additional information about the picture, which may not necessarily come from the alt text. http://krijnhoetmer.nl/irc-logs/whatwg/20080506#l-162

[09:42:00.0000] <Philip`>
alt="A painting of a white house with ..."

[09:42:01.0000] <Philip`>
What are UAs meant to do with 'kind'?

[09:42:02.0000] <Lachy>
same thing they're supposed to do with importantimage, but also say what kind of image it is

[09:46:00.0000] <Philip`>
If all they're doing is serialising the alt and kind values and some connecting words into a text stream (like "Image type: Painting. The house is ..."), why not have authors put that whole serialised text into the alt attribute (alt="Image type: Painting. The ...") instead of adding the complexity of an extra attribute?

[09:48:00.0000] <Lachy>
a better example would be the flowchart one from the spec:

[09:48:01.0000] <Lachy>
<img src="images/parsing-model-overview.png" alt="The network passes data to the Tokeniser stage..." kind="Flowchart">.  It would seem wrong to say alt="A flowchart illustrating how the network...".

[09:49:00.0000] <Lachy>
the UA doesn't need to read out the kind by default if there is alt text supplied. Only if the user asks for it.

[09:50:00.0000] <Philip`>
Why would the user want to know that the image a flowchart?

[09:50:01.0000] <Philip`>
Why would the user explicitly ask for it when it's going to say "no 'kind' was specified" for pretty much 100% of images?

[09:51:00.0000] <Philip`>
(I think your argument makes sense, but I'm not convinced yet :-) )

[09:52:00.0000] <Lachy>
it depends how the UA indicates its presence, or what kind of UA it is

[09:53:00.0000] <Philip`>
It seems it would be useful to work from concrete examples of what some UAs could/would do, rather than just presenting some metadata and hoping someone else will figure out what to do with it

[09:54:00.0000] <Lachy>
for a screen reader, it could use any aural indicator to to indicate the presence of more information (like, e.g. a sound effect in the background as it's reading the alt text (probably a bell or tone of some kind)

[09:55:00.0000] <Lachy>
for a text only UA like Lynx, it could include it by default in the textual output somehow.  e.g. [Image (Flowchart): The network....]

[09:58:00.0000] <Lachy>
I think screen readers should have done something similar things with other hidden metadata like longdesc. IIRC, users have to ask for it without knowing it's there in current implementations.

[10:00:00.0000] <Philip`>
Why would users of screen readers or Lynx want to know it's a flowchart?

[10:02:00.0000] <Lachy>
to help them decide if it's worth downloading and opening in external viewer, espcially if they want to zoom in on it to see it better

[10:07:00.0000] <Philip`>
I'm not sure why that would help them decide in a non-negligible number of non-contrived cases - I expect they would care about the information presented by the image (which they can find from the alt attribute) when deciding whether to look at it more closely, not about whether it's presented as a flowchart or a photo or a cartoon

[10:09:00.0000] <Philip`>
and it might be a photo of a flowchart, and the author will spend lots of time deciding what an appropriate 'kind' is, when actually it's a waste of time because it won't help people enough to be worthwhile

[10:11:00.0000] <Lachy>
it doesn't have to be used all the time

[16:38:00.0000] <Philip`>
/me 's hands are cold :-(

[16:39:00.0000] <jgraham>
Heh

[16:55:00.0000] <jwalden>
the latest opera beta didn't have new-style postMessage support, did it?


2008-05-10
[17:18:00.0000] <hsivonen>
I was reading about externality taxing in the CO2 context on the plane

[17:19:00.0000] <hsivonen>
the economics of it seem applicable to inaccessible Web pages

[17:19:01.0000] <hsivonen>
except tracking tax on those would be more difficult

[17:21:00.0000] <jwalden>
"on the plane"?

[17:21:01.0000] <jwalden>
er

[17:21:02.0000] <hsivonen>
jwalden: while flying back from Dublin

[17:21:03.0000] <jwalden>
nm me

[17:21:04.0000] <jwalden>
misparsing

[17:21:05.0000] <jwalden>
was binding "the CO2 context on the plane" as one unit

[17:22:00.0000] <hsivonen>
/me notes that Tim Harford writes about "moral posturing" in the absence of externality taxing

[17:23:00.0000] <hsivonen>
threatening people with lawsuits is a kind of an externality tax

[17:23:01.0000] <hsivonen>
but an uncertain one that applies only to big companies

[17:23:02.0000] <jwalden>
a hideously less efficient one, to be sure

[17:23:03.0000] <jwalden>
courts and lawsuits epitomize deadweight loss

[17:23:04.0000] <hsivonen>
yeah

[17:24:00.0000] <hsivonen>
I'm not a fan of lawsuit threatening as part of the accessibility advocate argument set

[17:24:01.0000] <othermaciej>
I think it would be hard to justify a large tax on inaccessible web pages on a Pigovian basis

[17:25:00.0000] <othermaciej>
in fact pure economic analysis might find the cost of making web pages accesible exceeds the likely economic benefit

[17:25:01.0000] <othermaciej>
I'm also not sure lack of accessibility can be modeled as a negative externality

[17:25:02.0000] <othermaciej>
rather, it is failure to provide a positive good to some people

[17:26:00.0000] <othermaciej>
pollution is a negative externality

[17:26:01.0000] <othermaciej>
giving free candy to some people but not others isn't

[17:26:02.0000] <hsivonen>
othermaciej: I agree that presence should be considered a positive thing, but you can't start subsidising anyone with an inane Web site for making it accessible

[17:28:00.0000] <othermaciej>
current approaches to accesibility in law involve neither subsidy nor tax

[17:28:01.0000] <othermaciej>
rather they rely on legal notions of public accomodation, and impose regulations on anything considered such

[17:29:00.0000] <jwalden>
/me is reasonably happy with how that's fallen out so far

[17:29:01.0000] <othermaciej>
well, it does create unintended consequences

[17:29:02.0000] <hsivonen>
I thought in the U.S. they relied on punitive damages acting as a probabilistic externality tex

[17:29:03.0000] <hsivonen>
tax

[17:29:04.0000] <jwalden>
I'm not sure what actually motivates orgs at the end of the day

[17:30:00.0000] <othermaciej>
hsivonen: the same could be said of most regulations

[17:30:01.0000] <jwalden>
government business is at least part of it

[17:30:02.0000] <hsivonen>
othermaciej: indeed

[17:30:03.0000] <othermaciej>
however the risk of damages does create unintended consequences

[17:30:04.0000] <othermaciej>
for instance, since the ADA passed, unemployment among the disabled is up

[17:30:05.0000] <hsivonen>
yeah, 508-based government sales are the carrot

[17:31:00.0000] <jwalden>
on the other hand, I seem to recall it was this channel that mentioned <http://www.nytimes.com/2008/01/20/magazine/20wwln-freak-t.html?_r=1&oref=slogin> once

[17:31:01.0000] <othermaciej>
government business seems like a better motivator

[17:31:02.0000] <gavin_>
hsivonen posted that link, I remember

[18:27:00.0000] <Hixie>
i love the guy who said that having <p> was bad because it wasn't machine checkable and proposed <div role=paragraph> as a solution

[18:27:01.0000] <othermaciej>
lol

[18:27:02.0000] <Lachy>
which mailing list was that on?

[18:29:00.0000] <hsivonen>
Lachy: public-html and wai-xtech

[18:29:01.0000] <Lachy>
found it

[18:31:00.0000] <Hixie>
vlad wants 'textStyle' instead of 'font', for consistency with <canvas> APIs instead of CSS

[18:31:01.0000] <Hixie>
opinions?

[18:31:02.0000] <Lachy>
I did some major restructuring of selectors api. It's now a lot more organised, but still have to fix up the nsresolver requirements

[18:32:00.0000] <Philip`>
Hixie: Makes sense to have 'text' in the name, for consistency with every other text-drawing property and method

[18:32:01.0000] <Philip`>
and also makes it clearer that it's a CSS style thing, not just a font name

[18:33:00.0000] <othermaciej>
Hixie: would the value of 'textStyle' be a font name?

[18:34:00.0000] <Hixie>
othermaciej: it would be the css 'font' property

[18:34:01.0000] <othermaciej>
I think it should be called 'font'

[18:34:02.0000] <othermaciej>
most graphics APIs describe the way to draw text with a "font", not a "text style"

[18:35:00.0000] <othermaciej>
whereas they do have notions like "line style" or whatever

[18:35:01.0000] <othermaciej>
so I do not buy the consistency argument

[18:35:02.0000] <Hixie>
ok in that case please reply to vlad's e-mail :-)

[18:35:03.0000] <Philip`>
People seem to use CSS properties like 'font-family' far more than they use 'font', so the analogy between the canvas API and CSS doesn't really work since people aren't familiar with 'font'

[18:36:00.0000] <othermaciej>
right, I forgot that 'font' is a shorthand

[18:36:01.0000] <othermaciej>
it would be more appropriate to provide individual properties for the pieces of that shorthand, for canvas

[18:36:02.0000] <othermaciej>
IMO

[18:36:03.0000] <Hixie>
ew

[18:36:04.0000] <Hixie>
we don't do that with fillStyle

[18:37:00.0000] <Lachy>
yeah, I think it should stay as font, for consistency with CSSOM property

[18:37:01.0000] <othermaciej>
Hixie: but fillStyle is not a shorthand, it's just overloaded for different types

[18:37:02.0000] <Philip`>
/me imagines very few people use the font CSSOM property, so that doesn't seem a compelling reason

[18:39:00.0000] <othermaciej>
Hixie: I will have to do some study to give a cogent reply

[18:39:01.0000] <Lachy>
Philip`, it's fairly common to see people change styles by using element.style.xxx = ...; for a whole variety of styles.

[18:39:02.0000] <Lachy>
although I can't say how frequently each property is used

[18:40:00.0000] <Philip`>
othermaciej: So instead of saying "ctx.font = '10pt Arial'" when you want to set the font and aren't sure of the current canvas state, you'd have to say "ctx.fontStyle = ctx.fontVariant = ctx.fontWeight = 'normal'; ctx.fontSize = '10pt'; ctx.fontFamily = 'Arial'"? That sounds quite painful...

[18:41:00.0000] <Lachy>
couldnt both .font as well as .fontFamily, .fontSize, etc. be provided like in CSSOM?

[18:42:00.0000] <Hixie>
the CSSOM needs to be optimised for reading

[18:42:01.0000] <Hixie>
this needs to be optimised for writing

[18:44:00.0000] <Lachy>
but if the author only wants to change one property, without affecting the others, then having to use the shorthand all the time would mean all properties would have to be explicitly specified, or they'd be inadvertenly reset to the defaults

[18:45:00.0000] <Lachy>
and changing, e.g. the font-weight requires specifying both the size and family at the same time

[18:45:01.0000] <Hixie>
that's the same argument as "what if you only want to change the Opacity component?"

[18:45:02.0000] <Hixie>
and the answer is, just implement the setter yourself

[18:45:03.0000] <Hixie>
instead of bloating the core API :-)

[18:46:00.0000] <Philip`>
Lachy: Out of lots of pages, I see approximately none setting .style.fontAnything

[18:47:00.0000] <Lachy>
Philip`, ok.

[18:47:01.0000] <Philip`>
(There's 5 setting style.font, 60 setting style.fontSize, 1465 setting style.display)

[18:47:02.0000] <Philip`>
(Not sure how many in total since my grep thing hasn't finished and it doesn't print any status information...)

[18:48:00.0000] <Philip`>
(*how many pages in total)

[18:49:00.0000] <Philip`>
(I really ought to compress all this data so it can be read from disk faster)

[18:49:01.0000] <Philip`>
(and also avoid having it spread over a hundred thousand separate files)

[18:53:00.0000] <othermaciej>
Hixie: my rough opinion after a bit of thought is that the current 'font' API is good as-is, but eventually canvas should be extended to have fonts as first-class objects too, so you can do things like change size or italics status or boldness level without having to do string pasting

[18:53:01.0000] <othermaciej>
Hixie: perhaps font objects could be added at the same time as path objects

[18:54:00.0000] <othermaciej>
Hixie: will try to send email along these lines soon

[18:54:01.0000] <Hixie>
k

[18:54:02.0000] <Hixie>
thanks

[19:31:00.0000] <Philip`>
Lachy: http://philip.html5.org/data/cssom-set-properties.txt shows the style properties that get set in inline scripts (or things that look quite like inline scripts)

[19:32:00.0000] <Philip`>
('font' is still pretty far down)

[19:34:00.0000] <Philip`>
(Insert obvious warning about biased sample etc)

[19:46:00.0000] <Philip`>
Hixie: http://www.whatwg.org/specs/web-apps/current-work/status-documentation.html has typos in "Once a section is interoperably implemented, s quite stable and unlikely to change significantly"

[20:02:00.0000] <roc>
anyone staying at Wild Palms tonight?

[20:10:00.0000] <roc>
grrr

[20:11:00.0000] <othermaciej>
what's wrong?

[20:14:00.0000] <roc>
I posted to the wrong channel

[02:59:00.0000] <Hixie>
http://www.w3.org/MarkUp/2008/xhtml-basic-11-implementation.html is funny

[02:59:01.0000] <Hixie>
they only tested one implementation

[02:59:02.0000] <Hixie>
they only had 3 tests

[03:00:00.0000] <Hixie>
the implementation they tested was an internal build

[03:00:01.0000] <Hixie>
they modified the official test to run it

[03:03:00.0000] <othermaciej>
lol

[03:04:00.0000] <othermaciej>
someone should let them know that mobile profiling is dead

[06:21:00.0000] <takkaria>
/me boggles people

[07:13:00.0000] <hsivonen>
Hixie: Jeremy Keith pointed me to http://ufxtract.com/testsuite/

[10:36:00.0000] <annevk>
back home, finally

[10:36:01.0000] <annevk>
GTA IV was in supply at the train station fortunately, so I'm all set

[10:36:02.0000] <annevk>
(on the other hand, the weather is quite nice around here...)

[13:07:00.0000] <virtuelv>
/me reads backlog across all of othermaciej's connects and disconnects

[13:07:01.0000] <virtuelv>
I almost died laughing reading that XHTML Basic test report

[13:08:00.0000] <Philip`>
virtuelv: http://krijnhoetmer.nl/irc-logs/whatwg/20080510 nicely hides the evidence of othermaciej's quite useless internet connection :-)

[13:09:00.0000] <virtuelv>
Philip`: so would my IRC client, if I just disabled that setting

[13:10:00.0000] <Philip`>
virtuelv: Does it apply that retroactively?

[13:11:00.0000] <virtuelv>
Philip`: i can't seem to find the setting directly in it

[13:11:01.0000] <virtuelv>
so I hardly think so

[13:12:00.0000] <virtuelv>
it doesn't, but I at least finally found the setting

[13:12:01.0000] <virtuelv>
nice bit is that XChat sets it on a per-channel basis

[13:35:00.0000] <jwalden>
postMessage has an early adopter: https://bugzilla.mozilla.org/show_bug.cgi?id=433183

[13:35:01.0000] <jwalden>
and if I may say so, they are absolutely, utterly insane

[13:40:00.0000] <Philip`>
jwalden: Would it still be insane if they were closely following the feature's development and ensured their code always worked reliably with it?

[13:42:00.0000] <jwalden>
only slightly less so

[13:42:01.0000] <jwalden>
what fraction of a fraction of a percent of people would even be able to take advantage of it?

[13:43:00.0000] <jwalden>
once someone ships it makes sense

[13:43:01.0000] <jwalden>
but not before, except experimentally

[13:46:00.0000] <jwalden>
actually, I take that back

[13:46:01.0000] <jwalden>
implementations update to the spec at different speeds

[13:46:02.0000] <jwalden>
so even if you do follow it, you lose

[13:47:00.0000] <annevk>
opera had postMessage quite early

[13:47:01.0000] <annevk>
though it was exposed on document

[14:14:00.0000] <Philip`>
jwalden: I assume the three million FF3 beta testers would be able to take advantage of it, which is not insignificant

[14:16:00.0000] <jwalden>
they can't take that much more advantage of it than they can of whatever the supported method is in 2, and there's still the changing-implementation-rate problem

[14:21:00.0000] <othermaciej>
facebook?!?!

[14:27:00.0000] <jwalden>
so it seems

[14:29:00.0000] <BenMillard>
annevk, I'm back from Sight City with 14.5 pages of notes (A5 lined paper)

[14:30:00.0000] <BenMillard>
a ZoomText developer seemed keen to participate in HTML5

[14:31:00.0000] <BenMillard>
other AT people seemed surprised that HTML5 even existed

[14:32:00.0000] <BenMillard>
they seemed generally positive about us wanting them to participate, though

[14:33:00.0000] <BenMillard>
the notes will eventually become a blog entry on Project Cerbera, somewhat like my entry for the HTMLWG F2F in Boston 2007

[16:04:00.0000] <Philip`>
/me wonders if there's some way he can make Firefox 3 not totally fail to render borders around input boxes

[16:05:00.0000] <Philip`>
It makes textareas unpleasantly invisible :-(


2008-05-11
[19:00:00.0000] <Philip`>
http://labs.adobe.com/technologies/spry/articles/best_practices/validating_spry.html

[19:01:00.0000] <Philip`>
It's great that you can write an HTML document with any arbitrary non-standard features and then have the document declare itself as valid

[19:36:00.0000] <jwalden>
that was the point, sadly

[19:44:00.0000] <Lachy>
It looks like they're using it for a valid purpose though. Being able to check that the proprietary features are used correctly is good for quality assurance.  But their advice is bad because they fail to realise the difference between HTML and XHTML

[19:45:00.0000] <Lachy>
and in HTML5's parsing, IIRC, all browsers will behave the way IE does and output the %SPRY; ]> at the top of the page

[19:47:00.0000] <Lachy>
so if authors want to use it, they would be better off advising it's use only before publishing, and then replacing it with the normal DOCTYPE

[02:31:00.0000] <Hixie>
annevk: yt?

[02:31:01.0000] <Hixie>
annevk: why does XHR not block Cookie, Cookie2, and Authorization headers?

[02:32:00.0000] <Hixie>
annevk: and if XHR shouldn't for same-domain, why does AC not block Authorization headers?

[02:32:01.0000] <Hixie>
or WWW-Authorization or whatever the header is called

[03:15:00.0000] <annevk>
they should probably be blocked

[03:15:01.0000] <annevk>
for AC it does block them though not explicitly

[03:15:02.0000] <annevk>
i.e., that's mentioned in the security section only

[03:15:03.0000] <Hixie>
that's very bad, given that there is a list of headers elsewhere

[03:15:04.0000] <Hixie>
that one could easily assume to be complete

[03:16:00.0000] <Hixie>
generally speaking, one should assume when writing a spec that the security section will be ignored and might as well be non-normative

[03:16:01.0000] <annevk>
yeah, please assume that the header story is incomplete at this point

[03:16:02.0000] <Hixie>
fwiw

[03:16:03.0000] <Hixie>
please fix it asap :-)

[03:16:04.0000] <Hixie>
i really would like this to go to cr before the f2f

[03:16:05.0000] <Hixie>
i'm working hard on getting the mozilla guys to unblock you

[03:16:06.0000] <annevk>
ok

[03:17:00.0000] <annevk>
i was planning on fixing it

[03:17:01.0000] <annevk>
cr seems unrealistic though :(

[03:17:02.0000] <Hixie>
why?

[03:17:03.0000] <Hixie>
i thought it was basically done

[03:17:04.0000] <Hixie>
didn't you ask to go to CR?

[03:17:05.0000] <Hixie>
oh, LC?

[03:17:06.0000] <Hixie>
LC i guess

[03:17:07.0000] <annevk>
yeah

[03:18:00.0000] <Hixie>
whichever

[03:18:01.0000] <Hixie>
the one that means you're done :-)

[03:24:00.0000] <annevk>
the main problem is that it's slightly unclear which headers to block and which not to block

[03:24:01.0000] <annevk>
and why

[03:24:02.0000] <Hixie>
how so?

[03:24:03.0000] <annevk>
i got these lists from implementors, but they didn't provide much rationale :)

[03:24:04.0000] <Hixie>
the list (with the addition of Cookie and WWW-Authorization) seems pretty solid to me

[03:25:00.0000] <Hixie>
and it's the same list as XHR1

[03:25:01.0000] <annevk>
yeah, but AC also has two additional whitelists, one for which headers can be exposed and one for which headers are always safe

[03:26:00.0000] <annevk>
(and the xhr1 list has outstanding questions as well)

[03:26:01.0000] <annevk>
(from reviewers)

[03:28:00.0000] <Hixie>
so reply to the questions :-)

[03:28:01.0000] <Hixie>
where are these whitelists defined?

[03:29:00.0000] <Hixie>
"white" isn't in the AC spec as far as i can tell

[03:29:01.0000] <annevk>
one is in xhr2 but jonas requested it to be moved to AC and one is in AC

[03:30:00.0000] <annevk>
http://dev.w3.org/2006/waf/access-control/#cross-site the first whitelist is just "accept" and "accept-language"

[03:30:01.0000] <annevk>
the other is slightly longer, under getResponseHeader

[03:31:00.0000] <Hixie>
ah

[03:31:01.0000] <Hixie>
having all these lists everywhere is confusing :-P

[03:31:02.0000] <annevk>
agreed

[03:32:00.0000] <Hixie>
so wait, you can't send any custom headers at all with XXX?

[03:32:01.0000] <annevk>
you can

[03:32:02.0000] <annevk>
just requires a preflight request for GET

[03:32:03.0000] <Hixie>
oh this is response headers

[03:32:04.0000] <Hixie>
nevermind

[03:33:00.0000] <Hixie>
i don't care about response headers :-)

[03:33:01.0000] <Hixie>
do you think the Cookie headers will be added to XHR1?

[03:34:00.0000] <annevk>
it was requested by adam barth or his collegue (security people), but no browser does it currently i believe

[03:35:00.0000] <Hixie>
that didn't answer my question :-)

[03:38:00.0000] <annevk>
i'm not sure what to do :)

[03:41:00.0000] <annevk>
for completeness the blacklist from xhr should probably be put in ac as well

[03:41:01.0000] <Hixie>
have you asked browser vendors what they'll do?

[03:42:00.0000] <Hixie>
the blacklist is in AC already

[03:42:01.0000] <Hixie>
it's identical but adds Cookie and Cookie2 (and not WWW-Authorisation, which is imho a serious error as that header should be in both specs)

[03:42:02.0000] <annevk>
i guess i should try e-mailing them directly

[03:43:00.0000] <annevk>
they're not all that responsive on the mailing list...

[03:44:00.0000] <Hixie>
yes, i often have to actually speak to them in person in fact

[03:44:01.0000] <Hixie>
that's one reason i wanted to move to the bay area -- more browser vendors here

[03:44:02.0000] <Hixie>
makes it easier to do my job

[03:44:03.0000] <Hixie>
:-)

[03:51:00.0000] <Hixie>
/me replies to cwilson's e-mail

[03:55:00.0000] <annevk>
i can see how that is convenient

[06:43:00.0000] <Philip`>
Mozilla needs better canvas regression tests :-(

[08:00:00.0000] <gsnedders>
jgraham: was <http://flickr.com/photos/gsnedders/2375643370/in/set-72157604322177094/> the photo to which you were referring on Friday?

[08:09:00.0000] <gsnedders>
Philip`: Sorry for keeping you out so long and making your hands cold (re what you said on Fri night)

[08:12:00.0000] <jgraham>
gsnedders: No. I'm not sure which photo it was

[08:12:01.0000] <gsnedders>
/me wonders what other photos there are online with any girl

[08:13:00.0000] <Philip`>
gsnedders: It was actually surprisingly warm until just before we left, so it wasn't really a problem :-)

[08:13:01.0000] <Philip`>
(though I am quite easily surprised by warmth here)

[08:14:00.0000] <gsnedders>
It was so humid on Saturday :(

[08:14:01.0000] <gsnedders>
Not a nice day for my grandmother's memorial service

[08:15:00.0000] <Philip`>
Are you heading back home now?

[08:15:01.0000] <gsnedders>
yeah

[08:15:02.0000] <gsnedders>
On a damned hot train

[08:16:00.0000] <gsnedders>
/me wonders if jgraham's thought that he was the first person to write ABNF on a train was right

[08:17:00.0000] <Philip`>
gsnedder⊙nns looks like SIRIUS

[08:17:01.0000] <gsnedders>
http://stuff.gsnedders.com/http-parsing.html#anchor8 contains a fair amount of ABNF now

[08:19:00.0000] <Philip`>
...though from their website it sounds like it's a unidirectional internet connection, so I don't know how they handle outgoing traffic, unless their site is just misleading

[08:21:00.0000] <gsnedders>
/me can report outgoing traffic works fine

[08:21:01.0000] <Philip`>
Maybe at much lower bandwidth than incoming?

[08:22:00.0000] <gsnedders>
Bandwidth fluctuates so much that's hard to judge

[08:22:01.0000] <Philip`>
/me wonders what traceroute says, since that ought to show the outgoing route (assuming it doesn't just fail to work)

[08:24:00.0000] <gsnedders>
Philip`: it stops at ims.icomera.com (192.168.0.1)

[08:24:01.0000] <gsnedders>
(trying to reach Google)

[08:26:00.0000] <Philip`>
Ah

[08:26:01.0000] <Philip`>
I suppose that's http://www.icomera.com/solutions/overview

[08:26:02.0000] <Philip`>
which seems to be using every wireless technology it can gets it hands on

[08:26:03.0000] <Philip`>
s/gets it/get its/

[08:26:04.0000] <gavin_>
Philip`: did some bug surface that your tests didn't catch?

[08:27:00.0000] <gavin_>
your tests that landed in Mozilla CVS, I mean

[08:28:00.0000] <gsnedders>
Well, as I said on Friday, it can't just use sats because it would break in tunnels (as the high frequency waves wouldn't diffract into the tunnels)

[08:28:01.0000] <Philip`>
gavin_: https://bugzilla.mozilla.org/show_bug.cgi?id=433235

[08:29:00.0000] <Philip`>
gavin_: I think I hadn't written that test when I was Mozillafying them

[08:29:01.0000] <gavin_>
ok

[08:30:00.0000] <gavin_>
I nominated it for blocking1.9, vlad will look into it I'm sure

[08:31:00.0000] <jgraham>
gsnedders: FWIW http://flickr.com/photos/gsnedders/2376735304/ isn't the photo I had in mind, but you do look a little... reanimated there

[08:32:00.0000] <Philip`>
gavin_: Okay, thanks - I don't expect it's a critical bug that anyone would really care about, but I'd like things to be bug-free anyway :-)

[08:33:00.0000] <Dashiva>
jgraham: An albino with a wig, maybe? :P

[09:36:00.0000] <gsnedders>
The heat of this train is really surreal.

[09:37:00.0000] <Philip`>
Open the window and stick your head out

[09:39:00.0000] <gsnedders>
Windows are only in existance at the end of the coach, and sticking your head out is quite possibly fatal

[09:40:00.0000] <jgraham>
"The heat of this train is really surreal." - it's causing all the clocks to melt?

[09:41:00.0000] <gsnedders>
Not quite.

[09:41:01.0000] <gsnedders>
But certainly this coach is > 30°C

[10:10:00.0000] <gsnedders>
jgraham: Having tagged all the photos of me, I still don't see what you were meaning at <http://flickr.com/photos/tags/gsnedders/>

[10:20:00.0000] <jgraham>
gsnedders: You've totally failed to consider the possibility that I imagined the whole thing and/or my memory is faulty

[10:21:00.0000] <gsnedders>
jgraham: No, I haven't. I just haven't said it yet.

[10:21:01.0000] <gsnedders>
jgraham: With photos of me, as someone with fairly light skin anyway, it's possible :P

[10:21:02.0000] <Philip`>
There's also the possibility that it is a malicious plot rather than an accidental falsehood

[10:22:00.0000] <gsnedders>
Philip`: I hadn't considered that, yet.

[10:31:00.0000] <gsnedders>
"Your cache administrator is root. " (from ims.nationalexpresswifi.train)

[10:32:00.0000] <Philip`>
/me discovers that he can't understand some of his canvas test cases, so he skips over them and assumes they're right

[10:33:00.0000] <Philip`>
(Fortunately they're undefined behaviour, so "right" is very subjective, and nobody can argue that my tests conflict with the spec)

[10:42:00.0000] <gsnedders>
It's raining in St Andrews (apparently)!

[10:43:00.0000] <gsnedders>
And all I have it a t-shirt :P

[10:43:01.0000] <gsnedders>
The first fine day next week is the first day of my exams. Typical.

[10:44:00.0000] <Philip`>
Liquid from the sky? That sounds wildly improbable

[10:44:01.0000] <gsnedders>
Well, look at the weather where you are Philip` 

[10:45:00.0000] <gsnedders>
(or, rather, how it was at 12:09Z, seeming that's the last time for which I have known good information)

[10:47:00.0000] <Philip`>
I can't see what the weather is like - I've closed my window blinds since it was too sunny and hot

[10:52:00.0000] <gsnedders>
Now we're in fog, so I'm just using that

[11:49:00.0000] <zcorpan_>
hsivonen: http://validator.nu/?doc=http%3A%2F%2Fwww.validome.org%2Fout%2Fena10&showsource=yes has confusing errors

[11:52:00.0000] <hsivonen>
zcorpan_: indeed. thanks

[11:54:00.0000] <zcorpan_>
hsivonen: http://validator.nu/?doc=http%3A%2F%2Fwww.validome.org%2Fout%2Fena13&showsource=yes could have more useful messages

[11:56:00.0000] <hsivonen>
zcorpan_: do you have a suggestion on how to handle it better?

[11:56:01.0000] <hsivonen>
(aside: in practice that one isn't much of a concern because UTF-16 isn't too popular for serialization)

[11:57:00.0000] <Dashiva>
hsivonen: You could search for a doctype in any poisition, and if you find one, report that as a possible error cause?

[11:57:01.0000] <zcorpan_>
hsivonen: sniff for a utf-16 BOM even if http content-type says something different

[11:57:02.0000] <hsivonen>
Dashiva: I don't follow.

[11:57:03.0000] <zcorpan_>
hsivonen: (although yeah, perhaps optimising utf-16 isn't worth it)

[11:57:04.0000] <hsivonen>
zcorpan_: makes sense

[11:59:00.0000] <Dashiva>
hsivonen: never mind, I misread the error messages

[12:00:00.0000] <zcorpan_>
hsivonen: http://validator.nu/?showsource=yes&doc=http://www.validome.org/out/ena2003 seems to truncate the document

[12:01:00.0000] <hsivonen>
zcorpan_: yeah, fatal error does

[12:01:01.0000] <Philip`>
hsivonen: Are there plans to change the validator output to tell people how to fix the errors, rather than just saying what the errors are and expecting them to figure out a solution themselves?

[12:03:00.0000] <hsivonen>
Philip`: for datatypes it already does. for other errors no tangible plans

[12:03:01.0000] <hsivonen>
Philip`: do you have ideas of how to correlate errors with advice in the cases where the advice would be non-obvious

[12:03:02.0000] <hsivonen>
?

[12:03:03.0000] <Philip`>
hsivonen: Non-obvious to who?

[12:04:00.0000] <hsivonen>
Philip`: good question

[12:04:01.0000] <zcorpan_>
hsivonen: what's the error in that document?

[12:04:02.0000] <Philip`>
When it says "Error: Text after & did not match an entity name.", I guess most people would find it entirely non-obvious that they have to write &amp; instead

[12:05:00.0000] <hsivonen>
zcorpan_: looks like a bug

[12:06:00.0000] <zcorpan_>
hsivonen: validome seems to think that http://validator.nu/?showsource=yes&doc=http://www.validome.org/out/ena2004 should have no errors

[12:07:00.0000] <hsivonen>
I really need to implement links to web-sniffer.net if only just for debugging

[12:09:00.0000] <hsivonen>
zcorpan_: appendix F of XML 1.0 first ed. considers UTF-16 without a BOM an error

[12:12:00.0000] <zcorpan_>
hsivonen: but not 4th ed?

[12:12:01.0000] <hsivonen>
zcorpan_: seems to be a case where the spec has changed since the parser was written

[12:12:02.0000] <zcorpan_>
hsivonen: ok

[12:13:00.0000] <zcorpan_>
hsivonen: might be good to issue a warning in that case

[12:13:01.0000] <hsivonen>
zcorpan_: hmm. the non-normative appendix changed

[12:14:00.0000] <hsivonen>
zcorpan_: the normative text still says: "Entities encoded in UTF-16 must and entities encoded in UTF-8 may begin with the Byte Order Mark"

[12:15:00.0000] <hsivonen>
zcorpan_: where does validome tell the expected outcome of tests?

[12:17:00.0000] <zcorpan_>
hsivonen: it doesn't, though it says whether the validome output is correct or not

[12:18:00.0000] <zcorpan_>
hsivonen: http://validator.nu/?showsource=yes&doc=http://www.validome.org/out/ena2016 and http://validator.nu/?showsource=yes&doc=http://www.validome.org/out/ena2017 are seemingly not supported (utf-32 with byte order 2143 and 3412 respectively)

[12:19:00.0000] <hsivonen>
I thought I had disabled UTF-32 altogether

[12:21:00.0000] <zcorpan_>
hsivonen: http://validator.nu/?showsource=yes&doc=http://www.validome.org/out/ena2014 works, is utf-32 byte order 1234

[12:22:00.0000] <hsivonen>
zcorpan_: I sent email about ena2004 to Validome Staff

[12:22:01.0000] <zcorpan_>
hsivonen: ok

[12:22:02.0000] <hsivonen>
Philip`: yeah, I guess I should improve those messages.

[12:23:00.0000] <hsivonen>
btw, did my extensibility email make sense?

[12:28:00.0000] <Philip`>
hsivonen: For case #3, I don't see why you say XML is a good fit and then say actually it's not good and you should use XML5 instead but actually that's not good either because it's a different syntax to HTML

[12:29:00.0000] <Philip`>
That seems to lead to the conclusion that XML is not a good solution

[12:32:00.0000] <hsivonen>
Philip`: I say XML. the first objection is really Sam's objection. the second objection is Dave's objection to my suggested solution to Sam's objection

[12:33:00.0000] <hsivonen>
Philip`: I'd be OK with leaving people with XML 1.0 in that case

[12:34:00.0000] <Philip`>
hsivonen: The second objection seems to be unresolved

[12:35:00.0000] <Philip`>
I don't think Facebook cares about being able to serialise all XML 1.0 infosets as FBML, so that seems to be an irrelevant point

[12:36:00.0000] <Philip`>
Also I expect they could easily want "<p>Blah <p>Blah" to be parsed sensibly, since that follows their users' understanding of how to write HTML, and XML5 wouldn't do that (as far as I'm aware)

[12:37:00.0000] <Philip`>
so they do still want one syntax instead of two

[12:40:00.0000] <hsivonen>
Philip`: ok

[12:41:00.0000] <hsivonen>
Philip`: it seems that I ended up replying to typical XML proponent concerns instead of developing my case properly

[12:47:00.0000] <zcorpan_>
hsivonen: validome expects http://validator.nu/?showsource=yes&doc=http://www.validome.org/out/ena6009 to emit an error

[12:47:01.0000] <zcorpan_>
hsivonen: (likewise for <table cellspacing>)

[12:50:00.0000] <hsivonen>
/me is puzzled by how the authors goes from "third" to "fourth" at http://www.bestkungfu.com/archive/date/2008/05/alt-and-the-flickr-defense/

[12:50:01.0000] <zcorpan_>
hsivonen: http://validator.nu/?showsource=yes&doc=http://www.validome.org/out/ena6017 has confusing messages

[12:51:00.0000] <hsivonen>
zcorpan_: yeah, the legacy schemas aren't too good :-(

[12:51:01.0000] <zcorpan_>
hsivonen: it says it uses the schema for html5 :)

[12:52:00.0000] <hsivonen>
zcorpan_: I meant ena6009

[12:53:00.0000] <hsivonen>
zcorpan_: but for the latter case, the legacy encoding decl datatype is bad

[12:53:01.0000] <zcorpan_>
hsivonen: indeed, i was thinking of the message about <frameset> not being allowed

[12:54:00.0000] <zcorpan_>
hsivonen: why isn't it using the schema for xhtml 1.0 frameset?

[12:58:00.0000] <hsivonen>
zcorpan_: probably because when I wrote the autodetection code, Frameset wasn't in the presets due to evilness or something along those lines

[12:59:00.0000] <hsivonen>
but yeah, I should fix that

[12:59:01.0000] <hsivonen>
in fact, I've made notes for everything you said except UTF-16 without BOM

[13:00:00.0000] <zcorpan_>
hsivonen: validome also expects an empty body to be valid in xhtml 1.0, which it is if you read the xhtml 1.0 spec pedantically, but i'd argue that that's a spec bug since it's invalid html4

[13:01:00.0000] <zcorpan_>
hsivonen: (coords='' checking is better when validating against html5 than xhtml 1.0)

[13:04:00.0000] <zcorpan_>
hsivonen: http://validator.nu/?showsource=yes&doc=http://www.validome.org/out/ena7001 doesn't emit an error

[13:05:00.0000] <zcorpan_>
hsivonen: (might have pointed this one out before: ) validome expects errors for http://validator.nu/?showsource=yes&doc=http://www.validome.org/out/ena7003 but i don't know why

[13:06:00.0000] <hsivonen>
what should be the error in ena7003

[13:06:01.0000] <hsivonen>
?

[13:07:00.0000] <hsivonen>
doh

[13:07:01.0000] <hsivonen>
/me sees "but i don't know why"

[13:07:02.0000] <zcorpan_>
hsivonen: http://validator.nu/?showsource=yes&doc=http://www.validome.org/out/ena3004 shouldn't have errors per html4 :)

[13:08:00.0000] <hsivonen>
zcorpan_: yeah, but this is HTML5 plus additional errors

[13:08:01.0000] <hsivonen>
in the parser

[13:08:02.0000] <hsivonen>
is the error in ena7001 that name and id aren't the same?

[13:09:00.0000] <hsivonen>
I disabled that check, because the Schematron was a perf issue

[13:10:00.0000] <zcorpan_>
hsivonen: yeah it is

[13:11:00.0000] <zcorpan_>
hsivonen: it seems that cdata parsing of <noframes> is incompatible with the frameset schema

[13:11:01.0000] <hsivonen>
nice

[13:12:00.0000] <zcorpan_>
hsivonen: (the parser also complains about </ in noframes)

[13:12:01.0000] <hsivonen>
:-(

[13:12:02.0000] <hsivonen>
so much brokenness on the legacy side

[13:14:00.0000] <zcorpan_>
html5 parsing rules yields confusing messages for http://validator.nu/?showsource=yes&doc=http://www.validome.org/out/ena3006

[13:16:00.0000] <zcorpan_>
/me notes that validome fails to detect http://www.validome.org/lang/en/get/http://www.validome.org/out/ena3008 even though the comparison page says validome gets it right

[13:23:00.0000] <zcorpan_>
hsivonen: validome expects an error for missing content-script-type http://validator.nu/?doc=http%3A%2F%2Fwww.validome.org%2Fout%2Fena3013&schema=http%3A%2F%2Fs.validator.nu%2Fxhtml10%2Fxhtml-strict.rnc+http%3A%2F%2Fs.validator.nu%2Fxhtml10%2Fxhtml.sch+http%3A%2F%2Fc.validator.nu%2Fall-html4%2F&showsource=yes (although xhtml 1.0 doesn't require it, if you read xhtml 1.0 just as pedantically as with empty <body>)

[13:24:00.0000] <zcorpan_>
(and content-script-type is completely useless anyway)

[13:28:00.0000] <hsivonen>
I'm not gonna add content-script-type check

[13:29:00.0000] <zcorpan_>
makes sense

[13:34:00.0000] <zcorpan_>
hsivonen: "I have to change my web-dev preset links (it should be automatically - my next request for ur validator.nu) " -- http://www.sitepoint.com/forums/showthread.php?p=3815150#post3815150

[13:36:00.0000] <zcorpan_>
hsivonen: i guess the request is that you look at the doctype to choose an appropriate xhtml 1.0 preset if the root namespace is xhtml

[13:37:00.0000] <zcorpan_>
hsivonen: (i know the current behavior is by design :) )

[13:40:00.0000] <hsivonen>
I'd rather look into whining about the doctype in a more comprehensible way

[13:41:00.0000] <hsivonen>
It's hard to balance this particular advocacy feature with user requests

[13:45:00.0000] <hsivonen>
fwiw, there's no mechanism in place for picking a present schema from XML doctype

[13:45:01.0000] <hsivonen>
it's not like I just disabled something

[13:45:02.0000] <zcorpan_>
yeah. saying "just use the preset select box" means "this validator is stupid and makes me jump through unnecessary hoops", and explaining the reasoning behind the behavior is way too much effort

[13:47:00.0000] <zcorpan_>
it seems that new comers don't check their own documents, instead they check documents from top sites they know have good markup

[13:47:01.0000] <zcorpan_>
which means xhtml documents, even if they use text/html themselves

[13:48:00.0000] <hsivonen>
skipped entity [dtd] needs to become something more comprehensible as the first change

[13:48:01.0000] <hsivonen>
like <blinquee>DTDs are legacy. Get rid of yours.</blinquee>

[13:49:00.0000] <zcorpan_>
lol <blinquee>

[13:49:01.0000] <zcorpan_>
/me files a bug for opera to implement <blinquee>

[13:49:02.0000] <zcorpan_>
not supporting it breaks this page! validator.nu

[13:50:00.0000] <hsivonen>
there at least was a demo implementing blinquee for Gecko years ago

[13:50:01.0000] <zcorpan_>
oh? didn't know

[13:51:00.0000] <Philip`>
https://bugzilla.mozilla.org/show_bug.cgi?id=163050 ?

[13:52:00.0000] <hsivonen>
Philip`: that's it

[14:21:00.0000] <hendry>
/me is back from Dublin

[15:02:00.0000] <Philip`>
Hmph, HTML5 is insufficiently strict about butts

[15:06:00.0000] <Dashiva>
<body-part />

[15:09:00.0000] <Lachy_>
Philip`, wtf?

[15:09:01.0000] <Philip`>
Dashiva: Not that kind of butt :-(

[15:09:02.0000] <Philip`>
You have a dirty mind

[15:09:03.0000] <Philip`>
It just says "The butt value means that the end of each line is a flat edge perpendicular to the direction of the line." which is unacceptably lacking any sort of "must"

[15:11:00.0000] <jgraham_>
Does anyone have a a suggestion for a not-too-inaccurate not-too-technical HTML tutorial I can point my mum at?

[15:13:00.0000] <gsnedders>
Talking to my parents about Friday night led to the inevitable, "Were they like what you expected?", which in turn lead to whether I was what you expected. So, Philip`, jgraham_?

[15:15:00.0000] <jgraham_>
/me doesn't really know

[15:15:01.0000] <gsnedders>
jgraham_: So you had no expectations? :P

[15:15:02.0000] <Philip`>
I never have expectations

[15:15:03.0000] <Philip`>
Life is much more fun when everything is a surprise

[15:16:00.0000] <gsnedders>
did I have hair more awesome than expected? :P

[15:17:00.0000] <jgraham_>
"Monday make up knowing that you've gotta go to school. You tell your Mum what to expect she says it's right out of the blue"

[15:17:01.0000] <gsnedders>
jgraham_: see my note on http://flickr.com/photos/jgraham/2479527700/?

[15:18:00.0000] <Lachy_>
gsnedders, is that you in the photo?

[15:18:01.0000] <gsnedders>
Lachy_: yeah

[15:18:02.0000] <jgraham_>
gsnedders: Have now :)

[15:18:03.0000] <gsnedders>
/me couldn't resist the "not us"

[15:18:04.0000] <Lachy_>
did you dye your hair? I thought you had blonde or light brown hair.

[15:19:00.0000] <gsnedders>
Nope.

[15:19:01.0000] <Lachy_>
really? I thought I'd seen a photo of you before. Maybe it was of someone else

[15:19:02.0000] <gsnedders>
If I ever dye my hair, it will be obvious.

[15:19:03.0000] <gsnedders>
By obvious, I meant really vibrant and unnatural (like green or blue)

[15:20:00.0000] <gsnedders>
Lachy_: http://flickr.com/photos/tags/gsnedders/

[15:20:01.0000] <gsnedders>
Lachy_: That goes back a while

[15:20:02.0000] <jgraham_>
gsnedders: Have you ever met David Perrett?

[15:20:03.0000] <gsnedders>
jgraham_: No

[15:20:04.0000] <gsnedders>
Name means nothing to me either

[15:21:00.0000] <Philip`>
http://www.st-andrews.ac.uk/~www_sp/people/resources/pics/dp.jpg ?

[15:21:01.0000] <jgraham_>
Ah, OK, nevermind then. He's a professor at St. Andrews that my parents know. He has dyed hair

[15:21:02.0000] <jgraham_>
Philip`: Yep

[15:22:00.0000] <zcorpan_>
hsivonen: xml parser bug http://validator.nu/?doc=http%3A%2F%2Fsimon.html5.org%2Ftest%2Fxml%2Fns-malformed%2F001.xml&showsource=yes

[15:22:01.0000] <gsnedders>
I'd expect my father knows him.

[15:22:02.0000] <gsnedders>
But he knows far too many people at uni

[15:24:00.0000] <gsnedders>
Lachy_: http://www.bebo.com/PhotoAlbumBig.jsp?MemberId=4376278632&PhotoAlbumId=5579156004&PhotoId=7209454479 — that's the oldest photo of me online

[15:24:01.0000] <gsnedders>
(I'll get more old photos on Flickr once I work out what they are of and/or label them)

[15:25:00.0000] <gsnedders>
Lachy_: It used to be a bit redder and glossier, but never really light

[15:26:00.0000] <Lachy_>
yeah, I think I just remembered wrongly. I remember seeing this one before http://www.bebo.com/PhotoAlbumBig.jsp?PageNbr=1&MemberId=4376278632&PhotoAlbumId=5579156004&PhotoId=5770850872

[15:26:01.0000] <gsnedders>
jgraham_: http://flickr.com/photos/jgraham/2479527704/ claims to be 8:58 BST

[15:26:02.0000] <gsnedders>
Lachy_: That just generally looks light

[15:27:00.0000] <gsnedders>
jgraham_: (and that's far too early)

[15:27:01.0000] <Philip`>
"Taken on May 9, 2008 at 8.58pm BST"

[15:27:02.0000] <Philip`>
/me points at the "pm" part

[15:27:03.0000] <gsnedders>
Still too early

[15:27:04.0000] <Philip`>
Oh

[15:27:05.0000] <gsnedders>
We weren't by Great St Mary's until later than that

[15:28:00.0000] <gsnedders>
We left at around 22:45+01, and I think we were there for around an hour

[15:28:01.0000] <gsnedders>
so likely it was actually 21:58+01

[15:28:02.0000] <Philip`>
That sounds plausible

[15:28:03.0000] <gsnedders>
I expect the photo on the Fen is equally wrong

[15:29:00.0000] <jgraham_>
I think my camera is set to GMT

[15:29:01.0000] <gsnedders>
jgraham_: That's my guess too

[15:29:02.0000] <gsnedders>
As I said, we should just set everything to GMT and be done with it :P

[15:29:03.0000] <jgraham_>
Though it's still not quite right and I don't know how you change the TZ

[15:29:04.0000] <Philip`>
Does Flickr just guess the timezone?

[15:30:00.0000] <jgraham_>
Conclusion: Metadata sucks.

[15:30:01.0000] <gsnedders>
How long were we at supper for?

[15:30:02.0000] <Philip`>
It probably took half an hour deciding where to go and then another half an hour deciding what to eat there :-p

[15:31:00.0000] <gsnedders>
/me looks at jgraham_ to know whether we're ready to order yet

[15:32:00.0000] <gsnedders>
I mean, he is the eldest :P

[15:34:00.0000] <gsnedders>
All my photos of him have come out badly, though

[15:36:00.0000] <jgraham_>
The EXIF format seems to have a optional timezone field by no way of storing timezones in the dates

[15:37:00.0000] <gsnedders>
Anyhow, I need to go sleep.

[15:37:01.0000] <gsnedders>
Two days of school, then study leave

[15:37:02.0000] <jgraham_>
goodnight

[15:37:03.0000] <gsnedders>
I'll deal with my photos from Cambridge… sometime.

[15:38:00.0000] <gsnedders>
/me needs to `rm to-do/*`

[15:39:00.0000] <Philip`>
The 'exif' tool says "Date and Time (original): 2008:05:09 20:58:31" and seems to indicate a lack of any timezone information anywhere

[15:40:00.0000] <Philip`>
so I assume Flickr is just making up bogus values for its own required 'timezone' metadata field


2008-05-12
[18:55:00.0000] <othermaciej>
Lachy_: do you know if any Opera preview builds support Selecotrs API?

[18:56:00.0000] <othermaciej>
(or does anyone else know?)

[18:56:01.0000] <othermaciej>
hi roc

[18:56:02.0000] <roc>
hi

[19:31:00.0000] <MikeSmith>
othermaciej: I am pretty sure that the Selectors API has not been implemented in Opera yet, even in snapshot builds

[19:35:00.0000] <othermaciej>
MikeSmith: ok, thanks

[19:35:01.0000] <othermaciej>
I'm writing up what was new in WebKit 3.1 and I want to mention which things are supported in other browsers.

[19:37:00.0000] <MikeSmith>
othermaciej: as far as things supported in other browsers, I believe the details in my XTech presentation are accurate (though I didn't bother to record citations)

[19:37:01.0000] <MikeSmith>
though I made on update to that since I did the presentation

[19:38:00.0000] <MikeSmith>
I added the fact that Mozilla/FF3 now supports the AppCache API

[19:38:01.0000] <MikeSmith>
which I didn't know until I read Mark Finkle's blog posting about it a few days ago

[19:42:00.0000] <othermaciej>
MikeSmith: I am testing as needed

[19:54:00.0000] <MikeSmith>
fucking openoffice

[19:54:01.0000] <MikeSmith>
/me tries hard to think of how any office suite could possibly be worse than openoffice

[19:55:00.0000] <MikeSmith>
great way to completely lock up my whole machine

[19:56:00.0000] <takkaria>
MikeSmith: it could not exist, for one

[19:57:00.0000] <MikeSmith>
openoffice should show a big warning every time you start it up: Don't be stupid enough to actually try to use this application to do anything useful.

[19:57:01.0000] <MikeSmith>
those guys should be ashamed of themselves with that thin

[19:58:00.0000] <jwalden>
I wouldn't go that far; on the other hand, I don't think it's something they can be especially proud of

[20:00:00.0000] <Philip`>
OpenOffice gets something like ten billion dollars less revenue each year than Microsoft Office, so it's not surprising that it doesn't work quite so well

[20:03:00.0000] <MikeSmith>
I don't use MS Office at all, so I don't really care how it compares. I just need it to work without it causing me to lose data and waste time

[20:03:01.0000] <MikeSmith>
it just should not be advertised as being stable and ready for production use

[20:03:02.0000] <MikeSmith>
because it's not

[20:13:00.0000] <roc>
used Word much?

[20:16:00.0000] <roc>
because I used to write papers in Word, and I kept running into this bug where Word would replace one of my figures with a big red X while it was repaginating. Apparently that bug has/had been around for years and years.

[20:20:00.0000] <MikeSmith>
roc: sounds pretty annoying

[20:20:01.0000] <MikeSmith>
I don't use Word, ever

[20:20:02.0000] <MikeSmith>
not for years now at least

[20:20:03.0000] <roc>
OK

[20:20:04.0000] <MikeSmith>
but yeah, when I did I can remember running into many bugs

[20:21:00.0000] <roc>
I'm just saying that Word has pretty bad bugs. It seems everyone got used to them.

[20:22:00.0000] <MikeSmith>
yeah, and/or figures out workarounds and uses those and eventually forgets there was a bug because they just use the workarounds. that's the problem with this kind of stuff -- it forces you to alter the way you use the app, to the point where you eventually don't even realize it

[20:22:01.0000] <roc>
right.

[20:22:02.0000] <roc>
it's a big barrier to entry

[20:24:00.0000] <MikeSmith>
I just find that working with any kind of office app is a huge time-waster.. end up needing to spend time trying to make all kinds of formatting tweaks

[20:24:01.0000] <roc>
ok, well, for you there's emacs and Latex :-)

[20:25:00.0000] <MikeSmith>
heh

[20:26:00.0000] <MikeSmith>
I usually try to stick to HTML and generate PDFs from that when needed 

[20:26:01.0000] <MikeSmith>
using Prince or something

[20:39:00.0000] <jwalden>
latex is a far worse time waster than an office app -- soo many ways to tweak stuff to be more semantic, change display characteristics, etc.

[22:11:00.0000] <Hixie>
wow, sam missed the point of your case #3

[22:12:00.0000] <Hixie>
i don't understand people's obsession with generic syntax extensibility

[22:12:01.0000] <Hixie>
given how small a part of the puzzle syntax is

[22:37:00.0000] <MikeSmith>
I'm not sure why distributed extensibility is seen as necessarily a work item that should be taken up by the HTML WG

[22:37:01.0000] <MikeSmith>
instead of by the CDF WG

[22:37:02.0000] <Hixie>
or the XMLWG

[22:37:03.0000] <Hixie>
oh wait they already solved it!

[22:38:00.0000] <roc>
well

[22:39:00.0000] <roc>
gah

[22:39:01.0000] <roc>
I can't remember what confidentiality level my information is

[22:39:02.0000] <Hixie>
w3c information?

[22:40:00.0000] <roc>
oh

[22:40:01.0000] <roc>
http://www.w3.org/2007/11/CDF_rechartering/CDF-proposed

[22:40:02.0000] <roc>
> Given the recent activity in non-XML HTML, the CDF WG will devote attention to inclusion of other languages in HTML as well as XHTML.

[22:40:03.0000] <Hixie>
haha

[22:40:04.0000] <Hixie>
good times

[22:40:05.0000] <Hixie>
now they can be pointless and ineffectual in twice as many syntaxes

[22:44:00.0000] <MikeSmith>
Hixie: do you remember when and why the "The HTML WG is encouraged to provide a mechanism to permit independently developed vocabularies such as Internationalization Tag Set (ITS), Ruby, and RDFa to be mixed into HTML documents." language was added to the draft charter?

[22:48:00.0000] <MikeSmith>
also, noting "Tim Berners-Lee described this incorporation of SVG and MathML without namespaces as horrific and the issue raiser completely concurs with the him." --

[22:48:01.0000] <MikeSmith>
noting that, it seems like what this issue should really have been titled is "Adding support for namespaces text/html"

[22:48:02.0000] <MikeSmith>
because that's effectively what it amounts to

[22:57:00.0000] <Hixie>
MikeSmith: because my advice was ignored and some of the architectural astronauts got their way, i imagine

[22:59:00.0000] <Hixie>
i think it's telling that no web app provider, and no web browser vendor, has asked for this feature

[23:02:00.0000] <MikeSmith>
I think it is telling that no large content providers or normal users have either

[23:06:00.0000] <othermaciej>
Hixie: what feature?

[23:06:01.0000] <othermaciej>
(I guess I should read the logs)

[23:06:02.0000] <Hixie>
namespaces

[23:06:03.0000] <othermaciej>
oh, CDF

[23:06:04.0000] <othermaciej>
oh, namespaces

[23:06:05.0000] <Hixie>
in text/html

[23:08:00.0000] <othermaciej>
CDF WG has not been very successful in engaging UA vendors

[23:08:01.0000] <othermaciej>
their main successes are writing down the obvious conclusions of other specs

[23:16:00.0000] <Hixie>
i was on the CDF WG

[23:16:01.0000] <Hixie>
i quit when i realised the group was clueless and doing pointless work

[23:19:00.0000] <MikeSmith>
I mentioned CDF just because it seems to me the group that is actually explicitly chartered to work on this very thing

[23:20:00.0000] <MikeSmith>
unlike the HTML WG, whose charter just says the group is "encouraged" to provide a mechanism for it

[23:21:00.0000] <MikeSmith>
and just as an aside, it's odd the degree to which some advocates of this have gone to avoid mentioning namespaces

[23:22:00.0000] <MikeSmith>
e.g., the HTML WG charter itself says "Whether this occurs through the extensibility mechanism of XML"

[23:43:00.0000] <Hixie>
MikeSmith: what the htmlwg charter says doesn't much matter to me in the context of this channel :-)

[00:22:00.0000] <hsivonen>
my humor/sarcasm/troll/seriousness detector is completely malfunctioning when it comes to the "move semantics into attributes" thread

[00:42:00.0000] <Hixie>
hsivonen: pretty sure he's being honest

[01:19:00.0000] <annevk>
hsivonen, I think in theory we can get away with two formats

[01:20:00.0000] <annevk>
If XML5 requires a new media type and all I'm not sure it's worth it

[01:32:00.0000] <hsivonen>
annevk: yeah, I should have rethought the point before replying.

[01:36:00.0000] <Hixie>
i hope someone sends a detailed reply to the e-mail justin sent me, because i don't want to do it :-P

[01:47:00.0000] <annevk>
Hixie, for the response headers, do you remember how you ended up with the whitelist?

[01:48:00.0000] <Hixie>
which one?

[01:48:01.0000] <annevk>
for getResponseHeader

[01:49:00.0000] <annevk>
"Cache-Control, Content-Language, Content-Type, Expires, Last-Modified, or Pragma"

[01:49:01.0000] <Hixie>
i came up with that?

[01:50:00.0000] <Hixie>
i probably looked at the spec and decided which were useful and safe

[01:50:01.0000] <annevk>
near the bottom of http://lists.w3.org/Archives/Public/public-webapi/2006Jun/0012.html

[01:51:00.0000] <annevk>
i wonder if it should be a blacklist instead

[01:52:00.0000] <annevk>
just like we have for setRequestHeader()

[01:52:01.0000] <Hixie>
blacklists are bad

[01:52:02.0000] <Hixie>
as a general rule

[01:52:03.0000] <Hixie>
from a security standpoint

[01:53:00.0000] <annevk>
true

[01:53:01.0000] <annevk>
maybe it's not so bad to limit this to just this

[01:57:00.0000] <hsivonen>
I read Joe Clark writing that the WCAG specifiers struggled with machine-checkability until realizing that their problem domain needs human checking

[01:58:00.0000] <hsivonen>
I wonder if there are citable sources for this other than Joe's inflammatorily titled ALA article

[01:58:01.0000] <annevk>
didn't he also state that the group was "dominated" by automatic accessibility checker companies?

[01:59:00.0000] <hsivonen>
IIRC something along those lines

[02:20:00.0000] <othermaciej_>
a whitelist is certainly less risky

[02:54:00.0000] <Lachy_>
/me finds it hard to believe Justin James' proposal is serious.

[02:55:00.0000] <Hixie>
curl: (6) Couldn't resolve host 'html5.validator.nu'

[02:55:01.0000] <Hixie>
that seems bad

[02:58:00.0000] <hsivonen>
Hixie: resolves for me from two autonomous systems

[02:59:00.0000] <Lachy>
hsivonen, re your email subject example, good mail clients warn before sending an email without a subject because most of the time it happens, it's an accidental omission that the sender would prefer to fix

[03:00:00.0000] <hsivonen>
Lachy: but the reason they prefer to fix it isn't on a spec's sayso but because they need it for communicating with the people they are seeking to communicate with

[03:00:01.0000] <Lachy>
although, some insert something like (no subject) by default

[03:00:02.0000] <Lachy>
ok

[03:01:00.0000] <hsivonen>
experience suggests that people write bad subjects if they are new to email or if they know they are in a position of power so that recipients will want to read the email based on sender name

[03:03:00.0000] <hsivonen>
now the problem accessibility advocates have is that alt is for the benefit of a third party when the author seeks to communicate in images with someone who will consume the images

[03:08:00.0000] <jgraham_>
FWIW, I often write bad subjects when e.g. emailing friends as the content of the mail is often just a collection of miscellany rather than one specific topic (and, honestly, I hope they would read it in any case ;) )

[03:09:00.0000] <jgraham_>
Mailing lists are a different audience and so are treated differently

[03:09:01.0000] <hsivonen>
jgraham_: I guess that falls under the case of getting read due to sender identity even if not necessarily due to a position of power

[03:09:02.0000] <jgraham_>
hsivonen: Yeah

[03:10:00.0000] <Lachy>
on high volume mailing lists, sometimes it's easier to read based on sender identity than based on the subject, especially when some people are known to write nonsense

[03:11:00.0000] <zcorpan>
Lachy: yeah. for instance, i never read your emails. ;)

[03:13:00.0000] <othermaciej>
hah

[03:13:01.0000] <othermaciej>
mean

[03:13:02.0000] <Lachy>
that's fine. I know some subscribers on some mailing lists who only read mine and a few others ;-)

[03:19:00.0000] <othermaciej__>
hsivonen: I just used validator.nu for the sake of blind conformance badge value; please forgive me

[03:21:00.0000] <hsivonen>
othermaciej: was the quality of your markup improved as a result?

[03:22:00.0000] <othermaciej>
hsivonen: not really, I just replaced spaces in some img src attributes with %20

[03:26:00.0000] <hsivonen>
was HyTime ever deployed anywhere?

[03:27:00.0000] <othermaciej>
what is HyTime?

[03:28:00.0000] <annevk>
spaces in <img src> should be allowed

[03:28:01.0000] <annevk>
either IRIs need to be cleaned up or this concept Hixie introduced but did not define should do it

[03:28:02.0000] <othermaciej>
HTML would have to make URI attribute values not be URIs in that case

[03:28:03.0000] <othermaciej>
(or IRIs)

[03:29:00.0000] <annevk>
it would just be some pre-processing

[03:29:01.0000] <zcorpan>
annevk: might slip through authoring mistakes in the form of misquoting

[03:30:00.0000] <zcorpan>
although now that " is not allowed in attribute names (and some other such checks), perhaps misquoted attributes would be catched anyway

[03:30:01.0000] <hsivonen>
othermaciej: HyTime is an ISO hypertext system based on SGML

[03:32:00.0000] <othermaciej>
literal space will work interoperably in attribute values of course, but not in many other URI contexts (if only because the space breaks attempts to autodetect them in text)

[03:32:01.0000] <othermaciej>
so I don't feel too bad about that requirement

[03:32:02.0000] <othermaciej>
on the other hand, I think it made my html less readable to use %20

[03:32:03.0000] <othermaciej>
and these are relative URIs in any case

[03:34:00.0000] <Hixie>
the biggest problem with spaces in href=""s that i've seen is that they risk getting wordwrapped

[03:34:01.0000] <Hixie>
and then who knows what'll happen

[03:39:00.0000] <hsivonen>
http://xml.coverpages.org/hytimeKimber9701.html

[03:39:01.0000] <hsivonen>
look for the paragraph starting with 'Technicaly, HyTime is an "enabling architecture".'

[03:41:00.0000] <Hixie>
HyTime is also a market failure.

[03:41:01.0000] <Hixie>
i wonder if there's a correlation

[03:42:00.0000] <othermaciej>
"it provides a set of 'meta-element types' that serve as templates or supertypes from which you derive your own element types with their own element type names and specialized attributes"

[03:42:01.0000] <othermaciej>
sounds sweet, sign me up

[03:43:00.0000] <othermaciej>
but it sounds like I should bring my space helmet

[03:45:00.0000] <hsivonen>
Hixie: do you have any reason why I shouldn't offer an option to get warnings whenever tags are inferred?

[03:46:00.0000] <Hixie>
none other than the usual "options are bad ui" thing

[03:46:01.0000] <Hixie>
if people want to be told when they use that syntax feature, then they should be able to obtain that information

[03:46:02.0000] <hsivonen>
yeah. I'm told I don't need to put it in the HTML5 UI as long as there's a magic bookmarkable URI parameter

[03:48:00.0000] <hsivonen>
the reasoning is that not having this feature is claimed to be an important evangelism blocker when advocating HTML5 to the XHTML-as-text/html crowd

[03:49:00.0000] <othermaciej>
hsivonen: I threw something at it with some (unintentional) errors and it seems to me the error recovery is not yet very good

[03:49:01.0000] <hsivonen>
othermaciej: 'it' being Validator.nu?

[03:49:02.0000] <othermaciej>
I forgot the close quote on some attributes

[03:49:03.0000] <othermaciej>
hsivonen: yeah

[03:50:00.0000] <hsivonen>
othermaciej: do you mean the recovery was bad or that the messages were bad?

[03:50:01.0000] <hsivonen>
(the recovery is per spec)

[03:51:00.0000] <othermaciej>
I mean that it gave lots of error messages for what was actually one mistake, and they created so much noise I could not spot my second mistake until I fixed the first

[03:51:01.0000] <othermaciej>
this is a UI complaint, not a standards compliance complaint

[03:51:02.0000] <hsivonen>
ok.

[03:52:00.0000] <hsivonen>
there seem to be two issues: 1) what the tokenizer spec says are errors (normative) and 2) how I phrase the errors (not normative; now phrased according to tokenization situation--not according to the probable user intent)

[03:52:01.0000] <hsivonen>
the old behavior with < breaking out of a tag and starting a new one was more intuitive

[03:54:00.0000] <hsivonen>
I wonder if it were worthwhile to stop reporting errors until there has been one transition to the data state

[03:55:00.0000] <hsivonen>
that would seem bad for general purpose perf and for Java byte code size, though...

[03:56:00.0000] <Hixie>
the spec only requires that you report one error if there is one and none if there are none, right?

[03:57:00.0000] <annevk>
spaces in URIs are problematic for ping= I guess

[03:57:01.0000] <hsivonen>
Hixie: yeah. but on the other hand, when I opt to stop reporting errors, people complain that the parser didn't keep on trucking

[03:57:02.0000] <Hixie>
hsivonen: well i just mean that you don't have to report exactly the right errors, if there are more useful ones you can report

[03:57:03.0000] <othermaciej>
does the spec require that if a conformance checker recovers from errors, it must report the same errors that the parsing algorithm would flag?

[03:58:00.0000] <othermaciej>
it seems to me that for usability's sake it's probably not good for the exact set of errors reported to be a normative requirement

[03:58:01.0000] <hsivonen>
I'm not sure, actually. It does require recovery to be per spec

[03:58:02.0000] <othermaciej>
since quality of error messages should be a potential point of competition

[03:59:00.0000] <Hixie>
Conformance checkers must report at least one parse error condition to the user if one or more parse error conditions exist in the document and must not report parse error conditions if none exist in the document. Conformance checkers may report more than one parse error condition if more than one parse error conditions exist in the document. Conformance checkers are not required to recover from parse errors.

[04:00:00.0000] <Hixie>
i can change the last sentence to allow arbitrary error recovery if you like

[04:00:01.0000] <othermaciej>
though in this particular case highlighting the likely actual authoring error and having the detailed error messages available in a distinguished form (indented, smaller text, disclosure triangle, something like that) would probably have been good enough

[04:01:00.0000] <othermaciej>
/me decides to run random pages on webkit.org through validator.nu

[04:02:00.0000] <othermaciej>
hmm the blog looks better than I thought

[04:02:01.0000] <othermaciej>
wrong doctype and it doesn't like the feed: URIs, and hyatt apparently would rather type <tt> than <code>

[04:04:00.0000] <hsivonen>
othermaciej: hyatt should use his editorship to make <tt> conforming :-)

[04:04:01.0000] <Lachy>
is <tt> commonly used?

[04:05:00.0000] <Hixie>
bed time

[04:05:01.0000] <Hixie>
nn

[04:05:02.0000] <hsivonen>
nn

[04:11:00.0000] <Philip`>
Lachy: http://canvex.lazyilluminati.com/survey/2007-07-17/analyse.cgi/tag/tt

[04:11:01.0000] <Philip`>
vs http://canvex.lazyilluminati.com/survey/2007-07-17/analyse.cgi/tag/code

[04:11:02.0000] <Philip`>
vs http://canvex.lazyilluminati.com/survey/2007-07-17/analyse.cgi/tag/samp etc

[04:11:03.0000] <hsivonen>
ouch. I may have missed conformance checking-relevant spec updates between 1478 and 1495

[04:12:00.0000] <Philip`>
(i.e. <tt> is on 30 pages out of 7739)

[04:12:01.0000] <Philip`>
(and you can decide on your own definition of "common" and whether <tt> counts)

[04:12:02.0000] <Lachy>
that's on more than <code>

[04:13:00.0000] <hsivonen>
yay. no changes annotated [c] in that range

[04:20:00.0000] <othermaciej>
one of those uses of <tt> is to create an invisible line spacer

[04:20:01.0000] <othermaciej>
on http://www.moytura.com/westport.htm

[04:20:02.0000] <othermaciej>
I should not be surprised I guess yet somehow I am

[04:21:00.0000] <othermaciej>
also a surprising number of these pages are in Polish

[04:22:00.0000] <othermaciej>
webkit.org homepage, the only error as html5 is the doctype

[04:23:00.0000] <Philip`>
http://validator.w3.org/check?uri=http%3A%2F%2Fwebkit.org%2F&charset=%28detect+automatically%29&doctype=Inline&group=0

[04:23:01.0000] <othermaciej>
Philip`: I'm forcing HTML5 mode

[04:23:02.0000] <othermaciej>
oh hmm

[04:23:03.0000] <Philip`>
Shouldn't duplicate IDs be an HTML5 error too?

[04:23:04.0000] <othermaciej>
yes

[04:23:05.0000] <hsivonen>
&group=0 ?

[04:24:00.0000] <Philip`>
hsivonen: That's just what it adds when I click the Check button

[04:24:01.0000] <othermaciej>
and it is definitely a duplicate id

[04:24:02.0000] <othermaciej>
hsivonen: does validator.nu have a bug tracker?

[04:25:00.0000] <othermaciej>
ok, found the bugzilla

[04:34:00.0000] <hsivonen>
Philip`: that's weird

[04:34:01.0000] <hsivonen>
Philip`: oh. sorry. it was w3.org

[05:08:00.0000] <othermaciej>
hsivonen: http://bugzilla.validator.nu/show_bug.cgi?id=184

[05:10:00.0000] <Philip`>
othermaciej: http://bugzilla.validator.nu/show_bug.cgi?id=149

[05:11:00.0000] <othermaciej>
darn, I thought I might have the first bug not filed by hsivonen 

[05:11:01.0000] <othermaciej>
(it looks like there are some others anyway, so I am not too sad that it's a dupe)

[06:06:00.0000] <Lachy>
wow, Node.lookupNamespaceURI has got some major interoperability issues :-(

[07:02:00.0000] <zcorpan>
Hixie: wrt setting reflecting booleans, it appears that IE and Opera have switched behavior with each other since last time I tested :-) 

[07:02:01.0000] <zcorpan>
Hixie: but safari and firefox still use the empty string

[07:41:00.0000] <hsivonen>
what's the difference between "cognitively impaired [user]" and "user with intellectual disabilities"?

[08:11:00.0000] <BenMillard>
hsivonen, cognitive impairments can include memory loss, emotional difficulties, mental health problems, the whole buffet of brain-related issues

[08:12:00.0000] <BenMillard>
hsivonen, intellectual disabilities isn't a term I've heard before but I imagine it's more specific...literacy and communication perhaps

[11:24:00.0000] <annevk>
alt madness continues :(

[11:24:01.0000] <Philip`>
The only way to resolve it is with a duel

[11:30:00.0000] <othermaciej>
I shouldn't have jumped into it

[11:47:00.0000] <webben>
re: the current alt discussion, it seems to me it's important to ascertain and distinguish between what UA+AT currently do (for backwards compatibility purposes) and what UA+AT should do (for ideal conformance).

[11:48:00.0000] <webben>
othermaciej and Faulkner are, AFAICT, both right. repetition is unwanted but differing navigational strategies may require a certain level of repetition in existing UA+AT.

[11:49:00.0000] <webben>
othermaciej: You mention how skipping the image might be useful for verbal output; I'm not sure that's necessarily the case (what about when you need to discover the image?), but in any case it seems to be something verbal UA+AT could implement using the proposed markup.

[11:50:00.0000] <webben>
e.g. skip the alt given when reading straight through the page.

[11:54:00.0000] <webben>
It's rather like the problem with link context, which is crucial when not navigating the page linearly, but potentially distracting when you do: http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2008-March/014258.html

[12:01:00.0000] <othermaciej>
webben: I don't see how the proposed markup would let you know that you can skip the alt

[12:02:00.0000] <othermaciej>
webben: I'm also not sure why navigating to the image has to be even available as an option, to a user consuming purely text

[12:02:01.0000] <webben>
othermaciej: If the aria-described-by or longdesc or whatever is defined as pointing to a full text alternative, and that alternative is in the current document, then the UA knows its reading the image alternative.

[12:02:02.0000] <othermaciej>
for a visually impaired but not blind user, or a cognitively disabled user, supplemental info would better expressed as title=""

[12:03:00.0000] <webben>
othermaciej: Navigating to images is important because images are content, and the user may wish to save the image/share it with others/edit it etc etc etc

[12:03:01.0000] <othermaciej>
webben: surely the same could be said for decorative images, and in that case alt="" is considered acceptable and will skip the image in many configurations

[12:03:02.0000] <webben>
othermaciej: decorative images aren't content; alt="" is fine.

[12:04:00.0000] <webben>
(yes, conceivably users might want access to decorative images, or CSS background-images, in both cases viewing or inspecting the source seems appropriate.)

[12:04:01.0000] <othermaciej>
visual UAs don't prevent users from saving decorative images, sharing them with others, or editing them (once saved)

[12:04:02.0000] <roc_>
it's kind of annoying when you're implementing a spec that has been implemented at least three times already and there are still basic ambiguities in the spec

[12:05:00.0000] <webben>
neither do non-visual UAs.

[12:05:01.0000] <webben>
the point is what's most useful to be alerted to.

[12:05:02.0000] <gsnedders>
othermaciej: OK, I _will_ use yacc (actually, I'll use PLY)

[12:05:03.0000] <othermaciej>
then I don't see why alt="" is ok for a decorative image but not for a non-decorative but redundant-with-the-text image

[12:06:00.0000] <othermaciej>
roc: what spec?

[12:06:01.0000] <webben>
because content is much more fundamental to general human communication and sharing than decorative images?

[12:06:02.0000] <othermaciej>
saving an image is not the common operation; you don't normally need to be alerted to it

[12:07:00.0000] <othermaciej>
unless for some reason screen reader users are much more likely to save images, which I grant is not impossible (don't really know)

[12:07:01.0000] <webben>
saving an image is only one of the scenarios I mentioned, but like I said: "You mention how skipping the image might be useful for verbal output; I'm not sure that's necessarily the case (what about when you need to discover the image?), but in any case it seems to be something verbal UA+AT could implement using the proposed markup."

[12:07:02.0000] <roc>
SVG flters, this time

[12:08:00.0000] <othermaciej>
lots of SVG has basic ambiguities in the spec

[12:08:01.0000] <othermaciej>
hardly specific to filters

[12:08:02.0000] <roc>
well yeah

[12:08:03.0000] <webben>
what's actually best for default behavior is perhaps best left to user testing by the UA+AT vendors and developers.

[12:08:04.0000] <roc>
I didn't say I was surprised :-)

[12:09:00.0000] <roc>
but you'd think they would have discovered interop problems and fixed the spec

[12:09:01.0000] <roc>
maybe there's a graphics programmer mindset that makes them all take the same decisions

[12:10:00.0000] <othermaciej>
webben: so you don't subscribe to the "behavior of current AT will never change, therefore the spec must cater to it" school of thought?

[12:11:00.0000] <webben>
othermaciej: The spec needs to cater to the behavior of current AT as a matter of backwards compatibility (it does this generally for browsers, and it's more important for AT because it does more harm to individuals who are locked into expensive AT if it doesn't); I think it also needs to look forward to the improvement of AT.

[12:11:01.0000] <webben>
"cater" here does not necessarily mean "optimise for"

[12:11:02.0000] <Philip`>
roc: There's quite a lot of bits in the canvas spec too, where the spec says pretty much nothing but everyone implements it the same way anyway

[12:12:00.0000] <Philip`>
(e.g. how you draw a stroked line)

[12:12:01.0000] <roc>
unfortunately if you try to specify that too much you will kill performance

[12:12:02.0000] <Philip`>
and I don't know if it's worth adding the necessary large amounts of complexity to the spec to describe precisely how to implement these things

[12:14:00.0000] <othermaciej>
webben: to me, speaking duplicate content seems like a pretty bad bug (perhaps as bad as occasionally reading gibberish filenames)

[12:14:01.0000] <othermaciej>
webben: and it is a bug that would occur in all existing and near-term foreseeable UA+AT combinations

[12:14:02.0000] <webben>
what's "near-term" here?

[12:15:00.0000] <othermaciej>
further, if this changes in some assistive technologies over time, this makes life harder for designers who actually want to create a good experience for e.g. screen reader users

[12:16:00.0000] <othermaciej>
there are probably few who care to the point of going beyond a simple checklist, and probably even fewer who test with one or more screen readers

[12:16:01.0000] <othermaciej>
so maybe that's not so important

[12:16:02.0000] <webben>
 /if/ it's a bug, I don't see why the implementation couldn't change to skip alt in certain situations, well, tomorrow e.g. with a commit to Orca or NVDA.

[12:18:00.0000] <othermaciej>
I'd love to work on a project that could give consistent one-day turnaround even just for simple bugs

[12:18:01.0000] <webben>
You've never seen a bug fixed in one day on WebKit?

[12:19:00.0000] <othermaciej>
I've seen plenty fixed in one day, a few one day after being reported, a tiny handful where that happend and the fix did not cause a regression, and 0 where the fix was shipped to end users that day

[12:20:00.0000] <jwalden>
bugs get fixed in one day, but only a very small portion of them

[12:20:01.0000] <webben>
a commit to Orca/NVDA != "fix was shipped to end users"

[12:20:02.0000] <othermaciej>
that's why I said "consistent one-day turnaround"

[12:20:03.0000] <webben>
I mean a source commit.

[12:20:04.0000] <othermaciej>
sure, some things could get fixed in one day, but very few things will and it is hard to control which

[12:20:05.0000] <Lachy>
othermaciej, would a DOMStringMap, which is partially defined in HTML5, work for the NSResolver?

[12:21:00.0000] <jwalden>
it'd be much easier if you could forego reviews, but that has its own problems :-)

[12:21:01.0000] <Lachy>
if there were a way for authors to create one.

[12:22:00.0000] <othermaciej>
Lachy: DOMStringMap looks like a spec for an implementation object with magical name lookup rules, not for a convenient author-provided static data map

[12:22:01.0000] <Lachy>
yeah, I guess.

[12:23:00.0000] <othermaciej>
for webkit we mandate reviews, passing all the tests, adding a regression test for any behavior change, and performance testing, but some things still manage to get fixed in a day

[12:23:01.0000] <othermaciej>
not many though!

[12:23:02.0000] <Philip`>
Is {'html':'http://...', 'svg':'http://...'} not an adequate convenient author-provided static data map?

[12:23:03.0000] <othermaciej>
and accessibility is an area that is so far lacking in regression testing in WebKit so scarier to work on

[12:24:00.0000] <othermaciej>
Philip`: Bjoern objected to it on the basis of (a) setters reintroducing possibility of random code execution (I say ignore them) (b) not scaling to non-JS languages (I don't care) and (c) not having an obvious way to represent the default namespace

[12:25:00.0000] <othermaciej>
I don't think you can name a JS property the empty string, but you could probably come up with a name that is not a valid namespace prefix and reasonably conveys defaultness

[12:25:01.0000] <othermaciej>
(perhaps '_' would do)

[12:26:00.0000] <Philip`>
Firefox seems happy enough with {'':1}

[12:27:00.0000] <Lachy>
Philip`, that can't work because of getters { get svg() { doSomething(); return "..." };

[12:27:01.0000] <Philip`>
Lachy: Is it not possible to just define it to ignore getters?

[12:28:00.0000] <Philip`>
If the property has a getter and is not just a raw value, then throw an exception or something, except express that in crazy ECMAScript notation with lots of double-square-brackets

[12:29:00.0000] <Philip`>
s/If/Or if/

[12:30:00.0000] <othermaciej>
ah, right, I meant getters, not setters

[12:30:01.0000] <Lachy>
the other solution that was suggested on public-webapi was to use a string like this "xht=http://www.w3.org/1999/xhtml svg=http://www.w3.org/2000/svg"

[12:30:02.0000] <othermaciej>
ES3 doesn't have getters per spec, ES4 may or may not have an in-language way to bypass them, the implementation can bypass in any case if it wants to

[12:32:00.0000] <Lachy>
the string is syntacally nicer than all the JS Object stuff, and it also doesn't suffer from any compatibility issues in non-ECMAScript langauges

[12:32:01.0000] <Philip`>
othermaciej: If there is no way for users to to bypass them, could implementations optimise the objects by not storing the hidden-by-getter properties at all?

[12:32:02.0000] <Philip`>
(which would break if there were special cases in the DOM that did require them to be bypassed)

[12:32:03.0000] <Lachy>
but I still need to check with Opera developer internally to see what they think of each solution

[12:32:04.0000] <othermaciej>
Philip`: what hidden-by-getter properties?

[12:33:00.0000] <othermaciej>
a property is either a regular property or it is programmatic

[12:33:01.0000] <othermaciej>
can't be both

[12:33:02.0000] <Philip`>
othermaciej: Ah, okay

[12:34:00.0000] <Philip`>
othermaciej: When you say "bypass", does that mean just detecting whether the property is regular or programmatic, as opposed to bypassing the getter in order to get at a regular property behind it?

[12:34:01.0000] <othermaciej>
Philip`: right, I mean getter properties are ignored for purposes of building the prefix mapping table

[12:35:00.0000] <othermaciej>
I think the string might be better though

[12:35:01.0000] <othermaciej>
certainly simplifies things specwise

[12:35:02.0000] <othermaciej>
though having to parse it will add maybe a tiny bit of overhead

[12:42:00.0000] <andersca>
Hixie: is there a suggested file extension for cache manifests

[12:43:00.0000] <jwalden>
/me doesn't see anything particularly onerous about a function, fw (little) iw

[12:43:01.0000] <jwalden>
but then again, I wouldn't touch namespaces if I could help it anyway, so I'm not the target audience

[12:47:00.0000] <annevk>
grmbl

[12:47:01.0000] <annevk>
http://lists.w3.org/Archives/Public/public-html/2008May/0228.html is pretty useless

[12:47:02.0000] <annevk>
he completely quotes Dave Singer out of context and then goes on with his alt= should be mandatory meme

[12:48:00.0000] <annevk>
they should really stop with their rhetoric and come with arguments...

[12:58:00.0000] <jgraham_>
Lachy: FWIW I think the string NSResolver is uglier than the object

[13:00:00.0000] <Lachy>
why?

[13:20:00.0000] <andersca>
Hixie?

[14:13:00.0000] <Hixie>
andersca: .manifest, i guess

[14:13:01.0000] <andersca>
Hixie: OK!

[14:14:00.0000] <andersca>
Hixie: from reading section 4.8, point 7 it seems that a resource can be stored in one or more application caches even if the browsing context is not associated with a cache

[14:15:00.0000] <Dashiva>
"We need to nail @alt to the mast!" ... "I keep an open mind"

[14:15:01.0000] <Dashiva>
@annevk and the rest of the people taking time to discuss it, you have my sympathy :)

[14:17:00.0000] <gsnedders>
jgraham_, Philip`: Do either of you know how common it is to change from Comp.Sci. to Nat.Sci. at the end of first year, especially among those who only did one science at a-level/equiv.?

[14:19:00.0000] <Hixie>
andersca: yeah, if a.html and b.html both reference different manifests that put c.html into their cache, then when you navigate to c.html, it'll be in two caches. But if in those caches it doesn't have a manifest="" attribute, then the document will first be loaded from one of those caches, and then the UA will realise it doesn't have a matching manifest, and will reload it from the network.

[14:21:00.0000] <jgraham_>
Lachy: Because the string isn't such a natural representation of the datatype which is basically a set of (key, value) pairs. Working with a string in code is harder than working with a structure that maps naturally onto the actual data. Also namespace URIs tend to be long and breaking a string over multiple lines is harder than doing the same with an object or other data structure — it's less likely that an IDE will have good indentatio

[14:21:01.0000] <jgraham_>
n for example.

[14:21:02.0000] <andersca>
Hixie: but also if the URL of c.html matches an opportunistically cached namespace

[14:21:03.0000] <andersca>
Hixie: right?

[14:21:04.0000] <jgraham_>
gsnedders: Not off the top of my head

[14:22:00.0000] <gsnedders>
jgraham_: looking on cam.ac.uk, it lists Maths & Phys. and Comp.Sci. as the two other routes into Nat.Sci. second year, but I can see no more

[14:22:01.0000] <Philip`>
gsnedders: Which version of CS? I believe the 25% option is for people who expect to continue into NatSci after the first year, while the 50% one is for people who expect to stay with CS, and I think they're adding the 75% soon for crazy people who really like CS

[14:23:00.0000] <jgraham_>
Basically if you want to do Physics + Maths Nat. Sci. in IB (second year) and you did reasonably well in 1A Physics/Maths no one will care about anything else

[14:23:01.0000] <gsnedders>
Philip`: I see only one on the website :P

[14:23:02.0000] <Hixie>
andersca: if you're offline, and you try to load a page as a fallback page for a top-level browsing context and it doesn't have a manifest="" matching the cache's manifest, then the load will be canceled and the browsing context will instead show an inline error message

[14:23:03.0000] <gsnedders>
Philip`: 50% it seems to be

[14:23:04.0000] <Philip`>
jgraham_: Is it possible to do entirely physics in IB?

[14:24:00.0000] <jgraham_>
gsnedders: For you 50% Comp. Sci. + Maths B + Physics seems right for the first year

[14:24:01.0000] <jgraham_>
(I think Maths B is the harder one)

[14:24:02.0000] <Hixie>
andersca: it is impossible for a page load to complete (or even get past the <html> implied start tag) in a top-level browsing context without there being an <html> start tag with a manifest="" pointing to the manifest of the cache from which it was loaded

[14:24:03.0000] <jgraham_>
Philip`: Yeah, you're basically expected to

[14:24:04.0000] <gsnedders>
Then I have the fun of choosing a college to apply to :P

[14:24:05.0000] <andersca>
Hixie: if I'm online, and load a page from the network and its url matches an opportunistic namespace 

[14:25:00.0000] <andersca>
Hixie: then that page needs to be stored regardless of whether the browsing context is associated with a cache or not

[14:25:01.0000] <jgraham_>
(I think technically the courses are Physics and Advanced Physics but few people do one and not the other)

[14:25:02.0000] <jgraham_>
(at least people who carry on with Physics in Pt II, and few people do IB physics but not Pt II physics)

[14:26:00.0000] <gsnedders>
(As I said on Friday, I don't meet the entrance requirements for Nat.Sci., so I can't apply for it (well, I can, it's just pointless))

[14:26:01.0000] <Philip`>
gsnedders: The http://www.cl.cam.ac.uk/admissions/undergraduate/intro/ site? That seems unclearer when I last looked, but it's mostly talking about what I'd call 50% and the "Alternative first year option" is 25%

[14:26:02.0000] <jgraham_>
gsnedders: The important thing about applying to a college is to remember that the criteria that you would apply in retrospect have little correlation with the information that is avaliable before you apply

[14:27:00.0000] <gsnedders>
Philip`: http://www.cam.ac.uk/admissions/undergraduate/courses/compsci/index.html

[14:27:01.0000] <gsnedders>
jgraham_: I'm aware. I should probably try and get criteria off my sister (who went to Selwyn)

[14:27:02.0000] <gsnedders>
Her reasons were silly :P

[14:28:00.0000] <Hixie>
andersca: hm yeah, step 7 of the navigate steps is broken

[14:28:01.0000] <gsnedders>
Philip`: That alternate option seems to be doing Nat.Sci. as a main subject, so applying for it, I think

[14:28:02.0000] <jgraham_>
The hard and fast rule is do not, under any circumstances, go with the "no preference" option

[14:28:03.0000] <andersca>
Hixie: for performance reasons it seems bad to do this :)

[14:28:04.0000] <gsnedders>
jgraham_: I've heard that time and time again, and quite frankly, I don't want to :)

[14:28:05.0000] <Philip`>
My criteria involved excluding colleges that were too far away, and excluding ones with names that sounded silly, etc, and then picking semi-arbitrarily from those that were left

[14:28:06.0000] <Hixie>
andersca: it's also broken in that if you try to fetch something that's oppcached and you get a 404, then you'll use the fallback, but if the fallback doesn't have a correct manifest="", it'll loop as it tries to use the network again

[14:29:00.0000] <andersca>
Hixie: good point

[14:29:01.0000] <gsnedders>
My sister wanted somewhere which could accommodate her for her entire degree, and taught a fair amount of it on site.

[14:29:02.0000] <jgraham_>
Philip`: That sounds quite like my criteria. Although I was also heavilly influenced by the fact that Jesus isn't on the tourist trail and had a sculpture exhibition on when I went to look around

[14:29:03.0000] <Hixie>
andersca: can you send mail?

[14:29:04.0000] <gsnedders>
Philip`: Would you, in retrospective, say King's was good?

[14:29:05.0000] <andersca>
Hixie: yes

[14:29:06.0000] <Hixie>
andersca: thanks

[14:30:00.0000] <Hixie>
andersca: woul it be helpful to you if i went and did appcache stuff next?

[14:30:01.0000] <jgraham_>
gsnedders: I think everywhere has accommodation for the whole degree now

[14:30:02.0000] <Philip`>
King's has the (seemingly quite rare) advantage of not having any stupid network restrictions, though I don't know if they'll change in the future

[14:30:03.0000] <andersca>
Hixie: I thought you already did :)

[14:30:04.0000] <jgraham_>
Having a tutor for you subject is a good idea

[14:30:05.0000] <gsnedders>
jgraham_: Well, my sister applied in '91 (before I was born)

[14:30:06.0000] <Hixie>
andersca: well you're about to send new feedback :-)

[14:30:07.0000] <andersca>
Hixie: ah :)

[14:30:08.0000] <andersca>
Hixie: I do have some feedback regarding dynamic entries too

[14:31:00.0000] <gsnedders>
Philip`: what sort of restrictions?

[14:31:01.0000] <Hixie>
there's apparently 19 e-mails in my appcache pile

[14:31:02.0000] <gsnedders>
jgraham_: you know anywhere where you could find that, apart from just nagging each college?

[14:31:03.0000] <Hixie>
mostly stuff from public-html

[14:31:04.0000] <Hixie>
so probably not useful

[14:31:05.0000] <Hixie>
for you

[14:32:00.0000] <Philip`>
gsnedders: I've been quite happy at King's so far, and I've never wished I was elsewhere :-)

[14:32:01.0000] <andersca>
Hixie: I think it's a bit weird that dynamic entries have some kind of order

[14:32:02.0000] <jgraham_>
gsnedders: It used to say in the prospectus

[14:32:03.0000] <Hixie>
andersca: well they need some sort of order to be enumerable, no?

[14:32:04.0000] <Hixie>
andersca: i don't have a strong opinion on that at all

[14:32:05.0000] <Hixie>
andersca: if you have a better idea, let me know, i'll use it

[14:32:06.0000] <Hixie>
andersca: put a note in the e-mail(s) you send reminding me that you'd like prompt responses and i'll reply when you send them

[14:33:00.0000] <andersca>
Hixie: will do!

[14:33:01.0000] <Hixie>
thanks

[14:33:02.0000] <Hixie>
bbiab

[14:33:03.0000] <gsnedders>
Another requirement for my sister was that she wouldn't be with people who were ridiculously rich who spent £10k on drink every weekend (as she rarely drank, and never to the extremes of many, due to not having the money)

[14:33:04.0000] <Philip`>
gsnedders: Some colleges have non-public IPs, proxies, firewalls, small bandwidth limitations, etc

[14:33:05.0000] <jgraham_>
gsnedders: Yeah it says in the prospectus

[14:33:06.0000] <gsnedders>
Philip`: Ah

[14:33:07.0000] <jgraham_>
"Director of studies"

[14:34:00.0000] <jgraham_>
not every college has them for evey subject

[14:34:01.0000] <jgraham_>
gsnedders: All the stuff about people from a certain class background being concentrated in a subset of colleges is mostly a myth

[14:35:00.0000] <jgraham_>
That is a bad reason for choosing a college

[14:35:01.0000] <gsnedders>
/me is quite tempted to apply for King's, but has no idea how well he'd manage to get in

[14:35:02.0000] <Philip`>
gsnedders: http://www.cl.cam.ac.uk/teaching/dos-list/ lists the DoSes of each college, though you can't read that page

[14:36:00.0000] <gsnedders>
Philip`: each of http://www.cam.ac.uk/admissions/undergraduate/colleges/ lists the DoSes

[14:36:01.0000] <jgraham_>
gsnedders: By going to the interview I expect

[14:36:02.0000] <gsnedders>
jgraham_: You really think I'd get in fine? :P

[14:37:00.0000] <jgraham_>
(seriously, I don't know what your grades are like but you have a lot of interest in your subject so you would stand a good chance)

[14:37:01.0000] <gsnedders>
jgraham_: (my grades are varied)

[14:37:02.0000] <jgraham_>
(disclaimer: I Am Not An Admission Officer)

[14:37:03.0000] <gsnedders>
(We will find out how I've done this year in August)

[14:37:04.0000] <jgraham_>
s/Admission/Admissions/

[14:38:00.0000] <gsnedders>
My parents know the admissions officer of Selwyn (totally unrelated to the fact my sister went there, too)

[14:39:00.0000] <gsnedders>
/me starts to enter an ISO-8601 date into the request form for a paper copy of the prospectus, then realises it says "dd/mm/yy"

[14:39:01.0000] <Philip`>
Apparently there were 223 CS applications in 2007, and 71 acceptances

[14:39:02.0000] <Philip`>
/me thought it used to be more like 5:1

[14:40:00.0000] <gsnedders>
See, that's what makes me worried. :P

[14:40:01.0000] <Philip`>
(200 male, 23 female)

[14:41:00.0000] <Philip`>
(which is more extreme than any other course)

[14:41:01.0000] <jgraham_>
(oh and if we're comparing netowrk access Jesus was very proud of the fact that it was the first college to put the internal network behind NAT but other than that has been very good to me, especially as a graduate where they have left me alone for 4 years without trying to charge me for any unwanted meals)

[14:42:00.0000] <gsnedders>
Philip`: Well, maybe that'll stop me tagging blog posts with lust :P

[14:42:01.0000] <jgraham_>
I guess the above makes more sense with s/network access/colleges/

[14:43:00.0000] <Philip`>
http://computing-info.ucam.org/Colleges might be either up-to-date or of highly questionable reliability

[14:43:01.0000] <jgraham_>
gsnedders: If that's your aim you should apply to be a medic, but definitely not a vet 

[14:43:02.0000] <gsnedders>
Really, it seems to be between Edinburgh or Cambridge (even though I will likely apply to at least York as well)

[14:43:03.0000] <gsnedders>
jgraham_: Why? What are they like?

[14:44:00.0000] <jgraham_>
gsnedders: According to a Varsity (student newspaper) survey I heard about second hand, Medics get the most sex, Vets the least

[14:44:01.0000] <gsnedders>
jgraham_: ah

[14:45:00.0000] <gsnedders>
jgraham_: So they test the theories they learn about spread of STDs, whereas vets are of the wrong species to do so?

[14:46:00.0000] <Philip`>
I think I saw a piece in another student newspaper from the person who happened to be hanging around when Varsity needed to fill some space and they dragged him into reporting that rather unreliable survey and he/she was surprised when it got attention in the national press

[14:46:01.0000] <Philip`>
s/him/him\/her/ (since I can't remember at all)

[14:47:00.0000] <gsnedders>
Why is it that all .ac.uk addresses are fussy about (www.)?

[14:47:01.0000] <gsnedders>
or rather, (www\.)?

[14:48:00.0000] <jgraham_>
gsnedders: Because they are evil

[14:48:01.0000] <Philip`>
Probably because they're old networks and older network administrators who resist any change, and in the old days they had to use different domain names to separate out HTTP and FTP and Telnet and Gopher and whatever

[14:49:00.0000] <jgraham_>
Philip`: I think it's university policy. At least I remember reading that once, although it may have changed

[14:49:01.0000] <gsnedders>
/me stops ordering paper prospectuses (mainly for the sake of my mother) having got to three

[14:50:00.0000] <gsnedders>
jgraham_: Where is the Jesus accommodation? Within the college site?

[14:50:01.0000] <jgraham_>
Philip`: I'm not surprised if the survey was wrong

[14:51:00.0000] <jgraham_>
gsnedders: Mostly, yes. Second years are mostly in college-owned houses on Jesus Lane / Malcolm street (right next to college) which is actually better than being in college

[14:51:01.0000] <Philip`>
The WWW might just be a passing fad, so why should they dedicate a high-level domain like "cam.ac.uk" to it?

[14:51:02.0000] <jgraham_>
I get the impression that Jesus has one of the best accommodation stories in the whole of Cambridge

[14:52:00.0000] <jgraham_>
(also, apparently theology was bottom)

[14:52:01.0000] <gsnedders>
/me re-reads the website, and realises he does actually meet the nat.sci. entrance requirements

[14:53:00.0000] <gsnedders>
(taking AH as equal to A-Level, which is untrue as AH is harder)

[14:53:01.0000] <jgraham_>
Some colleges have people in buildings far from the college or rooms above the bus station/gardies/etc.

[14:53:02.0000] <gsnedders>
jgraham_: Yeah, that's a large part of my sister's choice

[14:54:00.0000] <gsnedders>
/me will have far too much to do even over two days of open-days

[14:55:00.0000] <jgraham_>
gsnedders: Also, try to find out how the college deals with meals. Most have a "Kitchen Fixed Charge" that you have to pay every term

[14:55:01.0000] <jgraham_>
This can be in the form of meal vouchers (generally bad) or price discounts

[14:55:02.0000] <gsnedders>
Ergh. Fun. :\

[14:56:00.0000] <gsnedders>
/me expects he'll likely end up down for the 3rd/4th July open days, if either of you are around then :P

[14:56:01.0000] <Philip`>
Currently the KFC at King's is £85/term, in exchange for a discount over guest rates which only saves you money if you get two meals a day there

[14:56:02.0000] <jgraham_>
Caius has meal vouchers and some of the worst food in the known universe. But the students still eat there because otherwise they fell they've wasted their money

[14:56:03.0000] <annevk>
via MikeSmith: http://trac.webkit.org/changeset/33033

[14:56:04.0000] <annevk>
interesting

[14:56:05.0000] <Philip`>
(though I think there are potentially-implementable plans to change things a bit in the future)

[14:57:00.0000] <jgraham_>
Philip`: In Kings or in general?

[14:57:01.0000] <Philip`>
jgraham_: In King's

[14:57:02.0000] <jgraham_>
/me has an officemate who, as a grad, has to pay a KFC at Lucy and /hates/ them for it

[14:57:03.0000] <gsnedders>
jgraham_: Caius?

[14:58:00.0000] <jgraham_>
http://www.cai.cam.ac.uk/

[14:58:01.0000] <Philip`>
/me doesn't care about the KFC since it's quite minimal compared to other expenses

[14:58:02.0000] <gsnedders>
KFC? eww. :P

[14:58:03.0000] <Philip`>
and since I don't have anything else to spend my money on

[14:59:00.0000] <jgraham_>
I don't pay it atm because I don't live in college but I din't mind when I did live in college

[14:59:01.0000] <jgraham_>
But then I don't generally have an issue with taxation in exchange for services

[14:59:02.0000] <jgraham_>
As long as the services are worthwhile

[15:00:00.0000] <Philip`>
None of it is my money anyway, the government gave it to me :-)

[15:03:00.0000] <gsnedders>
jgraham_: Now I'm confused. You're at Jesus, right? Or am I confused?

[15:03:01.0000] <annevk>
Hixie, thought you were in favor of not having namespaces for the Selectors API? :)

[15:04:00.0000] <Hixie>
yeah but apparently they're not going away

[15:04:01.0000] <jgraham_>
gsnedders: Yeah, I'm at Jesus. 

[15:04:02.0000] <jgraham_>
But Caius food is legendary :)

[15:05:00.0000] <jgraham_>
(I ate there once when someone was trying to get rid of some meal tickets and got fed undercooked chicken with unpleasant BBQ sauce)

[15:06:00.0000] <gsnedders>
jgraham_: You're starting to make King's + KFC look _good_.

[15:06:01.0000] <gsnedders>
(and Edinburgh's catered accommodation look amazing)

[15:07:00.0000] <gsnedders>
/me heads off to bed, and warns jgraham_ and Philip` he'll be nagging them again :P

[15:08:00.0000] <jgraham_>
gsnedders: Everywhere has a KFC I think

[15:08:01.0000] <jgraham_>
It's just how they run it that changes

[15:08:02.0000] <gsnedders>
jgraham_: what _else_ do the colleges have, though?

[15:08:03.0000] <gsnedders>
Nothing?

[15:09:00.0000] <jgraham_>
gsnedders: KFC == Kitchen Fixed Charge, right?

[15:09:01.0000] <Philip`>
/me wonders if gsnedders is thinking 'KFC' means http://www.kfc.com/

[15:09:02.0000] <gsnedders>
Philip`: Yes ;P

[15:09:03.0000] <gsnedders>
/me is a n00b, remember?

[15:10:00.0000] <gsnedders>
Last day of teacher of this year of school tomorrow.

[15:10:01.0000] <Philip`>
Ah, it would be quite worrying if all the colleges had that kind of KFC :-p

[15:10:02.0000] <gsnedders>
Then no longer will I be in Secondary 5, and therefore no longer worthy of the WHATWG :P

[15:10:03.0000] <gsnedders>
Philip`: Yes, it would be. :P

[15:10:04.0000] <Philip`>
You would be no less worthy than anyone else here

[15:11:00.0000] <gsnedders>
/me hopes someone is moving into S5 here

[15:11:01.0000] <Philip`>
(assuming nobody else here is in Secondary 5, which seems a reasonable assumption)

[15:11:02.0000] <jgraham_>
Philip`: Indeed. I think even college food puts that kind of KFC to shame

[15:11:03.0000] <gsnedders>
Philip`: Hey! Scotland has a population smaller than London! Don't put down its education system!

[15:12:00.0000] <Philip`>
gsnedders: Anyway, good night :-)

[15:12:01.0000] <gsnedders>
(which is to say, it is unlikely there is anyone in S4)

[15:28:00.0000] <Philip`>
I wish browser developers didn't keep trying to release browsers

[15:28:01.0000] <Philip`>
It's a pain since there's never enough time to find bugs before they get shipped :-(

[15:28:02.0000] <othermaciej>
Philip`: there's always another release to fix 'em in

[15:29:00.0000] <gavin_>
hey, that bug fix might still yet make firefox 3

[15:30:00.0000] <roc>
I hope it doesn't, personally

[15:30:01.0000] <Philip`>
gavin_: But I found at least one more regression :-)

[15:30:02.0000] <annevk>
is firefox 3 still on wir schedule or is there a release month/date?

[15:31:00.0000] <gavin_>
"wir schedule"?

[15:31:01.0000] <gavin_>
oh, when it's ready

[15:31:02.0000] <gavin_>
there's no set release month/date

[15:31:03.0000] <gavin_>
so I guess it is still wir

[15:31:04.0000] <gavin_>
Philip`: filed yet?

[15:32:00.0000] <roc>
according to the platform blocker list, "when it's ready" is "now"

[15:32:01.0000] <roc>
for Gecko at least

[15:32:02.0000] <gavin_>
yes

[15:32:03.0000] <Philip`>
gavin_: Almost

[15:32:04.0000] <Hixie>
i hope someone replies to http://lists.w3.org/Archives/Public/public-html/2008May/0232.html and points out that none of the points raised in the five or six parent messages on that thread have been replied

[15:32:05.0000] <Hixie>
both boris and mjs were ignored

[15:33:00.0000] <annevk>
roc, gavin_, k, thanks

[15:33:01.0000] <gavin_>
rc1 is being built right now

[15:36:00.0000] <Philip`>
(gavin_: https://bugzilla.mozilla.org/show_bug.cgi?id=433397 )

[15:36:01.0000] <gavin_>
just saw that, yeah

[15:37:00.0000] <gavin_>
thanks

[15:37:01.0000] <Philip`>
/me ought to check more thoroughly for canvas regressions

[15:56:00.0000] <andersca>
Hixie: so

[15:56:01.0000] <andersca>
Hixie: let's say I'm loading a.html which has a cache manifest

[15:56:02.0000] <andersca>
Hixie: a.html is very slow to load, so I cancel the load

[16:00:00.0000] <Hixie>
yes?

[16:01:00.0000] <andersca>
what happens

[16:02:00.0000] <andersca>
a.html will of course not be in the cache as an implicit resource

[16:02:01.0000] <andersca>
should error be sent?

[16:02:02.0000] <Hixie>
the cache attempt fails

[16:08:00.0000] <Hixie>
actually that's related to an open issue

[16:08:01.0000] <Hixie>
i'll make sure it's clearer

[16:08:02.0000] <Hixie>
my intent was that it would cause the cache attempt or update attempt to fail

[16:11:00.0000] <andersca>
OK

[16:12:00.0000] <andersca>
Hixie: that makes sense

[16:14:00.0000] <Hixie>
glad you think so :-)

[16:15:00.0000] <Hixie>
that section needs an introductory section so bad

[16:25:00.0000] <andersca>
Hixie: it makes sense because it's easy to implement ;)

[16:31:00.0000] <Hixie>
Philip: please send feedback (to whatwg maybe...) pointing out that the spec should require something useful to be done with images with no alt in links

[16:31:01.0000] <Hixie>
or maybe with alt="" if that's the only contents

[16:31:02.0000] <Hixie>
or whatever the case was

[16:50:00.0000] <Philip`>
Hixie: That seems like just a user interface quality issue, with no effect on interoperability - why would the spec care about that?

[16:52:00.0000] <annevk>
whether or not to expose and potentially ignore what the author said seems like interop


2008-05-13
[17:05:00.0000] <Philip`>
/me hears a cat making really quite disturbing noises outside his window

[17:07:00.0000] <Dashiva>
This suggests your window is a blog

[17:08:00.0000] <Philip`>
It has enough security vulnerabilities to be a blog

[17:08:01.0000] <Philip`>
Someone ought to invent transparent bricks - those would be much better at keeping intruders out

[17:51:00.0000] <Hixie>
Philip`: fair enough

[18:31:00.0000] <Hixie>
wow, that thread was easy to deal with

[18:31:01.0000] <Hixie>
/me switches to sentences around to fix 13 e-mails

[18:32:00.0000] <Hixie>
two

[18:32:01.0000] <Hixie>
Lachy: the attribute name was intentionally stupid so that people would suggest new names

[18:34:00.0000] <Philip`>
I think the attribute should be named "!"

[18:34:01.0000] <Hixie>
well that would be exciting

[18:34:02.0000] <Philip`>
so you can write <img! src=... alt=..> for an important image

[18:34:03.0000] <Hixie>
in SO many ways

[18:35:00.0000] <Hixie>
well you couldn't do _that_

[18:35:01.0000] <Hixie>
you'd be creating an <img!/> element

[18:35:02.0000] <Philip`>
Whoops

[18:36:00.0000] <Hixie>
but i do like the idea of <img ! src='photo' alt=>

[18:36:01.0000] <Hixie>
'er

[18:36:02.0000] <Philip`>
That's spoilt my plan :-(

[18:36:03.0000] <Hixie>
but i do like the idea of <img ! src='photo' alt=Photo>

[18:37:00.0000] <Hixie>
the only things that end a tag name are >, /, whitespace, and EOF

[18:37:01.0000] <Philip`>
Hmm, it's even well-formed XML

[18:37:02.0000] <Philip`>
Oh

[18:37:03.0000] <Philip`>
No it's not

[18:37:04.0000] <Philip`>
I was accidentally testing it as text/html instead

[18:38:00.0000] <Philip`>
/me thought that seemed a bit odd

[18:59:00.0000] <Philip`>
annevk: I don't know if I've seen "interoperability" defined anywhere, so I'll make something up and say that interoperability is when authors can write pages that work in all UAs and developers can write UAs that work on all pages

[19:00:00.0000] <Philip`>
Then the linked-image case is interoperable regardless of how UAs handle <a><img alt=""></a>, because authors can write <a><img alt="Button label"></a> (which will work in all UAs) and developers can auto-generate some link text for <a><img alt=""></a> (which will work on all pages)

[19:02:00.0000] <Philip`>
(Not all authors *will* do that, and not all UAs *will* do t'other, but that isn't a lack of interoperability and it won't have negative effects beyond that page or that UA)

[19:05:00.0000] <Hixie>
andersca: spec updated to handle all feedback so far, though not what we discused earlier

[19:07:00.0000] <andersca>
Hixie: awesome - I'm fixing the case where you cancel the load of the implicit resource 

[19:07:01.0000] <Hixie>
that's in the spec now

[19:08:00.0000] <andersca>
oh, cool

[19:08:01.0000] <andersca>
/me reloads

[19:11:00.0000] <andersca>
Hixie: excellent

[19:11:01.0000] <andersca>
Hixie: what about step 2 of "If the resource being loaded was not loaded from an application cache, but it was loaded using HTTP GET or equivalent"

[19:12:00.0000] <andersca>
Hixie: if the implicit resource fails to load

[19:13:00.0000] <Hixie>
that's handled by step 21 of the application cache update process

[19:16:00.0000] <andersca>
Hixie: what if the cache manifest didn't change, then step 21 won't be reached

[19:18:00.0000] <Hixie>
andersca: oh you mean if this is a new implicit entry but the cache is already there...

[19:18:01.0000] <Hixie>
andersca: hmm

[19:18:02.0000] <andersca>
yes

[19:18:03.0000] <Hixie>
andersca: sure, will fix that

[19:18:04.0000] <andersca>
Hixie: what should happen in that case? 

[19:18:05.0000] <Hixie>
i guess it just doesn't get added

[19:18:06.0000] <andersca>
yeah

[19:19:00.0000] <andersca>
and there will be no error, since there's no update 

[19:23:00.0000] <andersca>
Hixie: maybe the document should no longer be associated with the cache in that case

[19:36:00.0000] <Hixie>
andersca: fixed the infinite loop

[19:36:01.0000] <andersca>
Hixie: excellent

[21:24:00.0000] <Hixie>
http://www.w3.org/2002/09/wbs/40318/dcp/ (as yet unannounced, i think, so don't vote on it yet) is going to be interesting. i wonder why that particular e-mail was chosen as a summary, for instance. and what weird effects there will be in casting the question in such a binary way.

[21:36:00.0000] <Hixie>
Lachy, annevk, anyone: do you remember how to get cvs access to dev.w3.org?

[21:37:00.0000] <heycam>
Hixie, i think you need to mail an ssh public key of yours to a Team person

[21:37:01.0000] <Hixie>
oh they still don't have a form for this?

[21:37:02.0000] <Hixie>
ok

[21:38:00.0000] <heycam>
not afaik

[21:38:01.0000] <Hixie>
do you remember what kind of key they want?

[21:38:02.0000] <Hixie>
1024 bit rsa?

[21:38:03.0000] <heycam>
don't know if they needed any specific key size / alg

[21:38:04.0000] <Hixie>
i seem to recall there was some problem with the first key i sent

[21:38:05.0000] <Hixie>
but this was years ago

[22:08:00.0000] <MikeSmith>
Hixie, heycam : only criterion for the key is that it needs to be SSH2

[22:08:01.0000] <MikeSmith>
SSH protocol 2

[22:09:00.0000] <Hixie>
k thanks

[22:09:01.0000] <MikeSmith>
it can rsa or dsa of any length

[00:11:00.0000] <annevk>
Philip`, you start getting content that relies on a particular UA behavior which is the problematic bit

[00:11:01.0000] <annevk>
Philip`, so it seems better to make it non-conforming

[00:59:00.0000] <zcorpan_>
/me considers setting up a filter to mark "alt" emails as read

[01:02:00.0000] <annevk>
hmm, all i tried to say was that HTML5 is not against "accessibility" and now i get ranted for that :(

[01:03:00.0000] <annevk>
seems like quite a hopeless case indeed

[02:26:00.0000] <jgraham_>
hsivonen: Thank you for pointing out that accessibility is clearly not a "cornerstone of the web" for any reasonable use of the metaphor. 

[02:27:00.0000] <jgraham_>
People really have to distinguish between what is actually true and what they would like to be true

[02:27:01.0000] <hsivonen>
indeed

[02:46:00.0000] <Hixie>
important="" isn't really the right name... we need something that means either:

[02:46:01.0000] <Hixie>
/me ponders

[02:47:00.0000] <Hixie>
i don't even know really want we want to convey

[02:47:01.0000] <Philip`>
You want to convey that the alt text is a description rather than an alternative presentation of the image

[02:47:02.0000] <Philip`>
(I think)

[02:48:00.0000] <Philip`>
Uh

[02:48:01.0000] <Philip`>
s/presentation/representation/

[02:49:00.0000] <Hixie>
not even really a description

[02:49:01.0000] <Philip`>
Hmm, but it should also convey that the image is non-decorative and no textual representation can exist, since those are the only cases where it's allowed

[02:50:00.0000] <Hixie>
we want to convey that the alt is that the image is known to be not decorative, and that alt is giving the category of image in question, not a replacement representation

[02:50:01.0000] <Hixie>
a textual representation might be able to exist, we just don't know what it is

[02:50:02.0000] <Philip`>
(I mean a description of the image as an object (e.g. "Photo"), rather than a description of the content of the photo)

[02:50:03.0000] <Hixie>
e.g. in the case of a photo uploaded by a blind user or hsivonen's image tool

[02:52:00.0000] <jgraham_>
If the text was a description of the image it could go in title

[02:52:01.0000] <jgraham_>
Oh you made that distinction

[02:54:00.0000] <Philip`>
Does it have to be an attribute, rather than e.g. <img src=... alt="Important image: Photo">?

[02:54:01.0000] <Philip`>
(which would not be machine-extractable information, but I'm not sure if machines need to extract this information)

[02:55:00.0000] <jgraham_>
What's the rationale for having this attribute at all? How would UAs use it? What problems would it fix?

[02:55:01.0000] <hsivonen>
Philip`: doesn't the heading and preceding paragraph in the image report already make the point that the images are important for the function of the tool?

[02:56:00.0000] <hsivonen>
Philip`: if a user can't use it but ends up wading through the content, how does "Important image: Photo" a dozen times help?

[02:57:00.0000] <jgraham_>
In particular how would this attribute deal with the problem of authoring tools adding bogus content to pass validation?

[02:58:00.0000] <Hixie>
the idea is that the attribute would have one effect on UAs

[02:59:00.0000] <Hixie>
namely, when the <img> element is being rendered as text instead of as an image, the image would be rendered in a way that clearly indicates that there was an image there

[02:59:01.0000] <Hixie>
instead of just silently replacing the image with text

[02:59:02.0000] <Hixie>
so <p>Hello <img src=world.png alt="World"> how are <img important src=you.png alt="Word">?</p> would render as something like:

[03:00:00.0000] <Philip`>
hsivonen: The alternative (i.e. <img src=...>) means the UA will have to apply heuristics, so it might ignore all the images or read all their filenames, or might ignore all spacer.gifs since they look decorative, which seems like a worse user experience than just telling the user that there's an image (as would happen with alt="Image")

[03:00:01.0000] <Hixie>
   Hello World how are [IMAGE: Word]?

[03:01:00.0000] <Hixie>
the idea being that you can still convey some sort of information, just not enough to act as a replacement

[03:01:01.0000] <hsivonen>
Philip`: are conceivable content-based heuristics worse?

[03:01:02.0000] <Philip`>
hsivonen: Users you can't see images would still find the image report tool useful to make sure they've added alt attributes to all their images (and if they've missed some then they could ask someone for suitable text)

[03:01:03.0000] <Hixie>
e.g. <img alt=Photo> is useless, but <img important alt=Photo> at least lets the user know there's an image there.

[03:01:04.0000] <Philip`>
s/you/who/

[03:01:05.0000] <hsivonen>
Philip`: are there notable UAs that read file name in the case of V.nu?

[03:01:06.0000] <jgraham_>
Hixie: So in fact this doesn't really help the AT case (since AFAIK they do that anyway), it just helps the text browser case

[03:02:00.0000] <Hixie>
jgraham_: I don't think current behaviour in ATs is really indicative of ideal behaviour in ATs, even in the face of existing legacy markup.

[03:02:01.0000] <hsivonen>
Philip`: yeah, but for that use case you only need to check if a particular heading appears in the report possible counting the rows in the table after it

[03:02:02.0000] <Hixie>
jgraham_: my limited experience suggests that ATs are horrificly bad at this point, and not in a manner gated by HTML's limitations.

[03:03:00.0000] <Philip`>
hsivonen: Lynx does

[03:03:01.0000] <Philip`>
hsivonen: e.g. http://validator.nu/?doc=http%3A%2F%2Fnews.bbc.co.uk&showimagereport=yes says "[_44646536_34e73f52-7689-477a-86d7-7d309cf9a480.jpg] Carlos Tevez, Ryan Giggs and Cristiano Ronaldo celebrate Manchester United's Premier League title victory From line 2817, column 73; to line 2817, column 344"

[03:03:02.0000] <jgraham_>
Hixie: Sure. But nevertheless they tend to supply this information with or without an additional attribute

[03:04:00.0000] <hsivonen>
Philip`: eww. is that the default now?

[03:04:01.0000] <Philip`>
hsivonen: I don't think I've set any non-default options

[03:04:02.0000] <Hixie>
jgraham_: possible

[03:04:03.0000] <Hixie>
jgraham_: but they don't have to

[03:05:00.0000] <Hixie>
jgraham_: especially if we provide this

[03:05:01.0000] <jgraham_>
Hixie: I'm just not sure what having an additional attribute buys you here

[03:05:02.0000] <Hixie>
jgraham_: yeah well neither am i really

[03:05:03.0000] <jgraham_>
Is this substantially different from the @noalt case, but with a semanticised name?

[03:05:04.0000] <Hixie>
jgraham_: i'm just trying to find a technically good solution to the problem that is being posited to exist with the current text

[03:06:00.0000] <Hixie>
jgraham_: noalt was defined to have no behaviour at all, and was defined to only be valid if alt= was missing

[03:06:01.0000] <Hixie>
jgraham_: so yes

[03:06:02.0000] <jgraham_>
I thought noalt was supposed to indicate a "critical content" image for which no alt text was available?

[03:07:00.0000] <Hixie>
yes but it was not defined to allow alt="" to be specified with it

[03:07:01.0000] <Philip`>
hsivonen: Any half-decent content-based heuristics ought to ignore <img src=spacer.gif>, which is bad in this case since it unexpectedly hides some of the image report's images from the user (when they might often want to know about the image so they can copy-and-paste the URI or something) - as the content producer you know more about whether the image is ignorable than the UA can know, and you have to use the alt attribute to communicate that knowled

[03:07:02.0000] <Hixie>
you couldn't do <img alt=photo noalt>

[03:07:03.0000] <Philip`>
s/$/ge/

[03:09:00.0000] <hsivonen>
Philip`: hmm. alt='image' would be truly ironic, but it would indeed make things better in Lynx

[03:09:01.0000] <jgraham_>
Hixie: So the @importantimage case would require alt and hence not address some of the cases where that is harmful e.g. a photo dragged into a WYSIWYG editor

[03:09:02.0000] <hsivonen>
but not in Safari+VO

[03:09:03.0000] <Hixie>
jgraham_: I don't know if it should require alt, necessarily

[03:10:00.0000] <Hixie>
jgraham_: but either alt, or important, or both, would need to be specified

[03:10:01.0000] <Philip`>
hsivonen: It would help in Opera+VO too, since that ignores images with no alt

[03:10:02.0000] <zcorpan>
Philip`: if the images in the image report were links, would UA still ignore spacer.gif?

[03:10:03.0000] <Philip`>
hsivonen: (at least while I'm believing that it's bad for the images to be ignored)

[03:10:04.0000] <Hixie>
jgraham_: though in the case of a wysiwyg editor, nothing that's been proposed so far would help, since you have no way to know if it's decorative or not

[03:10:05.0000] <Hixie>
jgraham_: and the spec currently doesn't allow you to not know that

[03:11:00.0000] <zcorpan>
Philip`: <a href=http://www.example.org/spacer.gif><img src=http://www.example.org/spacer.gif></a> is different from <img src=http://www.example.org/spacer.gif>

[03:11:01.0000] <Hixie>
fundamentally the problem with important="" boils down to the fact that you have to define the behaviour for when you have neither alt="" nor important=""

[03:11:02.0000] <jgraham_>
Hixie: The safe assumption is that it's not decorative

[03:12:00.0000] <Hixie>
jgraham_: not necessarily -- it would be annoying as all hell to have "image" read out a billion times

[03:12:01.0000] <Philip`>
zcorpan: The UA shouldn't ignore the whole link, but I don't see why it couldn't ignore the spacer.gif and treat that pattern the same as <a href></a> and convert it to text as "Empty link"

[03:12:02.0000] <jgraham_>
Hixie:  It basically seems like missing alt would map to important

[03:12:03.0000] <Hixie>
...and (continuing my earlier thought) you therefore end up making the default the same as including important="", and the attribute becomes mostly useless, unless we consider saying the kind of image to be a significant win

[03:13:00.0000] <Hixie>
right

[03:13:01.0000] <jgraham_>
s/missing alt/misgging alt and important/

[03:13:02.0000] <jgraham_>
s/gg/ss/

[03:13:03.0000] <Philip`>
s/ss/s/

[03:13:04.0000] <zcorpan>
Philip`: why would it be useful to do that?

[03:13:05.0000] <jgraham_>
Philip`: :-p

[03:14:00.0000] <Philip`>
zcorpan: Which "that"? The ignoring-the-image bit, or the rendering-empty-links-nonemptily bit?

[03:14:01.0000] <zcorpan>
Philip`: ignoring-the-image

[03:14:02.0000] <jgraham_>
Hixie: For the case where "image would be read out repeatedly, the user could just stop reading the page or engage some sort of skip images mode or something.

[03:14:03.0000] <Philip`>
zcorpan: <a href=...><img src=spacer.gif><img src=spacer.gif><img src=spacer.gif></a>

[03:15:00.0000] <zcorpan>
Philip`: i don't see your point

[03:15:01.0000] <Philip`>
zcorpan: You wouldn't really want to consider all the images to be significant - the only significant thing is that there's a link there

[03:15:02.0000] <Hixie>
jgraham_: at which point you've just assumed the images are decorative

[03:15:03.0000] <zcorpan>
Philip`: are there pages with 3 spacer.gifs in links?

[03:15:04.0000] <Philip`>
zcorpan: I don't know :-p

[03:15:05.0000] <Hixie>
zcorpan: almost certainly

[03:16:00.0000] <jgraham_>
Hixie: Only all remaining images. And it puts the user, rather than the author's editor, in control of their user experience

[03:16:01.0000] <Hixie>
jgraham_: you're assuming a competent user

[03:16:02.0000] <Hixie>
jgraham_: but sure

[03:16:03.0000] <Hixie>
jgraham_: i'm not sure that's the useful default, though

[03:17:00.0000] <Hixie>
jgraham_: depends which happens more often

[03:17:01.0000] <Philip`>
Users have a more direct incentive to act competently than authors do :-)

[03:17:02.0000] <jgraham_>
Basically I think that assuming images for which no information is available are decorative is worse information loss than assuming that they are important and risking the possibility that they are not

[03:17:03.0000] <Hixie>
jgraham_: you could also imagine a UA/AT saying "this page contains 45 unlabeled images." at the start of the read.

[03:17:04.0000] <jgraham_>
Hixie: Yeah

[03:17:05.0000] <Hixie>
jgraham_: (and then not saying where they are)

[03:19:00.0000] <zcorpan>
i think that noalt/importantimage/etc would be used incorrectly so much that the user experience will be the same or better if such attributes are ignored

[03:19:01.0000] <zcorpan>
frankly i think that alt is used incorrectly so much that UAs are better off doing image analysis to decide whether the image should be ignored or not

[03:20:00.0000] <jgraham_>
It would be good if tool authors could be convinced that accessibility is an important problem that needs to be solved by human-led analysis and so encouraged to provide tools for step-through analysis of the document's accessibility

[03:20:01.0000] <Hixie>
zcorpan: sure, sadly we're not there yet

[03:21:00.0000] <jgraham_>
If I were an accessibility expert I would spend my time doing that rather than trying to make one accessibility issue get proxied by a syntax requirement that can be caught by a validity check

[03:22:00.0000] <jgraham_>
(especially given how poor the proxy turns out to be)

[03:22:01.0000] <Hixie>
the accessibility people doing real work aren't visible in the standards process

[03:22:02.0000] <Hixie>
i know a lot of them at google

[03:23:00.0000] <Hixie>
you likely will only rarely hear of them

[03:24:00.0000] <jgraham_>
Hixie: That's a pity. I mean it's good that they're doing real work but bad that that takes up so much time they can't contribute some sanity to standards

[03:25:00.0000] <Hixie>
yeah well

[03:25:01.0000] <Hixie>
the same goes for most engineers, frankly

[03:26:00.0000] <jgraham_>
Yeah that's pretty noticable

[03:26:01.0000] <jgraham_>
/me should go

[03:26:02.0000] <Hixie>
later

[03:33:00.0000] <Philip`>
/me really doesn't like how his laptop's power cable has a loose connection

[03:33:01.0000] <Philip`>
Every time I move it, I worry about whether that's the last time I'll be able to charge my battery

[03:34:00.0000] <Hixie>
my dell had that

[03:34:01.0000] <Philip`>
It'd be nicer if things either worked or didn't work, rather than dangling on the edge

[03:34:02.0000] <Hixie>
and then one day smoke came out of it

[03:34:03.0000] <Philip`>
Oh, lovely

[03:34:04.0000] <Hixie>
they fedexed me a new one overnight

[03:36:00.0000] <Lachy>
Hixie, re your question about ssh and dev.w3.org, didn't you already have access? How have you been uploading the spec without it?

[03:36:01.0000] <Hixie>
not for me

[03:36:02.0000] <Lachy>
oh

[03:36:03.0000] <Hixie>
i got my access years and years ago

[03:37:00.0000] <Hixie>
hehay, i didn't know about rfc 4735

[03:37:01.0000] <Hixie>
example/* and */example media types

[03:43:00.0000] <hsivonen>
I wonder where the 'acrimonious tone' came from...

[03:47:00.0000] <Philip`>
(£30 for a new AC adapter? It's just a dumb lump of plastic and metal :-( )

[03:54:00.0000] <Lachy>
my understanding of important="" was that it would indicate that the image is important content and that the user should attempt to find out more about it through additional heuristics or image analysis, if the alt wasn't sufficient. But then I'm not sure how it's any different from <figure><img src="..."><legend>...</legend></figure>

[04:03:00.0000] <hsivonen>
looks like the meaning of "bolted on" is not agreed upon http://lists.w3.org/Archives/Public/wai-xtech/2008May/0099.html

[04:05:00.0000] <hsivonen>
I expect there's a Latin term for argument by shame

[04:05:01.0000] <Lachy>
I don't understand how it could be interpreted differently from being a feature that is not part of the actual content, but rather included along side it to address some issue

[04:06:00.0000] <hsivonen>
Lachy: if you twist your point of view far enough, an <img> element is just text with an image alternative

[04:06:01.0000] <Hixie>
uh

[04:06:02.0000] <Hixie>
wow

[04:07:00.0000] <Hixie>
that's quite the e-mail

[04:07:01.0000] <hsivonen>
Hixie: no offense to you intended, but what the spec used to say wasn't really how authors use <img>

[04:07:02.0000] <Hixie>
hsivonen: indeed

[04:07:03.0000] <Hixie>
that's why it changed :-)

[04:08:00.0000] <Hixie>
that e-mail reminds me of religous zealot e-mail

[04:08:01.0000] <Philip`>
hsivonen: What the spec says about almost all of HTML isn't how authors really use it

[04:08:02.0000] <Hixie>
"'god does not exist.' you are going to hell! the world is not perfect, but it definitely has a god!"

[04:09:00.0000] <Philip`>
/me sees that HTML 1.0 allowed but did not require ALT

[04:09:01.0000] <hsivonen>
Philip`: I think the syntax definition needs some more authoring practice alignment with border=0, etc.

[04:09:02.0000] <Philip`>
(at least judging by its '< A HREF="Go">< IMG SRC ="Button"> Press to start</A>' example)

[04:10:00.0000] <Philip`>
...although actually I'm just looking at an old draft

[04:10:01.0000] <Hixie>
Philip`: html1?

[04:10:02.0000] <Hixie>
Philip`: i thought "proposal" (the first version of html) didn't have <img> at all

[04:10:03.0000] <Hixie>
Philip`: and that mark anderseen invented it for mosaic

[04:10:04.0000] <Hixie>
is that wrong?

[04:10:05.0000] <Hixie>
or do you mean html+

[04:10:06.0000] <hsivonen>
/me defers replying, tries to get out of the alt tarpit by working on MathML parsing

[04:12:00.0000] <Philip`>
http://www.w3.org/MarkUp/draft-ietf-iiir-html-01.txt

[04:12:01.0000] <Philip`>
Google says that's "html 1.0", so who am I to disagree?

[04:12:02.0000] <hsivonen>
aside: I find that most of my work-related unhappiness comes from dealing with accessibility-related issues. And I think accessibility is a good thing.

[04:13:00.0000] <hsivonen>
I mean the alt thread and the aria-foo vs. aria:foo thing

[04:13:01.0000] <Philip`>
(http://www.ietf.org/rfc/rfc1866.txt seems to have optional ALT too)

[04:14:00.0000] <Lachy>
alt wasn't required until HTML 3.2, I think

[04:14:01.0000] <Lachy>
or maybe it was 4.0

[04:14:02.0000] <hsivonen>
I'd *so* walk away from the alt stuff if it wasn't about what people want my product to do

[04:14:03.0000] <Philip`>
/me is quite happy working on canvas stuff since it's so clearly inaccessible that nobody even cares any more

[04:16:00.0000] <Hixie>
wow, i've never seen this draft

[04:16:01.0000] <Hixie>
this adds a whole version of html to my previously assumed versions

[04:18:00.0000] <hsivonen>
Hixie: do you mean the IIIR draft? it has been cited on the WHATWG list.

[04:19:00.0000] <Hixie>
interesting

[04:19:01.0000] <Hixie>
must've not yet gotten to that e-mail, or assumed that e-mail to be referring to something else

[04:19:02.0000] <Hixie>
rel=interested is a hilarious (unscalable) idea

[04:20:00.0000] <hsivonen>
Hixie: I think I cited it in an <i> vs. <em> email

[04:20:01.0000] <Lachy>
what is rel=interested supposed to mean?

[04:20:02.0000] <Hixie>
oh i probably assumed it was just a copy of proposal, then

[04:20:03.0000] <Hixie>
Lachy: see near the end of that draft

[04:22:00.0000] <Lachy>
I don't see how that's not scalable. How hard can it be to notify any website that links to mine with rel=interested, when my page is updated? ;-)

[04:22:01.0000] <Lachy>
I suppose it's a like a poorly designed version of pingback

[04:23:00.0000] <Hixie>
wow, alt=" really is optional in that draft

[04:23:01.0000] <Hixie>
so optional that the draft doesn't even count it!

[04:23:02.0000] <Hixie>
"[IMG] has two attributes: SRC [...] ALIGN [...] ALT [...]"

[04:24:00.0000] <zcorpan>
obviously image alignment is more important than text alternative

[04:25:00.0000] <Lachy>
yeah, alt wasn't added later in RFC 1866, but still optinal until HTML 4.0. I guess that proves alt really is bolted on.

[04:25:01.0000] <Hixie>
man this spec is so vague it's ridiculous

[04:25:02.0000] <Hixie>
anyway

[04:25:03.0000] <Hixie>
back to making hml5

[04:25:04.0000] <Hixie>
html5

[04:27:00.0000] <zcorpan>
/me looks up expansions for hml

[04:27:01.0000] <zcorpan>
Human-Machine Language 5?

[04:27:02.0000] <zcorpan>
Hellfire Missile Launcher?

[04:28:00.0000] <zcorpan>
or Help Me Lord

[04:36:00.0000] <virtuelv>
Hixie: you were looking for a security model for File I/O?

[04:37:00.0000] <Philip`>
/me wonders why it's bad to use <progress> for a static gauge

[04:39:00.0000] <Lachy>
Philip`, because using progress sugests that the value will change in the future, and because common progress widgets UI looks different from a gauge

[04:40:00.0000] <Hixie>
virtuelv: ?

[04:40:01.0000] <virtuelv>
Lachy: yes, but if I say have a project tracker, and it says that some task is 80% complete

[04:40:02.0000] <Hixie>
virtuelv: in what context?

[04:40:03.0000] <virtuelv>
Hixie: the proposal sent to the web api mailing list?

[04:40:04.0000] <Hixie>
virtuelv: ah, yes, what about it?

[04:41:00.0000] <virtuelv>
I have a workable idea for it, I think

[04:41:01.0000] <Lachy>
virtuelv, then it's not really static. It may just take a while before it progresses beyond 80%

[04:41:02.0000] <Hixie>
let the wg know :-)

[04:41:03.0000] <virtuelv>
Lachy: the point is that it's not going to progress during the lifetime of the page

[04:42:00.0000] <Hixie>
Lachy: a progress bar might make sense for something that is a task competion amount (maybe.) but it's not appropriate for a disk space utilisation meter, e.g.

[04:42:01.0000] <Lachy>
Hixie, yeah, I know

[04:42:02.0000] <Hixie>
er, that was meant for Philip` 

[04:43:00.0000] <Lachy>
virtuelv, that doesn't really matter. It's still a progress indicator

[04:43:01.0000] <hsivonen>
as a practical matter, it matters that progress bar is animated to look busy on OS X and Ubuntu

[04:43:02.0000] <virtuelv>
Lachy: agreed, but it was somewhat unclear from Philip`s question

[04:44:00.0000] <hsivonen>
so if the project isn't completing while the users is waiting, you may want a gauge anyway

[04:46:00.0000] <Lachy>
are there non-animated progress indicators available in OSX or Ubuntu?

[04:46:01.0000] <hsivonen>
I don't know

[04:50:00.0000] <Lachy>
maybe we should have some way to indicate whether the progress bar is relatively static or dynamic, where a dynamic one is expected to update relatively quickly, and a static may stay the same for a while.

