2009-12-01
[16:26:00.0000] <Hixie_>
woah, the w3c archives archive across months now

[16:26:01.0000] <Hixie_>
sweet

[16:27:00.0000] <Hixie_>
othermaciej: can i link to http://esw.w3.org/topic/HTML/ChangeProposalTemplate from http://dev.w3.org/html5/status/issue-status.html ?

[16:27:01.0000] <othermaciej>
Hixie_: sure

[16:27:02.0000] <othermaciej>
Hixie_: that would be convenient and helpful in fact

[16:36:00.0000] <Hixie_>
done

[16:57:00.0000] <othermaciej>
Hixie_: have you had a chance to look at the I18N WG's proposal for multiple languages in meta, or Larry's for URLs?

[16:58:00.0000] <othermaciej>
Hixie_: no urgent deadline on those but your input could potentially save time if it allows dispensing with the formal bureaucracy

[17:01:00.0000] <MikeSmith>
Hixie_: I note that the spec says "The presence of a name attribute on an a element, if its value is not the empty string." .. so what should happen for the case of <a name=''> -- report it as an error?

[17:08:00.0000] <Hixie_>
othermaciej: do they have bugs filed?

[17:09:00.0000] <Hixie_>
MikeSmith: context?

[17:10:00.0000] <othermaciej>
Hixie_: Larry's thing is in the form of a Change Proposal

[17:11:00.0000] <othermaciej>
Hixie_: the I18N guys suggested a change in an informal way and said they talked to you about it

[17:11:01.0000] <othermaciej>
Hixie_: would you like that in the form of a bug?

[17:11:02.0000] <Hixie_>
i believe everything we spoke about in the i18n group has a bug already

[17:11:03.0000] <othermaciej>
Larry's thing has a bug as well as a Change Proposal I think (though the bug references a dated version)

[17:11:04.0000] <Hixie_>
right now my priority is getting the webapps draft to LC

[17:17:00.0000] <MikeSmith>
Hixie_: immediate context is http://bugzilla.validator.nu/show_bug.cgi?id=685 .. I'm trying to figure out what would be best to have validator.nu report for the case of <a name=''> . It currently just says, "Bad value for attribute name on element a: An ID must not be the empty string." My question is, what's the intent of the spec in distinguishing between the cases of <a name="foo"> and <a name=""> ? Is it useful for conforman

[17:17:01.0000] <MikeSmith>
ce checkers to report something different for those cases?

[17:18:00.0000] <Hixie_>
<a name="foo"> should warn that name="" is deprecated. <a name=""> should just say name="" is invalid, I think.

[17:18:01.0000] <Hixie_>
s/that name="" is deprecated/that id="" is considered a more reliable way of giving fragment identifier targets/

[17:19:00.0000] <Hixie_>
<a name=""> should be an error because these obsolete features are only allowed to ease transition, and there's nothing to transition if the attribute's value is "".

[17:19:01.0000] <MikeSmith>
OK, that makes sense

[17:19:02.0000] <MikeSmith>
for the <a name=""> case, v.nu now reports the warning, "The name attribute is obsolete. Consider putting an id attribute on the nearest container instead."

[17:20:00.0000] <Hixie_>
(why the nearest container?)

[17:20:01.0000] <MikeSmith>
so for the <a name=""> , we can have it report the error and also report that same warning

[17:20:02.0000] <MikeSmith>
Hixie_: I dunno. That's verbatim from your spec text.

[17:20:03.0000] <Hixie_>
that works too

[17:21:00.0000] <Hixie_>
the term "nearest container" is not in html5

[17:22:00.0000] <MikeSmith>
The logic of also reporting the warning is that somebody first gets the error, then changes their content to <a name=foo> and then gets a warning about, then that's just annoying -- they are going to say, you should have told me that the first time.

[17:22:01.0000] <MikeSmith>
Hixie_: hmm, I got it from somewhere.

[17:22:02.0000] <MikeSmith>
/me goes to look at current draft

[17:22:03.0000] <Hixie_>
agreed, the text for name="" should just say 'use id=""'

[17:22:04.0000] <Hixie_>
'...and give it a value!'

[17:23:00.0000] <MikeSmith>
erp, maybe I got that "nearest container" bit from Anne's HTML4-HTML5 differences doc 

[18:01:00.0000] <TabAtkins__>
MikeSmith: The 'nearest container' bit still makes sense, as it was very common to just use an empty <a name> at the top of a container to target it.

[18:01:01.0000] <MikeSmith>
TabAtkins__: OK

[18:02:00.0000] <MikeSmith>
so I guess it's fine to keep it in there and the wording can be refined later if needed

[18:02:01.0000] <TabAtkins__>
Yeah.

[18:03:00.0000] <Hixie_>
maybe it's referring to the case of an empty <a name=""></a> ?

[18:03:01.0000] <MikeSmith>
yeah, I reckon so

[18:04:00.0000] <MikeSmith>
that seems like the most common way it's actually seen in existing content

[18:15:00.0000] <erlehmann>
Hixie_, is it possible that the figure construct eventually will be changed back again or are all arguments coming in on the list now re-hashes of ye olde conflict ?

[18:16:00.0000] <Hixie_>
othermaciej would be a better person to answer that question

[18:17:00.0000] <othermaciej>
/me assumes erlehmann was asking about the arguments on the whatwg list

[18:17:01.0000] <othermaciej>
since there are not any coming in on public-html afaik

[18:18:00.0000] <erlehmann>
othermaciej, did you inherit the honors of supreme HTML5 commander from Hixie somehow ? ;)

[18:18:01.0000] <erlehmann>
because i mean the whatwg list.

[18:19:00.0000] <erlehmann>
hmm, better would be "i'm meaning", would it ?

[18:19:01.0000] <Hixie_>
oh, i thought the discussion was on a w3c bug

[18:19:02.0000] <othermaciej>
there's different discussion on a w3c bug

[18:19:03.0000] <Hixie_>
on the whatwg list, rehashing old arguments without introducing new data will not have any effect

[18:19:04.0000] <erlehmann>
oh, sorry

[18:19:05.0000] <Hixie_>
i haven't checked to see if recent threads introduce new data

[18:19:06.0000] <othermaciej>
someone suggested restricting the content model for <figure>

[18:20:00.0000] <othermaciej>
based on what I think were originally incorrect assumptions (though I am not sure why the argument still continues)

[18:21:00.0000] <erlehmann>
i am not sure too. dt / dd may be ugly, but it gets the job done.

[18:21:01.0000] <Hixie_>
i need to get through this webapps stuff then i'll be able to do a better job of keeping on threads

[18:21:02.0000] <Hixie_>
if anyone sees anyone being even remotely rude on the whatwg list, please let me know asap

[18:22:00.0000] <othermaciej>
the recent figure bug was orthogonal to dt/dd I think

[18:22:01.0000] <erlehmann>
oh, i was remotely rude today.

[18:22:02.0000] <erlehmann>
misunderstood TabAtkins, but we reconciled pretty fast.

[18:23:00.0000] <erlehmann>
othermaciej, what was it ? browsers aren't doing strange things with the dom again, are they ?

[18:24:00.0000] <othermaciej>
erlehmann: I'm actually not entirely sure what the motivation for the bug was

[18:25:00.0000] <erlehmann>
bet aesthetical reasons, hehe

[18:25:01.0000] <Hixie_>
/me is still unsure how to handle some of the htmlwg feedback (especially shelley's "remove XYZ" bugs)

[18:55:00.0000] <TabAtkins__>
Restricting <figure>'s content model won't work.  A common use of figures in, frex, academic journals is to hold tables.

[18:56:00.0000] <webben>
That bug would benefit from documentation of examples.

[18:59:00.0000] <TabAtkins__>
It's possible I'm misunderstanding the purpose of <figure>, of course.  My impression of it, though, was for content that is part of the main content, but whos exact location within the content is not important.

[18:59:01.0000] <TabAtkins__>
That is, after all, what figures are.  They're referred to, and are part of articles, but where in the article they're actually placed doesn't matter.

[19:00:00.0000] <TabAtkins__>
This is in opposition to <aside>, which is content that is only tangentially related to the main content, and could be removed without changing the meaning of the main content.

[19:00:01.0000] <webben>
that's part of what figures are.

[19:00:02.0000] <webben>
well, actually, I wouldn't define them like that

[19:01:00.0000] <webben>
often the main content requires the figure - the only thing that allows them to be removed is the reference that points to them.

[19:01:01.0000] <TabAtkins__>
Alternate conversation: Can anyone think of *any* sane reason why I'm being given a forbidden error when trying to run this .pl file, but if I change it to a .txt or .php file it works fine?

[19:01:02.0000] <TabAtkins__>
webben: True, but it's also very common practice to do so - to remove it from the main flow and insert a reference.

[19:02:00.0000] <TabAtkins__>
The point is that they represent additional data that is not part of the flow of the article.

[19:02:01.0000] <TabAtkins__>
But is still *part* of the article and is necessary for full understanding of it.

[19:02:02.0000] <webben>
I'm still not sure that's right.

[19:03:00.0000] <webben>
depends what "flow" refers to - they can be entirely essential to the flow of ideas.

[19:03:01.0000] <TabAtkins__>
Well, my definition happens to correspond pretty exactly to the current spec text, luckily enough.  ^_^

[19:04:00.0000] <TabAtkins__>
(I didn't actually realize it was that close.)

[19:04:01.0000] <erlehmann>
TabAtkins__, messed up .htaccess

[19:04:02.0000] <erlehmann>
??

[19:05:00.0000] <TabAtkins__>
erlehmann: Not that I can *tell*.  In fact, I have just *added* an htaccess, as there wasn't one there before.  All it contains is an "Options ExecCGI" line.

[19:05:01.0000] <erlehmann>
well, then …

[19:05:02.0000] <erlehmann>
no idea ^_^

[19:05:03.0000] <TabAtkins__>
(That would be a sane reason, though, although it's wrong.)

[19:07:00.0000] <TabAtkins__>
webben: Sometimes captioned images are indeed part of the normal flow of an article, I agree, and these are well-served by <figure> as well.

[19:13:00.0000] <webben>
but anyway I more meant - Shelley's claim is allowing tables and code listings and other text in "figure" is confusing because that's not what people call figures. If there's evidence to the contrary (as seems likely to me) would be good to cite specific examples. (I see only vague references in the ticket.)

[19:14:00.0000] <TabAtkins__>
I can go grab textbooks off my shelf that have code samples as figures.

[19:15:00.0000] <webben>
here's a specific example: http://tinyurl.com/ylrv4qe

[19:15:01.0000] <TabAtkins__>
Heh, that works.

[19:16:00.0000] <TabAtkins__>
And here's a table used as a figure: http://books.google.com/books?id=Fsz234KPVz8C&pg=PT120&dq=%22Figure+5%22&ei=I4wUS8raFJjQNPHJ2ZQL#v=onepage&q=%22Figure%205%22&f=false

[19:16:01.0000] <TabAtkins__>
I'll add these to the bugs.

[19:17:00.0000] <webben>
here's another data table labelled as a figure: http://tinyurl.com/yfksktb

[19:19:00.0000] <TabAtkins__>
Darn, I should have waited.  I just submitted it.

[19:19:01.0000] <TabAtkins__>
Still, it takes 30 seconds to find many examples.

[19:19:02.0000] <TabAtkins__>
So saying that using tables in figures will cause confusion is silly.

[19:20:00.0000] <webben>
or at least, saying that people don't put tables in "figures" is empirically false.

[19:21:00.0000] <webben>
(it might be a real practice, but still be confusing I guess)

[19:21:01.0000] <TabAtkins__>
True, but it's common enough that I highly doubt there is any confusion.

[19:29:00.0000] <TabAtkins__>
Okay, what are the reasons the cgi-bin folder on a linux box would be forbidden?  It has 755 permissions, and has both owner and group set to the values specified in apache2.conf

[20:17:00.0000] <Sirisian>
When you submit a review comment in the HTML 5 draft do people actually read those? Is that the most effective way to give feedback?

[20:25:00.0000] <MikeSmith>
Sirisian: you mean submitting it using the built-in form in the whatwg version of the spec?

[20:25:01.0000] <Sirisian>
yeah

[20:28:00.0000] <MikeSmith>
Sirisian: yeah, people do read all those. When you submit a comment that way, it actually causes a new issue to be opened in the W3C bugzilla

[20:29:00.0000] <Sirisian>
yeah I saw that. I added a comment about the canvas element. Just wanted to make sure that wasn't like a "this is for spelling mistakes" kind of thing

[20:30:00.0000] <Sirisian>
MikeSmith, so is it just one person in their spare time "Ian" that goes through and makes the changes?

[20:30:01.0000] <MikeSmith>
Sirisian: so you can also submit a comment directly using http://www.w3.org/Bugs/Public/enter_bug.cgi?product=HTML%20WG

[20:31:00.0000] <Sirisian>
yeah I saw that

[20:31:01.0000] <MikeSmith>
Sirisian: Ian is Hixie_ and yeah he is the one who actually makes the changes to the spec

[20:31:02.0000] <MikeSmith>
to that spec

[20:31:03.0000] <MikeSmith>
but he does it full time, not spare time

[20:32:00.0000] <Sirisian>
ah cool. I like that idea.

[20:32:01.0000] <MikeSmith>
there are quite a number of other people who follow all the bugzilla issues

[20:32:02.0000] <Sirisian>
yeah I noticed the issue rating system and such

[20:34:00.0000] <Sirisian>
Also this might seem silly, but is there something similar to the HTML 5 draft but for javascript and how browsers implement and interact with it? I've noticed over the years that browsers seems to use it slightly differently for input related things especially the mouse.

[20:35:00.0000] <MikeSmith>
Sirisian: not for Javascript itself, not that I know of

[20:37:00.0000] <MikeSmith>
but for existing APIs that get exposed to Javascript, there is some documentation in HTML5 that addresses improving interoperability of some of those

[20:37:01.0000] <MikeSmith>
and there's also the beginning of a "Web DOM Core" draft that attempts to document more

[20:38:00.0000] <MikeSmith>
Sirisian: http://simon.html5.org/specs/web-dom-core

[20:40:00.0000] <MikeSmith>
but if you mean differences in the way that browsers handle Javascript other than how they expose standard APIs, then that's more an issue of either bugs in particular browsers, or bugs/ambiguities in the Javascript spec that need to be fixed

[20:41:00.0000] <Sirisian>
nah not so much dom, but I was thinking of the mouse input and the event model associated with it. That's a good idea.

[20:41:01.0000] <Sirisian>
/me searches for some kind of spec

[20:41:02.0000] <erlehmann>
if anyone wants to test my wordpress plugin generating microdata licensing info I coded in the last few hours (thanks, foolip) => http://github.com/erlehmann/cc-figure

[21:00:00.0000] <MikeSmith>
Sirisian: as far at the event model, you might want to look at the DOM3 Events draft

[21:01:00.0000] <MikeSmith>
http://dev.w3.org/2006/webapi/DOM-Level-3-Events/html/DOM3-Events.html

[21:04:00.0000] <Sirisian>
ah thank you

[21:04:01.0000] <Sirisian>
I wasn't sure if it would be in the Dom stuff

[21:04:02.0000] <MikeSmith>
that document in part attempts to document how event handling actually works in current browsers

[21:04:03.0000] <MikeSmith>
so-called DOM Level Zero

[21:06:00.0000] <Sirisian>
Why?

[21:07:00.0000] <Sirisian>
Shouldn't it just list what the browsers should do?

[21:07:01.0000] <Sirisian>
Or is that what you meant?

[21:11:00.0000] <MikeSmith>
Sirisian: there would not be much point in it listing what browsers should do if browser instead actually did something different

[21:12:00.0000] <MikeSmith>
the DOM Level 0 stuff is basically what browsers have always done but that was never actually specified

[21:12:01.0000] <MikeSmith>
but there are some differences among browsers with regard to some of it

[21:14:00.0000] <MikeSmith>
so the value is speccing it at this point is that anybody developing a new browser would not need to reverse engineer it

[21:16:00.0000] <Sirisian>
Why is the spec so vague about the mouse stuff? Even the default action stuff is extremely vague. No wonder IE handles the 3 mouse buttons as 7 buttons and why opera has no prevent default in middle mouse or right mouse button (speaking of the context menu). It's not even defined. http://dev.w3.org/2006/webapi/DOM-Level-3-Events/html/DOM3-Events.html#events-Events-MouseEvent-button talk about vague

[21:18:00.0000] <MikeSmith>
Sirisian: the spec is still in development. If you find things it doesn't address but should, the best place to comment is on the public-webapps⊙wo mailing list

[21:18:01.0000] <Sirisian>
I don't do mailing lists. I just lose track of messages and get confused.

[21:26:00.0000] <MikeSmith>
Sirisian: OK, you can comment through bugzilla

[21:26:01.0000] <MikeSmith>
http://www.w3.org/Bugs/Public/enter_bug.cgi?product=WebAppsWG

[21:26:02.0000] <MikeSmith>
and select component "DOM3 Events"

[21:27:00.0000] <Sirisian>
Thank you for being pacient with :)

[21:31:00.0000] <MikeSmith>
Sirisian: "enlightened self interest" .. the easier it is for people to review and comment, the greater the chance of getting more good feedback and possibility of improvements

[21:32:00.0000] <Sirisian>
then why did they use bugzilla?

[21:33:00.0000] <MikeSmith>
Sirisian: point taken

[21:34:00.0000] <MikeSmith>
(if your point is that bugzilla is not an easy thing for most people to use)

[21:34:01.0000] <MikeSmith>
but some people would rather submit through bugzilla than join a mailing list

[21:34:02.0000] <Sirisian>
yeah that was my point. It took me a while to figure out how to view all of the issue. Someone told me about the product:html thing though and it made things easier.

[21:34:03.0000] <MikeSmith>
perhaps more specs should have the built-in comment form like the HTML5 spec does

[21:35:00.0000] <Sirisian>
yeah that's a clever setup they have there.

[21:36:00.0000] <MikeSmith>
they is Hixie, in this case

[21:37:00.0000] <Hixie_>
yeah i like it

[21:37:01.0000] <Hixie_>
we should do that more

[21:46:00.0000] <othermaciej>
I only wish the built-in comment form had a way for people to identify themselves if they choose

[21:51:00.0000] <MikeSmith>
othermaciej: yeah, me too

[21:51:01.0000] <MikeSmith>
though some people have just taken to including their name or initials in the text of the comment

[22:03:00.0000] <Hixie>
why does the originator matter?

[22:03:01.0000] <Hixie>
i ignore it anyway when dealing with feedback

[22:14:00.0000] <MikeSmith>
Hixie: it matters on that some people would like to actually identify themselves when submitting a comment

[22:15:00.0000] <MikeSmith>
yeah, I know they could just use bugzilla directly to do that

[22:15:01.0000] <MikeSmith>
but doing it from the built-in comment form is more convenient

[22:16:00.0000] <MikeSmith>
having it be an optional second input field in the form would seem to be the best of both worlds

[22:18:00.0000] <Sirisian>
What's that feature in browsers when you middle click? autoscroll?

[22:19:00.0000] <Hixie>
MikeSmith: i see that they do, but why do they do?

[22:19:01.0000] <MikeSmith>
Sirisian: my pointing device doesn't have a middle click

[22:20:00.0000] <MikeSmith>
Hixie: I guess in part to save the trouble of posting a comment to the bug later and needing to say, "I'm the one who submitted this comment."

[22:20:01.0000] <Sirisian>
I almost did that.

[22:20:02.0000] <MikeSmith>
and/or so that others know clearly who to contact if they want to get clarification outside of bugzilla -- e.g., on IRC

[22:21:00.0000] <MikeSmith>
so if Sirisian files a bug and I see it, I can ping him here to say, "Hey Sirisian, about that bug you just posted, I have a quick question.."

[22:22:00.0000] <Hixie>
ah

[22:22:01.0000] <othermaciej>
Hixie: when I see an IP address comment that's insightful, sometimes I wonder who it is

[22:22:02.0000] <Hixie>
well if anyone wants to submit a patch to do something different, i'm happy to let them have the code :-)

[22:23:00.0000] <othermaciej>
Hixie: I don't make IP address comments myself, because I want myself and others to be able to find them later

[22:23:01.0000] <Hixie>
(personally though i don't really care who wrote the comment :-) )

[22:25:00.0000] <Sirisian>
Hixie, so you're really the person that reads all the issues and makes the changes?

[22:25:01.0000] <Hixie>
yup

[22:25:02.0000] <Sirisian>
How did you get picked?

[22:37:00.0000] <Hixie>
Sirisian: i volunteered

[22:38:00.0000] <Sirisian>
ah cool

[22:40:00.0000] <Hixie>
should i call the web database spec Web SQL Database, Web SQL Database API, Web SQLite Database, or something else?

[22:47:00.0000] <nessy>
not SQLite - the others sound ok

[22:49:00.0000] <Mannerisky>
for what it's worth, I think WSQLDAPI is a long-enough acronym

[22:51:00.0000] <Hixie>
i went for Web SQL Database

[22:52:00.0000] <MikeSmith>
good

[22:53:00.0000] <MikeSmith>
I think the API part isn't necessary in spec titles if the Web qualifier it in the title

[22:53:01.0000] <MikeSmith>
e.g., Geolocation API could well be "Web Geolocation"

[22:53:02.0000] <MikeSmith>
maybe

[22:54:00.0000] <Hixie>
i just stick "Web" on all my specs so that they start with a W instead of an X

[22:54:01.0000] <MikeSmith>
heh

[22:54:02.0000] <MikeSmith>
just a long as you don't start starting them with "WS-", please

[23:30:00.0000] <MikeSmith>
one feature that's in git and mercurial that I will I really had with subversion is the way you can interactively select specific parts of changes to commit

[23:30:01.0000] <MikeSmith>
"git add --patch" and "hg record"

[23:31:00.0000] <MikeSmith>
hsivonen: another good reason for moving to mercurial or git ↑

[23:50:00.0000] <MikeSmith>
Hixie: did the spec previously recommend doing "a img, img[usemap] { border: 0; }" in place of using the border attribute?

[23:51:00.0000] <Hixie>
i don't think html5 ever recommended specific CSS rules

[23:52:00.0000] <MikeSmith>
hmm, this is another case of I wonder where it came from then -- validator.nu warning about it advises to do "a img, img[usemap] { border: 0; }" instead

[23:53:00.0000] <MikeSmith>
similar to the advice about putting an id on nearest container instead of a/@name

[23:53:01.0000] <Sirisian>
heh. https://connect.microsoft.com/IE/SearchResults.aspx?FeedbackType=2&Status=1&Scope=0&SortOrder=15&TabView=1

[23:53:02.0000] <Sirisian>
number 1 bug is that the canvas tag doesn't work.

[23:54:00.0000] <MikeSmith>
I guess hsivonen included that advice in the initial version of the warnings-reporter when he added it

[23:57:00.0000] <jgraham>
/me grumbles about getting 36 bugzilla emails in 9 hours, all from the same bug

[23:57:01.0000] <foolip>
w3c spec bug, perchance?

[23:58:00.0000] <Hixie>
this is a terrible article http://en.wikipedia.org/wiki/Push_technology

[23:58:01.0000] <jgraham>
foolip: Indeed

[00:00:00.0000] <foolip>
not that I'm a masochist, but is there a way to follow those bugs with manually CC:ing yourself on all those which interest you?

[00:00:01.0000] <foolip>
er, manually adding yourself to the CC list rather

[00:00:02.0000] <jgraham>
Yeah, ou can do that from the bugzilla page

[00:01:00.0000] <foolip>
jgraham: care to link to that particular flame war for my breakfeast entertainment?

[00:01:01.0000] <jgraham>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=8404

[00:02:00.0000] <jgraham>
Not sure if it is a flame war, I haven't actually read any of the mail yet

[00:02:01.0000] <foolip>
go <figure>

[00:02:02.0000] <Hixie>
can anyone find a reference to the term "long polling" before the HTTP technique?

[00:03:00.0000] <Hixie>
did someone just make that up for HTTP for no good reason? or does it actually mean something?

[00:03:01.0000] <foolip>
I thought it quite specifially referred to keeping HTTP connections open

[00:25:00.0000] <Sirisian>
I've never heard of it used except in the context of the HTTP thing. It's a common practice. I've used it before since it allows you to communicate back and forth with a client using ajax and pretend like it's a socket. Then again you can only keep the connection open for 30 seconds.

[00:26:00.0000] <Hixie>
if the term was invented for HTTP specifically, then "long polling" is a pretty dumb name for it

[00:27:00.0000] <Hixie>
then again, the web is full of dumb names

[00:27:01.0000] <Hixie>
DHTML, REST, AJAX, Comet...

[00:55:00.0000] <Philip`>
About identifying bugs: I do it so that I can later search for the bugs I filed and see if they were addressed satisfactorily or not

[00:57:00.0000] <Hixie>
yeah, but that just needs a unique string, it doesn't need an identifier

[01:01:00.0000] <Philip`>
A unique string is an identifier :-)

[01:03:00.0000] <Hixie>
i mean, a person identifier

[01:03:01.0000] <Hixie>
that is recognisable as such

[01:18:00.0000] <gsnedders>
We need some spec for normalizing web addresses

[01:19:00.0000] <gsnedders>
As there are limits to how much percent encoding normalization you can do

[01:19:01.0000] <gsnedders>
You can't, for example, normalize pct-encoded & and = without breaking sites

[01:20:00.0000] <Hixie>
when would this be done?

[01:20:01.0000] <gsnedders>
Comparing IRIs for caching purposes, for example

[01:20:02.0000] <Hixie>
seems like something the IRI spec should define; send mail to larry

[01:22:00.0000] <Lachy_>
gsnedders, how would it break sites if it's being done for comparison purposes?

[01:23:00.0000] <gsnedders>
Lachy_: You use the cache for the wrong page due to over-normalizing

[01:23:01.0000] <gsnedders>
So you display the wrong page to the user

[01:23:02.0000] <gsnedders>
Which seems _really_ bad.

[01:24:00.0000] <Lachy_>
ok, can you give an example of over normalising?

[01:24:01.0000] <gsnedders>
http://www.publicbroadcasting.net/mpr/.eventsfeed?cat=Rock%2C%20Roots%2C%20%26%20Other&start=10122009&end=12312010 --- decode the escaped ampersands in there, and you get a different resource

[01:25:00.0000] <Lachy_>
ah, ok

[01:25:01.0000] <gsnedders>
If you normalize before cache lookup/save, that can break badly.

[01:26:00.0000] <Lachy_>
normalising should never decode ampersands, so I don't get why you're calling it over-normalising. But I get what you mean anyway.

[01:26:01.0000] <gsnedders>
Lachy_: Why shouldn't it?

[01:26:02.0000] <gsnedders>
Lachy_: ampersands are just another character in the query string

[01:27:00.0000] <gsnedders>
Lachy_: Point me to the bit in the spec that says I shouldn't normalize pct-encoded ampersands.

[01:27:01.0000] <Lachy_>
well, on the server side, I guess.

[01:27:02.0000] <Lachy_>
on the client side, you'd never want %26 to be wrongly interpreted as & in a query sting

[01:28:00.0000] <Lachy_>
outside of the query string, decoding it would be fine though

[01:29:00.0000] <Lachy_>
similar issue with %3F for ?

[02:00:00.0000] <Philip`>
/me is disappointed to find that <!DOCTYPE html PUBLIC “-//WHATWG//HTML 20100401"> is almost but not quite quirky, by the rules in http://philip.html5.org/docs/quirks.txt

[02:01:00.0000] <Philip`>
s/“/"/

[02:02:00.0000] <Philip`>
(Disappointed because it was almost a practical demonstration of why making up random strings is a bad idea)

[06:45:00.0000] <Philip`>
Someone less lazy than me should suggest proposing <meta name="designed-for-specs" content="http://www.w3.org/TR/html5/"> or content="http://www.w3.org/TR/2009/WD-html5-20090825/ http://www.w3.org/TR/2009/WD-rdfa-in-html-20091015/" etc

[06:45:01.0000] <Philip`>
instead of trying to squeeze the information needlessly into the most complex syntactic feature of HTML5 with the most compatibility problems

[06:46:00.0000] <TabAtkins>
Okay.

[06:47:00.0000] <TabAtkins>
I can't play this "You're wrong", "That's your opinion", "You're wrong that it's an opinion" game with Shelley for too long.

[06:47:01.0000] <gsnedders>
TabAtkins: Can you stop filling up that mailbox

[06:48:00.0000] <Philip`>
It seems unlikely to be productive to say "You're wrong" to people

[06:48:01.0000] <TabAtkins>
Philip`: And when they're actually wrong?  And keep saying things that are wrong, and using those wrong things as justification for their arguments?

[06:48:02.0000] <Philip`>
Better to just state your arguments and reasoning, and trust that reasonable people will agree with you if you're right

[06:48:03.0000] <TabAtkins>
It's really hard to see what one can do except point out, "Uh, yeah, here are examples proving you wrong."

[06:48:04.0000] <Philip`>
(and trust that reasonable people will determine what goes in the spec)

[06:49:00.0000] <jgraham>
Philip`: It's not clear that the processed is biased in that way

[06:49:01.0000] <jgraham>
*process

[06:49:02.0000] <TabAtkins>
gsnedders: Sorry, but to be fair, I'm not the largest source of emails on that thread.

[06:50:00.0000] <Philip`>
/me looks at the thread again

[06:50:01.0000] <Philip`>
Oh, Shelley already basically said what I just said

[06:50:02.0000] <TabAtkins>
Philip`: It's just impossible to actually *argue* with someone who is saying things that are demonstrably wrong.  If you don't correct them, then *at best* you're shouting over them and hoping to reach the choir.

[06:51:00.0000] <annevk>
You could just ignore it

[06:51:01.0000] <Philip`>
So don't waste so much effort trying to argue with them :-)

[06:51:02.0000] <jgraham>
/me has now got 56 messages in 18 hours or so

[06:52:00.0000] <Philip`>
s/argue with/respond to/

[06:52:01.0000] <jgraham>
s/18/20

[06:52:02.0000] <jgraham>
From that bug

[06:54:00.0000] <TabAtkins>
annevk: Yeah, I could.  But then I'd have to address the arguments later on when they start wielding the process like a hacksaw.  And in the meantime it's possible that people are swayed over to the wrong side due to them not realizing that there are false arguments being employed.

[06:56:00.0000] <jgraham>
TabAtkins: You may as well ignore all bugs. Eventually the controversal ones will all get turned into change proposals at which point they have to be unignored

[06:57:00.0000] <jgraham>
Nothing you say upfront can possibly prevent that happening

[06:57:01.0000] <TabAtkins>
Can quirks mode trigger parsing changes?  That is, can a page produce a different DOM based on which mode it's in?

[06:57:02.0000] <annevk>
yes

[06:57:03.0000] <jgraham>
TabAtkins: Yes, there is one difference

[06:57:04.0000] <TabAtkins>
jgraham: Sigh, I guess you're right.  But they're Wrong On The Internet.

[06:57:05.0000] <TabAtkins>
annevk, jgraham: k thanks.

[06:58:00.0000] <TabAtkins>
Philip`: Suggestion made.

[07:02:00.0000] <AryehGregor>
TabAtkins, the process doesn't require anyone to talk to each other.  If the group doesn't come to consensus, you just have to write the proposals to convince the chairs, as far as I can tell.

[07:03:00.0000] <TabAtkins>
AryehGregor: Sure, it doesn't require it.  But discussion is the whole point of having a list.

[07:03:01.0000] <Philip`>
TabAtkins: Thanks

[07:04:00.0000] <AryehGregor>
TabAtkins, well, no.  Getting a better spec is the point of having the list.  Discussion is only a means to an end.

[07:04:01.0000] <TabAtkins>
Sorry, I'll clarify.  Discussion is typically a vital part of producing a better spec.

[07:14:00.0000] <AryehGregor>
Some types of discussion are.  Some discussion is just noise.

[07:17:00.0000] <daedb>
I can't believe there's actually a suggestion to allow iframe in figure, but not directly allow tables/code/etc in there...

[07:19:00.0000] <annevk>
currently those are allowed

[07:20:00.0000] <daedb>
I know, I'm reading the bug that wants to disallow them

[07:20:01.0000] <annevk>
oh that

[07:22:00.0000] <zcorpan_>
is "-//H//T M//L" a legal FPI?

[07:22:01.0000] <daedb>
I just stared in disbelief when I saw "another option would be to allow iframe"...

[07:23:00.0000] <annevk>
zcorpan_, does it matter?

[07:24:00.0000] <zcorpan_>
in other news the secret xml-stylesheet draft will hopefully soon be published as a FPWD

[07:24:01.0000] <Philip`>
daedb: Where was <iframe> suggested? (I only remember <object>)

[07:25:00.0000] <gsnedders>
zcorpan_: I think not, but I haven't a copy of the spec here

[07:25:01.0000] <Philip`>
Oh, in the bug

[07:25:02.0000] <daedb>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=8404#c28 "Another option is to add one more element: iframe"

[07:25:03.0000] <zcorpan_>
gsnedders: why? because "L" is not a legal language tag?

[07:26:00.0000] <zcorpan_>
or is "NONSGML" an allowed string and "T" is not?

[07:27:00.0000] <AryehGregor>
/me wants to see what happens when some of these Change Proposals actually get past the bureaucratic machinery and get resolved

[07:27:01.0000] <Philip`>
I usually use "figure" to express the semantics "this would cross over a column/page break if I inserted it inline"

[07:27:02.0000] <gsnedders>
zcorpan_: http://xml.coverpages.org/tauber-fpi.html

[07:30:00.0000] <zcorpan_>
public_text_class = 

[07:30:01.0000] <zcorpan_>
CAPACITY | CHARSET | DOCUMENT | DTD | ELEMENTS | ENTITIES | LPD | NONSGML | NOTATION | SHORTREF | SUBDOC | SYNTAX | TEXT

[07:30:02.0000] <Philip`>
Use "-////TEXT HT//ML"

[07:32:00.0000] <zcorpan_>
Maltese?

[07:32:01.0000] <Philip`>
Malayalam

[07:32:02.0000] <Philip`>
obviously

[07:32:03.0000] <zcorpan_>
ah

[07:32:04.0000] <Philip`>
Oh, how confusing

[07:33:00.0000] <Philip`>
http://www.w3.org/WAI/ER/IG/ert/iso639.htm puts ml next to Maltese in the 3-letter section

[07:33:01.0000] <Philip`>
but it's MT in the 2-letter section

[07:34:00.0000] <Philip`>
Guess that's a bug in the W3C page

[07:37:00.0000] <Philip`>
Anyway, I guess it doesn't matter since Maltese and Malayalam are equally accurate descriptions of the markup language

[07:45:00.0000] <foolip>
AryehGregor: had any luck writing a change proposal of your own?

[07:45:01.0000] <AryehGregor>
foolip, no, nothing much I want changed that the editor doesn't.

[07:46:00.0000] <foolip>
AryehGregor: I mean the counter-proposal, the "empty change proposal" as it were

[07:46:01.0000] <AryehGregor>
I haven't tried writing one yet.

[07:46:02.0000] <foolip>
k

[07:46:03.0000] <AryehGregor>
I've thought of it, but I don't have the time to do it justice, and am not really the most qualified person on any given issue that I've seen.

[07:47:00.0000] <foolip>
sorry, I'm being a retard, I mistook you for TabAtkins just ehn

[07:47:01.0000] <AryehGregor>
I'm mostly interested in what will happen when people figure out that the HTMLWG is more or less just a dictatorship of the chairs instead of the editor, in the end.  Just with a lot more politics and bureaucracy.

[07:47:02.0000] <Philip`>
foolip: Easy mistake to make, since they're both CamelCase names

[07:48:00.0000] <foolip>
Philip`: exactly!

[07:48:01.0000] <TabAtkins>
foolip: Sorry, I was at work until 10pm last night raging against this server.  I"m doing final troubleshooting with our network admin before we wipe the fucker and start over.

[07:48:02.0000] <TabAtkins>
It's planned for lunch today instead.

[07:48:03.0000] <foolip>
haha, we can always request a new deadline in a week if things are too busy

[07:48:04.0000] <TabAtkins>
If necessary, I'll do that.  But I think I can get it out today.

[10:00:00.0000] <TabAtkins>
I'm still not sure what it is that Lief is suggesting with this "caption" idea.

[10:00:01.0000] <TabAtkins>
In his examples he seems to just be renaming <figure> to <cap> and then throwing in <object> for some reason.

[10:30:00.0000] <foolip>
<hascaption><img><caption></hascaption> :)

[10:36:00.0000] <gsnedders>
IE creates a single tag for <span<span> right?

[10:36:01.0000] <gsnedders>
s/tag/element/

[10:37:00.0000] <gsnedders>
Whereas Firefox and WebKit create two (and Opera agrees with IE)

[11:19:00.0000] <hober>
TabAtkins: I'm usually unsure of what Lief is proposing in his emails, on any subject...

[11:19:01.0000] <TabAtkins>
I suppose that's true.

[11:29:00.0000] <othermaciej>
Lief often says that I proposed something which to me seems counter to the statements I actually made

[11:30:00.0000] <Philip`>
s/Lief/Leif/

[11:31:00.0000] <gsnedders>
You need the g flag.

[11:33:00.0000] <Philip`>
No I don't, these are magic regexps

[12:09:00.0000] <jgraham>
gsnedders: You still there?

[12:32:00.0000] <TabAtkins>
Dammit, perl is hard to learn when you're doing so by debugging a mildly complex working program.

[12:36:00.0000] <Philip`>
Easier than debugging a program that doesn't work

[12:37:00.0000] <Dashiva>
I didn't know it was possible to write perl that's only mildly complex

[12:40:00.0000] <TabAtkins>
This is a pretty straightforward cgi processing a form and sending an email.  It's only mildly complex.

[12:41:00.0000] <TabAtkins>
I also find it strange that when a cgi script fails you get a "file not found" message from apache.

[12:49:00.0000] <charles_>
Is there any way to get the data of a media element in javascript (html5)?

[12:49:01.0000] <charles_>
Something like getDataURL from canvas??

[12:51:00.0000] <Philip`>
charles_: You could draw a <video> onto a <canvas> and then use toDataURL

[12:51:01.0000] <Sonja>
hullo

[12:51:02.0000] <Philip`>
though I'm not sure whether it'd return the current frame or not

[12:51:03.0000] <charles_>
Philip, thanks but had audio in mind

[12:51:04.0000] <Philip`>
Oh, okay

[12:51:05.0000] <Philip`>
No way to access that

[12:52:00.0000] <charles_>
Seems to be lacking from standard and what if you wished to make an audio editor??