[04:51:00.0000] <Lachy>
and if there are non-animated progress bars available, they would be used for the static progress bars.

[04:53:00.0000] <hsivonen>
isn't a non-dynamic progress bar close enough to gauge for all practical purposes?

[04:53:01.0000] <Lachy>
I don't know. Maybe.

[04:56:00.0000] <Lachy>
Apple's HIG shows both active and inactive fill progress bars http://developer.apple.com/documentation/UserExperience/Conceptual/OSXHIGuidelines/XHIGControls/chapter_19_section_5.html#//apple_ref/doc/uid/TP30000359-TPXREF208

[04:57:00.0000] <Lachy>
but it doesn't explain when each one should be used.

[04:58:00.0000] <Hixie>
inactive = the window isn't focused

[04:58:01.0000] <Lachy>
oh, ok

[04:59:00.0000] <Hixie>
<meter> = capacity indicator on that page

[04:59:01.0000] <Philip`>
/me doesn't like forms that try to automatically capitalise addresses and result in "King'S College"

[05:00:00.0000] <Lachy>
so I guess my static/dynamic idea is no good. Ignore it.

[05:00:01.0000] <Hixie>
(we don't have relevance indicators or rating indicators or discrete capacity indicators or asynchronous progress indicators, as defined by that page)

[05:00:02.0000] <Hixie>
(though maybe we should)

[05:01:00.0000] <hsivonen>
I thought <meter> was meant to be a relevance indicator, too

[05:04:00.0000] <Lachy>
for the rating indicator in that HIG page, would that be best done <input type=range>, and somehow styled with XBL?

[05:04:01.0000] <Hixie>
hsivonen: yeah, in html5 <meter> acts as both

[05:04:02.0000] <Hixie>
Lachy: no idea

[05:04:03.0000] <Hixie>
4am. bed time.

[05:11:00.0000] <Hixie>
before i go to bed

[05:11:01.0000] <Hixie>
does anyone want to give a talk on html5 at @media 2008 in london at the end of may?

[05:11:02.0000] <Hixie>
Philip`? jgraham_?

[05:15:00.0000] <Hixie>
don't all reply at once :-)

[05:16:00.0000] <Hixie>
/me goes to bed

[05:26:00.0000] <BenMillard>
if JGraham goes and my travel is paid, we could show how HTML5 table accessibility has improved through research, prototypes and analysis

[05:26:01.0000] <Lachy>
hmm. Only 16 days to go till @media. I don't think I could put together a presentation and get approval and funding to go in that short time.

[06:33:00.0000] <hsivonen>
any suggestions on how to best handle *not* acknowledging the self-closing flag?

[06:34:00.0000] <hsivonen>
hmm. perhaps I should use break instead of early return and fall to a late return that inspects whether the flag has been cleared

[06:35:00.0000] <hsivonen>
yay for goto programming

[06:39:00.0000] <zcorpan>
goto is the shit. opera supports it in javascript too

[06:46:00.0000] <hsivonen>
if I reset the insertion mode in MathML, what do I set the secondary mode to?

[06:48:00.0000] <takkaria>
/me adds to the spam of public-html

[06:50:00.0000] <hsivonen>
takkaria: http://lists.w3.org/Archives/Public/wai-xtech/2008May/0098.html

[06:52:00.0000] <Lachy>
takkaria, this is what at least one blind photographer does http://my.opera.com/oedipus/albums/

[06:53:00.0000] <Lachy>
although, the alt attributes used repeat the tags, that's a problem with my.opera, which is out of his control

[07:05:00.0000] <takkaria>
hsivonen: that goes part of the way but doesn't quite address it

[07:08:00.0000] <hsivonen>
takkaria: it seems, though, that the WCAG 2.0 framers have come to the conclusion that there are use cases that should be allowed to exist even if they aren't fully accessible

[07:12:00.0000] <takkaria>
mm

[07:17:00.0000] <Philip`>
Hixie: @media seems too trendy for me :-p

[07:17:01.0000] <Philip`>
and I don't think I'd be any good at that kind of thing

[07:17:02.0000] <hsivonen>
Lachy: switching services is not out of the users control and advocates seem to be suggesting getting rid of bad tools, i.e. switching tools

[07:18:00.0000] <takkaria>
good lord, misunderstandings are easy to come by these days

[07:18:01.0000] <Philip`>
You could ask the toolmaker to improve the tools - switching away is not the only possible form of feedback

[07:19:00.0000] <hsivonen>
Philip`: sure. my point is that it comes down to choosing what you value more if the tool maker doesn't fix

[07:59:00.0000] <Lachy>
wow, Rob seemed to completely sidestep the issue instead of answering the question http://lists.w3.org/Archives/Public/public-html/2008May/0263.html

[08:17:00.0000] <takkaria>
there has to be some middle ground here somewhere such that the a11y people and the critical-content-altless people can stop talking past each other

[08:19:00.0000] <Philip`>
Perhaps all the middle ground has been turned into no man's land

[08:24:00.0000] <hsivonen>
the middle ground seems to be giving up on the distinction between alt not known and purely decorative and using alt='' for both cases

[08:26:00.0000] <zcorpan>
hsivonen: i think that could be reasonable, given that there is a non-trivial amount of content images on the web that have alt='', and so users would miss out if UA ignored them

[08:27:00.0000] <zcorpan>
(i don't have data but i've seen content images with alt='' more than content images with good alt, i think)

[08:27:01.0000] <hsivonen>
zcorpan: of course, the next question is: if no alt and alt='' mean the same thing, why bother with alt='' except for backwards compat and in order to have more error cases

[08:28:00.0000] <zcorpan>
hsivonen: indeed, i'd be fine with making no alt and alt='' equivalent and have alt optional :)

[08:30:00.0000] <Philip`>
/me thinks data would help, to determine roughly what "non-trivial amount" and "more" mean

[08:30:01.0000] <zcorpan>
Philip`: indeed

[08:31:00.0000] <zcorpan>
i might be completely off and hence my conclusions wrong

[08:38:00.0000] <takkaria>
hmm, webcams that take a snapshot every x minutes and upload it to the web are like blind photographers but remove humans from the equation entirely

[08:38:01.0000] <takkaria>
I think that might turn into my new example of choice

[08:39:00.0000] <hsivonen>
takkaria: whoever sets up the Webcam knows what general direction it is pointed at

[08:39:01.0000] <Philip`>
What if the webcam is mounted on a robot that is controlled remotely by visitors to the web site?

[08:40:00.0000] <takkaria>
hsivonen: not on a laptop

[08:41:00.0000] <takkaria>
I should start collecting these cases somewhere

[08:42:00.0000] <hsivonen>
the MathML stuff has an admirably small impact on tree construction code

[08:42:01.0000] <hsivonen>
most of the changes were related to adding a namespace parameter all over

[08:42:02.0000] <hsivonen>
which thankfully is the kind of stuff Eclipse is really good at

[08:45:00.0000] <mpt>
takkaria, http://www.geonet.org.nz/volcano/volcams/whiteisland/whiteisland-recent.html

[08:45:01.0000] <Lachy>
what we really need for the alt discussion to get somewhere is a study that looks at the reasons why people use alt="" or alt="bogus, validator friendly content" or whatever. The difficulty is in figuring out how to perform such a study

[08:46:00.0000] <Lachy>
maybe a carefully crafted survey, asking web developers a bunch of questions about the issue would work

[08:47:00.0000] <hsivonen>
Lachy: no, this is the kind of stuff where you have go look for revealed preferences from actual actions, because if you ask, people lie

[08:47:01.0000] <takkaria>
mpt: thanks

[08:47:02.0000] <Lachy>
yeah, but the problem is, how do you unambiguosly determine someone's reasons from simply looking at their output?

[08:48:00.0000] <Philip`>
Lachy: You can't, but you can't determine that by any other mechanism either

[08:48:01.0000] <mpt>
interview them afterwards

[08:48:02.0000] <mpt>
Welcome to sociology!

[08:49:00.0000] <takkaria>
/me attempts to pin down some concrete proposals rather than wordy long-paragaph replies

[08:49:01.0000] <Philip`>
Also, your assumption that people have reasons is not clearly valid

[08:52:00.0000] <Lachy>
Philip`, people always have reasons for doing something, even if that reason is that they didn't care and just picked one alternative randomly

[08:58:00.0000] <Lachy>
if I were to do a presentation on HTML5 at @media, which topics do you think people would like to hear about?

[08:58:01.0000] <Lachy>
it looks like the presentations are an hour long, so I could fit a lot in

[09:02:00.0000] <takkaria>
heh, moving the goalposts is popular sport on public-html

[09:03:00.0000] <hsivonen>
/me expects the famous 'axiomatic proof' to show up

[09:06:00.0000] <takkaria>
I wish I had the time to write a proper reply to Rob today, but it'l have to wait til tomorrow

[09:06:01.0000] <Lachy>
hsivonen, http://ln.hixie.ch/?start=1041735552&count=1 ?

[09:07:00.0000] <hsivonen>
Lachy: yeah

[09:08:00.0000] <Lachy>
fortunately, Shelbey hasn't joined the group yet. But who knows, she might.

[09:09:00.0000] <virtuelv>
hm, interesting image chosen for Hixie's last log entry

[09:11:00.0000] <Lachy>
virtuelv, the image is somewhat random (though it doesn't update too frequently).

[09:15:00.0000] <hsivonen>
can foreign elements ever participate in foster-parenting?

[09:23:00.0000] <zcorpan>
hsivonen: you mean in e.g. <table><math> ?

[09:30:00.0000] <zcorpan>
"you're not reading my questions." "you're not reading my answers."

[09:30:01.0000] <zcorpan>
/me wonders why he's still reading

[09:33:00.0000] <hsivonen>
zcorpan: yes

[09:34:00.0000] <hsivonen>
zcorpan: umm no

[09:34:01.0000] <hsivonen>
I mean stuff that occurs 'in foreign content'

[09:34:02.0000] <zcorpan>
ah

[09:40:00.0000] <virtuelv>
hsivonen: are your XTech slides anywhere online?

[09:42:00.0000] <hsivonen>
virtuelv: http://hsivonen.iki.fi/xtech2008/ but the server is not responding right now

[09:42:01.0000] <hsivonen>
actually it is responding but very slowly

[09:43:00.0000] <virtuelv>
hsivonen: that's ok, I just need something to paste into my trip report

[09:49:00.0000] <aroben>
/me wonders why it's <dialog> instead of <dialogue>

[09:52:00.0000] <hsivonen>
yay. pointing out a use case is undermining the work of the WG

[09:53:00.0000] <gavin>
wow

[09:54:00.0000] <hsivonen>
Hixie: I'm now even more convinced that 'in foreign content' should be a flag and not an item in the mode enumeration

[09:54:01.0000] <hsivonen>
for efficient impl

[09:55:00.0000] <hsivonen>
do we want to have the self-closing flag on end tags for the purpose of tokenizer tests?

[10:23:00.0000] <gsnedders>
Hixie: it isn't just atom:content that allows it. It's that, and Atom XHTML Text Constructs.

[10:24:00.0000] <hober>
i.e. atom:title, atom:summary, atom:content, atom;rights

[10:24:01.0000] <gsnedders>
/me was being lazy and not looking

[10:24:02.0000] <gsnedders>
But it's probably better to generically define it

[10:25:00.0000] <gsnedders>
atom:subtitle too

[10:25:01.0000] <hober>
right, by referencing the text construct dfn

[10:25:02.0000] <virtuelv>
hsivonen: you did a lightning talk too?

[10:26:00.0000] <hsivonen>
virtuelv: yeah

[10:27:00.0000] <virtuelv>
url?

[10:27:01.0000] <hsivonen>
virtuelv: the same directory

[10:28:00.0000] <hsivonen>
http://hsivonen.iki.fi/xtech2008/html5-parsing-lightning.pdf

[10:28:01.0000] <virtuelv>
ty

[10:28:02.0000] <virtuelv>
I never looked at the directory, since I assumed it was a direct link

[12:18:00.0000] <othermaciej>
ok the accessibility crowd needs to get their story straight

[12:19:00.0000] <othermaciej>
how can they cite each other's emails as supporting evidence when they are saying directly contradictory things?

[12:19:01.0000] <othermaciej>
am I losing my mind?

[12:24:00.0000] <Lachy>
othermaciej, which mails are you referring to?

[12:25:00.0000] <othermaciej>
Lachy: Robert Burns saying that alt is not supposed to be used for photos, and others citing his arguments

[12:25:01.0000] <othermaciej>
am I totally lacking in reading comprehension skills because that seems like the opposite of what AI54 says

[12:26:00.0000] <rammic>
Can anyone answer a stupid spec question for me? (4.7.5.1. Changes to the networking model) would seem to imply that if a cached application cannot be reached due to network error (downtime?), then the browser should load the cache version- right?

[12:28:00.0000] <Lachy>
yeah, I thought Rob's argument was a bit strange

[12:29:00.0000] <Lachy>
rammic, I'll check the spec and see how I interpret it.

[12:29:01.0000] <rammic>
Thanks. 

[12:31:00.0000] <andersca>
Hixie?

[12:32:00.0000] <Lachy>
rammic, I believe you're right. Though, I could be wrong, since I'm not very familar with that section of the spec.

[12:34:00.0000] <rammic>
Okay. I'm playing around with the Firefox implementation and found it wasn't working as expected... I hope failback is intended (otherwise kills several use cases).

[12:37:00.0000] <Philip`>
/me loves C, particularly when his program crashes with a non-existent backtrace because a function is somehow making %esp offset by 4 when it's called

[12:41:00.0000] <othermaciej>
ok it is starting to bug me a little that Steven Faulkner doesn't seem to even be trying to spell my name correctly

[12:45:00.0000] <Lachy>
najiec, I think it is you who is misspelling your name :-)

[12:59:00.0000] <Philip`>
(Yay for undeclared functions, and C guessing that it probably just takes int arguments and returns an int or something and who cares if that's going to crash your program)

[13:05:00.0000] <othermaciej>
-Werror

[13:05:01.0000] <Philip`>
Legacy codebase with hundreds of existing warnings :-p

[13:07:00.0000] <Philip`>
and I want my fork to differ as little as possible from the original version, to avoid future pain when merging changes, so I don't want to fix all the warnings in my copy

[13:09:00.0000] <gsnedders>
Philip`: I ended up trying that once. I rewrote all but one method in the end. :P

[13:19:00.0000] <annevk>
it's amusing that when html4all people are confused with HTML5, HTML5 is blamed

[13:19:01.0000] <annevk>
but when their propposal is confusing, the reviewer is blamed

[14:17:00.0000] <hsivonen>
argh. now armchair lawyering about copyright metadata is brought in

[14:21:00.0000] <Lachy>
I think alt text used in these Examples that Gez linked to would be better marked up as captions. http://emacspeak.sourceforge.net/raman/

[14:22:00.0000] <Lachy>
although, I guess they are links, so maybe not

[15:12:00.0000] <Hixie>
andersca: here now

[15:12:01.0000] <Hixie>
andersca: note btw there were some changes made yesterday that you will want to see

[15:13:00.0000] <andersca>
Hixie: oh?

[15:13:01.0000] <jgraham_>
othermaciej_: Steve consistently calls me "Graham" for some reason I don't quite understand

[15:14:00.0000] <Hixie>
andersca: yeah, look at the html5 tracker and at the recent checks with the webkit icon

[15:14:01.0000] <andersca>
Hixie: ice!

[15:14:02.0000] <andersca>
nice even

[15:15:00.0000] <Hixie>
andersca: as far as i know that resolves all the outstanding issues with that part of the spec

[15:19:00.0000] <gsnedders>
jgraham_: Would you rather myself and Philip` look at you at once?

[15:20:00.0000] <jgraham_>
gsnedders: That's unnerving. This is just faintly insulting (which is a pity because I think Steve is actually a nice guy)

[15:27:00.0000] <gsnedders>
Lachy: re your tweet: the solution to issue 42. :P (or, seriously, why defining error handling actually helps anyone when all the browsers are close enough anyway)

[15:28:00.0000] <Lachy>
awsome! Someone actually reads my twitter feed :-)

[15:29:00.0000] <gsnedders>
Lachy: Feel honoured. I follow very few :)

[15:36:00.0000] <Philip`>
/me wonders how much of http://googleblog.blogspot.com/2008/05/moving-to-unicode-51.html is the effect of Google's indexing system becoming less US-centric over time

[15:36:01.0000] <Hixie>
MikeSmith: help me out here. what would i need to do to get http://www.w3.org/2000/11/DOM-Level-2-errata updated?

[15:45:00.0000] <soypunk>
blah, I wish http://twitter.com/replies were visible for other users so that besides seeing my own I could also do http://twitter.com/lachy/replies

[15:45:01.0000] <soypunk>
i guess this is as good as it gets: http://quotably.com/Lachy/statuses/810546483

[15:51:00.0000] <mcarter>
in my pursuit of implementing cross-sub-domain SSE and TCPConnection for IE, I've noticed that sometimes switching document.domain causes some weird illegal access errors, but then calling CollectGarbage(); right after switching the domain sometimes fixes those errors. I don't suppose anyone might possibly know why that is?

[15:52:00.0000] <Hixie>
o_O

[15:54:00.0000] <Lachy>
soypunk, why do you need to be able to see replies to me?

[15:54:01.0000] <soypunk>
well replies to your inquiry...

[15:55:00.0000] <Lachy>
FWIW, you're the only one to respond via twitter so far

[15:55:01.0000] <Hixie>
i think for explaining <meter> better we should have some pictures of gagues

[15:55:02.0000] <Hixie>
gauges

[15:55:03.0000] <Hixie>
anyone wanna take screenshots of gauge controls? :-)

[15:55:04.0000] <soypunk>
Lachy: are you collecting this info for the authoring guide, a follow-up ALA piece, or other?

[15:55:05.0000] <Lachy>
other

[15:56:00.0000] <Lachy>
for 2 things actually

[15:56:01.0000] <smedero>
should really keep my html-wg/whatwg nicks the same.

[15:56:02.0000] <Lachy>
1. An interview I'm doing on thursday evening for a podcast, and a presentation that jgraham and I are doing at @media

[16:19:00.0000] <roc>
I guess I'd better alert dcamp to those offline spec changes

[16:20:00.0000] <Hixie>
ooo, i didn't realise mozilla had an implementation of the offline stuff that was in sync with the spec

[16:20:01.0000] <Hixie>
my bad

[16:20:02.0000] <roc>
you didn't?

[16:20:03.0000] <roc>
Dave rewrote it all to be in line with the spec

[16:20:04.0000] <roc>
well, hopefully

[16:20:05.0000] <Hixie>
i thought the mozilla stuff was an old implementation that was removed for ff3

[16:20:06.0000] <roc>
the old implementation was removed ... and a new one inserted :-)

[16:21:00.0000] <Hixie>
aah :-)

[16:21:01.0000] <Hixie>
/me updates the markers in the spec

[16:22:00.0000] <roc>
http://starkravingfinkle.org/blog/2008/05/firefox-3-offline-app-demo-part-2/

[16:22:01.0000] <roc>
> The biggest changes involve dropping support for our own <link rel="offline"> mechanism and supporting WHATWG manifests and application cache.

[16:22:02.0000] <Hixie>
nice

[16:23:00.0000] <othermaciej_>
is there a web interface that I can use to look at checkins of HTML5?

[16:23:01.0000] <roc>
I think you did know this at one point, since you wrote some tests and they failed on Firefox :-). but I think Dave fixed them

[16:23:02.0000] <Hixie>
othermaciej: yes, see the link at the top of the spec

[16:23:03.0000] <Hixie>
roc: ooooo, that rings a bell, yes

[16:24:00.0000] <Hixie>
man i need to keep better track of who does what

[16:24:01.0000] <othermaciej>
Hixie: thanks!

[16:25:00.0000] <gsnedders>
/me improves Hixie's bell

[16:32:00.0000] <vlad_>
annevk: ping

[16:32:01.0000] <Hixie>
('course it's 00:37 in his time zone, so he might not be around)

[16:33:00.0000] <vlad_>
damn you

[16:40:00.0000] <Philip`>
vlad_: http://html5.googlecode.com/svn/trunk/web-apps-tracker/web-apps-tracker - patches welcome ;-)

[16:45:00.0000] <hendry>
for vim users http://svn.natalian.org/projects/html5/html.vim


2008-05-14
[17:23:00.0000] <Hixie>
<script charset>

[17:23:01.0000] <andersca>
roc: looks like there is a bug in that offline app demo

[17:23:02.0000] <Hixie>
what do we think

[17:23:03.0000] <Hixie>
do we need it?

[17:23:04.0000] <roc>
andersca: yeah?

[17:23:05.0000] <Hixie>
or what

[17:23:06.0000] <andersca>
roc: I think todo-server.php should be added to the online whitelist

[17:24:00.0000] <Hixie>
i guess i should look at who is using it

[17:24:01.0000] <andersca>
roc: otherwise it will fail to load once the page has been cached

[17:24:02.0000] <roc>
andersca: better let Mark know

[17:24:03.0000] <andersca>
roc: will do!

[17:24:04.0000] <roc>
thanks

[17:37:00.0000] <Lachy>
LOL http://twitter.com/webconverger/statuses/810611186

[18:43:00.0000] <andersca>
Hixie?

[18:49:00.0000] <Hixie>
yo

[18:52:00.0000] <andersca>
If there is already an application cache identified by this manifest URI, and the most up to date version of that application cache contains a resource with the URI of the manifest, and that resource is categorised as a manifest, then: store the resource in the matching cache, categorised as an implicit entry, associate the Document with that cache, invoke the application cache update process, and abort these steps.

[18:52:01.0000] <andersca>
Hixie: was it resolved how to handle the load of the implicit resource being cancelled here?

[18:53:00.0000] <Hixie>
yes

[18:53:01.0000] <Hixie>
it's in the application cache update process algorithm

[18:53:02.0000] <Hixie>
step 9

[18:56:00.0000] <andersca>
Hixie: ah

[18:57:00.0000] <andersca>
Hixie: so the end result will be that the document will be associated with the cache

[18:57:01.0000] <andersca>
Hixie: but the implicit resource will not be in the cache

[18:58:00.0000] <Hixie>
yeah

[18:58:01.0000] <andersca>
Hixie: also, what if the load is canceled before the manifest has finished loading

[18:58:02.0000] <andersca>
cancelled

[18:58:03.0000] <Hixie>
step 8

[18:58:04.0000] <Hixie>
oops, there's a cross-reference error in that paragraph

[18:58:05.0000] <Hixie>
let me fix that

[19:00:00.0000] <Hixie>
ok where it says "run just to the caching failure steps" assume it says "run the cache failure steps" and links to the cache failure steps paragraph below

[19:00:01.0000] <Hixie>
i'm in the middle of an edit so i can't make the change right now

[19:03:00.0000] <andersca>
OK!

[19:03:01.0000] <andersca>
Hixie: thanks

[19:03:02.0000] <Hixie>
np

[19:04:00.0000] <Hixie>
search for "canceled" in that section, i added it explicitly in a few places (and in a few places, added it explicitly as NOT being something that would trigger failure mode)

[19:04:01.0000] <Hixie>
as to the end result if you cancel the document and it being associated with the cache anyway -- that seemed the most consistent thing to do

[19:05:00.0000] <Hixie>
it's what happens if the cache update fails, too

[19:29:00.0000] <andersca>
Hixie: yeah

[19:37:00.0000] <jwalden>
mcarter: SSE?

[19:40:00.0000] <Hixie>
<event-source>

[20:11:00.0000] <jwalden>
ah

[20:11:01.0000] <jwalden>
oh, server-sent events

[20:12:00.0000] <jwalden>
/me is dumb

[20:12:01.0000] <jwalden>
/me proposes <interlocution/>

[21:01:00.0000] <inimino>
foreach ($_FILES["pictures"]["error"] as $key => $error) {

[21:01:01.0000] <inimino>
    if ($error == UPLOAD_ERR_OK) {

[21:01:02.0000] <inimino>
        $tmp_name = $_FILES["pictures"]["tmp_name"][$key];

[21:01:03.0000] <inimino>
        $name = $_FILES["pictures"]["name"][$key];

[21:01:04.0000] <inimino>
        move_uploaded_file($tmp_name, "data/$name");

[21:01:05.0000] <inimino>
    }

[21:01:06.0000] <inimino>
erm, sorry

[21:01:07.0000] <inimino>
disregard that...

[21:33:00.0000] <MikeSmith>
Hixie: just now read back to your ping about the http://www.w3.org/2000/11/DOM-Level-2-errata.html doc

[21:34:00.0000] <MikeSmith>
cvs log shows plehegar was the one to make the latest edit to it

[21:35:00.0000] <othermaciej>
oh hmm I did not realize there were errata up

[21:35:01.0000] <MikeSmith>
I wonder whether maybe ownership of that doc should move the Web API WG

[21:36:00.0000] <othermaciej>
it should, if Web API WG doesn't own it already

[21:37:00.0000] <othermaciej>
though really I'd rather see a redone DOM Core 3 w/ conformance requirements stated properly and bogus non-browser-relevant stuff removed or moved to an appendix

[21:38:00.0000] <MikeSmith>
othermaciej: yeah, but short of that I reckon that we ought to keep that errata up to date

[21:39:00.0000] <MikeSmith>
unless/until we get something done that supersedes it

[21:39:01.0000] <othermaciej>
I guess DOM levels don't supersede each other

[21:39:02.0000] <othermaciej>
though when HTML5 comes out it should obsolete the older HTML DOM specs

[21:40:00.0000] <MikeSmith>
yeah

[21:44:00.0000] <MikeSmith>
For now, I'm assuming Hixie has a specific change that needs to be made. For record-keeping purposes, perhaps the process we should follow is to send a heads-up about it to the Web API mailing list, decide what the wording should be, record some resolution that we agreed it needed to be changed, then Doug or I can add it to that page.

[21:52:00.0000] <othermaciej>
MikeSmith: I vaguely recall that there was some minor inconsistency between DOM2 Core and DOM3 Core about exceptions

[21:52:01.0000] <othermaciej>
which should probably be errata'd

[21:52:02.0000] <othermaciej>
one of the dom 2 core test suite tests checks for it (though weirdly checks for the dom3 behavior)

[21:53:00.0000] <MikeSmith>
othermaciej: please write something up about it if/when you have time, send to public-webapi⊙wo

[21:53:01.0000] <othermaciej>
yeah, I'll have to find what the issue is

[21:56:00.0000] <Hixie>
MikeSmith: so you recommend mailing public-webapi⊙wo and what, getting wg consensus somewhow?

[21:57:00.0000] <Hixie>
i guess that works

[22:00:00.0000] <MikeSmith>
Hixie: yeah, if it's an obvious erratum, I think it's not so much a matter of getting consensus, it's more just having a record in the list archives that a new change was made and when

[22:01:00.0000] <Hixie>
k

[22:10:00.0000] <MikeSmith>
Hixie: btw, do you have a rough idea of when you might add section, header, footer, etc. to the parsing algorithm?

[22:11:00.0000] <Hixie>
probably the next time i work on the parser

[22:11:01.0000] <Hixie>
since i've added mathml now there's not much point delaying further

[22:11:02.0000] <MikeSmith>
OK, I ask just because of the issue of html5lib reporting "Undefined behaviour for end tag" for those

[22:12:00.0000] <Hixie>
yeah

[01:10:00.0000] <annevk>
vlad_, pong...

[01:10:01.0000] <annevk>
/me wonders if vlad_ is still awake, now it's 0:00 over there :D

[01:30:00.0000] <jwalden>
annevk: Opera needs to implement Date.now(), defined as function() { return new Date().getTime(); } -- I think you'd be the remaining non-IE browser without it given WebKit added support earlier today :-)

[01:31:00.0000] <weinig>
annevk: you haven't by any chance written any tests for the Access Control spec have you?

[01:32:00.0000] <Hixie>
so people want <script type="some proprietary stuff">...</script> to be a conforming way of smuggling proprietary data into a web app

[01:33:00.0000] <virtuelv>
jwalden: mind filing a bug on that, and give me the bug number?

[01:33:01.0000] <jwalden>
virtuelv: point me to a location and I will

[01:33:02.0000] <Hixie>
i wonder what to do abotu that

[01:34:00.0000] <virtuelv>
jwalden: https://bugs.opera.com/wizard

[01:35:00.0000] <annevk>
weinig, not yet, sorry

[01:35:01.0000] <weinig>
annevk: ok, no worries

[01:35:02.0000] <annevk>
ah, vlad requests that web-apps-tracker has an option to "strip" HTML from the diff

[01:36:00.0000] <annevk>
if anyone has time to implement that in a way that does not expose us to script injection attacks from Hixie, feel free :)

[01:36:01.0000] <Hixie>
btw anne

[01:37:00.0000] <Hixie>
i noticed the gears icons aren't showing properly

[01:39:00.0000] <annevk>
hmm, all those change requests to access control

[01:40:00.0000] <annevk>
if it doesn't stop i might argue for some more simplification of the server side syntax

[01:45:00.0000] <jwalden>
virtuelv: filed as bug 329986

[01:45:01.0000] <annevk>
Hixie, browsers support type=text/javascript1.3 too

[01:46:00.0000] <annevk>
Hixie, some do anyway

[01:46:01.0000] <Hixie>
really?

[01:46:02.0000] <Hixie>
oh

[01:46:03.0000] <Hixie>
well then

[01:46:04.0000] <annevk>
i think that makes more sense

[01:46:05.0000] <Hixie>
i can just strip all that crap then

[01:48:00.0000] <virtuelv>
jwalden: thanks

[01:48:01.0000] <jwalden>
virtuelv: I think that goes both ways -- speed and arithmetic usage improvements will be welcome :-)

[01:53:00.0000] <virtuelv>
jwalden: I presume (but can't actually promise, since it's not my call) that this will be trivial to add

[01:54:00.0000] <jwalden>
implementation-wise I have very little doubt it is, the question's resource allocation and prioritization

[01:55:00.0000] <annevk>
weinig, Hixie, since we have Access-Control-Origin, maybe having a space separated list of URIs from which requests are allowed (only checking the origin bits, similar to postMessage) is enough for the Access-Control / <?access-control?> case

[01:55:01.0000] <Hixie>
hm?

[01:55:02.0000] <Hixie>
what's the problem?

[01:56:00.0000] <annevk>
Access-Control: <http://opera.com:800> <https://w3.org/ignored>

[01:56:01.0000] <annevk>
it being quite complicated currently

[01:56:02.0000] <Hixie>
what's wrong with what we have now

[01:56:03.0000] <Hixie>
i thought we needed deny rules

[01:56:04.0000] <Hixie>
did we decide to get rid of the use cases for that?

[01:57:00.0000] <annevk>
deny rules are gone anyhow

[01:57:01.0000] <Hixie>
(and ignoring stuff in something security related is dangerous -- fail, don't ignore)

[01:57:02.0000] <Hixie>
they are?

[01:57:03.0000] <weinig>
annevk: exclude is the new deny

[01:57:04.0000] <Hixie>
so what does it look like now?

[01:57:05.0000] <Hixie>
that's what i meant by deny

[01:57:06.0000] <annevk>
oh ok

[01:57:07.0000] <weinig>
annevk: what about the wildcarding?

[01:58:00.0000] <annevk>
now we have Access-Control: allow <*.opera.com:*> <*.w3.org> exclude <evil.opera.com>

[01:58:01.0000] <annevk>
i guess it isn't too bad either

[01:59:00.0000] <Hixie>
i don't understand why we suddenly want to get rid of exclude rules

[01:59:01.0000] <annevk>
the main thing is that the majority case is probably whitelisting a few domains or using * and not doing anything this complex

[01:59:02.0000] <annevk>
and the real complex cases can be done using Access-Control-Origin

[02:00:00.0000] <Hixie>
well i'm all for making things simpler, but then i have to wonder why we didn't do them simple in the first place

[02:00:01.0000] <Hixie>
getting rid of "exclude" (and removing the "allow" keyword) is fine by me

[02:01:00.0000] <Hixie>
we still need the wildcards though

[02:01:01.0000] <annevk>
Hixie, wildcards for domains?

[02:01:02.0000] <Hixie>
protocols, domains, and ports

[02:01:03.0000] <Hixie>
same as now

[02:02:00.0000] <Hixie>
i'm just talking about just dropping the "exclude" part and the "allow" keyword

[02:03:00.0000] <annevk>
ok

[02:04:00.0000] <annevk>
i guess still allowing 'example.org' isn't too much trouble then either

[02:04:01.0000] <Hixie>
it's not trouble at all, you've already specified it

[02:04:02.0000] <annevk>
:)

[02:05:00.0000] <Hixie>
every time you make a change to a specification that isn't a change that fixes an actual error or that adds a new important feature, you lose a bit of credibility as a spec editor

[02:05:01.0000] <Hixie>
churn is bad

[02:05:02.0000] <Hixie>
people get frustrated if the spec keeps changing in non-obvious ways for apparently no reason

[02:07:00.0000] <Hixie>
especially when they have already implemented it

[02:07:01.0000] <annevk>
true, i'm just trying to figure out if we can reduce complexity because people have been complaining about that and i sort of agree it's quite complex

[02:07:02.0000] <Hixie>
i'd concentrate on getting the header story straightened out before worrying about that :-)

[02:08:00.0000] <annevk>
and given that the implementors (apart from WebKit so far) have been proposing changes as well during implementation...

[02:08:01.0000] <weinig>
I am not sure I see the benefit to users of removing exclude, as it is already optional

[02:08:02.0000] <annevk>
for instance, now Mozilla wants header/method opt-in

[02:08:03.0000] <annevk>
weinig, yeah, I guess i'll just keep all that in

[02:09:00.0000] <weinig>
ok

[02:09:01.0000] <Hixie>
did sicking ever send the second e-mail he said he would send?

[02:10:00.0000] <weinig>
annevk: what do you mean by header/method opt-in?

[02:10:01.0000] <annevk>
Hixie, no, he hasn't

[02:10:02.0000] <Hixie>
odd

[02:11:00.0000] <annevk>
weinig, http://lists.w3.org/Archives/Public/public-appformats/2008May/0034.html

[02:11:01.0000] <weinig>
ah, I list I am not on :)

[02:12:00.0000] <annevk>
access-control / xhr2 is currently developed on two separate lists unfortunately :(

[02:12:01.0000] <annevk>
once people get their act together it will be on one, but i'm not sure how long that'll take

[02:15:00.0000] <weinig>
will read the relevant back log of that list tomorrow 

[02:17:00.0000] <Philip`>
/me forgets why <dialog> can't just be spelt as <dl>

[02:19:00.0000] <takkaria>
/me boggles what happened to the alt thread overnight

[02:19:01.0000] <othermaciej>
annevk: hmm, I don't get how "Access-Control-Extra-Headers" addresses the stated problem

[02:20:00.0000] <othermaciej>
annevk: doesn't XHR2+AC already whitelist to a short list of headers?

[02:20:01.0000] <othermaciej>
or am I confused?

[02:20:02.0000] <othermaciej>
I guess I should get on public-appformats

[02:22:00.0000] <Hixie>
Philip`: different content models, different basic idea

[02:22:01.0000] <Hixie>
Philip`: a dialog is not a set of one-or-more-names/one-or-more-values tuples

[02:23:00.0000] <Hixie>
though maybe it should allow multiple names per "quote"

[02:25:00.0000] <annevk>
othermaciej, it whitelists a short list of headers for the request, but if you go outside that list and the server responds positively to a preflight request the headers will get through

[02:25:01.0000] <annevk>
(apart from the blacklisted headers of course)

[02:26:00.0000] <othermaciej>
annevk: ok I see

[02:26:01.0000] <Philip`>
Hixie: Since you removed the new bit about language=javascript1.x, where is that handled now?

[02:27:00.0000] <Hixie>
see mail to public-html

[02:27:01.0000] <Hixie>
it's handled by turning it into type=text/language/javascript1.x

[02:27:02.0000] <Hixie>
which browsers support already

[02:27:03.0000] <annevk>
well, some do

[02:27:04.0000] <annevk>
such as WebKit iirc

[02:28:00.0000] <Philip`>
I think my point was that it's not obvious to a UA implementor that they should support those in addition to text/javascript

[02:28:01.0000] <Hixie>
ah well maybe we should update the relevant rfc

[02:28:02.0000] <Hixie>
it needs updating anyway, right now it dumbly obsolete text/javascript

[02:28:03.0000] <Philip`>
The "Scripting languages" section already mentions "text/javascript", so it'd be helpful for implementors if that mentioned the typical aliases too

[02:28:04.0000] <Hixie>
and it doesn't define the e4x parameter, and it looks like the ES4 group is planning on making all kinds of mistakes with a version= parameter, too, which will also need defining

[02:29:00.0000] <Hixie>
yeah, that could work

[02:29:01.0000] <Hixie>
reply to the thread :-)

[02:29:02.0000] <Philip`>
Hixie: By "see mail to public-html", do you mean the mail which says "Done" when actually the final result to the spec was no change? :-)

[02:30:00.0000] <Hixie>
no, the reply to that one

[02:31:00.0000] <Philip`>
I don't see a reply

[02:31:01.0000] <Philip`>
...but that's not your problem

[02:31:02.0000] <Hixie>
http://lists.w3.org/Archives/Public/public-html/2008May/0292.html

[02:31:03.0000] <Philip`>
since it's just the list/Gmail being slow

[02:37:00.0000] <takkaria>
I'd forgotten quite how frustrating being a participant in the alt debate is

[02:37:01.0000] <Hixie>
i'd forgotten why i never reply in real time to threads

[02:37:02.0000] <annevk>
<interplay>

[02:37:03.0000] <Hixie>
i'm now relearning this with the <dialog> issue

[02:37:04.0000] <takkaria>
:)

[02:39:00.0000] <Hixie>
ok

[02:39:01.0000] <Hixie>
you may now vote on the next thing i work on

[02:39:02.0000] <Hixie>
should it be:

[02:39:03.0000] <Hixie>
<iframe> features

[02:39:04.0000] <Hixie>
or

[02:39:05.0000] <Hixie>
<video>

[02:39:06.0000] <Hixie>
you decide!

[02:39:07.0000] <takkaria>
video!

[02:40:00.0000] <takkaria>
I hope you're not expecting rationales

[02:41:00.0000] <annevk>
are we going to add <iframe> features for HTML5?

[02:41:01.0000] <annevk>
/me is silently hoping we push the sandbox issue out of HTML5

[02:41:02.0000] <annevk>
hah, so much for silently

[02:42:00.0000] <Hixie>
i'm expecting us to add iframe sandboxing features

[02:42:01.0000] <othermaciej>
I have heard someone from Microsoft claim that <iframe> is not a strong enough sandbox as-is and therefore iframe+postMessage was not a good enough solution for "mashups"

[02:43:00.0000] <othermaciej>
but he refused to say what the specific issues are

[02:43:01.0000] <othermaciej>
(why does this sound so familiar?)

[02:43:02.0000] <annevk>
parent.location.href = '...' is an issue

[02:43:03.0000] <Hixie>
in particular see the end of http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-May/011198.html for what i expect to look at adding

[02:43:04.0000] <annevk>
(from Douglas Crockford)

[02:44:00.0000] <othermaciej>
Hixie: <script sandbox="">?

[02:44:01.0000] <Hixie>
?

[02:44:02.0000] <othermaciej>
I hope that's not the proposal you expect to look at adding

[02:45:00.0000] <Hixie>
"the end of" was a key part of what i wrote just now :-)

[02:45:01.0000] <othermaciej>
then I have a lot of reading to do to get to the bottom!

[02:46:00.0000] <Hixie>
sorry :-)

[02:46:01.0000] <Hixie>
my bit starts from the line that just says "PROBLEMS"

[02:46:02.0000] <annevk>
you could just press "End"

[02:46:03.0000] <Hixie>
the rest is me mostly telling people their ideas suck

[02:46:04.0000] <Hixie>
(about 70% down)

[02:46:05.0000] <Hixie>
aybe 60%

[02:48:00.0000] <othermaciej>
Hixie: I think you dismissed the md5= possibility too readily - you just keep reading characters until the checksum matches, that works even if the content contains a malicious </sandbox>

[02:48:01.0000] <othermaciej>
faking the hash is a risk

[02:48:02.0000] <othermaciej>
another possibility is to require a psuedo-attribute on the end tag

[02:49:00.0000] <othermaciej>
<sandbox token="13498564123"><!-- that was a random number --> </sandbox token="13498564123">

[02:49:01.0000] <Hixie>
yeah i believe i have that feedback from you in the pile

[02:50:00.0000] <othermaciej>
I don't remember if I said this specific stuff before

[02:50:01.0000] <othermaciej>
though I vaguely remember this being discussed

[02:51:00.0000] <Philip`>
<sandbox13498564123><!-- now you don't even have to change the parser much --></sandbox13498564123>

[02:51:01.0000] <othermaciej>
client-side filtering of script could be done with less chance of time-of-check time-of-use errors, since browsers know exactly when they will run script

[02:51:02.0000] <othermaciej>
Philip`: or use a colon

[02:51:03.0000] <Philip`>
othermaciej: I assume this ought to work in XHTML too

[02:51:04.0000] <othermaciej>
<sandbox:13498564123><!-- now you don't even have to change the parser much --></sandbox:13498564123>

[02:52:00.0000] <othermaciej>
Philip`: I am not sure premature close is the same kind of problem in XHTML since it renders the document not well formed

[02:52:01.0000] <othermaciej>
Philip`: but I guess scripts could have run by the time you find that out

[02:52:02.0000] <annevk>
all this sandboxing seems rather fragile

[02:52:03.0000] <othermaciej>
and I'm sure Hixie won't want to invent a sandbox: namespace

[02:53:00.0000] <Hixie>
the biggest problem with a checksum thing is that if you get it wrong, your document ends up all in the sandbox

[02:53:01.0000] <Hixie>
well maybe no the biggest problem

[02:53:02.0000] <Hixie>
a big problem

[02:53:03.0000] <othermaciej>
annevk: basically the main benefit is that it lets sites be lazy and not do whitelist content filtering; and it avoids the possibility of a whitelist that is too lenient

[02:53:04.0000] <Philip`>
othermaciej: You should be able to take any HTML document and reserialise it as XHTML without any fancy transformations, even if some of the HTML features are redundant in XML, otherwise hsivonen will be unhappy

[02:54:00.0000] <Hixie>
not being able to provide this feature in xhtml is not imho any kind of blocker

[02:54:01.0000] <Hixie>
:-)

[02:54:02.0000] <othermaciej>
Hixie: my idea and Phillip's variant both do not have the possibility of getting a checksum wrong (though it is possible the magic random token will be copied wrong)

[02:54:03.0000] <Hixie>
othermaciej: how so?

[02:55:00.0000] <othermaciej>
there is no checksum

[02:55:01.0000] <othermaciej>
the token is arbitrary - the defense is that hostile content cannot predict it

[02:55:02.0000] <Philip`>
I assume XHTML5 would still want the safe script sandbox behaviour, even though the parser issue is only relevant for HTML

[02:55:03.0000] <Hixie>
othermaciej: oh the random number idea doesn't work at all, you're totally dependent on the site making unpredictable choices of random numbers

[02:56:00.0000] <Hixie>
othermaciej: given that we couldn't even depend on openssl's security experts making unpredictable ssl certs, i don't think we can rely on joe author making unpredictable magic numbers

[02:56:01.0000] <annevk>
and if people are allowed to edit their comments you would have to regenerate the magic number and you might forget to do that, etc.

[02:57:00.0000] <othermaciej>
generating a random number is certainly something you can fail at, but so is whitelist content filtering (the current best practice for injecting unknown content)

[02:57:01.0000] <zcorpan>
i know a name for <dialog>... <dl>

[02:57:02.0000] <zcorpan>
short for DiaLog

[02:57:03.0000] <zcorpan>
(or DiaLogue)

[02:57:04.0000] <othermaciej>
zcorpan: cute

[02:58:00.0000] <Hixie>
othermaciej: whitelist filtering is far easier to understand than random number generation, which is basically a black art

[02:59:00.0000] <Philip`>
Hixie: OpenSSL's security experts did it right, it was just the Debian developers who commented out all the entropy

[03:00:00.0000] <Hixie>
Philip`: for the recent case, yeah, but wasn't there a case long ago that was a real bug?

[03:00:01.0000] <Philip`>
Hixie: Oh, no idea

[03:00:02.0000] <Hixie>
Philip`: or alternatively, consider than win2k, winxp, and vista all had a system RNG that was proved predictable after something like 8 years

[03:00:03.0000] <othermaciej>
Hixie: well on Mac OS X you just call arc4random, but yeah you can screw it up

[03:01:00.0000] <othermaciej>
though a random token, unlike a checksum, would be easily fixable w/ no browser changes needed in case the algorithm is found to be vulnerable

[03:02:00.0000] <othermaciej>
though I guess it could turn comment spam into a CPU hogging attack generating all that entropy

[03:02:01.0000] <othermaciej>
anyway, I am not sure safe markup injection via a client-side feature is feasible, but if it were there would clearly be some advantages

[03:02:02.0000] <Philip`>
You'd probably generate the entropy on (uncached) page views, not on comment posting

[03:03:00.0000] <Philip`>
s/generate/use/

[03:03:01.0000] <othermaciej>
Philip`: I don't think you would want to generate a different random number every time, though maybe every time you generate the page if you cache static copies

[03:03:02.0000] <Philip`>
Fortunately thermodynamics guarantees that entropy won't decrease, so you'll never permanently run out

[03:05:00.0000] <othermaciej>
thank goodness

[03:06:00.0000] <zcorpan>
Hixie: experience with <address> should suggest that the element name matters more than its definition in terms of what authors will use it for

[03:06:01.0000] <othermaciej>
Hixie: styles cascading through a browsing context boundary (wait, do you mean cascading or inheriting or both) sounds like an interesting implementation challenge

[03:07:00.0000] <othermaciej>
Hixie: I do like <iframe noscript> and <iframe restrict> though (well, depending on specific restrictions)

[03:08:00.0000] <othermaciej>
Hixie: though restrict seems better for embedded "gadgets" than for comment fields

[03:08:01.0000] <Hixie>
yeah the stuff at the end is mostly just random ideas

[03:09:00.0000] <Hixie>
i haven't thought through exactly what the needs are

[03:11:00.0000] <othermaciej>
I think the real answer to the JSON case is cross-site XHR plus a native JSON API

[03:12:00.0000] <othermaciej>
trying to run scripts in the same browsing context but with special restrictions sounds like disasterville

[03:12:01.0000] <othermaciej>
I know you agree and all

[03:12:02.0000] <othermaciej>
I just had to say it

[03:15:00.0000] <jgraham>
Hixie: FWIW I think zcorpan is right, and I think enough people have complained about <dialog> to suggest it is a real issue

[03:15:01.0000] <Hixie>
what was the JSON case again?

[03:15:02.0000] <Hixie>
jgraham: the difference is that you could use <address> to include an address and nothing would go wrong. Using <dialog> for a dialog box makes no sense.

[03:16:00.0000] <zcorpan>
Hixie: makes about as much sense to me as using <div class=dialog>

[03:17:00.0000] <zcorpan>
Hixie: consider that dialog boxes would mostly be scripted and so would bypass validation

[03:17:01.0000] <jgraham>
Well nothing will happen but then nothing happens when you use <nav> for navigation either

[03:17:02.0000] <Hixie>
i don't understand what <div class="dialog"> would do either

[03:17:03.0000] <othermaciej>
Hixie: people embed semi-untrusted off-site script to transfer JSON data (the script is expected to just assign to a var)

[03:17:04.0000] <jgraham>
People are used to the idea that elements exist for no function other than "semantics"

[03:17:05.0000] <zcorpan>
Hixie: it wouldn't do anything except provide a hook for scripting and styling for the author to implement the dialog box

[03:21:00.0000] <zcorpan>
Hixie: (i'm just being a bitch here, i actually couldn't care less about the name)

[03:23:00.0000] <Hixie>
othermaciej: xhr2 is the solution to that

[03:24:00.0000] <Hixie>
jgraham: well i'd use a better name, but frankly i haven't seen one that wouldn't have worse problems than <dialog>

[03:24:01.0000] <Hixie>
zcorpan: :-)

[03:24:02.0000] <othermaciej>
Hixie: I mention native JSON API because ideally people would also stop using eval for this purpose (sometimes, awfully enough, eval without the proper preflight regexp)

[03:24:03.0000] <zcorpan>
Hixie: what's the problem with <dl>?

[03:26:00.0000] <Lachy>
Hixie, you could just allow <dl> to be used for dialog as well, and drop <dialog>

[03:26:01.0000] <jgraham>
Hixie: I know all the alternatives suck in their own special ways :(

[03:26:02.0000] <Hixie>
othermaciej: i recommend xml :-)

[03:26:03.0000] <othermaciej>
Hixie: no you don't

[03:26:04.0000] <Hixie>
zcorpan: it's inappropriate? it'd be like using <ol> for a set of paragraphs

[03:26:05.0000] <Hixie>
othermaciej: over json? sure it do

[03:26:06.0000] <Hixie>
i

[03:26:07.0000] <jgraham>
/me dislikes <dl> even more than <dialog>

[03:26:08.0000] <othermaciej>
I meant, generally speaking

[03:27:00.0000] <zcorpan>
Hixie: why is it inappropriate if the spec says it's appropriate?

[03:27:01.0000] <Hixie>
othermaciej: for proprietary sending data across servers, xml is often a fine thing

[03:27:02.0000] <Lachy>
how about <transcript>

[03:27:03.0000] <othermaciej>
I think JSON is nicer than XML for dumb data though to be fair adding a native API misses the point that it was just a built-in language construct

[03:27:04.0000] <jgraham>
<transcript> sounds pretty nice, although it's not strictly right

[03:27:05.0000] <Hixie>
zcorpan: because it would be dumb for hte spec to say it's appropriate. it'd be like using <ol> for lists and paragraphs, or something equally bogus

[03:28:00.0000] <Lachy>
nah, that's work too well

[03:28:01.0000] <Hixie>
othermaciej: well, maybe. not html5's problem, either way.

[03:28:02.0000] <othermaciej>
I'll agree with that

[03:35:00.0000] <Hixie>
i'm more tired than i should be at this time, so i'll just go to bed.

[03:35:01.0000] <Hixie>
nn!

[05:26:00.0000] <zcorpan>
annevk: can html5.org please link to the tracker?

[05:31:00.0000] <zcorpan>
i wonder if the requirement that <script src> elements must be empty is helpful or not

[05:32:00.0000] <zcorpan>
i've seen pages that use comments in <script src>, either <!--pseudo-comments--> or /* js comments */

[05:33:00.0000] <Dashiva>
It keeps the future open for allowing something there?

[05:34:00.0000] <gsnedders>
Dashiva: But that'd need be defined in HTML 2π, so requiring it to be empty in HTML 5 changes nothing

[05:34:01.0000] <zcorpan>
Dashiva: i don't think so -- html4 allows stuff there and people don't care whether the spec allows it or not anyway

[05:34:02.0000] <gsnedders>
/me really hopes the next revision _is_ 2π

[05:35:00.0000] <zcorpan>
Dashiva: what would you allow there in the future anyway?

[05:36:00.0000] <Dashiva>
zcorpan: Nested script elements for fallback

[05:36:01.0000] <zcorpan>
Dashiva: that doesn't work in legacy browsers so not particularly good fallback story

[05:37:00.0000] <zcorpan>
Dashiva: <script type=unknown>..</script> <script> fallback </script> works

[05:37:01.0000] <Dashiva>
zcorpan: Not if you support both scripts

[05:37:02.0000] <zcorpan>
Dashiva: they might be able to communicate with each other

[05:37:03.0000] <zcorpan>
Dashiva: so the new script can set a flag that it has executed

[05:37:04.0000] <Dashiva>
That's a big might

[05:38:00.0000] <zcorpan>
maybe but <script type=unknown> .. <script> fallback </script> </script> is guarenteed to *not* work

[05:39:00.0000] <Dashiva>
You're nesting the wrong way

[05:40:00.0000] <Dashiva>
Old browsers ignore the contents, so they use the outer element's src. New browsers check the contents for another script element and use that if possible

[05:40:01.0000] <zcorpan>
nesting doens't work either way

[05:41:00.0000] <zcorpan>
Dashiva: script is cdata parsing, can't have elements in there at all. changing that would break legacy content

[05:41:01.0000] <Philip`>
Dashiva: That might hurt with <script language=unknown>/* legacy content */ document.write('<script>'); </script>

[05:42:00.0000] <Dashiva>
Philip`: This is only with @src

[05:42:01.0000] <Philip`>
Ah

[05:43:00.0000] <zcorpan>
i'd rather design a new script language for the web in such a way that it's possible to communicate with javascript

[05:44:00.0000] <zcorpan>
(like flash can communicate with javascript, at least i think it can)

[05:44:01.0000] <annevk>
zcorpan, added

[05:44:02.0000] <zcorpan>
annevk: thanks

[05:47:00.0000] <zcorpan>
Dashiva: in practice, authors just put harmless stuff in there and the validator will complain about it, and they have to move the comment out of the script element, which seems pointless

[05:48:00.0000] <zcorpan>
Dashiva: moreover, if the legacy script is external, why not let the new script be external too and have another attribute for the new script?

[05:48:01.0000] <zcorpan>
<script src=fallback newsrc=new>

[05:48:02.0000] <Dashiva>
zcorpan: And duplicate all the other attributes as well?

[05:49:00.0000] <zcorpan>
Dashiva: what other attributes?

[05:50:00.0000] <Dashiva>
You'd want type, to make sure you actually support the new script.

[05:51:00.0000] <zcorpan>
why?

[05:51:01.0000] <zcorpan>
you know you support it since you support the newsrc='' attribute

[05:52:00.0000] <zcorpan>
of course this doesn't scale to many new scripting languages

[05:52:01.0000] <zcorpan>
btw, can vbscript talk to javascript?

[05:52:02.0000] <Dashiva>
Well, yeah, if you make one new src for each language, that'd work, but that's kinda crazy :)

[05:53:00.0000] <zcorpan>
let's solve this problem when a new scripting language actually is introduced :)

[05:54:00.0000] <Dashiva>
But until then, keeping our script elements neat and tidy doesn't hurt :P

[05:55:00.0000] <zcorpan>
Philip`: whenever you're bored enough, could you dig for data about pages that use <script src>..nonwhitespace..</script> ? :)

[06:01:00.0000] <zcorpan>
btw, <script src=my-interpreter id=myscript>my custom syntax script</script> could work

[06:02:00.0000] <zcorpan>
instead of <script type=my/custom>syntax script</script> <script src=my-interpreter></script>

[06:02:01.0000] <zcorpan>
(with id..)

[06:04:00.0000] <zcorpan>
although i guess that's less intiutive and so authors don't do it anyway, so might be pointless to allow it

[06:27:00.0000] <Philip`>
I would guess the most common case of content in srced scripts is when people write <script src="..." />[rest of page content]

[06:29:00.0000] <zcorpan>
Philip`: i guess it would be interesting to know how common that is too :)

[06:30:00.0000] <Philip`>
/me wishes Eclipse didn't crash immediately after loading

[06:56:00.0000] <Philip`>
/me has too many pages to process them all in a reasonable time :-(

[06:58:00.0000] <Philip`>
Ah, only 10424 left

[07:01:00.0000] <Philip`>
zcorpan: http://philip.html5.org/data/nonempty-scripts.html

[07:15:00.0000] <zcorpan>
Philip`: thanks!

[07:16:00.0000] <zcorpan>
ah, ";" is something i've seen as a way to prevent xslt to convert it to <script src='...'/>

[07:20:00.0000] <Philip`>
A couple of princeton.edu pages have a literal nbsp character inside their <script>

[07:20:01.0000] <Philip`>
I don't know if that's for the same kind of reason

[07:23:00.0000] <zcorpan>
i wonder if http://weather.noaa.gov/cgi-bin/iwszone?Sites=:njz024:njz023 put tags in script as a way to comment them out or if it actually was a mistake to put them there

[07:23:01.0000] <zcorpan>
they're duplicated before the script tag

[07:28:00.0000] <zcorpan>
hsivonen: in http://validator.nu/?doc=http%3A%2F%2Fwww.agl1985.org%2F&parser=html5&showsource=yes , group messages and look at "Error: Text not allowed in element script in this context."

[07:29:00.0000] <zcorpan>
hsivonen: notice that it seems to split up the text node and complain several times about the same <script> element containing text

[07:31:00.0000] <zcorpan>
i guess that page would be helped by the requirement since it has <script src><script src></script> where it was probably intended to have 2 script elements

[07:32:00.0000] <zcorpan>
but a conformance checker might be able to detect that anyway

[07:32:01.0000] <zcorpan>
and give a warning

[07:33:00.0000] <zcorpan>
i.e., if a script element contains "<script" that is not in an escaped text span

[07:36:00.0000] <zcorpan>
i guess authors would be stumped when the validator says to (re)move "/* ... This notice MUST stay intact for legal use ... */"

[08:15:00.0000] <hsivonen>
zcorpan: thanks. that's an undesirable but technically understandable feature of Jing

[08:15:01.0000] <hsivonen>
zcorpan: I'll see what I can do about it

[08:46:00.0000] <hsivonen>
<discourse> might actually be better than <dialog>

[08:50:00.0000] <Philip`>
Maybe it'd be better to drop the element entirely, since very few people will ever use it

[08:50:01.0000] <hsivonen>
oh sure

[08:51:00.0000] <Philip`>
When people are arguing over the colour of the bikeshed, the simplest solution is a bulldozer

[09:10:00.0000] <hsivonen>
which reminds me that I want the SVG parsing stuff unbulldozed

[09:10:01.0000] <shepazu>
then help out

[09:11:00.0000] <shepazu>
we aren't bulldozing, we are working on a proposal

[09:11:01.0000] <zcorpan>
/me would like to have a list of requirements and a list of objections to the proposal that was in the spec

[09:13:00.0000] <zcorpan>
or rather i think it would be helpful for the interested parties

[09:14:00.0000] <Philip`>
shepazu: Is that work happening anywhere open/public?

[09:15:00.0000] <shepazu>
http://www.w3.org/Graphics/SVG/WG/wiki/SVG_in_text-html

[09:16:00.0000] <shepazu>
mind, I've been on vacation for the last 2 weeks

[09:17:00.0000] <shepazu>
and I prefer the previous revision

[09:19:00.0000] <Philip`>
"Should be able to take a conforming SVG document and paste its contents into an HTML document and have it be the same DOM" - is that talking about at least one SVG document (which would be useless) or all SVG documents (which would be impossible because of the DTD stuff)?

[09:20:00.0000] <Philip`>
I suppose it should be talking about the subset of SVG documents that can be generated by Inkscape and Illustrator, i.e. unprefixed element names and no fancy DTD stuff except some namespace strings (I hope)

[09:20:01.0000] <Lachy>
Yay! it's about time the W3C started investigating the copyright licence for the authoring guide. :-)

[09:21:00.0000] <Lachy>
http://lists.w3.org/Archives/Public/public-html/2008May/0295.html

[09:22:00.0000] <Lachy>
I don't understand what he means by this point: "that this is a copyright and not a licensing issue"

[09:22:01.0000] <Lachy>
what's the difference?

[09:22:02.0000] <hsivonen>
shepazu: the main SVG WG objection that I am aware of is an objection that I think is misguided and should be fixed by removing the objection

[09:22:03.0000] <shepazu>
hsivonen: and the rest of us differ

[09:23:00.0000] <hsivonen>
shepazu: so how would I help out?

[09:23:01.0000] <shepazu>
in a telcon, brb

[09:50:00.0000] <zcorpan>
why is it a requirement that it produces the same DOM? HTML and XHTML have slightly different DOMs. shouldn't it be a requirement that the image renders the same as it would in XML?

[09:51:00.0000] <zcorpan>
e.g. it would render the same even if RDF doesn't survive

[09:54:00.0000] <zcorpan>
"Should be able to provide some sort of fallback mechanism for the SVG-in-HTML so that UAs that don’t know how to handle these SVG fragments will display the fallback."

[09:54:01.0000] <zcorpan>
i think that's orthogonal to text/html integration since it equally applies to application/xhtml+xml

[09:55:00.0000] <annevk>
can't even require rendering the same way

[09:55:01.0000] <annevk>
quirks mode baby

[09:55:02.0000] <zcorpan>
although text/html parsing can do some tricks to help, e.g. support cdata sections in some places but not others

[09:55:03.0000] <Philip`>
zcorpan: The most relevant UAs that support application/xhtml+xml also support SVG, so that's a mostly theoretical concern, whereas fallback for SVG-in-text/html is a practical concern

[09:55:04.0000] <Lachy>
why does quirks mode have to have any effect upon the rendering of SVG?

[09:56:00.0000] <zcorpan>
Lachy: case sensitivity of class

[09:57:00.0000] <Lachy>
but does the case insenstivity need to apply to SVG elements?  Why can't it just apply to HTML elements?

[09:58:00.0000] <zcorpan>
Lachy: in firefox, opera and safari, it applies to svg as well

[09:59:00.0000] <zcorpan>
Lachy: it doesn't need to

[09:59:01.0000] <Lachy>
ah, crap :-(

[09:59:02.0000] <zcorpan>
but consistency is nice ;)

[09:59:03.0000] <zcorpan>
i think class should be case insensitive in getElementsByClassName and querySelector too

[09:59:04.0000] <zcorpan>
in quirks mode

[10:00:00.0000] <Lachy>
zcorpan, mail public-webapi about that

[10:00:01.0000] <zcorpan>
Lachy: ok

[10:00:02.0000] <Lachy>
oh, but that probably doesn't need to be specced in there

[10:01:00.0000] <Lachy>
if the case insensitivity is defined in HTML5, and Selectors defines the semantics of the class selector, it can be safely left up to those specs.

[10:01:01.0000] <Lachy>
but send mail anyway, just incase I'm wrong about it.

[10:01:02.0000] <zcorpan>
yeah i concluded the same (that html5 can define it)

[10:01:03.0000] <zcorpan>
but ok

[10:05:00.0000] <Lachy>
oh, crap :-(

[10:06:00.0000] <Lachy>
IE8 doesn't even support querySelector in quirks mode

[10:06:01.0000] <zcorpan>
sad but not very surprising

[10:07:00.0000] <Lachy>
true, since they're using the IE5.5 quirks mode engine.

[10:09:00.0000] <Lachy>
hmm. WebKit won't uppercase class names at all in quirks mode.

[10:10:00.0000] <Lachy>
<p class="FOO"> won't be matched by document.querySelector(".FOO"); (but it does match if both are lowercase)

[10:10:01.0000] <Lachy>
/me goes to file a bug

[10:10:02.0000] <Philip`>
zcorpan: About <script src>: Did you count number of script elements, or number of pages?

[10:12:00.0000] <Philip`>
/me notes that he doesn't guarantee his output doesn't interleave elements from different pages, since it depends on what the thread scheduler decided to do

[10:43:00.0000] <takkaria>
ha!

[10:43:01.0000] <takkaria>
the TAG wants to make people use aria: rather than aria-

[10:46:00.0000] <Philip`>
The cost-benefit analysis seems to be missing that the 'colon' approach is a "new paradigm for 'namespace'"

[10:47:00.0000] <krijn>
Iria:tating :/

[10:47:01.0000] <Philip`>
where the new paradigm involves using getAttribute instead of getAttributeNS, and not requiring xmlns, and not allowing the prefix to be variable

[10:48:00.0000] <Philip`>
i.e. whatever it is, it's not XML Namespaces

[10:49:00.0000] <Philip`>
(but is close enough to confuse people into thinking that it is)

[10:49:01.0000] <BenMillard>
takkaria, krijn and Philip` are discussin the Q&A blog entry here, I take it? http://www.w3.org/QA/2008/05/syntax_for_aria_costbenefit_an.html

[10:50:00.0000] <krijn>
I have no idea, people should not notice my comments in here ;)

[10:50:01.0000] <Philip`>
BenMillard: I think so, via http://lists.w3.org/Archives/Public/public-html/2008May/0308.html

[10:50:02.0000] <takkaria>
that's what I was referring to, yes

[10:50:03.0000] <smedero>
there's also the last TAG minutes: http://www.w3.org/2001/tag/2008/05/08-minutes

[10:51:00.0000] <BenMillard>
looks like the author of the Q&A entry pasted the HTML output of a word processor document instead of conforming with the Q&A style and structure

[10:52:00.0000] <Philip`>
The colon thing would make more sense in a world that was going to transition away from text/html and to application/xhtml+xml, since then we'd end up with a relatively clean architecture and everything

[10:52:01.0000] <Philip`>
but that's not going to happen, and we'd be stuck with a colon mess in text/html

[10:53:00.0000] <takkaria>
Philip`: you think it's not going to happen, I daresay that the TAG think/hope it will...

[10:54:00.0000] <BenMillard>
from the minutes they consider this part of tagSoupIntegration-54 :(

[10:54:01.0000] <BenMillard>
so they look dead set on deprecating text/html as quickly as possible...still

[10:55:00.0000] <Philip`>
takkaria: There's a petabyte of legacy content that disagrees

[10:55:01.0000] <takkaria>
Philip`: your argument is invalid (that people haven't yet does not mean that they will) as well you know

[10:55:02.0000] <takkaria>
however true the conclusion is. :)

[10:56:00.0000] <zcorpan_>
Hixie: shouldn't it be degrade gracefully rather than fail gracefully, since the application still works in the given example?

[10:56:01.0000] <Philip`>
takkaria: The argument is that even if everyone in the world started producing XHTML, there's just too much HTML for it to ever go away

[10:59:00.0000] <takkaria>
hmm, the minutes are interesting

[10:59:01.0000] <Philip`>
(Of course that's untrue since there won't still be humans using HTML a million years from now; but I would hope at that point the transition is to something far more radical than XHTML)

[11:01:00.0000] <takkaria>
hmm, the minutes are interesting reading

[11:01:01.0000] <takkaria>
about the SVG WG: "I think they're the obvious case to lean on for a uniform solution. They're in XML and they already have a story that says it's ok to add new attributes in foreign namespaces anywhere you like."

[11:06:00.0000] <BenMillard>
takkaria, I agree

[11:07:00.0000] <shepazu>
takkaria: he said that about the SVG community, not the SVG WG

[11:08:00.0000] <takkaria>
shepazu: ah, my bad

[11:10:00.0000] <shepazu>
and I'm not sure I agree with him... it's now implemented, for better or worse, and that's the important thing

[11:12:00.0000] <zcorpan_>
Philip`: i counted number of script elements

[11:15:00.0000] <MikeSmith>
takkaria: btw, any update on your plans for HTML5 parsing implementation?

[11:15:01.0000] <takkaria>
MikeSmith: yeah, I have a position to start

[11:15:02.0000] <takkaria>
MikeSmith: as in, I was accepted

[11:16:00.0000] <takkaria>
I have a printout of the relevant bits of the spec now too

[11:16:01.0000] <takkaria>
it's exam period though, so I'm only really playing around with the existing code in my breaks

[11:18:00.0000] <hsivonen>
Philip`: Henry and I disagree on which is worse: colon but no NS processing or having a non-colon delimiter

[11:19:00.0000] <hsivonen>
Philip`: I think colon without NS processing is worse than going to non-NS processing using an untainted delimiter

[11:19:01.0000] <hsivonen>
people with Member access may be interested in http://lists.w3.org/Archives/Member/tag/2008May/0026.html

[11:20:00.0000] <Philip`>
People without Member access could be interested too :-)

[11:20:01.0000] <hsivonen>
yeah :-(

[11:21:00.0000] <MikeSmith>
Bobby "Blue" Bland had a great song called "Members Only"

[11:24:00.0000] <MikeSmith>
and Jimi Hendrix had a great song called "Castles Made of Sand"

[11:25:00.0000] <takkaria>
yeah, that's a fantastic song

[11:26:00.0000] <takkaria>
I'm trying to work out what could possibly end the permathread

[11:26:01.0000] <MikeSmith>
good luck with that

[11:26:02.0000] <takkaria>
I think some people just have yet to be convinced that there are cases where alt text is genuinely not available

[11:26:03.0000] <Philip`>
I imagine the one thing that won't end the thread is posting to it :-)

[11:26:04.0000] <takkaria>
/me grins

[11:27:00.0000] <Philip`>
Of course the other thing that won't end it is not posting to it

[11:28:00.0000] <takkaria>
the straw men make having a decent debate hard :(

[11:28:01.0000] <takkaria>
IME, this kind of thing is best resolved by sitting down and negotiating one-on-one, at least in real life

[11:29:00.0000] <Philip`>
You can defeat the straw men by starting a flamewar

[11:29:01.0000] <takkaria>
personally I find the a11y people including personal insults all the time to be on par with trying to start a flamewar

[11:30:00.0000] <Philip`>
I think "all the time" is quite an exaggeration

[11:31:00.0000] <Philip`>
and "the a11y people" is quite a term which I've forgotten

[11:31:01.0000] <Philip`>
Uh, a generalisation?

[11:33:00.0000] <takkaria>
well, of the posts I've made on that topic since I started it, I think most of the replies to me included some implications that I was thoughtless, naive, ignorant, presumptious etc

[11:33:01.0000] <takkaria>
(not all of the above in any one post, but one of them in most)

[11:34:00.0000] <MikeSmith>
I can imagine that some might say that the whole should-alt-be-required discussion is massive waste of time that could have been completely prevented by some better judgement on the part of the editor

[11:37:00.0000] <Philip`>
MikeSmith: Better technical judgement, or better appease-everyone-so-they-stop-arguing judgement?

[11:38:00.0000] <MikeSmith>
neither

[11:43:00.0000] <takkaria>
I increasingly think that asking a photographer-for-a-living to provide alt text for e.g. a portfolio website is actually insane, because the purpose of the portfolio is to show photographs

[11:43:01.0000] <Dashiva>
At least the strawmen are accessible strawmen :)

[11:44:00.0000] <hsivonen>
this is a reply to me but I don't understand its point: http://lists.w3.org/Archives/Public/public-html/2008May/0303.html can someone help me understand it?

[11:45:00.0000] <Philip`>
hsivonen: Not me, I'm afraid

[11:45:01.0000] <takkaria>
/me adds that posts to his list of reasons top-posting is bad

[11:45:02.0000] <hsivonen>
what's a 'lambda user'?

[11:46:00.0000] <Philip`>
takkaria: He could have put his reply at the bottom and it would not have made any more sense

[11:46:01.0000] <takkaria>
Philip`: no, but as a consequence of not top-posting, people tend to actually reply to points more than just go off on one

[11:47:00.0000] <Dashiva>
Philip`: The opposite to top-posting isn't bottom-posting, but context-posting

[11:47:01.0000] <takkaria>
that is, I believe inline quotes make people actually think about what they want to say. :)

[11:47:02.0000] <Philip`>
Dashiva: That seems an odd definition of "opposite"

[11:47:03.0000] <BenMillard>
Dashiva, it is the opposite in terms of position but not in terms of usefulness

[11:47:04.0000] <Philip`>
I'd just say it's one of several alternatives

[11:48:00.0000] <Philip`>
(and the non-context-based ones of those alternatives make messages hard to understand)

[11:48:01.0000] <Dashiva>
Philip`: Well, top-posting is really just outside-posting with the reply put at the bottom by the client as default

[11:49:00.0000] <Philip`>
Hmm, how about two-column email? Put the original in the left and line up your replies on the right

[11:50:00.0000] <Dashiva>
I've tried doing that in lecture notes. Code in one column, and comments to it in the other

[11:50:01.0000] <Dashiva>
The comment column got really easily out of sync, when a preceding comment linewrapped or similar. Needs quality UI to work :)

[11:50:02.0000] <hsivonen>
/me now sees the TAG posted the ARIA stuff to the WG

[11:50:03.0000] <hsivonen>
sigh

[11:51:00.0000] <Dashiva>
Philip`: Then again, you are the person who didn't fix Hixie's graph, so surely you can not do it

[11:53:00.0000] <Philip`>
Dashiva: I'm too busy to not do anything useful now, since I'm waiting to see if an order delivery status will change from "With Delivery Driver" to something more useful before I get bored and go home, and that takes up all my concentration

[11:54:00.0000] <Philip`>
(The delivery driver has had it for 11 hours now, so I don't see why it's taking them so long...)

[11:54:01.0000] <hsivonen>
I notice that the "cost/benefit analysis" was not amended to take into account feedback I gave f2f

[11:55:00.0000] <MikeSmith>
hsivonen: so please take some time to e-mail HT and point that out

[11:55:01.0000] <MikeSmith>
instead of bitching about it here

[11:56:00.0000] <hsivonen>
ok. here we go again

[11:57:00.0000] <MikeSmith>
hsivonen: yeah, I understand. I just don't know what the alternative is

[11:58:00.0000] <MikeSmith>
what I do know is that the alternative of me communicating to others about it by proxy is not practical alternative

[12:07:00.0000] <hsivonen>
MikeSmith: sure. I'm just not particularly happy about finding myself writing email about this again after thinking I already communicated f2f

[12:09:00.0000] <MikeSmith>
hsivonen: there are many things these days that I find myself needing to do that I'm not particularly happy about

[12:10:00.0000] <Philip`>
("Delivered: 14/05/2008 10:50:45" - oh, thanks for telling me eight hours later)

[12:14:00.0000] <Dashiva>
Well, I for one am pleased nobody has pulled the formal complaint thing again. 

[12:14:01.0000] <Philip`>
They probably noticed that it wasn't very effective last time

[12:29:00.0000] <Dashiva>
Ooh, bad voiceover diss in that mail

[12:29:01.0000] <Dashiva>
It's too bad AT doesn't compete on who's the best rather than who's the worst :)

[12:32:00.0000] <hsivonen>
MikeSmith: email sent

[12:34:00.0000] <takkaria>
hsivonen: you have a bias against making your own content accessible, did you know?

[12:35:00.0000] <MikeSmith>
hsivonen: thanks. as always

[12:35:01.0000] <hsivonen>
takkaria: do I?

[12:36:00.0000] <takkaria>
hsivonen: according to a reply to your last-but-one post on public-html

[12:44:00.0000] <zcorpan_>
hsivonen: do you agree with my <script src> analysis? would it make sense to implement the suggested checks?

[13:04:00.0000] <zcorpan_>
hsivonen: hmm, i never understood it as dispatching on qname and ignoring namespace, but rather support both {null, 'aria-foo'} and {...aria namespace..., 'foo'}

[13:05:00.0000] <zcorpan_>
hsivonen: although he's just been talking about the colon, not how it'd actually be implemented, so it's a bit unclear

[13:05:01.0000] <zcorpan_>
er

[13:05:02.0000] <zcorpan_>
s/aria-foo/aria:foo/

[13:12:00.0000] <virtuelv>
why can't we have aria*foo or aria _m_O~o_m_foo

[13:12:01.0000] <zcorpan_>
virtuelv: they are not well-formed in xml, unfortunately

[13:12:02.0000] <virtuelv>
pity

[13:12:03.0000] <zcorpan_>
indeed

[13:20:00.0000] <virtuelv>
I particularily like _m_O~o_m_

[13:20:01.0000] <virtuelv>
on an unrelated note: How far is CSS 2.1 from exiting CR state?

[13:21:00.0000] <zcorpan_>
a couple of thousand test cases away maybe? i don't know what happens with css these days

[13:21:01.0000] <virtuelv>
(provided we could drop the aural stuff, and say "CSS 3 module" for it

[13:22:00.0000] <virtuelv>
heh, which is only optional in the spec, I see

[13:22:01.0000] <zcorpan_>
speaking of which, i should check the status of the magic body stuff

[14:09:00.0000] <Hixie>
what's the law that hsivonen refers to sometimes to the effect that a technology ends up having as many components as working groups?

[14:10:00.0000] <Philip`>
Sounds like http://en.wikipedia.org/wiki/Conway's_Law

[14:10:01.0000] <hsivonen>
that's it

[14:11:00.0000] <Hixie>
aha!

[14:12:00.0000] <Hixie>
it's not in the list of adages on wikipedia

[14:12:01.0000] <Hixie>
/me adds it

[14:12:02.0000] <Philip`>
It's in http://en.wikipedia.org/wiki/List_of_adages_named_after_people

[14:12:03.0000] <Philip`>
Clearly there needs to be a list of lists of adages

[14:12:04.0000] <Hixie>
heh

[15:16:00.0000] <Dashiva>
Pragmatism, the lost art

[15:27:00.0000] <weinig>
annevk: ping

[15:37:00.0000] <gsnedders>
jgraham__: how do you get what Element.textContent would using lxml?

[15:40:00.0000] <hsivonen>
following implication arrows in the wrong direction seems to be too common

[15:40:01.0000] <jgraham__>
gsnedders: Recursion

[15:40:02.0000] <gsnedders>
jgraham__: k. so there is no simpler way :(

[15:40:03.0000] <jgraham__>
(or similar)

[15:41:00.0000] <jgraham__>
gsnedders: I don't know of a simpler way. There might be one I guess

[15:41:01.0000] <jgraham__>
hsivonen: Gez?

[15:46:00.0000] <Dashiva>
So much struggle to gain a tiny potential bit of accessibility for conforming documents, at the cost of so many non-conforming documents...

[15:47:00.0000] <gsnedders>
jgraham__: OK, I give in. lxml is quick enough :)

[15:48:00.0000] <Philip`>
gsnedders: If it's quick enough, your data set is too small

[15:48:01.0000] <gsnedders>
Philip`: I'm just operating on a couple month old copy of HTML 5

[15:48:02.0000] <gsnedders>
Philip`: And I doubt anything larger will be using it

[15:49:00.0000] <hsivonen>
jgraham__: among others

[15:50:00.0000] <Philip`>
gsnedders: What about HTML6?

[15:50:01.0000] <gsnedders>
/me wonders whether implementing DOM on top of lxml would be sane

[15:50:02.0000] <Philip`>
Given the way HTML5 is developed, the spec can never get smaller, because that would involve being less precise or dropping features

[15:51:00.0000] <Philip`>
so you need to be sufficiently forward-looking when developing tools

[15:51:01.0000] <takkaria>
Philip`: unless Hixie gets cloned and he can split out bits HTML5 into another spec, whilst still editing it all

[15:51:02.0000] <Philip`>
else you'll get the same problem that the W3C tools get when trying to process the HTML5 spec

[15:51:03.0000] <zcorpan_>
Philip`: it could drop non-normative stuff without being less precise or dropping features

[15:51:04.0000] <takkaria>
it's a possibility, anyway

[15:52:00.0000] <Philip`>
zcorpan_: Oh, yes

[15:52:01.0000] <Philip`>
/me wonders what the ratio of normative to non-normative text is

[15:52:02.0000] <zcorpan_>
Philip`: in terms of file size, it could also use less whitespace and less markup and less comments

[15:52:03.0000] <gsnedders>
Philip`: No, you get the problem of the real spec gen when you try and parse it multiple times

[15:52:04.0000] <Philip`>
(or, equivalently, the ratio of informative to uninformative text)

[15:53:00.0000] <zcorpan_>
Philip`: you could count paragraphs that contain normative keywords

[15:54:00.0000] <Hixie>
Philip`: dropping features happens

[15:59:00.0000] <Philip`>
zcorpan_: Hmm, only 36%

[15:59:01.0000] <Philip`>
(counting <p> elements that match /\b(must|required|should|recommended|may|optional)\b/i)

[15:59:02.0000] <gsnedders>
jgraham__: I'm getting a lot of errors with latest lxml and html5lib SVN

[15:59:03.0000] <zcorpan_>
Philip`: i meant "paragraphs" as defined in html5! not <p> elements! ;)

[16:00:00.0000] <Hixie>
you'd have to count sentences, not paragraphs

[16:00:01.0000] <Hixie>
and unfortunately some sentences make entire sections normative

[16:00:02.0000] <Hixie>
e.g. the algorithms introduced by "must do this"

[16:00:03.0000] <Hixie>
or the entities table

[16:00:04.0000] <gsnedders>
jgraham__: Actually, if I svn up, then it works. Remembering to do that helps :P

[16:01:00.0000] <Philip`>
zcorpan_: Get a browser to implement document.getParagraphs() and then I'll count that

[16:01:01.0000] <zcorpan_>
Philip`: if you file a bug on that i can buy a beer to the relevant developer to get it implemented in opera

[16:03:00.0000] <gsnedders>
16.032 seconds on my 2.4GHz Core 2 Duo laptop to parse html5 (the source version, not the postprocessed version) into a lxml structure

[16:03:01.0000] <gsnedders>
my only issue is the lack of support of XPath 2 :P

[16:04:00.0000] <Philip`>
gsnedders: If you save it as XML, how long does it take to parse it back again?

[16:04:01.0000] <gsnedders>
Philip`: next to nothing

[16:04:02.0000] <Hixie>
uh

[16:04:03.0000] <Hixie>
that's a long time

[16:04:04.0000] <Hixie>
that's probably about as long as the current processor

[16:05:00.0000] <Philip`>
Hixie: You should write the spec as XHTML

[16:05:01.0000] <jgraham__>
gsnedders: For postprocessing, saving it to a python pickle is possibly even faster than saving to XML

[16:05:02.0000] <Hixie>
Philip`: people should write better parsers :-P

[16:05:03.0000] <gsnedders>
Philip`: The bottleneck is html5parser.py:141(normalizedTokens)

[16:05:04.0000] <jgraham__>
gsnedders: ?! Really?

[16:06:00.0000] <jgraham__>
Surely the bottleneck is in the input stream?

[16:06:01.0000] <gsnedders>
jgraham__: I'm doing this locally, so no

[16:06:02.0000] <Philip`>
Seems it'd be quickest to have an external HTML-to-XHTML tool written in a real language, and then Python code can quickly load it and do whatever fancy processing it wants

[16:07:00.0000] <jgraham__>
gsnedders: Even for local disk access I expect getChar or charsUntill to take most of the time

[16:07:01.0000] <hsivonen>
zcorpan_: I haven't yet developed an opinion about the script issue, although I don't like the suggestion as a knee-jerk reaction

[16:07:02.0000] <jgraham__>
Because they are slooow and get called a lot of times

[16:07:03.0000] <gsnedders>
jgraham__: charsUtil takes 6 secs

[16:08:00.0000] <hsivonen>
zcorpan_: as for dispatch on qName, I'm very convinced based on information not in email that dispatch on qName is being put forth

[16:08:01.0000] <jgraham__>
So that's 1/3 of the time. How much does notmalizeToken take?

[16:08:02.0000] <jgraham__>
normalize

[16:08:03.0000] <hsivonen>
based on f2f info that is

[16:09:00.0000] <gsnedders>
/me reruns it, so he can just pastebin it

[16:09:01.0000] <gsnedders>
jgraham__: http://pastebin.ca/1018429

[16:09:02.0000] <gsnedders>
Now, back to working for my English exam in 12.75 hours

[16:10:00.0000] <gsnedders>
s/12/9/

[16:11:00.0000] <jgraham__>
gsnedders: You're reading the wrong column

[16:11:01.0000] <gsnedders>
/me looks again

[16:11:02.0000] <gsnedders>
seemingly yes :)

[16:11:03.0000] <jgraham__>
The first column is the interesting one

[16:12:00.0000] <gsnedders>
and charsUntil is the only significant one

[16:12:01.0000] <jgraham__>
(the third column includes time in code called by that function, so for normalizedTokens that's the whole tokenizer

[16:12:02.0000] <gsnedders>
jgraham__: yeah, I know

[16:12:03.0000] <gsnedders>
jgraham__: I was looking quickly :)

[16:12:04.0000] <gsnedders>
(too quickly)

[16:12:05.0000] <jgraham__>
Did you get this from parse.py?

[16:13:00.0000] <jgraham__>
/me wonders why the output isn't sorted

[16:13:01.0000] <Philip`>
The first column seems misleading since you could make the bottleneck not be the bottleneck by splitting it into lots of small functions

[16:13:02.0000] <gsnedders>
jgraham__: see the top line

[16:13:03.0000] <jgraham__>
gsnedders: Ah, OK

[16:14:00.0000] <jgraham__>
Philip`: Still if a single function is taking 1/3 of the total runtime is is a bottleneck...

[16:14:01.0000] <gsnedders>
Hixie: Can I just ignore your edge case until we have a html5 parser as a python extension? :P

[16:17:00.0000] <Hixie>
/me invites the tag to review html5 so that they don't come back in three years saying "you didn't invite us to review it! that's why we're sending you this feedback so late!"

[16:17:01.0000] <Hixie>
gsnedders: depends if you want me to use your script or not :-P

[16:18:00.0000] <gsnedders>
/me doesn't have strong feelings either way

[16:18:01.0000] <gsnedders>
/me just used you to get a subset of features that is actually used :P

[16:25:00.0000] <Lachy>
jgraham__, yt?

[16:25:01.0000] <jgraham__>
Lachy: Yep

[16:25:02.0000] <Lachy>
re the title for our presentation

[16:26:00.0000] <Lachy>
HTML5: From Use Cases to Solutions was ok, but I'd prefer to find something a little better

[16:26:01.0000] <jgraham__>
Lachy: Yeah, I agree it's not great

[16:28:00.0000] <jgraham__>
Do you have any ideas?

[16:28:01.0000] <Lachy>
these are the areas the talk will cover, and we need a title that sort of reflects them:

[16:28:02.0000] <Lachy>
1. Introduction to HTML5 - discuss the aims and objectives (5 min)

[16:28:03.0000] <Lachy>
2. The design principles (10 min)

[16:28:04.0000] <Lachy>
3. Constructing web sites with HTML5 - use cases, markup and DOM APIs (20 min)

[16:28:05.0000] <Lachy>
4. Talk about the HTML WG, WHATWG and the community surrounding it (10 min)

[16:31:00.0000] <Lachy>
Understanding HTML5: [insert subtitle]

[16:32:00.0000] <Hixie>
i recommend "HTML5"

[16:32:01.0000] <Lachy>
that's a bit boring

[16:33:00.0000] <Hixie>
if the excitement of your talk is dependent on the excitement of your title, you've got bigger problems :-)

[16:33:01.0000] <jgraham__>
Hixie: Whilst the title isn't really important for making the talk good, it might be important for making people turn up

[16:34:00.0000] <jgraham__>
This conference seems to be for people who value aesthetics

[16:34:01.0000] <Hixie>
"HTML5: Sex, Drugs, and Rock and Roll"?

[16:35:00.0000] <Lachy>
LOL

[16:35:01.0000] <jgraham__>
Glad it's worked like that for you, Hixie

[16:35:02.0000] <jgraham__>
:)

[16:36:00.0000] <jgraham__>
Something about moving the web forward?

[16:37:00.0000] <gsnedders>
"No sex, no drugs, no rock, no roll, when it comes to today"

[16:38:00.0000] <Hixie>
"HTML5: Why everyone else is wrong and we're right"

[16:38:01.0000] <Hixie>
you'd probably get people in for that one

[16:38:02.0000] <gsnedders>
"HTML5: It's not XHTML2"

[16:38:03.0000] <Lachy>
"HTML5: It's not that bad"

[16:38:04.0000] <Philip`>
"HTML5: Free money if you come to our talk"

[16:39:00.0000] <jgraham__>
How I learnt to stop worrying and love HTML5

[16:39:01.0000] <Lachy>
The HTML5 Generation

[16:39:02.0000] <jgraham__>
Or maybe just HTML5: How I learnt to stop worrying and love endless tedious arguments about alt

[16:39:03.0000] <gsnedders>
"HTML5: Not quite yet 2π"

[16:40:00.0000] <Dashiva>
jgraham__: ow

[16:40:01.0000] <Lachy>
HTML5: More Bikesheds than Anyone Else

[16:40:02.0000] <Dashiva>
How about just sticking to the classic? Understanding HTML5: 5 > 2

[16:40:03.0000] <Hixie>
we by far don't have more bikesheds than anyone else

[16:40:04.0000] <Hixie>
html5 has fewer bikesheds than most wgs i've been involved in

[16:40:05.0000] <Hixie>
by some quite wide margin

[16:40:06.0000] <Lachy>
yeah, I guess

[16:41:00.0000] <takkaria>
HTML5: Architectually Inconsistent, as declared by the TAG

[16:41:01.0000] <Dashiva>
What's your top three list, Hixie?

[16:41:02.0000] <jgraham__>
No architecture astronauts please, we're HTML5

[16:41:03.0000] <Hixie>
Dashiva: everyone else is tired for first spot. :-P

[16:41:04.0000] <zcorpan_>
"HTML5: More Powerplants than Anyone Else"

[16:41:05.0000] <Philip`>
How ironic that http://code.google.com/doctype/ has a broken doctype

[16:41:06.0000] <Lachy>
"Painting HTML5: Design, Development and the Community"

[16:41:07.0000] <takkaria>
2005: An HTML5 Odyssey

[16:42:00.0000] <gsnedders>
"HTML5: Hixie is God"

[16:42:01.0000] <Dashiva>
HTML5: A whitespace odyssey?

[16:42:02.0000] <jgraham__>
Lachy: That is almost good

[16:42:03.0000] <Lachy>
takkaria, it's 2008 now :-)

[16:42:04.0000] <takkaria>
I must have missed the year change again

[16:42:05.0000] <jgraham__>
HTML5: Hixie's Text Markup Language

[16:42:06.0000] <Hixie>
hah

[16:42:07.0000] <Dashiva>
+1

[16:43:00.0000] <gsnedders>
jgraham__: No, s/Text/Tough/

[16:43:01.0000] <jgraham__>
Toy?

[16:43:02.0000] <Dashiva>
Terrible?

[16:43:03.0000] <jgraham__>
Terrible?

[16:43:04.0000] <gsnedders>
/me expects we can come up with loads

[16:43:05.0000] <Dashiva>
Wait, there's that word

[16:43:06.0000] <gsnedders>
/me falls asleep

[16:43:07.0000] <Dashiva>
Someone used it in a mail... transigent!

[16:43:08.0000] <takkaria>
HTML5: With Added RFC2119 Goodness

[16:44:00.0000] <Philip`>
Dashiva: Is that really a word?

[16:44:01.0000] <Lachy>
HTML5: You MUST Come"

[16:44:02.0000] <Dashiva>
Philip`: It's intransigent without in

[16:44:03.0000] <Hixie>
bbl

[16:44:04.0000] <Lachy>
are you sure you didn't mean transient?

[16:44:05.0000] <Philip`>
The closest thing in my dictionary is transignification

[16:44:06.0000] <Dashiva>
Lachy: http://www.merriam-webster.com/dictionary/intransigent

[16:45:00.0000] <Dashiva>
I find the meaning fits really well too

[16:46:00.0000] <Lachy>
oh good, cause one definition of transient is "Decaying with time"

[16:46:01.0000] <Dashiva>
Hixie's Transsiberian Markup Language

[16:47:00.0000] <Lachy>
HTML5: Building the Foundation of the Web

[16:47:01.0000] <Philip`>
"Satanism manifests itself in different countries under various forms and names, such as Nihilism, Intransigentism, Petrolean Communism."

[16:48:00.0000] <Philip`>
Uh, does anyone happen to know what 'Petrolean' is?

[16:48:01.0000] <jgraham__>
Lachy: It would be good to get the idea across that we're not just going to blather on for an hour about the process but we're going to give people an idea of how they can use bits of HTML5 today

[16:48:02.0000] <Dashiva>
Venezuela?

[16:48:03.0000] <Dashiva>
HTML5: What will it do for you lately?

[16:49:00.0000] <j4_james>
 suck

[16:49:01.0000] <Lachy>
Dashiva, ask not what HTML5 can do for you, ask what you can do for HTML5!

[16:50:00.0000] <Dashiva>
Lachy: Sure, but that's after you get them into the room

[16:50:01.0000] <jgraham__>
HTML The Second Coming

[16:51:00.0000] <jgraham__>
/me wonders what fraction of the audience you could offend with that

[16:52:00.0000] <Dashiva>
You could always smooth it over with Second Parsing or something like that

[16:52:01.0000] <Lachy>
nothing wrong with offending people. It helps get people talking.

[16:52:02.0000] <Philip`>
The "browser compatibility" tables in e.g. http://code.google.com/docreader/#p(doctype)s(doctype)t(DialogElement) seem totally bogus

[16:52:03.0000] <Dashiva>
Again, after you get them into the room :)

[16:54:00.0000] <jgraham__>
So are we any closer to a serious idea?

[16:55:00.0000] <Dashiva>
"HTML5: Who, what and why"

[16:56:00.0000] <Lachy>
I liked "HTML5: Building the Foundation of the Web" (maybe s/Building/Developing/)

[16:56:01.0000] <Dashiva>
That sounds really like "replacing what we already had with something new and better" rather than "formalizing what we have and making it better" to me

[16:57:00.0000] <Lachy>
How about just "HTML5 Development"

[16:57:01.0000] <takkaria>
how about HTML5? :)

[16:58:00.0000] <jgraham__>
Lachy: My main problem with "Building the foundation of the web" is that it doesn't communicate the practical aspect of how to use it

[16:58:01.0000] <jgraham__>
HTML 5: A kickstart guide

[16:59:00.0000] <jgraham__>
Getting your hands dirty with HTML5

[16:59:01.0000] <Lachy>
yeah, that one is ok


2008-05-15
[17:00:00.0000] <jgraham__>
Great. Now all we need is a short summary of the talk 

[17:00:01.0000] <Lachy>
or how about we use this channel's topic: HTML5: Please leave your sense of logic at the door, thanks!

[17:01:00.0000] <jgraham__>
I would love to go with that but I don't know what effect it would have

[17:01:01.0000] <Lachy>
hmm, true. getting your hands dirty is fine

[17:02:00.0000] <jgraham__>
Given there has been quite a lot of negative blogging about HTML5 it might be best not to play up to that image (unless we could make it work for us, but given the short timescales and the logistical difficulties, it seems risky)

[17:02:01.0000] <Lachy>
ok, for the summary, we can just write 1 or 2 paragraphs that expand on that list of topics

[17:03:00.0000] <jgraham__>
Shall we both have a go and then edit them together somehow

[17:03:01.0000] <jgraham__>
Or google docs?

[17:03:02.0000] <Lachy>
sure, how does google docs help us work together?

[17:04:00.0000] <jgraham__>
(or write first then set up a collaborative document to edit)

[17:05:00.0000] <Lachy>
yeah, that would be better

[18:43:00.0000] <Hixie>
/me finally dives into the day's alt nonsense

[18:44:00.0000] <othermaciej>
oy vay

[18:44:01.0000] <Dashiva>
If you don't say something in 30 minutes, we'll send in a search time

[18:44:02.0000] <Dashiva>
*team

[18:45:00.0000] <othermaciej>
one day of that thread burned me out

[18:45:01.0000] <roc>
I haven't read a single message of it, but I can feel the pain from here

[18:49:00.0000] <Dashiva>
It's surreal at times. "We must make @alt mandatory to encourage good behavior via people desiring conformance" and then "Pages that aren't hand-written and don't get author input must resign themselves to being non-conforming"

[18:49:01.0000] <Lachy>
that alt thread is ridiculous. There were around 40-50 posts on it just today, and I just can't keep up with it

[18:49:02.0000] <Hixie>
the entire concept of requiring pages to be non-conforming

[18:49:03.0000] <Hixie>
is utterly non-sensical

[18:50:00.0000] <Hixie>
it's like having a minimum speed limit sign with a value higher than the speed limit sign next to it

[18:51:00.0000] <Dashiva>
The other thing that bugs me is that there seems to be an implicit assumption that AT aren't allowed to use the same techniques on non-conforming pages and conforming pages

[18:51:01.0000] <Hixie>
or a four way intersection with every exit marked with the "no entry" sign

[18:53:00.0000] <Dashiva>
(But if you hate people driving cars, that's not so unlikely ;)

[18:55:00.0000] <Philip`>
Hixie: It's sensical in some cases - e.g. if I want to make a page with tables for layout, then it's required to be non-conforming - so how is the distinction drawn between things that can be done in a conforming way and things that can't?

[18:56:00.0000] <Hixie>
you're not allowed to make a page with tables for layout

[18:56:01.0000] <Hixie>
and you should never do so

[18:57:00.0000] <Hixie>
that differs from this case, where there are people suggesting that in some cases, omitting alt (or giving bad alternate text) is what should happen

[18:57:01.0000] <Hixie>
despite making it non-conforming

[18:58:00.0000] <Philip`>
Ah, so it would be consistent if they said that when you want to make a page with omitted alt you should actually not make that page at all

[18:58:01.0000] <Hixie>
such an argument would not be persuasive

[18:59:00.0000] <Hixie>
but it would at least be consistent

[19:01:00.0000] <Philip`>
/me thinks "conforming" vs "non-conforming" is not a particularly useful way to look at pages - it's much more useful to look at a list of ways you can improve the quality of your page, and you should be disappointed when there are no suggestions left (because you've reached the limits of the conformance criteria)

[19:05:00.0000] <Hixie>
Philip`: yeah, maybe.

[19:05:01.0000] <Hixie>
Philip`: not sure how to formalise that though

[19:09:00.0000] <Philip`>
Require that conformance checkers have a friendly popup character saying "Hey! You might want to write &amp; instead of & here, to make sure it will never confuse a browser. Click here for a detailed explanation!" instead of a red box saying "Error: Text after & did not match an entity name."

[19:09:01.0000] <Philip`>
That would help people think of it as advice rather than admonishment

[19:11:00.0000] <Hixie>
ye-es

[19:11:01.0000] <Dashiva>
Wouldn't that just raise even more resistance, for "watering out" the concept of conformance?

[19:11:02.0000] <Philip`>
Hixie: You don't sound convinced :-(

[19:17:00.0000] <roc>
why don't you just have the checker transform the document to make it transforming

[19:17:01.0000] <roc>
to make it conforming, I mean

[19:27:00.0000] <mpt>
because that wouldn't improve the Web

[19:27:01.0000] <mpt>
because if a checker can do it, a browser can do it too

[19:29:00.0000] <roc>
the author can check the transformed document to make sure it still matches the author's intent

[19:29:01.0000] <roc>
browsers can't do that

[19:30:00.0000] <mpt>
aha

[19:31:00.0000] <mpt>
That's something that could be tested -- what percentage of authors would bother to check the autocorrected output

[19:32:00.0000] <roc>
I'm assuming they'd at least view the page

[19:32:01.0000] <Lachy>
roc, there are utilities that do such things. e.g. HTML Tidy

[19:33:00.0000] <Philip`>
If it presented the autocorrected document as a big lump of code, they'd be unlikely to check it all; I imagine it wouldn't be nearly so bad if it printed a suggested correction (or several) beside each error message, so authors could use that as a guide for fixing their code semi-manually

[19:35:00.0000] <mpt>
And for the particular issue of alt text, showing them the page won't help (unless you show it lynx-style)

[19:36:00.0000] <mpt>
(btw, alt="[A picture of a cat looking at pictures of cats]"? oh dea)

[19:36:01.0000] <mpt>
r

[19:38:00.0000] <Philip`>
/me guesses that IE8 stopping using alt for tooltips will do a lot to decrease the usage of alt

[19:40:00.0000] <mpt>
iirc Microsoft's reasoning for doing alt= as tooltip in the first place was to increase use of alt=

[19:40:01.0000] <mpt>
(in IE3, I think it was)

[19:41:00.0000] <mpt>
at which it undoubtedly succeeded, but with the wrong sort of contents

[19:41:01.0000] <Dashiva>
Oh, what wicked webs we weave

[19:41:02.0000] <mpt>
and it's been a long, slow walk back

[19:42:00.0000] <Dashiva>
Maybe some more focus should be put on @importantimage (or @realalt)

[19:42:01.0000] <Philip`>
It does make sense to make metadata visible, so that people actually notice it and use it

[19:42:02.0000] <mpt>
title= is metadata. alt= is data. ;-)

[19:43:00.0000] <Philip`>
I'd say the data is the image :-)

[19:43:01.0000] <mpt>
That's alternative data

[19:43:02.0000] <Dashiva>
Is invisible data worse or less bad than invisible metadata?

[19:43:03.0000] <Philip`>
The alt text is describing the image data, hence it's metadata

[19:43:04.0000] <Philip`>
(for a loose sense of 'describing')

[19:44:00.0000] <mpt>
Minus 5 points for using "describing" and "alt text" in the same sentence

[19:44:01.0000] <Philip`>
Dashiva: Invisible data is usually SEO spam, so it's bad

[19:46:00.0000] <Dashiva>
wget needs an AI mode to detect which URLs with get parameters are duplicates, and which are successive pages of multipage content

[19:48:00.0000] <Dashiva>
(I just now stopped a gallery download at 600 MB, and discovered only 20 MB of it was images)

[19:57:00.0000] <mpt>
/me imagines a world full of blind people, and search engines that only understand images, not text

[19:58:00.0000] <Philip`>
/me can't imagine who is publishing those images in such a world

[19:58:01.0000] <Dashiva>
mpt: Couldn't you just take a screenshot of the text you wanted and search for that? :)

[20:03:00.0000] <mpt>
Dashiva, right, and then Google gets confused by all the SEO spam images that contain that word, while human visitors see only the text

[20:04:00.0000] <mpt>
(or hear, rather)

[20:07:00.0000] <Philip`>
Hmph, the page I found a while ago with a 10KB alt attribute on an image has been changed and no longer has that image :-(

[20:17:00.0000] <mpt>
A picture is worth a thousand decabytes?

[20:34:00.0000] <Lachy>
I have to find a way to work some of these into the presentaion :-) http://icanhascheezburger.com/tag/html/

[01:39:00.0000] <Hixie>
wow, the e-mail i just sent was a reply to 41 e-mails of spiralling confusion

[01:48:00.0000] <hsivonen>
Hixie: IIRC Maciej said Safari uses QuickTime on Windows

[01:48:01.0000] <Hixie>
the details don't matter in this particular instance

[01:50:00.0000] <mcarter>
I wonder how many people in here reside somewhere near silicon valley? I don't suppose you guys ever get together for an html5 lunch?

[01:51:00.0000] <Hixie>
i'm in mountain view and would be up for #whatwg social outings, so long as i don't have to organise them :-)

[01:51:01.0000] <Hixie>
it hasn't happened so far though

[01:51:02.0000] <Hixie>
a number of us are in the area

[01:51:03.0000] <mcarter>
hmm, I'm in mountain view more than half the time

[01:52:00.0000] <mcarter>
I'll see about putting together a wiki somewhere for a social meeting

[01:52:01.0000] <mcarter>
i finally conquered IE tonight.. that is, just got sse working in IE with only a small caveat or two for a server-side implementer

[01:53:00.0000] <Hixie>
nice

[01:54:00.0000] <mcarter>
there's apparently no way to keep IE from downloading things that don't have content-type "text/html" or "text/plain", as far as I can tell

[01:54:01.0000] <Hixie>
as in it pops up a dialog box?

[01:54:02.0000] <Hixie>
when doing what?

[01:54:03.0000] <Hixie>
surely image/jpeg doesn't get downloaded

[01:54:04.0000] <mcarter>
well, i meant as an iframe src

[01:55:00.0000] <Hixie>
yeah i always hate it when browsers cause you to download stuff when an _iframe_ has content that isn't supported

[01:55:01.0000] <Hixie>
that's just dumb

[01:55:02.0000] <Hixie>
though i may have made html5 require that behaviour. oops. i'll have to check that out.

[01:55:03.0000] <mcarter>
i never noticed it until i tried to implement this sse stuff for IE

[01:56:00.0000] <mcarter>
I can't think of any other way to do it then to locally poll a plain text iframe, and parse the contents

[01:58:00.0000] <mcarter>
you can expect that feedback for TCPConnection on friday... particularly now that I've got a working demo

[01:58:01.0000] <mcarter>
and with that, i'm off to bed (do you ever sleep?)

[02:00:00.0000] <Hixie>
cool, nn

[02:00:01.0000] <Hixie>
i sleep around 4am - 1pm

[02:13:00.0000] <Philip`>
Hixie: "the ratio of the correct rendered width of each pixel to the actual width of each pixel in the image (i.e., the multiple by which the video's intrinsic width must be multiplied to obtain the rendered width that gives the correct aspect ratio)" - that seems like an abuse of the word "must"

[02:14:00.0000] <Hixie>
oops

[02:14:01.0000] <Hixie>
will fix

[02:25:00.0000] <Lachy>
Hixie, which video formats require pixelratio="" in the markup?

[02:26:00.0000] <Hixie>
h.263, at least

[02:26:01.0000] <Lachy>
ok.

[02:26:02.0000] <Hixie>
youtube also suggested it so that misencoded videos (very common apparently) can be fixed up without reencoding them (since that's really expensive)

[02:26:03.0000] <Lachy>
I thought all that info would have been specified in the container format, if not in the codec itself.

[02:26:04.0000] <Hixie>
yeah, me too

[02:27:00.0000] <Hixie>
i pushed back on it quite a bit but apparently anamorphic videos are quite common and common formats don't handle them

[02:27:01.0000] <Lachy>
wow. I thought only DVDs used anamorphic content, since they needed to scale to fit 4:3 and 16:9 TVs.

[02:27:02.0000] <Hixie>
Philip`: fixed

[02:28:00.0000] <Hixie>
intentionally anamorphic video is common in professional workflows

[02:28:01.0000] <Lachy>
ok

[02:28:02.0000] <Hixie>
when i was an amateur projectionist at university we had to deal with anamorphic lenses all the time for both 16mm and 35mm movies

[02:28:03.0000] <Hixie>
but pixelratio is more intended for misencoded videos

[02:29:00.0000] <Hixie>
they also asked for things like fixing up brightness and contrast, but i figured css filters were a better solution to that

[02:29:01.0000] <Hixie>
the last thing they really asked for which i thought made sense was a way to detect if the UA was throttling the download speed

[02:30:00.0000] <Lachy>
I thought youtube reencoded all videos anyway, since it has to convert them all to the right format for flash

[02:30:01.0000] <othermaciej>
how are you supposed to detect that? (the UA just admitting it?)

[02:30:02.0000] <Hixie>
right

[02:30:03.0000] <othermaciej>
why do they want to know that?

[02:30:04.0000] <Hixie>
video.throttling or video.isDownloadThrottled were the two proposals i had for attribute names

[02:31:00.0000] <othermaciej>
is it more useful info than whether a firewall or a traffic shaping kernel module are throttling the download speed?

[02:31:01.0000] <Hixie>
othermaciej: the spec allows UAs to artificially throttle downloads to prevent videos from being downloaded until the user wants them

[02:32:00.0000] <othermaciej>
Hixie: I guess I am asking what the use case for this information is

[02:32:01.0000] <Hixie>
othermaciej: but the argument is that video players (such as youtube's) aren't going to want to be sitting there saying "bandwidth too low! stalled!" when in fact it's just the UA that's blocking the download

[02:32:02.0000] <Hixie>
since they'd look silly :-)

[02:32:03.0000] <Hixie>
better to be able to say "download paused" in some way

[02:33:00.0000] <othermaciej>
in WebKit what we would likely do is load right away but not play the video until the user switches to that tab (if autoplay is set and it is in a background tab)

[02:34:00.0000] <Hixie>
(actually the throttling feedback was from the google video team's experience, not youtube. but there's little practical difference these days.)

[02:35:00.0000] <Hixie>
othermaciej: right but some people want to be able to tell their UA to just not download the video at all yet, or to stop downloading it

[02:35:01.0000] <Hixie>
the example given in this e-mail was a 3 hour clip on google video, where the user wants to stop the download after 5 minutes are downloaded

[02:35:02.0000] <Hixie>
so he right clicks and chooses "pause download" or whatever the UA (probably opera) provides

[02:36:00.0000] <Hixie>
now google video's player doesn't want to say "eep! your connection died!" it just wants to say "(download paused)"

[02:36:01.0000] <othermaciej>
so is throttled just supposed to represent stopping/pausing the load, or does anyone actually want to keep loading but at reduced bandwidth?

[02:37:00.0000] <othermaciej>
(not sure we care about a way to stop short of closing the page but that seems like at least conceivably a good idea)

[02:38:00.0000] <Hixie>
well the spec already provides the bits-per-second-over-the-last-few-seconds information (bufferingRate) so you could distinguish the two cases (paused vs throttled) by seeing if that number was non-zero

[02:38:01.0000] <Hixie>
<Hixie> well the spec already provides the bits-per-second-over-the-last-few-seconds information (bufferingRate) so you could distinguish the two cases (paused vs throttled) by seeing if that number was non-zero

[02:39:00.0000] <Hixie>
and i'd be shocked if opera didn't want to support precise control over the download rate

[02:39:01.0000] <othermaciej>
sorry about the flaky connection here

[02:39:02.0000] <othermaciej>
really? does opera have download rate controls for other things?

[02:39:03.0000] <Hixie>
opera has ui to do everything under the sun

[02:39:04.0000] <othermaciej>
/me tries not to actually use Opera's UI when testing in it

[02:42:00.0000] <Hixie>
ok bufferingThrottled is probably the value that most closely fits in with the existing API

[02:43:00.0000] <Hixie>
i wish other video sites had sent feedback

[02:46:00.0000] <Hixie>
crap, that's all the video feedback i had

[02:46:01.0000] <Hixie>
that means <iframe> is next

[02:48:00.0000] <Hixie>
good

[02:48:01.0000] <Hixie>
ak

[02:49:00.0000] <hsivonen>
http://lists.w3.org/Archives/Public/wai-xtech/2008May/0163.html

[03:00:00.0000] <Hixie>
hsivonen: http://1997.webhistory.org/www.lists/www-talk.1993q1/0262.html

[03:00:01.0000] <Hixie>
http://1997.webhistory.org/www.lists/www-talk.1993q1/0182.html specifically

[03:01:00.0000] <Hixie>
wow, guido was involved in the web early

[03:01:01.0000] <hsivonen>
Hixie: thanks

[03:02:00.0000] <Hixie>
tbl's original reply to <img>: http://1997.webhistory.org/www.lists/www-talk.1993q1/0186.html

[03:03:00.0000] <othermaciej>
shades of his turn to the dark side

[03:03:01.0000] <Hixie>
that message contains the first example of alternate text, btw, as far as i can tell from his vague description of how it should work, except that it is terible alternate text if so...

[03:05:00.0000] <Hixie>
i don't think he realised that's what it was for though, or at least he didn't mention it was for that

[03:06:00.0000] <Hixie>
thank goodness we didn't go down THIS path http://1997.webhistory.org/www.lists/www-talk.1993q1/0202.html

[03:07:00.0000] <othermaciej>
img is so much better than either of those options

[03:07:01.0000] <Hixie>
ok i can firmly say that there is absolutely nothing in that thread talking about alternatives to visual media

[03:07:02.0000] <othermaciej>
(though with benefit of hindsight we might prefer if it had spelled out "image" and not been a void element)

[03:08:00.0000] <Hixie>
ooo, while we're at it: DanC saying that "most of the technical issues in the HTML spec have been ironed out": http://1997.webhistory.org/www.lists/www-talk.1993q1/0041.html

[03:09:00.0000] <Hixie>
/me notes that FIFTEEN YEARS LATER we're still fixing issues that were relevant and not resolved back then

[03:11:00.0000] <Philip`>
http://1997.webhistory.org/www.lists/www-talk.1993q3/0985.html

[03:11:01.0000] <Philip`>
Oh, http://1997.webhistory.org/www.lists/www-talk.1993q3/0461.html seems earlier

[03:13:00.0000] <Philip`>
/me doesn't see any earlier references on that site

[03:17:00.0000] <Hixie>
so it took 6 months for someone to suggest alt="", and even then it was for text browsers not blind users

[03:17:01.0000] <Hixie>
heh

[03:17:02.0000] <Hixie>
bolt-on indeed

[03:17:03.0000] <Philip`>
http://1997.webhistory.org/www.lists/www-talk.1993q3/0995.html would have been a nicer way of doing it

[03:18:00.0000] <Hixie>
yep

[03:18:01.0000] <Hixie>
the funny thing about that (which is what <object> does, btw) is that there is no way to omit vs not omit the alt attriute

[03:19:00.0000] <Hixie>
which means there's no way to distinguish inaccessible images from decorative images

[03:20:00.0000] <Lachy>
Hixie, re this comment from a few days ago "<Hixie> i think for explaining <meter> better we should have some pictures of gagues", I have such images that I used in a previous presentation, if you want them.

[03:20:01.0000] <Hixie>
sure!

[03:21:00.0000] <Lachy>
they're in these powerpoint slides. http://lachy.id.au/dev/presentation/developing-with-html5/ - I can send you the originals when I get home later this evening

[03:32:00.0000] <tomg>
ohh, http://code.google.com/doctype/ is nice

[03:35:00.0000] <Philip`>
/me didn't see much on it other than invalid HTML, failure to work in Opera, incorrect test results, and advocacy of yet another JavaScript library, but maybe he was looking too cynically :-)

[03:36:00.0000] <Hixie>
ooo, ppt opens in keynote

[03:36:01.0000] <Hixie>
Philip`: send comments to mark pilgrim :-)

[03:37:00.0000] <Philip`>
Hixie: I don't have any comments much more constructive than "the site doesn't work, and when it does work it's wrong" :-p

[03:37:01.0000] <Hixie>
:-)

[03:37:02.0000] <tomg>
iWork is brilliant 

[03:37:03.0000] <hsivonen>
notes that Google is breaking the Web by moving addressing after the hash

[03:38:00.0000] <Hixie>
Lachy: i only saw one <meter> in that ppt and it was a rating example, not a gauge :-)

[03:39:00.0000] <Hixie>
hsivonen: yeah, i'm not a huge fan of that either

[03:40:00.0000] <Lachy>
I have others that weren't used in the presentation. I may have a gauge at home.

[03:40:01.0000] <Lachy>
if not, I'll make one.

[03:40:02.0000] <Hixie>
cool :-)

[03:41:00.0000] <hsivonen>
I wonder how Google search is going to index Google docreader using public HTTP spidering

[03:42:00.0000] <hsivonen>
that is, if someone else writes a similar reader, what happens to Web search?

[03:42:01.0000] <Philip`>
http://www.google.com/search?q=inurl%3Acode.google.com%2Fdocreader

[03:42:02.0000] <Philip`>
I think the answer to how it'll index it is "not very well"

[03:42:03.0000] <Hixie>
you don't have to use the ajax reader

[03:42:04.0000] <Hixie>
there's a separate view of the wiki that isn't ajax

[03:43:00.0000] <hsivonen>
Hixie: what if people out there link to the ajax reader?

[03:43:01.0000] <Hixie>
then the links don't contribute pagerank

[03:43:02.0000] <Hixie>
i didn't say it was a good idea :-)

[03:43:03.0000] <Hixie>
i was just providing further information

[03:43:04.0000] <hsivonen>
ok :-)

[03:44:00.0000] <Philip`>
Is there a way to access the non-AJAX version without disabling JS or copying-and-pasting the non-JS URL?

[03:45:00.0000] <Hixie>
i don't believe there are any links on http://code.google.com/doctype/ that don't rewrite themselves to point to the ajax reader, no

[03:45:01.0000] <Hixie>
but i don't know

[03:45:02.0000] <Hixie>
open in new tab seems to work

[03:46:00.0000] <tomg>
sigh. could someone throw away the british rail network please. 

[03:46:01.0000] <Philip`>
/me notes that Googlebot follows <a href="window.location='foo'"> links, and seemingly no other search engine does

[03:47:00.0000] <Philip`>
Uh

[03:47:01.0000] <Philip`>
<a href="javascript:...">

[03:48:00.0000] <Hixie>
how cunning of us

[03:48:01.0000] <Philip`>
as well as following any strings in <script>s that start with a '/' character, even if they're commented out

[03:49:00.0000] <Philip`>
(where 'strings' means quote-delimeted strings)

[03:49:01.0000] <Philip`>
s/incorrect spelling/correct spelling/

[03:49:02.0000] <Hixie>
really?

[03:49:03.0000] <Hixie>
that seems like a bug

[03:49:04.0000] <Hixie>
do you have a test page showing this?

[03:52:00.0000] <virtuelv>
Philip`: does that mean it's possible to DoS googlebot?

[03:54:00.0000] <Philip`>
Hixie: http://canvex.lazyilluminati.com/bottest.html and http://canvex.lazyilluminati.com/bottest2.html and http://canvex.lazyilluminati.com/bottest3.html

[03:55:00.0000] <hsivonen>
I noticed that XTech caused me to forget to rerun the monthly V.nu stats for April

[03:55:01.0000] <Philip`>
and my logs show requests from Googlebot for http://canvex.lazyilluminati.com/misc/bottest-requests.txt

[03:57:00.0000] <Hixie>
Philip`: that's really odd

[03:57:01.0000] <Hixie>
Philip`: i've seen the code that picks urls and i don't understand how you re getting those results

[03:58:00.0000] <Philip`>
Oh, there's also NextGenSearchBot following all the links except for /1 on bottest3.html

[03:58:01.0000] <Philip`>
but that's the only one

[03:59:00.0000] <hsivonen>
hmm. I think Gez's latest email can lead to constructive discussion

[04:00:00.0000] <hsivonen>
but before I reply, I think I should write some code 

[04:00:01.0000] <hsivonen>
because otherwise the mailing list will starve my code writing tasks

[04:00:02.0000] <Philip`>
Hixie: Well, don't ask me :-p

[04:04:00.0000] <hsivonen>
it's a shame that Google doctype doesn't have Opera 9.5 in its test matrix

[04:05:00.0000] <hsivonen>
/me didn't know <plaintext> had @align

[04:05:01.0000] <tomg>
yes

[04:05:02.0000] <Hixie>
hsivonen: it's a wiki :-D

[04:06:00.0000] <Hixie>
though, is 9.5 out yet?

[04:06:01.0000] <othermaciej>
Google doctype says some things that seem wrong

[04:06:02.0000] <hsivonen>
Hixie: FF3 isn't either

[04:06:03.0000] <tomg>
or IE8 :)

[04:07:00.0000] <othermaciej>
http://code.google.com/docreader/#p(doctype)s(doctype)t(DocumentParentWindowProperty)

[04:07:01.0000] <othermaciej>
I don't think it is correct to claim that no browser implements it

[04:07:02.0000] <othermaciej>
or rather, if that is true, why is it documented at all?

[04:08:00.0000] <othermaciej>
but I am pretty sure IE does actually in fact support it

[04:09:00.0000] <tomg>
as does Opera

[04:09:01.0000] <othermaciej>
also it seems odd that the DOM reference includes vendor extensions but the CSS reference seemingly does not

[04:09:02.0000] <othermaciej>
(except for broken vendor extensions that aren't prefixed)

[04:10:00.0000] <Hixie>
hsivonen: fair enough

[04:11:00.0000] <Hixie>
othermaciej: it's a wiki, feel free to edit it :-)

[04:11:01.0000] <othermaciej>
so 'accelerator', 'behavior' and 'layer-background-image' are included but not -webkit-box-shadow

[04:11:02.0000] <tomg>
can you add pages?

[04:11:03.0000] <Hixie>
othermaciej: i assume CSS stuff will be added in due course

[04:11:04.0000] <Hixie>
tomg: sure

[04:11:05.0000] <tomg>
oh yes

[04:11:06.0000] <othermaciej>
I certainly do not have the personal resources to do research on CSS properties

[04:11:07.0000] <Hixie>
othermaciej: as far as i know there's no policy to exclude prefixed extensions

[04:11:08.0000] <tomg>
that'll be the New Page link then ;)

[04:11:09.0000] <Hixie>
othermaciej: yeah join the club :-)

[04:12:00.0000] <tomg>
it'll be good to document Opera and WebKit stuff 

[04:12:01.0000] <tomg>
they are both really lacking in documentation 

[04:12:02.0000] <othermaciej>
it would be handy if it were marked which are part of what standard (if any) in addition to listing browser support

[04:13:00.0000] <othermaciej>
is there a bug tracker other than "it's a wiki, you can edit it"?

[04:13:01.0000] <tomg>
heh

[04:13:02.0000] <tomg>
yes it really needs a mass-import of DOM references and APIs

[04:14:00.0000] <Philip`>
Is there any HTML element for which the browser compatibility table is correct?

[04:14:01.0000] <othermaciej>
because I am willing to file bugs but not willing to do the research to personally fix them

[04:14:02.0000] <Philip`>
tomg: Why import, rather than just linking to MDC?

[04:14:03.0000] <Philip`>
/me prefers not having too many sources for the same information

[04:15:00.0000] <tomg>
er, because the MDC is for Gecko?

[04:16:00.0000] <othermaciej>
funny that someone made automated test cases for all of these but did not make the compat tables match

[04:16:01.0000] <tomg>
you'd have thought with making a cross-browser reference you'd want to at least start with a page about every CSS extension

[04:17:00.0000] <Philip`>
tomg: The MDC is (mostly) for standards, and the DOM references and APIs should be for standards too

[04:17:01.0000] <othermaciej>
tomg: wow, the html element tables are so wrong it is scary

[04:17:02.0000] <tomg>
yes

[04:18:00.0000] <othermaciej>
I love the one for the <video> element

[04:18:01.0000] <tomg>
errrrrrrrr

[04:18:02.0000] <tomg>
that needs some amending 

[04:19:00.0000] <othermaciej>
/me wonders if "giant repository of incomplete and largely wrong information" is a useful service

[04:19:01.0000] <othermaciej>
but then again, people find wikipedia useful

[04:20:00.0000] <Philip`>
Are all the compatibility tables hard-coded into the wiki pages, so there's no way to run tests automatedly and update the tables?

[04:20:01.0000] <Hixie>
othermaciej: yeah, there's an issues database

[04:20:02.0000] <tomg>
nothing supports <video> does it>

[04:20:03.0000] <othermaciej>
Safari 3.1 supports <video>

[04:20:04.0000] <tomg>
oh.

[04:20:05.0000] <tomg>
yes

[04:20:06.0000] <othermaciej>
also supports styling it

[04:21:00.0000] <Hixie>
othermaciej: http://code.google.com/p/doctype/issues/list

[04:21:01.0000] <tomg>
that's the only one

[04:21:02.0000] <Hixie>
http://code.google.com/p/doctype/issues/entry to file a new one

[04:21:03.0000] <othermaciej>
Hixie: thanks

[04:22:00.0000] <Hixie>
Philip`: thanks for the links, i have learnt something new about google that i didn't before. :-)

[04:30:00.0000] <hsivonen>
Hixie: you aren't acking the self-closing flag on isindex. Why is that?

[04:30:01.0000] <Hixie>
isindex isn't valid anyway

[04:30:02.0000] <hsivonen>
so why produce one more error?

[04:30:03.0000] <Hixie>
you don't have to show it :-)

[04:31:00.0000] <Hixie>
if you want the spec changed send mail

[04:31:01.0000] <Hixie>
i don't really have a strong opinion

[04:31:02.0000] <Hixie>
i just didn't want to ack anything that wasn't valid

[04:31:03.0000] <hsivonen>
making the delta between normative and shown errors bigger is not nice for interoperable test cases

[04:31:04.0000] <Hixie>
(there's no real good reason for such wants)

[04:31:05.0000] <hsivonen>
you ack wbr

[04:31:06.0000] <Hixie>
doesn't wbr come along with a bunch of valid elements?

[04:32:00.0000] <Hixie>
as far as i recall i just did whatever was the smallest change

[04:34:00.0000] <hsivonen>
email sent

[04:35:00.0000] <annevk>
wbr should become valid anyway :)

[04:35:01.0000] <annevk>
or maybe just nobr, hmm

[04:35:02.0000] <hsivonen>
that too

[04:36:00.0000] <hsivonen>
speaking of test cases and the self-closing flag, we need to change the tokenizer test format

[04:36:01.0000] <hsivonen>
for backwards compat, we probably should make self-closing appear in JSON only when true

[04:36:02.0000] <annevk>
wtf is up with this ARIA debate

[04:37:00.0000] <annevk>
"Next steps for the ARIA syntax discussion"

[04:37:01.0000] <Hixie>
i think the next step is called "shipping"

[04:37:02.0000] <hsivonen>
lol

[04:38:00.0000] <hsivonen>
also, do we want test cases to track the self-closing flag on end tag tokens?

[04:38:01.0000] <annevk>
yeah, maybe it's better to not reply for a month or so than put energy into this

[04:39:00.0000] <Hixie>
i was telling hsivonen earlier that my recommendation is just to reply to the tag e-mail, quoting just the part that suggests a colon, and say, basically: "we have a requirement that the same script access to the attribute value work in XML and HTML without syntactical differences in the markup. this doesn't address that requirement."

[04:39:01.0000] <Hixie>
in no more than one paragraph

[04:39:02.0000] <Hixie>
without any fluff

[04:42:00.0000] <annevk>
"But I am also in the small minority of HTML authors who have read the spec and try to follow it." after he just said that <b> and <i> are obsolete and you need to use <span> instead...

[04:43:00.0000] <othermaciej>
if Google actually uses this bogus UA check code internally then I am very afraid: http://code.google.com/docreader/#p(doctype)s(doctype)t(ArticleUserAgent)

[04:46:00.0000] <hsivonen>
did someone already put hendry's vim script on the WHATWG wiki?

[04:48:00.0000] <Hixie>
othermaciej: what's wrong with it?

[04:48:01.0000] <othermaciej>
well, I filed a bunch of things I found wrong here: http://code.google.com/p/doctype/issues/detail?id=8

[04:49:00.0000] <othermaciej>
I would expect there are also mistakes relating to browsers where I am less familiar with the UA string

[04:49:01.0000] <annevk>
Hixie, the blacklist is in AC as well for the theoretical case of another spec wanting to do something similar to XHR

[04:50:00.0000] <Hixie>
annevk: like what?

[04:50:01.0000] <annevk>
another type of XMLHttpRequest

[04:51:00.0000] <Hixie>
othermaciej: oh, those aren't major issues

[04:51:01.0000] <Hixie>
othermaciej: i thought you'd found major problems :-)

[04:51:02.0000] <annevk>
I don't think we should have that, but the spec could handle it in theory

[04:51:03.0000] <othermaciej>
using indexOf() to implement the check is pretty bad

[04:51:04.0000] <othermaciej>
as is identifying Safari as "khtml"

[04:51:05.0000] <annevk>
that's why the spec also defines the preflight request itself, etc.

[04:51:06.0000] <Hixie>
annevk: i don't understand why you'd ever want to allow all headers on an author-controlled cross-domain request

[04:52:00.0000] <Hixie>
othermaciej: *shrug* it doesn't break anything though

[04:52:01.0000] <annevk>
Hixie, but then we'd be stuck with Accept and Accept-Language ...

[04:53:00.0000] <Hixie>
othermaciej: i agree it's not perfect, don't get me wrong; but it's not stop-ship or anything

[04:53:01.0000] <Hixie>
annevk: so? we are anyway, no?

[04:53:02.0000] <annevk>
the current proposal handles arbitrary headers

[04:53:03.0000] <annevk>
though it requires a preflight request for those not in the whitelist nor in the blacklist

[04:53:04.0000] <annevk>
in the GET case

[04:54:00.0000] <othermaciej>
it's also pretty sloppy code in general

[04:54:01.0000] <othermaciej>
certainly not worthy of being held up as an example

[04:55:00.0000] <annevk>
hmm

[04:55:01.0000] <annevk>
Jonas' second case might be solved by simply converting \ to /

[04:56:00.0000] <annevk>
which I think is what some UAs already do...

[04:56:01.0000] <othermaciej>
if the utility functions are really used by google then they make good fodder for possibly APIs to add as native

[04:59:00.0000] <hendry>
hsivonen: not from the recent changes page i can see

[05:01:00.0000] <Hixie>
<Hixie> othermaciej: i agree it's not perfect, don't get me wrong; but it's not stop-ship or anything

[05:01:01.0000] <Hixie>
i agree it's not good sample code, certainly

[05:02:00.0000] <Hixie>
but this is just the first release :-)

[05:02:01.0000] <othermaciej>
I am guessing (perhaps wrongly) that since this stuff has goog.whatever all over it that it comes from some internal code snippet repository

[05:02:02.0000] <othermaciej>
thus I hope fixes make it back to the original

[05:03:00.0000] <othermaciej>
Hixie: I have found much more bogus UA checks in Google apps before though, so I guess it could be worse

[05:03:01.0000] <Hixie>
annevk: wait so you _can_ set arbitrary headers? i thought it was excluively whitelist limited.

[05:03:02.0000] <Hixie>
othermaciej: i don't know if this stuff is from a particular widely used framework, actually

[05:03:03.0000] <Hixie>
particularly

[05:04:00.0000] <BenMillard>
Lachy, on 13th May 2008 you wrote "what we really need for the alt discussion to get somewhere is a study that looks at the reasons why people use alt="" or alt="bogus, validator friendly content" or whatever. The difficulty is in figuring out how to perform such a study" -- http://krijnhoetmer.nl/irc-logs/whatwg/20080513#l-546

[05:04:01.0000] <annevk>
Hixie, yes you can

[05:04:02.0000] <Hixie>
annevk: oh. then nevermind, the blacklist does take effect.

[05:04:03.0000] <annevk>
Hixie, hence the proposal from Jonas I guess to restrict it somehow to only what the server allows

[05:04:04.0000] <BenMillard>
Lachy, the approach I used for Collections of Interesting Data Tables could fit that? http://sitesurgeon.co.uk/tables/

[05:04:05.0000] <othermaciej>
(I recall one trying to identify Safari based on having document.contains and lacking some method on element, which broke when we moved contains() from Node to Element to match IE)

[05:05:00.0000] <Hixie>
othermaciej: heh

[05:05:01.0000] <othermaciej>
Hixie: the sad thing is that it was then using this weird way of detecting Safari to make assumptions about completely unrelated things, like which editing APIs are present

[05:06:00.0000] <Hixie>
that's sad

[05:06:01.0000] <othermaciej>
I think someone got the idea that capability testing is better than UA testing but the message got garbled in transmission

[05:06:02.0000] <Hixie>
yeah really

[05:15:00.0000] <Lachy>
they were probably trying to work around the problem of browsers lying about user agent strings. But such detection is only necessary when you actually want to use the APIs your testing for.

[05:18:00.0000] <Lachy>
BenMillard, it could work. But, as I said a few minutes after that, how do you determine an authors reasons based solely upon what they did?

[05:20:00.0000] <hsivonen>
Lachy: you hire an economist :-)

[05:20:01.0000] <Lachy>
if the hypothesis is that validators insisting upon the alt attribute encourage people to simply make the validator happy, then what predictions can be made from that, which can be tested?

[05:20:02.0000] <Dashiva>
Lachy: That some useful images will have alt=""

[05:21:00.0000] <Dashiva>
(and alt="image" and alt="alt" and alt="stop bothering me stupid validator I hate accessibility" :)

[05:21:01.0000] <BenMillard>
Lachy, e-mailing every author whose markup I collect to ask why they did it that way might actually work, although I understand hsivonen's reservations about that

[05:22:00.0000] <BenMillard>
but what authors actually do is more relevant than why they do it, imho

[05:24:00.0000] <Lachy>
BenMillard, no, why they do it is more relevant when trying to defend a position that is itself a reason for why

[05:24:01.0000] <Lachy>
we already know roughly what authors do

[05:25:00.0000] <hsivonen>
you could ask people why in the hope that one of them does not lie

[05:25:01.0000] <hsivonen>
then you mitigate the risk of you not inventing all possible reasons

[05:26:00.0000] <hsivonen>
but once you have an array of possible reasons, you need to test if the actual behavior matches

[05:26:01.0000] <Hixie>
<iframe src="data:text/html,<!DOCTYPE HTML> <p>Hello <p title=&quot;cruel&quot; <p>World"></iframe> 

[05:27:00.0000] <Hixie>
that wouldn't be too bad, you'd only have to escape &, ", and maybe '.

[05:27:01.0000] <BenMillard>
hsivovnen, that makes sense

[05:27:02.0000] <Hixie>
/me is looking for ways to use iframe to embed comments without an http request and without it being ugly and unmaintainable (e.g. base64)

[05:28:00.0000] <Hixie>
othermaciej: dude, what's up with your network? i'd go batty if my ssh connections kept dropping

[05:28:01.0000] <BenMillard>
lachy & hsivonen, it's a question of whether the sponsorship I hope to get will cover that kind of work. hopefully it will, if it's useful to HTML5 accessibility

[05:30:00.0000] <Lachy>
maybe, instead of just asking "Why do you insert alt text?", which is very subject to lying, what if we instead presented a page to web developers and asked them to fix it, without actually telling them we're looking specifically at alt text.

[05:31:00.0000] <BenMillard>
Lachy, I don't understand what you and hsivonen mean by authors lying in this context

[05:32:00.0000] <Lachy>
We include a whole bunch of different errors, including several missing alt, then they have to fix the errors in the best way they can..

[05:32:01.0000] <Hixie>
Lachy: what are we trying to learn?

[05:32:02.0000] <Lachy>
we're trying to learn what alt text use and *why* they use it.

[05:33:00.0000] <Hixie>
ah, interesting

[05:33:01.0000] <Hixie>
i predict most use it because they learnt that they were supposed to

[05:33:02.0000] <Hixie>
and that they forget to do so half the time :-)

[05:33:03.0000] <Hixie>
and have no idea how to do a good job

[05:33:04.0000] <Lachy>
so if validation is all they're trying to achieve, then we would expect authors to simply insert alt="" or alt="bogus". But if they actually care about accessibility, we would get real alt text

[05:34:00.0000] <BenMillard>
Hixie & Lachy, we can make educated guesses easily enough

[05:34:01.0000] <BenMillard>
part of my commercial work is retrofitting accessibility and training authors; usually they don't know alt even exists

[05:34:02.0000] <Hixie>
Lachy: anecdotally, it seems that they try (and mostly fail) to use it for its intended purpose

[05:35:00.0000] <Hixie>
BenMillard: indeed

[05:36:00.0000] <BenMillard>
I think Lachy is suggesting a more scientific (with a small "s") way to figure out what reasons are common and uncommon, along with typical levels of alt text quality

[05:36:01.0000] <Lachy>
there's still a few problems with my ideas. How do we set up appropriate controls and elimiate as many other issues as possile

[05:36:02.0000] <Lachy>
*possible

[05:36:03.0000] <Lachy>
BenMillard, I'm trying to make it as scientific as possible. Obviously, it needs improvement

[05:38:00.0000] <BenMillard>
Lachy, selection bias is an issue. anything that requires voluntary work (like correcting a sample page) will get bloggers, students, a few standardistas and maybe some HTML4all type peeps but no mainstream authors

[05:39:00.0000] <annevk>
hmm, copying over what HTML5 currently says does not seem like a good solution

[05:39:01.0000] <annevk>
and there's nothing else

[05:40:00.0000] <Lachy>
BenMillard, the sample only needs to consist of people who actually care about validation. People who don't care about validation are irrelevant, since they won't be affected by what the validator says.