[12:52:01.0000] <Philip`>
There have been discussions about creating some kind of 'audio canvas' for editing

[12:52:02.0000] <charles_>
Philip, there are other uses as well

[12:52:03.0000] <Philip`>
but nothing concrete yet

[12:52:04.0000] <Dashiva>
The usual way is to spec simple features first, and return and add complex ones later

[12:53:00.0000] <charles_>
THANKS, I'd love to know about it and will keep an eye open

[12:53:01.0000] <Sonja>
will html5 make it possible for me to write a little roguelike browser game with Ruby without needing to use Flash or othetr RIA?

[12:53:02.0000] <charles_>
For me the toDataURL would be all I need

[12:54:00.0000] <charles_>
Sonja, as a game developer, I'd say YES

[12:54:01.0000] <Sonja>
cool

[12:54:02.0000] <Sonja>
the person in control of html5 is Ian Hickson, right?

[12:54:03.0000] <charles_>
Sonja, canvas for the well canvas and audio for sound effects

[12:54:04.0000] <Dashiva>
Sonja: That should be possible now

[12:54:05.0000] <Dashiva>
It's just animated ascii art

[12:55:00.0000] <Sonja>
Dashiva how would i animate the ascii ? using ajax ?

[12:55:01.0000] <charles_>
Sonja, you could even do it with custom fonts if you wish

[12:55:02.0000] <Sonja>
ah, like tile fonts

[12:55:03.0000] <Dashiva>
Sonja: Yes

[12:55:04.0000] <charles_>
Sonja, ajax is communication, you would use a timer event

[12:56:00.0000] <TabAtkins>
Heh, doing a roguelike with actual monospace text would be fun.  Nuts to <canvas>, ascii is where it's at.

[12:56:01.0000] <charles_>
Sonja, if it is a standalone game, you could load the entire thing and then play no need for ajax unless you really need it

[12:57:00.0000] <charles_>
Sonja, with custom font loaded that is most of the battle

[12:57:01.0000] <TabAtkins>
Just load it as the textContent of a <pre> and you're done.

[12:57:02.0000] <charles_>
TabAtkins, need anim with timer for so fun stuff

[12:57:03.0000] <TabAtkins>
Though I suppose you may want to color the characters.  In that case you just wrap them in spans and such.

[12:58:00.0000] <TabAtkins>
charles_: Yeah, of course, but on each frame you're just replacing the textContent.

[12:58:01.0000] <charles_>
or divs with fixed position

[12:58:02.0000] <charles_>
or spans of course

[12:58:03.0000] <TabAtkins>
Sure, but that's making things too difficult.

[12:58:04.0000] <Sonja>
each time the player hits an arrow key or a letter, it would load a new html page?

[12:59:00.0000] <charles_>
Sonja, nope, capture input events via javascript and away you go

[12:59:01.0000] <Sonja>
sounds like i need to learn javascript then. cool!

[12:59:02.0000] <Sonja>
and i can make javascript talk to ruby where the actual game is written in

[13:00:00.0000] <charles_>
Sonja, that would be the core of the game, using javascipt you could change the fonts all over the place and anim them too

[13:00:01.0000] <TabAtkins>
Sonja: Yeah, definitely learn javascript.  Anything else would be silly in terms of latency.  Learn jQuery to make it easy.

[13:00:02.0000] <TabAtkins>
And then yes, you can use ajax to talk back to the ruby program on the server and receive instructions.

[13:00:03.0000] <Sonja>
thanks. you guys are all very smrt and friendly!!

[13:00:04.0000] <charles_>
Sonja, maybe later ruby or lua or etc

[13:01:00.0000] <charles_>
Do it for a living and so have no choice ;)

[13:02:00.0000] <charles_>
Sonja, it is very important to do games in html5 (browser) because if you can do games you can do apps and that is a goal of html5

[13:03:00.0000] <charles_>
Thanks all will look for the audio canvas and see what I can find.

[13:03:01.0000] <Philip`>
It's important to do games because games are fun

[13:04:00.0000] <Dashiva>
You could make a javascript roguelike proxy

[13:04:01.0000] <Philip`>
You don't need to justify it with any other reasoning :-p

[13:04:02.0000] <Dashiva>
That just sends raw key events and receives a string array containing the screen

[13:06:00.0000] <Sonja>
neat. this makes javascript look like ruby. http://github.com/jedediah/prettyscript

[13:06:01.0000] <Philip`>
Just use a JS Telnet emulator

[13:07:00.0000] <jcranmer>
you say that as if it's a good thing

[13:08:00.0000] <Sonja>
haha aw

[13:08:01.0000] <Dashiva>
>Blocks are translated to anonymous functions (yay!)

[13:09:00.0000] <Dashiva>
Sounds efficient

[13:13:00.0000] <TabAtkins>
Not their fault javascript doesn't allow direct syntax hacking.

[13:13:01.0000] <TabAtkins>
All your higher-order functions have to be done with closures.

[13:13:02.0000] <jgraham>
/me always thought games were just designed as stress tests for implementations

[13:13:03.0000] <Sonja>
what are some good books to learn javascript/jquery from scratch?

[13:14:00.0000] <jgraham>
TabAtkins: You are veering dangerously close to lisp weenie territory :)

[13:14:01.0000] <TabAtkins>
jgraham: I am so far into that territory it isn't even funny.

[13:14:02.0000] <jgraham>
TabAtkins: No need o embaras yourself in public though ;)

[13:15:00.0000] <TabAtkins>
There's nothing wrong with being a lisp weenie.  We have to keep the rest of you imperative-lovers honest.

[13:16:00.0000] <jgraham>
/me always had the impression that lisp allowed for imperative-style programming

[13:16:01.0000] <jgraham>
at least compared to, say, haskell

[13:16:02.0000] <jgraham>
Sonja: Supposedly "javascipt: the good parts'" is OK

[13:17:00.0000] <Sonja>
thanks

[13:17:01.0000] <jgraham>
At least I seem to recall people sying that

[13:17:02.0000] <jgraham>
/me has never really read a book about javascript

[13:18:00.0000] <jgraham>
Sonja: I think there is a John Resig book about jQuery if you care about that in particular

[13:18:01.0000] <TabAtkins>
jgraham: Yeah, it certainly allows it.  It just really hurts to do so.

[13:19:00.0000] <TabAtkins>
jgraham: And then there's things that people *call* imperative, like LOOP, which aren't.  LOOP's just a delicious DSL.

[13:19:01.0000] <Sonja>
well TabAtkins suggested jquery "to make it easy"

[13:19:02.0000] <jgraham>
And he is releasing "Secrets of the Javacript Ninja" sometime. Which could be good. Or terrible

[13:20:00.0000] <jgraham>
Sonja: jQuery is mainly useful as a low-pain alterntaive to direct DOM manipulation

[13:20:01.0000] <TabAtkins>
Sonja: Yeah, the thing is that while javascript is a great language, the DOM (the functions you use to communicate with/manipulate the page) is *not*.  jQuery papers all that over with a beautiful, consistent syntax.

[13:20:02.0000] <TabAtkins>
I've never learned DOM scripting, and don't intend to.  jQuery is just too beautiful.

[13:21:00.0000] <jgraham>
The DOM is indeed eye-bleedingly ugly but it is not exactly hard

[13:22:00.0000] <TabAtkins>
jgraham: I'll note, though, that I'm only a few steps away from being a Haskell weenie too.  I keep dipping my toes in the pool, and then backing out and just hacking the functionality I want into my lisp utils file instead (like lazy lists, woo!).

[13:23:00.0000] <jgraham>
/me is about three chpters in to "Real World Haskell" 

[13:23:01.0000] <jgraham>
Which is really fun but not something I spend enough time on

[13:23:02.0000] <TabAtkins>
It feels like http://learnyouahaskell.com/ is an awesome resource.

[13:23:03.0000] <Sonja>
cool

[13:23:04.0000] <TabAtkins>
In addition to being ridiculous in general.

[13:23:05.0000] <Sonja>
so i'd learn javascript and jquery instead of javascript and dom

[13:24:00.0000] <TabAtkins>
Sonja: Yeah.

[13:24:01.0000] <Sonja>
you guys are such a n00b-friendly channel. thanks! <3

[13:24:02.0000] <jgraham>
TabAtkins: That is clearly _why inspired. Anything that is _why inspired is inherently good. Even thouh I know no ruby...

[13:25:00.0000] <jgraham>
"This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 Unported License because I couldn't find a license with an even longer name. "

[13:26:00.0000] <Sonja>
_why quit the internets! :-O

[13:26:01.0000] <Sonja>
jgraham haha

[13:27:00.0000] <jgraham>
Sonja: indeed :(

[13:27:01.0000] <Sirisian>
Does w3c have their IRC server/channel?

[13:28:00.0000] <TabAtkins>
Sirian: Yeah, irc://irc.w3.org:80

[13:28:01.0000] <TabAtkins>
And then #htmlwg, though that's mostly dead but for the telecons.

[13:28:02.0000] <TabAtkins>
^^^ I meant Sirisian.

[13:29:00.0000] <Sirisian>
oh 80. I was trying 6665 like it said on a random site

[13:29:01.0000] <Philip`>
You mean #html-wg

[13:29:02.0000] <TabAtkins>
jgraham: Yeah, definitely _why-inspired, and definitely great for that.  It pulls off the ridiculous properly as well.  You can't just be silly and expect to be good.

[13:29:03.0000] <TabAtkins>
Philip`: Sure.

[13:29:04.0000] <Sirisian>
TabAtkins, is it down?

[13:30:00.0000] <TabAtkins>
No?  I'm there right now, in the #css channel.

[13:30:01.0000] <TabAtkins>
Oh, *but* a bunch of people just got kicked a few minutes ago.

[13:31:00.0000] <TabAtkins>
So you may not have a route to it.

[13:31:01.0000] <Sirisian>
TabAtkins, ah I got it. My IRC client didn't understand when I was using a semi-colon. Definitely wanted a space. Thank you

[13:31:02.0000] <Sirisian>
er colon*

[13:32:00.0000] <TabAtkins>
Sirisian: Ah, k.  You can just drop the irc:// in the front, and just do a /server irc.w3.org:80

[13:32:01.0000] <TabAtkins>
Or wait, you mean the port colon?

[13:32:02.0000] <TabAtkins>
Weird.

[13:32:03.0000] <Sirisian>
x-chat is odd. /server irc.w3.org 80

[13:32:04.0000] <TabAtkins>
Strange.

[14:01:00.0000] <TabAtkins>
othermaciej: Where should I ask for an extension on writing a change proposal for issue 76?  I'm writing it tonight, but I want to run it past foolip first, and don't want to run the risk of overrunning the tomorrow deadline.

[14:02:00.0000] <othermaciej>
TabAtkins: in an email reply to the call for counter-proposals

[14:02:01.0000] <othermaciej>
TabAtkins: request for extension will probably be viewed more favorably if you link a rough draft and say you need some extra time to finalize it (with a specific deadline for finishing)

[14:03:00.0000] <TabAtkins>
Cool, I'll do so tonight.

[14:03:01.0000] <othermaciej>
TabAtkins: thank you sir

[14:17:00.0000] <TabAtkins>
Argh she has no idea what a cowpath is why god why.

[14:17:01.0000] <Sirisian>
wow who doesn't know what that is? >_>

[14:18:00.0000] <TabAtkins>
It's over on the htmlwg list.

[14:18:01.0000] <Philip`>
TabAtkins: A cowpath is when you lay out a long strip of concrete for the cows to walk down, even if the cows almost never bother doing so

[14:18:02.0000] <Dashiva>
Well, if you don't know the term it's not that outlandish to assume cowpath = any path that has been walked, as opposed to paved roads

[14:19:00.0000] <Philip`>
and even if the cows that do walk down it get lost and end up falling into a ditch

[14:20:00.0000] <TabAtkins>
Laughing at that mental image probably makes me a bad person.

[14:47:00.0000] <shelleyp>
TabAtkins, you don't like me. Fine. But you have to question if you're responding to me because you disagree, or just because its me

[14:47:01.0000] <TabAtkins>
shelleyp: I respond because I disagree.  My like or dislike of you has nothing to do with it.

[14:48:00.0000] <shelleyp>
I find that unlikely, when I see the petty sniping that goes on in this group

[14:48:01.0000] <shelleyp>
You have a problem with what I write on the email list, take it up with me, face to face

[14:49:00.0000] <TabAtkins>
We've been over this before.  I do challenge you on something when it's important.  When it's not and I'm merely annoyed, there's no need to.

[14:50:00.0000] <shelleyp>
Then you can take it up with me via email. But I'm tired of having people point out to me the latest nasty little quip about me on this children's hour called an IRC channel

[14:50:01.0000] <TabAtkins>
Then... don't listen about the IRC channel?  It's not like anyone's insulting you.  There aren't low blows.

[14:51:00.0000] <othermaciej>
/me notes irony of shelleyp's last statement

[14:51:01.0000] <TabAtkins>
What, that the reason I don't take it up in email is precisely because it's not important enough to take up in email?

[14:52:00.0000] <TabAtkins>
Or the nasty little quip followed by "children's hour called an IRC channel"?

[14:53:00.0000] <othermaciej>
the fact that she used a nasty little quip to complain about nasty little quips

[14:53:01.0000] <TabAtkins>
Perhaps I should point it out to her.  Nah, she says she's tired of that.

[14:56:00.0000] <TabAtkins>
shelleyp: You'll read this later, so I can address it to you here.  I've discussed with you before (over email!) the differences between high-latency, high-cost communications like email and low-latency, low-cost communications like IRC.  Some things are appropriate in one that are not appropriate in the other.  I route all my important communications through email where I can collect my...

[14:56:01.0000] <TabAtkins>
...thoughts best, so feel free to ignore me in IRC.


2009-12-02
[20:31:00.0000] <MikeSmith>
/me peruses http://wiki.commonjs.org/wiki/Binary/D

[21:12:00.0000] <MikeSmith>
othermaciej: very cool to see Wildfox message about SVG filters support

[21:13:00.0000] <othermaciej>
MikeSmith: oh?

[21:13:01.0000] <MikeSmith>
on webkit-dev

[21:46:00.0000] <erlehmann>
SVG filters are PURE GOLD

[21:46:01.0000] <erlehmann>
i liek the whole displacement mapping thing best

[21:50:00.0000] <shepazu>
erlehmann: is that because of mariocart? http://www.tapper-ware.net/stable/PerspectiveWith6LinesOfXML-CanvasEnhanced/index.xhtml

[21:52:00.0000] <erlehmann>
shepazu, that would be one example. but animated displacement maps (gif files) are fine too.

[01:22:00.0000] <gsnedders>
http://www.w3.org/2001/tag/doc/content-to-apps.html

[01:24:00.0000] <Philip`>
2001? That's an old document

[01:25:00.0000] <gsnedders>
Since when did dates in W3C URIs mean anything?

[01:26:00.0000] <Philip`>
Since humans started extracting obvious patterns and meaning from text that they read

[01:28:00.0000] <Hixie>
annevk: yt? what spec should define how the 'abort' event fires when the "unload a document" algorithm for the xhr's browsing context's document is run, if any?

[01:47:00.0000] <annevk>
hmm, prolly XHR :/

[01:48:00.0000] <annevk>
well, which abort event?

[01:48:01.0000] <Hixie>
apparently (I haven't tested) 'abort' fires on XHR objects that are pending when the document is unloaded

[01:50:00.0000] <Hixie>
i really need a better term for "first script"

[02:01:00.0000] <annevk>
responsible script?

[02:01:01.0000] <annevk>
or something like that

[02:01:02.0000] <othermaciej>
what's special about the first script?

[02:03:00.0000] <othermaciej>
I see

[02:03:01.0000] <othermaciej>
the similar concept in WebKit is "dynamic global object"

[02:04:00.0000] <othermaciej>
the dynamic global object is the one where a particular nested chunk of script execution began

[02:06:00.0000] <othermaciej>
"first script" is not a great term, since an event listener is a function, not a script

[02:07:00.0000] <othermaciej>
"code entry-point" seems like a needlessly obscure term for a "function"

[02:08:00.0000] <othermaciej>
functions are also not necessarily associated at all with a script, in the sense of a <script> element anyway

[02:08:01.0000] <othermaciej>
it would be better to skip right to the global object from the function and bypass the concept of "script"

[02:19:00.0000] <Hixie>
code entry-point doesn't mean function

[02:19:01.0000] <Hixie>
e.g. <script>/*this is a code entry-point*/</script>

[02:19:02.0000] <Hixie>
it's not really to do with the global object either

[02:20:00.0000] <Hixie>
e.g. the script's base URL can be different for two scripts with the same global object

[02:22:00.0000] <othermaciej>
how?

[02:23:00.0000] <Hixie>
<script xml:base="a">...</script><script xml:base="b">...</script>

[02:24:00.0000] <othermaciej>
do browsers actually use the "script's base URL" as defined there, rather than the base URL of the script's global object's browsing context's active document?

[02:24:01.0000] <Hixie>
oh actually that's not true, my bad

[02:25:00.0000] <othermaciej>
what doing relative URL resolution for window.open for instance?

[02:25:01.0000] <Hixie>
no you're right

[02:25:02.0000] <Hixie>
the spec just has script's base URL as being a shorthand way of saying that

[02:25:03.0000] <Hixie>
i suppose i could remember the global object instead of the script and call it the dynamic global object

[02:26:00.0000] <othermaciej>
I'm not sure "dynamic global object" will be more clear to people necessarily, that was just a cite of our usage

[02:26:01.0000] <othermaciej>
we contrast it with the "lexical global object" which is the one where the currently running function was initially defined

[02:27:00.0000] <Hixie>
yeah i don't know if it's better enough to be worth it

[02:27:01.0000] <Hixie>
"first script" is confusing people though

[02:27:02.0000] <jgraham>
fwiw "dynamic global object" as a set of words doesn't really convey anything to me

[02:27:03.0000] <othermaciej>
I'm just saying that the "script" notion does not really add a useful abstraction, in terms of what actually needs to happen

[02:27:04.0000] <Hixie>
jgraham: that's a good thing, compared to "first script", which conveys something wrong. :-)

[02:28:00.0000] <othermaciej>
"outermost global object" would perhaps be better in context, though it might connote outermost frame instead of outermost in the sense of execution nesting

[02:28:01.0000] <Hixie>
othermaciej: it's a useful abstraction because there are three kinds of global objects (Window, WorkerGlobalScope, and the empty object)

[02:28:02.0000] <Hixie>
is there some term referring to the first stack frame?

[02:28:03.0000] <Hixie>
which we could adopt here?

[02:29:00.0000] <Hixie>
othermaciej: (though it could certainly be phrased in a different way than it is now)

[02:31:00.0000] <othermaciej>
too verbose but likely unambiguous: outermost script execution entry point

[02:31:01.0000] <Hixie>
entry script?

[02:31:02.0000] <Hixie>
might be better than first script, at any rate

[02:32:00.0000] <othermaciej>
"first script

[02:32:01.0000] <othermaciej>
" sounds too much like the first one that appears in the page

[02:32:02.0000] <Hixie>
right

[02:33:00.0000] <othermaciej>
is there any way in which an impotent script's base URL can be used?

[02:33:01.0000] <othermaciej>
I think the answer is no so the empty object case doesn't matter

[02:34:00.0000] <othermaciej>
anyway, "entry" might have less chance of producing confusion

[02:36:00.0000] <Hixie>
there isn't, and indeed its script's base URL is explicitly not set currently

[02:37:00.0000] <Hixie>
anyway, i'll leave this stuff for now

[02:37:01.0000] <Hixie>
websockets next, and then i'm ready for webapps LCs

[02:37:02.0000] <Hixie>
only 61 websocket e-mails

[02:54:00.0000] <Hixie>
/me looks at http://dev.w3.org/2009/dap/camera/Overview.html and wonders why <input type=image> as specced in html5 isn't enough

[02:54:01.0000] <Hixie>
(for what that api can do)

[03:16:00.0000] <Lachy>
Hixie, presumably, you mean, <input type=file>?

[03:16:01.0000] <Hixie>
uh yes

[03:26:00.0000] <virtuelv>
Hixie: because not everything is about grabbing a static image, deferring any control over the photo/film process to a native implementation

[03:27:00.0000] <virtuelv>
(I don't neccesarily completely agree with the approach in the spec, but <input type="file"> would be horribly insufficient, if you were to reimplent something like Layar, or the google turn-by-turn app from Android 2.0

[03:27:01.0000] <Hixie>
certainly <input type=file> isn't the ultimate solution for everything

[03:28:00.0000] <Hixie>
i just meant for doing what was in the api spec

[03:30:00.0000] <virtuelv>
two problems (which can probably be overcome

[03:30:01.0000] <virtuelv>
1) Media data from a device is not "files"

[03:30:02.0000] <virtuelv>
the main purpose of an API is to grab the data, not neccesarily for upload, but for later processing

[03:31:00.0000] <Hixie>
<input type=file> exposes a URL the same way the API does, these days

[03:31:01.0000] <virtuelv>
2) <input type="file"> offers little control over the user interface, and would force the user to make an informed choice

[03:32:00.0000] <Hixie>
the api doesn't give much control either

[03:32:01.0000] <Hixie>
as written today

[05:53:00.0000] <glen1>
anyone know what the <figure/> tag mean

[05:56:00.0000] <annevk>
see the spec

[06:00:00.0000] <glen1>
the w3c page crashes my browser lol

[06:00:01.0000] <glen1>
is it for pictures and stuff

[06:01:00.0000] <foolip>
try the multipage version, it may crash your browser less

[06:01:01.0000] <glen1>
lol html5 sucks

[06:03:00.0000] <Philip`>
That was insightful

[06:03:01.0000] <annevk>
IRC always is

[06:04:00.0000] <foolip>
does anyone know how to set up irssi to show part messages for people who have spoken in the last x minutes, but ignore the rest?

[06:05:00.0000] <foolip>
I assume we have many irssi users here

[06:05:01.0000] <Philip`>
/me uses irssi but has no idea

[06:07:00.0000] <murr5y>
dare you do a version ctcp? :)

[06:40:00.0000] <foolip>
TabAtkins: well written, hope it helps

[06:40:01.0000] <TabAtkins>
Thanks, yeah.

[06:53:00.0000] <zcorpan_>
Hixie: not sure i follow the new text on interface objects in workers... should ImageData be visible to a worker?

[06:55:00.0000] <annevk>
it seems so

[06:56:00.0000] <annevk>
in particular because of "and the interfaces of any objects made accessible through APIs"

[07:00:00.0000] <zcorpan_>
and it's made accessible because you can postMessage an instance of an ImageData?

[07:08:00.0000] <annevk>
thinking about it more I guess it's not quite clear

[07:08:01.0000] <annevk>
well, it depends on how you interpret things

[07:08:02.0000] <annevk>
ImageData is made accessible through an API...

[08:00:00.0000] <virtuelv>
Hixie: yt?

[09:10:00.0000] <jgraham>
Argh

[09:12:00.0000] <Philip`>
?

[09:13:00.0000] <jgraham>
/me has has had even more enough of the bug 8404 discussion than before

[09:15:00.0000] <jgraham>
It is like an initially minor itch that once scratched became pussy and started to bleed

[09:15:01.0000] <virtuelv>
oh well, I'm advocating <input type="file" for camera capture now

[09:16:00.0000] <virtuelv>
there's however a bit there about selecting a camera on clicking,

[09:16:01.0000] <virtuelv>
good approach for microdata?

[09:34:00.0000] <Philip`>
jgraham: I suggest amputation

[09:38:00.0000] <TabAtkins>
jgraham: I suggest writing the word as "puss-y" to avoid ambiguous connotations.

[09:47:00.0000] <adactio>
Hmmm... I just had a random thought... shouldn't the datalist element allow the optgroup element?

[09:55:00.0000] <TabAtkins>
<datalist> doesn't exist anymore, does it?

[09:56:00.0000] <myakura>
TabAtkins: I think that's <datagrid> which doesn't exist. I believe <datalist> does.

[10:01:00.0000] <adactio>
Datalist is alive and well (and implemented in Opera).

[10:02:00.0000] <jgraham>
TabAtkins: Yeah I noticed tht afterwards

[10:23:00.0000] <BenMillard>
evening all, long time since I've dropped by here :)

[10:24:00.0000] <BenMillard>
heh, it was "285 days ago" according to this: http://gavinsharp.com/irc/whatwg.html

[10:24:01.0000] <webben>
evening BenMillard :)

[10:25:00.0000] <BenMillard>
webben, I recognise your name (and several others)

[10:25:01.0000] <BenMillard>
I'm on a one-week holiday from my full-time website developer job

[10:26:00.0000] <BenMillard>
I'm going to apply to work at Opera, so I can help refine HTML5 and resume studying how authors use HTML, but I'm not sure if they have a job opening which fits that

[10:26:01.0000] <BenMillard>
"Core Technology Testers (Web technology experts)" is the closest I've found: http://www.opera.com/company/jobs/opening/67/

[10:30:00.0000] <BenMillard>
jgraham, I've been using your Document Map add-on for Firefox during the past few months and really like it!

[10:30:01.0000] <BenMillard>
jgraham, I'm using 0.6.1 - is that the most recent version?

[10:30:02.0000] <webben>
BenMillard: Good luck! :)

[11:17:00.0000] <miketaylr>
and the troll of the day award goes to: http://twitter.com/slicknet/status/6276901024

[11:17:01.0000] <miketaylr>
http://twitter.com/slicknet/status/6278180943

[11:56:00.0000] <foolip>
/me reads http://edward.oconnor.cx/2009/12/RDFa-and-the-DOM

[12:04:00.0000] <Philip`>
On the subject of text vs (DOM|Infoset|any sane model), see also http://github.com/msporny/rdfa-test-suite/commit/a27e4f4500a854024b0e3914af13f6227d54d805

[12:04:01.0000] <foolip>
wow, that was actually very well worth reading

[12:05:00.0000] <Philip`>
which seems to suggest SAX parsers may get confused by thinking markup like &lt;string&gt; is identical to <string>

[12:05:01.0000] <foolip>
(Edward O'Connor, not the git commit...)

[12:06:00.0000] <foolip>
hmm, a well written commit message though, I never bother writing that much

[12:07:00.0000] <Philip`>
It's a good explanation of the rationale behind the change, I just think the rationale is completely wrong and misguided :-)

[12:08:00.0000] <foolip>
I'm not sure I understand though

[12:08:01.0000] <Philip`>
(and caused by thinking the raw byte stream is important)

[12:08:02.0000] <foolip>
why would seeing <strong> in a text node cause you to think that's an XML literal?

[12:09:00.0000] <Philip`>
http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2009Nov/0099.html - "The first was an issue that would occur on SAX-based processors that did not have access to the raw byte stream and thus wouldn't be able to tell if the encoded "<strong>" value should be interpreted as an XML Literal or not."

[12:09:01.0000] <Philip`>
foolip: It wouldn't

[12:11:00.0000] <foolip>
I just don't know enough about how XML Literals are encoded or how SAX parsers work to understand

[12:11:01.0000] <foolip>
but that's OK, because I have packing to do and am flying to Vietnam on Friday

[12:11:02.0000] <foolip>
yay me

[12:14:00.0000] <Philip`>
They shouldn't be encoded at all - the idea is that when you're converting an element to a value, if it has child elements then you serialise the element's content into an XMLLiteral string, else you use the content as a typeless string

[12:14:01.0000] <Philip`>
which is nice and easy when you consider XML documents to be a tree/stream of elements and text nodes

[12:14:02.0000] <foolip>
right

[12:15:00.0000] <foolip>
so how do "SAX-based processors that did not have access to the raw byte stream" see XML?

[12:15:01.0000] <foolip>
last time I used SAX for anything I was feeding it XML as a string and got parse events

[12:16:00.0000] <Philip`>
but if you don't have that model then even the concept of "serialise the element's content" makes no sense, so it's sort of defined as copying the characters contained within the element, except munging it a bit to add namespace declarations, and also it needs canonicalisation, and the specification is really fuzzy

[12:16:01.0000] <foolip>
(or something along those lines)

[12:16:02.0000] <Philip`>
SAX processors see a stream of startElement and endElement and characters events

[12:16:03.0000] <Philip`>
(or something very similar to that)

[12:16:04.0000] <foolip>
oh, on that end

[12:16:05.0000] <foolip>
I was thinking backwards

[12:17:00.0000] <foolip>
of course

[12:18:00.0000] <foolip>
sounds like a problem very much not specific to RDFa though

[12:18:01.0000] <Philip`>
I don't think there's a problem at all

[12:19:00.0000] <Philip`>
Or, I don't think there's a problem at all with SAX

[12:19:01.0000] <Philip`>
or with writing RDFa processors on top of a SAX processor

[12:19:02.0000] <Philip`>
The only problem is the RDFa spec

[12:20:00.0000] <foolip>
what does it actually say to do produce an XML Literal?

[12:20:01.0000] <Philip`>
(which is mostly specific to RDFa, not other specs)

[12:20:02.0000] <foolip>
pretend it has the raw byte stream, I presume

[12:20:03.0000] <foolip>
oh well

[12:21:00.0000] <Philip`>
"The value of the [XML literal] is a string created by serializing to text, all nodes that are descendants of the [current element], i.e., not including the element itself, and giving it a datatype of rdf:XMLLiteral."

[12:22:00.0000] <Philip`>
plus errata: "For the avoidance of doubt, this means in part that the current default namespace of each descendant element MUST also be included in the emitted XML literal."

[12:22:01.0000] <Philip`>
plus errata: "Unfortunately, the requirements for such a serialized string are not completely clear when it comes to XML Namespaces. For the avoidance of doubt, in future versions of the RDFa Syntax Recommendation, we expect to indicate that the value of the XML literal SHOULD be serialized Exclusive Canonical XML as defined in [XMLEXC14N]. We advise current implementations to start supporting this use model now, since other participants in their XML tool ch

[12:22:02.0000] <hober>
foolip: thanks

[12:23:00.0000] <foolip>
sounds like it assumes a DOM

[12:23:01.0000] <Philip`>
Oh, so it does actually talk about serializing nodes

[12:23:02.0000] <foolip>
hober: oh there you are :)

[12:25:00.0000] <Philip`>
Discussions about the namespace/c14n requirements and RDFa-in-text/html seem to indicate that many people don't believe in the DOM-like model, though

[12:26:00.0000] <Philip`>
Anyway, I should probably stop complaining about this :-p

[12:26:01.0000] <foolip>
We probably all should

[12:27:00.0000] <foolip>
but while we're on the topic, I found this amusing: http://www.merttol.com/articles/web/understanding-rdfa.html

[12:27:01.0000] <foolip>
(see my comment at end)

[12:29:00.0000] <foolip>
I just hope that this whole RDFa issue doesn't have people hating *RDF*

[12:30:00.0000] <Philip`>
Not at all

[12:31:00.0000] <Philip`>
RDF is hated for its own unique reasons

[12:32:00.0000] <Dashiva>
RDF is too vague to be a proper target for hating

[12:34:00.0000] <foolip>
down with graphs!

[12:34:01.0000] <Philip`>
Down with data structures!

[12:34:02.0000] <Philip`>
Down with data!

[12:36:00.0000] <foolip>
PNG instead of HTML5?

[12:37:00.0000] <BenMillard>
foolip, that would put me and aobut 1,000,000 other people out of a job :(

[12:37:01.0000] <BenMillard>
s/aobut/about/

[12:37:02.0000] <Dashiva>
Poor Bobby McFerrin, doomed to never be a unique individual

[12:41:00.0000] <Dashiva>
Maybe RDFa will usher in a new age of less graphy RDF, one where almost all statements use string literals 

[12:43:00.0000] <Philip`>
Bonus points if you have RDFa embedded in markup inside string literals

[12:44:00.0000] <Sirisian_>
workmad3, why are you here?

[12:44:01.0000] <workmad3>
Sirisian_: just hovering :)

[13:13:00.0000] <TabAtkins>
othermaciej: Well, turns out I already had an account, but it doesn't *look* like I have editting privileges.  I guess I'll wait for MikeSmith to show up in irc.

[13:17:00.0000] <othermaciej>
TabAtkins: I think Laura Carlson also has the power to enable editing - not sure who else

[13:17:01.0000] <othermaciej>
TabAtkins: I'm not really sure why the wiki is restricted in the first place - I'm not aware of any vandalism problems

[13:21:00.0000] <Philip`>
Maybe there aren't any vandalism problems because it's restricted

[13:23:00.0000] <Dashiva>
Or maybe maciej doesn't consider offers on cheap airline tickets vandalism

[13:25:00.0000] <othermaciej>
TabAtkins: looks like you should already have editing rights according to the list

[13:25:01.0000] <othermaciej>
TabAtkins: in fact supposedly both of us have the right to grant editing rights - perhaps that is just done by editing the <http://esw.w3.org/topic/ESWEditorsGroup> page

[13:48:00.0000] <TabAtkins>
othermaciej: Hrm.  It's telling me I don't have permission to edit this page, where "this page" is the new one I'm trying to create.

[13:50:00.0000] <othermaciej>
TabAtkins: unfortunately I don't really understand how the Wiki is managed

[13:50:01.0000] <TabAtkins>
Yeah, me neither.  Shrug.

[13:50:02.0000] <othermaciej>
TabAtkins: I could paste in a copy of the document for you but I'm hesitant to put it in a place you can't edit

[13:51:00.0000] <TabAtkins>
You could paste it in, with a link to the location on my server for "latest version"?

[13:57:00.0000] <othermaciej>
your account is named TabAtkins?

[13:57:01.0000] <TabAtkins>
Yup

[14:08:00.0000] <TabAtkins>
othermaciej: Now that I think about it, I editted the Taxi page before TPAC.  So I do have editting rights.  Perhaps I just don't have page creation rights.

[14:08:01.0000] <TabAtkins>
If you can create the page for me I think it likely that I can edit it.

[14:10:00.0000] <TabAtkins>
(Which would explain why I don't see a "Create Page" button around anywhere, and had to manually put a new page in the url.)

[15:02:00.0000] <othermaciej>
TabAtkins: ok - what would you like the page to be called?

[15:02:01.0000] <TabAtkins>
I was attempting to call it Issue76CounterProposal, but use whatever you think is best.

[15:11:00.0000] <Philip`>
That naming convention doesn't scale well if there's more than two proposals

[15:11:01.0000] <hober>
Issue76StatusQuoChangeProposal?

[15:13:00.0000] <Lachy>
someone needs to put an end to this annoying figure vs. table discussion. The arguments are just getting repeated over and over again.


2009-12-03
[01:21:00.0000] <gsnedders>
Hmm, maybe Anolis should give dt ids as well

[01:33:00.0000] <zcorpan_>
"The WG has CONSENSUS to publish (in XML date space) http://www.w3.org/XML/Group/2009/09/xml-stylesheet.html" - http://www.w3.org/mid/CF83BAA719FD2C439D25CBB1C9D1D30211992D7C⊙Hppc

[01:35:00.0000] <jgraham>
Is consensus a magic word that means something different in all caps?

[01:35:01.0000] <zcorpan_>
don't think so, it's just all caps to draw attention to it, like ACTION and ISSUE

[01:36:00.0000] <annevk>
what does XML data space mean?

[01:36:01.0000] <annevk>
gsnedders, maybe just dts where the only child is dfn

[01:36:02.0000] <zcorpan_>
date space

[01:36:03.0000] <zcorpan_>
dunno

[01:38:00.0000] <jgraham>
Presumably it means "with a URI of the form http://www.w3.org/XML/Group/YYYY/MM/

[01:38:01.0000] <jgraham>
"

[01:39:00.0000] <zcorpan_>
hmm yeah probably

[01:44:00.0000] <annevk>
oh, so no FPWD?

[01:44:01.0000] <annevk>
hmm

[01:45:00.0000] <zcorpan_>
"... (with minor edits to the SotD and pubrules issues) as a FPWD of this PER."

[01:47:00.0000] <darobin>
heya

[01:47:01.0000] <darobin>
does anyone know what license the acid3 test is under?

[01:48:00.0000] <darobin>
I'm talking to someone who'd like to ship an offline test suite and they'd like to know

[01:58:00.0000] <annevk>
you'd have to email Hixie_ I think

[02:02:00.0000] <Philip`>
darobin: http://krijnhoetmer.nl/irc-logs/whatwg/20080507#l-11 - "<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!"

[02:05:00.0000] <Lachy>
Hixie_, in the Microdata examples section, several examples use itemtype="http://example.org/animals#cat", but then in the API section, you have another cats example using var cats = document.getItems("http://example.com/feline");

[02:05:01.0000] <Lachy>
Is it intentional that these examples use different URLs?

[02:28:00.0000] <virtuelv>
is there a particular reason why negative values for r0/r1 in createRadialGradient throws instead of being clamped to 0?

[02:31:00.0000] <Hixie_>
Lachy: off hand, no idea

[02:31:01.0000] <Hixie_>
virtuelv: off hand, probably that's what browsers did? dunno.

[02:32:00.0000] <virtuelv>
they don't all, and there is interop problems

[02:32:01.0000] <virtuelv>
gecko and webkit doesn't throw

[02:32:02.0000] <Hixie_>
hey, the validator stopped complaining about caption or whatever it was complaining about, sweet

[02:32:03.0000] <Hixie_>
virtuelv: ah

[02:32:04.0000] <virtuelv>
Hixie_: also the same for fillRect

[02:33:00.0000] <Hixie_>
maybe they said they'd follow the spec? i dunno

[02:33:01.0000] <Hixie_>
see if there's bugs filed on them

[02:33:02.0000] <Hixie_>
it's been a few weeks since i looked at the canvas stuff

[02:35:00.0000] <zcorpan_>
does opera throw?

[02:37:00.0000] <Lachy>
/me will file a bug about the differing URLs after lunch, since it makes more sense for readers to keep using the familiar URL, rather than suddenly using a new random and unrelated one

[02:39:00.0000] <jgraham>
zcorpan_: Yes

[02:39:01.0000] <jgraham>
Gecko seems to use abs(r)

[02:40:00.0000] <jgraham>
Webkit seems to clamp to 0

[02:41:00.0000] <zcorpan_>
abs seems unexpected

[02:41:01.0000] <zcorpan_>
or?

[02:41:02.0000] <jgraham>
Well anything is unexpected really

[02:42:00.0000] <zcorpan_>
negative value is unexpected, a particular behavior when using a negative value doesn't need to be unexpected :)

[02:44:00.0000] <jgraham>
I think either the gecko or webkit behaviour is fine. Throwing is evidently not (unless gecko and webkit are both prepared to change)

[02:45:00.0000] <darobin>
Philip`: thanks for the pointer!

[02:45:01.0000] <zcorpan_>
2 out of 3 will have to change regardless

[02:47:00.0000] <jgraham>
Yeah but the opera behaviour causes sites to just stop working. The gecko + webkit behaviours don't cause hard failure which, for some of the more decorative uses of canvas, is more important than the exact rendering

[02:51:00.0000] <zcorpan_>
"The behavior of the 'auto' value is user agent-dependent, but should cause a scrolling mechanism to be provided for overflowing boxes." - http://www.w3.org/TR/CSS21/visufx.html#overflow

[02:52:00.0000] <zcorpan_>
i had forgotten that the css spec sucks

[02:59:00.0000] <Philip`>
/me wonders if Opera is going to start drawing the radial gradient circles at roughly the positions the user specified instead of making up something completely weird :-)

[02:59:01.0000] <Philip`>
(See e.g. diagram in http://lists.w3.org/Archives/Public/public-canvas-api/2009OctDec/0022.html)

[02:59:02.0000] <Philip`>
((but ignore the "Correct interpretation?" because that's wrong, it should match WebKit))

[03:00:00.0000] <Philip`>
s/WebKit/Safari/

[03:04:00.0000] <jgraham>
/me also notes that it is totally non obvious that "INDEX_SIZE_ERR" would mean "value outside allowed range"

[03:04:01.0000] <jgraham>
Of course it is probably far too late to change that in general

[03:05:00.0000] <Philip`>
Seems like the alternative is making up a new exception

[03:06:00.0000] <Philip`>
which is harder work than reusing one that already has "size" in its name and is therefore not wholly inappropriate for size-related things

[03:06:01.0000] <Philip`>
and just ignore the bit where it says "index"

[03:06:02.0000] <Hixie_>
man the ietf submission system is annoying

[03:06:03.0000] <Philip`>
/me would hope browsers would print informative error messages to their error log before throwing the exception

[03:08:00.0000] <jgraham>
Philip`: Well if you already know to ignore the misleading part then it is hardly misleading

[03:08:01.0000] <jgraham>
It should really just throw RangeError() but I guess that would be too wasy

[03:08:02.0000] <jgraham>
*easy

[03:10:00.0000] <Philip`>
But that wouldn't work with such important cases as using the API from Java

[03:12:00.0000] <Hixie_>
othermaciej: yt?

[03:12:01.0000] <othermaciej>
Hixie_: aye

[03:12:02.0000] <jgraham>
Yes, like I said it would be too easy to have an API that works well with the only language avaliable in web clients

[03:12:03.0000] <Hixie_>
othermaciej: are the issues on http://dev.w3.org/html5/status/issue-status.html that say "Review Change Proposal" issues that will at some point get a deadline for others to provide change proposals?

[03:13:00.0000] <Hixie_>
othermaciej: or are they things for which it is too late to provide alternatives?

[03:13:01.0000] <othermaciej>
Hixie_: they will all at some point get a deadline - and all unlisted issues will get a deadline to submit a Change Proposal in the first place

[03:13:02.0000] <Hixie_>
ah ok

[03:14:00.0000] <Hixie_>
so -76 is the only one for which a deadline has passed?

[03:14:01.0000] <othermaciej>
Hixie_: none of them are too late to act on, though the chairs have not set any firm deadline for counter-proposals on any of them

[03:14:02.0000] <Hixie_>
any idea what the timeline will be for -76?

[03:15:00.0000] <othermaciej>
deadlines are about to pass for 7 and 63, which will cause them to be closed without prejudice if no one submits anything by tomorrow

[03:15:01.0000] <othermaciej>
for 76 a deadline has passed, and we now have two proposals on the table

[03:15:02.0000] <othermaciej>
I have not yet discussed with the other chairs what to do with the two proposals now

[03:16:00.0000] <Hixie_>
k

[03:17:00.0000] <othermaciej>
at the very least we'd probably give the Change Proposal authors some time to update their rationales in response to counter-arguments and each other's proposals

[03:59:00.0000] <Hixie_>
oh dear, i just remembered i have a noon lunch tomorrow

[03:59:01.0000] <Hixie_>
time to run to bed. nn.

[03:59:02.0000] <Lachy>
haha

[03:59:03.0000] <gsnedders>
"Oh damn. I have to go to bed at a vaguely sensible time."

[03:59:04.0000] <jgraham>
/me wonders if running to bed doesn't just make it harder to sleep

[04:00:00.0000] <gsnedders>
jgraham: Well, if it tires him out, then it might help.

[04:00:01.0000] <jgraham>
But if it just releases a bunch of endorphins?

[04:00:02.0000] <jgraham>
(sp?)

[04:01:00.0000] <annevk>
just crash land in a coma

[04:01:01.0000] <annevk>
works all the time

[04:01:02.0000] <annevk>
:p

[04:01:03.0000] <gsnedders>
Hey, just because I don't go out to Russian caberets!

[04:01:04.0000] <gsnedders>
(sp?)

[04:16:00.0000] <zcorpan_>
"number, for the credit card number and security code" - http://24ways.org/2009/have-a-field-day-with-html5-forms

[04:17:00.0000] <annevk>
http://twitter.com/lawouach/status/6301296435 what exactly is sad?

[04:18:00.0000] <zcorpan_>
that the spec wasn't perfect before the edit

[04:19:00.0000] <annevk>
or it's worse as a result?

[04:19:01.0000] <annevk>
/me hasn't been able to come up with much else either

[04:21:00.0000] <Philip`>
Or that the spec has to tell you how to write a server?

[04:23:00.0000] <annevk>
maybe, though it's just non-normative

[04:27:00.0000] <Philip`>
Or maybe that sockets and servers are being defined in the spec that's meant to be for a document markup language?

[04:30:00.0000] <Philip`>
Trying to extract understanding from Twitter is fun but, I guess, futile

[04:30:01.0000] <annevk>
zcorpan_, ouch, using type=number for that is so wrong

[04:31:00.0000] <zcorpan_>
commented

[04:31:01.0000] <annevk>
/me wonders why there was no technical review

[06:22:00.0000] <foolip>
/me prints TabAtkins reply and frames it

[06:23:00.0000] <foolip>
(not really)

[06:23:01.0000] <TabAtkins>
Haha.

[06:23:02.0000] <foolip>
anyway, I was writing a reply, but a shit one in comparison, so thanks for sparing me the effort

[06:23:03.0000] <TabAtkins>
You're welcome.  I was afraid mine was kind of shit anyway, but I guess my review made it better.

[06:24:00.0000] <murr4y>
mailing list? :p

[06:24:01.0000] <Philip`>
"Please think of the environment before printing this email"

[06:24:02.0000] <foolip>
if this were a forum the moderators would now lock the thread and make a decision

[06:25:00.0000] <Philip`>
/me feels sorry for people who work in an environment where they have to add that kind of comment to their emails

[06:25:01.0000] <foolip>
Philip`: you've seent that in real mails?

[06:25:02.0000] <Philip`>
Yes

[06:29:00.0000] <jgraham>
I think I have seen that, but I am genuinely unsure if it happened to me or if I just remember Philip` talking about it before

[06:30:00.0000] <Philip`>
http://lists.w3.org/Archives/Public/public-html-comments/2008Dec/0005.html is one

[06:31:00.0000] <Philip`>
(Most I've seen have been non-public mailing-list ones)

[06:31:01.0000] <jgraham>
WHat kind of comment syntax is */ /* anyway?

[06:32:00.0000] <Philip`>
(often preceded by giant disclaimers of copyright and liability, and followed by untrimmed top-posting quoting of the entire thread)

[06:32:01.0000] <zcorpan_>
jgraham: close multiline comment, open multiline comment

[06:32:02.0000] <Philip`>
jgraham: Bold and italic and green, converted from HTML to text

[06:48:00.0000] <MikeSmith>
TabAtkins: are you able to edit any ESW wiki pages yet?

[06:49:00.0000] <TabAtkins>
Let me check.

[06:51:00.0000] <TabAtkins>
MikeSmith: Nope.  Tried http://esw.w3.org/topic/HTML/ChangeProposals/KeepMicrodata, but it's still marked immutable.

[06:51:01.0000] <TabAtkins>
As before, I *can* edit MeetingTaxis.

[06:57:00.0000] <Philip`>
It's not just cached or something?

[07:03:00.0000] <MikeSmith>
TabAtkins: I have asked the systeam to check into

[07:03:01.0000] <MikeSmith>
the real solution for this is for me to get a Mediawiki set up for us, and for us to quit using the ESW wiki

[07:04:00.0000] <MikeSmith>
at this point, the systems team is not enthusiastic about responding to tickets for the ESW Wiki

[07:04:01.0000] <Lachy>
MikeSmith, can the W3C mirror the WHATWG wiki?

[07:04:02.0000] <MikeSmith>
TabAtkins: are you in Houston?

[07:04:03.0000] <TabAtkins>
MikeSmith: Yup.

[07:05:00.0000] <TabAtkins>
Well, Richmond/Rosenburg, but that's just outside Houston.

[07:05:01.0000] <MikeSmith>
Lachy: I'm sure it could but I would reckon it wouldn't

[07:05:02.0000] <karlushi>
annevk, is there a bug in dragonfly or Opera? For HTML resources, when the cache on server side is well done (working with Firefox, safari, etc), Dragonfly displays 200 OK and saying the browser sent Cache-Control: no-cache. (No 304 Not Modified) - Opera/9.80 (Macintosh; Intel Mac OS X; U; fr) Presto/2.2.15 Version/10.00 

[07:05:03.0000] <MikeSmith>
TabAtkins: somebody contacted me about speaking about HTML5 at an event in Dallas in May

[07:06:00.0000] <MikeSmith>
they have a small budget and I think could cover all or part of your travel expenses

[07:06:01.0000] <karlushi>
/me is downloading a more recent version of Opera to check if the problem persists

[07:07:00.0000] <TabAtkins>
MikeSmith: Hmm.  Sounds interesting.  When is it?

[07:07:01.0000] <MikeSmith>
May 2-5

[07:07:02.0000] <MikeSmith>
75-minute slot

[07:08:00.0000] <Lachy>
MikeSmith, it just sucks to have two wikis, when the WHATWG wiki is already up and running, and should be good enough for the HTMLWG, if it weren't for those who insist on having it under the w3.org domain for political reasons

[07:08:01.0000] <TabAtkins>
If I could get up in May 2nd, it would be great.  Later in the week, I'm trying to conserve vacation for my anniversary.

[07:08:02.0000] <MikeSmith>
OK

[07:08:03.0000] <TabAtkins>
And I'm already planning on 'getting sick' during the CSS ftf in March.

[07:08:04.0000] <MikeSmith>
so I

[07:08:05.0000] <annevk>
karlushi, likely dragonfly

[07:08:06.0000] <MikeSmith>
...will put you in contact with the organizer

[07:08:07.0000] <TabAtkins>
MikeSmith: Sounds good.

[07:09:00.0000] <MikeSmith>
OK, will let you know when I hear back from him

[07:10:00.0000] <karlushi>
oooh http://opera.ca/ looks like Opera. ouch

[07:11:00.0000] <Madness>
Gh, red O, bad choice.

[07:17:00.0000] <jgraham>
/me is sorely tempted to mock TabAtkins for living in a country where <20 days of vacation a year is normal

[07:17:01.0000] <TabAtkins>
jgraham: Mock away.  I get *10*.

[07:18:00.0000] <jgraham>
But decides that would be cruel

[07:18:01.0000] <annevk>
10? whoa

[07:18:02.0000] <TabAtkins>
And that's... not unusual.

[07:18:03.0000] <jgraham>
In all seriousness I don't knoe how you deal with that

[07:18:04.0000] <TabAtkins>
2 weeks is pretty standard.

[07:19:00.0000] <TabAtkins>
I do it by MANNING UP AND BEING 'MURRICAN FUCK YEAH

[07:19:01.0000] <annevk>
I barely manage with 25days :)

[07:19:02.0000] <annevk>
I'm told Brazilian gov employees get upwards of 2months

[07:20:00.0000] <TabAtkins>
I'm thinking of asking for vacation instead of money for my xmas bonus this year.

[07:20:01.0000] <TabAtkins>
because i'm going to blow a week on my anniversary, and it'll take me *until* my anniversary to earn that much.

[07:20:02.0000] <Lachy>
2 weeks is fine in countries where unused days automatically accrue in following years

[07:21:00.0000] <jgraham>
Lachy: For some definition of "fine" that means "unbearable"

[07:22:00.0000] <gsnedders>
Lachy: That still means an average of two weeks per year

[07:22:01.0000] <TabAtkins>
Lachy: Erm, what gsnedders said.

[07:22:02.0000] <Lachy>
it depends how much you're paid, and whether you can afford to compensate with some unpaid days.

[07:22:03.0000] <Lachy>
Though, I agree, it's not ideal

[07:22:04.0000] <TabAtkins>
In many cases, 'unpaid days' aren't really an option either.

[07:22:05.0000] <jgraham>
Oh well if you are allowed to swap money for vacation that is different

[07:22:06.0000] <TabAtkins>
That's just... not something you do.

[07:23:00.0000] <jgraham>
But what TabAtkins said

[07:23:01.0000] <jgraham>
Presumably this is why so many Americans don't even have a passport (supposedly)

[07:23:02.0000] <Lachy>
I think I will end up with 2 unpaid days this year, just due to the dates I had to fly to get the cheaper flights

[07:24:00.0000] <Lachy>
and cause it's a 24 hour flights to Australia, which takes up 2 whole days there and back

[07:33:00.0000] <TabAtkins>
jgraham: Yup, so few of us get enough vacation to justify traveling outside the country.  Also, our country's just so much *bigger* than any of your European countries, there's a ton you can do within the nation that would require you to leave the country.

[07:34:00.0000] <TabAtkins>
"You" as in you specifically, a european.

[07:36:00.0000] <jgraham>
(To be fair you can travel between a lot of european countries without a passport but I guess many people have still been outside Europe)

[07:36:01.0000] <jgraham>
(I don't know how the physical size compares though)

[07:36:02.0000] <JvA>
I don't remember where I read it, but about a year ago I read in a paper that Europeans and American scientists had concluded that a person need about 3-4 weeks of coherent vacation to 100% relax, and let the body recover from the that year's stress.

[07:37:00.0000] <JvA>
And that was one of the reason's most of west Europe, China etc have 4-6 weeks vacation. (if people in China always get their 4 weeks is another matter)

[07:37:01.0000] <jgraham>
In sweden a lot of people take blocks of 3-4 weeks often in the summer

[07:38:00.0000] <jgraham>
In fact it seems quite common to take four weeks just spend it in your summerhouse 

[07:38:01.0000] <jgraham>
Which I don't really understand

[07:38:02.0000] <jgraham>
s/often/generally/

[07:39:00.0000] <JvA>
jgraham: We have like 2 months of warm sun, so we "måste passa på" when the sun is up ;)

[07:39:01.0000] <Lachy>
jgraham, http://www.wolframalpha.com/input/?i=Area+of+europe+and+USA - USA is much bigger than Europe

[07:40:00.0000] <Philip`>
But lots of it is big and empty and boring

[07:40:01.0000] <JvA>
jgraham: I think many families with children do this: 3-4 weeks during the summer, 1 week to some sunny place during the winter. Ever been to Thailand during the winter? You'll find 10% of the Nordic population there.

[07:41:00.0000] <JvA>
Swedes are probably crazy.

[07:41:01.0000] <Lachy>
Philip`, Wolfram Alpha can't tell me the area of just the interesting parts of the USA.

[07:42:00.0000] <jgraham>
JvA: I met a bunch of people who aren't really "families with children" doing the same sort of thing

[07:42:01.0000] <jgraham>
Lachy: That is less than a factor of 2

[07:42:02.0000] <jgraham>
/me wonders how it defines "europe"

[07:43:00.0000] <jgraham>
"Europe with Russia" is much bigger than the USA and also has a large number of boring desert bits :)

[07:44:00.0000] <Lachy>
only part of Russia is in the continent of Europe though

[07:44:01.0000] <Lachy>
so I assume the calculation either excluded russia completely, or did it based on the actual size of the continent

[07:45:00.0000] <JvA>
jgraham: It's called "the industry vacation". Many companies still make 3-4 weeks of vacation mandatory in the summer.

[07:45:01.0000] <jgraham>
JvA: I noticed that. I would still tend to go somewhere outside Sweden myself (at least for some of the time)

[07:46:00.0000] <jgraham>
Because y'know Sweden is great, but the world is full of interesting places

[07:52:00.0000] <JvA>
jgraham: Agreed, but you speak English. Traveling "alone" (everything but charter airlines) might be quite a new concept for many Swedes?

[07:53:00.0000] <JvA>
+ the younger generation probably can't afford it anyhow. 25-35% of the people < 27 is out of work while many of the others study.

[07:53:01.0000] <JvA>
Anyhow, I should really stop talking about this in #whatwg :p

[07:54:00.0000] <jgraham>
The sample of Swedes I know is heavily biased toward the young and employed though

[08:00:00.0000] <gsnedders>
jgraham: Yeah, interesting places? Like Sweden :P

[08:07:00.0000] <TabAtkins>
Ah, there's nothing like using simulation to decide not to play a game of chance.

[08:28:00.0000] <Dashiva>
http://news.cnet.com/8301-13577_3-10407330-36.html

[08:28:01.0000] <Dashiva>
I am disappointed by the lack of "I am as high as a kite"

[09:49:00.0000] <MikeSmith>
TabAtkins: 

[09:50:00.0000] <MikeSmith>
you should be able to edit now

[09:50:01.0000] <MikeSmith>
apparently the problem as that your actual Wiki username is "Tab Atkins", not TabAtkins

[09:50:02.0000] <TabAtkins>
Indeed, now I can.

[09:50:03.0000] <TabAtkins>
Ah yeah, I keep forgetting about that.

[09:51:00.0000] <TabAtkins>
Every time I log in I try to do TabAtkins first.

[09:51:01.0000] <TabAtkins>
IIRC, the signup form wasn't clear that the name it was asking for would be my username, and then I didn't think it was worth it to sign up for another one just to get a WikiText name.

[09:59:00.0000] <MikeSmith>
TabAtkins: this seems to be a not-uncommon problem with users and MoinMoin

[10:00:00.0000] <Dashiva>
MoinMoin itself is a problem :)

[10:03:00.0000] <MikeSmith>
yeah

[14:19:00.0000] <Hixie>
wait, shelley thinks my opinion of what should be in and not in html5 is a compelling argument?

[14:20:00.0000] <Hixie>
i thought my opinion was worthless to her

[14:20:01.0000] <zcorpan_>
http://forums.whatwg.org/viewtopic.php?t=4177 - hmm, what's the attack vector?

[14:21:00.0000] <zcorpan_>
does it matter that a page can read path cookies from a 404 or 403?

[14:22:00.0000] <Hixie>
can't see why it would

[14:22:01.0000] <Hixie>
or how it could

[14:22:02.0000] <Hixie>
surely 404 pages should be allowed to use cookies?

[14:23:00.0000] <Hixie>
TabAtkins: it would be good to update your change proposal to consisely convey the points made in the recent threads -- let me know if you would like to talk about how to do that

[14:23:01.0000] <zcorpan_>
yeah, but should a parent page be able to access them if the parent page uses a different path?

[14:23:02.0000] <othermaciej>
Hixie: don't worry - I still think your opinion is worthless, if not backed by sound technical arguments

[14:24:00.0000] <Hixie>
me too

[14:24:01.0000] <Hixie>
i'm just confused that shelley seems to pick and chose when my opinion is "compelling" based purely on whether it matches her own opinion

[14:24:02.0000] <TabAtkins>
Hixie: Yeah, I'll be doing so.  I'm taking the day off of that thread, though, to let things settle.  I figure one post per day in contentious threads is a good policy.

[14:24:03.0000] <Hixie>
TabAtkins: sounds far too reasonable for public-html. ;-)

[14:24:04.0000] <Hixie>
zcorpan_: how could you stop it?

[14:25:00.0000] <TabAtkins>
You're right.  I'll commence spamming with ALL CAPS immediately!

[14:25:01.0000] <zcorpan_>
Hixie: apparently ie8 blocks reading cookie for 404 (haven't tested myself though)

[14:25:02.0000] <Hixie>
so 404 pages can't use cookies in IE?

[14:25:03.0000] <Hixie>
weird

[14:30:00.0000] <nessy>
really?

[14:30:01.0000] <nessy>
strange...

[14:36:00.0000] <othermaciej>
that seems not terribly useful, since the cookies already got sent to the server

[14:37:00.0000] <othermaciej>
oh I guess the theoretical scenario is different parts of the same origin not trusting each other?

[14:38:00.0000] <zcorpan_>
yes

[14:39:00.0000] <zcorpan_>
i'm can't imagine a reasonable attack vector right now, but otoh i'm kind of too tired to think properly right now

[15:42:00.0000] <Lachy>
othermaciej, how long before we call for a WG vote on this microdata/rdfa issue, so we can hurry up and drop RDFa and move on with more important stuff?

[15:45:00.0000] <zcorpan_>
ah, so remy implemented sessionStorage with window.name, as i suggested a few months ago that someone should do

[15:45:01.0000] <zcorpan_>
(in http://24ways.org/2009/breaking-out-the-edges-of-the-browser )

[15:45:02.0000] <othermaciej>
Lachy: we already published RDFa as an FPWD - I don't think a vote to end-of-life it and convert it to a Note would be productive at this time (though maybe it will turn out that way eventually)

[15:45:03.0000] <Lachy>
it seems to me that everyone is already pretty firm on their opinion about whether or not microdata should be split, and so there's little point in letting this drag on much longer

[15:46:00.0000] <othermaciej>
Lachy: agreed, I don't want the discussion to go forever

[15:46:01.0000] <Lachy>
othermaciej, I know that. IMHO, the way we'll deal with RDFa is just to block it from getting to CR, and since microdata covers all the use cases, it can go to Note later

[15:46:02.0000] <othermaciej>
Lachy: but I would at least like to give the authors of the Change Proposals the opportunity to update them in response to counter-arguments

[15:47:00.0000] <Lachy>
but for now, letting them play around with their own spec for a while doesn't seem to be causing much harm, other than the competition between the two specs

[15:47:01.0000] <Lachy>
fair enough

[15:48:00.0000] <Lachy>
othermaciej, it might be nice if you could set some sort of deadline for getting the change proposals finalised. Maybe a week or so.

[15:49:00.0000] <Lachy>
we should try to avoid having a vote during holiday times while people are away, so sooner rather than later would be better.

[15:49:01.0000] <othermaciej>
Lachy: I think all of the Chairs are pretty keen to get at least one issue all the way through the end of the process

[15:49:02.0000] <Hixie>
rdfa will likely get its own wg, as i understand it

[15:50:00.0000] <Hixie>
so it's very likely to go to REC unless you really want to go out of your way to stop it (which seems pointless)

[15:50:01.0000] <othermaciej>
they do plan to form their own WG at some point, yes

[15:50:02.0000] <Lachy>
don't they already have an RDF working group?

[15:50:03.0000] <Hixie>
RDFa is no more appropriate for the RDF WG than microdata is

[15:51:00.0000] <Philip`>
Lachy: No, just the XHTML2 WG

[15:51:01.0000] <Lachy>
huh? Didn't they initally create it?

[15:51:02.0000] <Hixie>
what?

[15:51:03.0000] <Lachy>
/me isn't too up to date on the history of RDF and RDFa, so never mind

[15:51:04.0000] <Hixie>
/me confoosed

[15:52:00.0000] <Hixie>
RDF has nothing to do with RDFa, politically, as i understand it

[15:52:01.0000] <Hixie>
RDFa was an XHTML2 thing

[15:52:02.0000] <othermaciej>
RDFa was created by a joint task force of the XHTML2 WG and one of the RDF-related WGs

[15:52:03.0000] <othermaciej>
It is intended that in the future there will be a WG for RDFa specifically which could continue to work on the XHTML and HTML syntaxes for it down the line

[15:53:00.0000] <Dashiva>
What is the future work supposed to be? Once you have a proper serialization of RDF, isn't that the end of it?

[15:53:01.0000] <Philip`>
It was intended that work on RDFa will continue after the XHTML2 WG is shut down, regardless of whether it's in a W3C WG or not (though preferably it would be) (as far as I understand)

[15:54:00.0000] <Philip`>
Dashiva: Errata on old versions, development of new versions with improvements based on past experience, development of test suites, promotion of adoption, etc

[15:54:01.0000] <Philip`>
I guess

[15:55:00.0000] <Philip`>
(Specific things that have been discussed include RDFa 1.1 and a DOM interface)

[15:55:01.0000] <othermaciej>
I think the bottom line is that blocking HTML+RDFa as a spec or as a technology is not likely to be a fruitful avenue to pursue

[15:55:02.0000] <Hixie>
blocking any spec or technology is pretty pointless and petty

[15:56:00.0000] <Hixie>
as i learnt from trying to block xforms and then having the w3c try to block html5 :-)

[15:56:01.0000] <Lachy>
Hixie, the problem with RDFa in HTML ever getting to REC is that it would indicate that implementers have actually bothered to implement it, meaning clueless people could start using it

[15:56:02.0000] <Hixie>
no it wouldn't

[15:57:00.0000] <Lachy>
Why? it can't get to REC without having been implemented.

[15:57:01.0000] <Hixie>
RDFa already has plenty enough implementations to convince the W3C to let it go to REC

[15:59:00.0000] <Lachy>
but there's no implementations of the HTML specific processing stuff yet, like dealing with namespaces in html

[15:59:01.0000] <Hixie>
your faith in the w3c following their process is admirable

[15:59:02.0000] <Philip`>
Almost all the implementations handle non-well-formed-XML content


2009-12-04
[16:00:00.0000] <Philip`>
and if you don't poke at the edges too hard then they all interoperate

[16:00:01.0000] <Philip`>
(Best not to poke in the middle either, or it might fall apart)

[16:00:02.0000] <Hixie>
Lachy: http://www.w3.org/TR/SVGTiny12/ is a REC. That's all I have to say, really.

[16:01:00.0000] <Lachy>
haha

[16:01:01.0000] <zcorpan_>
but it has a huuge testsuite! like 800 tests!

[16:02:00.0000] <Lachy>
woah!

[16:04:00.0000] <nessy>
http://www.bitflash.com/prod_playerSVGT.html <- google tells me that is a svgtiny 1.2 implementation

[16:05:00.0000] <nessy>
so is this http://www.tinyline.com/products.html

[16:05:01.0000] <zcorpan_>
it seems two experimental builds of opera were used for the implementation report

[16:05:02.0000] <nessy>
so they seem to have a reason to have svgtiny1.2 at REC, no?

[16:06:00.0000] <Lachy>
In other news, my US work visa application is finally seeing some progress. It has now been received by the USCIS and will soon begin going through the review process.

[16:06:01.0000] <Lachy>
I should hear more in 15 days

[16:06:02.0000] <nessy>
congrats!

[16:10:00.0000] <Hixie>
nessy: about the same level of reason to have html5 at a REC :-)

[16:10:01.0000] <Hixie>
with a test suite of 8000 tests you really can't learn much about interoperability of a spec as large as svg

[16:11:00.0000] <Hixie>
i'd estimate you need 20,000 or more, and that's counting things like checking hundreds of different colour values for 'color' as one test

[16:11:01.0000] <nessy>
if all the products pass these tests, shouldn't they be highly interoperable?

[16:11:02.0000] <nessy>
ah I see

[16:12:00.0000] <nessy>
is there a test suite for html5?

[16:12:01.0000] <Hixie>
not that i know of

[16:12:02.0000] <nessy>
that would be … millions of tests?

[16:12:03.0000] <Hixie>
hopefully at least tens of thousands, yes

[16:12:04.0000] <nessy>
it'd be really difficult I would imagine

[16:12:05.0000] <Hixie>
hence my i estimate it'll take a decade

[16:12:06.0000] <nessy>
better to have a validator!

[16:13:00.0000] <nessy>
yeah, I've always agreed with that estimate - at least for getting it into a "finished" stated

[16:13:01.0000] <nessy>
which doesn't mean it's not usable beforehand

[16:13:02.0000] <nessy>
plenty of products are in beta for a long time :)

[16:15:00.0000] <Hixie>
indeed

[16:17:00.0000] <zcorpan_>
woot, html5 says "behavior is undefined"

[16:17:01.0000] <Hixie>
html5 doesn't! websocket protocol does!

[16:17:02.0000] <zcorpan_>
same thing

[16:17:03.0000] <Hixie>
hah

[16:17:04.0000] <Hixie>
it actually says that the next layer protocol spec is responsible for it

[16:28:00.0000] <zcorpan_>
"I favor microdata, but I would much prefer that microdata be dropped and (some version of) RDFa be in the main spec than that both exist indefinitely in parallel." - same reasoning about html5 and xhtml2 (a few years ago)?

[16:28:01.0000] <Hixie>
microdata _is_ "some version of RDFa"

[16:29:00.0000] <Hixie>
i just took rdfa and tweaked it until it didn't have the problems that made it inappropriate for the use cases that people asked for

[16:29:01.0000] <Hixie>
(that's why it shared the same attribute names initially)

[16:29:02.0000] <zcorpan_>
just like with wf2?

[16:30:00.0000] <Hixie>
yeah, very similar to how WF2 was an attempt to make XForms work in HTML

[16:31:00.0000] <Lachy>
Has there been much interest from the microformats community to start looking into defining formats utilising microdata?

[16:33:00.0000] <TabAtkins>
Lachy: tantek worked on the Microdata vocabs, which are straight from microformats.

[16:35:00.0000] <Lachy>
yeah, I meant other than those in the vocabs spec, which are just based on older microformats

[16:35:01.0000] <TabAtkins>
You don't really have to do much of anything to use a microformat with the microdata syntax.  They already have the tree-structure defined.

[16:35:02.0000] <Lachy>
that's good.

[16:36:00.0000] <Hixie>
you have to do _some_ work to define it properly

[16:36:01.0000] <Hixie>
but then that's work you'd have to do to define them properly for class=""/rel="" too

[16:38:00.0000] <Lachy>
Is the plan to one day encourage people to phase out the older class-based hCard, in favour of the microdata version?

[16:38:01.0000] <Lachy>
in fact, is there some problem being solved by mapping hCard into microdata?

[16:39:00.0000] <zcorpan_>
"It seems you can access contentWindow, but trying to access contentWindow.document (for IE's "special" 404 pages) raises an "access denied" exception, so it's not exactly a restriction on document.cookie, but on document itself..."

[16:40:00.0000] <TabAtkins>
Lachy: Yeah.  The class/rel syntax used by microformats atm requires custom parsing rules for every microformat, and still has a few notable failures (like the inability to nest vcards, frex).  Microdata gives you a single consistent parsing model to use, and solves nesting problems.

[16:41:00.0000] <Hixie>
i expect the class/rel-based microformats will continue to exist -- they do work, and are deployed

[16:41:01.0000] <hober>
Lachy: there are some notes here: http://microformats.org/wiki/html5

[16:41:02.0000] <Lachy>
Hixie, yeah, I don't expect support will ever be dropped entirely.

[16:45:00.0000] <Lachy>
hober, the wiki still mentions the <itemref> element, which was dropped.

[16:46:00.0000] <webben>
Lachy: "yeah, I meant other than those in the vocabs spec, which are just based on older microformats" ... it's worth noting that new microformats are actually pretty rare.

[16:46:01.0000] <webben>
so I wouldn't necessarily expect to see lots of new formats expressed in microdata in a short space of time.

[16:47:00.0000] <Lachy>
webben, yeah, I'm aware of how frequently they create new formats

[16:48:00.0000] <hober>
Lachy: wikis are for updating :)

[16:48:01.0000] <Lachy>
hober, I would do it, but that would require me to remember what my microformats wiki username and password was

[16:56:00.0000] <erlehmann>
did someone say microdata ? can it be appropriate to add RDFa stuff to my cc-figure-generator so both is contained in the generated HTML fragments ?

[16:59:00.0000] <Lachy>
erlehmann, what is your cc-figure-generator?

[16:59:01.0000] <Lachy>
also, I think wasting time on adding RDFa to anything is not useful, especially if it already uses microdata

[17:00:00.0000] <erlehmann>
Lachy, a wordpress plugin for embedding cc-licensing microdata http://github.com/erlehmann/cc-figure/

[17:00:01.0000] <erlehmann>
but i plan to make a stand-alone generator page

[17:01:00.0000] <erlehmann>
that just displays the code. actually, its simple to make this change.

[17:01:01.0000] <Lachy>
do you have a working demo so I can see the result?

[17:09:00.0000] <erlehmann>
Lachy, give me 15 minutes and i'll make a new commit containing a self-contained xhtml demo page.

[17:10:00.0000] <zcorpan_>
having both rdfa and microdata for the same data seems likely to confuse consumers that support both

[17:10:01.0000] <zcorpan_>
if there are such consumers

[17:11:00.0000] <erlehmann>
zcorpan_, confuse ? could it ?

[17:11:01.0000] <zcorpan_>
well, it would probably generate two sets of the same data

[17:15:00.0000] <zcorpan_>
also, the markup would be ridiculously verbose if it used both microdata and rdfa :)

[17:15:01.0000] <zcorpan_>
(just one of them is too verbose for my taste)

[17:18:00.0000] <mikekelly>
is there going to be a way of using HTTP Authorization that doesn't suck?

[17:19:00.0000] <Hixie>
you'd have to ask the http working group

[17:21:00.0000] <mikekelly>
I'm sorry can you rephrase that, I don't speak emo

[17:22:00.0000] <Hixie>
/me gives mikekelly a hug

[17:22:01.0000] <mikekelly>
:D

[17:22:02.0000] <mikekelly>
seriously though I have no idea what you're getting at

[17:24:00.0000] <mikekelly>
oh wait hold on

[17:24:01.0000] <mikekelly>
no actually please can you explain what it is you think needs to be done to make that work?

[17:24:02.0000] <mikekelly>
the only real problem is that there's no way to persist the Authorization header

[17:25:00.0000] <mikekelly>
so if you push an XHR request through with the Authorization header set

[17:25:01.0000] <mikekelly>
it'll work once and then get dropped

[17:25:02.0000] <mikekelly>
+ you can't wipe the header either so you can't logout

[17:25:03.0000] <Hixie>
this channel is mostly where we discuss HTML, not HTTP, I'm afraid

[17:25:04.0000] <Hixie>
well that and such things as unicorns

[17:25:05.0000] <Hixie>
and narwhals

[17:25:06.0000] <mikekelly>
I like them

[17:25:07.0000] <Hixie>
and the lack of logic in the land

[17:26:00.0000] <mikekelly>
Ian you fucking ROCK

[17:26:01.0000] <Hixie>
why thank you

[17:26:02.0000] <Hixie>
wait, are you saying i DO rock, or that i AM a rock

[17:26:03.0000] <Hixie>
because the latter seems less awesome

[17:27:00.0000] <mikekelly>
it was actually a question I  don't punctuate, well

[17:27:01.0000] <mikekelly>
see what I did there?

[17:27:02.0000] <Hixie>
are you asking if i perform reproductive activities with minerals??

[17:28:00.0000] <mikekelly>
possibly

[17:28:01.0000] <Hixie>
that seems less appropriate even than saying i am a rock!

[17:28:02.0000] <Hixie>
though i will give you points for lack of logic

[17:28:03.0000] <mikekelly>
you're the expert.

[17:28:04.0000] <Hixie>
in lack of logic? sadly yes. :-(

[17:29:00.0000] <mikekelly>
cheer up

[17:29:01.0000] <Hixie>
so what brings you to our fair channel at such random intervals?

[17:30:00.0000] <Hixie>
do you grace us with your presence as a way to alleviate your boredom?

[17:30:01.0000] <mikekelly>
preeetty much

[17:31:00.0000] <erlehmann>
Lachy, http://github.com/erlehmann/cc-figure/blob/master/cc-figure-generator-standalone.xhtml

[17:32:00.0000] <erlehmann>
since the raw version is served with the wrong mime type, just download it http://github.com/erlehmann/cc-figure/raw/master/cc-figure-generator-standalone.xhtml

[17:32:01.0000] <mikekelly>
Hixie: so.. in your humble opinion..

[17:32:02.0000] <mikekelly>
what do you think will happen wrt HTTP Auth

[17:33:00.0000] <Hixie>
i don't think about such things

[17:34:00.0000] <mikekelly>
whats the word on the street

[17:34:01.0000] <mikekelly>
I want the down lo Ian

[17:35:00.0000] <GarethAdams|Home>
the word on the street is "SLOW"

[17:35:01.0000] <Hixie>
i would have thought you would be more knowledgable about the word on the street... after all, i'm stuck up here in this ivory tower

[17:35:02.0000] <GarethAdams|Home>
at least on my street

[17:35:03.0000] <Hixie>
GarethAdams|Home: not "PED XING"?

[17:35:04.0000] <GarethAdams|Home>
not in the UK

[17:35:05.0000] <mikekelly>
yeah but I get it man

[17:35:06.0000] <mikekelly>
it's a nice tower

[17:35:07.0000] <mikekelly>
good for you

[17:35:08.0000] <Hixie>
GarethAdams|Home: ah, yeah, the UK has FAR saner typographical conventions for roads

[17:36:00.0000] <GarethAdams|Home>
http://www.saferroadsderbyshire.org.uk/images/roadSlow_tcm28-110980.jpg

[17:36:01.0000] <Hixie>
GarethAdams|Home: you should see road signs in the US, it's a dog-gone _disaster_

[17:36:02.0000] <erlehmann>
"WHATWG", "typographical conventions for roads" — DOES NOT COMPUTE

[17:36:03.0000] <GarethAdams|Home>
(that;s not my road, just an example)

[17:36:04.0000] <Hixie>
GarethAdams|Home: they have individuals road signs with more than 4 fonts on them for less than one sentence!