[05:40:01.0000] <hsivonen>
Lachy: unless you are looking for spillover effects

[05:40:02.0000] <BenMillard>
Lachy, that doesn't quite match my experience.

[05:40:03.0000] <Lachy>
but there's still selection bias problems if we inadvertenly get a lot of people who care a lot about accessibility

[05:41:00.0000] <zcorpan>
couldn't it be possible to search in forums for people who ask about the alt attribute in the context of validation? although that only catches authors who validate *and ask*, not authors who validate and "fix" without asking

[05:41:01.0000] <Lachy>
so we would need some way of measuring the experience of each subject

[05:41:02.0000] <zcorpan>
but could still be interesting to look into

[05:41:03.0000] <BenMillard>
Lachy, I find websites (particularly in the public section) who "need" to comply with standards, hence validate, but whose developers don't really care about validation or accessibility

[05:41:04.0000] <Lachy>
cool, then they might be the kind of people we should target

[05:42:00.0000] <BenMillard>
zcorpan, I've considered collecting tutorials, forum topics, blog entries, national standards and suchlike to analyse where they contradict and where they line up

[05:45:00.0000] <Hixie>
this is one case where it would be important to define the predictions and how to determine whether they are right or not before collecting any data, i think

[05:45:01.0000] <zcorpan>
BenMillard: ok, cool

[05:45:02.0000] <Lachy>
Hixie, indeed

[05:46:00.0000] <Hixie>
ok bed time

[05:46:01.0000] <Hixie>
nn

[05:48:00.0000] <hsivonen>
nn 

[05:49:00.0000] <annevk>
Hixie, the loadstart thing was called out some time ago

[05:49:01.0000] <annevk>
Hixie, and you said back then it didn't match the semantics of HTML5 video that well so you used a different name intentionally

[05:54:00.0000] <annevk>
/me wonders if the other Philip TAYLOR realizes he's alerting a string

[05:54:01.0000] <annevk>
(rather than testing any of the assertions made)

[06:01:00.0000] <annevk>
(i was wrong about Hixe saying something about loadstart above)

[06:03:00.0000] <Lachy>
a more accurate test case: javascript:<!--alert("FAIL");%0Aalert("PASS");

[06:04:00.0000] <zcorpan>
annevk: btw, some of the tests on tc.labs rely on reparsing

[06:07:00.0000] <Lachy>
HTML4.01 states "The JavaScript engine allows the string "<!--" to occur at the start of a SCRIPT element, and ignores further characters until the end of the line. JavaScript interprets "//" as starting a comment extending to the end of the current line. This is needed to hide the string "-->" from the JavaScript parser."

[06:07:01.0000] <Lachy>
but it doesn't make any normative reference to any spec for JS or ECMAScript

[06:08:00.0000] <Lachy>
so that statement was probably just based on defacto behaviour back then too

[06:09:00.0000] <annevk>
zcorpan, ah, I guess I forgot about that when writing the tests, feel free to fix

[06:09:01.0000] <othermaciej>
the ES4 folks were reluctant to add an actual specification for <!-- comments

[06:09:02.0000] <othermaciej>
or at least Brendan was

[06:09:03.0000] <Lachy>
oh. why?

[06:10:00.0000] <Lachy>
is it supported in any JavaScript engines that are used outside of HTML?

[06:10:01.0000] <annevk>
I think it's because ES has use cases beyond the Web and therefore Web influences should be limited or something

[06:10:02.0000] <Lachy>
what about in ActionScript (which I think is based on ECMAScript)

[06:10:03.0000] <annevk>
I'm not entirely sure I like that approach. Fortunately CSS does define handling of <!-- and -->

[06:12:00.0000] <Lachy>
if it's not specced in ECMAScript and only applies to HTML+JS implementations, then maybe HTML5 should require support for it somehow

[06:16:00.0000] <annevk>
I think it is supported in the ES runtime as it works in external files too

[06:17:00.0000] <Lachy>
annevk, which ES runtime implementation are you referring to?

[06:19:00.0000] <hsivonen>
I'm thinking I may not want to be specwise correct on the handling of xmlns attributes

[06:22:00.0000] <Lachy>
hsivonen, in what way do you want to handle them differently?

[06:22:01.0000] <Lachy>
do you want to flag them with warnings?

[06:23:00.0000] <hsivonen>
Lachy: I think I want to drop them on the floor and synthetize ns mappings upon ns context change

[06:23:01.0000] <hsivonen>
which is what I do with <html> now

[06:34:00.0000] <Lachy>
how does that differ from the spec? I thought xmlns was ignored in text/html for the purposes of determining an element's namespace?

[06:34:01.0000] <Philip`>
Opera 9.2 doesn't treat <!-- the same as //

[06:34:02.0000] <Philip`>
(but 9.5 does seem to)

[06:34:03.0000] <Lachy>
although, I'm not really familiar with the mathml parsing section

[06:41:00.0000] <hsivonen>
Lachy: the spec says you put the xmlns attributes into the DOM instead of throwing them away and generating new clean ones

[06:52:00.0000] <annevk>
replied to the TAG crap

[06:56:00.0000] <annevk>
http://www.w3.org/mid/op.ua64w7nk64w2qv⊙aooc

[06:59:00.0000] <hsivonen>
annevk: thank you!

[07:04:00.0000] <zcorpan>
annevk: indeed, thanks

[07:05:00.0000] <zcorpan>
Philip`: how does 9.2 treat <!-- differently from // ?

[07:07:00.0000] <Philip`>
zcorpan: Hmm, I was mistaken - I forgot that <!-- resulted in the </script> being escaped and not ending the script element, which in Opera 9.2 causes the script to not run

[07:08:00.0000] <MikeSmith>
annevk: thanks for taking time to write the TAG reply 

[07:08:01.0000] <zcorpan>
Philip`: ok

[07:10:00.0000] <Philip`>
zcorpan: Opera 9.2/9.5 is a bit funny on http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cscript%3E%0D%0A%3C!--%20foo%0D%0A%3C!--%20bar%0D%0A%2F%2F%20--%3E%0D%0A%3C%2Fscript%3E%0D%0A

[07:10:01.0000] <Philip`>
since the 'foo' line disappears from the DOM view

[07:13:00.0000] <zcorpan>
Philip`: oh wow, that's a bug

[07:13:01.0000] <zcorpan>
Philip`: thanks

[07:13:02.0000] <Philip`>
zcorpan: Crazy browser :-)

[07:16:00.0000] <hsivonen>
I think I've now implemented MathML tree construction...

[07:16:01.0000] <hsivonen>
perhaps I should sync test cases now...

[07:18:00.0000] <zcorpan>
Philip`: most browserse are :)

[07:20:00.0000] <MikeSmith>
hsivonen: thanks also (for TAG reply)

[07:21:00.0000] <hsivonen>
MikeSmith: you're welcome

[07:22:00.0000] <zcorpan>
i wonder if i should reply saying that the analysis doesn't bring up anything that wasn't considered 6 months ago, afaict

[07:22:01.0000] <zcorpan>
it claims to have new discoveries, but i don't see them

[07:22:02.0000] <zcorpan>
(other than a firefox bug that is unrelated to the colon)

[07:23:00.0000] <zcorpan>
(http://lists.w3.org/Archives/Public/www-tag/2008Apr/0242.html )

[07:24:00.0000] <MikeSmith>
zcorpan: yeah, it would actually be helpful to point that out. lest others be left with the impression that the original decision that led to the implementations was not very carefully considered

[07:26:00.0000] <MikeSmith>
btw, for those who care, there's another chat with the Internet Explorer team at 10am US/Pacific today

[07:27:00.0000] <MikeSmith>
Chris Wilson said he will be on that

[07:27:01.0000] <MikeSmith>
http://blogs.msdn.com/ie/archive/2008/05/13/may-chat-with-the-ie-team-on-thursday.aspx

[07:28:00.0000] <Lachy>
MikeSmith, what timezone is US Pacific?

[07:28:01.0000] <MikeSmith>
PST

[07:28:02.0000] <MikeSmith>
GMT -5 I think

[07:28:03.0000] <Lachy>
nevermind, the blog gives the UTC time as well

[07:28:04.0000] <hsivonen>
/me thinks it's PDT this time of year

[07:29:00.0000] <Lachy>
it says 17:00UTC, so that makes it -07:00

[07:29:01.0000] <MikeSmith>
GMT -8 actually

[07:29:02.0000] <Lachy>
must be DST

[07:29:03.0000] <MikeSmith>
yeah

[07:29:04.0000] <MikeSmith>
you guys are right

[07:29:05.0000] <Lachy>
ah, that's right. Pacific is on the west coast in the USA.

[07:30:00.0000] <Lachy>
I'm so used to the pacific being on the east where I'm from :-)

[07:30:01.0000] <BenMillard>
shortest distance is probably going North from here (UK)

[07:30:02.0000] <MikeSmith>
anyway, the MS online chat app they use for those chats blocks Opera and Safari and does not seem to work correctly in Opera or Safari even when you do UA header spoofing

[07:31:00.0000] <Lachy>
ok. I can use Firefox for it

[07:31:01.0000] <MikeSmith>
but it doesn't block Firefox or Minefiel

[07:31:02.0000] <MikeSmith>
Minefield

[07:31:03.0000] <MikeSmith>
Lachy: yeah, it seems to work as expected in FF3 and Minefield

[07:32:00.0000] <Lachy>
how long do they normally go for?

[07:32:01.0000] <Philip`>
/me finally manages to find the IE8 Tech Beta Survey

[07:32:02.0000] <Philip`>
They do quite a good job of hiding these things

[07:32:03.0000] <Lachy>
I have my podcast interview at 20:00 my time (UTC+2), and I need to do some revision beforehand.

[07:32:04.0000] <Lachy>
so I may have to skip it

[07:33:00.0000] <Philip`>
Hmm, Microsoft is nicer than Google since their survey admits that Opera exists

[07:34:00.0000] <MikeSmith>
Lachy: I think they last for one hour

[07:35:00.0000] <aaronlev>
othermaciej: do you know if your developers looking at ARIA have taken a look at the ARIA implementor's guide we're working on?

[07:36:00.0000] <Philip`>
http://www.microsoft.com/windowsxp/expertzone/chats/chatroom.aspx?ctl=hlp&lang=en-US says it ought to work in Safari 1.2

[08:24:00.0000] <Dashiva>
ow

[08:24:01.0000] <Dashiva>
Lisa called you Anna, annevk :)

[08:26:00.0000] <Philip`>
Dashive: That's an easy mistake to make

[08:35:00.0000] <annevk>
/me wonders what setAttribute() bug Opera has

[08:37:00.0000] <Dashiva>
That table he made seems very cavalier about dismissing consistency with html

[08:37:01.0000] <Philip`>
annevk: http://www.w3.org/XML/2008/04/ARIA-Testing/ - "contrary to e.g. the definition of GetAttributeNode in the DOM specification, it matches the argument to GetAttribute against the local name, rather than the nodeName, of the attribute nodes it checks."

[08:37:02.0000] <annevk>
it seems PFWG members also have some weird notion of what they're working on "Losing the  namespaces architectures seems to brake a lot of the original aim of what we set out to do."

[08:38:00.0000] <annevk>
it seems to me they set out to create a low-level accessibility API

[08:40:00.0000] <annevk>
Philip`, ah, that's probably a bug, though DOM Level 3 Core clearly states that using getAttribute() does not give predictable results in namespace aware environments

[08:42:00.0000] <annevk>
Dashiva, I replied to that e-mail on www-archive

[08:42:01.0000] <annevk>
Given Henry's latest I've the feeling I'm wasting my time though

[08:43:00.0000] <annevk>
He clearly lives in another world where namespaces are so important that pragmatism can simply be ignored and complex solutions are preferable over simple ones.

[09:12:00.0000] <gsnedders>
/me guesses he has broken his local copy of html5lib seeming nothing passes

[09:13:00.0000] <annevk>
hsivonen might have made some changes to the tests from what I've read

[09:14:00.0000] <virtuelv>
what happens if <script async> does a document.write?

[09:17:00.0000] <annevk>
same as injecting some script into the document with document.write does now

[09:19:00.0000] <virtuelv>
annevk: that's not at all clear

[09:19:01.0000] <MikeSmith>
/me happy to discover that Google Doctype is directly accessible through http://code.google.com/p/doctype/wiki/Welcome (as alternative to use DocReader interface)

[09:20:00.0000] <virtuelv>
annevk: my point being, given that I have a DOM that is "script async, foo1, foo2, foo3"

[09:20:01.0000] <virtuelv>
and assume the script does a document.write, and the document has parsed foo2 but not foo3

[09:20:02.0000] <virtuelv>
does the write insert the content before foo1 or after foo2?

[09:25:00.0000] <MikeSmith>
Google Code wikis don't provide any way to view change history for a particular wiki page?

[09:25:01.0000] <annevk>
ok, I'm going through the 460 checkins now and summarizing them

[09:26:00.0000] <soypunk>
MikeSmith: I think the only way you can view the history of a wiki page is through SVN

[09:27:00.0000] <soypunk>
example: http://code.google.com/p/doctype/source/browse/wiki/ADisabledAttribute.wiki

[09:27:01.0000] <soypunk>
(well that's their web interface to SVN anyway...)

[09:28:00.0000] <soypunk>
(so I guess you have two routes. web or actually looking at SVN diffs.)

[09:29:00.0000] <soypunk>
/me grrs

[09:30:00.0000] <soypunk>
seems to be in Colloquy where it doesn't update the IRC nickname preferences. :/ 

[09:30:01.0000] <soypunk>
seems to be a bug.

[09:35:00.0000] <gsnedders>
I have something that mostly works to replace charsUntil, jgraham__ 

[10:04:00.0000] <gsnedders>
jgraham__: http://code.google.com/p/html5lib/issues/detail?id=69 — that's a "someone else can do this" issue :P

[10:15:00.0000] <annevk>
anyone know how to add a contact to your skype list?

[10:15:01.0000] <Lachy>
annevk, there should be a + button at the bottom of the window

[10:16:00.0000] <Lachy>
though it might be different on linux from what it is on mac

[10:16:01.0000] <annevk>
ah, indeed

[10:16:02.0000] <annevk>
that dialog looked so complex, i thought it was for something else :)

[10:18:00.0000] <MikeSmith>
annevk: which 460 changes you talking about?

[10:19:00.0000] <annevk>
http://html5.org/tools/web-apps-tracker?limit=460

[10:19:01.0000] <MikeSmith>
annevk: might want to take a look at this too:

[10:19:02.0000] <annevk>
http://www.w3.org/html/wg/html5/diff/#changelog

[10:19:03.0000] <MikeSmith>
http://dev.w3.org/html5/pubnotes/FPWDdiff-colored.html

[10:22:00.0000] <annevk>
interesting stuff

[10:27:00.0000] <virtuelv>
baby lolcats, document.write has interop issues

[10:34:00.0000] <virtuelv>
This is a mess I wish I hadn't started testing

[10:36:00.0000] <Philip`>
virtuelv: The world will love you for it

[10:37:00.0000] <virtuelv>
Philip`: let's just say that in a test set of three browsers, no two are entirely consistent with each other

[10:37:01.0000] <virtuelv>
I just wish we could forbid all use of document.write outright

[10:37:02.0000] <Philip`>
virtuelv: If you only got three different results, that's pretty good going

[10:37:03.0000] <virtuelv>
Banner advertising is destroying the web

[10:37:04.0000] <virtuelv>
Philip`: in some cases I do, yes

[10:38:00.0000] <Philip`>
It's worse when browsers are not even consistent with themselves

[10:44:00.0000] <zcorpan>
annevk: s/formet/format/

[10:44:01.0000] <zcorpan>
annevk: s/dataSet/dataset/

[10:44:02.0000] <zcorpan>
btw, shouldn't pixelratio be on <video> as well as <source>?

[10:45:00.0000] <zcorpan>
and shouldn't <audio><source pixelratio> be disallowed?

[10:46:00.0000] <zcorpan>
Hixie: ^

[10:56:00.0000] <virtuelv>
Philip`: it sucks even more than you can imagine

[11:02:00.0000] <annevk>
zcorpan, fixed those offline

[11:11:00.0000] <virtuelv>
there is enough wtf's in this thing to create an acid test all on its own

[11:34:00.0000] <zcorpan_>
Hixie: vLinkColor should be vlinkColor and likewise for aLinkColor

[11:52:00.0000] <Philip`>
/me sees "Q: [120]  How are you doing with ACID3 ? A: Not really a relevant test for us right now. We're very focused on improving our standards compliance in CSS, HTML and DOM; the Acid3 test chases browser bugs.

[11:52:01.0000] <Philip`>
Our number has improved from IE7, but we're not really focused on it. If the author gets around to writing a guide to it, we might be able to break it apart more quickly."

[11:53:00.0000] <gsnedders>
Philip`: Where's that?

[11:54:00.0000] <Philip`>
gsnedders: The IE8 Experts chat thing

[11:54:01.0000] <gsnedders>
Philip`: Ah. that thing that claims to work in Saf1.2 and later but not your current browser when using Saf3.1.1

[11:54:02.0000] <hober>
link?

[11:55:00.0000] <Philip`>
hober: http://www.microsoft.com/windowsxp/expertzone/chats/default.mspx

[11:55:01.0000] <Philip`>
but there's only a minute left

[11:55:02.0000] <Philip`>
gsnedders: You should get yourself a proper browser, like Minefield

[11:55:03.0000] <gsnedders>
Philip`: It only lists Saf1.2 or later as being supported on OS X :P

[11:56:00.0000] <gsnedders>
Philip`: But Minefield works (Fx 2.0 doesn't)

[11:57:00.0000] <zcorpan_>
acid3 doesn't need a guide for breaking it apart quickly

[11:59:00.0000] <gsnedders>
Philip`: Not that I can find anywhere to report a bug

[14:09:00.0000] <Hixie>
not a single reply to my e-mail to the tag

[14:09:01.0000] <Hixie>
/me is sad

[14:12:00.0000] <gsnedders>
/me gives Hixie a hug in compensation

[14:12:01.0000] <gsnedders>
(not that I expect Hixie wants a hug, but hey)

[14:14:00.0000] <Hixie>
well they could at least tell me i'm being rude or something

[14:15:00.0000] <Philip`>
Hixie: The email you sent yesterday? I would have thought they'd at least need a teleconference to set an action to respond, and another to then agree on the response

[14:15:01.0000] <Hixie>
aah

[14:15:02.0000] <Hixie>
good point

[14:15:03.0000] <Hixie>
i keep forgetting that people put up artificial barriers to progress

[14:30:00.0000] <gsnedders>
jgraham__: OK, you win. //comment()[normalize-space(.) = 'toc'] takes 0.021s

[14:30:01.0000] <gsnedders>
And I don't think I'll be using anything any more complex

[14:52:00.0000] <hsivonen>
my display broke... now do I buy from Apple or do I shop for alternatives...

[14:53:00.0000] <jgraham__>
gsnedders: Do you have any idea if the regexp thing actually is faster?

[14:54:00.0000] <gsnedders>
jgraham__: I dunno. Due to the insane number of error it causes, it may well be just that which makes it slower currently

[14:54:01.0000] <jgraham__>
hsivonen: For a desktop machine, I assume?

[14:55:00.0000] <jgraham__>
I read somewhere that apple displays have poor colour saturation

[14:56:00.0000] <jgraham__>
OTOH, buying displays in general is hard because the one piece of information that gives you some idea of the panel characteristics is the one that no one reveals

[14:57:00.0000] <hsivonen>
jgraham__: laptop actually but used like a desktop because apple doesn't have a desktop without integrated display between mini and pro

[14:57:01.0000] <hsivonen>
but yeah, I'm buying a desktop screen

[14:57:02.0000] <jgraham__>
That makes sense.

[14:58:00.0000] <jgraham__>
How much do you care about quality? If you care about size more than colour reproduction I expect non-Apple is the way forward. Otherwise it's a bit harder

[15:03:00.0000] <gsnedders>
jgraham__: Apple's actual desktop monitors are meant to be very good

[15:03:01.0000] <gsnedders>
But expensive.

[15:04:00.0000] <jgraham__>
gsnedders: I agree that they're good (and expensive). I did read somewhere that they have relatively poor colour saturation but that might not actually be true

[15:04:01.0000] <gsnedders>
I'd go for Dell's UltraSharp range — they use (in the 20", 24", and 30" cases) the same LCD panels as the Apple monitors, and tend to be otherwise just as good if not better, as well as being far cheaper

[15:04:02.0000] <gsnedders>
jgraham__: I've never seen that in any review, nor has it been my experience in the one I had :P

[15:05:00.0000] <gsnedders>
jgraham__: What they've used in the latest iMacs suffered from that, though, IIRC

[15:05:01.0000] <hsivonen>
I like quality but I don't need the best possible panel. at some point price and the looks of the case matter more

[15:05:02.0000] <Hixie>
/me got an apple cinema display for his mini

[15:05:03.0000] <Hixie>
and i'm very happy with it

[15:05:04.0000] <jgraham__>
Well I might be confused

[15:05:05.0000] <hsivonen>
the easy thing with apple is that you only need to pick a size

[15:05:06.0000] <Hixie>
sadly the mini can't drive the biggest cinema displays, but oh well

[15:05:07.0000] <gsnedders>
hsivonen: Well, the Apple one certainly looks better than the Dell one, though the Dell one is far from ugly.

[15:06:00.0000] <jgraham__>
(I ended up buying a HP LP2065 which isn't as big as I would like but has nice colours and is slightly affordable)

[15:06:01.0000] <gsnedders>
(It looks nothing special for a monitor, but it isn't easy to find huge flaws with though)

[15:06:02.0000] <hsivonen>
thanks. I'll look up Apple, Dell and HP

[15:06:03.0000] <gsnedders>
Hixie: Silly Apple not putting dual-link DVI on it!

[15:07:00.0000] <gsnedders>
Though I doubt the onboard graphics could cope with the resolution :)

[15:07:01.0000] <Hixie>
(we use dell ones at work, they're pretty good, lots of ports)

[15:07:02.0000] <jgraham__>
I /think/ I have a Dell Ultrasharp at work (an oldish 24" widescreen one) but its a little bright and hard to adjust

[15:07:03.0000] <gsnedders>
/me just re-implemented textContent for lxml: return u"".join(Element.xpath("child::text()")) :P

[15:08:00.0000] <jgraham__>
gsnedders: That's a more intelligent way of doing it than I thought of 

[15:08:01.0000] <gsnedders>
jgraham__: XPath is really rather powerful. It saddens me that it doesn't support XPath 2.0, though :P

[15:09:00.0000] <gsnedders>
jgraham__: It's the first way I thought of doing it without manually looping over it (which I expect will be a thousandth of a second quicker)

[15:10:00.0000] <jgraham__>
gsnedders: I just used recursion, which is probably slower as you have to manually check for comment nodes 

[15:15:00.0000] <gsnedders>
There isn't even a function to lowercase a string in XPath 1.0

[15:17:00.0000] <gsnedders>
is @id case sensitive in HTML?

[15:17:01.0000] <hsivonen>
gsnedders: use translate()

[15:17:02.0000] <gsnedders>
hsivonen: Yeah, I know

[15:20:00.0000] <gsnedders>
Hmm, a string containing both ' and " cannot be represented in XPath 1.0

[15:26:00.0000] <Lachy>
My podcast interview went well earliet

[15:27:00.0000] <Lachy>
*earlier

[15:27:01.0000] <Hixie>
cool

[15:27:02.0000] <Hixie>
what did you talk about?

[15:27:03.0000] <Hixie>
is it online yet?

[15:27:04.0000] <Lachy>
A lot of it was based upon the ALA article I wrote before

[15:27:05.0000] <Lachy>
not yet

[15:27:06.0000] <gsnedders>
work around for today: xpath_string = u"concat('" + u"', \"'\", '".join(id.split("'")) + u"')"

[15:28:00.0000] <Lachy>
I talked a lot about the community and empahsised that anyone can get involved with the blog, forum, mailing lists, etc.

[15:28:01.0000] <Lachy>
Then I discussed a few features that are available in browsers today or are being implemented

[15:29:00.0000] <Lachy>
like canvas and cross document messaging

[15:32:00.0000] <Lachy>
Then I did an interview with annevk about selectors api for standardssuck.org

[15:36:00.0000] <Hixie>
Lachy: cool

[15:36:01.0000] <Hixie>
i'm so glad y'all are doing these talks and interviews and stuff

[15:37:00.0000] <Lachy>
me too

[15:37:01.0000] <Hixie>
i'd never have the time to edit the spec if i had to do it :-)

[15:38:00.0000] <Lachy>
oh, I got some feedback from the podcast guy...

[15:38:01.0000] <Lachy>
He would like to see a page maintained that describes the current state of implementations of HTML5 features

[15:39:00.0000] <Lachy>
it would be good if there were someone here capable of maintaining such a page

[15:39:01.0000] <Hixie>
the annotations in the spec have that info

[15:39:02.0000] <smedero>
There is this: http://wiki.whatwg.org/wiki/Implementations_in_Web_browsers

[15:39:03.0000] <Hixie>
you could probably write a page that uses those dynamically

[15:40:00.0000] <Lachy>
yeah, but the spec isn't such a friendly place to send people looking for implementation info

[15:40:01.0000] <Hixie>
sure but i mean we could make a separate page

[15:40:02.0000] <Hixie>
that presented the same info

[15:40:03.0000] <Hixie>
using the same back-end protocol

[15:40:04.0000] <Philip`>
Would be nice to have a basic test case for each feature so you can easily see which ones your browser supports

[15:40:05.0000] <Hixie>
and it could even support editing the info

[15:40:06.0000] <Lachy>
ok

[15:40:07.0000] <Hixie>
the annotation system supports listing test cases

[15:40:08.0000] <Hixie>
and demos

[15:40:09.0000] <Hixie>
too

[15:41:00.0000] <Lachy>
what would it take to write such a page? Is there some sort of API for it?

[15:43:00.0000] <Philip`>
http://dev.w3.org/html5/spec/toc-status.html / http://dev.w3.org/html5/spec/Makefile uses the API

[15:44:00.0000] <Hixie>
i don't think it's documented, but yes, there's an API

[15:45:00.0000] <Lachy>
cool. Anyone have time to write the script for it?

[15:46:00.0000] <hober>
I might give it a go this weekend, but don't quote me on that.

[15:47:00.0000] <Lachy>
/me makes a note to quote hober on the weekend about it.

[15:47:01.0000] <hober>
damn.


2008-05-16
[17:54:00.0000] <aruner>
Hixie, what's the history of the blocked header list in XHR in the spec.?  How was that list arrived at?  

[17:54:01.0000] <Hixie>
bjoern sent a detailed reply regarding this recently

[17:55:00.0000] <Hixie>
in other news i am especially amused that sunava just delayed another MONTH

[17:55:01.0000] <aruner>
/me kicks his mail client for retrieving messages slowly

[17:56:00.0000] <Dashiva>
Hixie: But otherwise it's going well. She's quoting properly, at least 

[17:56:01.0000] <Hixie>
aruner: http://lists.w3.org/Archives/Public/public-webapi/2008May/0189.html

[17:56:02.0000] <Hixie>
Dashiva: i believe sunava is a he, am i wrong?

[17:57:00.0000] <Hixie>
annevk: yt?

[17:57:01.0000] <Dashiva>
I'm just guessing, so you're probably right

[17:57:02.0000] <aruner>
/me sighs.  Though I *do* think that their general design principles are good (be conservative, err on the side of ultra safe) I think specifics help us understand their notions on threat modeling.

[17:58:00.0000] <Dashiva>
Flickr seems to support you

[17:58:01.0000] <Hixie>
aruner: their design principles resulted in xdr, which has at least three known security bugs, so... :-)

[17:59:00.0000] <aruner>
/me grins

[18:00:00.0000] <aruner>
Dashiva: Sunava Dutta is a he

[18:00:01.0000] <Dashiva>
He and anne should make a club

[18:00:02.0000] <Hixie>
i live in the bay area

[18:01:00.0000] <Hixie>
i don't make any judgements regarding gender

[18:01:01.0000] <aruner>
Hixie: are you putting in an appearance at the F2F in Redmond?  I'm not sure.

[18:01:02.0000] <Hixie>
as it's not uncommon for people to change gender...

[18:01:03.0000] <Hixie>
aruner: i'd like the spec to be done by then.

[18:01:04.0000] <Hixie>
aruner: in which case i won't turn up

[18:01:05.0000] <Dashiva>
I try to use she for females because some get insulted by 'he' as gender-neutral

[18:01:06.0000] <Hixie>
aruner: we'll see nearer the time

[18:01:07.0000] <Philip`>
You can make a pretty good guess at someone's gender based solely on the fact that they're posting to a web specification mailing list

[18:02:00.0000] <Hixie>
sad but true

[18:02:01.0000] <Dashiva>
That's asking for a beatdown, Philip` ;)

[19:21:00.0000] <Hixie>
so i wonder how to define sandboxes

[19:22:00.0000] <Hixie>
such that if you have document A with iframe containing document B with iframe containing document C

[19:22:01.0000] <Hixie>
that:

[19:23:00.0000] <Hixie>
1. if A, B, and C are from the same origin, B can't play with C

[19:23:01.0000] <Hixie>
hm wait

[19:23:02.0000] <Hixie>
i was going to say that if B and C have hte same origin but aren't the same as A they should be able to talk

[19:23:03.0000] <Hixie>
but no, that would still allow problems

[19:23:04.0000] <Hixie>
so we really do want unique origins...

[19:23:05.0000] <Hixie>
hmm

[19:23:06.0000] <Hixie>
that makes it easier

[19:24:00.0000] <Hixie>
HMM.

[19:24:01.0000] <Hixie>
/me ponders

[19:26:00.0000] <othermaciej>
so I assume you are thinking sandboxes that still have script enabled?

[19:26:01.0000] <othermaciej>
(since for noscript iframes this would not matter)

[19:27:00.0000] <othermaciej>
maybe I should step back and ask what use cases you are targetting

[19:27:01.0000] <Hixie>
actually same-origin matters even without scripts

[19:27:02.0000] <Hixie>
a bunch of things use the origin concept

[19:27:03.0000] <Hixie>
but yes, i am considering scripts too

[19:27:04.0000] <othermaciej>
like frame targetting?

[19:27:05.0000] <Hixie>
yeah, that sometimes uses origin

[19:27:06.0000] <othermaciej>
I think the desired restrictions depend on the use cases

[19:27:07.0000] <othermaciej>
ones I can imagine:

[19:28:00.0000] <othermaciej>
- blog comments

[19:28:01.0000] <othermaciej>
- user-authored content in a profile on a social network site

[19:28:02.0000] <othermaciej>
- "widgets" embedded from off-site

[19:28:03.0000] <othermaciej>
- "widgets" hosted by the provider of the page containing them

[19:28:04.0000] <othermaciej>
- ads

[19:29:00.0000] <Hixie>
comments and other user-contributed content, gadgets (same site and other site; including ads), and (not sandboxing but still <iframe> related) client-side includes are the main ones

[19:29:01.0000] <othermaciej>
ads are complicated because at least some ad networks consider it important to support frameless flash ads that overlay the content

[19:29:02.0000] <othermaciej>
but anything that can overlay the content is not properly sandboxed by definition

[19:30:00.0000] <Hixie>
there are indeed various scenarios with varying levels of sandboxing

[19:30:01.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/sandbox has my current strawman proposal (cut and pasted from the start of the e-mail i'm writing; note that i've used past tense because it'll be past tense once i send the mail, but the spec isn't actually updated yet)

[19:30:02.0000] <othermaciej>
for comments you would like them to be able to size naturally and seem to be inline as far as the end-user is concerned, but they should not be able to css position outside their area

[19:32:00.0000] <Hixie>
yeah the comments would be <iframe sandbox="unrestricted-origin" seamless src="data:text/html,(small html document)"></iframe>

[19:33:00.0000] <othermaciej>
I'm thinking about whether inheriting styles and applying outer stylesheets may be a security risk

[19:33:01.0000] <Hixie>
in certain cases it is

[19:33:02.0000] <Hixie>
in others not so much

[19:33:03.0000] <Hixie>
i can't see how it could be in a script-limited blog comment case

[19:34:00.0000] <othermaciej>
I'm thinking for the comment use case and the user-generated content on hosted profile page use case

[19:34:01.0000] <Hixie>
oh i have to disble forms too while i'm at it

[19:34:02.0000] <othermaciej>
and also plugins probably

[19:34:03.0000] <Hixie>
yeah

[19:34:04.0000] <othermaciej>
(including Java)

[19:35:00.0000] <Hixie>
updated sandbox file

[19:35:01.0000] <othermaciej>
I am nervous about the performance impact of using an iframe per comment with hundreds of comments

[19:35:02.0000] <Hixie>
yeah me too

[19:35:03.0000] <Hixie>
i'm not sure what else to do though

[19:36:00.0000] <othermaciej>
navigating the parent should be restricted (dunno if that falls out of other things), including back/forward

[19:36:01.0000] <othermaciej>
window sizing and positioning should be restricted

[19:37:00.0000] <Hixie>
any browsing context navigation outside the sandbox is blocked

[19:37:01.0000] <othermaciej>
does embedding further subframes in the content need to be restricted?

[19:37:02.0000] <Hixie>
though that does mean that links in comments won't work

[19:38:00.0000] <othermaciej>
or do they just inherit the sandbox restrictions already in effect?

[19:38:01.0000] <Hixie>
i don't see why it should be, but they certainly are also sandboxed

[19:38:02.0000] <Hixie>
and they're not in the same sandbox, either

[19:39:00.0000] <othermaciej>
links in blog comments not working is kind of a bummer

[19:39:01.0000] <othermaciej>
perhaps there should be a way to turn that on, since it can't be used as an annoyance in the noscript case

[19:39:02.0000] <Hixie>
yeah

[19:40:00.0000] <Hixie>
comments would now be <iframe sandbox="unrestricted-origin force-links-to-parent" seamless src="data:text/html,(small html document)"></iframe>

[19:40:01.0000] <Hixie>
it's getting long

[19:41:00.0000] <othermaciej>
the keywords are a bit verbose

[19:41:01.0000] <Hixie>
it's a security thing, i'm worried about making them too simple

[19:41:02.0000] <Hixie>
i'm especially worried about making them sound like they _increase_ security

[19:41:03.0000] <Hixie>
when all the keywords decrease it somewhow

[19:41:04.0000] <othermaciej>
perhaps there should be two addtributes

[19:42:00.0000] <othermaciej>
sandbox="" or restrict="" that turns everything off, and an allow="" attribute to turn things back on selectively

[19:42:01.0000] <othermaciej>
or sandbox-exceptions=""

[19:43:00.0000] <Hixie>
well "force-links-to-parent" isn't really an exception or a restriction, it just changes the rules to send all user-activated links to the parent browsing context

[19:43:01.0000] <othermaciej>
you could achieve that with <base target="parent"> and allowing targetted links to parent only

[19:43:02.0000] <othermaciej>
a bit more work but makes it strictly an exception, not a behavior change

[19:43:03.0000] <Hixie>
(<iframe seamless sandbox allow="..." src="data:text/html,(small html document)"></iframe> makes sense, but i'm not sure what the keywords would be)

[19:44:00.0000] <othermaciej>
though if it is not a restriction or an exception perhaps it should be a separate attribute like seamless

[19:44:01.0000] <Hixie>
well the idea is that you can allow whatever random content in the sandbox as you want

[19:44:02.0000] <Hixie>
so i'm not sure i want to be poking <base> into it

[19:44:03.0000] <Hixie>
ooo

[19:44:04.0000] <Hixie>
wait!

[19:44:05.0000] <Hixie>
it can just be part of seamless' behaviour!

[19:44:06.0000] <othermaciej>
for the comment use case presumably the data: URL is generated by software and the user's comment does not have <html> or <body>?

[19:44:07.0000] <othermaciej>
oh interesting

[19:45:00.0000] <Hixie>
comments would now be <iframe sandbox="unrestricted-origin" seamless src="data:text/html,(small html document)"></iframe> again

[19:45:01.0000] <othermaciej>
is it ever useful to allow gadgets to navigate the parent or open a window?

[19:45:02.0000] <Hixie>
probably, but i want to start with as small a set of allowed behaviours as possible

[19:51:00.0000] <othermaciej>
I guess for gadgets, anything forbidden can be selectively exposed via postMessage

[19:51:01.0000] <Hixie>
eah

[19:51:02.0000] <Hixie>
y

[19:53:00.0000] <Hixie>
maybe seamless should base its origin determination on the "real" origin, not the origin affected by sandbox=""

[19:53:01.0000] <othermaciej>
Hixie: i think the seamless navigation rule should only apply to navigations triggered by the content of the iframe

[19:53:02.0000] <Hixie>
as opposed to what?

[19:53:03.0000] <othermaciej>
Hixie: if the parent doc navigates a seamless iframe it should not navigate itself

[19:53:04.0000] <Hixie>
hm, that makes things more complex

[19:53:05.0000] <Hixie>
why would you use target="" on a seamless iframe?

[19:53:06.0000] <Hixie>
why not just change .data?

[19:53:07.0000] <Hixie>
or .src

[19:54:00.0000] <Hixie>
or whatever

[19:54:01.0000] <othermaciej>
changing .src is a navigation

[19:54:02.0000] <Hixie>
not per spec

[19:54:03.0000] <Hixie>
oh actually it is

[19:54:04.0000] <Hixie>
nevermind

[19:54:05.0000] <Hixie>
i'm thinking of <object>

[19:54:06.0000] <othermaciej>
well put it this way, changing iframe.src, changing iframe.contentWindow.location, and targetted link aimed at iframe all go at the same code path

[19:54:07.0000] <othermaciej>
at least in WebKit

[19:55:00.0000] <Hixie>
yeah

[19:55:01.0000] <Hixie>
spec too

[19:55:02.0000] <Hixie>
ok

[19:55:03.0000] <Hixie>
well

[19:55:04.0000] <Hixie>
i could just do it for <a> links

[19:55:05.0000] <othermaciej>
knowing the source frame of a navigation is not a big deal

[19:55:06.0000] <Hixie>
i guess

[19:55:07.0000] <othermaciej>
we always have to know anyway

[19:55:08.0000] <Hixie>
though what do you call the source if someone from another iframe clicks a link using .click() ?

[19:56:00.0000] <Hixie>
or what if frame A calls .click() on a link in frame B which has a target="" to A, which is seamless?

[19:56:01.0000] <othermaciej>
I would say only cases where the source frame is the seamless frame itself should get retargetted

[19:57:00.0000] <Hixie>
sure but what is the source frame in that example?

[19:58:00.0000] <othermaciej>
I don't know in general what a cross-frame click() call on a link considers the source frame

[19:58:01.0000] <othermaciej>
presumably something has to define which it is in any case

[19:59:00.0000] <othermaciej>
I don't think it matters which is chosen from security POV

[19:59:01.0000] <othermaciej>
since A has to have unrestricted origin in your scenario and thus could just modify B to do something bad

[20:01:00.0000] <Hixie>
well i've never had to define it before now :-)

[20:01:01.0000] <Hixie>
but i agree it doesn't really matter much

[20:01:02.0000] <Hixie>
so long as it is defined

[20:02:00.0000] <Hixie>
i guess i'll just add an "argument" to the "navigate" algorithm and make all the call sites say what the source is

[20:04:00.0000] <Hixie>
/me defines the term "source browsing context"

[20:04:01.0000] <Hixie>
(better terms welcome)

[20:05:00.0000] <Hixie>
hm, this causes <meta http-equiv=refresh> inside a seamless iframe to navigate the parent iframe

[20:06:00.0000] <Hixie>
oh hey, we do need this, to define .referer better

[20:06:01.0000] <Hixie>
:-)

[21:54:00.0000] <MikeSmith>
Hixie: why's the body element included in the "Obsolete elements, attributes, and APIs" section?

[01:20:00.0000] <othermaciej>
hey web expert people

[01:21:00.0000] <othermaciej>
is there any way in Firefox to do the equivalent of either word-wrap: break-word (breaks lines in the middle of long words that would overflow the box) or text-overflow: ellipsis (ellipsizes lines that are too long)?

[01:28:00.0000] <jwalden>
othermaciej: 3 you could use &shy; or the Unicode codepoint, I suspect

[01:28:01.0000] <othermaciej>
jwalden: between every character?

[01:28:02.0000] <jwalden>
wait

[01:28:03.0000] <jwalden>
that's what that does?

[01:28:04.0000] <othermaciej>
someone was asking me how to handle a long word in a fixed-width box gracefully

[01:29:00.0000] <jwalden>
"gracefully"

[01:29:01.0000] <othermaciej>
(url to be specific)

[01:29:02.0000] <othermaciej>
overflow: visible makes it stick out of the box

[01:29:03.0000] <othermaciej>
overflow: hidden silently truncates

[01:29:04.0000] <othermaciej>
overflow: scroll adds a scrollbar

[01:29:05.0000] <jwalden>
last I recall we were breaking URLs, actually; that may have changed

[01:29:06.0000] <othermaciej>
I suggested word-wrap break-word to wrap the word in the middle of the box

[01:29:07.0000] <othermaciej>
or text-overflow: ellipsis to truncate but with an ellipsis at the end

[01:30:00.0000] <othermaciej>
neither of those work in FF though

[01:30:01.0000] <jwalden>
I'd do &shy; at the / boundaries myself

[01:31:00.0000] <jwalden>
I don't remember anything particular that would work

[01:31:01.0000] <othermaciej>
(I just used a long word instead of a URL as my test case though so I could not tell if there was magical URL breaking, but I assume no or they would not have complained of this bug)

[01:32:00.0000] <jwalden>
so data:text/html,%3Cdiv%20style=%22width:%2010em;%20border:%201px%20dotted%20green;%22%3Ehttp://localhost:8888/tests/browser/microformats/test/test_Microformats_hCard.html%3C/div%3E does wrapping for me in 3+ source code, I think 3 does the same

[01:32:01.0000] <roc>
<wbr> or a zero-width unicode space?

[01:33:00.0000] <roc>
there is magical URL breaking in Firefox 3

[01:33:01.0000] <othermaciej>
I was testing with an FF3 beta but possibly an old one

[01:33:02.0000] <roc>
although it's not terribly magical, it works pretty well

[01:34:00.0000] <othermaciej>
I think it it desired for this to work for non-URL overflow too or overflow of a single url component (the column in question is not very wide)

[01:35:00.0000] <othermaciej>
and not sure that putting &shy; or <wbr> or zero-width space between every single character will fly

[01:35:01.0000] <othermaciej>
but I guess I can pass it along

[01:35:02.0000] <roc>
it doesn't have to be every single character, I guess

[01:40:00.0000] <othermaciej>
(both the things I mentioned work in IE as well as Safari fwiw, might be worth adding one or both to Gecko at some point)

[01:40:01.0000] <roc>
yeah, I should implement word-break

[01:40:02.0000] <roc>
text-overflow is totally underspecified

[01:41:00.0000] <roc>
and Webkit, at least, does weird things with it sometimes. Probably IE too although i test with IE less

[01:41:01.0000] <othermaciej>
word-wrap: break-word != word-break

[01:41:02.0000] <roc>
oh right

[01:42:00.0000] <othermaciej>
it is handy though cause it does what you expect in an editable area when your line is too long and is all one word (or at least what I expect)

[01:42:01.0000] <roc>
"Shaping characters are still shaped as if the word were not broken" ... that is weird

[01:42:02.0000] <othermaciej>
yeah, you would want arabic chars to still take medial form

[01:42:03.0000] <othermaciej>
not terminal or initial

[01:42:04.0000] <roc>
I guess that's the easy to implement way

[01:43:00.0000] <othermaciej>
it also seems like the correct way (at least for Arabic)

[01:43:01.0000] <roc>
but not for optional ligatures

[01:43:02.0000] <othermaciej>
presumably you can't break mid-ligature

[01:43:03.0000] <roc>
why not?

[01:43:04.0000] <othermaciej>
because it is one glyph

[01:43:05.0000] <othermaciej>
I mean

[01:43:06.0000] <othermaciej>
you could break between the chars

[01:43:07.0000] <othermaciej>
but the glyph is on one line or the next

[01:44:00.0000] <roc>
you can cut it in half

[01:44:01.0000] <roc>
we have to do that for selection and ligatures that cross span boundaries with different styles

[01:47:00.0000] <othermaciej>
if you can tell the difference between things that are ligature-like and things that are like Arabic I think you would be in your rights to undo the ligature, but I think breaking on one side of the glyph or the other should do, as perfecting the breakpoint when you already broke in the middle of a word seems less than essential

[01:47:01.0000] <roc>
right

[01:47:02.0000] <roc>
that's what I'd do

[01:47:03.0000] <annevk>
Hixie, am now

[01:48:00.0000] <annevk>
/me reads up on the sandbox discussion in the backlog

[02:16:00.0000] <annevk>
MikeSmith, I think it's there to define obsolete parts of the body element

[02:24:00.0000] <MikeSmith>
annevk: OK

[02:28:00.0000] <annevk>
I think it's confusing

[03:01:00.0000] <MikeSmith>
annevk: for the publication-notes doc, I wrote a short context description of that section as "contains descriptions of the obsoleted applet element and obsolete parts of DOM interface for the body element."

[03:19:00.0000] <MikeSmith>
so the table in section 8.6, "Named character references" was removed completely

[03:19:01.0000] <MikeSmith>
http://www.w3.org/html/wg/html5/#named

[03:19:02.0000] <MikeSmith>
wondering if we want to published without that

[03:19:03.0000] <MikeSmith>
if so, a note should probably be added there

[03:20:00.0000] <annevk>
that seems like a bug in the W3C script

[03:20:01.0000] <MikeSmith>
ah

[03:21:00.0000] <MikeSmith>
so still something that'll need to get fixed before we publish

[03:21:01.0000] <othermaciej>
annevk: I think Hixie's sandbox proposal is converging (at least I mostly liked the last thing I saw before I left work)

[03:22:00.0000] <othermaciej>
annevk: I fear the idea of using an iframe per comment will just not fly performance-wise (even with data: URLs)

[03:22:01.0000] <annevk>
yeah, it's such a hack

[03:22:02.0000] <annevk>
i doubt people will use it over better sandboxing on the server using real html5 parsers and tools

[03:23:00.0000] <othermaciej>
but you do need visual sandboxing I think (not let the content to overflow) as well as script, though I guess overflow: hidden does that

[03:23:01.0000] <othermaciej>
(does overflow: hidden hide relative, absolute or fixed position children that stick out?)

[03:23:02.0000] <othermaciej>
annevk: his design seems ok for gadgets/widgets embedded in web pages

[03:23:03.0000] <othermaciej>
or ads

[03:24:00.0000] <othermaciej>
not sure it flies for other use cases

[03:45:00.0000] <hsivonen>
Does Safari 3.1 support closed captiong for video?

[03:51:00.0000] <hsivonen>
othermaciej: Does Safari 3.1 support closed captiong for video?

[03:51:01.0000] <othermaciej>
oddly someone just asked me that

[03:51:02.0000] <othermaciej>
I have not tested

[03:51:03.0000] <hsivonen>
ok. thanks

[03:51:04.0000] <othermaciej>
I do not know how you turn it on in QuickTime

[04:05:00.0000] <annevk>
oh lol

[04:06:00.0000] <annevk>
Robert Burns is joining the ARIA discussion

[04:06:01.0000] <annevk>
guess I'm now more than done with that :)

[04:07:00.0000] <hendry>
/me wonders if there is a resource describing each test in Acid3 or perhaps some sort of source control access. best i've found is http://en.wikipedia.org/wiki/Acid3

[04:08:00.0000] <othermaciej>
oh no, Robert Burns

[04:08:01.0000] <othermaciej>
is there any clusterfuck of a discussion that he can't make worse?

[04:10:00.0000] <hsivonen>
sigh

[04:10:01.0000] <Dashiva>
hendry: You mean to show what spec part they test?

[04:10:02.0000] <annevk>
i wonder if i should reply

[04:11:00.0000] <annevk>
he's totally wrong, but i know that when i point that out he finds some way to turn the argument around

[04:13:00.0000] <Dashiva>
annevk: I suppose the question is if anyone else listens to him, and needs clarification :)

[04:13:01.0000] <annevk>
yeah, deleted the e-mail

[04:13:02.0000] <othermaciej>
" In my view the null namespace should not exist in a namespaced document processed by a namespace aware application. Instead unprefixed attributes should be attached to the same namespace as their parent elements."

[04:14:00.0000] <hsivonen>
he is not wrong about XMLNS having a design bug

[04:14:01.0000] <hsivonen>
it's just a bug that is too late to fix

[04:14:02.0000] <othermaciej>
he seems to think namespaces already work this way

[04:15:00.0000] <hsivonen>
many namespace advocates do

[04:15:01.0000] <othermaciej>
also seems to think adding namespaces (with : syntax) in HTML5 would solve anything with regards to compat issues with the proposal

[04:17:00.0000] <annevk>
the whole point about ARIA is that it doesn't require architectural changes

[04:17:01.0000] <annevk>
so arguing we first need to make architectural changes misses the point

[04:18:00.0000] <annevk>
but pointing that out to Robert Burns doesn't work based on past experience

[04:18:01.0000] <hsivonen>
I have to go buy a new display with a powered USB hub right now.

[04:18:02.0000] <hsivonen>
I need 3 USB devices to get work done and the MacBook can power only 2

[04:20:00.0000] <othermaciej>
however Robert Burns does have the ability to increase confusion of third parties to the discussion if not corrected

[07:57:00.0000] <hsivonen>
(I got an Apple display. thanks for the recommendations)

[08:04:00.0000] <annevk>
http://lists.w3.org/Archives/Public/public-forms/2008May/att-0022/2008-05-14.html#topic3 "Steven Pemberton: There was a huge amount of RDFa."

[08:04:01.0000] <annevk>
(over XTech)

[08:05:00.0000] <hsivonen>
annevk: different room, different amount of RDFa, I guess

[08:06:00.0000] <annevk>
Not sure what's happening to the Forms Task Force

[08:07:00.0000] <annevk>
Also seems they continue with copying HTML features. It seems though that they try to copy from the HTML4 specification rather than from implementations...

[08:10:00.0000] <hendry>
Dashiva: er, OK (sorry missed this IRC window somehow)

[09:01:00.0000] <Philip`>
http://doctype.googlecode.com/svn/trunk/tests/html/elements/plaintext-element-appears-in-dom-test.html - that's not quite right :-(

[09:59:00.0000] <JJ08>
Are there currently any problems with the <video> element in safari?

[10:00:00.0000] <JJ08>
i can get it to show the player but the .mpg wont play. Also the poster image covers the play controls sometimes.

[12:14:00.0000] <virtuelv>
/me curses at Google's newfound love for not making websites work in browsers instead of named product

[12:14:01.0000] <virtuelv>
+s

[12:14:02.0000] <JJ08>
lol. Sounds like your enjoying yourself

[12:15:00.0000] <virtuelv>
not at all

[12:15:01.0000] <virtuelv>
http://www.google.com/translate_t

[12:16:00.0000] <virtuelv>
they have opted for creating a rather ugly, custom widget instead of a standard <select>

[12:16:01.0000] <virtuelv>
continuing down that path, they might just as well create their own runtime, only viewable with the google client, on the google os

[12:16:02.0000] <MikeSmith>
the docreader thing looks like of a like a frankenstein monster that escaped from the lab and now out terrorizing the villagers

[12:17:00.0000] <virtuelv>
running on Google hardware

[12:17:01.0000] <virtuelv>
MikeSmith: yeah, and now they ruined translate as well

[12:17:02.0000] <virtuelv>
http://www.google.com/translate_t

[12:17:03.0000] <JJ08>
lol.

[12:17:04.0000] <JJ08>
i see what you mean now.

[12:17:05.0000] <virtuelv>
what's the use of standards when everything gets replaced with JavaScript that barely works anywhere

[12:17:06.0000] <JJ08>
its not very "standard" at all.

[12:18:00.0000] <JJ08>
oh, see you beat me to it

[12:18:01.0000] <JJ08>
and at my college Javascript is disabled .... so that goes to pot

[12:18:02.0000] <takkaria>
it's that bloody google web toolkit thing

[12:18:03.0000] <virtuelv>
if you turn off js, it sort of works

[12:19:00.0000] <virtuelv>
Sincerely, GWT should die

[12:19:01.0000] <JJ08>
there turning into Apple (even though i just spent over £1000 on one)

[12:19:02.0000] <JJ08>
tightly controlling everything they own

[12:21:00.0000] <JJ08>
googles source code is just one big line. They forgot to use "/n". lol

[12:21:01.0000] <virtuelv>
a single character costs google

[12:22:00.0000] <virtuelv>
removing redundant whitespace is IMHO acceptable for stuff that's production code, and has undergone proper QA

[12:22:01.0000] <JJ08>
i suppose if you put it like that.

[12:23:00.0000] <Lachy>
jgraham__, we're presenting on day 1 http://www.vivabit.com/atmedia2008/london/schedule/

[12:24:00.0000] <Lachy>
assuming the schedule doesn't change again between now and then

[12:24:01.0000] <virtuelv>
<DIV class="goog-inline-block goog-menu-button-caption" unselectable="on">Spanish</DIV>

[12:24:02.0000] <JJ08>
i wanted to go to @media. I had all the cash sorted then i had to move house! Bummer, well pissed off i was

[12:38:00.0000] <Dashiva>
When this is all over, we'll have to make medals for the people who risked mind and sanity in the alt wars

[12:46:00.0000] <smedero>
Dashiva: That or we'll all lose our minds and start implementing Dmitry Turin's HTML 6.

[12:46:01.0000] <Dashiva>
I don't see how that'll be much worse than our current heading

[12:46:02.0000] <Dashiva>
/me mumbles

[12:47:00.0000] <smedero>
I kinda wish he would email the list now with how he intends to support alt on his new picture input type: http://html6.by.ru/site/html60/en/author/inputpic_eng.htm

[12:47:01.0000] <smedero>
oh, i'm just being cruel now.

[12:47:02.0000] <smedero>
:(

[12:48:00.0000] <JJ08>
wtf is HTML 6?

[12:50:00.0000] <virtuelv>
JJ08: Dmitry Turin's private proposal for next-generation HTML

[12:50:01.0000] <virtuelv>
I'm amazed at how much people here actually know about it

[12:50:02.0000] <virtuelv>
http://html6.by.ru/site/html60/en/author/label_eng.htm

[12:51:00.0000] <JJ08>
i am taking this as a joke. This looks idiotic

[12:51:01.0000] <virtuelv>
<style>

[12:51:02.0000] <virtuelv>
  form { next: execute }

[12:51:03.0000] <virtuelv>
</style>

[12:51:04.0000] <virtuelv>
<body>

[12:51:05.0000] <virtuelv>
  <form action="program.exe">

[12:51:06.0000] <smedero>
You can also see his musings on SQL 5: http://sql50.euro.ru/site/sql50/en/author/index_eng.htm

[12:51:07.0000] <virtuelv>
</body>

[12:51:08.0000] <virtuelv>
Value next=execute orders to execute received information in local operational system (mainly value next=execute is intended for those cases, when form's address specifies into local file system).

[12:51:09.0000] <JJ08>
what the hell is this: <a href="./c.htm#:body/h2[@@~"part of header"]+h3[5]"> text </a>

[12:51:10.0000] <virtuelv>
http://html6.by.ru/site/html60/en/author/transfer_eng.htm

[12:51:11.0000] <JJ08>
is this some random guy that has set this uo?

[12:51:12.0000] <JJ08>
*up

[12:52:00.0000] <smedero>
Dmitry pops into various standards related mailing list from time to time and proposes one of his concepts.

[12:52:01.0000] <JJ08>
have any actually been used?

[12:52:02.0000] <smedero>
not that I know of, no.

[12:53:00.0000] <JJ08>
he looks serious about what he is trying to preach.

[12:53:01.0000] <smedero>
to be fair to Dmity, he at least provides examples and occasionally use cases.

[12:53:02.0000] <smedero>
however insane they may be

[12:54:00.0000] <virtuelv>
actually, I have seen a user-js implementation of his XPath-links

[12:57:00.0000] <virtuelv>
and the idea itself, for being able to link into specific locations of a document that doesn't have relevant anchors or id's has merit

[12:59:00.0000] <JJ08>
Maybe he will come up with something beneficial. I'll give him credit because he provides examples and is obviously working on ways to improve. Its more than what im doing.

[12:59:01.0000] <Dashiva>
It's not that his ideas necessarily are bad, but most of them are too different from everything we already have to fit in

[13:00:00.0000] <Dashiva>
He seems like he really likes stream processing, maybe we could get him to work on event-source :)

[13:01:00.0000] <JJ08>
but how can someone be considered credible if he isn't considering 'standards'. His source code is evidence to that.

[13:01:01.0000] <Lachy>
Hixie, yt?

[13:01:02.0000] <Dashiva>
JJ08: Well, our alt advocate friends are carrying that banner high themselves, so why can't Dmitry? :)

[13:03:00.0000] <JJ08>
any links to these 'friends' ? I know your telling the truth, id like to see thats all

[13:04:00.0000] <jgraham__>
Lachy: Yeah, they moved the progrmme around a lot at the same time. I wonder if Steve Faulkner will be around on Day 1

[13:04:01.0000] <jgraham__>
s/at the same time/ at the same time as scheduling us/

[13:05:00.0000] <Dashiva>
JJ08: If you can teach me how to search message contents in thunderbird, I'd probably find it.

[13:05:01.0000] <takkaria>
Dashiva: top right-hand corner search box, click on the magnifying glass, select "entire message"

[13:06:00.0000] <Dashiva>
Oh, swell

[13:06:01.0000] <Dashiva>
... but it seems it gladly hits the server instead of using the local copy.

[13:06:02.0000] <takkaria>
you using IMAP?

[13:06:03.0000] <Dashiva>
Yes

[13:06:04.0000] <JJ08>
lol, i cant :( So what your saying is that these people who are involved in the development of HTML 5 are not even following standards?

[13:06:05.0000] <smedero>
the public-html archive search works fairly well: http://lists.w3.org/Archives/Public/public-html/

[13:07:00.0000] <Dashiva>
JJ08: No, not like that

[13:07:01.0000] <takkaria>
Dashiva: IMAP has a built-in search command, so Tb uses that by default

[13:07:02.0000] <Dashiva>
Just that even advocates sometimes fail to go the extra mile in every single thing

[13:07:03.0000] <JJ08>
sorry. I get you now.

[13:07:04.0000] <Lachy>
jgraham__, yeah, Steve should be around.

[14:12:00.0000] <Lachy>
My podcast for boagworld.com that I recorded yesterday will most likely appear around the 4th of June

[14:42:00.0000] <mcarter>
good afternoon

[14:43:00.0000] <mcarter>
Hixie, I don't suppose you ever wrote that TCPConnection server perl script for your train set that I can test my implementation against?

[16:27:00.0000] <virtuelv>
http://www.regdeveloper.co.uk/2008/05/16/html5_xhtml_vendors_lockin/

[16:30:00.0000] <Lachy>
The article says: "For sophisticated cross-platform applications, there already exist several proprietary solutions (such as Mozilla's XUL, Adobe's Flash, or Microsoft's Silverlight)," the group said.

[16:30:01.0000] <Lachy>
Who in the group said that?

[16:30:02.0000] <takkaria>
that seems like a very confused bit of reporting

[16:30:03.0000] <Hixie>
the spec says that

[16:31:00.0000] <Hixie>
so much for the spec not being a position paper. :-D

[16:31:01.0000] <Hixie>
hahahahaha

[16:32:00.0000] <Hixie>
i love the picture of chris wilson with a caption "Microsoft's Wilson: modular HTML 5"

[16:32:01.0000] <Hixie>
that's hilarious

[16:36:00.0000] <Dashiva>
The picture reminds me of those british judges with wigs

[16:36:01.0000] <Dashiva>
Poor Chris

[16:37:00.0000] <Lachy>
LOL

[16:39:00.0000] <takkaria>
the IBM article that links to I haven't read before

[16:39:01.0000] <takkaria>
has some good lines

[16:39:02.0000] <takkaria>
"The HTML V5 specification was written using good communication with browser vendors ... The team is sceptical about official W3C approval, though. The FAQ doesn't even try to give a serious answer about the expected date of approval."

[16:40:00.0000] <Lachy>
IIRC, that IBM article wasn't very good

[16:40:01.0000] <takkaria>
it's not, no

[16:41:00.0000] <Lachy>
wow. I wrote that FAQ, and I certainly did try to give a very serious answer to that question

[16:42:00.0000] <Hixie>
i thought the answer in the faq was pretty accurate

[16:42:01.0000] <Lachy>
though, saying 15 years generally doesn't go down well with people and it works better if you emphasise that features can be used as soon as browsers implement them, rather than waiting for the finished spec

[16:43:00.0000] <Hixie>
it took us 18 years to get where we are now

[16:43:01.0000] <Lachy>
that's what I did in the podcast yesterday, and the host responded quite well to that answer

[16:43:02.0000] <Hixie>
and html5 is specifying all of that and enough to take us forward

[16:43:03.0000] <Hixie>
and they think 15 years to get two perfect implementations is too long? :-)

[16:43:04.0000] <Hixie>
some people have issues

[16:44:00.0000] <roc>
I've started having terrifying thoughts that I'm going to be working on Web browsers for the rest of my life

[16:45:00.0000] <Lachy>
re the modularisation, they fail to realise that there may be a lot to write specs for, but there's too few editors to do it.

[16:45:01.0000] <Hixie>
roc: why is working on the cutting edge of one of humanity's biggest achievements terrifying? :-)

[16:45:02.0000] <Hixie>
Lachy: who is "they"?

[16:46:00.0000] <roc>
I bet Pharoah said the same thing to the slaves building his pyramids

[16:46:01.0000] <Lachy>
the people who are pushing for splitting the spec up into modules

[16:46:02.0000] <Lachy>
chris wilson, included

[16:46:03.0000] <Hixie>
roc: :_P

[16:47:00.0000] <Hixie>
roc: i heard from a student who attended the talk you gave at stanford btw, remind me to slip you a fiver when we next meet :-)

[16:47:01.0000] <roc>
ooh, which one?

[16:47:02.0000] <Lachy>
oh, at least MikeSmith does: So for now, nothing changes as far as the W3C is concerned. Smith noted that while modularization of specification is a good idea in principle, the consortium needs a commitment from one or more member organizations to take over a part of the specification. “We don’t have that; nobody has stepped up,” he said.

[16:47:03.0000] <Lachy>
http://www.sdtimes.com/content/article.aspx?ArticleID=32067 (2nd last paragraph)

[16:47:04.0000] <Hixie>
roc: you gave more than one?

[16:47:05.0000] <roc>
no, which student?

[16:47:06.0000] <Hixie>
oh

[16:47:07.0000] <Hixie>
andy

[16:47:08.0000] <Hixie>
works on indexing at google

[16:48:00.0000] <Hixie>
andy hochhaus

[16:48:01.0000] <roc>
ok, don't know him


2008-05-17
[18:14:00.0000] <mcarter>
Hixie, whats the rationale behind the framing of each payload in the TCPConnection spec?

[02:57:00.0000] <zcorpan_>
hmm, perhaps the offline-webapps should expand a bit what the difference is between localStorage["status"] = "Idling."; and window["status"] = "Idling.";

[05:20:00.0000] <Dashiva>
/me wonders if anyone has been offended yet

[05:32:00.0000] <takkaria>
Dashiva: what by?

[05:37:00.0000] <Dashiva>
'primary'

[05:37:01.0000] <jgraham__>
By hsivonen suggesting that cameras are mainly designed to take pictures, you mean?

[05:37:02.0000] <Dashiva>
Yes. The nerve!

[05:41:00.0000] <takkaria>
I'm glad someone finally said it, Rob scared me off from replying to the alt thread again

[05:43:00.0000] <Dashiva>
takkaria: It was brought up once before, and it got some pretty nasty replies

[06:44:00.0000] <hsivonen>
sometimes the alt thread makes me wonder if I'm suffering from metacognitive miscalibration

[06:44:01.0000] <hsivonen>
I take the comments above as indication that I might not be. thanks

[06:50:00.0000] <MikeSmith>
I see mention of "fast-path DOM" at http://wiki.mozilla.org/Mozilla_2/Work_List

[06:51:00.0000] <MikeSmith>
wondering just what that might mean

[07:36:00.0000] <Philip`>
MikeSmith: Judging by e.g. http://markmail.org/message/sudwtkjlsrkf47zx it seems to be about optimising the JS-DOM interface by avoiding the XPCOM overhead

[07:37:00.0000] <MikeSmith>
Philip`: cool, thanks for the pointer

[07:44:00.0000] <hsivonen>
I think I've now said all I have to say about alt

[07:45:00.0000] <hsivonen>
now I'm just going to wait and see what if anything WAI is going to tell the HTML WG to do

[07:49:00.0000] <MikeSmith>
hsivonen: hope it will free up more time for you work on validator.nu

[07:49:01.0000] <MikeSmith>
btw, do you have a v.nu TODO list online somewhere?

[07:49:02.0000] <hsivonen>
MikeSmith: there's the bugzilla, but it doesn't have all TODO on it

[07:50:00.0000] <hsivonen>
and yeah, I want to do a lot more coding

[07:51:00.0000] <MikeSmith>
what's the status on the work you've been doing to improve the error output from jing?

[07:51:01.0000] <MikeSmith>
is that done, or you still need to do more work?

[07:51:02.0000] <hsivonen>
MikeSmith: the status is that I'm blocking on George Bina pushing out a new upstread snapshot

[07:51:03.0000] <MikeSmith>
ok

[07:52:00.0000] <MikeSmith>
yeah, I remember you saying that a while back

[07:52:01.0000] <MikeSmith>
what's the current likelihood of George folding back in your changes upstream?

[07:53:00.0000] <hsivonen>
MikeSmith: I think the likelihood isn't too great

[07:53:01.0000] <hsivonen>
we'll see

[08:06:00.0000] <hsivonen>
I think not replying on the mailing list is a key skill

[08:07:00.0000] <hsivonen>
the trouble I have is that I'm not sure how badly the alt stuff is going to come back and bite Validator.nu if I don't explain what I'm doing

[08:07:01.0000] <hsivonen>
that is, that I'm not seeking to be evil here

[08:18:00.0000] <Dashiva>
Geez

[08:19:00.0000] <hsivonen>
Dashiva: ?

[08:19:01.0000] <Dashiva>
I'm pondering a dilemma. What would they choose between a) always perfect alt, everything else in wcag intentionally violated and b) optional alt

[08:20:00.0000] <takkaria>
they would spin round the question such that they didn't have to answer. :)

[08:21:00.0000] <Dashiva>
Yes, I expect most answers would be "We want both, and a pony"

[08:21:01.0000] <hsivonen>
Dashiva: and you'd be immoral for asking :-/

[08:22:00.0000] <Dashiva>
That's okay, I was cited in the original formal complaint, so my innocence is lost

[08:23:00.0000] <hsivonen>
I think it isn't okay. it creates and unhealthy working environment and polarizes people needlessly

[08:29:00.0000] <Dashiva>
I guess I still have problems taking it 100% seriously. It's just so... absurd... at times.

[08:32:00.0000] <hsivonen>
perhaps I'm taking it too seriously, but it really bothers me that discussions about different ways to address an issue have a constant undertone that people who try to do something they believe is for the better are immoral, closet bigots, etc.

[08:36:00.0000] <Philip`>
hsivonen: I don't think anyone ever seeks to be evil - they just see themselves from a twisted perspective, and then act in a way that they consider perfectly rational :-)

[08:37:00.0000] <Dashiva>
We definitely need people like you (and more of them).

[08:38:00.0000] <Dashiva>
Can't let it turn into a war of attrition, where they proclaim "victory" just because everyone gives up trying to reason with them

[08:40:00.0000] <jgraham__>
Gez would be a lot more convincing if he didn't argue in tautologies

[08:41:00.0000] <jgraham__>
"<img> without alt is defined as an incomplete structure incomplete structures are bad, therefore img without alt is bad"

[08:41:01.0000] <jgraham__>
is totally logical but pretty pointless

[08:45:00.0000] <takkaria>
I think some people just don't accept presenting an image as a valid thing to want to do in HTML

[08:52:00.0000] <hsivonen>
jgraham__: whether it is bad isn't the whole story. it's whether the alternative is worse

[08:55:00.0000] <Dashiva>
Haha. I just imagined a conspiracy theory that would explain it all: They have to make sure non-conforming documents are horribly inaccessible, because if non-conforming documents were possible to consume, it would hurt the push for mandatory accessibility.

[09:08:00.0000] <jgraham__>
hsivonen: I know. 

[09:08:01.0000] <hsivonen>
jgraham__: yeah. my point was that the tautology wasn't the only problem

[09:09:00.0000] <jgraham__>
Sure. 

[09:10:00.0000] <jgraham__>
FWIW I'm not sure that the Authoring Tool Accessibility Guidelines are very helpful

[09:11:00.0000] <hsivonen>
I should probably reread them, but I don't feel like doing it right now. 

[09:11:01.0000] <jgraham__>
For example they seem to encourage KompoZer-style dialogs that make it very easy to generate bad alt text

[09:12:00.0000] <MikeSmith>
hsivonen: I'm not sure I'd say you're taking it too seriously, but I would wonder whether the time and energy any and everybody is spending on it might be more productively spent on other things

[09:12:01.0000] <jgraham__>
Although they do say that it should be possible to check the accessibility of a document, they don't seem to take the stance that this should be a human-led process

[09:13:00.0000] <hsivonen>
MikeSmith: I think it's clear that it would be more productive for *everyone* to drop the issue.

[09:13:01.0000] <hsivonen>
MikeSmith: the part I'm worried about is only some people dropping it and design getting done by being out-shouted

[09:15:00.0000] <MikeSmith>
hsivonen: whatever decision might possibly be reached right now would not be the final decision about it anyway

[09:15:01.0000] <jgraham__>
That's a legitimate fear given that e.g. John Foliot has stated that he views the situation as trench warfare

[09:15:02.0000] <MikeSmith>
I mean I can see this discussion will get re-opened when we go to last call

[09:18:00.0000] <hsivonen>
the cost of having the courtesy to discuss stuff instead of just being rude and not talking certainly has gone up lately

[09:19:00.0000] <BenMillard>
I have a message rule which automatically marks as read all messages with subject line mentioning any of various keywords

[09:19:01.0000] <BenMillard>
sadly, I added "alt" to that list a few months ago

[09:19:02.0000] <MikeSmith>
it seems to me it was well known (or should have been well known) that this and other contentious changes to authoring-conformance criteria related to accessibility would provoke just the kinds of reactions just we are seeing now

[09:21:00.0000] <Dashiva>
MikeSmith: It would indeed be wonderful if someone could convince them to leave the issue alone at least until LC

[09:22:00.0000] <MikeSmith>
Dashiva: would be wonderful, but it's not going to happen, of course

[09:25:00.0000] <BenMillard>
MikeSmith, what are you getting at with the "it was well known" comment? it looks like you're suggesting such changes set out to be controversial rather than being genuine attempts at good design...although I may be reading too much between the lines

[09:25:01.0000] <MikeSmith>
BenMillard: you're reading too much between the lines

[09:26:00.0000] <BenMillard>
phew :)

[09:27:00.0000] <BenMillard>
I've not seen much review of data table accessibility since that section got updated

[09:28:00.0000] <BenMillard>
maybe people like what we did (?!)

[09:29:00.0000] <Dashiva>
Would be nice

[09:31:00.0000] <BenMillard>
Lachy & jgraham__, does your presentation show people how to contribute productively to HTML5's development?

[09:31:01.0000] <jgraham__>
BenMillard: I want to cover that for sure

[09:32:00.0000] <jgraham__>
I was planning to use you as an example :)

[09:32:01.0000] <BenMillard>
jgraham, cool :)

[09:34:00.0000] <MikeSmith>
jgraham__: does it give examples of how to not contribute productively? ;)

[09:34:01.0000] <jgraham__>
MikeSmith: Were it only that I were that brave

[09:34:02.0000] <MikeSmith>
heh

[09:34:03.0000] <BenMillard>
I have a dream where I give a presentation about the path I took from my bad initial reactions to HTML5, through the process of studying data tables and hanging around here in IRC, meeting WHATWG people at the W3C thing last year, to the understanding I have now for how the work is done

[09:35:00.0000] <MikeSmith>
jgraham__: It reminds me of something .. here in Japan, there are many signs in public (on trains and such) that have cartoon illustrations of bad/discouraged behavior

[09:36:00.0000] <MikeSmith>
e.g., illustration of a woman doing her makeup on the train

[09:36:01.0000] <MikeSmith>
and statement being basically, Don't do this.

[09:36:02.0000] <jgraham__>
MikeSmith: I remember those; when I was in Japan we wore our rucksacs on the train which was a social faux-pas

[09:37:00.0000] <jgraham__>
Someone pointed us at the pictures

[09:37:01.0000] <MikeSmith>
I wonder if we could come up with some illustrations of bad behavior that we'd like to discourage in discourse about HTML

[09:37:02.0000] <MikeSmith>
jgraham__: yeah, problem with rucksacs is that the trains are so crowded

[09:37:03.0000] <MikeSmith>
so bags ended up bumping into peopel

[09:38:00.0000] <BenMillard>
MikeSmith, flamethrowers, trolls, sheep and bandwagons come to mind?

[09:38:01.0000] <MikeSmith>
BenMillard: ah yes

[09:38:02.0000] <jgraham__>
MikeSmith: The train wasn't particularly crowded and I was pretty annoyed that people seed to think it was OK to just push me out of the way when it was easier to go round me

[09:38:03.0000] <jgraham__>
But such is the joy of different cultures

[09:38:04.0000] <MikeSmith>
I also imagine a "painting the bikeshed" illustration

[09:38:05.0000] <jgraham__>
Especially when you're sleep deprived

[09:39:00.0000] <jgraham__>
s/seed/seemed/

[09:39:01.0000] <MikeSmith>
jgraham__: ah yeah, the pushing-out-of-the-way thing is hard to get used to

[09:40:00.0000] <MikeSmith>
anywhere else, somebody doing it would likely end up getting a punch in the nose

[09:40:01.0000] <jgraham__>
heh

[09:41:00.0000] <MikeSmith>
One way i've found to educate people about it is to push back

[09:41:01.0000] <MikeSmith>
but unfortunately, some of the main culprits are old ladies

[09:41:02.0000] <MikeSmith>
so that doesn't work so well for that case

[09:42:00.0000] <MikeSmith>
or rather, works a bit too well

[09:43:00.0000] <hsivonen>
MikeSmith: I once considered writing observations about how to really annoy CSS WG members on www-style

[09:43:01.0000] <jgraham__>
Japan is a fascinating place. It has very different cultural norms to almost everywhere else in the "developed world"

[09:43:02.0000] <hsivonen>
MikeSmith: but then I decided not to write it considering that my wannabe article from years ago offended people

[09:44:00.0000] <MikeSmith>
well, some people are just waiting to be offended, actively looking around for something to get offended about

[09:44:01.0000] <MikeSmith>
so sometimes nice to oblige them

[09:45:00.0000] <Philip`>
Other times people just write offensive things ;-)

[09:45:01.0000] <MikeSmith>
with some token of offensiveness to keep them busy

[09:45:02.0000] <Dashiva>
Is the css wg still trying to decide if "public" includes logging IRC?

[09:45:03.0000] <annevk>
No, that's settled now

[09:46:00.0000] <jgraham__>
I remember that one of the www-style flamewars made me realise that working groups / mailing lists for WGs are not able to deal with deliberate disruptive influences well.

[09:47:00.0000] <jgraham__>
It's too easy to tarpit everyone 

[09:47:01.0000] <MikeSmith>
jgraham__: aye

[09:48:00.0000] <jgraham__>
BenMillard (or anyone I guess): do you have anything particular about the community that you think should go in the @media talk?

[09:48:01.0000] <MikeSmith>
I guess that issue you mentiong is one of the biggest issues with trying to do purely consensus-driven decision making

[09:49:00.0000] <Dashiva>
BenMillard: Any chance of that dream becoming reality?

[09:50:00.0000] <jgraham>
MikeSmith: The interesting thing is that the WHATWG sidestepped the issue by not maintaining an illusion of decision making being consensus driven

[09:50:01.0000] <MikeSmith>
jgraham: the record of browser implementation progress that has been made with parts of HTML5 seems worth mentioning in your talk

[09:50:02.0000] <MikeSmith>
how long is the talk?

[09:50:03.0000] <jgraham>
MikeSmith: 1 hour. I think we'll aim for about 45+15

[09:51:00.0000] <jgraham>
(the 15 being for questions)

[09:51:01.0000] <jgraham>
So the plan is:

[09:51:02.0000] <MikeSmith>
have you seen the slides I did for my presentation at XTech?

[09:51:03.0000] <jgraham>
1. Introduction

[09:51:04.0000] <jgraham>
2. Design Principles

[09:51:05.0000] <jgraham>
3. Demo how HTML5 features can be used in projects today

[09:51:06.0000] <jgraham>
4. Communit

[09:52:00.0000] <jgraham>
y

[09:52:01.0000] <jgraham>
MikeSmith: Yeah, I saw the slides. We will crib from them for 3 I guess

[09:52:02.0000] <MikeSmith>
cool

[09:53:00.0000] <hsivonen>
jgraham: I have a couple of community observations both of which may not be politically correct

[09:53:01.0000] <MikeSmith>
I think the record of implementation progress for HTML5 has been exceptional compared to other specs

[09:53:02.0000] <MikeSmith>
and that is always worth mentioning

[09:54:00.0000] <hsivonen>
jgraham: 1) the WHATWG community formed when WHATWG was seen as illegitimate by many, so it attracted people who looked at the substance instead of well-known logo

[09:54:01.0000] <MikeSmith>
especially to counter the "The spec will take 15 years to complete??!!" stuff

[09:54:02.0000] <hsivonen>
jgraham: 2) "running code" process vs. consensus process

[09:55:00.0000] <MikeSmith>
jgraham: I also have some general HTML5 overview slides here:

[09:56:00.0000] <MikeSmith>
http://www.w3.org/2008/Talks/04-24-smith/ http://www.w3.org/2008/Talks/04-24-smith/index.pdf

[09:56:01.0000] <MikeSmith>
feel free to crib from those too, if you find anything useful

[09:56:02.0000] <jgraham>
MikeSmith: Thanks

[09:56:03.0000] <MikeSmith>
when I made them, I felt free to crib from previous presentations that annevk and zcorpan had done :)

[09:57:00.0000] <jgraham>
hsivonen: I agree with both your observations

[09:57:01.0000] <hsivonen>
jgraham: do you have my ten points on file?

[09:57:02.0000] <jgraham>
hsivonen: No, I don't think so

[09:58:00.0000] <hsivonen>
jgraham: http://pastebin.ca/992541

[09:58:01.0000] <BenMillard>
Dashiva, seems unlikely as nobody has ever heard of me :|

[09:59:00.0000] <Dashiva>
If you make it, they will ask ;)

[10:01:00.0000] <hsivonen>
regarding consensus process, see the first paragraph of http://en.wikipedia.org/wiki/Management_by_perkele (the content in the third para and onwards is questionable)

[10:01:01.0000] <BenMillard>
jgraham__, "09 You can participate." giving specific URLs, such as to Hixie's message with 6 steps to joining HTMLWG, as well as an irc:// URI for this channel and other useful logistical things would be valuable

[10:02:00.0000] <BenMillard>
until you're already in the community, finding where the community takes place is almost impossible

[10:03:00.0000] <BenMillard>
contributions don't need to be Grand Scientific Experiments; Mythbusters "science with a small s" is infinitely more useful than speculation

[10:04:00.0000] <Dashiva>
Just documenting status quo can be helpful, and that's largely just someone having to put in the hours

[10:05:00.0000] <BenMillard>
Dashiva, exactly

[10:07:00.0000] <MikeSmith>
hsivonen: thanks for the pointer. I remember somebody mentioning this term when we were at XTech

[10:12:00.0000] <BenMillard>
by the way, my detailed notes from Sight City 2008 are coming together here: http://projectcerbera.com/blog/2008/05#day10

[10:12:01.0000] <BenMillard>
I've got to document 9th May 2008 then it's done, I think

[10:13:00.0000] <BenMillard>
people I talked to mostly had no idea HTML5 was being developed

[10:13:01.0000] <BenMillard>
they seemed happy about us wanted them to be involved, though

[10:23:00.0000] <Philip`>
BenMillard: You have a link to http://projectcerbera.local/ which doesn't work

[10:35:00.0000] <jgraham>
hsivonen: Thanks for the links

[11:06:00.0000] <MikeSmith>
jgraham: another suggestion to consider, fwiw: was thinking about whatever publisher it was that put out the "Missing Manual" series of tech books

[11:07:00.0000] <MikeSmith>
the HTML5 spec could be seen as the "missing" spec for HTML

[11:07:01.0000] <Philip`>
(Apparently Pogue Press / O'Reilly)

[11:07:02.0000] <MikeSmith>
the spec defines how conformant implementations must behave 

[11:08:00.0000] <MikeSmith>
the spec that UA/browser implementors have been missing for the last 10+ years

[11:08:01.0000] <Philip`>
That sounds like a good idea - publish the HTML5 spec as a book, such that it's slightly cheaper than printing hundreds of pages on your own laser printer, and use the money to fund future development

[11:09:00.0000] <jgraham>
Philip`: Who would buy the HTML spec as a book?

[11:09:01.0000] <Philip`>
jgraham: People who would otherwise print it out, and realise that the book is cheaper

[11:09:02.0000] <jgraham>
Philip`: e.g.?

[11:09:03.0000] <Philip`>
jgraham: I believe I've heard Chris Wilson talking about the spec as being an inch thick, so I assume he's printed it

[11:10:00.0000] <annevk>
marcosc prints specs

[11:10:01.0000] <jgraham>
Oh well. 

[11:10:02.0000] <jgraham>
Still a market of 10-100 people isn't really going to make the book cheape

[11:10:03.0000] <jgraham>
r

[11:11:00.0000] <annevk>
true

[11:11:01.0000] <jgraham>
I think the time is about right for the first html5-specific book aimed at web-devs though

[11:12:00.0000] <Philip`>
/me remembers there was a Perl 6 book two or three years ago

[11:13:00.0000] <Philip`>
That's already totally obsolete, and Perl 6 isn't even implemented yet, so HTML 5 is in a far better position

[11:14:00.0000] <MikeSmith>
there are services that can print paperback books these days on-demand

[11:16:00.0000] <MikeSmith>
hey, there are some people that hate some of the ideas in the HTML5 spec so much (e.g., handling on non-well-formed content) that maybe there is a market for printing the spec on rolls of toilet paper

[11:17:00.0000] <Philip`>
The HTML5 spec-annotation system should be extended so you can vote for how much you like a section on a scale of 1 to 5

[11:18:00.0000] <Philip`>
then we can automatically extract the most suitable sections for those applications such as toilet paper

[11:18:01.0000] <MikeSmith>
heh

[11:18:02.0000] <MikeSmith>
we're doing some great productive brainstorming here

[11:20:00.0000] <MikeSmith>
but I think Doug Crockford is way ahead of us all, as usual

[11:20:01.0000] <MikeSmith>
with his  JavaScript: The Good Parts book

[11:21:00.0000] <MikeSmith>
great cross-selling opportunity for O'Reilly/Amazon or whoever

[11:21:01.0000] <MikeSmith>
 JavaScript: The Good Parts book bundled with  JavaScript: The Bad Parts toilet paper

[11:22:00.0000] <Philip`>
/me thinks that's an awesome idea

[15:35:00.0000] <Hixie>
mcarter__: otherwise you don't know when to fire the event on the other side

[15:36:00.0000] <mcarter__>
Hixie, why not fire an event every time any data comes through?

[15:40:00.0000] <Hixie>
BenMillard: you can always give your talk as a video blog and put it on youtube :-)

[15:41:00.0000] <BenMillard>
Hixie, interesting idea

[15:41:01.0000] <Hixie>
mcarter__: because if side A sends two strings to side B, you want two events on side B

[15:41:02.0000] <Hixie>
mcarter__: not one, or three

[15:42:00.0000] <BenMillard>
Hixie, it's a question of time and money and equipment, where I have little of any

[15:42:01.0000] <mcarter__>
Hixie, yeah, but protocols generally already take care of framing... and if you're inventing a new protocol, you can just make it \r\n based if you want simple framing

[15:43:00.0000] <mcarter__>
Hixie, what i mean, is that our transport protocol can not have framing, and you can effectively do the same thing in your client by parsing the incoming data

[15:43:01.0000] <Hixie>
mcarter__: if you frame on newlines then you can't send multiline text, and TCP/IP doesn't have framing at a more useful level than "byte" as far as I can tell.

[15:43:02.0000] <Hixie>
mcarter__: i don't want to require that every js author implement a byte buffer to collect strings, that's just annoying fluff that we should take care for them

[15:44:00.0000] <Hixie>
mcarter__: adding framing also means we can later expand the protocol to support both strings and arbitrary binary blobs

[15:44:01.0000] <Hixie>
mcarter__: e.g. when we implement Blobs as proposed in Web API

[15:44:02.0000] <Philip`>
BenMillard: In case you don't have psychic powers and didn't see it while you weren't here: http://krijnhoetmer.nl/irc-logs/whatwg/20080517#l-368

[15:45:00.0000] <mcarter__>
Hixie, I can imagine a scenario where you want to modify an existing server, say a jabber server, just by making it understand the initial handshake, then leaving the rest of its code untouched. With the framing added to the data transport, its not as easy as that

[15:47:00.0000] <Hixie>
mcarter: that wasn't one of the use cases i was considering, but in any case if you're going to that kind of lengths adding a simple translator on the end isn't a big deal

[15:47:01.0000] <mcarter>
perhaps not

[15:49:00.0000] <BenMillard>
Philip`, oops, thanks for that!

[15:50:00.0000] <Hixie>
http://www.openajax.org/runtime/wiki/Feature_Requests_Summary_Page and the pages it link to are hard to read for some reason

[15:50:01.0000] <Hixie>
i don't know why

[15:51:00.0000] <Hixie>
i think it just has too muh fluff

[15:51:01.0000] <Hixie>
instead of being simple and direct

[15:51:02.0000] <Hixie>
i just feel like i'm wading in mud when i read it

[15:51:03.0000] <Hixie>
which is really sad

[15:51:04.0000] <Hixie>
because this would be really important feedback

[15:51:05.0000] <BenMillard>
Hixie, the "Active Feature Requests" duplicates the first column in the third column, only difference being the third column items are linked

[15:52:00.0000] <Hixie>
yeah, for example

[15:52:01.0000] <Hixie>
and the wiki pages themselves have about 50% boilerplate

[15:53:00.0000] <Hixie>
so it's like wading through generic material to get to the actual data

[15:53:01.0000] <BenMillard>
Hixie, they use <th> for the normal headers and <td colspan><strong> for spanned ones, annoyingly :(

[15:54:00.0000] <BenMillard>
I see ordered lists written using (1) foo (2) bar rather than <ol>

[15:54:01.0000] <BenMillard>
I always find that hard to read

[15:54:02.0000] <mcarter>
are those artifacts of mediawiki, or are they deliberately putting those in?

[15:54:03.0000] <Hixie>
yeah

[15:55:00.0000] <BenMillard>
oh, just realised the entire 2nd column in that table just says "coming later" :D :D

[15:55:01.0000] <BenMillard>
so really, it's just a set of <h3><ul> and not a table at all (yet)

[15:56:00.0000] <BenMillard>
Hixie, you aren't going mad. that site has made unhelpful choices for presenting its content

[15:56:01.0000] <Hixie>
glad to hear it's not just me :-)

[15:59:00.0000] <BenMillard>
Hixie, I like how it talks about "heroic JavaScript" in contrast to the "vastly insufficient" CSS: http://www.openajax.org/runtime/wiki/Better_UI_Layout

[15:59:01.0000] <BenMillard>
/me imagines Superman's outfit with a J prefixed to the S

[16:06:00.0000] <Hixie>
heh

[16:52:00.0000] <Hixie>
i think <title> should be made optional for documents intended to be used inside iframes

[16:59:00.0000] <takkaria>
sounds sensible


2008-05-18
[17:01:00.0000] <Philip`>
What's wrong with using <title></title>? That'd allow validators to still tell you when you forgot to put a title in a normal page

[17:03:00.0000] <hasather>
Philip`: agreed

[17:14:00.0000] <Hixie>
Philip`: well would <title></title> be acceptable on a normal page?

[17:14:01.0000] <Hixie>
maybe i should only allow it to be omitted in documents written in data: URIs in <iframe>s

[17:18:00.0000] <Philip`>
Hixie: It would be no less acceptable than <title>----</title> or anything else that doesn't make sense, and the spec shouldn't require pages to make sense, so it shouldn't require against <title></title>

[17:20:00.0000] <Philip`>
(Actually, I don't think I believe that)

[17:21:00.0000] <Philip`>
(but still I don't think the spec can require that you don't write <title></title>, because then people will just write <title> </title> instead, so the effect is the same)

[17:21:01.0000] <Dashiva>
I don't see how <title></title> can be any worse than <title> </title>

[17:22:00.0000] <Dashiva>
... right

[17:22:01.0000] <Hixie>
and i don't see how omitting it altogether is any worse than either of those

[17:23:00.0000] <Hixie>
:-)

[17:23:01.0000] <Philip`>
The practical effect if it's allowed to be omitted is that validators won't tell you when it's omitted, which is bad since usually you don't want to omit it

[17:23:02.0000] <Dashiva>
/me sees ugly parallels to @alt on the horizon

[17:23:03.0000] <Hixie>
you also don't want to leave it blank :-)

[17:23:04.0000] <Hixie>
but what do you think of my other suggestion?

[17:24:00.0000] <Philip`>
and if validators do complain, and you really don't want a title, it's not at all harmful to enter <title></title>

[17:24:01.0000] <Hixie>
that is, only allowing it to be omitted in documents written in data: URIs in <iframe>s

[17:24:02.0000] <Philip`>
Making conformance depend on the context of a document seems like a really bad idea

[17:24:03.0000] <Hixie>
aww

[17:24:04.0000] <Dashiva>
Hixie: I'm crazy enough to say that nobody is going to validate an iframe data:URI anyhow :)

[17:24:05.0000] <Hixie>
i would hope that hsivonen will make his validator check any data: URIs for conformance

[17:24:06.0000] <Hixie>
including nested ones :-)

[17:24:07.0000] <Philip`>
hsivonen was complaining about conformance depending on resources that the document links to, so I don't imagine he'd be happy if it depended on resources linking to the document :-)

[17:24:08.0000] <Hixie>
in due course eventually of course

[17:25:00.0000] <Hixie>
you know, when he adds CSS support, etc

[17:25:01.0000] <Dashiva>
I just don't think the use case is worth the special casing

[17:25:02.0000] <Hixie>
so the use case here is making it easy to use <iframe>s to sandbox blog comments

[17:26:00.0000] <Dashiva>
And I posit that such comments are autogenerated, and therefore can autogenerate <title>Comment by DudeMan</title>

[17:26:01.0000] <Hixie>
I want to allow, e.g.: <iframe seamless sandbox src="data:text/html,I don't like you!"></iframe>

[17:26:02.0000] <Philip`>
Including non-quirks mode stuff?

[17:26:03.0000] <Hixie>
and <iframe seamless sandbox src="data:text/html,Please buy our <a href='http://example.com/'>porn</a>."></iframe>

[17:26:04.0000] <Dashiva>
haha

[17:26:05.0000] <Hixie>
i guess we would require <!DOCTYPE HTML>

[17:27:00.0000] <Dashiva>
I don't think I ever saw porn spam that didn't claim to be free

[17:27:01.0000] <Philip`>
If you're already requiring at least the "data:text/html,<!DOCTYPE HTML>' boilerplate, it doesn't seem hard to add '<title></title>' too

[17:27:02.0000] <Hixie>
i was just showing the embedded markup, don't bikeshed my example :-P

[17:28:00.0000] <Hixie>
Philip`: yeah, i guess

[17:28:01.0000] <Dashiva>
data:text/bikeshed;blue

[17:28:02.0000] <Hixie>
Philip`: but that's sad

[17:28:03.0000] <Philip`>
and everybody would want to write 'data:text/html,<!DOCTYPE HTML><html><head><title></title></head><body>...</body></html>' anyway because they don't like optional tags

[17:29:00.0000] <Hixie>
Philip`: i could get away with making the doctype optional too, and making "sandbox" force strict mode maybe

[17:31:00.0000] <Dashiva>
You're trying to make poor henry cry, aren't you

[17:32:00.0000] <Hixie>
people aren't going to like boilerplate before every comment

[17:32:01.0000] <Hixie>
hm

[17:32:02.0000] <Hixie>
actually

[17:33:00.0000] <Hixie>
i'm going to have to make the quirks mode inherit anyway

[17:33:01.0000] <Hixie>
because the style sheets will be mixed in from the parent

[17:33:02.0000] <Philip`>
I wouldn't mind boilerplate - that's what programming languages have variables for, so I'll only ever have to write it once

[17:34:00.0000] <Dashiva>
I wouldn't put it past authors to add stylesheets to the data uri too

[17:35:00.0000] <Hixie>
that's fine i guess

[17:35:01.0000] <Philip`>
Is it an Opera bug that data:text/html,a#b results in "a"?

[17:35:02.0000] <takkaria>
the prospect of data urls inside data urls inside data urls terrifies me

[17:36:00.0000] <Hixie>
Philip`: i believe not

[17:36:01.0000] <Hixie>
takkaria: i've used them often

[17:36:02.0000] <takkaria>
"often"? :)

[17:36:03.0000] <Hixie>
e.g. for graphics in style sheets in html documents

[17:36:04.0000] <Hixie>
well, you know

[17:36:05.0000] <Philip`>
Hixie: Oh; then doesn't that mean you'll needs lots of fiddly escaping to make sure the constructed data: URIs work right?

[17:36:06.0000] <Hixie>
sometimes

[17:36:07.0000] <Philip`>
and then they won't be human-readable any more, so you might as well use base-64

[17:37:00.0000] <Hixie>
Philip`: you'll need to escape two characters, #, and the character used for quoting the attribute

[17:38:00.0000] <Hixie>
Philip`: if people forget to quote the ", then it'll be quickly obvious (long before security problems, probably)

[17:38:01.0000] <Hixie>
oh and %s i guess

[17:38:02.0000] <Philip`>
Hixie: And &

[17:38:03.0000] <Hixie>
and &

[17:38:04.0000] <Hixie>
hm

[17:38:05.0000] <Hixie>
crap

[17:38:06.0000] <Hixie>
four characters is a problem

[17:39:00.0000] <Hixie>
we could introduce a brand new way of embedding data, but then it'll not be backwards compatible...

[17:40:00.0000] <Philip`>
Hmm, isn't non-backward-compatibility a good thing here? Otherwise a site using <iframe sandbox src="data:..."> will expose users in old browsers to all sorts of security issues that the sandbox was meant to prevent

[17:40:01.0000] <Hixie>
well, the sandbox is an extra level of security

[17:41:00.0000] <Hixie>
in the transition period it doesn't add much

[17:41:01.0000] <Philip`>
and anyway data:text/html,... won't be compatible with IE6/7, and maybe not with IE8 either, so in practical terms there's no backward compatibility

[17:41:02.0000] <Hixie>
oh, true

[17:41:03.0000] <Hixie>
i keep forgetting how much IE sucks

[17:41:04.0000] <Dashiva>
<sandbox> here we go

[17:42:00.0000] <Hixie>
we're using <iframe>s for many other reasons

[17:42:01.0000] <Hixie>
but we don't have to use src="" on <iframe> for embedding data inline

[17:43:00.0000] <Philip`>
Hixie: If the feature adds any security, some authors will start relying on that security (perhaps without considering all the implications in legacy/buggy UAs), and then some users (of legacy/buggy UAs) will suffer because of that

[17:44:00.0000] <Philip`>
which doesn't seem like good security design

[17:44:01.0000] <Dashiva>
So like @safesrc then

[17:44:02.0000] <Philip`>
Dashiva: I thought the point was that the source was *un*safe, hence needing all this protection around it :-)

[17:45:00.0000] <Hixie>
Philip`: yeah well

[17:46:00.0000] <Hixie>
Philip`: good security design in this case is somewhat mutually exclusive with what people actually want

[17:46:01.0000] <Dashiva>
Philip`: It's a safe you put the src in

[17:47:00.0000] <Hixie>
but what i was actually thinking of is a custom parse mode for an iframe "attribute"

[17:47:01.0000] <Philip`>
<iframe src=<<EOF>

[17:48:00.0000] <Hixie>
<iframe data=", and then anything but " will be treated literally, "" will be treated as a single ", and a single " will close it

[17:48:01.0000] <Philip`>
(Yay here-docs)

[17:48:02.0000] <Hixie>
heredocs are hard to escape

[17:49:00.0000] <Hixie>
you have to be able to guarentee that you have a non-predictable random number generator

[17:50:00.0000] <Philip`>
What's the problem with using base64? These comment systems are always going to be generated dynamically, and all languages have a trivial way to base64-encode some data, and then it's easy to see that you've got it right (rather than having to worry about whether there's a character you forgot to escape)

[17:50:01.0000] <Hixie>
people were not happy with stuff you couldn't look at in view source

[17:51:00.0000] <Hixie>
when i suggested that

[17:51:01.0000] <Philip`>
Browser developers need to make 'view source' cleverer so it auto-expands base64

[17:52:00.0000] <Hixie>
browser developers have bigger fish to fry

[17:52:01.0000] <Dashiva>
Fancy encodings are a pickle

[17:52:02.0000] <Dashiva>
Take percent-encoding in URLs when displaying the URL in the status bar

[17:53:00.0000] <Philip`>
<iframe data="normal HTML attributes with escaped &amp; and &quot; and not % or #"> seems easier to understand, and is compatible with all existing HTML parsers and editors and syntax highlighters and libraries and everything

[17:53:01.0000] <Hixie>
that might work

[17:54:00.0000] <Hixie>
though the failure mode for not doign &amp; isn't obvious

[17:54:01.0000] <Philip`>
It's only two characters to escape, or $cgi->escapeHTML(...) or [% ... | html %] or whatever

[17:55:00.0000] <Hixie>
yeah

[17:55:01.0000] <Philip`>
The failure mode for not doing &amp; is not a security issue

[17:56:00.0000] <Philip`>
...so it's still annoying but not fatally so

[17:56:01.0000] <Hixie>
sure it is

[17:56:02.0000] <Philip`>
Not doing &quot; is a more severe problem but it's also pretty much impossible to miss when you get it wrong

[17:57:00.0000] <Hixie>
oh hm actually no it isn't, you're right

[17:57:01.0000] <Hixie>
yeah forgetting to quote " is no issue

[17:57:02.0000] <Hixie>
othermaciej_ needs to do something about his network

[17:57:03.0000] <Hixie>
i'd go insane if i had a network like that

[18:00:00.0000] <Philip`>
Hixie: Perhaps he *has* gone insane

[18:00:01.0000] <Hixie>
true

[18:02:00.0000] <Hixie>
an attribute might work... hmm...

[18:02:01.0000] <Hixie>
what should we call it

[18:02:02.0000] <Hixie>
data="" is bad due to <object data> being a URI

[18:02:03.0000] <Philip`>
<iframe src2="...">

[18:02:04.0000] <Hixie>
<iframe markup=""> ?

[18:03:00.0000] <Hixie>
and markup="" can override src=""

[18:03:01.0000] <Philip`>
content

[18:03:02.0000] <Hixie>
that way you can set both for legacy UAs

[18:03:03.0000] <Hixie>
content="" is good

[18:03:04.0000] <Hixie>
but it takes markup

[18:03:05.0000] <Hixie>
and <meta content> doesn't

[18:03:06.0000] <Dashiva>
Is there anything that takes markup at the moment?

[18:03:07.0000] <Hixie>
no

[18:03:08.0000] <Philip`>
Yes

[18:03:09.0000] <Hixie>
other than href="" and src="" and data=""

[18:04:00.0000] <Philip`>
The contents of most elements takes markup

[18:04:01.0000] <Hixie>
element markup

[18:04:02.0000] <Hixie>
in attributes

[18:04:03.0000] <Hixie>
is what i assume Dashiva meant

[18:04:04.0000] <Dashiva>
Yes

[18:05:00.0000] <Hixie>
i wish we could use the actual content of <iframe>s but the <!-- parsing behaviour is just inane

[18:05:01.0000] <Philip`>
What's the fallback behaviour in current UAs for <iframe not-src-attribute>text</iframe>?

[18:06:00.0000] <Hixie>
same as src=about:blank

[18:06:01.0000] <Philip`>
like, could you put a stripped plain-text version of the comment in there, to have it work (though uglier) for all users?

[18:06:02.0000] <Philip`>
Oh

[18:07:00.0000] <Philip`>
/me doesn't like the attribute name "markup" because it doesn't make him realise that it's giving the content of the iframe

[18:07:01.0000] <Hixie>
yeah

[18:07:02.0000] <Hixie>
i don't like it either

[18:08:00.0000] <Hixie>
<iframe data value content> are all taken already for other uses and would be confusing

[18:08:01.0000] <Hixie>
<iframe source> is too much like src

[18:08:02.0000] <Dashiva>
src2 is too much like http? ;)

[18:08:03.0000] <Hixie>
src2 doesn't convey why it s different from src

[18:09:00.0000] <Hixie>
<iframe subdocument=""> ?

[18:09:01.0000] <Hixie>
<iframe doc=""> ?

[18:09:02.0000] <Philip`>
Leave it unspecified for now and then use implementation experience to determine what name to settle on

[18:09:03.0000] <Dashiva>
That sounds like something

[18:09:04.0000] <Hixie>
Philip`: :-)

[18:09:05.0000] <Hixie>
Philip`: that doesn't work so well for things like this

[18:10:00.0000] <Hixie>
<iframe html=""> ?

[18:10:01.0000] <Hixie>
i wonder how to deal with this xhtml vs html issue here

[18:10:02.0000] <Philip`>
What about XHTML?

[18:10:03.0000] <Philip`>
Good point!

[18:11:00.0000] <Hixie>
/me would be happy to not deal with it :-D

[18:12:00.0000] <Lachy>
what's the purpose of this new attribute for iframe??

[18:12:01.0000] <Hixie>
Lachy: a way of including arbitrary blog comments sandboxed

[18:13:00.0000] <Dashiva>
srcml?

[18:13:01.0000] <Lachy>
ok, I'll go read the irc logs...

[18:14:00.0000] <Hixie>
heh, not sure it'll help :-)

[18:14:01.0000] <Hixie>
doc="" and html="" are the main contenders i think

[18:15:00.0000] <Lachy>
regarding this: <Hixie> i think <title> should be made optional for documents intended to be used inside iframes 

[18:16:00.0000] <Lachy>
doesn't assistive technology read out the title of documents in frames?

[18:16:01.0000] <Hixie>
not when the seamless="" attribute is specified

[18:16:02.0000] <Philip`>
Lachy: It's not too late to pretend you never mentioned accessibility here

[18:16:03.0000] <Lachy>
what's the seamless attr?

[18:17:00.0000] <Dashiva>
Probably to make it not look like an iframe

[18:17:01.0000] <Hixie>
Lachy: an attribute to make iframes work like client-side includes

[18:19:00.0000] <takkaria>
Dmitry would be happy

[18:21:00.0000] <Philip`>
Can you do <!doctype html><iframe seamless src=header.html></iframe>blah blah<iframe seamless src=footer.html></iframe>?

[18:21:01.0000] <Hixie>
yup

[18:21:02.0000] <Hixie>
and clicking on links in header.html will navigate the parent frame

[18:22:00.0000] <Philip`>
/me can't decide whether that's a good thing or a bad thing

[18:22:01.0000] <Hixie>
why would it be bad?

[18:22:02.0000] <Hixie>
people have been asking for it for decades

[18:22:03.0000] <Hixie>
even in the 70s people were saying html should support it

[18:22:04.0000] <Dashiva>
The idea is to have in-document content with outside-document security, isn't it?

[18:23:00.0000] <Hixie>
though i think that may have just been some poor research on the behalf of time travellers

[18:23:01.0000] <Philip`>
People have been asking for flying cars for decades and it'd still be a terrible idea because it'd be abused far too much

[18:23:02.0000] <Hixie>
ground cars are a bad idea too, but we still have those :-)

[18:23:03.0000] <Hixie>
Dashiva: that's another part of the current work i'm doing, but it's not seamless=""

[18:23:04.0000] <Hixie>
Dashiva: that's the sandbox="" part

[18:24:00.0000] <Philip`>
We're stuck with the legacy of ground cars, but that's no reason to make things worse by making cars that far harder to drive

[18:24:01.0000] <Philip`>
s/ / are/

[18:26:00.0000] <Hixie>
flying cars don't have to be manual-drive

[18:26:01.0000] <Hixie>
they could be automated

[18:27:00.0000] <Dashiva>
Flying cars are better than people strapping jet engines on their ground cars

[18:27:01.0000] <Hixie>
(actually that would be easier than making ground cars automated)

[18:27:02.0000] <Philip`>
I think there's more problems in a flying car than the manual gearbox :-p

[18:28:00.0000] <Philip`>
I wouldn't like to be stuck in rush hour on a windy day

[18:28:01.0000] <Hixie>
(first, flying automated vehicles are far easier than ground vehicles, and the AI is far more advanced today; and second, there's little legacy to deal with)

[18:28:02.0000] <Hixie>
gotta go

[18:28:03.0000] <Hixie>
bbl

[18:29:00.0000] <Philip`>
Also, automated cars always lock you in and kill you

[18:29:01.0000] <Dashiva>
Is that a rule?

[18:29:02.0000] <Philip`>
Yes

[18:29:03.0000] <Philip`>
Just watch TV

[18:29:04.0000] <Dashiva>
I remember the cab I took in Japan locked the doors

[18:29:05.0000] <Dashiva>
But it wasn't automated

[18:29:06.0000] <Dashiva>
And it took me to the dentist instead of killing me

[18:29:07.0000] <Philip`>
Did poison gas start hissing out of the dashboard?

[18:30:00.0000] <Dashiva>
No, and it was cheap too

[18:30:01.0000] <Philip`>
Oh

[18:30:02.0000] <Philip`>
I guess real life doesn't live up to the expectations provided by TV :-(

[18:30:03.0000] <Dashiva>
The driver could have been an android, though.

[18:39:00.0000] <takkaria>
Doctor Who had the automated-car-locks-in-and-kills-people thing recently

[18:40:00.0000] <Dashiva>
I don't think it's really limited to cars. Automated houses tend to do the same, don't they?

[18:41:00.0000] <takkaria>
not had as much experience with them

[18:48:00.0000] <Philip`>
takkaria: That's the one I was thinking of :-)

[04:31:00.0000] <annevk>
http://lists.w3.org/Archives/Public/public-xhtml2/2008May/0049.html ...

[05:08:00.0000] <hsivonen>
Hixie: Validator.nu should check data URIs for conformance but not nested ones

[05:09:00.0000] <hsivonen>
(and, yes, Hixie's data URIs are non-conforming for having spaces)

[05:12:00.0000] <hsivonen>
Hixie: "should" as in "I think it already does, if it doesn't, it's a bug"

[05:12:01.0000] <hsivonen>
that is, data URIs are checked for data URI level conformance

[05:12:02.0000] <hsivonen>
the payload is not checked 

[07:04:00.0000] <MikeSmith>
http://blog.whatwg.org/vim-checker

[07:04:01.0000] <MikeSmith>
just published

[07:06:00.0000] <Lachy>
wow! Do people still use vim?

[07:12:00.0000] <MikeSmith>
Lachy: only smart people :)

[07:29:00.0000] <MikeSmith>
the term "Grouping content" doesn't seem to be defined anywhere is the spec

[07:29:01.0000] <MikeSmith>
only as the title of what used to be the "Prose" elements section

[07:31:00.0000] <MikeSmith>
but since what was called "prose content" is now "flow content", maybe the "Grouping content" section should be titled "Flow content" ..

[07:34:00.0000] <MikeSmith>
hmm, no

[07:34:01.0000] <MikeSmith>
I see that all "phrasing content" is also by definition "flow content"

[07:36:00.0000] <MikeSmith>
maybe "Grouping content elements are those used to group together runs of phrasing content into logical structures such as paragraphs and lists."

[08:48:00.0000] <hendry>
MikeSmith: did you see http://natalian.org/archives/2008/05/17/vim-web-ide/

[08:48:01.0000] <hendry>
i should have cut the screencast into 3 parts. as the validator.nu part I screwed up.

[08:49:00.0000] <hendry>
Lachy: if you're in London we could meet up. I won't be going to @media mind

[08:55:00.0000] <MikeSmith>
hendry: saw it but dint watch the video yet

[08:59:00.0000] <hendry>
MikeSmith: video editing is such a pain in Debian. Cinerella does not work for me.

[09:00:00.0000] <MikeSmith>
I have long ago given up on trying to do much with video or audio under Linux

[09:00:01.0000] <MikeSmith>
even playing

[09:00:02.0000] <Philip`>
I used Blender for video editing once

[09:00:03.0000] <hendry>
heh, well... there must be some good video editing web application

[09:00:04.0000] <Philip`>
It actually sort of worked, once I learned what actions would crash it

[09:00:05.0000] <hendry>
(flash application mind)

[09:01:00.0000] <hendry>
Philip`: heh

[09:01:01.0000] <Philip`>
and also I had to compress my input videos because it didn't like 2GB files, if I remember correctly

[09:02:00.0000] <MikeSmith>
avoiding video and audio problems under Linux was one of the best side of effects for me of moving away from Linux-only laptop to a MacBook and running Debian and Windows on it under VMs

[09:03:00.0000] <Philip`>
and it was a bit akward to use for a minute-long video with only about eight video tracks, so I wouldn't want to use it for anything more complex

[09:04:00.0000] <hendry>
MikeSmith: i have no problems with playback

[09:13:00.0000] <Philip`>
Should <a href="1.2.3.4:5"> be interpreted as an absolute or relative URI?

[09:14:00.0000] <Philip`>
(Firefox thinks relative; Opera thinks absolute, which breaks a link on http://ms.aybab2u.com:8080/ )

[09:15:00.0000] <takkaria>
relative, blatantly

[09:16:00.0000] <Philip`>
IE thinks "invalid syntax error"

[09:18:00.0000] <Dashiva>
Philip`: It's the old ": means schema", isn't it?

[09:19:00.0000] <Dashiva>
No, port

[09:20:00.0000] <Philip`>
I don't know what it is, except that it's annoying that it breaks :-)

[09:21:00.0000] <Dashiva>
Probably isn't relevant, actually, since it dealt with manually entering an address

[09:29:00.0000] <MikeSmith>
hendry: got your mail, changes made

[09:31:00.0000] <hendry>
MikeSmith: thanks

[09:32:00.0000] <hendry>
/me wonders if it's possible to link 2mins into a video /myvid.ogg?min=2

[09:34:00.0000] <hdh>
alternate video/audio streams too

[09:37:00.0000] <hendry>
can't find anything on youtube. i think it was possible on google video.

[09:40:00.0000] <MikeSmith>
/me looks around for kfish

[09:40:01.0000] <MikeSmith>
this is what annodex is for, I guess

[09:41:00.0000] <MikeSmith>
Xiph

[09:41:01.0000] <hdh>
henry: looks like it broke sometime ago http://groups.google.com/group/google-video-general/browse_thread/thread/f7645bab2595ce8c

[09:43:00.0000] <hendry>
hdh: good hunting

[09:43:01.0000] <hdh>
thx, I searched the help center

[09:50:00.0000] <hdh>
is there a google custom search (or similar) for all html5 sources (mailing-list, irc, the spec)?

[09:54:00.0000] <hendry>
hdh: not that I know of. good idea. start it and invite me :)

[09:56:00.0000] <hdh>
whatwg.org, html5.org, krijnhoetmer.nl/irc-logs/*, *.w3.org/html/wg/html5/*; any more to add?

[09:57:00.0000] <hdh>
lists.w3.org/Archives/Public/www-html/*, lists.w3.org/Archives/Public/www-html-testsuite/*

[09:58:00.0000] <MikeSmith>
hdh: http://www.w3.org/html/

[10:03:00.0000] <hdh>
http://www.google.com/coop/cse?cx=012969598209138263813:exzxsykptac

[10:09:00.0000] <hendry>
hdh: I would prefer it only searched http://www.whatwg.org/specs/web-apps/current-work/multipage/ not http://www.whatwg.org/specs/web-apps/current-work/

[10:10:00.0000] <Philip`>
Does Google index whatwg.org/issues?

[10:11:00.0000] <hdh>
heh, I can't change any setting (the page says "Save failed")

[10:11:01.0000] <Dashiva>
Philip`: You mean if it indexes the contents of the script? :)

[10:11:02.0000] <Philip`>
Looks like it probably doesn't, in which case there's http://canvex.lazyilluminati.com/misc/cgi/issues.cgi/ , though that shouldn't actually have anything that's not on the mailing lists

[10:12:00.0000] <Philip`>
Dashiva: I mean indexing the data that script gets from its custom API via XHR :-)

[10:12:01.0000] <Dashiva>
I'm pretty sure we'll know very soon if the google spider starts executing scripts :)

[10:13:00.0000] <Philip`>
It does statically follow link-like strings in scripts

[10:13:01.0000] <Philip`>
but it's very dumb about it

[10:14:00.0000] <Philip`>
so it's not quite at the level of embedding a JS interpreter and DOM implementation

[10:18:00.0000] <hendry>
hdh: i had the same problems in firefox. switch to webkit.

[10:21:00.0000] <hsivonen>
/me is mildly amused to see <tt> instead of <code> on the WHATWG wiki

[10:27:00.0000] <hsivonen>
hmm. I wonder if I should just implement the SVG stuff now that I'm looking at the MathML stuff anyway...

[10:30:00.0000] <hsivonen>
yeah, I should

[10:32:00.0000] <hsivonen>
we need to extend the tree builder test format to cover SVG and MathML

[10:42:00.0000] <hdh>
I have set the CSE to allow volunteers

[10:55:00.0000] <hdh>
"Status: Error: OK (403)." <-- the error message when login to annotate system fails, that OK looks weird

[11:53:00.0000] <Philip`>
hsivonen: "Validator.nu should check data URIs for conformance but not nested ones (and, yes, Hixie's data URIs are non-conforming for having spaces)" - should it be checking that in http://validator.nu/?doc=data%3Atext%2Fhtml%2Ca+b too?

[11:53:01.0000] <Philip`>
(Sadly WF2 causes functional regressions when implemented, so you can't test the validator like that in Opera)

[11:56:00.0000] <hsivonen>
Philip`: what's your opinion? should I make it whine about the URI in that case?

[11:56:01.0000] <hsivonen>
Philip`: when I wrote the code, I figured that in this case Validator.nu is acting as a general app accepting data URIs rather than as a checker

[11:57:00.0000] <Philip`>
hsivonen: That sounds sensible - it's more useful if the validator works than if it doesn't work

[11:58:00.0000] <Philip`>
Then again, it's a validator so it's useful if it tells you when you're doing something invalid

[11:59:00.0000] <Philip`>
So, uh, I don't know

[12:04:00.0000] <Philip`>
/me discovers that a todo list which he's never going to read in the future is actually still useful - when the effort of writing the todo entry is comparable to the effort of actually doing the task, it encourages him to just do the task

[14:49:00.0000] <Hixie>
hsivonen: i meant check the payload. i was thinking of making spaces legal, but a new attribute is a better idea.

[15:17:00.0000] <Philip`>
Does anyone happen to know what license the Creative Commons licenses themselves are under?

[15:17:01.0000] <Philip`>
(e.g. is it allowed to take the CC license text and modify it a bit?)

[15:26:00.0000] <bradee-oh>
Hixie: around?

[15:32:00.0000] <Dashiva>
Philip`: I don't see how a licence would need licensing. Patents, maybe, but copyright?

[15:51:00.0000] <Philip`>
Dashiva: It needs copyright licensing as much as any other textual document needs it

[15:52:00.0000] <Philip`>
The GPL explicitly allows copying and distribution but not modification (presumably to prevent the confusion of not-quite-GPL licenses based on it), but I can't find any similar information about CC licenses

[15:55:00.0000] <Dashiva>
Sure, but there's a significant difference between the specific text of the GPL preamble and saying "You are allowed to copy my stuff if you give props"

[15:56:00.0000] <annevk>
Hmm, MikeSmith' post makes me want to try Vim

[15:56:01.0000] <Dashiva>
annevk: What are you using atm?

[15:58:00.0000] <annevk>
komodo and gedit

[15:59:00.0000] <annevk>
and whatever else has syntax highlighting and reasonable undo/redo history

[16:05:00.0000] <annevk>
I wonder how difficult it would be to write a plugin for Dreamweaver that uses Validator.nu

[16:05:01.0000] <annevk>
or for Komodo for that matter

[16:06:00.0000] <jgraham>
Komodo would be pretty easy I think.

[16:06:01.0000] <jgraham>
Plugins are javascript

[16:07:00.0000] <jgraham>
Has someone integrated validator.nu into emacs yet?

[16:08:00.0000] <Dashiva>
annevk: How about getting it into opera? ;)

[16:17:00.0000] <Philip`>
jgraham: Someone must have written a JVM in Emacs Lisp, so you could just run the validator in that

[16:18:00.0000] <jgraham>
Philip`: Always the one with the over complex solution :)

[16:25:00.0000] <Hixie>
bradee-oh: vaguely

[16:27:00.0000] <bradee-oh>
Hixie: I emailed my question to whatwg...  but it's kind of bewildering, I was hoping you could clear up some concerns in short fashion, if you're awake and all that stuff ;)

[16:28:00.0000] <Hixie>
as you note in your e-mail, that mailing list won't affect the webidl spec

[16:28:01.0000] <Hixie>
but yes, [XXX] is for making a [[Delete]] function work

[16:28:02.0000] <Hixie>
but how does the WebIDL spec differ from what we had?

[16:29:00.0000] <Hixie>
the intent was not to not change anything substantial

[16:29:01.0000] <bradee-oh>
Hixie: well, the problem is... the passage we had about enumerating the keys of a storage object... I can't find how the WebIDL spec keeps that functionality in place.

[16:29:02.0000] <bradee-oh>
maybe I'm misreading WebIDL or just glazing over the appropriate section

[16:29:03.0000] <Hixie>
oh hm

[16:30:00.0000] <Hixie>
/me looks at webidl and es3

[16:31:00.0000] <Hixie>
uh yeah, oops

[16:31:01.0000] <Hixie>
webidl doesn't seem to support this

[16:31:02.0000] <Hixie>
will fix

[16:31:03.0000] <bradee-oh>
Hixie: woohoo!

[16:31:04.0000] <bradee-oh>
Hixie++

[16:31:05.0000] <bradee-oh>
though to be fair...

[16:31:06.0000] <bradee-oh>
Hixie-- for the original removal  ;)

[16:32:00.0000] <bradee-oh>
Hixie++ // for being such a super guy

[16:32:01.0000] <bradee-oh>
thanks!

[16:32:02.0000] <Hixie>
heycam: we have a problem with webidl

[16:33:00.0000] <Hixie>
heycam: there doesn't seem to be a way to say that all the members should be DontEnum but that a certain set of properties should be added (that aren't in the IDL) for the purposes of for-in enumeration

[16:34:00.0000] <Hixie>
bradee-oh: the fix might take a while, just assume it will happen. if you really need the spec fixed (e.g. because someone is arguing with you) then let me know in a couple of days if i haven't yet done it

[16:34:01.0000] <Hixie>
bradee-oh: (i'm in the middle of some other edits)

[16:35:00.0000] <bradee-oh>
Hixie: assurances that it will happen is good enough for me.  I'll check up on it later this week.  thanks!

[16:35:01.0000] <Hixie>
np

[16:38:00.0000] <heycam>
Hixie, yeah i just saw that mail from bradee-oh 

[16:38:01.0000] <Hixie>
cool


2008-05-19
[19:29:00.0000] <MikeSmith>
annevk: fwiw, off-the-shelf highlighting in vim is quite good and pretty easily configurable, and it's undo/redo mechanism is significantly better than any other editor I've used

[19:30:00.0000] <MikeSmith>
it has unlimited undo, and a command, "earlier", that take a time argument - to take you back to the state of the buffer that point in time

[19:30:01.0000] <MikeSmith>
http://vimdoc.sourceforge.net/htmldoc/undo.html#:earlier

[19:34:00.0000] <heycam>
annevk, done the web idl abstract rewording you asked for earlier

[02:38:00.0000] <hendry>
/me wonders if there is a campaign to get rid of XSLT from browsers

[02:39:00.0000] <Philip`>
/me hopes not, since he has some code that uses it

[02:41:00.0000] <Dashiva>
What use is client-side XSLT apart from outsourcing the processing cycles to the browser? :)

[02:42:00.0000] <Philip`>
The XML data could come from a remote server you don't have control over, so you can't do the XSLT on the server side

[02:43:00.0000] <Philip`>
or you could want to transform a single XML file in several different ways, and not want to have the client download several slightly different XML files instead of just one

[02:46:00.0000] <annevk>
The remote server argument is slightly bogus without Access Control

[02:46:01.0000] <Philip`>
But it's less bogus with it :-)

[02:48:00.0000] <Dashiva>
Besides, if you want XSLT on remote server content you can pull it to your local server once and do XSLT there, leaving you with local requests too. Double win :D

[02:49:00.0000] <annevk>
Dashiva, it wouldn't be user-specific anymore in that case

[02:49:01.0000] <Philip`>
Dashiva: That's kind of nasty for performance and cost and security

[02:50:00.0000] <Philip`>
Also it requires too much cleverness on the server - it's much nicer when you can just serve static files

[03:02:00.0000] <Philip`>
hendry: About http://natalian.org/archives/2008/05/15/10-years-of-xml/ "Near impossible to diff (and nicely version control)" - there's research like http://www.seas.upenn.edu/~harmony/ that seems kind of relevant for that (though I've no idea how well it'd work in practice)

[03:04:00.0000] <hendry>
well I still think XSLT is unneccessary bloat in the browser

[03:04:01.0000] <hendry>
i've personally since very few websites use it. i wonder if there are some stats of how much of the "web" uses that publishing technique

[03:04:02.0000] <MikeSmith>
hendry: It's a moot point now

[03:05:00.0000] <MikeSmith>
and I reckon we've got a lot of other examples of unnecessary bloat in browsers

[03:06:00.0000] <MikeSmith>
anyway, it gives developers another choice

[03:12:00.0000] <MikeSmith>
hendry: a rant about XML can't be complete without specific mention of Draconian error handling

[03:13:00.0000] <hendry>
MikeSmith: that's Hixie's rant :)

[03:13:01.0000] <MikeSmith>
dunno if you've seen this:

[03:13:02.0000] <MikeSmith>
http://esw.w3.org/topic/HTML/DraconianErrorHandling

[03:15:00.0000] <hendry>
MikeSmith: ok ok, i'll link it :)

[03:16:00.0000] <hsivonen>
In the early days of Atom, I disagreed with markp on XML error handling.

[03:16:01.0000] <hsivonen>
I was wrong. He was right.

[03:17:00.0000] <MikeSmith>
hendry: fwiw, judging popularity of XML by how many people attended one session at XTech 2008 doesn't seem like any kind of real indicator at all (since there were only 100 or so total attendees at XTech anyway)

[03:19:00.0000] <MikeSmith>
hsivonen: yeah, Mark was the real pioneer in stating very clearly and emphatically and publicly about why it didn't and couldn't work

[03:19:01.0000] <hsivonen>
(although writing a liberal feed parser isn't the full answer. We still need a spec on how to be liberal in an interoperable way.)

[03:19:02.0000] <MikeSmith>
hsivonen: true

[03:19:03.0000] <Philip`>
The only reliable way to measure popularity is Google search result counts, and then you'll find that XML is 98% as popular as God

[03:19:04.0000] <MikeSmith>
heh

[03:20:00.0000] <MikeSmith>
/me heads off to write a rant about God

[03:20:01.0000] <MikeSmith>
speaking of God, the real rant I'd write if I had the time is, Please don't get religous

[03:21:00.0000] <MikeSmith>
especially don't get religious about technologies

[03:21:01.0000] <MikeSmith>
either way, good or bad

[03:21:02.0000] <MikeSmith>
because nobody really cares about your religious beliefs 

[03:22:00.0000] <Philip`>
What do you mean by "religious"?

[03:22:01.0000] <Lachy_>
MikeSmith, it would be better if people didn't get religious about anything at all, not just technology. Even religion

[03:22:02.0000] <MikeSmith>
Lachy_: amen

[03:22:03.0000] <zcorpan>
Lachy: what if being religious makes one happier?

[03:22:04.0000] <Philip`>
/me assumes people tend not to literally worship technologies

[03:23:00.0000] <toolskyn>
Philip`: according to googlefight.com, xml is actually more popular than god

[03:23:01.0000] <MikeSmith>
Philip`: religous = dogmatic support/rejection of particular ideas/technologies/whatever

[03:23:02.0000] <Lachy>
zcorpan, getting high makes some people happier. That doesn't make it a good thing to do.

[03:24:00.0000] <Philip`>
What if you being religious makes everyone else happier (because e.g. it discourages you from harming them)?

[03:24:01.0000] <Dashiva>
Crushing your enemies, having them driven before you, and hearing the lamentations of their women also makes some people happy

[03:24:02.0000] <zcorpan>
Lachy: getting high has long-term physical damage, i don't see being religious having such consequences

[03:24:03.0000] <MikeSmith>
Lachy: getting high is a good thing to do, actually

[03:24:04.0000] <Lachy>
Philip`, the argument from morality you're partially using is bogus

[03:24:05.0000] <MikeSmith>
more people should do it

[03:25:00.0000] <MikeSmith>
I have personal empirical evidence that getting high is a good thing

[03:25:01.0000] <jgraham>
Being religious might have long term damage to society

[03:26:00.0000] <hsivonen>
there seems to be a clash between screen readers and hand ergonomics

[03:26:01.0000] <Lachy>
zcorpan, religion gets in the way of progress, particularly science and education, causes more wars than anything else in history, and is overall a big waste of time.

[03:26:02.0000] <hsivonen>
the default key bindings of both Orca and NVDA suck on a keyboard that doesn't have an Insert key

[03:26:03.0000] <MikeSmith>
Lachy: religion unites people toward common goald

[03:26:04.0000] <MikeSmith>
goal

[03:26:05.0000] <MikeSmith>
goalsssss

[03:26:06.0000] <MikeSmith>
and against common enemies

[03:26:07.0000] <Dashiva>
MikeSmith: Uniting with the common goal of not uniting with other people? ;)

[03:27:00.0000] <MikeSmith>
e.g. against XML, or against "bad HTML"

[03:28:00.0000] <MikeSmith>
or against "lazy authors" and in support of... whatever  the alternative is ("educating authors"?)

[03:28:01.0000] <Philip`>
Lachy: It seems to have encouraged progress in architecture, given how most major building projects in the past have either been for war or religion

[03:28:02.0000] <Dashiva>
In support of requiring a HTML author's licence

[03:29:00.0000] <MikeSmith>
Dashiva: great idea

[03:29:01.0000] <MikeSmith>
we should issue those

[03:29:02.0000] <MikeSmith>
like 007 license to kill

[03:30:00.0000] <MikeSmith>
instead, license to write HTML that actually works in browsers

[03:30:01.0000] <MikeSmith>
and that parses interoperably

[03:30:02.0000] <Dashiva>
With licence to parse unambiguously

[03:30:03.0000] <Philip`>
If you get caught with overlapping tags then you get 3 points on your license

[03:31:00.0000] <hsivonen>
Philip`: you mean like <i <b> >? :-)

[03:31:01.0000] <Dashiva>
That would be an interesting mental exercise, write up penalty guidelines for a HTML licence

[03:31:02.0000] <toolskyn>
we should have one of these IE 'friendly warnings' about code written by a unlicenced author

[03:31:03.0000] <Dashiva>
How many points for not quoting an attribute containing spaces?

[03:32:00.0000] <Philip`>
hsivonen: No :-p

[03:32:01.0000] <Philip`>
hsivonen: (but it's not overlapping elements, since elements are just part of a tree and there's no concept of overlapping, and I can't remember a better term)

[03:32:02.0000] <Dashiva>
interleaved?

[03:32:03.0000] <Dashiva>
incestuous?

[03:33:00.0000] <MikeSmith>
"incestuous" is a much better term -- more intriguing

[03:33:01.0000] <Philip`>
I suppose "misnested" works

[03:33:02.0000] <Dashiva>
It does cover the "My parent is my sibling" thing

[03:34:00.0000] <hsivonen>
hmm. I getting tree builder test failures again

[03:34:01.0000] <Philip`>
("... [this algorithm] has been dubbed the "adoption agency algorithm" (in contrast with other possibly algorithms for dealing with misnested content, which included the "incest algorithm", the "secret affair algorithm", and the "Heisenberg algorithm").")

[03:34:02.0000] <hsivonen>
presumably, there has been a reason for the tests to change

[03:35:00.0000] <hsivonen>
daytime soap opera algorithm

[03:35:01.0000] <MikeSmith>
the "red-headed stepchild" algorithm

[03:37:00.0000] <MikeSmith>
the red-headed stepchild metaphor is maybe a good metaphor for how some view the work on spec'ing text/html and Web apps

[03:38:00.0000] <hsivonen>
/me doesn't like hair-colorist algorithm names

[03:38:01.0000] <hsivonen>
http://pastebin.ca/1022593

[03:38:02.0000] <hsivonen>
those are the errors I'm getting

[03:43:00.0000] <hsivonen>
hmm. I've made marquee scoping...

[03:43:01.0000] <hsivonen>
but still </div> closes the marquee...

[03:43:02.0000] <hsivonen>
hmm...

[03:45:00.0000] <MikeSmith>
I have a request that if somebody here (Philip`? or of jgraham is listening) who has commit access to the html5lib sources could consider, I'd appreciate it

[03:46:00.0000] <MikeSmith>
some time last year, Mark P. commented out this line in html5parser.py

[03:47:00.0000] <MikeSmith>
sys.stderr.write("Warning: Undefined behaviour for start tag %s"%name)

[03:47:01.0000] <MikeSmith>
but for whatever reason didn't comment out this one:

[03:47:02.0000] <MikeSmith>
sys.stderr.write("Warning: Undefined behaviour for end tag %s"%name)

[03:48:00.0000] <MikeSmith>
wondering of somebody could comment that out in the source

[03:48:01.0000] <MikeSmith>
or at least put at "\n" on the end of it if it's kept in

[03:53:00.0000] <Philip`>
MikeSmith: Is there an easy way to make that message get printed?

[03:54:00.0000] <MikeSmith>
Philip`: put a <section> element into your source

[03:54:01.0000] <MikeSmith>
or any other new element that Hixie hasn't yet added to the parsing algorithm

[03:54:02.0000] <MikeSmith>
e.g., <header>

[03:55:00.0000] <Philip`>
Ah, yes

[03:56:00.0000] <Philip`>
MikeSmith: Committed

[03:56:01.0000] <MikeSmith>
Philip`: cheers

[03:56:02.0000] <Philip`>
(since it looks like http://canvex.lazyilluminati.com/html5lib/changeset/988 just forgot about the end tag case)

[03:57:00.0000] <MikeSmith>
Philip`: I guess I can actually see some value in having that message emitted, even for non-debugging purposes

[03:58:00.0000] <MikeSmith>
but for consistency at least, if it's already commented out for the start-tag case, seems it is right to have it out for the end-tag case

[03:58:01.0000] <hsivonen>
hmm. I have a string interning bug with special legacy names like marquee and listing

[03:58:02.0000] <Philip`>
MikeSmith: I'm not sure I see the value, since there's nothing a user of the library can do about the issue

[03:59:00.0000] <Philip`>
(and since it parses pretty much like the user would expect)

[03:59:01.0000] <MikeSmith>
Philip`: OK

[04:00:00.0000] <Philip`>
But someone could point out the value to me :-)

[04:01:00.0000] <hsivonen>
so is there a one-stop place for finding all the parsing algorithm-sensitive obsolete elements?

[04:05:00.0000] <Philip`>
hsivonen: Like extracting all the element name strings listed in the spec?

[04:05:01.0000] <hsivonen>
Philip`: yeah

[04:05:02.0000] <hsivonen>
Philip`: I extracted all the names that the non-parsing part of the spec defines as conforming

[04:06:00.0000] <hsivonen>
now I manually added marquee, nobr, wbr, listing and plaintext because those failed test cases

[04:09:00.0000] <hsivonen>
/me landed a huge pile of refactored and new parser code

[04:10:00.0000] <Philip`>
hsivonen: applet, basefont, bgsound, big, center, dir, font, frame, frameset, hidden, image, isindex, listing, marquee, nobr, noembed, noframes, plaintext, s, spacer, strike, tt, u, wbr, xmp

[04:10:01.0000] <hsivonen>
Philip`: thank you

[04:11:00.0000] <Philip`>
are the things I see as /"[a-z0-9]+"/ in the tree constructor spec as of some months ago that are not in the list of conforming elements

[04:11:01.0000] <Philip`>
(though I'm not totally positive I didn't accidentally miss any)

[04:12:00.0000] <hsivonen>
"hidden"???

[04:12:01.0000] <Philip`>
Uh

[04:12:02.0000] <hsivonen>
is that a false positive on the attribute?