[17:36:05.0000] <GarethAdams|Home>
lol

[17:36:06.0000] <foolip>
erlehmann: is assuming that something without an extension is for download great?

[17:36:07.0000] <mikekelly>
you're wasting my time here. do you mind?

[17:36:08.0000] <foolip>
erlehmann: also, showing the markup live while typing would be nice :)

[17:36:09.0000] <Hixie>
GarethAdams|Home: and they misspell words like "through" on purpose, even when not necessary

[17:37:00.0000] <Hixie>
GarethAdams|Home: it's quite the zoo

[17:37:01.0000] <GarethAdams|Home>
thru?

[17:37:02.0000] <Hixie>
"thru" "xing" "ped"

[17:37:03.0000] <GarethAdams|Home>
lol

[17:37:04.0000] <GarethAdams|Home>
"sidewalk"

[17:37:05.0000] <Hixie>
i like how the "slow" on your road is too big to fit the lane it's in

[17:37:06.0000] <GarethAdams|Home>
maybe that one's actually sensible, actually

[17:37:07.0000] <erlehmann>
foolip, if you can tell me a quick way to get th MIME type from the HTTP headers in javascript, go one. commit code or something.

[17:37:08.0000] <mikekelly>
Hixie

[17:37:09.0000] <mikekelly>
we need you to be on the case for this Auth thing

[17:38:00.0000] <mikekelly>
the fate of the world is in your hands

[17:38:01.0000] <Hixie>
you mean i get to decide what to do with http auth?

[17:38:02.0000] <Hixie>
sweet

[17:38:03.0000] <Hixie>
can we dump it altogether?

[17:38:04.0000] <foolip>
erlehmann: I cannot

[17:38:05.0000] <erlehmann>
i smell a new CSSquirrel comic

[17:39:00.0000] <erlehmann>
foolip, that is actually the one thing (besides more css files) i am missing here

[17:39:01.0000] <foolip>
erlehmann: I would just assume it's an img, or possible have a radio button for choosing?

[17:39:02.0000] <mikekelly>
Hixie: why would you suggest that was a good idea?

[17:40:00.0000] <mikekelly>
actually you know what don't even bother

[17:40:01.0000] <foolip>
erlehmann: but <img> certainly seems a sane default, no?

[17:40:02.0000] <erlehmann>
foolip, no un-necessary switches and bolts.

[17:40:03.0000] <erlehmann>
foolip, what if i want to present a torrent ?

[17:40:04.0000] <foolip>
erlehmann: as a figure?

[17:40:05.0000] <erlehmann>
got me there :p

[17:41:00.0000] <erlehmann>
yes, img may be a sane default. as you can see i am not even handling SVG in <img>

[17:41:01.0000] <mikekelly>
have a super evening chaps

[17:41:02.0000] <erlehmann>
if you fix it, i'll commit it. but now i'm gonna make a new, simple stylesheet

[17:41:03.0000] <erlehmann>
i like css

[17:43:00.0000] <Hixie>
mikekelly: oh noes, you're not leaving us are you?

[17:50:00.0000] <foolip>
erlehmann: http://pastebin.com/d20904a86

[17:51:00.0000] <foolip>
don't have time to figure out how to do a git merge request, I'm heading to the airport

[17:51:01.0000] <erlehmann>
foolip, this is MADNESS

[17:51:02.0000] <erlehmann>
i'll sleep a night about thit

[17:52:00.0000] <erlehmann>
but thanks

[17:52:01.0000] <erlehmann>
i can attribute it to you nonetheless

[17:52:02.0000] <foolip>
erlehmann: ok, I don't mind much, just a suggestion

[17:52:03.0000] <erlehmann>
i'd prefer mime type detection via http request

[17:53:00.0000] <erlehmann>
that would also protect against typos

[17:53:01.0000] <foolip>
you'd need the php for that

[17:53:02.0000] <erlehmann>
because 4xx codes etc. would be DETECTED

[17:53:03.0000] <erlehmann>
:/

[17:53:04.0000] <erlehmann>
can't the javascript just pull the headers ? ;_;

[17:53:05.0000] <foolip>
javascript can't do cross-origin and even if it were I don't know how to get the mimetype

[17:54:00.0000] <foolip>
so server-side is the only way right now

[17:54:01.0000] <foolip>
but I shall be off now

[17:54:02.0000] <foolip>
good luck :)

[17:56:00.0000] <erlehmann>
foolip, a php origin-redirecting-script would be a lot hackish for a small corner case

[17:57:00.0000] <erlehmann>
good traveling

[17:57:01.0000] <erlehmann>
or what you say there

[18:01:00.0000] <Hixie>
othermaciej: http://www.w3.org/mid/4d2fac900912031743q23ced9d7nd2e44b63a9706191⊙mgc

[18:01:01.0000] <othermaciej>
Hixie: that URL's not working for me

[18:01:02.0000] <Hixie>
latest e-mail from mark miller to the new security list

[18:02:00.0000] <Hixie>
/mid/ usually has some lag, it'll work in a bit

[18:03:00.0000] <othermaciej>
I think I just replied to it, if it's what I think it is

[18:03:01.0000] <Hixie>
k

[18:04:00.0000] <erlehmann>
Hixie, this raises the issue of a 4xx code for content that will be there in the future ;(

[18:04:01.0000] <erlehmann>
i meant ;)

[18:04:02.0000] <othermaciej>
Hixie: Mark Miller sure hates origin-based security, eh?

[18:04:03.0000] <othermaciej>
/me wonders if Mark will also object to postMessage()

[18:06:00.0000] <Hixie>
mark probably doesn't hate origin-based security any more than i do, he's just less willing to keep extending it

[18:06:01.0000] <Hixie>
personally i don't see an alternative, i don't think we can retrofit caps onto the web

[18:08:00.0000] <othermaciej>
I don't think it's a very good design, but I think it's more sensible to do the best we can with it than to try to fight it

[18:09:00.0000] <othermaciej>
since we can't entirely remove the concept of origin without badly breaking the web

[18:10:00.0000] <Hixie>
indeed

[18:10:01.0000] <Hixie>
i can certainly identify with his views; there's something to be said for not encouraging people to fight CSRF the way that Origin: suggests we fight it

[18:11:00.0000] <othermaciej>
Hixie: if you meant this message then I replied: <http://lists.w3.org/Archives/Public/public-web-security/2009Dec/0040.html>

[18:11:01.0000] <Hixie>
but imho the alternatives are just too complicated given the way the web works

[18:11:02.0000] <othermaciej>
well, a secret token defense is potentially more sound, but you have to pretty much build it yourself, and you have to ensure that your secret tokens are actually unpredictable, not replayable, etc

[18:11:03.0000] <Hixie>
and so the decision facing authors is not just origin-based csrf protection vs caps-based security, but between no security, weak security, and complicated security

[18:12:00.0000] <Hixie>
and the latter is too expensive in practice for many authors

[18:12:01.0000] <othermaciej>
deploying both an Origin defense and a secret-token defense is better than either alone

[18:12:02.0000] <Hixie>
indeed

[18:12:03.0000] <othermaciej>
and Origin is extremely cheap to add when you discover you forgot to protect a resource against CSRF, since it can be tacked on by an application firewall

[18:12:04.0000] <Hixie>
but i expect he fears that people will pick "weak" over "complicated" if offered both

[18:12:05.0000] <Hixie>
and that people will pick "complicated" over "none" if offered those choices only

[18:13:00.0000] <Hixie>
i fear they'll pick "none", personally

[18:13:01.0000] <Hixie>
which is why i'm ok with "weak"

[18:13:02.0000] <othermaciej>
well, at present "complicated" vs "none" is exactly the choice you get

[18:13:03.0000] <Hixie>
indeed

[18:13:04.0000] <othermaciej>
and we see some people in each bucket

[18:13:05.0000] <othermaciej>
I'm not even sure of the scenarios where Origin fails as a CSRF defense

[18:14:00.0000] <othermaciej>
I'm pretty sure it works for the basic scenario of a Cookie-authenticated form, even if backed by nothing else

[18:14:01.0000] <othermaciej>
secret token CSRF defense has its own risks, precisely because of being complicated (in that it's easy to get it wrong)

[18:15:00.0000] <Hixie>
origin's failures aren't as a csrf defense but as a generic defense -- i.e. the danger is that it'll be misused to do other things

[18:15:01.0000] <othermaciej>
so I don't even entirely agree with the framing, let alone the conclusion

[18:16:00.0000] <othermaciej>
I'm not sure what else you'd use it as a defense for (other than opening a hole to allow limited cross-site XHR, but that use is totally orthogonal to the use in the HTML5 spec)

[18:17:00.0000] <othermaciej>
I know his worry about Origin-for-CORS is about the possibility of mistakenly using Origin as the sole defense in cases where it is not sound to use it that way

[18:17:01.0000] <Hixie>
right

[18:17:02.0000] <othermaciej>
not sure about Origin-as-CSRF-defense

[18:17:03.0000] <Hixie>
well it's the same feature, effectively

[18:17:04.0000] <Hixie>
so presumably if it can be misused, it can be misused, regardless of which spec mentions it

[18:18:00.0000] <Hixie>
btw, the htmlwg process document says that "without prejudice" means the issue can't be reopened, not that it can't block LC, as far as I can tell

[18:18:01.0000] <othermaciej>
if CORS did not exist and you had only AnonXHR, then even if HTML5 still used it for CSRF defense there would be no way to abuse it without the same way

[18:18:02.0000] <Hixie>
(though the latter is indeed implied by the former, obviously)

[18:18:03.0000] <Hixie>
othermaciej: true

[18:18:04.0000] <Hixie>
i guess

[18:19:00.0000] <othermaciej>
Hixie: "An issue that is closed without prejudice in this way can only be re-raised with approval of the Chairs."

[18:19:01.0000] <Hixie>
right

[18:19:02.0000] <othermaciej>
so it's not that it can't be reopened, but I did forget to mention that approval of the chairs is needed

[18:25:00.0000] <othermaciej>
I corrected that technicality

[19:29:00.0000] <Heimidal>
is anyone aware of a CDN/datastore that supports the Access-Control-Allow-Origin header? S3 (and therefore CF) doesn't seem to...

[19:54:00.0000] <doublec>
Heimidal, I'm not aware of any

[20:38:00.0000] <erlehmann>
anyone here has tips on figure styling ?

[21:08:00.0000] <TabAtkins>
erlehmann: In what way?

[21:09:00.0000] <erlehmann>
TabAtkins, just hints. i'm somewhat restless, eager to produce CSS while the world sleeps till it is time to attend university.

[21:10:00.0000] <erlehmann>
and media wiki figure styling is somewhat lacking

[21:11:00.0000] <TabAtkins>
I was gonna offer tips, but then I realized I only have tips on <details> styling.

[21:12:00.0000] <TabAtkins>
Anyway, I recommend figure{display:table} figure>[caption]{display:table-caption;caption-side:bottom;} figure>[caption]:first-child{caption-side:top;}

[21:12:01.0000] <TabAtkins>
And then a border and a light background.

[21:16:00.0000] <erlehmann>
pretty generic idea ;)

[21:17:00.0000] <TabAtkins>
Indeed!

[21:17:01.0000] <TabAtkins>
My <details> ideas are better.

[21:19:00.0000] <erlehmann>
tellme

[21:20:00.0000] <erlehmann>
and tell me they aren't :focus tricks ;)

[21:20:01.0000] <TabAtkins>
http://www.xanthir.com/etc/details

[21:20:02.0000] <TabAtkins>
No, they're not.

[21:20:03.0000] <TabAtkins>
They're css polygon tricks

[21:23:00.0000] <erlehmann>
not remotely stimulating

[21:24:00.0000] <TabAtkins>
Bah!

[21:26:00.0000] <TabAtkins>
Okay, fine, I just came up with this.  I find it decent:

[21:26:01.0000] <TabAtkins>
figure{

[21:26:02.0000] <TabAtkins>
 float: left;

[21:26:03.0000] <TabAtkins>
 border: 2px solid black;

[21:26:04.0000] <TabAtkins>
 -moz-border-radius: 8px;

[21:26:05.0000] <TabAtkins>
 -webkit-border-radius: 8px;

[21:26:06.0000] <TabAtkins>
 background: #c00;

[21:26:07.0000] <TabAtkins>
 margin: 1em;

[21:26:08.0000] <TabAtkins>
 padding: 2em 1em 1em;

[21:26:09.0000] <TabAtkins>
 position: relative;

[21:26:10.0000] <TabAtkins>
} 

[21:27:00.0000] <TabAtkins>
figure > [caption] {

[21:27:01.0000] <TabAtkins>
 position: absolute;

[21:27:02.0000] <TabAtkins>
 width: 80%;

[21:27:03.0000] <TabAtkins>
 top: -20px;

[21:27:04.0000] <TabAtkins>
 left: -10px;

[21:27:05.0000] <TabAtkins>
 font-size: smaller;

[21:27:06.0000] <TabAtkins>
 border: 3px outset gray;

[21:27:07.0000] <TabAtkins>
 background: white;

[21:27:08.0000] <TabAtkins>
 color: black;

[21:27:09.0000] <TabAtkins>
 padding: 2px;

[21:27:10.0000] <TabAtkins>
}

[21:27:11.0000] <TabAtkins>
Use <img src="http://www.xanthir.com/etc/FAT_PONY.gif> as the figure contents, obviously.

[21:27:12.0000] <GPHemsley>
/me points at PasteBin and glares in TabAtkins' direction.

[21:28:00.0000] <TabAtkins>
Bah again.

[01:59:00.0000] <gsnedders>
Hixie: hixie.ch is down

[01:59:01.0000] <Hixie>
that's odd, i'm ssh'ed into it and it's working fine

[01:59:02.0000] <Hixie>
hm, so it is

[02:00:00.0000] <Hixie>
btw on a totally unrelated note will someone please explain to me why apache needs 20MB per thread to do NOTHING?

[02:02:00.0000] <zcorpan_>
20MB is always nice to have, just in case

[02:02:01.0000] <Philip`>
Does "nothing" include e.g. running a whole Perl interpreter inside each thread via mod_perl?

[02:03:00.0000] <Hixie>
no

[02:03:01.0000] <jgraham>
Is it actually 20Mb per thread or is it one of those things where it is really 20Mb that is mostly shared?

[02:04:00.0000] <Philip`>
/me never quite understands where all the memory goes

[02:04:01.0000] <gsnedders>
/dev/null?

[02:04:02.0000] <Hixie>
i never know how to work it out

[02:04:03.0000] <Hixie>
so dunno

[02:04:04.0000] <Hixie>
(^jgraham)

[02:04:05.0000] <jgraham>
Hixie: Me neither

[02:05:00.0000] <zcorpan_>
Philip`: it's natural to forget things

[02:05:01.0000] <Hixie>
/me wonders wtf is going on on his server

[02:05:02.0000] <Hixie>
there's nothing apparently doing anything unusual, but my memory usage is through the roof

[02:06:00.0000] <Hixie>
based on the memory usage and the number of apache servers, i'm thinking that this is 20MB per-thread

[02:06:01.0000] <zcorpan_>
http://microdata.freebaseapps.com/ looks nice

[02:06:02.0000] <Philip`>
(Memory usage per-thread doesn't make much sense, since all threads in a process share a single address space)

[02:07:00.0000] <Hixie>
per forked process, i mean

[02:07:01.0000] <Hixie>
zcorpan_: they were involved in the usability study, iirc

[02:08:00.0000] <Philip`>
I get values like VIRT=35532 RES=15m SHR=3212 SWAP=18m CODE=332 DATA=13m for my Apaches

[02:08:01.0000] <Philip`>
Don't know what that translates to in real memory usage, though

[02:08:02.0000] <Hixie>
RES=20m is more normal for me

[02:09:00.0000] <Hixie>
i wonder how badly the wiki and blog and so on would be affected if i turned off mod_php

[02:09:01.0000] <Philip`>
You could always reduce the number of Apache processes

[02:09:02.0000] <zcorpan_>
Hixie: btw i think the forums don't cache images

[02:09:03.0000] <gsnedders>
Hixie: Totally broken?

[02:10:00.0000] <Hixie>
Philip`: unfortunately, that's out of my control

[02:10:01.0000] <Philip`>
(unless there's such heavy load that it really needs more than a handful)

[02:10:02.0000] <Philip`>
Oh, oka

[02:10:03.0000] <Philip`>
y

[02:11:00.0000] <Hixie>
i tried turning off mod_php, so i expect things to break

[02:11:01.0000] <Hixie>
i'll turn it back on after i see what it does

[02:11:02.0000] <Philip`>
/me likes Cacti since it draws pretty graphs and makes it obvious when something unusual is happening like heavy memory usage or load average

[02:11:03.0000] <Hixie>
i figure things can't break worse than they are right now (not doing anything at all)

[02:11:04.0000] <Hixie>
Philip`: dreamhost has pretty graphs too, but i can't work out why the memory graph is through the roof, so it doesn't help that much :-)

[02:12:00.0000] <Hixie>
/me increases his memory allocation to 4GB to see if that helps

[02:13:00.0000] <gsnedders>
Hixie: Because PHP is evil?

[02:13:01.0000] <Hixie>
no reason to think it's php

[02:13:02.0000] <Philip`>
It helps because you can look at the pretty graphs instead of worrying about the problem

[02:13:03.0000] <Philip`>
/me wonders why anyone would ever need 4GB on a server, unless they've got zillions of scripts and databases

[02:14:00.0000] <Hixie>
to debug memory overallocation bugs :-)

[02:14:01.0000] <Philip`>
Oh, right, good point

[02:16:00.0000] <Hixie>
oh, hm, i can't update the spec if the server's down, heh

[02:17:00.0000] <Hixie>
ok that didn't help, let's see what happens if i disable apache altogether

[02:17:01.0000] <Hixie>
can't make it worse than now...

[02:22:00.0000] <Hixie>
ok that closed apache but then it didn't come back up

[02:22:01.0000] <Hixie>
time to reboot the server

[02:26:00.0000] <mikekelly>
I'm back now

[02:26:01.0000] <mikekelly>
hi Ian

[02:27:00.0000] <mikekelly>
I meant to ask you

[02:27:01.0000] <mikekelly>
are you planning on bringing out Hixie action figures?

[02:28:00.0000] <mikekelly>
that would be pretty sweet.

[02:31:00.0000] <timz>
finally some progress :-) http://www.spiegel.de/netzwelt/web/0,1518,664475,00.html

[03:42:00.0000] <zcorpan_>
does resolving a url drop the fragment? that seems wrong, no?

[03:43:00.0000] <gsnedders>
Resolving a URL according to what?

[03:44:00.0000] <zcorpan_>
> It seems "Resolving Web addresses" 

[03:44:01.0000] <zcorpan_>
> will drop fragment component according to 

[03:44:02.0000] <zcorpan_>
> http://www.w3.org/html/wg/href/draft.html

[03:46:00.0000] <gsnedders>
How? What drops it? As far as I can see it shouldn't.

[03:46:01.0000] <zcorpan_>
ok

[03:47:00.0000] <Hixie>
why would it drop the fragment?

[03:47:01.0000] <Hixie>
wouldn't that mean we could never use fragments...?

[03:49:00.0000] <Philip`>
It does drop the base URL's fragment

[03:49:01.0000] <Philip`>
but that's not too exciting

[03:50:00.0000] <gsnedders>
Well, if you resolve "" to the base URL it shouldn't

[03:51:00.0000] <Philip`>
Why shouldn't it?

[03:52:00.0000] <Philip`>
RFC 3986 always uses the non-base URL's fragment

[03:53:00.0000] <gsnedders>
Oh, my memory is just wrong.

[03:53:01.0000] <zcorpan_>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/328 - firefox drops the fragment, opera doesn't, and chrome ignores the base altogether...?

[03:55:00.0000] <zcorpan_>
/me sees Hixie's server works again

[03:56:00.0000] <Hixie>
oh the fragment in the base should be dropped, sure

[03:56:01.0000] <Hixie>
i thought you meant in the url being resolved

[03:56:02.0000] <Hixie>
the resolving algorithm is defined in the uri drafts, iirc

[03:58:00.0000] <zcorpan_>
i did, since that's what 鵜飼文敏 was talking about in the email about Web Sockets URL

[03:59:00.0000] <Hixie>
your test is not testing whether the url being resolved should lose its fragment identifier

[04:00:00.0000] <Hixie>
but maybe you're just talking about several things and i'm not keeping up :-)

[04:00:01.0000] <zcorpan_>
indeed :)

[04:03:00.0000] <zcorpan_>
Hixie: just don't make fragment in Worker constructor throw; it's useful for having different code paths

[04:03:01.0000] <Hixie>
how do you mean?

[04:04:00.0000] <zcorpan_>
page a does new Worker('worker.js#a') and page b does new Worker('worker.js#b'), and worker.js uses switch(location.search) { ... }

[04:04:01.0000] <zcorpan_>
er

[04:04:02.0000] <zcorpan_>
location.hash

[04:11:00.0000] <Hixie>
oh, workers

[04:12:00.0000] <Hixie>
nothing's changed for workes

[04:12:01.0000] <Hixie>
workers

[04:13:00.0000] <zcorpan_>
good

[04:16:00.0000] <zcorpan_>
http://www.w3.org/mid/Pine.LNX.4.62.0912041218080.5629⊙hdc - any hints of which specs those might be? :)

[04:17:00.0000] <Hixie>
depends what people want

[04:17:01.0000] <zcorpan_>
rdfa?

[04:17:02.0000] <annevk>
range

[04:17:03.0000] <annevk>
dom *

[04:17:04.0000] <annevk>
basically :)

[04:18:00.0000] <zcorpan_>
you could take over dom core

[04:18:01.0000] <annevk>
where * is range, events, core

[04:18:02.0000] <annevk>
a rewrite of CSS would be even more awesome, but so much work :/

[04:19:00.0000] <Hixie>
things on my list include stream api, filesystem, execcommand, push notifications, push storage updates, range, dom core, html6, some css3 things, traversal, user interaction events

[04:19:01.0000] <Hixie>
not dom events, shepazu is doing that

[04:19:02.0000] <Hixie>
not cssom, anne is doing that

[04:20:00.0000] <zcorpan_>
oh yes, would be great with a better spec for execcommand

[04:20:01.0000] <annevk>
tbh a stream api would be sweet too

[04:21:00.0000] <Hixie>
annevk: not css2.2, css2.1 is one of the best specs we have right now and so at the bottom of the list of specs that need rewriting :-P

[04:21:01.0000] <Hixie>
oh websql also should be on that list

[04:22:00.0000] <zcorpan_>
will html6 be generated from the same source as html5?

[04:22:01.0000] <Hixie>
whatever i work on will almost certainly be from the whatwg source document, yeah

[04:22:02.0000] <Hixie>
my workflow is so highly optimised at this point that it'd be stupid for me to try to do anything else

[04:22:03.0000] <Hixie>
(i tried with workers, and it was a huge mistake)

[04:24:00.0000] <annevk>
Hixie, "best spec" I say :)

[04:26:00.0000] <Hixie>
aw

[04:26:01.0000] <Hixie>
/me sad

[04:27:00.0000] <annevk>
heh

[04:27:01.0000] <annevk>
the problem is that there's very little hook for specs that work on top of it

[04:27:02.0000] <annevk>
such as CSSOM View and CSSOM

[04:27:03.0000] <annevk>
e.g. take getPropertyPriority ...

[04:28:00.0000] <Hixie>
it has issues, sure

[04:28:01.0000] <zcorpan_>
but that's just an issue for annevk

[04:29:00.0000] <Hixie>
i've run into that one too actually

[04:29:01.0000] <Hixie>
for canvas in particular

[04:37:00.0000] <hsivonen>
Hixie: will all HTML6 edits be #ifdefed like all Trident bug fixes?

[04:42:00.0000] <Hixie>
bug fixes won't be, new features would be

[04:42:01.0000] <Hixie>
if i do html6

[04:42:02.0000] <Hixie>
which i don't think i will

[04:42:03.0000] <Hixie>
though it is an option

[04:43:00.0000] <jgraham>
/me wonders whether Hixie means "someone else will edit HTML6" or "HTML will move away from version numbers"

[04:43:01.0000] <jgraham>
(or both)

[04:43:02.0000] <Hixie>
neither

[04:43:03.0000] <Hixie>
i meant, "if new features get added to html in the next couple of quarters"

[04:43:04.0000] <jgraham>
Yes, I guess that was option 4

[04:43:05.0000] <Hixie>
(and not added to html5)

[04:44:00.0000] <annevk>
oh man

[04:44:01.0000] <annevk>
getComputedStyle is fucked

[04:44:02.0000] <gsnedders>
Are you just realizing this?

[04:44:03.0000] <annevk>
for 'width' it returns the used value unless the element is not rendered (different from not in the DOM) in which case it returns the computed value

[04:45:00.0000] <annevk>
though not quite the computed value, since all non-percentages and non-keywords are normalized to pixels, which is not an absolute unit

[04:47:00.0000] <annevk>
gsnedders, not really, I realized it two years ago, which is why I postponed working on this

[04:49:00.0000] <annevk>
also, it's not consistent, e.g. for background-position you always get the computed value, normalized to pixels

[04:49:01.0000] <annevk>
so percentages for a rendered document do not get converted to pixels

[04:50:00.0000] <annevk>
the result of all this is http://dev.w3.org/csswg/cssom/#resolved-values

[04:50:01.0000] <annevk>
i now have to reverse engineer every CSS property...

[04:54:00.0000] <annevk>
heh, hsivonen sure is on top of the DOCTYPE discussion :)

[04:55:00.0000] <Dashiva>
Spec design would be so much easier with a time machine

[04:56:00.0000] <annevk>
would be kind of boring to use it for that

[04:56:01.0000] <Dashiva>
Well, you wouldn\t need a complete time machine

[04:56:02.0000] <gsnedders>
Blatantly a better use would be to go to multiple classes at the same time

[04:57:00.0000] <Dashiva>
Like, say, a box where you insert a spec, and it outputs the most updated version of the spec from e.g. 10 years later

[04:57:01.0000] <annevk>
that would leave me with just travelling...

[04:57:02.0000] <annevk>
now you mention it... :p

[05:02:00.0000] <adactio>
Can I get a quick sanity check in case I go posting something stupid to the list/bug tracker? In this table: http://www.whatwg.org/specs/web-apps/current-work/multipage/the-input-element.html#input-type-attr-summary

[05:02:01.0000] <adactio>
Shouldn't it include the "number" type?

[05:02:02.0000] <adactio>
(probably in the same column as "range")

[05:04:00.0000] <annevk>
number is included

[05:05:00.0000] <annevk>
column before

[05:05:01.0000] <annevk>
difference is that you can require a number and a number can be readonly

[05:05:02.0000] <Philip`>
gsnedders: Sorry, Harry Potter references are not permitted in here

[05:10:00.0000] <virtuelv>
Philip`: you missed the memo

[05:10:01.0000] <virtuelv>
only one entertainment item may be disallowed at any gien time

[05:11:00.0000] <virtuelv>
and from now until eternity, it's twilight

[05:11:01.0000] <annevk>
omg that movie was terrible

[05:11:02.0000] <annevk>
still can't believe I went

[05:11:03.0000] <gsnedders>
Philip`: Sorry, it's jgraham's influance on me

[05:11:04.0000] <annevk>
though I fell asleep during Harry Potter too (the more recent movies)

[05:11:05.0000] <adactio>
annevk: thank you for the sanity check. Curse these eyes.

[05:12:00.0000] <annevk>
ah, it's still early :)

[05:12:01.0000] <gsnedders>
Twilight is good if you're a teenage girl who still has some semblance of romantic ideals

[05:12:02.0000] <jgraham>
gsnedders: So ideal for you then?

[05:12:03.0000] <gsnedders>
Indeed.

[05:19:00.0000] <Lachy>
I'm having a hard time figuring out what benefits Jirka perceives in Larry's DOCTYPE proposal, given he agrees that about:legacy-compat already addresses the use case he seems concerned about.

[05:19:01.0000] <Philip`>
Maybe he was unaware of about:legacy-compat

[05:20:00.0000] <Lachy>
I thought he was one of the main proponents pushing for us to add it, so I'm sure he knew about it

[05:29:00.0000] <hsivonen>
annevk: I'm not on top of the doctype discussion in the sense that I don't understand what the expected utility of the proposed change is

[05:40:00.0000] <Lachy>
I don't think anyone really understand what the expected utility of the change is

[05:41:00.0000] <Lachy>
I think we're just getting pro-versioning people jumping up in support, without actually thinking about what they want it for or whether this actually addresses their needs.

[05:43:00.0000] <annevk>
argh, it's even more complicated

[05:43:01.0000] <annevk>
it also depends on the type of box

[05:43:02.0000] <annevk>
though the results are not quite consistent there

[06:04:00.0000] <annevk>
if at least Firefox/WebKit did the same thing it might be somewhat easier:/

[06:05:00.0000] <Philip`>
Isn't it usually easier when they differ, since it indicates a lack of compatibility requirements?

[06:08:00.0000] <annevk>
maybe...

[06:11:00.0000] <Dashiva>
It's good for showing there isn't a lot of content depending on it, but it's still bad if you want to introduce a new feature depending on it, since you have to wait for old browsers to fade away

[06:13:00.0000] <annevk>
the annoying thing is when you reach a solution that means every impl has to change

[06:40:00.0000] <Hixie>
7am really is WAY past my bedtime...

[06:43:00.0000] <TabAtkins>
Dude, Hixie, gtfo.

[06:43:01.0000] <murr4y>
maybe you need to redefine your bedtime ;p

[06:44:00.0000] <TabAtkins>
He lives in california.  7am is ridiculous for him.  ^_^

[06:54:00.0000] <TabAtkins>
My dilemna: Should I (a) point out how many things Shelley wants us to stop talking about, despite never 'quashing' conversation, (b) desperately argue that going this meta signals that we have nothing better to talk about, or (c) ignore the whole lot.

[06:54:01.0000] <TabAtkins>
I'm thinking (c) is the best choice.  Meta breeds meta.

[06:55:00.0000] <AryehGregor>
This is why dictatorship is a good idea for things like specs.

[06:55:01.0000] <AryehGregor>
Otherwise people will talk forever about nothing.

[06:56:00.0000] <jgraham>
TabAtkins: d) employ the "one post per day per thread" rule

[06:56:01.0000] <TabAtkins>
jgraham: Sounds like a plan.

[06:56:02.0000] <TabAtkins>
Sounds like the type of plan I would have come up with yesterday, in fact.

[06:56:03.0000] <jgraham>
Happily I used my post to that thread already today

[06:58:00.0000] <TabAtkins>
Well, I'm talking about the ubermeta "HTML6" thread that just started.  But I think I'll widen my rule to "per topic" rather than "per thread".  Meta is meta is meta.

[06:58:01.0000] <zcorpan_>
/me employs the "zero post per day per thread" rule for threads that have too much indentation in m2

[06:59:00.0000] <jgraham>
/me actually had to stop using an indented view of mail for that reason

[07:00:00.0000] <zcorpan_>
indented threads are great because you can determine from the indentation pattern whether it's worth reading

[07:05:00.0000] <tobyink>
I can't seem to find an answer to this in the HTML5 draft - how to determine the language of an element if xml:lang or lang are set, but to an invalid language code?

[07:06:00.0000] <tobyink>
Do I treat this as lang="" i.e. language has been set to null?

[07:06:01.0000] <tobyink>
Or do I treat it as a missing attribute and take the language from the parent element?

[07:06:02.0000] <gsnedders>
"If the resulting value is not a recognized language code, then it must be treated as an unknown language (as if the value was the empty string)."

[07:07:00.0000] <tobyink>
Ah - well spotted - thanks.

[09:23:00.0000] <AryehGregor>
Is there any point in "If your download does not start automatically, click here" pages?

[09:26:00.0000] <TabAtkins>
It may just be an artefact at this point.  They all use an <iframe> loading the content, which should work everywhere.  At least it gives you something to look at while your browser decodes what the download is supposed to be.

[09:26:01.0000] <AryehGregor>
I thought they're pointless, but opera.com does it too.

[09:26:02.0000] <AryehGregor>
Well, pointless and/or attempts to expose you to more advertising.

[09:27:00.0000] <Dashiva>
Well, sometimes the connection itself fails

[09:27:01.0000] <Dashiva>
Then it's nice to have a link to click for a second try

[09:28:00.0000] <AryehGregor>
You can just click whatever link would have gotten you to the "If your download does not start automatically" page.

[09:36:00.0000] <Dashiva>
Yeah, but that's less obvious to the user

[09:45:00.0000] <Dashiva>
It always hurts a little when people mention separate DOM and HTML specs as a good idea

[10:50:00.0000] <karlushi>
http://confoo.ca/en/session#Web%20Standard

[10:51:00.0000] <karlushi>
mpilgrim will be presenting html5 in montreal

[10:53:00.0000] <mpilgrim>
indeed

[10:53:01.0000] <mpilgrim>
for some definition of "html5"

[10:58:00.0000] <karlushi>
I'll try to be there

[10:59:00.0000] <mpilgrim>
me too

[11:01:00.0000] <karlushi>
it really depends on the snow storm

[11:01:01.0000] <mpilgrim>
why don't i ever get invited to canada in august

[11:02:00.0000] <karlushi>
it's too hot ;)

[11:32:00.0000] <AryehGregor>
http://www.mediawiki.org/wiki/Special:Code/MediaWiki/59741

[11:32:01.0000] <AryehGregor>
Maybe we can try switching to HTML5 again within a couple of months.

[11:32:02.0000] <AryehGregor>
Albeit with an XHTML 1.0 Strict doctype.

[11:35:00.0000] <AryehGregor>
/me goes to read his standards-list mail . . . putting it in a separate inbox is awesome

[12:16:00.0000] <AryehGregor>
"I don't see broad implementation of HTML5 in user agents or the community."

[12:16:01.0000] <AryehGregor>
Um, really?

[12:38:00.0000] <AryehGregor>
/me abides by the "one post per thread per day" rule

[14:55:00.0000] <jgraham>
AryehGregor: FWIW I think I almost entirely disagree with all the points that you made by email

[14:56:00.0000] <jgraham>
But I am loathe to reply there because I don't think having a long drawn out discussion of this is helping anyone

[14:58:00.0000] <jgraham>
I don't think that standards organisations should be compelled to produce a unique tehnology for each use case; that was the thinking that led to HTML being abandoned in the first place

[14:59:00.0000] <Dashiva>
And there are factual errors too

[14:59:01.0000] <Dashiva>
XHTML2 wasn't shut down until quite a while after HTML was resumed

[14:59:02.0000] <jgraham>
I don't think it is possible to go from where RDFa is today and make it "good" by my criteria without making it "bad" from the point of view of the people who currently like it

[15:00:00.0000] <Dashiva>
This "one technology for one thing" issue also completely misses the fact that RDFa was made only for XHTML2

[15:00:01.0000] <Dashiva>
It's like saying don't make jet engines, we'll just put car engines on our planes, they do the same thing

[15:01:00.0000] <jgraham>
the fact that multiple competing standards leads to uncertianty is bad but I expect it to go away over time once it becomes clear whether a technology has real traction

[15:02:00.0000] <jgraham>
I don't think competition is a waste of resources if it leads to a better end result

[15:03:00.0000] <TabAtkins>
jgraham: Agreed, but only up to the point that it's clear that one result is better.  At that point spending resources on the lesser branches is a waste.

[15:04:00.0000] <jgraham>
and I think that given the choice today between using RDFa and using nothing, many people would chose to use nothing i.e. RDFa would not win by default

[15:04:01.0000] <jgraham>
TabAtkins: We are far far away from that point

[15:05:00.0000] <TabAtkins>
We're far away from knowing whether RDFa or Microdata is better for HTML?

[15:05:01.0000] <TabAtkins>
I'd strongly disagree with that.

[15:06:00.0000] <TabAtkins>
Since Microdata is "RDFa, but with better names and some tuneups based on actual use-cases".

[15:06:01.0000] <Dashiva>
And less horrible ideas

[15:06:02.0000] <othermaciej>
we're far away from having agreement in the HTML Working Group that it's clear that one technology is better (even though many individual WG members feel one or the other is clearly better)

[15:07:00.0000] <TabAtkins>
True, agreement in the HTMLWG is still far off.

[15:07:01.0000] <TabAtkins>
On the other hand, I'm somewhat pessimistic of either group actually shifting.

[15:07:02.0000] <jgraham>
TabAtkins: I agree that it is better. But we're far from being able to show that it is better by being more successful

[15:07:03.0000] <othermaciej>
if Microdata had significant deployment success that outpaced RDFa, it would be easier to make the case

[15:08:00.0000] <othermaciej>
or likewise if RDFa had a lot more deployment success and use by data mining tools than it does now, while Microdata failed to get traction

[15:08:01.0000] <othermaciej>
I don't expect either of those things to happen in time for Last Call though

[15:08:02.0000] <Dashiva>
Besides, it seems all parites have acknowledged that RDFa 1.0 is lacking 

[15:09:00.0000] <Dashiva>
So we can't even make a valid comparison until they hash out 1.1

[15:11:00.0000] <jgraham>
/me is still of the opinion that nothing should happen before LC

[15:11:01.0000] <TabAtkins>
I am interested in what they come out with for 1.1.  What I've heard of so far isn't addressing any of my concerns.

[15:11:02.0000] <TabAtkins>
jgraham: I'd be fine with that.  It's the issue-raising people that are trying to force decisions now.  I'm on record as being fine with "wait until it's obvious".

[15:12:00.0000] <Dashiva>
Well, we can't reach LC until the issues are resolved...

[15:15:00.0000] <mpilgrim>
othermaciej: yt?

[15:16:00.0000] <mpilgrim>
any plans to support <input type=color> beyond https://bugs.webkit.org/show_bug.cgi?id=28966 ?