[04:12:03.0000] <Philip`>
Yes

[04:13:00.0000] <Philip`>
I removed most of the false positives but missed that one

[04:13:01.0000] <hsivonen>
ok. thanks

[04:14:00.0000] <hsivonen>
Philip`: thanks. my list lacked bgsound and noembed still

[04:15:00.0000] <hsivonen>
fortunately, the additions didn't break my magic hash function

[04:16:00.0000] <Philip`>
/me sees that the grass in his garden is now significantly taller than a cat lying down

[04:16:01.0000] <jwalden>
/me wonders why freenode just dropped like a rock

[04:16:02.0000] <hsivonen>
/me didn't notice freenode dropping

[04:16:03.0000] <Philip`>
jwalden: Just looked like a minor netsplit to me

[04:16:04.0000] <Philip`>
11:16 -!- Netsplit wolfe.freenode.net <-> irc.freenode.net quits: mpt, jwalden, weinig

[04:16:05.0000] <jwalden>
interesting

[04:17:00.0000] <jwalden>
from my pov everyone else dropped but those three :-)

[04:17:01.0000] <mpt>
we're special

[04:18:00.0000] <Philip`>
/me wonders if IRC networks all use a spanning tree between servers, rather than something more robust

[05:05:00.0000] <zcorpan>
othermaciej: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3Chtml%20test%3Da%3E%3Cscript%3Edocument.documentElement.setAttributeNS(''%2C'test'%2C'b')%3C%2Fscript%3E

[05:05:01.0000] <zcorpan>
othermaciej: results in duplicate attributes in webkit

[05:07:00.0000] <zcorpan>
for some reason i thought the dom didn't check for duplicate attributes with different prefixes

[05:09:00.0000] <othermaciej_>
zcorpan: if you make the namespace parameter null it works, so the bug seems to be treating empty string and null differently

[05:09:01.0000] <othermaciej>
"works" == adds the attribute only once

[05:18:00.0000] <Philip`>
Hmm, I have a site which uses TinyMCE so non-technical users can edit bits of content, and I've been asked to remove the part of the TinyMCE image upload dialog that asks for alt text because it's too confusing for users

[05:20:00.0000] <hsivonen>
argh. my token name list is lacking the ruby stuff

[05:21:00.0000] <jmb>
Philip`: is that merely an interface issue, or is the concept of alt text causing the confusion?

[05:22:00.0000] <Philip`>
jmb: I'm not quite sure what the distinction is - do you mean could the problem be fixed by changing the UI to still ask for alt text but in a different way (as opposed to fixing the interface to simply not ask for it at all)?

[05:23:00.0000] <jmb>
that's the distinction I'm interested in, yes

[05:25:00.0000] <zcorpan>
Philip`: perhaps there should be a button "No images view" which replaces all images with editable boxes

[05:25:01.0000] <Philip`>
I can't think of any way the UI could ask for alt text without having to explain to users the concepts like the purpose of it and why they should bother, and I can't think of any way to explain that without confusing or scaring the users who are just wanting to stick an image in there

[05:25:02.0000] <zcorpan>
with an explanation that search engines and blind people will see that text instead of the images

[05:26:00.0000] <Philip`>
zcorpan: That sounds too hard to implement - I can cope with changing dialog boxes, but I don't want to fiddle with the TinyMCE code in any non-trivial way

[05:26:01.0000] <hsivonen>
Philip`: I think a big problem with the whole alt debate is wanting to build tool UI for it but not wanting to explain to users what it is about

[05:27:00.0000] <hsivonen>
which is why the V.nu Image Report actually mentions accessibility in the UI text

[05:27:01.0000] <zcorpan>
Philip`: but i think it would be better UI than dialog boxes

[05:28:00.0000] <Philip`>
(In this case, it's unlikely that many people will ever view the content that was created, and the creators usually won't care about search engines or blind people - they just get the satisfaction of making their own text appear on the site)

[05:29:00.0000] <zcorpan>
if they don't care then forcing them will produce bad alt text anyway

[05:29:01.0000] <hsivonen>
argh. the SVG stuff requires storing both the lowercase name and the case-corrected name on the stack

[05:29:02.0000] <hsivonen>
bad

[05:29:03.0000] <Philip`>
zcorpan: My definition of "better" includes "possible to implement without excessive effort" :-)

[05:29:04.0000] <Philip`>
hsivonen: Can you store just the case-corrected name and use case-insensitive comparison operators in appropriate places?

[05:29:05.0000] <hsivonen>
Philip`: that would kill perf

[05:30:00.0000] <hsivonen>
Philip`: actual tree builders won't need to keep the case-corrected string around, but to implement streaming SAX, it has to stay around

[05:31:00.0000] <hsivonen>
so another round of changing interfaces it is...

[05:31:01.0000] <MikeSmith>
hsivonen: while recognizing the general principle of trying to avoid additional options/pragmas whenever possible, I continue to think a pretty effective UI for the specific alt case (and perhaps only the alt case) would be an option saying, "Don't show me error messages about missing alt attributes."

[05:31:02.0000] <zcorpan>
hsivonen: would making the tokenizer case-preserving and having the parser do case-insensitive checks kill perf?

[05:32:00.0000] <hsivonen>
zcorpan: I'm not sure

[05:32:01.0000] <hsivonen>
zcorpan: probably not if the tree builder dispatches on magic enums instead of strings

[05:33:00.0000] <hsivonen>
currently, mine dispatches on interned string identity

[05:33:01.0000] <hsivonen>
but I want to change it to use magic numbers instead

[05:33:02.0000] <hsivonen>
I'm just unsure if today is the right time to do it

[05:33:03.0000] <hsivonen>
that is, will Hixie change stuff in a way that requires me to reassign the magic numbers

[05:34:00.0000] <zcorpan>
probably :P

[05:38:00.0000] <Philip`>
That sounds like a sensible answer whenever the question is "will Hixie cause pain to an implementor like me"

[05:39:00.0000] <hsivonen>
that is, are there elements that are safe to coalesce into one enum value?

[05:40:00.0000] <hsivonen>
td and th seem to always go together

[05:40:01.0000] <hsivonen>
and thead, tbody and tfoot

[05:40:02.0000] <hsivonen>
but are those all?

[05:40:03.0000] <hsivonen>
oh and h1-h6

[05:40:04.0000] <Philip`>
What about things like b and i?

[05:41:00.0000] <hsivonen>
now I have to run some regexps on my code to find out

[05:41:01.0000] <Philip`>
Or: If you coalesce h1-h6, how do you handle <h1>foo</h6> ?

[05:41:02.0000] <hsivonen>
Philip`: I'm thinking about coalescing what the spec always has under one case

[05:42:00.0000] <Philip`>
Ah

[05:42:01.0000] <Philip`>
/me could probably derive that automatically from his OCamlified version of the spec, except it's old and incomplete and currently broken :-(

[05:43:00.0000] <hsivonen>
I could derive it almost automatically if I hadn't forgotten Perl as much as I have :-(

[05:44:00.0000] <Philip`>
TIMTOWTDI, so if you forget several ways then there's usually still another way to solve any problem with Perl

[05:47:00.0000] <Dashiva>
Philip`: And then you have two problems.

[05:49:00.0000] <Philip`>
Dashiva: That's what one-liners are for - once you've solved your original problem using Perl, you just discard the Perl into the murky depths of your .bash_history and it's no longer a problem

[05:50:00.0000] <Lachy>
LOL, it's funny when comment spammers try to post on-topic comments to the whatwg blog, yet fail miserably :-)

[05:50:01.0000] <Lachy>
this is in the moderation queue for the Vim article. "Your article is very interesting, Checking means verify the file. Thank you for giving this information."

[05:53:00.0000] <Philip`>
Blogs should just not accept URLs in comments at all

[05:53:01.0000] <Dashiva>
It's too bad spam bot authors have no sense of humor, or they could use "Your ideas are intriguing and I wish to subscribe to your newsletter."

[05:53:02.0000] <Philip`>
(including not having an explicit "website" field for commenters)

[05:58:00.0000] <zcorpan>
Philip`: then spammers will just write plain text URLs

[05:59:00.0000] <zcorpan>
Philip`: and people who actually want to link to relevant things it will just be annoying to deal with plain text URLs

[05:59:01.0000] <Dashiva>
We'll avoid the whole target/window.open spectacle, though :)

[06:12:00.0000] <zcorpan>
what Julian Reschke suggests for xhr seems to go against the priority of constituencies principle

[06:13:00.0000] <zcorpan>
http://www.w3.org/mid/48314BE6.9060708⊙gd

[06:13:01.0000] <hsivonen>
sigh

[06:20:00.0000] <hsivonen>
Hixie: any particular reason why pre and listing don't go together when breaking out of foreign lands?

[06:21:00.0000] <zcorpan>
hsivonen: because pre is common and listing is not?

[06:21:01.0000] <hsivonen>
yeah. I'm going to send email

[06:23:00.0000] <Lachy>
Philip`, URLs and email addresses provide a convenient mechanism for giving automatic approval or rejection of repeat commenters, based on their prior comment history

[06:23:01.0000] <Lachy>
it also allows commenters to be reliably identified and help them become known if they're seen to consistently write good comments

[06:24:00.0000] <Philip`>
Lachy: But I could trivially write a comment and use someone else's email address and URL

[06:24:01.0000] <Lachy>
surprisingly, few spammers actually seem to do that

[06:24:02.0000] <Philip`>
so "reliably identified" seems untrue

[06:25:00.0000] <Lachy>
although there are times where trolls will falsely post as someone else, it doesn't happen frequently enough to be of great concern

[06:25:01.0000] <Lachy>
I only know it to have happened once to me in all my years of reading and posting comments on blogs

[06:26:00.0000] <zcorpan>
readers should be able to report bad comments

[06:26:01.0000] <zcorpan>
i think that results in less overhead to deal with comments

[06:26:02.0000] <Philip`>
Then spammers will start spamming the 'report bad comment' feature

[06:27:00.0000] <Lachy>
rating systems like that are used on slashdot, youtube, digg, etc.

[06:27:01.0000] <Lachy>
They're not entirely successful

[06:27:02.0000] <Philip`>
Some much less successful than others

[06:27:03.0000] <zcorpan>
ok

[06:27:04.0000] <Philip`>
(The quality of comments on Slashdot does seem somewhat higher than on Youtube)

[06:28:00.0000] <Lachy>
I'm surprised slashdot has done little to actually filter out the goatse spam.

[06:28:01.0000] <Lachy>
Philip`, that's because the intelligence of slashdot readers is probably, on average, greater than that of your average youtube user

[06:30:00.0000] <Philip`>
Lachy: It seems to have pretty much died out by itself, just by being modded to -1 all the time so normal users will never see it

[06:31:00.0000] <Lachy>
I always have all comments shown. I don't like comments being automatically hidden, since it messes up threading and makes me wonder what I'm missing out on

[06:32:00.0000] <Philip`>
You're missing out on the goatse spam :-p

[06:32:01.0000] <Lachy>
sometimes the troll comments are funny

[06:38:00.0000] <hsivonen>
/me identified 52 magic enum items for the tree builder

[06:38:01.0000] <hsivonen>
of which 14 group more than one element name

[06:46:00.0000] <annevk>
implementing SVG for demonstration purposes?

[06:46:01.0000] <hsivonen>
annevk: I'm preparing to

[06:47:00.0000] <hsivonen>
annevk: I'm doing architectural changes that make the SVG case tweaking stuff not suck quite that badly for perf

[06:52:00.0000] <gsnedders>
Internet at home is broken :(

[06:53:00.0000] <Philip`>
Try your neighbours' internets

[06:54:00.0000] <gsnedders>
All require passwords (in terms of wireless, at least)

[06:54:01.0000] <gsnedders>
/me is in café in town

[06:55:00.0000] <Philip`>
They might be using WEP, which doesn't really count as a password ;-)

[06:55:01.0000] <gsnedders>
:)

[06:55:02.0000] <Lachy>
gsnedders, most people use very simple passwords. It shouldn't be too hard to guess, or even crack

[06:55:03.0000] <gsnedders>
Lachy: too much effort :(

[06:56:00.0000] <gsnedders>
it makes no sense, because I get to SHOWTIME, before it fails to get any further.

[06:56:01.0000] <Lachy>
is it less effort to walk to the cafe, than it is to take 5 minutes to crack a password?

[06:56:02.0000] <gsnedders>
And BT being typically BT are refusing to do anything about it and are blaming us.

[06:56:03.0000] <gsnedders>
Lachy: It means trying to crack the password. :P

[06:57:00.0000] <gsnedders>
Lachy: and it's only 50m away from school

[06:58:00.0000] <Philip`>
A couple of wireless routers around my house are unsecured and have login/password "admin"/"admin"


2008-05-20
[18:19:00.0000] <MikeSmith>
Hixie: editorial -

[18:19:01.0000] <MikeSmith>
"The reversed attribute is a boolean attribute. If present, it indicates that the list is an ascending list. If the attribute is present, the list is a descending list."

[18:20:00.0000] <MikeSmith>
2nd sentence should start "If absent,"

[18:52:00.0000] <MikeSmith>
hmm, the algorithm for determining what an object element represents has a requirement to check the value of a "classid" attribute

[18:52:01.0000] <MikeSmith>
but nowhere is such an attribute defined

[18:53:00.0000] <MikeSmith>
I guess it's meant to be a content attribute

[18:53:01.0000] <MikeSmith>
but it's not in the content model for the object element

[18:54:00.0000] <Philip`>
The content model only lists conforming attributes, and classid is non-conforming

[18:54:01.0000] <MikeSmith>
ah

[18:55:00.0000] <MikeSmith>
Philip`: this is some MS proprietary attribute?

[18:55:01.0000] <Philip`>
so it's the same as how e.g. <isindex> is used in the parsing algorithm, but is not defined anywhere

[18:55:02.0000] <MikeSmith>
OK

[18:55:03.0000] <Dashiva>
Oh boy, more alt fun

[18:57:00.0000] <Dashiva>
If the designers of these tools cannot at this time 

[18:57:01.0000] <Dashiva>
make their tools ATAG-compliant

[18:57:02.0000] <MikeSmith>
seems like for those kinds of cases, to avoid confusion (mine), the spec should explicitly mention that those are things that are not conformant for authoring purposes and so not documented in the spec

[18:57:03.0000] <Dashiva>
That line is such a wonderful thing

[18:57:04.0000] <Philip`>
MikeSmith: I think it's mostly used for ActiveX GUIDs, but I can see some people suggesting things like classid="java:Example.class" too but I don't know if that's implemented anywhere

[18:58:00.0000] <Philip`>
MikeSmith: How you can document something to say that it's not documented in the spec? :-p

[18:58:01.0000] <MikeSmith>
heh, true I suppose

[18:59:00.0000] <hsivonen>
I misread the commit log and thought Hixie made classid conforming :-(

[18:59:01.0000] <MikeSmith>
I think it would be appropriate for all instances of those kinds of things to be enclosed in <blink> tags

[18:59:02.0000] <hsivonen>
I think making classid non-conforming at this time does not make sense, even though it is product-specific

[19:00:00.0000] <Philip`>
MikeSmith: I agree it'd be nice to be explicit when things mentioned in the spec are non-conforming, because otherwise it's not at all clear that the spec author didn't just forget to add that attribute to the list

[19:00:01.0000] <MikeSmith>
Philip`: exactly

[19:01:00.0000] <hsivonen>
also, to me it would make sense for non-IE browsers to map the Flash classid to an NPAPI Flash plug-in, etc.

[19:04:00.0000] <csarven>
re: http://www.whatwg.org/specs/web-apps/current-work/multipage/section-global.html#embedding "So here we're saying that microformats shouldn't use this... is that what we want?" -- I suppose, but, I don't believe microformats is particularly interested in "Embedding custom non-visible data". One of its main goals is to mark "visible" data.

[19:04:01.0000] <csarven>
I don't know who wrote that or if there is a backlog for it

[19:05:00.0000] <Philip`>
csarven: Hixie wrote it

[19:14:00.0000] <csarven>
microformats is still open to finding a better way to do datatime-design-pattern. For a second there I thought data-* could assist that but realised it is probably not the right fit.

[19:14:01.0000] <csarven>
"User agents must not derive any implementation behavior from these attributes or values. Specifications intended for user agents must not define these attributes to have any meaningful values." -- I suppose this is to prevent authors from creating custom markup and relying on some UA behaviour.

[19:16:00.0000] <takkaria>
csarven: datetime-design-pattern is solved with <time>, no?

[19:16:01.0000] <csarven>
I don't know. Can you give me an example using <time>?

[19:18:00.0000] <csarven>
takkaria I see.

[19:18:01.0000] <takkaria>
csarven: <time datetime="2007-08-20T23:30Z">20 Aug</time>

[19:19:00.0000] <csarven>
Certainly.

[19:20:00.0000] <csarven>
As far as I can tell. The existing <abbr> approach would probably need to me replaced with <time>

[19:20:01.0000] <takkaria>
I believe that was one of the reasons <time> was invented

[19:22:00.0000] <Philip`>
As far as I'm aware, that was the only reason <time> exists

[19:23:00.0000] <csarven>
Then I'm not sure if data-* would at all be a use to microformats.

[19:24:00.0000] <Philip`>
If there was something equivalent to data-* but without the 'implementations must ignore these attributes' requirement, would that be useful to microformats?

[19:29:00.0000] <annevk>
csarven, I suggested to Hixie that data-* must not be used by UAs

[19:29:01.0000] <annevk>
csarven, this excludes UA extensions such as versioning and also excludes microformats

[19:29:02.0000] <annevk>
csarven, if microformats feel they need some attribute other than what they already got (such as a solution for time -> <time>) I think they can easily negotiate that with the HTML WG

[19:30:00.0000] <csarven>
Philip` Not necessarily because of the "visible" bit. data-* almost falls on meta-data.

[19:30:01.0000] <csarven>
annevk Right on.

[19:30:02.0000] <csarven>
<time> is perfect IMHO

[19:31:00.0000] <annevk>
Jacques Distler: "HTML5lib sucks." :-(

[19:32:00.0000] <annevk>
Oh, he's complaining about performance...

[19:33:00.0000] <annevk>
I think that's more or less a non-goal at this point. We need bindings to a C implementation to get that going strong...

[19:33:01.0000] <annevk>
/me adds a comment

[19:33:02.0000] <smedero>
Anne: are you talking about this post? http://golem.ph.utexas.edu/~distler/blog/archives/001689.html 

[19:36:00.0000] <annevk>
yeah

[21:12:00.0000] <andersca>
Hixie: what was the rationale for making the dynamic entries ordered? 

[21:28:00.0000] <hsivonen>
do C extensions to Ruby have to use UTF-8?

[21:29:00.0000] <hsivonen>
C extensions to Python should use UTF-16 or UTF-32, right?

[21:29:01.0000] <jgraham>
hsivonen: I think that's right (the python bit)

[21:30:00.0000] <jgraham>
http://docs.python.org/api/unicodeObjects.html

[21:30:01.0000] <hsivonen>
the difficult bit with recasting the V.nu tokenizer into UTF-8 C would be that REPLACEMENT CHARACTER would take more space than \0 and the tokenizer expects them to fit into the same space

[21:31:00.0000] <jgraham>
If that's the only difficult bit it doesn't sound too bad

[21:31:01.0000] <jgraham>
:)

[21:37:00.0000] <takkaria>
I should mention I'm being paid to work on a C html5 parser this summer

[21:37:01.0000] <hsivonen>
takkaria: what's your implementation strategy? from scratch? adapting an existing implementation?

[21:39:00.0000] <jgraham>
takkaria: I mentioned that on Jacques' blog

[21:39:01.0000] <jgraham>
Is it a Summer of Code thing?

[21:39:02.0000] <jgraham>
IIRC you are using lxml compatible interfaces?

[21:39:03.0000] <jgraham>
sorry libxml2

[21:43:00.0000] <takkaria>
yes, summer of code

[21:44:00.0000] <takkaria>
and there will be a libxml binding, yes

[21:44:01.0000] <takkaria>
hsivonen: well, jmb (on this channel) has started, so it's not entirely from scratch

[22:09:00.0000] <Hixie>
to anyone who sent feedback over the past few days, please send mail

[22:10:00.0000] <Hixie>
andersca: well it has to have some stable order for enumeration

[22:10:01.0000] <Hixie>
andersca: apart from that, i don't really cre

[22:10:02.0000] <andersca>
Hixie: OK

[00:18:00.0000] <Hixie>
/me commits <iframe seamless>

[00:21:00.0000] <hober>
Hixie: formatting error

[00:21:01.0000] <hober>
I see "he seamless/code> attr..."

[00:21:02.0000] <Hixie>
yeah i'm on it

[00:25:00.0000] <Hixie>
ok fixed those

[00:25:01.0000] <Hixie>
that was a mess :-)

[00:25:02.0000] <Hixie>
ok let's see

[00:26:00.0000] <Hixie>
i guess sandbox="" is next

[00:38:00.0000] <MikeSmith>
Hixie: small but important change:

[00:38:01.0000] <MikeSmith>
"The reversed attribute is a boolean attribute. If present, it indicates that the list is an ascending list. If the attribute is present, the list is a descending list."

[00:38:02.0000] <MikeSmith>
2nd sentence should start "If absent,"

[00:38:03.0000] <MikeSmith>
can send you mail about it if you want

[00:40:00.0000] <Hixie>
actually i'm between edits right now so i can change it

[00:40:01.0000] <Hixie>
but in general e-mail is definitely better

[00:40:02.0000] <Lachy>
Hixie, when the src of a seamless iframe is changed to a document that is not the same origin (either via the dom or <a href=... target=the-iframe>), is that expected to change from a seamless frame to a normal iframe?

[00:53:00.0000] <jwalden_>
http://pastebin.mozilla.org/436577

[00:53:01.0000] <jwalden_>
I'm not getting a message saying it'll appear after moderation or anything like that

[00:53:02.0000] <jwalden_>
so I don't know what's up with it

[00:54:00.0000] <jwalden_>
just redirects back to the main blog page

[00:56:00.0000] <Hixie>
/me tries to comment too

[00:57:00.0000] <Hixie>
wfm

[00:58:00.0000] <Hixie>
(i used safari 3)

[00:58:01.0000] <Hixie>
(trunk

[00:58:02.0000] <Hixie>
)

[00:59:00.0000] <Hixie>
one day i'll stop working on browsers and then i'll have to pick a preferred browser that i use everywhere, instead of switching browsers every 2 minutes...

[01:00:00.0000] <gavin_>
when you have to do something on the web that's not testing browsers, do you really just randomly pick one?

[01:01:00.0000] <gavin_>
you don't care about any of the stuff that browsers keep track of for you?

[01:04:00.0000] <Hixie>
gavin_: i use multiple machines and i nuke my profiles regularly, so no

[01:08:00.0000] <Hixie>
jwalden: wfm using safari trunk; might be a browser problem.

[01:08:01.0000] <Hixie>
jwalden: good comment btw

[01:08:02.0000] <jwalden>
Firefox 2?  that would be pernicious

[01:08:03.0000] <Hixie>
oh i assumed you were using ff trunk

[01:08:04.0000] <jwalden>
/me tries Safari 3.latest

[01:08:05.0000] <Hixie>
afk, brb

[01:08:06.0000] <jwalden>
I'm lame

[01:13:00.0000] <jwalden>
would have been nice for MSFT to point out their implementation isn't up to latest spec, but I'm guessing that's just their paranoia about saying what they're doing before they do it :-\

[01:14:00.0000] <jwalden>
or rather, to say that anyone who experiments should expect future breakage

[01:16:00.0000] <Hixie>
jwalden: when you're in an abusive relationship, and your spouse has been beating you for 15 years, flinching when they raise their hand to hug you isn't paranoia

[01:16:01.0000] <Hixie>
it's good sense

[01:17:00.0000] <MikeSmith>
jwalden: "Our policy gives our customers more predictability in their planning."

[01:17:01.0000] <MikeSmith>
http://blogs.msdn.com/ie/archive/2008/05/13/may-chat-with-the-ie-team-on-thursday.aspx#8502193

[01:18:00.0000] <Hixie>
MikeSmith: that answer is especially stupid given microsoft's policy of announcing everything years before they have implemented it

[01:18:01.0000] <MikeSmith>
predictability meaning, I guess, that their customers can consistently plan on them not giving any information about anything until they ship it

[01:18:02.0000] <MikeSmith>
yeah

[01:19:00.0000] <Hixie>
(my favourite example being winfs, which iirc was announced around when Windows 95 was in beta)

[01:19:01.0000] <Hixie>
(13 or 14 years ago)

[01:20:00.0000] <MikeSmith>
announce ahead when it's convenient and useful (even vaporware), don't announce anything else ahead unless it gives some potential market advantage to do so

[01:20:01.0000] <jwalden>
I think there's a double standard here, tho, of announcing something (so committing to it being there) before it's ready and not announcing it until it's complete

[01:20:02.0000] <roc>
Hixie, I think you're mixing up WinFS with the "Cairo" "OLE file system"

[01:20:03.0000] <roc>
I think they cancelled the project at least once in between :-)

[01:21:00.0000] <Hixie>
roc: right

[01:21:01.0000] <Hixie>
roc: i just mean their "next generation" file system to come after NTFS which is somehow "revolutionary"

[01:22:00.0000] <Hixie>
the actual project has, as you say, been announced and canceled and reannounced many times since the original announcement

[01:22:01.0000] <MikeSmith>
I think in part that policy is just a general kneejerk policy of unnecessary secrecy -- of keeping everything in development (across all their product lines) secret by default

[01:22:02.0000] <Hixie>
i really do think the abusive relationship analogy is appropriate here, btw, right down to people being willing to forgive them any time they claim to have turned over a new leaf

[01:22:03.0000] <MikeSmith>
it reminds me of the Bush administration secrecy policy

[01:23:00.0000] <MikeSmith>
which has resulted in more stuff being made secret in the last eight years than in the previous 40 or something years combined

[01:23:01.0000] <Hixie>
(and right down to the feeling i have of watching the relationship -- i've seen abusive relationships and it's pretty much exactly the same feeling)

[01:23:02.0000] <Hixie>
MikeSmith: but the point is they DON'T keep everything secret!

[01:24:00.0000] <MikeSmith>
yeah, same as Bush admin

[01:24:01.0000] <MikeSmith>
they leak stuff when they want to

[01:24:02.0000] <roc>
Silverlight is the best current example of this

[01:24:03.0000] <MikeSmith>
like the "Valerie Plame is a CIA agent" thing

[01:24:04.0000] <jwalden>
everybody does that, a way of saying something without saying it

[01:24:05.0000] <roc>
its feature set has been announced for years and it hasn't even shipped yet

[01:24:06.0000] <jwalden>
it's kinda smarmy, but it's common practice

[01:25:00.0000] <Hixie>
i thought silverlight 2 had shipped

[01:25:01.0000] <Hixie>
am i falling prey to the fud too?

[01:25:02.0000] <roc>
don't think so

[01:25:03.0000] <Hixie>
so what is it that microsoft.com asks me to install every time i go there?

[01:25:04.0000] <Hixie>
oh sl2 is in beta

[01:25:05.0000] <Hixie>
ok

[01:25:06.0000] <MikeSmith>
while I can see a policy like that making sense for their other products -- or for other classes of products in general -- I think it makes a lot less sense for IE, or for a browser product in general

[01:26:00.0000] <MikeSmith>
because for one thing they're not selling it, nor is anybody else selling a desktop browser

[01:37:00.0000] <MikeSmith>
Hixie: btw, you got me in the Acks twice, as "Michael Smith" and "Mike Smith".. not sure I merit being in there at all (since the stuff I've commented on has mostly just been minor editorial things) but if you do have me in there, "Michael(tm) Smith" would be preferred (to make it clear which Michael Smith among the nation of Michael Smiths it is)

[01:39:00.0000] <Hixie>
(tm) or &trade; ? :-)

[01:40:00.0000] <Lachy>
MikeSmith, what's the purpose of the (tm) in your name?

[01:41:00.0000] <Hixie>
he just said what the purpose was

[01:41:01.0000] <Hixie>
"to make it clear which Michael  Smith among the nation of Michael Smiths it is

[01:41:02.0000] <Hixie>
"

[01:42:00.0000] <Lachy>
ok, but why use TM? Is it your middle initials or something?

[01:42:01.0000] <Hixie>
"trademark"

[01:42:02.0000] <Hixie>
/me upgrades Lachy's sense of humour to the latest security update :-P

[01:43:00.0000] <Lachy>
yeah, I get that it means trademark. But I don't get the relevance

[01:44:00.0000] <Hixie>
it's irony

[01:44:01.0000] <Hixie>
but i don't know how to explain it

[01:45:00.0000] <Hixie>
like, "I'm going to go and eat now, because I'm Hungry(tm)"

[01:45:01.0000] <MikeSmith>
Lachy: what Hixie said :)

[01:47:00.0000] <Lachy>
ok, so I guess there's no logical or rational explanation for it and I just need to accept it :-(

[01:49:00.0000] <Hixie>
it's irony because his name is so common, yet he claims it as a trademark

[01:49:01.0000] <Hixie>
it's perfectly rational, but i don't know how to explain it :-)

[01:49:02.0000] <Hixie>
MikeSmith: spec is updated btw

[01:50:00.0000] <Lachy>
ok, that's better explanation.

[01:53:00.0000] <MikeSmith>
Hixie: cheers

[01:53:01.0000] <MikeSmith>
saw your checkins

[01:56:00.0000] <Lachy>
/me wonders why he didn't receive twitter updates for the recent spec checkins

[01:56:01.0000] <Hixie>
did twitter get them?

[01:56:02.0000] <Lachy>
yeah, the iframe seamless one is there, but I didn't receive it

[01:58:00.0000] <Lachy>
ah, it's twitters crappy jabber implementation that stops sending me twitter updates sometimes

[02:01:00.0000] <Hixie>
i haven't been able to get the jabber to work, like, ever

[02:01:01.0000] <Lachy>
what do you use to receive updates?

[02:02:00.0000] <Philip`>
/me would have assumed Hixie already knew about updates to the spec without having to get Twitter notifications :-p

[02:02:01.0000] <Hixie>
Lachy: i don't follow twitter

[02:03:00.0000] <Hixie>
i only use it to get direct twitters when i'm expecting them (and then i just look at the site, half the time i don't get e-mail notifications)

[02:04:00.0000] <Hixie>
other than that, for me, it's a write-only medium

[02:09:00.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/#xxxoriginpleasesuggestabettername

[02:09:01.0000] <Hixie>
just for the record

[02:09:02.0000] <Hixie>
that is not a proposal for the name

[02:09:03.0000] <Hixie>
and if anyone implements that keyword i will personally hunt them down

[02:24:00.0000] <Lachy>
Hixie, add sandbox to "The DOM attributes src, name, and seamless must reflect the content attributes of the same name. "

[02:26:00.0000] <Hixie>
will do

[02:26:01.0000] <Hixie>
haven't got that far yet

[02:26:02.0000] <Lachy>
ok.

[02:26:03.0000] <Lachy>
will there be a DOMTokenList attribute for the sandbox?

[02:28:00.0000] <roc>
it would be useful if the spec document was published in multiple versions with different stability

[02:29:00.0000] <Hixie>
roc: hm, interesting idea

[02:29:01.0000] <Hixie>
roc: how would that work?

[02:29:02.0000] <roc>
not sure

[02:29:03.0000] <jwalden>
html5-mm

[02:29:04.0000] <roc>
but right now, for authors and implementors its hard to tell what is set in stone and what is likely to change

[02:29:05.0000] <Lachy>
roc, why would that be useful?

[02:29:06.0000] <Lachy>
oh

[02:30:00.0000] <Hixie>
roc: well, the sections do have annotations

[02:30:01.0000] <Hixie>
roc: which should make it relatively easy

[02:30:02.0000] <roc>
and what is likely to get them hunted down :-)

[02:30:03.0000] <Hixie>
roc: but the presentation could do with some work, true

[02:40:00.0000] <roc>
not all the sections have annotations

[02:41:00.0000] <roc>
and it's hard to see what is covered by each annotation, especially if you happen to jump to the middle of the document and there's no annotation visible. I guess that's what you mean by the presentation needing work.

[02:45:00.0000] <Hixie>
roc: yeah

[02:45:01.0000] <Hixie>
roc: i'm slowly adding annotations

[02:45:02.0000] <Hixie>
roc: any help in either that or making the presentation better would be welcome

[02:49:00.0000] <jwalden>
Hixie: what's the "tests: ##" thing supposed to mean in the annotation boxes?  some whatwg/w3-maintained test suite or something?

[02:49:01.0000] <Hixie>
links (one URI per line) to any tests you know of

[02:50:00.0000] <jwalden>
ah

[02:52:00.0000] <jwalden>
I suppose that doesn't work so well for cross-domain tests

[02:53:00.0000] <MikeSmith>
Hixie: the XXX convention in the spec is something you consistently use to indicates a placeholder ("we're not done here yet") thing?

[02:53:01.0000] <Hixie>
jwalden: ?

[02:53:02.0000] <Hixie>
MikeSmith: yes

[02:53:03.0000] <Hixie>
the blue line on the chart shows the count of XXX markers

[02:53:04.0000] <MikeSmith>
aha

[02:54:00.0000] <MikeSmith>
wondered what that blue line was

[02:54:01.0000] <MikeSmith>
/me is remined of Errol Morris film "The Thin Blue Line"

[02:54:02.0000] <jwalden>
er, yeah, I suppose you could set up tests on a bunch of real servers

[02:54:03.0000] <jwalden>
would be a pain, but you could do it, I guess

[02:56:00.0000] <Hixie>
not that much of a pain

[02:56:01.0000] <Hixie>
i do it for a bunch of tests on hixie.ch

[02:56:02.0000] <Hixie>
they use another.domain.libpr0n.com as the other domain

[02:56:03.0000] <Hixie>
just one symlink

[02:57:00.0000] <Hixie>
for the record, browsers all suck

[02:57:01.0000] <Hixie>
i just wasted like 30 minutes trying to debug something in safari using the safari web developer tools

[02:57:02.0000] <Hixie>
which frankly are crappy compared to firebug and other things available for firefox

[02:57:03.0000] <Hixie>
but god, safari loads the html5 spec so much faster than firefox

[03:01:00.0000] <Hixie>
(safari also reloads way more than necessary when all i do is change a fragment identifier, and doesn't reload the tabs when it crashes)

[03:01:01.0000] <Hixie>
(and occasionally it seems to get wedged in a state where it won't download any pages)

[03:02:00.0000] <Hixie>
but it feels much lighter than firefox

[03:02:01.0000] <Hixie>
at least on mac

[03:54:00.0000] <Hixie>
bbl

[03:58:00.0000] <MikeSmith>
I'm trying to figure out what the main use cases are the the Opera File I/O proposal would enable

[03:59:00.0000] <MikeSmith>
I mean what specific feature it would bring to users that existing applications can't

[03:59:01.0000] <MikeSmith>
features

[09:45:00.0000] <annevk>
Hixie, "an descending"

[10:13:00.0000] <annevk>
Hixie, seamless is not a boolean on the IDL

[11:05:00.0000] <hendry>
i need to draw a graph. what to library to use. seems to be so many nowadays.

[12:07:00.0000] <Philip`>
hendry: Gnuplot!

[12:08:00.0000] <Philip`>
/me has almost figured out how to use that, so he doesn't want to waste the investment by switching to anything else

[12:16:00.0000] <Lachy>
hendry, it depends what you need to draw the graph for and where you intend to use it.

[12:16:01.0000] <MikeSmith>
/me reads jwalden comment on IE Blog posting about cross-document messaging

[12:16:02.0000] <MikeSmith>
http://blogs.msdn.com/ie/archive/2008/05/19/enabling-mash-ups-in-internet-explorer-8-with-cross-document-messaging.aspx#8520209

[12:23:00.0000] <MikeSmith>
by way of summary at Ajaxian:

[12:23:01.0000] <MikeSmith>
http://ajaxian.com/archives/ie-8-and-cross-document-messaging

[12:24:00.0000] <MikeSmith>
/me wonders how Dion Almaer makes time to code and still blog & tweet as much as he does

[12:50:00.0000] <hendry>
Lachy: on the Web with Firefox 3. need to plot a simple approximate 2d graph to demonstrate an idea. I guess inkscape is probably best. I love xfig but it doesn't do SVG I guess.

[12:52:00.0000] <Lachy>
hendry, yeah, you could use SVG or <canvas>.

[12:52:01.0000] <Lachy>
If you want it generated dynamcially from a table of data, you could use PlotKit

[12:52:02.0000] <Philip`>
Or PNG

[12:53:00.0000] <Philip`>
Avoid overcomplication :-)

[13:15:00.0000] <hsivonen>
http://joeclark.org/access/webaccess/WCAG/response1_Understanding-WCAG2.html interesting points

[13:30:00.0000] <Dashiva>
hsivonen: There's also the whole "wcag process isn't accessible" thing that nobody wants to talk about :)

[13:31:00.0000] <Philip`>
"If data were understandable by themselves, we wouldn’t make a chart." seems untrue - we'd make a chart if the chart was more understandable than the data, even if the data was already understandable

[13:32:00.0000] <Lachy>
"A slider? That requires a mouse, not to mention full vision. " - How does a slider require a mouse? It can be used by keyboard users if implemented correctly

[13:32:01.0000] <hsivonen>
Dashiva: I don't have a diagnosed learning disability, but I find WCAG2 and ATAG2 hard to follow, because they are so abstract and then lawyerly on top of that

[13:32:02.0000] <Lachy>
maybe that was true when the document was written, and there was no way to add a slider to a web page accessibly

[13:33:00.0000] <hsivonen>
Lachy: it's from 2006

[13:33:01.0000] <Lachy>
ok

[13:39:00.0000] <Dashiva>
I thing I've been wondering is if (e.g.) red-green colorblind people are more likely to think "red looks like green", "green looks like red", "red and green are different, but I can't tell them apart" or "red and green are two names for the same color"

[14:02:00.0000] <MikeSmith>
/me notes Updated: 2006.05.23 at the bottom of that article

[14:33:00.0000] <Lachy>
http://css-tricks.com/html-5-vs-xhtml-2-an-article-roundup-and-poll/

[14:35:00.0000] <Lachy>
also note the poll at the bottom of the right column. XHTML2 (33%, 33 votes) is ahead of HTML5 (18%, 17 votes)

[14:40:00.0000] <Philip`>
Lachy: The people have spoken - HTML5 is doomed :-(

[14:42:00.0000] <Lachy>
indeed. I'm going to quit now and join the XHTML2 group before it's too late.

[14:52:00.0000] <Dashiva>
Darn, I wish someone told me the future of the web was decided on a first-come, first-vote, 50-votes-only basis

[14:53:00.0000] <hsivonen>
Dashiva: cf. The Workshop :-)

[15:39:00.0000] <Lachy>
Dashiva, there were more votes. But most people voted don't care.


2008-05-21
[18:08:00.0000] <MikeSmith>
mcarter_: have you done a write-up anywhere yet on your implementation related to the connection API?

[18:10:00.0000] <takkaria>
PlotKit is interesting, but also lacking in a bunch of features

[18:10:01.0000] <takkaria>
for example, vertical gridlines

[19:28:00.0000] <Hixie>
annevk: please send feedback to the list, not the irc channel :-)

[19:43:00.0000] <gsnedders>
This internet access saga is truly bizarre (but it works again now, so hey)

[19:44:00.0000] <annevk>
Hixie, ok, e-mailed you + www-archive

[19:50:00.0000] <gsnedders>
It feels weird sending emails to Hixie without www-archive :P

[20:55:00.0000] <Hixie>
annevk: thanks

[21:34:00.0000] <annevk>
geez

[21:34:01.0000] <annevk>
twitter is slower than ever (besides when it's unresponsive)

[21:59:00.0000] <Lachy>
Woo Hoo! This book is finally available for pre-order. http://www.amazon.com/Ultimate-HTML-Reference-Ian-Lloyd/dp/0980285887/

[21:59:01.0000] <Lachy>
I was one of the technical reviewers for it as it was being written

[22:04:00.0000] <Philip`>
Lachy: What makes it ultimater than the thousands of other HTML books? :-)

[22:05:00.0000] <Lachy>
cause I reviewed it and know it's good :-)

[22:05:01.0000] <Philip`>
That kind of appeal to authority is not very convincing :-p

[22:07:00.0000] <Lachy>
For one reason, unlike many other books that teach people to use XHTML, this one actually teaches both HTML and XHTML and gives good advice about the use of both.

[22:08:00.0000] <Lachy>
there's a public beta online here http://reference.sitepoint.com/html

[22:10:00.0000] <Lachy>
hmm. It seems to be missing entity references.

[22:10:01.0000] <Lachy>
oh well, that can be added to the next revision

[22:19:00.0000] <Lachy>
Hixie, yt?

[22:21:00.0000] <Lachy>
Hixie, the entity reference table in the spec is missing the <tbody> after the first row

[22:23:00.0000] <Hixie>
w3c or whatwg?

[22:25:00.0000] <Lachy>
whatwg

[22:25:01.0000] <Lachy>
I didn't check the w3c copy of the spec

[22:25:02.0000] <Lachy>
I only noticed in the whatwg copy cause I was copying the table for use in the html authoring guide

[22:25:03.0000] <annevk>
w3c too iirc

[22:25:04.0000] <Lachy>
I'm going to rewrite it to make it more author friendly

[22:26:00.0000] <Hixie>
oh, i see what you're saying

[22:26:01.0000] <Hixie>
odd

[22:26:02.0000] <Hixie>
if possible, can you send mail?

[22:26:03.0000] <Hixie>
i'm in the middle of a painful edit

[22:26:04.0000] <Lachy>
sure

[22:26:05.0000] <Hixie>
thanks

[22:30:00.0000] <Lachy>
Hixie, do you use a script to generate that table?

[22:30:01.0000] <Hixie>
yes

[22:30:02.0000] <Lachy>
Can I get a copy of it, so I can modify it to generate a more appropriate table for the author guide.

[22:31:00.0000] <Hixie>
it's one of my first ever python scripts, so i dunno :-)

[22:31:01.0000] <Lachy>
LOL

[22:32:00.0000] <Hixie>
http://junkyard.damowmow.com/322

[22:32:01.0000] <Hixie>
that script is then called using this bash magic:

[22:32:02.0000] <Hixie>
wget -o /dev/null -N http://www.w3.org/2003/entities/2007xml/unicode.xml

[22:32:03.0000] <Hixie>
[ unicode.xml -nt entities-unicode.inc ] && nice -n 19 python .entity-processor.py > .new-entities-unicode.inc

[22:32:04.0000] <Hixie>
[ -s .new-entities-unicode.inc ] && mv -f .new-entities-unicode.inc entities-unicode.inc

[22:32:05.0000] <Hixie>
cat entities-* | sort | uniq > entities.inc

[22:32:06.0000] <Lachy>
thanks

[22:33:00.0000] <Hixie>
and entities-* matches the entities-unicode.inc file mentioned above and entities-legacy.inc

[22:33:01.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/entities-legacy.inc

[22:33:02.0000] <Hixie>
which is a hard-coded list of entities

[22:35:00.0000] <annevk>
Hixie, doing more work on sandboxing?

[22:35:01.0000] <Hixie>
yeah

[22:35:02.0000] <Hixie>
the http://www.whatwg.org/specs/web-apps/current-work/index-diff file shows what i have so far

[22:39:00.0000] <annevk>
Hixie, for seamless it should probably say something about initial containing block and such too

[22:39:01.0000] <Hixie>
othermaciej: the problem with making item() available, etc, is that you lose the ability to use the data-item="" attribute

[22:39:02.0000] <Hixie>
it does

[22:39:03.0000] <Hixie>
unless i messed up the commit

[22:40:00.0000] <Hixie>
search for "initial containing block" in html5 and there are two occurances, both doing what you just asked for :-)

[22:40:01.0000] <annevk>
it sets the height tozero...

[22:40:02.0000] <Hixie>
right

[22:41:00.0000] <annevk>
it doesn't entirely remove it

[22:41:01.0000] <annevk>
hmm

[22:41:02.0000] <Hixie>
you can't remove it

[22:41:03.0000] <Hixie>
that would break css in pretty fundamental ways

[22:41:04.0000] <Hixie>
it's like saying we should remove the root element

[22:41:05.0000] <Hixie>
what would it even mean?

[22:41:06.0000] <annevk>
it means that the initial containing block of the nested browsing context is that of the browsing context

[22:41:07.0000] <annevk>
well, that's what I thought it would specify

[22:42:00.0000] <othermaciej>
Hixie: yeah, I can see why it could cause conflicts

[22:42:01.0000] <Hixie>
annevk: wow, that would be weird

[22:43:00.0000] <annevk>
now absolutely positioning stuff still goes against the nested browsing context initial containing block, which is not what you want (not always anyway)

[22:43:01.0000] <Hixie>
why wouldn't it be what you want?

[22:44:00.0000] <Hixie>
it's still an iframe

[22:44:01.0000] <Hixie>
you can't render outside the iframe

[22:44:02.0000] <Hixie>
if you could you'd be open to all kinds of weird hacks

[22:44:03.0000] <annevk>
if you could you'd have the client-side includes people have been looking for

[22:45:00.0000] <Hixie>
i don't think client side include requests have really been looking for ways to render outside the box

[22:45:01.0000] <Hixie>
they mostly just want to include their navigation links or header or footer

[22:50:00.0000] <annevk>
btw, in the short definition for the sandbox attribute you should note that giving it values means you lift restrictions

[22:51:00.0000] <roc>
I prefer what Hixie currently has

[22:52:00.0000] <roc>
we don't want to make handling of the IFRAMEd content too weird

[22:52:01.0000] <annevk>
Hixie, maybe split "xxx_origin_please_suggest_a_better_name" into "dom-access" and "storage-access"

[23:02:00.0000] <Hixie>
annevk: sandbox="dom-access" doesn't convey what it does, since even with sandbox="", DOM access is still possible, just not from other documents, even if their URIs are in the same origin

[23:02:01.0000] <Hixie>
annevk: and storage access (and XHR, etc) is allowed if scripting is enabled, it just doesn't work (since the document isn't in the expected origin)

[23:04:00.0000] <annevk>
localStorage would actually still work and end up in void, or not?

[23:04:01.0000] <annevk>
maybe just "default-origin" or some such then?

[23:06:00.0000] <annevk>
Hixie, btw, these two comments make it much more clear for some reason what the attribute is supposed to do then the note in the spec

[23:06:01.0000] <Hixie>
yeah i dunno how to explain it

[23:08:00.0000] <annevk>
maybe localStorage should fail if the origin is an identifier and not a tuple

[23:12:00.0000] <annevk>
hmm, StorageEvent.uri should be StorageEvent.url ...

[23:14:00.0000] <roc>
one day I'll have to figure out what the difference between a URL and a URI is. Wouldn't want to embarrass myself in public

[23:14:01.0000] <Dashiva>
Don't forget IRI

[23:14:02.0000] <roc>
oh yeah, I discovered those recently. Mad

[23:14:03.0000] <hober>
and URN

[23:16:00.0000] <Philip`>
"URI" is the term you use when talking to standards people so you look clever, and "URL" is the term you use when talking to normal people so you're understood

[23:17:00.0000] <andersca>
lol

[23:19:00.0000] <annevk>
<input type=url> was at one point type=uri...

[23:20:00.0000] <Hixie>
i plan to change the spec to use the term "url"

[23:20:01.0000] <Hixie>
and just define "url" in the terminology section as being an iri

[23:21:00.0000] <Hixie>
i expect this will cause the nitpickers to have enough heart attacks to keep them off our back for a few months

[23:21:01.0000] <Hixie>
(actually, not just an iri. it will depend on the character encoding in weird ways.)

[23:21:02.0000] <Dashiva>
Hixie: Also define uri as url, then

[23:22:00.0000] <annevk>
hmm, URL5 :)

[23:22:01.0000] <Hixie>
Dashiva: what does that mean?

[23:28:00.0000] <Dashiva>
Hixie: Just in case there's a URI left in there somewhere, it should be defined :)

[23:30:00.0000] <annevk>
othermaciej, there's a difference between style.foo = x and storage.foo = x

[23:30:01.0000] <annevk>
othermaciej, for style.foo to work foo has to bedefined on the CSS2Properties interface

[23:31:00.0000] <annevk>
othermaciej, storage.foo will always work (though maybe not when foo matches some existing member)

[23:32:00.0000] <Hixie>
Dashiva: it's easy to ensure there isn't :-)

[23:32:01.0000] <Hixie>
/me is utterly tangled up in a mess of defining scripting enabled/disabled on a per-browsing-context basis instead of globally

[23:33:00.0000] <othermaciej>
annevk: sure, it is a predefined list instead of completely open-ended, but CSS styles do have the concept of adding and removing properties, and those properties are reflected directly as JS-level properties

[23:33:01.0000] <Hixie>
i agree with anne that it's not the same

[23:33:02.0000] <Hixie>
"'opacity' in style" is always true

[23:33:03.0000] <Hixie>
it doesn't become false when the property is removed

[23:34:00.0000] <roc>
"removeProperty" is a misnomer from the JS point of view

[23:34:01.0000] <Hixie>
i'm not wedded to the 'delete' syntax, but i think your argument that it's the wrong semantic is not valid

[23:34:02.0000] <othermaciej>
the style declaration example is not to illustrate that it is the wrong semantic

[23:35:00.0000] <othermaciej>
it is to illustrate that, in general, you can reasonably have custom getters and setters without custom deleters

[23:35:01.0000] <othermaciej>
and this will not confuse people

[23:35:02.0000] <Hixie>
i don't think it would confuse people

[23:35:03.0000] <othermaciej>
that was roc's argument

[23:35:04.0000] <othermaciej>
that if you don't have custom delete, you shouldn't even have custom get and set

[23:35:05.0000] <Hixie>
i just think that in this particular case, it makes sense to offer custom delete

[23:36:00.0000] <roc>
it's custom *add* that I really think deserves a matching custom delete

[23:36:01.0000] <Dashiva>
I assume setting it to undefined to delete was considered and rejected, what was the reasoning?

[23:36:02.0000] <Hixie>
setting it to "undefined" would set it to the string "undefined"

[23:36:03.0000] <othermaciej>
set and add have the same syntax in JS

[23:36:04.0000] <Hixie>
as currently defined

[23:37:00.0000] <Hixie>
othermaciej: style.foo = 'baz' is matched by delete style.foo, for non-valid values of 'foo'

[23:37:01.0000] <roc>
of course they have the same syntax, that's part of the problem ehre

[23:38:00.0000] <othermaciej>
Hixie: or indeed, for any object it is true that when the setter does nothing special, delete works normally

[23:39:00.0000] <Hixie>
right

[23:39:01.0000] <Hixie>
so why is this different?

[23:39:02.0000] <othermaciej>
well, it's not usually the case that when the setter *does* do something special, delete still works normally

[23:40:00.0000] <othermaciej>
(or beyond that, delete does something correspondingly special)

[23:40:01.0000] <Hixie>
i don't think we've ever really had an equivalent interface before

[23:41:00.0000] <Hixie>
except maybe NamedNodeMap, which imho is poorly designed

[23:41:01.0000] <Hixie>
but like i said, i don't really care much about this

[23:42:00.0000] <othermaciej>
I don't care that much about the semantic argument, except that I think having custom delete is a small improvement, and supporting it correctly is potentially a nontrivial burden

[23:42:01.0000] <Dashiva>
Are there any common cases where you'd be deleting stuff often?

[23:43:00.0000] <Dashiva>
I figure if you're setting a property for storage you're likely to keep it. Maybe it'll be unused (e.g. set to nothing) in periods, but the use for the property isn't going to disappear

[23:43:01.0000] <othermaciej>
I will admit to being a little biased, in that the crazy dynamicity of JS makes me want to cling to any shred of statically determined behavior

[23:43:02.0000] <othermaciej>
I think delete is not going to be nearly as common as set/add or get, but it seems plausible

[23:44:00.0000] <othermaciej>
(in the style case too I would guess you set a lot more often than you remove)

[23:44:01.0000] <Hixie>
well, the ie/ff argument is a strong one -- what we really should be asking is whether the ff team and the ie team plan to remove the custom getting when they switch from globalStorage to localStorage

[23:44:02.0000] <roc>
I don't really know why custom delete would be a big burden if you already support custom add

[23:44:03.0000] <Dashiva>
Because add is actually set?

[23:44:04.0000] <roc>
in syntax it is

[23:45:00.0000] <othermaciej>
custom code running at set time is something we already have to live with

[23:45:01.0000] <roc>
but implementation-wise, not really

[23:45:02.0000] <roc>
or not necessarily

[23:45:03.0000] <othermaciej>
both due to DOM APIs and due to supporting the Mozilla getter/setter extension to ECMAScript

[23:45:04.0000] <roc>
BTW I almost always see style.opacity = ''; rather than removeProperty

[23:46:00.0000] <Dashiva>
Because they're equivalent in practice

[23:46:01.0000] <jgraham>
Dashiva: It seems like there would be lots of cases where you wanted to persist some property for a certian amount of time (e.g. till the network becomes available) and then stop persisting it

[23:46:02.0000] <Dashiva>
jgraham: Sure, and later when the network becomes unavailable again, you'd want to persist that same property with a new value

[23:46:03.0000] <othermaciej>
custom code at delete time (in particular, JS executing as a result of the event dispatch) would be new

[23:46:04.0000] <othermaciej>
it's not impossible to handle

[23:47:00.0000] <othermaciej>
but every additional bit of operator overloading makes the language harder to optimize

[23:47:01.0000] <jgraham>
Dashiva: In that sort of case I would tend to delete stuff rather than set it to some magic value (e.g. the empty string)

[23:47:02.0000] <Dashiva>
jgraham: Because the case is very rare, right?

[23:48:00.0000] <roc>
I don't think 'delete' is going to make any material difference to the optimizability of JS

[23:48:01.0000] <jgraham>
No just because it feels like cleaning up your storage area is the right thing to do

[23:48:02.0000] <Dashiva>
Yes, and you want to clean up because you don't expect to use it in the near future

[23:49:00.0000] <jgraham>
I guess that depends what near future means

[23:49:01.0000] <Dashiva>
What I'm trying to claim is that if you use a property often enough that you'd benefit from delete shorthand, you're using it often enough that you're likely to be just nullifying the value instead

[23:51:00.0000] <othermaciej>
roc: every statement that can run arbitrary code invalidates many assumptions

[23:51:01.0000] <othermaciej>
currently when you see this code:

[23:51:02.0000] <othermaciej>
var x = 1;

[23:51:03.0000] <roc>
since 'set' and 'get' are already in that class

[23:51:04.0000] <othermaciej>
delete z.y;

[23:51:05.0000] <othermaciej>
you can assume x has the same value afterwards

[23:51:06.0000] <roc>
that game is already over

[23:51:07.0000] <Hixie>
othermaciej: no you can't

[23:51:08.0000] <Hixie>
othermaciej: z.y might run a getter

[23:52:00.0000] <Dashiva>
(Only on z, right?)

[23:52:01.0000] <roc>
I think you're going to have to stretch really hard to find a realistic testcase where optimizing around deletes is important and there are no gets/sets in the way

[23:53:00.0000] <jgraham>
Dashiva: I don't have a strong opinion on that. I guess my Python bias makes me think that delete is a reasonable syntax for this case but it's dangerous to apply reasoning from one language in another

[23:53:01.0000] <othermaciej>
Hixie: ok, if z is a local variable or parameter you can

[23:53:02.0000] <roc>
even if you come up with one, the problem is still completely dwarfed by gets/sets, so I still find it hard to care

[23:54:00.0000] <othermaciej>
roc: sure, I don't think it will be common, but the other crazy dynamic things in the language suck so much that on principle I am against additional overloading

[23:54:01.0000] <Hixie>
othermaciej: well, get IE and FF (or at least one of them) to not support delete when they switch to localStorage

[23:55:00.0000] <othermaciej>
I don't really want to engage in a majority vote power struggle, I just wanted to express the downsides of overloading the delete operator (particularly in a way that runs JS code)

[23:55:01.0000] <Hixie>
othermaciej: (right now you can't tell from the spec that 'delete' is supposed to do anything, it's just an [XXX] marker)

[23:55:02.0000] <roc>
I appreciate your principles but it really is too late to apply them to JS

[23:55:03.0000] <othermaciej>
it's not as bad as overloading * or && would be

[23:55:04.0000] <Lachy>
how do I see the output from the assert statements in python? Is there some debug mode I need to turn on or something?

[23:56:00.0000] <othermaciej>
roc: it is certainly possible to infer useful knowledge from some of the operators that are still not overloaded

[23:56:01.0000] <jgraham>
Lachy: You don't unless the expression evaluates to False

[23:56:02.0000] <Hixie>
othermaciej: my point is that even if i agreed with you, i wouldn't be able to do anything about it at this point really

[23:56:03.0000] <Hixie>
othermaciej: since the spec doesn't actually explicitly mention this at all right now

[23:56:04.0000] <Lachy>
oh, ok. The python docs aren't clear about that at all

[23:56:05.0000] <roc>
othermaciej: yeah, absolutely

[23:57:00.0000] <jgraham>
if they eval to False you'll get an AssertionError raised

[23:57:01.0000] <Hixie>
othermaciej: (also, note that overloading [[Delete]] is explicitly supported by the ES spec, unlike overloading the * or && operators)

[23:57:02.0000] <roc>
other areas of the language aren't as goofy as properties

[23:57:03.0000] <othermaciej>
(I would again point to * and && as examples, boy would it suck if those depended on the input types)

[23:57:04.0000] <roc>
well

[23:57:05.0000] <roc>
* already does, if you're optimizing hard

[23:59:00.0000] <Hixie>
/me starts seeing the light at the end of his scripting-disabled-per-browsing-context tunnel of doom

[00:00:00.0000] <Hixie>
hm

[00:00:01.0000] <Hixie>
i wonder if parsing changes if i set designMode

[00:00:02.0000] <othermaciej>
for * you know the result is a number and you know it always converts to number and multiplies as number - skipping conversions or optimizing for a subset of numbers based on additional knowledge can indeed help, but it is not as bad as +

[00:01:00.0000] <othermaciej>
Hixie: I have made my case, I am not really inclined to lobby implementors beyond that

[00:01:01.0000] <othermaciej>
(I would reserve that for more important issues)

[00:01:02.0000] <Dashiva>
othermaciej: Implement it yourself then ;)

[00:02:00.0000] <roc>
obviously we should have a new mozDelete JS operator

[00:02:01.0000] <Hixie>
othermaciej: at least one of the vendors hasn't even heard your case (microsoft aren't on the whatwg list)

[00:02:02.0000] <othermaciej>
Dashiva: WebKit already implements localStorage

[00:02:03.0000] <annevk>
/me gets a steady stream of html4all e-mails written by RB

[00:02:04.0000] <Hixie>
othermaciej: so i'm not sure what you're expecting to have happen

[00:02:05.0000] <annevk>
/me thinks it's time to go to bed

[00:03:00.0000] <Dashiva>
othermaciej: Hmm. It's hard to not implement delete any more than you already have, then.

[00:04:00.0000] <othermaciej>
Brady actually implemented delete but I'd like to encourage him to remove it

[00:05:00.0000] <Hixie>
sigh, ie8 throws an exception saying you can't get document.designMode if you try to set it (not get it) while parsing

[00:05:01.0000] <othermaciej>
Hixie: I am not expecting anything, I am just declining the action item of trying to persuade other implementors to do or not do anything specific when they implement localStorage (beyond stating the facts)

[00:05:02.0000] <takkaria>
RB has an interesting idea of use-cases

[00:05:03.0000] <Hixie>
othermaciej: ok

[00:06:00.0000] <othermaciej>
Hixie: but I can forward the thread to public-html if you think that would be helpful

[00:08:00.0000] <Hixie>
othermaciej: i don't really know what would be helpful. the spec will have to require [Delete] support of the browsers support it, though

[00:08:01.0000] <Hixie>
othermaciej: so at the end of the day, it _is_ a majority vote

[00:09:00.0000] <Dashiva>
/me wonders about the IE extensions to storage

[00:23:00.0000] <othermaciej>
Hixie: honestly I am more worried about whether Firefox and IE will transition to localStorage at all, or if we'll end up having to also implement globalStorage in WebKit

[00:24:00.0000] <othermaciej>
hopefully IE can since they are still just a beta but who knows?

[00:24:01.0000] <Hixie>
i hope ff does, since they were the reason we changed the api

[00:25:00.0000] <Hixie>
i have little hope of anything convenient to the health of the web coming from the IE team

[00:25:01.0000] <Hixie>
i mean, they couldn't even implement globalStorage without embrace-and-extending it

[00:25:02.0000] <Hixie>
even if they don't mean to, they seem to just act that way out of habbit

[00:27:00.0000] <othermaciej>
it does seem kind of habitual

[00:40:00.0000] <Dashiva>
othermaciej: Should it really say prefix twice? "This class is designed to carry a prefix but match exclusively on the prefix and namespaceURI"

[01:39:00.0000] <Philip`>
Hixie: Is it meant to be non-conforming if scripts cause an invalid state at any point during their execution, or only when they've finished? (e.g. if you do "img = document.createElement('img'); document.body.appendChild(img); img.src = '...'" is it non-conforming because there's (temporarily) an invalid state (with a srcless img)?)

[01:40:00.0000] <Hixie>
uh let me fix that hold on

[01:44:00.0000] <MikeSmith>
/me reads starts a read-through of r1643 <iframe sandbox> diff

[01:45:00.0000] <Philip`>
Good luck with that :-)

[01:45:01.0000] <Philip`>
/me just skimmed it to see if there were any obvious changes like an ASCII art elephant graphic

[01:48:00.0000] <Hixie>
Philip`: fixed

[01:51:00.0000] <Philip`>
Hixie: Thanks

[01:52:00.0000] <MikeSmith>
/me notes that http://dev.w3.org/cvsweb/html5/spec/Overview.html.diff?r1=1.830&r2=1.831&f=h provides side-by-side colored diff for those who like such things (I like the html5.org web-apps tracker view better myself...)

[01:53:00.0000] <Philip`>
Hixie: It seems unfair to only exempt automated conformance checkers from interpreting the author's intent - non-automated ones might be equally unable to interpret intent in some cases

[01:53:01.0000] <Dashiva>
I don't like how it selects the entire line with the background color

[01:55:00.0000] <Dashiva>
Lots of font tags there, ho

[01:55:01.0000] <Dashiva>
/me smiles at <font face="Helvetica,Arial" size="-1">&nbsp;</font>

[01:55:02.0000] <Philip`>
e.g. I might submit a document written in Lojban to a non-automated conformance checking service, and they could point out lots of non-machine-checkable issues (like if I used <b> instead of <h1>) but couldn't possibly check all the issues (like if the alt text was a decent alternative to the image)

[01:57:00.0000] <Philip`>
I don't like the idea of conformance checkers anyway, I'd prefer conformance chess

[02:00:00.0000] <Dashiva>
If you lose, your page is deleted?

[02:02:00.0000] <Hixie>
Philip`: why couldn't a human determine if the page had good alt text?

[02:03:00.0000] <Hixie>
or rather, not good, but spec-compliant alt text

[02:03:01.0000] <Dashiva>
If he doesn't know the language

[02:04:00.0000] <Hixie>
one would need to know the language to correctly assess conformance of a document

[02:04:01.0000] <Hixie>
just like one would need more than 2 bytes of memory to correctly assess conformance of a document

[02:04:02.0000] <Dashiva>
/me sees a challenge!

[02:05:00.0000] <Dashiva>
How much cache do I get?

[02:05:01.0000] <Hixie>
none

[02:05:02.0000] <Dashiva>
Darn

[02:05:03.0000] <Hixie>
2 bytes memory total

[02:05:04.0000] <Hixie>
including registers

[02:05:05.0000] <Hixie>
and I/O

[02:05:06.0000] <Dashiva>
Address space?

[02:05:07.0000] <Hixie>
one bit address space, 0 addresses the first byte, 1 addresses the second.

[02:07:00.0000] <Hixie>
you actually could do it in a conforming way -- just invoke the "User agents may impose implementation-specific limits on otherwise unconstrained inputs" rule to say that you only check documents up to 2 bytes

[02:07:01.0000] <Hixie>
and always say it'ss non-conforming

[02:07:02.0000] <Hixie>
but that's not the point! :-P

[02:07:03.0000] <Dashiva>
With enough disk space, and enough exams to procrastinate, I'm sure I could cook up something insane

[02:07:04.0000] <Hixie>
no disk space, you have 2 bytes total

[02:07:05.0000] <Hixie>
for everything

[02:08:00.0000] <Hixie>
program, data, cache, registers, everything

[02:08:01.0000] <Dashiva>
I have this vague feeling we've deviated from the initial problem statement somewhat

[02:08:02.0000] <Hixie>
including the video RAM

[02:08:03.0000] <Hixie>
:-)

[02:08:04.0000] <Hixie>
my point was just that there are minimum requirements for determining conformance

[02:08:05.0000] <Hixie>
one of those is (for non-automated checkers) knowign the language

[02:19:00.0000] <Philip`>
The service wouldn't be certifying conformance, it would just be helping me improve the quality of my document, so it's useful even if it isn't complete

[02:20:00.0000] <Hixie>
sure

[02:20:01.0000] <Hixie>
web browsers are useful even when they're not conforming too

[02:20:02.0000] <Hixie>
conformance isn't a minimum requirement for usefulness

[02:21:00.0000] <Hixie>
not by a long shot

[02:23:00.0000] <Philip`>
/me goes to bed :-)

[02:23:01.0000] <Hixie>
nn :-)

[04:28:00.0000] <Hixie>
hm, <iframe sandbox> for the first time introduces the concept of a document's origin changing while the document is up

[04:29:00.0000] <Hixie>
you could do   var e = iframe.contentDocument.documentElement; iframe.setAttribute('sandbox', '');

[04:29:01.0000] <Hixie>
at that point, you can navigate 'e' fine, but if you try to read anything from e.parentNode, you get an exception

[04:29:02.0000] <Hixie>
good times

[04:34:00.0000] <jruderman>
doesn't document.domain do that too?

[04:36:00.0000] <Hixie>
oh, i guess so

[04:36:01.0000] <Hixie>
yeah, fair enough

[04:37:00.0000] <jwalden>
thankfully that's the only other one

[04:37:01.0000] <Hixie>
though that one just changes the "effective origin" not the true origin

[04:37:02.0000] <jwalden>
effective origin is what matters, no?

[04:37:03.0000] <Hixie>
er, "effective script origin"

[04:37:04.0000] <jwalden>
for sec

[04:38:00.0000] <Hixie>
depends what for

[04:38:01.0000] <Hixie>
per spec document.domain doesn't affect where you can take images from for <canvas>, for instance

[04:39:00.0000] <Hixie>
and the rules for which browsing contexts you can navigate aren't affected by document.domain per spec

[04:39:01.0000] <jwalden>
oh, yeah

[04:41:00.0000] <Hixie>
and if you meta refresh to an http uri that redirects to a javascript: uri, the javascript: uri compares the original origin to the redirecting uri's origin to decide whether to run or not (and then runs in the same global context, but fails almost immediately since it can't get to anything on the Window or Document objects since it has a different effective script origin)

[04:41:01.0000] <Hixie>
hm, maybe that is a bug

[04:41:02.0000] <Hixie>
anyway

[04:41:03.0000] <Hixie>
the Web security model is going to be promising job security for us for years to come

[04:42:00.0000] <jruderman>
hah

[04:47:00.0000] <roc>
hmm

[04:47:01.0000] <roc>
changing the principal on a document actually sounds very scary

[04:47:02.0000] <roc>
that might be something to avoid if we can

[04:56:00.0000] <jwalden>
I think it's impossible to do that

[04:57:00.0000] <roc>
do it, or avoid it?

[04:57:01.0000] <jwalden>
avoid changing the principal

[04:58:00.0000] <roc>
we could just say that dynamic changes to 'sandbox' don't have any effect until a new document is loaded in the iframe

[04:58:01.0000] <jwalden>
that doesn't address document.domain

[04:58:02.0000] <jwalden>
but that seems like a good idea anyway

[05:17:00.0000] <Hixie>
roc: yeah, maybe

[05:17:01.0000] <Hixie>
roc: once i've sent out the e-mail where i say i added this, please do send any feedback you have :-)

[05:17:02.0000] <Hixie>
it'll likely be a few hours or days yet though

[05:18:00.0000] <roc>
our feedback on that should come from boris

[05:20:00.0000] <Hixie>
that's good too :-)

[05:20:01.0000] <Hixie>
this is all highly speculative stuff

[05:20:02.0000] <Hixie>
no rush or anything on any of this

[05:20:03.0000] <Hixie>
(sandbox="" or seamless="")

[05:21:00.0000] <roc>
these are features people have been asking for for years

[05:21:01.0000] <Hixie>
yeah

[05:21:02.0000] <roc>
I guess you saw the Mozilla bug

[05:21:03.0000] <Hixie>
indeed

[05:22:00.0000] <Hixie>
speccing this stuff was basically impossible until recent weeks, because it depends on so much of the previously unspecified stuff in really core ways

[05:22:01.0000] <Hixie>
i mean, it depends on the navigation model, on multiple windows, on how scripts run, on same-origin... it's got its tentacles all over the stuff that didn't ever get acknowledged before HTML5

[05:23:00.0000] <Hixie>
these features may in fact be the first real author-visible tangible benefits of having this spec be this detailed

[05:24:00.0000] <jwalden>
/me fights the urge to make a bad pun on the mailing list

[05:25:00.0000] <roc>
why fight it

[05:25:01.0000] <jwalden>
because I'm not sure I have much more that's useful to say at the same time

[05:25:02.0000] <jwalden>
at least not if we wanted the test harness to be usable in current browsers

[07:51:00.0000] <hsivonen>
"Require that Henri do the impossible by checking that scripts never cause non-conforming states to occur." great

[09:02:00.0000] <annevk>
hsivonen, his various definitions of compatible are pretty strict

[09:02:01.0000] <annevk>
hsivonen, probably needs more ~ usage :)

[09:07:00.0000] <annevk>
" 

[09:07:01.0000] <annevk>
The difficult, important, and often overlooked part of a version identification strategy is specifying the meaning and interpretation when a consumer encounters a version identifier it does not know about."

[09:07:02.0000] <annevk>
that's actually not just true for versioning :)

[09:11:00.0000] <hsivonen>
I think it is good that the doc acks that version identifiers aren't necessary, but it would be better to say that they are actually harmful

[09:12:00.0000] <annevk>
yes

[09:13:00.0000] <annevk>
maybe we should comment

[09:22:00.0000] <Hixie>
one day y'all will realise the true value of TAG findings...

[09:23:00.0000] <Hixie>
hsivonen: it wasn't really a major change to what we understood the spec to say anyway, i just hadn't realised that we hadn't made conformance checkers check for all conformance states

[09:23:01.0000] <Hixie>
hsivonen: it doesn't really affect you at all

[09:24:00.0000] <othermaciej>
is it greater than the value of TAG losings?

[09:26:00.0000] <hsivonen>
Hixie: as I said before, Dave Orchard's presentation about compatibility was good stuff before he got to applying it to XSD (it was XSD's fault that it wasn't as good from there on)

[09:27:00.0000] <hsivonen>
Hixie: I was just pointing out to annevk where the same reasoning is written down

[09:34:00.0000] <hsivonen>
aargh. I accidentally introduced a lot of bugs with Eclipse's refactoring tools

[09:36:00.0000] <roc>
Undo

[09:37:00.0000] <hsivonen>
roc: that would break a lot of good stuff too

[09:37:01.0000] <roc>
Redo :-)

[09:37:02.0000] <hsivonen>
fortunately, I have a way of identiying the problematic points

[09:37:03.0000] <Philip`>
Right click -> Remove Bugs (ctrl+alt+shift+r)

[10:38:00.0000] <Hixie>
what were the attributes we'd come up with for replacing data: URIs on <iframe>s?

[10:39:00.0000] <annevk>
html=""

[10:39:01.0000] <annevk>
markup=""

[10:39:02.0000] <annevk>
content=""

[10:39:03.0000] <hsivonen>
aren't data URIs more compatible?

[10:40:00.0000] <hsivonen>
or do we want incompatibility for security?

[10:40:01.0000] <Hixie>
data: URIs require too much thinking

[10:40:02.0000] <Hixie>
you have to escape too many characters (three four to be minimally safe, more to be compliant)

[10:41:00.0000] <Hixie>
and you don't get much backcompat anyway

[10:41:01.0000] <Hixie>
since IE doesn't support them

[10:42:00.0000] <hsivonen>
Hixie: by the time anyone ships a sandbox, IE8, Firefox, Safari and Opera together will be a big legacy

[10:42:01.0000] <Hixie>
a new attribute only really desperately needs one character escaped (the delimiter -- typically " => &quot) and for correctness the ampersand (& => &amp;) though the latter isn't security-critical, and the former is very obvious if you get it wrong

[10:44:00.0000] <annevk>
ah crap, <meta> has stolen content="" already it seems

[10:44:01.0000] <Hixie>
IE8 doesn't support <iframe src="data:text/html,..."> either

[10:44:02.0000] <Hixie>
at least not in beta1

[10:44:03.0000] <Hixie>
(acid3 tries to take care of that, but that's not likely to be seen until IE9)

[10:44:04.0000] <Hixie>
(and that's assuming a major IE8.5 between 8 and 9)

[10:47:00.0000] <annevk>
I like content="" the best

[10:48:00.0000] <Hixie>
"src", "source", "src2", "content", "value", and "data" are all vetoed

[10:48:01.0000] <Hixie>
doc="" is currently leading the pack

[10:48:02.0000] <Hixie>
with html="" a close second

[10:48:03.0000] <hsivonen>
will it accept only HTML fragments or XML too?

[10:48:04.0000] <Hixie>
HTML

[10:49:00.0000] <annevk>
that violates some principle you had initially

[10:49:01.0000] <Hixie>
which one?

[10:49:02.0000] <annevk>
that uas wouldn't be required to support both parsers if they didn't want to

[10:49:03.0000] <Hixie>
UAs wouldn't be

[10:49:04.0000] <annevk>
so XML UAs may ignore doc=""?

[10:49:05.0000] <hsivonen>
time to disassemble my code and count the bytes again...

[10:49:06.0000] <Hixie>
i dunno :-)

[10:50:00.0000] <Hixie>
doc="" doesn't exist

[10:50:01.0000] <Hixie>
it's just an idea

[10:50:02.0000] <Hixie>
i don't really want to skew this just to make the xml crowd happy

[10:50:03.0000] <Hixie>
since in practice nobody will use the xml version

[10:51:00.0000] <annevk>
htmldoc=""

[10:51:01.0000] <Hixie>
or type=""

[10:51:02.0000] <Hixie>
there are various options

[10:51:03.0000] <annevk>
lets not overload type even more :)

[10:52:00.0000] <Hixie>
does <iframe> have a type=""?

[10:52:01.0000] <Lachy>
Hixie, I think it does in HTML4

[10:52:02.0000] <Hixie>
cool

[10:52:03.0000] <Hixie>
well then

[10:52:04.0000] <Hixie>
we just use that

[10:52:05.0000] <Lachy>
/me checks...

[10:52:06.0000] <annevk>
doesn't seem like it...

[10:52:07.0000] <annevk>
does have longdesc :)

[10:53:00.0000] <Lachy>
actually, it doesn't

[10:53:01.0000] <hsivonen>
w00t. only 7933 bytes

[10:54:00.0000] <Lachy>
what problem is this new attribute trying to solve?

[10:54:01.0000] <Hixie>
if it doesn't have it, then i say we're fine with using it for our piurposes

[10:54:02.0000] <Hixie>
which new attribute?

[10:54:03.0000] <Lachy>
the one you're discussing adding to iframe

[10:54:04.0000] <Hixie>
doc="" or type=""?

[10:55:00.0000] <Hixie>
doc="" solves the problem of there not being any good safe way to embed content inline in a sandbox

[10:55:01.0000] <annevk>
making type conforming will led to confusion when it is not used together with doc

[10:55:02.0000] <annevk>
lets call it doctype=""

[10:55:03.0000] <Hixie>
type="" solves the theoretical problem of not being able to decide whether doc="" is html or xml or text/plain or png...

[10:55:04.0000] <Hixie>
annevk: :-P

[10:56:00.0000] <Lachy>
ok, so it's a backwards incompatible replacement for data URIs that provides security in new browsers without compromising older ones.

[10:57:00.0000] <Hixie>
data: URIs don't really provide what we need here, so it's not a replacement for them

[10:57:01.0000] <Hixie>
data: URIs are too complicated to correctly escape to really be a good security solution

[10:57:02.0000] <Lachy>
I don't really like the idea of embedding markup within an attribute

[10:58:00.0000] <Hixie>
do you have a better idea?

[10:58:01.0000] <Lachy>
but I have no alternative solution

[10:58:02.0000] <Hixie>
i agree that it's suboptimal

[11:00:00.0000] <annevk>
alternatively we could put it inside <iframe> ...

[11:01:00.0000] <annevk>
though maybe that's more icky for authors than just escaping "

[11:01:01.0000] <Lachy>
is the proposal do be able to include fragments of HTML in them like this: <iframe doc="<p>hello <em>world!</em>">, and then that creates a document within the frame using that fragment?

[11:01:02.0000] <Hixie>
content inside <iframe> has REALLY obscure escaping semantics

[11:01:03.0000] <Hixie>
Lachy: yes

[11:01:04.0000] <Hixie>
Lachy: well, that's a variant of hte proposal

[11:01:05.0000] <Hixie>
i'm writing it up in more detail in the e-mail i'm sending

[11:02:00.0000] <Lachy>
I assume authors won't need to include <!DOCTYPE html> in the fragment?

[11:02:01.0000] <Hixie>
we could make that implicit, probably

[11:02:02.0000] <Lachy>
and it would only allow elements that are valid within <body>

[11:02:03.0000] <Hixie>
hy?

[11:03:00.0000] <Hixie>
why, even

[11:03:01.0000] <Lachy>
hm. I guess that's not necessary

[11:04:00.0000] <othermaciej>
putting unescaped markup in a quoted string seems like it makes it easy to accidentally terminate the attribute value early

[11:04:01.0000] <othermaciej>
(or maliciously)

[11:05:00.0000] <Hixie>
othermaciej: luckily, that is trivially detectable, and easy to solve

[11:05:01.0000] <Hixie>
othermaciej: even without security attacks, people will use " or ' in their comments and thus it'll be caught quickly

[11:06:00.0000] <Hixie>
othermaciej: and the delimiter is the only character you absolutely must escape

[11:06:01.0000] <othermaciej>
I thought your major argument against some approaches to inline sandboxing was that they still require (minimal) escaping

[11:07:00.0000] <Hixie>
my objections were that they required non-trivial or easy-to-get-wrong escaping.

[11:07:01.0000] <Hixie>
people would be motivated to get this right for reasons other than just security

[11:08:00.0000] <othermaciej>
escaping case-insensitive </sandbox> is not very hard, though certainly possible to get wrong

[11:08:01.0000] <Hixie>
since getting it wrong would break the entire page the moment someone used a quote or apostrophe (as appropriate)

[11:08:02.0000] <Hixie>
othermaciej: sure but people don't type </sandbox> in as a matter of course

[11:08:03.0000] <Hixie>
othermaciej: so there it becomes _only_ a security problem

[11:09:00.0000] <Hixie>
othermaciej: and since it's more than one character, you end up with problems with over-long utf-8 sequences, embedded nulls, all kinds of stuff like that

[11:09:01.0000] <Hixie>
othermaciej: none of which affects a single quote mark

[11:09:02.0000] <Hixie>
well the overlong utf-8 might i guess

[11:09:03.0000] <othermaciej>
you have to normalize in any case

[11:10:00.0000] <othermaciej>
(unicode normalize)

[11:10:01.0000] <othermaciej>
(or reject input that is not in normal form already)

[11:10:02.0000] <Hixie>
who does?

[11:11:00.0000] <Hixie>
as far as i can tell, with the doc="" idea everyone or anyone can forget to normalise and you're still ok

[11:11:01.0000] <othermaciej>
whoever is escaping quotes

[11:12:00.0000] <othermaciej>
(i.e. whoever processes the input after receiving it from an untrusted source but before putting it in the attribute value)

[11:12:01.0000] <Hixie>
why?

[11:12:02.0000] <annevk>
escaping </sandbox> would not cut it btw, consider </sandbox >

[11:13:00.0000] <othermaciej>
would escaping "</sandbox" cut it?

[11:13:01.0000] <othermaciej>
Hixie: to successfully escape denormal close quotes?

[11:13:02.0000] <othermaciej>
or am I missing the point of your "overlong utf-8" remark?

[11:13:03.0000] <Philip`>
You could easily forget to escape "</sandbox" case-insensitively

[11:14:00.0000] <annevk>
case-insensitive </sandbox might, but then so would </iframe i suppose

[11:15:00.0000] <annevk>
still, " seems a lot easier

[11:16:00.0000] <othermaciej>
it is easier, but the attribute approach has the big downside that you need an iframe per item to be sandboxed, and iframes are heavyweight

[11:17:00.0000] <othermaciej>
so if we can countenance some escaping of untrusted content being required, I'd rather see a version that is usable for hundreds or thousands of items

[11:17:01.0000] <Philip`>
Heavyweightedness is just an implementation bug :-)

[11:17:02.0000] <Hixie>
othermaciej: i'm confused. I'm talking about encoding U+0022 as, e.g., 0xC0 0xA2.

[11:17:03.0000] <Hixie>
othermaciej: which has nothing to do with unicode normalisation forms

[11:18:00.0000] <othermaciej>
Hixie: then I apologize for using the terminology incorrectly

[11:18:01.0000] <annevk>
othermaciej, ah yeah, creating hundreds of document objects is costly

[11:18:02.0000] <annevk>
not really sure how to do something smarter than that though

[11:18:03.0000] <othermaciej>
Philip`: some of the heavyweightedness of iframe is effectively required by specs

[11:18:04.0000] <Hixie>
othermaciej: all the workable sandboxing approaches i've seen so far invoke new browsing contexts (iframes)

[11:18:05.0000] <annevk>
afaict with <sandbox> you have the same issue, no?

[11:19:00.0000] <Hixie>
othermaciej: the only differences between the working proposals i've seen is how they encode the data

[11:19:01.0000] <Hixie>
othermaciej: i don't understand how to do origin isolation without a separate browsing context

[11:19:02.0000] <othermaciej>
Hixie: I agree you can't do a sandbox with any scripting enabled without a separate browsing context

[11:19:03.0000] <othermaciej>
(not feasibly anyway)

[11:19:04.0000] <othermaciej>
I think you can do a noscript sandbox without an extra browsing context or document

[11:20:00.0000] <Hixie>
it's not just scripting

[11:20:01.0000] <Hixie>
e.g. target="" and forms would need very complicated and subtle changes to be done safely without an iframe

[11:21:00.0000] <Hixie>
and i'm not confident that i could spec those precisely enough to avoid introducing security bugs into the spec

[11:22:00.0000] <Hixie>
and i'm certainly not willing to take on the responsibility for spec-mandated security holes in humanity's most widely deployed application platform :-)

[11:22:01.0000] <Hixie>
i can't afford the legal fees, for one

[11:23:00.0000] <annevk>
i hope you have some protection against that regardless :)

[11:23:01.0000] <othermaciej>
I think for the use case of the inline sandbox (blog comments or the like) it would be ok to totally ban targetted links and form elements (parser discards them)

[11:23:02.0000] <othermaciej>
and if the spec is found to require security holes it will be ignored on that count as I am sure you know

[11:23:03.0000] <annevk>
MessageEvent.uri comes to mind :)

[11:23:04.0000] <annevk>
though that's more a privacy issue

[11:23:05.0000] <Hixie>
othermaciej: not before multiple shipping implementations are out there, possibly with content depending on it (e.g. look how far we've gotten with blocking cross-origin <script> or form submissions)

[11:24:00.0000] <Philip`>
Hixie: Just add a statement "UAs must not contain any security holes" - if the spec also requires certain features which are security holes, then any UA trying to implement HTML5 will necessarily be non-conforming and not your problem any more

[11:24:01.0000] <Philip`>
For blog comments, could you put the entire comments section in a single sandbox?

[11:24:02.0000] <Philip`>
(to avoid the issues with having hundreds of separate sandboxes on a single page)

[11:25:00.0000] <Hixie>
othermaciej: (i'm not at all convinced i could catch all cases -- e.g so far your proposal has not covered banning <meta http-equiv=refresh>)

[11:25:01.0000] <othermaciej>
Philip`: often you want un-sandboxed content near each comment

[11:25:02.0000] <Hixie>
having one sandbox might work

[11:25:03.0000] <othermaciej>
Philip`: so not sure that works

[11:25:04.0000] <Hixie>
but there is that issue

[11:25:05.0000] <othermaciej>
Hixie: is <meta http-equiv=refresh> allowed in the body?

[11:26:00.0000] <Hixie>
nope, but the parser doesn't care about that

[11:26:01.0000] <othermaciej>
I don't really mean allowed

[11:26:02.0000] <othermaciej>
I mean, does it work when in the body

[11:26:03.0000] <Hixie>
sure

[11:26:04.0000] <Hixie>
i'm not sure how you'd disable script on a per-element basis, either

[11:26:05.0000] <annevk>
it could work if the parser just dropped certain elements during a certain mode

[11:27:00.0000] <annevk>
"in sandbox insertion mode"

[11:27:01.0000] <annevk>
or something

[11:27:02.0000] <Hixie>
e.g. you'd have to somehow handle <img src="foo.cgi"> redirecting to a javascript: URI

[11:27:03.0000] <othermaciej>
Hixie: redirecting to a javascript: URI is not allowed in any case

[11:27:04.0000] <Hixie>
it is per spec at the moment

[11:27:05.0000] <othermaciej>
(it would be a security hole)

[11:27:06.0000] <Hixie>
and it is nicely and safely defined

[11:28:00.0000] <Hixie>
(it runs in a void execution context and its return value is used as the image data)

[11:28:01.0000] <othermaciej>
disallowing redirect to javascript: is perfectly safe

[11:28:02.0000] <Hixie>
(unless the origins match in which case it runs in the context of the page, iirc)

[11:28:03.0000] <hsivonen>
/me wants to see an accessibility advocate type alt text for hundreds of vacation photos

[11:28:04.0000] <othermaciej>
I don't see the point of allowing it, context notwithstanding

[11:29:00.0000] <Philip`>
annevk: Then some people would want to handle <!--[if IE]><script>alert('haxxored')</script><![endif]--> too, and it'd seem a bit dangerous

[11:29:01.0000] <Hixie>
well it's just one example of what the spec does currently allow which we would have to handle somehow

[11:29:02.0000] <othermaciej>
I think the spec just shouldn't allow that

[11:29:03.0000] <othermaciej>
but good point

[11:29:04.0000] <Hixie>
with a browsing context, handling it is trivial

[11:29:05.0000] <Hixie>
you just disable scripting.

[11:30:00.0000] <othermaciej>
well, you have to make sure not to create a special-case void scripting context there either

[11:30:01.0000] <othermaciej>
I wouldn't assume that just automatically follows

[11:30:02.0000] <annevk>
Philip`, I'm assuming conforming parsers

[11:31:00.0000] <Lachy>
Hixie, how can U+0022 be encoded as 0xC0 0xA2? That doesn't seem to work

[11:32:00.0000] <Hixie>
Lachy: lax UTF-8 decoders that do no sanity checking would decode that to U+0022

[11:32:01.0000] <Philip`>
annevk: That seems like an impractical assumption

[11:34:00.0000] <annevk>
Philip`, your practical assumption has security holes

[11:37:00.0000] <Philip`>
annevk: My assumption that some people won't implement the parser precisely, combined with your suggestion to enforce security rules in the parser, is what has security holes

[11:41:00.0000] <annevk>
Hixie, you went from attribute DOMString to boolean DOMString

[11:43:00.0000] <annevk>
you want attribute boolean

[11:43:01.0000] <Hixie>
lol

[11:43:02.0000] <Hixie>
i suck

[11:59:00.0000] <annevk>
/me wonders which browser Hixie used

[12:00:00.0000] <hsivonen>
chaals++ on wai-xtech

[12:00:01.0000] <Hixie>
safari trunk

[12:00:02.0000] <Hixie>
i may have js disabled, i didn't check

[12:00:03.0000] <Hixie>
i've been doing tests that involve disabling js recently

[12:00:04.0000] <annevk>
k

[12:00:05.0000] <Dashiva>
A gallery that doesn't work without JS is pretty sad

[12:01:00.0000] <annevk>
i wrote it myself for myself so i don't really care about that

[12:01:01.0000] <annevk>
though cross browser would be nice

[12:01:02.0000] <Dashiva>
:)

[12:01:03.0000] <annevk>
at some point i guess i will use pushState() and such

[12:01:04.0000] <Hixie>
if it's for yourself, interop isn't important and worrying about conformance seems mostly pointless

[12:02:00.0000] <Hixie>
let alone accessibility

[12:34:00.0000] <hsivonen>
Hixie: do you have a full list of known (including non-conforming HTML legacy) attributes for HTML, MathML and SVG?

[12:59:00.0000] <hsivonen>
Philip`: do you have a full attribute list by any chance?

[13:02:00.0000] <hsivonen>
/me is glad Hixie didn't add xml:id to the adjustable attributes

[13:04:00.0000] <annevk>
I'm pretty sure that was a conscious decision

[13:04:01.0000] <hsivonen>
me too

[13:05:00.0000] <annevk>
hsivonen, full attribute list is pretty hard to determine, no?

[13:05:01.0000] <hsivonen>
annevk: perhaps, but presumable Hixie already did part of the work in order to find the camelCased ones

[13:10:00.0000] <annevk>
hsivonen, oh, I think that list is complete, but doesn't include SVG 1.2

[13:19:00.0000] <hsivonen>
annevk: I need a list that includes non-camelcase SVG and HTML (incl legacy) and MathML

[13:23:00.0000] <takkaria>
Doug Crockford's JS book is well-written

[14:17:00.0000] <Philip`>
hsivonen: I don't

[14:18:00.0000] <hsivonen>
Philip`: ok. thakn

[14:18:01.0000] <hsivonen>
s

[14:24:00.0000] <hsivonen>
dealing with xmlns cruft in the V.nu parser is painful

[14:26:00.0000] <MikeSmith>
/me contemplates making joke about how maybe some people would say namespaces should have been titled pain-spaces, but reconsiders...

[15:04:00.0000] <Dashiva>
I wonder if Dmitry would have better luck with his proposals if he didn't invent new words for every single term

[15:07:00.0000] <MikeSmith>
using "Man of good will," as a salutation is nice, though

[15:08:00.0000] <MikeSmith>
the spirit of optimism in it

[15:08:01.0000] <Lachy>
it's werid and I don't like it.

[15:08:02.0000] <MikeSmith>
Lachy: c'mon

[15:09:00.0000] <Philip`>
http://lists.xml.org/archives/xml-dev/200805/msg00087.html

[15:09:01.0000] <Philip`>
Not sure where else it's been posted

[15:09:02.0000] <MikeSmith>
I think "quixotic" is the word I'm looking for

[15:10:00.0000] <MikeSmith>
Lachy: please consider adding Cervantes "Don Quixote" to your reading list, if you haven't read it yet

[15:10:01.0000] <MikeSmith>
before or after "At Swim-Two-Birds"

[15:11:00.0000] <Lachy>
MikeSmith, it's far too formal for a place that is more comfortable when treated as a relatively informal discussion channel in most cases

[15:12:00.0000] <MikeSmith>
yeah, I suppose it is if you take it too seriously

[15:13:00.0000] <takkaria>
windows mobile is one of the worst OSes I've ever ever used

[15:13:01.0000] <MikeSmith>
but it's a quite appropriate way to address the world if you're battling windmills

[15:13:02.0000] <MikeSmith>
takkaria: amen to that

[15:14:00.0000] <Philip`>
MikeSmith: You think it's optimistic to imply there's one man in the HTML WG who is of good will?

[15:15:00.0000] <MikeSmith>
Philip`: ah geez.. I really doubt he means to imply that

[15:16:00.0000] <Philip`>
takkaria: Windows, Linux and OS X are three of the worst desktop OSes I've ever used

[15:16:01.0000] <MikeSmith>
Windows Mobile in a special class by itself

[15:16:02.0000] <Philip`>
(though they're also three of the best)

[15:16:03.0000] <hsivonen>
gotta love all the special attentition lang/xml:lang requires...

[15:17:00.0000] <MikeSmith>
if you want a good laugh, try some javascript timing tests on a windows mobile device some time

[15:17:01.0000] <MikeSmith>
or any kind of timing test

[15:18:00.0000] <Philip`>
or any other kind of test

[15:18:01.0000] <MikeSmith>
heh

[15:18:02.0000] <takkaria>
if you want two hours of frustration, try to send a file that's about 0.5MB smaller than the amount of space you have free on your storage card via bluetooth

[15:19:00.0000] <takkaria>
because it seems to accept the file, ask if you want to save it, and when you answer "yes", delete it since it's worried about free space

[15:20:00.0000] <takkaria>
and AFAICT, there's no way to tell it to save on the 1GB microSD card which has more than enough free space for fifty copies of the same file

[15:20:01.0000] <takkaria>
</rant>

[15:21:00.0000] <hsivonen>
/me tries to trade memory for execution speed with the attribute silliness

[15:23:00.0000] <takkaria>
hmm, Gez seems to think that the entirely emotive reply to Hixie about his lack of empathy was appropriate in a technical context

[15:26:00.0000] <Dashiva>
And this from the same crowd who insist that we absolutely must take into consideration AT five versions outdated because the users don't upgrade :)

[15:27:00.0000] <Philip`>
The best thing about the alt discussion is that the really long threads look pretty in Gmail when it shows the colour-coded names of authors

[15:28:00.0000] <takkaria>
I also emailed Gez privately just because emailing the list seemed like it was clogging up people's inboxes

[15:28:01.0000] <takkaria>
sigh

[15:30:00.0000] <hsivonen>
aaaargh. it's crazy how much code I have to write to support both HTML5 DOMs and XML 1.0 + Namespaces infosets

[15:31:00.0000] <hsivonen>
Namespaces--

[15:31:01.0000] <Philip`>
/me suggests dropping XML support

[15:32:00.0000] <Philip`>
XML isn't even trendy any more, so there's no reason to bother with it

[15:32:01.0000] <hsivonen>
Philip`: that would mean violating the SAX API all the time and dropping generic DOM support and dropping XOM support

[15:33:00.0000] <Philip`>
Hmm, I suppose that could be cosidered a bit of a pain

[15:35:00.0000] <hsivonen>
the thing is that only the TreeBuilder knows if xlink:href needs to be dropped on the floor or kept when the parser is coercing the data into a proper infoset

[15:48:00.0000] <Philip`>
/me isn't sure "tangents of tangents" has a particularly useful geometric meaning :-(

[15:48:01.0000] <takkaria>
Philip`: sorry

[15:49:00.0000] <Philip`>
since a tangent is a straight line touching some other line, so a tangent of a tangent would just be the tangent itself :-p

[15:49:01.0000] <takkaria>
luckily I wasn't using it geometrically

[15:49:02.0000] <hsivonen>
Philip`: only in 2D

[15:49:03.0000] <Philip`>
Indeed - in that case it does make sense :-)

[15:49:04.0000] <takkaria>
hmm, this discussion would then be a tangent of a tangent of a tangent...

[15:50:00.0000] <Philip`>
hsivonen: s/straight line/(N-1)-dimensional plane in an N-dimensional space/

[15:51:00.0000] <Philip`>
/me hopes he didn't get that entirely wrong

[15:51:01.0000] <hsivonen>
the attribute names are much worse than the element names


2008-05-22
[01:08:00.0000] <MikeSmith>
yay, krijnh is back

[01:17:00.0000] <krijnh>
Yay :)

[01:17:01.0000] <krijnh>
Sorry for being down

[01:17:02.0000] <krijnh>
Are the logs reachable from the outside?

[01:26:00.0000] <krijnh>
Grr

[01:27:00.0000] <MikeSmith>
annevk, krijnh: I have incomplete logs from last few days

[01:27:01.0000] <krijnh>
MikeSmith: can you mail them to me?

[01:28:00.0000] <krijnh>
/me hopes they're in the same format

[01:28:01.0000] <MikeSmith>
krijnh: OK, but I reckon probably Hixie or another screen user has more complete ones

[01:28:02.0000] <MikeSmith>
anyway, will send you what I have for now

[01:28:03.0000] <krijnh>
My connection is pretty shitty btw.. I think I'll disconnect a lot today

[01:30:00.0000] <MikeSmith>
krijnh: can I send you the complete logs and let you figure out which parts are needed? my logs are cumulative XChat logs 

[01:30:01.0000] <krijnh>
Depends on the format

[01:30:02.0000] <MikeSmith>
e.g., whatwg one is 8.1MB

[01:30:03.0000] <MikeSmith>
or I can just post them at URL for you to download

[01:31:00.0000] <MikeSmith>
xchat format

[01:31:01.0000] <krijnh>
No idea what that is ;)

[01:31:02.0000] <krijnh>
You can also ftp it to me

[01:32:00.0000] <krijnh>
User: up / Pass: load

[01:32:01.0000] <MikeSmith>
krijnh: example of format:

[01:32:02.0000] <MikeSmith>
May 22 16:36:08 <krijnh>        No idea what that is ;)

[01:32:03.0000] <krijnh>
One moment, I'll see how that's handled

[01:33:00.0000] <MikeSmith>
will ftp them to you

[01:33:01.0000] <krijnh>
http://krijnhoetmer.nl/irc-logs/whatwg/20080520

[01:34:00.0000] <krijnh>
Probably acceptable :)

[01:34:01.0000] <MikeSmith>
[[

[01:34:02.0000] <Hixie>
hsivonen: i agree with raman and would go even further and require that the <input> element be used to make checkboxes, so... i don't plan to reply to the e-mail in question

[01:34:03.0000] <MikeSmith>
mput 01_freenode-#whatwg.log? y

[01:34:04.0000] <MikeSmith>
421 Service not available, remote server has closed connection

[01:34:05.0000] <MikeSmith>
]]

[01:35:00.0000] <krijnh>
Huh

[01:36:00.0000] <krijnh>
Shitty connection, shitty ftp software, sorry :)

[01:36:01.0000] <annevk>
I didn't quite get raman, the <span> example is quite different from the <input> one

[01:37:00.0000] <annevk>
The <input> one represents a control where the <span> example represents something that fakes a control to AT

[01:37:01.0000] <krijnh>
MikeSmith: mail then?

[01:42:00.0000] <MikeSmith>
krijn: http://people.w3.org/mike/logs/

[01:42:01.0000] <hsivonen>
Hixie: I agree that <input type=checkbox> is preferable. ARIA just seems necessary considering what people out there do (including GWT it seems)

[01:43:00.0000] <Hixie>
annevk: (re xhr/ac) ah. i recommend just making a decision and replying to the e-mail -- not replying leaves people without the ability to help you since we don't know that you're stuck :-)

[01:43:01.0000] <Hixie>
othermaciej: what on that list (http://lists.w3.org/Archives/Public/www-tag/2008May/0087.html) isn't architectural?

[01:44:00.0000] <othermaciej>
Hixie: " The use of imperative definitions rather than abstract definitions with  the requirement of black-box equivalence in implementations"

[01:45:00.0000] <Hixie>
othermaciej: ah, yeah, i figured i should include that since it seems like the kind of thing the tag would complain about

[01:45:01.0000] <annevk>
i replied to raman

[01:45:02.0000] <hsivonen>
Hixie: as I understand it, people want to make checkboxes that look like they want and they don't like the way the native ones look

[01:45:03.0000] <othermaciej>
Hixie: ok I will grant they may mistake it for architecture

[01:45:04.0000] <hsivonen>
Hixie: and they want to do it in a way that doesn't require all browser vendors to participate for the solution to work

[01:46:00.0000] <hsivonen>
Hixie: it's pretty ironic that with ARIA, the top 4 vendors *are* participating all of a sudden

[01:46:01.0000] <Hixie>
othermaciej: my goal here is to preclude the tag from coming back in 2 years and telling us we have to do X and that it's ok that they are sending the feedback late because we didn't ask for feedback earlier (as they recently did with aria)

[01:47:00.0000] <annevk>
othermaciej, the TAG has in fact given comments on imperative definitions to quite some extent on the Access Control specification

[01:47:01.0000] <Hixie>
hsivonen: seems to me the assumptions underlying the ARIA work have been invalidated

[01:47:02.0000] <hsivonen>
Hixie: so better CSS and XBL2 don't fit the requirements

[01:47:03.0000] <annevk>
othermaciej, it went so far that one of them started drafting text for doing it in a different way

[02:05:00.0000] <hsivonen>
Hixie: actually, it may be that all browser vendors participated precisely because ARIA was set up in such a way that no vendor could block it by not participating

[02:05:01.0000] <Hixie>
annevk: i hope they try to do that with the "navigate" algorithm

[02:05:02.0000] <othermaciej>
annevk: the TAG seems to be a force of randomness

[02:05:03.0000] <Hixie>
hsivonen: maybe

[02:05:04.0000] <annevk>
implementation ARIA is much more trivial than finding a solution for customizing form controls

[02:05:05.0000] <annevk>
And also more short term as authors can use it directly without having to worry about legacy

[02:05:06.0000] <Hixie>
i can't wait for people to start using aria on pages that don't need it

[02:05:07.0000] <annevk>
To be clear, I agree that it's far from optimal. At this point it just seems sort of convenient for everyone...

[02:05:08.0000] <othermaciej>
I feel like for every part of ARIA we implement, we have to tithe by adding appropriate equivalent HTML5 functionality and/or needed additional form control stylability

[02:05:09.0000] <othermaciej>
but I think even with the right high-level semantic elements, a low-level accessibility feature like ARIA makes sense

[02:05:10.0000] <annevk>
Hixie, so I think AC is good as is, apart from the warning for IIS servers which I will add today

[02:05:11.0000] <othermaciej>
back shortly

[02:05:12.0000] <Hixie>
annevk: cool

[02:05:13.0000] <annevk>
Hixie, XHR2 has some of the issues XHR1 has which makes it more problematic

[02:05:14.0000] <Hixie>
annevk: solve the xhr1 issues. :-)

[02:05:15.0000] <annevk>
I initially thought we needed the method/header whitelist from the server but Bjoern convinced me that Jonas didn't make sense...

[02:05:16.0000] <annevk>
Hixie, indeed :)

[02:14:00.0000] <hsivonen>
can anyone think of notable legacy HTML attributes that aren't conforming in either of HTML 4.01 Transitional or HTML5?

[02:14:01.0000] <Hixie>
hsivonen: <marquee speed>

[02:14:02.0000] <hsivonen>
Hixie: thanks

[02:14:03.0000] <Hixie>
actually that's not a valid attribute

[02:14:04.0000] <Hixie>
wikipedia says <marquee scrolldelay>

[02:14:05.0000] <hsivonen>
/me tries to find Philip`'s marquee stats

[02:14:06.0000] <Hixie>
<img lowsrc> probably was never valid either

[02:14:07.0000] <Lachy__>
hsivonen, <table height>

[02:14:08.0000] <hsivonen>
Lachy__: height is a conforming name in other contexts

[02:14:09.0000] <hsivonen>
I'm looking for unique strings

[02:14:10.0000] <hsivonen>
Hixie: yeah, I need to add lowsrc

[02:14:11.0000] <Hixie>
what's this for?

[02:14:12.0000] <Hixie>
if it's for ui, you probably just should look at common attribute names in stats

[02:14:13.0000] <Hixie>
i expect <Script languaje=""> is used more than <img lowsrc>

[02:14:14.0000] <hsivonen>
Hixie: for efficient magic handling of attribute names

[02:14:15.0000] <hsivonen>
in the parser

[02:14:16.0000] <annevk>
http://www.eskimo.com/~bloo/indexdot/html/tagpages/attributes/data.htm

[02:14:17.0000] <annevk>
http://www.eskimo.com/~bloo/indexdot/html/tagpages/attributes/editing.htm

[02:14:18.0000] <Hixie>
given the effort you're expending on optimisation work, i expect your parser to be able to parse the html5 spec in milliseconds :-P

[02:14:19.0000] <hsivonen>
Hixie: so far benchmarking suggests that having magic knowledge of element names was a win

[02:14:20.0000] <annevk>
onbounce, onfinish, etc.

[02:15:00.0000] <annevk>
from http://www.eskimo.com/~bloo/indexdot/html/tagpages/attributes/events.htm

[02:15:01.0000] <hsivonen>
Hixie: I need this code for some attributes anyway in order not to make the common cases suck

[02:15:02.0000] <hsivonen>
Hixie: once the code is there, I might as well put less common cases on the fast track, too

[02:15:03.0000] <Hixie>
cool

[02:16:00.0000] <annevk>
Hixie, add <ruby> support if you don't want to do parsing :)

[02:17:00.0000] <Hixie>
<ruby> support is going to be a pain because the css side doesn't support what i need

[02:18:00.0000] <Hixie>
iirc

[02:18:01.0000] <Hixie>
either that or i couldn't work out how to do IE compat on parsing, or something

[02:18:02.0000] <Hixie>
there was some complication

[02:19:00.0000] <hsivonen>
but doing this attribute work in a way that isn't very naïve is surprisingly tedious

[02:19:01.0000] <hsivonen>
I think I'm going to suspend this task in order to avoid starving higher-layer conformance checking work

[02:20:00.0000] <annevk>
Hixie, I thought it was the parsing and that you decided an approximation would be enough

[02:20:01.0000] <Hixie>
probably

[02:21:00.0000] <Hixie>
ruby is on my list

[02:21:01.0000] <krijn>
MikeSmith: downloaded your logs, will put them online today

[02:22:00.0000] <Hixie>
it's about half way down my priority list, whether i sort by number of e-mails, number of e-mails per thread, age of feedback, or whatever

[02:24:00.0000] <MikeSmith>
krijn: OK

[02:25:00.0000] <hsivonen>
annevk: thanks for the bloo links. is anyone actually using the IE4 dataformatas stuff?

[02:25:01.0000] <krijn>
And still 855 mails behind on public-html :/

[02:26:00.0000] <hsivonen>
I hope we could go back in time and make SVG use hyphens instead of camelCase

[02:28:00.0000] <hsivonen>
getting attributes right is such a mess

[02:28:01.0000] <hsivonen>
and most of the mess is due to Namespaces and xml:lang

[02:28:02.0000] <hsivonen>
compared to those, the camelCase fixup is a lesser mess

[02:28:03.0000] <roc_>
hmm

[02:29:00.0000] <roc_>
SVG *does* use hyphens

[02:29:01.0000] <roc>
I guess it uses both :-(

[02:29:02.0000] <hsivonen>
roc: SVG uses many things

[02:29:03.0000] <Hixie>
SVG is as bad as HTML

[02:29:04.0000] <hsivonen>
roc: the hyphen is benign

[02:29:05.0000] <Hixie>
it uses runtogether, camelCase, hyphen-ated, and mixtures of those

[02:29:06.0000] <roc>
I actually prefer camelCase

[02:29:07.0000] <hsivonen>
xlink:foo and the camelCase suck really badly for me

[02:30:00.0000] <Hixie>
oh and with:namespace, too

[02:30:01.0000] <hsivonen>
I mean if I end up writing lot of corner case code *anyway* I think I should make it performant, too.

[02:30:02.0000] <Hixie>
html at least has the defence of having been designed by many people

[02:30:03.0000] <Hixie>
svg is inconsistent for no good reason

[02:30:04.0000] <hsivonen>
but then there are so many situations that need different handling

[02:30:05.0000] <hsivonen>
aargh

[02:31:00.0000] <hsivonen>
I've decided I'm going to drop some marginal elegance and correctness

[02:31:01.0000] <roc>
hyphens in names need to be mangled when you convert them to DOM attribute names etc

[02:31:02.0000] <hsivonen>
roc: good point. But *my* code doesn't deal with that. :-)

[02:32:00.0000] <Hixie>
html is case-insensitive, so anything with uppercase letters becomes a mess

[02:32:01.0000] <annevk>
hsivonen, dunno

[02:32:02.0000] <annevk>
Hixie, SVG is designed by many people too I think :(

[02:33:00.0000] <Hixie>
yeah but they're in one wg, so that's no excuse

[02:33:01.0000] <roc>
yeah, not so much

[02:33:02.0000] <annevk>
Hixie, hmm, maybe Policy-Path is a bigger issue, do you think it would still be worth it if it was only for the entire domain?

[02:34:00.0000] <Hixie>
probably not

[02:35:00.0000] <hsivonen>
so we have: 1) plain attribute foo, 2) camelCase in no namespace fooBar, 3) legacy colons: xlink:href, 4) lang vs. xml:lang, 5) default namespace talisman: xmlns, 6) XLink talisman xmlns:xlink, 7) other prefix cruft: xmlns:foo, 8) the id attribute

[02:35:01.0000] <hsivonen>
that's pretty insane

[02:35:02.0000] <hsivonen>
and then some of those have contextual XML mappability constraints

[02:36:00.0000] <hsivonen>
I think the count goes up to at least 11 different situations

[02:37:00.0000] <hsivonen>
and only 3 of them aren't the fault of Namespaces

[02:38:00.0000] <Hixie>
so i have an e-mail here saying that GBK should be treated like GB2312

[02:38:01.0000] <hsivonen>
every time I write code that touches this area, I can't but think that Namespaces was a huge mistake

[02:38:02.0000] <Hixie>
but wikipedia says GBK is a superset of 2312

[02:38:03.0000] <Hixie>
*confused*

[02:38:04.0000] <hsivonen>
Hixie: GB2312 should be treated as GBK

[02:38:05.0000] <Hixie>
ok that makes more sense

[02:39:00.0000] <hsivonen>
Hixie: Gecko does something a bit weirder which amounts to that as a black box and WebKit simply aliases it, IIRC

[02:39:01.0000] <gsnedders>
hsivonen: yeah, that's right

[02:40:00.0000] <Hixie>
i wonder how to test if euc-kr and win-949 are being treated the same

[02:41:00.0000] <hsivonen>
Hixie: I'm so glad you didn't add xml:id to the fixups

[02:41:01.0000] <Hixie>
that was intentional

[02:41:02.0000] <hsivonen>
it would raised the case count from 11 to 13

[02:41:03.0000] <hsivonen>
oops. 14 actually

[02:41:04.0000] <krijn>
hsivonen: is there a way to use validator.nu on a page with http authentication?

[02:42:00.0000] <hsivonen>
krijn: not directly.

[02:42:01.0000] <krijn>
Tried putting in the uri, but that's deprecated :)

[02:42:02.0000] <hsivonen>
krijn: but you can retrieve the bytes and the HTTP header yourself and POST it to V.nu

[02:43:00.0000] <Philip`>
hsivonen: If this is just for performance, why do you care about specified / legacy attributes rather than about the most commonly used attributes?

[02:43:01.0000] <hsivonen>
I don't want to ask people to send credentials to my server

[02:43:02.0000] <krijn>
hsivonen: I understand

[02:44:00.0000] <hsivonen>
Philip`: you are right. I should also care about the most commonly used

[02:44:01.0000] <Philip`>
hsivonen: Why "also", rather than "only"?

[02:45:00.0000] <annevk>
so access control could help with validating authenticated sites (though you'd still be sharing data with henri one way or another)

[02:45:01.0000] <hsivonen>
Philip`: theory, I guess

[02:46:00.0000] <Philip`>
hsivonen: http://www.dia.wa.gov.au/ has a dataformatas attribute (though it doesn't seem to be used correctly or usefully)

[02:48:00.0000] <Philip`>
hsivonen: I could get a list of the most commonly used attribute names, if that'd be useful

[02:48:01.0000] <hsivonen>
Philip`: it would be useful, yes

[02:48:02.0000] <Philip`>
(http://www.oyak.com.tr/ uses a dataformatas too)

[02:52:00.0000] <hsivonen>
Hixie: making xmlns talismans allowed contextually adds to the attribute craziness pain

[02:53:00.0000] <Hixie>
i imagine

[02:56:00.0000] <annevk>
actually, Hixie, how is Policy-Path vulnerable in the cases Jonas and Bjoern mention?

[02:56:01.0000] <hsivonen>
Namespaces is such a white elephant

[02:57:00.0000] <annevk>
I don't quite get it, because it requires value equivalence it should be more safe

[02:57:01.0000] <hsivonen>
makes things harder for authors to grok and makes thing harder to implement

[02:57:02.0000] <Hixie>
annevk: if you OPTIONS http://example.com/foo/ and it says that there is a policy-path for all of /foo/

[02:58:00.0000] <Hixie>
annevk: and you then do a POST to http://example.com/foo/bar and due to a misconfiguration on the server that is treated as equivalent to http://example.com/baz

[02:58:01.0000] <annevk>
oh right, it's about the requests after that, never mind :(

[02:59:00.0000] <hsivonen>
Hixie: I'm so glad you aren't trying to emulate the AVNormalize stuff

[02:59:01.0000] <Hixie>
annevk: of course you could similarly argue that HTTP in general is a security risk because what happens if when you do a GET to http://example.com/test, the server, being misconfigured, instead does an rm -rf of all the data in the directory for example.org which happens to be hosted on the same virtual host

[02:59:02.0000] <Hixie>
hsivonen: ?

[02:59:03.0000] <hsivonen>
Hixie: making the attribute *value* change contextually, too

[02:59:04.0000] <Hixie>
hsivonen: o_O

[03:00:00.0000] <hsivonen>
/me referst to XML spec fragment ids: http://www.w3.org/TR/REC-xml/#AVNormalize

[03:00:01.0000] <annevk>
true

[03:00:02.0000] <annevk>
grmbl

[03:00:03.0000] <annevk>
i don't like having to resolve these issues as i don't really know the right answer

[03:01:00.0000] <Hixie>
well someone has to resolve them :-)

[03:01:01.0000] <Hixie>
/me has this problem all the time with html5

[03:01:02.0000] <Hixie>
hsivonen: oh, attribute value normalisation

[03:01:03.0000] <Hixie>
hsivonen: yeah, screw that, who cares :-)

[03:02:00.0000] <Hixie>
annevk: in practice if you make the wrong choice, someone tells you, and you change it

[03:02:01.0000] <Hixie>
annevk: it's pretty simple :-)

[03:03:00.0000] <annevk>
well, if only one person speaks up, such as Julian, it's more tricky, because i often disagree with him :)

[03:03:01.0000] <annevk>
but the XHR issues are more trivial anyway

[03:03:02.0000] <annevk>
except for dealing with him

[03:03:03.0000] <Philip`>
hsivonen: Is it more useful to know the number of pages each attribute appears on, or the total number of times the attribute appears on all pages?

[03:04:00.0000] <hsivonen>
Philip`: I suppose the latter in principle but the former may be better in order to avoid weird selection biases

[03:10:00.0000] <Philip`>
http://canvex.lazyilluminati.com/survey/2007-07-17/analyse.cgi/attr/t gets somewhat skewed by Topix

[03:11:00.0000] <hsivonen>
Philip`: should I add csobj to the element list?

[03:12:00.0000] <Hixie>
annevk: well, if you need backup don't hesitate to let me know you want a reply

[03:13:00.0000] <Hixie>
and i'll deal with it :-)

[03:13:01.0000] <annevk>
heh, thanks

[03:15:00.0000] <Philip`>
hsivonen: http://philip.html5.org/data/attr-count-pages.txt http://philip.html5.org/data/attr-count-total.txt

[03:15:01.0000] <Hixie>
aw man

[03:15:02.0000] <Hixie>
marginheight

[03:15:03.0000] <hsivonen>
Philip`: thanks

[03:15:04.0000] <Hixie>
totally didn't think of that

[03:16:00.0000] <hsivonen>
Hixie: if you make it conforming, it's easier to remember :-)

[03:16:01.0000] <Hixie>
i've already made style="" conforming, don't push your luck :-P

[03:18:00.0000] <Philip`>
hsivonen: http://philip.html5.org/data/tag-count-pages.txt http://philip.html5.org/data/tag-count-total.txt if you want to see where csobj is

[03:18:01.0000] <hsivonen>
Philip`: thank you

[03:19:00.0000] <Philip`>
(The page counts are still strongly biased by a few sites that have thousands of pages in the list)

[03:19:01.0000] <hsivonen>
I know the Web is weird and still these stats tend to surprise me in some ways

[03:20:00.0000] <Philip`>
(but most of the pages are still from sites with only one page in the list, if I remember correctly)

[03:21:00.0000] <Hixie>
hsivonen: how so?

[03:21:01.0000] <hsivonen>
Hixie: like claris cruft

[03:21:02.0000] <Hixie>
ah

[03:22:00.0000] <Hixie>
NYT is a big enough site that i see NYT-specific cruft in my data

[03:22:01.0000] <Hixie>
<nyt_copyright> and the like

[03:22:02.0000] <Hixie>
and ebay is big enough that it skews certain things, like the most common absolute url for <img src=""> is a 1x1 pixel GIF on ebay

[03:23:00.0000] <hsivonen>
I wonder what CDN load that gif causes

[03:29:00.0000] <Philip`>
Most EBay pages seem to only get a dozen viewers, so it shouldn't be as much load as the occurrence count would suggest

[03:30:00.0000] <Philip`>
/me realises that when he said "start tag" he was totally wrong, since he forgot the distinction between the tokeniser and the SAX-mode parser

[03:32:00.0000] <Philip`>
Hixie: s/occurance/occurrence/ in latest checkin

[03:33:00.0000] <Hixie>
thx

[04:16:00.0000] <zcorpan>
styles on html and body in the outer document would apply to each seamless iframe

[04:19:00.0000] <zcorpan>
perhaps doc='' should be an "in body" fragment

[04:20:00.0000] <zcorpan>
<iframe innerhtml=''>

[04:20:01.0000] <Philip`>
<iframe body=''>

[04:22:00.0000] <zcorpan>
yeah body is better

[04:22:01.0000] <zcorpan>
optionally used together with type=''

[04:23:00.0000] <zcorpan>
wonder if there's a case where one'd want to have a full html document and not just in body

[04:23:01.0000] <zcorpan>
and not the xml syntax

[04:25:00.0000] <Philip`>
Seamless framesets!

[04:25:01.0000] <hsivonen>
zcorpan: style sheet link?

[04:25:02.0000] <hsivonen>
that's something from outside the body

[04:25:03.0000] <hendry>
can't one assume utf-8 nowadays? http://googleblog.blogspot.com/2008/05/moving-to-unicode-51.html

[04:26:00.0000] <zcorpan>
hsivonen: <style scoped>@import works though

[04:26:01.0000] <zcorpan>
hsivonen: but yeah

[04:26:02.0000] <hendry>
i was just thinking why <meta charset="UTF-8"> is really needed

[04:26:03.0000] <hsivonen>
hendry: nope

[04:26:04.0000] <hendry>
hsivonen: reasoning? :)

[04:26:05.0000] <hsivonen>
hendry: we are talking about parsing a DOMString that is already UTF-16

[04:26:06.0000] <zcorpan>
hendry: i have set my browser to fall back to utf-8 and it breaks a number of pages

[04:27:00.0000] <hendry>
zcorpan: for example?

[04:29:00.0000] <hendry>
another silly question. how do I track the actual HTML rendering of this change? http://html5.org/tools/web-apps-tracker?from=1660&to=1661 in http://www.whatwg.org/specs/web-apps/current-work/multipage/

[04:29:01.0000] <Hixie>
how do you spell "cyclable"? As in, something that can be cycled.

[04:29:02.0000] <hsivonen>
Re: earlier ODF links: I think Adobe made a mistake when it tried to protect its direct Acrobat sales by asking MS not to build PDF export into Office by default

[04:30:00.0000] <Hixie>
hendry: how do you mean?

[04:30:01.0000] <Philip`>
Hixie: You look on Google to see if many other people have spelt it that way, and if they have then it's probably adequately legitimate :-)

[04:31:00.0000] <Hixie>
doesn't work

[04:31:01.0000] <hendry>
Hixie: i want a link of that diff to the particular section of http://www.whatwg.org/specs/web-apps/current-work. so i can get more context.

[04:31:02.0000] <Hixie>
"cyclable" is a common french word

[04:31:03.0000] <Hixie>
hendry: oh

[04:31:04.0000] <Hixie>
hendry: just search for the added text

[04:32:00.0000] <zcorpan>
hendry: e.g. https://ladda.telenor.se/topup.asp

[04:33:00.0000] <Philip`>
Hixie: Alternatively, you spell it whatever way you think looks sensible and then see if anyone suggests that's wrong and justifies an alternative :-)

[04:34:00.0000] <hendry>
Hixie: searching around http://www.whatwg.org/specs/web-apps/current-work/ is hard work for my Thinkpad ;) but OK. Another silly question. How do I see the revision number on http://www.whatwg.org/specs/web-apps/current-work/

[04:34:01.0000] <annevk>
hendry, svn.whatwg.org/webapps/

[04:35:00.0000] <Philip`>
hendry: You can use http://html5.org/tools/web-apps-tracker?from=1660&to=1661&context=20 if you want more context

[04:37:00.0000] <hendry>
was expecting a little SVN r1663 printed somewhere on the spec

[04:37:01.0000] <Hixie>
hendry: use a better browser :-)

[04:37:02.0000] <hendry>
Philip`: ah thanks. though I still prefer to read er "rendered HTML". 

[04:39:00.0000] <annevk>
since Hixie makes markup mistakes about every week, reading the diffs is easier :p

[04:39:01.0000] <hsivonen>
markup is tough :-)

[04:41:00.0000] <Hixie>
hendry: the one on the site isn't in subversion

[04:41:01.0000] <Hixie>
hendry: it's the working copy

[04:41:02.0000] <Hixie>
hendry: so it's always rX where X = N+1 where N is rHEAD

[04:45:00.0000] <gsnedders>
Hixie: http://trac.webkit.org/browser/trunk/WebCore/platform/text/TextCodecICU.cpp#L61 if you haven't already see that

[04:47:00.0000] <Hixie>
gsnedders: anything specific?

[04:47:01.0000] <gsnedders>
Hixie: No, just the general list of encodings there

[04:47:02.0000] <Hixie>
seems to match basically what's in the spec now

[04:48:00.0000] <Hixie>
except for TIS-620

[04:55:00.0000] <annevk>
has anyone checked whether this alias stuff also applies to text/plain and text/xml ?

[04:56:00.0000] <Hixie>
i imagine it applies to text/plain, yes

[04:56:01.0000] <Hixie>
really we should just get IANA updated

[04:57:00.0000] <annevk>
that's what i was getting at

[04:57:01.0000] <Hixie>
i'm happy to remove this stuff once IANA is updated. :-)

[04:58:00.0000] <gsnedders>
But are we going to get it updated? Unlikely.

[04:58:01.0000] <gsnedders>
annevk: text/xml certainly

[05:20:00.0000] <Philip`>
/me disapproves of the Americanization like s/serialise/serialize/ :-p

[05:21:00.0000] <hsivonen>
annevk: should the "I've read the changes!" button take me to the next diff?

[05:22:00.0000] <Hixie>
could someone who has an implementation of the encoding sniffer let me know if my latest checkin is correct?

[05:23:00.0000] <Hixie>
Philip`: me too, but i have to be consistent at least

[05:25:00.0000] <annevk>
hsivonen, no

[05:25:01.0000] <hsivonen>
ok.

[05:25:02.0000] <annevk>
hsivonen, maybe someone could add a "next diff" thingie...

[05:25:03.0000] <hsivonen>
reading diff from my mailbox then

[05:25:04.0000] <annevk>
hsivonen, it stores a cookie for your next visit

[05:30:00.0000] <hsivonen>
/me finds that feature additions aren't always annotated [c]

[05:32:00.0000] <hsivonen>
apparently <script charset> and data-* escaped my script for this reason

[05:32:01.0000] <Hixie>
yeah i'm pretty flaky

[05:33:00.0000] <Hixie>
i often don't really know what to annotate things with

[05:33:01.0000] <Hixie>
sorry about that

[05:35:00.0000] <hsivonen>
/me notes language attribute participates in processing but is not conforming

[05:36:00.0000] <hsivonen>
I wonder if I've missed anything else

[05:36:01.0000] <hsivonen>
my cursory reading suggests no

[05:38:00.0000] <hsivonen>
Hixie: did you make type attribute required for inline scripts? http://bugzilla.validator.nu/show_bug.cgi?id=186

[05:40:00.0000] <annevk>
if he did, it's a bug

[05:40:01.0000] <Hixie>
not intentionally, why?

[05:40:02.0000] <Hixie>
oh

[05:40:03.0000] <Hixie>
i made it required if it's not javascript

[05:40:04.0000] <Hixie>
or something like that

[05:40:05.0000] <hsivonen>
Hixie: "When used to include script data, the script data must be embedded inline, the format of the data must be given using the type attribute"

[05:40:06.0000] <Hixie>
shouldn't affect you

[05:40:07.0000] <Hixie>
right

[05:41:00.0000] <hsivonen>
Hixie: not as clear as it could be

[05:41:01.0000] <hsivonen>
I take it that there's no actionable change in http://bugzilla.validator.nu/show_bug.cgi?id=186 then?

[05:41:02.0000] <Hixie>
send mail saying what's confusing :-)

[05:41:03.0000] <hsivonen>
ok

[05:41:04.0000] <Hixie>
hsivonen: depends what you are testing

[05:42:00.0000] <Hixie>
hsivonen: if the type="" is application/xml or some such, you should probably test xml well-formedness or something, if you want to be perfect and cool

[05:42:01.0000] <Hixie>
but if you're treating script as a black box for now, then no, probably not

[05:43:00.0000] <hsivonen>
Hixie: black box for now

[05:43:01.0000] <hsivonen>
Hixie: except it is a text-based black box

[05:44:00.0000] <hsivonen>
Hixie: I don't support theoretical XML tree-based languages

[05:46:00.0000] <hsivonen>
/me sees Hixie defined iWeb as legitimate

[05:48:00.0000] <Hixie>
hm?

[05:48:01.0000] <Hixie>
iWeb does a lot of bad things

[05:48:02.0000] <Hixie>
i doubt i've made its output valid

[05:48:03.0000] <hsivonen>
"However, WYSIWYG tools are legitimate. WYSIWYG tools should use elements they know are appropriate, and should not use elements that they do not know to be appropriate. This might in certain extreme cases mean limiting the use of flow elements to just a few elements, like div, b, i, and span and making liberal use of the style attribute."

[05:49:00.0000] <Hixie>
something to that effect has always been there, no?

[05:50:00.0000] <hsivonen>
not sure

[05:50:01.0000] <Hixie>
i mean, we used to have a whole _section_ justifying wysiwyg editors

[05:50:02.0000] <annevk>
yeah, before it was <font>

[05:50:03.0000] <hsivonen>
I wasn't complaining :-)

[05:55:00.0000] <Philip`>
Hmm, IE has much stricter charset parsing than other browsers

[05:57:00.0000] <Hixie>
ok i'm going to bed now

[05:57:01.0000] <Hixie>
nn

[07:22:00.0000] <Lachy>
/me regrets jumping in to the alt debate again :-(

[07:24:00.0000] <hsivonen>
I'm so glad I promised myself that I stay out of it

[08:05:00.0000] <takkaria>
I love Rob Burns

[08:06:00.0000] <takkaria>
he appears to be applying literary deconstruction to Namespces in XML

[08:06:01.0000] <Dashiva>
We'll just redefine everything and have a wizard handle the backwards compatability problems

[08:07:00.0000] <MikeSmith>
yes

[08:07:01.0000] <MikeSmith>
Dashiva: there you have stumbled upon the solution to all our problems 

[08:07:02.0000] <MikeSmith>
a wizard or magic fairy

[08:09:00.0000] <Philip`>
/me grabs the fairy and bites its head off

[08:12:00.0000] <Dashiva>
I think a message saying "Can we please keep the 'how to represent missing alt data' and 'is missing alt data conforming' issues separate, and not start arguing about one in a thread actually making progress on the other?"

[08:12:01.0000] <Dashiva>
would be in order about now...

[08:13:00.0000] <annevk>
i like it how RB puts the blame for everything on these horrible implementors

[08:13:01.0000] <takkaria>
/me bingles Philip TAYLOR (Ret'd) for his latest opine

[08:13:02.0000] <Dashiva>
takkaria: Feel free the write the mail outlined above ;)

[08:15:00.0000] <takkaria>
Dashiva: the three-line posts that he makes are often not replied to, so I'll just sit and hope that's the case here

[08:18:00.0000] <MikeSmith>
/me is nostalgic for Philip TAYLOR (Webmaster) .. and for that brief moment in time when we had Philip TAYLOR (Webmaster, Ret'd) 

[08:21:00.0000] <takkaria>
I think he's very often wrong, but at least he's short and to the point. :)

[08:26:00.0000] <Lachy>
The wizard and magic fairy solution is intriguing.

[08:48:00.0000] <Lachy>
hmm. It looks like the machine-checkable vs. non-machine-checkable conformance criteria debate is yet another bikeshed to avoid.

[08:49:00.0000] <Philip`>
It's not a bikeshed, since it's a significant issue

[08:49:01.0000] <annevk>
someone just mentioned the semantic web in that context

[09:42:00.0000] <takkaria>
I'm not quite sure how JJ's machine-checkable idealism would be implemented, even after his previous posts on the issue

[09:42:01.0000] <takkaria>
cos it seems to me that if he had his way, HTML5 would require much more verbosity and have little default styling, such that no-one would actually write it over HTML4

[09:49:00.0000] <gsnedders>
takkaria: Implication is that it has to have _no_ default styling

[09:49:01.0000] <gsnedders>
takkaria: Thereby going against the principle of backwards compat.

[09:51:00.0000] <takkaria>
gsnedders: that too :)

[09:52:00.0000] <takkaria>
my point is really that the increased verbosity it would require would mean that you need more bytes to get the same effect so no one would actually use it

[09:53:00.0000] <Dashiva>
And someone would then implement <body default-style> to use HTML4 styles :)

[09:53:01.0000] <Philip`>
That's why browsers have to remove support for HTML4 features so that everyone will rapidly transition into the exciting new world of the semantic web

[09:56:00.0000] <takkaria>
someone should probably point him towards XHTML2...

[09:57:00.0000] <annevk>
HT doesn't seem to be willing to accept reality just yet... :(

[10:02:00.0000] <annevk>
I wonder if HT is suggesting that browsers dispatch on nodeName rather than localName + namespaceURI

[10:03:00.0000] <annevk>
That's near insane

[10:05:00.0000] <Philip`>
Seems kind of odd to throw out the whole namespace URI thing just so you can keep something that looks the same as the colon syntax

[10:06:00.0000] <Dashiva>
It's to preserve the integrity of namespaces that we're rewriting namespaces

[10:06:01.0000] <Dashiva>
Isn't that obvious?

[10:07:00.0000] <takkaria>
the first rule of namespaces is that you do not follow Namespaces

[10:07:01.0000] <Dashiva>
What's the second rule?

[10:07:02.0000] <Lachy>
there is no second rule

[10:08:00.0000] <Dashiva>
Is that a rule?

[10:09:00.0000] <Philip`>
It's more of a guideline

[10:12:00.0000] <takkaria>
except for the evil implementors

[10:15:00.0000] <Dashiva>
The vast browser-wing conspiracy

[10:17:00.0000] <Philip`>
Does there exist a list of conspiracies that were actually real, and worked successfully for quite a while?

[10:17:01.0000] <Dashiva>
Yes

[10:17:02.0000] <Dashiva>
I read an article about it yesterday, actually

[10:17:03.0000] <Philip`>
It'd be interesting to see how vast a conspiracy could get, and how the probability of keeping it secret varies with size

[10:19:00.0000] <takkaria>
Mozilla is especially to blame for the browser conspiracy.  they only look like they're open

[10:20:00.0000] <Philip`>
But Opera is closed source and for-profit so it must be transmitting all your browsing details to the NSA

[10:21:00.0000] <Dashiva>
Don't forget that they all send your browsing data to these so-called "anti-phising" services

[10:21:01.0000] <Philip`>
Firefox doesn't (by default) :-p

[10:22:00.0000] <Dashiva>
That's what they want you to think

[10:22:01.0000] <Philip`>
/me assumes Opera does, but isn't sure

[10:22:02.0000] <Dashiva>
It just means they hide it

[10:22:03.0000] <Philip`>
Dashiva: Where do they hide it?

[10:22:04.0000] <Dashiva>
ICMP

[10:22:05.0000] <Philip`>
...

[10:22:06.0000] <takkaria>
they actually install a trojan network stack on Windows machines so you can't detect it

[10:22:07.0000] <Dashiva>
The payload of PING can be used for data transfer just fine

[10:22:08.0000] <Philip`>
takkaria: I'm running Linux

[10:23:00.0000] <takkaria>
Philip`: they don't do so well there

[10:24:00.0000] <Philip`>
Also I'll claim that I'm running Gentoo and compiled Firefox from source, so I wouldn't be affected by secret hidden code in the distributed binary

[10:24:01.0000] <Dashiva>
And you read the source? :)

[10:25:00.0000] <Philip`>
No, but millions of other eyes have

[10:26:00.0000] <Dashiva>
Yeah, and those millions of eyes totally caught that debian vuln last week ;)

[10:26:01.0000] <Philip`>
(Actually I used Gentoo's mozilla-firefox-bin package which downloads the official Mozilla binary instead, because I didn't fancy waiting an hour for it to compile after every minor version upgrade...)

[10:26:02.0000] <Philip`>
Dashiva: Indeed, they did catch it last week, which proves that vulnerabilities always get noticed :-)

[10:26:03.0000] <Dashiva>
In other news, Henry is quite optimistic about HTML5 it seems

[14:03:00.0000] <Dashiva>
"Oh, sure, everyone on the entire internet uses flickr, but it's only one site so we'll ignore it"

[14:04:00.0000] <Lachy>
I would respond, but I don't want to fuel the fire any more

[14:05:00.0000] <Lachy>
John just doesn't realise there is a difference between choosing not to optimise for a situation and choosing not to deal with it at all

[14:05:01.0000] <Dashiva>
I guess specs are an edge case too, then :)

[14:06:00.0000] <Lachy>
every site is an edge case, since it's only one out of billions :-)

[14:07:00.0000] <jgraham>
Hey it's marginally more sane that Rob Burns explaining why the sky is green if you squint at it just right

[14:07:01.0000] <Lachy>
but it's not just Flickr. It's photobucket, my opera, .mac and every other site that publishes user generated content

[14:07:02.0000] <jgraham>
I should emphasise the *marginally* there :)

[14:07:03.0000] <Lachy>
jgraham, it depends on the defintion of "green".

[14:08:00.0000] <Dashiva>
Don't come here with your fancy "facts", Lachy. We don't need any of those.

[14:09:00.0000] <Lachy>
and it also depends on whether you're referring to the whole sky at all times during the day, or the brief moment it transistions from blue to orange at sunset

[14:09:01.0000] <hober>
yes, we all need to mind the /topic :)

[14:13:00.0000] <jgraham>
Well didn't his last email basically say that the fact that the sky is blue should be ignored because the process for deciding sky colour should favour whatever colour he likes and he likes green therefore the sky is green

[14:15:00.0000] <Lachy>
I haven't read that mail yet. Is it a recent one or old one?

[14:19:00.0000] <Philip`>
Lachy: I think it's several of them

[14:22:00.0000] <tndH>
apparently Boris Zbarsky is underestimating implementors...

[14:24:00.0000] <gavin_>
Boris Zbarsky is an implementor

[14:25:00.0000] <tndH>
yeah, not sure if Robert Burns is taking that into account though.

[14:25:01.0000] <Philip`>
It's not impossible for people to underestimate themselves :-)

[14:27:00.0000] <othermaciej>
maybe Boris is just being humble

[14:27:01.0000] <jwalden>
I don't think so

[14:30:00.0000] <syp_>
jwalden: hello, did you see my reply to your mail about the browser tests?

[14:30:01.0000] <jwalden>
syp_: I noticed it when skimming through email, problem being it got caught at the end of an existing thread and thus didn't show up as visible -- sec, I'll read

[14:31:00.0000] <syp_>
yes, I replied on the implementors list while the original was on the other list

[14:38:00.0000] <Hixie>
am i totally misunderstanding cwilson's attack vector in waf

[14:39:00.0000] <Hixie>
or is he just describing something that's a subset of a much bigger problem

[14:40:00.0000] <annevk>
he is, and that was already pointed out some time ago when Maciej brought that issue up

[14:41:00.0000] <annevk>
having said that, I'm not too comfortable yet with DNS rebinding / man-in-the-middle / etc.

[14:42:00.0000] <annevk>
I should put a filter in place for RB

[14:43:00.0000] <annevk>
I get annoyed when reading his e-mails and that's not good

[14:43:01.0000] <Hixie>
hehe

[14:44:00.0000] <othermaciej>
from last time I don't think there is a DNS rebinding vulnerability but I guess I'll have to slog through his email

[14:44:01.0000] <Hixie>
he's not very detailed and i already replied

[14:44:02.0000] <Hixie>
but if you can find something i missed, pelase let me know

[14:45:00.0000] <annevk>
though Chris understanding of the vulnarability is different from Maciej's

[14:45:01.0000] <annevk>
Maciej's was about being able to do POST after a DNS rebinding attack on the OPTIONS

[14:46:00.0000] <annevk>
Chris seems to be talking about exposing credentials which seems to imply he didn't even read the proposal carefully

[14:46:01.0000] <Hixie>
yeah

[14:47:00.0000] <othermaciej>
I think Chris is describing what I described, but in a more vague and handwavey way

[14:47:01.0000] <othermaciej>
"This enables the vector of DNS attacks - the idea being that between those two connections, an attacker could insert themselves in to the stream.  (Actually, more likely it would be the other way around - an attacker would insert themselves into the stream, give back "it's okay to do x-domain", then release and let the real site give back data."

[14:47:02.0000] <othermaciej>
I don't remember how we concluded that this was not a real vulnerability

[14:48:00.0000] <othermaciej>
Hixie: DNS rebinding works in reverse

[14:48:01.0000] <othermaciej>
Hixie: the DNS infrastructure does not have to be compromised to do a DNS rebinding attack

[14:48:02.0000] <jgraham>
/me considers replying to Justin James, realises that insanity is doing the same thing over and over but expecting different results, decides not to bother

[14:48:03.0000] <Hixie>
it does against HTTP 1.1 servers

[14:49:00.0000] <Hixie>
Host: safeguards against DNS rebinding from other names

[14:49:01.0000] <othermaciej>
checking the Host header is indeed a full defense against DNS rebinding

[14:49:02.0000] <othermaciej>
if you check it

[14:49:03.0000] <othermaciej>
which many (most) existing servers don't

[14:49:04.0000] <Hixie>
if you don't check it, XHR is the _least_ of your problems

[14:49:05.0000] <annevk>
"RE: Work that would be required to revert Internet Explorer 8 to  'aria:'" wtf!

[14:49:06.0000] <othermaciej>
checking the Host header is indeed a defense against the DNS rebinding attack against the preflight chec

[14:49:07.0000] <annevk>
why is the W3C so fucked up?

[14:50:00.0000] <Hixie>
if you don't check Host:, the simplest attack vector is just to serve the JS file from hostile.com and the rest of the resource from victim.hostile.com

[14:50:01.0000] <Hixie>
no need for XHR

[14:52:00.0000] <annevk>
(fortunately Chris is not interested in the idea)

[14:52:01.0000] <Lachy>
how does checking Host: prevent an attacker from intercepting a request and serving a hostile JS file?

[14:53:00.0000] <othermaciej>
ok let me do the short lecture on DNS Rebinding

[14:53:01.0000] <othermaciej>
DNS rebinding works like this

[14:53:02.0000] <othermaciej>
you access http://attacker.com/

[14:53:03.0000] <othermaciej>
attacker.com has a very short DNS expiration time

[14:54:00.0000] <othermaciej>
a little while after access, a DNS change propagates so that attacker.com points to the same IP as victim.com

[14:54:01.0000] <othermaciej>
attacker.com loads some victim.com resources

[14:54:02.0000] <othermaciej>
in general this only works when the only access control to victim.com is network position (for example it is behind a firewall)

[14:54:03.0000] <othermaciej>
since you end up sending attacker.com cookies, not victim.com cookies

[14:55:00.0000] <othermaciej>
if victim.com checks the Host header, then even this is ineffective

[14:55:01.0000] <othermaciej>
for the same reason, a DNS rebinding attack against the method check preflight would be ineffective, because attacker.com cookies would be sent for the POST, not victim.com cookies

[14:55:02.0000] <othermaciej>
so you could only do what XDR would always let you do (POST without credetials)

[14:56:00.0000] <jwalden>
Host: checking is required by HTTP/1.1 -- are servers just non-compliant?

[14:56:01.0000] <Dashiva>
You can have a * host that catches everything

[14:57:00.0000] <Hixie>
it's not so much that com cookies would be sent for the POST, so much as a DNS rebinding attack against the method check preflight would not grant you any more access than simply doing a 1995-era cross-site <form> submit

[14:57:01.0000] <annevk>
jwalden, are browsers just non-compliant? madness!

[14:57:02.0000] <othermaciej>
is anything compliant to anything?

[14:57:03.0000] <Lachy>
othermaciej, I thought the problem was if attacker.com impersonated victim.com. So that when a user loads victim.com in their browser and then attacker.com uses DNS rebinding to intercept some of the requests and send back malicious files

[14:57:04.0000] <jwalden>
well, browsers didn't have a spec

[14:57:05.0000] <jwalden>
and you have to check to do vhosting correctly

[14:58:00.0000] <jwalden>
when you have it set up to serve many origins from one server

[14:58:01.0000] <annevk>
everything is probably inherently imperfect due to human nature

[14:58:02.0000] <othermaciej>
Lachy: yeah, that's not DNS Rebinding

[14:58:03.0000] <Lachy>
ok

[14:58:04.0000] <annevk>
that's man-in-the-middle iirc

[14:59:00.0000] <othermaciej>
Lachy: that would require compromising DNS itself, and if you have that, you don't need any form of cross-site access

[14:59:01.0000] <othermaciej>
(running a malicious DNS server on an open wireless network would be one form of this attack)

[14:59:02.0000] <Hixie>
if you can mitm the last router, you can just inject whatever you want straight into the http stream

[14:59:03.0000] <othermaciej>
generally SSL protects against that, unless the user blindly clicks through on invalid certs

[14:59:04.0000] <Hixie>
(poking with dns would be far more complicated)

[15:00:00.0000] <annevk>
oh, i guess man-in-the-middle is more a malicious proxy

[15:00:01.0000] <Lachy>
othermaciej, the problem is the users do blinding accept invalid certs

[15:00:02.0000] <Lachy>
*blindly

[15:00:03.0000] <Hixie>
ff3 makes it nigh on impossible to click through an invalid cert

[15:00:04.0000] <othermaciej>
Lachy: I'd really like to make Safari always reject self-signed certs but I am not sure marketing would agree

[15:00:05.0000] <Hixie>
othermaciej: just do what ff3 does :-)

[15:00:06.0000] <jwalden>
it's just five clicks or so :-)

[15:01:00.0000] <jwalden>
through arcane security UI :-)

[15:01:01.0000] <Hixie>
five pretty unobvious clicks

[15:01:02.0000] <Hixie>
yeah

[15:01:03.0000] <Hixie>
i've had to do it numerous times :-)

[15:01:04.0000] <jwalden>
"feature, not bug" :-)

[15:03:00.0000] <annevk>
if it happens often for the user he'll make a habbit out of it and you'll have the same problem

[15:03:01.0000] <gsnedders>
othermaciej: I've got a couple of bug reports through SP not accepting self-signed certs, but something like Saf has the advantage of being able to prompt the user at least

[15:03:02.0000] <othermaciej>
we do prompt

[15:03:03.0000] <othermaciej>
prompts are kinda useless

[15:05:00.0000] <gsnedders>
othermaciej: Yeah, well all something like SP can do is either outright accept it or outright refuse it.

[15:06:00.0000] <Hixie>
i'm pretty sure i just saw an e-mail on this alt thread claiming 50 years of experience

[15:06:01.0000] <Hixie>
the web has only been around for 18...

[15:07:00.0000] <Dashiva>
Accessible file cabinets, maybe?

[15:08:00.0000] <gsnedders>
I can't even claim 18 years of experience of life

[15:08:01.0000] <Dashiva>
Not even counting the pre-birth months?

[15:08:02.0000] <gsnedders>
/me wonders how many under 18 y/os there are

[15:08:03.0000] <gsnedders>
Dashiva: Not even

[15:09:00.0000] <gsnedders>
(I was born on 19920420)

[15:09:01.0000] <Dashiva>
Well, then... get off my lawn!

[15:10:00.0000] <gsnedders>
I'm not on  a lawn, yet alone your lawn.

[15:10:01.0000] <Dashiva>
You're a young'un, you're always on my lawn

[15:16:00.0000] <Philip`>
Dashiva: I think you need to do some gardening, since lawns really shouldn't look like carpet tiles

[15:16:01.0000] <Dashiva>
I thought the British prided themselves on lawns that were suitable for any kind of activity

[15:17:00.0000] <Philip`>
Only croquet is permitted

[15:17:01.0000] <jgraham>
And even that only during the summer

[15:19:00.0000] <gsnedders>
Hahaha

[15:19:01.0000] <gsnedders>
See, that's why you to college that actually allows you to use the lawn.

[15:19:02.0000] <gsnedders>
(Apart from for croquet)

[15:20:00.0000] <Lachy>
he must have been referring to the time when the web was implemented over snail mail and alt text was written on the back of photos called "postcards"

[15:21:00.0000] <gsnedders>
Oh, the passage we had in the French exam stated something like (to translate the French): "She posts a blog on the internet…"

[15:21:01.0000] <gsnedders>
Unlike all these people who blog on paper

[15:21:02.0000] <Dashiva>
gsnedders: intranet

[15:21:03.0000] <Dashiva>
duh

[15:22:00.0000] <hendry>
 /away sssssssssleep

[15:22:01.0000] <gsnedders>
I should do that, too.

[15:22:02.0000] <annevk>
/me tries campaigning against <timerange>

[15:22:03.0000] <gsnedders>
Especially seeming I have a physics exam and Hixie is "suggesting" I do physics at uni

[15:23:00.0000] <Philip`>
Sleep is a waste of time that could be used for last-minute revision

[15:23:01.0000] <jgraham>
Last minute revision is a waste of time that could be used for sleep

[15:23:02.0000] <annevk>
what's up with all these linux security updates

[15:23:03.0000] <gsnedders>
Philip`: Apart from when you have CFS and would therefore fall asleep during the three hour exam

[15:23:04.0000] <gsnedders>
annevk: Linux is insecure?

[15:24:00.0000] <annevk>
tthe changelogs are also completely incomprehensible

[15:24:01.0000] <annevk>
i can't believe i talked my mom into this, though I guess Windows updates are just as bad

[15:25:00.0000] <Philip`>
The Windows updates just hide all the technical details, which isn't really an improvement

[15:25:01.0000] <gsnedders>
OS X is reporting 0KB free on /. I don't think that' good.

[15:25:02.0000] <Philip`>
gsnedders: Slashdot has run out of space?

[15:25:03.0000] <jgraham>
annevk: This is ubuntu, right? I think I have had three batches of security updates recently including the original SSL one

[15:26:00.0000] <annevk>
indeed

[15:26:01.0000] <gsnedders>
Philip`: No, on root, followed by full-stop

[15:26:02.0000] <jgraham>
Philip`: When slashdot runs out of space all the intertubes run out of space </bagpuss>

[15:27:00.0000] <annevk>
I wonder what my mom will think of it

[15:27:01.0000] <annevk>
if she even notices the difference :)

[15:27:02.0000] <jgraham>
annevk: I guess security updates tend to come in batches as one hole is noticed any then everyone else goes on a security audit for a bit

[15:30:00.0000] <gsnedders>
/me is rather amazed by the fact he's only ever got one security hole found in SP. There have to be more!

[15:31:00.0000] <jgraham>
gsnedders: Maybe you have too few users for anyone to care :)

[15:31:01.0000] <gsnedders>
jgraham: We have plenty of users :)

[15:32:00.0000] <Philip`>
gsnedders: Maybe you published the wrong email address for reporting security vulnerabilities, so all the reports are getting lost

[15:32:01.0000] <gsnedders>
Philip`: No

[16:36:00.0000] <Lachy>
re this post about fragment identifiers http://lists.w3.org/Archives/Public/public-html/2008May/0509.html ...

[16:37:00.0000] <Lachy>
couldn't xpointer just be fixed up and defined in terms of the DOM, so that it would work the same in HTML as it would with XHTML?

[16:37:01.0000] <Hixie>
xpointer is not resilient to changes

[16:37:02.0000] <Hixie>
which is imho a pretty important requirement for this idea

[16:38:00.0000] <Dashiva>
xpathref!

[16:38:01.0000] <annevk>
same issue

[16:38:02.0000] <Hixie>
hsivonen: i'm not replying to your mail "Re: A comment to character encoding declaration" earlier today since i've sent mail on those subjects already

[16:38:03.0000] <Hixie>
let me know if i missed something

[16:38:04.0000] <Hixie>
oh actually nevermind

[16:38:05.0000] <Hixie>
i'll just reply

[16:41:00.0000] <Lachy>
would it be possible to fix xpointer so that it is resilliant to changes? Since it already exists, and if this turns out to be a valid use case to address, wouldn't it be better to improve what we have instead of defining somethng totally new?

[16:43:00.0000] <annevk>
i don't think anything would be resilliant to changes other than id=

[16:43:01.0000] <annevk>
and even that isn't so maybe that point is moot

[16:44:00.0000] <annevk>
anyway, one university guy suggesting a feature does not a use case make

[16:44:01.0000] <Dashiva>
There is a userjs for it :)

[16:47:00.0000] <Hixie>
ok what spec defines the syntax of mime media types?

[16:47:01.0000] <Hixie>
rfc2045 doesn't seem to define them separate from cotent-type headers

[16:47:02.0000] <Hixie>
and doesn't seem to define quoted-string at all

[16:48:00.0000] <Hixie>
hmm, rfc2616 (http) seems to define them

[16:48:01.0000] <Hixie>
i can use that


2008-05-23
[17:03:00.0000] <Hixie>
Philip`: yt?

[17:06:00.0000] <Hixie>
eh nm

[17:06:01.0000] <Hixie>
well

[17:06:02.0000] <Hixie>
when you get back, could you add some tests to http://philip.html5.org/demos/html/charset-parsing/ ?

[17:06:03.0000] <Hixie>
specifically:

[17:07:00.0000] <Hixie>
<meta http-equiv="content-type" content="text/html; charset='...'">

[17:07:01.0000] <Hixie>
<meta http-equiv="content-type" content='text/html; charset="..."'>

[17:07:02.0000] <Hixie>
<meta http-equiv="content-type" content='text/html; test="; charset="..."'>

[17:08:00.0000] <Hixie>
<meta http-equiv="content-type" content='text/html; test="test"; charset="..."'>

[17:08:01.0000] <Hixie>
<meta http-equiv="content-type" content='text/html; test="; charset=...'>

[17:08:02.0000] <Hixie>
<meta http-equiv="content-type" content='text/html; test="test"; charset=...'>

[17:09:00.0000] <Hixie>
<meta http-equiv="content-type" content="text/html; test='; charset=...">

[17:09:01.0000] <Hixie>
<meta http-equiv="content-type" content="text/html; test='test'; charset=...">

[17:10:00.0000] <annevk>
/me wonders why JB didn't join the FTF himself

[17:10:01.0000] <Philip`>
Hixie: Hello

[17:10:02.0000] <Hixie>
hello

[17:10:03.0000] <Philip`>
/me updates his list

[17:12:00.0000] <annevk>
heh, your file names are hex numbered

[17:15:00.0000] <Philip`>
Hixie: Updated

[17:15:01.0000] <Philip`>
annevk: Oops, that was unintentional

[17:15:02.0000] <Philip`>
I think I intended %02d, but I wrote %02x

[17:16:00.0000] <Hixie>
Philip`: thanks

[17:16:01.0000] <annevk>
would %03d be three digits?

[17:16:02.0000] <Philip`>
Hmm, my default font isn't great at distinguishing " and ''

[17:16:03.0000] <Hixie>
something seems to be wrong

[17:16:04.0000] <Philip`>
annevk: Yes (zero-padded)

[17:17:00.0000] <Philip`>
Hixie: Could you be slightly more specific?

[17:17:01.0000] <Hixie>
i'm trying to figure out what is wrong :-)

[17:17:02.0000] <Philip`>
Do you just have a vague sense of unease, and no idea what the problem is?

[17:17:03.0000] <hober>
annevk: IIRC, at the time, it was believed that ((both wgs should have chairs on the ftf) xor (neither wg should have chairs on the ftf))

[17:17:04.0000] <Hixie>
the lines appear to be duplicated

[17:18:00.0000] <Hixie>
rom the first <meta charset="text/html; charset=iso-8859-1">: not utf-8 to the end seems to just be a duplication of the earlier set

[17:18:01.0000] <Philip`>
/me notes that he gets a lot of "service temporarily unavailable" when loading the charset page in IE, but that's Dreamhost's fault

[17:18:02.0000] <Hixie>
nevermind

[17:18:03.0000] <Hixie>
cache problem

[17:19:00.0000] <Philip`>
Ah

[17:19:01.0000] <Philip`>
I added the new ones kind of near the top, so all the others got renumbered

[17:19:02.0000] <annevk>
hmm, seems that implementations simply scan for "charset="

[17:20:00.0000] <Hixie>
ok webkit is clearly just doing that, indeed

[17:20:01.0000] <Hixie>
IE, not so much

[17:20:02.0000] <annevk>
Opera and Firefox seem to be doing that too

[17:21:00.0000] <annevk>
(that's the two I tested, I didn't check IE or WebKit)

[17:21:01.0000] <Hixie>
firefox uses the same algorithm on charset="" as content=""

[17:21:02.0000] <Hixie>
webkit does what html5 says

[17:21:03.0000] <Hixie>
for charset=""

[17:22:00.0000] <Hixie>
opera freaked out on me the first time, let's try again

[17:22:01.0000] <Hixie>
ok yeah opera's scrollbar can't handle that page, wtf

[17:22:02.0000] <annevk>
"<meta charset="text/html; charset=utf-8">: not utf-8 " is what I get in Firefox

[17:22:03.0000] <Hixie>
opera does the same as webkit except it checks the http-equiv attribute

[17:22:04.0000] <Hixie>
what build?

[17:23:00.0000] <annevk>
"Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9b5pre) Gecko/2008031504 Minefield/3.0b5pre"

[17:23:01.0000] <Hixie>
weird

[17:24:00.0000] <Hixie>
also weird, IE8 supports charset='...' but not charset="...", the latter of which is the only valid one...

[17:24:01.0000] <annevk>
so my Firefox builds seems to have a different algo for charset=

[17:25:00.0000] <annevk>
I think both Opera and IE check http-equiv

[17:27:00.0000] <Hixie>
there are now three occurances of the term "willful violation" in the spec

[17:30:00.0000] <annevk>
spec lawyers rejoice

[17:32:00.0000] <Dashiva>
If you violate a should for a valid reason, is it still considered a violation?

[17:35:00.0000] <annevk>
nope

[17:39:00.0000] <annevk>
http://signified.com.au/a-poem-element-for-html5/

[17:41:00.0000] <MikeSmith>
/me awakens and confronts long scrollback buffer for #whatwg in irc client

[17:42:00.0000] <annevk>
MikeSmith, we don't have membership fees, we just want all your time :)

[17:44:00.0000] <Dashiva>
MikeSmith: I could bring in a comic chat client to make amusing comic strip representations of the logs

[17:51:00.0000] <MikeSmith>
/me completes reading of scrollback

[17:52:00.0000] <MikeSmith>
/me records data point: 6 hours of #whatwg scrollback took about 5 minutes to read

[17:54:00.0000] <Hixie>
so about 1 minute an hour

[17:54:01.0000] <Hixie>
not bad

[17:55:00.0000] <Philip`>
Took much longer for us to write all that

[17:55:01.0000] <MikeSmith>
annevk: re: time -- I find that the time I choose to invest in #whatwg discussions and reading usually pays better returns than time I'm obligated to spend on some other things

[17:56:00.0000] <MikeSmith>
Dashiva: yes, comic chat client, please.. could recruit the Wizard and/or Magic Fairy to do that, maybe

[17:57:00.0000] <Dashiva>
Microsoft already made it, many years ago

[17:57:01.0000] <MikeSmith>
cheers to Microsoft for that

[18:02:00.0000] <MikeSmith>
btw, 'twas I that suggested to Erik Wilde that he join the HTML WG and post his thoughts on fragments IDs

[18:02:01.0000] <MikeSmith>
though I do recognize the heinous-ness of xpointer will prevent it from being practical for this case

[18:06:00.0000] <MikeSmith>
and on the topic of joining the HTML WG and public-html list, questions I get from people now and then are, Can I get the list messages in digest form, or do you have a "no mail" option, or [some other way that they can avoid getting all the list mail but still remain in the group]

[18:06:01.0000] <MikeSmith>
I'm not sure what is the right thing to tell people

[18:07:00.0000] <Hixie>
"join the whatwg list, it has both of those options"

[18:08:00.0000] <MikeSmith>
Hixie: yeah, I always tell people that anyway

[18:08:01.0000] <Hixie>
:-)

[18:10:00.0000] <MikeSmith>
about public-html, I usually say, Good opportunity/test-case for you to learn about the client-side filtering capabilities in your mail client .. or to switch to a mail client that has good filtering

[18:33:00.0000] <MikeSmith>
http://lists.w3.org/Archives/Public/www-archive/2008May/0056.html

[18:33:01.0000] <MikeSmith>
saved for re-use later

[18:33:02.0000] <MikeSmith>
my "boilerplate response about mailing-list traffic volume"

[18:33:03.0000] <MikeSmith>
which took roughly 15 minutes for me to think up and write

[18:35:00.0000] <MikeSmith>
longer that I thought it would, but 15 minutes less than it will take me next time I get the same question

[18:36:00.0000] <MikeSmith>
I guess I should put that (in another form) up on the HTML WG website somewhere

[18:38:00.0000] <jacobolus>
MikeSmith: I'm not sure “you'll get tons of spam, and it will be a good life/technical lesson to deal with it” is really so compelling

[18:51:00.0000] <MikeSmith>
jacobolus: patches welcome

[18:54:00.0000] <jacobolus>
MikeSmith: how about “each email on the list contains a pearl of wisdom which will brighten your life: if you find the amount of traffic overwhelming, it will only be because you are insufficiently appreciative of your good fortune”

[18:58:00.0000] <MikeSmith>
jacobolus: about pearls of wisdom, I already have a site for automatically dispensing those:

[18:58:01.0000] <MikeSmith>
http://logopoeia.com/wisdom/

[18:58:02.0000] <jacobolus>
whoa! Our DNA was altered to conceal the cosmic secret of timbral magnetism raster mathematics.

[19:00:00.0000] <jacobolus>
5 steps include: “Memorize homeopathic law,” and “Make time law your guide.”

[19:00:01.0000] <jacobolus>
I want to know what “homeopathic law” and “time law” are, so I can follow these

[19:14:00.0000] <MikeSmith>
so I think I have come to terms with twitter and found what is for me at least a way to use it productively that doesn't require a lot of time:

[19:14:01.0000] <MikeSmith>
http://twitter.com/sideshowbarker

[19:16:00.0000] <MikeSmith>
mainly as a personal+social bookmarking mechanism

[19:31:00.0000] <Hixie>
anne and mjs should read para2 of http://lists.w3.org/Archives/Member/w3c-html-cg/2008AprJun/0082.html

[19:32:00.0000] <Hixie>
pity that the guy anne cited above on poems doesn't buy my line that <p> is defined as meaning "stanza"

[19:43:00.0000] <Dashiva>
cg?

[19:49:00.0000] <Hixie>
coordination group

[02:07:00.0000] <hsivonen>
what's the relationship of the proposed zip: URI scheme and the existing jar: URI scheme?

[02:11:00.0000] <annevk>
jar: opens up the ZIP file and then does addressing while zip: would be an internal thing afaict

[02:17:00.0000] <Hixie>
othermaciej, annevk: para 2 of http://lists.w3.org/Archives/Member/w3c-html-cg/2008AprJun/0082.html

[02:18:00.0000] <othermaciej>
awesome

[02:18:01.0000] <annevk>
yeah, I read that before I want to bed

[02:19:00.0000] <annevk>
I also wrote something in this channel "/me wonders why JB didn't join the FTF himself"

[02:19:01.0000] <Hixie>
ah

[02:23:00.0000] <othermaciej>
I need to turn my list of possible architectural consistency principles into some sort of draft

[02:24:00.0000] <othermaciej>
I have so much (English) writing to do

[02:24:01.0000] <othermaciej>
also need to justify our ES4 votes

[02:24:02.0000] <othermaciej>
and reply on Hixie's thread about forward-compatible syntax for ES4

[02:24:03.0000] <othermaciej>
and call bullshit on chris wilson's security comments

[02:24:04.0000] <othermaciej>
ugh

[02:24:05.0000] <othermaciej>
gonna be a long weekend

[02:24:06.0000] <Hixie>
you're not supposed to do this stuff on the weekend

[02:25:00.0000] <Hixie>
you're supposed to run pedestrians down in GTA4 on the weekend

[02:25:01.0000] <othermaciej>
but I don't have time to get anything done during the week

[02:25:02.0000] <Hixie>
ah well

[02:25:03.0000] <Hixie>
neither do i

[02:25:04.0000] <othermaciej>
the wonders of management

[02:25:05.0000] <othermaciej>
actually I coded a lot this week

[02:25:06.0000] <othermaciej>
but making JavaScript run really fast is actually kind of fun

[02:27:00.0000] <Hixie>
is JS execution a bottleneck on the web?

[02:27:01.0000] <roc>
yes

[02:27:02.0000] <Hixie>
i thought dom calls were by far the bottleneck in js execution

[02:28:00.0000] <roc>
sometimes they are

[02:28:01.0000] <Hixie>
interesting

[02:28:02.0000] <roc>
that's the great thing about the Web. everything's a bottleneck, depending on what you look at

[02:29:00.0000] <Hixie>
heh

[02:30:00.0000] <othermaciej>
Hixie: both can be

[02:30:01.0000] <othermaciej>
our DOM has somewhat less room to speed it up

[02:30:02.0000] <othermaciej>
at least the core DOM

[02:30:03.0000] <MikeSmith>
so what good benchmarking mechanisms are there for measuring DOM performance?

[02:30:04.0000] <MikeSmith>
I know about those for measuring JS perf

[02:31:00.0000] <othermaciej>
I don't know if any are great

[02:31:01.0000] <annevk>
entities are safe for now, nice!

[02:31:02.0000] <othermaciej>
this is a clean microbenchmark: http://www.hixie.ch/tests/adhoc/perf/dom/artificial/core/001.html

[02:31:03.0000] <MikeSmith>
othermaciej: thanks

[02:31:04.0000] <othermaciej>
testing the non-querySelector non-XPath mode of JS selector libraries is a more realistic use pattern

[02:31:05.0000] <othermaciej>
though one we are hoping to obsolete

[02:32:00.0000] <othermaciej>
quirksmode.org also has some DOM perf tests but not rolled up nicely

[02:32:01.0000] <Hixie>
the ones in http://www.hixie.ch/tests/adhoc/perf/dom/real-world/ should be better, they're based on actual web sites

[02:32:02.0000] <othermaciej>
(all differnet ways of building a table DOM - amusingly innerHTML wins in every browser)

[02:34:00.0000] <Hixie>
someone really should make a job out of comparing browsers on a suite of benchmarks on reference hardware with care to making the tests fair, and then publishing hte results in graphical form regularly

[02:34:01.0000] <othermaciej>
doing one operation (with the same args) a bunch of times may not be representative of DOM performance generally speaking

[02:34:02.0000] <Hixie>
true

[02:34:03.0000] <othermaciej>
since caching means that mixed operations are often a tougher workload

[02:35:00.0000] <othermaciej>
Hixie: that would be nice - I am not enjoying resorting to making our own benchmarks

[02:35:01.0000] <MikeSmith>
holy god

[02:35:02.0000] <othermaciej>
but the market for independent browser benchmarks seems to have died

[02:35:03.0000] <Hixie>
yeah

[02:35:04.0000] <MikeSmith>
othermaciej: I can see why you might prefer that benchmark :)

[02:35:05.0000] <othermaciej>
also a lot of the perf/dom/real-world tests are really repaint tests

[02:36:00.0000] <othermaciej>
MikeSmith: if you compared WebKit to other browsers on that, I think the results it shows are real - we have done quite well since before we ever saw that benchmark

[02:36:01.0000] <Hixie>
a lot of the real world dom work relies on painting

[02:36:02.0000] <MikeSmith>
othermaciej: wow

[02:36:03.0000] <Hixie>
but yes, those tests are just a start

[02:36:04.0000] <MikeSmith>
extremely impressive

[02:36:05.0000] <othermaciej>
Hixie: sure, repaint performance is really important

[02:37:00.0000] <othermaciej>
separating things is kind of nice, at least for browser developers, though I am not sure what level of perf info granularity is interesting to content authors or end users

[02:38:00.0000] <othermaciej>
MikeSmith: this is the quirksmode test I was thinking of: http://www.quirksmode.org/dom/innerhtml.html

[02:38:01.0000] <othermaciej>
(his posted numbers are old though you gotta try it yourself)

[02:39:00.0000] <MikeSmith>
othermaciej: thanks

[02:51:00.0000] <othermaciej>
Hixie: I can also tell you that on mobile devices JS execution is even more likely to be the bottleneck in web page performance (though not as much as the network itself)

[02:53:00.0000] <MikeSmith>
othermaciej: that's interesting.. why is that?

[02:53:01.0000] <MikeSmith>
why in particular on mobile

[02:53:02.0000] <MikeSmith>
because CPU-bound?

[02:54:00.0000] <MikeSmith>
othermaciej: and by "more" you mean specifically vs. DOM performance?

[02:54:01.0000] <roc>
cpus are slow

[02:54:02.0000] <othermaciej>
well I can't tell you that much about the mobile device I know most about, but yes, the CPUs tend to be less powerful

[02:54:03.0000] <Hixie>
othermaciej: interesting

[02:54:04.0000] <roc>
and they don't speculate well

[02:54:05.0000] <othermaciej>
ARM has practically no branch prediction

[02:54:06.0000] <Hixie>
that explains all the wild rumours

[02:54:07.0000] <othermaciej>
what wild rumors?

[02:55:00.0000] <Hixie>
about the iphone!

[02:55:01.0000] <roc>
the JS performance work we did for Firefox 3 sped up mobile by a greater factor than on PCs

[02:55:02.0000] <MikeSmith>
dual-core iPhone coming

[02:55:03.0000] <othermaciej>
there was a rumor that JS is slow on the iPhone?

[02:55:04.0000] <Hixie>
if they speculated better, the rumours would have been more accurate!

[02:55:05.0000] <MikeSmith>
heh

[02:55:06.0000] <roc>
hoho

[02:55:07.0000] <roc>
don't give up your day job

[02:56:00.0000] <MikeSmith>
/me just now gets Hixie joke.. (I'm slow on teh uptake)

[02:56:01.0000] <Hixie>
/me goes back to rejecting suggestions

[02:56:02.0000] <roc>
cool, he took my advice!

[02:57:00.0000] <Hixie>
i wish screen sharing on mac didn't turn the host's screen back on

[02:58:00.0000] <Hixie>
i want to change the music my mac mini is playing without turning on the massive light that is the screen

[02:58:01.0000] <Hixie>
since the lights are off in the room

[02:58:02.0000] <roc>
I'm a little bit afraid that CPUs might migrate to lots of dumb cores, with minimal speculation, no out-of-order, no wide-issue. That would really suck for JS performance

[02:58:03.0000] <roc>
but I don't think Intel thinks much about JS performance

[02:59:00.0000] <annevk>
they prolly should...

[03:00:00.0000] <Hixie>
well, given JS' importance in the coming years, whoever _does_ think about it may end up having an advantage

[03:00:01.0000] <othermaciej>
that would suck for performance of any language that is not easily parallelizable

[03:00:02.0000] <othermaciej>
which is... almost all of them

[03:00:03.0000] <othermaciej>
the hardware people are going multicore because they can't make the cores much faster any more, at least not to the same degree

[03:00:04.0000] <roc>
it sucks less for languages where you can do more in the compiler

[03:00:05.0000] <othermaciej>
not cause they think any software people love inventing how to parallelize

[03:01:00.0000] <othermaciej>
it's certainly not great for C

[03:01:01.0000] <roc>
and it sucks more for languages where compilation depends on generating fast paths padded with lots of predictable guard code

[03:01:02.0000] <othermaciej>
it might be ok for Java since Java loves to spawn a hojillion threads

[03:02:00.0000] <MikeSmith>
/me is reminded by this conversation of Knuth recent interview comments on multicore architecture - http://www.informit.com/articles/article.aspx?p=1193856

[03:02:01.0000] <othermaciej>
branches are already the hot spots in most x86 code (well at least code that doesn't die from register pressure)

[03:02:02.0000] <othermaciej>
I doubt intel plans to make it worse

[03:02:03.0000] <othermaciej>
multicore is just the harware guys giving up on moore's law

[03:02:04.0000] <roc>
no

[03:02:05.0000] <othermaciej>
no one is ballsy enough to try to build, say, an async CPU

[03:02:06.0000] <roc>
Moore's law is still giving them transistors

[03:02:07.0000] <othermaciej>
well ok, moore's law as popularly misunderstood

[03:02:08.0000] <othermaciej>
they have more transistors but can't turn them into more serial speed

[03:03:00.0000] <roc>
yeah.

[03:03:01.0000] <roc>
Moore's law never applied to heat dissipation

[03:04:00.0000] <Hixie>
an async cpu?

[03:04:01.0000] <othermaciej>
you can build logic without a clock

[03:04:02.0000] <othermaciej>
but it's very hard to get right

[03:04:03.0000] <othermaciej>
taking out all the clock lines and the need for synchronization can make it much faster though

[03:04:04.0000] <othermaciej>
(in theory)

[03:05:00.0000] <othermaciej>
without more power dissipation (since you took out all the wires to spread the clock signal around the chip)

[03:05:01.0000] <othermaciej>
maybe memristors will make a good low-power substrate for logic somehow

[03:08:00.0000] <roc>
we may end up with chips with hundreds of billions of transistors, but if you turn on more than 5% of them the computer melts

[03:09:00.0000] <roc>
which would make application-specific hardware feasible

[03:09:01.0000] <roc>
JS hardware

[03:09:02.0000] <roc>
mmmmmm

[03:10:00.0000] <othermaciej>
I'm sure that will be even more successful than the Java CPU

[03:10:01.0000] <roc>
hehe

[03:12:00.0000] <Hixie>
othermaciej: hm, interesting

[03:17:00.0000] <Philip`>
I think I've heard of some ARM CPU that can execute JVM bytecode directly - does anybody use that?

[03:19:00.0000] <takkaria>
there was an async ARM multiprocessor system for a while called Hydra

[03:19:01.0000] <hsivonen>
annevk: yes, HT is suggesting dispatch on nodeName

[03:20:00.0000] <annevk>
lol

[03:21:00.0000] <annevk>
the podcast Marcos and I made on this is live now, http://standardssuck.org/aria-in-html5

[03:21:01.0000] <annevk>
it's too long and probably too boring to watch, but it was fun to make :)

[03:23:00.0000] <Lachy>
ha! nice response to the comment on youtube http://www.youtube.com/watch?v=tr0ra5hbFfo

[03:29:00.0000] <annevk>
"Views: 209" I don't really buy that

[03:31:00.0000] <Lachy>
/me couldn't stop watching it ;-)

[03:32:00.0000] <Hixie>
anything there i don't know already? :-)

[03:33:00.0000] <annevk>
Hixie, I don't think so

[03:37:00.0000] <gsnedders>
/me expects there are quite a few nice bits in it, but he can't be bothered to watch it

[03:38:00.0000] <gsnedders>
(I have an exam starting in 2.3 hours)

[04:00:00.0000] <Hixie>
/me dons his asbestos suit again

[04:00:01.0000] <hsivonen>
annevk: there's no cabal!

[04:05:00.0000] <othermaciej>
Hixie: on the other hand, allowing any vendor to extend the platform is how we ended up with <img> and <canvas>

[04:06:00.0000] <annevk>
"and suggesting that XML could somehow be replaced by HTML is like saying that JSON could somehow be replaced by Python." hmm

[04:06:01.0000] <othermaciej>
(but I am not gonna get into the extensibility debate)

[04:06:02.0000] <hsivonen>
/me notes that some people prefer using python over json for configuration files

[04:06:03.0000] <othermaciej>
(and I'm not sure I really want to cite <canvas> as a positive example)

[04:06:04.0000] <annevk>
othermaciej, <canvas> is the reason Apple now proposes stuff before implementing

[04:07:00.0000] <othermaciej>
ok lemme put it this way

[04:07:01.0000] <Hixie>
othermaciej: <img> was proposed to the community first

[04:07:02.0000] <annevk>
though maybe I don't need to tell you that :)

[04:07:03.0000] <Hixie>
othermaciej: <canvas> was a mitigated disaster

[04:07:04.0000] <othermaciej>
<canvas> sucks in details, but unlike <marquee> it does not suck in its very concept

[04:07:05.0000] <othermaciej>
it would be nice if it were better but it isn't in the "should not even exist" category

[04:07:06.0000] <hsivonen>
othermaciej: you are being culturally insensitive

[04:08:00.0000] <othermaciej>
hsivonen: to IE developers?

[04:08:01.0000] <Hixie>
othermaciej: <canvas> as a technology is fine. as an example of how we want the language extended process-wise, it's a terrible example.

[04:08:02.0000] <hsivonen>
othermaciej: no, to Chinese Web authors

[04:08:03.0000] <Philip`>
/me wonders why so many people with high-profile sites still choose to use <marquee> if it sucks

[04:08:04.0000] <Hixie>
othermaciej: we're still dealing with the effects of decisions apple made without consulting with the community

[04:08:05.0000] <Hixie>
othermaciej: e.g. the lack of a Path object

[04:09:00.0000] <othermaciej>
really? a Path object is easy to add

[04:09:01.0000] <othermaciej>
the parts of <canvas> that suck are the things that are terrible but that can't be taken away

[04:09:02.0000] <othermaciej>
however, if Apple had tried to follow W3C process, <canvas> would not exist

[04:09:03.0000] <Hixie>
adding a path object would likely either result in duplication of a chunk of the api or a very odd api

[04:09:04.0000] <othermaciej>
WHATWG process is certainly more conducive to stuff actually happening

[04:10:00.0000] <Hixie>
othermaciej: but anyway

[04:10:01.0000] <othermaciej>
Hixie: it's not uncommon for graphics context APIs to both have an implicit context path and explicit paths

[04:10:02.0000] <othermaciej>
(that is in fact how CoreGraphics works for instance)

[04:10:03.0000] <Hixie>
othermaciej: even if canvas was a brilliant example of perfection, it's still the case that we don't want individuals and vendors unilaterally extending text/html without community involvement

[04:10:04.0000] <Philip`>
othermaciej: Maybe that's because they all had the same lack of foresight when designing the initial API, and then had to remain backward compatible? :-)

[04:11:00.0000] <Hixie>
heh

[04:11:01.0000] <othermaciej>
no, it's just convenient to have an implicit path

[04:11:02.0000] <othermaciej>
same as it is convenient to be able to draw a rect directly, even though you could use lineTo 

[04:12:00.0000] <othermaciej>
I don't think the implicit path is an API flaw

[04:12:01.0000] <othermaciej>
lack of explicit paths is an (easily rectified) API flaw

[04:13:00.0000] <othermaciej>
transforms affecting the creation of the implicit path is just plain weird but I can't remember if that is our fault or Mozilla's ultimately

[04:13:01.0000] <annevk>
I wonder what happens with a WG vote

[04:13:02.0000] <annevk>
on that issue

[04:14:00.0000] <hsivonen>
annevk: the WG votes to do something that has not been shown to be doable

[04:14:01.0000] <othermaciej>
globalAlpha / globalCompositeOperation is not a great design (better to push and pop transparency layers)

[04:15:00.0000] <othermaciej>
the fact that strokeStyle and fillStyle take random assortments of objects, some of which may be strings but none of which may be a CSSOM color is pretty daft

[04:15:01.0000] <Philip`>
/me still likes transforms affecting path creation, because otherwise it's impossible to do things like stoked ellipses

[04:15:02.0000] <hsivonen>
I'd love to see a vote where voting Yes required supplementing the vote with a concrete proposal

[04:15:03.0000] <annevk>
hsivonen, yeah, I'm not really sure how the numbers turn out, but I think people will request namespaces

[04:16:00.0000] <Philip`>
s/stoked/stroked/

[04:16:01.0000] <Hixie>
supplementing the vote with a concrete proposal is easy and such proposals have been put forward many times

[04:16:02.0000] <Hixie>
supplementing the vote with a concrete WORKING proposal is what would be interesting

[04:16:03.0000] <Hixie>
but i've no idea how to require that

[04:17:00.0000] <Philip`>
Require each proposal to be shipped in a web browser with at least 1% market share before you're allowed to vote 'yes' with that proposal

[04:17:01.0000] <othermaciej>
that's a harsh requirement

[04:19:00.0000] <Hixie>
haha

[04:19:01.0000] <Hixie>
that'd be awesome

[04:19:02.0000] <annevk>
otoh, you need to have some buy in from implementors

[04:19:03.0000] <hsivonen>
I wonder if data-* could be supported using NVDL or if I need to escape outside of schemas

[04:19:04.0000] <othermaciej>
anyway, I am certainly not a fan of unilateral extensions as the first option

[04:19:05.0000] <othermaciej>
and in WebKit the only area where we have had any recently afaik is CSS

[04:19:06.0000] <Hixie>
vendor-prefixed ones in css are ok, given css's fallback model

[04:19:07.0000] <Hixie>
and you are working with the csswg on those

[04:19:08.0000] <annevk>
RB continues misunderstanding how namespaces work

[04:20:00.0000] <hsivonen>
does anyone know if NVDL has a human-readable (non-ISO) spec too?

[04:20:01.0000] <Hixie>
at least, insofar as they are letting you work with them

[04:20:02.0000] <annevk>
he's a fine example of why we should keep them outside of HTML :)

[04:20:03.0000] <hsivonen>
annevk: indeed

[04:20:04.0000] <hsivonen>
/me doesn't like ISO drafting rules

[04:21:00.0000] <annevk>
hsivonen, oh schemas don't have part of attribute name wildcards?

[04:21:01.0000] <annevk>
/me foresees trouble getting data-* into SVG

[04:21:02.0000] <hsivonen>
annevk: RELAX NG doesn't

[04:22:00.0000] <hsivonen>
annevk: RELAX NG allows you to wildcard the whole attribute or the local name but not part of the local name

[04:22:01.0000] <annevk>
data-* is one of the things I hope everyone implement quickly and we can advocate a lot

[04:23:00.0000] <annevk>
so that people start switching over and we get less issues introducing new attributes

[04:24:00.0000] <hsivonen>
/me wonders what ISO specs *aren't* fast-tracked

[04:25:00.0000] <Philip`>
C++?

[04:25:01.0000] <hsivonen>
perhaps

[04:27:00.0000] <hsivonen>
comparing OASIS and ISO versions of RELAX NG specs is a great example of how ISO makes specs unapproachable

[04:28:00.0000] <hsivonen>
bah. I'm just going to write a SAX filter that throws data-* out

[04:30:00.0000] <hsivonen>
I wonder how to best expose the filter in the generic UI...

[04:32:00.0000] <hsivonen>
I wonder if I should model data-* as a magic namespace

[04:33:00.0000] <hsivonen>
/me decides to unify namespace dropping code and data-* dropping code

[04:34:00.0000] <hsivonen>
no doubt this will make people scream

[04:36:00.0000] <hsivonen>
/me temporarily undoes the decision

[04:38:00.0000] <hsivonen>
because that would allow data-* on non-HTML elements as a side effect

[04:39:00.0000] <hsivonen>
generic is hard

[04:41:00.0000] <Lachy>
is allowing data-* on non-HTML elements necessarily a bad thing?

[04:41:01.0000] <hsivonen>
Lachy: not if I can blame it on Hixie :-)

[04:41:02.0000] <hsivonen>
Lachy: currently I can't AFAICT

[04:42:00.0000] <Lachy>
it's not really that much different from allowing aria-* on generic elements

[04:42:01.0000] <annevk>
i'd like data-* to be generic like aria-*

[04:42:02.0000] <Hixie>
there is one major difference

[04:42:03.0000] <hsivonen>
Lachy: I don't allow aria-* on generic elements

[04:42:04.0000] <Hixie>
in that no spec allows it yet :-)

[04:42:05.0000] <annevk>
 though maybe we should have some more experience with data-* first

[04:42:06.0000] <hsivonen>
Lachy: and technically the two are *very* different in V.nu

[04:42:07.0000] <Lachy>
Hixie, that's only a minor barrier to get across

[04:43:00.0000] <annevk>
putting it in the spec is, dealing with the storm of comments might be more tough

[04:43:01.0000] <hsivonen>
as of ARIA 1.0, the aria-foo attributes can be finitely enumerated

[04:43:02.0000] <hsivonen>
the data-* attributes cannot, because they are countably infinite

[04:43:03.0000] <Lachy>
filing such comments in the 'ignore' folder is an effective way to deal with them

[04:45:00.0000] <hsivonen>
also, aria-foo is more sensitive to validation than just throwing away, which is basically what data-* needs

[04:45:01.0000] <annevk>
I don't like that MathML strips whitespace and such in attributes

[04:46:00.0000] <hsivonen>
annevk: as implemented, it doesn't

[04:46:01.0000] <hsivonen>
annevk: spec bug, IMO

[04:46:02.0000] <hsivonen>
annevk: V.nu deliberately violates the MathML spec here

[04:49:00.0000] <annevk>
are they fixing it?

[04:49:01.0000] <Hixie>
see mail from david just now

[04:49:02.0000] <Hixie>
wait

[04:50:00.0000] <Hixie>
what do you mean it strips whitespace?

[04:50:01.0000] <Hixie>
during parsing?

[04:52:00.0000] <annevk>
Hixie, type=" checkbox " is not a checkbox in HTML, for similar constructs in MathML that would be a "checkbox"

[04:53:00.0000] <hsivonen>
according to spec that is--not according to Gecko

[04:54:00.0000] <hsivonen>
annevk: I got the impression they didn't want to fix the spec

[04:58:00.0000] <Hixie>
annevk: ah

[04:58:01.0000] <Hixie>
how silly

[04:59:00.0000] <Hixie>
/me reads the mail in his AAA-Productivity folder

[04:59:01.0000] <Hixie>
there was nothing there that my life would have been worse if i had missed reading it

[04:59:02.0000] <Hixie>
glad my filters are still working well

[04:59:03.0000] <annevk>
is that like really good or is it some vague remark to accessibility?

[05:00:00.0000] <annevk>
ah, guess that explains it :)

[05:00:01.0000] <othermaciej>
I assume it is a hack for a mail client that insists on alphabetizing IMAP folders or something

[05:00:02.0000] <Hixie>
it's a carefully named folder

[05:01:00.0000] <annevk>
othermaciej, then you'd have used ZZZ-x :)

[05:01:01.0000] <Lachy>
I don't see an aaa-productivity folder on whatwg.org/issues/ - what's in it?

[05:01:02.0000] <Hixie>
which fully shows the importance i place on its contents

[05:01:03.0000] <Hixie>
in case someone should stumble across it one day

[05:01:04.0000] <Hixie>
i wouldn't want them insulted

[05:01:05.0000] <othermaciej>
heh

[05:02:00.0000] <Hixie>
Lachy: all mail sent from: or to: someone who sends, ah, very important e-mails

[05:02:01.0000] <Hixie>
and who causes people to send even more important e-mails

[05:03:00.0000] <roc>
I'm pretty sure we trim whitespace when interpreting MathML attributes

[05:03:01.0000] <roc>
at least in most cases

[05:09:00.0000] <hsivonen>
roc: http://lists.w3.org/Archives/Public/www-math/2007Dec/0008.html

[05:10:00.0000] <hsivonen>
roc: I haven't retested with Firefox 3

[05:11:00.0000] <Hixie>
2188 e-mails to go

[05:11:01.0000] <Hixie>
i guess i'll do that tomorrow

[05:11:02.0000] <Hixie>
bed time no

[05:11:03.0000] <Hixie>
w

[05:11:04.0000] <annevk>
tomororw, huh? hah

[05:12:00.0000] <annevk>
g'night

[05:12:01.0000] <roc>
hsivonen: oh well

[05:15:00.0000] <Hixie>
kittens, why are you all still arguing this aria- thing

[05:16:00.0000] <hsivonen>
/me isn't

[05:16:01.0000] <annevk>
/me is not a kitten

[05:17:00.0000] <annevk>
but yeah...

[05:17:01.0000] <othermaciej>
you should just give up and join the Cult of the Sacred:Colon

[05:18:00.0000] <othermaciej>
don't worry about why:you are using:it, just use:it every:where

[05:18:01.0000] <annevk>
it:starts making:sense null:now

[05:18:02.0000] <Hixie>
anne: i meant "kittens" as a word of exasperation, not a reference to you :-P

[05:18:03.0000] <Hixie>
like people say "jesus"

[05:18:04.0000] <Hixie>
anyway

[05:18:05.0000] <annevk>
i know :D

[05:18:06.0000] <Hixie>
really going to bed now :-)

[05:18:07.0000] <Hixie>
nn

[05:18:08.0000] <othermaciej>
me too

[05:19:00.0000] <annevk>
see you guys tonight

[06:53:00.0000] <Philip`>
/me notices that he has zero pages using &lang; or &rang;

[07:38:00.0000] <Lachy>
http://my.opera.com/core/blog/selectors-api

[07:38:01.0000] <Lachy>
that was posted yesterday. It's Opera's new blog for the Core department

[07:42:00.0000] <Dashiva>
/me pokes Lachy with a [] instead of new Array()

[07:44:00.0000] <Lachy>
both work

[07:44:01.0000] <Lachy>
Is there a reason to prefer [] over new Array();?

[07:44:02.0000] <Philip`>
But one is horrifically ugly

[07:44:03.0000] <MikeSmith>
Lachy: is this hardcore or softcore?

[07:44:04.0000] <Lachy>
which one?

[07:44:05.0000] <MikeSmith>
or slowcore?

[07:45:00.0000] <MikeSmith>
Lachy: nuttin' ... was just making a dumb joke

[07:45:01.0000] <Dashiva>
Lachy: new Array is slower, longer, and error-prone when the array is initialized to contain a single number

[07:45:02.0000] <MikeSmith>
Lachy: you work for Lars-Erik?

[07:45:03.0000] <Philip`>
Lachy: The one that doesn't look like Perl or Python

[07:46:00.0000] <Lachy>
MikeSmith, Wilhelm is my manager

[07:47:00.0000] <Lachy>
Lars Erik is the manager of the whole core dept. I think

[07:48:00.0000] <Lachy>
Philip`, anything that doesn't look like Perl has to look good :-)

[07:48:01.0000] <Philip`>
Lachy: There really needs to be a link from the individual posts back to the blog's front page

[07:48:02.0000] <Philip`>
(like making the heading image be a link back)

[07:49:00.0000] <Philip`>
(Also pretty much every other Opera blog has that same stupid usability bug)

[07:49:01.0000] <Lachy>
there are a lot of things that my.opera needs to improve. I'll pass that onto the appropriate people

[07:49:02.0000] <Dashiva>
That's because it's the default, probably :)

[07:51:00.0000] <Philip`>
Dashiva: I thought I remembered hearing that new Array(...) was significantly faster than [...] (in Firefox)

[07:51:01.0000] <Lachy>
I changed new Array() to [].

[07:52:00.0000] <Lachy>
hmm. I think I need to write a test to see which one really is faster.

[07:52:01.0000] <Philip`>
var t=new Date();for(var i=0;i<100000;++i)var z=[i,i,i,i,i,i];alert(new Date()-t)

[07:53:00.0000] <Philip`>
var t=new Date();for(var i=0;i<100000;++i)var z=new Array(i,i,i,i,i,i);alert(new Date()-t)

[07:53:01.0000] <Philip`>
In Firefox 3, the former takes 270ms, the latter takes 130ms

[07:53:02.0000] <Philip`>
but unless you're initialising a hundred thousand arrays you're not going to notice the difference so you should use [...] because it's so much prettier

[07:54:00.0000] <Philip`>
The same code in Opera 9.5 is the other way around, taking 430ms and 540ms

[07:56:00.0000] <Philip`>
In Safari 3.0 it's 160ms/130ms

[07:58:00.0000] <Philip`>
In IE6 it's roughly 950ms for both

[07:58:01.0000] <Philip`>
If you really care about performance, nobody uses Opera so you should use new Array()

[07:58:02.0000] <Dashiva>
/me wonders if Philip` forgot about IE

[07:59:00.0000] <Philip`>
?

[07:59:01.0000] <Dashiva>
IE7 is faster with [] than new Array

[08:00:00.0000] <Philip`>
Oh

[08:01:00.0000] <Dashiva>
And for a proper test, you have to check with 0 and large_number_value initial values too :)

[08:01:01.0000] <MikeSmith>
/me finally actually reads Lachy's Selectors API writeup at Opera core blog

[08:01:02.0000] <MikeSmith>
Lachy: nice article

[08:01:03.0000] <MikeSmith>
short but sweet

[08:01:04.0000] <Lachy>
thanks

[08:02:00.0000] <Philip`>
In IE7 on totally different hardware, both take about 1600ms

[08:02:01.0000] <Philip`>
Dashiva: Do you have evidence for your claim? :-)

[08:03:00.0000] <MikeSmith>
Lachy: is this the first public acknowledgment of the Selectors API support in Opera?

[08:04:00.0000] <Philip`>
Lachy: I think the first sentence in that article is too complex and hard to parse, but otherwise it looks good :-)

[08:05:00.0000] <Lachy>
Philip`, I don't see any problems with "The Selectors API specification currently being worked on within the WebAPI working group at the W3C defines DOM APIs designed to make it possible to select elements within the document using Selectors."

[08:05:01.0000] <Lachy>
would it work better if there were commas placed in it at appropriate places?

[08:06:00.0000] <MikeSmith>
The Selectors API specification (currently being worked on within the WebAPI working group at the W3C) defines DOM APIs designed to make it possible to select elements within the document using Selectors.

[08:06:01.0000] <Philip`>
Maybe "The Selectors API specification, currently being ... at the W3C, defines DOM APIs that make it possible ..."

[08:07:00.0000] <Lachy>
fixed

[08:09:00.0000] <Philip`>
I'd probably s/designed to make it possible/that make it possible/ too because there's too many verbs at the moment :-)

[08:09:01.0000] <Philip`>
but I could just be wrong

[08:09:02.0000] <Philip`>
Anyway, the commas solve the problem where I misparsed the first half of the sentence when first reading it

[08:12:00.0000] <Philip`>
"Our implementation also partially supports the namespace resolver features" - is there any information on how partial that is?

[08:13:00.0000] <Lachy>
I can't remember what ours is missing.

[08:16:00.0000] <Lachy>
actually, I think, as currently specced, it fully supports the NSResolver

[08:17:00.0000] <Lachy>
the only issue is that our implementation doesn't deal with DOM modifications by the resolver according to the spec, but that's due to the fact that I haven't determined what to do about that yet

[08:17:01.0000] <Lachy>
although, please file bugs if you find them

[08:22:00.0000] <Philip`>
/me finds a site which gives an ASP NullReferenceException if he tries to log in with Opera 9.2, but works fine in 9.5

[08:24:00.0000] <Philip`>
(which is kind of annoying because I don't want to use 9.5 yet)

[08:33:00.0000] <Philip`>
Oh, whoops, it was only failing in Opera 9.2 because I disabled referrer logging there

[08:54:00.0000] <annevk>
Despite the flaws, Ubuntu is quite nice. Installed it within 30 minutes on a new PC and it seems that my mom is able to use it :)

[08:55:00.0000] <annevk>
It complained about something related to the wireless card and drivers, but that worked so I'm not really sure what it was about...

[08:55:01.0000] <annevk>
Maybe they are "not free"

[09:14:00.0000] <Philip`>
/me wonders if installing Ubuntu from CD experiences the problems with OpenSSL

[09:14:01.0000] <mpt>
Philip`, yes, please make installing updates the first thing you do after restarting

[09:15:00.0000] <mpt>
We will have new CD images Real Soon Now

[09:17:00.0000] <Philip`>
(Fortunately I only have one Ubuntu server which is older than that bug so its keys are fine, and an Ubuntu VM image which is probably affected but is only run locally and accessed with serial-console-over-UDP instead of SSH)

[10:33:00.0000] <MikeSmith>
done.

[10:33:01.0000] <MikeSmith>
http://dev.w3.org/html5/pubnotes/

[10:33:02.0000] <MikeSmith>
up to date, as of r1687 today

[10:34:00.0000] <MikeSmith>
corrections/additions welcome

[10:34:01.0000] <Philip`>
/me wonders what HTML5 has to do with pubs

[10:34:02.0000] <MikeSmith>
heh

[10:35:00.0000] <MikeSmith>
I will take a rest f2f a while now

[10:35:01.0000] <Philip`>
"It primarily document changes" - documents

[10:37:00.0000] <Philip`>
"You can find the source for the current version of this document in the [W3C source repository]." - that links to the latest version, not to the source or to a source repository browser

[10:37:01.0000] <MikeSmith>
Philip`: document/documents changes made

[10:39:00.0000] <Philip`>
"In the “Conformance checkers” subsection, the text of a statement was updated to now read (added text highlighted:" - missing ')'

[10:39:01.0000] <MikeSmith>
Philip`: source link fixed

[10:40:00.0000] <Philip`>
Validators complain because you need something like <blockquote><p><q>... and are missing the <p>, I think

[10:41:00.0000] <MikeSmith>
missing paren now added

[10:41:01.0000] <Philip`>
It might be kind of nice if the references to spec sections were links

[10:41:02.0000] <MikeSmith>
Philip`: yeah, I need to fix that blockquote part

[10:41:03.0000] <MikeSmith>
karl pointed out the validity problem

[10:42:00.0000] <MikeSmith>
the source is in text/html and validates as HTML5 fine

[10:42:01.0000] <MikeSmith>
problem is that I need to transform it to XHTML per W3C pubrules

[10:42:02.0000] <MikeSmith>
I will make turn the references to spec sections into links tomorrow

[10:42:03.0000] <Philip`>
<code class="element">embed</code> has kind of ugly blue-on-blue styling - might be nice to use the same styles as the spec

[10:44:00.0000] <MikeSmith>
yeah, I can switch that styling any time

[10:44:01.0000] <MikeSmith>
spec just uses orangered for everything

[10:45:00.0000] <Philip`>
The comma in "Section 2.2, Elements" looks unnatural to me - a colon or dot would seem more normal. (But I'm just being very picky, so feel free to ignore me ;-) )

[10:47:00.0000] <hsivonen>
annevk: I liked particularly the second half of the first episode of standards suck

[10:48:00.0000] <Philip`>
MikeSmith: "Some changes were made Within the “Dynamic markup insertion in XML” subsection in this section" - s/W/w/

[10:48:01.0000] <hsivonen>
annevk: just curious: what kind of decision process did you have for choosing the video host?

[10:49:00.0000] <hsivonen>
annevk: in particular, did you knowingly reject blip.tv?

[10:49:01.0000] <MikeSmith>
Philip`: yeah, I will probably globally replace the comma thing

[10:50:00.0000] <MikeSmith>
Philip`: made W->w change

[10:50:01.0000] <Philip`>
MikeSmith: "A placeholder for a “URLs” subsection was added, with editorial notes about what its intended will be." - lack of grammar near end of sentence

[10:50:02.0000] <Philip`>
or lack of a word or something

[10:50:03.0000] <MikeSmith>
/me thinks he probably meant to put "intended purpose"

[10:51:00.0000] <MikeSmith>
made it "intended purpose"

[10:53:00.0000] <Philip`>
"the phrase “valid browsing context name” was emended to become" - I've never heard the word "emend" before :-)

[10:55:00.0000] <MikeSmith>
Philip`: yeah, any alternatives to that word welcome

[10:55:01.0000] <gsnedders>
MikeSmith: corrected?

[10:55:02.0000] <MikeSmith>
it is any actual real dictionary word

[10:56:00.0000] <gsnedders>
MikeSmith: improved, becoming

[10:56:01.0000] <MikeSmith>
"refined" is probably best

[10:57:00.0000] <gsnedders>
/me isn't a very good thesaurus

[10:57:01.0000] <MikeSmith>
but I think I used "emend" to avoid overuse of "refined"

[10:58:00.0000] <Philip`>
MikeSmith: "In this section, an statement was added" s/n//

[10:58:01.0000] <gsnedders>
Philip`: that leaves "I this section, an statement was added"

[10:59:00.0000] <Philip`>
gsnedders: No, these are context-sensitive regexps that only match in the intended location

[10:59:01.0000] <gsnedders>
Philip`: Ah.

[11:00:00.0000] <Philip`>
gsnedders: They are not mechanically computable, but that's okay because we're not computers

[11:01:00.0000] <MikeSmith>
Philip`: an->a changed

[11:04:00.0000] <Philip`>
MikeSmith: "This section concerns elements used to mark up content at the “text level” (as opposed to the “sectioning” or “grouping” levels." - missing ')', brain stack overflow

[11:05:00.0000] <MikeSmith>
Philip`: :) fixed

[11:07:00.0000] <MikeSmith>
one thing, I call the DOM formalisms "interface definitions" throughout.. most of the time, seems like we just call each instance of those "an IDL" .. which I've never really liked and which I don't think casual readers are going to understand

[11:08:00.0000] <Philip`>
MikeSmith: It might be helpful to group changes, e.g. the removal of automatic cross referencing results in a lot of elements having "In this section, the following statement was removed completely: “the title attribute has special semantics on this element when used with the dfn element”."

[11:08:01.0000] <Philip`>
which is kind of repetitive, and not immediately obvious that it's repetitive because it's always the same change

[11:08:02.0000] <MikeSmith>
Philip`: yeah, I thougt about that, but it sorta comes to a high-level choice of either grouping them by related change or grouping them by section

[11:09:00.0000] <MikeSmith>
... and it seemed like strict grouping by section would be easier

[11:09:01.0000] <MikeSmith>
... and I'm lazy so I went with that

[11:09:02.0000] <Philip`>
so perhaps all the later elements should just say "Affected by the <a href=#xref-removal>removal of automatic cross referencing</a>" and point back at the bit in dfn

[11:10:00.0000] <MikeSmith>
OK, yeah, that might work better

[11:10:01.0000] <Philip`>
Primarily grouping by section seems good; it just might be nice in a few places to cut across that sectioning system

[11:11:00.0000] <MikeSmith>
true, yeah

[11:11:01.0000] <Philip`>
(The SVN logs are there for people who want things grouped by changes)

[11:13:00.0000] <takkaria>
MikeSmith: nice changelog!

[11:14:00.0000] <Philip`>
"the difficulties of attempting to mark up edits that cross implied paragraphs" occurs twice in adjacent sections, which seems unnecessarily repetitive; perhaps the first paragraph's parenthetical comment could be removed

[11:16:00.0000] <Philip`>
"The prefatory text for one of the examples was updated to read:" - that's another word I don't think I've ever heard before :-p

[11:17:00.0000] <Philip`>
(That's not a bad thing, since it gives me a chance to expand my vocabulary :-) )

[11:17:01.0000] <MikeSmith>
takkaria: thanks

[11:17:02.0000] <MikeSmith>
Philip`: "introductory text" is maybe less pedantic

[11:18:00.0000] <Philip`>
"* an iframe name, ... * an iframe sandbox, ..." - that should probably say "attribute" somewhere

[11:19:00.0000] <Philip`>
Also that list is missing '.'s at the ends of sentences

[11:20:00.0000] <Philip`>
and "conformance criteria were added in relation to behavior when the sandboxed plugins browsing context flag is set" is missing a capital letter and '.' which makes it inconsistent with the rest of its list

[11:20:01.0000] <MikeSmith>
Philip`: was intentional not to refer to with "attribute" .. to mean that those are abstract characteristics associated with the iframe

[11:20:02.0000] <MikeSmith>
... but perhaps that doesn't make much sense

[11:23:00.0000] <Philip`>
MikeSmith: Ah, that sounds reasonable

[11:23:01.0000] <Philip`>
MikeSmith: "This specification does not define a mechanism for interacting with third-party handlers, as it is expected to be user-agent-specific. Some UAs might opt to support a plugin mechanism such as the Netscape Plugin API; others may use remote content convertors or have built-in support for certain types."

[11:23:02.0000] <MikeSmith>
Philip`: I fixed the punctuation/caps stuff

[11:23:03.0000] <Philip`>
^ that text was modified a bit, e.g. to remove the spelling errors, not just moved

[11:23:04.0000] <Philip`>
(and pubnotes has the old pre-modified version)

[11:23:05.0000] <MikeSmith>
OK

[11:24:00.0000] <MikeSmith>
I guess that was changed in the last few days

[11:24:01.0000] <Philip`>
(which confused me since I was looking for the spelling error in the spec and couldn't find it)

[11:24:02.0000] <MikeSmith>
will grabe the latest

[11:25:00.0000] <MikeSmith>
Philip`: you mean the text in 3.12.4 ?

[11:25:01.0000] <Philip`>
"In the “User interface” subsection, an instance of the phrase “if scripting is disabled” was changed to read, “if the media element is without script." - missing closing quote

[11:25:02.0000] <MikeSmith>
the "embed element" section?

[11:25:03.0000] <Philip`>
MikeSmith: Yes

[11:26:00.0000] <MikeSmith>
oK

[11:26:01.0000] <MikeSmith>
I think I will remove those blockquote excerpts from the embed section 

[11:26:02.0000] <MikeSmith>
and just refer to them instead

[11:27:00.0000] <MikeSmith>
fixed the missing quote char

[11:31:00.0000] <Philip`>
"The text of the section was revised to remove references to the colSpan and rowSpan DOM attributes and to instead reference the headers DOM attribute." - not sure what the exact changes were, but I'd guess it should be "also" rather than "instead"

[11:32:00.0000] <Philip`>
"the conformance requirements for the td and th elements now state they they implement interfaces that inherit from that interface" - s/they they/that they/

[11:35:00.0000] <MikeSmith>
Philip`: change "headers" part to "and to add references to"

[11:35:01.0000] <MikeSmith>
"they they"->"that they" fixed

[11:36:00.0000] <Philip`>
"(Note that the “XXX4“ name is simply a placeholder; it is not intended that the method will be implemented with that literal name.)" - that's a shame :-(

[11:36:01.0000] <MikeSmith>
heh

[11:37:00.0000] <MikeSmith>
I am reasonably sure that if I didn't put that, somebody would post a message saying, What is it with the name of this XXX4 method?

[11:37:01.0000] <Philip`>
"The word “origin” was replaced by “value” in the sentence that now reads, “If the scheme is ‘file’, then the user agent may return a UA-specific value”, some instances of the phrase “the same as the origin” were refined to read “equal to the origin”." - should that be a semicolon before the "some"?

[11:39:00.0000] <MikeSmith>
Philip`: made it a separate list item now

[11:40:00.0000] <Philip`>
"the addition of a statements related to the source browsing context" - s/s//

[11:40:01.0000] <Philip`>
or s/a // or whatever

[11:41:00.0000] <Philip`>
"The following not was added: The above algorithm is a willful violation of the HTTP specification." - s//e/ unless someone already mentioned that one

[11:41:01.0000] <Philip`>
"Also in he “Storing name/value pairs” subsection" - s//t/

[11:41:02.0000] <MikeSmith>
"a statements" fixed

[11:42:00.0000] <MikeSmith>
not->note fixed already

[11:42:01.0000] <Philip`>
"This section remains largely unchanged, though some significant changes were made to to — among those, the following:" - to to?

[11:43:00.0000] <Philip`>
"In the subsection that defines the “” link type, a statement was added that Only one link is created for the set of one or more “up” keywords and, if present, the “index” keyword." - is that really meant to be “”?

[11:43:01.0000] <MikeSmith>
toot toot fixed

[11:45:00.0000] <MikeSmith>
Philip`: "" link type should have been "UP" .. fixed that

[11:45:01.0000] <Philip`>
"asyncronous"

[11:46:00.0000] <MikeSmith>
"asyncronous" fixed

[11:46:01.0000] <Philip`>
"the range “U+0030 DIGIT ZERO .. U+0039 DIGIT NINE” as added" - s/as/was/

[11:47:00.0000] <Philip`>
"statement were added"

[11:49:00.0000] <MikeSmith>
"as added" fixed

[11:49:01.0000] <Philip`>
"to include named character references commonly used in MathML documents" - they're not "commonly used", they're just "potentially used" or "supported" or something

[11:50:00.0000] <MikeSmith>
"statement were" fixed

[11:51:00.0000] <MikeSmith>
Philip`: OK, I guess I will change that to "for potential use in MathML"

[11:51:01.0000] <Philip`>
Hixie: Section 10 (#no) says "This section is non-normative." but then 10.4 gives normative requirements

[11:51:02.0000] <BenMillard>
MikeSmith, I remember asking about a "no mail" option last year: http://lists.w3.org/Archives/Public/public-html/2007Jun/thread.html#msg1011 and http://lists.w3.org/Archives/Public/public-html/2007Jul/thread.html#msg77

[11:52:00.0000] <MikeSmith>
made it "of potential use"

[11:52:01.0000] <MikeSmith>
BenMillard: so what was the resolution of that?

[11:52:02.0000] <BenMillard>
MikeSmith, it seems W3C doesn't think you can participate in a group unless you are on the mailing list

[11:52:03.0000] <MikeSmith>
BenMillard: I guess that kind of makes sense

[11:53:00.0000] <Dashiva>
You could always sign up with an address you don't read mail for ;)

[11:53:01.0000] <BenMillard>
MikeSmith: written another way, it seems W3C thinks you can only participate if you are on the mailing list

[11:53:02.0000] <Philip`>
MikeSmith: Oh, I think I got to the end

[11:53:03.0000] <MikeSmith>
Philip`: cool. how long did it take?

[11:53:04.0000] <Philip`>
/me checks his IRC history

[11:54:00.0000] <Philip`>
Um, too long :-p

[11:54:01.0000] <MikeSmith>
heh

[11:54:02.0000] <MikeSmith>
~70mins maybe

[11:54:03.0000] <Philip`>
(More than an hour, though not entirely with full concentration)

[11:54:04.0000] <MikeSmith>
and you know this material pretty well already

[11:54:05.0000] <hsivonen>
looks like Marcos walked right into a bikeshed with widget: :-(

[11:55:00.0000] <MikeSmith>
Philip`: so I reckon it would be a good 90-minute read for somebody not familiar with it

[11:55:01.0000] <MikeSmith>
hsivonen: yeah, dunno that he knew what he'd be in for

[11:55:02.0000] <Philip`>
MikeSmith: If they're trying to read and understand all of it, that sounds about reasonable

[11:55:03.0000] <MikeSmith>
... I should have thought about it & warned him

[11:57:00.0000] <MikeSmith>
Philip`: thanks much

[11:58:00.0000] <Philip`>
MikeSmith: I'm not going to bother checking whether you missed any changes, but it looks quite comprehensive, so hopefully it will be helpful to people :-)

[11:59:00.0000] <Philip`>
MikeSmith: It'd be nice to summarise the high-level changes, for lazy people, like "Added support for inline MathML in the HTML syntax" and "Added simple text drawing API to <canvas>" etc

[11:59:01.0000] <BenMillard>
MikeSmith, it's certainly exhaustive. I think very few people will read all of it, or even any significant portion of it, but I can well imagine people skimming through to find out the history of their pet issue(s)

[12:00:00.0000] <MikeSmith>
Philip`: annevk did that already in a separate doc, so I will just make the build script pull that out from cvs and include it

[12:00:01.0000] <MikeSmith>
BenMillard: yeah, that's the use case

[12:00:02.0000] <MikeSmith>
mainly

[12:01:00.0000] <Philip`>
MikeSmith: Ah, right

[12:01:01.0000] <MikeSmith>
though annevk summary list is missing a few things that ought to be in it, I think

[12:02:00.0000] <MikeSmith>
... though I can't think of specifically what at the moment

[12:02:01.0000] <BenMillard>
MikeSmith, annevk's doc has been updated during this month, so maybe what was missing has been added now?

[12:03:00.0000] <BenMillard>
(for the record, talking about this doc: http://dev.w3.org/html5/html4-differences/ I take it?)

[12:03:01.0000] <MikeSmith>
BenMillard: saw the latest a couple days back and remember not seeing something i thought should be there

[12:03:02.0000] <BenMillard>
ah, ok

[12:26:00.0000] <hsivonen>
so it seems data-* is the first feature Hixie has introduced that RELAX NG is really bad at *allowing*

[12:27:00.0000] <hsivonen>
After pondering this, I think I'm going to make the preset HTML5 schemas magic and fake it

[12:27:01.0000] <hsivonen>
So I think at this point, I'm not going to expose this feature to custom schemas

[12:30:00.0000] <hsivonen>
hendry: where did you get an instance of the Iris browser to run? is it shipping with some device already?

[13:58:00.0000] <takkaria>
nice, a post from the IE team wanting to implement HTML5 behaviour in at least a bit of their parser

[14:37:00.0000] <Lachy>
takkaria, it's unfortnate he missed the part in the spec where the answer is clearly spelled out for them

[14:37:01.0000] <Lachy>
it took me just a few minutes to find the appropriate section talking about the stack of open elements.

[14:38:00.0000] <Philip`>
That only helps if you know you're looking for the stack of open elements

[14:38:01.0000] <Philip`>
which is not an obvious thing to know when you're trying to work out what happens when a script changes its ancestor nodes

[14:38:02.0000] <Lachy>
Philip`, he should have. Hixie told him to look at that section http://blogs.msdn.com/ie/archive/2008/04/23/what-happened-to-operation-aborted.aspx#8423021

[14:41:00.0000] <takkaria>
Lachy: it is unfortunate, but it's still dialogue

[14:42:00.0000] <Lachy>
yeah, it's certainly good that they're trying to fix it.

[14:59:00.0000] <Hixie>
/me replies to travis

[15:03:00.0000] <takkaria>
Lachy: my main point being, they're asking *before* they implement rather than after

[15:05:00.0000] <Hixie>
as far as i can tell they're asking after they tried and failed :-)

[15:19:00.0000] <Lachy>
Hixie, that's a really good explanation.  Maybe you should consider adding it to the spec to illustate how the algorithm works.

[15:21:00.0000] <Hixie>
hm, good idea

[15:23:00.0000] <Lachy>
and a similar example for the adoption agency algorithm would help too

[15:23:01.0000] <Hixie>
yeah i plan on adding diagrams and stuff once the section is stable

[15:25:00.0000] <takkaria>
in SVG, I would hope :P

[15:36:00.0000] <annevk>
hsivonen, I'd have to ask Marcos. I don't really know

[15:36:01.0000] <annevk>
I hope we start publishing the original files as well as I don't have Flash :)

[15:40:00.0000] <annevk>
first spec I managed to get to CR involves namespaces: http://www.w3.org/TR/2008/CR-css3-namespace-20080523/

[15:40:01.0000] <annevk>
I hope everyone appreciates the irony just as much as I do

[16:37:00.0000] <annevk>
Lachy, don't argue with RB

[16:38:00.0000] <annevk>
Lachy, web standards 101 :)

[16:38:01.0000] <Lachy>
annevk, I know. I just couldn't resist

[16:39:00.0000] <Lachy>
I'm sorry

[16:39:01.0000] <Lachy>
But I like how his only response was to claim that he didn't make the assumption, which he clearly must have given his claims


2008-05-24
[17:10:00.0000] <MikeSmith>
Hixie: if/when you have some time for review/sanity-check:

[17:10:01.0000] <MikeSmith>
http://dev.w3.org/html5/pubnotes/

[17:11:00.0000] <Hixie>
jesus

[17:12:00.0000] <Hixie>
looks fine to me, in my 2 minute scan

[17:12:01.0000] <Hixie>
and spot checks

[17:12:02.0000] <Hixie>
i expect will people reasoning for many of the changes

[17:13:00.0000] <Hixie>
er

[17:13:01.0000] <Hixie>
data corruption error on output

[17:13:02.0000] <Hixie>
retrying

[17:13:03.0000] <jwalden>
will want?

[17:13:04.0000] <Hixie>
i expect people will want reasoning for many of the changes

[17:13:05.0000] <Hixie>
which is to say, i think nobody will ever be happy :-)

[17:13:06.0000] <Hixie>
but wow, that's a big document

[17:13:07.0000] <Hixie>
did i really do that much in the last 4 months?

[17:14:00.0000] <MikeSmith>
Hixie: I take "nobody will ever be happy" as a given :)

[17:14:01.0000] <Hixie>
indeed

[17:14:02.0000] <jwalden>
MikeSmith: 6.4, target of the event is the window on which it was called, not that window's document

[17:14:03.0000] <Hixie>
me too :-)

[17:15:00.0000] <jwalden>
assuming the version diffed here got that change

[17:15:01.0000] <MikeSmith>
jwalden: thanks, will edit it now

[17:16:00.0000] <MikeSmith>
jwalden: version diffed there is the current version, up to date with the most recent checkin

[17:16:01.0000] <jwalden>
okay

[17:16:02.0000] <jwalden>
it would have, then

[17:16:03.0000] <annevk>
/me deletes www-archive spam from MikeSmith :)

[17:16:04.0000] <jwalden>
there was a couple-day window where it could have been okay

[17:19:00.0000] <annevk>
Lachy, standardssuck.com and standardssuck.net now work

[17:20:00.0000] <Lachy>
LOL

[17:22:00.0000] <Lachy>
I expected you to just set up a redirect

[17:22:01.0000] <Lachy>
<meta name="refresh" content="0;url=http://standardssuck.org"/> should work ;-)

[17:23:00.0000] <annevk>
Redirect 301 ... too, but that's boring :)

[17:23:01.0000] <annevk>
(actually, dreamhost provides the Redirect 301 ... option without having to set up some kind of hosting thingie around it)

[17:25:00.0000] <MikeSmith>
annevk: sorry, probably I should send those to member-only version of www-archive so as to spam fewer people

[17:25:01.0000] <annevk>
MikeSmith, neh, this is way better

[17:25:02.0000] <annevk>
we're a public group, the more we can drag into the open the better

[17:31:00.0000] <annevk>
btw, Ubuntu 8.04 is shipping with Firefox beta 5... isn't that kind of defeating the purpose of having final versions?

[17:32:00.0000] <Philip`>
Google taught us that we should store our critical data in beta services

[17:33:00.0000] <Philip`>
/me apparently has precisely 16384 messages on "Google(TM) Mail BETA"

[17:35:00.0000] <takkaria>
annevk: it'll be updated to the final when that comes out

[17:36:00.0000] <annevk>
takkaria, sure, but users are getting it before "wir"

[17:36:01.0000] <Philip`>
How quickly do users switch to the latest version of Ubuntu? (i.e. how many will get the pre-release version of Firefox?)

[17:37:00.0000] <annevk>
through Ubuntu 8.04

[17:37:01.0000] <Philip`>
/me would assume lots of people migrate immediately when the new version is released, and then it'd quickly drop off until there's just a steady rate of new users

[17:38:00.0000] <annevk>
all the people getting a new PC with Ubuntu will have it

[17:39:00.0000] <annevk>
seems people are already discussing this elsewhere: http://ubuntuforums.org/showthread.php?t=734176

[17:40:00.0000] <takkaria>
I can't be the only person that finds it funny that people are calling photo-sharing sites like Flickr an edge-case

[17:40:01.0000] <takkaria>
(and immensely frustrating)

[17:40:02.0000] <Philip`>
annevk: https://bugzilla.mozilla.org/show_bug.cgi?id=421482#c32 too

[17:41:00.0000] <Philip`>
takkaria: Flickr is one site out of tens of millions, which seems a bit edgy

[17:42:00.0000] <Philip`>
and I don't imagine there are hundreds of thousands of other photo sharing sites

[17:43:00.0000] <annevk>
Philip`, it seems that photo sharing in general is an edge case scenario according to certain people

[17:43:01.0000] <othermaciej>
Flickr is one very popular site with millions of pages

[17:43:02.0000] <jgraham>
Photo sharing must be one of the most common ways that most people use the web

[17:43:03.0000] <annevk>
Though maybe someone should ask to be sure...

[17:44:00.0000] <othermaciej>
#39 on Alexa top 500

[17:44:01.0000] <annevk>
Won't be me though, as I still consider myself lucky that I didn't get any replies to the latest e-mail I threw into that pit

[17:44:02.0000] <othermaciej>
ImageShack is #34

[17:44:03.0000] <jgraham>
it's not just flickr, pretty much all social networking sites have some sort of photo functionality

[17:44:04.0000] <othermaciej>
photobucket is #26

[17:46:00.0000] <jgraham>
Doesn't facebook (#8) have photo sharing?

[17:47:00.0000] <jgraham>
Also presumably Myspace

[17:47:01.0000] <takkaria>
yeah, both of those do

[17:47:02.0000] <Philip`>
I imagine this is the point where trendy people mention the Long Tail - the top few sites are insignicant compared to the hundreds of billions of other web pages, so we shouldn't optimise for the former at the expense of the latter

[17:47:03.0000] <annevk>
Matt also seems to skew the statistics a bit by including the trivial iconic images Flickr serves into his comparisons...

[17:48:00.0000] <annevk>
Philip`, is there evidence that the Long Tail doesn't do photo sharing?

[17:48:01.0000] <jgraham>
(e.g. via gallery and so on)

[17:51:00.0000] <Philip`>
annevk: I have no idea; but Flickr and ImageShack and Photobucket and Facebook and MySpace aren't counter-evidence since they're not part of the tail at all

[17:52:00.0000] <othermaciej>
Philip`: that's not what the Long Tail means

[17:52:01.0000] <othermaciej>
Philip`: that's getting it backwards

[17:52:02.0000] <othermaciej>
the power law distribution means the top few sites get as much traffic/attention as the whole long tail combined

[17:54:00.0000] <jgraham>
othermaciej: I think I remember a survey that backed up that hypothesis (although I think it was one of those "ask people what they do" type setups which are horribly biased)

[17:54:01.0000] <othermaciej>
jgraham: I'm sure anyone with raw traffic stats could prove it, but even just knowing that the distribution is a power law proves it

[17:55:00.0000] <jgraham>
othermaciej: Yeah I know, but how do you know that the distribution is a power law?

[17:56:00.0000] <othermaciej>
all available evidence seems consistent with that hypothesis

[17:56:01.0000] <othermaciej>
it may be that every available sample that has been disclosed is biased

[17:57:00.0000] <MikeSmith>
hsivonen: I thought hendry may have built his Iris browser from source at George Staikos code.staikos.net, but browsing there now, I don't see any subdirectory for the Iris stuf

[17:57:01.0000] <jgraham>
I agree if you can actually measure it you can see if it's a power law or not. 

[17:57:02.0000] <othermaciej>
I just had coffee with George Staikos

[17:57:03.0000] <othermaciej>
I think their WebKit ports are open source but the browser is not

[17:58:00.0000] <Hixie>
photos are an edge case

[17:58:01.0000] <Hixie>
haha

[17:58:02.0000] <Hixie>
that's so funny

[17:58:03.0000] <othermaciej>
jgraham: power law distribution can be detected from samples, the question is just whether the samples are biased

[17:58:04.0000] <MikeSmith>
othermaciej: next time you see George, please tell him I said Hi and that I'm very glad I no longer have to be involved with the CABForum :)

[17:59:00.0000] <jgraham>
The fraction of time people spend on various sites is not obviously a trivial thing to measure

[17:59:01.0000] <Philip`>
othermaciej: All the samples would be heavily biased towards the head of the distribution, and you'd probably have too little data about the tail to work out the distribution

[17:59:02.0000] <Philip`>
(Actually, that's probably totally untrue)

[17:59:03.0000] <MikeSmith>
othermaciej: http://www.torchmobile.com/blog/?p=5 : he says, "Our source code is now imported into the git server at git://code.staikos.net/ alongside the QtWebKit and Torch Mobile Qt WebKit code,"

[18:02:00.0000] <Lachy>
Google Image Search and other similar services is another example

[18:03:00.0000] <othermaciej>
maybe I am wrong

[18:04:00.0000] <othermaciej>
Philip`: broad studies of blogs show that, at least there, the power law goes very deep

[18:04:01.0000] <othermaciej>
the hypothesis that the low end goes off the power law is pretty much eliminated in that case

[18:04:02.0000] <othermaciej>
it could be that for the web in general it is different but I doubt it

[18:09:00.0000] <jgraham>
othermaciej: Interestingly there seem to be a lot of people who claim the powerlaw thing is well supported by evidence but the actual evidence is thin on the ground. The only things I could find related to the number of links pointing to different blogs and the pageviews per month on the Sun website

[18:10:00.0000] <jgraham>
s/is thin/seems to be thin/

[18:10:01.0000] <jgraham>
I would be very surprised if the powerlaw thing doesn't hold for the web at large but I'd like to see the data proving it

[18:14:00.0000] <annevk>
geez, 2 AM

[18:14:01.0000] <othermaciej>
does alexa distribute raw data for their to sites list?

[18:14:02.0000] <othermaciej>
if so you could see if it fits

[18:14:03.0000] <othermaciej>
I have seen technorati's raw data for all the blogs they track and it is a very good fit

[18:15:00.0000] <annevk>
power laws are everywhere :)

[18:16:00.0000] <MikeSmith>
/me finishes reading Dave Singer <timerange> proposal

[18:21:00.0000] <jgraham>
othermaciej: I guess if I wanted to be really difficult I would argue that the navigation model for blogs (largely based around links from a pool of other blogs) isn't the same as the navigation model for ebay, amazon and google (based much more around search)

[18:21:01.0000] <jgraham>
But I don't really want to be difficult and should really sleep instead :)

[18:39:00.0000] <MikeSmith>
/me chuckles at ggaren http://trac.webkit.org/changeset/34091 "Removed terrible lie from ChangeLog"

[18:40:00.0000] <Dashiva>
Maybe that line itself is a lie, and it actually removed a truth

[18:42:00.0000] <MikeSmith>
Dashiva: "All Cretans are liars"

[18:42:01.0000] <Dashiva>
That's what she said

[18:42:02.0000] <MikeSmith>
heh

[18:43:00.0000] <MikeSmith>
hmm, opera dev build making the fan on my mikebook run continuously

[18:44:00.0000] <Hixie>
i don't understand what parse mode webkit and gecko use for innerHTML of <td>

[18:44:01.0000] <Hixie>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3Ea%3Ccaption%3Eb%3Ctable%3E%3Ctd%3EFAIL%3C%2Ftable%3E%0A%3Cscript%3Edocument.getElementsByTagName('td')%5B0%5D.innerHTML%20%3D%20'a%3Ccaption%3Eb'%3C%2Fscript%3E

[18:44:02.0000] <Dashiva>
MikeSmith: Check CPU load

[18:46:00.0000] <MikeSmith>
Dashiva: just up-installed to new build & trying that now

[18:49:00.0000] <Dashiva>
I have recurring cases of random CPU spiking for seconds or minutes myself, so it could be what you had

[18:58:00.0000] <MikeSmith>
Cpu(s):  1.0%us, 21.2%sy,  0.0%ni, 76.2%id,  1.3%wa,  0.0%hi,  0.3%si,  0.0%st

[18:58:01.0000] <MikeSmith>
  8693 mike      20   0  307m 121m  18m S 16.7 24.0   2:23.09 opera

[18:58:02.0000] <MikeSmith>
which don't look bad

[18:58:03.0000] <MikeSmith>
yet still it seems to be causing my fan to kick in

[19:05:00.0000] <Philip`>
16.7% CPU load seems pretty heavy...

[19:23:00.0000] <Hixie>
annevk: yt?

[19:32:00.0000] <Hixie>
anne: did you ever reply to this one? http://lists.w3.org/Archives/Public/public-webapi/2008May/0171.html

[20:06:00.0000] <Lachy>
Hixie, if an author uses <col irrelevant="">, does that implicitly make all the cells in that column irrelevant too?

[21:29:00.0000] <MikeSmith>
/me ponders the idea of a commit hook that for each change to the spec would find the nearest ancestor <h2> for each part of the change, and record that data in the diff

[21:32:00.0000] <MikeSmith>
gnu diff as that built-in behavior that for C code, it tells you which function the change occurred in

[21:32:01.0000] <MikeSmith>
"-p  --show-c-function : Show which C function each change is in"

[21:33:00.0000] <MikeSmith>
extending that to handle looking for <h2> might be something reasonably do-able and useful

[21:34:00.0000] <MikeSmith>
or <hN>

[23:05:00.0000] <Hixie>
MikeSmith: i'd be happy to replace the diff i use if you can code that up :-)

[23:05:01.0000] <Hixie>
Lachy: no

[23:20:00.0000] <MikeSmith>
/me downloads the GNU diff sources to take a look

[23:22:00.0000] <MikeSmith>
Hixie: I'm also exploring the idea of adding to the W3C CVS webview interface a way to generate HTML diffs between any two arbitrary versions of an HTML page maintained in W3C CVS space

[23:22:01.0000] <MikeSmith>
that also should be pretty straightforward to set up

[23:22:02.0000] <MikeSmith>
we have an online HTML Diff interface already:

[23:22:03.0000] <MikeSmith>
http://www.w3.org/2007/10/htmldiff

[23:23:00.0000] <MikeSmith>
it'd mostly be a matter of setting up the CVS webview to be able to pass to the URLs like the following

[23:23:01.0000] <MikeSmith>
http://dev.w3.org/cvsweb/~checkout~/html5/spec/Overview.html?rev=1.877&content-type=text/plain

[23:24:00.0000] <MikeSmith>
http://dev.w3.org/cvsweb/~checkout~/html5/spec/Overview.html?rev=1.878&content-type=text/plain

[23:24:01.0000] <MikeSmith>
pairs of revisions

[23:24:02.0000] <Hixie>
yeah i use that to generate index-diff

[00:12:00.0000] <MikeSmith>
othermaciej: yt?

[00:12:01.0000] <othermaciej_>
MikeSmith: yeah, what's up?

[00:14:00.0000] <MikeSmith>
othermaciej: about http://www.hixie.ch/tests/adhoc/perf/dom/artificial/core/001.html in iPhone

[00:14:01.0000] <MikeSmith>
Safari on iPhone

[00:14:02.0000] <MikeSmith>
I guess you know this already

[00:15:00.0000] <MikeSmith>
that test doesn't seem work unless the count is cranked down a bit

[00:16:00.0000] <MikeSmith>
e.g.: http://www.pureanarchy.com/hixietest.html

[00:17:00.0000] <othermaciej_>
MikeSmith: iPhone has a JS execution time limit and I would not be surprised if that test blows through it

[00:17:01.0000] <MikeSmith>
aha

[00:17:02.0000] <MikeSmith>
so I tried it on Opera Mobile on my own handset here

[00:18:00.0000] <MikeSmith>
and it also didn't work, so yeah, I'd guess it's the time limit

[00:18:01.0000] <MikeSmith>
other kinda weird thing about the test results with the adjusted-count page at URL above

[00:19:00.0000] <MikeSmith>
is the Index test result

[00:20:00.0000] <MikeSmith>
which is much slower relative to the other test than it is in Webkit running on my macbook

[00:21:00.0000] <othermaciej_>
how recent is your build?

[00:22:00.0000] <othermaciej_>
in Safari 3.1 the Index time is plenty fast to me, even more so in a trunk, but indeed it used to be much slower before 3.1

[00:22:01.0000] <MikeSmith>
othermaciej_: my desktop webkit is the latest nightly

[00:23:00.0000] <othermaciej>
I just tried it in a fresh build of trunk

[00:23:01.0000] <othermaciej>
MikeSmith: http://paste.lisp.org/display/61193

[00:23:02.0000] <othermaciej>
those are my results

[00:24:00.0000] <MikeSmith>
yeah, those are in line with what I get on my machine

[00:24:01.0000] <MikeSmith>
what I meant was the results in Safari on the iPhone

[00:25:00.0000] <MikeSmith>
on iPhone, the Index test is much slower relative to the other tests

[00:25:01.0000] <MikeSmith>
on desktop, it is one of the fastest tests

[00:25:02.0000] <othermaciej>
it's probably got something closer to the Safari 3 code

[00:25:03.0000] <MikeSmith>
on iPhone it is the slowest]

[00:26:00.0000] <MikeSmith>
OK

[00:26:01.0000] <othermaciej>
where the index test was slow

[00:26:02.0000] <othermaciej>
due to a bug in caching for nodeLists

[00:26:03.0000] <MikeSmith>
ah

[02:34:00.0000] <annevk>
lol, somebody did indeed bring up all the other sites out there

[02:34:01.0000] <annevk>
how very predictable

[02:35:00.0000] <annevk>
Hixie, replied

[02:35:01.0000] <Hixie>
hm, you rejected it :-(

[02:36:00.0000] <Hixie>
Sec-Origin would be really cool because if browsers included it with everything, it could be used to reject CSRFs more reliably than Referer

[02:37:00.0000] <Hixie>
oh well

[02:39:00.0000] <Hixie>
annevk: what's the status on the header stuff? did you end up making any decisions on that?

[02:42:00.0000] <annevk>
Hixie, it's easy to change, just comment

[02:42:01.0000] <Hixie>
i hope adam will

[02:42:02.0000] <annevk>
Hixie, though shouldn't we just call it Origin then?

[02:42:03.0000] <Hixie>
(he's got a better handle on the implications)

[02:42:04.0000] <Hixie>
(or collin, maybe)

[02:43:00.0000] <Hixie>
i guess Sec-* isn't safe in old XHRs either huh

[02:43:01.0000] <annevk>
it will be going forward, but isn't now

[02:43:02.0000] <annevk>
maybe in WebKit it is already :)

[02:43:03.0000] <Hixie>
that's the main reason to use Sec-

[02:44:00.0000] <annevk>
yeah

[02:44:01.0000] <annevk>
Hixie, I think I just want to leave AC as is

[02:45:00.0000] <annevk>
Hixie, apart from naming changes such as dropping "Access-Control-" from "Access-Control-Origin"

[02:45:01.0000] <Hixie>
cool

[02:45:02.0000] <Hixie>
last call time then?

[02:45:03.0000] <annevk>
for AC, yes, although maybe I should more explitly call out the problems with Access-Control-Policy-Path

[02:46:00.0000] <annevk>
in the security section for authors

[02:46:01.0000] <Hixie>
i'd be so happy if you could make the f2f pointless

[02:46:02.0000] <Hixie>
i don't want to travel :-)

[02:48:00.0000] <annevk>
i'll reply to my own e-mail

[02:48:01.0000] <annevk>
for the origin thingie

[02:48:02.0000] <Hixie>
heh k

[02:49:00.0000] <othermaciej>
Hixie: what is Sec-Origin and why would it be more reliable than Referer?

[02:49:01.0000] <Hixie>
origin wouldn't have the path information, so it wouldn't be stripped by https->http posts, for instance

[02:50:00.0000] <Hixie>
though i guess it could still be stripped by over-zealous privacy advocates

[02:50:01.0000] <Hixie>
and so you'd still have to assume its absence meant same-origin

[02:50:02.0000] <Hixie>
<Hixie> and so you'd still have to assume its absence meant same-origin

[02:50:03.0000] <Hixie>
adam can explain his idea better

[02:51:00.0000] <othermaciej>
If path info is the only reason to strip Referer then it would be a simple matter to include Referer but remove path info from it

[02:51:01.0000] <othermaciej>
but I am not sure it is the only reason

[02:55:00.0000] <annevk>
if that's the case maybe it should take magical values "same-origin" and "non same-origin" so you can at least know what is happening in case the URI is not exposed

[02:56:00.0000] <annevk>
though maybe middleware will just strip it out

[02:56:01.0000] <Hixie>
does hsivonen have a page anywhere that exposes his parser?

[02:56:02.0000] <Hixie>
oh nm

[02:56:03.0000] <Hixie>
http://parsetree.validator.nu/

[02:57:00.0000] <Hixie>
hm

[02:57:01.0000] <Hixie>
doesn't work

[02:57:02.0000] <annevk>
he could add <textarea> now I think

[02:57:03.0000] <Hixie>
http://parsetree.validator.nu/?doc=http%3A%2F%2Fjunkyard.damowmow.com%2F323&submit=Print+Tree

[02:59:00.0000] <annevk>
hsivonen, ^^

[02:59:01.0000] <Hixie>
he has mail about this already

[02:59:02.0000] <Hixie>
so do you in fact

[02:59:03.0000] <Hixie>
:-)

[03:00:00.0000] <annevk>
i'm sort of doubtful we can avoid the F2F btw

[03:00:01.0000] <annevk>
we'll see how it goes I guess

[03:00:02.0000] <Hixie>
that's not the right attitude :-)

[03:00:03.0000] <Hixie>
i thought you thought you were ready for LC months ago but that mozilla was the only thing blocking you?

[03:00:04.0000] <Hixie>
we've gotten rid of that blocker now

[03:01:00.0000] <annevk>
we did?

[03:02:00.0000] <annevk>
hmm, parsetree does work for other stuff, e.g. http://parsetree.validator.nu/?doc=http%3A%2F%2Fannevankesteren.com%2F

[03:06:00.0000] <Hixie>
mozilla has sent all the feedback they have and will not be blocking XHR/AC on the grounds of cookies being insecure, yes

[03:06:01.0000] <Hixie>
at least, that's the assurance i have received

[03:08:00.0000] <annevk>
i think i'll take your IIS and Bjoern's Apache example and put them in the spec

[03:09:00.0000] <annevk>
maybe example.org yours first :)

[03:10:00.0000] <Hixie>
what did i use? :-)

[03:10:01.0000] <annevk>
microsoft.com

[03:10:02.0000] <Hixie>
hah

[03:11:00.0000] <Hixie>
i wish example.com and example.org didn't have the same name

[03:12:00.0000] <annevk>
there's something.invalid

[03:12:01.0000] <annevk>
wait, something.example

[03:12:02.0000] <Hixie>
yeah but that looks invalid :-)

[03:12:03.0000] <Hixie>
and .example looks to long

[03:12:04.0000] <annevk>
true, they don't give us much choice for describing same-origin and all that

[03:13:00.0000] <annevk>
(i did use hello-world.example and such fwiw, because I can't be bothered to try setting a new standard there)

[03:13:01.0000] <Hixie>
heh

[03:14:00.0000] <annevk>
the reason IE puts <noframes> in <head> btw is because IE ignores </head>

[03:17:00.0000] <Hixie>
yeah I just put it in head because it was the least change to the spec

[03:18:00.0000] <Hixie>
the alternative was putting it in between <head> and <body>, and that seemed far, far more complex

[03:26:00.0000] <annevk>
Hixie, btw, hsivonen also raised the EOF issue and he has a longer list of elements he'd like to see added (besides h4)

[03:26:01.0000] <Hixie>
if he sent mail, i'm sure i'll get to it

[03:27:00.0000] <annevk>
the batch processing is over?

[03:27:01.0000] <Hixie>
i use batch processing when there are enough mail to batch process

[03:27:02.0000] <Hixie>
and when they are all sent to the same list

[03:55:00.0000] <MikeSmith>
/me would also be really happy to avoid traveling

[03:56:00.0000] <MikeSmith>
let's have all f2f meetings in Tokyo

[03:56:01.0000] <MikeSmith>
you can stay in my apartment

[03:56:02.0000] <MikeSmith>
I have room for 1.5 people

[03:56:03.0000] <Hixie>
i recommend having the meetings on irc and e-mail

[03:56:04.0000] <roc>
amen

[03:57:00.0000] <MikeSmith>
word

[04:03:00.0000] <gsnedders>
MikeSmith: who you planning on cutting in two?

[04:04:00.0000] <MikeSmith>
gsnedders: you don't want two of me

[04:04:01.0000] <MikeSmith>
trust me on that

[04:04:02.0000] <annevk>
ok, I think AC is ready for LC

[04:05:00.0000] <annevk>
I think it's better for XHR2 to wait until at least all feedback on XHR1 LC is in

[04:05:01.0000] <MikeSmith>
annevk: so let's start LC process for AC

[04:05:02.0000] <MikeSmith>
we will need ArtB

[04:06:00.0000] <annevk>
I said in "Re: Moving forward with XHR2 and AC" that it is ready for LC

[04:06:01.0000] <annevk>
should we get ArtB to formally ask the WG?

[04:07:00.0000] <MikeSmith>
I think we need him to do that

[04:08:00.0000] <MikeSmith>
likely you are actually more familiar with W3C WG process stuff around this than me

[04:08:01.0000] <MikeSmith>
but I don't think it can be done without the chair

[04:08:02.0000] <gsnedders>
/me passes MikeSmith a n00b badge

[04:09:00.0000] <annevk>
MikeSmith, from my recollection we need WG agreement (either implicit or explicit the latter being preferred)

[04:10:00.0000] <MikeSmith>
/me tosses n00b badge into his pile with shitbird and jaggass badges

[04:10:01.0000] <MikeSmith>
annevk: I guess

[04:11:00.0000] <MikeSmith>
but getting that is a lot more of potential fubar situation than just getting artb on board

[04:12:00.0000] <gsnedders>
When does LC on XHR1 end?

[04:12:01.0000] <gsnedders>
(i.e., by when do I need to get off my lazy ass?)

[04:13:00.0000] <annevk>
Monday June 2

[04:13:01.0000] <gsnedders>
The day of my final exam. Heh.

[04:16:00.0000] <MikeSmith>
annevk: at this point it's clear that we have to extend the WAF WG charter til end of June at least anyway

[04:18:00.0000] <MikeSmith>
so getting a WD out for *public* review in that window of time would be great

[04:19:00.0000] <MikeSmith>
I was going to say that I would honestly not want to make getting that done contingent on consensus of the WG

[04:19:01.0000] <MikeSmith>
but on second thought

[04:19:02.0000] <MikeSmith>
..

[04:20:00.0000] <MikeSmith>
most of the lets-slow-this-down comments

[04:20:01.0000] <MikeSmith>
have come from "the public"

[04:20:02.0000] <MikeSmith>
not from active members of the group

[04:21:00.0000] <MikeSmith>
e.g., the "I'm commenting on this as an individual, not representing my employer"

[04:21:01.0000] <MikeSmith>
which is a lie/BS to begin with

[04:23:00.0000] <MikeSmith>
and then the assorted other people-who-by-very-deliberate-choice-are-not-members-of-the-WG with whatever agenda it is they have 

[04:23:01.0000] <MikeSmith>
anyway, I say too much

[04:23:02.0000] <MikeSmith>
as usua

[04:23:03.0000] <MikeSmith>
sual

[04:23:04.0000] <annevk>
so I just remembered we need to settle on naming of Access-Control-Origin before publication

[04:23:05.0000] <MikeSmith>
usual

[04:23:06.0000] <MikeSmith>
annevk: yeah.

[04:23:07.0000] <annevk>
but hopefully that can be resolved within a few days

[04:23:08.0000] <MikeSmith>
decide.

[04:24:00.0000] <MikeSmith>
annevk: resolve it.

[04:24:01.0000] <annevk>
well, I'd like to hear the opinion of Adam / Collin first

[04:24:02.0000] <MikeSmith>
annevk: yes

[04:24:03.0000] <MikeSmith>
they are among the few

[04:24:04.0000] <MikeSmith>
the very few

[04:24:05.0000] <annevk>
they're quite awesome from what I've seen :)

[04:24:06.0000] <MikeSmith>
yep

[04:24:07.0000] <MikeSmith>
they have earned it

[04:24:08.0000] <MikeSmith>
through merit

[04:25:00.0000] <MikeSmith>
they speak rarely

[04:25:01.0000] <MikeSmith>
but when they do..

[04:25:02.0000] <MikeSmith>
they have earned the right to be listened to carefully

[04:26:00.0000] <Hixie>
annevk: when does XHR1 LC end?

[04:27:00.0000] <Dashiva>
That's the sage's quandry. The less you speak, the more awed are people when you actually do, but the less often you're able to dispense wisdom :)

[04:27:01.0000] <Hixie>
adam and collin are awesome

[04:27:02.0000] <MikeSmith>
Dashiva: actually, speaking ain't worth nearly as much as other things

[04:28:00.0000] <MikeSmith>
Adam and Collin walk the walk

[04:29:00.0000] <MikeSmith>
as in (just for one example) coding/patching Webkit code around some of this tuff

[04:29:01.0000] <MikeSmith>
stuff

[04:29:02.0000] <annevk>
Hixie, Mon June 2

[04:30:00.0000] <Hixie>
is there any feedback you have received on xhr1 and not yet dealt with in xhr2?

[04:31:00.0000] <annevk>
yes, what to do with send(document) if document can't be serialized for instance

[04:31:01.0000] <annevk>
well, some people disagree with how it's handled, and apparently it matches neither Firefox or IE, but I think everyone is doing something else...

[04:32:00.0000] <annevk>
another issue is URI/IRI confusion

[04:32:01.0000] <annevk>
maybe I should put them in the issue tracker

[04:32:02.0000] <Hixie>
i recommend dealing with that now then, and then once you've handled all that, pushing to get XHR2 into LC, even if XHR1 hasn't finished LC yet -- then if there are major issues, rerelease XHR2 as a second last call quickly before its first LC is over.

[04:32:03.0000] <annevk>
that would at least give an overview

[04:33:00.0000] <Hixie>
that would, i think, be the most efficient way of getting all this to CR

[04:33:01.0000] <Hixie>
you generally want to avoid putting yourself in a position of having feedback you can't deal with (e.g. by waiting on someone else's input)

[04:33:02.0000] <MikeSmith>
annevk: fwiw, I think a 2nd LC maybe basically an inevitability anyway

[04:34:00.0000] <annevk>
Hixie, I agree, but URIs/IRIs are very confusing and it's not at all clear what the right solution is to me :)

[04:34:01.0000] <annevk>
Hixie, I think I might need the HTML5 URL stuff

[04:35:00.0000] <Hixie>
read the specs :-)

[04:35:01.0000] <Hixie>
and test implementations :-)

[04:35:02.0000] <Hixie>
that's the way to make things clearer :-)

[04:36:00.0000] <MikeSmith>
/me records milestone of checkin number 1700 against HTML5 spec source, notes that at current rate, checkin 2000 will be reached within 2 monhts

[04:37:00.0000] <Dashiva>
MikeSmith: Have you seen the issues graph?

[04:37:01.0000] <Hixie>
the number of checkins per day depends more on the latency of the various servers i use and on my mood than on how much work i do

[04:37:02.0000] <annevk>
MikeSmith, btw, aside from sandboxing and details like document.charset I'd be interested in hearing what html5-diff doesn't cover currently

[04:37:03.0000] <MikeSmith>
Dashiva: yep

[04:38:00.0000] <Hixie>
because if the computers aren't working with me, or if i'm being too productive, i batch changes into bigger changesets

[04:38:01.0000] <Hixie>
(too productive for the spec regen script to keep up, i mean)

[04:38:02.0000] <Hixie>
annevk: i've no idea :-)

[04:38:03.0000] <MikeSmith>
annevk: read through my doc. I lack the time to enumerate what may be missing

[04:39:00.0000] <annevk>
k, because you thought it missed something you'd like to see there...

[04:39:01.0000] <annevk>
MikeSmith, the chair needs to be involved in LC stuff most definitely btw, e-mailing chairs⊙wo and such

[04:40:00.0000] <Hixie>
annevk: oh?

[04:41:00.0000] <annevk>
Hixie, not you, MikeSmith :)

[04:41:01.0000] <Hixie>
oh

[04:41:02.0000] <Hixie>
oh, oops, the line got highlighted cos it included the word "html5" and i assumed you were talking to me!

[04:41:03.0000] <MikeSmith>
annevk: yeah. but ArtB is generally supportive of moving work forward rather than allowing it to be bogged down or tar-babied

[04:41:04.0000] <annevk>
hah

[04:41:05.0000] <Hixie>
nm :-)

[04:42:00.0000] <Dashiva>
MikeSmith: It's a bit worrying how the issues value in the graph doesn't seem to change even though the revision number goes up and up :)

[04:42:01.0000] <MikeSmith>
and Art is not particularly happy about the fact that the grand WG merger has now taken 7+ months to not happen

[04:43:00.0000] <MikeSmith>
so in the mean time, we hav things to get done

[04:44:00.0000] <MikeSmith>
Dashiva: /me points to Hixie about that.. 

[04:45:00.0000] <Hixie>
Dashiva: the XXX count will go up until the e-mails hit near zero

[04:45:01.0000] <MikeSmith>
annevk: Access Control actually getting implemented in some UAs will be a huge help

[04:45:02.0000] <Hixie>
Dashiva: i'm not spending any time tracking down XXXs

[04:45:03.0000] <Hixie>
Dashiva: i just add them when i see something wrong as i'm editing other things

[04:45:04.0000] <MikeSmith>
++ to Webkit and and Sam Weinig and othermaciej for helping get AC actually implemented

[04:46:00.0000] <MikeSmith>
annevk: maybe you can do something about getting some Opera support behind it

[04:46:01.0000] <Hixie>
gotta love people who have clearly not yet worked out the scale of the web

[04:46:02.0000] <Hixie>
"1.7% is nothing!"

[04:46:03.0000] <MikeSmith>
Hixie: heh

[04:46:04.0000] <Hixie>
uh no. welcome to the web.

[04:47:00.0000] <Hixie>
1.7% is probably more pages than you have ever looked at it in your entire life.

[04:47:01.0000] <Hixie>
and more people than you've met or dealt with ever.

[04:47:02.0000] <MikeSmith>
we need better need ways of modding down all the noise

[04:47:03.0000] <Hixie>
i mean, elements used only on 0.01% of pages can still have giant interop issues

[04:48:00.0000] <MikeSmith>
Aaron Leventhal has some good ideas about how to deal with representing "truth structures" in public discussions

[04:49:00.0000] <MikeSmith>
... hope we can make time to actually do some of that

[04:50:00.0000] <Dashiva>
Hixie: Yeah, you told me that before. I'm just enjoying the perspective of someone who doesn't know :)

[04:50:01.0000] <MikeSmith>
e.g., I never want to see any message about something that has proven to be false

[04:50:02.0000] <MikeSmith>
... or that at least there is general agreement is false

[04:51:00.0000] <MikeSmith>
what I want is: bubble up to my attention the arguments that have some degree of agreement as being viable

[04:51:01.0000] <MikeSmith>
and please god give me a way to drown out the noise

[04:51:02.0000] <MikeSmith>
e.g., RB

[04:51:03.0000] <Hixie>
Dashiva: :-P

[04:52:00.0000] <Hixie>
Dashiva: actually i do occasionally deal with XXXs. Mostly when my mail server goes down. :-)

[04:52:01.0000] <roc>
a friend of mine is working on a social Web site for representing truth structures

[04:52:02.0000] <roc>
it's doomed

[04:52:03.0000] <Dashiva>
Hixie: Do you still do that, or did it stop when you got GTA4?

[04:52:04.0000] <roc>
BTW doesn't Jonas deserve a ++ for getting AC implemented too?

[04:53:00.0000] <Hixie>
Dashiva: you can see where i got GTA4 on that chart

[04:53:01.0000] <Hixie>
Dashiva: it's just before the last hump before the last big fall

[04:53:02.0000] <Hixie>
Dashiva: (april 29th)

[04:53:03.0000] <annevk>
roc, yeah

[04:53:04.0000] <annevk>
sicking++

[04:53:05.0000] <Hixie>
Dashiva: took me out for a week :-)

[04:53:06.0000] <Dashiva>
Hehe, yeah

[04:54:00.0000] <annevk>
roc, been quite useful in discussions so far when people complain it's too hard

[04:54:01.0000] <MikeSmith>
roc: yeah, Jonas deserves ++ for many many things

[04:54:02.0000] <annevk>
MikeSmith, e-mailed Art + you

[04:54:03.0000] <MikeSmith>
annevk: thanks

[04:55:00.0000] <MikeSmith>
general comment: the way things are now, we have N different ones of us taking time to set up our own private filtering

[04:55:01.0000] <MikeSmith>
Hixie gots his AAA-important folder and filter

[04:56:00.0000] <MikeSmith>
I got my my less-tactfully-worded bozo filter

[04:56:01.0000] <MikeSmith>
we could really benefit from some way to share our bozo rules

[04:56:02.0000] <Hixie>
it's called "aaa-productivity"

[04:56:03.0000] <Hixie>
for extra irony

[04:57:00.0000] <MikeSmith>
/me is taking a moment for a belly laugh

[04:57:01.0000] <MikeSmith>
anyway, a pipe dream

[04:57:02.0000] <Dashiva>
If you can get away with it without being accused of censoring and favoritism, excellent

[04:58:00.0000] <MikeSmith>
we will continue to be plagued by the likes of RB I guess

[04:58:01.0000] <MikeSmith>
Dashiva: not censoring, not favortism

[04:58:02.0000] <MikeSmith>
not at all

[04:58:03.0000] <MikeSmith>
that't the pont

[04:58:04.0000] <MikeSmith>
point 

[04:59:00.0000] <MikeSmith>
the whole "wisdom of the crowds" ideal

[04:59:01.0000] <MikeSmith>
modding it collectively

[04:59:02.0000] <Dashiva>
MikeSmith: I know, but do you think the same people who create all this noise would stand silent, regardless of rational arguments?

[04:59:03.0000] <MikeSmith>
which may or may not actually work in practice

[05:00:00.0000] <MikeSmith>
Dashiva: we don't need a way for them to stand silent

[05:00:01.0000] <MikeSmith>
we just need a way to mod them down

[05:00:02.0000] <Hixie>
nn

[05:00:03.0000] <MikeSmith>
Hixie: otsukare-sama and o-yasumi

[05:00:04.0000] <Dashiva>
Does Hixie dream of electric spec writers?

[05:02:00.0000] <Dashiva>
MikeSmith: I guess I'm a bit pessimistic with regard to factionism

[05:02:01.0000] <MikeSmith>
not factionism

[05:12:00.0000] <MikeSmith>
annevk: see my reply

[05:17:00.0000] <MikeSmith>
/me turns to reading http://lists.w3.org/Archives/Public/public-html/2008May/0598.html

[05:18:00.0000] <MikeSmith>
hsivonen: I can't say I'm particularly happy about the completely arbitrary requirement that charset encoding markup declaration must occur within first 512 bytes

[05:19:00.0000] <MikeSmith>
real-worl documents may have comments much longer than that

[05:20:00.0000] <annevk>
what's wrong with <!DOCTYPE html><meta charset=x><!-- really really long comment --> ?

[05:20:01.0000] <MikeSmith>
Hixie reply about "multimegabyte comment" seems to me like a red herring

[05:20:02.0000] <MikeSmith>
or whatever other rhetorical term is appropriate here

[05:21:00.0000] <MikeSmith>
there are in practice no "multimegabyte comments"

[05:21:01.0000] <MikeSmith>
so not even sure sure wtf he is talking about there

[05:21:02.0000] <annevk>
and even if they are, there's been no reason given they can't occur after a <meta>

[05:21:03.0000] <takkaria>
gsnedders: re: http5, for some sites (like royalmail.com) require you to send an Accept: header or they don't serve up any content

[05:22:00.0000] <annevk>
<meta> can have a reparsing effect, longs comments won't

[05:22:01.0000] <MikeSmith>
but 512 seems a totally arbitrary and effectively counter-productive criterion to me

[05:22:02.0000] <annevk>
well, if browsers do a prescan, it's likely over the first packet or something, which is 512 bytes

[05:23:00.0000] <gsnedders>
MikeSmith: HTML 5 has some really large comments in it :P

[05:24:00.0000] <MikeSmith>
and then there's the criterion that meta@charset must be the 1st child element of head.. I really don't care why that was added, but it it is an absolutely broken requirement

[05:24:01.0000] <gsnedders>
takkaria: FWIW, I'm currently only working on defining parsing. Things like that are currently out of scope — HTTP5 is a purely theoretical spec that could be based upon the parsing spec.

[05:24:02.0000] <gsnedders>
takkaria: But feel free to send me an email about it

[05:24:03.0000] <gsnedders>
takkaria: That's the sort of thing I have informatively mentioned currently

[05:24:04.0000] <takkaria>
gsnedders: k, will do

[05:26:00.0000] <annevk>
MikeSmith, that's to prevent re execution of scripts and such

[05:27:00.0000] <MikeSmith>
annevk: I see. but I guess you know that getting authors to comply with it in practice is, well... optimistic

[05:29:00.0000] <MikeSmith>
... along with a good number of instances of autoring-conformance criteria in the current draft that basically are completely arbitrary from the POV of authors

[05:30:00.0000] <annevk>
I thought we tried to reduce those

[05:30:01.0000] <MikeSmith>
annevk: trying is not doing

[05:31:00.0000] <annevk>
Sorry, did reduce those...

[05:31:01.0000] <annevk>
Anyway, maybe you should comment on the draft?

[05:31:02.0000] <gsnedders>
takkaria: Sorry, I'm a bit busy rediscovering the awesomeness of NFS:HP2

[05:34:00.0000] <billyjack>
annevk: speaking of arbitrariness 

[05:34:01.0000] <billyjack>
I'm told that the guy who maintains http://logopoeia.com/wisdom/

[05:35:00.0000] <billyjack>
.. that there are some bits of wisdom there (easter eggs)

[05:35:01.0000] <billyjack>
... that if you care to take the time

[05:35:02.0000] <annevk>
isn't "NORMALISATION" with a "Z" in the US?

[05:35:03.0000] <billyjack>
... if you reload that a few tims

[05:35:04.0000] <billyjack>
times

[05:36:00.0000] <billyjack>
e.g. 10 times or so

[05:36:01.0000] <billyjack>
you might discover some real bits of wisdom

[05:36:02.0000] <annevk>
heh

[05:36:03.0000] <billyjack>
wisdom wisdom which you would really not want to paste into this channel 

[05:36:04.0000] <billyjack>
or some other channels

[05:37:00.0000] <billyjack>
but wisdom for your own personal enlightenment

[05:37:01.0000] <billyjack>
like Zen

[05:37:02.0000] <billyjack>
annevk: yeah, Z

[05:37:03.0000] <annevk>
"I live in Tokyo and work as a software non-executive."

[05:38:00.0000] <billyjack>
Hixie has this weird thing for wanting to spell everyhing with S's 

[05:38:01.0000] <gsnedders>
annevk: Technically, it should be z in en-gb too. It is in en-gb-oed.

[05:38:02.0000] <billyjack>
it's nouts

[05:38:03.0000] <billyjack>
nuts

[05:38:04.0000] <annevk>
I think he fixed some of that recently

[05:38:05.0000] <billyjack>
annevk: yeah, he did in many places

[05:39:00.0000] <billyjack>
though I think "Tokenisation" remains

[05:39:01.0000] <billyjack>
in truth, I couldn't care less myself about the UK vs. US spelling

[05:39:02.0000] <gsnedders>
en-gb-x-hixie is defined to use -ize and not -ise, FWIW

[05:40:00.0000] <billyjack>
I think he changed it based on feedback from others

[05:40:01.0000] <gsnedders>
(as -ise is a made up thing from the 20th cent. — -ize is technically correct)

[05:41:00.0000] <billyjack>
maybe in regard to search engines that are too dumb to know that Tokenisation and Tokenization mean exactly and precisely the same thing.. i dunno

[05:41:01.0000] <billyjack>
this is one of the beautiful things about Japanese

[05:41:02.0000] <billyjack>
(the need to not deal with Tokenisation vs Tokenization

[05:42:00.0000] <gsnedders>
billyjack: Google does some normalization

[05:42:01.0000] <billyjack>
I know G does

[05:42:02.0000] <billyjack>
lesser engines seem to have more trouble..

[05:48:00.0000] <jgraham>
gsnedders: You're missing the whole language evolution thing

[05:48:01.0000] <MikeSmith>
gsnedders: btw, so have seen mention here about your university plans and such

[05:48:02.0000] <MikeSmith>
... curious as to what your actual plans are

[05:48:03.0000] <gsnedders>
MikeSmith: Congrats on being able to read!

[05:48:04.0000] <gsnedders>
:P

[05:48:05.0000] <gsnedders>
jgraham: Hey! I want to read anglo-saxon (sp?)!

[05:48:06.0000] <jgraham>
(like pluralising with s is a made-up since roman times thing iirc

[05:48:07.0000] <jgraham>
)

[05:48:08.0000] <MikeSmith>
I hate reading

[05:49:00.0000] <MikeSmith>
plurals are a mostly unnecessary language construct

[05:50:00.0000] <MikeSmith>
you can almost always get away without plurals and still avoid ambiguity

[05:50:01.0000] <MikeSmith>
(works fine in Japanese at least, for the most part)

[05:51:00.0000] <jgraham>
MikeSmith: It's just style. Japanese has counters or something instead, right?

[05:51:01.0000] <jgraham>
s/style/language style/

[05:51:02.0000] <MikeSmith>
counters yeah

[05:52:00.0000] <jgraham>
/me can't actually remember how those work

[05:52:01.0000] <MikeSmith>
every language has counters

[05:52:02.0000] <MikeSmith>
e.g, "four *sheets* of paper" 

[05:52:03.0000] <MikeSmith>
"four *pieces* of paper"

[05:53:00.0000] <MikeSmith>
nothing unique about those in Japanese, really

[05:53:01.0000] <MikeSmith>
four glasses of bear

[05:53:02.0000] <MikeSmith>
beer

[05:53:03.0000] <jgraham>
MikeSmith: Right. I seem to recall that they are a bigger deal in Japanese than in English

[05:53:04.0000] <MikeSmith>
yeah

[05:53:05.0000] <gsnedders>
MikeSmith: In short, as I'm running off, Cambridge's Comp.Sci. BA, Edinburgh's Comp. Phys. MPhys, or York's Theoretical Phys. MPhys

[05:53:06.0000] <MikeSmith>
they are a bit more

[05:53:07.0000] <jgraham>
In the same way that morphology is a bigger deal in Romance languages

[05:53:08.0000] <MikeSmith>
yup

[05:54:00.0000] <MikeSmith>
but in Japanese the general rule is that you can get by without much of that

[05:54:01.0000] <MikeSmith>
and still be "correct"]

[05:55:00.0000] <MikeSmith>
e.g., a grammatically correct sentence requires only a verb

[05:55:01.0000] <MikeSmith>
and no explii

[05:55:02.0000] <MikeSmith>
explicitly-stated subject

[05:56:00.0000] <jgraham>
I don't know any Japanese but it seems like it would be a pretty nice language except for a) Kanji and b) the social structures baked into the language

[05:56:01.0000] <MikeSmith>
kanji is a major major PITA, that's for sure

[05:56:02.0000] <MikeSmith>
but it is not absolutely intellectually difficult

[05:57:00.0000] <MikeSmith>
learning kanji is just brute-force memorization

[05:57:01.0000] <MikeSmith>
basically

[05:57:02.0000] <jgraham>
Right, but learning 10,000 hard-to-draw symbols need not be intellectually difficult to be a major problem :)

[05:57:03.0000] <MikeSmith>
social constructs are a totally different matter

[05:58:00.0000] <MikeSmith>
but for the most part non-native speakers get a "by" on much of those

[05:58:01.0000] <jgraham>
/me has difficulty drawing 26 alphabetic characters with sufficient uniqueness

[05:59:00.0000] <jgraham>
I guess it ends up being just another way that your foreignness is made obvious though

[05:59:01.0000] <MikeSmith>
jgraham: the amount of information that are represented by the complete set of kanji in normal use pales in comparison to the set of information you have in your mind about many other subject areas

[06:00:00.0000] <MikeSmith>
I promise you you that

[06:00:01.0000] <MikeSmith>
and for the most part you don't need to write them

[06:00:02.0000] <MikeSmith>
don't need to hand-write them

[06:00:03.0000] <Lachy>
memorising 10,000 different symbols seems like it would be impossible

[06:01:00.0000] <MikeSmith>
you just need to be able to real them

[06:01:01.0000] <MikeSmith>
Lachy: not impossible 

[06:01:02.0000] <MikeSmith>
not by a long shot

[06:01:03.0000] <Lachy>
on average, how many of them do most japenese people know?

[06:01:04.0000] <MikeSmith>
you have in your mind now many sets of info that are orders of magnitude larger than the set of kanji

[06:02:00.0000] <MikeSmith>
Lachy: minimal adult literacy is on the order of 2000 kanji

[06:02:01.0000] <jgraham>
MikeSmith: I think for one brief part of my life I could recognise the Kanji for Railway Station

[06:02:02.0000] <MikeSmith>
駅

[06:02:03.0000] <MikeSmith>
eki

[06:03:00.0000] <MikeSmith>
part of that character means "horse"

[06:03:01.0000] <Lachy>
how do you type those characters without having 10,000 keys on the keyboard?

[06:03:02.0000] <jgraham>
Yeah, I remember the word. Getting on buses and going "Eki?" was a frequent part of my trip to Japan

[06:03:03.0000] <Lachy>
or even just a few thousand, given modifier keys?

[06:03:04.0000] <MikeSmith>
that's the other thing with many kanji -- they can be decomposed into familiar parts

[06:04:00.0000] <jgraham>
Lachy: There's a system where you type ascii and combinations are turned into kanji

[06:04:01.0000] <jgraham>
I think

[06:04:02.0000] <MikeSmith>
Lachy: invest some time in learning about CJK input modes

[06:04:03.0000] <MikeSmith>
:)

[06:04:04.0000] <MikeSmith>
talk to Wilhelm

[06:04:05.0000] <Lachy>
so not only do japenese people need to learn their own symbols, but they need to learn to type them phonetically in ASCII?

[06:04:06.0000] <MikeSmith>
Lachy: yep

[06:05:00.0000] <MikeSmith>
my daughter who is just 10 years old is now learning for the first time how to represent Japanese in the "roman" alphabet

[06:06:00.0000] <Lachy>
wouldn't it be easier if they just learned a language that used the roman alphabet, like Engrish?

[06:06:01.0000] <MikeSmith>
and the majority of people in japan type Japanese in roman chars on their computer keyboards

[06:07:00.0000] <jgraham>
Lachy: The effort of learning a whole new language seems huge compared to learning a different representation of a familiar grammar

[06:07:01.0000] <MikeSmith>
so if I want to type 駅, I type "e"+"ki" to get えき, then use the IME to change that to 駅

[06:08:00.0000] <MikeSmith>
Lachy: roman alphabet sucks in many ways

[06:08:01.0000] <Lachy>
in what way?

[06:08:02.0000] <MikeSmith>
alphabets in general are deficient

[06:08:03.0000] <MikeSmith>
Lachy: for one thing -

[06:09:00.0000] <MikeSmith>
take the the symbol, 人

[06:09:01.0000] <MikeSmith>
it means the same thing in Chinese and Japenese

[06:09:02.0000] <MikeSmith>
Japanese

[06:10:00.0000] <MikeSmith>
though its sound/reading is different

[06:10:01.0000] <Lachy>
I asked about the roman alphabet though.

[06:10:02.0000] <MikeSmith>
but you don't need to know the the sound/reading/pronunciation

[06:10:03.0000] <jgraham>
MikeSmith: Are you claiming that is a feature?

[06:10:04.0000] <MikeSmith>
jgraham: yes, absolutely

[06:11:00.0000] <MikeSmith>
it is a superior way to convery information

[06:11:01.0000] <Lachy>
MikeSmith, why is it a feature for the same symbol to be pronounced differently? I call that a bug

[06:11:02.0000] <MikeSmith>
a pure symbol

[06:11:03.0000] <jgraham>
(lots of words in english and, say German have similar alphabetic representations but different pronunciations)

[06:11:04.0000] <MikeSmith>
Lachy: same meaning

[06:11:05.0000] <jgraham>
MikeSmith: I think I tend to disagree

[06:11:06.0000] <MikeSmith>
that's why

[06:11:07.0000] <Lachy>
same meaning should have same sound

[06:11:08.0000] <MikeSmith>
Lachy: no

[06:12:00.0000] <annevk>
/me disagrees too, though not on the sound/reading basis but because 9000 symbols is way harder to grasp than 26 letters

[06:12:01.0000] <Lachy>
Why? It makes communicating aurally harder

[06:12:02.0000] <MikeSmith>
jgraham: you would disagree less if you were a Japanese speaker not fluent in Chinese and found yourself trying to read signs and menus and such in China

[06:12:03.0000] <jgraham>
The problem with using one symbol per concept is that the number of concepts is much larger than the reasonable number of symbols. Therefore you have to start combining symbols and soon enough you have an alphabet

[06:13:00.0000] <MikeSmith>
jgraham: no, not really

[06:13:01.0000] <MikeSmith>
the majority of written communication in Japanese is done in pure symbols

[06:13:02.0000] <Lachy>
MikeSmith, I agree it's good to use the same symbol for the same meaning. I just think it's a bug that it's pronounced differently

[06:14:00.0000] <Lachy>
isn't all *written* communication done with symbols?

[06:14:01.0000] <MikeSmith>
Lachy: I encourage you to get Chinese people and Japanese to pronounce those symbols in the same way

[06:15:00.0000] <jgraham>
In what way? I would have thought that you could either have a simple alphabet which makes learning easy but allows more combinations of symbol or a more complex alphabet which makes learning hard but leads to fewer different combinations to represent the same concept

[06:15:01.0000] <MikeSmith>
Lachy: yeaha, all written communication is done in symbols, but some symbols are better than others

[06:15:02.0000] <Lachy>
I would. But I'm having enough trouble trying to get Norwegian's to stop mispronouncing their own letters. :-)

[06:15:03.0000] <jgraham>
I don't think optimising for Japanese speakers reading Chinese menus is the right optimisation

[06:15:04.0000] <MikeSmith>
as far as alphabets, Korean Hangul is absolutely a better alphabet than anything else

[06:16:00.0000] <annevk>
/me finds http://www.logoi.com/notes/pictograph.html

[06:16:01.0000] <MikeSmith>
and roman alphabet is relatively very ppor

[06:16:02.0000] <myakura>
/me wonders he can read and understand chinese menus if it's simplified

[06:16:03.0000] <Lachy>
I find it annoying the, for example, the letter 'a' is prounced differently in norwegian

[06:17:00.0000] <annevk>
the letter 'a' is pronounced differently all over the place

[06:17:01.0000] <Lachy>
I know. It's annoying

[06:17:02.0000] <jgraham>
MikeSmith: I certainly don't claim that the Roman alphabet is perfect (I was also using the word alphabet to include things like Kanji that perhaps are strictly not alphabets)

[06:17:03.0000] <annevk>
Lachy, it's annoying not everyone speaks the same language, yes...

[06:17:04.0000] <MikeSmith>
Lachy: Norwegian pronunciation is a mess.. you have some people who are fairly understandable, then you have those like Arve who do that sing-song thing 

[06:17:05.0000] <Lachy>
but 'i' and 'y' are even more annoying here. They're pronounced exactly the same in norwegian, even though the norewegians try to claim otherwise

[06:18:00.0000] <jgraham>
Lachy: In what context? 

[06:18:01.0000] <MikeSmith>
btw, myakura knows infinitely more about this than me

[06:18:02.0000] <jgraham>
I mean English isn't phonetic so it's not like we have a way of pronouncing "a"

[06:18:03.0000] <jgraham>
I think that phonetic writing systems make more sense than pictoral ones

[06:19:00.0000] <jgraham>
(although some Euopean languages manage the phonetic thing much better)

[06:20:00.0000] <MikeSmith>
jgraham: my 10 year old daughter seems to have no trouble learning two different Japanese phonetic alphabets while at the same time having learned to write and read hundreds of kanji and now learning the roman alphabet

[06:21:00.0000] <Lachy>
MikeSmith, that's because children still have the ability to learn new languages, which adults generally lose

[06:21:01.0000] <MikeSmith>
ture

[06:22:00.0000] <MikeSmith>
true

[06:22:01.0000] <myakura>
yeah

[06:22:02.0000] <jgraham>
MikeSmith: Picking a 10 year old isn't really fair because below the age of ~10 humans are built for learning language. Above that age we seem to be much worse at it (compare foreign language ability in the UK where people start learning at 11 to countries where they start younger)

[06:22:03.0000] <jgraham>
Hmm Lachy said that but shorter and faster

[06:23:00.0000] <Lachy>
plus, children's hearing is generally better so they can pick up the slight nuances in the sounds easier than adults

[06:24:00.0000] <MikeSmith>
jgraham: yeah, I will concede that. problem is that, as you point out, educational system don't start to teach languages to children until they are past the point of optimal language learning

[06:24:01.0000] <Lachy>
so while a kid may not have difficulty differentiating between 'i' and 'y' in norweign, I certainly can't.