[15:23:00.0000] <othermaciej>
mpilgrim: I'm here

[15:24:00.0000] <othermaciej>
mpilgrim: I think we'd want to use a proper color picker control at some point

[15:24:01.0000] <othermaciej>
mpilgrim: no specific timeline for that, but definitely we do plan to do it in general

[15:26:00.0000] <mpilgrim>
thanks

[15:27:00.0000] <mpilgrim>
one thing i do miss about mac os x is the awesome crayon color picker :)

[15:28:00.0000] <mpilgrim>
btw, the iphone's dynamic keyboard-in-response-to-html5-input-types features heavily in the upcoming "web forms" chapter of dive into html5

[15:28:01.0000] <Dashiva>
My main requirement for a color picker is that it doesn't make it near impossible to get pure colors (in the corners and edges)

[15:29:00.0000] <mpilgrim>
dunno if you had anything to do with that, but please pass along my gratitude to whoever thought of it

[15:29:01.0000] <mpilgrim>
it's far and away the coolest thing i've learned about html5 forms so far

[15:32:00.0000] <othermaciej>
mpilgrim: yeah, that's a neat use of it for sure

[15:34:00.0000] <othermaciej>
Dashiva: I'm not sure what you mean

[15:35:00.0000] <Dashiva>
The color picker in Aion Online only reacted to clicks, you couldn't click and drag to the edge, so it was really hard to hit edge/corner values like pure black

[15:35:01.0000] <othermaciej>
Dashiva: but the OS X color picker dialog gives you a wide range of ways to pick colors, including by RGB clides, CMYK sliders, hue/saturation wheel plus luminosity slider, picking from an image, or selecting one of a number of predefined colors from a box of crayons

[15:36:00.0000] <othermaciej>
(also you can drag a color onto the color well control w/o even popping up a dialog)


2009-12-05
[16:22:00.0000] <TabAtkins>
othermaciej: Sorry for the curtness on the lists.  >_<  Really didn't intend to be rude; I just shouldn't have responded at all.  It was an obviously provocative statement that didn't deserve a response.

[16:25:00.0000] <othermaciej>
TabAtkins: just do your best to be courteous, please, even if you feel provoked for whatever reason

[16:26:00.0000] <TabAtkins>
Yeah, I know.  I'm sorry I let myself go.  I'll watch myself more in the future.e

[16:29:00.0000] <mpilgrim>
png optimization baffles me

[16:30:00.0000] <TabAtkins>
That it works?

[16:35:00.0000] <mpilgrim>
that i've optimized the same images with about six different tools, and each one seems to find a way to make each image slightly smaller than that last round

[16:35:01.0000] <TabAtkins>
You've discovered a perfect compressor.

[16:35:02.0000] <TabAtkins>
Keep going until the file is just 1 bit.

[16:36:00.0000] <mpilgrim>
apparently there's an element of randomness to it as well

[16:37:00.0000] <mpilgrim>
there are far too many ways to express the same image to try them all

[16:37:01.0000] <mpilgrim>
so it's a sort of monte carlo method

[16:38:00.0000] <mpilgrim>
which is quite maddening for those of us who are obsessive about making our sites load more quickly

[16:38:01.0000] <mpilgrim>
because, you know, you could always run it again, and maybe it'll be smaller

[16:38:02.0000] <mpilgrim>
you can't ever be sure that you're done

[16:39:00.0000] <Rik`_>
mpilgrim: do you know about http://imageoptim.pornel.net/ ?

[16:39:01.0000] <Rik`_>
it will take try several tools and pick the best

[17:32:00.0000] <mpilgrim>
Rik`: oh goodie, another tool to try

[17:32:01.0000] <mpilgrim>
you're not helping my neuroses

[17:33:00.0000] <Rik`>
mpilgrim: hopefully, this will be the last one

[17:46:00.0000] <erlehmann>
without dialog, what sensible markup options for an interview are there ?

[18:09:00.0000] <Hixie>
it has been brought to my attention that i was incorrect yesterday in saying RDFa was the work of the XHTML2 WG

[18:09:01.0000] <Hixie>
it was actually a joint project of the XHTML2 WG and the Semantic Web Deployment Working Group

[18:09:02.0000] <Hixie>
(I'd never heard of the Semantic Web Deployment Working Group before)

[18:28:00.0000] <TabAtkins>
erlehmann:  There are many sensible options.  The spec's suggest <p> and <b> method would work just fine.  Ideally I might mark up the speaker with a <hx> and give it display:run-in.

[01:54:00.0000] <Lachy>
I don't get how Shelley can complain about the size of the spec being too large, and yet get all worked up about people suggesting that some features will end up being deferred to a future revision of HTML.

[02:15:00.0000] <Hixie>
it's probably similar to how she thinks that the "most compelling" argument for something she agrees with is my opinion, but that if she disagrees with something, that my opinion is worthless

[02:52:00.0000] <Hixie>
sorry, "a compelling", not necessarily the "most compelling"

[02:53:00.0000] <jgraham>
I don't get why people are so happy to let her dictate the terms of all discussion

[03:00:00.0000] <roc>
Hixie: isn't that common? If my enemy agrees with me *even though* he is evil, that is clearly strong evidence of the universal appeal of my position. On other hand, if my enemy disagrees with me, then that is only because he is evil.

[03:01:00.0000] <Hixie>
roc: I didn't say it was uncommon, but that doesn't make it acceptable

[07:50:00.0000] <mpilgrim>
i wonder if julian realizes he just walked into a trap

[07:50:01.0000] <mpilgrim>
probably not, since that's why they call them "traps"

[07:51:00.0000] <mpilgrim>
IMG isn't controversial *now*, but it was most certainly controversial when it was introduced (by a browser vendor, against the working group's wishes)

[07:52:00.0000] <mpilgrim>
it also wasn't in the first several drafts of XHTML 2, probably because XHTML 2 was TBL's recreation of what he thought HTML always should have been

[07:52:01.0000] <annevk2>
back back

[07:52:02.0000] <annevk2>
discussions on public-html are still somewhat painful to read

[07:53:00.0000] <mpilgrim>
references: http://lists.w3.org/Archives/Public/public-html/2009Dec/0173.html and http://lists.w3.org/Archives/Public/public-html/2009Dec/0174.html

[07:54:00.0000] <mpilgrim>
and TBL never wanted an IMG element

[07:54:01.0000] <annevk2>
I'm guessing some of the participants are not very happy with life

[07:56:00.0000] <mpilgrim>
/me is very happy with life, but finds that reading public-html distracts him from getting real work done

[07:56:01.0000] <mpilgrim>
like writing a book

[07:56:02.0000] <mpilgrim>
unlike irc, which isn't distracting at all :)

[07:57:00.0000] <annevk2>
IRC is nice

[07:58:00.0000] <annevk2>
I've never really seen significant process made on public-html discussions involving "two camps" with mutually incompatible goals

[07:58:01.0000] <annevk2>
I stopped participating in them, mostly

[08:00:00.0000] <annevk2>
The other reason is that most discussions feel like damage control, which I also tried to do in the CDF WG, and was no fun at all and mostly pointless in the end anyway

[08:01:00.0000] <annevk2>
Maybe we should just split the group and have two editions of HTML5. After a decade we can study the then contemporary practices and go to REC

[08:02:00.0000] <mpilgrim>
i'm pretty sure that's inevitable anyway

[08:02:01.0000] <mpilgrim>
maybe shelley et. al. could call their version "HTML5 Plus"

[08:03:00.0000] <annevk2>
sounds fine to me

[08:03:01.0000] <annevk2>
hopefully then everyone can be somewhat more happy instead of all this hostile crap

[08:04:00.0000] <mpilgrim>
i miss XHTML 2

[08:04:01.0000] <mpilgrim>
it provided a nice safe playground where all these like-minded people could play without hurting anyone

[08:04:02.0000] <timz>
mpilgrim: they'd better call it HTML5less

[08:05:00.0000] <mpilgrim>
/me wonders if timz knows the history of "HTML Plus"

[08:05:01.0000] <timz>
hehe no :-)

[08:05:02.0000] <timz>
i'm having a look

[08:06:00.0000] <daedb>
this http://www.w3.org/MarkUp/HTMLPlus/htmlplus_1.html ?

[08:08:00.0000] <timz>
haha, yes that would probably fit her/their ideas of what HTML5 should be

[08:09:00.0000] <mpilgrim>
the W3C has a long history of fucking up HTML by going off into la-la land for a few years while the world evolves without them

[08:09:01.0000] <daedb>
/me has never heard of HTML Plus before...

[08:09:02.0000] <mpilgrim>
HTML+ was one such effort (between HTML 1.0 and HTML 2.0)

[08:09:03.0000] <mpilgrim>
HTML 2.0 was basically "hey, let's see what browsers have actually implemented while we weren't looking"

[08:09:04.0000] <mpilgrim>
then they made HTML 3.0, which was similarly delusional

[08:10:00.0000] <mpilgrim>
then they threw that away and made HTML 3.2, which was basically "hey, let's see what else browsers have actually implmeented while we were busy masturbating"

[08:10:01.0000] <timz>
haha

[08:11:00.0000] <mpilgrim>
then they made HTML 4.0, which shockingly wasn't completely horrible (though it was horribly underspecified)

[08:11:01.0000] <timz>
you've been around all of those discussions or you just red the "history of html, a different perspective"

[08:11:02.0000] <mpilgrim>
then they gave up on HTML altogether, to the point of ignoring all W3C process and refusing to issue errata or integrate existing errata into point releases

[08:12:00.0000] <mpilgrim>
and started whole hog on XHTML

[08:12:01.0000] <mpilgrim>
XHTML 1.0 is literally HTML 4 + a few useless slashes

[08:13:00.0000] <mpilgrim>
oh, and a new MIME type and draconian error handling, which nobody uses (hi Sam)

[08:13:01.0000] <mpilgrim>
then XHTML 1.1, which nobody uses

[08:13:02.0000] <timz>
so is HTML5 a w3c idea or a community (vendors) idea ?

[08:13:03.0000] <mpilgrim>
then XHTML 2, which they spent almost a decade on and finally put out to pasture this year

[08:14:00.0000] <mpilgrim>
HTML5 was a community-led effort of browser vendors, standards wonks (like Hixie), and other interested parties

[08:14:01.0000] <mpilgrim>
all formerly quite active within the W3C

[08:14:02.0000] <mpilgrim>
they begged and pleaded with the W3C to let them work on HTML5 (or 4.1, or whatever, anything but XHTML 2)

[08:14:03.0000] <mpilgrim>
the W3C said no

[08:15:00.0000] <mpilgrim>
so they formed their own working group (WHATWG)

[08:15:01.0000] <mpilgrim>
and started from scratch (since the license of the HTML 4 specification forbids derivative works by non-W3C organizations)

[08:16:00.0000] <Lachy>
so then it seems this HTML+RDFa effort may just be the next round of crap to be attempted during the overall development of HTML, that will likely eventually die out like the rest.

[08:16:01.0000] <mpilgrim>
HTML+RDFa doesn't appear to have any chance at all of being a good spec until they redefine RDFa in terms of the DOM

[08:16:02.0000] <mpilgrim>
anything else is simply insane

[08:17:00.0000] <mpilgrim>
that's not to say people won't use it

[08:17:01.0000] <mpilgrim>
google rich snippets, etc.

[08:17:02.0000] <timz>
with limited vocabs that is

[08:17:03.0000] <mpilgrim>
but defining a new syntax "on top of" HTML in terms of a byte stream is just insane

[08:19:00.0000] <daedb>
heh, HTML+ had a <person> element :)

[08:20:00.0000] <mpilgrim>
http://diveintomark.org/archives/2009/11/02/why-do-we-have-an-img-element is fun related reading on the history of HTML

[08:20:01.0000] <Dashiva>
RDFa being based on the source text kinda makes me think of people parsing HTML with regex

[08:20:02.0000] <mpilgrim>
it (and the rest of this stuff) will form the basis of the introductory chapter of http://diveintohtml5.org/

[08:20:03.0000] <Dashiva>
It seems to work until you start using it for anything significant

[08:21:00.0000] <mpilgrim>
indeed

[08:21:01.0000] <daedb>
<RENDER TAG="PROPNAME" STYLE="I">... that's... interesting :O

[08:22:00.0000] <timz>
mpilgrim: "but defining a new syntax "on top of" HTML", you are referring to RDFa ?

[08:22:01.0000] <mpilgrim>
yes

[08:23:00.0000] <mpilgrim>
it would appear that the current HTML+RDFa effort has zero chance of producing a good specification

[08:23:01.0000] <mpilgrim>
and if RDFa sees enough implementation, we'll need to do an RDFa5 retro-spec in a few years

[08:23:02.0000] <timz>
because it so explicitly bound to xml and rdf ?

[08:24:00.0000] <mpilgrim>
which will consist of reverse engineering google's implementation (or whoever ends up being the dominant consumer of RDFa on the public web)

[08:24:01.0000] <timz>
hmm, sounds like a good argument for keeping microdata

[08:24:02.0000] <Dashiva>
Hmm, yeah

[08:24:03.0000] <mpilgrim>
i'm not a fan of microdata either

[08:25:00.0000] <mpilgrim>
it's a decent enough solution to the stated problem

[08:25:01.0000] <Lachy>
mpilgrim, that's not surprising. I read the HTML+RDFa spec, and it's clearly been written by people with little clue about how to actually write specifications

[08:25:02.0000] <mpilgrim>
i'm just not a fan of the stated problem

[08:25:03.0000] <timz>
hehehe

[08:25:04.0000] <Dashiva>
I'm sure there will be a decent amount of complaints that Google is wilfully sabotaging RDFa by having a non-perfect implementation

[08:25:05.0000] <mpilgrim>
when people say "i want to extend HTML in arbitrary, private, and probably proprietary ways," the appropriate answer is "fuck you"

[08:27:00.0000] <Dashiva>
Extensibility has become one of those words that doesn't mean anything anymore

[08:28:00.0000] <Dashiva>
Like when the topic was making sure HTML5 was easily extensible by HTML6, it turned into a complaint about lack of DE

[08:28:01.0000] <Lachy>
mpilgrim, if their intention is for those extensions to appear on the public web and have 3rd party implmentations and websites using them, then that sure is the right answer

[08:29:00.0000] <Lachy>
Dashiva, even the concept of distributed extensibility doesn't mean much in practice. It's basically become an alias for full namespace support and RDFa

[08:30:00.0000] <Lachy>
and that's despite all the extension points already in HTML5 that allow 3rd party extensions, without using either namespaces or RDFa

[08:31:00.0000] <mpilgrim>
Lachy: haven't you heard? those aren't "real" extensibility points

[08:31:01.0000] <mpilgrim>
Paging True Scotsman, party of 1... True Scotsman, party of 1

[08:31:02.0000] <Lachy>
the problem seems to be that proponents believe those to be the ultimate solution and are unwilling to accept any alternative solution, regardless of any evidence presented, nor evaluation of them against the use cases and problems they solve

[08:36:00.0000] <timz>
http://diveintomark.org/archives/2009/11/02/why-do-we-have-an-img-element <-- good read mpilgrim :-)

[08:37:00.0000] <mpilgrim>
'twas fun to research and write, too

[08:39:00.0000] <Dashiva>
http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2009Dec/0022.html

[08:39:01.0000] <Dashiva>
Validation is overrated, I guess

[08:42:00.0000] <annevk2>
Removing the test seems like a very poor solution to the problem...

[08:43:00.0000] <daedb>
http://www.w3.org/MarkUp/HTMLPlus/htmlplus_21.html Wow, HTML+ has both <img> and <image>... did the alt attribute not exist when that was written?

[08:44:00.0000] <annevk2>
alt was invented for lynx at some point, not sure when

[08:44:01.0000] <mpilgrim>
the alt attribute was first added in HTML 4.0

[08:45:00.0000] <mpilgrim>
lol

[08:45:01.0000] <mpilgrim>
i'd forgotten this salient little detail:

[08:45:02.0000] <mpilgrim>
when the XHTML 2 WG did finally re-introduce the <img> element,

[08:45:03.0000] <mpilgrim>
they gave it a different content type than <img> in XHTML 1 and HTML 4

[08:46:00.0000] <mpilgrim>
they made it a non-void element

[08:46:01.0000] <mpilgrim>
the child content was the fallback content

[08:47:00.0000] <mpilgrim>
which is actually a great idea, but (i'm told) it royally breaks existing web content

[08:47:01.0000] <annevk2>
yeah, it seems they admitted compat was needed, but then at the same time did not

[08:47:02.0000] <annevk2>
or something

[08:47:03.0000] <mpilgrim>
"Also, <img> was reintroduced to XHTML 2, no doubt after a vigorous and healthy debate in which all parties treated each other with mutual respect.  But now it has a different content model than <img> in XHTML 1 and HTML 4, which just goes to show that mutual respect is for chumps."

[08:47:04.0000] <mpilgrim>
http://microformats.org/discuss/mail/microformats-discuss/2005-October/001783.html

[08:56:00.0000] <daedb>
Cool, HTML+ had an element for figures :)

[09:27:00.0000] <Lachy>
daedb, so did HTML3

[09:40:00.0000] <daedb>
Lachy: Hopefully it will stay this time.

[09:41:00.0000] <Lachy>
daedb, it might get dropped again, since we still can't resolve the caption element issue, and the current dt/dd model sucks

[09:53:00.0000] <timz>
what's the primary argument for not using caption ?

[09:53:01.0000] <timz>
in figure

[09:54:00.0000] <Lachy>
timz, compatibility with legacy parsing issues

[09:55:00.0000] <timz>
okay, so html5 documents have to be compatible with older parsers ?

[09:56:00.0000] <Dashiva>
No, html5 parsers have to be compatible with older documents

[09:56:01.0000] <Dashiva>
But the reverse is of course desirable

[10:07:00.0000] <daedb>
Lachy: I really hope not, it's one of my favourite new elements.

[13:32:00.0000] <zcorpan_>
http://www.w3.org/XML/2009/12/xml-stylesheet/

[14:40:00.0000] <AryehGregor>
jgraham, given the choice between using RDFa and using nothing, many people would use nothing.  Except when RDFa is the only thing that works with CC license readers, or Rich Snippets, or . . .

[14:41:00.0000] <AryehGregor>
I'm not worried about individual authors saying "Let's use RDFa, that sounds great!"  I'm worried about organizations with some clout basing new de facto standards on it.

[14:41:01.0000] <AryehGregor>
Thus forcing everyone to use it.

[14:41:02.0000] <AryehGregor>
That's how it's worked its way into MediaWiki.

[14:42:00.0000] <AryehGregor>
I couldn't really argue with "But you have CC license-scraping bots and this would be great for Commons."

[14:42:01.0000] <AryehGregor>
(Wikimedia Commons, that is)

[14:42:02.0000] <AryehGregor>
RDFa is the only Creative Commons-approved way to embed license metadata, AFAIK.

[14:45:00.0000] <hsivonen>
AryehGregor: FWIW, I think committees don't get to vote to kill specs. they only get to vote on whether they allocate their resources to developing a given technology

[14:46:00.0000] <hsivonen>
for example, OASIS doesn't get to vote OOXML out of existence

[14:46:01.0000] <AryehGregor>
Well, the W3C could still endorse one or the other exclusively.

[14:46:02.0000] <hsivonen>
and the W3C didn't successfully vote HTML out of existence

[14:46:03.0000] <hsivonen>
AryehGregor: sure

[14:46:04.0000] <AryehGregor>
So could the HTMLWG.

[14:46:05.0000] <hsivonen>
yes

[14:46:06.0000] <AryehGregor>
That would have an impact.

[14:47:00.0000] <AryehGregor>
Just like disbanding the XHTMLWG had an impact.  We get to say "No, don't use that, it's obsolescent."

[14:47:01.0000] <AryehGregor>
I doubt it will happen, though.

[14:47:02.0000] <AryehGregor>
/me shakes fist dramatically at the W3C

[14:47:03.0000] <hsivonen>
it had a publicity impact

[14:47:04.0000] <hsivonen>
but XHTML2 was already dead 

[14:47:05.0000] <AryehGregor>
Yes, that's my point.

[14:47:06.0000] <hsivonen>
or s still very much alive depending on point of view

[14:48:00.0000] <AryehGregor>
If the W3C said "Hey, we decided RDFa is a bad idea, use microdata instead", then we could say to Creative Commons "Hey, why don't you encourage people to use microdata instead of RDFa?  The W3C says so!"

[14:48:01.0000] <hsivonen>
AryehGregor: I observe that the W3C didn't stop work on XForms when it invited HTML5 (incl. Web Forms 2.0) in

[14:49:00.0000] <AryehGregor>
And I think that's silly too.  :)

[14:49:01.0000] <AryehGregor>
(plus, I'm sure that will die soon enough, unless it has significant non-HTML users)

[14:51:00.0000] <hsivonen>
XForms got baked into ODF 1.0

[14:51:01.0000] <hsivonen>
some approximation of RDFa got baked into ODF 1.1, according to rumors

[14:52:00.0000] <AryehGregor>
Yay.

[14:53:00.0000] <AryehGregor>
Oh well, none of this can possibly be more horrible than all the legacy HTML currently has to deal with.

[14:53:01.0000] <jgraham>
/me wouldn't put money on that

[14:54:00.0000] <jgraham>
(I am about to go sleep otherwise I might try to say something more useful)

[14:57:00.0000] <AryehGregor>
It's okay, it's unlikely there's anything useful to say in this discussion, except "Leave them alone for now and let God sort it out."

[15:05:00.0000] <Philip`>
Seems unfair to leave our mess for Him to clean up

[15:07:00.0000] <AryehGregor>
Well, it's really His mess, since He's the owner of the entire universe, after all.

[15:07:01.0000] <AryehGregor>
So everything is His.

[15:07:02.0000] <AryehGregor>
Also, it's His fault for telling us we can't murder people for promoting bad standards.

[15:11:00.0000] <Philip`>
That seems a little extremist

[15:11:01.0000] <AryehGregor>
Yes, I agree, it's just as likely they'd have murdered us, so we can't really blame Him on that score.

[15:12:00.0000] <AryehGregor>
We can blame the people responsible for the other standards, though.  :)

[15:12:01.0000] <Philip`>
I don't think people who promote what you consider bad standards are that extremist either :-p

[15:19:00.0000] <zcorpan_>
"Sure; but if I've scheduled for it, I can deal with a hundred comments in 

[15:19:01.0000] <zcorpan_>
a day -- that's not a problem." -- http://www.w3.org/mid/Pine.LNX.4.62.0912050854010.5629⊙hdc

[15:22:00.0000] <zcorpan_>
re rjavascript, i guess it would be useful with a linter for js performance pitfalls in modern implementations

[15:22:01.0000] <AryehGregor>
Haha, Google Site Performance is telling me my page is performing poorly because I don't have gzip compression enabled for http://www.google-analytics.com/urchin.js.

[15:22:02.0000] <AryehGregor>
\o/

[15:22:03.0000] <jgraham>
Die with die

[15:23:00.0000] <jgraham>
/me obviously got distracted from going to bed

[15:23:01.0000] <zcorpan_>
or maybe browsers should have a way to log in the error console when the js engine has to do something expensive

[15:23:02.0000] <jgraham>
/me was talking about the "with" statement in case it wasn't obvious

[15:24:00.0000] <jgraham>
zcorpan_: It seems like something Dragonfly et. al. could tell you (depending on how much attaching a debugger affects the actual code)

[15:25:00.0000] <jgraham>
(But it would be nice if they could do per-call profiling with big warnings for things that hit slow paths)

[15:26:00.0000] <Philip`>
"slow paths" include falling off the JIT, which might be really quite difficult to explain to an average script developer

[15:26:01.0000] <Dashiva>
I'm a bit rusty on the edges, but isn't there some overlap between rjavascript and stuff like caja?

[15:27:00.0000] <zcorpan_>
Philip`: no need to explain why it's slow, just state that it's slow and suggest fast alternatives

[15:27:01.0000] <Philip`>
(There's a thing that gives JIT profiling information for code in SpiderMonkey but it's not the most user-friendly thing imaginable, by necessity)

[15:28:00.0000] <jgraham>
Philip`: That was basically what I had in mind hen I said "slow paths". I'm not sure why it would be hard to explain to people sufficiently motivated to profile their code

[15:30:00.0000] <jgraham>
/me really is going to bed now

[15:32:00.0000] <Philip`>
jgraham: It seems hard because they'd first need to understand what a tracing JIT is and how it works

[15:33:00.0000] <Philip`>
http://blog.mozilla.com/dmandelin/2009/02/26/tracevis-performance-visualization-for-tracemonkey/

[15:37:00.0000] <Philip`>
It's easy to suggest "don't use with" and "don't use eval", but I don't really see how you could automatically give suggestions for the more complex problems that occur

[15:40:00.0000] <Philip`>
(It would still be nice, just maybe not possible)

[15:56:00.0000] <AryehGregor>
/me hasn't figured out what "with" actually does yet, but gets the impression that it's cool and trendy and is vaguely disappointed if he shouldn't use it in JS for performance reasons

[15:56:01.0000] <Dashiva>
It's horrible and evil

[15:56:02.0000] <AryehGregor>
But also cool and trendy?

[15:56:03.0000] <Dashiva>
No

[15:57:00.0000] <Dashiva>
It was cool and trendy a few years ago, when having the shortest bookmarklets possible was a goal

[15:57:01.0000] <Dashiva>
And it sometimes appears in golfing, I suppose?


2009-12-06
[16:31:00.0000] <AryehGregor>
"with" sometimes appears in golfing?

[16:33:00.0000] <AryehGregor>
I mean, I imagine it appears in many contexts, being a common English word.

[16:33:01.0000] <Dashiva>
Code golf

[17:52:00.0000] <hsivonen>
I find it most odd that HTTP gurus oppose the use of POST for ping on safety grounds

[17:53:00.0000] <othermaciej>
what are the safety grounds?

[17:53:01.0000] <Hixie>
if anyone ever understand the objection, please let me know, so i can fix it

[17:53:02.0000] <othermaciej>
can ping produce POSTs that can't be produced by posting a form?

[17:53:03.0000] <hsivonen>
othermaciej: see Roy's Change Proposal

[17:53:04.0000] <Hixie>
i've tried many times to understand the objection, but not been able to

[17:54:00.0000] <hsivonen>
/me avoids summarizing in order to avoid missummarizing

[17:55:00.0000] <othermaciej>
hsivonen: ah, when you say "safety grounds" you just mean the official HTTP classification of what methods are labeled "safe"

[17:57:00.0000] <hsivonen>
othermaciej: yes

[17:57:01.0000] <othermaciej>
I wonder if it is a relevant consideration that a submit button can be styled to look just like a link

[17:57:02.0000] <othermaciej>
(let alone the fact that a POST can be issued programmatically from an onclick handler)

[17:58:00.0000] <Hixie>
http://damowmow.com/playground/demos/http/002/ is relevant

[17:58:01.0000] <othermaciej>
what Content-Type does the ping post use?

[17:58:02.0000] <Hixie>
note that the test.html file on that page has no script

[17:59:00.0000] <Hixie>
yet it can trick people (even highly experienced http experts) into causing POSTs to be submitted

[17:59:01.0000] <othermaciej>
clever, that one doesn't even require form controls to be stylable

[18:00:00.0000] <Dashiva>
othermaciej: It doesn't seem that anything is a relevant consideration except the notion that POST is unsafe

[18:00:01.0000] <Hixie>
i guess if POST is unsafe and POST is part of HTTP that means HTTP is unsafe!

[18:02:00.0000] <Dashiva>
Only in browsers, because browsers never do anything right

[02:19:00.0000] <mitnavn>
Is there a wrapping element for HTML 5? I need a container of all my elements to set a side width, should I just use a div?

[02:19:01.0000] <Hixie>
yes

[02:19:02.0000] <Hixie>
<div> is that element

[02:20:00.0000] <mitnavn>
Thanks

[03:09:00.0000] <gsnedders>
jgraham: yt?

[03:39:00.0000] <jgraham>
gsnedders: Yes

[07:04:00.0000] <Dashiva>
How many times does someone have to repeat obvious falsehoods and be corrected before you assume bad faith?

[07:06:00.0000] <Steve^>
6 or 7

[07:13:00.0000] <virtuelv>
http://qdb.us/295884

[11:08:00.0000] <Dashiva>
http://www.android.com/ has a <link rel="stylesheet" href="//google.com/css/inlay.css">

[11:08:01.0000] <Dashiva>
I wonder if that href type is going to catch on

[11:10:00.0000] <hsivonen>
Dashiva: I expect redirect from http to https catch on first

[11:12:00.0000] <Dashiva>
How does that look?

[11:14:00.0000] <hsivonen>
Dashiva: just a 301 respons with an https URL in the Location header

[11:24:00.0000] <webben>
hsivonen: How does the server know to offer a redirect to an HTTPS alternative? Referer header?

[11:26:00.0000] <Philip`>
Dashiva: Slashdot has used that style for ages

[11:28:00.0000] <hsivonen>
webben: by seeing which port was used for the request

[11:30:00.0000] <hsivonen>
/me wishes the chairs would point out to Shelley that the HTML WG charter covers application in addition to documents

[11:30:01.0000] <webben>
hsivonen: But if you serve up HTML response over SSL, but link to a resource with http:// won't it use port 80 when requesting that resource?

[11:31:00.0000] <webben>
s/it/the UA/

[11:31:01.0000] <hsivonen>
webben: I meant not serving up HTML as http but redirecting to https at that point

[14:06:00.0000] <boblet>
hey all, re: abbr/acronym, is there a way now for authors to indicate whether the term should be read as a word (like FUD) or as initials (like CIA)? Or is that up to AT?

[14:10:00.0000] <zcorpan>
boblet: it's up to the AT, as it has always been

[14:11:00.0000] <boblet>
zcorpan: ok, thanks. I was under the impression that historically that was the distinction between abbr/acronym. good to know

[14:13:00.0000] <Philip`>
As far as I'm aware it should be the same as for non-marked-up abbreviations, and the same as for all non-abbreviated collections of characters too

[14:14:00.0000] <Philip`>
i.e. it's the AT's job to work out how to read text

[14:25:00.0000] <boblet>
another random question; it’s fine to have an id and a class with the same name, right? Spec doesn’t forbid it (id value unique amongst all ids in subtree), but wondering if eg browser bug or other gotcha

[14:26:00.0000] <zcorpan>
it's fine

[14:26:01.0000] <boblet>
thanks zcorpan 

[15:12:00.0000] <zcorpan>
so anyone's got any comments on the xml-stylesheet draft?

[15:59:00.0000] <AryehGregor>
Ooh, Chrome has neat UI for untrusted SSL certs: it writes "https" in red, but has a little slash through it too.

[15:59:01.0000] <AryehGregor>
Very noticeable.


2009-12-07
[16:00:00.0000] <zcorpan>
not as noticeable as in ie where the whole address bar is red and there's a red warning box next to the address bar

[16:02:00.0000] <AryehGregor>
Hmm, true.  Looks neater, though.

[16:07:00.0000] <hsivonen>
also not as noticeable as having to click through a few dialogs.

[16:09:00.0000] <AryehGregor>
Yes, I should really have said "cure" or something.

[16:09:01.0000] <AryehGregor>
"cute"

[16:09:02.0000] <AryehGregor>
It's rather less noticeable than some browsers' warnings.

[16:09:03.0000] <AryehGregor>
Which is fair enough, since 99.8% of the time, it will just be a site with a bad cert and not someone evil.

[16:10:00.0000] <AryehGregor>
Freaking out just encourages users to ignore security warnings, since they rightfully conclude that browser implementers are paranoid.

[16:11:00.0000] <AryehGregor>
Which they all seem to be, without exception.  I haven't ever seen a sane response from a browser to something like a bad SSL certificate.

[16:13:00.0000] <AryehGregor>
An appropriate response when visiting a new site would probably be a prominent visual cue in the URL bar, but with no extra screen to click through unless you try to submit a form or something.

[16:14:00.0000] <doublec>
is that what chrome does?

[16:15:00.0000] <AryehGregor>
No, Chrome gives a moderately scary page you have to click through.

[16:15:01.0000] <AryehGregor>
They let you do it in one click, but it's still overkill if you're visiting some random wiki or whatever.

[16:15:02.0000] <AryehGregor>
https://circumflex.us/wiki/ is not a security threat.

[16:16:00.0000] <AryehGregor>
Yeah, it would be nice if you could just fail on a bad cert, but in practice they're almost all non-malicious, so let's not try to misrepresent reality to users, shall we?

[16:17:00.0000] <AryehGregor>
There's some thing going on in public-webapps where they want to allow sites to force SSL only, and that's great.

[16:17:01.0000] <AryehGregor>
SSL only with hard failure on cert errors, obviously.

[16:17:02.0000] <zcorpan>
a visual cue is easily missed by users, so it'd be pretty bad in the 0.2% cases where the site is evil and steals all your money (or whatever)

[16:18:00.0000] <doublec>
zcorpan, users will be trained by then to automatically click the scary dialogs though

[16:18:01.0000] <doublec>
the dialogs do more bad than good. Since the majority of the time the user will ignore them, they'll do it for the actual bad case too

[16:19:00.0000] <AryehGregor>
zcorpan, 1) You could have the page only pop up on form submissions.  2) Overstating severity in almost all cases will cause users to just ignore you in the long term, it's never the right answer.

[16:20:00.0000] <zcorpan>
what if the evil site has a form but the script prevents submission and steals login credentials?

[16:21:00.0000] <AryehGregor>
Then I guess you're screwed.  See (2).  Of course, you can add other heuristics too, like give scary errors only if the user has previously visited the site with a valid SSL certificate.

[16:21:01.0000] <AryehGregor>
That will help prevent obvious MITM against amazon.com or such.

[16:21:02.0000] <doublec>
that's a good option

[16:21:03.0000] <AryehGregor>
Also, you could pop up the warning as soon as the user starts typing in a form.

[16:22:00.0000] <AryehGregor>
The point is, though, just popping up Yet Another Warning to click through isn't a good option at all.  At least AFAIK.  Has anyone done usability research on this?

[16:22:01.0000] <zcorpan>
you could trick the user to think he's typing in a form but there's actually no form so the browser can't know when to pop up the warning

[16:23:00.0000] <AryehGregor>
Yes, then you're screwed, as I said.

[16:23:01.0000] <Philip`>
Perhaps the solution is that everyone should stop using SSL, unless they're something serious like a bank

[16:24:00.0000] <AryehGregor>
It reminds me of an article I read a while back by some guy.  He remarked that as a Windows GUI user, he was trained to just ignore all warnings, because he figured they wouldn't let him do anything really unsafe.  Then he tried Linux for a while.  After a while, for some reason he wanted to run fsck on a mounted filesystem, or such.  It said "WARNING, this filesystem seems to be mounted, running fsck will probably destroy all your data.  Are

[16:24:01.0000] <AryehGregor>
 you sure?"  He said yes and lost all his data, and from then on learned that Linux command line utilities actually give accurate warnings.

[16:24:02.0000] <Philip`>
to avoid the situation where broken certificates on random mailing lists and bugzillas you find when searching on Google result in you learning to ignore broken certificates

[16:24:03.0000] <AryehGregor>
Except that actually he didn't, because this article was about how he tried out btrfs as his root filesystem even though it was clearly marked as usable only for development, and he found out it wasn't usable.

[16:25:00.0000] <AryehGregor>
Philip`, a broken cert is actually more secure than no cert.  It doesn't prevent an active attacker, but it prevents passive sniffing.

[16:25:01.0000] <AryehGregor>
Also, if you whitelist it the first time, you'll be informed again if it changes.

[16:26:00.0000] <Philip`>
/me is partially responsible for proliferation of insecure certificates, since he set up a self-signed HTTPS thing on his server primarily so that SVN clients could get safely through proxies

[16:27:00.0000] <doublec>
Philip`: your certificate isn't insecure though, right?

[16:27:01.0000] <AryehGregor>
He said it's self-signed.

[16:27:02.0000] <doublec>
It just doesn't prove your identify

[16:27:03.0000] <AryehGregor>
Well, the same is true for any cert, typically.

[16:28:00.0000] <doublec>
true

[16:28:01.0000] <AryehGregor>
Also also, in practice, users will often visit SSL-secured sites via ordinary HTTP (e.g., typing "amazon.com" into URL bar), so you can spoof some percentage even without raising SSL warnings.

[16:30:00.0000] <Philip`>
doublec: I guess it encrypts securely, assuming the command-line string I copied from a random website is set up correctly

[16:30:01.0000] <doublec>
haha, good point

[16:31:00.0000] <Philip`>
but it's self-signed, it's not set up with the same domain name it's being used on, and I won't notice or care enough to update it once it reaches the expiry date

[16:31:01.0000] <doublec>
and that you didn't use a version of an os with a bad openssl library

[16:31:02.0000] <AryehGregor>
Proposed UI #2: Flat-out refuse SSL access with a bad cert to a manually-maintained list of high-profile e-commerce sites, with no opt-out.  For other sites, on the first visit have a non-blocking warning in the top, saying something like "This site may not be secure; don't submit sensitive information to it. [More Info] [X]"

[16:31:03.0000] <AryehGregor>
Your list can safely include any large e-commerce site, I'd hope, so it can be pretty large.

[16:32:00.0000] <hsivonen>
I've been unable to instruct myself or people who turn to me for advice how to tell apart innocent cert expiration and an attack

[16:32:01.0000] <hsivonen>
seems cost-effective to always assume innocent expiration

[16:32:02.0000] <AryehGregor>
Check the hash on the root CA and manually compare it to the correct hash obtained by Googling.

[16:33:00.0000] <AryehGregor>
Expiration or just misconfiguration.

[16:33:01.0000] <AryehGregor>
SSL is a PITA to configure right.

[16:33:02.0000] <AryehGregor>
Especially if you have multiple hosts on the same machine.

[16:33:03.0000] <AryehGregor>
There was a time when https://google.com would return a cert error.

[16:34:00.0000] <hsivonen>
I don't know what hash "the hash" is

[16:34:01.0000] <AryehGregor>
There you go: https://amazon.com/

[16:34:02.0000] <AryehGregor>
Hmm, does Chrome expose cert details anywhere?

[16:35:00.0000] <hsivonen>
as far as I'm aware, I've never been MITM'ed but I've seen a zillion cert failures

[16:35:01.0000] <hsivonen>
just clicking through is rational economics

[16:36:00.0000] <AryehGregor>
Yes, precisely.  And by misrepresenting this basic fact, all the major browsers are contributing to user contempt for browser warnings.

[16:36:01.0000] <AryehGregor>
I've had similar experiences with Chrome's malware-site-blacklist.

[16:36:02.0000] <hsivonen>
(I have, as a matter of principle, spent non-trivial time double-checking failures, though)

[16:37:00.0000] <AryehGregor>
It repeatedly brought up a giant scary red page on my site because someone hotlinked an image hosted on a blacklisted domain.

[16:37:01.0000] <hsivonen>
http://hsivonen.iki.fi/extended-uncertainty/

[16:38:00.0000] <Philip`>
/me wonders if a more usable system for preventing realistic attacks is to simply proxy all traffic through a trusted server with a hard-coded properly-maintained certificate, rather than relying on a zillion sites to each implement it themselves

[16:38:01.0000] <AryehGregor>
In Firefox, Page Info -> Security -> View Certificate -> Details, I think "Subject's Public Key" for the highest-level certificate is what's important.

[16:38:02.0000] <AryehGregor>
But I'm not sure.

[16:38:03.0000] <AryehGregor>
Philip`, how would those sites authenticate to the server?

[16:39:00.0000] <Philip`>
since that'd at least work for the situation where you're connecting through an untrusted wireless network

[16:39:01.0000] <Philip`>
AryehGregor: They wouldn't

[16:39:02.0000] <Philip`>
It's unlikely anyone will MITM between the proxy server and the web site

[16:40:00.0000] <AryehGregor>
Hmm.

[16:40:01.0000] <AryehGregor>
You could hack a router, it's not necessarily very hard.

[16:40:02.0000] <AryehGregor>
But harder than setting up a free wireless access point, yes.

[16:41:00.0000] <AryehGregor>
I would really like to see some statistics on how often MITM attacks actually occur, and what users do in the face of different types of error pages.

[16:41:01.0000] <AryehGregor>
I suspect that MITM attacks are *vanishingly* rare in practice.

[16:41:02.0000] <AryehGregor>
Like say less than 1 in 10,000 cert errors.

[16:41:03.0000] <AryehGregor>
Or even less.  I've never even *heard* of an actual MITM attack.

[16:41:04.0000] <AryehGregor>
I mean, on SSL on the public web etc.

[16:42:00.0000] <AryehGregor>
The sort we're talking about, like someone steals your bank password.

[16:42:01.0000] <Philip`>
You can download easy-to-use tools to break into wireless networks and capture traffic, and as far as I'm aware there's nothing quite so trivial for breaking into remote internet routers

[16:43:00.0000] <AryehGregor>
Yes, it would be harder.

[16:44:00.0000] <Philip`>
MITMing doesn't need to be made impossible, it just needs to be made hard enough that people won't attack that way and will use phishing or trojans or whatever

[16:44:01.0000] <Philip`>
(which I presume is what they already do)

[16:44:02.0000] <zcorpan>
mpilgrim: in the forms chapter where you talk about type=email, it seems you've forgotten about form validation

[16:44:03.0000] <AryehGregor>
Ultimately it should be made impossible if we can.  Making it unattractive is a good start, though.

[16:45:00.0000] <zcorpan>
mpilgrim: opera won't submit the form if the user types in an invalid email address

[16:45:01.0000] <AryehGregor>
I don't think routing everything through a central server is a practical solution, though.

[16:45:02.0000] <Philip`>
AryehGregor: Have lots of decentralised servers :-)

[16:46:00.0000] <Philip`>
They just need to be connected to infrastructure that's reasonably well maintained and not entirely trivial to break into

[16:50:00.0000] <zcorpan>
earlier opera versions also supported autofilling type=email with addresses from the address book, but that was removed to prevent phishing the user's address book with clever tricks, iirc

[16:51:00.0000] <hsivonen>
zcorpan: :-(

[16:51:01.0000] <hsivonen>
zcorpan: has type=url autocompletion been removed, too?

[16:52:00.0000] <zcorpan>
no

[16:53:00.0000] <zcorpan>
i guess it could be used to find out which bank the user uses

[16:53:01.0000] <AryehGregor>
You can already do that by looking at the link color.

[16:54:00.0000] <zcorpan>
but then you need to guess the url

[16:54:01.0000] <zcorpan>
with type=url you can have a fun game that's controlled with up and down keys and enter

[16:55:00.0000] <AryehGregor>
Just try the URLs of every major bank.  You can do lots per second, with no user interaction.

[16:55:01.0000] <zcorpan>
yeah

[16:55:02.0000] <AryehGregor>
Have you seen those cool demos where you just visit a web page and let it run for a while and it gives you a square with the favicons of all the sites you visit?

[16:55:03.0000] <zcorpan>
no

[16:56:00.0000] <zcorpan>
anyway, bed time

[16:56:01.0000] <daedb>
AryehGregor: do you have a link for that favicon demo?

[16:57:00.0000] <AryehGregor>
Sadly, no.

[17:10:00.0000] <AryehGregor>
Is anyone other than Gecko planning on switching to HTML5 parsing anytime soon?

[22:18:00.0000] <Hixie>
/me tries to summon TabAtkins

[03:19:00.0000] <gsnedders>
Should crossOriginIframe.contentDocument.foobar throw SECURITY_ERR?

[03:20:00.0000] <gsnedders>
Yes, because everything goes to the getter

[03:49:00.0000] <WWFan>
i almost a SECURITY_ERR once

[04:40:00.0000] <zcorpan_>
Hixie: hmm, so should the ApplicationCache interface object be accessible to a shared worker, but not a dedicated worker?

[04:40:01.0000] <Hixie>
currently yes

[04:43:00.0000] <gsnedders>
Hixie: As far as I can tell, crossOriginIframe.contentDocument[1] shouldn't throw SECURITY_ERR. Is that intended?

[04:44:00.0000] <Hixie>
off-hand, no idea whatsoever

[04:44:01.0000] <gsnedders>
What about crossOriginIframe.contentDocument.foobar?

[04:45:00.0000] <gsnedders>
For that I can read the spec either way as to whether it can throw.

[04:45:01.0000] <Hixie>
i do not feel comfortable answering security-level questions without several hours of getting familiar with the subject matter again :-)

[04:47:00.0000] <gsnedders>
:)

[04:58:00.0000] <gsnedders>
So, per HTML 5, everything in crossOriginIframe.contentDocument that is in Document should throw.

[04:58:01.0000] <gsnedders>
This only happens in IE on the real web.

[04:59:00.0000] <gsnedders>
At least in a simple case like http://software.hixie.ch/utilities/js/live-dom-viewer/saved/329

[04:59:01.0000] <Hixie>
what do other UAs do?

[05:02:00.0000] <gsnedders>
Everything else gives the element

[05:02:01.0000] <gsnedders>
i.e., HTMLHtmlElement

[05:05:00.0000] <Hixie>
safari seems to be showing another document than the one returned through the dom

[05:06:00.0000] <gsnedders>
Yet in most cases all browsers throw

[05:06:01.0000] <gsnedders>
What's going on with that simple case?

[05:06:02.0000] <Hixie>
timing maybe?

[05:06:03.0000] <Hixie>
yes, timing

[05:06:04.0000] <Hixie>
i bet that's the about:blank page

[05:06:05.0000] <gsnedders>
Ah

[05:07:00.0000] <Hixie>
before it loads example.com

[05:07:01.0000] <gsnedders>
Right

[05:07:02.0000] <gsnedders>
Duh, that makes sense

[05:07:03.0000] <Hixie>
man you're making me work on stuff

[05:11:00.0000] <gsnedders>
The fact that iframe.contentDocument should throw is implicit from the fact that it is accessing a member of a Window object?

[05:19:00.0000] <salty-horse>
is there an extra word here (marked with ***)? -- "It indicates that ***the*** no referrer information is to be leaked when following the link." -- http://www.whatwg.org/specs/web-apps/current-work/multipage/links.html#link-type-noreferrer

[05:19:01.0000] <Hixie>
yes!

[05:19:02.0000] <Hixie>
can you file a bug? just type basically what you just typed here into the text field at the bottom of the screen on that page

[05:19:03.0000] <Hixie>
and hit the "submit review comments" button

[05:20:00.0000] <salty-horse>
sure. it's also quoted in mark pilgrim's dive into html 5, so I'll file it someplace else too :)

[05:20:01.0000] <Hixie>
:-)

[05:20:02.0000] <Hixie>
thanks!

[05:22:00.0000] <salty-horse>
filed at http://www.w3.org/Bugs/Public/show_bug.cgi?id=8448

[05:26:00.0000] <Hixie>
thanks!

[06:40:00.0000] <zcorpan_>
Hixie: why is there a naming difference between Navigator and WorkerLocation? why not WorkerNavigator, or Location?

[06:59:00.0000] <Hixie>
zcorpan_: which Navigator?

[07:12:00.0000] <zcorpan_>
Hixie: in workers

[07:13:00.0000] <Hixie>
dunno

[07:13:01.0000] <Hixie>
probably just a msitake

[07:13:02.0000] <Hixie>
mistake

[07:14:00.0000] <zcorpan_>
wonder if it's worth making them consistent

[07:14:01.0000] <zcorpan_>
btw, are interface objects that are defined in the workers spec support to be available in the non-worker context?

[07:15:00.0000] <zcorpan_>
s/support/supposed/

[07:18:00.0000] <jgraham>
Hixie: Is it known that the permalink feature on Live Dom Viewer gices a 503?

[07:18:01.0000] <jgraham>
*gives

[07:19:00.0000] <Hixie>
zcorpan_: no

[07:19:01.0000] <Hixie>
it is not known

[07:19:02.0000] <Hixie>
what's 503??

[07:21:00.0000] <gsnedders>
Service Not Available

[07:21:01.0000] <zcorpan_>
Hixie: certainly Worker and SharedWorker should be available. it's not clear from the spec which interface objects and constructors should be available to the non-worker context

[07:21:02.0000] <Hixie>
zcorpan_: send mail to whatwg, i'll fix it. in general, only worker and sharedworker are supposed to be outside workers

[07:21:03.0000] <zcorpan_>
ok

[07:23:00.0000] <Hixie>
jgraham: wfm

[07:23:01.0000] <Hixie>
jgraham: what browser?

[07:23:02.0000] <Hixie>
crap, i just noticed i missed tab

[07:24:00.0000] <gsnedders>
Only some URLs

[07:24:01.0000] <gsnedders>
http://software.hixie.ch/utilities/js/live-dom-viewer/?<script>%0Avar%20re%20%3D%20/a/g;%0Avar%20li%20%3D%20re.lastIndex;%0Are.exec('aa');%0Aw(li);%0A</script>

[07:24:02.0000] <gsnedders>
mod_security?

[07:24:03.0000] <Hixie>
escape the <

[07:24:04.0000] <Hixie>
TabAtkins!

[07:24:05.0000] <TabAtkins>
Hixie!

[07:24:06.0000] <Hixie>
gsnedders: hm, maybe not

[07:24:07.0000] <Hixie>
gsnedders: dunno

[07:25:00.0000] <gsnedders>
That doesn't matter

[07:33:00.0000] <Philip`>
/me wonders if Hixie is having fun on public-html

[08:13:00.0000] <othermaciej>
public-html is more fun than a barrel of monkeys

[08:15:00.0000] <Hixie>
i don't know if "fun" is the word i would use

[08:15:01.0000] <Hixie>
but it's unclear to me how to make progress on the change proposal other than active debate at this point

[08:16:00.0000] <jgraham>
votes?

[08:17:00.0000] <Hixie>
i can't trigger a vote

[08:17:01.0000] <Philip`>
Dictatorial decision-making?

[08:17:02.0000] <jgraham>
(I guess the debate might be useful for making the votes more informed, but it might also just cause people to give up on public-html)

[08:18:00.0000] <jgraham>
Hixie: Sure. I don't think that there is anything that you can do. 

[08:19:00.0000] <Hixie>
Philip`: tried that already :-)

[08:19:01.0000] <Hixie>
jgraham: well then why are the chairs not doing anything, if there's nothing we can do?

[08:20:00.0000] <jgraham>
Hixie: No idea. Ask them

[08:21:00.0000] <Hixie>
othermaciej: how can we make progress?

[08:21:01.0000] <Hixie>
(since apparently people don't think my responding to shelley's e-mails is helping)

[08:21:02.0000] <othermaciej>
Hixie: the Chairs will do something, though I can't tell you what yet

[08:22:00.0000] <Hixie>
that's what you said last week :-)

[08:22:01.0000] <othermaciej>
Hixie: we meet every week for an hour on Tuesdays

[08:22:02.0000] <Hixie>
if a weekly telecon prevents out of band progress, that's a sign that the telecon is ineffective at helping make progress

[08:23:00.0000] <othermaciej>
actually we mostly just talk on IRC

[08:23:01.0000] <othermaciej>
not that it's relevant either way

[08:23:02.0000] <othermaciej>
it has helped us to get things completed that we couldn't finish by email

[08:24:00.0000] <othermaciej>
if we haven't stated a path forward by Wednesday you are welcome to point at me and laugh

[08:25:00.0000] <Hixie>
it just seems sad that it's been a month and a bit since WHATWG got to LC, and a month since TPAC, and the group has still not made any decisions this year

[08:25:01.0000] <Philip`>
/me thinks Hixie's latest email was helpful to read, since it explained most of the relevant reasoning and means he doesn't have to bother reading or remembering what was said in the rest of the thread, but he doesn't imagine it'll stop the debate

[08:26:00.0000] <Philip`>
(I guess it'd be a useful thing to point at if there's ever a vote or something)

[08:28:00.0000] <othermaciej>
indeed, we haven't made any formal decisions yet (though some issues are now overdue to close without prejudice)

[08:28:01.0000] <othermaciej>
we do have many things in the pipeline

[08:28:02.0000] <othermaciej>
it would be good to get at least one issue properly fully decided before the end of the year

[08:29:00.0000] <othermaciej>
then I'll think the number of things in the pipeline is a positive sign for progress rather than just changing to a different state of quagmire

[08:31:00.0000] <gsnedders>
/me finally gets around to reading email backlog… fun…

[08:33:00.0000] <annevk2>
gsnedders, I agree...

[08:33:01.0000] <annevk2>
/me is also working on that

[08:34:00.0000] <Philip`>
/me wonders if the "Moratorium on the spec-splitting discussion" on public-html from November 2008 has expired yet

[08:35:00.0000] <gsnedders>
AFAIK it hasn't.

[08:42:00.0000] <Philip`>
/me likes how <p property="ex:test">test <!-- test --></p> gives him three different outputs from current RDFa implementations

[08:55:00.0000] <timz>
is there a quick overview of what elements can be the contents of other elements or do i have to explore section 4 of the spec.

[09:00:00.0000] <JonathanNeal>
Is hgroup a block level element?

[09:02:00.0000] <timz>
formatBlock candidate.

[09:03:00.0000] <timz>
but, hehe, i don't think says anythin aabout being a block element , sorry

[09:05:00.0000] <JonathanNeal>
formatBlock candidate means "could be block soon, yo." 

[09:05:01.0000] <JonathanNeal>
?

[09:05:02.0000] <TabAtkins>
JonathanNeal: Since it wraps block-level elements, I'm going with "yes".

[09:06:00.0000] <TabAtkins>
"formatBlock candidate" means you can use it as the second argument to the "formatBlock" command, which wraps something in whatever you give as the second argument.

[09:06:01.0000] <TabAtkins>
So that lends itself to the "hgroup is a block" theory as well.

[09:06:02.0000] <Philip`>
JonathanNeal: What do you mean by "block level element"?

[09:07:00.0000] <TabAtkins>
Philip`: Default styles, in the CSS sense.

[09:07:01.0000] <JonathanNeal>
TabAtkins, I agree.

[09:07:02.0000] <Philip`>
If you care about default styles, look in the default styles section of the spec

[09:07:03.0000] <Philip`>
http://www.whatwg.org/specs/web-apps/current-work/multipage/rendering.html#display-types

[09:08:00.0000] <JonathanNeal>
Philip`, I meant in a "what you see" css sense, yes.

[09:08:01.0000] <Philip`>
"...hgroup... { display: block; }"

[09:08:02.0000] <JonathanNeal>
I love http://www.whatwg.org/specs/web-apps/current-work/multipage/rendering.html#display-types I was looking for that page earlier today.

[09:10:00.0000] <JonathanNeal>
I keep seeing only a few when news articles come out like http://www.techradar.com/news/internet/everything-you-need-to-know-about-html5-654828

[09:11:00.0000] <JonathanNeal>
These informative blog postings about html5, and I think "isn't there a page already out there they could reference to get this right?"

[09:11:01.0000] <Philip`>
They should reference http://www.whatwg.org/specs/web-apps/current-work/ :-)

[09:24:00.0000] <JonathanNeal>
blink { text-decoration: blink; }

[09:24:01.0000] <JonathanNeal>
 lol

[15:17:00.0000] <salty-horse>
nitpick: shouldn't the FAQ/Blog/.. links in whatwg.org use header tags instead of <em>?  http://gsnedders.html5.org/outliner/process.py?url=http%3A%2F%2Fwhatwg.org

[15:32:00.0000] <salty-horse>
is there an html5 way to specify that an <article> with a summarized post in a blog's main page and the permalink of the post page have the same content?


2009-12-08
[01:49:00.0000] <zcorpan_>
isn't chairship in html wg also invite-only?

[01:49:01.0000] <jgraham>
zcorpan_: Didn't hixie say that?

[01:50:00.0000] <zcorpan_>
dunno, still have 24 emails to read

[03:23:00.0000] <annevk3>
I guess I should review Uniform Messaging at some point...

[03:26:00.0000] <Hixie>
does it affect xbl2, <canvas>/<img>, and <video>'s future use of CORS?

[03:29:00.0000] <annevk3>
They say they want to replace CORS with it

[03:30:00.0000] <annevk3>
And solve all the same use cases...

[03:30:01.0000] <annevk3>
I don't think that is possible given what I know about the design and I believe supporting UM on top of CORS is trivial, but at some point I ought to make sure that is all true and point it out I guess...

[03:31:00.0000] <gsnedders>
"That's what she says/But the words of an eager lover should be treated like words written on wind and swift flowing water"

[03:32:00.0000] <Hixie>
good thing i didn't integrate anything with CORS yet i guess

[03:33:00.0000] <Hixie>
i wonder if UM also affects the CORS-like thing in websockets

[03:34:00.0000] <annevk3>
I would guess Mark is not in favor of that either since it is origin-based

[03:45:00.0000] <salty-horse>
the second example here uses a really odd notation for tags closing the one on the previous line: "><p" -- http://www.whatwg.org/specs/web-apps/current-work/multipage/sections.html#the-section-element

[03:45:01.0000] <salty-horse>
and the zork example here also has this notation, and it's wrong (closes a tag twice): http://www.whatwg.org/specs/web-apps/current-work/#the-pre-element

[03:46:00.0000] <salty-horse>
it also appears here: http://www.whatwg.org/specs/web-apps/current-work/#a-group-of-images-that-form-a-single-larger-picture-with-no-links

[03:46:01.0000] <annevk3>
weird syntax is on purpose

[03:46:02.0000] <annevk3>
for bugs you should prolly file bugs

[04:01:00.0000] <salty-horse>
filed: http://www.w3.org/Bugs/Public/show_bug.cgi?id=8455

[04:01:01.0000] <Philip`>
erlehmann: You can just copy-and-paste the URL and add "multipage" before the # and it'll give you the proper version

[04:03:00.0000] <salty-horse>
Philip`, nice!

[04:13:00.0000] <annevk3>
what also worries me a bit is that studying further threads such as "Re: CSRF vulnerability in Tyler's GuestXHR protocol?" have unanswered questions

[04:13:01.0000] <annevk3>
it's not really clear to me everything has been thought through entirely

[04:14:00.0000] <Hixie>
i haven't been paying enough attention to really know either way

[04:15:00.0000] <murr4y>
i just sent my first mailing list mail ever :p

[04:15:01.0000] <murr4y>
am i supposed to recieve it back from the list?

[04:15:02.0000] <Hixie>
:-)

[04:15:03.0000] <Hixie>
depends on your settings and your mail service provider

[04:16:00.0000] <Hixie>
which list?

[04:16:01.0000] <murr4y>
help

[04:17:00.0000] <Hixie>
what was the subject?

[04:19:00.0000] <murr4y>
it was: Validator error: End tag for body seen but there wereunclosed elements

[04:19:01.0000] <murr4y>
my name is ali

[04:19:02.0000] <murr4y>
just checked my settings, i am set to recieve my own posts

[04:19:03.0000] <murr4y>
maybe it takes a few minutes, guess i was a bit impatient :p

[04:20:00.0000] <Hixie>
ah yes, i saw your mail a while back

[04:20:01.0000] <Hixie>
dunno why you wouldn't have received it

[04:20:02.0000] <Hixie>
it did make it to the list :-)

[04:20:03.0000] <Philip`>
/me thought the default was to not receive your own messages

[04:20:04.0000] <Dashiva>
Is Roy really arguing that we shouldn't add anything that can already be done in JS?

[04:20:05.0000] <annevk3>
murr4y, see http://lists.whatwg.org/pipermail/help-whatwg.org/2009-December/date.html

[04:21:00.0000] <murr4y>
ah, nice :)

[04:23:00.0000] <Dashiva>
... and he doesn't seem to understand the point of CR

[04:23:01.0000] <annevk3>
Hixie, so the network events thingie I copied from you, but I guess they don't need to be queued if the "same-origin request event rules" themselves are a task (and they are for the async scenario)

[04:24:00.0000] <annevk3>
might actually simplify a few things...

[04:24:01.0000] <Hixie>
annevk3: yeah if you're in a task, and if the task completes quickly (doesn't spin the event loop or pause or anything), then you can fire events synchronously within the task

[04:25:00.0000] <Hixie>
also so long as you never execute script cross-origin

[04:26:00.0000] <annevk3>
so why are image load events queued rather than dispatched directly?

[04:26:01.0000] <annevk3>
do impls really do it that way?

[04:26:02.0000] <Hixie>
dispatched directly from what?

[04:26:03.0000] <annevk3>
dispatched directly from the fetch task if the complete condition has been met

[04:26:04.0000] <annevk3>
sorry, available condition

[04:27:00.0000] <Hixie>
i guess that could be done too

[04:27:01.0000] <annevk3>
copying that is how XHR ended up with this bug

[04:27:02.0000] <annevk3>
:/

[04:27:03.0000] <Hixie>
heh

[04:27:04.0000] <Hixie>
shouldn't copy without understanding :-)

[04:28:00.0000] <annevk3>
well, I thought I did

[04:28:01.0000] <Hixie>
:-)

[04:29:00.0000] <annevk3>
and I asked for review months ago on that change saying I was not entirely sure

[04:29:01.0000] <annevk3>
of course it happens just after we publish a Last Call WD

[04:29:02.0000] <annevk3>
/me is getting annoyed with that

[04:29:03.0000] <jgraham>
salty-horse: Yes that outliner probably is broken gsnedders has one that works

[04:30:00.0000] <gsnedders>
Well, it might work. I'm not promising anything.

[04:31:00.0000] <Hixie>
annevk3: eh, writing specs takes time. you can't rush reviews; q.f. my e-mail about hy i think web storage etc should have 6 months lc review at minimum

[04:31:01.0000] <Hixie>
annevk3: don't forget XHR was originally set to go to CR only by 2012 :-)

[04:34:00.0000] <annevk3>
:p

[04:36:00.0000] <annevk3>
Hixie, I guess

[04:36:01.0000] <annevk3>
Hixie, the change has been in for about 6 months by now though

[04:36:02.0000] <annevk3>
first TR/ draft with it was in August

[04:37:00.0000] <zcorpan_>
does anyone know a wysiwyg editor with css3 and html5 support?

[04:37:01.0000] <annevk3>
but yeah, stuff takes ages

[04:37:02.0000] <Hixie>
zcorpan_: live dom viewer? :-)

[04:38:00.0000] <zcorpan_>
heh

[04:42:00.0000] <Dashiva>
Microdata isn't ready to compete against RDFa on many levels [...] published W3C REC status

[04:42:01.0000] <Dashiva>
We can't let Microdata become a REC because it's not a REC

[04:43:00.0000] <Hixie>
yeah the arguments for removing microdata don't seem all that rational to me

[04:43:01.0000] <Hixie>
i don't really get it

[04:43:02.0000] <Hixie>
it's like people want it killed but don't dare suggest that

[04:43:03.0000] <Hixie>
which seems pretty silly

[04:43:04.0000] <Hixie>
given how many things we've killed so far in html5

[04:43:05.0000] <Hixie>
it's hardly a controversial thing to suggest

[04:45:00.0000] <zcorpan_>
hasn't that been suggested too?

[04:45:01.0000] <Hixie>
only by shelley, i think

[04:45:02.0000] <Dashiva>
They do suggest that too, but that's harder to achieve

[04:45:03.0000] <Hixie>
the arguments for dropping it seem much more rational to me

[04:45:04.0000] <zcorpan_>
maybe it's the next step of the agenda

[04:46:00.0000] <Dashiva>
Once it's a separate spec, proposal to publish as note and abandon it

[04:46:01.0000] <Dashiva>
It's surprising how many fans of the w3c process are coming out and saying "You must have implementations before LC" in various situations lately

[04:50:00.0000] <annevk3>
"whatever works"

[05:34:00.0000] <Hixie>
anyone understand what http://www.w3.org/Bugs/Public/show_bug.cgi?id=8108 is asking?

[05:36:00.0000] <annevk3>
sounds somewhat similar to the transparency for events thingie someone was asking about

[05:36:01.0000] <annevk3>
i.e. something to be addressed with extensions to pointer-events

[05:36:02.0000] <annevk3>
but it's quite vague so maybe I'm wrong :)

[05:39:00.0000] <jgraham>
It sounds like it was auto-translated

[05:40:00.0000] <jgraham>
I guess it means something about whether objects should revieve events if they are transparent

[05:44:00.0000] <WWFan>
i wish browsers would be faster with their implementations :(

[05:45:00.0000] <Hixie>
k, thanks

[05:45:01.0000] <Hixie>
WWFan: they'd be more buggy if they were faster, probably :-)

[05:52:00.0000] <Philip`>
/me wonders if Hixie could easily make his Bugzilla-posting script wrap text at the same width as Bugzilla itself wraps text, so it doesn't end up with individual words on lines by themselves

[05:52:01.0000] <Hixie>
the posting script does wrap text to the same width that bugzilla does, because it's bugzilla doing the wrapping

[05:52:02.0000] <Hixie>
i just send an unwrapped string

[05:53:00.0000] <Philip`>
Oh

[05:53:01.0000] <Philip`>
Maybe you should wrap text manually then, since Bugzilla seems to get it wrong :-)

[05:53:02.0000] <Philip`>
(Well, not manually, but in the posting script)

[05:54:00.0000] <Philip`>
It is slightly aesthetically displeasing and therefore a grave concern

[05:56:00.0000] <Hixie>
i recommend contacting the bugzilla authors or giving me perl code that, given a string $text, mutates $text to your satisfaction

[05:57:00.0000] <Hixie>
(and has no other side-effects other than CPU and memory usage)

[05:57:01.0000] <Dashiva>
Darn, I was hoping to do it on the GPU

[05:58:00.0000] <Dashiva>
Submit a comment, render the result, do graphical analysis, delete and retry if it doesn't match the requirements for an acceptable result

[05:58:01.0000] <Hixie>
i don't think i have a GPU on this machine

[06:03:00.0000] <Philip`>
use Text::Wrap;

[06:03:01.0000] <Philip`>
$Text::Wrap::columns = 79;

[06:03:02.0000] <Philip`>
# ...

[06:03:03.0000] <Philip`>
$text = wrap('', '', $text);

[06:10:00.0000] <Hixie>
i have updated the script accordingly but not tested it

[06:10:01.0000] <Hixie>
so someone should file a (real) bug to see if still works

[06:12:00.0000] <Hixie>
annevk3: i sent a comment on UM. It might be helpful for you, when you review UM, to go through the old requirements for CORS (like being able to handle very large lists of origins to say yes to, such that the header can't include a list of origins), and make sure that UM matches them.

[06:13:00.0000] <annevk3>
my understanding is that UM does not work with origins at all

[06:14:00.0000] <Philip`>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=8457 - it didn't die or anything

[06:14:01.0000] <Philip`>
so that's good

[06:15:00.0000] <Philip`>
(though that particular bug doesn't help see if it's wrapping correctly)

[06:29:00.0000] <annevk3>
my MacBook crashed...

[06:29:01.0000] <annevk3>
(again)

[06:29:02.0000] <annevk3>
(three times so far today, so much fail)

[06:32:00.0000] <annevk3>
Hixie, so yeah, I think the design is based around some concept of tokens, but the details of that have not been explained fully I think

[06:32:01.0000] <annevk3>
Hixie, but maybe I'm missing something

[07:18:00.0000] <Hixie>
annevk3: i couldn't find any tokens in their proposal, just a single on/off flag

[07:19:00.0000] <annevk3>
yup

[07:19:01.0000] <annevk3>
I guess Tyler still has the idea to transmit such data through URLs

[07:20:00.0000] <Hixie>
i guess i'll wait to see what his reply is

[07:20:01.0000] <annevk3>
afaict the proposal is just not complete enough as to how it ought to function, but hopefully a reply to your email will shed some light on that

[07:20:02.0000] <annevk3>
right :)

[07:21:00.0000] <Hixie>
this JSON thing in public-rdfa-in-xhtml-tf seems to somewhat miss the point of the object of prefixes

[07:22:00.0000] <salty-horse>
reasking from yesterday: is there an html5 way to specify that an <article> with a summarized post in a blog's main page and the permalink of the post page have the same content?

[07:22:01.0000] <Hixie>
and this thread on public-web-security that assumes that it's ok to embed arbitrary hostile css seems to be also somewhat missing the point

[07:22:02.0000] <Hixie>
salty-horse: can you elaborate?

[07:23:00.0000] <salty-horse>
I'll try rephrasing: is there a semantic way to mark a post's "read more" link in a blog's main page? a way to say that link is a permalink to the same content you see now?

[07:24:00.0000] <annevk3>
rel=bookmark?

[07:25:00.0000] <annevk3>
http://whatwg.org/html5#link-type-bookmark

[07:25:01.0000] <salty-horse>
that's it :) thanks!

[07:25:02.0000] <Hixie>
rel=bookmark is how you label a permalink, yeah

[07:26:00.0000] <salty-horse>
and it relates to the ancestor <article> which is exactly what I asked :)

[07:29:00.0000] <Hixie>
i could've sworn i'd already written the text for what 8111 is asking

[07:32:00.0000] <Hixie>
"a new version of opera was downloaded, but the update was not applied"

[07:32:01.0000] <Hixie>
well... apply it then

[07:42:00.0000] <Philip`>
Hixie: The line-wrapping seems to not be applied to messages like http://www.w3.org/Bugs/Public/show_bug.cgi?id=8110#c1 :-(

[07:43:00.0000] <Philip`>
so words like "If" and "please" still appear on lines by themselves

[07:43:01.0000] <Philip`>
at least with my particular browser and font and font size

[07:44:00.0000] <annevk3>
hmm, wfm

[07:44:01.0000] <Philip`>
(I was concerned about that rather than about the initial bug reports)

[07:44:02.0000] <Philip`>
annevk3: I guess you don't have my particular browser and font and font size :-)

[07:45:00.0000] <annevk3>
guess not

[07:45:01.0000] <Philip`>
It works for me if I zoom in or out, but not when I just look at it normally

[07:53:00.0000] <Hixie>
Philip`: that doesn't go through my script, i use safari to edit the bugs

[07:54:00.0000] <Hixie>
only comments by contributor⊙wo are scripted

[07:56:00.0000] <ttepasse>
Lines of poetry or source code I want to individually manipulate via the dom. The HTML5 way would be <pre><span>...</span><br>...</pre>?

[07:57:00.0000] <Dashiva>
If the lines are entities, sounds like a list

[07:57:01.0000] <Philip`>
Hixie: Oh, I'd assumed it was somewhat automated

[08:00:00.0000] <ttepasse>
<pre><ol> is sadly not allowed.

[08:00:01.0000] <Dashiva>
You wouldn't need <pre>

[08:01:00.0000] <Hixie>
ttepasse: lines of poetry (unless it has unusual formatting) should use <p><span>...</span><br><span>...</span></p>

[08:01:01.0000] <Hixie>
ttepasse: (the <span>s being needed only if you want to do DOM manipulation)

[08:02:00.0000] <Hixie>
ttepasse: source code would just be <pre><code><span>...</span>\n<span>...</span></code></pre> (where \n is a line break, not a literal "\n")

[08:02:01.0000] <Hixie>
(and again, <span>s here just for DOM manipulation)

[08:02:02.0000] <ttepasse>
Frustrating, but ok.

[08:02:03.0000] <Hixie>
why frustrating?

[08:03:00.0000] <ttepasse>
<line> is more expressive.

[08:03:01.0000] <Dashiva>
Not much difference between <line> and <span><br>

[08:03:02.0000] <Hixie>
if it wasn't for <br> already existing, <line> would probably be better

[08:03:03.0000] <Hixie>
but we already have <br>, so that's water under the bridge

[08:04:00.0000] <annevk3>
would <line> have white-space:nowrap ?

[08:04:01.0000] <annevk3>
or pre-line or some such

[08:04:02.0000] <Hixie>
dunno, haven't thought about it

[08:05:00.0000] <annevk3>
/me was never quite sure how <l> would work and whether it would be the quite the same as <br>

[08:06:00.0000] <ttepasse>
I'd imagine white-space: normal for poetry and white-space:pre for XHTML2s "blockcode > l"

[08:06:01.0000] <Hixie>
ok wtf. <table rules=all> has an effect on _padding_??

[08:09:00.0000] <othermaciej>
Hixie: good point re the public-web-security thread

[08:09:01.0000] <Hixie>
actually make that border-spacing

[08:09:02.0000] <othermaciej>
Hixie: you can do a lot of damage, security-wise, just by moving controls around on the page, making them invisible, injecting images or text content....

[08:10:00.0000] <Hixie>
othermaciej: yeah!

[08:10:01.0000] <Hixie>
othermaciej: if you accept arbitrary input, you have to whitelist, it's not rocket science

[08:10:02.0000] <othermaciej>
I'm not sure if they are trying to defend against deliberate self-injection of arbitrary CSS from an untrusted source, or just mitigate the damage if it occurs accidentally (on a site that successfully prevents script injection)

[08:10:03.0000] <annevk3>
the latter

[08:11:00.0000] <annevk3>
hmm, I guess it can apply to both

[08:11:01.0000] <annevk3>
but it seemed like they were walking about injection scenarios

[08:11:02.0000] <annevk3>
talking, not walking

[08:11:03.0000] <othermaciej>
annevk3, Hixie: I think UM is probably worth supporting as an alternate XHR mode (it's useful for sandboxing toolkits like Caja), but it's not at all obvious how one would apply it to video

[08:11:04.0000] <annevk3>
or <img>

[08:11:05.0000] <annevk3>
or EventSource

[08:12:00.0000] <annevk3>
etc.

[08:12:01.0000] <othermaciej>
like, if your secret token is in the page then anyone can get it using "View Source"

[08:12:02.0000] <annevk3>
but UM can be one extra flag in the current XHR spec rather than a whole separate spec

[08:12:03.0000] <Hixie>
isn't UM a subset of CORS?

[08:12:04.0000] <Hixie>
specifically, basically the subset IE implements?

[08:12:05.0000] <othermaciej>
annevk3: I don't mind if there's a separate spec for what the flag does, assuming that's extensive enough to be more than a few sentences

[08:12:06.0000] <annevk3>
you'd always exclude credentials and set Origin to null

[08:13:00.0000] <Hixie>
why would a site opt out of including Origin

[08:13:01.0000] <annevk3>
given that all the infrastructure is already in XHR/CORS it seems like a waste, but sure

[08:13:02.0000] <Hixie>
if it knows to opt out of it, why would it use it

[08:13:03.0000] <Hixie>
the point of the proposal seems to be to remove the risk that someone might use Origin at all

[08:13:04.0000] <annevk3>
well yeah, it's not clear we need it at all if we decide CORS is the way to go

[08:13:05.0000] <Hixie>
so allowing people to disable Origin is unlikely to satisfy anyone

[08:15:00.0000] <Hixie>
sweet jesus

[08:15:01.0000] <Creap>
why is <table border> still in HTML5?

[08:15:02.0000] <Hixie>
i'm 4-for-4 on uninteroperable behaviour for mixing CSS with table frames=all

[08:15:03.0000] <Hixie>
er, rules=all

[08:17:00.0000] <othermaciej>
Hixie: ability to *opt* out of Origin does serve one use case, namely allowing content sandboxed by something like Caja to still do XHR

[08:18:00.0000] <annevk3>
you would get that automatically with <iframe sandbox>, fwiw

[08:19:00.0000] <Hixie>
othermaciej: that still wouldn't be safe for same-origin requests

[08:20:00.0000] <othermaciej>
(assuming non-sandboxed content on that site does send Origin with its XHR requests)

[08:20:01.0000] <othermaciej>
(and checks it)

[08:21:00.0000] <othermaciej>
annevk3: I am not sure what problem if any people doing Caja-like things would have with <iframe sandbox> - they do dislike regular iframes since those let a bit too much power leak through (e.g. ability to navigate the top frame)

[08:21:01.0000] <Hixie>
this would also mean we should add new flags for EventSource, WebSocket, <video>, <img>, etc

[08:22:00.0000] <Hixie>
which i don't think makes sense

[08:22:01.0000] <othermaciej>
good point - the Caja model is not compatible with protecting such things purely via Origin without addition of some new flag

[08:22:02.0000] <Hixie>
i think if you're doing something like caja you have to build the entire platform from the ground up, you can't do half the job and then rely on the browser's XHR or WebSocket, etc.

[08:23:00.0000] <othermaciej>
(for EventSource or WebSocket it is plausible, at least they are APIs)

[08:23:01.0000] <othermaciej>
Caja currently presumably does not pass through EventSource or WebSocket at all

[08:24:00.0000] <othermaciej>
but if they allow <img> and <canvas>, then Caja-tized content (or whatever the term is) can read the pixels of all of the embedding site's images

[08:37:00.0000] <TabAtkins>
Creap: The same reason tons of other retarded old presentational attributes are in HTML5 - because sites use them and rely on them working.  <table border> is *not* valid HTML5, though - the spec says not to use it, but specifies how browsers should treat it when it does appear.

[08:37:01.0000] <Creap>
ok

[08:37:02.0000] <Hixie>
oh sorry creap, missed your question

[08:37:03.0000] <Hixie>
thanks tab

[08:37:04.0000] <TabAtkins>
np

[08:38:00.0000] <Creap>
I don't see how a site would rely on border for _working_ though?

[08:38:01.0000] <Creap>
border=0 for table layouts?