[06:25:00.0000] <MikeSmith>
Lachy: in Japanese, the troublesome sounds are "r" and "l"

[06:25:01.0000] <myakura>
and 'b' and 'v'

[06:25:02.0000] <MikeSmith>
ah yeah

[06:25:03.0000] <MikeSmith>
video vs. bideo

[06:26:00.0000] <myakura>
i don't think i can easily differentiate them

[06:26:01.0000] <MikeSmith>
myakura: I think you can :)

[06:28:00.0000] <MikeSmith>
myakura: in terms of pronunciation and other things, it's far easier for me to communicate with you face to face than is it for me to communicate with many "native" English speakers

[06:28:01.0000] <MikeSmith>
e.g., I was in Dublin a couple weeks ago...

[06:28:02.0000] <Lachy>
MikeSmith, what's your native language?

[06:28:03.0000] <myakura>
ah

[06:29:00.0000] <MikeSmith>
And in several instances in Dublin, I found myself, in the "nod my head and pretend I understand" situation

[06:30:00.0000] <MikeSmith>
Lachy: I was born in Oklahoma and lived for the first 3 years of my life in Japan

[06:31:00.0000] <takkaria>
hehe, Dublin

[06:31:01.0000] <MikeSmith>
... but my native language is essentially English, I guess

[06:32:00.0000] <MikeSmith>
with large amounts of mexican/central-american spanish thrown in

[06:32:01.0000] <myakura>
MikeSmith: i don't think i can easily communicate with people with brogue either

[06:32:02.0000] <MikeSmith>
and french-Canadian

[06:32:03.0000] <MikeSmith>
myakura: you and me both :)

[06:33:00.0000] <MikeSmith>
e.g., a guy in Dublin was talking about "cooking a stick"

[06:33:01.0000] <MikeSmith>
... and I nod my head

[06:33:02.0000] <MikeSmith>
... yeah, cooking a stick (what the hell is this guy talking about)

[06:34:00.0000] <MikeSmith>
... but apparently the word "steak" is pronounced much closer to "stick" in Dublin

[06:35:00.0000] <Lachy>
MikeSmith, did the guy have an irish accent?

[06:35:01.0000] <MikeSmith>
Lachy: heh, yeah

[06:35:02.0000] <MikeSmith>
... that would be an understatement

[06:35:03.0000] <MikeSmith>
He was Irish

[06:35:04.0000] <Lachy>
where I come from, steak is pronounced the same as stake

[06:36:00.0000] <MikeSmith>
I personally reckon the Irish to be among the great peoples of the earth

[06:36:01.0000] <Philip`>
Lachy: That causes terrible problems during vampire attacks

[06:37:00.0000] <MikeSmith>
... in terms of their collective sense of humor and love for life and poetry and music

[06:37:01.0000] <MikeSmith>
... and just about anything else you could imagine

[06:37:02.0000] <takkaria>
and catholicism

[06:37:03.0000] <MikeSmith>
takkaria: yeah

[06:37:04.0000] <MikeSmith>
... and except speaking English

[06:38:00.0000] <MikeSmith>
... which is is no way their natural native language

[06:39:00.0000] <MikeSmith>
Lachy: steak is not stake in your language

[06:39:01.0000] <MikeSmith>
... you guys love to stretch out vowels to ridiculous extremes

[06:40:00.0000] <Lachy>
MikeSmith, what?

[06:40:01.0000] <MikeSmith>
more like "stai-aike"

[06:40:02.0000] <Lachy>
MikeSmith, no. Aussies don't do that

[06:41:00.0000] <MikeSmith>
oK

[06:41:01.0000] <Lachy>
maybe some Americans who have bastardised our language do that

[06:42:00.0000] <MikeSmith>
well, there are a treasure trove of other examples of OZ embrace-and-extend of pronunciation that I can't think of right now

[06:42:01.0000] <MikeSmith>
... not that I'm saying it's bad

[06:42:02.0000] <MikeSmith>
I like the sound of it,

[06:42:03.0000] <MikeSmith>
the cadence

[06:43:00.0000] <MikeSmith>
e.g. the way that chaals talks

[06:43:01.0000] <Lachy>
MikeSmith, we're proud of our Aussie slang!

[06:44:00.0000] <MikeSmith>
the way that Arve talks both in English and Norski, I like the sound

[06:44:01.0000] <Lachy>
chaals doesn't speak like a typcial aussie

[06:44:02.0000] <MikeSmith>
Lachy: oK, well, here's another example:

[06:44:03.0000] <Lachy>
he speaks like a victorian

[06:44:04.0000] <MikeSmith>
David Storey

[06:45:00.0000] <Lachy>
(Victoria is a state in Aus, where I think Chaals is from)

[06:45:01.0000] <MikeSmith>
I challenge you to translate wtf David is talking about when he slips into his Geordi-speak

[06:46:00.0000] <Lachy>
what is Geordi-speak?

[06:46:01.0000] <MikeSmith>
(which he does pretty often)

[06:46:02.0000] <MikeSmith>
Lachy: Newcastle, Northumberland

[06:46:03.0000] <MikeSmith>
the north-north

[06:47:00.0000] <MikeSmith>
of that tiny little counry

[06:47:01.0000] <MikeSmith>
country

[06:47:02.0000] <MikeSmith>
which is a very different world than the south

[06:48:00.0000] <takkaria>
my sister lived by newcastle for a few years, it's not too bad

[06:48:01.0000] <MikeSmith>
people in the North in england are quite a bit different than those you find  in London and environs

[06:49:00.0000] <MikeSmith>
I like that Newcastle accent... just that I can't figure out what the hell they are saying most of the time

[06:49:01.0000] <MikeSmith>
... seems to involve dropping most vowels

[06:49:02.0000] <MikeSmith>
and saying things under your breath

[06:50:00.0000] <jgraham>
glottal stops!

[06:50:01.0000] <jgraham>
/me goes to eat lunch

[06:50:02.0000] <MikeSmith>
jgraham: yep

[06:50:03.0000] <takkaria>
in the north we like our glottal stops. :)

[06:50:04.0000] <takkaria>
/me is from manchester

[06:50:05.0000] <MikeSmith>
takkaria: aha

[06:51:00.0000] <MikeSmith>
/me estimation of takkaria goes up a few notches

[06:51:01.0000] <takkaria>
heh

[06:52:00.0000] <takkaria>
though I'm from south manchester, not north, so my accent is reasonably understandable

[06:52:01.0000] <MikeSmith>
manchester, for one thing, has produced many innovative/exceptional musicians/songwriters/drunkards over the last 15 years or more

[06:55:00.0000] <MikeSmith>
and as far as accents, "the sound of Ian Brown", for good or bad

[06:55:01.0000] <MikeSmith>
and many others

[06:56:00.0000] <takkaria>
manchester has a lot to recommend it

[06:56:01.0000] <takkaria>
though also far too many unsigned musicians

[06:57:00.0000] <MikeSmith>
takkaria: you could say the same about many places... Berlin, for one

[06:57:01.0000] <MikeSmith>
... or Birmingham even

[06:57:02.0000] <MikeSmith>
(me is a huge fan of Broadcast, from Birmingham)

[06:58:00.0000] <takkaria>
I say it about manchester because I'm one of them and it makes it very hard to get noticed by anyone at all. :)

[06:58:01.0000] <MikeSmith>
takkaria: you got some high expectations to compete with there ..

[07:00:00.0000] <takkaria>
it's true

[07:00:01.0000] <MikeSmith>
e.g., Johnny Marr.. what can be said there/

[07:00:02.0000] <MikeSmith>
?

[07:01:00.0000] <MikeSmith>
even Morrissey

[07:01:01.0000] <MikeSmith>
nutjob

[07:01:02.0000] <MikeSmith>
but brilliant

[07:01:03.0000] <MikeSmith>
and never quits

[07:02:00.0000] <takkaria>
on a related note, Andy Rourke recommended us to a promoter once and we got a good gig as a result

[07:02:01.0000] <MikeSmith>
I'm intrigued

[07:02:02.0000] <MikeSmith>
who is "us" ?

[07:03:00.0000] <takkaria>
my band, System Fault

[07:03:01.0000] <MikeSmith>
/me notes that he worked in record stores during most of the time he was in university

[07:03:02.0000] <MikeSmith>
takkaria: first impression: extremely dumb name

[07:04:00.0000] <MikeSmith>
change that

[07:04:01.0000] <takkaria>
yeah, it's bad, isn't it?

[07:04:02.0000] <MikeSmith>
yeah, really

[07:04:03.0000] <MikeSmith>
unequivocally

[07:04:04.0000] <MikeSmith>
hmm,

[07:05:00.0000] <takkaria>
it was chosen in, er, earlier days

[07:05:01.0000] <MikeSmith>
so which of thsese four ruffians is you?

[07:05:02.0000] <takkaria>
I've been lobbying for a name change for quite a while but other members aren't having it

[07:05:03.0000] <MikeSmith>
aha

[07:05:04.0000] <MikeSmith>
#4

[07:06:00.0000] <MikeSmith>
Guitar, Synth & Backing Vox

[07:06:01.0000] <takkaria>
yeah, that's the one

[07:06:02.0000] <MikeSmith>
takkaria: do more than lobby

[07:06:03.0000] <MikeSmith>
that name is millstone

[07:08:00.0000] <MikeSmith>
and "synth"... well, 

[07:08:01.0000] <gsnedders>
MikeSmith: Want any more specific answers about uni now I return?

[07:09:00.0000] <MikeSmith>
takkaria: "keyboard wizardy" .. "fingers" 

[07:09:01.0000] <MikeSmith>
please anything but "synth"

[07:10:00.0000] <takkaria>
MikeSmith: synth is where it's at here in the UK

[07:11:00.0000] <MikeSmith>
gsnedders: yeah, if you got the time,. the whole point of this is to try to document for "casual readers" what changes have been made, at a high lelve

[07:11:01.0000] <MikeSmith>
level

[07:11:02.0000] <MikeSmith>
takkaria: that will pass

[07:12:00.0000] <MikeSmith>
try breaking some new ground in terms of the way you describe your skills

[07:12:01.0000] <MikeSmith>
... that almost always pays off

[07:12:02.0000] <takkaria>
really?

[07:13:00.0000] <MikeSmith>
takkaria: this seems obvious to me.. but I guess I may come from a different plane

[07:13:01.0000] <MikeSmith>
takkaria: just as a data point

[07:13:02.0000] <gsnedders>
MikeSmith: Huh?

[07:13:03.0000] <takkaria>
MikeSmith: well, your advice is valued, as someone who really isn't UK-centric at all

[07:13:04.0000] <MikeSmith>
... do you now what the "Amen break" means?

[07:14:00.0000] <takkaria>
of course

[07:14:01.0000] <MikeSmith>
takkaria: OK, good

[07:14:02.0000] <MikeSmith>
gsnedders: sorry, misunderstood

[07:15:00.0000] <MikeSmith>
gsnedders: I think you should weigh Hixie advice against other considerations

[07:16:00.0000] <MikeSmith>
e.g., focused coursework in CS

[07:16:01.0000] <MikeSmith>
as opposed to physics

[07:16:02.0000] <MikeSmith>
(I say this as someone who minored in physics myself at university)

[07:17:00.0000] <MikeSmith>
... and physical sciences in genera;

[07:17:01.0000] <MikeSmith>
general

[07:18:00.0000] <MikeSmith>
in particular, there is a good argument for spending a lot of time at university learning the math fundamentals underlying CS

[07:18:01.0000] <jgraham>
/me doesn't understand what MikeSmith means about focussed coursework

[07:18:02.0000] <gsnedders>
MikeSmith: Most of the advice I've got agrees with Hixie though :P

[07:19:00.0000] <MikeSmith>
learning while you have the luxury of time

[07:19:01.0000] <Lachy>
gsnedders, advice about what?

[07:19:02.0000] <gsnedders>
MikeSmith: In the case of the comp.phys. at Edi, in first year you do the two introductory CS courses

[07:19:03.0000] <gsnedders>
Lachy: uni

[07:21:00.0000] <Lachy>
gsnedders, so how did you go in your physics exam? Was that the one you had yesterday?

[07:21:01.0000] <MikeSmith>
jgraham: algorithms

[07:21:02.0000] <MikeSmith>
(in so many words)

[07:21:03.0000] <gsnedders>
Lachy: Yeah. I thought the paper was hard, but so did everyone. I think I did quite well in it though.

[07:21:04.0000] <MikeSmith>
basics

[07:21:05.0000] <MikeSmith>
structures

[07:21:06.0000] <gsnedders>
(There again, "quite well" in my world means a low A)

[07:21:07.0000] <MikeSmith>
math

[07:22:00.0000] <Lachy>
physics is easy. It's just remembering a bunch of formulas and knowing a few constants

[07:22:01.0000] <jgraham>
I understand that learning CS basics is more likely to happen on a CS course but I'm still not clear on what "focused coursework" means

[07:22:02.0000] <Philip`>
gsnedders: Do you know how the maths content varies between all the courses?

[07:22:03.0000] <Lachy>
at least, it was for me in high school

[07:22:04.0000] <jgraham>
Lachy: I like to think that it's actually harder than that :)

[07:23:00.0000] <MikeSmith>
Lachy: physics is not easy

[07:23:01.0000] <MikeSmith>
at a certain level it is

[07:23:02.0000] <MikeSmith>
like many things

[07:23:03.0000] <Lachy>
MikeSmith, I know.

[07:23:04.0000] <gsnedders>
Philip`: In the case of Edi. and pure maths, the maths in comp.phys. is identical to that of phys. (but I don't know what that is)

[07:24:00.0000] <MikeSmith>
beyond basics, physics and math grow into the realm of the non-intuitive 

[07:24:01.0000] <MikeSmith>
that's the problem for most people

[07:25:00.0000] <jgraham>
gsnedders: In general physics courses will have you doing "continuous" maths (e.g. vector calculus, differential equations)

[07:25:01.0000] <jgraham>
whereas CS courses have you doing discrete maths

[07:25:02.0000] <jgraham>
which I am less sure about

[07:25:03.0000] <jgraham>
so I'll let someone else describe it in case I am wrong

[07:26:00.0000] <MikeSmith>
jgraham: I guess I mean investing time over 3-4 years in learning the math and fundamentals behind modern CS while at university

[07:26:01.0000] <MikeSmith>
... instead of having to take/make time to learn it later

[07:26:02.0000] <jgraham>
MikeSmith: I guess that assumes that your end goal is to know CS

[07:26:03.0000] <Philip`>
There's quite a bit of partly-continuous maths like probability and Fourier transform stuff in CS here

[07:27:00.0000] <jgraham>
In which case taking CS is a good idea :)

[07:27:01.0000] <MikeSmith>
jgraham: that is a worthy end goal, as far as I'm concerned

[07:27:02.0000] <jgraham>
MikeSmith: As someone who knows less CS than he would like, I agree, but I would also say the same about Physics

[07:27:03.0000] <MikeSmith>
this gets back in part to the idea of having a common framework for discussion

[07:28:00.0000] <jgraham>
Or indeed Mathematics itself

[07:28:01.0000] <Philip`>
(AI in particular had lots of integral equations and stuff)

[07:28:02.0000] <MikeSmith>
jgraham: well, we here are not here because of common interest in Physics

[07:29:00.0000] <jgraham>
MikeSmith: Sure. But gsnedders might be here because he is interested in web stuff but still be more interested in Physics

[07:29:01.0000] <jgraham>
Therefore CS would be the wrong choice for him

[07:29:02.0000] <gsnedders>
/me would have more enthusiasm to be doing something like comp.phys. than CS

[07:30:00.0000] <jgraham>
/me thinks that the combination of computers and differential equations is the root of all evil

[07:30:01.0000] <gsnedders>
And from a lot of what I've seen CS is badly taught at a heckuva lot of places, which doesn't add to the enthusiasm

[07:30:02.0000] <Philip`>
/me guesses jgraham wouldn't want to do electronic engineering :-)

[07:30:03.0000] <Lachy>
what does CS stand for?

[07:30:04.0000] <gsnedders>
Lachy: Computer Science

[07:30:05.0000] <takkaria>
depending on your uni, switching courses can be more or less easy, but if you find you're on the wrong one, then changing shouldn't be too bad

[07:31:00.0000] <Lachy>
ok. We call that IT where I come from

[07:31:01.0000] <jgraham>
FWIW I still maintain that, were he to get in the Cambridge CS course is ideal for gsnedders because he would do both Physics and CS in the first year

[07:31:02.0000] <MikeSmith>
part of the context for what I say is that here in Japan, I sometimes (often) meet "engineers" who have gone through university course in CS and seem to me to be lacking in understanding of some fundamentals that any graduate with a CS degree from a major university in europe or n. america would be assumed to have

[07:31:03.0000] <gsnedders>
And I could always change to doing Physics in second year, too

[07:31:04.0000] <jgraham>
gsnedders: Indeed. Hence ideal

[07:32:00.0000] <gsnedders>
/me has a Cam prospectus next to him

[07:32:01.0000] <jgraham>
Also, I would worry that a computational physics course wouldn't teach much in the way of actual CS

[07:32:02.0000] <jgraham>
Though I could be wrong

[07:33:00.0000] <gsnedders>
It's the only one of the three that has arrived. There again, I bet the other two unis haven't even done their '09 prospectuses yet :P

[07:33:01.0000] <takkaria>
gsnedders: which unis are you looking into?

[07:33:02.0000] <gsnedders>
http://www.ed.ac.uk/studying/undergraduate/finder/degree.php?id=0,4,F355 — that has a shortish description of it

[07:33:03.0000] <gsnedders>
takkaria: Cam, Edi, York

[07:33:04.0000] <MikeSmith>
jgraham: based just on anecdotal datea, I guess i should also note that some of the cleverest/most-effective developers I personally know are people who did not major in CS at university

[07:34:00.0000] <Philip`>
Lachy: The impression I have is that IT here is about things like "how do I make an Access database to store my business's client list", whereas CS is about things like "how do I model a database using relational algebra, and how do I implement it efficiently taking account of disk geometry" :-)

[07:34:01.0000] <takkaria>
gsnedders: I would imagine they all have their 09 prospectuses sorted, I'm pretty sure Manchester has

[07:35:00.0000] <gsnedders>
takkaria: Then Edi and York are just really slow at sending it :)

[07:35:01.0000] <Lachy>
Philip`, I covered all of that stuff in my Bachelor of IT course that I did at uni.

[07:35:02.0000] <gsnedders>
It's hardly as if it has a long way to come from either, especially Edi.

[07:36:00.0000] <jgraham>
Lachy: Which uni?

[07:36:01.0000] <gsnedders>
http://www.ph.ed.ac.uk/undergraduate/degrees/structure/ — that has far more information about the course content for the comp.phys.

[07:37:00.0000] <takkaria>
gsnedders: they'll be receiving hundreds and hundreds of requests for prospectuses though, so some delay is inevitable

[07:37:01.0000] <gsnedders>
takkaria: And I guess Cam's propaganda department is good :)

[07:38:00.0000] <Lachy>
jgraham, Charles Sturt University in Bathurst, NSW, Australia

[07:38:01.0000] <Lachy>
csu.edu.au

[07:39:00.0000] <Philip`>
takkaria: I don't see why it'd take long to print only hundreds of copies of something, and the demand should be entirely predictable based on previous years

[07:41:00.0000] <jgraham>
gsnedders: That course seems to be mostly physics with a bit of computing. I would try and find the detailed syllabus to check if the "computing" bits are solving differential equations using fortran 77

[07:42:00.0000] <takkaria>
Philip`: I would guess sending out prospectii to colleges and the like would take precedence, and perhaps "hundreds" was a little on the low side

[07:42:01.0000] <takkaria>
Philip`: but you have a fair point

[07:43:00.0000] <jgraham>
gsnedders: have you seen http://www.inf.ed.ac.uk/teaching/degrees/cs_ph.html?

[07:43:01.0000] <gsnedders>
takkaria: I know that "we" (my school) only get the Scottish ones, FWIW

[07:43:02.0000] <gsnedders>
jgraham: yes

[07:44:00.0000] <Philip`>
jgraham: It points to http://www.ph.ed.ac.uk/undergraduate/degrees/tasters/compsim2.html which menions Java which is probably a bit more sensible than Fortran :-)

[07:44:01.0000] <Philip`>
s//t/

[07:44:02.0000] <gsnedders>
jgraham: Yeah, it's certainly Java that's used

[07:45:00.0000] <Philip`>
Clearly it would be better to use ML

[07:47:00.0000] <gsnedders>
Philip`: Haskell _may_ come in to it, I can't remember which of the CS courses you do

[07:48:00.0000] <takkaria>
common lisp > ML

[07:48:01.0000] <gsnedders>
I'm sure I found something better about the course before

[07:49:00.0000] <gsnedders>
Yeah, Haskell comes into it

[07:49:01.0000] <gsnedders>
<http://www.inf.ed.ac.uk/teaching/courses/inf1/1Aguide.html>

[07:51:00.0000] <gsnedders>
<http://www.drps.ed.ac.uk/06-07/dpts/SCE_FINAL/147.html> and <http://www.ph.ed.ac.uk/teaching/progspecs/ProgSpecCP.pdf>

[07:52:00.0000] <Philip`>
"Applicable Mathematics" - is that implying the rest of mathematics is inapplicable?

[07:52:01.0000] <gsnedders>
Philip`: My thoughts exactly :)

[07:53:00.0000] <gsnedders>
Oh, Fortran does come into it

[07:53:01.0000] <Philip`>
"Parallel Fortran in Physics"

[07:53:02.0000] <gsnedders>
But that was inevitable.

[07:54:00.0000] <gsnedders>
MikeSmith: How much of CS is just maths without doing much application of it? Or does that vary massively?

[07:57:00.0000] <gsnedders>
Philip`: ^^

[07:59:00.0000] <gsnedders>
http://www.drps.ed.ac.uk/08-09/dpts/SCE_FINAL/147.html — that's '08–'09, which is more up to date, and is the latest available

[08:02:00.0000] <gsnedders>
It seems for the comp.phys. you can do the norm CS courses too as your outside courses

[08:10:00.0000] <Philip`>
gsnedders: I can't think of much maths I did that wasn't leading to some kind of application - I suppose probability was kind of left on its own, but other maths lead to databases and cryptography and digital signal processing and automated theorem proving and proving program correctness and whatever, and so pretty much all the maths was done with that kind of final goal in mind and was rarely taught as a pure maths course

[08:12:00.0000] <Philip`>
though I could be misremembering, particularly since I gave up on some of the maths ones since they were too hard and I was too lazy :-)

[08:15:00.0000] <Philip`>
The Denotation Semantics course taught me that "⊤" was called "top" and "⊥" was "bottom" but I'm not sure I managed to follow much else of it :-/

[08:15:01.0000] <Philip`>
*Denotational

[08:17:00.0000] <takkaria>
grr, I want my IRC client to do UTF-8

[08:17:01.0000] <Lachy>
takkaria, which IRC client are you using?

[08:18:00.0000] <takkaria>
irssi via SSH

[08:18:01.0000] <Philip`>
irssi via (screen via) SSH does UTF-8 fine for me, though its font fails to draw ⊥/⊤ except as boxes :-(

[08:19:00.0000] <takkaria>
hmm, I'm just getting ? instead of boxes

[08:21:00.0000] <Philip`>
Maybe your font doesn't have the box glyph either

[08:22:00.0000] <takkaria>
I seem to be using a unicode font OK

[08:22:01.0000] <takkaria>
oh, I'm not

[08:22:02.0000] <takkaria>
that explains a lot

[08:53:00.0000] <Lachy>
hey, does anyone have any ideas for some images I could use to represent each of the design principles in the slides?

[08:54:00.0000] <Lachy>
it can be a bit abstract if there's nothing that directly conveys the concept

[09:06:00.0000] <Dashiva>
Priority of constituencies could be a pyramid

[09:09:00.0000] <Lachy>
Dashiva, interesting idea

[09:11:00.0000] <Lachy>
FYI, whatwg blog article was just republished here http://css.dzone.com/news/html-5-reverse-ordered-lists - I hope it gets some intersting comments

[09:13:00.0000] <Dashiva>
I'm tempted to suggest an crossed out ivory tower, or in one of those "do not" traffic signs, but I can't decide which one it should be :)

[09:14:00.0000] <Lachy>
for which principle?

[09:15:00.0000] <Lachy>
I have this, or something similar, for pave the cowpaths http://www.flickr.com/photos/adactio/298913307/

[09:19:00.0000] <takkaria>
hm, looks like html5 doesn't describe how to handle whitespace in <a href="">

[09:22:00.0000] <Lachy>
takkaria, does it need to define any special handling?

[09:23:00.0000] <Lachy>
do you mean within the element's text content, or its attributes, or something else?

[09:23:01.0000] <takkaria>
within the href="" attribute in particular

[09:24:00.0000] <Lachy>
like handling spaces as %20?

[09:24:01.0000] <takkaria>
yeah

[09:24:02.0000] <takkaria>
and trimming leading spaces

[09:24:03.0000] <takkaria>
Fx3 seems to ignore all but one trailing space too

[09:25:00.0000] <takkaria>
(in the DOM)

[09:25:01.0000] <takkaria>
but then following that link strips all trailing and leading spaces

[09:28:00.0000] <takkaria>
I'd imagine it's required for interopability too

[09:29:00.0000] <gsnedders>
Yeah, it certainly is.

[09:30:00.0000] <gsnedders>
WebKit's URL class always strips leading/trailing spaces, no matter where it comes from

[09:32:00.0000] <Dashiva>
Lachy: Support Existing Content, Solve Real Problems, or Handle Errors

[09:33:00.0000] <Dashiva>
Maybe all three? :)

[09:34:00.0000] <Lachy>
for support existing content, I think I'm going with a bridge.

[09:35:00.0000] <Lachy>
Either this one http://www.flickr.com/photos/hb2/288721287/ or this one http://www.flickr.com/photos/andyinnyc/421695304/

[09:37:00.0000] <takkaria>
could someone with Safari around check to see what the href="" attribute looks like in the DOM at http://tinyurl.com/5wx9wa?

[09:40:00.0000] <Lachy>
takkaria, href=" http://google.com/search?q=hello world "

[09:41:00.0000] <takkaria>
thanks

[10:23:00.0000] <takkaria>
it's interesting how the number of private mails Hixie has is roughly the same as it was last October

[10:23:01.0000] <takkaria>
(for html5 issues, anyway)

[11:41:00.0000] <Lachy>
damn. Searching flickr for photos to use in my presentation that use creative commons licences, but don't use share-alike (cause copyleft licences suck), is difficult.

[11:47:00.0000] <jgraham>
Lachy: Copyleft would just mean that we had to make the slides CC licensed too, right? that seems acceptable

[11:47:01.0000] <jgraham>
(Well CC SA specifically)

[11:48:00.0000] <Lachy>
jgraham, I'd prefer to make it public domain and use CC by or by-nc licenced photos

[11:51:00.0000] <jgraham>
I don't mind either way but if not making it SA makes it worse it seems worth making it SA. I'm struggling to think of any situation in which not-SA would be a significant advantage

[11:52:00.0000] <jgraham>
Like when would someone try to distribute a derived-work under a more restrictive license?

[11:58:00.0000] <Lachy>
maybe I could get away with using share-alike licenced photos, since it's possible to mix public domain with anything, and using an SA licence doesn't on some parts doesn't change that.

[11:58:01.0000] <Lachy>
it would just have to be clear that photos use their own licences anyway, none of which are more restrictive than SA

[11:59:00.0000] <Lachy>
but I'd prefer to avoid that whole issue since IANAL and copyright licences are confusing 

[12:00:00.0000] <Philip`>
It just means you can license the composite work only as CC, which is alright if it's composed of non-CC things that still allow to license them as CC

[12:00:01.0000] <Philip`>
(The only problem is when you're using something that requires the composite to be licensed as CC, and something else that requires it to be licensed as something else different)

[12:02:00.0000] <Philip`>
(*different and incompatible)

[12:02:01.0000] <Lachy>
I want everything but the photos to be free from copyright, so if someone took out the photos and replaced them with their own, there would be no restrictions

[12:03:00.0000] <Philip`>
You could just not care about copyright at all, and use whatever pictures look good, because it's not like anyone is going to get arrested for presentation slide piracy :-)

[12:04:00.0000] <Lachy>
since the slides will be published on the web afterwards, I'd rather avoid any possible complications

[12:05:00.0000] <Lachy>
any suggestions for what to use for Degrade Gracefully?

[12:05:01.0000] <jgraham>
Lachy: IANAL but IIRC it's pretty hard to public domain your work. Like harder than saying "this is public domain". So you're better off picking a suitable license that grants people the permissions that you want to give them.

[12:06:00.0000] <Lachy>
jgraham, it's not hard to say something is in the public domain

[12:07:00.0000] <jgraham>
Lachy: The problem is, I think, that in certain jurisdictions just saying that doesn't make it so

[12:09:00.0000] <jgraham>
Lachy: see http://en.wikipedia.org/wiki/Wikipedia:Granting_work_into_the_public_domain for example

[12:09:01.0000] <Lachy>
so you can just say "This is in the Public Domain. You are free to use this for any purpose, without any restrictions"

[12:10:00.0000] <jgraham>
The second status there has some weight yes. Although you should probably be more clear

[12:10:01.0000] <Lachy>
http://creativecommons.org/licenses/publicdomain/

[12:11:00.0000] <Lachy>
but I don't care that much. If someone doesn't understand what "Public Domain" means, or lives in a stupid juristiction, that's not my problem. They need to fix their country's laws.

[12:12:00.0000] <Dashiva>
What matters is that Lachy isn't going to prosecute anyone :)

[12:13:00.0000] <jgraham>
Dashiva: It also matters if anyone can prosecute Lachy

[12:14:00.0000] <jgraham>
Lachy: Note that it says on the page before "Please note that the Public Domain Dedication may not be valid outside of the United States"

[12:14:01.0000] <jgraham>
(in the UK Public Domain doesn't mean anything, apparently)

[12:15:00.0000] <jgraham>
So it seems like the safest thing to do is to take the part of the PD license that reads "the Work may be freely reproduced, distributed, transmitted, used, modified, built upon, or otherwise exploited by anyone for any purpose, commercial or non-commercial, and in any way, including by methods that have not yet been invented or conceived."

[12:15:01.0000] <jgraham>
and just use that

[12:19:00.0000] <Lachy>
fine. I'll adjust my site-wide copyright licence to say something like that http://lachy.id.au/about/copyright

[13:35:00.0000] <annevk>
takkaria, did you test Opera 9.5??

[14:42:00.0000] <gsnedders>
/me moves SP inline with HTML 5's stuff on character encoding names

[14:51:00.0000] <Lachy>
I have images for most design principles now. I just need Media Indpendence and Well Defined Behaviour

[14:51:01.0000] <Lachy>
any suggestions?

[14:52:00.0000] <gsnedders>
Lachy: I want to say some sort of crash for Well Defined Behaviour

[14:52:01.0000] <hdh>
is the cowpath suitable for well-defined?

[14:52:02.0000] <hdh>
I don't have the link here to check the licence

[14:52:03.0000] <Lachy>
hdh, I'm using a cowpath for Pave the Cowpaths.

[14:52:04.0000] <Lachy>
gsnedders, I've got a train crash for Handle Errors

[14:53:00.0000] <Philip`>
That ought to be a dragon

[14:53:01.0000] <gsnedders>
Lachy: Someone being mad?

[14:53:02.0000] <Philip`>
(Guess it's hard finding photos of one, though)

[14:53:03.0000] <gsnedders>
(mad in a silly way)

[14:53:04.0000] <Lachy>
Philip`, a dragon for what and why?

[14:54:00.0000] <Lachy>
I thought of finding a photo of a sign with lots of rules on it, but couldn't find an appropriate one

[14:54:01.0000] <Philip`>
Lachy: For (non-)draconian error handling

[14:54:02.0000] <annevk>
Looking at the SVG minutes there hasn't been much progress on SVG in HTML


2008-05-25
[01:08:00.0000] <Hixie>
areas i could focus on next if no-one has any specific requests: ruby, abbr, alt, links, focus, usemap, sections

[01:43:00.0000] <annevk>
Hixie, did you fix the <p> <span> <ol> thingie?

[01:43:01.0000] <annevk>
I guess nobody e-mailed about that

[01:45:00.0000] <MikeSmith>
Hixie: what would need to be done for sections? 

[01:52:00.0000] <Hixie>
dunno, haven't looked in that folder :-)

[01:52:01.0000] <Hixie>
annevk: what issue?

[01:53:00.0000] <annevk>
Hixie, nesting <ol> inside <p> through means of a <span>

[01:53:01.0000] <Hixie>
it's already in the spec

[01:53:02.0000] <Hixie>
it's commented out

[01:53:03.0000] <Hixie>
search for XXXSPAN

[01:54:00.0000] <annevk>
the spec has "<p class="big-issue">This doesn't match browsers.</p>"

[01:54:01.0000] <annevk>
'<p><i><div><p>'

[01:54:02.0000] <annevk>
so maybe that's different

[01:56:00.0000] <annevk>
seems to be similar anyway

[01:56:01.0000] <annevk>
Hixie, in the parsing section there doesn't seem to be XXXSPAN

[01:57:00.0000] <Hixie>
oh i misunderstood which issue you meant

[01:57:01.0000] <Hixie>
yes

[01:57:02.0000] <Hixie>
i have no idea how to fix that particular issue

[01:57:03.0000] <Hixie>
and intend to not fix it unless someone can come up with a better solution :-)

[01:59:00.0000] <annevk>
you're hoping browsers will change?

[02:00:00.0000] <Hixie>
yes

[02:04:00.0000] <Hixie>
hm

[02:05:00.0000] <Hixie>
maybe it's just a matter of not closing ps unless the current element is a p

[02:10:00.0000] <MikeSmith>
Hixie: I vote for having a sane ruby spec

[02:24:00.0000] <Hixie>
no, it's more complicated than that, because i'd have to make <p> not look through non-formatting non-phrasing elements

[02:48:00.0000] <annevk>
ruby is fine with me too

[02:49:00.0000] <annevk>
i hope browsers will implement it at some point

[02:51:00.0000] <jgraham>
Is it allowed to ask you not to work on alt or abbr but to not mind otherwise? (because working on alt or abbr is a waste of everyone's time since there are people who will object whatever the spec says, so we will have to revisit the issue again anyway)

[02:52:00.0000] <annevk>
instead of asking that, you could just vote for the other options

[02:52:01.0000] <annevk>
:)

[02:52:02.0000] <jgraham>
(unless by "work on alt" you mean "organise some proper non-automated study into the way that @alt is used on the internet and how that varies with the other properties of the page" rather than "change the spec text about alt")

[02:52:03.0000] <Lachy>
jgraham, what's the problem with abbr? who would object?

[02:53:00.0000] <jgraham>
Lachy: There is a dormant prema-thread about abbr, no?

[02:53:01.0000] <jgraham>
s/prema/perma/

[02:55:00.0000] <Lachy>
oh, you mean that one where some people want to preserve the theoretical differences between abbr and acronym and add more attributes for fine grained control over them?

[02:56:00.0000] <Hixie>
jgraham: i have to reply to the feedback at some point

[02:56:01.0000] <Lachy>
the only good solution is to make abbr and acronym synonymous, since people use them interchangebly anyway.

[02:57:00.0000] <Lachy>
or just leave it as is with only abbr

[02:57:01.0000] <Hixie>
annevk: in fact, as far as i can tell, what the spec says is closer to what IE does than what the other browsers do

[02:57:02.0000] <Hixie>
it's also what firefox does

[02:58:00.0000] <jgraham>
Hixie: I know. But replying only has the effect of generating more feedback so if you prioritize those areas then you will never get anything else done

[02:58:01.0000] <annevk>
Hixie, my Firefox allows <p><span><ol>

[02:58:02.0000] <jgraham>
Lachy: Yeah that thread

[02:59:00.0000] <Hixie>
annevk: yes, it has a special case for <span>, but only <span>

[03:00:00.0000] <annevk>
Hixie, works for <i> too

[03:00:01.0000] <Hixie>
not for me

[03:00:02.0000] <annevk>
<!DOCTYPE html><p><i><p>x</p></i></p>x

[03:00:03.0000] <annevk>
is what I used

[03:00:04.0000] <annevk>
<!DOCTYPE html><p><i><ol><li>test</li></ol></i></p>x

[03:01:00.0000] <annevk>
works too

[03:01:01.0000] <annevk>
but it starts breaking down if you leave out some end tag

[03:01:02.0000] <annevk>
quite messy

[03:02:00.0000] <Hixie>
oh, you're including end tags

[03:02:01.0000] <Hixie>
well then sure

[03:02:02.0000] <Hixie>
firefox has a differet mode for that

[03:02:03.0000] <Hixie>
if all the tags match it just does exactly what the markup is most of hte time

[03:02:04.0000] <annevk>
interesting

[03:03:00.0000] <Hixie>
in any case what the spec does is still closer to what IE does

[03:03:01.0000] <Hixie>
even for <span>

[03:03:02.0000] <Hixie>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%3Cstyle%3Ep%20{%20border%3A%20solid%20fuchsia%3B%20padding%3A%200.5em%3B%20}%20div%20{%20border%3A%20solid%20teal%3B%20padding%3A%200.5em%3B%20}%3C%2Fstyle%3E%0A...%3Ctable%3E%3Ctr%3E%3Ctd%3E%3Cp%3E%3Ci%3E%3Cdiv%3E%3Cp%3E%3C%2Ftable%3E%0A...%3Ctable%3E%3Ctr%3E%3Ctd%3E%3Cp%3E%3Cspan%3E%3Cdiv%3E%3Cp%3E%3C%2Ftable%3E%0A

[03:05:00.0000] <annevk>
ok

[03:08:00.0000] <Hixie>
in fact for this:

[03:08:01.0000] <Hixie>
<!DOCTYPE html>...<p><span><p>...</p></span></p>...

[03:08:02.0000] <Hixie>
the DOM that HTML5 gives is as close to what IE does as you can get without the non-tree DOM thing

[03:08:03.0000] <Hixie>
and much cloer than what other browsers do

[03:09:00.0000] <Hixie>
in fact that particular case matches webkit pretty well

[03:10:00.0000] <Hixie>
and opera's rendering doesn't match its DOM

[03:10:01.0000] <Hixie>
/me goes to remove the issue marker

[03:12:00.0000] <annevk>
wow, video codec contains a lot of e-mails

[03:13:00.0000] <annevk>
Hixie, "Error loading folder contents: . Let Hixie know."

[03:14:00.0000] <annevk>
(stuff still works, so I'm not sure what's going on)

[03:15:00.0000] <Hixie>
weird

[03:15:01.0000] <Hixie>
let me know if it can be reproduced

[03:16:00.0000] <annevk>
after a refresh it's no longer there

[03:16:01.0000] <Hixie>
probably just memory issues on my server

[03:16:02.0000] <annevk>
btw, there's WF3-geo and geolocation

[03:16:03.0000] <Hixie>
yeah

[03:16:04.0000] <Hixie>
i'm not looking there until webapi decides what they're doing

[03:16:05.0000] <Hixie>
or until the forms tf completes

[03:17:00.0000] <annevk>
not really sure what to do there

[03:17:01.0000] <annevk>
Forms WG still seems fixed on the idea of bringing XForms to HTML

[03:18:00.0000] <Hixie>
well since that's not in the forms tf charter, that should be easy to deal with

[03:20:00.0000] <annevk>
the latest was http://lists.w3.org/Archives/Public/public-forms-tf/2008May/0003.html

[03:20:01.0000] <Hixie>
/me goes to look at his ruby data

[03:20:02.0000] <Hixie>
yeah i saw

[03:25:00.0000] <annevk>
I note that nobody replied to http://lists.w3.org/Archives/Public/public-forms-tf/2008Apr/0023.html

[03:27:00.0000] <annevk>
maybe it would be best to draft a new text document taking guidelines 1-7 from maciej and propose to adopt that

[03:27:01.0000] <othermaciej>
I might have time this weekend

[03:28:00.0000] <annevk>
k

[03:28:01.0000] <annevk>
it probably doesn't have to be HTML per se, you could just do it like the charter document

[03:30:00.0000] <Hixie>
you mean you actually want to do the work the tf chartered itself to do?

[03:30:01.0000] <Hixie>
that's a radical idea

[03:30:02.0000] <Hixie>
it'll never catch on

[03:32:00.0000] <annevk>
hmm, seems that my browser couldn't handle the sarcasm

[03:34:00.0000] <annevk>
Hixie, for your issues graph, it would be nicer if the horizontal line snapped to the nearest graph

[03:37:00.0000] <Hixie>
yeah

[03:38:00.0000] <Hixie>
so according to my data

[03:38:01.0000] <Hixie>
there are a lot of stray </rt> elements

[03:38:02.0000] <Hixie>
but not many stray </rp> elements

[03:39:00.0000] <Hixie>
despite there being roughly the same amount of both

[03:40:00.0000] <Hixie>
as in, 2.3% vs 0.017%

[03:40:01.0000] <Hixie>
that's a very odd result

[03:41:00.0000] <Hixie>
hmm, </ruby> has about the same problem

[03:41:01.0000] <Hixie>
as </rt>

[03:41:02.0000] <Hixie>
i clearly need to do this again but get sample urls

[04:36:00.0000] <Hixie>
oh that's hilarious

[04:36:01.0000] <Hixie>
http://blog.kansai.com/kasutera7966/132

[04:37:00.0000] <Hixie>
it has two occurances of <ruby> ... <rt> ... </ruby>, which is fine as far as that goes

[04:37:01.0000] <Hixie>
but then at the end of the main bit of text there are two orphan </rt> tags

[04:37:02.0000] <Hixie>
clearly some software somewhere was like "ok well i'm closing these damnit!"

[04:41:00.0000] <annevk>
that's some interesting markup right there :)

[04:43:00.0000] <Hixie>
so uh

[04:43:01.0000] <Hixie>
how do you get IE to actually render hte ruby

[04:45:00.0000] <annevk>
<!doctype html> <ruby> test <rt> test </rbuy>

[04:45:01.0000] <annevk>
wfm

[04:46:00.0000] <annevk>
though you probably want </ruby> there :)

[04:47:00.0000] <annevk>
/me uses IE6

[04:48:00.0000] <annevk>
I don't think IE does "advanced" ruby

[04:48:01.0000] <annevk>
the rbc and rtc elements

[04:49:00.0000] <annevk>
s/advanced/complex/

[04:49:01.0000] <Hixie>
http://jyosui.blogzine.jp/blog/cat7074652/index.html

[04:49:02.0000] <Hixie>
^ are you sure you want us to support ruby?

[04:50:00.0000] <Hixie>
woah

[04:50:01.0000] <Hixie>
it only works in quirks mode for me

[04:52:00.0000] <Hixie>
<ruby><rb>ハクナ・マタタ</rb><rt>くよくよするな</rp></ruby>

[04:53:00.0000] <Hixie>
good work matching tags there http://d.hatena.ne.jp/keyword/%A5%E9%A5%A4%A5%AA%A5%F3%A1%A6%A5%AD%A5%F3%A5%B0

[04:53:01.0000] <annevk>
afaict IE ignores <rp> and <rb>

[04:54:00.0000] <Hixie>
yes

[04:54:01.0000] <annevk>
and <rt> only functions if it has a non-whitespace previous sibling 

[04:55:00.0000] <Hixie>
ok so here is what i propose:

[04:55:01.0000] <Hixie>
<ruby> and </ruby> act like any random phrasing element tags

[04:56:00.0000] <Hixie>
<rp> and <rt> pop elements up to the nearest in-scope <ruby>, if any

[04:56:01.0000] <Hixie>
</rp> and </rt> act like normal phrasing element end tags

[04:56:02.0000] <Hixie>
<ruby> represents a run of one or more ruby annotations

[04:57:00.0000] <Hixie>
each annotation consists of some phrasing content followed by an <rt> element or a run of <rp>-<rt>-<rp> elements

[04:57:01.0000] <Hixie>
<rp> elements get ignored altogether

[04:58:00.0000] <Hixie>
the renderer takes all content of the <ruby> element and splits it into groups of  not-rt, rt

[04:58:01.0000] <Hixie>
for each of those groups it renders the not-rt as the base and the rt as the ruby text

[04:59:00.0000] <Hixie>
/me notes this in the spec for tomorrow

[04:59:01.0000] <annevk>
</ruby> should close <rp> / <rt>

[05:00:00.0000] <Hixie>
it would, if treated as i described above

[05:00:01.0000] <annevk>
without error?

[05:00:02.0000] <Hixie>
oh you don't want it to be an error?

[05:00:03.0000] <Hixie>
sure, we can add that

[05:00:04.0000] <annevk>
cool

[05:00:05.0000] <annevk>
/me likes lazy markup

[05:02:00.0000] <myakura>
http://tinyurl.com/6b7s4w (babelfish'd from ja to en) mentions that ie8's ruby support.. wonder how that'll be done..

[05:04:00.0000] <Hixie>
http://www.google.com/translate?u=http%3A%2F%2Fwww.exconn.net%2FBlogs%2Fwindows%2Farchive%2F2008%2F03%2F09%2F23648.aspx&hl=en&ie=UTF8&sl=ja&sl=es&tl=en&tl=en

[05:04:01.0000] <Hixie>
mildly better

[05:05:00.0000] <annevk>
the Google version spells Ruby wrong (Rubi)

[05:06:00.0000] <Hixie>
i noticed that :-)

[05:06:01.0000] <Hixie>
you can suggest a better translation :-)

[05:06:02.0000] <Hixie>
hover over it

[05:10:00.0000] <Hixie>
placeholder: http://www.whatwg.org/specs/web-apps/current-work/#the-ruby

[05:10:01.0000] <Hixie>
i'll fill in the details tomorrow

[05:11:00.0000] <Hixie>
MikeSmith: i don't envy you your job

[05:12:00.0000] <MikeSmith>
heh

[05:12:01.0000] <MikeSmith>
you referring to my survey mail, I guess

[05:13:00.0000] <Hixie>
yeah

[05:14:00.0000] <Hixie>
if you do have people leave the group because of the quorum thing, please let me know so i can invite them to send feedback through the whatwg -- i imagine there might well be people who want to send feedback but don't care about the voting nonsense

[05:14:01.0000] <Hixie>
(me, for instance)

[05:15:00.0000] <MikeSmith>
I think at least a couple of those people who have not responded are not actually interested in the HTML work

[05:15:01.0000] <MikeSmith>
or maybe just interested in the parts of it that relate to forms

[05:16:00.0000] <MikeSmith>
I'd wonder whether they would have plans to ever send feedback and have a discussion about that even

[05:18:00.0000] <MikeSmith>
but yeah, if they do decide to leave, I will let you know, and will myself point out to them that the whatwg list is a great place for them to submit any feedback they might have

[05:18:01.0000] <annevk>
hmm, <rt> closes <rp>, but <rp> does not close <rt>

[05:19:00.0000] <annevk>
</ruby> closes <rt><rp>

[05:19:01.0000] <Hixie>
annevk: yeah, IE's handling is seriously borked

[05:20:00.0000] <Hixie>
MikeSmith: cool

[05:20:01.0000] <Hixie>
what's our current quorum level?

[05:20:02.0000] <MikeSmith>
we are short one response

[05:21:00.0000] <Hixie>
ah

[05:21:01.0000] <MikeSmith>
ChrisW told me on Friday that he would respond by the deadline, but he didn't (he's traveling and I would guess he may have been without connectivity)

[05:21:02.0000] <Hixie>
wow, no negative votes at all

[05:21:03.0000] <Hixie>
impressive

[05:21:04.0000] <Hixie>
ok

[05:21:05.0000] <Hixie>
bed time

[05:21:06.0000] <Hixie>
nn

[05:21:07.0000] <MikeSmith>
'night

[05:23:00.0000] <MikeSmith>
hmm, is there not any kind of svn webview interface for whatwg svn? or some way that I can get a URL for a particular revision of the spec?

[05:23:01.0000] <MikeSmith>
I know about http://svn.whatwg.org/webapps/ but as far as I can see, it only shows the latest rev

[05:25:00.0000] <annevk>
MikeSmith, http://html5.org/tools/web-apps-tracker ?

[05:26:00.0000] <annevk>
oh, I see

[05:26:01.0000] <annevk>
no

[05:26:02.0000] <MikeSmith>
ah, OK

[05:26:03.0000] <annevk>
should be quite trivial to make that though

[05:27:00.0000] <MikeSmith>
yeah, I would think so, since the web-apps-tracker needs to be able to get arbitrary revisions anyway

[05:51:00.0000] <gsnedders>
MikeSmith: All it does, IIRC, is run svn diff

[05:52:00.0000] <annevk>
/me reads http://www.w3.org/2001/tag/2008/05/21-minutes.html ...

[05:53:00.0000] <annevk>
"TBL: I don't resent +xml, but I think text/html should migrate smoothly to XML over time."

[05:56:00.0000] <Philip`>
MikeSmith: "we are short one response" - I (or, rather, Dashiva's script) count 12 Member responses, which is exactly how many are needed

[05:57:00.0000] <Dashiva>
There could be a bug :)

[05:57:01.0000] <Philip`>
(Oxford Brookes University, Apple, Inc., Mozilla Foundation, Opera Software, Mitsue-Links Co., Ltd., Google, Inc., Nokia Corporation, AOL LLC, International Webmasters Association / HTML Writers Guild, W3C/MIT, Boeing Company, University of Innsbruck)

[05:58:00.0000] <Philip`>
It's definitely cutting it quite fine, though

[05:59:00.0000] <MikeSmith>
I clearly can't count :)

[06:00:00.0000] <Philip`>
That's understandable, since you don't have 12 fingers ;-)

[06:01:00.0000] <annevk>
Maybe MikeSmith counted W3C * 3?

[06:02:00.0000] <MikeSmith>
I think I just miscounted

[06:03:00.0000] <annevk>
the only thing is that html5-diff needs to be updated presumably because of sandboxing and ruby

[07:49:00.0000] <takkaria>
mhm, those TAG minutes are interesting

[07:49:01.0000] <takkaria>
"the reason we're not seeing a strong example of the need for distributed extensibility is that the lack of namespace-based extensibility in HTML bounded the discussion"

[08:28:00.0000] <gsnedders>
Philip`: Why not just wrap anything < 0 or > 10 the max. bounds?

[08:28:01.0000] <Lachy>
I need to work out what questions to ask the audience in the presentation. This is what I have so far...

[08:28:02.0000] <Lachy>
Who has never heard of HTML 5 before?

[08:28:03.0000] <Lachy>
Who has, but doesn’t know much about it?

[08:28:04.0000] <Lachy>
Who thinks they know more about it than we do?

[08:28:05.0000] <gsnedders>
Who thinks HTML 5 is stupid because XHTML is the future?

[08:29:00.0000] <Lachy>
heh

[08:29:01.0000] <Philip`>
Who thinks alt should be mandatory?

[08:29:02.0000] <Lachy>
I don't want to start a flame war in the middle of a presentation

[08:30:00.0000] <Philip`>
:-(

[08:30:01.0000] <Philip`>
(Why are you asking the audience questions at all?)

[08:30:02.0000] <MikeSmith>
"Who among you would like to challenge us to a fist fight after this presentation?"

[08:31:00.0000] <Lachy>
Philip`, as a way to get the audience more involved and interested

[08:31:01.0000] <Lachy>
if you have better suggestions, let me know

[08:31:02.0000] <MikeSmith>
"Who among you is without sin and would like to cast the first stone?"

[08:31:03.0000] <Philip`>
/me has no useful experience at giving presentations, so he has no suggestions

[08:32:00.0000] <Lachy>
MikeSmith, what relevance does that have to HTML5?

[08:33:00.0000] <MikeSmith>
Lachy: I'm brainstorming here

[08:34:00.0000] <MikeSmith>
""Who among you would possibly have some red-haired ses or similar that you would like to share with us in appreciation for the fact that we have taken the time to come here and educate you about HTML5?"

[08:34:01.0000] <Lachy>
ses?

[08:34:02.0000] <Lachy>
did you mean sis, as in sister?

[08:35:00.0000] <MikeSmith>
Lachy: nope, ses

[08:35:01.0000] <Lachy>
then what is it?

[08:35:02.0000] <MikeSmith>
teh red-haired variety

[08:35:03.0000] <Lachy>
?

[08:36:00.0000] <MikeSmith>
Lachy: I dunno. something I heard about on the Internet somewhere

[08:36:01.0000] <Philip`>
"an ancient Chinese plucked zither"?

[08:37:00.0000] <Philip`>
Google finds "red-haired ses" once on the internet, at www.marijuanapassion.com...

[08:41:00.0000] <MikeSmith>
Lachy: I recommend mocking other people/groups in your presentation, as much as possible

[08:42:00.0000] <gsnedders>
Lachy: Especially that gsnedders guy.

[08:43:00.0000] <Philip`>
/me recommends not doing so, because it'll just make the audience think you're arrogant and/or wrong :-p

[08:44:00.0000] <Philip`>
(unless you're just confirming their pre-existing prejudices, in which case it's alright)

[08:45:00.0000] <Lachy>
gsnedders, I'll use that funny looking photo of you in the presentation when we're talking about the community

[08:45:01.0000] <gsnedders>
Lachy: jgraham's photo?

[08:45:02.0000] <gsnedders>
Lachy: Or what?

[08:46:00.0000] <Lachy>
gsnedders, yeah, that one will do

[08:46:01.0000] <Lachy>
http://flickr.com/photos/jgraham/2479527700/

[08:46:02.0000] <gsnedders>
I was meaning http://flickr.com/photos/jgraham/2479527704/

[08:46:03.0000] <gsnedders>
seeming that's also me

[08:47:00.0000] <Lachy>
no, people can't see your funny hair style in that one. The aim is to get people to laugh

[08:47:01.0000] <Lachy>
:-P

[08:47:02.0000] <gsnedders>
Lachy: If you want my hair, there are better photos :P

[08:48:00.0000] <Dashiva>
Oh, oh!

[08:48:01.0000] <gsnedders>
Lachy: http://flickr.com/photos/gsnedders/2396366112/?

[08:48:02.0000] <Dashiva>
You can have a "Hi, I'm a HTML5 document. And I'm a XHTML2 document" thing!

[08:48:03.0000] <gsnedders>
Lachy: Just look through photos tagged with gsnedders 

[08:50:00.0000] <MikeSmith>
gsnedders sporting the anglo-fro

[08:53:00.0000] <gsnedders>
I still think I look really stupid in http://flickr.com/photos/gsnedders/614426881/

[08:55:00.0000] <Lachy>
gsnedders, here's the incomplete slide http://lachy.id.au/temp/gsnedders.jpg

[08:56:00.0000] <gsnedders>
hah!

[08:56:01.0000] <Lachy>
anyone else want their photo included?

[08:57:00.0000] <gsnedders>
I was more being silly, but hey, I'm there now :P

[08:57:01.0000] <Lachy>
gsnedders, I know. But the more photos we get to make this presentation interesting, the less we have to say make it interesting

[08:58:00.0000] <MikeSmith>
Lachy: feel free to include this:

[08:58:01.0000] <MikeSmith>
http://flickr.com/photos/sideshowbarker/2067348343/

[08:58:02.0000] <Lachy>
who is that?

[08:58:03.0000] <gsnedders>
Lachy: I still think http://flickr.com/photos/jgraham/2479527704/ is a better photo of me

[08:59:00.0000] <Lachy>
gsnedders, no, photos from before your ascension are better :-)

[09:00:00.0000] <Dashiva>
Lachy: It's probably mike!

[09:02:00.0000] <gsnedders>
I should probably get me own photos from Cambridge online

[09:04:00.0000] <MikeSmith>
Lachy: I'm told that picture is of the W3C Special Missions Subsection Junior Co-Chief

[09:08:00.0000] <Lachy>
MikeSmith, does that mean it is or isn't you?

[09:09:00.0000] <MikeSmith>
here's picture of me wearing Antal Lakner's "Passive-Dress Double-Gravity Suit System"

[09:09:01.0000] <MikeSmith>
http://flickr.com/photos/sideshowbarker/88963458/

[09:09:02.0000] <MikeSmith>
that suit weighs 40kg

[09:09:03.0000] <MikeSmith>
"A new development from INERS, the Passive Dress puts to the test one of the human body's fundamental functions - holding itself in position against the gravitational force of the planet where we live."

[09:09:04.0000] <MikeSmith>
Lachy: I am that I am.

[09:10:00.0000] <MikeSmith>
אהיה אשר אהיה

[09:10:01.0000] <MikeSmith>
ehyeh asher ehyeh

[09:11:00.0000] <Lachy>
MikeSmith, have you started typing in tongues?

[09:13:00.0000] <MikeSmith>
Yes, I've filled with Holy Ghost electricity matrix astrochemistry

[09:15:00.0000] <Lachy>
/me finds http://people.w3.org/mike/

[09:20:00.0000] <Lachy>
/me concludes that the photo of the drunk mobile phone user is Mike

[09:20:01.0000] <Lachy>
MikeSmith, do you have a high res version of this photo without the purple hat that's in the flickr version? http://sideshowbarker.net/about/

[09:22:00.0000] <MikeSmith>
Lachy: http://flickr.com/photos/sideshowbarker/1582027053/sizes/o/

[09:22:01.0000] <MikeSmith>
that's the highest res I got

[09:23:00.0000] <Lachy>
that'll do

[09:24:00.0000] <Lachy>
btw, what does "keitai" mean?

[09:24:01.0000] <Dashiva>
cell phone

[09:30:00.0000] <MikeSmith>
Lachy: what Dashiva said

[09:30:01.0000] <MikeSmith>
actually it literally means "mobile"

[09:31:00.0000] <Lachy>
MikeSmith, yeah, I translated "cell phone" to "mobile" anyway ;-)

[09:35:00.0000] <Dashiva>
And I translated from mobile to cell phone

[09:36:00.0000] <Dashiva>
MikeSmith: Does anyone say the full keitai denwa anymore?

[09:38:00.0000] <Dashiva>
Or more practically, is 'keitai' in any other use that would require disambiguation?

[09:39:00.0000] <MikeSmith>
Dashiva: yeah, keitai is still used some other ways

[09:39:01.0000] <Lachy>
Dashiva, are you an American?

[09:39:02.0000] <MikeSmith>
and do still here "keitai denwa" sometimes

[09:39:03.0000] <MikeSmith>
and especically, still see it written a lot

[09:40:00.0000] <MikeSmith>
e.g., in signs on trains and such that say, "Don't talk on your keitai denwa on the train."

[09:40:01.0000] <Dashiva>
Lachy: No

[09:41:00.0000] <Dashiva>
Lachy: My personal experience is that people who use 'mobile' understand 'cell', but the opposite is not always true. So just to be safe :)

[09:43:00.0000] <MikeSmith>
using the phrase "cell phone" is a good way to demonstrate lack of clue

[09:59:00.0000] <MikeSmith>
technical people at mobile operators/carriers use the word "terminal" to refer to keitai

[09:59:01.0000] <MikeSmith>
because to them, that's all it is (a data terminal)

[10:11:00.0000] <Philip`>
Do they use the term "non-terminal" for anything specific?

[12:53:00.0000] <hsivonen>
MikeSmith: I think long comments are bad for perf in general. If people still want very long comments, surely they can put them after the charset meta?

[12:55:00.0000] <hsivonen>
MikeSmith: so far it seems that 512 is a practical magic value. if implementations will have a cut-off, I think we are better off if the cut-off is predictable i.e. written down

[12:56:00.0000] <hsivonen>
MikeSmith: I'm less fond of the first child of head req.

[13:09:00.0000] <Lachy>
hsivonen, what's wrong with the first child of head requirment?

[13:10:00.0000] <hsivonen>
Lachy: as long as the prescan catches the meta, non-first child is harmless

[13:10:01.0000] <hsivonen>
so the byte location is what really matters

[13:11:00.0000] <Lachy>
but isn't it easier to get people to put meta as the first child of head, than to get them to count bytes?

[13:11:01.0000] <hsivonen>
MikeSmith: is "sing-song" Norwegian the same kind of "singing" as in Swedish spoken in Sweden?

[13:11:02.0000] <hsivonen>
Lachy: yes

[13:57:00.0000] <hsivonen>
/me reads TAG minutes on Oauth and OpenID

[14:06:00.0000] <hsivonen>
Lachy: what's the keynote theme that you are using?

[14:12:00.0000] <Lachy>
hsivonen, it's called Industrial

[14:15:00.0000] <hsivonen>
Lachy: thanks. I didn't realize it was one of the built-in ones

[16:22:00.0000] <Hixie>
Dashiva: nah, the audience wouldn't be able to hold their suspension of disbelief if you implied that there was an actual xhtml2 document somewhere

[16:23:00.0000] <Dashiva>
Surely they've made one or two for examples

[16:23:01.0000] <Lachy>
wow, it appears the last time the XHTML2 WG actually publised a draft of XHTML2 was back in 2006.

[16:26:00.0000] <webben>
Lachy: that's of the main document, I believe.

[16:27:00.0000] <Lachy>
webben, yeah, I'm aware they've been working on other drafts, but still, for a draft that's clearly inadequate, you'd think they'd spend a bit more time on it

[16:28:00.0000] <webben>
it's possible they do; but haven't published anything, I guess.

[16:29:00.0000] <webben>
it's also possible that there are bigger inadequacies in the modules that they're concentrating on

[16:29:01.0000] <annevk>
hmm, sickings comments are not really helping

[16:33:00.0000] <Hixie>
so i'm amused that two days after the TAG discussed issue 41, i closed it

[16:35:00.0000] <annevk>
weinig, othermaciej, do you guys have similar concerns to Mozilla with respect to Access Control or does it seem fine from a security perspective?

[16:35:01.0000] <Lachy>
if anyone wants to take a look at the current (incomplete) progress of my presentation slides, they're up here. http://lachy.id.au/temp/html5.zip (uses Keynote format)

[16:35:02.0000] <Lachy>
feedback appreciated

[16:36:00.0000] <annevk>
can you export to HTML or something?

[16:36:01.0000] <Lachy>
I can do PDF

[16:36:02.0000] <annevk>
i guess that works

[16:36:03.0000] <Hixie>
pdf would be nice

[16:36:04.0000] <Hixie>
i even have keynote and pdf would be nice :-)

[16:36:05.0000] <Lachy>
but you miss out on the cool animations with PDF :-)

[16:37:00.0000] <annevk>
somehow that doesn't strike me as a negative

[16:37:01.0000] <Hixie>
the animations are presumably not part of the substance :-)

[16:37:02.0000] <Lachy>
it's just slide transitions, so no biggy

[16:41:00.0000] <Lachy>
uploading html5.pdf now, will be done in about 3 minutes

[16:41:01.0000] <Lachy>
brb

[16:42:00.0000] <annevk>
Hixie, btw, can the <ruby> stuff easily be extended to cover complex ruby going forward?

[16:42:01.0000] <annevk>
Hixie, I heard rumors about at least Microsoft wanting to implement that

[16:43:00.0000] <Hixie>
well it wouldn't fall back well but sure

[16:43:01.0000] <Hixie>
we could add an attribute to <rt>

[16:44:00.0000] <Lachy>
done. http://lachy.id.au/temp/html5.pdf

[16:44:01.0000] <annevk>
btw, <q> and <rt> + <rp> are slightly inconsistent

[16:44:02.0000] <Hixie>
<ruby>a<rt>A</rt>b<rt>B</rt>c<rt>C</rt><rt span=3>123</ruby>

[16:44:03.0000] <annevk>
(not that it matters much)

[16:45:00.0000] <Hixie>
i'd have to see what the use cases really were first though

[16:47:00.0000] <Dashiva>
Like compound words where the pronounciation of two symbols is not the sum of the parts?

[16:50:00.0000] <annevk>
I wonder why IE doesn't do <rb>

[16:53:00.0000] <annevk>
Hixie, fair enough

[16:53:01.0000] <Hixie>
it doesn't do <rb> because <rb> is dumb

[16:53:02.0000] <Hixie>
same reason the html5 spec won't do <rb>

[16:54:00.0000] <annevk>
isn't that like saying that <dt> is dumb?

[16:54:01.0000] <Lachy>
I'm off to bed. If you've got any comments about the presentation slides, leave them in here and I'll read them in the morning. Good night.

[16:54:02.0000] <Hixie>
yes

[16:54:03.0000] <Hixie>
though at least in the case of <dt> it could separate multiple terms

[16:54:04.0000] <annevk>
oh true, <dt> actually has a use case

[16:55:00.0000] <annevk>
this is true for <rb> in case of complex ruby as well it seems

[16:55:01.0000] <Hixie>
not really, as shown in the idea above

[16:55:02.0000] <annevk>
so you can have several things in content order

[16:55:03.0000] <annevk>
in your example above they wouldn't be in content order, so less optimal fallback

[16:55:04.0000] <Hixie>
the way they do it in xhtml ruby falls back poorly

[16:56:00.0000] <Hixie>
eh?

[16:56:01.0000] <Hixie>
they'd be _more_ in content order

[16:57:00.0000] <annevk>
meh, you're right

[16:57:01.0000] <annevk>
geez, that's a huge pdf

[16:59:00.0000] <annevk>
heh, I like the slide that says "Solve Real Problems" and has a rubix cube on it

[16:59:01.0000] <annevk>
rubik's, apparently


2008-05-26
[17:00:00.0000] <annevk>
oh, "secure by design" and the Chinese wall is a funny too

[17:02:00.0000] <Hixie>
Lachy: what font is that?

[17:03:00.0000] <othermaciej>
annevk: we're not sure yet about all the security details (have not read the moz comments for instance)

[17:03:01.0000] <annevk>
othermaciej, k

[17:04:00.0000] <othermaciej>
we'll try to comment soon

[17:04:01.0000] <othermaciej>
now that we have an implementation and all

[17:04:02.0000] <annevk>
i'm fine with you guys just implementing the spec :)

[17:05:00.0000] <annevk>
/me wonders if RB and DT have been introduced to each other already

[17:05:01.0000] <annevk>
/me thinks they'd be a perfect match

[17:06:00.0000] <Dashiva>
<dt>a</dt><rt>A</rt> ?

[17:06:01.0000] <annevk>
Dashiva, RB and DT are not elements, but people :)

[17:07:00.0000] <Hixie>
http://software.hixie.ch/utilities/js/live-dom-viewer/ie8.html?...%3Cruby%3E%3Col%3E%3Cli%3E%3Cp%3Eaaa%3Crt%3Ebbb%3Cli%3Eccc%3C%2Fruby%3E...

[17:08:00.0000] <Dashiva>
Oh, those ones

[17:08:01.0000] <Hixie>
i have no idea what to do with that markup

[17:08:02.0000] <Hixie>
i can't see any way to get the IE rendering and the IE DOM at the same time

[17:08:03.0000] <Hixie>
or any thing resembling either

[17:09:00.0000] <annevk>
Dashiva, see html4all

[17:10:00.0000] <Dashiva>
Did dmitry start posting there?

[17:10:01.0000] <annevk>
Dashiva, no RB is proposing Unicode extensions

[17:11:00.0000] <Hixie>
annevk: any ideas for that ruby problem?

[17:11:01.0000] <Dashiva>
Hixie: Isn't it just the <rt> blissfully ignoring other tags?

[17:11:02.0000] <Hixie>
how do you mean?

[17:12:00.0000] <annevk>
Hixie, maybe <rt> should only do something special in case <ruby> is parent

[17:12:01.0000] <Dashiva>
Wait, what does a red LI mean?

[17:12:02.0000] <annevk>
Hixie, for both parsing and styling

[17:12:03.0000] <Hixie>
red LI means that the LI had two parents

[17:12:04.0000] <Hixie>
but that's another problem

[17:12:05.0000] <Hixie>
i'm just talking about the rendering

[17:12:06.0000] <annevk>
Hixie, in that case it would remain compatible with other browsers, at least

[17:12:07.0000] <Hixie>
annevk: the other browsers do nothing with ruby

[17:13:00.0000] <Dashiva>
Hixie: The rendering looks to me like  1. aaa (sup bbb) \n 2. (sup ccc) ...

[17:13:01.0000] <annevk>
Hixie, exactly, and since <rt> is not a child of <ruby> that would not change here

[17:13:02.0000] <Dashiva>
So apart from making the list counter rubytext, isn't it just an open <rt> going through?

[17:13:03.0000] <Hixie>
annevk: people using ruby are mostly worrying about IE, i'd wager.

[17:13:04.0000] <Hixie>
annevk: so compatibility with IE is what matters

[17:14:00.0000] <Hixie>
Dashiva: my point is how do we decide what to make the rt ruby-annotate?

[17:14:01.0000] <Dashiva>
Hixie: the bs or the cs?

[17:15:00.0000] <Hixie>
http://software.hixie.ch/utilities/js/live-dom-viewer/ie8.html?...%3Cruby%3E%20aa%20%3Col%3E%3Cli%3E%20bb%20%3Cem%3E%20cc%20%3Crt%3E%20ddddd%20%3C%2Frt%3E%20ee%20%3C%2Fem%3E%20ff

[17:15:01.0000] <webben>
Hixie: Would probably be worth confirming that. (I'd tend to assume the same thing, but there is that Firefox extension for Ruby as well.)

[17:15:02.0000] <othermaciej>
Hixie: the IE rendering is probably more important than the IE DOM, for Ruby

[17:15:03.0000] <Hixie>
othermaciej: greed

[17:16:00.0000] <Hixie>
a

[17:17:00.0000] <webben>
Hixie: Have you pulled up any Ruby in the Google caches? Do we know what current use is depending on?

[17:17:01.0000] <Hixie>
most ruby seems simple

[17:18:00.0000] <Hixie>
maybe we can get away with not being compatible with IE here

[17:18:01.0000] <Hixie>
and just being compatible with most of hte markup

[17:18:02.0000] <annevk>
http://software.hixie.ch/utilities/js/live-dom-viewer/ie8.html?...%3Cruby%3E%20x%20%3Crt%3E%20z%20%3Cdiv%3E%20y%20%3C%2Fdiv%3E%20%3C%2Frt%3E%20%3C%2Fruby%3E%20...

[17:18:03.0000] <annevk>
it seems like a CSS matter

[17:19:00.0000] <annevk>
that is, having block level elements inside <rt> will do funny stuff

[17:19:01.0000] <Hixie>
ok well we'll go for something simple and see what problems we get

[17:23:00.0000] <Dashiva>
Lachy: Pave the cowpaths should be more explicit about only paving GOOD cowpaths

[17:26:00.0000] <Dashiva>
Lachy: Existing content should mention that it means actual useful content, not just standards

[17:27:00.0000] <Dashiva>
Lachy: Priority doesn't say the list is actually in order

[17:28:00.0000] <Hixie>
wow, i can do the entire ruby parsing rules with one new entry in the in-body mode, and two new entires in the "generate implied end tags" list.

[17:28:01.0000] <Hixie>
and that's it.

[17:29:00.0000] <Hixie>
that was easier than i expected

[17:30:00.0000] <Dashiva>
Awesome

[18:02:00.0000] <Hixie>
not surprising microsoft think html5 should be split into smaller specs

[18:02:01.0000] <Hixie>
have you tried opening html5 in IE recnetly?

[18:02:02.0000] <Hixie>
sheesh

[18:03:00.0000] <othermaciej>
heh

[18:17:00.0000] <Philip`>
Which IE?

[18:18:00.0000] <Philip`>
IE8 seems particularly bad on the spec - it freezes for around ten seconds when moving the mouse over a link

[18:19:00.0000] <Hixie>
IE8b1, yes

[18:20:00.0000] <Philip`>
You shouldn't expect beta software to actually work

[18:20:01.0000] <Philip`>
Actually, you shouldn't expect any software to work

[18:21:00.0000] <Philip`>
It's good to get that warm feeling of pleasant surprise when you find some that really does work

[18:24:00.0000] <roc>
IE8b1 is particularly bad (apparently)

[18:36:00.0000] <Hixie>
annevk: ok, check out the spec, is it ok?

[18:40:00.0000] <Dashiva>
Lachy: Well-defined behavior and handle errors should probably be adjacent, so you can point out that both "correct" and "incorrect" documents are consistent

[18:42:00.0000] <Dashiva>
Lachy: Media independence sounds kinda like starting with a desktop and "adding support" for other things. But that could be just me.

[19:16:00.0000] <Hixie>
areas i could focus on next if no-one has any specific requests: abbr, alt, links, focus, usemap, sections

[19:17:00.0000] <othermaciej>
I am tempted to request alt just to watch the fireworks

[19:22:00.0000] <Dashiva>
Hixie: Let's leave alt for the end, it's going to take up 90% of LC anyhow :)

[19:23:00.0000] <Dashiva>
Is links <a> or <link>?

[19:28:00.0000] <MikeSmith>
Hixie: how about URLs?

[19:28:01.0000] <MikeSmith>
http://www.whatwg.org/specs/web-apps/current-work/#urls

[22:18:00.0000] <Lachy>
Hixie, the font is Helvetica Neue Light

[22:20:00.0000] <MikeSmith>
Lachy: what day is your presentation?

[22:20:01.0000] <Lachy>
Thursday morning at 11:00

[22:20:02.0000] <MikeSmith>
can you plan to Skype it out?

[22:21:00.0000] <Lachy>
maybe

[22:21:01.0000] <Lachy>
it will be podcast afterwards though

[22:22:00.0000] <MikeSmith>
ah

[22:22:01.0000] <MikeSmith>
OK

[22:23:00.0000] <MikeSmith>
Lachy: so what do you plan to say if you get the "how long with it be before there's a final rec for HTML5" question?

[22:25:00.0000] <Lachy>
I plan to focus on how authors can use features as soon as they are implemented and shipped in browsers

[22:26:00.0000] <MikeSmith>
Lachy: also always worth pointing out that the record of implementation progress on parts of the spec is exceptional

[22:26:01.0000] <MikeSmith>
that is, it's getting implemented at a much more rapid pace than just about any other spec-in-progress that anybody'd want to compare it with

[22:27:00.0000] <Lachy>
good point

[22:28:00.0000] <MikeSmith>
Lachy: I think it's possible that you may have some reporters or other media people approach you there and ask you about the schedule and other things

[22:28:01.0000] <MikeSmith>
e.g., splitting up the spec

[22:29:00.0000] <MikeSmith>
many reporters already have in mind pretty much what they want to write about and will try to steer you into saying something that supports whatever message that is

[22:29:01.0000] <MikeSmith>
so it pays to be a bit careful what you say to them

[22:30:00.0000] <MikeSmith>
e.g., before responding to some question that they ask, ask them what *they* think

[22:33:00.0000] <MikeSmith>
put them on the defensive instead of letting them put you on the defensive

[22:33:01.0000] <Lachy>
ok

[22:34:00.0000] <MikeSmith>
/me gets a Reboot invitation in his inbox

[22:35:00.0000] <MikeSmith>
Lachy: maybe you or annevk or somebody in Europe can consider proposing an HTML5 talk for Reboot

[22:35:01.0000] <MikeSmith>
annevk did one last year

[22:35:02.0000] <Lachy>
when is Reboot?

[22:35:03.0000] <MikeSmith>
event in Copenhagen each year

[22:36:00.0000] <MikeSmith>
http://www.reboot.dk/index.php

[22:36:01.0000] <MikeSmith>
26-27 June

[22:37:00.0000] <MikeSmith>
un-conference

[22:38:00.0000] <Lachy>
maybe. It depends if Opera will be willing to send me to that one as well

[22:46:00.0000] <Lachy>
MikeSmith, what's the purpose of this http://dev.w3.org/html5/pubnotes/ ? Is that a summary of the spec changes since the last WD?

[22:47:00.0000] <MikeSmith>
Lachy: yes

[22:48:00.0000] <MikeSmith>
Lachy: also, I intend http://dev.w3.org/html5/pubnotes/ to also be a record that we maintain going forward

[22:48:01.0000] <MikeSmith>
so each time a change is made to the spec, we also update that

[22:49:00.0000] <MikeSmith>
with separate sections for each WD

[22:49:01.0000] <MikeSmith>
that is, after we publish the next WD this week or next, the stuff that is there for the FPWD will become a static/historical section

[22:50:00.0000] <MikeSmith>
and we start a new section, which becomes a static section after we publish the next-next WD (3 months or so from now)

[22:51:00.0000] <Lachy>
so the one document will contain all the changes for all WDs, instead of starting a new document for each?

[22:55:00.0000] <MikeSmith>
Lachy: well, that's what I've been planning at least

[22:56:00.0000] <MikeSmith>
it seems better to me to have on running record

[22:56:01.0000] <MikeSmith>
because our WD dates are actually just arbitrary

[22:56:02.0000] <Lachy>
it might be easier if each on was on a separate page at least, since otherwise the one page will grow excessively large after just a few WDs

[22:57:00.0000] <MikeSmith>
well, that's also a complaint against the spec itself

[22:57:01.0000] <MikeSmith>
the pubnotes doc is still only 155K

[22:57:02.0000] <Lachy>
ok

[22:57:03.0000] <MikeSmith>
it has a long way to go before it approaches the 2.4MB size of the spec

[22:59:00.0000] <Lachy>
since the W3C gives each one a dated URI anyway, and each section will remain static, you could just link to all the previous versions and then replace the doc with the latest changes only

[23:05:00.0000] <MikeSmith>
Lachy: yeah, I suppose

[23:05:01.0000] <MikeSmith>
we'll see what we do after we get the next WD out

[00:29:00.0000] <MikeSmith>
Philip`: about your offline-webapps message: seems to me like the serialization should just be referred to as "text/html", with "HTML 5" and "HTML5" both just meaning the abstract language

[00:30:00.0000] <MikeSmith>
any distinction that might try to be made between "HTML 5" and "HTML5" is not going to be intuitive to readers

[01:06:00.0000] <hsivonen>
Lachy: slides 8 through 11 are PowerPointish--not Keynoteish

[01:07:00.0000] <hsivonen>
Lachy: I don't understand the photo for existing content

[01:07:01.0000] <hsivonen>
Lachy: I like the rubik's cube

[01:07:02.0000] <hsivonen>
DOM tree, too

[01:08:00.0000] <hsivonen>
Lachy: I don't understand slide 34 without speech

[01:09:00.0000] <hsivonen>
Lachy: slide 33 needs a warning against deploying content features before any browser implements that particular feature

[01:09:01.0000] <hsivonen>
the photo of the rosetta stone has low res

[02:18:00.0000] <Lachy>
hsivonen, the photo for Support Existing Content is of a bridge support. If you have a better idea, let me know

[02:20:00.0000] <hsivonen>
Lachy: I don't have a better idea.

[02:21:00.0000] <hsivonen>
Lachy: the photo made me think it was illustrating bridging instead of the support wires

[02:22:00.0000] <othermaciej>
so you guys have pictures for the HTML Design Principles?

[02:22:01.0000] <othermaciej>
is there gonna be a presentation about them?

[02:23:00.0000] <Lachy>
othermaciej, jgraham and I are presenting the desing principles as part of our presentation at @media

[02:24:00.0000] <othermaciej>
neat

[02:24:01.0000] <othermaciej>
I remember when we wrote down the first wiki version of that document, based on an IRC conversation

[02:24:02.0000] <Lachy>
you can download the current incomplete slides here http://lachy.id.au/temp/

[02:24:03.0000] <othermaciej>
now even people who hated them originally cite them as canon

[02:24:04.0000] <Lachy>
Either in keynote format (html5.zip) or PDF

[02:25:00.0000] <othermaciej>
sure, I am interested to see

[02:25:01.0000] <Lachy>
hsivonen, do you have any ideas of how to make slides 8-11 less like powerpoint?

[02:27:00.0000] <othermaciej>
slide 9 has incorrect parallel construction

[02:27:01.0000] <hsivonen>
Lachy: by turning each bullet point into a slide with one or two keywords from the bullet point on the slide using the "Title - Centered" slide master

[02:27:02.0000] <othermaciej>
Document/Standardise/Develop vs. Defining

[02:28:00.0000] <othermaciej>
hah, the pictures are great

[02:30:00.0000] <hsivonen>
Handle Errors is especially great

[02:34:00.0000] <hsivonen>
Lachy: your slides are missing CC attribution and license legalese

[02:34:01.0000] <hsivonen>
for the photos that is

[02:38:00.0000] <Lachy>
hsivonen, I'll add the licence stuff at the end

[02:39:00.0000] <hsivonen>
/me doesn't like it that RNG list {} cannot contain interleave

[02:56:00.0000] <annevk>
MikeSmith, yeah, I thought about going to reboot, it's pretty cool, but there doesn't seem to be much activity going on for this year yet...

[02:57:00.0000] <hsivonen>
let's see if I can still deploy a new instance of V.nu after all these massive parser changes...

[02:57:01.0000] <annevk>
Hixie, the ruby stuff is not checked in yet, fwiw, but it looks ok

[03:07:00.0000] <annevk>
Hixie, links / definition for URL would be cool

[03:07:01.0000] <annevk>
hsivonen,are you already up to date with the new stuff?

[03:10:00.0000] <hsivonen>
annevk: I am not

[03:15:00.0000] <hsivonen>
annevk: http://bugzilla.validator.nu/buglist.cgi?component=HTML+parser&bug_status=NEW&bug_status=ASSIGNED

[03:17:00.0000] <annevk>
MikeSmith, ok, submitted a proposal: http://www.reboot.dk/artefact-4161-en.html

[03:18:00.0000] <hsivonen>
annevk: what's the length of a presentation at reboot?

[03:18:01.0000] <annevk>
25 - 30 minutes I believe

[03:19:00.0000] <hsivonen>
oh great. the JDK did roughly what HTML5 requires now but I had to code around it to get the behavior that the spec used to require a while ago (for charset name matching)

[03:21:00.0000] <hsivonen>
"For instance, "GB_2312-80" and "g.b.2312(80)" are considered equivalent names. " that's just scary

[03:21:01.0000] <annevk>
and weird

[03:21:02.0000] <hsivonen>
I wonder if making  scary

[03:21:03.0000] <annevk>
and annoying to test

[03:21:04.0000] <hsivonen>
argh

[03:21:05.0000] <hsivonen>
copy paste error

[03:22:00.0000] <hsivonen>
I wonder if making g.b.2312(80) an error would make sense

[03:25:00.0000] <hsivonen>
aargh. It looks like I regressed comment tokenization and the test cases didn't catch it

[03:42:00.0000] <gsnedders>
hsivonen: It's what WebKit does, and there is stuff that relies upon it, sadly :(

[03:43:00.0000] <gsnedders>
Oh typical. I'm in Cophenhagen a few days after reboot

[03:45:00.0000] <hsivonen>
hmm. nice. I can't reproduce the comment problem with a minimized test case

[03:46:00.0000] <hsivonen>
so it looks like I have to run a debugger with the huge spec file

[06:39:00.0000] <annevk>
hsivonen, textarea input for http://parsetree.validator.nu/ ?

[06:43:00.0000] <hsivonen>
annevk: I suppose I have to build a proper UI for it, yes :-(

[06:44:00.0000] <annevk>
hmm, no need for proper UI

[06:44:01.0000] <annevk>
copying http://james.html5.org/parsetree.html wfm

[06:44:02.0000] <annevk>
ideally it uses GET

[06:52:00.0000] <Lachy>
anyone have an idea for a cool/funny blog name I can use for an example site in my presentation?

[06:52:01.0000] <Lachy>
Something more creative than "My Blog" would be nice

[06:53:00.0000] <annevk>
Cat Pictures for Everyone

[06:53:01.0000] <Lachy>
perfect!

[06:54:00.0000] <Lachy>
hmm. a Web 2.0 version of that would be better

[06:57:00.0000] <annevk>
C4t 3xch4ng3!

[06:59:00.0000] <Lachy>
what do you think of this? http://www.h-master.net/web2.0/image/(reflect)Cat%20LovrBETA.png

[06:59:01.0000] <annevk>
wfm, I won't be there :p

[07:01:00.0000] <Lachy>
that will give me a perfect opportunity to put in some of those LOL cats with HTML on them

[07:03:00.0000] <Philip`>
MikeSmith: Renaming the "HTML5" syntax to "text/html" seems too ugly, particularly since we'd have to rename "XHTML5" to "application/xhtml+xml" too

[07:04:00.0000] <MikeSmith>
Philip`: I think "text/html" is as easy to pronounce as "HTML5"

[07:05:00.0000] <Philip`>
MikeSmith: But it's harder to type, which is more important than pronunciation

[07:06:00.0000] <annevk>
HTML, XHTML, and HTML5

[07:07:00.0000] <annevk>
since HTML and XHTML don't have versions, that should be ok

[07:07:01.0000] <annevk>
HTML5 can also be written as HTML 5 for backwards compatibility

[07:07:02.0000] <Lachy>
MikeSmith, when and where did you suggest calling it text/html? In a mail or somewhere in IRC?

[07:30:00.0000] <MikeSmith>
Lachy: here

[07:32:00.0000] <Philip`>
"07:34 < MikeSmith> Philip`: about your offline-webapps message: seems to me like the serialization should just be referred to as "text/html", with "HTML 5" and "HTML5" both just meaning the abstract language"

[07:35:00.0000] <Dashiva>
Lachy: Do the <i> one!

[07:36:00.0000] <Lachy>
Dashiva, specifically which one?

[07:36:01.0000] <Dashiva>
http://icanhascheezburger.com/2008/03/24/funny-pictures-i-i/

[07:39:00.0000] <Lachy>
why does photoshop have to be so damn compliated?!

[07:40:00.0000] <Dashiva>
Because if everyone can do it, it's no longer cool

[07:41:00.0000] <Lachy>
yeah, but it sucks cause I can't even figure out how to change the colour of a layer

[07:58:00.0000] <Lachy>
bugger photoshop, it's uselss. I'll use Fireworks instead, cause at least it lets me work with the image easily

[07:58:01.0000] <Lachy>
(or maybe it's just the user that's useless)

[08:04:00.0000] <takkaria>
it just has a steep learning curve

[08:05:00.0000] <Lachy>
takkaria, yeah, unfortunately, I don't have enough time to learn how to use it and produce some mockup websites by this afternoon, or tomorrow at the latest

[08:39:00.0000] <Philip`>
http://www.jpl.nasa.gov/news/phoenix/images.php?fileID=6332 - "This image ... was taken between 2008-05-26T00:19:03.098 and 2008-05-26T00:19:03.274 local time" - I think that's the first time I've ever seen someone seriously use millisecond datetimes

[08:39:01.0000] <Philip`>
The phrase "local time" is unfortunately ambiguous when they're on Mars, though

[08:46:00.0000] <annevk>
heh

[09:02:00.0000] <Lachy>
I need a caption for that LOL cat image, for demonstrating figure/legend

[09:12:00.0000] <Lachy>
Here's my sample blog design. Does it look ok? http://lachy.id.au/temp/blog.png

[09:13:00.0000] <gsnedders>
oh noes! it's da mow mow!

[09:13:01.0000] <Dashiva>
It's possible to read the text, you gottam ake it smaller :P

[09:13:02.0000] <Lachy>
I made the text bigger since it needs to be readable by the audience

[09:14:00.0000] <Lachy>
obviously, if it were a real site, I'd make the text 8pt Arial

[09:16:00.0000] <Dashiva>
Well, maybe put some text in the entry

[09:16:01.0000] <Lachy>
Dashiva, like what?

[09:16:02.0000] <Lachy>
Hixie, is it ok that I stole Astrophy from you?

[09:17:00.0000] <Dashiva>
Lachy: Dunno. I just figured it'd be more bloglike with some text :)

[09:18:00.0000] <Lachy>
I'm not that creative. But if you think of something appropriate, I'll add it

[09:21:00.0000] <Dashiva>
Make a pun on Kitten HTML and KHTML!

[09:23:00.0000] <myakura>
Fork it then call one WebKitten :)

[09:23:01.0000] <myakura>
well, don't fork :)

[09:23:02.0000] <annevk>
hehe

[09:25:00.0000] <gsnedders>
myakura: Spoon it?

[09:26:00.0000] <Lachy>
/me adds credit to Hixie for the cat logo

[09:26:01.0000] <myakura>
gsnedders: http://www.doknowevil.net/wp-content/uploads/2007/05/omg_spoon_cat.jpg

[10:05:00.0000] <annevk>
Lachy, if you need people to laugh: http://img515.imageshack.us/img515/8726/jobsco0.gif

[11:16:00.0000] <Lachy>
LOL. It's funny how RB talks about the problems caused by moving an attribute in the null namespace from one element to another, then admits in the following paragraph that the problem doesn't exist. http://lists.w3.org/Archives/Public/public-html/2008May/0613.html

[11:19:00.0000] <Lachy>
/me resists the temptation to respond, since it's unlikely to achieve anything

[11:20:00.0000] <Dashiva>
I'm just wondering what he wants to accomplish

[11:20:01.0000] <Dashiva>
Surely he realizes that it's years too late to change it

[11:21:00.0000] <Lachy>
I'm not sure. But the problems he's describing aren't real problems, and even if they were, his proposal suffers from additional problems beyond the back compat issues

[11:39:00.0000] <gsnedders>
Lachy: Can I see the full slide that I'm in, yet?

[11:42:00.0000] <Dashiva>
gsnedders: Worried? :)

[11:50:00.0000] <Lachy>
gsnedders, not yet. that section isn't finished yet

[12:14:00.0000] <annevk>
Dashiva, RB believes Namespaces for XML, namespaces in the DOM, etc. can still be changed

[12:48:00.0000] <annevk>
according to Wikipedia "rubi" is alternative to "ruby"

[12:53:00.0000] <Dashiva>
Probably a direct transliteration from Japanese

[13:24:00.0000] <gsnedders>
http://htmlpurifier.org/demo.php?html=%EF%BF%BF — The return of U+FFFF!

[14:30:00.0000] <gsnedders>
/me hopes <http://hg.gsnedders.com/spec-gen/file/tip/src/specGen/processes/xref.py> is a decent xrefing impl.

[14:32:00.0000] <Philip`>
Eww, tabs for indentation :-(

[14:34:00.0000] <gsnedders>
Spaces annoy me. It means I can't decide how much indentation something has — with tabs I can decide how wide they should be.

[14:35:00.0000] <Hixie>
http://www.nytimes.com/2008/05/26/technology/26firefox.html?th&emc=th

[14:35:01.0000] <hsivonen>
I suppose I should wait a while to see how feedback on ruby turns out

[14:35:02.0000] <Hixie>
"Microsoft waited five years before releasing the sixth version of

[14:35:03.0000] <Hixie>
Internet Explorer in 2006. Dean Hachamovitch, general manager of Microsoft's

[14:35:04.0000] <Hixie>
Internet Explorer group, says the company was focused on plugging security

[14:35:05.0000] <Hixie>
holes during that time."

[14:35:06.0000] <Hixie>
FIVE YEARS to plug security holes?

[14:35:07.0000] <Hixie>
exactly HOW BUGGY was their browser???

[14:36:00.0000] <gsnedders>
hah. They're still claiming that?

[14:36:01.0000] <Philip`>
gsnedders: Tabs make e.g. that hg page look broken and ugly, because the class contents are only indented by one space

[14:36:02.0000] <gsnedders>
Most of IE team wasn't even working on IE then.

[14:36:03.0000] <othermaciej>
claiming what?

[14:36:04.0000] <gsnedders>
Philip`: I know, it sucks. :(

[14:37:00.0000] <gsnedders>
Philip`: But it's only the web UI of a SCM :P

[14:37:01.0000] <gavin_>
http://krijnhoetmer.nl/irc-logs/whatwg/20080526#l-408

[14:37:02.0000] <othermaciej>
ah, a creative phrasing

[14:37:03.0000] <Philip`>
gsnedders: The solution is to use spaces, because then there won't be nasty problems when people view your code with various tools :-p

[14:37:04.0000] <gsnedders>
I hate lxml's .tail. It's a pain in the ass to deal with.

[14:37:05.0000] <Dashiva>
gsnedders: You have 'get term and strip' twice in the code

[14:38:00.0000] <Philip`>
/me actually tends to use whatever his text editor has by default, which varies depending on which text editor he uses

[14:38:01.0000] <othermaciej>
"the company was focused on plugging security holes during that time" ---> "the product was in maintenance mode and all they did was plug security holes as they came up"

[14:38:02.0000] <gsnedders>
Philip`: That's even worse :P

[14:38:03.0000] <gsnedders>
Dashiva: But on different things! :P

[14:39:00.0000] <Dashiva>
gsnedders: Make a function :)

[14:40:00.0000] <gsnedders>
Dashiva: Hey! I'm more concerned with getting something that works!

[14:40:01.0000] <gsnedders>
Dashiva: This is the first real project I've taken on in Python, FWIW

[14:41:00.0000] <gsnedders>
Dashiva: But yeah, it probably should be :P

[14:41:01.0000] <gsnedders>
[in a function]

[14:42:00.0000] <takkaria>
I wonder how many bytes are wasted a year by including licences at the top of every file

[14:42:01.0000] <Dashiva>
gsnedders: Maybe emit a warning if you encounter a second dfn for the same term?

[14:43:00.0000] <gsnedders>
Dashiva: I have no infrastructure to do such things!

[14:43:01.0000] <gsnedders>
/me wonders what the "real" spec-gen does

[14:43:02.0000] <Dashiva>
infrastructure?

[14:43:03.0000] <takkaria>
gsnedders: then start building it

[14:43:04.0000] <gsnedders>
it links to the last dfn for the term

[14:43:05.0000] <Philip`>
Run the real spce-gen on your set of test cases to see what happens :-)

[14:43:06.0000] <Philip`>
s/ce/ec/

[14:44:00.0000] <gsnedders>
Philip`: But I, um, don't, um, have, um, access to something that's W3C member only…

[14:44:01.0000] <gsnedders>
<p><dfn id=foo>foo</dfn><dfn id=foo0>foo</dfn><a href="#foo0">foo</a> is the output, FWIW

[14:44:02.0000] <Dashiva>
gsnedders: If you have no way to produce output, you can always raise an exception. Very XML and stuff.

[14:44:03.0000] <gsnedders>
with <p><dfn>foo</dfn><dfn>foo</dfn><span>foo</span> as input

[14:45:00.0000] <Philip`>
I imagine Hixie would love to have the really quite slow build proces abort halfway through with a fatal error message

[14:45:01.0000] <Dashiva>
Philip`: I know! Isn't it awesome?

[14:46:00.0000] <gsnedders>
Philip`: No, less than half way through

[14:46:01.0000] <gsnedders>
Philip`: Because the W3C and WHATWG copies of the spec go through the spec-gen separately

[14:47:00.0000] <Hixie>
i'd rather have draconian error handling here than error recovery

[14:47:01.0000] <Hixie>
either way i have to fix the error

[14:47:02.0000] <gsnedders>
/me shrugs

[14:47:03.0000] <Hixie>
but with error recovery i might check in the wrong thing if i don't notice

[14:47:04.0000] <gsnedders>
I can do draconian easily enough

[14:47:05.0000] <annevk>
please make id based on the title attribute in some way

[14:48:00.0000] <gsnedders>
annevk: it already is

[14:48:01.0000] <Hixie>
in fact right now the spec gen script really annoyingly does all kinds foweird fixups in the face of errors

[14:48:02.0000] <Hixie>
so much though that i had to add a step that uses henri's validator and then also add a step that shows me the diff post-spec-gen so i can check it didn't fuck anything up

[14:48:03.0000] <gsnedders>
annevk: http://hg.gsnedders.com/spec-gen/file/tip/src/specGen/utils.py#l38 if you're interested

[14:48:04.0000] <annevk>
gsnedders, <dfn title=x-x>x</dfn> ends up as <dfn title=x-x id=x>x</dfn>

[14:48:05.0000] <annevk>
oh, you changed that, ok

[14:49:00.0000] <gsnedders>
Hixie: Yeah, I've seen plenty of odd things. But with HTML parsers (in the plural) like it uses, that's to be expected :(

[14:50:00.0000] <Dashiva>
/me wonders how long until someone takes Hixie's comments out of context

[14:51:00.0000] <Hixie>
i've given up worrying about that

[14:51:01.0000] <Dashiva>
But these ones are really bizarro world

[14:52:00.0000] <gsnedders>
yay! http://www.w3.org/TR/xpath works but http://www.w3.org/TR/xpath/ doesn't. TRs aren't consistent about needing a / or not :(

[14:55:00.0000] <gsnedders>
/me thinks we need chtml5lib

[14:56:00.0000] <Philip`>
Everyone thinks we need chtml5lib

[14:57:00.0000] <Hixie>
aw man, anne sent a messgae about ruby to the painters list

[14:57:01.0000] <gsnedders>
hah!

[14:58:00.0000] <takkaria>
I wonder what kind of performance gain html5lib would get if just the tokeniser was replaced with C

[14:58:01.0000] <gsnedders>
http://hg.gsnedders.com/spec-gen/file/tip/src/specGen/utils.py#l72 — XPath is fun. Strings cannot contain both ' and ". Trying dealing with arbitrary input and making it safe to be used in XPath :\

[14:59:00.0000] <gsnedders>
/me wonders why he doesn't just do a simple replace there

[15:01:00.0000] <Philip`>
gsnedders: You should probably write that like 'u"concat('', '%s')" % (blah)' since that's easier to read than concatenating three strings

[15:02:00.0000] <gsnedders>
/me just made it look a bit saner without doing that

[15:02:01.0000] <gsnedders>
But yeah, that is a good idea

[15:03:00.0000] <gsnedders>
/me stops clearing that bit up for a day

[15:06:00.0000] <annevk>
Hixie, keeping them deliberately out of the loop would be bad

[15:07:00.0000] <annevk>
It was in reply to RB's response here: http://www.w3.org/2002/09/wbs/40318/wd-html5-may/results

[15:09:00.0000] <othermaciej>
annevk: is he really worth replying to?

[15:10:00.0000] <annevk>
Probably not, but that was just what triggered it, pointing out new additions on public-html should be done anyway I think

[15:11:00.0000] <othermaciej>
I think you are right

[15:12:00.0000] <Philip`>
annevk: It would have been clearer if your email said which of Robert Burns' concerns you were referring to, since he seems to have more than one

[15:12:01.0000] <takkaria>
RB has some very big concerns. :)

[15:13:00.0000] <annevk>
he has some issues, yeah :p

[15:14:00.0000] <gsnedders>
And his reasons for disagreeing actually make no sense, according to my built-in parser.

[15:22:00.0000] <annevk>
Philip`, feel free to add more info

[15:48:00.0000] <weinig>
annevk: regarding the Access-Control spec and XHR2, I can't tell if you are supposed to send request headers set by the user using xhr.setRequestHeader

[15:50:00.0000] <annevk>
weinig, they are included, though it might require a preflight request in case of GET requests

[15:51:00.0000] <annevk>
weinig, also, some are blacklisted

[15:51:01.0000] <annevk>
http://dev.w3.org/2006/webapi/XMLHttpRequest-2/#send and http://dev.w3.org/2006/waf/access-control/#cross-site

[15:51:02.0000] <weinig>
is perhaps looking at the wrong version of the spec

[15:51:03.0000] <weinig>
/me is indeed

[15:52:00.0000] <weinig>
annevk: thanks

[15:52:01.0000] <annevk>
the latest published drafts were not really clear on this

[15:52:02.0000] <weinig>
annevk: I was looking at the published stuff

[15:53:00.0000] <annevk>
weinig, feel free to e-mail next time something is not clear as our timezones are not really compatible :)

[15:53:01.0000] <weinig>
annevk: will do, thanks

[15:53:02.0000] <annevk>
weinig, I remember you trying to ping me before

[15:56:00.0000] <Dashiva>
I'm positively surprised by many responses to the publishing poll

[15:57:00.0000] <Dashiva>
It seems the "WDs need not be perfect" idea has really taken hold

[15:59:00.0000] <annevk>
It will be interesting when we go for LC

[16:00:00.0000] <Philip`>
Even for people who think there shouldn't be a WD yet, there already is one and nothing is lost by updating it

[16:00:01.0000] <Philip`>
annevk: Define "interesting" ;-)

[16:00:02.0000] <Dashiva>
Philip`: Interesting times interesting

[16:01:00.0000] <Dashiva>
Glazman's comment seems a bit strange, though. Is a small change to @alt enough to make it not seem like a new version of html?

[16:04:00.0000] <takkaria>
I like Laura Carlson's response -- "we need more people from outside the WG to comment on how wrong changing the syntax requirements for alt are"

[16:06:00.0000] <annevk>
that's not what she said

[16:06:01.0000] <takkaria>
no, but that's what the bits between the lines say. :)

[16:06:02.0000] <Philip`>
http://www.bellabotanica.com/ - excellent use of ruby elements

[16:06:03.0000] <Philip`>
...<br /><img src="images/group01.jpg"></span><rb></br>...

[16:07:00.0000] <Dashiva>
I was thinking... how many <rt> are there out there that are typos of <tr>?

[16:07:01.0000] <Dashiva>
I kept writing <tr> when I was making examples with <rt>

[16:08:00.0000] <annevk>
I'm not sure <table-row> and <ruby-text> would've been better

[16:14:00.0000] <Philip`>
/me finds quite a bit of ruby markup in "kids" pages on hiroshima.jp and nagasaki.jp

[16:14:01.0000] <takkaria>
Philip`: urls?

[16:15:00.0000] <Philip`>
takkaria: http://www.pref.nagasaki.jp/toukei/kids/ / http://www.city.otake.hiroshima.jp/kids/

[16:15:01.0000] <Philip`>
Oh, and http://www.city.kawasaki.jp/25/25zinken/home/kidspage/ too

[16:16:00.0000] <Philip`>
Is it a real correlation between ruby and pages-for-kids?

[16:16:01.0000] <takkaria>
from what I know of what ruby is used for, doesn't seem an unreasonable conjecture

[16:19:00.0000] <Philip`>
Is http://za455.myweb.hinet.net/ Chinese?

[16:20:00.0000] <Philip`>
(The others I seen all looked like Japanese, though I didn't check closely)

[16:20:01.0000] <Philip`>
s//'ve/

[16:21:00.0000] <takkaria>
I don't know, but it has spans with lang="en-us" in it :)

[16:22:00.0000] <Dashiva>
Philip`: It's less needed in pages-for-adults on the internet, since you can copy-and-paste into a dictionary

[16:26:00.0000] <Philip`>
Hmm, great, IE8 doesn't support Ruby in IE8-mode

[16:26:01.0000] <takkaria>
hah

[16:26:02.0000] <Philip`>
Oh, actually, it does something with it, but not anything correct

[16:28:00.0000] <Philip`>
It does the right thing with font sizes and hiding <rp>s, but doesn't position stuff

[16:33:00.0000] <Philip`>
Hmm, great, IE8 doesn't render <xmp> as display:block in IE8-mode

[16:33:01.0000] <Philip`>
Are they intentionally trying to make IE8-mode incompatible with the web?

[16:34:00.0000] <roc>
the thought crossed my mind, but it wouldn't make sense

[16:36:00.0000] <takkaria>
is this the IE8 beta they released a couple of months ago still?

[16:37:00.0000] <Philip`>
Yes

[16:38:00.0000] <Dashiva>
Release fast, release often

[16:52:00.0000] <Philip`>
<ruby style="ruby-align:rightvertical"><span lang="EN-US" style="font-family: 新細明體"><font size="4">僅</font></span>...

[16:52:01.0000] <Philip`>
The web is great

[16:52:02.0000] <Dashiva>
I'm in awe

[16:54:00.0000] <Philip`>
<FONT size="-1"><RP><RT></FONT><FONT size="+1" color="#0000ff"><B>TEL03-3968-887１ 　 FAX03-3968-8873</B></FONT><RP></RUBY><BR>

[16:55:00.0000] <Philip`>
Not quite sure what's going on there


2008-05-27
[17:27:00.0000] <Philip`>
hsivonen: http://validator.nu/?doc=http://philip.html5.org/demos/html/ruby/wild-examples.html - why is it trying to parse the contents of the <xmp> elements?

[17:35:00.0000] <roc>
the next IE beta is apparently next quarter

[17:36:00.0000] <Philip`>
i.e. between a few days and three months away?

[17:36:01.0000] <Philip`>
/me wonders when quarters start

[17:36:02.0000] <Philip`>
Oh, probably a month plus a few days

[17:37:00.0000] <Philip`>
Silly 1-based month numbering system

[17:37:01.0000] <roc>
that's just how we do things on our planet

[19:31:00.0000] <Dashiva>
/me is tempted to bang head against wall

[19:32:00.0000] <Dashiva>
Maybe the alt discussion has as goal to induce self-inflicted injuries on watchers, to eliminate them from further participation

[22:06:00.0000] <Hixie>
annevk: i don't really mind you mentioning it on public-html

[22:07:00.0000] <Hixie>
it's just that every time anyone brings anything up on the public-html list, people always start asking things for the sake of purity instead of practical reasons

[22:07:01.0000] <Hixie>
e.g. the numerous questions about why <rb> isn't in the spec

[22:07:02.0000] <Hixie>
clearly none of the people who asked that have ever used ruby before

[23:08:00.0000] <MikeSmith>
Hixie: hey, it worked: "Discuss the difference between rel='' on <link> and <a>. (whatwg r1706)"

[23:08:01.0000] <MikeSmith>
thanks

[23:08:02.0000] <Hixie>
nice

[01:05:00.0000] <hsivonen>
Philip`: I have no idea why the contents of <xmp> are parsed. Do we have tree builder test coverage for xmp?

[01:40:00.0000] <annevk>
Hixie, seems your collegue did the same thing on the WHATWG list so it doesn't really matter :)

[01:40:01.0000] <annevk>
(and cc'ed the public-html list in the process...)

[01:45:00.0000] <hsivonen>
/me wonders how high-troughput apps deal with the eventual performance of server HotSpot

[01:49:00.0000] <hsivonen>
/me sees a 2D barcode on the desktop version of Nico Nico

[01:50:00.0000] <hsivonen>
Philip`: it seems that the xmp problem was fixed as a side effect of the major parser refactoring: http://validator.nu/?doc=http://philip.html5.org/demos/html/ruby/wild-examples.html

[02:03:00.0000] <annevk>
hsivonen, your parenthetical doesn't make sense (last e-mail)

[02:05:00.0000] <hsivonen>
annevk: hmm. looks like my turned itself off in the middle of the sentence :-(

[02:05:01.0000] <hsivonen>
see it happened again. whoa

[02:06:00.0000] <annevk>
It doesn't make much sense for "data-" to be conforming. You can't access it using dataset...

[02:06:01.0000] <annevk>
hmm, time for a new one?

[02:20:00.0000] <Philip`>
Hixie: Why isn't <rb> in the spec, for the sake of making existing pages that use ruby correctly with <rb> be conforming? :-)

[02:21:00.0000] <annevk>
Has anyone written a library for .dataset yet?

[02:21:01.0000] <annevk>
People are struggling with it: http://forum.mootools.net/viewtopic.php?pid=49569

[02:22:00.0000] <annevk>
Other people hacking to add support for data-: http://creativepony.com/journal/2008/05/27/hack-markaby-to-let-you-use-a-html5-data-attribute/

[02:23:00.0000] <Philip`>
hsivonen: Okay, thanks

[02:24:00.0000] <annevk>
Philip`, <rb> isn't in the spec because there's no need for it

[02:25:00.0000] <Philip`>
Hixie: s/affilated/affiliated/ in recent checkin

[02:27:00.0000] <Philip`>
annevk: It's needed for avoiding gratuitously making existing content non-conforming

[02:34:00.0000] <annevk>
Philip`, most content I've seen so far isn't conforming anyway :)

[02:40:00.0000] <Philip`>
annevk: Many are pretty close to being (machine-checkably) conforming, so it seems unhelpful to add unnecessary differences

[02:41:00.0000] <hsivonen>
would dropping <rb> and </rb> in the parser be bad?

[02:42:00.0000] <annevk>
yes, rb { color:red } and such

[02:42:01.0000] <Philip`>
That would be confusing, because of <ruby><rb style="font-family: whatever">stuff</rb>...</ruby>

[02:42:02.0000] <Philip`>
though it's effectively what IE does already

[02:42:03.0000] <Philip`>
(so people have to use <span> instead)

[02:42:04.0000] <annevk>
ruby::base { }

[02:42:05.0000] <annevk>
ruby::ruby-base { }

[02:43:00.0000] <hsivonen>
annevk: did you make those up just now or are they from the CSS draft?

[02:43:01.0000] <annevk>
made them up

[02:43:02.0000] <Philip`>
That seems unpleasantly complex

[02:44:00.0000] <Philip`>
It took me years to even work out you could have selectors like "parent child", and :: is much more confusing :-p

[02:44:01.0000] <annevk>
the alternative is <rb>, which makes the markup more complex

[03:48:00.0000] <Hixie>
Philip`: <rb> is non-conforming to help authors who may get trapped by trying to style <rb> without success

[03:50:00.0000] <annevk>
so dataset[""] = x doesn't work either?

[03:50:01.0000] <Philip`>
Hixie: By that logic you should make all of ruby non-conforming, to help authors who may get trapped by trying to use it when a reasonably significant number of users use browsers which don't yet support it

[03:51:00.0000] <Hixie>
annevk: per spec, it should "work"

[03:52:00.0000] <Hixie>
Philip`: yeah i didn't buy it either. the real reason it's not conforming is because i don't want to add a useless element to the spec, and ruby right now is in very early stages, so i'm willing to sacrifice a few authors today for the good of authors in the coming decades.

[03:53:00.0000] <annevk>
it does make implementing ruby slightly more complex

[03:53:01.0000] <Philip`>
Is there a known reason why the original ruby spec has <rb>?

[03:54:00.0000] <annevk>
apparently it's feasible to implement ruby using the inline table model from CSS and some to set a row as a baseline (Opera has a feature for that)

[03:54:01.0000] <Philip`>
annevk: Do you mean the presence of <rb> makes it more complex?

[03:54:02.0000] <annevk>
no, the absence

[03:55:00.0000] <Philip`>
annevk: You'd have to cope with people who don't use <rb> anyway, so its conformingness seems irrelevant

[03:56:00.0000] <annevk>
unless rendering is defined without requiring pseudo elements :)

[03:56:01.0000] <annevk>
but yeah

[04:09:00.0000] <Philip`>
annevk: Existing content omits <rb>, so, for compatibility, rendering would have to be defined without requiring explicit <rb> elements

[04:10:00.0000] <Hixie>
/me tries to support anne in the xhr debacle

[04:10:01.0000] <Philip`>
(I suppose they could be implied, like tbody etc, but that seems kind of nasty)

[04:11:00.0000] <annevk>
Hixie, I think that's going ok now. I just need to get a few outstanding things resolved and then I simply ask for WG resolutions on each because I can't work it out with Julian and then we can go to CR (or another LC, not sure)

[04:12:00.0000] <Hixie>
WG resolutions?

[04:12:01.0000] <annevk>
Philip`, seems the main problem is the rb/rt pairing in the current draft

[04:13:00.0000] <Hixie>
why can't you just make a decision and move on?

[04:13:01.0000] <Hixie>
you'll always have some peopel unhappy

[04:13:02.0000] <Hixie>
html5 has all kinds of things julian disagrees with :-)

[04:13:03.0000] <annevk>
Hixie, yes, but in order to suggest that he can raise a formal objection, i need support of the WG or something

[04:13:04.0000] <annevk>
Hixie, that seems like the simplest solution to me

[04:14:00.0000] <Hixie>
/me doesn't plan on trying to get "support of the WG" before telling people to raise formal objections :-)

[04:14:01.0000] <Hixie>
/me doesn't really care about formal objections at all in fact :-)

[04:15:00.0000] <annevk>
Hixie, ah, I was wondering whether to reply to that e-mail from sicking asking for more info

[04:17:00.0000] <Hixie>
seriously though, if you've considered all the input and come to a conclusion, and he just disagrees with your conclusion but agrees that there is no information you aren't aware of, then asking the wg isn't going to do any good

[04:17:01.0000] <Hixie>
groups of people make random decisions

[04:18:00.0000] <Hixie>
the more people the more random

[04:18:01.0000] <Hixie>
random decisions aren't what's best for hte web

[04:18:02.0000] <annevk>
I had this debacle once in the CSS WG, I don't want it again

[04:19:00.0000] <annevk>
I don't like it when chairs of WGs go all mad on me

[04:19:01.0000] <Hixie>
eh you'll get used to it

[04:20:00.0000] <Hixie>
i guess i'm in the lucky position of being able to just pull out of the w3c altogether if it gets too crazy

[04:20:01.0000] <Hixie>
we really need a patent policy setup for whatwg

[04:20:02.0000] <Hixie>
that way we wouldn't need the w3c at all

[04:23:00.0000] <gsnedders>
Silly W3C.

[04:23:01.0000] <gsnedders>
"The only person who counts is the editor. If you disagree with the editor, go tell your mummy on him."

[04:25:00.0000] <Hixie>
if enough people disagree with the editor, the editor gets replaced and is out of a job

[04:25:01.0000] <Hixie>
but you'll always find people who disagree

[04:25:02.0000] <Hixie>
anyway. bed time. nn.

[04:25:03.0000] <gsnedders>
Nah. Everyone agrees with me.

[04:25:04.0000] <gsnedders>
night, Hixie 

[04:59:00.0000] <Philip`>
/me really doesn't like when online purchase systems show a payment page with an iframe containing (what claims to be) a page from his bank asking for his number and password for confirmation

[05:00:00.0000] <Philip`>
since I can't tell where the page is actually from, and even when I can tell I just find it's www.securesite.co.uk and can't tell how legitimate that is

[05:01:00.0000] <Philip`>
(I'm not even sure if there's an easy way to tell that it's HTTPS)

[05:03:00.0000] <annevk>
the banking system in the Netherlands uses redirects for that

[05:04:00.0000] <Philip`>
Oh, actually, securesite.co.uk is something totally unrelated - the legitimate one is securesuite.co.uk

[05:06:00.0000] <Philip`>
(http://ambrand.com/2006/09/06/is-securesuitecouk-a-phishing-scam/ etc)

[05:50:00.0000] <gsnedders>
Philip`, jgraham: Are either of you involved in the uni open days on 3rd/4th July?

[05:51:00.0000] <Philip`>
gsnedders: I'm not planning to be

[05:53:00.0000] <Philip`>
(though I'm also not planning not to be)

[05:54:00.0000] <gsnedders>
Philip`: :)

[05:54:01.0000] <gsnedders>
/me will be there

[05:54:02.0000] <Philip`>
(Depends on whether someone from our group cares enough to do a demo or something)

[05:55:00.0000] <gsnedders>
I need to decide whether to have a look at the Nat.Sci. stuff or not

[05:56:00.0000] <Philip`>
(Oh, I guess we don't care enough, since the email said to reply by 2nd May if interested in demoing)

[05:57:00.0000] <Philip`>
Any particular NatSci stuff?

[05:59:00.0000] <gsnedders>
Philip`: Well, Physics may be a good guess seeming that's what I've talked about

[06:00:00.0000] <gsnedders>
Philip`: and seeming I haven't done any other pure sciences in years (I regret not doing Chem., but it's too late now)…

[06:03:00.0000] <Philip`>
gsnedders: Ah, sounds reasonable

[06:04:00.0000] <gsnedders>
The other two prospectuses (York and Edi) arrived today, FWIW

[06:04:01.0000] <Philip`>
/me still hasn't worked out how to extract useful information from open days

[06:04:02.0000] <Philip`>
Both today? Sounds like a conspiracy

[06:24:00.0000] <gsnedders>
Philip`: Well, it still stands true that Cam's propaganda department is better

[06:25:00.0000] <Philip`>
Hmm, I haven't physically used my computer at the CL for so long that I've forgotten the login password :-/

[06:37:00.0000] <gsnedders>
Philip`: "password"?

[06:39:00.0000] <Philip`>
gsnedders: http://en.wikipedia.org/wiki/Password

[06:41:00.0000] <gsnedders>
Philip`: Why do I think you deliberately misinterpreted what I meant?

[06:41:01.0000] <Philip`>
(Turns out that actually I did remember the password, but the magic networking stuff was broken so it didn't recognise it until I rebooted (and it takes forever for the computer to reboot because it's got lots of magic networking stuff to set up))

[06:41:02.0000] <Philip`>
gsnedders: I didn't know how to correctly interpret what you wrote

[06:42:00.0000] <Philip`>
(and still don't)

[06:42:01.0000] <gsnedders>
Philip`: The string password is your password.

[06:42:02.0000] <Philip`>
Oh

[06:42:03.0000] <Philip`>
It isn't :-p

[06:47:00.0000] <Philip`>
/me now has Synergy working, so he can use his laptop's mouse and keyboard to control his desktop computer (which has 4x as much screen area, which is sometimes handy)

[06:47:01.0000] <gsnedders>
Philip`: See, this is why I have a 17" laptop :P

[06:49:00.0000] <Philip`>
I've got a 15" laptop, which usually isn't too small and isn't too hard to carry around :-)

[06:49:01.0000] <Philip`>
but it's not so good when I want a dozen virtual machine shell windows at once

[06:51:00.0000] <virtuelv>
/me has a 12.1" laptop

[06:51:01.0000] <virtuelv>
(And an external 20" widescreen monitor)

[06:53:00.0000] <gsnedders>
I just don't carry mine around much :)

[06:53:01.0000] <gsnedders>
And I have a nice rucksack to carry it in now :)

[07:26:00.0000] <gsnedders>
I'm about to make myself look really stupid. Time for me to try and write C.

[07:28:00.0000] <gsnedders>
I want to see how much I can speed html5lib up by just rewriting inputstream.charsUntil in C

[07:36:00.0000] <annevk>
algirht, now I need to move all XHR1 changes to XHR2

[09:03:00.0000] <gsnedders>
I'm failing badly at getting charsUntil any quicker

[09:07:00.0000] <Philip`>
Which part of charsUntil is slow?

[09:08:00.0000] <Philip`>
like, is it doing lots of complex calculation itself, or is it just spending its time pulling characters out of Python data structures?

[09:09:00.0000] <gsnedders>
Philip`: As far as I can see, the former

[09:09:01.0000] <gsnedders>
well, it's hard to say :P

[09:11:00.0000] <Philip`>
Try removing all the functionality from the function and see how that affects performance :-)

[09:12:00.0000] <Philip`>
and if it's still just as slow, then it must be the interface with surrounding code that's slow, in which case I guess you can't do much to optimise that function by itself

[09:13:00.0000] <gsnedders>
Philip`: There isn't really much functionality to remove, without completely affecting what the function has to do. Removing some of the stuff would make it far quicker as it was operating on less data. :P

[09:14:00.0000] <annevk>
just rewriting it in C would help

[09:15:00.0000] <Philip`>
You could remove all the stuff that implements the "until" thing, and run it with test data that doesn't depend on the "until" thing working, and see how much effect that has

[09:15:01.0000] <annevk>
and then we need it to ship with python by default

[09:31:00.0000] <Philip`>
Seems like charsUntil can be made ~30% faster just by using a list instead of a deque

[09:31:01.0000] <Philip`>
(and storing the string reversed, so all the operations happen on the tail of the list)

[09:32:00.0000] <Philip`>
(except for readChunk but that can be made to only be called when the list is nearly empty, I think)

[09:37:00.0000] <Philip`>
(Also it could maybe be made much faster by storing a list of strings, instead of a list of characters)

[09:37:01.0000] <Philip`>
/me might look at that later tonight

[09:38:00.0000] <gsnedders>
Philip`: How would that allow you to make it quicker, thoguh

[09:38:01.0000] <gsnedders>
*though

[09:38:02.0000] <Lachy>
oh crap! I just lost a whole heap of slides cause Keynote died :-(

[09:38:03.0000] <Lachy>
I

[09:38:04.0000] <Lachy>
I'll have to remake them now :-(

[09:38:05.0000] <Philip`>
gsnedders: Which "that" do you mean?

[09:38:06.0000] <gsnedders>
Philip`: str not char

[09:38:07.0000] <Philip`>
Lachy: Doesn't all office software have autosave?

[09:38:08.0000] <Philip`>
gsnedders: It'd let you run regexps

[09:39:00.0000] <Philip`>
and depending on how Python stores lists, it might be hugely more memory-efficient (hence cache-efficient and everything)

[09:39:01.0000] <gsnedders>
Philip`: http://code.google.com/p/html5lib/issues/detail?id=69

[09:40:00.0000] <Philip`>
/me wishes Opera didn't insist on opening patch files with "less"

[09:40:01.0000] <gsnedders>
Philip`: What if it opened it with "more"?

[09:40:02.0000] <hsivonen>
Lachy: no Time Machine?

[09:40:03.0000] <Philip`>
gsnedders: It'd let you run regexps without wasting loads of time calling 'join'

[09:41:00.0000] <Philip`>
(and without wasting memory etc)

[09:41:01.0000] <takkaria>
are strings immutable in python?

[09:41:02.0000] <Dashiva>
Philip`: Maybe you shouldn't set it to use less then :)

[09:42:00.0000] <Philip`>
Dashiva: I never did, since that'd be stupid :-)

[09:42:01.0000] <Philip`>
takkaria: Yes

[09:42:02.0000] <Lachy>
hsivonen, no. The weird thing is I thought I had saved it regularly, but I lost everything since last night

[09:42:03.0000] <gsnedders>
Philip`: It probably means changing a heckuva lot though

[09:42:04.0000] <Lachy>
Philip`, obviously keynote doesn't, though I wish it would

[09:43:00.0000] <Philip`>
gsnedders: I can't see any references other than a dozen mentions of self.queue in inputstream.py, and it should be feasible to optimise things somewhat without changing its interface

[09:46:00.0000] <Lachy>
at least this time I can remake it faster since I don't have to make decisions again, just remember what I had done, and I still have all the images

[10:09:00.0000] <annevk>
gsnedders, the real solution here is a port of hsivonen's parser to C++ and make a Python binding for that and get it shipped with Python 2.x

[10:10:00.0000] <annevk>
gsnedders, though maybe that's too complex for now

[10:10:01.0000] <gsnedders>
annevk: Yeah, sure. That gets a quick parser. But how about when you want to iterate over every element in HTML 5? libxml (and lxml therefore too) is the only thing reasonably quick currently

[10:25:00.0000] <annevk>
http://www.w3.org/Submission/2008/01/

[10:31:00.0000] <gsnedders>
HTTP is really odd.

[10:32:00.0000] <gsnedders>
I find it very interesting that IIS (or HTTP.sys, really) is so much stricter than Apache

[10:33:00.0000] <gsnedders>
Even at a basic level, Apache accepts "HTTP / 1.1"  while IIS sends 400

[10:34:00.0000] <gsnedders>
it is, arguably, allowed by RFC2616

[10:34:01.0000] <gsnedders>
implied lws sux.

[10:39:00.0000] <Dashiva>
annevk: So they want to push a patented algorithm into web fonts, or?

[10:41:00.0000] <annevk>
beats me

[10:41:01.0000] <annevk>
with the recent advanced of Opera and Safari, and soon Mozilla, I think EOT is pretty much dead

[10:41:02.0000] <annevk>
advancements*

[10:42:00.0000] <takkaria>
except for people who care about IE

[10:46:00.0000] <gsnedders>
/me reports a bug on CFNetwork

[10:51:00.0000] <Lachy>
any suggestions for what I should put on the slide about <video>?  We're probably going to show an actual demo video in a browser for this as well

[10:51:01.0000] <Philip`>
/me discovers that the bottleneck in his code is "for i in range(self.chunkOffset, len(self.chunk)):", so he rewrites it as an ugly "i = ...; while i < ...; ...; i += 1" and cuts 40% off his execution time

[10:52:00.0000] <Philip`>
(I would have thought Python wouldn't be so incredibly dumb as to fail to optimise that very common case)

[10:53:00.0000] <Dashiva>
Nobody uses indexes anymore, Philip`. It's all lists. ;)

[10:53:01.0000] <Philip`>
The whole point of this code is that lists are slow and I want to avoid them :-p

[10:53:02.0000] <Lachy>
Philip`, do you have any other cool <canvas> demos we could show off in our presentation?

[10:54:00.0000] <hober>
Philip`: xrange is your friend

[10:54:01.0000] <Lachy>
I might just use your game at canvex.lazyilluminati.com

[10:55:00.0000] <Philip`>
hober: Aha, thanks, that's not noticeably slower than the manual loop, which is nice :-)

[10:55:01.0000] <Philip`>
(though it'd still be nicer if Python did that optimisation automatically)

[10:56:00.0000] <Philip`>
Lachy: Canvex is about the only significant <canvas> thing I've done

[10:56:01.0000] <Lachy>
Philip`, ok.

[10:56:02.0000] <Philip`>
Lachy: There's http://www.p01.org/releases/DHTML_contests/files/20lines_twinkle/ which might be interesting

[10:56:03.0000] <Lachy>
Philip`, could you zip up all the files for me so I can easily download and use it offline?

[10:56:04.0000] <annevk>
Lachy, ask p01, he made that

[10:57:00.0000] <Philip`>
Lachy: Also http://glimr.rubyforge.org/cake/canvas.html with various stuff

[10:57:01.0000] <Philip`>
Lachy: What annevk said :-)

[10:58:00.0000] <Lachy>
annevk, yeah, I will. But was asking Philip` for his canvex game

[10:58:01.0000] <Philip`>
Lachy: Oh, I misinterpreted you in the same way

[11:00:00.0000] <Philip`>
Lachy: http://canvex.lazyilluminati.com/83/83.zip is an almost recent copy of hopefully all the necessary files

[11:00:01.0000] <Philip`>
(I think it's just missing the Safari logo in the outdoors bit, because the game didn't work in Safari back then)

[11:03:00.0000] <Lachy>
Philip`, it's reporting a well formedness error in ./textures/misc/ui_bottom.svgz

[11:03:01.0000] <Philip`>
Lachy: That's because Firefox is stupid, I think

[11:04:00.0000] <Philip`>
Either run it on a local web server, or modify play.xhtml to refer to .svg instead of .svgz

[11:04:01.0000] <Philip`>
...and if the .zip doesn't include .svg files then gzip -d the .svgz ones

[11:04:02.0000] <Lachy>
what's the .svgz file?

[11:04:03.0000] <Philip`>
gzip-compressed SVG

[11:04:04.0000] <Lachy>
ok, I can run it from a webserver

[11:05:00.0000] <Philip`>
Might need to make sure the web server is configured to send the right content-type/encoding for .svgz - I'm not sure if they do it by default

[11:06:00.0000] <Philip`>
(The problem with using file:/// is that Firefox guesses the content-type but not -encoding for .svgz files, so it tries parsing the compressed data as XML)

[11:06:01.0000] <Lachy>
what's the apache directive for setting the content-encoding?

[11:08:00.0000] <Lachy>
Philip`, what's the purpose of the graph I get when I press Cmd+T, and why have you overridden my Open New Tab command?

[11:08:01.0000] <Philip`>
AddEncoding gzip .svgz

[11:08:02.0000] <Philip`>
Lachy: It shows the framerate, and I overrode your new tab command because I'm cruel and unthoughtful

[11:33:00.0000] <gsnedders>
ARGH!

[11:33:01.0000] <gsnedders>
/me goes mad at RFC2616 using a totally undefined form of BF

[11:33:02.0000] <takkaria>
someone stamp on your foot?

[11:33:03.0000] <gsnedders>
*BNF

[11:53:00.0000] <Lachy>
oh, damn. running Windows in VMWare on Mac, going to http://localhost/ doesn't see the server running on the Mac, it tries to find the server running on Windows. :-(

[11:55:00.0000] <hober>
ipconfig /all in a terminal, look at the IP for 'default gateway', and point the browser at that

[11:55:01.0000] <hober>
IIRC

[11:56:00.0000] <gsnedders>
Lachy: Change the network type

[11:57:00.0000] <Lachy>
thanks hober

[11:59:00.0000] <Lachy>
gsnedders, ?

[12:00:00.0000] <gsnedders>
Lachy: IIRC, if you go to Settings -> Network (on the VM) one of the options should make localhost the host

[12:00:01.0000] <Lachy>
hober, re http://krijnhoetmer.nl/irc-logs/whatwg/20080515#l-1206 - did you get a chance to follow up on that?

[12:02:00.0000] <hober>
not yet; work's been pretty crazy

[12:03:00.0000] <hober>
hoping to take a look at it this weekend

[12:04:00.0000] <Lachy>
gsnedders, none of the prefs in there seem to help

[12:10:00.0000] <Philip`>
Lachy: You could use your real machine rather than a VM

[12:16:00.0000] <Lachy>
Philip`, I need to use the windows build because we don't have a suitable mac build that supports video

[12:16:01.0000] <Philip`>
Oh, how annoying - I can easily make charsUntil ten times faster on my artifical test case, but it makes about zero different in practice :-(

[12:17:00.0000] <Philip`>
*difference

[12:17:01.0000] <Philip`>
Lachy: Ah, okay

[12:21:00.0000] <Philip`>
...and when parsing the spec (with parse.py --no-html), inputStream.py is only about 30% of the total execution time anyway, so there's little room for global improvement anyway

[12:23:00.0000] <Philip`>
/me gives up instead

[12:26:00.0000] <gsnedders>
Philip`: That's why I was trying with the spec. That's not so artificial.

[12:26:01.0000] <Philip`>
Actually I'm cheating and testing with the first 1K or 10K lines of the spec, since I'm lazy

[12:27:00.0000] <gsnedders>
Philip`: parser.parse() with the lxml etree builder makes charsUntil 70% of the total time

[12:31:00.0000] <Windstoss>
Hi there!

[12:32:00.0000] <Windstoss>
Does Web Applications 1.0 either directly or indirectly refer to the Selectors API or XmlHttpRequest of the W3C Web API WG?

[12:33:00.0000] <Philip`>
gsnedders: When I run parse.py --profile --no-html --treebuilder=lxml on the first 10K lines of the spec, it spends cumtime=0.819 out of total 5.982 in charsUntil

[12:34:00.0000] <Philip`>
(and 1.507 in total in inputstream.py, vs 2.116 in tokenizer.py and 1.019 in html5parser.py)

[12:35:00.0000] <Windstoss>
This Colloquy sort of IRC Client is strange :(

[12:35:01.0000] <Philip`>
Windstoss: It does

[12:35:02.0000] <Philip`>
Windstoss: e.g. http://www.whatwg.org/specs/web-apps/current-work/multipage/section-origin.html has an [XHR] reference

[12:36:00.0000] <Philip`>
Windstoss: (Doesn't refer to the Selectors API as far as I can see, though)

[12:37:00.0000] <Windstoss>
Philip`: thanks! :)

[12:37:01.0000] <Windstoss>
Philip`: I'm writing a thesis about HTML 5 and was wondering, if the Selectors API is something seperate

[12:38:00.0000] <gsnedders>
Philip`: That's totally different to what I get from cProfile…

[12:38:01.0000] <Philip`>
Windstoss: I think the only relation is that similar people are working on both

[12:38:02.0000] <gsnedders>
Philip`: 5.214 out of 17.312 using python parse.py --no-html --profile --treebuilder=lxml ../../html5

[12:39:00.0000] <gsnedders>
Philip`: (that's on an entire several month old .src.html copy of html5)

[12:39:01.0000] <Windstoss>
Philip`: you mean those pragmatic, we do whats necessary sorts of people? ;)

[12:41:00.0000] <Philip`>
gsnedders: Whoops, I was using my modified inputstream.py... With the original it's more like 1.542 out of 6.628 in charsUntil

[12:41:01.0000] <Philip`>
but that's still only ~25%

[12:41:02.0000] <Philip`>
Oh, which is not much different to what you saw

[12:41:03.0000] <Philip`>
(but is quite different to 70%)

[12:42:00.0000] <gsnedders>
I got 70% from cProfile

[12:43:00.0000] <Philip`>
Even cutting execution time by 70% wouldn't make html5lib significantly faster, compared to the orders of magnitude from writing the whole thing in Java/C/etc

[12:44:00.0000] <Philip`>
/me gets depressed and goes home

[12:44:01.0000] <gsnedders>
Sure, but some of us don't know Java/C :P

[12:44:02.0000] <Philip`>
That's why you start collecting vast amounts of money to encourage a Java/C programmer to write it for you

[12:45:00.0000] <gsnedders>
Philip`: Or I just learn C :)

[12:46:00.0000] <gsnedders>
HTTP is _so_ odd.

[13:11:00.0000] <Dashiva>
/me wonders if JF realizes he's arguing that valid HTML is useless

[13:26:00.0000] <takkaria>
I love facebook

[13:26:01.0000] <takkaria>
in terms of availability, it's getting to be worse than myspace

[14:24:00.0000] <Hixie>
is w3.org down for anyone else?

[14:24:01.0000] <billmason>
Yes.

[14:25:00.0000] <Hixie>
i swear the w3.org services are down more often than whatwg.org

[14:26:00.0000] <hober>
w3.org is working for me

[14:26:01.0000] <hober>
at least, browsing/reading mailing list archives

[14:27:00.0000] <Hixie>
those are on lists.w3.org

[14:27:01.0000] <Hixie>
different server

[14:27:02.0000] <Philip`>
www.w3.org (in France) doesn't work for me

[14:27:03.0000] <Philip`>
Oh wait, maybe it does

[14:27:04.0000] <Philip`>
except not over the web, and traceroute goes funny

[14:28:00.0000] <Hixie>
funny how?

[14:29:00.0000] <Philip`>
It just gets to inria-nice.cssi.renater.fr and then starts printing *s

[14:29:01.0000] <Philip`>
which is not particularly funny

[14:30:00.0000] <Hixie>
ah

[14:30:01.0000] <Hixie>
that probably just means the next host is down

[14:30:02.0000] <Hixie>
(or blocking ICMP)

[14:30:03.0000] <smedero>
it is back up

[14:30:04.0000] <Hixie>
so it is

[14:30:05.0000] <Philip`>
ping worked, so it was probably just filtered

[14:30:06.0000] <Lachy>
/me is glad the spec is mirrored! :-)

[14:30:07.0000] <Hixie>
Lachy: sadly, w3.org's mailing list search is not :-(

[14:31:00.0000] <Lachy>
it's ok. My email on dreamhost is down again as well

[14:31:01.0000] <Hixie>
heh

[14:31:02.0000] <Lachy>
so I hadn't noticed

[14:32:00.0000] <Lachy>
any idea what photo I could use to represent video on the <video> slide?

[14:32:01.0000] <Hixie>
a dog on a skateboard

[14:32:02.0000] <Lachy>
I'm looking for a painting canvas, possibly sitting on an easel for use on the <canvas> slide

[14:32:03.0000] <takkaria>
cinefilm rolls, or something of that nature

[14:33:00.0000] <Philip`>
Lachy: A screenshot from http://www.youtube.com/watch?v=oHg5SJYRHA0

[14:33:01.0000] <Lachy>
Hixie, any particular reason, or did you just pick something random?