[08:38:02.0000] <Hixie>
TabAtkins: did you see manu's updated change proposal? not sure if there's anything that needs adding to yours, it didn't seem to make any particularly strong new points, but i might have missed something

[08:40:00.0000] <TabAtkins>
How strange.

[08:41:00.0000] <Hixie>
Creap: many sites use border, especially if they were written in the 90s

[08:44:00.0000] <TabAtkins>
Creap: More specifically, a browser which ignored <table border> entirely would render a lot of sites in a visibly uglier way than other browsers.  So they'd be at a competitive disadvantage.  HTML5 helps by defining precisely what they need to do when they see that attribute, since they'll have to handle it anyway.  This way they don't have to spend QA time and money reverse-engineering...

[08:44:01.0000] <TabAtkins>
...other browsers.

[08:59:00.0000] <Hixie_>
/me tries to remember what the devapi list is called

[09:00:00.0000] <Hixie_>
ah, public-device-apis

[09:02:00.0000] <salty-horse>
googling for "html5 outline" gives http://james.html5.org/outliner.html as the first result. since that doesn't work, maybe that page could redirect to http://gsnedders.html5.org/outliner/ ?

[09:31:00.0000] <Hixie_>
/me sends his ideas to public-device-apis to see if anyone has any suggestions for UI

[09:51:00.0000] <annevk3>
can't you just have something simple? that when you hit accept you see a little screen displaying whatever your camera aims at and maybe some audio animation for the mic

[09:51:01.0000] <annevk3>
and some visible stop button around there too

[09:51:02.0000] <annevk3>
maybe sound separately isn't even needed, it's pretty clear that it goes with video

[09:51:03.0000] <annevk3>
at least in other contexts, e.g. skype

[09:54:00.0000] <annevk3>
oh, nobody is here

[09:54:01.0000] <annevk3>
netsplits ftw

[10:12:00.0000] <annevk3>
this split sucks

[10:13:00.0000] <annevk3>
I guess I should go have dinner instead of complaining

[10:24:00.0000] <annevk3>
Hixie_, see logs for some thoughts

[10:24:01.0000] <annevk3>
netsplit ftl

[10:24:02.0000] <Hixie_>
looking...

[10:25:00.0000] <Hixie_>
the problem is the "hit accept" part

[10:26:00.0000] <Hixie_>
also, you don't want a separate monitor window when you're doing video chat

[10:26:01.0000] <Hixie_>
you want your camera to be shown p-i-p in the incoming video's output

[10:26:02.0000] <Hixie_>
which is trivial with css and <video>

[10:27:00.0000] <Hixie_>
the problem is how to do the bit where the page asks for video, and the user accepts it

[10:27:01.0000] <Hixie_>
(or alternatively, the bit where the user tells the page he's ready to use video)

[10:29:00.0000] <TabAtkins>
Hrm.  Does <video> allow different-origin @src to be used?

[10:30:00.0000] <TabAtkins>
My skim of that section doesn't see anything disallowing it.

[10:30:01.0000] <annevk3>
p-i-p?

[10:30:02.0000] <TabAtkins>
picture-in-picture

[10:30:03.0000] <Hixie_>
TabAtkins: yes, but we'll have to limit that to no-metadata-only later when we add metadata reading

[10:31:00.0000] <annevk3>
hmm, I suppose

[10:31:01.0000] <TabAtkins>
Hixie: That sucks.  That means anyone can drain my bandwidth by hotlinking my videos.  ;_;

[10:31:02.0000] <TabAtkins>
Image hotlinking is bad enough already.

[10:32:00.0000] <annevk3>
have there been issues with what Flash is using?

[10:32:01.0000] <Hixie_>
TabAtkins: they can anyway by using iframe

[10:32:02.0000] <TabAtkins>
I dunno what flash is using.

[10:32:03.0000] <Hixie_>
annevk3: flash uses a modal prompt, i believe.

[10:32:04.0000] <TabAtkins>
Hixie: Yeah, but then I have to have a page amenable to that, it's not as pretty, and it's more effort.

[10:33:00.0000] <Hixie_>
hm?

[10:33:01.0000] <TabAtkins>
More effort for the hotlinker, that is.

[10:33:02.0000] <TabAtkins>
While just doing <video src="xanthir.com/coolvid"> is trivial.

[10:34:00.0000] <AryehGregor>
Also will probably break, unless http:// became optional and I didn't notice.

[10:34:01.0000] <AryehGregor>
(is it optional?)

[10:34:02.0000] <TabAtkins>
Gah, technicalities.

[10:34:03.0000] <TabAtkins>
No, I'd need //xanthir.com

[10:34:04.0000] <AryehGregor>
It can't be, that's legitimately baseURL/xanthir.com/...

[10:35:00.0000] <Hixie_>
TabAtkins: well when we add teh cross-origin protection for metadata, we'll also add the ability to block cross-origin loads altgoether

[10:35:01.0000] <AryehGregor>
Which Firefox already does, right?

[10:35:02.0000] <Hixie_>
TabAtkins: but i thought we should wait for anne to go through the pain of figuring out how cors is to work before using it

[10:36:00.0000] <TabAtkins>
Hrm.  Is it appropriate to make shutting down linking a proactive choice?  CORS is all about doing the opposite.

[10:37:00.0000] <othermaciej>
TabAtkins: will you have a chance to update your counter-proposal soon? Would Thursday be a reasonable deadline?

[10:38:00.0000] <TabAtkins>
othermaciej: Yeah, I'm planning to do it at lunch today.  Meant to do it yesterday, but ended up on the phone with shepazu instead.

[10:38:01.0000] <othermaciej>
wfm

[10:42:00.0000] <annevk3>
Hixie_, blocking altogether is not a feature of CORS

[10:42:01.0000] <annevk3>
I'm not all a fan of the @font-face crowd abusing it for that

[10:42:02.0000] <annevk3>
but maybe that's just me

[10:42:03.0000] <Hixie_>
sure it is, if the resource's allowed origin doesn't match the actual origin, that blocks it

[10:43:00.0000] <TabAtkins>
That's a same-origin restriction, and has nothing to do with CORS.

[10:43:01.0000] <annevk3>
it would just block for the purposes of sharing metadata

[10:43:02.0000] <TabAtkins>
CORS then just lifts that restriction.

[10:43:03.0000] <annevk3>
the idea is that CORS only lifts restrictions

[10:43:04.0000] <Hixie_>
TabAtkins: right, but if you start with the assumption that the default is not to block but to allow playback, just with metadata disabled, and then you apply CORS, you get what i described

[10:44:00.0000] <annevk3>
no you don't

[10:44:01.0000] <annevk3>
CORS applied on cross-origin images would not suddenly stop displaying external images

[10:44:02.0000] <Hixie_>
othermaciej: any idea if window.iTunes.waveform turns out to have been a good idea or not? (seems few if any LPs actually use it, is that just because they didn't want to, or does it have issues?)

[10:44:03.0000] <annevk3>
well, we could do that, but that was not the intent

[10:44:04.0000] <Hixie_>
annevk3: that depends on what the spec that applies CORS says

[10:44:05.0000] <TabAtkins>
CORS on external images would do basically nothing, right, anne?

[10:45:00.0000] <TabAtkins>
Since they already fully share themselves cross-origin?

[10:45:01.0000] <annevk3>
TabAtkins, it would stop tainting canvas

[10:45:02.0000] <Hixie_>
annevk3: anyway, this is all academic; eventually we'll need this three-mode thing, but not yet

[10:45:03.0000] <TabAtkins>
Ah, yes.

[10:45:04.0000] <Hixie_>
and then we can either use CORS, or modify CORS, or use something else, or whatever

[10:46:00.0000] <othermaciej>
Hixie_: I'm not sure what that is

[10:47:00.0000] <annevk3>
it seems I was right that UM requires a redesign of every feature

[10:47:01.0000] <othermaciej>
Hixie_: is that some extension to visualize the currently playing song?

[10:47:02.0000] <annevk3>
fun times

[10:47:03.0000] <annevk3>
really dinnertime now :)

[10:47:04.0000] <othermaciej>
Hixie_: I can ask some of the people who worked on that stuff

[10:47:05.0000] <Hixie_>
othermaciej: it exposes the amplitude of the current playback audio, split into 128 buckets, as far as i can tell

[10:48:00.0000] <Hixie_>
othermaciej: it would be useful to know what their experience was, in case we should just add it to HTMLMediaElement

[10:48:01.0000] <othermaciej>
Hixie_: amplitude, not FFT frequency analysis?

[10:48:02.0000] <Hixie_>
dunno

[10:48:03.0000] <Hixie_>
i assume it's the magnitude of each of 128 frequency ranges

[10:49:00.0000] <Hixie_>
but really no idea

[10:50:00.0000] <othermaciej>
I see, that would make sense as data for a visualizer

[12:19:00.0000] <zcorpan_>
Philip`: could you add Excalibur Logotype to web font optimizer?

[12:25:00.0000] <Philip`>
zcorpan_: No, since it doesn't seem to claim to have a license that allows modification and distribution

[12:25:01.0000] <Philip`>
(It doesn't actually seem to claim to have any license at all, as far as I can tell)

[12:26:00.0000] <zcorpan_>
bummer

[12:26:01.0000] <Philip`>
You could use the offline http://bitbucket.org/philip/font-optimizer/src/ tool though

[12:27:00.0000] <Philip`>
which should be easy to use (in theory) and means the licensing is entirely your responsibility

[12:29:00.0000] <zcorpan_>
too lazy at the moment :)

[13:31:00.0000] <TabAtkins>
I know it's invalid, but setting a boolean attribute to something other than "" or the element's name still sets it, correct?

[13:31:01.0000] <TabAtkins>
So, frex, <script async=true> would still make an async script?

[13:31:02.0000] <Philip`>
Yes

[13:31:03.0000] <Philip`>
All that matters is whether the attribute is present

[13:32:00.0000] <TabAtkins>
K, just making sure.  All the invalid-but-specified things are always tucked away in places that I can't find easily.

[13:32:01.0000] <Philip`>
(<script async=false> would make it async too, which is a reason to not encourage use of async=true)

[13:33:00.0000] <Philip`>
http://www.whatwg.org/specs/web-apps/current-work/multipage/common-microsyntaxes.html#boolean-attributes

[13:33:01.0000] <Philip`>
That covers the invalid cases

[13:35:00.0000] <TabAtkins>
The previous paragraph should make it clear, but it's easy to be paranoid.

[14:24:00.0000] <Sirisian_>
curious is rounded corners part of HTML5? I don't get why moz and webkit have their own style tags.

[14:25:00.0000] <Hixie_>
that's css

[14:25:01.0000] <Sirisian_>
ah yes

[14:25:02.0000] <Sirisian_>
whoops

[14:53:00.0000] <annevk3>
TabAtkins, you should point out that the IDL attribute is a boolean and so technically should not be set to a string

[14:54:00.0000] <annevk3>
TabAtkins, so script.async = true is perfectly fine

[14:54:01.0000] <TabAtkins>
I don't actually know that, so I won't point it out.

[14:54:02.0000] <annevk3>
and script.async = 'async' is not

[14:54:03.0000] <annevk3>
but script.setAttribute("async", "async") is

[14:54:04.0000] <annevk3>
etc.

[14:54:05.0000] <annevk3>
it seems the commenter is confused on that

[14:55:00.0000] <TabAtkins>
Nod, I am too.  jQuery abstracts it enough that I don't have to care about things.

[14:55:01.0000] <annevk3>
I see...

[14:55:02.0000] <TabAtkins>
So the idl attribute *can* be unset with =false?  Or no?

[14:56:00.0000] <annevk3>
it can, yes

[14:56:01.0000] <annevk3>
it's a boolean

[14:56:02.0000] <TabAtkins>
kk

[14:56:03.0000] <TabAtkins>
DOM is confusing.  >_<

[14:56:04.0000] <annevk3>
see the HTMLScriptElement interface

[14:56:05.0000] <annevk3>
it's very simple

[14:57:00.0000] <annevk3>
setting the IDL attribute to false will in fact remove the content attribute

[14:57:01.0000] <TabAtkins>
So those are the names I should use when talking about this?  "IDL attribute" and "content attribute"?

[14:58:00.0000] <annevk3>
it's what HTML5 uses

[14:58:01.0000] <TabAtkins>
kk

[14:58:02.0000] <annevk3>
you might see DOM attribute being referred to as property as well because that's what it is in ECMAScript, but that's somewhat confusing given CSS

[14:59:00.0000] <TabAtkins>
"DOM attribute" = "IDL attribute"?

[14:59:01.0000] <jgraham>
me does't think "very simple" is really true

[15:00:00.0000] <jgraham>
When script.async = false does exactly the opposite to script.setAttribute("async", false)

[15:00:01.0000] <annevk3>
TabAtkins, ah sorry, used to be DOM attribute, now is IDL attribute

[15:01:00.0000] <annevk3>
jgraham, they're completely different things

[15:01:01.0000] <jgraham>
annevk3: Yes. That's why it's not "very simple"

[15:05:00.0000] <annevk3>
mu

[15:09:00.0000] <annevk3>
(hopefully mu works for DOM arguments too)

[15:09:01.0000] <annevk3>
(maybe it should be "du")

[15:11:00.0000] <TabAtkins>
script.async="false" would still make it true, right?  (Almost certain of this, but want to double-check before I send the email.)

[15:11:01.0000] <Philip`>
That sounds like an issue for WebIDL

[15:11:02.0000] <annevk3>
that's because a non-zero string evaluates to true in ECMAScript

[15:12:00.0000] <TabAtkins>
That's what I thought.

[15:14:00.0000] <Philip`>
http://dev.w3.org/2006/webapi/WebIDL/#es-boolean

[15:14:01.0000] <Philip`>
so it just defers to ES's ToBoolean

[15:15:00.0000] <Philip`>
(which does what annevk3 says)

[15:16:00.0000] <annevk3>
annevk3 is awesome

[15:26:00.0000] <annevk2>
hmm

[15:27:00.0000] <annevk2>
it's prolly no longer worth pointing out that SGML allowed omission of the attribute if the value itself gave enough context...

[15:39:00.0000] <annevk2>
I've learned history is important, but I always wonder if knowing (some of) the technical details of SGML is really helping me...

[15:40:00.0000] <Philip`>
Is your life not immeasurably enriched by knowing it's the attribute name rather than the value that is optional?

[15:41:00.0000] <annevk2>
So far I've not experienced that unfortunately

[15:41:01.0000] <annevk2>
Hopefully that is yet to come

[15:48:00.0000] <jwalden>
𝅘𝅥𝅮 the best is yet to come 𝅘𝅥𝅮

[15:52:00.0000] <annevk2>
Track 18 of Metal Gear Solid? :)

[15:53:00.0000] <annevk2>
(see also http://en.wikipedia.org/wiki/Metal_Gear_Solid_Original_Game_Soundtrack for those who do not have the CD)

[15:54:00.0000] <annevk2>
(Wikipedia is awesome)


2009-12-09
[16:00:00.0000] <TabAtkins>
I need those soundtracks.  MGS music is awesome.

[18:38:00.0000] <Dashiva>
othermaciej: "We will not do any numerical counting all the votes." ... "To the extent we apply any numerical analysis to the results"

[18:40:00.0000] <othermaciej>
Dashiva: belt and suspenders!

[18:40:01.0000] <othermaciej>
or were you pointing out my typo?

[18:40:02.0000] <Dashiva>
No, just wondering which one is closer to reality :)

[18:41:00.0000] <othermaciej>
the former

[18:42:00.0000] <othermaciej>
the latter I stated in case anyone gets worked up about, say, all 10 of the opera reps responding

[18:43:00.0000] <Dashiva>
The vast Opera-wing conspiracy

[18:44:00.0000] <Dashiva>
But okay, answer accepted

[20:42:00.0000] <erlehmann>
tach

[21:54:00.0000] <bfulgham_>
Does anyone know if there is a proposal for animating transitions on visibility so that 'hidden' could accordion the surround blocks?

[21:55:00.0000] <bfulgham_>
Or rather, animating on "display" to allow accordion or other effects?

[22:08:00.0000] <smaug>
bfulgham_: you're talking about css transitions? That discussion should happen on www-style⊙wo

[22:08:01.0000] <bfulgham_>
smaug: thanks

[00:55:00.0000] <Philip`>
http://software.intel.com/en-us/articles/xml-parsing-accelerator-with-intel-streaming-simd-extensions-4-intel-sse4/ - intriguing

[00:55:01.0000] <Philip`>
html5lib would be much faster if it was rewritten to use SSE assembly instead of Python

[00:57:00.0000] <annevk2>
do it!

[00:57:01.0000] <annevk2>
the problem is of course that the code will only run on a proprietary processor

[00:58:00.0000] <Philip`>
/me wonders what a non-proprietary processor would be

[00:59:00.0000] <annevk2>
one with a standardized programming API?

[00:59:01.0000] <annevk2>
I suppose it might still be proprietary in that case

[01:00:00.0000] <Philip`>
Is there such a thing that exists today and is used?

[01:01:00.0000] <Philip`>
Everyone just uses x86 and copies Intel (except occasionally when Intel copies AMD) so it seems the closest thing to a standard :-)

[01:15:00.0000] <Hixie>
http://www.google.com/search?hl=en&safe=off&esrch=RTSearch&gl=us&tbo=1&tbs=mbl:1&q=html5&aq=f&oq=&aqi=g10

[01:15:01.0000] <Hixie>
woo

[02:44:00.0000] <annevk2>
Hixie, so you wanna combine the UI for streaming video/audio with geolocation, etc.?

[02:44:01.0000] <annevk2>
and most other things pages can ask for going forward, such as e.g. digital camera harddrive

[02:44:02.0000] <Hixie>
only if that is the best solution

[02:45:00.0000] <Hixie>
i don't know what the best solution is

[02:45:01.0000] <annevk2>
how would you combine that with the existing API?

[02:45:02.0000] <annevk2>
yeah I guess

[02:45:03.0000] <annevk2>
meh

[02:45:04.0000] <Hixie>
i wasn't really thinking of geolocation

[02:45:05.0000] <Hixie>
did tab update his proposal?

[02:46:00.0000] <annevk2>
for microdata? haven't seen anything

[02:48:00.0000] <nessy>
ohh! microdata! I wonder how that combines with video metadata

[03:45:00.0000] <Hixie>
what time is the htmlwg telecon these days?

[03:47:00.0000] <AryehGregor>
How is the telecon conducted?  IRC?

[03:47:01.0000] <jgraham>
Telephone

[03:47:02.0000] <AryehGregor>
Amazing.

[03:47:03.0000] <jgraham>
Hence the "tele"

[03:47:04.0000] <AryehGregor>
"tele" means "distant".

[03:47:05.0000] <AryehGregor>
You know, like "telecommunications", and "teleport".

[03:47:06.0000] <jgraham>
Only in greek (or something)

[03:48:00.0000] <AryehGregor>
Okay, anyway.  I guess Hixie will have to borrow someone's phone.  :)

[03:48:01.0000] <gsnedders>
television is vision over telephone?

[03:48:02.0000] <jgraham>
Gah

[03:48:03.0000] <jgraham>
Hixie: 6pm CET on a Thursday 

[03:48:04.0000] <AryehGregor>
Telepathy is detecting thoughts via telephone.

[03:48:05.0000] <Hixie>
jgraham: thanks

[03:48:06.0000] <Hixie>
/me changes to CET

[03:49:00.0000] <jgraham>
/me has never heard of a telecon that was not conducted by telephone

[03:49:01.0000] <Hixie>
/me changes back to PDT

[03:50:00.0000] <Philip`>
It's 1700Z according to two out of three repetitions of the time in the last reminder email

[03:51:00.0000] <Hixie>
figured i should at least be aware that i'm missing it this week since it's plausible i'll be awake at that time

[03:51:01.0000] <Hixie>
(though not near a phone unless i get some meeting room booked)

[03:52:00.0000] <AryehGregor>
Teleconferences can also be by video.

[03:53:00.0000] <Hixie>
we call those video conferences

[03:53:01.0000] <Hixie>
and they're much better than telecons though still almost as much of a waste of time

[03:56:00.0000] <AryehGregor>
What's wrong with IRC?

[03:57:00.0000] <Hixie>
or e-mail?

[03:57:01.0000] <Hixie>
beats me

[03:57:02.0000] <AryehGregor>
E-mail is a lot slower, it has definite disadvantages.

[03:57:03.0000] <AryehGregor>
Although also advantages.

[03:57:04.0000] <Hixie>
e-mail is far faster at making progress than telecons

[03:57:05.0000] <Hixie>
than scheduled telecons, i should say

[03:58:00.0000] <Hixie>
scheduled telecons cause working groups to only work immediately before and during a telecon

[03:58:01.0000] <Hixie>
so it becomes impossible for the group to make decisions except during the meeting

[03:58:02.0000] <Hixie>
which is ridiculous

[03:58:03.0000] <Hixie>
you can see this e.g. from the way the chairs of the htmlwg are only able to make decisions on tuesdays

[03:58:04.0000] <Hixie>
scheduled irc meetings would have the same problem

[03:59:00.0000] <AryehGregor>
Isn't it an HTMLWG policy that decisions are only made by e-mail and such?  There's something about allowing asynchronous communication.

[03:59:01.0000] <Hixie>
yes, but subgroups (like the chairs) can arrange for their own decisions to be synchronous if they want

[04:00:00.0000] <Hixie>
so long as membership in the group isn't a requirement to influence decisions

[04:02:00.0000] <AryehGregor>
/me is glad to see that there will be no actual votes, just decisions by a cabal of three chairs instead of one editor

[04:03:00.0000] <AryehGregor>
Although theoretically the decision will be what causes the least strong objections instead of the one they think is technically best, so it's somewhat different.

[04:05:00.0000] <Philip`>
I guess "least strong objections" means you should threaten to shout loudly and file FOs etc if you don't get your way, so that your objections carry more weight

[04:06:00.0000] <Hixie>
yeah i'm curious to know how they intend to weigh rationales if not in the same way that i did when examining the issue as the editor

[04:06:01.0000] <AryehGregor>
Yes, that sounds like the strategically appropriate response to such a policy.

[04:06:02.0000] <AryehGregor>
Presumably they're allowed to ignore you if they think you're faking, though.

[04:07:00.0000] <jgraham>
Hixie: It is not impossible that they will weight rationales in the same way that you did but come to a different conclusion

[04:07:01.0000] <AryehGregor>
Honestly, no one can really fault them too much no matter what decision they make on something like microdata, since obviously they can't please everyone.

[04:07:02.0000] <jgraham>
But I agree that is not what they have stated they will do

[04:08:00.0000] <Hixie>
surely if they get a different conclusion then by definition they've used a different method of weighing arguments :-)

[04:08:01.0000] <Hixie>
i'm not saying there's anything wrong with that

[04:08:02.0000] <Hixie>
i'm just wondering what their method will be

[04:08:03.0000] <AryehGregor>
Even if they secretly just pick whatever solution they like, people will probably still end up being more satisfied because they were told it's a compromise rather than "I think you're wrong, too bad".

[04:09:00.0000] <Hixie>
(in fact i'd say that they _should_ use a different method, since otherwise the whole escalation process is pointless since they'd always agree with me)

[04:09:01.0000] <AryehGregor>
It wouldn't be pointless politically even if they always agreed with you, only technically.

[04:10:00.0000] <Hixie>
technically is what i care about

[04:10:01.0000] <AryehGregor>
Then why are you having anything to do with the W3C?  :)

[04:10:02.0000] <jgraham>
Hixie: Well they can use the same _method_ but get different results due to different inputs

[04:10:03.0000] <Hixie>
AryehGregor: patent policy and to get microsoft's feedback

[04:10:04.0000] <Hixie>
AryehGregor: in retrospect, as i've said before, i think we would have been better off just getting our own patent policy

[04:11:00.0000] <jgraham>
e.g. you might tend to weigh arguments that correspond with your own experience more highly

[04:11:01.0000] <AryehGregor>
Well, but that's basically political.

[04:11:02.0000] <AryehGregor>
I'd think the WHATWG with its own patent policy would be at a significant disadvantage without Microsoft on board.

[04:11:03.0000] <Philip`>
The three chairs should meet on a wind-swept heath and drop the arguments into a cauldron and then dance around it until a conclusion reveals itself to them

[04:12:00.0000] <Hixie>
jgraham: i guess if you don't consider the weighing function part of the weighing function... :-)

[04:12:01.0000] <Hixie>
AryehGregor: in practice we've gotten very little useful input (or indeed any input) from microsoft

[04:12:02.0000] <AryehGregor>
Yes, but they at least claim they'll implement HTML5 at some point, which is something.

[04:13:00.0000] <AryehGregor>
I really meant from an implementation perspective, not feedback.

[04:13:01.0000] <AryehGregor>
It's not much good if Microsoft doesn't implement it.

[04:13:02.0000] <Hixie>
i don't think the w3c affected that really

[04:13:03.0000] <AryehGregor>
Maybe not.  Too late now, I guess.

[04:14:00.0000] <AryehGregor>
Hixie, please tell your superiors at Google that they need to kill Microsoft faster, thanks.

[04:14:01.0000] <jgraham>
Hixie: The exact weighing function isn't exactly part of the "method"

[04:14:02.0000] <Hixie>
google has no intention of killing microsoft, we actually wish microsoft would compete better to give us more of a run for our money

[04:14:03.0000] <Hixie>
but that's another story

[04:15:00.0000] <Hixie>
jgraham: fair enough

[04:15:01.0000] <Philip`>
/me imagines a Google monopoly would be about as bad as a Microsoft monopoly

[04:15:02.0000] <gsnedders>
But they "do no evil"! :P

[04:15:03.0000] <AryehGregor>
Philip`, you mean like all the terrible effects we've been seeing from their monopoly on search?

[04:16:00.0000] <jgraham>
Google monopoly? Do you go round a little board and buy up different websites and build adverts on them to collect money?

[04:16:01.0000] <AryehGregor>
Slightly lower market share than MS in the OS market, but still.

[04:16:02.0000] <Hixie>
AryehGregor: google has nowhere near a monopoly on search

[04:17:00.0000] <Philip`>
Google doesn't seem to have shown any reluctance to develop and deploy their own private web technologies that are largely focused on the needs of other parts of Google, without much input from other vendors

[04:18:00.0000] <AryehGregor>
Well, that's called "not being a bunch of hippie open-source freaks", not "being monopolistic".  :)

[04:19:00.0000] <AryehGregor>
As long as they don't try to lock you in to anything.

[04:19:01.0000] <AryehGregor>
Which generally they don't.

[04:19:02.0000] <AryehGregor>
There's some tie-in between their services, but usually pretty weak.

[04:19:03.0000] <AryehGregor>
Okay, I was supposed to leave ten minutes ago, bye.

[04:19:04.0000] <Hixie>
later

[04:40:00.0000] <boblet>
interesting—HTML5-style charset declaration isn’t stronger than user-declared character encoding in FF3.5.5, but http-equiv style one is

[04:41:00.0000] <Hixie>
yeah okcool.de has a comment to that effect

[04:42:00.0000] <Hixie>
i haven't been able to get more info on it -- do you have test cases showing this?

[04:43:00.0000] <boblet>
Hixie: is that to me? I’ll up one…

[04:43:01.0000] <Hixie>
yes :-)

[05:19:00.0000] <boblet>
Hixie: crap—testing methodology fail. Sorry. Here’s the question that made me check it: http://doctype.com/doesnt-html-5-meta-charset-tag-work

[05:19:01.0000] <boblet>
Hixie: I couldn’t reproduce the error

[05:29:00.0000] <boblet>
Can <nav> be used for site search or pre/next page links? I think no, but heard yes for search…

[05:31:00.0000] <boblet>
(I’ve also seen “read more…” links marked as <nav> btw ;-)

[05:32:00.0000] <daedb>
I can see <nav> being used for prev/next page links, but I wouldn't use it for search or read more links...

[05:53:00.0000] <Philip`>
boblet: Seems odd

[05:53:01.0000] <Philip`>
Could be his HTML/text editor detects the meta http-equiv and saves as UTF-8 in that case, and ISO-8859-1 otherwise, or something

[05:54:00.0000] <Philip`>
/me doesn't know what else it would be, especially since he says he's setting the HTTP Content-Type charset too

[05:56:00.0000] <Philip`>
boblet: I hope <nav> is okay for next/previous, since the multipage HTML5 spec uses it for that

[06:12:00.0000] <boblet>
Philip`: yeah that Doctype.com q is peculiar. hopefully I’ll hear back

[06:16:00.0000] <boblet>
Philip`: Looking at the HTML5 multipage spec, next/prev page links are part of a larger nav block. Do you think it’d apply on eg the meter/noscript links on http://dev.w3.org/html5/markup/nav.html ?

[06:17:00.0000] <boblet>
Philip`: Mike’s used div.nav span.nav-prev and span.nav-next

[06:19:00.0000] <daedb>
boblet: I'd probably use <nav> for those links...

[06:20:00.0000] <Hixie>
boblet: weird. there's a comment in the source of okcool.de that talks about this also, and i can't get anyone to explain to me why.

[06:21:00.0000] <Philip`>
boblet: I imagine he used that because the W3C pubrules don't like HTML5 markup

[06:23:00.0000] <Hixie>
who's doing webidl these days?

[06:37:00.0000] <Hixie>
indeed, is there a bug tracker for webidl?

[06:37:01.0000] <boblet>
Hixie: re: charset, maybe it’s just a problem somewhere else (server header, tools…)

[06:37:02.0000] <Hixie>
maybe...

[06:38:00.0000] <boblet>
Philip`: ok that makes sense

[06:38:01.0000] <Hixie>
hmm, webidl has a bug component in bugzilla, but not bugs

[06:38:02.0000] <boblet>
I’m surprised that <nav> applies to prev/next links, as they’re often just a single link (not a block), and there’s already rel="next/prev"

[06:39:00.0000] <Philip`>
Judging by the relevant code (around http://mxr.mozilla.org/mozilla1.9.1/source/intl/chardet/src/nsMetaCharsetObserver.cpp#319 (for Firefox 3.5)), it looks like http-equiv charset and meta charset are reported indistinguishably, so it shouldn't be possible for one to work and not the other

[06:39:01.0000] <boblet>
Hixie: would you use <nav> for site search form?

[06:40:00.0000] <Hixie>
would you have a "skip navigation" link for a site search form?

[06:40:01.0000] <Hixie>
Philip`: yeah, i never found any distinctions when doing the testing way back when

[06:41:00.0000] <Philip`>
(That does indicate <meta charset=utf8 foo=bar won't work, though)

[06:41:01.0000] <Philip`>
Uh, <meta charset=utf8 foo=bar>

[06:45:00.0000] <Hixie>
oh?

[06:46:00.0000] <Philip`>
Hmm, but that's not what happens in practice

[06:47:00.0000] <Philip`>
(The GetCharsetFromCompatibilityTag is only in the 'else' of a thing that checks for the number of attributes)

[06:51:00.0000] <Philip`>
Maybe that's not the code that's actually used

[06:51:01.0000] <boblet>
Hixie: I wouldn’t, but I can see some would. Yeah, that’s the nub

[06:52:00.0000] <Hixie>
boblet: i think it'd be fine to use <nav> for that, but also fine not to

[06:53:00.0000] <Hixie>
basically, don't use <nav> unless you think <section> would also be appropriate, with an <h1>Navigation</h1>.

[06:53:01.0000] <boblet>
Hixie: I really love that a lot of the structural element usage is almost up to authors. That’s what most people hate, of course, but still :)

[06:53:02.0000] <Hixie>
it'd be extremely difficult, and only mildly useful, to be more specific than we already are

[06:53:03.0000] <Philip`>
Whatever code Firefox 3.5 is using, it also accepts things like <meta notcharset="utf-8">

[06:53:04.0000] <Hixie>
and we're already waaaay more specific than html4 ever was

[06:54:00.0000] <Philip`>
so it can't just be the attribute-based extracter

[06:54:01.0000] <boblet>
makes it harder to write best practice info, but it’s much more interesting

[06:55:00.0000] <Hixie>
best practice is "don't use <div> and don't use class=''", basically

[06:55:01.0000] <Hixie>
but often you have to use one or the other

[06:55:02.0000] <Hixie>
since html isn't _that_ expressive, even with all the new stuff

[06:56:00.0000] <boblet>
Hixie: why not class? that seems to becoming a best practice with eg @stubbornella’s CSS performance info

[06:58:00.0000] <boblet>
I’ve actually started using her “h2 .h2 {}” style classes, then in html <section><h1 class="h2">

[06:59:00.0000] <Hixie>
o_O

[06:59:01.0000] <Hixie>
why would you do that

[07:00:00.0000] <boblet>
not as pure HTML but better than “section section h1, article section h1 {}”

[07:00:01.0000] <Hixie>
oh well, yes, we need to figure out a better css selector solution for nested section headers

[07:00:02.0000] <Philip`>
boblet: Why not use <h2>?

[07:00:03.0000] <Hixie>
i meant theoretical best practice in the future, not best practice now

[07:01:00.0000] <boblet>
also I think class="h2" has semantic meaning

[07:01:01.0000] <Hixie>
/me wonders when breakfast service is _supposed_ to start at his cafe

[07:01:02.0000] <boblet>
Philip`: HTML5-style headings—reset to <h1> with each section

[07:01:03.0000] <Hixie>
no, you can use whatever level you want

[07:01:04.0000] <Hixie>
<section><h2></h2></section> and <section><h1></h1></section> mean the same thing

[07:02:00.0000] <boblet>
well, I think it was maintain correct heading levels taking nesting into account, or reset to <h1> each time you start a new section, no?

[07:02:01.0000] <Hixie>
aha, 8am

[07:02:02.0000] <Hixie>
another hour

[07:02:03.0000] <Hixie>
no, it resets inside each <section>

[07:03:00.0000] <Hixie>
the highest <hx> within each section (er, lowest, i guess, closest to <h1>) is the heading

[07:03:01.0000] <boblet>
but not to <h1>? interesting

[07:03:02.0000] <Hixie>
well the best practice is you should have only one <hx> per <Section>

[07:03:03.0000] <Hixie>
so it doesn't make any difference

[07:03:04.0000] <boblet>
so you could skip levels, eg <h1></h1> <section><h3></h3>…

[07:04:00.0000] <Hixie>
you can use    <body><h6></h6><section><h3></h3><nav><h1></h1></nav></section></body>   means the same as   <body><h2></h2><section><h4></h4><nav><h6></h6></nav></section></body>

[07:04:01.0000] <Hixie>
...which means the same as   <body><h1></h1><section><h1></h1><nav><h1></h1></nav></section></body>

[07:04:02.0000] <Hixie>
each section does its own heading hierarchy

[07:05:00.0000] <boblet>
but wouldn’t that mean that it’s easier to just start at <h1> for each new section?

[07:05:01.0000] <boblet>
assuming you’re doing crazy stuff like applying styles via classes? :)

[07:05:02.0000] <Hixie>
apparently not, since you use classes :-)

[07:05:03.0000] <boblet>
haha

[07:05:04.0000] <Hixie>
<h1 class="h2"> is not as easy as <h2> :-)

[07:06:00.0000] <boblet>
hrm… I must ponder this more :) at this rate I’ll never release this site

[07:07:00.0000] <workmad3>
and even though <body><h1></h1><section><h2></h2></section></body> means the same thing no matter what level of heading you use to a machine, there are still ways of doing it that are clearer to people :)

[07:07:01.0000] <Hixie>
indeed

[07:08:00.0000] <workmad3>
and clearest to a human is probably either <h1> all the time or <hx> appropriate to the sectioning level...

[07:08:01.0000] <Hixie>
agreed

[07:08:02.0000] <workmad3>
(which I believe are the two examples given in the spec :) )

[07:09:00.0000] <boblet>
You might want to check out http://j.mp/6FMCDi — @stubbornella’s OOCSS presentation

[07:10:00.0000] <Hixie>
will do

[07:10:01.0000] <boblet>
it’s quite a different way to approach things. It’s definitely changed some of my perceptions

[07:11:00.0000] <boblet>
workmad3: I guess <h1 class="h2"> is just the first one with a cop-out for easier CSS selector writing

[07:11:01.0000] <boblet>
(at least, that’s what I thought)

[07:11:02.0000] <workmad3>
boblet: I personally see that as an abomination :)

[07:12:00.0000] <boblet>
hahaha

[07:12:01.0000] <workmad3>
and besides, why use a <h1> there... it's *much* simpler to use a <div>... so then you have <div class="h1">, <div class="em">, <div class="font">...

[07:13:00.0000] <boblet>
I can definitely see how on a large popular site writing long CSS selector chains would gradually kill you tho

[07:13:01.0000] <Hixie>
woah, that suggestion on slide 68 (the <h1 class="h2"> thing) is crazy

[07:13:02.0000] <Hixie>
if you're finding that you're applying h2 styles to an h3, then your semantics are wrong

[07:13:03.0000] <boblet>
Hixie: you’re killing me here

[07:13:04.0000] <boblet>
:D

[07:13:05.0000] <Hixie>
the rest of it so far is good

[07:14:00.0000] <Hixie>
i also disagree with slide 92

[07:14:01.0000] <boblet>
I think it‘s to address the ballooning of selectors problem; like specifying each case of clearfix rather than applying a .clearfix or .group style

[07:15:00.0000] <Hixie>
i think styling elements is the way to go, and styling classes should be the exception

[07:15:01.0000] <Hixie>
but that might just be that i think you should define defaults first, which seems to be the same, but said differently

[07:15:02.0000] <boblet>
well, she does say “(unless defining defaults)”

[07:17:00.0000] <Hixie>
yeah

[07:17:01.0000] <Hixie>
i strongly agree with what she says at the end

[07:17:02.0000] <Hixie>
i don't understand why we don't have better tools

[07:18:00.0000] <boblet>
@stubbornella addresses that too http://j.mp/7i7Dfe — CSS Wish List

[07:19:00.0000] <boblet>
adding programatic concepts like mixins to CSS

[07:19:01.0000] <workmad3>
boblet: check out sass :)

[07:20:00.0000] <boblet>
haha—planning to do so for this site (have it open in a tab somewhere here)

[07:20:01.0000] <workmad3>
sass does have mixins... and the ability to do maths, etc. in css files

[07:21:00.0000] <workmad3>
and it gets compiled down to normal CSS and can even be minimised, etc. :)

[07:22:00.0000] <boblet>
workmad3: yeah I’m looking forward to checking out stylesheet size differences (why I didn’t start with it)

[07:22:01.0000] <boblet>
ok thanks for the food for thought all. Bed calls

[07:22:02.0000] <workmad3>
I need chocolate and coffee myself

[07:22:03.0000] <workmad3>
but then it's only 3:30 p.m. for me :)

[07:23:00.0000] <boblet>
Will hopefully have an HTML5 site with the HTML5 articles I’ve written so far on it tomorrow to announce

[07:23:01.0000] <boblet>
later

[07:26:00.0000] <ray>
h1 class="h2"?