[14:33:02.0000] <Lachy>
LOL!

[14:33:03.0000] <Hixie>
that's like the canonical youtube video

[14:33:04.0000] <Hixie>
(the dog on the skateboard)

[14:33:05.0000] <Dashiva>
I thought the canonical youtube video was a cat doing nothing

[14:33:06.0000] <Hixie>
a screenshot from the rick roll video would be pretty funny to

[14:33:07.0000] <takkaria>
/me curses Philip`

[14:33:08.0000] <Lachy>
yeah, a photo of Rick Astley would be awesome

[14:34:00.0000] <Hixie>
o

[14:34:01.0000] <Dashiva>
So will the <video> part be a rickroll?

[14:35:00.0000] <Lachy>
I should get a copy of that video for the presentation

[14:35:01.0000] <Lachy>
though, I was intending to use the video of me skiing over a bus

[14:36:00.0000] <Lachy>
http://itsnotpossible.typepad.com/trashfan/astley1.jpg :-)

[14:36:01.0000] <Philip`>
Someone managed to make that video play during a lecture (on 'Introduction to Security') here recently, by sneaking in at night and booting the lecture theatre's computer into Linux to copy a script into Windows' Startup folder to play the video when the lecturer was logged in, which was a useful demonstration of the need for more careful security :-)

[14:38:00.0000] <Lachy>
LOL

[14:39:00.0000] <smedero>
when I'm looking to break up a text heavy presentation involving HTML, I often nab a photo from the visual HTML jokes flickr pool: http://www.flickr.com/groups/htmljokes/pool/page5/

[14:41:00.0000] <Lachy>
smedero, I've already got 2 LOL cats, a page of Corporate Ipsum, a Rick Astley photo (possibly a video too)

[14:41:01.0000] <smedero>
sounds like you've got it covered. :D

[14:43:00.0000] <Lachy>
I can't find a decent picture of a canvas on an easel with an appropriate licence. :-(

[14:43:01.0000] <Lachy>
maybe I should just go for some famous painting

[14:44:00.0000] <gsnedders_mibbit>
And me! Don't forget me!

[14:44:01.0000] <gsnedders_mibbit>
(of what's in the presentation)

[14:44:02.0000] <Lachy>
gsnedders_mibbit, you're still in there

[14:45:00.0000] <Lachy>
though I still have to fit the drunk MikeSmith photo in somewhere too

[14:47:00.0000] <Dashiva>
I wonder if there's a lorem ipsum equivalent for IDE designers, that uses lorem ipsum generator code

[14:48:00.0000] <hdh>
they can just ROT13 the IDE's source

[14:51:00.0000] <Lachy>
Dashiva, where would an IDE designer user lorem ipsum?

[14:54:00.0000] <Lachy>
I now have a painting of FSM on the canvas slide

[14:55:00.0000] <Lachy>
the one Michelangelo painted on the Sistine Chapel :-)

[14:55:01.0000] <Philip`>
Lachy: I hope you're bearing in mind that HTML is Serious Business

[14:56:00.0000] <Lachy>
Philip`, yeah, just gotta keep people interested so they'll listen to the serious stuff

[15:25:00.0000] <takkaria>
hmm, RB accuses Hixie of ignoring some browser implementers

[15:31:00.0000] <Philip`>
It seems true that there hasn't been much input from e.g. Lynx developers, or those people who make mobile browsers that don't work as well as Opera/WebKit, or AT developers who build on top of existing browsers

[15:31:01.0000] <takkaria>
tis hardly something Hixie can help though

[15:32:00.0000] <takkaria>
I dunno, maybe I should just killfile RB and let him be someone else's problem

[15:32:01.0000] <Lachy>
I like how, when we attempted to address RB's concerns, he raised the bar and bitched some more http://lists.w3.org/Archives/Public/public-html/2008May/0622.html

[15:33:00.0000] <Lachy>
like, complaining it was added to a later draft, instead of the revision the survey was about, kind of misses the point of asking for someting to be added

[15:34:00.0000] <takkaria>
the comments about consensus-building are interesting

[15:34:01.0000] <takkaria>
(in the WG questionnaire)

[15:34:02.0000] <Philip`>
Perhaps his concerns could be addressed by saying anything that isn't already mentioned in the spec (including the various placeholder notes) isn't going to be added in the future, so he doesn't have to worry about major features not being present yet

[15:36:00.0000] <takkaria>
http://html4all.org/mailman/archives/list_html4all.org/2008-May/000873.html terrifies me

[15:37:00.0000] <Lachy>
maybe I should start complaining more loudly. There are still features that I've asked for in the past which still haven't been added.

[15:37:01.0000] <Philip`>
takkaria: In terms of the web, that seems to be the wrong place to concentrate effort - approximately nobody is going to use such complex markup, and most of those people will get it wrong, so there's a much better return on investment by developing decent heuristics for pronouncing un-marked-up text

[15:39:00.0000] <takkaria>
Philip`: I knew that already, but maybe now you'll be quoted on the private html4all list

[15:39:01.0000] <takkaria>
RB is, I think, worse than Dmitry for these kinds of proposals now

[15:40:00.0000] <Philip`>
takkaria: That private list is only used for administrative details so I'm sure they wouldn't conduct the kind of private conversations there that they used to conduct on list⊙ho

[15:40:01.0000] <takkaria>
ah

[15:40:02.0000] <takkaria>
http://html4all.org/mailman/archives/list_html4all.org/2008-May/000856.html and the messages immidiately after it are similarly astronauty proposals

[15:40:03.0000] <takkaria>
brb, anyway, popping to tesco

[15:41:00.0000] <Lachy>
what is tesco?

[15:41:01.0000] <Philip`>
There's nothing necessarily bad about having markup for pronunciations of text, in certain contexts; it's just not appropriate for the web, which is what HTML5 is focussing on

[15:41:02.0000] <Philip`>
Lachy: http://en.wikipedia.org/wiki/Tesco

[15:41:03.0000] <Philip`>
(Big supermarket)

[15:42:00.0000] <Lachy>
wow, didn't realise they were so big

[15:42:01.0000] <Lachy>
surprising that I hadn't heard of them before

[15:43:00.0000] <Philip`>
"over £1 in every £7 [...] of UK retail sales is spent at Tesco"

[15:46:00.0000] <Lachy>
how can I represent offline web apps with a diagram?

[15:47:00.0000] <Philip`>
Gmail logo + unplugged network cable?

[15:47:01.0000] <Lachy>
hmm. A photo of an unplugged network cable would be nice

[15:47:02.0000] <Lachy>
that's something I can easily take a photo of myself, so even better

[15:55:00.0000] <annevk>
Lachy, you want http://www.flickr.com/photos/f8dy/347023948/

[15:59:00.0000] <Lachy>
ah, right. I think I saw a better one of those last week

[16:01:00.0000] <hober>
takkaria: my linguist wife said, re: the link that terrifies you, "uugggggggh. this guy needs to go back to Lojban land."

[16:01:01.0000] <Lachy>
wtf? I don't get this one. http://www.flickr.com/photos/jotape_es/384293705/in/pool-htmljokes

[16:02:00.0000] <takkaria>
hober: heh

[16:03:00.0000] <takkaria>
Lachy: http://en.wikipedia.org/wiki/Font_(disambiguation)

[16:03:01.0000] <annevk>
a whole new WG just for geolocation?

[16:04:00.0000] <Lachy>
takkaria, oh, so it's old word for fountain

[16:04:01.0000] <Lachy>
which still doesn't make that much sense, since I would call that a bubbler

[16:04:02.0000] <annevk>
imagine if we needed separate WGs for HTML syntax, SQL storage, persistent storage, offline/online events, offline application cache, sandboxing, "origin", etc.

[16:04:03.0000] <Philip`>
/me has never heard the term "bubbler" for fountains before

[16:05:00.0000] <Lachy>
http://en.wikipedia.org/wiki/Bubbler

[16:05:01.0000] <Philip`>
(Actually, I'm not sure I've heard the term "bubbler" ever used seriously for anything)

[16:05:02.0000] <Philip`>
Crazy foreigners :-p

[16:06:00.0000] <hober>
Philip`: heh. /me speaks en_US-x-boston

[16:14:00.0000] <Philip`>
Is it intentional that I get 13 errors when running the html5lib tests?

[16:15:00.0000] <Dashiva>
Hixie isn't shiny enough, he doesn't reflect the WG

[16:17:00.0000] <takkaria>
let's replace him with an inanimate piece of shiny silver

[16:18:00.0000] <Hixie>
i can tell i'm going to have fun when i read today's htmlwg mail

[16:18:01.0000] <Dashiva>
No, just the one by RB, Hixie :)

[16:18:02.0000] <Dashiva>
Then again, that one mail is plenty

[16:19:00.0000] <Dashiva>
Well, there's also the different accessibility guys contradicting each other left and right, but that's nothing new

[16:20:00.0000] <takkaria>
and MMM's latest post is amusing

[16:21:00.0000] <Dashiva>
"Do you expect UAs to implement acutal useful features? This is an outrage!"

[16:21:01.0000] <Dashiva>
(Not part of today's mail)

[16:21:02.0000] <annevk>
Hixie, do you know if sicking is just replying to the AC questions one by one?

[16:22:00.0000] <annevk>
Hixie, he had objections to all three, but he's only starting a new thread on one

[16:22:01.0000] <Hixie>
no idea

[16:22:02.0000] <annevk>
sigh

[16:22:03.0000] <Hixie>
i propose you ignore feedback that has had no reasoning

[16:22:04.0000] <Hixie>
and say so

[16:23:00.0000] <Philip`>
(Hmm, my html5lib now parses the spec (and discards the output) 15% faster than before, and doesn't fail any tests that it didn't fail yesterday)

[16:23:01.0000] <annevk>
sure, i just don't want to get bits and piece feedback for the next few weeks

[16:24:00.0000] <annevk>
Mozilla has surprisingly similar stalling tactics to Microsoft, though I guess it's not intentional

[16:24:01.0000] <Dashiva>
Maybe MS should file a patent

[16:24:02.0000] <annevk>
Actually, I know it's not intentional (at least until now), sorry about that

[16:25:00.0000] <Philip`>
Nobody apologises when accusing Microsoft of anything :-p

[16:27:00.0000] <Dashiva>
Philip`: Well, there's enough reasonable accusations to make that there's seldom a need to make unreasonable ones ;)

[16:28:00.0000] <annevk>
It's kind of frustrating though. sicking says he'll give feedback the same day. Then waits three days and provides only a third of what he promised.

[16:29:00.0000] <gavin_>
have you asked him about it?

[16:30:00.0000] <annevk>
Hixie did

[16:48:00.0000] <Lachy>
What picture should I use to represent software like html5lib and validator.nu?

[16:51:00.0000] <Lachy>
are there any other software projects I should mention besides those two?

[16:51:01.0000] <gsnedders>
Don't _ever_ forget to install Boot Camp 2.1 before upgrading to XP SP3 on Apple hardware.

[16:51:02.0000] <gsnedders>
/me has just spent hours digging himself out of that hole

[16:52:00.0000] <Dashiva>
Lachy: The test case projects?

[16:53:00.0000] <Lachy>
Dashiva, stuff that's relevant to develpers

[16:53:01.0000] <Lachy>
test cases are only relevant to implementers

[16:54:00.0000] <gsnedders>
Lachy: I'm not relevant to developers :P

[16:54:01.0000] <Dashiva>
I suppose most of the other language implementations are personal projects

[16:54:02.0000] <Lachy>
gsnedders, your photo is the main attraction ;-)

[16:55:00.0000] <gsnedders>
Lachy: :D

[16:55:01.0000] <takkaria>
Lachy: if you want html5lib as a library, then maybe a library...

[16:55:02.0000] <takkaria>
if you mean html5lib as a suite of tools, I dunno

[16:55:03.0000] <Lachy>
I'd prefer something that represents it as a development tool

[16:56:00.0000] <Dashiva>
A toolbox is a bit cliche :)

[16:56:01.0000] <takkaria>
bricks and mortar are a traditional kind of development tool icon

[16:56:02.0000] <takkaria>
at least where I come from

[16:56:03.0000] <Hixie>
lachy: http://www.buytikitorches.com/images/funnel-copper_1.jpg

[16:56:04.0000] <Hixie>
with random garbage pouring in the top

[16:56:05.0000] <Hixie>
and tiny beautiful trees coming out of the bottom

[16:57:00.0000] <Lachy>
Hixie, assume I don't have the time or skills to make such modifications

[16:57:01.0000] <Philip`>
http://www.accessibility.nl/files/images/tag-soup2-035.jpg

[16:57:02.0000] <Dashiva>
accessibility isn't a tag!

[16:58:00.0000] <Dashiva>
"Waiter, there're buzzwords in my tag soup"

[16:58:01.0000] <Lachy>
hah :-D

[16:58:02.0000] <Philip`>
When's Google going to make image search that actually works, and shows me tag soup instead of American 'football' players with tins of CHUNKY Savory Pot Roast and suchlike?

[16:59:00.0000] <Hixie>
Dashiva: i'm sure there are pages with <accessibility> tags!

[16:59:01.0000] <Dashiva>
<accessibility><html>...</html></accessibility> solves everything

[16:59:02.0000] <takkaria>
no, making alt mandatory solves everything


2008-05-28
[17:00:00.0000] <Hixie>
dashiva: though ironically, and possible presciently, it gets parsed as <html><accessibility>...</accessibility></html> :-)

[17:01:00.0000] <Dashiva>
I did have fun at one time with <invisible>

[17:01:01.0000] <Lachy>
Dashiva, you forgot the level="AAA" attribugte

[17:01:02.0000] <Lachy>
*attribute

[17:01:03.0000] <Lachy>
maybe this http://en.wikipedia.org/wiki/Image:HTML.svg

[17:01:04.0000] <Lachy>
if I edit it to use an HTML5 DOCTYPE

[17:02:00.0000] <Dashiva>
You put <invisible>The content of this page is hidden to protect the author's rights.</invisible> followed by 100 newlines at the top of the document

[17:02:01.0000] <Hixie>
heh

[17:03:00.0000] <Philip`>
http://www.newtacoma.com/ has some <hidden input="...">, which is peculiar

[17:03:01.0000] <Philip`>
Looks like a failed SEO attempt

[17:07:00.0000] <Hixie>
i still haven't finished dealing with e-mail

[17:07:01.0000] <Hixie>
sheesh

[17:07:02.0000] <Hixie>
today is an especially bad day for e-mail

[17:09:00.0000] <Dashiva>
There was one mail where someone claimed to read all the public-html mail, and also claimed there had never been a mail from you about @alt alternatives

[17:09:01.0000] <Hixie>
i saw that

[17:09:02.0000] <Hixie>
that was funny

[17:09:03.0000] <Dashiva>
I wonder if anyone's going to call him out on that

[17:09:04.0000] <takkaria>
/me notes hixie now has a zzz folder as well as an aaa-productivity folder :)

[17:10:00.0000] <takkaria>
I was going to, but I couldn't find the post that wasn't made

[17:11:00.0000] <Hixie>
i've had zzz for a while, though until recently it only had notes from myself to myself in there (which don't appear on the issues list)

[17:11:01.0000] <Hixie>
zzz is for things that i don't want to deal with until after we're in CR

[17:11:02.0000] <Hixie>
or at least near going to CR

[17:11:03.0000] <Dashiva>
takkaria: This one? http://lists.w3.org/Archives/Public/public-html/2008May/0073.html

[17:11:04.0000] <Philip`>
/me never noticed that part of the email, since he gave up reading it long before that point

[17:12:00.0000] <takkaria>
Dashiva: yeah, that one

[17:14:00.0000] <Philip`>
Since nobody raised any decent objections to that proposal, it is clearly perfect and should be implemented

[17:15:00.0000] <webben>
objections were raised in this channel (mainly along the lines of what attribute to use)

[17:15:01.0000] <webben>
IIRC

[17:16:00.0000] <annevk>
oh crap, it's 1:20

[17:16:01.0000] <Philip`>
webben: Ah, I'm not counting minor syntax details, because if I did count them then my statement would be proved false

[17:16:02.0000] <annevk>
/me blames HTML5

[17:16:03.0000] <webben>
using some other attribute to square the circle does seem to meet most of the raised concerns

[17:16:04.0000] <Dashiva>
At some point, we'll have to get a shed just tok keep all our bikesheds in

[17:16:05.0000] <Hixie>
HTML5 takes the blame for time advancing

[17:16:06.0000] <Hixie>
anything of the size of HTML5 is bound to have effects on the time space continuum

[17:17:00.0000] <webben>
Well, it does include a TIME element, doesn't it? ;)

[17:17:01.0000] <annevk>
in this case it was more having to deal with comments from shepers because of fear of the html5 dependency

[17:17:02.0000] <annevk>
and how it would badly affect vendors

[17:17:03.0000] <Hixie>
you'd think he'd let the vendors raise their own concerns

[17:17:04.0000] <annevk>
apparently he doesn't trust me when i say that vendors have requested that so i actually had to dig up an e-mail from maciej

[17:18:00.0000] <Dashiva>
The vendors need to start using that backroom channel nobody gets to see :)

[17:18:01.0000] <annevk>
would be nice if he read the thread about it first before starting a new one on the same subject

[17:19:00.0000] <Philip`>
Dashiva: We'd have to teach them the secret handshake first

[17:19:01.0000] <annevk>
Dashiva, moving everything to a less bureaucratic forum would be fine with me, but WHATWG doesn't have a PP yet

[17:20:00.0000] <takkaria>
Philip`: hmm, somehow lolzapolooza doesn't work when typed...

[17:20:01.0000] <Hixie>
yeah we really need a patent policy

[17:20:02.0000] <takkaria>
uh, that backfired

[17:20:03.0000] <Hixie>
annevk: get your lawyer people to write down what they would want a patent policy to cover, and mail it to me

[17:20:04.0000] <annevk>
PP involves lawyers, that's scary

[17:21:00.0000] <Hixie>
opera's lawyers aren't too bad

[17:22:00.0000] <Dashiva>
It was a bit disconcerting to get an office in the middle of legal. Suddenly there are suits all around you

[17:22:01.0000] <Dashiva>
Go one floor up, and it's shorts and t-shirts

[17:24:00.0000] <Philip`>
Several people where I work don't even wear socks or shoes, which doesn't seem a great idea to me given the stuff I know I've dropped on the floor

[17:25:00.0000] <annevk>
maybe you should tell them :)

[17:28:00.0000] <annevk>
(that was in reply to Philip`, not Hixie, fwiw)

[17:29:00.0000] <annevk>
/me tries to be extra careful now he mentioned lawyers

[17:31:00.0000] <roc>
dbaron: did you make any progress on the spurious menu command firing bug?

[17:32:00.0000] <dbaron>
roc, no... but did you mean to ask that in #developers ?

[17:32:01.0000] <roc>
yes I did!

[17:33:00.0000] <MikeSmith>
lastlog mike

[17:33:01.0000] <MikeSmith>
oops

[17:33:02.0000] <MikeSmith>
heh

[17:34:00.0000] <MikeSmith>
"<Lachy> though I still have to fit the drunk MikeSmith photo in somewhere too"

[17:34:01.0000] <Dashiva>
Checking if we talk behind your back? :)

[17:35:00.0000] <annevk>
hehe, we could use that as a gimmick

[17:35:01.0000] <annevk>
"works on your mobile: <picture>"

[17:35:02.0000] <MikeSmith>
Dashiva: making my IRC client show me if anybody was trying to ping me while I was asleep..

[17:41:00.0000] <Hixie>
annevk: btw, i could live with the hardcoded client-side limitation on the path component containing "..\"

[17:42:00.0000] <takkaria>
I accidentally send \0s to IRC far too often, I wonder if I can make my client stop me

[17:43:00.0000] <Hixie>
how??

[17:43:01.0000] <takkaria>
I'm not sure, I think I manage it by hitting the wrong button when switching tabs in irssi

[17:44:00.0000] <annevk>
yeah, it seemed sort of reasonable to me too, although it's amazingly ugly and i wonder if it'll ever fly

[17:44:01.0000] <Hixie>
why wouldn't it fly?

[17:44:02.0000] <Hixie>
html5 has far worth things

[17:44:03.0000] <Hixie>
worse

[17:44:04.0000] <annevk>
it may also be a slippery slope, not sure

[17:44:05.0000] <Hixie>
well if you find yourself going down the slope, just say no and remove it

[17:45:00.0000] <annevk>
i'd like to hear from the webkit guys if they feel this is needed

[17:45:01.0000] <othermaciej_>
if we feel what is needed?

[17:46:00.0000] <othermaciej_>
the crazy IIS workaround?

[17:46:01.0000] <othermaciej_>
I haven't studied it enough to know

[17:46:02.0000] <Hixie>
yeah

[17:46:03.0000] <Hixie>
not much to study really

[17:46:04.0000] <Dashiva>
When Doug says "commercial browser vendors", who does he mean?

[17:46:05.0000] <Lachy>
gsnedders, first draft of your slide http://lachy.id.au/temp/gsnedders.jpg :-)

[17:46:06.0000] <annevk>
(another option would be to simply disable the whole feature for IIS based on the Server response header from the OPTIONS request)

[17:47:00.0000] <annevk>
(but that seems worse)

[17:47:01.0000] <shepazu>
Dashiva: like, mobile vendors... Opera, for one :)

[17:47:02.0000] <othermaciej>
Dashiva: probably vendors of per-unit licensed mobile browsers

[17:47:03.0000] <othermaciej>
though I don't think Opera specifically objects to referencing HTML5

[17:48:00.0000] <Dashiva>
Yeah, I didn't see Opera objecting to that...

[17:48:01.0000] <othermaciej>
also I am not sure why charging a license fee is related to needing to market something, or how that relates to specification cross-references

[17:48:02.0000] <annevk>
I put it in the spec...

[17:48:03.0000] <annevk>
so far, no viable alternative has been propposed, so this is a non-issue

[17:52:00.0000] <annevk>
I'm also a bit confused with shepazu asking me to provide a reference for where implementors have requested that, and he, spokesman of the commercial browser vendors, has not done the same...

[17:52:01.0000] <annevk>
/me -> bed

[17:53:00.0000] <shepazu>
annevk: it's a simple question

[17:54:00.0000] <shepazu>
where do your requirements come from?

[17:54:01.0000] <annevk>
public-webapi

[17:54:02.0000] <Philip`>
Hmm, Google Code does autolinking between issues and commit messages, but it doesn't seem to do clever stuff like in Trac where you can close an issue via a commit message :-(

[17:54:03.0000] <Philip`>
(*autolinking of references of the form "issue 123" and "r123")

[17:55:00.0000] <annevk>
shepazu, I hope you followed the pointers I provided and snipped in your reply

[17:56:00.0000] <shepazu>
annevk: of course, but it's not much of a "requirement", and doesn't refute removal from HTML5

[17:58:00.0000] <Dashiva>
I'm a bit surprised that there are commercial browser vendors who don't plan on implementing HTML5

[17:59:00.0000] <annevk>
shepazu, nothing does, it's just that nobody has done that yet

[17:59:01.0000] <annevk>
shepazu, and because nobody has done that yet, there's no incentive to change references

[17:59:02.0000] <annevk>
shepazu, as the current references serve implementors best

[18:05:00.0000] <Philip`>
/me discovers the 'pv' tool, which is quite neat

[18:05:01.0000] <Philip`>
since I can use it to e.g. feed a big HTML file to html5lib, and get a progress bar and estimated-time-to-completion and everything

[18:09:00.0000] <Dashiva>
If Hixie added alt="_none" as the missing value to the spec today, what would the responses be? Discuss

[18:10:00.0000] <Philip`>
I'd complain that it doesn't degrade gracefully

[18:10:01.0000] <shepazu>
alt="_blank"

[18:10:02.0000] <Dashiva>
Would you? Would you really revive the whole debacle when it might be over?

[18:10:03.0000] <Philip`>
Nobody would want their page littered with "_none" tooltips

[18:10:04.0000] <Philip`>
Dashiva: Yes, since I'd like to have the best possible solution :-)

[18:11:00.0000] <annevk>
Dashiva, the whole point of the debacle is that people have different views; if one party just gives in to the other of course it would be over...

[18:11:01.0000] <annevk>
anyway, should really go to bed now

[18:12:00.0000] <Dashiva>
It's a dual-sided question, I suppose. Who would think it was a good solution, and who would realize it's a bad solution yet remain silent.

[18:13:00.0000] <Philip`>
Only crazy people would think it's a good solution ;-)

[18:14:00.0000] <Dashiva>
Yet people are suggesting it on this very day :)

[18:14:01.0000] <Philip`>
I never suggested that people are not crazy

[18:15:00.0000] <Dashiva>
But if you believe there are crazy people on the list, should you not feel obliged to attempt to rectify this?

[18:15:01.0000] <Philip`>
I don't feel any need to argue against people who are suggesting crazy things, only against the crazy things that are currently in the spec

[18:16:00.0000] <Philip`>
(unless it seems likely that the crazy suggestions are going to be put into the spec, in which case it'd be good to pre-emptively argue against the craziness)

[18:17:00.0000] <Dashiva>
Even though the crazy people are claiming a lot of bandwidth and other finite resources?

[18:18:00.0000] <Philip`>
"You are currently using 102 MB (1%) of your 6767 MB." - that's equivalent to, like, 5 pence of bandwidth? I'm not too concerned about that particular resource

[18:19:00.0000] <Philip`>
Arguing against crazy people won't stop the arguments - only arguing against the spec's craziness and getting that fixed will stop them

[18:19:01.0000] <Dashiva>
I meant mental bandwidth, e.g. how much of the mailing list discussion is about it :)

[18:20:00.0000] <Philip`>
(I think the spec's view on alt is crazy, since it forces information loss by making important images and lazy-author images indistinguishable)

[18:22:00.0000] <Hixie>
it would be an interesting exercise to have a spec editor just agree to every request

[18:22:01.0000] <Hixie>
and see what you end up with

[18:22:02.0000] <Dashiva>
Hixie: Remember, you aren't reflecting the WG

[18:22:03.0000] <Dashiva>
That means everyone!

[18:22:04.0000] <Hixie>
i hope i'm not!

[18:23:00.0000] <Philip`>
You should reflect the majority of the WG, by doing nothing whatsoever

[18:23:01.0000] <Hixie>
hah

[18:23:02.0000] <roc>
"alt" is supposed to be shown in tooltips now?

[18:23:03.0000] <roc>
I thought I won that battle 7 years ago

[18:23:04.0000] <Hixie>
the spec in fact says it is not

[18:24:00.0000] <Philip`>
roc: No, but pages need to degrade gracefully in IE6/7/etc

[18:25:00.0000] <Philip`>
Then again, you can just do <img src=... alt="_none" title=""> to avoid the tooltip, though then it'll look ugly while the page is loading and displays the image placeholder thing

[18:43:00.0000] <Hixie>
i really don't understand this obsession from shepazu and julian about not referencing html5

[18:44:00.0000] <shepazu>
"obsession"?

[18:44:01.0000] <shepazu>
I'm not obsessed, I just don't think it's a good idea to normatively reference something so early in the Rec track, with years to go

[18:45:00.0000] <Hixie>
btw john wanted me to remind you all that he did apologize for not remembering my post proposing to make alt="" mandatory

[18:45:01.0000] <Dashiva>
shepazu: It's even harder to reference implementations

[18:46:00.0000] <Hixie>
shepazu: why not?

[18:46:01.0000] <Hixie>
shepazu: it's by far the most detailed and complete definition so far

[18:46:02.0000] <Hixie>
shepazu: by all previous standards of quality, html5 is already far beyond REC

[18:46:03.0000] <Philip`>
/me assumes shepazu does not believe the timeline in the HTML WG charter, that suggests CR by 2008 Q3 and REC by 2010 Q3

[18:47:00.0000] <shepazu>
Hixie: not in terms of stability

[18:47:01.0000] <Hixie>
shepazu: REC doesn't mean jack with respect to stability, just look at CSS2

[18:47:02.0000] <Hixie>
shepazu: or HTML4

[18:47:03.0000] <shepazu>
Philip`: the editor doesn't believe it, why should I?

[18:47:04.0000] <Philip`>
shepazu: I think that's quite sensible :-)

[18:47:05.0000] <shepazu>
:)

[18:48:00.0000] <Dashiva>
I think more and more people are caring less and less about REC :)

[18:48:01.0000] <Hixie>
also it's not like the stuff anne is referencing can actually change

[18:48:02.0000] <Lachy>
damn, I really should finish reading the whole thread before responding :-(

[18:48:03.0000] <othermaciej>
it would be great to switch to a better reference if there were one, but currently there isn't

[18:48:04.0000] <Hixie>
since it's all just describing actual implementations

[18:48:05.0000] <othermaciej>
referencing a document that doesn't yet contain the right info does not seem like a wise move

[18:48:06.0000] <othermaciej>
if the goal is to improve progress

[18:48:07.0000] <shepazu>
Dashiva: I think your sample may not be representative :)

[18:49:00.0000] <Dashiva>
shepazu: Really? I'd think CSS2.1 and CSS3 would be elsewhere on the track if so

[18:49:01.0000] <shepazu>
I'm not proposing dereferencing it, I'm proposing allocating resources to redirect it 

[18:50:00.0000] <shepazu>
Dashiva: CSS is not a sane measure :)

[18:50:01.0000] <shepazu>
that's a good example of a bad example

[18:51:00.0000] <shepazu>
but that WG is getting its act together recently

[18:51:01.0000] <Dashiva>
Well, the other example being that HTML was left on the backburner for how many years? And javascript, to close the trinity.

[18:53:00.0000] <Hixie>
wow, http://www.w3.org/mid/OF7B780188.FAF5B9D6-ON85257456.00453CE0-86257456.0045F764⊙uic was totally ignored

[18:53:01.0000] <Hixie>
so much for caring about the blind

[18:55:00.0000] <Philip`>
Why does the lack of responses within 12 hours to an email that didn't raise any obvious questions indicate that people don't care about the blind?

[18:56:00.0000] <Hixie>
well, lots of other e-mails got replies

[18:56:01.0000] <Philip`>
That's not answering the question at all :-p

[18:57:00.0000] <Hixie>
e-mails that, just like that one, suggested allowing alt to be optional

[18:57:01.0000] <Hixie>
so why did they get replies and not this one?

[18:58:00.0000] <Philip`>
That one doesn't suggest allowing alt to be optional, as far as I can tell

[18:58:01.0000] <Hixie>
hey, nobody picked up JF on his false claim that html5 requires a DTD

[18:58:02.0000] <Philip`>
(I interpreted "DTD" as a typo for "doctype")

[18:58:03.0000] <Hixie>
ah, maybe

[18:58:04.0000] <Hixie>
he's saying that flickr is irrelevant to accessibility, basically

[18:59:00.0000] <Hixie>
(at least for the totally blind)

[18:59:01.0000] <Dashiva>
And that valid html is useless, since google is doing fine

[18:59:02.0000] <Philip`>
That email seems to just suggest that it's not particularly interesting what alt text Flickr uses, so it could use alt="" or alt="_none" or alt="14097651074651076.jpeg" or [nothing] and we should instead consider the effect that the conformance requirements will have on other web pages

[19:00:00.0000] <Hixie>
fair enough

[19:00:01.0000] <Philip`>
(*That email == the 0F7B one)

[19:00:02.0000] <Philip`>
(Uh, OF7B)

[19:00:03.0000] <Hixie>
(of course we do still have to consider flickr in non-blind cases, e.g. text browsers)

[19:00:04.0000] <Hixie>
(and other threads go into that in more detail)

[19:02:00.0000] <Hixie>
http://www.w3.org/mid/012301c8c056$7d03ca60$6d01a8c0⊙se is probably the clearest statement of john's opinion that i have seen so far

[19:02:01.0000] <Philip`>
"300 Multiple Choices" - that's far too many to choose from :-(

[19:03:00.0000] <Dashiva>
<movie title/>

[19:04:00.0000] <Hixie>
woot, i finally got to an e-mail actually discussing my proposal

[19:04:01.0000] <Dashiva>
Hixie: As I read it, he's got two concerns. One is that there must be some kind of presence, and the other like RB that he wants different image roles.

[19:04:02.0000] <Dashiva>
Anything I'm missing?

[19:04:03.0000] <Hixie>
not sure what you mean by either of those

[19:05:00.0000] <Philip`>
"Complete this sentence: This is ______: (A) orange (B) an elephant (C) madness (D) Greece"

[19:06:00.0000] <Dashiva>
Philip`: If you had said my name in that line, it would've been orange.

[19:07:00.0000] <Dashiva>
Hixie: As in, he wants there to be an explicit statement of missing alt data. And, like RB, he wants to subdivide the missing alt data images into several classes.

[19:07:01.0000] <Hixie>
Dashiva: ah. well that's what importantimage="" does.

[19:08:00.0000] <Hixie>
wish i could come up with a better name for it

[19:08:01.0000] <Philip`>
importantimage only subdivides into two classes

[19:08:02.0000] <Philip`>
which isn't quite the same

[19:09:00.0000] <Philip`>
(I'm not sure how any UA would make use of the extra subdivisions, though)

[19:10:00.0000] <Dashiva>
I wonder if <img src missingalt="personal photo"> would be acceptable, or if it would need alt="" (some value, not necessarily empty) as well

[19:11:00.0000] <Hixie>
Philip`: no it changes the alt="" attribute into something that categorises (freeform) the image

[19:12:00.0000] <Hixie>
Dashiva: my proposal is <img src alt="personal photo" important>

[19:12:01.0000] <Hixie>
well, src="..."

[19:13:00.0000] <Philip`>
Dashiva: <img alt="personal photo" almostmissingalt> conveys the same information but degrades better in existing UAs, and means you don't have to worry about what <img alt="foo" missingalt="bar"> means

[19:13:01.0000] <Philip`>
Hixie: Ah, I assumed it was just a boolean attribute

[19:13:02.0000] <Lachy>
Hixie, is the intended purpose of importantimage="" only for images where there is no good alt text, and whatever is there is a just a poor substitue, or for all images that are important content, regardless of their alt text quality?

[19:14:00.0000] <Hixie>
it's only for the images that right now can have alt="" omitted

[19:14:01.0000] <Hixie>
it says "the alt attribute isn't an alternative, it just tells you what kind of photo it is"

[19:14:02.0000] <Dashiva>
Philip`: I'm not sure if it's better or not, to give legacy browsers the same information-less alt text over and over. As in, I'm not sure either way.

[19:14:03.0000] <Lachy>
ok. Then that makes the name importantimage really bad.

[19:14:04.0000] <Hixie>
yes

[19:15:00.0000] <Hixie>
it allows UAs to change from rendering the image as "%s" where %s = alt, to rendering the image as "[IMAGE: %s]" where %s = alt and "[IMAGE: ... ]" is rendered in some UA-specific way

[19:15:01.0000] <Hixie>
different voice, different color, etc

[19:15:02.0000] <Hixie>
showing an icon

[19:15:03.0000] <Dashiva>
onlyimage, missingalt

[19:15:04.0000] <Philip`>
Dashiva: It's not informationless - it shows that the image is a personal photo, and not a graph or whatever, and it ensures legacy browsers won't pretend the image doesn't exist or display it as "[10496518746051746506.jpeg]"

[19:16:00.0000] <Dashiva>
Philip`: If the values are fine-grained enough

[19:16:01.0000] <Hixie>
<img src="..." alt-just-say-what-kind-of-image-it-is-and-doesn't-provide-a-replacement alt="photo">

[19:17:00.0000] <Hixie>
maybe that attribute name is a bit long

[19:17:01.0000] <Dashiva>
descriptive-alt :D

[19:17:02.0000] <Hixie>
it's not descriptive

[19:17:03.0000] <Hixie>
that's the problem!

[19:17:04.0000] <Hixie>
nondescript="" maybe

[19:17:05.0000] <Dashiva>
alternate-alt

[19:17:06.0000] <Philip`>
Seems more useful if it is descriptive, rather than just saying what kind it is

[19:17:07.0000] <Hixie>
the point is it can't be descriptive

[19:17:08.0000] <Hixie>
you don't know what the image is!

[19:20:00.0000] <Philip`>
You can use metadata like Flickr titles, which has a non-zero chance of describing the photo better than just saying "photo", and UAs can automatically detect and remove repetition if they've recently read the same text from the page title/heading/etc

[19:21:00.0000] <Hixie>
uh huh

[19:21:01.0000] <Philip`>
(That's much more plausible than them containing advanced image analysis algorithms :-p )

[19:22:00.0000] <Dashiva>
I think the feature is very different depending on how specific the alt-replacing value is

[19:22:01.0000] <Dashiva>
If it's just "image" or "photo" or "icon", compared to "personal photo" and "graph or chart"

[19:23:00.0000] <Dashiva>
The latter case borders on descriptive, but would it ever occur in practice?

[19:23:01.0000] <Hixie>
could be either

[19:23:02.0000] <Philip`>
Nobody's going to use this new attribute so it doesn't need to be particularly easy to type

[19:23:03.0000] <Dashiva>
There's a point

[19:24:00.0000] <Hixie>
(i'm not sure how you could end up with it being a graph or chart)

[19:24:01.0000] <Dashiva>
Hixie: If it's a middle management chart sharing photo site

[19:24:02.0000] <Hixie>
fair enough

[19:24:03.0000] <othermaciej>
how about calling the attribute noalt

[19:24:04.0000] <othermaciej>
that is at least brief

[19:24:05.0000] <Dashiva>
Or some kind of CMS where you know the usage of a subarea of the content

[19:24:06.0000] <othermaciej>
and all it would be is a conformance checker silencer

[19:25:00.0000] <Hixie>
having an attribute noalt="" which requires alt="" to be present would at a very minimum make us the butt of several jokes over the next few years.

[19:25:01.0000] <Lachy>
othermaciej, noalt would be somewhat confusing for authors because, in this case, it would still require authors to provide alt="something"

[19:25:02.0000] <othermaciej>
I don't think it should require alt to be present

[19:25:03.0000] <Hixie>
oh, then you're proposing something else

[19:25:04.0000] <Dashiva>
othermaciej: Then you're making a different proposal

[19:25:05.0000] <othermaciej>
yeah

[19:26:00.0000] <Hixie>
noalt="" as just an alternative to alt="" is just dumb imho

[19:26:01.0000] <Dashiva>
Do you plan <img src noalt="personal photo">?

[19:26:02.0000] <Hixie>
it doesn't solve anything

[19:26:03.0000] <Philip`>
othermaciej: If (alt XOR noalt) was required, that would incentivise inappropriate insertion of noalt just to silence conformance checkers, whereas requiring (alt OR (alt AND importantimage)) would remove that harmful incentive since you'd have to come up with alt text anyway

[19:26:04.0000] <Hixie>
and adds a number of new error conditions

[19:26:05.0000] <othermaciej>
I think the conformance checker insisting that you say <img src="foo.jpg" very-long-terrible-attribute-name alt="something useless"> then no one will respect the conformance checker

[19:26:06.0000] <Hixie>
that's why we're looking for a nice short name for very-long-terrible-attribute-name

[19:26:07.0000] <othermaciej>
Philip`: it would create an incentive to add alt="" just as now

[19:26:08.0000] <Philip`>
othermaciej: How could the conformance checker insist on that? All it knows is whether you've got an alt attribute or not

[19:27:00.0000] <Philip`>
and it has no idea whether you need the very-long-etc attribute or not

[19:27:01.0000] <Philip`>
so it has to assume you don't

[19:27:02.0000] <othermaciej>
in that case it wouldn't solve the problem of incentive to add alt="" incorrectly

[19:27:03.0000] <othermaciej>
though I will agree it does not give materially more or less incentive than now

[19:27:04.0000] <Lachy>
but why would authors bother using <img alt="something" importantimage=""> when they can just as easily use <img alt="something">, especially since most authors won't notice the difference unless they test in non-visual browsers

[19:27:05.0000] <Dashiva>
Are there use cases for this attribute in hand-written content (and not counting templateish html for generated content)

[19:28:00.0000] <Philip`>
Dashiva: I could hand-write <img src="random-image.cgi" alt="???">

[19:28:01.0000] <Dashiva>
That counts as templateish in my book

[19:28:02.0000] <othermaciej>
the only semi-sane reason to insist on mandatory alt, even if it is bad, is to make it harder to accidentally omit alt

[19:29:00.0000] <Dashiva>
Philip`: Basically, are there cases where you'd be writing the attribute name often

[19:29:01.0000] <othermaciej>
if the way to say "I omitted alt on purpose" is longer to type than alt="", then authors will either ignore it or use alt="" anyway, which was the original problem the spec wanted to solve

[19:30:00.0000] <Hixie>
so we need a solution that's 6 letters or less

[19:30:01.0000] <Hixie>
badalt?

[19:30:02.0000] <hdh>
not-alt(ernate) sounds better than noalt, being a bool

[19:30:03.0000] <Philip`>
Do they have to be ASCII letters?

[19:30:04.0000] <Hixie>
yes

[19:30:05.0000] <Philip`>
:-(

[19:30:06.0000] <Hixie>
:-P

[19:30:07.0000] <Dashiva>
alt0

[19:31:00.0000] <Dashiva>
Hmm... x-alt, maybe

[19:31:01.0000] <Philip`>
<img src=... alt="Some kind of photo, I don't know really" altsucks>

[19:32:00.0000] <Hixie>
the effect i would expect this attribute to have on visual (graphical) browsers is that when they have images disabled, they'd show icons only for images with this attribute set (and maybe also those without alt="" set at all)

[19:32:01.0000] <Hixie>
so maybe an attribute name that plays on that?

[19:33:00.0000] <Philip`>
And they'd show the alt text (but no icon) for other images?

[19:33:01.0000] <Hixie>
<img src="..." alt="photo" icon>  is confusing since the image isn't an icon...

[19:33:02.0000] <Hixie>
Philip`: right

[19:36:00.0000] <Philip`>
othermaciej: Hmm, I thought it was more about making sure conscientious authors are able to write conforming content - I don't see how we can do anything with lazy authors, since all the useful things we could do would be requiring the authors to give us more information, which they won't because they're lazy

[19:36:01.0000] <Philip`>
(Nearly all authors (except the most conscientious) will just write non-conforming content anyway, so we can't do anything to them)

[19:37:00.0000] <othermaciej>
Philip`: well, there's two different issues for conscientious authors

[19:38:00.0000] <othermaciej>
ok let me backtrack

[19:39:00.0000] <othermaciej>
is it better for an author who is serving a truly random unknown image to say <img alt="image" src="random.cgi" altsucks> or  <img alt="" src="random.cgi" altsucks> or <img src="random.cgi" noalt>?

[19:39:01.0000] <othermaciej>
(though perhaps arguably all of these may be worse than saying nothing)

[19:40:00.0000] <othermaciej>
I would argue that saying "image" is worse in current AT, since it is likely worse than the localized image indicator, and alt="" is worse in current AT because it will make many screen readers skip the image entirely

[19:40:01.0000] <Hixie>
with my proposed attribute alt="" would be required to be non-empty

[19:40:02.0000] <othermaciej>
future AT would presumably ignore alt="" on such images so it won't be a benefit in that case

[19:41:00.0000] <Philip`>
othermaciej: The first option seems to produce the best output in current UAs, since the second is often entirely skipped and the user will never know there was an image, and the third is sometimes read as "random.cgi" (which is uglier than "image") or as "image" (which is no better than "image")

[19:41:01.0000] <othermaciej>
but alt="image" may well still be worse than letting the AT do its normal image identification heuristic

[19:41:02.0000] <othermaciej>
Philip`: reading as "image" is better because it is localized

[19:41:03.0000] <othermaciej>
and can use a distinguished voice

[19:41:04.0000] <Philip`>
It seems hard to be much worse than ATs' normal image identification heuristics :-)

[19:42:00.0000] <Dashiva>
I think we all know existing AT is not quite optimal in many ways

[19:42:01.0000] <Hixie>
we're not targetting today's ATs

[19:42:02.0000] <Hixie>
(by definition)

[19:42:03.0000] <Hixie>
though it is important that contain degrade well in today's ATs

[19:42:04.0000] <Hixie>
which is the main argument against making alt="" optional

[19:42:05.0000] <othermaciej>
future ATs would not benefit from saying alt="image" and may well do worse than if it is not specified (unless they hardcode a big list of alt text that is too generic to be helpful)

[19:43:00.0000] <Philip`>
Future UAs can read <img alt="image" altsucks> in a distinguished voice, and it will be localised consistently with the rest of the page's textual content

[19:43:01.0000] <Dashiva>
Only image is spelled correctly

[19:43:02.0000] <Dashiva>
:)

[19:44:00.0000] <Dashiva>
As I see it, AT needs to cope with images missing alt. That's a fact of life. So if we give them a missing alt and the magic future-supported new attribute, they will do no worse than they would with most images, and they will do better in the future

[19:44:01.0000] <othermaciej>
<img alt="image"> would degrade worse than <img> with no alt attribute in at least some current AT (e.g. VoiceOver)

[19:44:02.0000] <Philip`>
othermaciej: In what ways is it worse?

[19:45:00.0000] <Philip`>
(At least by default, VO says "image" in the same voice as normal page content, if I remember correctly)

[19:45:01.0000] <othermaciej>
because it is not localized to the user's OS language

[19:45:02.0000] <Philip`>
/me has no idea how many people reconfigure those voices

[19:45:03.0000] <Philip`>
Is that much of a problem, given that the entire page is not localised to the user's OS language either?

[19:46:00.0000] <othermaciej>
(localizing to the page content is not the same; we don't translate all the menus to French if you visit a French web page on a German system)

[19:47:00.0000] <Philip`>
If you don't understand French, you might not understand it saying "image" in a French accent (or whatever the translation is), but you also wouldn't understand anything else on the page so the image part is a relatively trivial inconvenience

[19:48:00.0000] <Dashiva>
The utility of this attribute in the first place is rather minimal

[19:48:01.0000] <Hixie>
http://junkyard.damowmow.com/325 btw if people want to play around with various UAs

[19:48:02.0000] <Philip`>
You'd just want to be able to find the <img src=german-flag.gif alt=Deustch> and then you'd be happy

[19:49:00.0000] <Philip`>
Hixie: Try them in <a href>s too, since that gives different behaviour

[19:49:01.0000] <Dashiva>
But ironically enough, I have a design & usability exam in a few hours, so I gotta sleep. :)

[19:49:02.0000] <Hixie>
can't right now. may later.

[19:49:03.0000] <Hixie>
bbiab

[19:49:04.0000] <gavin_>
"image" is "image" in french

[19:51:00.0000] <Philip`>
Oh, bed is a good idea

[19:54:00.0000] <Philip`>
About attributes names: For consistency with <img src alt>, it ought to be a hard-to-understand few-letter abbreviation

[19:57:00.0000] <hdh>
mpa, missing-proper-alt, just not from one word

[19:58:00.0000] <Philip`>
/me fails to think of an expansion for "lol" that can fit in front of "cat(egory)"

[20:17:00.0000] <MikeSmith>
Lachy: you there?

[23:20:00.0000] <Dashiva>
Philip`: "lazy or lacking"

[00:53:00.0000] <MikeSmith>
Lachy: ping me when you back around

[00:53:01.0000] <MikeSmith>
please

[01:07:00.0000] <Lachy>
ping

[01:07:01.0000] <Lachy>
MikeSmith, 

[01:07:02.0000] <MikeSmith>
Lachy: hei

[01:07:03.0000] <MikeSmith>
wanted to ask if you had seen Thomas Roessler's HTML5 slides yet

[01:07:04.0000] <Lachy>
no

[01:08:00.0000] <MikeSmith>
http://www.owasp.org/index.php/AppSecEU08_HTML5

[01:08:01.0000] <MikeSmith>
PDF slides: http://www.w3.org/2008/Talks/0521-owasp-html5-tlr/0521-owasp-html5-tlr.pdf

[01:09:00.0000] <MikeSmith>
he has some fun images in there

[01:09:01.0000] <MikeSmith>
maybe some that you could repurpose

[01:14:00.0000] <Lachy>
nothing suitable in there

[01:14:01.0000] <Lachy>
I'm looking for a spirit level or something similar to represent validation

[01:14:02.0000] <Lachy>
and maybe a hammer or some other tool for html5lib

[01:41:00.0000] <hsivonen>
Lachy: Re: photo representing tools: http://flickr.com/photos/geishaboy500/100043823/

[01:44:00.0000] <Lachy>
hsivonen, nice!

[01:47:00.0000] <virtuelv>
heh, what's this supposed to mean? http://twitter.com/mollydotcom/statuses/821528080

[01:57:00.0000] <hsivonen>
whoa. lots of irc log about noalt

[01:58:00.0000] <annevk>
I skipped it

[01:58:01.0000] <annevk>
just like some of the e-mails :)

[01:59:00.0000] <othermaciej>
so can anyone help me find a good CC-licensed (or otherwise reusable) photo of a beaker or flask of acid (or some chemical that could pass for such)?

[01:59:01.0000] <othermaciej>
I could use a presentation picture for "acid tests"

[01:59:02.0000] <hsivonen>
fwiw, I agree with othermaciej that the only semi-sane reason for noalt is helping forgetful authors, but I think the Image Report already covers that case

[01:59:03.0000] <othermaciej>
I think the image report covers it too

[01:59:04.0000] <othermaciej>
hsivonen: although, there is value to a warning that you can drive down to 0

[01:59:05.0000] <othermaciej>
compared to a long list you must review on every change

[02:00:00.0000] <othermaciej>
(that's why WebKit builds with -Werror even though it causes us no end of pain in central builds for Mac OS X)

[02:00:01.0000] <Philip`>
hsivonen: Forgetful authors will forget to turn the image report on, or if they do turn it on then they'd forget to look at it

[02:01:00.0000] <hsivonen>
Philip`: will they forget to run a validator at all?

[02:01:01.0000] <hsivonen>
(that's what I forget occasionally when updating an existing page)

[02:02:00.0000] <Philip`>
(and it doesn't provide any immediate feedback like "you forgot to put alt on 3 images here and here and here", so it takes a reasonable amount of effort and concentration to work out where you've gone wrong)

[02:02:01.0000] <hsivonen>
(leading to comical stuff like validator documentation being invalid)

[02:02:02.0000] <hsivonen>
the Image Report has a group for missing alt that you can drive to zarro and links to source locations

[02:03:00.0000] <Philip`>
hsivonen: Not all people will forget to validate at all, and some of those who remember will fail to use the fancy options

[02:04:00.0000] <othermaciej>
hsivonen: but that's no good if you have cases where missing alt is the right thing

[02:04:01.0000] <Philip`>
hsivonen: It takes effort and concentration to find that group, since it looks the same as all the other groups of images

[02:04:02.0000] <othermaciej>
hsivonen: thus, noalt to tag such locations

[02:04:03.0000] <hsivonen>
othermaciej: yeah, that's a semi-sane use case for noalt

[02:05:00.0000] <Philip`>
<!-- #pragma warning(disable:missing-alt) --><img src=random.cgi>

[02:05:01.0000] <hsivonen>
Philip`: shouldn't authors maximize the effort and concentration on accessibility??????

[02:05:02.0000] <zcorpan_>
i don't see why missing alt is worse than alt='' when the right thing should be alt='something else'

[02:06:00.0000] <Hixie>
it's not

[02:06:01.0000] <Hixie>
it's better :-)

[02:06:02.0000] <zcorpan_>
both authors and tools emit alt='' out of cargo cult fasion

[02:06:03.0000] <Philip`>
hsivonen: Yes, but we should maximise the return for a given level of (typically quite low) effort

[02:07:00.0000] <hsivonen>
Philip`: sure ( ?????? means :-) )

[02:08:00.0000] <hsivonen>
what does XXX4 mean on tlr's slide of window props?

[02:10:00.0000] <MikeSmith>
hsivonen: referring to http://www.whatwg.org/specs/web-apps/current-work/#xxx4index

[02:10:01.0000] <MikeSmith>
I guess

[02:10:02.0000] <othermaciej>
it is surprisingly hard to find a good free picture of a beaker or flask of acid

[02:10:03.0000] <Philip`>
(Also the Image Report has way too many words in it, partly because the alt requirements are too hard to explain, and so people simplify to "alt is optional" (which isn't what we want since in almost all cases alt is required))

[02:11:00.0000] <hsivonen>
MikeSmith: thanks

[02:11:01.0000] <MikeSmith>
hsivonen: placeholder method name from Window interface

[02:11:02.0000] <MikeSmith>
hsivonen: np

[02:11:03.0000] <hsivonen>
sometimes the Web is so weird that I'm not sure if there's an implemented property named XXX4

[02:11:04.0000] <MikeSmith>
othermaciej: we should consider other image metaphors for acid

[02:12:00.0000] <MikeSmith>
blotter-paper, windowpane

[02:12:01.0000] <MikeSmith>
Sandox Labs ampule

[02:13:00.0000] <annevk>
lol, XXX4 is a reference to Web IDL for something that's not defined yet

[02:13:01.0000] <othermaciej>
MikeSmith: I don't want people to get the impression that this is an electric kool-aid acid text

[02:14:00.0000] <othermaciej>
now I want to add window.XXX4 to WebKit

[02:14:01.0000] <othermaciej>
but I need to figure out what is should do

[02:14:02.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/#xxx4index

[02:14:03.0000] <Hixie>
it's the index getter on Window

[02:15:00.0000] <othermaciej>
maybe this picture for "acid test": http://www.dkimages.com/discover/previews/967/65020227.JPG ?

[02:15:01.0000] <othermaciej>
or maybe this one looks friendlier: http://www.koh-development.com/Images/erlenmeyer_flask.jpg

[02:15:02.0000] <Hixie>
use the pictures on acidtests.org/images

[02:15:03.0000] <Hixie>
:-)

[02:15:04.0000] <MikeSmith>
othermaciej: that first one says "noxious odor" to me

[02:16:00.0000] <MikeSmith>
othermaciej: second one says, "beaker full of brandy"

[02:16:01.0000] <othermaciej>
how about this: http://z.about.com/d/chemistry/1/0/B/T/flask.jpg

[02:16:02.0000] <othermaciej>
does that one say, "this web stuff is totally scientific"?

[02:17:00.0000] <hsivonen>
tlr has an interesting photo choice for introducing the HTML WG

[02:17:01.0000] <othermaciej>
or this: http://www.koh-development.com/Images/filtration_flask.jpg

[02:18:00.0000] <annevk>
heh, that presentation is great

[02:18:01.0000] <annevk>
(the one from thomas)

[02:18:02.0000] <hsivonen>
(the photo introducing the WG is http://www.flickr.com/photos/lassmatazz/325437141 )

[02:19:00.0000] <annevk>
hsivonen, isn't that a photo about what will happen with registerContentHandler ?

[02:19:01.0000] <annevk>
hsivonen, web sites fighting over who will get the media type?

[02:20:00.0000] <hsivonen>
annevk: oh

[02:20:01.0000] <annevk>
hsivonen, could be your way as well I suppose

[02:20:02.0000] <annevk>
/me thought http://www.flickr.com/photos/nathansnider/497536082/ was introducing HTML5

[02:28:00.0000] <MikeSmith>
fwiw, while taking a look at the following page I came across an issue that seems to me to illustrate to authors one of unintended consequences of serving XHTML content as text/html

[02:28:01.0000] <MikeSmith>
page is this:

[02:29:00.0000] <MikeSmith>
http://www.w3.org/TR/xmlschema-1/

[02:29:01.0000] <MikeSmith>
I looked at it because David Storey from Opera pointed out a CSS mistake in it

[02:29:02.0000] <MikeSmith>
http://www.w3.org/TR/xmlschema-1/#c-selector-xpath

[02:29:03.0000] <MikeSmith>
the table there

[02:30:00.0000] <MikeSmith>
has negative text-indent set on it

[02:30:01.0000] <MikeSmith>
which causes the text to overflow the left edge of the table

[02:30:02.0000] <MikeSmith>
in Opera and in Webkit

[02:31:00.0000] <MikeSmith>
but in looking at that, the other thing I noticed was duplicate <a name>s with the same ID values all over the place

[02:32:00.0000] <MikeSmith>
in Web Inspector

[02:32:01.0000] <othermaciej>
blech

[02:34:00.0000] <MikeSmith>
http://www.w3.org/2008/05/dup-a-name.png

[02:34:01.0000] <MikeSmith>
so I filed a bug:

[02:34:02.0000] <MikeSmith>
https://bugs.webkit.org/show_bug.cgi?id=19282

[02:34:03.0000] <Philip`>
http://code.google.com/p/aza/source/browse/trunk/SocialHistory/SocialHistory.js - that detecting-visited-links thing is catching on

[02:34:04.0000] <MikeSmith>
knowing pretty much what the answer would

[02:35:00.0000] <Philip`>
"The best part is that this information leak will never be plugged because it’s a fundamental feature of the browser.", apparently

[02:35:01.0000] <annevk>
MikeSmith, it's an Appendix C violation

[02:35:02.0000] <MikeSmith>
Philip`: hmm, that's Aza Raskin's stuff

[02:36:00.0000] <Philip`>
MikeSmith: Indeed

[02:36:01.0000] <MikeSmith>
annevk: yeah. I guess it must have been generated from xmlspec source

[02:37:00.0000] <MikeSmith>
so maybe same problem will occur with anything generated from xmlspc

[02:37:01.0000] <annevk>
there are other specs with the same issue, yeah

[02:38:00.0000] <annevk>
or some issue like that anyway

[02:38:01.0000] <annevk>
W3C is just as much responsible for TAG soup as the rest of the Web :p

[02:38:02.0000] <MikeSmith>
heh

[02:39:00.0000] <MikeSmith>
well, particular end result of that problem, in Webkit at least, is duplicate IDs (and fragment identifiers) all the hell all over the place in the DOM

[02:40:00.0000] <annevk>
maybe you guys should deploy validator.nu ;)

[02:40:01.0000] <Philip`>
/me had the same problem when his spec-annotation code produced <span style="background:red">%s</span> and then %s = '' and then it was serialised as XML and interpreted as HTML, so he had to write an HTML serialiser instead

[02:41:00.0000] <Philip`>
The problems tend to be annoyingly subtle

[02:43:00.0000] <hsivonen>
annevk: validator.nu doesn't catch duplicate IDs in text/html at the moment

[02:46:00.0000] <annevk>
the specific issue is <a/> not duplicate IDs

[02:46:01.0000] <hsivonen>
oh ok

[02:48:00.0000] <Lachy>
I need one more picture for the How to Contribute slide, and I'm done!

[02:50:00.0000] <Lachy>
what's an animal that's typically known for working together in a community, or could otherwise convey teamwork?

[02:50:01.0000] <Philip`>
A human?

[02:50:02.0000] <Lachy>
Philip`, other than a human

[02:50:03.0000] <hsivonen>
Lachy: an ant

[02:50:04.0000] <hsivonen>
Lachy: but I read somewhere that ants are actually pretty chaotic

[02:51:00.0000] <Philip`>
Teamwork is often chaotic :-)

[02:51:01.0000] <MikeSmith>
Lachy: a jackal

[02:51:02.0000] <Lachy>
what's a jackal?

[02:51:03.0000] <hsivonen>
Lachy: a lone assassin

[02:52:00.0000] <MikeSmith>
pack of jackals

[02:52:01.0000] <hsivonen>
the lone gunmen?-)

[02:53:00.0000] <Lachy>
http://en.wikipedia.org/wiki/The_Lone_Gunmen ?

[02:53:01.0000] <Philip`>
http://www.flickr.com/photos/kasimetcalfe/118471837/

[02:54:00.0000] <hsivonen>
Lachy: http://en.wikipedia.org/wiki/The_Jackal_(fictional_character)

[02:54:01.0000] <Hixie>
i wouldn't liken this community to ants

[02:55:00.0000] <Lachy>
what about bees?

[02:55:01.0000] <Philip`>
But they're so cute

[02:56:00.0000] <Hixie>
nor bees

[02:56:01.0000] <Hixie>
sharks, maybe...

[02:56:02.0000] <roc>
piranhas

[02:57:00.0000] <Hixie>
yeah, or piranhas

[02:57:01.0000] <Hixie>
do they school?

[02:57:02.0000] <annevk>
http://upload.wikimedia.org/wikipedia/en/6/69/New_FOXHOUND.jpg

[02:57:03.0000] <roc>
"It is not rare to find individuals with one eye missing due to a previous attack."

[02:57:04.0000] <annevk>
though not everyone might get the reference

[02:59:00.0000] <Lachy>
annevk, I don't get it

[02:59:01.0000] <Lachy>
what about http://en.wikipedia.org/wiki/Image:School_of_reef_fish_at_Rapture_Reef%2C_French_Frigate_Shoals.jpg

[03:00:00.0000] <Lachy>
http://commons.wikimedia.org/wiki/Image:School_of_Chaetodon_ulietensis.jpg

[03:01:00.0000] <roc>
Piranhas seem more appropriate, for their tendency to mutilate and cannibalize each other

[03:02:00.0000] <MikeSmith>
biker gang

[03:03:00.0000] <Lachy>
roc, I could only find photos of individual piranhas

[03:03:01.0000] <hsivonen>
roc: we don't do that

[03:04:00.0000] <roc>
the alt-fest sounds pretty painful for everyone

[03:05:00.0000] <othermaciej>
how about beavers?

[03:05:01.0000] <othermaciej>
they cooperate in a way that doesn't exactly signal hive mind

[03:05:02.0000] <othermaciej>
and are productive

[03:05:03.0000] <Lachy>
I almost forgot, I need to find a higher resolution photo of the rosetta stone with an appropriate licence

[03:06:00.0000] <othermaciej>
I still haven't found good acid

[03:06:01.0000] <othermaciej>
a picture thereof I mean

[03:06:02.0000] <othermaciej>
not gonna complain about drug sourcing here

[03:10:00.0000] <hsivonen>
roc: I think the alt and the ARIA colon threads are making people careful to avoid discussing accessibility topics

[03:12:00.0000] <annevk>
Lachy, http://www.oceanlab.abdn.ac.uk/blog/wp-content/pw-tight-family-cri.jpg

[03:13:00.0000] <hsivonen>
annevk: are those orcas?

[03:14:00.0000] <annevk>
they're described as whales on the page

[03:16:00.0000] <annevk>
i found them on

[03:16:01.0000] <annevk>
http://www.oceanlab.abdn.ac.uk/blog/?p=236

[03:17:00.0000] <hsivonen>
should I make the XML side of Validator.nu match character encoding names according to the HTML5 rules?

[03:24:00.0000] <hsivonen>
are SVG WG meeting minutes public these days?

[03:26:00.0000] <Lachy>
This one will do http://flickr.com/photos/nathanphilpot/458372179/

[03:26:01.0000] <annevk>
should be, http://lists.w3.org/Archives/Public/public-svg-wg/latest

[03:27:00.0000] <hsivonen>
annevk: thanks. so many lists to choose from

[03:28:00.0000] <annevk>
hsivonen, I think that's their new list but they seem to be using the member-only one as well sometimes

[03:45:00.0000] <gsnedders>
Lachy: LOL

[03:46:00.0000] <gsnedders>
Lachy: (that's re: my slide)

[03:47:00.0000] <gsnedders>
Philip`: That commit took around six seconds off parsing HTML 5's source (as opposed to the postprocessed index)

[03:47:01.0000] <gsnedders>
Philip`: (on my local machine, taking it down to 11s)

[03:55:00.0000] <Lachy>
gsnedders, I changed it. I'm just going to say that funny bit

[03:55:01.0000] <Lachy>
I'm publishing the slides now

[04:03:00.0000] <Lachy>
http://lachy.id.au/dev/presentation/hands-on-html5/hands-on-html5.pdf

[04:04:00.0000] <Lachy>
keynotes slides are still uploading

[04:04:01.0000] <gsnedders>
31.4MB!?

[04:05:00.0000] <Lachy>
yeah. If I zip the PDF, then it only reduces it to 30.4

[04:05:01.0000] <Lachy>
the keynote slides are only 14

[04:06:00.0000] <Lachy>
http://lachy.id.au/dev/presentation/hands-on-html5/hands-on-html5.zip

[04:07:00.0000] <gsnedders>
Lachy: Where are you giving it, anyway?

[04:07:01.0000] <Hixie>
i see you agreed with my Gill Sans suggestion :-)

[04:08:00.0000] <Hixie>
i think your "solve real problems" picture is ironic

[04:08:01.0000] <Hixie>
and should have a "no" symbol over it :-)

[04:08:02.0000] <Hixie>
(circle with a slash)

[04:10:00.0000] <Lachy>
it's supposed to convey the concept of solving things, not be technically accurate :-)

[04:12:00.0000] <gsnedders>
Lachy: html5lib only has complete parsers in Python and Ruby, and has a incomplete tokeniser in C. There is no C++

[04:12:01.0000] <Dashiva>
It seems Lachy ignored all my comments *sob*

[04:12:02.0000] <Lachy>
ok, I'll fix that

[04:13:00.0000] <gsnedders>
ah. @media. Mols was trying to con me into going there

[04:13:01.0000] <Lachy>
gsnedders, I just changed it to "Python, Ruby, C"

[04:13:02.0000] <Dashiva>
Are there supposed to be two identical kitteh blog slides?

[04:14:00.0000] <Lachy>
they're not identical

[04:14:01.0000] <Lachy>
oh, I see. the PDF printed each individual step, rather than the whole slide

[04:14:02.0000] <Dashiva>
I can't tell the difference between the two first in the pdf.

[04:14:03.0000] <Lachy>
I'll fix that later

[04:14:04.0000] <Dashiva>
oh

[04:22:00.0000] <hsivonen>
Hixie: do you intend the HTML5 encoding matching to be normative over XHTML5?

[04:27:00.0000] <hsivonen>
Lachy: "C++" under "html5lib"?

[04:28:00.0000] <Hixie>
gsnedders: is she the one who is working with microsoft?

[04:29:00.0000] <Hixie>
or is this another molly

[04:29:01.0000] <Hixie>
i get them all so confused

[04:29:02.0000] <Hixie>
hsivonen: um

[04:29:03.0000] <Hixie>
hsivonen: elabroate?

[04:30:00.0000] <hsivonen>
Hixie: the step where punctuation is ignored when finding a decoder to instantiate

[04:30:01.0000] <hsivonen>
Hixie: is that supposed to apply Web platform wide

[04:30:02.0000] <hsivonen>
to XML, CSS, text/plain, etc.?

[04:30:03.0000] <gsnedders>
Hixie: yes, that Molly

[04:30:04.0000] <gsnedders>
Hixie: as in "molly" on IRC :P

[04:31:00.0000] <gsnedders>
Hixie: and molly.com

[04:31:01.0000] <Hixie>
hsivonen: oh. dunno. what do UAs do? Does XML define anything?

[04:31:02.0000] <Hixie>
or CSS?

[04:31:03.0000] <hsivonen>
Hixie: I thought you'd have it figured out :-)

[04:32:00.0000] <Hixie>
i just wrote the html5 spec, not the others :-P

[04:32:01.0000] <Hixie>
it probably should apply platform-wide, but the validator should complain about anything not matching the real names in any case

[04:32:02.0000] <hsivonen>
ok

[04:37:00.0000] <Lachy>
hsivonen, yeah, I know. I looked at the html5lib repo and saw python, ruby and C, but I accidentally wrote C++ instead.

[04:43:00.0000] <hsivonen>
I think I'll just let the punctuation weirdness fall under the "not the preferred name" error

[04:46:00.0000] <annevk>
"Who thinks they know more about it than we do?" is slightly arrogant

[04:46:01.0000] <annevk>
and makes me want to come over and make fun of you :p

[04:46:02.0000] <hsivonen>
/me agrees it is arrogant

[04:48:00.0000] <Dashiva>
Remove "thinks they"

[04:48:01.0000] <Dashiva>
And then they can see if it's anne before mocking them

[04:58:00.0000] <annevk>
hsivonen, I think the SVG WG is going to proppose using an XML parser in text/html in some way

[04:58:01.0000] <Lachy>
hsivonen, annevk, what should I use instead?

[04:58:02.0000] <annevk>
(glancing over their minutes)

[04:58:03.0000] <annevk>
Lachy, just drop it

[04:58:04.0000] <Lachy>
ok

[04:59:00.0000] <Dashiva>
annevk: That'll be fun

[04:59:01.0000] <hsivonen>
annevk: I'm eager to see their implementation

[05:00:00.0000] <hsivonen>
annevk: also, I'm eager to see an explanation how it solves a Real Problem

[05:01:00.0000] <Hixie>
annevk: if they manage to find a way to do that while addressing the constaints we have to work with, that'd be great!

[05:01:01.0000] <Hixie>
i totally failed to do so

[05:01:02.0000] <othermaciej>
having an implementation is merely an implementation issue

[05:02:00.0000] <othermaciej>
no need to be concerned with that sort of thing

[05:02:01.0000] <annevk>
Hixie, I'm personally not that intrigued by the idea of doubling the complexity of the HTML parser, but we'll see

[05:03:00.0000] <othermaciej>
annevk: obviously a worthwhile tradeoff for the benefit of dracanoian error handling in text/html, plus the all-important ability to cut & paste source fragments into illustration tools

[05:03:01.0000] <Hixie>
if they seriously find a way to do it while not violating the constratints, i'm definitely interested

[05:03:02.0000] <Hixie>
however, i'm 99% sure that's not possible, so that may be a moot point

[05:04:00.0000] <Hixie>
i wonder if i wrote down the constraints anywhere convenient

[05:04:01.0000] <Hixie>
other than irc logs...

[05:05:00.0000] <annevk>
I think they're scattered around on varoius wiki pages, e-mail archives and IRC

[05:05:01.0000] <Hixie>
pity

[05:07:00.0000] <annevk>
given  that including a new vocabulary is something we don't expect to happen often it's probably not much of an issue to repeat them now and then, though having them somewhere on a wiki would be preferred of course

[05:14:00.0000] <Hixie>
http://krijnhoetmer.nl/irc-logs/whatwg/20080417 has the constraints

[05:19:00.0000] <zcorpan>
Hixie: could you click on one or more of those yellow boxes so i know where to look?

[05:22:00.0000] <zcorpan>
/me found

[05:23:00.0000] <annevk>
starts about here: http://krijnhoetmer.nl/irc-logs/whatwg/20080417#l-238

[05:26:00.0000] <Lachy>
I've fixed all the issues with the slides and uploaded new copies

[05:27:00.0000] <Hixie>
http://wiki.whatwg.org/wiki/Constraints_for_New_Vocabularies

[05:28:00.0000] <annevk>
/me downloads a new copy

[05:28:01.0000] <gsnedders>
Someone will be happy about <http://hg.gsnedders.com/http-parsing/rev/dc816eb2bc3a> :P

[05:30:00.0000] <annevk>
Lachy, it's not too clear on where people should go to join this stuff, e.g. whatwg.org / w3.org/html

[05:32:00.0000] <Hixie>
well crap. i can't find why i added rel=contact.

[05:33:00.0000] <hsivonen>
Hixie: XFN?

[05:33:01.0000] <Hixie>
nope, that's the problem

[05:33:02.0000] <Hixie>
it clashes with xfn

[05:33:03.0000] <Hixie>
i need to work out whether to nuke it altogether, or rename it

[05:34:00.0000] <hsivonen>
I've finally seen a reasonable use case for a subset of XFN

[05:34:01.0000] <annevk>
hmm, it's not in http://www.w3.org/TR/html4/types.html#type-links

[05:34:02.0000] <annevk>
though I remember using it way before XFN

[05:34:03.0000] <hsivonen>
social network portability by scraping XFN+hCard-enabled profile pages

[05:36:00.0000] <Hixie>
oh well, i guess i'll comment it out

[05:37:00.0000] <Hixie>
oh wait, i see why i have it

[05:37:01.0000] <Hixie>
it's the 8th mode used rel= value

[05:37:02.0000] <annevk>
s/mode/most/

[05:39:00.0000] <hsivonen>
does WordPress emit it on blogrolls or something?

[05:39:01.0000] <Lachy>
annevk, I added this as the second last slide just before the credits http://lachy.id.au/temp/links.jpg

[05:40:00.0000] <Lachy>
is that ok?

[05:40:01.0000] <Lachy>
or should I left align it like the others?

[05:40:02.0000] <Lachy>
or center align the other links for blog, etc.

[05:41:00.0000] <annevk>
that's great

[05:41:01.0000] <annevk>
though'd put it as the last slide

[05:42:00.0000] <Lachy>
ok

[05:48:00.0000] <hsivonen>
hmm. WAI really like to use "Best Practices" instead of "Guide" or "Primer"

[05:51:00.0000] <Lachy>
those changes have been made and are uploading now. if anyone finds any more problems with the slides, email me and I'll try and fix them tonight or tomorrow morning. 

[05:51:01.0000] <Lachy>
gotta go. bye

[05:54:00.0000] <gsnedders>
annevk: is there any way to get something like responseHTML?

[05:56:00.0000] <annevk>
it's called responseXML

[05:57:00.0000] <gsnedders>
annevk: Without using XML? :P

[05:57:01.0000] <annevk>
responseXML is a misnomer for responseDocument

[05:57:02.0000] <gsnedders>
what does IE do if it gets something for XHR that's app/xhtml+xml?

[05:58:00.0000] <annevk>
I forgot

[05:58:01.0000] <gsnedders>
annevk: See, I'm trying to work from XHR1 — that makes it clear that responseXML is XML :P

[06:00:00.0000] <gsnedders>
bbiab

[06:04:00.0000] <Hixie>
ah, our platform is so quirky

[06:04:01.0000] <Hixie>
responseXML contains HTML, innerHTML contains XML...

[06:05:00.0000] <annevk>
and no vendor lock in

[06:05:01.0000] <annevk>
it's amazing that it survives

[06:06:00.0000] <annevk>
wait, in retrospect, that's contradictory :)

[06:10:00.0000] <Hixie>
the img-alt folder has risen to the top of the list of folders sorted by number of e-mails, ignoring folders that i am intentionally delaying handling of

[06:10:01.0000] <Hixie>
(e.g. wf2, rendering, aria)

[06:11:00.0000] <Hixie>
dom-focus is the next biggest non-delayed folder

[06:11:01.0000] <Hixie>
and that's mostly about tabindex and friends

[06:11:02.0000] <Hixie>
no idea what to do with that...

[06:12:00.0000] <annevk>
I thought there was a proposal for having some kind of focus model

[06:12:01.0000] <Hixie>
i guess i'll find out when i read the folder

[06:13:00.0000] <annevk>
Hixie, use 'set' for unordered and 'list' for ordered?

[06:13:01.0000] <Hixie>
/me points at the topic

[06:13:02.0000] <Hixie>
:-P

[06:13:03.0000] <hsivonen>
Hixie: tabindex is considered part of the aria folder these days, isn't it?

[06:13:04.0000] <Hixie>
aria is about what you report to the AT, not what the non-AT-user sees

[06:14:00.0000] <annevk>
Hixie, meh, that's what Python does and it's quite understandable :)

[06:14:01.0000] <Hixie>
:-)

[06:16:00.0000] <Hixie>
actually it seems i've already resolved tabindex="" in the spec

[06:16:01.0000] <Hixie>
these comments are just minor editorial things, mostly

[06:16:02.0000] <Hixie>
or requests for new features

[06:18:00.0000] <annevk>
the e-mail I thought of is titled "Focus management" from May 2007

[06:21:00.0000] <Hixie>
yeah that's just new ideas and fixes

[06:22:00.0000] <hsivonen>
hmm. does one become an implementor by writing a Cocoa app that embeds WebKit?

[06:22:01.0000] <Hixie>
how to do notifications (toasts, bouncing the dock icon, etc) is also in this folder

[06:22:02.0000] <Hixie>
i wonder how to do that without opening it up to abuse

[06:23:00.0000] <Hixie>
hsivonen: if you just use it as a black box and write pages for it, you become an author and likely an unimportant one, since that's not an interop case)

[06:23:01.0000] <annevk>
hsivonen, yeah...

[06:23:02.0000] <Hixie>
hsivonen: if you write a browser with it, you become an implementor

[06:24:00.0000] <Hixie>
hsivonen: if you actually hack on webkit itself, you are an implementor.

[06:24:01.0000] <annevk>
/me wonders if we're investing too much time discussing pointless e-mails

[06:25:00.0000] <annevk>
probably not

[06:30:00.0000] <Hixie>
stress relief :-)

[06:30:01.0000] <Hixie>
ok bed time

[06:30:02.0000] <Hixie>
nn

[06:30:03.0000] <hsivonen>
nn

[06:38:00.0000] <annevk>
g'night

[06:41:00.0000] <Philip`>
/me sees http://lists.w3.org/Archives/Public/public-html-diffs/

[06:41:01.0000] <hsivonen>
who is "TZ" in http://lists.w3.org/Archives/Public/public-svg-wg/2008AprJun/0057.html ?

[06:42:00.0000] <Philip`>
/me assumes it's a bad list to subscribe to, if it's going to get ~2.5MB diff-marked emails for every commit

[06:43:00.0000] <hsivonen>
the minuted claims about Gecko's parsing behavior don't look right

[06:44:00.0000] <annevk>
Tony Zlatinski I think

[06:45:00.0000] <annevk>
(samsung)

[06:45:01.0000] <hsivonen>
annevk: thanks

[06:47:00.0000] <MikeSmith>
Philip`: about the diffs list, yeah, you probably would not want to subscribe to thaat

[06:47:01.0000] <MikeSmith>
but it may be moot

[06:47:02.0000] <MikeSmith>
for the time being at least

[06:47:03.0000] <roc>
hum

[06:48:00.0000] <MikeSmith>
because I've not actually been able to deliver to it

[06:48:01.0000] <roc>
what Opera developer is in favour of switching to XML parsing in the middle of an HTML document

[06:48:02.0000] <MikeSmith>
not able to deliver the diffs to it (because I'm running into delivery issues with the W3C mail server)

[06:48:03.0000] <Philip`>
Do they also want to switch back to an HTML parser for doing inline HTML in SVG?

[06:49:00.0000] <roc>
I can tell you for sure that we do not switch between our HTML parser and our XML parser

[06:50:00.0000] <annevk>
I think their idea is having something like the <style> or <script> element for XML

[06:51:00.0000] <Philip`>
Like IE's <xml>?

[06:52:00.0000] <annevk>
I guess so, yes

[06:53:00.0000] <gsnedders_mibbit>
they've sent anything yet?

[06:53:01.0000] <annevk>
I believe the aforementioned Tony is tasked with writing a proposal

[06:53:02.0000] <Philip`>
/me wonders how many of their concerns would be satisfied by using the spec's old SVG-in-HTML parsing behaviour, but with conformance requirements such that any conforming SVG-in-HTML is trivially copy-and-pasteable into well-formed XML

[06:53:03.0000] <annevk>
From their minutes, anyway

[06:54:00.0000] <shepazu>
Philip`: that was indeed my initial thought, and it's possible that that's a way forward

[06:56:00.0000] <shepazu>
Philip`: but that's a rather big long road to walk (with roadblocks and minefields), and while it's worth looking into, a certain time pressure's been put on the SVG WG (the looming threat of "well, if they don't propose something soon, we'll put this back in")

[06:57:00.0000] <roc>
oh boy

[06:58:00.0000] <Philip`>
shepazu: XML-style error handling seems likely to face far more roadblocks and minefields than tightening the conformance requirements would

[06:58:01.0000] <shepazu>
note that that's not why we're looking at this alternative... Samsung has direct implementation experience that we're drawing from

[06:58:02.0000] <roc>
just propose something and the whatwgistas will debug it for you

[06:59:00.0000] <shepazu>
roc, I've been proposing something for months, and that's not the case.... it's met with unreasonable resistence

[06:59:01.0000] <shepazu>
resistance

[06:59:02.0000] <annevk>
well, iirc, it didn't meet the constraints: http://wiki.whatwg.org/wiki/Constraints_for_New_Vocabularies

[06:59:03.0000] <shepazu>
... case in point.

[06:59:04.0000] <annevk>
anyway, back to merging xhr1 / xhr2

[07:00:00.0000] <roc>
shepazu: the constraints are unreasonable?

[07:01:00.0000] <shepazu>
roc: not necessarily, but the interpretation seems to be

[07:02:00.0000] <shepazu>
I don't see how my proposal fails any of those constraints

[07:02:01.0000] <Philip`>
shepazu: Do you have a link to your proposal?

[07:02:02.0000] <roc>
I'm not well enough informed to comment on the substance

[07:02:03.0000] <roc>
unfortunately

[07:03:00.0000] <shepazu>
http://wiki.whatwg.org/wiki/Extensions#Proposal_2:_Extensibility_Element

[07:04:00.0000] <shepazu>
Philip`: but the bit about adding an element is not necessary to the proposal (merely to allowing fallback content, which I also think is necessary)

[07:05:00.0000] <Philip`>
shepazu: That doesn't "Fail gracefully in the face of ignorant author copy/paste", because someone will copy-and-paste the <ext> into the top of their document and test it in a legacy UA and it will break nastily in <ext>-supporting UAs

[07:05:01.0000] <shepazu>
Philip`: explain how?

[07:06:00.0000] <Philip`>
/me realises he didn't read the whole of the proposal

[07:08:00.0000] <roc>
interesting

[07:08:01.0000] <Philip`>
shepazu: How they would copy-and-paste it, or how it would break nastily?

[07:08:02.0000] <shepazu>
the latter

[07:09:00.0000] <Philip`>
Depending on how the error handling is defined, it would probably result in parts of the page disappearing (e.g. if the HTML content after an unclosed <ext><svg>... is misinterpreted as SVG)

[07:10:00.0000] <shepazu>
so, there are 2 parts to my proposal: one, the <ext> element (and the fallback element child), and the parsing aspects (which say, essentially, parse this in HTML5 but with XML constraints)

[07:10:01.0000] <shepazu>
Philip`: no, the parser should be able to handle that, just as it does overlapping elements, etc., in the current HTML5 spec

[07:10:02.0000] <Philip`>
/me has to go for an hour

[07:11:00.0000] <roc>
I have to go to sleep

[07:11:01.0000] <roc>
but it seems like this wiki page is where energy should be focused

[07:11:02.0000] <roc>
for now

[07:12:00.0000] <roc>
regarding "Note also that the fallback idea doesn't work."

[07:15:00.0000] <roc>
the main danger seems to be that SVG script nodes would be executed by legacy UAs

[07:15:01.0000] <shepazu>
interesting

[07:15:02.0000] <roc>
which isn't that big a deal IMHO

[07:16:00.0000] <shepazu>
without an SVG context, most of those scripts would merely fail silently, no?

[07:16:01.0000] <roc>
probably

[07:16:02.0000] <shepazu>
still, worth noting

[07:16:03.0000] <roc>
but no approach to this problem is going to be able to avoid this

[07:16:04.0000] <shepazu>
right

[07:17:00.0000] <shepazu>
we want to concentrate the most effort on enabling good functionality, not recovering funky stuff

[07:17:01.0000] <roc>
some weird stuff might happen with <font> tags too

[07:17:02.0000] <shepazu>
though of course we have to consider failure cases, as well...

[07:17:03.0000] <roc>
but I think you should probably ditch the <fallback> idea since it's a bit leaky

[07:17:04.0000] <roc>
and it seems that resource constraints on speccing this out are an issue

[07:18:00.0000] <shepazu>
roc: I'd like to talk more when you have more time

[07:18:01.0000] <shepazu>
not sure what you mean by "leaky"

[07:18:02.0000] <roc>
I mean, no approach to this can be foolproof

[07:18:03.0000] <roc>
except, perhaps, putting all the SVG markup in an attribute

[07:19:00.0000] <roc>
which I do not recommend :-)

[07:19:01.0000] <shepazu>
heh

[07:20:00.0000] <annevk>
<iframe doc=... type=text/xml></iframe> ...

[07:20:01.0000] <roc>
exactly

[07:20:02.0000] <annevk>
we might get that anyway btw for sandboxing

[07:20:03.0000] <roc>
right

[07:21:00.0000] <roc>
I don't think of that as a satisfactory solution here though :-)

[07:21:01.0000] <annevk>
agreed

[07:21:02.0000] <roc>
shepazu: in my ignorance, it seems to me that your proposal could be made to work, i.e. that Hixie's comments could be addressed

[07:22:00.0000] <shepazu>
roc: well, at least you admit that you're ignorant ;)  humility is an important personal attribute :)

[07:22:01.0000] <shepazu>
jk

[07:22:02.0000] <roc>
as long as things are defined so that any malformed content, such an unexpected tag end, bails out of <ext> immediately

[07:22:03.0000] <shepazu>
roc: that was my intent

[07:22:04.0000] <roc>
you're absolutely right :-)

[07:23:00.0000] <shepazu>
:)

[07:24:00.0000] <roc>
so one comment that would really help to address is "What elements/attributes are known?"

[07:24:01.0000] <roc>
and probably "What are you considering an error in tree construction?"

[07:24:02.0000] <roc>
since for example

[07:24:03.0000] <roc>
given <p>foo bar <ext>baz qux    <p>foo bar

[07:25:00.0000] <roc>
it would be good to bail out of <ext> when we hit the next <p>

[07:25:01.0000] <roc>
I should also point out that implementing full-on XML parsing inside the HTML parser could be painful, despite what TZ may think

[07:26:00.0000] <roc>
 /could be/would be/

[07:26:01.0000] <annevk>
this bailing out would require reparsing in some way to not lose content

[07:27:00.0000] <roc>
in which case?

[07:27:01.0000] <annevk>
that's probably not acceptable

[07:27:02.0000] <annevk>
but it's not really clear to me how the XML parsing works

[07:28:00.0000] <annevk>
(when nested inside text/html)

[07:28:01.0000] <annevk>
say you have <p><ext>X</p><p>Y</p>

[07:28:02.0000] <annevk>
what would happen?

[07:29:00.0000] <roc>
I would leave <ext> at the </p>

[07:29:01.0000] <roc>
the XML parser would parse "X" as a DOM text node

[07:29:02.0000] <roc>
then you would leave <ext> mode and parse </p> as HTML

[07:29:03.0000] <roc>
I guess retokenization would be necessary

[07:30:00.0000] <roc>
but this is the discussion that you and shepazu should be having, instead of the pistols at dawn impression I'm getting

[07:30:01.0000] <annevk>
<p><ext><x>...</p>

[07:30:02.0000] <shepazu>
probably

[07:30:03.0000] <annevk>
roc, heh

[07:31:00.0000] <annevk>
there's also issues with decoding, incorrect byte sequences, etc., incremental display

[07:31:01.0000] <shepazu>
roc: I believe I've been acting in good faith... this is something I've been asking for for a while

[07:31:02.0000] <roc>
I'm not accusing anyone of bad faith

[07:31:03.0000] <annevk>
someone putting <ext> at the start of a well-formed fragment and changing the semantics of all the elements it includes

[07:32:00.0000] <shepazu>
annevk: that's an interesting point

[07:32:01.0000] <annevk>
(elements might end up in the wrong namespae, <image> would no longer be treated as <img>, etc.)

[07:32:02.0000] <roc>
incremental display?

[07:32:03.0000] <shepazu>
roc: progressive rendering

[07:32:04.0000] <shepazu>
(I assume)

[07:33:00.0000] <annevk>
yeah, depending on specifics that could be affected

[07:33:01.0000] <annevk>
the fallback proposal is pretty badly for that

[07:33:02.0000] <shepazu>
annevk: it's possible, but not a necessity

[07:34:00.0000] <shepazu>
annevk: the fallback proposal wouldn't have any effect there

[07:34:01.0000] <annevk>
sure, it's not really clear to me what exactly the use cases are people envision with this

[07:34:02.0000] <annevk>
shepazu, i think it would

[07:34:03.0000] <annevk>
shepazu, document sitll downloading, you're displaying some XML, at the end you suddenly hit a parse error...

[07:35:00.0000] <roc>
I think the fallback proposal should be tossed over the side, it is a lot less important than the problem we're trying to solve here

[07:38:00.0000] <shepazu>
roc: as an author, I disagree that it's unimportant... there were many times I'd have liked to provide a PNG or text fallback if SVG wasn't supported

[07:38:01.0000] <annevk>
isn't the assumption that going forward all UAs support SVG natively?

[07:38:02.0000] <roc>
you can probably cobble together some <switch> and <img xmlns="..."> to get fallback that actually works

[07:39:00.0000] <roc>
and not only work in legacy UAs that don't support <ext>, but also in UAs that support <ext> but not SVG

[07:40:00.0000] <annevk>
that's seems like a rather odd case to optimize for imo

[07:41:00.0000] <roc>
I'm not saying we should optimize for it

[07:42:00.0000] <roc>
just commenting that it (probably) falls naturally out of shepazu's direction

[07:42:01.0000] <shepazu>
annevk: I'm not assuming that IE will support SVG (though it would be great if it did)... and in the meantime, I'd like for authors to use SVG with confidence that the user will see *something*

[07:43:00.0000] <shepazu>
roc: <switch> only works if the UA supports <switch>, which is a bad fallback assumption :)

[07:43:01.0000] <annevk>
for IE people can use some library that plugs into silverlight in the meantime

[07:43:02.0000] <annevk>
assuming IE will implement <ext> but not SVG also seems like an odd case to optimize for

[07:43:03.0000] <roc>
optimizing for UAs that support <ext> and SVG but not <switch> seems like a very silly thing to do

[07:43:04.0000] <shepazu>
annevk: the <ext> proposal doesn't assume that <ext> will be implemented... it works anyway

[07:44:00.0000] <roc>
or even thinking about them for one second

[07:44:01.0000] <hsivonen>
shepazu: <desc> can be used for fallback

[07:46:00.0000] <shepazu>
hsivonen: that's not the intended use (it breaks the semantics), and it's not robust anyway

[07:46:01.0000] <shepazu>
there can be many <desc> elements in a file

[07:46:02.0000] <shepazu>
svg file

[07:51:00.0000] <annevk>
yeah, <desc> is prolly best for that if needed

[07:53:00.0000] <zcorpan>
<svg><fallback><img>

[07:55:00.0000] <roc>
so

[07:55:01.0000] <shepazu>
zcorpan: that makes it a feature of SVG, not of HTML, and it would need to be added as a feature of every language that is intended to work in HTML (mathml, etc.)... and I don't think it solves any problems that <ext><fallback> doesn't

[07:56:00.0000] <roc>
shepazu: your intent is that browsers would do strict XML conformance checking on the <ext> fragment, right?

[07:57:00.0000] <shepazu>
roc: depends what you mean by "strict"

[07:57:01.0000] <shepazu>
but generally, yes

[07:57:02.0000] <zcorpan>
shepazu: right but fallback would be nice to have in xhtml too, not just text/html

[07:57:03.0000] <shepazu>
zcorpan: I'd like <ext> to be added to XHTML, too

[07:58:00.0000] <zcorpan>
shepazu: ah

[07:58:01.0000] <shepazu>
roc, until such time as there is an infrastructure of parsers and toolchains that understand a looser parser for XML, which I think is likely

[07:59:00.0000] <shepazu>
roc: it's a matter of deployment strategies, in part

[07:59:01.0000] <roc>
wouldn't you consider that a disaster?

[07:59:02.0000] <shepazu>
consider what a disaster?

[08:01:00.0000] <roc>
loose XML parsing

[08:01:01.0000] <roc>
anyway

[08:01:02.0000] <shepazu>
if inkscape, and validators, and XML generators and parsers all had a looser parser that recovers particular validity and certain well-formedness "errors", and if it were widely deployed, I would be totally comfortable with having SVG use that syntax

[08:02:00.0000] <shepazu>
just not as it stands

[08:02:01.0000] <roc>
it seems that the biggest problem you're going to have is convincing people that browser should parse the <ext> fragment in strict XML syntax

[08:02:02.0000] <shepazu>
yeah :)

[08:03:00.0000] <shepazu>
well, not all people, just certain ones :)

[08:03:01.0000] <roc>
it seems that your wiki proposal doesn't even mention that

[08:03:02.0000] <roc>
directly

[08:03:03.0000] <shepazu>
hmmm... yeah, maybe I assumed that, it had been my stance for a while and it was talked about on IRC... I'll add it

[08:03:04.0000] <shepazu>
thanks

[08:04:00.0000] <roc>
in fact it says "Inside you can use XML or another format. "

[08:04:01.0000] <hsivonen>
shepazu: I wouldn't be too surprised if down the road Inkscape had an HTML5 parser

[08:05:00.0000] <annevk>
parsers are cheap

[08:05:01.0000] <hsivonen>
annevk: depends on how contorted they are

[08:05:02.0000] <roc>
hsivonen: is that true?

[08:05:03.0000] <roc>
:-)

[08:06:00.0000] <hsivonen>
roc: parsers are not exactly cheap :-)

[08:06:01.0000] <roc>
so until people agree on whether it's permissible or required to have strict XML parsing inside HTML, there's not much point in discussing <ext>

[08:06:02.0000] <roc>
so I advise you to focus on that...

[08:06:03.0000] <roc>
now I really have to sleep because North Americans are waking up and giving me work to do

[08:06:04.0000] <annevk>
well, HTML parsers until now have been pretty expensive, but now that it is properly defined they're relatively cheap

[08:07:00.0000] <roc>
I hope that's true

[08:07:01.0000] <annevk>
roc, hah

[08:07:02.0000] <roc>
we may not really know until a major browser implements HTML5 parsing

[08:07:03.0000] <annevk>
(re: sleep/work)

[08:07:04.0000] <hsivonen>
annevk: chances are that an HTML5 parser will be cheap for Inkscape

[08:08:00.0000] <hsivonen>
annevk: however, for Mozilla not as cheap

[08:09:00.0000] <hsivonen>
anyway, I think the SVG WG's copy-paste concern is much better solved by a serializer than by complicating parsing, in my opinion

[08:10:00.0000] <hsivonen>
I think Hixie's solution is remarkably elegant and (relatively) low-impact on the parser

[08:11:00.0000] <gsnedders>
annevk: I would throw a fatal error when <ext> doesn't start with an element

[08:11:01.0000] <gsnedders>
I'd rather <ext> was special like <script>: the only thing that ended it was another <ext>

[08:11:02.0000] <annevk>
hsivonen, true, if you have an existing HTML parser it might be relatively expensive on the short term

[08:11:03.0000] <gsnedders>
This is why I really expect making <ext> CDATA easier

[08:11:04.0000] <gsnedders>
<try> <except>? :P

[08:11:05.0000] <gsnedders>
(And no, that isn't a serious suggestion)

[08:11:06.0000] <annevk>
the first bit was?

[08:16:00.0000] <shepazu>
http://www.securityfocus.com/bid/29386/exploit

[08:23:00.0000] <Philip`>
shepazu: Hixie is concerned about cases like http://www.cocopahrv.com/map.html where extension elements like <svg> contain HTML elements that must still be treated as HTML elements by the error recovery system, which is a bit of a pain

[08:24:00.0000] <Philip`>
and I'm not sure how that could be handled correctly by only using abort-on-XML-ill-formedness (instead of abort-on-HTML-elements)

[08:26:00.0000] <Philip`>
(particularly since people seem to dislike unwinding and reparsing upon detecting errors (like with unclosed <script> elements))

[08:27:00.0000] <Philip`>
(It'd still be possible to do abort-on-XML-ill-formedness-and-on-HTML-elements and have that work, though)

[08:30:00.0000] <Philip`>
(though if it was actual real XML-ill-formedness, that'd be irritating since it'd prevent copying output from SVG tools directly into HTML, because SVG tool output sometimes has crazy namespace doctype stuff and you couldn't copy the doctype into the HTML document)

[08:31:00.0000] <hsivonen>
does the SVG WG have other stated issues with Hixie's approach except for copy-paste?

[08:32:00.0000] <annevk>
prolly namespaces

[08:32:01.0000] <hsivonen>
annevk: do you mean round-tripping Inkscape cruft?

[08:33:00.0000] <annevk>
yeah, or author created namespaces to work around the lack of data-

[08:33:01.0000] <Philip`>
http://www.w3.org/Graphics/SVG/WG/wiki/SVG_in_text-html has various SVG WG comments, in case you haven't seen that

[08:33:02.0000] <Philip`>
(Hmm, that page looks far less ugly now that I have appropriate fonts (Gill Sans?) installed)

[08:34:00.0000] <hsivonen>
annevk: it seems to me that adding that feature to Hixie's design would be less disruptive than requiring an XML tokenizer mode

[08:34:01.0000] <hsivonen>
"SVG should remain XML when inline in HTML." is not a use case

[08:34:02.0000] <hsivonen>
as a requirement, it lacks rationale

[08:35:00.0000] <annevk>
i suppose, it's not really clear to me at all, so far it's all rather vague and people seem very focused on specific solutions rather than what we actually need (likely on both sides)

[08:35:01.0000] <Philip`>
We need to compete with VML-in-text/html

[08:36:00.0000] <hsivonen>
"Should allow for unrestricted growth of the SVG language by the SVG specifications" having legacy restrictions comes with the text/html territory. that's tough, but that's the way it is

[08:39:00.0000] <Philip`>
gsnedders: How do you parse the spec in 11 seconds? It seems to take me at least 17, and your computer shouldn't be 50% faster than mine :-(

[08:39:01.0000] <Philip`>
(on the latest version of 'source')

[08:40:00.0000] <hsivonen>
I disagree with even more stuff under "Use Cases and Requirements Discussion"

[08:40:01.0000] <gsnedders>
Philip`: I has leet computer.

[08:41:00.0000] <gsnedders>
Philip`: I just run `python parse.py --no-html --profile --treebuilder=lxml ../../html5`

[08:41:01.0000] <gsnedders>
/me shrugs

[08:43:00.0000] <Philip`>
gsnedders: Hmm, if I run that then it takes minutes and says "25.673 CPU seconds"

[08:44:00.0000] <hsivonen>
/me strongly disagrees with "Already used for SVG, so small cost of implementation " under "XML Parser"

[08:45:00.0000] <gsnedders>
/me tries time … to see if that agrees with the number of CPU seconds

[08:47:00.0000] <Philip`>
(I'm using a C2D 2.0GHz, with frequency scaling locked at maximum, so it ought to be pretty hard to get 50% better single-threaded performance, unless it's just because my python is slow or something)

[09:31:00.0000] <lightyear>
good evening everybody

[09:32:00.0000] <lightyear>
who can I ask some questions about the python html5lib?

[09:34:00.0000] <Philip`>
lightyear: It may be best to just ask the question here, and somebody will reply if they know the answer

[09:35:00.0000] <zcorpan>
are the requirements on sizes='' testable?

[09:35:01.0000] <lightyear>
k

[09:36:00.0000] <lightyear>
I was wondering if there is any nice API to search something in a tree that I get from html5lib

[09:36:01.0000] <jgraham>
lightyear: It depends which treebuilder you use

[09:36:02.0000] <jgraham>
I suggest using lxml with python

[09:36:03.0000] <lightyear>
like lets say, I want to have the element "html->body->div" with the attribute "class" set to ".my_class"

[09:37:00.0000] <jgraham>
If you use lxml you can use xpath to express that constraint

[09:37:01.0000] <lightyear>
I was using beautifulsoup before but that fails with a "maximum recursion depth exceeded" on my content

[09:37:02.0000] <lightyear>
lxml... sounds good

[09:37:03.0000] <jgraham>
like /html/body/div[@class='my_class' or something

[09:38:00.0000] <jgraham>
lightyear: Is the maximum recursion depth being exceeded only when you try to find stuff?

[09:38:01.0000] <lightyear>
yeah... that sound pretty good

[09:39:00.0000] <lightyear>
jgraham: no, when I try to parse it with the bsoup parser

[09:39:01.0000] <jgraham>
lightyear: Oh, that may be a html5lib bug. Can you file an issue and attach a testcase for which that happens please

[09:39:02.0000] <Philip`>
http://code.google.com/p/html5lib/issues/detail?id=59 ?

[09:40:00.0000] <lightyear>
read this, so I guess it the problem...

[09:40:01.0000] <lightyear>
but that one works here...

[09:40:02.0000] <jgraham>
Philip`: Could be. I was secretly hoping people weren't hitting the issue. I guess I need to prioritize fixing it then

[09:42:00.0000] <jgraham>
lightyear: It would be really useful to see a testcase if your issue is distinct from that one

[09:42:01.0000] <lightyear>
I try to parse http://tvshack.net

[09:42:02.0000] <lightyear>
the homepage

[09:42:03.0000] <hsivonen>
I'm tempted to write a TreeBuilder subclass for the GWT DOM wrapper

[09:43:00.0000] <jgraham>
On a differnt topic, has anything interesting happened in the last 3 days? The email subject lines on public-html don't look promising...

[09:43:01.0000] <jgraham>
lightyear: Thanks

[09:43:02.0000] <zcorpan>
Hixie: was adding ruby now in response to robert burns in http://www.w3.org/2002/09/wbs/40318/wd-html5-may/results ? :)

[09:44:00.0000] <jgraham>
(I guess Ruby is quite interesting)

[09:45:00.0000] <Philip`>
"Date: 2008-05-26 03:12:10 -0700; [cit] (2) <ruby> support."

[09:45:01.0000] <Philip`>
"Robert Burns: last responded on 26, May 2008 at 16:09 (UTC)"

[09:46:00.0000] <Philip`>
zcorpan: Presumably it wasn't in response, since it was done six hours earlier :-)

[09:46:01.0000] <zcorpan>
Philip`: ah, so good timing them :)

[09:46:02.0000] <zcorpan>
then

[09:46:03.0000] <lightyear>
jgraham: that is the way I can reproduce it everytime: http://pastebin.com/m4cd3b2be

[09:46:04.0000] <Philip`>
jgraham: There was, uh, some discussion of alt

[09:47:00.0000] <jgraham>
Philip`: That's why I said interesting ;)

[09:47:01.0000] <Philip`>
jgraham: Also I made inputstream a bit faster, which may be handy

[09:47:02.0000] <jgraham>
Philip`: Wow, cool. 

[09:48:00.0000] <Philip`>
(but the effect is only tens of percents, not orders of magnitude)

[09:49:00.0000] <hsivonen>
GWT's JDK class library is much more limited than I thought :-(

[09:50:00.0000] <jgraham>
Still, nice to have :) Did you consider pre-populating the regexp cache with common regexps? I guess that might not make much difference since you have to compile them at least once anyway

[09:51:00.0000] <Philip`>
I'm not sure how that would help, since it would not decrease (and might increase) the number of regexp compilations, and anyway they take about zero time to compile

[09:52:00.0000] <jgraham>
It would make excepting the try/except less common which afaik is relatively slow

[09:54:00.0000] <Philip`>
There's only six regexps ever used (at least when parsing the spec), so that case seems pretty rare, although I wonder if precompiling ['A', 'B', ...] to /[A-Za-z]/ would make the regexp execution faster...

[09:55:00.0000] <takkaria>
I imagine it would, assuming python uses the pcre library

[09:56:00.0000] <Philip`>
Um, it would help if the regexp cache didn't totally ignore 'opposite'

[10:01:00.0000] <Philip`>
Hmm, /[A-Za-z]/ makes no measurable difference according to my measurements

[10:05:00.0000] <jgraham>
takkaria: I think python has its own regexp engine

[10:10:00.0000] <takkaria>
evidence I should learn python, really, so I don't keep on giving out counterfactual advice

[10:11:00.0000] <hsivonen>
Philip`: do you have experience with GWT?

[10:14:00.0000] <hsivonen>
it seems that the Validator.nu parser is not quite ready for GWT use, but with a little refactoring it could become suitable for use inside the GWT

[10:29:00.0000] <Philip`>
hsivonen: None at all

[10:32:00.0000] <hsivonen>
ok

[10:33:00.0000] <hsivonen>
it seems awesome in principle, but in practice it would be a lot more awesome if it supported a larger subset of the JDK class library

[10:36:00.0000] <Philip`>
gsnedders: Oh, whoops, I'd limited my CPU to 1.67GHz a while ago when trying to avoid it overheating while upgrading Gentoo - it goes a bit faster when I remove that limit

[10:39:00.0000] <gsnedders>
Philip`: That doesn't help.

[10:43:00.0000] <Philip`>
Also it seems I can save ~10% in the tokeniser by improving the entity matcher

[10:47:00.0000] <Philip`>
/me commits that and wonders how much it really helps

[10:49:00.0000] <Philip`>
Hmm, seems like around 7% in parsing the spec

[10:50:00.0000] <gsnedders>
The only problem with trying with the spec is there are very few people who actually care about paring the spec.

[10:50:01.0000] <gsnedders>
(or anything of the size)

[10:50:02.0000] <gsnedders>
I mean, one of Hixie's requirements for the spec-gen was speed. "speed is the most important."

[10:51:00.0000] <gsnedders>
Philip`: It seems to have regressed here

[10:51:01.0000] <gsnedders>
Philip`: 12.5s ow

[10:51:02.0000] <gsnedders>
*now

[10:51:03.0000] <Philip`>
It's not particularly non-representative of all HTML content, and the performance should be proportional to shorter documents

[10:51:04.0000] <Philip`>
gsnedders: Oh. That's not good

[10:51:05.0000] <Philip`>
Does it give the same result when you run multiple times?

[10:52:00.0000] <gsnedders>
/me is just trying a second time

[10:52:01.0000] <gsnedders>
Back down to 11s

[10:52:02.0000] <gsnedders>
/me shrugs

[10:52:03.0000] <gsnedders>
Probably just CPU throttling

[10:52:04.0000] <Philip`>
Ah

[10:52:05.0000] <Philip`>
That's why I disable the frequency scaling :-)

[10:53:00.0000] <Philip`>
('cpufreq-set -g performance' etc)

[10:53:01.0000] <gsnedders>
On what? A laptop?

[10:53:02.0000] <Philip`>
Yes

[10:53:03.0000] <gsnedders>
/me cares too much about battery life :)

[10:53:04.0000] <gsnedders>
My next laptop (that I'll probably get end of next summer) will be specifically bought for its battery life

[10:54:00.0000] <Philip`>
That's why I only disable the frequency scaling temporarily while doing performance testing, and also have the AC power plugged in all the time :-)

[10:54:01.0000] <gsnedders>
I'm lazy :)

[10:54:02.0000] <Philip`>
(It resets automatically from 'performance' to 'ondemand' whenever I suspend the machine, so it doesn't matter if I forget)

[10:55:00.0000] <Philip`>
(Oops, no, I think it resets whenever I put the AC adapter in/out)

[10:56:00.0000] <Philip`>
(since that's what /etc/acpi/actions/ac_adapter.sh is set to do)

[10:58:00.0000] <Philip`>
gsnedders: Is battery life now considered more valuable than a 17" screen?

[10:58:01.0000] <gsnedders>
Philip`: Certainly.

[10:59:00.0000] <gsnedders>
If I was to buy one now, I'd likely get the Dell XPS M1330

[11:12:00.0000] <gsnedders>
/me stares

[11:12:01.0000] <gsnedders>
Mail.app is at 784MB RAM usage

[11:12:02.0000] <gsnedders>
and increasing at 1MB/s

[11:42:00.0000] <zcorpan_>
btw, has something interesting happened in this space since friday?

[11:42:01.0000] <zcorpan_>
/me hasn't been around

[11:43:00.0000] <Dashiva>
Yes

[11:43:01.0000] <Dashiva>
Log reading recommended

[11:44:00.0000] <zcorpan_>
Dashiva: can you summarize? :P

[11:44:01.0000] <Dashiva>
If I could, I would :)

[11:45:00.0000] <zcorpan_>
then i'm going to assume that nothing interesting happened and do other things :)

[11:46:00.0000] <Dashiva>
There was some constructive talk on @alt, for one.

[11:47:00.0000] <zcorpan_>
that's cool

[11:47:01.0000] <zcorpan_>
although i think i should try to ignore alt discussions altogether

[11:47:02.0000] <Dashiva>
Commentary on the SVG group's plans to use XML for SVG in HTML

[11:48:00.0000] <zcorpan_>
oh that might be interesting

[11:48:01.0000] <Dashiva>
A little ruby here and there. Feedback on Lachy's upcoming slideshow.

[11:49:00.0000] <zcorpan_>
k

[11:49:01.0000] <zcorpan_>
thanks

[11:53:00.0000] <hober>
I think I missed the constructive @alt discussion... it got completely drowned in the, err, other parts of the @alt threads

[11:55:00.0000] <takkaria>
there were a good five or six posts which advanced the discussion

[11:55:01.0000] <takkaria>
I still don't see it ever being resolved though

[12:14:00.0000] <gsnedders>
I need to blog something before this month ends

[12:34:00.0000] <takkaria>
heh, a whole string of proposals from RB have just hit the issue tracker

[12:35:00.0000] <hsivonen>
takkaria: also http://www.w3.org/mid/AFB93528-A1CA-49EC-B342-5A8CD2B4080B⊙rc

[12:35:01.0000] <Dashiva>
/me checks

[12:35:02.0000] <smedero>
gregory rosmaita opened a few too.

[12:36:00.0000] <takkaria>
smedero: he opened them for RB

[12:36:01.0000] <smedero>
ahh, I see now

[12:37:00.0000] <takkaria>
hsivonen: oh, lord, I wonder if the tag and xhtml2 know what they're in for

[12:39:00.0000] <Dashiva>
"I, Robert - One man saw it coming."

[12:42:00.0000] <takkaria>
I find the summaries and proposals RB puts forward thoroughly confusing

[12:43:00.0000] <takkaria>
I'm never quite sure what he sets out to achieve because he never gives a real problem statement

[12:44:00.0000] <jgraham>
Please someone make it stop

[12:45:00.0000] <Philip`>
takkaria: The problems are that his proposed features are not supported

[12:46:00.0000] <takkaria>
/me grins

[12:47:00.0000] <Philip`>
They're nice problems since they have an easy obvious solution (i.e. supporting the proposed features)

[12:47:01.0000] <takkaria>
also, putting all these things on the wiki makes it very hard to discuss them

[12:49:00.0000] <Philip`>
At least the wiki pages do give problem statements

[12:53:00.0000] <Philip`>
Do Python profilers exist with timings per line/statement/etc rather than per method?

[13:19:00.0000] <takkaria>
the entry for ISSUE-49 on the wiki appears to contain parts of another proposal

[13:29:00.0000] <annevk>
damn, he hijacked ISSUE-42

[13:32:00.0000] <gsnedders>
The answer to the first question in the physics exam last week was (I hope) 42.

[13:32:01.0000] <gsnedders>
Therefore, I now know the ultimate question!

[13:33:00.0000] <jmb>
is it any good? :)

[13:33:01.0000] <takkaria>
if it was on a physics paper, I doubt it

[13:33:02.0000] <gsnedders>
It's quite long

[13:34:00.0000] <gsnedders>
It's all about how far Q is from P, given a car breaking in the area to a complete rest with a given mass and initial speed, IIRC

[13:34:01.0000] <takkaria>
mm, I miss physics

[13:35:00.0000] <gsnedders>
I need to decide whether to do phys at uni

[13:36:00.0000] <Philip`>
Utilising the uncertainty principle, you could simultaneously do and also not do physics

[13:36:01.0000] <takkaria>
but only if gsnedders was in a quantum state and we weren't observing him :)

[13:36:02.0000] <gsnedders>
/me headdesks

[13:37:00.0000] <takkaria>
as soon as we observed him he'd have done one or the other

[13:37:01.0000] <Philip`>
As soon as we observe him, we just get entangled with him

[13:37:02.0000] <Philip`>
so it's only a problem if somebody then observes us

[13:38:00.0000] <takkaria>
I ended up doing philosophy

[13:38:01.0000] <takkaria>
don't do that

[13:38:02.0000] <Dashiva>
ISSUE-50 (associate-attributions-citations-quotations-and-references)

[13:38:03.0000] <Dashiva>
/me winces

[13:39:00.0000] <Philip`>
/me avoids thinking about it

[13:39:01.0000] <gsnedders>
takkaria: Don't worry, I wasn't planning on it

[13:41:00.0000] <Dashiva>
takkaria: You'd have to make him completely oblivious to himself as well, since he's macroscopic enough to be his own observer

[13:45:00.0000] <Philip`>
/me wonders if calling somebody "macroscopic" could be considered an insult

[13:51:00.0000] <takkaria>
hm, I'm going to reply to RB's use-cases for one of the issues

[13:51:01.0000] <takkaria>
let's see how this goes

[13:51:02.0000] <Philip`>
Uh oh

[13:57:00.0000] <annevk>
interesting, the TAG starts talking about solving real problems

[13:58:00.0000] <annevk>
that could be a slippery slope for them :D

[13:59:00.0000] <takkaria>
where?

[13:59:01.0000] <takkaria>
/me runs to look

[14:05:00.0000] <annevk>
http://www.pacificspirit.com/blog/2008/05/28/xri_solves_what_real_problems is one thing (linked from www-tag)

[14:13:00.0000] <takkaria>
well, that's sent.  probably the worst decision I've ever made to reply to that, but there we go

[14:15:00.0000] <annevk>
you mean so far?

[14:15:01.0000] <Dashiva>
Your sacrifice will not be forgotten

[14:15:02.0000] <annevk>
:p

[14:16:00.0000] <annevk>
takkaria, btw, non-hierarchy use case is the bible iirc

[14:17:00.0000] <takkaria>
oh?

[14:17:01.0000] <Dashiva>
the reason why 

[14:17:02.0000] <Dashiva>
they want to why they want to

[14:17:03.0000] <Dashiva>
Isn't the bible very hierarchical?

[14:17:04.0000] <takkaria>
I edit my own mail, did you know?

[14:17:05.0000] <takkaria>
I still think it's nuts to try and put non-hierarchical data in a hierarchical data format, regardless of use-cases

[14:18:00.0000] <takkaria>
it's like wanting to embed binary video data in an Atom feed

[14:18:01.0000] <annevk>
Dashiva, I remember being explained it overlapping things

[14:18:02.0000] <Dashiva>
I'd think hyperlinks would handle that

[14:18:03.0000] <annevk>
I haven't looked into a bible in a while

[14:21:00.0000] <smedero>
http://lists.w3.org/Archives/Public/public-html/2008May/0690.html

[14:22:00.0000] <smedero>
he's a busy man today

[14:23:00.0000] <annevk>
fortunately e-mail clients have filters

[14:24:00.0000] <annevk>
he prepared this stuff sometime in advance (see html4all archives)

[14:25:00.0000] <Dashiva>
What on earth is IDENT supposed to do?

[14:27:00.0000] <Dashiva>
It looks like class v2

[14:27:01.0000] <takkaria>
will he stop bloody pretending that wikis are good places to propose things?

[14:28:00.0000] <Dashiva>
You should tell him so, if you feel strongly about it

[14:28:01.0000] <annevk>
Dashiva, I thought it would be some DTD thing, but I can't find it

[14:28:02.0000] <Dashiva>
The wiki page says it would only need to be unique within siblings

[14:28:03.0000] <takkaria>
Dashiva: I already have

[14:28:04.0000] <Dashiva>
So like some kind of local subtree id... but that removes the whole purpose of having a _unique_ value

[14:29:00.0000] <annevk>
Dashiva, seems best to not invest much time in this

[14:30:00.0000] <Dashiva>
Well, some time at least. Wouldn't want to dismiss him only because of precedent :)

[14:32:00.0000] <takkaria>
you say that, but...

[14:35:00.0000] <takkaria>
I can imagine that few people reply to any/most of his points and then he complains in six months time that the editor hasn't done what he said

[14:36:00.0000] <takkaria>
anyway, this isn't productive, bbl

[14:36:01.0000] <annevk>
that's ok, it's not possible to do what everyone wants it seems

[14:40:00.0000] <takkaria>
I lied about being back later

[14:40:01.0000] <takkaria>
the idea of providing a mechanism for describing the semantics of styling is amazing

[14:46:00.0000] <Dashiva>
Maybe there should be a dogfood requirement for proposals

[14:46:01.0000] <Dashiva>
E.g. if you want a new attribute, you have to use that attribute yourself for 6 months first

[14:57:00.0000] <Hixie>
woot, my VP is publicly pimping HTML5: http://www.techcrunch.com/2008/05/28/live-from-google-io/

[14:58:00.0000] <Dashiva>
Hixie: I think RB means "page authors" when he says implementors

[14:58:01.0000] <Hixie>
well then he's misusing the word

[14:58:02.0000] <Dashiva>
Specifically "me"

[14:59:00.0000] <csarven>
Is there a way to mix multiple encodings in the same document? e.g., Document being UTF-8 and a component in ISO58859-1

[14:59:01.0000] <Hixie>
i'm not ignoring rob burns

[14:59:02.0000] <Hixie>
heck i even have a special filter set up to prioritise his e-mails!

[14:59:03.0000] <Hixie>
it marks them AAA-Productivity!

[14:59:04.0000] <Hixie>
what more can he possibly want

[15:00:00.0000] <csarven>
*ISO8859-1

[15:00:01.0000] <Dashiva>
I can't seem to find that folder in /issues/, Hixie :)

[15:01:00.0000] <Philip`>
csarven: There isn't

[15:01:01.0000] <Hixie>
it's a prioritisation level that happens before sorting feedback into feedback issue folders

[15:02:00.0000] <csarven>
Philip` Thanks

[15:02:01.0000] <Philip`>
csarven: Convert everything to UTF-8 on the server, and then the world will be happy :-)

[15:03:00.0000] <csarven>
That's almost the case but there is a 3rd party component which is in iso8859-1 :S

[15:03:01.0000] <Philip`>
Can't you add some kind of filter to translate that component's output?

[15:04:00.0000] <Dashiva>
What kind of component? iframe? json?

[15:04:01.0000] <csarven>
I probably could

[15:04:02.0000] <csarven>
Just an external JavaScript

[15:04:03.0000] <annevk>
Hixie, WHATWG was also pimped on the google blog

[15:04:04.0000] <annevk>
well, mentioned

[15:05:00.0000] <Hixie>
oh?

[15:05:01.0000] <Philip`>
csarven: As in <script src="some-iso-8859-1-file.js"></script>?

[15:05:02.0000] <Hixie>
oh today you mean

[15:05:03.0000] <csarven>
Yes

[15:05:04.0000] <Hixie>
yes

[15:06:00.0000] <annevk>
http://googleblog.blogspot.com/2008/05/happy-birthday-google-gears.html

[15:06:01.0000] <Hixie>
yeah

[15:06:02.0000] <Hixie>
that was today :-)

[15:06:03.0000] <Hixie>
part of the same push, i imagine

[15:06:04.0000] <Philip`>
csarven: I guess <script src="..." charset="iso-8859-1"></script> might work for that

[15:06:05.0000] <Hixie>
as in, part of the Google I/O stuff

[15:06:06.0000] <annevk>
i see

[15:07:00.0000] <csarven>
Philip` More like <script>document.write("<script>..)</script>

[15:07:01.0000] <annevk>
you there as well?

[15:07:02.0000] <Hixie>
nah

[15:07:03.0000] <annevk>
seems that Chris Wilson is around

[15:07:04.0000] <annevk>
(from twitter)

[15:07:05.0000] <Hixie>
Gears had several announcements today, though, and so it makes sense they'd mention HTML5 in that context

[15:07:06.0000] <Philip`>
csarven: Hmm, I'm not entirely certain what the problematic situation is

[15:11:00.0000] <csarven>
Philip` http://www.lebelage.ca/argent_et_droits/vos_droits/ -- Contest links, bottom right

[15:12:00.0000] <gsnedders>
/me is trying to write the exact kind of email he hates writing

[15:13:00.0000] <Hixie>
in the latest tag minutes tbl says ignoring rdf for html5 is unacceptable

[15:13:01.0000] <Hixie>
http://www.w3.org/mid/m2ej7mwhed.fsf⊙nc

[15:14:00.0000] <Hixie>
/me looks forward to saying that technologies have to prove themselves _before_ html5's design is warped to handle them

[15:14:01.0000] <Hixie>
aha, there's the google alert for the google blog mentioning whatwg :-)

[15:17:00.0000] <Philip`>
csarven: Ah - are you allowed to change it to <script>document.write("<script charset='iso-8859-1' ... ?

[15:17:01.0000] <Philip`>
/me doesn't actually have any idea whether browsers respect <script charset>

[15:18:00.0000] <Hixie>
they do

[15:19:00.0000] <Philip`>
Does it take precedence over Content-Type charset?

[15:20:00.0000] <roc>
hmm

[15:20:01.0000] <Hixie>
it does whatever the spec says, roughly

[15:20:02.0000] <Hixie>
though brosers differ in various complicated ways

[15:21:00.0000] <gsnedders>
/me wonders whether the HTTP of "HTTP/1.1" needs to be case-insensitive

[15:22:00.0000] <gsnedders>
In everything except CFNetwork (WebKit/OS X) and HTTP.sys (IIS) it is

[15:22:01.0000] <Philip`>
Ah, I missed the part of the spec where it said what to do with charset

[15:22:02.0000] <csarven>
Philip` Hixie Yes, it works! :)

[15:23:00.0000] <roc>
shepazu: you arond?

[15:23:01.0000] <Hixie>
gsnedders: is it really case insensitive? or is it ignored?

[15:24:00.0000] <gsnedders>
Hixie: case-insensitive

[15:24:01.0000] <csarven>
Although, I don't know how stable that is cross-browsres as Hixie says

[15:24:02.0000] <gsnedders>
Hixie: I think, from what I can see

[15:24:03.0000] <gsnedders>
Hixie: In Fx it is certainly

[15:25:00.0000] <annevk>
I thought you could also omit it

[15:25:01.0000] <gsnedders>
annevk: In what? Responses? Only if you're HTTP/0.9, which isn't used any more, so is irrelevant

[15:26:00.0000] <gsnedders>
/me discovers for compat. you can't use HTTP/5.0 :(

[15:27:00.0000] <annevk>
yeah, in the response

[15:27:01.0000] <Hixie>
...and that's why version numbers are bad. :-)

[15:27:02.0000] <gsnedders>
HTTP/1.5 it'll have to be :P

[15:27:03.0000] <Philip`>
How about "HTTP/1.<backspace><backspace>5.0"?

[15:28:00.0000] <gsnedders>
Philip`: 400 Bad Request

[15:28:01.0000] <Philip`>
Hmph

[15:28:02.0000] <gsnedders>
/me wonders

[15:28:03.0000] <annevk>
do we need a new version number?

[15:29:00.0000] <annevk>
don't browsers have rules when you omit it altogether?

[15:29:01.0000] <gsnedders>
Should I keep non-strict parsers (which are meant to be used for responses) as close to RFC2616 as possible while still being useful in the real world, or just go for lax as hell?

[15:30:00.0000] <gsnedders>
annevk: In Fx at least if it doesn't start with "HTTP/" it is treated as HTTP/0.9

[15:30:01.0000] <gsnedders>
annevk: Which is to say that the entire response is the body

[15:30:02.0000] <annevk>
gsnedders, what does "lax as hell" mean? compatible with the real world?

[15:31:00.0000] <annevk>
gsnedders, I see, so we need HTTP/ as a flag at the start to indicate headers are part of the response?

[15:31:01.0000] <gsnedders>
annevk: It doesn't throw any sort of error, even when the real world doesn't rely on it, and it isn't allowed as part of RFC2616

[15:32:00.0000] <gsnedders>
annevk: Currently if you send something really odd to a non-strict parser you can get a fatal error (it really has to be really mucked up to get that to happen, though)

[15:32:01.0000] <annevk>
gsnedders, does that happen with browsers?

[15:33:00.0000] <gsnedders>
annevk: In some, yes

[15:33:01.0000] <annevk>
if it has to be something really odd you might as well never get a fatal error...

[15:34:00.0000] <gsnedders>
/me wonders what causes it currently

[15:36:00.0000] <gsnedders>
It's far easier to cause when you're dealing with a request and a strict parser.

[15:37:00.0000] <gsnedders>
annevk: Basically the response line needs to be invalid at the start (i.e., the "HTTP/1.1 200" part)

[15:38:00.0000] <gsnedders>
annevk: Or if it doesn't have a blank line before the body

[15:38:01.0000] <gsnedders>
annevk: I think that's all that currently causes a fatal error in a non-strict response parser

[15:39:00.0000] <annevk>
wow, http://www.builderau.com.au/program/html/soa/HTML-5-A-change-in-course-straight-for-the-iceberg/0,339028420,339289373,00.htm?feed=rss is quite confused

[15:40:00.0000] <takkaria>
we need a crack PR team to get blog posts like that sorted out

[15:43:00.0000] <Hixie>
annevk: that's old

[15:44:00.0000] <Hixie>
annevk: came out in january

[15:44:01.0000] <Hixie>
i do like the bit that says that html5 is a 15 year step backwards

[15:44:02.0000] <Hixie>
that basically puts the web back at where it started

[15:44:03.0000] <Hixie>
which would be impressive

[15:44:04.0000] <gsnedders>
and HTTP/0.9!

[15:44:05.0000] <gsnedders>
I mean, that didn't even have headers, or status codes :P

[15:44:06.0000] <annevk>
Hixie, maybe they republished some content

[15:45:00.0000] <Hixie>
yeah

[15:56:00.0000] <roc>
finally! http://ajaxian.com/archives/announcing-ajax-libraries-api-speed-up-your-ajax-apps-with-googles-infrastructure

[16:01:00.0000] <Philip`>
The internet is hard because all the functionality and data is distributed and distributed systems are hard, so it's good that Google is aiming to save us by providing a single centralised source for all services and data

[16:01:01.0000] <roc>
yes it is

[16:01:02.0000] <Philip`>
We won't need internet connections at all, we can just have a Google connection

[16:03:00.0000] <Philip`>
Single points of failure that can break vast swathes of the web are always good too

[16:03:01.0000] <roc>
yeah. http://webaccelerator.google.com/

[16:04:00.0000] <annevk>
that product is nice, it exposes incorrect usage of GET

[16:05:00.0000] <Philip`>
/me wonders how the hosting of ajax.googleapis.com compares to e.g. Akamai

[16:06:00.0000] <roc>
perhaps comparable service levels, but AFAIK random people can't use Akamai for free

[16:07:00.0000] <roc>
actually, at least early on, random people *could* Akamize their URLs, because the sales staff were selling service faster than they could whitelist referrers, so they'd just track down heavy unauthorized usage after the fact. Not sure if that's changed

[16:10:00.0000] <Philip`>
Hmm, Akamai seems to be directly peered to all the ISPs I can test, and ajax.googleapis.com looks like it is too (except with four more hops and no reverse DNS)

[16:12:00.0000] <annevk>
http://six03.com/blog/html5/

[16:19:00.0000] <Hixie>
/me looks at notifications APIs

[16:24:00.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/notes

[16:25:00.0000] <Philip`>
OS notifications apply to windows, which seems pretty useless when the notifier is probably in a non-visible tab in that window

[16:27:00.0000] <othermaciej>
one interesting form of notification you are missing is badging the dock icon

[16:27:01.0000] <Hixie>
good point

[16:27:02.0000] <othermaciej>
on mac

[16:27:03.0000] <Hixie>
though that wouldn't make much sense if you have shared tabs

[16:27:04.0000] <othermaciej>
it is common to badge with unread counts

[16:28:00.0000] <othermaciej>
but it might make sense in a window-per-app model like Fluid

[16:28:01.0000] <othermaciej>
I'm not sure how you would reconcile it with use in a general browser (probably not allow it, or maybe badge the tab instead)

[16:28:02.0000] <Hixie>
yeah

[16:29:00.0000] <roc>
I don't think Fluid or Prism actually restrict the app to one window, so they'd probably need to implement something flexible

[16:29:01.0000] <roc>
or provide a configuration option

[16:30:00.0000] <othermaciej>
I believe Fluid has some form of dock icon badging

[16:30:01.0000] <Philip`>
Presumably notifications are useful only for web apps which the user uses a lot, not for random web pages, so would it be reasonable to require it to be opt-in per site? (preferably with the site offering the option to the user, then calling some script function which makes the browser non-modally ask the user for confirmation, or whatever)

[16:31:00.0000] <roc>
the browser could show the notification in the tab, with some UI that lets it move out to the dock or taskbar or whatever

[16:31:01.0000] <othermaciej>
it would be annoying to let a random tab bounce the browser dock icon

[16:31:02.0000] <Philip`>
/me doesn't fancy ads making any part of his browser UI start bouncing and flashing

[16:31:03.0000] <Hixie>
the proposal i wrote up in that file suggests that the notifications would only appear inside the window until the user clicked a button on a notification to opt that origin into being able to send them system-wide

[16:31:04.0000] <roc>
yeah, that sounds good

[16:32:00.0000] <Hixie>
not sure what to do about bouncing the dock icon, i think maybe we should not provide that separately and just make the notifications bounce the icon if that origin is opted in to the side-wide thing

[16:32:01.0000] <Hixie>
it seems bad form to ask for the user's attention without saying why anyway

[16:32:02.0000] <Philip`>
That doesn't sound good since I have one browser window and dozens of tabs, and I don't want all of them to be able to distract me when I'm working on something totally unrelated in a different tab

[16:33:00.0000] <Hixie>
s/window/tab/

[16:33:01.0000] <Philip`>
(even if it's in the same window)

[16:33:02.0000] <Philip`>
Ah, okay

[16:36:00.0000] <Hixie>
gears' version of this has displayAtTime, displayUntilTime, and addAction() APIs

[16:37:00.0000] <Hixie>
i wonder what the use case for those was

[16:37:01.0000] <Hixie>
/me shall have to contact them

[16:37:02.0000] <Philip`>
Do those get scheduled and displayed even if you've navigated away from the page that added them?

[16:38:00.0000] <Philip`>
(e.g. you could visit your calendar page once, and it'd remind you of all events for the rest of the day)

[16:38:01.0000] <Philip`>
(at least until your browser crashed and restarted and then you forgot to revisit the calendar to reset the scheduled notifications and missed your appointments)

[16:38:02.0000] <roc>
that wouldn't be a great approach

[16:38:03.0000] <roc>
meetings could change online and you'd get stale data

[16:39:00.0000] <annevk>
what was the whatwg.org link again for the expected timeline?

[16:40:00.0000] <Philip`>
annevk: http://www.whatwg.org/specs/web-apps/current-work/TIMETABLE ?

[16:40:01.0000] <annevk>
ah

[16:40:02.0000] <annevk>
I was trying current-work/timeline

[16:40:03.0000] <annevk>
and also TIMELINE

[16:40:04.0000] <annevk>
but not table

[16:41:00.0000] <Philip`>
I tried Google instead :-)


2008-05-29
[17:08:00.0000] <Hixie>
yeah I'm not sure I follow the use case for the scheduled notifications

[17:09:00.0000] <Hixie>
i'm baffled by why some people seem to want range.insertNode() to not insert the node into the range, but only when the range is collapsed

[17:10:00.0000] <annevk>
Hixie, calendar apps?

[17:11:00.0000] <Hixie>
yeah but see what roc said

[17:12:00.0000] <annevk>
presumably if that happened the app would update the notification?

[17:12:01.0000] <roc>
either the app is running in a browser context or it isn't

[17:12:02.0000] <roc>
if it is, it doesn't need the timed events

[17:13:00.0000] <roc>
if it isn't, there's a problem with staleness

[17:16:00.0000] <Philip`>
/me attempts to write feedback on arc(), but then realises he already did it a month ago

[18:31:00.0000] <andersca>
Hixie?

[18:37:00.0000] <Lachy_>
jgraham, yt?

[18:45:00.0000] <Hixie>
andersca: here

[19:11:00.0000] <andersca>
Hixie: just sent mail to the whatwg list

[19:16:00.0000] <Hixie>
cool

[19:16:01.0000] <Hixie>
what's the priority?

[19:18:00.0000] <Hixie>
andersca: does the fallback thing help?

[19:19:00.0000] <andersca>
Hixie: I don't think so, since it requires the namespace to have the same domain as the manifest

[19:20:00.0000] <Hixie>
ah indeed

[19:20:01.0000] <andersca>
Hixie: also, it would be nice to be able to handle the mt[0-9].google.com case too

[19:20:02.0000] <andersca>
I'd say the priority is "medium"

[19:20:03.0000] <Hixie>
my guess would be that i would just make the onlien whitelist a list of uri prefixes

[19:21:00.0000] <Hixie>
and you'd list the 10 google.com hosts in the file

[19:21:01.0000] <andersca>
yeah, that sounds good

[19:21:02.0000] <Hixie>
feel free to experiment with doing that and mail your experiences to the list :-)

[19:21:03.0000] <andersca>
will do, thanks!

[00:22:00.0000] <BenMillard>
today I decided to skim some recent IRC logs and public-HTML messages

[00:22:01.0000] <BenMillard>
I think ignoring almost all public-HTML e-mail was the right choice for me, albeit a tough one

[00:23:00.0000] <BenMillard>
I also think ignoring nearly everything from this channel is the right choice for me

[00:23:01.0000] <BenMillard>
not because either are of poor quality or whatever

[00:24:00.0000] <BenMillard>
it's simply that the time I'd spend tracking it all wouldn't make me more useful or productive

[00:25:00.0000] <BenMillard>
in more positive news, Mozilla are now funding me to do more work along the lines of Collections of Interesting Data Tables: http://blog.hecker.org/2008/05/28/mozilla-foundation-activities-week-ending-20080523/

[00:31:00.0000] <othermaciej>
BenMillard: that's cool

[00:36:00.0000] <BenMillard>
othermaciej, thanks :)

[00:36:01.0000] <BenMillard>
this was the proposal I gave them: http://projectcerbera.com/web/study/html5-research

[01:16:00.0000] <hsivonen>
Dashiva: paragraphs in the Bible can span over verse boundaries but if you were using normal CSSish counters, it would be natural in the markup sense to make verses siblings of each other without interrupting paragraphs

[01:18:00.0000] <BenMillard>
Hixie, ^ about Mozilla funding. Current thoughts are: http://projectcerbera.com/blog/2008/05#day29

[01:19:00.0000] <Hixie>
hello

[01:19:01.0000] <Hixie>
yeah i saw the announcement -- someone mentioned it in #htmlwg. congratulations!

[01:19:02.0000] <Hixie>
/me looks at your blog

[01:20:00.0000] <Hixie>
i'd love to see data on the img alt stuff, jgraham would be a great person to coordinate with on that, he has some good ideas of what to look for

[01:20:01.0000] <Hixie>
(though it's going to be tough to draw conclusions about what is most likely to get authors to do the most accessible thing from _any_ data, sadly)

[01:22:00.0000] <BenMillard>
hixie, cheers. I'll probably start by adding the 30 or so tables I didn't get add to the first study

[01:22:01.0000] <BenMillard>
and then analyse how frequently header+header relationships and <td scope> are used

[01:24:00.0000] <Hixie>
cool

[01:24:01.0000] <BenMillard>
jgraham and I previosuly coordinated on data table stuff, so doing that in parallel could work

[01:24:02.0000] <Hixie>
awesome

[01:24:03.0000] <Hixie>
one thing i'd caution against, looking at this list, is to not do anything that could be automated somehow -- if there's something that could be automated that you think needs doing, just let me know and i'll get some google resources to do it

[01:25:00.0000] <Hixie>
probably a better use of all our resources

[01:25:01.0000] <Hixie>
the most valuable thing you can bring to these studies is the human judgement aspect

[01:25:02.0000] <BenMillard>
hixie, yeah I can't compete with Google (or Philip` or gnsedders) when it comes to total numbers

[01:26:00.0000] <Hixie>
also one thing i can really help with is getting you lists of sample urls that have some markup feature

[01:27:00.0000] <BenMillard>
that would be another string to my bow :)

[01:28:00.0000] <BenMillard>
although I think it's more interesting to work from the other direction; finding types of content and then check the markup

[01:28:01.0000] <Hixie>
indeed

[01:28:02.0000] <BenMillard>
both ways are useful, though

[01:28:03.0000] <BenMillard>
I intend to spend maybe a week writing down every URL I come across in daily life to help target "normal" websites

[01:29:00.0000] <Hixie>
i could also get you a list of pages with high page rank or something, if you want

[01:29:01.0000] <hsivonen>
BenMillard: congratulations for the grant

[01:30:00.0000] <BenMillard>
hsivonen, thanks!

[01:30:01.0000] <BenMillard>
Hixie, these are all very cool offers and would be really helful

[01:30:02.0000] <BenMillard>
but right now, I need to make some websites to cover the money I've lost not making anything for the past 2 months!

[01:30:03.0000] <Hixie>
:-)

[01:30:04.0000] <Hixie>
feel free to contact me whenever

[01:31:00.0000] <Hixie>
give me a few days to get you data, it takes time to go through billions of files

[01:32:00.0000] <BenMillard>
I'll add a link to these logs to my blog, then when I return from making sites I can refresh my memory with all these options

[01:32:01.0000] <BenMillard>
it's really quite exciting!

[02:01:00.0000] <bosky101>
hi. does anyone remember the site/url where someone(i think it was ian hixie) gave information of css classnames heuristics based on a large crawl

[02:03:00.0000] <hsivonen>
bosky101: http://code.google.com/webstats/

[02:05:00.0000] <bosky101>
brilliant! thanks

[03:58:00.0000] <annevk>
hmm, the weather is terrible

[03:58:01.0000] <Philip`>
I disagree

[03:58:02.0000] <Hixie>
yeah, hasn't been above 25C all week

[03:58:03.0000] <Hixie>
i had to wear a coat today!

[03:59:00.0000] <annevk>
until now it was good, now it rains badly

[04:06:00.0000] <Hixie>
It feels so weird, writing all these "should"s

[04:06:01.0000] <roc>
It's freezing here

[04:06:02.0000] <roc>
down to 9C

[04:06:03.0000] <Hixie>
i guess it's coming on to winter in your neck of the woods

[04:07:00.0000] <roc>
definitely

[04:08:00.0000] <annevk>
freezing at 9C, hah

[04:08:01.0000] <Philip`>
Today is quite sunny - it should reach 20C if we're lucky

[04:08:02.0000] <annevk>
a few years back it reached -10 / -15 in Oslo at some points

[04:09:00.0000] <Hixie>
it reached -20C while i was there, for a few hours at night, irrc

[04:09:01.0000] <Hixie>
iirc

[04:09:02.0000] <roc>
I was joking. When I lived in Pittsburgh it reached -20C a few times

[04:09:03.0000] <Hixie>
-20C is stupid cold

[04:09:04.0000] <annevk>
brr

[04:09:05.0000] <Hixie>
/me hugs the bay area

[04:09:06.0000] <Hixie>
we have two seasons here

[04:09:07.0000] <Hixie>
spring and summer

[04:09:08.0000] <hsivonen>
Is it warm on the inland side of the Bay?

[04:09:09.0000] <hsivonen>
San Francisco is *cold*

[04:09:10.0000] <roc>
yes

[04:09:11.0000] <roc>
yes it is

[04:10:00.0000] <Hixie>
SF doesn't count as the bay area, insofar as weather goes

[04:11:00.0000] <roc>
I'm actually amazed it doesn't get colder in Oslo than Pittsburgh

[04:11:01.0000] <roc>
I thought those Norse hackers were good because they couldn't go outside for six months

[04:12:00.0000] <annevk>
gulfstream

[04:12:01.0000] <hsivonen>
If I create a fresh document instance in JS and add <script> elements, will they run? if so, in what context?

[04:12:02.0000] <annevk>
well, it does get pretty dark in the winter most of the day :)

[04:12:03.0000] <hsivonen>
If I importNode script elements to the current HTML document, will they run?

[04:12:04.0000] <Hixie>
hsivonen: yes, and in that document's

[04:12:05.0000] <Hixie>
hsivonen: importNode will work if the <Script>s haven't already run

[04:12:06.0000] <hsivonen>
Hixie: is there a way to tell a document object to be safe and not run scripts?

[04:13:00.0000] <Hixie>
not before html5, no

[04:13:01.0000] <Hixie>
<iframe sandbox> will do that though

[04:13:02.0000] <Hixie>
in new uas

[04:13:03.0000] <hsivonen>
Hixie: how does your live dom viewer handle script security from user input?

[04:13:04.0000] <roc>
the stupid irony of the Bay Area is that they have great weather but most of the population gains nothing due to late sleep patterns and/or insane work habits

[04:13:05.0000] <Hixie>
hsivonen: that domain has no cookies

[04:14:00.0000] <Hixie>
roc: hah, so true

[04:14:01.0000] <Hixie>
/me saw dawn yesterday :-/

[04:15:00.0000] <annevk>
Hixie's working hours work for me :)

[04:15:01.0000] <annevk>
gives a reasonable Europe overlap

[04:15:02.0000] <Hixie>
that's partly why i do it

[04:15:03.0000] <Hixie>
at least, that's my excuse

[04:15:04.0000] <Hixie>
i could do it as easily if i got up early :-P

[04:17:00.0000] <hsivonen>
I find it bad for my health that people start having interesting discussions when it is late in the evening here

[04:17:01.0000] <hsivonen>
this time works much better for me

[04:17:02.0000] <Hixie>
heh

[04:18:00.0000] <Hixie>
hm

[04:18:01.0000] <Hixie>
the notification API has one problem

[04:18:02.0000] <Hixie>
if you open multiple instances of a calendar app

[04:19:00.0000] <Hixie>
and they all fire notifications for each event...

[04:19:01.0000] <annevk>
why would you do that?

[04:20:00.0000] <Hixie>
do what?

[04:21:00.0000] <Philip`>
They could use localStorage to synchronise between all the tabs

[04:21:01.0000] <Hixie>
seems unlikely that they'd bother

[04:21:02.0000] <Hixie>
i guess we can just tell uas to coallesce identical messages

[04:22:00.0000] <annevk>
open multiple tabs of the same app

[04:23:00.0000] <Philip`>
If I get two identical email messages, close enough together that their notifications will overlap, I'd probably want to receive two notifications, else I might miss the second message

[04:23:01.0000] <Hixie>
i do that sometiems with calender, either by mistake or to compare things

[04:23:02.0000] <Hixie>
if you get two identical messages, why would you care about the second one?

[04:24:00.0000] <roc>
"Missile launched!"

[04:24:01.0000] <Philip`>
*two different email messages but with identical subject lines so the subset of data used in the notifications is identical

[04:24:02.0000] <Hixie>
hmm

[04:26:00.0000] <Philip`>
(Get one notification, switch to tab and read email message, get second notification but the UA coalesces it with the first so I can't tell there's been another message)

[04:26:01.0000] <hsivonen>
do iframes let the parent document remove the documentElement and create a new one?

[04:26:02.0000] <Hixie>
hsivonen: in principle, yes, in practice it depends on the UA

[04:27:00.0000] <Hixie>
Philip`: how about just having it say "x2" "x3" "x4" ?

[04:27:01.0000] <hsivonen>
Hixie: do you happen to remember if Gecko/WebKit/Opera allow removal and reinsertion of the document element?

[04:27:02.0000] <Hixie>
no idea

[04:27:03.0000] <roc>
Hixie: yeah, that's like Unix kernel messages

[04:27:04.0000] <hsivonen>
ok

[04:27:05.0000] <hsivonen>
thanks

[04:28:00.0000] <roc>
Error: not a typewriter (4 times)

[04:28:01.0000] <hsivonen>
I'm trying to figure out if I should bother to make an HTML5 live DOM viewer using GWT and the Validator.nu parser

[04:28:02.0000] <roc>
sick

[04:29:00.0000] <hsivonen>
roc: why sick? isn't it cool compiler tech?

[04:29:01.0000] <Hixie>
sure, go for it :-)

[04:29:02.0000] <hsivonen>
I find that my code is one regexp removal away from GWT compat

[04:29:03.0000] <roc>
those things are not mutually exclusive

[04:29:04.0000] <Philip`>
If the coalescence only happened to notifications from different tabs, I suppose that'd avoid the missed-email problem except in quite rare cases

[04:30:00.0000] <Philip`>
so it probably wouldn't be a real problem then

[04:31:00.0000] <roc>
the 4x/4 times feature is nice for abusive code and for buggy code too

[04:31:01.0000] <Philip`>
/me hopes the notification thing would include UI for disabling notifications from abusive sites

[04:32:00.0000] <Philip`>
(as well as the UI for enabling it first)

[04:32:01.0000] <roc>
nah

[04:32:02.0000] <roc>
well

[04:33:00.0000] <roc>
in Firefox the obvious thing would be to start by showing it in an infobar top-of-tab

[04:33:01.0000] <annevk>
i believe the design is that to make it system wide you first need to grant the site access somehow

[04:33:02.0000] <roc>
and give you options to make it go away or promote to desktop level

[04:33:03.0000] <annevk>
what roc said

[04:34:00.0000] <Philip`>
I'm thinking about what happens when unknowingly grant permission and then the site abuses the feature and you want to stop it

[04:34:01.0000] <Hixie>
actually i have this whole plan for how to deal with abuse

[04:34:02.0000] <Hixie>
the spec describes it, littered with "should"s

[04:34:03.0000] <roc>
one of the most interesting things we're doing on the Web is figuring out ways to let apps do everything they want without trust and without driving users mad with approve/deny decisions

[04:35:00.0000] <Hixie>
indeed

[04:36:00.0000] <Philip`>
You just a generic Cancel/Allow dialog box with a "[ ] remember this decision" checkbox

[04:36:01.0000] <Philip`>
s//need/

[04:36:02.0000] <roc>
no

[04:36:03.0000] <roc>
no no no

[04:36:04.0000] <Philip`>
Alas :-(

[04:38:00.0000] <Hixie>
/me puts Philip` in the brig for 24 hours for even joking about suggesting that

[04:43:00.0000] <annevk>
so is Android just software or also hardware?

[04:43:01.0000] <zcorpan>
Hixie: since "textarea" isn't in the list of tags that escape foreign lands, whether to do case fixup or not has nothing to do with clashing with html since it'll be in the svg namespace anyway (re http://www.w3.org/mid/Pine.LNX.4.62.0805230003340.12911⊙hdc )

[04:44:00.0000] <Hixie>
annevk: android is a software platform

[04:44:01.0000] <annevk>
so where is the hardware from?

[04:45:00.0000] <Hixie>
(in fact, it's an open source linux-based OS distribution, though the source hasn't been released yet aiui)

[04:45:01.0000] <Hixie>
the hardware comes from any hardware manufacturer, most likely those in the Open Handset Alliance: http://www.openhandsetalliance.com/oha_members.html

[04:46:00.0000] <Hixie>
look under "Handset Manufacturers"

[04:46:01.0000] <Hixie>
though typically handsets are sold through network operators ("Mobile Operators" on that list)

[04:47:00.0000] <Hixie>
(i mean, typically for any platform, not just android)

[04:48:00.0000] <annevk>
I meant the hardware used in the demos, but I guess your answer works :)

[04:49:00.0000] <Hixie>
not sure about particular demos, they probably mentioned it in the relevant talks though

[04:51:00.0000] <hsivonen>
so Microsoft wasn't able to break the coupling of Java-the-language and Java-the-class-library, but now Google is doing it with GWT and Android

[05:00:00.0000] <roc>
IBM also did it

[05:00:01.0000] <hsivonen>
true

[05:00:02.0000] <hsivonen>
actually, the only desktop Java app I run regularly is Eclipse

[05:00:03.0000] <hsivonen>
Swing lost

[05:02:00.0000] <hsivonen>
with Eclipse, though, the full J2SE class library is still there

[05:02:01.0000] <hsivonen>
not so with GWT which doesn't even have proper java.io

[05:02:02.0000] <roc>
I was thinking of other things

[05:02:03.0000] <hsivonen>
oh

[05:13:00.0000] <annevk>
/me looks at the notifications API

[05:13:01.0000] <Hixie>
don't yet

[05:14:00.0000] <Hixie>
it'll be done momentarily

[05:15:00.0000] <annevk>
in the IDL it's listed under "// modal user prompts"

[05:15:01.0000] <annevk>
i guess modal should be removed there

[05:16:00.0000] <Hixie>
valid

[05:17:00.0000] <Hixie>
ok the spec is regenned

[05:18:00.0000] <annevk>
ah yeah, it differs a lot :)

[05:19:00.0000] <annevk>
it says "of abused" instead of "or abused"

[05:19:01.0000] <Philip`>
Hixie: http://www.whatwg.org/specs/web-apps/current-work/multipage/ is broken

[05:20:00.0000] <Hixie>
it happens

[05:20:01.0000] <annevk>
also, should it be phrased in a way that does not require it to be visual? "show" makes me think it's visual only

[05:21:00.0000] <Hixie>
it'll fix itself the next time it regens

[05:21:01.0000] <Hixie>
happens every 24h

[05:21:02.0000] <annevk>
is "sandboxed notifications browsing context flag" a a new <iframe sandbox> feature?

[05:22:00.0000] <Hixie>
reload

[05:26:00.0000] <annevk>
still has the typo

[05:27:00.0000] <annevk>
(second paragraph, s/of/or/)

[05:30:00.0000] <Hixie>
oops, missed your comment above

[05:30:01.0000] <Hixie>
was meant to be "if" not "or" :-)

[05:31:00.0000] <annevk>
they can be annoying without being abused too!

[05:31:01.0000] <annevk>
/me doesn't care about working of such things though

[05:34:00.0000] <Hixie>
hehe

[05:34:01.0000] <annevk>
wording, I mean, meh

[05:37:00.0000] <Philip`>
"By default no origin should be flagged as such" - that's abuse of RFC2119 language

[05:39:00.0000] <virtuelv>
Hixie: isn't showNotification a bit overengineered?

[05:39:01.0000] <Hixie>
how so?

[05:39:02.0000] <virtuelv>
I'd say that content and callback is sufficient

[05:39:03.0000] <virtuelv>
leave the title to the UA

[05:39:04.0000] <Hixie>
(it's a lot simpler than, say, the version gears has)

[05:39:05.0000] <virtuelv>
and drop the subtitle

[05:40:00.0000] <virtuelv>
  void showNotification(in DOMString msg, in Function callback); is what Opera has in the widget spec

[05:40:01.0000] <Hixie>
for gmail we want notifications like:

[05:40:02.0000] <Hixie>
| Canvas line style comments

[05:40:03.0000] <Hixie>
| Philip Taylor

[05:40:04.0000] <virtuelv>
Hixie: try to remain compatible with Growl?

[05:40:05.0000] <Hixie>
| The spec says: "The lineCap attribute

[05:40:06.0000] <Philip`>
"user agents may allow users to whitelist flags or groups of flags as being trusted notification sources" - should that be s/flags/origins/g?

[05:40:07.0000] <Hixie>
| defines the type of endings ...

[05:41:00.0000] <Hixie>
Philip`: er yes.

[05:41:01.0000] <virtuelv>
Hixie: while you're at it: window.getAttention()

[05:42:00.0000] <Hixie>
what would that do?

[05:42:01.0000] <virtuelv>
typically, try to grab attention to the window, without grabbing focus

[05:42:02.0000] <virtuelv>
example: change tab color, flash task bar, etc

[05:42:03.0000] <virtuelv>
no arguments, no return value

[05:42:04.0000] <Hixie>
that seems far too abuseable

[05:43:00.0000] <virtuelv>
abuseable, how?

[05:44:00.0000] <Philip`>
Hixie: "all subdomains and ports of *.example.org" - that should be "all subdomains and ports of example.org", or "all ports of any domain matching *.example.org" or something

[05:44:01.0000] <Hixie>
i don't want every porn site i go to to be flashing its task bar button at me

[05:44:02.0000] <Hixie>
Philip`: yeah... hmm

[05:44:03.0000] <Philip`>
Hixie: "Then, mail.example.com and calendar.example.com would both be able to show notifications" - s/.com/.org/

[05:44:04.0000] <Hixie>
oops

[05:45:00.0000] <annevk>
apparently IE has it, from https://bugzilla.mozilla.org/show_bug.cgi?id=293077

[05:46:00.0000] <Hixie>
IE has (had?) window.focus()

[05:46:01.0000] <Hixie>
which i even worse

[05:46:02.0000] <annevk>
oops, yeah

[05:46:03.0000] <annevk>
it does

[05:46:04.0000] <virtuelv>
Hixie: but you want them to display "Free porn" popups?

[05:46:05.0000] <annevk>
I think Opera has that too and Gmail abuses it which is annoying

[05:47:00.0000] <Hixie>
virtuelv: as defined, they can't, since hte popups are limited to the rendering area of the tab.

[05:47:01.0000] <Hixie>
virtuelv: (unless i click the button to trust the site)

[05:48:00.0000] <Hixie>
also, the real question is when would you want to grab attention without having something to grab attention about? if you have somethign to tell the user, use a notification.

[05:48:01.0000] <Hixie>
it'll do the getAttention() thing if the user trusts your site

[05:48:02.0000] <virtuelv>
Hixie: for instance a web-based IRC client that notifies you whenever someone mentions your name

[05:48:03.0000] <virtuelv>
like XChat does, for instance

[05:49:00.0000] <Philip`>
Hixie: "a manner consisted with the platform conventions" - s/consisted/consistent/

[05:49:01.0000] <Hixie>
why wouldn't you want it to show a notification with the line that was said

[05:49:02.0000] <Hixie>
oops, fixed, thanks

[05:50:00.0000] <Philip`>
Hixie: Should the first "should" in "Otherwise, the notification should be rendered in a manner consisted with the platform conventions for system-wide notifications, but it should be rendered within ..." be "should not"?

[05:50:01.0000] <Hixie>
no?

[05:50:02.0000] <Hixie>
why would we ever recommend against using platform conventions?

[05:51:00.0000] <Philip`>
It seems weird to say it should be consistent with platform conventions (which are always outside the browser window) and also that it should be inside the window

[05:51:01.0000] <Hixie>
i guess i could remove the first bit

[05:52:00.0000] <Philip`>
The latter "should" should be more important than the first one, so I'm not sure the first one is needed at all

[05:52:01.0000] <Hixie>
k

[05:52:02.0000] <virtuelv>
Hixie: because that line is far too intrusive

[05:52:03.0000] <virtuelv>
I just need to know that something has happened

[05:53:00.0000] <Hixie>
i'd want to know what happened

[05:53:01.0000] <virtuelv>
for instance, when someone mentions my name in a different tab in xchat, someone has mentioned my name

[05:53:02.0000] <Hixie>
i hate that i have to go look at my irc client to establish that they just said "nn"

[05:53:03.0000] <annevk>
what happens if only whitespace was passed?

[05:53:04.0000] <Hixie>
anyway

[05:53:05.0000] <Hixie>
that's moot

[05:53:06.0000] <Hixie>
because like i said

[05:53:07.0000] <Hixie>
getAttention() is far too abusable

[05:54:00.0000] <Hixie>
annevk: you get a blank notification with just the favicon and title of the page, i guess

[05:54:01.0000] <Philip`>
Hixie: "if a site contains a gadget of a mail application in a sandboxed iframe and that frame triggers a notification upon the receipt of a new e-mail message, that notification would be displayed on top of the gadget only." - the last bit seems wrong, since the notification would be displayed on the top-level browsing context instead of just the gadget

[05:55:00.0000] <Philip`>
Hixie: s/burried/buried/

[05:56:00.0000] <virtuelv>
what is the point of a notification if it can't escape the viewport for the page it originated from, btw?

[05:56:01.0000] <Hixie>
no, if it's sandboxed it'll be displayed on the iframe itself

[05:56:02.0000] <Hixie>
(fixed buried)

[05:57:00.0000] <Hixie>
virtuelv: the point is to show the user that the site supports notifications so that it can opt into showing them system-wide

[05:57:01.0000] <virtuelv>
I realise that you can trust the notification, but I'd need to know that something has happened in the tab to switch to it, I don't actually need the notification

[05:57:02.0000] <Hixie>
s/it/he/

[05:57:03.0000] <Hixie>
see the spec, even normal (not trusted) notifications make the tab caption get highlighted in some way

[05:57:04.0000] <Philip`>
Hixie: Is the sandboxed iframe a top-level browsing context?

[05:58:00.0000] <virtuelv>
Hixie: which is what getAttention() would do as well

[05:58:01.0000] <Hixie>
Philip`: no, how could it be?

[05:58:02.0000] <Hixie>
virtuelv: but there's no way to opt in with getAttention()

[05:58:03.0000] <Hixie>
virtuelv: and it's just a subset of showNotification()

[05:58:04.0000] <Philip`>
Hixie: It could have been defined that way without me noticing :-)

[05:59:00.0000] <virtuelv>
Hixie: more precisely, it's the empty subset of showNotification

[05:59:01.0000] <Hixie>
Philip`: ah :-)

[05:59:02.0000] <Hixie>
virtuelv: eh?

[05:59:03.0000] <Philip`>
Hixie: In that case, the spec just says "it should be rendered within the top-level browsing context of the browsing context associated with the script execution context of the script that invoked the method" and says nothing about it being displayed just on the sandboxed iframe

[06:00:00.0000] <Hixie>
look at the paragraph immediately above that one

[06:00:01.0000] <virtuelv>
empty: no title, subtitle, text or callback

[06:00:02.0000] <virtuelv>
or think of it as a subset of focus() that doesn't try to grab the window focus

[06:00:03.0000] <Philip`>
Hixie: Oh, right, I missed/forgot that bit

[06:01:00.0000] <virtuelv>
I believe most WM's have a method called getAttention these days

[06:01:01.0000] <Hixie>
virtuelv: that would show a notification with the page title and favicon (and maybe domain name, and maybe time, and some buttons).

[06:01:02.0000] <Hixie>
virtuelv: most WMs assume that the user is running trusted code. The Web is different.

[06:01:03.0000] <Hixie>
anyway i should go to bed

[06:01:04.0000] <Hixie>
thanks for the review Philip`

[06:02:00.0000] <virtuelv>
Hixie: and that is more information than I would need to show

[06:02:01.0000] <Hixie>
virtuelv: i disagree :-)

[06:03:00.0000] <virtuelv>
Hixie: you have reinvented alert() asynchronously in this case, albeit a tab-modal one

[06:03:01.0000] <annevk>
it's not modal

[06:04:00.0000] <Philip`>
Hixie: Did you see my first comment?

[06:04:01.0000] <Philip`>
('"By default no origin should be flagged as such" - that's abuse of RFC2119 language')

[06:04:02.0000] <Hixie>
Philip`: nope, missed that one too. will fix.

[06:05:00.0000] <Hixie>
virtuelv: yup, pretty much. except it's not modal, and is transient, and isn't locked to the page.

[06:05:01.0000] <Philip`>
Hixie: "the user agent could just shown one" - s/shown/show/

[06:06:00.0000] <Hixie>
thanks, will fix that too

[06:06:01.0000] <Philip`>
Hixie: Why remove duplicate notifications that are from the same browsing context? That case seems most likely to be intentional (e.g. two identical emails) since there's only one instance of the application in that browsing context

[06:07:00.0000] <ralphm>
Hi! I'm toying with adding link elements pointing to XMPP end points, but I am not sure what rel values to use.

[06:07:01.0000] <Hixie>
Philip`: because having two notifications that are the same is dumb and ugly :-)

[06:08:00.0000] <ralphm>
Example: I have a page that is a representation of a resource, and want to offer an XMPP Publish-Subscribe endpoint that could be subscribed to, to receive updates when the resource changes.

[06:09:00.0000] <Hixie>
annevk: btw that bug is what caused me to start speccing this in the first place (it's the top e-mail in the dom-focus folder)

[06:09:01.0000] <ralphm>
Any ideas?

[06:09:02.0000] <Hixie>
i'm afraid i don't really know what an xmpp end point is

[06:09:03.0000] <Hixie>
but what about rel="xmpp-end-point"?

[06:10:00.0000] <ralphm>
Hixie: you do know about XMPP itself? I was actually thinking of rel="alternate feed"

[06:11:00.0000] <Hixie>
don't use two keywords

[06:11:01.0000] <Hixie>
it'll create two links

[06:12:00.0000] <ralphm>
yeah, intentionally. I read the description of the different link rels here: http://www.whatwg.org/specs/web-apps/current-work/multipage/section-links.html#feed

[06:12:01.0000] <Hixie>
oh, i see, you want to use the existing "alternate feed" definition

[06:12:02.0000] <ralphm>
yes

[06:12:03.0000] <Hixie>
so it's just an RSS page?

[06:12:04.0000] <Hixie>
except not RSS?

[06:12:05.0000] <ralphm>
it is not a page

[06:13:00.0000] <Hixie>
then don't use rel="alternate feed" :-)

[06:13:01.0000] <ralphm>
it would point to e.g. xmpp:pubsub.example.org?;node=test

[06:13:02.0000] <ralphm>
why not?

[06:13:03.0000] <Hixie>
(sounds complicated)

[06:13:04.0000] <Hixie>
rel="alternate feed" points to a syndication feed document

[06:14:00.0000] <Hixie>
which happens to contain the same content as the current document

[06:14:01.0000] <MikeSmith>
/me says, cool to see ralphm on #whatwg

[06:14:02.0000] <ralphm>
well, that XMPP resource provides that, except not as a document as such

[06:14:03.0000] <annevk>
ralphm, I wouldn't use feed, feed is for RSS/Atom

[06:15:00.0000] <annevk>
ralphm, just mint xmpp-end-point or something

[06:15:01.0000] <ralphm>
you would get Atom entry documents sent over XMPP asynchronously whenever the resource changes

[06:15:02.0000] <Hixie>
lordy

[06:15:03.0000] <mpt>
/me tries to imagine what the "I trust this Web site to show me notifications outside the browser window" would look like

[06:16:00.0000] <Hixie>
mpt: a little icon that looks a bit like the windows "restore window" button icon, at the top left of the toast notification thingy

[06:17:00.0000] <mpt>
Or maybe like the "Undock" button for palettes in various art programs

[06:17:01.0000] <ralphm>
I would say the end result for the user would be similar, it is a feed, a client (given XMPP powers) would be able to subscribe to the feed and give the user updates, except without polling.

[06:18:00.0000] <ralphm>
so that's why I thought 'alternate feed' would be appropriate, although I understand it was not designed with XMPP in mind, per se.

[06:18:01.0000] <Hixie>
mpt: right

[06:19:00.0000] <Hixie>
ralphm: what's the use case for this? just changing to a push model instead of pull?

[06:20:00.0000] <ralphm>
Hixie: for this particular example, yes

[06:21:00.0000] <Hixie>
wouldn't that mean you couldn't have more than 65000 subscribers?

[06:21:01.0000] <ralphm>
also, it could be a feed for just one particular thing, like one blog item or one person's profile page

[06:21:02.0000] <ralphm>
Hixie: huh what?

[06:21:03.0000] <ralphm>
why?

[06:21:04.0000] <Hixie>
ralphm: isn't there a limit to how many clients can connect to an xmpp server at once?

[06:22:00.0000] <Hixie>
i guess they don't each need their own outgoing port

[06:22:01.0000] <ralphm>
Hixie: XMPP has a distributed server model, it scales pretty well

[06:23:00.0000] <ralphm>
also, the attempt to subscribe could point to a repeater of the same notifications, so that eventually scales infinitely

[06:23:01.0000] <ralphm>
and way better than Atom/RSS polling

[06:25:00.0000] <Hixie>
i guess it's just hte idea of sending html embedded in xml embedded in xml that has me worried

[06:25:01.0000] <Hixie>
anyway

[06:25:02.0000] <Hixie>
the main problem with using "feed" is that current era blog readers wouldn't be able to handle the xmpp: protocol

[06:25:03.0000] <Hixie>
and so you'd still want them to get the http: one

[06:26:00.0000] <ralphm>
yes, I would provide one of those too

[06:26:01.0000] <Hixie>
but the UA would in practie not know which to send to the blog reader

[06:26:02.0000] <Hixie>
so i'd recommend just minting a new pair of keywords like "xmpp-feed" and "main-xmpp-feed"

[06:26:03.0000] <ralphm>
also, we tend to design for the future, I hope

[06:27:00.0000] <Hixie>
sadly designing for the future without taking the present into account tends to lead to systems that never get adopted

[06:27:01.0000] <ralphm>
yeah, like annevk suggested. Sure I can do that, but I am wondering if we should define a new rel if the semantics are mostly identical, except for mapping to a different protocol scheme

[06:28:00.0000] <hsivonen>
ralphm: are you the person who discussed XMPP integration with annevk and me at XTech after lightning talks?

[06:28:01.0000] <ralphm>
hsivonen: yeah

[06:28:02.0000] <Hixie>
ralphm: well, it's not really the semantics that matter, it's how they are handled in the wild

[06:28:03.0000] <hsivonen>
ralphm: ok. (good to connect the threads)

[06:30:00.0000] <ralphm>
Hixie: a quick scan of web browser reveals that ie, safari seem to ignore the link (seems ok to me) and ff does show it, but nothing happens when you act on it (e.g. clicking on the feed icon)

[06:30:01.0000] <Hixie>
ralphm: if you do want to use feed/alternate feed, i recommend finding a dozen or so different blog readers, and testing them with a variety of different ways of linking the feeds, as picked from a big sample of blogs (e.g. the last 50 blogs mentioned on digg and reddit, and feeds from sites like facebook and twitter) and seeing what they do when you add xmpp: to the mix

[06:30:02.0000] <Hixie>
ralphm: well you need to test the browsers in conjunction with various blog reader software and blog reader sites

[06:31:00.0000] <Hixie>
ralphm: and you need to test not just how xmpp: works, but how the fallback to http: works when xmpp: is specified but not supported

[06:31:01.0000] <Hixie>
ralphm: since if that doesn't work very reliably, people just won't use it

[06:32:00.0000] <Hixie>
anyway, i have to go to bed now

[06:32:01.0000] <Hixie>
it being 25 to 6 :-)

[06:32:02.0000] <ralphm>
Hixie: yeah, that seems fair. Night and thanks

[06:32:03.0000] <Hixie>
np

[06:32:04.0000] <Hixie>
nn

[06:40:00.0000] <ralphm>
If anyone else has ideas or suggestions, that'd be most welcome. We want to draft a specification for auto discovery of XMPP entities and publish-subscribe topics within the XMPP Standards Foundation.

[07:29:00.0000] <gavin_>
annevk: is there a way for me to followup on a bug I filed on Opera a long time ago?

[07:30:00.0000] <annevk>
nr?

[07:30:01.0000] <gavin_>
I don't remember

[07:30:02.0000] <gavin_>
and I can't find any email

[07:30:03.0000] <annevk>
e-mail?

[07:30:04.0000] <annevk>
address

[07:30:05.0000] <gavin_>
not sure of that either... gavin.sharp⊙gc, or maybe gavin/gsharp⊙mc

[07:32:00.0000] <annevk>
bit busy, will look

[07:32:01.0000] <MikeSmith>
Hixie: why was the name "Doctype" chosen for Google Doctype?

[07:32:02.0000] <gavin_>
thanks

[07:34:00.0000] <hsivonen>
now instead of having to compete against w3cschools and Zeldman, I have to compete against Google itself for the search term "doctype"

[07:35:00.0000] <hsivonen>
in fact, Google Doctype pushed me off the first result page on google.fi

[07:35:01.0000] <annevk>
gavin_, ok, found something, it's marked fixed

[07:36:00.0000] <hsivonen>
still there on google.com

[07:36:01.0000] <annevk>
gavin_, it's about setting Opera as the default browser

[07:37:00.0000] <gavin_>
yeah, that's the one

[07:37:01.0000] <gavin_>
cool

[07:37:02.0000] <gavin_>
is it in shipped builds?

[07:38:00.0000] <gavin_>
I suppose I should test

[07:38:01.0000] <annevk>
gavin_, should be in 9.5 builds i guess

[07:38:02.0000] <gavin_>
alright, thanks

[07:38:03.0000] <annevk>
gavin_, according to my data it was fixed over a year ago :)

[07:39:00.0000] <gavin_>
yeah, I reportred it a awhile ago and kinda forgot

[07:39:01.0000] <gavin_>
it was causing trouble with our profile importer

[07:39:02.0000] <annevk>
hereby thanks for the report :)

[07:42:00.0000] <hsivonen>
Hixie: is your Live DOM Viewer under an Open Source license?

[08:05:00.0000] <Philip`>
hsivonen: Hixie hasn't sued me for copying and modifying it

[08:17:00.0000] <hsivonen>
Philip`: not good enough :-(

[08:53:00.0000] <Lachy>
Our presentation went well this morning

[08:55:00.0000] <annevk>
thanks for telling us that!

[08:55:01.0000] <annevk>
:p

[08:55:02.0000] <hsivonen>
do Gecko/WebKit/Opera support DOM Level 3 setUserData?

[08:56:00.0000] <gsnedders>
/me hopes he didn't look too stupid

[08:59:00.0000] <Philip`>
gsnedders: That's why you should avoid ever having photos published :-p

[08:59:01.0000] <gsnedders>
Philip`: I should take a photo of you when you don't realise next time I see you :)

[09:11:00.0000] <annevk>
Philip`, did you add a regression testcase for the html5lib thing you fixed?

[09:13:00.0000] <Philip`>
annevk: Yes (test_stream.py / test_newlines2)

[09:14:00.0000] <Philip`>
(The Ruby implementation doesn't seem to be affected, as far as I can tell)

[09:21:00.0000] <gsnedders>
PLY doesn't seem able to easily cope with http-parsing. :(

[09:23:00.0000] <gsnedders>
comment           = "(" *( ctext / quoted-pair / comment ) ")" would be really horrible

[09:24:00.0000] <gsnedders>
It has no sort of repetition outwith inside tokens

[10:18:00.0000] <Philip`>
"If the WG wants to recommend UI for browsers, it should do so in a seperate document, so as to not confuse what is required for interoperability." - how does that relate to the recent notification API, whose definition is pretty much all about UI?

[11:19:00.0000] <gsnedders>
Oh dear.

[11:19:01.0000] <gsnedders>
/me concludes there isn't any lexer in any language he knows that can be used for HTTP while keeping quite close to ABNF

[11:22:00.0000] <Philip`>
Is there not an ABNF lexer implementation?

[11:22:01.0000] <gsnedders>
No

[11:22:02.0000] <gsnedders>
But we've been over this before :)

[11:22:03.0000] <Philip`>
Have we?

[11:23:00.0000] <Philip`>
/me 's memory is failing

[11:23:01.0000] <gsnedders>
Yeah, when I was on the train going outh.

[11:23:02.0000] <gsnedders>
*south

[11:23:03.0000] <gsnedders>
There's http://simpleparse.sourceforge.net/ which does EBNF

[11:23:04.0000] <Philip`>
Does http://www.coasttocoastresearch.com/home.php not exist or work or something?

[11:24:00.0000] <gsnedders>
There was some issue with it

[11:24:01.0000] <Philip`>
/me goes away

[11:25:00.0000] <gsnedders>
Ah yeah.

[11:25:01.0000] <gsnedders>
I couldn't work out how the hell you used it

[11:27:00.0000] <gsnedders>
/me works it out!

[12:17:00.0000] <hsivonen>
takkaria: http://html4all.org/mailman/archives/list_html4all.org/2008-May/000892.html

[12:18:00.0000] <Dashiva>
I wonder what he imagines those marks would be used for if there was no implementation

[12:38:00.0000] <Philip`>
Dashiva: They are used for semantics, I guess

[12:41:00.0000] <gsnedders>
Hixie: Could you possibly look at <http://simplepie.org> from a Google IP and say if it looks as if it's been hacked — apparently it does from Y! IPs (done by IPs, and not UA, seemingly)

[12:45:00.0000] <Philip`>
gsnedders: http://66.102.9.104/search?q=cache:tAPdA3FmK6cJ:simplepie.org/

[12:46:00.0000] <gsnedders>
Philip`: Sure, but that's six days old. I have no idea how old this seeming hack is, and I can't find any trace of it in the sp.o code

[12:47:00.0000] <Philip`>
"After the neotropical steroid, designer of the clearer prilosec it intoxicated interstate bubbly and recapitulative; satirically mottled." - surely you can't fix the hack and destroy such poetry

[12:47:01.0000] <Philip`>
Ah

[12:49:00.0000] <gsnedders>
http://66.102.9.104/search?q=cache:tAPdA3FmK6cJ:simplepie.org/&hl=en&strip=1 — that shows it all :\

[12:51:00.0000] <gsnedders>
/me can't find the cause at all :\

[12:54:00.0000] <Philip`>
Checked the web server configuration too?

[12:55:00.0000] <Philip`>
It looks like most of the links in the page are to other similarly compromised sites

[12:57:00.0000] <gsnedders>
Philip`: I don't have access to that

[13:01:00.0000] <takkaria>
yay, I got on html4all

[13:01:01.0000] <takkaria>
I must be part of the cabal now

[13:05:00.0000] <gsnedders>
/me wonders if he is

[13:06:00.0000] <takkaria>
RB even managed to link to the wrong post of mine

[13:06:01.0000] <gsnedders>
http://html4all.org/pipermail/list_html4all.org/2007-September/000424.html

[13:06:02.0000] <smedero>
gsnedders: For what it is worth, I just had someone on the Y! internal network look at simplepie.org and they don't see the spammy version. I can email you the frontpage HTML source if you like... but I don't see any of the residuals of one of the wordpress spam hacks.

[13:07:00.0000] <takkaria>
it makes a good point

[13:07:01.0000] <Philip`>
shepazu: http://google.com/search?q=cache:Yof6TxEwk3YJ:schepers.cc/ - your site seems a bit spammy

[13:07:02.0000] <gsnedders>
smedero: I just emailed the guy I've been dealing with, having removed what I thought might be the cause

[13:08:00.0000] <Philip`>
shepazu: Same problem on http://www.svg-whiz.com/ which looks like yours

[13:08:01.0000] <shepazu>
Philip`: spammy?

[13:10:00.0000] <Philip`>
shepazu: "Couldn't you behoove on polygon duet, so you could squib outshoots? Her senegas stand-up as scrape as surpassing eugenias. "Transitionally, I'm bussing out", equidistant diethylpropion hcl 75mg tuberous sclerosis 2." and so on

[13:11:00.0000] <Philip`>
(plus links to other sites that have the same issue when viewed via the Google cache)

[13:11:01.0000] <shepazu>
Philip`: I'm not seeing that...

[13:12:00.0000] <Philip`>
shepazu: Using the Google cache link?

[13:12:01.0000] <gsnedders>
smedero: Could you get them to look once more?

[13:13:00.0000] <smedero>
yeah I just nudged him... though now his status says "lunch!". :/ bums.

[13:14:00.0000] <shepazu>
Philip`: suggestions?

[13:19:00.0000] <smedero>
/me heads to lunch as well

[13:21:00.0000] <zcorpan_>
krijnh: your linkification regexp fails on your own url at http://krijnhoetmer.nl/irc-logs/wai-aria/20080528

[13:22:00.0000] <Philip`>
shepazu: No idea - ask gsnedders when he's worked out why he has the same problem :-)

[13:22:01.0000] <shepazu>
thanks for the pointer

[13:22:02.0000] <Philip`>
/me writes a script to see how big this network is

[13:23:00.0000] <Philip`>
Oh, that doesn't work since Google blocks my cache-spider

[13:24:00.0000] <annevk>
is my site affected in some way? my page rank dropped from 7 to 5 for no apparent reason

[13:24:01.0000] <Philip`>
shepazu: http://www.vectoreal.com/ was in the list before Google started hating me, though

[13:24:02.0000] <annevk>
"To highlight the insanity of the WhatWG" classic

[13:25:00.0000] <shepazu>
/me wishes he had more time to figure this crud out

[13:27:00.0000] <Philip`>
shepazu: It looks like a pretty widespread problem :-(

[13:28:00.0000] <shepazu>
Philip`: to be honest, I don't understand what the issue is...

[13:28:01.0000] <shepazu>
why is archive showing spam when the site itself isn't?

[13:29:00.0000] <shepazu>
for the record, though, I am totally against squibbing offshoots

[13:29:01.0000] <Philip`>
shepazu: It probably detects Googlebot (via IP or something) and modifies the response content

[13:29:02.0000] <annevk>
clever

[13:30:00.0000] <shepazu>
is this something I can or need to do something about?

[13:30:01.0000] <Philip`>
I think it replaces everything between the first <p> and last </p>

[13:30:02.0000] <gsnedders>
Philip`: In the sp.o case it doesn't

[13:30:03.0000] <annevk>
shepazu, you should look in your templates whether there's some hack in there and probably change passwords and all

[13:31:00.0000] <shepazu>
ah, ok, thanks anne

[13:31:01.0000] <Philip`>
shepazu: It seems to be that your site/server is hacked in some way, which is not good, but I haven't got a clue how to fix it :-/

[13:31:02.0000] <hsivonen>
are all those sites running a vulnerable version of WordPress?

[13:31:03.0000] <annevk>
if they're very clever they also hacked the new password thingie to e-mail them

[13:31:04.0000] <annevk>
so a clean install might be safest

[13:32:00.0000] <Philip`>
gsnedders: Yes it does

[13:32:01.0000] <gsnedders>
Philip`: OK, maybe it does. But that, I don't think, is what is causing it.

[13:32:02.0000] <hsivonen>
IIRC, Technorati stopped indexing old WP installs due to something like this

[13:32:03.0000] <gsnedders>
hsivonen: sp.o is running the latest release

[13:32:04.0000] <Philip`>
gsnedders: "<div class="blogimage"><img src="/images/feature_feed.png" alt="[Feed Icon]" /></div> <p><strong><em>SimplePie is a very ..." --> "<div class="blogimage"><img src="/images/feature_feed.png" alt="[Feed Icon]" /></div> "I suppose laterad", barracked corpus."

[13:32:05.0000] <hsivonen>
ah. ok

[13:32:06.0000] <Philip`>
shepazu: Is svg-whiz.com using Wordpress?

[13:33:00.0000] <shepazu>
I did have my blog hacked recently, but svg-whiz doesn't have a blog, it's just HTML

[13:33:01.0000] <Philip`>
Aha

[13:33:02.0000] <gsnedders>
hsivonen: WP isn't even touch the home page

[13:33:03.0000] <Philip`>
Better question: Your site is hosted on Dreamhost, isn't it?

[13:33:04.0000] <Philip`>
/me thanks bsmedberg for noticing that :-)

[13:33:05.0000] <gsnedders>
Mine is, and I'm oh-so-tempted to blame them :P

[13:34:00.0000] <annevk>
oh crap, are all dreamhost sites affected?

[13:34:01.0000] <annevk>
or was this from some time ago? in which case i'm not

[13:35:00.0000] <Philip`>
annevk: The Google cache things are from the past few days, so it can't have been that long ago

[13:36:00.0000] <gsnedders>
It still happens for some people at Y!, I know

[13:36:01.0000] <Philip`>
shepazu: In that case, I think you don't need to do anything, since it's all Dreamhost's fault

[13:36:02.0000] <shepazu>
evil dreamhost!

[13:37:00.0000] <shepazu>
(though I do like them in general)

[13:37:01.0000] <Philip`>
Possible temporary fix: Remove all <p> tags from your pages ;-)

[13:37:02.0000] <shepazu>
Philip`: done, I only use SVG semantics on my sites ;P

[13:40:00.0000] <gsnedders>
shepazu: I don't like them in general :P

[13:47:00.0000] <hsivonen>
does JS intern new strings by default?

[13:47:01.0000] <hsivonen>
/me is surprised by http://code.google.com/p/google-web-toolkit/source/browse/releases/1.5/user/super/com/google/gwt/emul/java/lang/String.java?r=2940#488

[13:49:00.0000] <Philip`>
hsivonen: As far as I'm aware, there's no way to distinguish the interned vs non-interned cases in JS

[13:49:01.0000] <Dashiva>
"CHECKSTYLE_OFF: This class has special needs."

[13:50:00.0000] <hsivonen>
Philip`: why is the impl. I referenced return new String(this); instead of return this, then?

[13:51:00.0000] <Philip`>
hsivonen: No idea

[13:52:00.0000] <annevk>
hsivonen, it seems to invoke intern() and not return new String(this)

[13:53:00.0000] <annevk>
euh, wait

[13:53:01.0000] <annevk>
it seems to not do anything

[13:53:02.0000] <hsivonen>
annevk: /*-{ ... }-*/ is JSNI syntax--not a comment

[13:53:03.0000] <Philip`>
Is "String" JS's String, or is it a magic GWT class variable?

[13:54:00.0000] <annevk>
hsivonen, interesting

[13:54:01.0000] <hsivonen>
Philip`: String in Java scope is java.lang.String and String inside /*-{ ... }-*/ is JavaScript String as far as I can tell

[13:55:00.0000] <Philip`>
Hmm, too much magic for me to understand :-p

[13:55:01.0000] <Philip`>
(I was wondering what type 'this' was, but then I decided it's probably safer to not know)

[13:56:00.0000] <annevk>
Dean Edwards about GWT / JSNI: "This is basically an admission that in the browser environment, JavaScript is king. If you ultimately need to fall back to JavaScript to fix your abstraction layer then your abstraction layer is leaking. JavaScript is king in the browser and GWT is for cowards."

[13:56:01.0000] <hsivonen>
Philip`: this seems to be the JS String

[13:57:00.0000] <hsivonen>
annevk: that's like saying that JNI is an admission that C is king

[13:57:01.0000] <hsivonen>
which it is

[13:57:02.0000] <hsivonen>
but HLLs are still useful

[13:57:03.0000] <annevk>
I thought the last bit was funny

[14:21:00.0000] <gsnedders>
smedero: Where is the guy you got to test it?

[14:22:00.0000] <smedero>
San Fran

[14:22:01.0000] <smedero>
Flickr offices

[14:23:00.0000] <gsnedders>
smedero: thx.

[14:23:01.0000] <Philip`>
/me tries to update the spec splitter to use lxml, since that's a bit faster than minidom

[14:24:00.0000] <gsnedders>
Philip`: Don't forget .tail!

[14:24:01.0000] <gsnedders>
Philip`: That is the most horrible thing ever.

[14:25:00.0000] <Philip`>
Oh, that looks yucky

[14:29:00.0000] <Philip`>
Hmm, 13s to parse the (postprocessed) spec, when using lxml and psyco - that doesn't seem too awful

[14:30:00.0000] <Philip`>
(~16.5s without psyco)

[14:31:00.0000] <hsivonen>
Google should write a Python to JavaScript compiler, too

[14:32:00.0000] <Philip`>
That sounds fun with e.g. Python's yield statement

[14:34:00.0000] <othermaciej>
you can rewrite generators to make the state explicit

[14:34:01.0000] <othermaciej>
or rewrite to continuation passing style

[14:35:00.0000] <Philip`>
Can you do that without having to rewrite the caller too?

[14:38:00.0000] <gsnedders>
peh. get me a decent lexer in Python first :P

[14:47:00.0000] <gsnedders>
(I need recursiveness and repetition at the parser generator level)

[14:47:01.0000] <hsivonen>
http://lists.w3.org/Archives/Public/www-tag/2008May/0206.html

[14:50:00.0000] <hsivonen>
http://lists.w3.org/Archives/Public/www-tag/2008May/0204.html

[14:52:00.0000] <annevk>
www-tag is funny these days

[14:54:00.0000] <Philip`>
Argh, lxml's child iteration seems to be skipping over the one element I want to find :-/

[15:27:00.0000] <hsivonen>
I had an idea: 1) doc.write <math></math>. 2) Sniff what namespace it was put in. 3) Remove it from the DOM. 4) If it wasn't in the MathML namespace, doc.write a script tag for an HTML5 parser in JS. 

[15:27:01.0000] <hsivonen>
5) doc.write <plaintext style='display:none'>

[15:28:00.0000] <hsivonen>
6) attach the tree builder of the JS parser to the html, head and body elements of the current doc

[15:28:01.0000] <hsivonen>
7) take the textContent of the plaintext element

[15:28:02.0000] <hsivonen>
8) Remove plaintext

[15:28:03.0000] <hsivonen>
9) feed its textContent to the parser

[15:28:04.0000] <hsivonen>
feasible?

[15:29:00.0000] <hsivonen>
even subsequent doc.write could work if browsers allow document.write to be replaced with a JavaScript function

[15:30:00.0000] <Hixie>
sounds pretty sick to me :-)

[15:31:00.0000] <Hixie>
what license do you waht the live dom viewer in?

[15:31:01.0000] <hsivonen>
Hixie: MIT license would be the easiest

[15:33:00.0000] <Hixie>
ok, it's available under the MIT license

[15:34:00.0000] <hsivonen>
Hixie: thank you

[15:34:01.0000] <Hixie>
np

[15:35:00.0000] <Philip`>
hsivonen: I did a (very primitive) parser (for parsing a non-HTML language) like that a while ago, and I think IE had some difficulties (which I didn't bother debugging) but otherwise it seemed to work

[15:36:00.0000] <hsivonen>
Philip`: cool

[15:36:01.0000] <hsivonen>
Philip`: IE doesn't have createElementNS anyway

[15:36:02.0000] <Philip`>
http://canvex.lazyilluminati.com/misc/sexp.html

[15:37:00.0000] <hsivonen>
so I ignored IE in my GWT tree builder

[15:41:00.0000] <Hixie>
i'm kinda starting to like the colon

[15:41:01.0000] <Hixie>
for aira

[15:42:00.0000] <Hixie>
it would actually do more harm to xml namespaces than anything else i've tried to do

[15:45:00.0000] <annevk>
fortunately internal consistency is more important than harming XML namespaces

[15:46:00.0000] <Hixie>
i dunno, it would in fact harm xhtml too

[15:47:00.0000] <annevk>
to continue this little debate, harming other languages in order to make HTML look better doesn't seem the best approach

[15:48:00.0000] <annevk>
(especially if you're meanwhile compromising the HTML syntax somewhat)

[15:48:01.0000] <Hixie>
no no, only to make xml fail

[15:49:00.0000] <Hixie>
we still have e.g. svg and mathml in text/html

[15:49:01.0000] <annevk>
i meant that using the colon would not be consistent with HTML

[15:49:02.0000] <Hixie>
true

[15:50:00.0000] <Hixie>
well except for xlink

[15:50:01.0000] <Hixie>
in svg

[15:51:00.0000] <annevk>
:)

[15:52:00.0000] <annevk>
very unfortunate

[15:52:01.0000] <annevk>
MathML uses it too it seems, in some way

[15:52:02.0000] <Hixie>
the colon? how?

[15:54:00.0000] <annevk>
xlink:href

[15:54:01.0000] <Hixie>
oh?

[15:54:02.0000] <Hixie>
not in text/html it doesn't :-)

[15:54:03.0000] <Hixie>
oh well i guess it does actually

[15:54:04.0000] <annevk>
yeah it does

[15:55:00.0000] <annevk>
also in MathML3

[15:55:01.0000] <annevk>
and 2 probably

[15:55:02.0000] <annevk>
it also has xml:space, xml:lang and xml:id (I commented on that one)

[15:55:03.0000] <Hixie>
oh well

[15:55:04.0000] <annevk>
I also see my:color and my:background in the attributes list...

[15:55:05.0000] <annevk>
maybe I should review MathML

[15:55:06.0000] <Hixie>
xml:space does nothing, as far as i can tell, but text/html supports it

[15:55:07.0000] <Hixie>
xml:id isn't supported

[15:56:00.0000] <Hixie>
xlink in general i don't expect to actually be supported

[15:56:01.0000] <Hixie>
though i guess that's up to the UA

[15:56:02.0000] <Hixie>
(xlink being incompatible with svg makes xlink less attractive, so i guess most UAs won't do it)

[15:56:03.0000] <Hixie>
(if they do SVG)

[15:56:04.0000] <annevk>
yeah, makes more sense to simply ignore XLink

[15:57:00.0000] <annevk>
i hope they change SVG slightly while still feasible to not require namespaces for references

[15:57:01.0000] <annevk>
by introducing href= and src= attributes where appropriate

[15:59:00.0000] <Hixie>
i actually hope they don't

[15:59:01.0000] <Hixie>
redundant attributes cause all kinds of problems

[15:59:02.0000] <Hixie>
you have to define priority, you have to test that the priority works even with DOM manipulation, etc

[16:00:00.0000] <Hixie>
it works differently for new and old UAs, etc

[16:00:01.0000] <Hixie>
and it adds no functionality

[16:00:02.0000] <Hixie>
just look at the trouble that charset="" has been

[16:01:00.0000] <annevk>
that's a different case, though I suppose that if we get SVG in text/html without the namespaces requirement it's sort of acceptable

[16:01:01.0000] <annevk>
"MathML uses the namespace URI http://www.w3.org/ns/mathml-cd for the XML encoding of MathML content dictionaries."

[16:01:02.0000] <annevk>
-- http://www.w3.org/TR/MathML3/chapter8.html#contm.mcd.format

[16:03:00.0000] <Hixie>
ooooh

[16:03:01.0000] <Hixie>
david singer just proposed something to me by e-mail that may solve the alt text issue

[16:03:02.0000] <Hixie>
how about instead of a new attribute to say when the alt text is an alternative vs being something just saying what kind of image it is

[16:04:00.0000] <Hixie>
we say that the alt attribute is an alternative unless it starts and ends with a [ and a ] ?

[16:04:01.0000] <Hixie>
alt='photo' is the alt text for a png of the word "photo", alt='[photo]' is the alt text for a photo with no description

[16:04:02.0000] <Hixie>
the UA can still do all the magic with alt, _and_ it's probably quite compatible with legacy usage

[16:04:03.0000] <Hixie>
and it works great in legacy uas

[16:05:00.0000] <Philip`>
How does it work in legacy AT?

[16:05:01.0000] <annevk>
what does <img> mean?

[16:05:02.0000] <Philip`>
(Pronouncing "left square bracket image right square bracket" would be quite irritating)

[16:05:03.0000] <Hixie>
<img> is non-conforming and doesn't mean anything, and UAs use heuristics to decide how to render it

[16:05:04.0000] <Hixie>
Philip`: well let's see

[16:06:00.0000] <Hixie>
/me grabs out his Jaws 8 install from under the bed

[16:06:01.0000] <annevk>
leaving it out still seems nicer

[16:07:00.0000] <Hixie>
leaving it out has issues with legacy content and doesn't convey all the information it could

[16:07:01.0000] <Hixie>
(actually saying what kind of image it is)

[16:07:02.0000] <Philip`>
0.6% of my alt values are of the form "[...]"

[16:08:00.0000] <Hixie>
you can get data faster than i can :-)

[16:08:01.0000] <Philip`>
That's because I can just use grep and perl locally :-)

[16:09:00.0000] <Hixie>
:-)

[16:09:01.0000] <Philip`>
http://philip.html5.org/data/alt-in-square-brackets.txt

[16:10:00.0000] <annevk>
some of those seem like real replacements

[16:10:01.0000] <annevk>
"[Reset Text Size]"

[16:10:02.0000] <Philip`>
(These are total numbers of occurrences, rather than number of pages)

[16:10:03.0000] <annevk>
or "[E-Mail]"

[16:11:00.0000] <annevk>
but maybe that's not too problematic

[16:11:01.0000] <Philip`>
Given legacy content, UAs can't tell that e.g. alt="[Search]" is any different to alt="Search"

[16:12:00.0000] <Philip`>
so either the difference is important, in which case that syntax won't work; or the difference is not important, in which case we don't need any syntax at all and you can just put the not-really-alternative text in the alt attribute directly

[16:13:00.0000] <annevk>
otoh, it's only 0.6% of the images, that's a small percentage according to some people, right? :D

[16:14:00.0000] <Hixie>
Philip`: we can't add _any_ syntax without breaking some images. However, that's moot, since even just keepign alt="" as defined in HTML4 breaks a huge number of images.

[16:14:01.0000] <Philip`>
That's huge compared to how many people will bother adopting whatever requirements we specify for their alt-less critical images :-p

[16:14:02.0000] <Hixie>
hm maybe this should only apply to images that aren't in lnks

[16:14:03.0000] <Hixie>
links

[16:14:04.0000] <Philip`>
so UAs would be better off ignoring the square brackets entirely, because most of the time they'll only be there for tooltip-prettiness instead of for HTML5-spec-conformance

[16:14:05.0000] <Hixie>
that might solve 90% of our problems here

[16:15:00.0000] <Hixie>
Philip`: initially maybe

[16:16:00.0000] <Philip`>
Hixie: I imagine it would take a very long time for a feature that has pretty much no effect for any users to reach 0.6% adoption

[16:16:01.0000] <Hixie>
it has a small effect, it shows an icon for hte image when the user has images disabled. :-)

[16:17:00.0000] <Philip`>
Users don't have images disabled

[16:17:01.0000] <Philip`>
and if they do then they're crazy and don't count

[16:19:00.0000] <Philip`>
and you're assuming some UAs will implement that icon-hiding behaviour, which seems a dangerous assumption since if they thought it was a good idea then they'd probably have hidden the image icons already

[16:22:00.0000] <Philip`>
lxml's text/tail thing is actually quite ugly

[16:24:00.0000] <Hixie>
firefox does hide the image icons

[16:24:01.0000] <Hixie>
has for years

[16:28:00.0000] <Philip`>
Oh, okay

[16:33:00.0000] <Philip`>
It's slightly odd that alt="[..." (8215 matches) is so much more common than alt="{..." (214) and alt="<..." (388) and alt="_..." (285)

[16:35:00.0000] <Hixie>
maybe we should use { for this then

[16:44:00.0000] <Philip`>
http://philip.html5.org/data/alt-first-char.txt

[16:45:00.0000] <Dashiva>
/me wonders what the data will be used for

[16:45:01.0000] <Philip`>
/me shrugs

[16:56:00.0000] <Philip`>
Hmm, html5lib serialiser is slow :-(


2008-05-30
[17:00:00.0000] <annevk>
I almost replied to html4all's latest e-mail

[17:00:01.0000] <annevk>
meh

[17:40:00.0000] <Hixie>
http://lists.w3.org/Archives/Public/www-archive/2008May/0105.html

[17:46:00.0000] <Dashiva>
"Again, we need an editor who uses his judgment in editing the draft. However, we need that judgment to be informed by the opinions of others in the WG."

[17:46:01.0000] <Dashiva>
He needs to make up his mind

[17:50:00.0000] <Hixie>
http://www.secinfo.com/d11MXs.2WQ9.d.htm is probably a good test case for html5 browsers

[17:51:00.0000] <Hixie>
(my own html5 parser runs out of element id bits parsing it)

[17:51:01.0000] <Hixie>
(so that means it has more than 64k elements)

[17:51:02.0000] <Hixie>
hsivonen's parser has a fatal error on line 236

[17:51:03.0000] <Hixie>
validator.w3.org returns a 500

[17:51:04.0000] <Hixie>
the html5 dumper just returns no data

[17:52:00.0000] <Hixie>
html5lib dumper, i mean

[17:52:01.0000] <Dashiva>
What is that thing...

[17:55:00.0000] <Dashiva>
4.5 MB. gEBTN says 103430 elements, among them 58 tables, 39093 cells (and one <th>).

[17:59:00.0000] <Philip`>
html5lib seems happy to parse it, and it only takes a minute

[18:37:00.0000] <Hixie>
http://gareth53.blogspot.com/2008/05/atmedia2008-day-1-notes.html

[18:37:01.0000] <Hixie>
very, very rough notes

[18:37:02.0000] <Hixie>
"HTML5 by Jame Hunt & Lachlan Young"

[18:37:03.0000] <Hixie>
"html and xhtml - smae syntax, dif vocab"

[18:37:04.0000] <Hixie>
some amusing typos there :-)

[18:37:05.0000] <Hixie>
but the rest of the notes are pretty good

[18:37:06.0000] <Hixie>
bbl

[18:43:00.0000] <Philip`>
Hmm, the spec splitter runs in 20s on my laptop but 105s on my web server :-(

[18:53:00.0000] <Dashiva>
Double capitals are a plague on modern society

[18:53:01.0000] <Dashiva>
"move from IE^ to IE&"

[19:29:00.0000] <takkaria>
/me sighs at RB

[22:51:00.0000] <roc>
annevk, shepazu: you around?

[22:52:00.0000] <shepazu>
a little

[22:52:01.0000] <shepazu>
hitting the sack soon, kinda tired

[00:18:00.0000] <annevk>
/me is now

[00:32:00.0000] <othermaciej>
so did Google announce anything interesting at Google I/O?

[00:32:01.0000] <othermaciej>
I heard something about Gears announcements

[00:32:02.0000] <othermaciej>
(#whatwg is my lazyweb)

[00:35:00.0000] <hdh>
Hixie said something about his VP publicly endorse HTML5, there was a transcript link too

[00:35:01.0000] <annevk>
Gears and Android I saw

[00:35:02.0000] <hdh>
http://www.techcrunch.com/2008/05/28/live-from-google-io/

[00:36:00.0000] <othermaciej>
I heard they announced various sites will use Gears, I wonder if those will work with browsers that implement the relevant HTML5 features

[00:37:00.0000] <annevk>
I saw somewhere that they believed (maybe hoped) Gears would eventually become obsolete, so I hope so, but I'm not sure

[00:44:00.0000] <hdh>
the implement status box is covering part of the spec text, how can I make it go away?

[00:44:01.0000] <annevk>
by using the w3.org version

[00:47:00.0000] <hdh>
user stylesheet FTW

[00:55:00.0000] <hsivonen>
Hixie: wouldn't [] in alt degrade ungracefully?

[00:56:00.0000] <hsivonen>
Hixie: also, if a validator complained about missing alt, how would you make the author insert [] instead of e.g. file name?

[01:05:00.0000] <hsivonen>
http://parsetree.validator.nu/?doc=http%3A%2F%2Fwww.secinfo.com%2Fd11MXs.2WQ9.d.htm WFM without error 500

[01:05:01.0000] <hsivonen>
looks like the HTTP connection times out, though

[01:06:00.0000] <hsivonen>
(Validator.nu has deliberate short timeouts)

[01:09:00.0000] <annevk>
it prints IOException at the end

[01:11:00.0000] <hsivonen>
annevk: I'm guessing it's a timeout

[01:14:00.0000] <annevk>
"In the meaning of technical in the W3C policies it is certainly a technical question whether the document is published or not."

[01:14:01.0000] <annevk>
-- RB

[01:14:02.0000] <annevk>
-- http://www.w3.org/2002/09/wbs/40318/wd-html5-may/results

[01:14:03.0000] <annevk>
/me laughs

[01:18:00.0000] <gavin_>
haha

[02:25:00.0000] <annevk>
hsivonen, http://ajaxian.com/archives/gwt-15-release-candidate-announced

[02:25:01.0000] <hsivonen>
annevk: that's what I'm using

[02:27:00.0000] <othermaciej>
so if you combine GWT and Rhino on Rails, it seems that Google's emerging strategy is to run Java on the client and JavaScript on the server

[02:28:00.0000] <othermaciej>
(someone should tell them they've got it backwards)

[02:31:00.0000] <Hixie>
GWT compiles the Java language code to JS for the client

[02:32:00.0000] <hsivonen>
GWT's <super-source/> needs *much* better docs

[02:32:01.0000] <Hixie>
so really it'd just be using JS as a bytecode :-)

[02:35:00.0000] <othermaciej>
I wonder if it would be worth making a restricted-JS-to-JS optimizer

[02:35:01.0000] <othermaciej>
I mean, I know Google has one internally

[02:35:02.0000] <othermaciej>
but available to the general public

[02:35:03.0000] <othermaciej>
for example JS implementations cannot soundly do inlining in the general case, but you could pre-inline simple functions if the author promises not to use arguments or caller properties on functions

[02:37:00.0000] <Hixie>
i believe the Caja project has something like that

[02:37:01.0000] <othermaciej>
(though I am not sure what else if anything is better done as a pre-pass; maybe if the author makes some promises about not doing weirdshit with getters and setters you can do some common subexpression elimination)

[02:37:02.0000] <othermaciej>
Caja is a restricted dialect, is it not?

[02:37:03.0000] <othermaciej>
oh

[02:37:04.0000] <Hixie>
yeah but the optimiser might be separate, i'm not sure

[02:37:05.0000] <othermaciej>
it's restricted in a more severe way than what I have in mind

[02:38:00.0000] <Hixie>
do you have any opinions on whether it is worth adding a feature to make tabindex scoped in html5?

[02:38:01.0000] <hsivonen>
aaronlev: ^

[02:38:02.0000] <othermaciej>
it's an interesting idea

[02:38:03.0000] <othermaciej>
though I tend to think 0 and -1 are the only tabindex values anyone should ever use

[02:39:00.0000] <othermaciej>
I guess scoping makes it potentially possible to use other values in a non-broken way

[02:39:01.0000] <aaronlev>
scoped in what way?

[02:40:00.0000] <othermaciej>
aaronlev: I assume what Hixie means is that you can have a container element and set up the tab order inside it, but without affecting tab order of things before the container or after the container

[02:41:00.0000] <othermaciej>
which would solve the problem that tabindex is currently all-or-nothing per page

[02:42:00.0000] <aaronlev>
yeah that's a problem, but i'm not sure we'd want to use tabindex with a new meaning

[02:42:01.0000] <aaronlev>
otherwise what about older browsers

[02:42:02.0000] <aaronlev>
another problem is reading order for a screen reader

[02:42:03.0000] <aaronlev>
it's like tabindex but it can link containers that aren't necessarily focusable

[02:42:04.0000] <aaronlev>
i should say it = aria-flowto

[02:43:00.0000] <othermaciej>
in an ideal world, document order would be set up right for reading/tabbing order and you use CSS to visually reorder things if needed

[02:43:01.0000] <aaronlev>
yes

[02:43:02.0000] <othermaciej>
but that does not work well with some table setups

[02:43:03.0000] <annevk>
what are the use cases?

[02:44:00.0000] <othermaciej>
annevk: making a reusable control that drops into any page, which has internal focusable items but their logical tab order is not document order

[02:44:01.0000] <aaronlev>
for flowto or modified tabindex?

[02:44:02.0000] <othermaciej>
perhaps a grid control where you want tab to go vertically first, then to the next column, etc

[02:44:03.0000] <aaronlev>
for flowto, you might have a sidebar for an article, and want to decide where the best point in the article to branch off to the sidebar is

[02:45:00.0000] <aaronlev>
because navigation is not just abotu focusable objects

[02:45:01.0000] <othermaciej>
sidebars are an inherently nonlinear form of document organization

[02:45:02.0000] <aaronlev>
yes things like PDF have dealt with this for years

[02:45:03.0000] <aaronlev>
you have to decide a reading order

[02:45:04.0000] <aaronlev>
also there is printing

[02:45:05.0000] <annevk>
sounds like XBL / datagrid

[02:46:00.0000] <annevk>
and menu

[02:46:01.0000] <aaronlev>
i remember similar problems turning page layout files into something for braille embossers

[02:46:02.0000] <aaronlev>
in that case the doc is a set of pages with each page having a set of objects

[02:46:03.0000] <aaronlev>
kind of a mess

[02:46:04.0000] <othermaciej>
annevk: XBL doesn't give you anything to control tab order

[02:46:05.0000] <othermaciej>
(afaik)

[02:46:06.0000] <aaronlev>
i don't know, does HTML have any goal to go into more of a word processing / publishing space?

[02:46:07.0000] <othermaciej>
annevk: the problem is that tabindex as specified currently basically can't be used

[02:47:00.0000] <othermaciej>
because everything with an explicit tabindex comes before all other things

[02:47:01.0000] <othermaciej>
HTML doesn't have goals, it's just a markup language

[02:48:00.0000] <aaronlev>
othermaciej: don't you know what i mean, the most active leaders in html-wg

[02:48:01.0000] <othermaciej>
I imagine some people have goals to use HTML for more word-processy things

[02:48:02.0000] <aaronlev>
you guys are some of them

[02:48:03.0000] <aaronlev>
thanks

[02:48:04.0000] <othermaciej>
I think a lot of the action in the HTML space right now is about using it as an application format

[02:48:05.0000] <othermaciej>
I would say Web applications are a more interesting target than Web documents right now

[02:48:06.0000] <aaronlev>
another interesting use case is key navigable diagrams

[02:49:00.0000] <aaronlev>
if i'm in an org chart what's the navigation order

[02:49:01.0000] <othermaciej>
if the org chart is done in SVG, you can make document order anything you want and use that to control the tab order

[02:49:02.0000] <aaronlev>
changing tabindex in a non-compatible way isn't as good as implementing something new which takes care of the needs completely

[02:50:00.0000] <othermaciej>
I'm not sure that when Hixie says "scoped tabindex" he necessarily literally means to reuse the tabindex attribute

[02:50:01.0000] <othermaciej>
of course, we just had a long conversation about it without him clarifying what he meant

[02:50:02.0000] <annevk>
othermaciej, XBL + nav-index from CSS prolly does

[02:50:03.0000] <aaronlev>
the truth is, within many containers the last active element is in the tab order, but as a whole the container is just in the tab order once

[02:50:04.0000] <othermaciej>
annevk: nav-index?

[02:50:05.0000] <aaronlev>
within the container directional keys may be used to navigate

[02:50:06.0000] <othermaciej>
annevk: I don't think CSS is the right place to specify tab order

[02:50:07.0000] <aaronlev>
tabbing is a limited way to think about navigation really

[02:51:00.0000] <annevk>
othermaciej, why not, if you reorder items, you might want to change tab order too

[02:51:01.0000] <hsivonen>
I'm inclined to think tabindex shouldn't change for compat reasons

[02:51:02.0000] <othermaciej>
tab ordering things is tied to the content

[02:51:03.0000] <hsivonen>
and that authors should always use 0 or -1

[02:52:00.0000] <othermaciej>
and is behavioral, not presentational

[02:52:01.0000] <othermaciej>
and if nav-index is not itself scoped, it does not solve the problem

[02:52:02.0000] <annevk>
othermaciej, if I position one link before the order in one layout and the other way around in another, I might want to change tab order as well

[02:52:03.0000] <hsivonen>
it seems to me that in foo-index, "index" is the problem

[02:52:04.0000] <othermaciej>
I can imagine you could implement scoped tabindex in a way that older browsers would see tabindex="0"

[02:52:05.0000] <hsivonen>
and ARIA's IDREF-based flow makes more sense

[02:53:00.0000] <othermaciej>
that is true

[02:53:01.0000] <othermaciej>
a navigation order should be about next/previous items

[02:53:02.0000] <othermaciej>
not absolute index

[02:53:03.0000] <annevk>
hsivonen, there's directional navigation properties as well

[02:54:00.0000] <annevk>
http://www.w3.org/TR/css3-ui/#keyboard

[02:54:01.0000] <othermaciej>
nav-index is broken in the same way as tabindex

[02:55:00.0000] <othermaciej>
up/down/left/right do not substitute for previous or next

[02:55:01.0000] <othermaciej>
and I still think defining keyboard navigation in CSS does not make sense

[02:55:02.0000] <othermaciej>
even though CSS can change the visual order

[02:55:03.0000] <hsivonen>
does C++ have something that roughly matches the semantics of Java's static { ... } block?

[02:56:00.0000] <hsivonen>
i.e. code that runs once when a .dll/.so/.dylib is loaded?

[02:56:01.0000] <othermaciej>
if you tell me what Java's static { } block does I can probably tell you

[02:56:02.0000] <othermaciej>
oh

[02:56:03.0000] <othermaciej>
nothing that you should actually use

[02:56:04.0000] <othermaciej>
if you have an extern or static global variable of class type, the constructor runs on startup

[02:57:00.0000] <othermaciej>
but the order is ill-defined

[02:57:01.0000] <othermaciej>
and it may or may not work right in shared libraries on various systems

[02:57:02.0000] <hsivonen>
:-(

[02:57:03.0000] <othermaciej>
it is better to have all entry points call an initIfNeeded() type function

[02:57:04.0000] <hsivonen>
ok

[02:58:00.0000] <othermaciej>
(I think order of global constructor calls is defined within a translation unit but not between translation units)

[02:58:01.0000] <hsivonen>
or perhaps if the V.nu parser was ported to C++, the static blocks could be replaced with some tormented preprocessor/template code

[02:59:00.0000] <hsivonen>
I use static {} blocks to initialize a couple of sorted arrays with magic data in them

[02:59:01.0000] <hsivonen>
I *could* generate the arrays in the right order ahead of compile time

[03:00:00.0000] <hsivonen>
but then I'd need an additional script to keep in sync

[03:00:01.0000] <othermaciej>
for cases of magic data arrays

[03:00:02.0000] <othermaciej>
what I like to do is generate the source code for the array constant from a template file by script

[03:00:03.0000] <othermaciej>
(we do that for CSS properties, tag names, attribute names, CSS property keywords, JS language keywords, etc)

[03:01:00.0000] <hsivonen>
ok. thanks. I'll do that if the need arises

[03:10:00.0000] <Hixie>
sorry, got distracted

[03:10:01.0000] <Hixie>
this would be a new attribute

[03:10:02.0000] <Hixie>
which modifies the behaviour of tabindex

[03:11:00.0000] <Hixie>
so that numbers greater than 0 only consider other tabindex attributes within that same scope

[03:11:01.0000] <Hixie>
i tend to think that maybe it's not worth it

[03:11:02.0000] <hsivonen>
Hixie: what's the compat story?

[03:12:00.0000] <Hixie>
none really, this likely wouldn't be used until this generation was obsolete

[03:12:01.0000] <zcorpan>
hsivonen: in my experience, generally, pages that use tabindex today have bad usability in today's browsers anyway

[03:12:02.0000] <Hixie>
aha, zcorpan! i was looking for you earlier

[03:12:03.0000] <Hixie>
this was originally your diea

[03:12:04.0000] <Hixie>
idea

[03:13:00.0000] <zcorpan>
hey Hixie

[03:13:01.0000] <othermaciej>
Hixie: I think a way to specify next element in the tab order by ID would be more useful, if any such thing is to be added

[03:14:00.0000] <othermaciej>
(and then you could also specify tabindex="0" if needed or not give a tabindex, for compat)

[03:14:01.0000] <othermaciej>
but I am not sure it is critical in any case

[03:15:00.0000] <Hixie>
yeah

[03:16:00.0000] <zcorpan>
Hixie: i think source order is a powerful enough tool to specify tab order, and browsers would probably be more usable on legacy content if all positive values were treated as 0

[03:17:00.0000] <othermaciej>
zcorpan: tables are about the only case where you may want a tab order that is not natural source order, and it's hard to fix the source order w/o breaking the layout

[03:17:01.0000] <Hixie>
yeah. the spec allows that, iirc

[03:17:02.0000] <othermaciej>
(at least so far as I can think of)

[03:17:03.0000] <Hixie>
it seems like this might be a CSS thing anyway

[03:17:04.0000] <Hixie>
css3 ui does try to take this responsibility

[03:17:05.0000] <zcorpan>
othermaciej: true

[03:17:06.0000] <aaronlev>
Should be able to point to id of a non-focusable container

[03:18:00.0000] <aaronlev>
as next thing to navigate to

[03:18:01.0000] <othermaciej>
it should be able to point to anything, and tabbing will find the next item in document order from there

[03:18:02.0000] <zcorpan>
othermaciej: wasn't there a proposal to swap the direction of tables so that rows become columns?

[03:19:00.0000] <othermaciej>
zcorpan: I do not recall such a thing but I have not followed every proposal

[03:19:01.0000] <othermaciej>
that particular proposal sounds like it would have a lousy compat story

[03:20:00.0000] <aaronlev>
othermaciej: after tabbing past the end of the container pointed to, should tabbing continue after the container? Or should it continue on from the original point that branched to the container

[03:20:01.0000] <aaronlev>
iow

[03:20:02.0000] <aaronlev>
i might want to say, tab through all the stuff over there before continuing on here

[03:20:03.0000] <aaronlev>
more like a subroutine than giving up control

[03:21:00.0000] <othermaciej>
sounds dubious

[03:21:01.0000] <othermaciej>
if you want that, the target container can point back to after you

[03:21:02.0000] <othermaciej>
tabbing into the same container from two different points does not sound like good UI

[03:22:00.0000] <othermaciej>
Cocoa just lets you specify the previous item, the next item, and whether any given item is keyboard focusable

[03:22:01.0000] <othermaciej>
it works pretty well

[03:24:00.0000] <aaronlev>
othermaciej: but that's not sufficient for documents

[03:24:01.0000] <aaronlev>
because for screen readers the reading order of doc text needs to match the nav order or it gets confusing

[03:24:02.0000] <aaronlev>
so it's better if each container points to the next

[03:24:03.0000] <aaronlev>
plus it's fewer things to change

[03:24:04.0000] <aaronlev>
for the author

[03:25:00.0000] <othermaciej>
I just suggested each thing pointing to the next

[03:25:01.0000] <othermaciej>
I don't think I understand what is different about your proposal

[03:25:02.0000] <othermaciej>
(other than being about more than just tab order)

[03:26:00.0000] <aaronlev>
it might not be different, just IRC communication ambiguity

[03:26:01.0000] <othermaciej>
that is possible

[03:26:02.0000] <othermaciej>
back in a few

[03:54:00.0000] <hsivonen>
boohoo. I get unending unresponsive script dialogs when I try to run the script GWT compiled

[03:56:00.0000] <hsivonen>
hah. The script crashes Drosera.

[04:03:00.0000] <zcorpan>
Error: = in an unquoted attribute value.

[04:03:01.0000] <zcorpan>
At line 234, column 38

[04:03:02.0000] <zcorpan>
istrant.asp?CIK=30908

[04:03:03.0000] <zcorpan>
http://validator.nu/?doc=http%3A%2F%2Fwww.secinfo.com%2Fd11MXs.2WQ9.d.htm

[04:03:04.0000] <zcorpan>
hsivonen: looks like a false positive

[04:05:00.0000] <hsivonen>
zcorpan: per spec http://www.whatwg.org/specs/web-apps/current-work/#attribute4

[04:05:01.0000] <zcorpan>
hsivonen: yes i know

[04:05:02.0000] <zcorpan>
hsivonen: still, false positive in the sense that it wasn't a mistake

[04:05:03.0000] <Hixie>
hsivonen: drosera crashes a lot for me

[04:06:00.0000] <Hixie>
anyway bed time

[04:06:01.0000] <Hixie>
nn

[04:08:00.0000] <annevk>
zcorpan, hmm yeah, escaping = in URIs would be annoying

[04:09:00.0000] <zcorpan>
annevk: the straightforward "fix" is to use quotes

[04:10:00.0000] <annevk>
meh

[04:34:00.0000] <Philip`>
zcorpan: Forbidding = also detects errors like <img src=whatever.gifalt=something>, which wouldn't be caught by forbidding " and '

[04:34:01.0000] <zcorpan>
Philip`: true

[04:34:02.0000] <annevk>
you'd catch that one by looking at the page

[04:35:00.0000] <annevk>
(though not for <a rel=xtitle=foo> I suppose)

[04:35:01.0000] <othermaciej>
drosera is obsolete

[04:36:00.0000] <hsivonen>
othermaciej: interesting. has the replacement been shipped?

[04:36:01.0000] <othermaciej>
hsivonen: yeah, debugging is built into the Web Inspector now

[04:36:02.0000] <othermaciej>
the latest WebKit nightlies don't even include Drosera

[04:36:03.0000] <hsivonen>
oh. I was unaware

[04:36:04.0000] <hsivonen>
thanks

[04:37:00.0000] <othermaciej>
we haven't widely announced it yet, but we should soon

[04:37:01.0000] <othermaciej>
the new debugger is much more stable and performance

[04:37:02.0000] <othermaciej>
*performant

[04:38:00.0000] <zcorpan>
Philip`: perhaps we should pages that have = in unquoted attributes to get an idea whether it helps or hurts

[04:38:01.0000] <zcorpan>
s//look at/

[04:41:00.0000] <zcorpan>
yay.. http://lists.w3.org/Archives/Public/w3c-wai-ig/2008AprJun/0087.html

[04:44:00.0000] <zcorpan>
(aria editor's drafts are now public)

[05:18:00.0000] <Philip`>
zcorpan: http://www.rottentomatoes.com/m/changing_lanes/ -- <iframe valign=top WIDTH= HEIGHT= MARGINWIDTH=0 ...>

[05:19:00.0000] <Philip`>
http://www.ahaus-online.de/ -- <table width=100%cellpadding="4" cellspacing="1">

[05:19:01.0000] <zcorpan>
Philip`: ooh interesting

[05:20:00.0000] <Philip`>
Most of the cases look like <a href=foo?bar=baz>

[05:22:00.0000] <annevk>
<table width> is not conforming anyway :)

[05:22:01.0000] <Philip`>
zcorpan: If you want some extremely rough raw data, see http://philip.html5.org/data/some-unquoted-attributes-with-equals.txt

[05:23:00.0000] <Philip`>
(That's all lines (for the first n pages before I got bored and stopped it) matching /<[a-zA-Z0-9:]+[^"'>]+\s[a-zA-Z0-9:]+\s*=\s*[^\s>"']+=/ which seems to kind of work alright as an approximation)

[05:24:00.0000] <Philip`>
http://www.specialtybottle.com/: <INPUT type=hidden value=BGCOLOR=#FFFFFF name=nonUPS_body>

[05:24:01.0000] <zcorpan>
Philip`: don't you think the value was intended to be "BGCOLOR=#FFFFFF"?

[05:25:00.0000] <zcorpan>
Philip`: the other two cases would get caught by other errors

[05:25:01.0000] <Philip`>
Oh

[05:25:02.0000] <Philip`>
That's possible

[05:26:00.0000] <Philip`>
http://www.gencat.net:8000/osial/owa/p01.dad_ens?cod=1708650006: <TD VALIGN=TOP BGCOLOR=6A92E3><B><FONT FACE=ARIALSIZE=1 color="#FFCC66">DADES GENERALS</FONT></B></TD>

[05:26:01.0000] <Philip`>
<table width= otherstuff=etc> seems surprisingly common

[05:27:00.0000] <zcorpan>
seems people think that whitespace starts a new attribute

[05:27:01.0000] <zcorpan>
perhaps that should be banned? <p class= foo>?

[05:28:00.0000] <gsnedders>
I should probably blog something this month before this month ends.

[05:29:00.0000] <annevk>
i don't think so

[05:30:00.0000] <Philip`>
gsnedders: For any particular reason?

[05:30:01.0000] <gsnedders>
Philip`: Because I normally blog at least once per month :P

[05:32:00.0000] <zcorpan>
<TABLE width=240 height=300 border=0<cellspacing=0 cellpadding=0 bgcolor="083a81">

[05:35:00.0000] <Philip`>
Far too many people use <table> :-/

[05:35:01.0000] <zcorpan>
<FONT face=Arial,color=#ffffcc>

[05:38:00.0000] <Philip`>
(and <font>)

[05:39:00.0000] <Philip`>
(and uppercase tag names)

[05:39:01.0000] <Philip`>
(and unquoted attributes)

[05:39:02.0000] <Philip`>
(XHTML is much prettier, except for all the closing-tag stuff)

[05:43:00.0000] <zcorpan>
<a class=chlnk href onclick=this.style.behavior='url(#default#homepage)';this.setHomePage('http://www.elarcadenoe.org'); style=CURSOR:hand>

[05:44:00.0000] <zcorpan>
i think that's the first time i've seen <a href> in the wild

[05:45:00.0000] <virtuelv>
hm, new favicon for Google

[05:49:00.0000] <Philip`>
zcorpan: You haven't been looking at many pages if that's the first <a href> you've seen :-p

[05:50:00.0000] <zcorpan>
Philip`: i mean as opposed to <a href=''>

[05:50:01.0000] <zcorpan>
or <a href="">

[05:50:02.0000] <Philip`>
I mean that too

[05:50:03.0000] <zcorpan>
ok

[05:51:00.0000] <Philip`>
But I can cheat and use grep instead of actually looking at pages :-)

[05:51:01.0000] <zcorpan>
:P

[05:51:02.0000] <Philip`>
Hmm, maybe it's not actually extremely common

[05:52:00.0000] <Philip`>
http://philip.html5.org/data/some-minimised-href.txt is some (limited to those where the href is the first attribute)

[05:53:00.0000] <zcorpan>
still not many pages

[05:53:01.0000] <Philip`>
That's because my computer's too slow and hasn't finished checking all the pages I've got :-(

[05:54:00.0000] <zcorpan>
hmm, <meta name="keywords" content="<meta name=verify-v1 content=B+yjuq6j1MLG1uzPsO3Z2gR2ECWibgtutlpJeA/3g7k= />">

[05:54:01.0000] <Philip`>
http://dementia.seesaa.net/: <a href><a href="http://dementia.seesaa.net/article/9909416.html">FmÇÌè`</a><br>

[05:55:00.0000] <Philip`>
(http://philip.html5.org/data/some-minimised-href.txt now updated a bit)

[06:00:00.0000] <Philip`>
http://www.pscelex.com/: <a href HREF="quotes.html" class="psc1">Quotes</a>&nbsp;&nbsp;&nbsp;

[06:00:01.0000] <Philip`>
The HTML5 parser breaks that

[06:00:02.0000] <Philip`>
but maybe that's okay since Firefox breaks it already

[06:01:00.0000] <Philip`>
...and so does IE

[06:01:01.0000] <Philip`>
(but not Opera)

[06:02:00.0000] <Philip`>
(Uh... Not Opera 9.2)

[06:02:01.0000] <Philip`>
(but 9.5 does break it)

[06:04:00.0000] <annevk>
9.5 has bug fixes :)

[06:05:00.0000] <Philip`>
And bugs, since it says there's two 'href' attributes on the element :-p

[06:05:01.0000] <Philip`>
(both with value "")

[06:06:00.0000] <Philip`>
(though innerHTML shows the two attributes with distinct values)

[06:07:00.0000] <Philip`>
zcorpan: http://philip.html5.org/data/some-minimised-href.txt is now finished updating, showing all lines matching /(?i)<a\s*href\b\s*[>a-z]/

[06:08:00.0000] <Philip`>
so I guess you were right that it's not particularly common :-(

[06:08:01.0000] <zcorpan>
http://tinyurl.com/5e2c7s - trimmed list (removed rows with src=, href=, action=, rows that were just js, and rows that contained URLs with = )

[06:09:00.0000] <zcorpan>
uh

[06:09:01.0000] <zcorpan>
why doesn't it work

[06:09:02.0000] <Philip`>
Works for me

[06:09:03.0000] <Philip`>
assuming 'it' == that tinyurl

[06:10:00.0000] <Philip`>
and not in Firefox

[06:10:01.0000] <zcorpan>
http://preview.tinyurl.com/5e2c7s works 

[06:12:00.0000] <Philip`>
<META name= robots= content "INDEX, FOLLOW" > is impressively wrong

[06:53:00.0000] <zcorpan>
Philip`: yeah

[06:57:00.0000] <zcorpan>
hsivonen: <!DOCTYPE html><title></title><embed x "y src=''> validates in v.nu

[07:02:00.0000] <zcorpan>
uh, i can't spell to majority

[07:05:00.0000] <zcorpan>
hsivonen: trying to validate the empty string using the textarea doesn't seem to work (the textarea becomes disabled)

[07:06:00.0000] <Philip`>
zcorpan: foo=="bar" wouldn't have shown up because my regexp didn't match that

[07:08:00.0000] <zcorpan>
Philip`: ah

[07:09:00.0000] <Philip`>
but if I do search specifically for <something something== then I find mostly

[07:09:01.0000] <Philip`>
  if (p<1 || p==(val.length-1)) errors+='- '+nm+' must contain an e-mail address.\n';

[07:09:02.0000] <Philip`>
except for:

[07:09:03.0000] <Philip`>
http://www.epinions.com/game-Titles-Consoles-SonyPlaystation-All-Brave_Fencer_Musashi:     <span class=lkr>Enter your ZIP code to see if there is tax & shipping &nbsp;</span><span class="rkr"><input name="zip" type="text" value="Your ZIP" maxlength="9" size="10" style="color:#7D7D7D" onfocus="javascript:clearDefault(this)">&nbsp;&nbsp;<input name=="zip_submit" align="middle" type="submit" value="Enter"></span><br>

[07:09:04.0000] <Philip`>
http://www.angelfire.com/movies/bcrudup/: <font color==#9370D8>

[07:09:05.0000] <Philip`>
http://www.delhieducation.net/: <td colspan==2 height=10>

[07:10:00.0000] <Philip`>
http://www.watsonvillehs.net/:  <TD BACKGROUND=="http://www.watsonvillehs.net/images/departments07/home_bw.jpg" ALIGN="CENTER" >

[07:10:01.0000] <zcorpan>
so it does occur, that's good to know

[07:10:02.0000] <Philip`>
http://www.ottosell.de/jds/jdsworks.htm:                <meta name=="revisit-after" content="20 days">

[07:10:03.0000] <Philip`>
http://www.caribbean-travel.com/airnegril/:     <P align=left><FONT color=red face=Arial size==+2><STRONG>In US Toll

[07:10:04.0000] <Philip`>
and more

[07:10:05.0000] <zcorpan>
feel free to reply that to the list :)

[07:12:00.0000] <mpt>
Philip`, stop it, you're depressing

[07:12:01.0000] <mpt>
:-)

[07:13:00.0000] <Philip`>
mpt: Don't blame me, blame the web :-)

[07:26:00.0000] <gsnedders>
Damned web.

[07:26:01.0000] <gsnedders>
We need Web 2.0!

[07:26:02.0000] <gsnedders>
Oh, wait.

[07:26:03.0000] <gsnedders>
Web π!

[07:28:00.0000] <zcorpan>
gsnedders: is Web π where authors don't make mistakes?

[07:28:01.0000] <gsnedders>
zcorpan: Yeah, and where XHTML 2.0 and XML validating processors are used

[07:29:00.0000] <zcorpan>
gsnedders: then hsivonen would be out of business

[07:30:00.0000] <gsnedders>
zcorpan: True, but Web π is an idealistic dream, and nothing more.

[07:43:00.0000] <takkaria>
I think it's a bad ideal, too

[08:17:00.0000] <Philip`>
gsnedders: http://benjamin.smedbergs.us/blog/2008-05-30/google-cache-only-spam/

[08:21:00.0000] <gsnedders>
Philip`: Yeah, he emailed us too

[08:23:00.0000] <takkaria>
hmm, http://robert.accettura.com/blog/2008/05/29/rebreaking-the-web/

[08:28:00.0000] <annevk>
Rome wasn't built in a day

[08:28:01.0000] <gsnedders>
build('Rome', 86400);

[08:28:02.0000] <gsnedders>
Done.

[08:31:00.0000] <takkaria>
I'm just not sure what he proposes as an alternative

[08:31:01.0000] <annevk>
no

[09:23:00.0000] <annevk>
the stronger people advocate modularization, the more I start to think it's not that easy

[09:40:00.0000] <gsnedders>
School now officially sucks

[09:40:01.0000] <gsnedders>
I am no longer in Secondary 5.

[09:40:02.0000] <gsnedders>
I am now in Secondary 6 :(

[10:33:00.0000] <hsivonen>
is there a reason why <!DOCTYPE html><title></title><embed x "y src=''> should not validate per spec?

[10:41:00.0000] <annevk>
yeah "y is not conforming

[10:44:00.0000] <hsivonen>
annevk: which part of the spec says that "y is not conforming?

[10:45:00.0000] <Philip`>
hsivonen: http://www.whatwg.org/specs/web-apps/current-work/multipage/section-writing0.html#attributes1

[10:45:01.0000] <Philip`>
"Attribute names must consist of one or more characters other than the space characters, U+0000 NULL, U+0022 QUOTATION MARK ("), U+0027 APOSTROPHE ('), U+003E GREATER-THAN SIGN (>), U+002F SOLIDUS (/), and U+003D EQUALS SIGN (=) characters, the control characters, and any characters that are not defined by Unicode."

[10:46:00.0000] <hsivonen>
Philip`: but the writing section isn't normative for conformance checking

[10:46:01.0000] <hsivonen>
Philip`: it's just a restatement of something else

[10:46:02.0000] <hsivonen>
(potentially a buggy restatement)

[10:47:00.0000] <Philip`>
If it was just a restatement of a normative requirement, it shouldn't say "must", I would have thought

[10:49:00.0000] <hsivonen>
time to send email, I guess

[10:55:00.0000] <annevk>
whoa, why does Mozilla want to make Access Control so damn complicated :( 

[10:58:00.0000] <hsivonen>
whoa! when did Hixie put " and ' back as tokenizer parse errors in the attribute name state?

[10:59:00.0000] <hsivonen>
hmm. I even have it implemented

[11:00:00.0000] <hsivonen>
the problem is in the previous state

[11:00:01.0000] <hsivonen>
/me is puzzled

[11:02:00.0000] <hsivonen>
hmm. it only happens if the previous attribute is a boolean attribute

[11:03:00.0000] <hsivonen>
email sent

[11:05:00.0000] <hsivonen>
I need to go through all the intra-tag parse errors and add probable causes to the messages

[11:07:00.0000] <hsivonen>
annevk: it seems to me that the AC cookie issue could be mitigated by instructing servers to make same-site-XHR duplicate the cookie value in a custom header

[11:09:00.0000] <hsivonen>
annevk: it might be worthwhile to add a Note about verifying that a request is same-site by checking for a custom header that duplicates the cookie

[11:09:01.0000] <hsivonen>
annevk: since non-same-origin site could not have read the cookie in order to set the custom header

[11:10:00.0000] <hsivonen>
(trick from the GWT security guide)

[12:09:00.0000] <hsivonen>
http://www.oreillynet.com/xml/blog/2008/05/ubiquity_xforms.html

[12:14:00.0000] <hsivonen>
what's the relationship of Backplane Ltd. and x-port.net Ltd.?

[12:15:00.0000] <hsivonen>
they seem to share the address and the managing director

[12:15:01.0000] <annevk>
hsivonen, if it is a real issue suggesting workarounds for servers is probably not a good solution

[12:16:00.0000] <annevk>
hsivonen, because people won't do it

[12:16:01.0000] <annevk>
I'm not really convinced it's a real issue though

[12:16:02.0000] <hsivonen>
annevk: Google seems to be serious about suggesting cookie duplication as an anti-CSRF measure

[12:17:00.0000] <hsivonen>
it's already needed for protecting same-site XHR endpoints that could be fooled by cross-site form POST or the like

[12:43:00.0000] <Philip`>
Handy hint: When upgrading GRUB on Gentoo, check whether it says it's going to make your system unbootable unless you do the MBR setup process again

[13:32:00.0000] <hsivonen>
w00t. I have the V.nu parser runnig as JS

[13:32:01.0000] <hsivonen>
it needs a setTimeout event pump before anyone can actually use it for anything useful

[13:40:00.0000] <Philip`>
:-o

[14:24:00.0000] <hsivonen>
so firebug crashes on me when I try to debug an unresponsive script

[14:25:00.0000] <hsivonen>
and Web Inspector in WebKit beach balls on unresponsive script and doesn't offer to debug it

[14:29:00.0000] <hsivonen>
looks like there'd be a use case for Canvas 3D: http://googlesystem.blogspot.com/2008/05/google-earth-browser-plugin.html

[14:49:00.0000] <kingryan>
hsivonen: how are you running it as javascript?

[14:50:00.0000] <hsivonen>
kingryan: compiled it with GWT with a JS entry point

[14:50:01.0000] <hsivonen>
however, the tokenizer eof() method goes into an infinite loop in JS but not in Java bytecode

[14:50:02.0000] <Philip`>
Does GWT produce readable JS code?

[14:50:03.0000] <hsivonen>
kingryan: with the browser's native DOM as the tree impl

[14:51:00.0000] <hsivonen>
Philip`: it can produce semi-readable code which is how I figured what loop it gets stuck into

[14:51:01.0000] <hsivonen>
Philip`: by default, the output is *totally* unreadable

[15:04:00.0000] <Hixie>
hsivonen: if you can find a way to reproduce it in a small test file, i can report the bug

[15:05:00.0000] <hsivonen>
Hixie: do you mean small HTML input or a minimized Java case?

[15:05:01.0000] <Hixie>
small gwt input; minimised java case

[15:05:02.0000] <hsivonen>
Hixie: that would be *hard*

[15:05:03.0000] <hsivonen>
Hixie: I'll try to debug this a bit more in the morning

[15:05:04.0000] <Hixie>
yeah, i figure it would be

[15:06:00.0000] <Hixie>
realistically i doubt the gwt team would have the time to track down the bug if they had to learn your code too :-)

[15:06:01.0000] <hsivonen>
yeah :-(

[15:06:02.0000] <hsivonen>
I'll see if the method behaves the same way in isolation

[15:07:00.0000] <Hixie>
it's an open source project though right? there should be a public bug reporting form if you want to just attach the entire project with steps to reproduce

[15:07:01.0000] <Hixie>
also does it do the same thing in other UAs?

[15:07:02.0000] <hsivonen>
yeah, there's a public bug tracker

[15:07:03.0000] <hsivonen>
same problem both in Gecko and WebKit

[15:09:00.0000] <Hixie>
k

[15:09:01.0000] <Hixie>
probably a fundamental compiler bug then unless it's actually a bug in your code somehow

[15:09:02.0000] <hsivonen>
it works in the hosted mode

[15:10:00.0000] <hsivonen>
which suggests it is a compiler bug

[15:10:01.0000] <Hixie>
yep

[15:12:00.0000] <hsivonen>
nn

[15:26:00.0000] <smedero>
"TBL: I feel it is important that the HTML5 spec be split into smaller chunks." http://www.w3.org/2001/tag/2008/05/21-minutes

[15:26:01.0000] <Hixie>
i think everyone agrees with that

[15:26:02.0000] <Philip`>
It's fun trying to install Python 2.3 modules on Gentoo, since I have to make the /usr/bin/python symlink point to python2.3 so they get installed for 2.3, but Gentoo's package manager relies on Python 2.4 features :-(

[15:27:00.0000] <zcorpan_>
Hixie: i don't think it's particularly important

[15:27:01.0000] <Philip`>
jgraham / annevk / etc: Is it know that several tests like "32-bit hex numeric entity" fail in Python 2.3?

[15:27:02.0000] <Philip`>
s/know/known/

[15:27:03.0000] <Philip`>
(...in html5lib)

[15:27:04.0000] <Philip`>
("OverflowError: string/unicode conversion")

[15:32:00.0000] <bzed>
jgraham: hey :) any news on a html5lib release?

[15:32:01.0000] <zcorpan_>
hsivonen: i agree that always allowing = in unquoted attributes seems bad

[15:33:00.0000] <zcorpan_>
hsivonen: src and href (and action) are not the only attributes that take uris in practice, <option value> in particular was pretty common

[15:35:00.0000] <zcorpan_>
hsivonen: (or rather, letting the authoring mistakes in go uncatched seems bad)

[15:35:01.0000] <zcorpan_>
s/in //

[15:37:00.0000] <zcorpan_>
hsivonen: moreover, uris were not the only cases of deliberate use of =

[15:40:00.0000] <annevk>
Philip`, I forgot whether we cared about Python 2.3

[15:41:00.0000] <annevk>
Philip`, I guess we might

[15:41:01.0000] <Philip`>
annevk: It's weird since it works fine when I run test_tokenizer.py, but a few of those tokeniser tests fail when I run runtests.py :-/

[15:42:00.0000] <Philip`>
Does an earlier test do something funny to the Python environment?

[15:47:00.0000] <Philip`>
Oh, test_parser.py does warnings.simplefilter("error")

[15:54:00.0000] <Philip`>
Hmm, the problem is too non-obvious and hard to reproduce :-(

[15:54:01.0000] <Philip`>
/me therefore gives up on it

[16:19:00.0000] <Philip`>
Why are EBCDIC encoding not a "must not support"? They seem to have the same security problems, if you write a U+0014 and convince the user's browser to decode it as EBCDIC to get a '<'

[16:19:01.0000] <Philip`>
*encodings

[16:20:00.0000] <Philip`>
*same security problems as UTF-7 etc

[16:31:00.0000] <Philip`>
s/U+0014/U+004C/

[16:49:00.0000] <Hixie>
the problem with UTF-7 is that IE autodetected it too eagerly, so non-UTF-7 pages let people smuggle in UTF-7 stuff that got treated as such

[16:49:01.0000] <Hixie>
does it do that with EBCDIC?

[16:50:00.0000] <Hixie>
the only difference really is that EBCDIC has more than zero users

[16:50:01.0000] <Hixie>
probably not much more than zero, but probably enough that we don't want to make them non-conforming

[16:55:00.0000] <Philip`>
I have no idea how much support/usage EBCDIC has

[16:57:00.0000] <gsnedders>
The fun of EBCDIC is all the different codepages

[16:59:00.0000] <Philip`>
It just seems like UTF-7 and EBCDIC have the same inherent technical problems, so they should both be handled the same, i.e. either both permitted (but require that they don't get autodetected) or both forbidden


2008-05-31
[17:00:00.0000] <Philip`>
(Presumably both still suffer from the "Web browsers have an awesome easter egg! Just click the Encoding menu then select EBCDIC and there's a cool space fighter game.                    (EBCDIC-encoded <script>...)" attack)

[17:07:00.0000] <Hixie>
any non-ASCII encoding has that problem

[17:08:00.0000] <Hixie>
it seems to be very easy to massively confuse browsers by playing with .focus() and onfocus=""

[17:08:01.0000] <Hixie>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%20%0A...%3Cform%3E%3Cinput%20name%3Da%20onfocus%3D%22value%2B%3D1%3Bd.name%3D%27a%27%3Bname%3D%27d%27%3Bb.focus()%22%3E%0A%3Cinput%20name%3Db%20onfocus%3D%22value%2B%3D1%3Bc.focus()%22%3E%0A%3Cinput%20name%3Dc%20onfocus%3D%22value%2B%3D1%3Bd.focus()%22%3E%0A%3Cinput%20name%3Dd%20onfocus%3D%22value%2B%3D1%3Ba.focus()%22%3E

[17:08:02.0000] <gsnedders>
Hixie: Therefore, we should require everyone to use ASCII!

[17:09:00.0000] <Hixie>
e.g. in firefox, click the fourth input

[17:09:01.0000] <Hixie>
firefox is the most graceful

[17:09:02.0000] <Hixie>
opera just locks the page

[17:10:00.0000] <gsnedders>
Saf behaves like Op

[17:10:01.0000] <gsnedders>
Fun.

[17:10:02.0000] <annevk>
man, amazon.co.uk sucks

[17:10:03.0000] <gsnedders>
Browsers are silly

[17:10:04.0000] <Hixie>
safari and ie hang altoether

[17:10:05.0000] <Hixie>
opera doesn't hang altogether

[17:10:06.0000] <gsnedders>
annevk: That's because it includes .uk :)

[17:10:07.0000] <Hixie>
just the page

[17:10:08.0000] <annevk>
I login, select a set of books, go to the order pages, at that point it tells me I can't get them shipped to NL

[17:10:09.0000] <gsnedders>
heh

[17:11:00.0000] <Hixie>
ok IE hangs the system

[17:11:01.0000] <Hixie>
not just the browser

[17:11:02.0000] <Hixie>
jesus

[17:11:03.0000] <Hixie>
/me reboots his vm

[17:11:04.0000] <gsnedders>
Awesomeness.

[17:11:05.0000] <Hixie>
bbl

[17:11:06.0000] <gsnedders>
bye

[17:12:00.0000] <annevk>
the last time I ordered something from them it never arrived

[17:12:01.0000] <annevk>
and I didn't get my money back, so maybe it was silly to look again

[17:12:02.0000] <gsnedders>
I've had misprinted books come

[17:13:00.0000] <gsnedders>
Sent it back, got a replacement with the same problem

[17:15:00.0000] <gsnedders>
/me should probably put in a late application to start uni this year

[17:15:01.0000] <gsnedders>
/me just wants to get the hell away from school

[17:22:00.0000] <Hixie>
amazon's tech support were always very good when i complained about stuff

[17:25:00.0000] <Hixie>
does .blur() simply undo the focus regardless of the element it is invoked on?

[17:43:00.0000] <annevk>
Hixie, not in Opera

[17:45:00.0000] <annevk>
doesn't seem to happen in Firefox either

[17:47:00.0000] <Hixie>
woah, IE is the only browser that does onfocus/onblur asynchronously

[17:59:00.0000] <Hixie>
i think i'm just gonna spec focus and blur to be sensible

[17:59:01.0000] <Hixie>
and not worry too much about compat

[17:59:02.0000] <Hixie>
since the browsers mostly hang...

[18:29:00.0000] <Hixie>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%3Cbody%20onload%3D%22document.forms%5B0%5D.a.focus()%22%3E%0A...%3Cform%3E%0A%3Ctextarea%20name%3Dt%20cols%3D100%20rows%3D3%3E%3C%2Ftextarea%3E%3Cp%3E%0A%3Cinput%20name%3Da%20onblur%3D%22t.value%2B%3D'ab%20'%3B%20b.focus()%3B%22%3E%0A%3Cinput%20name%3Db%20onfocus%3D%22t.value%2B%3D'bf1%20'%3Ba.focus()%3Bt.value%2B%3D'bf2%20'%3B%22%20onblur%3D%22t.value%2B%3D'bb%20'%3B%22%3E

[18:29:01.0000] <Hixie>
wtf is safari doing

[18:32:00.0000] <Dashiva>
Making your job more interesting

[18:39:00.0000] <Hixie>
safari is clearly buggy here

[18:42:00.0000] <othermaciej>
that test case makes me cry

[18:47:00.0000] <Hixie>
well if you want to fix it wait a few days, i'm writing a spec for it

[18:54:00.0000] <othermaciej>
honestly, I don't care all that much what happens when nodes focus each other from blur and focus handlers in complicated ways, but I am sure there are sites depending on it

[18:54:01.0000] <Hixie>
i doubt it, IE hangs.

[18:55:00.0000] <othermaciej>
we don't have the worst behavior, then :-)

[18:55:01.0000] <Hixie>
you hang in certain situations too

[18:55:02.0000] <Hixie>
though IE's hang at one point required me to reboot my vm

[18:55:03.0000] <Hixie>
so...

[18:58:00.0000] <jwalden>
haha

[19:58:00.0000] <Hixie>
i love that browsers have FOUR different events for focusing

[19:58:01.0000] <Hixie>
focus, activate, focusin, and DOMFocusIn

[23:26:00.0000] <Lachy>
good morning everyone

[23:34:00.0000] <hdh>
hi

[00:59:00.0000] <hsivonen>
Hixie: do you have examples of Web pages and browsers that use/support EBCDIC?

[03:08:00.0000] <Hixie>
hsivonen: nope

[03:13:00.0000] <Hixie>
and this is why <video> doesn't have a fullscreen option:

[03:13:01.0000] <Hixie>
http://www.bunnyhero.org/2008/05/10/scaring-people-with-fullscreen/

[03:17:00.0000] <annevk>
/me doesn't have Flash

[03:25:00.0000] <Hixie>
probably wise

[03:48:00.0000] <gsnedders>
Hmm… http://www.drps.ed.ac.uk/08-09/dpts/SCE_FINAL/205.html

[03:49:00.0000] <gsnedders>
That just seems too mixed for its own good

[05:48:00.0000] <hsivonen>
Hixie: if there's no EBCDIC to be found, wouldn't it be a good idea to ban EBCDIC altogether? i.e. tell implementors not to add it

[06:12:00.0000] <hsivonen>
Hixie: OK. I debugged it. It's a compiler bug. I have default: break eofloop; that the compiler just loses on the way.

[06:13:00.0000] <hsivonen>
I'll file a bug

[06:33:00.0000] <annevk>
So I set up the Google Webmaster Tools and it turns out that Google can't handle 410 responses

[06:34:00.0000] <annevk>
It claims these are "HTTP errors"

[06:35:00.0000] <annevk>
Though http://www.google.com/support/webmasters/bin/answer.py?answer=83045&hl=en suggests this is an intentional message

[06:35:01.0000] <annevk>
/me submits feedback

[06:37:00.0000] <Dashiva>
It seems pretty odd to suggest replacing 410 with a redirect, considering 410 means there is nothing to redirect to...

[06:38:00.0000] <annevk>
indeed, I mentioned Mark Pilgrim in my response, in the hope that he can explain that to them :)

[06:39:00.0000] <hsivonen>
Hixie: it turned out to be a known bug: http://code.google.com/p/google-web-toolkit/issues/detail?id=2069

[06:44:00.0000] <hsivonen>
well, at least now I've reached the point in my life where program misbehavior is really a compiler bug and not my bug

[06:45:00.0000] <annevk>
I suppose it's not a happy moment?

[06:45:01.0000] <annevk>
RB madness: http://lists.w3.org/Archives/Public/public-html-wg-issue-tracking/2008May/

[06:46:00.0000] <hsivonen>
annevk: it's not a happy moment in the sense that it's a bug that I can't easily fix

[06:47:00.0000] <hsivonen>
annevk: it is happy in the sense that I'm not blaming my own mistakes on a compiler without a reason

[06:48:00.0000] <hsivonen>
hmm. actually, now that I think of it, I've already hit a known javac bug in V.nu before

[06:48:01.0000] <Dashiva>
annevk: He was gone for 6 months? I never noticed...

[06:49:00.0000] <hsivonen>
but the javac case wasn't program misbehavior but javac failing to compile altogether

[06:49:01.0000] <hsivonen>
anyway, Eclipse's built-in compiler seems to be the least buggy of the three

[06:49:02.0000] <hsivonen>
at least in the areas that I use

[06:53:00.0000] <jgraham_>
Rob-Burns-spam really discourages me from spending what little free time I have at the moment reading public-html email

[06:54:00.0000] <jgraham_>
It's like he's set out to DOS the WG

[06:54:01.0000] <annevk>
I suggest you just skip his e-mail

[06:54:02.0000] <annevk>
It's not productive to let what you want to do be blocked by a single person

[06:54:03.0000] <jgraham_>
annevk: Yeah, but it's hard to do because he sucks in sensible people

[06:55:00.0000] <jgraham_>
(I have a big block of unread email from him that I intend to ignore)

[06:55:01.0000] <annevk>
Oh, I'd suggest skipping those threads altogether (about namespaces and ARIA)

[06:55:02.0000] <annevk>
They contain no new information / silly suggestions

[06:56:00.0000] <jgraham_>
The ARIA thread is good as a warning of what happens when you let the TAG back down below the atmosphere

[06:56:01.0000] <jgraham_>
:)

[06:57:00.0000] <jgraham_>
Or at least what happens when they try to solve ground-level problems from their positon way up in space

[06:57:01.0000] <annevk>
The TAG is good in describing architecture, not in inventing it :)

[06:57:02.0000] <hsivonen>
/me expects countless hours and email messages to be spent on distributed extensibility over the next year

[06:58:00.0000] <Philip`>
Is Henry S. Thompson representing the views of the whole TAG in that discussion?

[06:58:01.0000] <Dashiva>
That would be scary

[06:59:00.0000] <annevk>
s/year/decade/

[06:59:01.0000] <jgraham_>
On an entirely different topic and just because I'm too lazy to find somewhere more sensible to ask, has anyone ever experienced problems getting Evince to display the correct glyphs for certian characters? I have a document that works fine in kpdf and acroread but has all the capital greek letters replaced with other things in evince

[07:00:00.0000] <jgraham_>
But only on certian machines

[07:00:01.0000] <annevk>
Also on a slightly different topic, how did the presentation go?

[07:00:02.0000] <Philip`>
Does the PDF have the fonts embedded in it?

[07:00:03.0000] <jgraham_>
Philip`: Mostly, yes (the cmr fonts seem to be embedded)

[07:00:04.0000] <jgraham_>
annevk: It was good, I think. Either people liked it or they lied to me :)

[07:01:00.0000] <annevk>
Any tough questions?

[07:01:01.0000] <hsivonen>
jgraham_: how did you divide talking time with Lachy?

[07:02:00.0000] <jgraham_>
We ran out of time for questions. Afterwards some people seemed concerned that we were introducing too many new sectioning elements for no reason

[07:02:01.0000] <Philip`>
Does anybody ever say "I thought your talk was a bit boring / superficial / pointless / poorly presented / etc" to anybody?

[07:02:02.0000] <jgraham_>
hsivonen: Kind of randomly. He did most of the design principles stuff, with some interjections from me. I did more of the new features stuff with some slides from him

[07:03:00.0000] <jgraham_>
Philip`: There are /so many/ people I would like to say that to, but am too polite :)

[07:03:01.0000] <jgraham_>
(Simon Willison seemed to be mainly asleep in the front row)

[07:03:02.0000] <Philip`>
/me would imagine that would be considered impolite, but otherwise everyone will always get a biased view of how their presentations went

[07:03:03.0000] <annevk>
http://pdkm.spaces.live.com/blog/cns!D1DDEC9FF002FB8C!872.entry didn't like the talk at least... complaining about mailing lists and academia

[07:03:04.0000] <annevk>
but spaces.live.com prolly says enough

[07:04:00.0000] <hsivonen>
Philip`: occasionally the audience pointedly disagrees with the presenter

[07:04:01.0000] <jgraham_>
Yeah, I think we spent too long on design principles

[07:04:02.0000] <hsivonen>
Philip`: happened at XTech with Steven Pemberton's talk

[07:05:00.0000] <Philip`>
hsivonen: Disagreeing with presenter doesn't make it a bad presentation - it's good if things are thought-provoking and get people interested and involved :-)

[07:05:01.0000] <jgraham_>
So even being on spaces.live.com it isn't an entirely unfair criticism :)

[07:05:02.0000] <hsivonen>
a Flickr guy made notes in the audience during the presentation and then flushed his counter-arguments at the end

[07:06:00.0000] <jgraham_>
(I should note for posterity that it is in no way Lachy's fault that we sepnt too long on that section)

[07:06:01.0000] <annevk>
http://blog.lylo.co.uk/2008/05/31/media-london-2008-round-up-day-1/ mainly complaints about HTML5 features likely taking a long time

[07:07:00.0000] <annevk>
http://gareth53.blogspot.com/2008/05/atmedia2008-day-1-notes.html has notes on "HTML5 by Jame Hunt & Lachlan Young"

[07:08:00.0000] <annevk>
http://blog.charlvn.za.net/2008/05/media-2008.html wants videos because he couldn't attend

[07:08:01.0000] <hsivonen>
Netscape 4 went away eventually, too

[07:08:02.0000] <annevk>
(that's about all blogsearch.google.com gives me)

[07:08:03.0000] <jgraham_>
I think the 'it's not going to happen anytime soon thing' is a bit unfortunate. We had a problem doing the video demo (note to others: VMWare doen't like being run on an external display)

[07:08:04.0000] <gsnedders>
/me is curious how he went down

[07:08:05.0000] <hsivonen>
I think stopping progress due to IE6 being still around doesn't make sense

[07:09:00.0000] <jgraham_>
gsnedders: afterwards somebody mentioned that it was amazing that you were involved given your age

[07:09:01.0000] <annevk>
I guess the point is more that the audience of @media expects to get things they can do right away rather than information of something that comes in five years.

[07:09:02.0000] <gsnedders>
jgraham_: Heh

[07:09:03.0000] <annevk>
Not that we should do something else :)

[07:10:00.0000] <jgraham_>
I think if we'd had more demos it would be more obvious that some of this stuff is avaliable either now or in the very near future

[07:10:01.0000] <Philip`>
(I suppose disagreement sometimes does just indicate the presentation wasn't very good - I was at a talk by Richard Stallman about his suggestion for a new copyright system, and many people seemed to disagree with him and he just dismissed most of the arguments (because e.g. it's clearly trivial to have a single source collecting donations for musicians, and poll the global population to see what music they listen to and distribute the donations fairly

[07:11:00.0000] <jgraham_>
I plan to post to blog.whatwg.org once Lachlan has put the slides up on his site with links to the demos that we put together but that didn't make the talk

[07:11:01.0000] <Philip`>
...in proportion to sqrt(popularity))

[07:11:02.0000] <gsnedders>
jgraham_: As long as you got my current age right, and not how old I was last month :P

[07:11:03.0000] <jgraham_>
gsnedders: You're 12, right? :-p

[07:11:04.0000] <annevk>
Philip`, hah

[07:12:00.0000] <hsivonen>
Philip`: single source is already implemented on a national level in many places

[07:12:01.0000] <gsnedders>
jgraham_: Yeah, certainly :P

[07:12:02.0000] <hsivonen>
Philip`: the fair distribution part isn't

[07:12:03.0000] <jgraham_>
(seriously I said you were 16, which I think is right unless you get older much fater than normal people)

[07:12:04.0000] <hsivonen>
(it seems to me that the national single sources are quite happy with the distribution being unfair)

[07:12:05.0000] <jgraham_>
s/fater/faster/

[07:12:06.0000] <Philip`>
hsivonen: It's never implemented as a replacement for purchasing music, as far as I'm aware

[07:13:00.0000] <hsivonen>
(e.g. by pocketing the royalties for airplay of foreign music)

[07:13:01.0000] <hsivonen>
Philip`: ah

[07:13:02.0000] <gsnedders>
jgraham_: (Yeah, I don't seem to get much older)

[07:14:00.0000] <hsivonen>
even though it doesn't license me to do whatever with music, I already have to pay a slice in the price of storage media to the music protection money collector

[07:14:01.0000] <annevk>
the national thing is more about playing music in stores / work place / etc.

[07:14:02.0000] <annevk>
not about buying CDs

[07:14:03.0000] <hsivonen>
annevk: in Finland it is also about individuals copying music

[07:15:00.0000] <annevk>
like a tax?

[07:15:01.0000] <hsivonen>
annevk: yes, like a tax

[07:15:02.0000] <hsivonen>
annevk: but they still want to ban copying of course and just pocket the tax

[07:15:03.0000] <annevk>
interesting, and everyone pays?

[07:16:00.0000] <hsivonen>
annevk: everyone except those who have the time and energy to get their blank storage from Estonia

[07:17:00.0000] <annevk>
the video game industry is much less affected by this, but maybe that's also because long time storage is not important for video games

[07:17:01.0000] <Philip`>
If I remember correctly (which I may well not be), RMS was suggesting that all music (and other artistic creations) would be uncopyrightable, and the public would choose to pay a voluntary tax that would be distributed to all artists

[07:17:02.0000] <hsivonen>
annevk: the guy who writes copyright law in Finland used to work for the music royalty collectors and his wife has a stake in a book royalty collection society

[07:17:03.0000] <annevk>
and there's some amount of hardware lock in with video game systems

[07:18:00.0000] <annevk>
hsivonen, hehe

[07:22:00.0000] <hsivonen>
It sucks that unelected officials run the show and the elected ministers can't control them

[07:24:00.0000] <hsivonen>
If I star an issue on Google Code, will it send notifications to Gmail?

[07:25:00.0000] <jgraham_>
hsivonen: Yeah in theory I think

[07:26:00.0000] <jgraham_>
/me hasn't worked out why issues in the html5lib issue tracker aren't sent to the mailing list

[07:26:01.0000] <jgraham_>
(although I'm probably overlooking something obvious)

[07:26:02.0000] <hsivonen>
/me sets up a forward from gmail to normal email

[07:27:00.0000] <jgraham_>
bzed: On the subject of html5lib I can try and make a release in the next couple of days. I think we don't have any serious blockers at this point

[07:27:01.0000] <hsivonen>
/me is a bit discouraged by the GWT compiler bug not having an owner

[07:28:00.0000] <Philip`>
jgraham_: There are 13 test failures when I run it, probably due to self-closing tag stuff though I've not looked in any detail

[07:29:00.0000] <jgraham_>
Philip`: If the test faliures are related to the self closing tag stuff we should just skip those tests for now. If they're more serious, I blame you for trying to make our oerformance suck less ;)

[07:30:00.0000] <jgraham_>
(seriously though the regexp stuff is cool)

[07:30:01.0000] <Philip`>
jgraham_: Those tests failed before I made any changes, so they're not my fault, which is why I didn't bother trying to fix them :-)

[07:32:00.0000] <gsnedders>
/me blames Philip` anyway :P

[07:34:00.0000] <annevk>
jgraham_, did you add the mailing list as user to the project and starred the items while logged in as the mailing list?

[07:35:00.0000] <annevk>
jgraham_, I haven't actually tried this feature myself, but it seems it might work that way

[07:37:00.0000] <bzed>
jgraham_: I'm not sure how far the debian lenny freeze is away, so that would be good

[07:40:00.0000] <Philip`>
jgraham_: By the way, have you looked at issue 70 at all? (I was wondering if I should try fixing it, since I can see where the problem is)

[07:47:00.0000] <jgraham_>
Philip`: No, I haven't made time to look yet. If you know where the problem is I'm quite confident that you'll fix it faster than I can

[07:48:00.0000] <annevk>
/me wants a C -> Python version of html5lib and a browser impl

[07:49:00.0000] <Philip`>
/me tries to work out which of the tree-construction/tests[1-7].dat files to add new tests to, and fails

[07:49:01.0000] <jgraham_>
annevk: On the admin page, under Email notifications of project activity will automatically be sent to the following email addresses. We have html5lib-discuss⊙gc under "All issue changes" but AFAICT issue changes don't get sent there

[07:51:00.0000] <gsnedders>
How can you check whether an ECMAScript variable is a string?

[07:52:00.0000] <annevk>
jgraham_, did we add codesite-noreply⊙gc to html5lib-discuss?

[07:52:01.0000] <annevk>
jgraham_, and to html5lib-commits?

[07:53:00.0000] <annevk>
/me wonders who owns html5lib-discuss

[07:53:01.0000] <jgraham_>
annevk: Ah, that could be it

[07:53:02.0000] <annevk>
(seems html5lib-commits is ok)

[07:54:00.0000] <jgraham_>
gsnedders: Can't you use foo instanceof String

[07:55:00.0000] <jgraham_>
?

[07:55:01.0000] <gsnedders>
Seemingly yes

[07:55:02.0000] <gsnedders>
Hmm… Everything goes bizarre in WebKit, without it causing errors

[07:56:00.0000] <jgraham_>
annevk: I own it, but under an odd gmail account I only use for backup of my university email

[07:57:00.0000] <annevk>
I still have this plan of integrating all my non-work accounts into gmail one day...

[07:59:00.0000] <gsnedders>
Saf returns odd things from getResponseHeader(header)

[07:59:01.0000] <gsnedders>
(or at least when you try and run typeof of it)

[08:00:00.0000] <gsnedders>
It returns typeof xhr.getResponseHeader("connection") == "object" for some requests :\

[08:01:00.0000] <gsnedders>
Actually…

[08:01:01.0000] <gsnedders>
Maybe they're treated as HTTP/0.9

[08:01:02.0000] <jgraham_>
annevk: Added you as an owner for html5lib-discuss

[08:01:03.0000] <gsnedders>
yup. that seems to be it.

[08:02:00.0000] <hsivonen>
annevk: isn't <em> to <i> what <object> is to <embed>?

[08:03:00.0000] <annevk>
in my mind <em> and <embed> are the subsets

[08:03:01.0000] <hsivonen>
ah

[08:05:00.0000] <jgraham_>
hsivonen: I don't have a problem with you changing the tests

[08:06:00.0000] <hsivonen>
jgraham_: ok. thanks

[08:10:00.0000] <hdh>
heh, I'm running cdouble's firefox, and everything after <video> gets treated as fallback content, I have to read the rest in source

[08:12:00.0000] <gsnedders>
jgraham_: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cscript%3E%0Avar%20foo%20%3D%20%22bar%22%3B%0Aalert(foo%20instanceof%20String)%3B%0A%3C%2Fscript%3E

[08:14:00.0000] <Philip`>
Hmm, BeautifulSoup's 'replace' method is a bit broken - if you have "foo<br>foo" and want to replace the second 'foo' node, you can't, because it's indistinguishable from the first one :-/

[08:14:01.0000] <hsivonen>
Hixie: do you have pending update for the meta prescan algorithm? that is, is now a good time to sync implementation?

[08:17:00.0000] <jgraham_>
gsnedders typeof foo == "string"?

[08:18:00.0000] <gsnedders>
jgraham_: Yeah, that's what I eventually found. Seems to work, at least

[08:20:00.0000] <gsnedders>
/me realises one test is broken because XMLHttpRequest forbids him from testing it

[08:21:00.0000] <gsnedders>
/me blames annevk for making his life hard

[08:21:01.0000] <gsnedders>
/me thinks Opera is odd

[08:21:02.0000] <gsnedders>
(in terms of what it returns from XMLHttpRequest)

[08:22:00.0000] <annevk>
meh

[08:24:00.0000] <Philip`>
Oh, and BeautifulSoup's 'extract' method is also broken, because it uses 'remove' which has the same problem as 'replace'

[08:24:01.0000] <Philip`>
So, uh, that's a pain

[08:25:00.0000] <gsnedders>
annevk: Opera seems to always return an empty string when you try and see what you got for the "connection" header

[08:25:01.0000] <annevk>
what does beautifyfulsoup do that html5 doesn't?

[08:26:00.0000] <annevk>
gsnedders, file a bug?

[08:26:01.0000] <annevk>
gsnedders, also, I suggest to use some HTTP sniffer instead of XHR, prolly more reliable

[08:26:02.0000] <gsnedders>
annevk: But how do I do automated tested from browsers? :(

[08:28:00.0000] <Philip`>
annevk: ?

[08:28:01.0000] <gsnedders>
/me wishes you could actually view your own bugs on Opera

[08:28:02.0000] <gsnedders>
I may well have reported this months ago

[08:31:00.0000] <jgraham_>
annevk: The point is that the BeautifulSoup interface for python has an infinite-recursion bug that Philip` is trying to fix

[08:31:01.0000] <Philip`>
gsnedders: You did, about six months ago

[08:31:02.0000] <Philip`>
(303305)

[08:31:03.0000] <jgraham_>
BeautifulSoup isn't really designed to be used the way that html5lib uses it, unfortunatly

[08:32:00.0000] <jgraham_>
I remember it was somewhat non-trivial to get it working in the first place

[08:32:01.0000] <Philip`>
jgraham_: The infinite recursion is a bug in html5lib; the problem is I encountered a new bug when trying to fix it, and I'm not sure how to cleanly fix that one

[08:32:02.0000] <jgraham_>
Oh, OK :)

[08:32:03.0000] <Philip`>
since that's a problem with BeautifulSoup's API being unusable

[08:32:04.0000] <gsnedders>
Philip`: You have access? Heh. That's cheating! :P

[08:33:00.0000] <jgraham_>
I meant Beautifulsoup interface for html5lib btw

[08:34:00.0000] <jgraham_>
(was the bug just in BS code or in all html5lib?)

[08:34:01.0000] <Philip`>
Of course the real problem is that Python strings are immutable

[08:34:02.0000] <jgraham_>
s/code/html5lib code/

[08:35:00.0000] <Philip`>
jgraham_: The infinite recursion was the BS part of html5lib, since appendChild(TextNode(NavigableText(...))) was just calling itself

[08:36:00.0000] <jgraham_>
I can see how that would be bad

[08:37:00.0000] <Philip`>
Since strings are immutable, this code has to splice a new string node into the tree in place of the old one, but I can't use the BS API because that doesn't work when you have two equivalent nodes within an element, and it's hard to do manually since there's loads of references to fix up :-(

[08:38:00.0000] <Philip`>
(Actually, maybe it's not as bad as I'm thinking it is...)

[08:39:00.0000] <gsnedders>
hmm… this is an odd case.

[08:40:00.0000] <gsnedders>
Safari/CFNetwork are stricter than anything else, and I expect it may well be a case of it just not being so old

[08:40:01.0000] <gsnedders>
(and therefore not having to deal with what the web was like in 2000/2001)

[08:42:00.0000] <Philip`>
Hmm, tests pass - that means I don't have enough tests

[08:43:00.0000] <gsnedders>
Yeah, I'm in a similar state with Fx and Saf :P

[08:43:01.0000] <gsnedders>
s/Saf/IE/

[08:45:00.0000] <Philip`>
Is it bad that html5lib can take quadratic time to parse a document?

[08:45:01.0000] <Philip`>
(because "a</a>a</a>a</a>..." does quadratic-cost string concatenation)

[08:46:00.0000] <annevk>
as long as the C version doesn't do it :)

[08:47:00.0000] <Philip`>
That would require treebuilders to not be generating Python data structures

[08:48:00.0000] <Philip`>
or else to have a postprocessing step that takes a tree and coalesces adjacent text nodes

[08:48:01.0000] <Philip`>
(which wouldn't work once html5lib has scripting support)

[08:50:00.0000] <jgraham_>
Couldn't you use a list in the adoption agency and then join things later to avoid some of the concatenation? I haven't thought this throught at-all so I could well be wrong

[08:50:01.0000] <Philip`>
This is nothing to do with the adoption agency

[08:51:00.0000] <Philip`>
It's just to do with adjacent text nodes

[08:51:01.0000] <hsivonen>
the V.nu parser in the coalescing mode just makes a guess about how large a coalescing buffer to allocate

[08:51:02.0000] <hsivonen>
SaxTree isn't coalescing

[08:55:00.0000] <Philip`>
hsivonen: What does the coalescing buffer do?

[08:55:01.0000] <Philip`>
and how does it handle input like foo<table><td>bar</a>baz</td>quux ?

[08:56:00.0000] <hsivonen>
Philip`: hmm. I'm not sure.

[08:59:00.0000] <Philip`>
Hmm, no browser coalesces fooquux

[08:59:01.0000] <Philip`>
I can't think of any way to handle this using immutable strings without getting quadratic performance, which is bad

[09:00:00.0000] <Philip`>
(I guess you can handle barbaz more reasonably by having an appendable buffer, though only Firefox seems to do that in practice)

[09:01:00.0000] <Philip`>
(*only Firefox and IE)

[09:06:00.0000] <hsivonen>
if no browser coalesces those nodes, it's not an interop issue, so we don't need to worry about it (except perhaps the spec should tell implementors not to worry about it)

[09:07:00.0000] <Philip`>
Half the browsers don't even put quux as a text node outside the table

[09:11:00.0000] <annevk>
how do you tell what browsers do? (especially IE / Opera) did you make performance tests to quantify your assumptions?

[09:15:00.0000] <Philip`>
annevk: Oops, I meant only Firefox and IE seem to produce a single coalesced text node in the output DOM (whereas Opera and WebKit create multiple nodes)

[09:15:01.0000] <Philip`>
I have no idea how they implement it internally

[09:15:02.0000] <Philip`>
(but I would hope they do it sanely)

[09:17:00.0000] <Philip`>
(If I do foo<script id=i>i=document.getElementById('i');i.parentNode.removeChild(i)</script>bar then Firefox does produce two text nodes, so it looks like it's detecting consecutive text appends in the parser, rather than detecting appending to a node that already ends in text)

[09:52:00.0000] <hsivonen>
cross-site RPC by abusing GET: http://groups.google.com/group/Google-Web-Toolkit-Contributors/browse_thread/thread/94c18c4ec158070c/

[09:56:00.0000] <Philip`>
Aaargh

[09:56:01.0000] <Philip`>
I don't like BeautifulSoup :-(

[09:57:00.0000] <Philip`>
x<table 1><table 2>x parses correctly, but x<table><table>x doesn't

[09:58:00.0000] <Philip`>
because insertBefore(second <table>) does a lookup of <table> and finds the first one

[10:00:00.0000] <Philip`>
Oh, actually that's mostly html5lib's fault

[10:30:00.0000] <Philip`>
Is it ever possible to have to insert an element before another element, when that 'another element' is not the last child of its parent? (when there's no scripting)

[11:59:00.0000] <hsivonen>
http://www.w3.org/DesignIssues/diagrams/arch/follow.svg

[11:59:01.0000] <hsivonen>
somehow protein looks out of place

[12:02:00.0000] <Philip`>
Is there a version of that image that doesn't freeze for a second each time I click a scrollbar?

[12:02:01.0000] <Philip`>
Bonus points if the shadows don't get cut off near the right edge

[12:04:00.0000] <hsivonen>
Philip`: WFM in Firefox 3

[12:05:00.0000] <Philip`>
Not FM, though :-(

[12:05:01.0000] <Philip`>
(or at least not acceptably)

[12:06:00.0000] <Philip`>
Safari in Wine renders it almost fast enough, though probably just because it misses the shadows

[12:06:01.0000] <hsivonen>
Philip`: it seems to suffer from the OmniGraffle/SVG clipping bogosity that poisons SVG filter clipping 

[12:06:02.0000] <Philip`>
hsivonen: I thought FF3 had workarounds for that

[12:07:00.0000] <Philip`>
but maybe I'm remembering wrong

[12:07:01.0000] <hsivonen>
like I said, WFM in Firefox 3

[13:47:00.0000] <gsnedders>
Now, to blog something this month…

[16:14:00.0000] <jgraham>
For the benefit of whoever marked the logs, I should also note that there were several people I spoke to at @media who were very positive about the new sectioning elements

[16:15:00.0000] <gsnedders>
/me always wonders who actually does the marking

[16:16:00.0000] <Philip`>
Probably Googlebot

[16:16:01.0000] <gsnedders>
Ah. That makes sense.

[16:34:00.0000] <Hixie>
hsivonen: re EBCDIC, yeah, maybe you're right... Ask the lists and the blog if anyone is using it?

[16:34:01.0000] <gsnedders>
Hixie: Can't you look in the Google Cache for the usage of charsets?

[16:54:00.0000] <gsnedders>
I guess I'm failing to blog anything this month