[07:30:00.0000] <miketaylr>
<div class="span">

[07:30:01.0000] <ray>
html5 has a way to set the charset besides meta http-equiv?

[07:31:00.0000] <Hixie>
you can just say <meta charset="...">

[07:31:01.0000] <Hixie>
(or use HTTP headers)

[07:33:00.0000] <ray>
that's nice

[07:33:01.0000] <Hixie>
it even works in existing browsers

[07:33:02.0000] <ray>
meta http-equiv=blahblahblah was the only thing i couldn't type from memory

[07:34:00.0000] <Hixie>
you could type the DOCTYPE from memory? :-)

[07:34:01.0000] <ray>
:)

[07:34:02.0000] <Hixie>
i'm impressed :-)

[07:37:00.0000] <Hixie>
on an unrelated note, i'm amazed that my last e-mail to shelley actually convinced her

[07:38:00.0000] <Hixie>
i guess it pays to be thorough with each e-mail and to show how one is contradicting oneself rather than just replying to the latest comment each time, ignoring earlier ones

[07:43:00.0000] <jgraham>
/me would like to dispell the notion that only people who don't think the use cases should be addressed support microdata

[07:43:01.0000] <Philip`>
/me wonders where Shelley said she was convinced

[07:44:00.0000] <Hixie>
Philip`: she didn't, but she didn't reply to my e-mail, which is the only sign i've ever seen of her admitting she was wrong on anything

[07:44:01.0000] <Philip`>
It seems an equally likely hypothesis is that she had already made all the points she wanted to make and was fed up with going in circles and therefore stopped

[07:46:00.0000] <Hixie>
that'd be a first, if so

[07:46:01.0000] <Philip`>
jgraham: If you do that, you could also dispel the notion that the use cases are synonymous with the Semantic Web community

[07:47:00.0000] <Hixie>
anyone got a suggestion for a good e-mail that summarises the storage mutex issues?

[07:48:00.0000] <Philip`>
/me doesn't care about Semantic Webs or Linked Data or RDF or anything, but still thinks it's probably useful to have a way to easily encode structured data in a web page so screen-scrapers are easier to write

[07:49:00.0000] <Hixie>
that's basically google's position too

[07:50:00.0000] <jgraham>
Philip`: I habe a draft of an email that says more or less that that (I wrote it before complaining on IRC even) but I am never convinced that actually sending email is a good idea

[07:50:01.0000] <jgraham>
*have

[07:53:00.0000] <gsnedders>
/me grumbles about not having a clearly defined difference between es-discuss (where discussion is most about ES5) and es5-discuss (where discussion is mostly about ES5)

[07:53:01.0000] <Philip`>
gsnedders: Kind of like public-html and whatwg?

[07:53:02.0000] <Philip`>
(on good days)

[07:53:03.0000] <gsnedders>
But both mailing lists are on the same server.

[07:53:04.0000] <Philip`>
(where public-html discussion is mostly about HTML5 and not about HTML WG processes)

[07:54:00.0000] <Philip`>
s/good/rare/ perhaps

[07:55:00.0000] <jgraham>
gsnedders: Well es-discuss is mostly about es.next these days

[07:55:01.0000] <jgraham>
But I think there are too many mailing lists

[07:55:02.0000] <Hixie>
woo, breakfast time

[07:55:03.0000] <Hixie>
bbl

[08:59:00.0000] <Philip`>
/me thinks the HTML WG should have a poll to determine what poll options to provide in a second poll

[08:59:01.0000] <TabAtkins>
Philip`: I'm not sure that's a good idea.  Can we have a poll about it?

[09:00:00.0000] <gsnedders>
TabAtkins: I don't think a poll for that is worthwhile. Poll?

[09:02:00.0000] <TabAtkins>
Poll?  Poll poll poll-poll.

[09:50:00.0000] <jgraham>
othermaciej: It would be nice if the issues list had the name or shortname of each issue somewhere

[09:50:01.0000] <othermaciej>
jgraham: good point

[09:51:00.0000] <othermaciej>
jgraham: I'll look into that when I get into the office

[09:51:01.0000] <jgraham>
othermaciej: Thanks

[09:52:00.0000] <gsnedders>
othermaciej: For the poll on microdata, is it one vote per W3C Member or one vote per WG representive

[09:53:00.0000] <jgraham>
gsnedders: It is not a poll

[09:53:01.0000] <othermaciej>
gsnedders: I have a three-part answer for that

[09:53:02.0000] <othermaciej>
1) it's not a vote

[09:54:00.0000] <othermaciej>
2) multiple representatives of an organization are free to each state their response

[09:54:01.0000] <othermaciej>
3) if we *do* somehow count (which we don't intend to), we will weight all responses from a single organization as one vote

[09:54:02.0000] <gsnedders>
/me had the understanding of the W3C process that it didn't matter how the outcome of the poll was judged, just whether it was a formal WG vote or not, jgraham 

[09:55:00.0000] <othermaciej>
it is not a formal Vote

[09:55:01.0000] <jgraham>
/me wonders how that works if different respondents from the same organisation say different things

[09:55:02.0000] <othermaciej>
jgraham: might end up counting as 0.5 in favor of each option - but it likely won't come up, since we don't plan to count

[09:56:00.0000] <jgraham>
othermaciej: Sure :)

[09:56:01.0000] <othermaciej>
we just don't want people to complain or be suspicious if there are, say, 7 Apple responses or 10 Opera responses

[10:11:00.0000] <Philip`>
Ooh, impressive, 3 out of 3 of the times listed in the new telecon reminder email are correct and consistent

[10:11:01.0000] <Philip`>
(Too bad the date in the subject is wrong)

[12:49:00.0000] <MattCampbell>
Which html5lib tree implementation is best for round-tripping HTML? That is, I want to parse some HTML, possibly do some manipulations, and then serialize the tree back to HTML.

[12:50:00.0000] <MattCampbell>
I'm talking about the Python implementation.

[12:55:00.0000] <gauthierm>
if an autoplaying video is removed from the document using removeChild(), should it keep playing?

[13:07:00.0000] <kinetik>
gauthierm: yes

[13:08:00.0000] <gauthierm>
kinetik: thanks. Bug in FF 3.5 then.

[13:08:01.0000] <kinetik>
gauthierm: "Media elements must not stop playing just because all references to them have been removed; only once a media element to which no references exist has reached a point where no further audio remains to be played for that element (e.g. because the element is paused, or because the end of the clip has been reached, or because its playbackRate is 0.0) may the element be garbage collected."

[13:08:02.0000] <gauthierm>
ah sorry, I meant NOT a bug in FF 3.5

[13:09:00.0000] <kinetik>
i think there actually is a bug where the media is stopped early in some cases

[13:09:01.0000] <kinetik>
which should be fixed in 3.6

[13:09:02.0000] <gauthierm>
it is a bit weird that audio can keep playing with no way for the user to stop it

[13:12:00.0000] <Philip`>
I guess it helps if you want to write a page that has sound effects, since you don't need to manually keep all the audio elements in the page and then clean them up once they've finished playing

[13:13:00.0000] <gauthierm>
kinetik: is the bug you're referring to triggered when removing a media element that is currently playing?

[13:14:00.0000] <gauthierm>
FF seems to only continue playing if the media is set to autoplay and the element is removed before it starts playing.

[13:14:01.0000] <gauthierm>
WebKit stops when the element is removed in both cases.

[13:22:00.0000] <jgraham>
MattCampbell: It shouldn't matter in theory

[13:22:01.0000] <kinetik>
gauthierm: there was no self-reference to the media when playing, so it'd die at some random point after it was removed from the document and no longer referenced

[13:22:02.0000] <kinetik>
gauthierm: https://bugzilla.mozilla.org/show_bug.cgi?id=518659#c6

[13:23:00.0000] <jgraham>
I always use lxml although there are some bugs with that in corner cases (some of the XMLness checking code is known-wrong, it can't represent all doctypes, etc.)

[13:24:00.0000] <jgraham>
but it is the best tested

[13:24:01.0000] <MattCampbell>
If lxml is the best tested, then I'll use that.

[13:24:02.0000] <jgraham>
beautifulsoup should be avoided

[13:28:00.0000] <gauthierm>
kinetik: I'm a bit confused. Here is my test case: http://labs.silverorange.com/files/mozilla/video-testcase/test-case.html When the video starts playing and you click remove, audio stops instantly. If you click remove before the video loads (still a black box) it continues buffering and then starts playing in the background.

[13:29:00.0000] <MattCampbell>
How tolerant is html5lib's parser with regard to malformed markup?  Can it handle any markup that a current browser can?

[13:31:00.0000] <Philip`>
MattCampbell: It will parse any sequence of bytes and never stop with a fatal error (excepting some of the bugs with lxml etc)

[13:32:00.0000] <Philip`>
and will recover from errors in a way that should be sufficiently compatible with current browsers that future browsers will be happy to adopt the same parsing algorithm

[13:32:01.0000] <kinetik>
gauthierm: that's because removing an active media element causes it to pause

[13:35:00.0000] <kinetik>
er, so, my original answer was incorrect, sorry

[13:35:01.0000] <MattCampbell>
jgraham: Mainly out of curiosity, why should BeautifulSoup be avoided?

[13:36:00.0000] <Philip`>
MattCampbell: It lacks features (like namespace support) and has quite a few bugs that we don't how to work around, if I remember correctly

[13:38:00.0000] <MattCampbell>
So it seems that html5lib + lxml.etree might be suitable for screen-scraping apps, which is what BeautifulSoup was intended for.

[13:41:00.0000] <Philip`>
It ought to work for that

[13:41:01.0000] <Philip`>
If not, please file bugs :-)

[13:42:00.0000] <Philip`>
(A while ago the BeautifulSoup author sounded interested in replacing its parser with html5lib, which would be good if you prefer that API, but I don't know if anything's happened with that)

[13:43:00.0000] <MattCampbell>
I have no particular API preference; since lxml seems to be the most tested, I'll use that.

[13:46:00.0000] <Philip`>
Some of the bugs with lxml may cause fatal errors if you try parsing lots of random real pages, but the bugs ought to be easy to fix and then it'll be robust

[13:47:00.0000] <gauthierm>
kinetik: ah, that makes more sense :) There is a bug in FF then because the networkState is NETWORK_LOADING when the element is removed but FF is NOT acting as if the pause() method was called. I will file on bugzilla.

[13:47:01.0000] <kinetik>
gauthierm: great, thank you

[14:20:00.0000] <Hixie>
TabAtkins: man, i hope you're able to keep track of all the change proposal changes in doing your updates

[14:20:01.0000] <Hixie>
i can barely work out what the proposal is anymore

[14:21:00.0000] <TabAtkins>
It's, well… It's a good thing I procrastinate.

[14:22:00.0000] <Hixie>
heh

[14:22:01.0000] <Hixie>
you have just a few more hours, i guess :-)

[14:28:00.0000] <othermaciej>
you have until 9 AM Pacific Time tomorrow, basically (i.e. the time of the telecon)

[14:29:00.0000] <TabAtkins>
The hope is that from the time I start to the time I turn it in, no more change proposals.

[14:31:00.0000] <hober>
TabAtkins: some feedback on your counter proposal

[14:31:01.0000] <hober>
TabAtkins: for what it's worth, here's what I think is one of the more compelling reasons for having features in HTML like <time>, the microdata attributes, etc.

[14:32:00.0000] <hober>
TabAtkins:  HTML's existing extensibility mechanisms (class="" etc.) are pretty good, but could use some enhancing to better support groups like the Microformats community (who build on top of HTML.)

[14:32:01.0000] <hober>
TabAtkins:  <time>, microdata, etc. simply are that enhancing of HTML's extenisbility mechanisms over what is in HTML4.

[14:32:02.0000] <hober>
/me didn't even mention RDFa

[14:32:03.0000] <hober>
:)

[14:33:00.0000] <TabAtkins>
Heh, but it's sort of dishonest then.  RDFa could also be used to do that.  Microdata just does it better and more easily.

[14:33:01.0000] <hober>
... which is why we should incorporate Microdata into HTML5 and not RDFa, no? :)

[14:34:00.0000] <TabAtkins>
It's why we should *support* Microdata.  It doesn't, by itself, mean we should keep it in HTML5.

[14:34:01.0000] <TabAtkins>
But it is a good support for the actual reasons.

[14:35:00.0000] <hober>
Hmm. No, I think the point I'm trying to make *does* support "keeping microdata in HTML5", in much the same way that my argument supports "keeping class='' in HTML5"

[14:36:00.0000] <TabAtkins>
Not if there was compelling evidence that Microdata could serve the same noble purpose in its own separate spec.

[14:37:00.0000] <hober>
I have an insufficiently powerful imagination, I guess: I can't see the difference between Microdata and the other extensibility points of HTML. No one actually wants class='' to be in a different spec, right?

[14:38:00.0000] <hober>
/me studiously ignores the XHTML Role Attribute Module (because what is role='' if not "class='' 2: class harder"?)

[14:40:00.0000] <TabAtkins>
Oh, don't worry.  I don't see a difference either.  I'm just saying.

[15:15:00.0000] <Hixie>
i really don't understand why we would keep class="" in html5 if we didn't keep microdata, personally


2009-12-10
[16:05:00.0000] <zcorpan_>
kinetik: "When a media element is removed from a Document, if the media element's networkState attribute has a value other than NETWORK_EMPTY then the user agent must act as if the pause() method had been invoked."

[16:05:01.0000] <zcorpan_>
http://www.whatwg.org/specs/web-apps/current-work/multipage/video.html#video

[16:06:00.0000] <kinetik>
yes, i mentioned that later

[16:06:01.0000] <zcorpan_>
oh, indeed

[16:06:02.0000] <kinetik>
i was confused originally

[16:10:00.0000] <zcorpan_>
maybe i should write a tutorial or something for html5 video and other things i've doing qa for

[16:11:00.0000] <zcorpan_>
however i'm a bit disconnected to real-world use cases

[16:12:00.0000] <zcorpan_>
i guess i could look at flash sites or desktop apps for inspiration

[16:22:00.0000] <AryehGregor>
That's the problem, isn't it?  I mean, the people who can write good tutorials are the ones who spend their time actually building websites and/or writing tutorials, while the people who know WTF they're talking about are the ones who spend their time writing specs.  The two sets have a fairly small intersection.

[16:30:00.0000] <AryehGregor>
Tsk, tsk, hacks.mozilla.org suggesting people use multiple="true".  http://hacks.mozilla.org/2009/12/w3c-fileapi-in-firefox-3-6/

[16:34:00.0000] <AryehGregor>
It also forces Courier for form input instead of respecting my default monospace font.

[18:00:00.0000] <AryehGregor>
/me wonders why people misspell his name when it's five letters and easily copy-pastable

[18:00:01.0000] <AryehGregor>
I can understand mispronouncing it, but . . .

[18:07:00.0000] <AryehGregor>
/me doesn't care too much, once let a professor call him "Awry" for a semester without correcting him

[18:09:00.0000] <deltab>
that page also defines a function getByteAt that returns "the Unicode value of the character at the given index"

[18:12:00.0000] <deltab>
oh, the API does seem to work that way

[18:12:01.0000] <TabAtkins>
What a wonderfully named function.

[18:14:00.0000] <TabAtkins>
Ah, nice.  Turns out Manu has made his argument for removing Microdata much, much weaker.

[18:15:00.0000] <deltab>
hmm, so there's no way to get bytes directly from a Blob?

[18:16:00.0000] <Dashiva>
TabAtkins: Doesn't that imply someone else (or he) could just restate the previous, stronger arguments?

[18:17:00.0000] <TabAtkins>
Dashiva: The 'strong' arguments were based on fairness.  The remove-Microdata crowd has rebuked that very line of argument pretty thoroughly.

[18:18:00.0000] <Dashiva>
Do they have actual authority to do that, though?

[18:18:01.0000] <TabAtkins>
Debunking themselves?  I don't see why not.

[18:18:02.0000] <TabAtkins>
It would look pretty silly to come back and say they really meant it was a good argument.

[18:19:00.0000] <TabAtkins>
(It wasn't, but it was better than what's there now.)

[18:19:01.0000] <Dashiva>
No, someone else I mean. "Those people hate microdata, but there are some of us who really believe in fairness and such..."

[18:20:00.0000] <TabAtkins>
Sure.  It's still a crap argument.  Technologies don't succeed or fail based on fairness.

[18:20:01.0000] <Dashiva>
Fair enough (pun noticed)

[18:22:00.0000] <TabAtkins>
/me didn't intend to be punny. ?_?

[18:22:01.0000] <Dashiva>
Oh, the pun was that fairness being irrelevant was "fair enough"

[18:25:00.0000] <TabAtkins>
omg

[18:26:00.0000] <MikeSmith>
AryehGregor: I guess you probably saw bugmail a while back related to issues you raised regarding spurious errors for wikipedia.org validation

[18:26:01.0000] <MikeSmith>
I think everything you reported is now fixed in the source

[18:27:00.0000] <MikeSmith>
you will still get a bunch of errors reported, but they are real errors

[18:29:00.0000] <MikeSmith>
like "Bad value bat-smg for attribute lang"

[18:30:00.0000] <MikeSmith>
and about 5 other instances of language tags of the form "foo-bar"

[18:30:01.0000] <TabAtkins>
Microdata has an experimental FF impl, doesn't it?

[18:31:00.0000] <TabAtkins>
Of the DOM api, I mean.

[18:32:00.0000] <TabAtkins>
/me doesn't want to read back through the permathread to find out if he's remembering correctly.

[18:42:00.0000] <MikeSmith>
/me finds http://meta.wikimedia.org/wiki/Language_code#Subdomains_that_do_not_conform_valid_ISO_639_language_code

[18:50:00.0000] <MikeSmith>
TabAtkins: what's the status on your CSS gradient syntax proposal?

[18:50:01.0000] <TabAtkins>
MikeSmith: I've got a minor change I need to make so that it interacts sanely with background-size (right now you have to guess to come up with implementable behavior).

[18:51:00.0000] <TabAtkins>
And then shepazu and I have been seriously discussing aligning CSS and SVG's notions of gradients, so it's possible that further changes may arrive in the near future.

[18:51:01.0000] <TabAtkins>
(We should see something about that on the www-style list soon.)

[18:51:02.0000] <MikeSmith>
cool

[18:57:00.0000] <MikeSmith>
/me finds CSS WG issue 179 (Add SVG equivalents to gradients proposal.)

[18:57:01.0000] <othermaciej>
hi MikeSmith!

[18:57:02.0000] <MikeSmith>
othermaciej: hey man

[18:58:00.0000] <othermaciej>
how are things?

[18:58:01.0000] <MikeSmith>
good good

[18:58:02.0000] <MikeSmith>
one thing I wanted to ask you about is if SVG filters have been flipped on in nightlies yet

[19:00:00.0000] <MikeSmith>
plus I found a crasher - https://bugs.webkit.org/show_bug.cgi?id=32357

[19:02:00.0000] <othermaciej>
MikeSmith: filters are on in trunk - I don't know if they have made it to a nightly yet

[19:02:01.0000] <othermaciej>
thanks for reporting the crasher

[19:10:00.0000] <TabAtkins>
MikeSmith: Yeah, issue 179 is actually impossible!  Turns out that several of the radial gradients can't be reproduced in SVG.

[19:10:01.0000] <othermaciej>
interesting

[19:11:00.0000] <TabAtkins>
Yeah, which is why we're planning on aligning our abstractions so that we can express the same things using roughly the same concepts.

[19:11:01.0000] <MikeSmith>
TabAtkins: well, kind of a bummer to hear that

[19:11:02.0000] <MikeSmith>
(I meant bummer to hear that they can't be reproduced)

[19:12:00.0000] <TabAtkins>
Eh, it just means that we need to lift SVG a bit, because I think the cases that it's not hitting are useful.

[19:15:00.0000] <shepazu>
yeah, I think there is a pretty easy way to make SVG versions of most of the gradients, but some of them are not really possible in the general case... should be easy to add them to SVG... I am not sure I like some of the CSS syntax, but that's relatively unimportant

[19:16:00.0000] <shepazu>
specifically, SVG doesn't allow the author to explicitly declare whether the radial gradient is elliptical or circular

[19:17:00.0000] <shepazu>
(though it's possible to get both results)

[19:21:00.0000] <TabAtkins>
That latter part is due to SVG having a slightly crazy notion of image servers that isn't sufficiently developed and general, while CSS doesn't have such an idea at all and has so far relied on url() images all working basically the same.

[19:22:00.0000] <TabAtkins>
I want to both generalize SVG's image-server concept, and upgrade CSS to also use it.

[19:25:00.0000] <TabAtkins>
Also: omg I'm stopping reading all the issue-76 threads.  They are entirely uninteresting.  >_<

[19:25:01.0000] <TabAtkins>
Also: now I need to go write a change proposal for that one dt/dd issue.  Not a counter-proposal, at least.

[19:31:00.0000] <othermaciej>
TabAtkins: thanks for updating your Change Proposal

[19:32:00.0000] <TabAtkins>
np.  There are several points that Manu specifically rebutted in my change proposal, and while I'd like to re-rebut them, I don't believe they're actually relevant to the proposal, so I left the mout.

[21:40:00.0000] <Lachy>
more calls for draconian error handling in HTML, but still no new or valid arguments to back it up. http://rebuildingtheweb.com/en/error-messages/

[22:23:00.0000] <JonathanNeal>
Hey guys!

[22:25:00.0000] <JonathanNeal>
Just the other day, I was asked to downgrade our entire website to HTML4 because FCKEditor was trashing HTML5 elements.  Have you heard about that?  I got around it by "protecting" the specific tags themselves, which means they don't show up in the WYSIWYG but they do show up in the code.

[22:26:00.0000] <MikeSmith>
I though the FCKEditor guy was following the HTML5 work

[22:27:00.0000] <JonathanNeal>
1. FCKEditor has since become CKEditor.  2 After several pages of google + searches in their forums will guide you to 0 results of FCKEditor modified to use HTML5.

[22:27:01.0000] <JonathanNeal>
Also, the demo of CKEditor trashes HTML5 elements as well, fwiw.

[22:28:00.0000] <boblet>
http://dev.fckeditor.net/ticket/4556

[22:28:01.0000] <JonathanNeal>
Try "<article><h1>Foo</h1><p>Bar</p></article>" in http://ckeditor.com/demo

[22:28:02.0000] <boblet>
“type changed from Bug to New Feature”

[22:29:00.0000] <JonathanNeal>
new feature as in upcoming feature?

[22:29:01.0000] <boblet>
not sure if that is an attempt at comedy or an indication that it’s in the works

[22:29:02.0000] <MikeSmith>
I think it's simply an indication that it is a low priority

[22:29:03.0000] <MikeSmith>
this is not specific to new HTML5 elements, right?

[22:30:00.0000] <boblet>
JonathanNeal: you could try the old <div class="new-html5-element-name"> in the meantime

[22:30:01.0000] <MikeSmith>
I mean, it does this for any elements it doesn't know, right?

[22:30:02.0000] <JonathanNeal>
I tried messing with the source, checking for every place they mentioned "div" or "address" and cloned those properties for "article" and "section" but it still didn't work.

[22:30:03.0000] <MikeSmith>
an editor that wraps unknown elements in <p> elements would seem to be broken

[22:30:04.0000] <JonathanNeal>
MikeSmith, yes you're right.  And FCKEditor has a "protected tags" function, but it does not actually work.

[22:31:00.0000] <boblet>
/me thought that FCKEditor was more … apt :|

[22:31:01.0000] <MikeSmith>
good times

[22:31:02.0000] <boblet>
as a name

[22:31:03.0000] <MikeSmith>
why in the world does it have this <p>-wrapping misfeature to begin with?

[22:31:04.0000] <JonathanNeal>
But hey, despite it all, I protected the source of <(all new html 5 elements and /)> and we have http://vm-20.liferay.com/

[22:32:00.0000] <JonathanNeal>
It's still in progress, but man the thing sings to the outliner sometimes.

[22:32:01.0000] <boblet>
MikeSmith: because it’s FCKEd…itor, of course

[22:34:00.0000] <JonathanNeal>
boblet, lol yes, I tried to get help with the bug by making this demonstration in pictures @ http://sandbox.thewikies.com/fck-that/

[22:35:00.0000] <MikeSmith>
WTFEditor

[22:36:00.0000] <MikeSmith>
marketing slogan: It's full of surprises!

[22:37:00.0000] <boblet>
the state of WYSIWYG is just sad. I really hope someone will do one that only supports modern browsers and doesn’t use MSWord as interface inspiration

[22:37:01.0000] <boblet>
but I’m not holding my breath

[22:40:00.0000] <JonathanNeal>
I enjoy http://en.wikipedia.org/wiki/Fckeditor#Name_change

[22:42:00.0000] <JonathanNeal>
But seriously, it doesn't look like there is a solution right now, so the fact that we have it working and playing nicely is a dream come true, or at least a waking dream for now.

[00:45:00.0000] <erlehmann>
I dont get why JonathanNeal didn't just change the source code.

[00:46:00.0000] <erlehmann>
Of the editor.

[01:57:00.0000] <annevk2>
/me likes the idea of just using <input> for video et al

[02:11:00.0000] <nessy>
annevk2: how and what for?

[02:14:00.0000] <annevk2>
as a way of dragging video input streams and other input streams to

[02:14:01.0000] <annevk2>
or selecting them

[02:14:02.0000] <annevk2>
rather than an api that gives some kind of dialog for the task

[02:15:00.0000] <csarven>
If any of the terms 'external', 'internal', 'inline' are defined in the HTML4 or 5 spec, where are they?

[02:16:00.0000] <annevk2>
if they are, within a <dfn> :)

[02:18:00.0000] <csarven>
http://www.w3.org/Terms.html#database mentions the difference between external and internal wrt to the document

[02:21:00.0000] <zcorpan_>
"1992/updated Apr-95,"

[02:23:00.0000] <csarven>
=)

[03:25:00.0000] <mitnavn>
Hey, I'm messing with the HTML 5 drag and drop. Does anyone know how to cancel the original action of the drop, so only my own function will run? return false; doesn't do it.

[03:26:00.0000] <zcorpan_>
e.preventDefault() ?

[03:26:01.0000] <mitnavn>
Shouldn't return false do it then? Anyway, let me try.

[03:27:00.0000] <mitnavn>
Yeah it doesn't do it. I'm dragging an image to a textarea, I want to prevent the source URL of being added to the textarea.

[03:32:00.0000] <zcorpan_>
what happens if you DELETE the green player?

[03:32:01.0000] <zcorpan_>
(sorry, public-html reference)

[03:33:00.0000] <annevk2>
he no longer plays

[03:38:00.0000] <jgraham>
/me is surprised that he doesn't get the reference

[03:38:01.0000] <jgraham>
Must be a thread I am not reading or something

[03:39:00.0000] <jgraham>
/me wonders if a whole change proposal is needed to say that <fltcap> is a stupid element name

[03:51:00.0000] <Dashiva>
"If a technical argument favors microdata, it is also a political argument."

[03:52:00.0000] <Hixie>
mitnavn: try canceling some of the other events... might be ondrop that's doing it or something

[03:52:01.0000] <Hixie>
jgraham: actually a www-archive reference

[03:53:00.0000] <zcorpan_>
oh, right

[03:53:01.0000] <zcorpan_>
/me doesn't pay attention to which list he's reading

[03:57:00.0000] <Dashiva>
What is fltcap supposed to mean?

[03:58:00.0000] <Dashiva>
figure-like-thing caption?

[03:58:01.0000] <jgraham>
10 points to anyone who guesses without reading the proposal

[03:58:02.0000] <jgraham>
(and we all know what points mean...)

[04:03:00.0000] <Hixie>
figure... < caption?

[04:03:01.0000] <Hixie>
i give up

[04:03:02.0000] <Philip`>
It's nearly an anagram of flatpack, which might be a clue

[04:05:00.0000] <Philip`>
It's no harder to guess the meaning of fltcap than of dl/dt/dd anyway

[04:05:01.0000] <Hixie>
dt and dd don't stand for anything in <figure>

[04:05:02.0000] <Hixie>
they just happen to already work

[04:06:00.0000] <Hixie>
it'll be interesting to see what happens with <figure>

[04:06:01.0000] <Hixie>
because personally that's one of the things that i dislike the most about html5 right now

[04:06:02.0000] <Hixie>
i much preferred <legend>

[04:07:00.0000] <Dashiva>
Does anyone know the answer?

[04:07:01.0000] <Hixie>
/me doesn't remember

[04:07:02.0000] <Hixie>
it's been a while since i read the proposal

[04:12:00.0000] <daedb>
I think it was "floating caption" or something like that

[04:14:00.0000] <MikeSmith>
why is it again that <legend> isn't the right solution for the long term?

[04:14:01.0000] <Hixie>
because people care more about the short term, sadly

[04:15:00.0000] <MikeSmith>
that doesn't seem like a terrifically compelling reason

[04:16:00.0000] <Philip`>
/me searches the mailing list and can't actually find any explanation of what fltcap means

[04:16:01.0000] <MikeSmith>
I would people 10 or 20 years from now are not going to be especially happy about that rationale

[04:16:02.0000] <MikeSmith>
*would think

[04:16:03.0000] <jgraham>
daedb wins although I guess he looked at the proposal

[04:16:04.0000] <daedb>
I liked <legend> just because it's a single element that can actually be pronounced.

[04:17:00.0000] <jgraham>
MikeSmith: But "don't worry about the short term" thinking leads to XHTML 2

[04:17:01.0000] <daedb>
jgraham: I actually did remember it, no cheating :p

[04:17:02.0000] <jgraham>
daedb: But had you previously looked?

[04:17:03.0000] <Philip`>
daedb: It could be pronounced wrongly and cause people to think it's about feet

[04:17:04.0000] <MikeSmith>
jgraham: sterling logic there

[04:17:05.0000] <daedb>
jgraham: Yeah, weeks ago

[04:18:00.0000] <jgraham>
Philip`: Too much I'm Sorry I Haven't A Clue?

[04:18:01.0000] <Philip`>
I thought there was a Design Principle about designing features so that they work as well as possible in legacy browsers or something along those lines

[04:19:00.0000] <Hixie>
jgraham: you can't ignore the longterm either

[04:19:01.0000] <Hixie>
jgraham: one has to find a balance

[04:19:02.0000] <jgraham>
MikeSmith: The point is that doing stuff that people can use now is important because otherwise they get frustrated and ignore the solution. And in 10 years time everyone is so used to the workaround that no one bothers with the "official" solution anymore

[04:19:03.0000] <daedb>
Philip`: It keeps reminding me of felt-tip pens for some reason.

[04:20:00.0000] <jgraham>
or they cargo-cult the idea that the element is broekn long after the actual brokenness has vanished

[04:20:01.0000] <MikeSmith>
that too

[04:20:02.0000] <Hixie>
/me thinks we should either use <legend> or drop the elements for now and add them later when the parsers are updated

[04:20:03.0000] <Hixie>
which shouldn't take more than a few years, i think

[04:20:04.0000] <Hixie>
5-10 at the most

[04:21:00.0000] <MikeSmith>
in the short term people can use whatever they're using now for marking up figures

[04:21:01.0000] <daedb>
Dropping is not an option imo

[04:21:02.0000] <jgraham>
I think that they are really useful elements today <details> in particular

[04:21:03.0000] <Philip`>
jgraham: I thought ISIHAC involved making up homonyms, not mispronunciations

[04:21:04.0000] <MikeSmith>
yeah, definitely significant demand for <details>

[04:22:00.0000] <daedb>
/me cares more about <figure> than <details>

[04:22:01.0000] <MikeSmith>
on another topic, does anybody have a clue where to report GTK+ bugs?

[04:22:02.0000] <Philip`>
so they'd find some joke about ends of ledges

[04:22:03.0000] <jgraham>
Philip`: Yeah I guess

[04:23:00.0000] <jgraham>
but this is IRC so obviously you have to adapt the rules

[04:23:01.0000] <Philip`>
MikeSmith: http://www.gtk.org/development.html

[04:23:02.0000] <daedb>
Whatever happened to the <p caption> idea? I kinda like that as an alternative.

[04:23:03.0000] <jgraham>
/me notes that he also thinks that <figure> is worthy and should go in HTML5

[04:23:04.0000] <Philip`>
jgraham: You'd have to adapt the title too, since on the internet people don't apologise for being clueless

[04:24:00.0000] <jgraham>
We need I'm Sorry I Haven't a Change Proposal That Meets All The Formal Criteria

[04:25:00.0000] <jgraham>
(But Will Work On It Promptly)

[04:36:00.0000] <zcorpan_>
can't we make hixie go "hmm that's not a bad idea actually. ok done." for <p caption>? :)

[04:37:00.0000] <Hixie>
captions can have multiple paragraphs and lists and stuff

[04:37:01.0000] <zcorpan_>
<div caption>

[04:37:02.0000] <Hixie>
the whole point of <div> is it has no semantics

[04:37:03.0000] <jgraham>
It feels like an ugly hack

[04:38:00.0000] <jgraham>
/me just wishes that English had another word that meant "caption"

[04:38:01.0000] <annevk>
<legend>

[04:38:02.0000] <zcorpan_>
can we use words from another language?

[04:38:03.0000] <Philip`>
<subtitle>

[04:39:00.0000] <Hixie>
<legend> is a fine word

[04:39:01.0000] <Philip`>
<caption2>

[04:39:02.0000] <zcorpan_>
c4p710n

[04:39:03.0000] <daedb>
<captionator>

[04:40:00.0000] <Dashiva>
<caption2:captionharder>

[04:40:01.0000] <jgraham>
Well I think the first problem is the assumption that <figure> and <details> should have the same solution

[04:41:00.0000] <Hixie>
i also think <table> and <fieldset> should have the same solution

[04:42:00.0000] <Hixie>
(or at least three of the four, since two of them can't ever be made the same)

[04:42:01.0000] <Hixie>
if we had <di>, i'd want it to be the same too

[04:42:02.0000] <csarven>
<caption> could trigger a table layout in legacy browsers

[04:42:03.0000] <csarven>
<legend> is better

[04:42:04.0000] <jgraham>
I don't understand why you want them all to be the same

[04:42:05.0000] <Hixie>
that was pretty much my reasoning, yes

[04:43:00.0000] <jgraham>
Well I understand why you would want <table><caption> to be the same as <figure><caption> but that isn't possible

[04:43:01.0000] <jgraham>
But <details> seems rather different to <figure>

[04:44:00.0000] <Hixie>
jgraham: because i think that having parallel constructs be similar leads to better usability (untested hypothesis)

[04:45:00.0000] <daedb>
Does <label> have any bad issues that prevent usage in <figure>/<detail>?

[04:45:01.0000] <jgraham>
I suspect having identical constructs for dissimmilar things leads to poorer usability (and I think these things are somewhat dissimmilar)

[04:45:02.0000] <Hixie>
daedb: yeah, <label> has all kinds of issues

[04:45:03.0000] <jgraham>
daedb: You can't use <label> in <figure> because it prevents you putting form controls inside the caption

[04:45:04.0000] <jgraham>
and has legacy issues

[04:45:05.0000] <daedb>
sigh, I suspected as much...

[04:46:00.0000] <Hixie>
jgraham: i don't think that having them be the same is as strong a requirement as our solution not having ugly names like "fltcap"

[04:46:01.0000] <Hixie>
jgraham: but i certainly wouldn't want to have them all be <x> <xcaption/> ... </x> for every value of x

[04:47:00.0000] <Hixie>
jgraham: if there are names that are significantly superior to <legend> but that only work for one of them, i wouldn't be opposed to exploring that

[04:47:01.0000] <jgraham>
Yes, I agree that fltcap is a non-starter

[04:48:00.0000] <jgraham>
And I agree that having n names is probably too many for the relevant n

[04:48:01.0000] <csarven>
<details> seems to be a superset of <figure>

[04:48:02.0000] <jgraham>
csarven: Huh?

[04:48:03.0000] <jgraham>
/me doesn't see that at all

[04:48:04.0000] <Hixie>
i think they're pretty distinct

[04:49:00.0000] <csarven>
<details> is way too generic

[04:49:01.0000] <Hixie>
i think <details> has more in common with <fieldset> than <figure>

[04:49:02.0000] <jgraham>
I agree with that

[04:50:00.0000] <jgraham>
(It's a pity that fieldset is so screwed up or we could just have <fieldset details> and not invent anything new)

[04:50:01.0000] <zcorpan_>
<fieldset closed>

[04:50:02.0000] <annevk>
what's up with the <something semantic> construct?

[04:50:03.0000] <daedb>
<key>?

[04:51:00.0000] <annevk>
did it start with <time>?

[04:51:01.0000] <daedb>
/me is looking at legend synonyms on wiktionary...

[04:52:00.0000] <zcorpan_>
<img usemap> was before <time>

[04:52:01.0000] <jgraham>
annevk: Well for fieldset it makes sense since <details> is a special kind of <fieldset>

[04:52:02.0000] <zcorpan_>
and ismap

[04:52:03.0000] <annevk>
zcorpan_, usemap was a pointer

[04:52:04.0000] <annevk>
but yeah...

[04:53:00.0000] <zcorpan_>
can we make all <fieldset>s closeable?

[04:53:01.0000] <zcorpan_>
i.e. say that <fieldset> is our <details>

[04:54:00.0000] <zcorpan_>
but only if it has a legend

[04:54:01.0000] <annevk>
maybe <fieldset controls>

[04:54:02.0000] <Philip`>
/me has made a site using lots of fieldsets that have checkboxes in their legends, that cause their content to expand/collapse

[04:55:00.0000] <Philip`>
(and it looked like it worked reasonably nicely)

[07:30:00.0000] <AryehGregor>
MikeSmith, yeah, I noticed that, thanks.  Changing the subdomains will be a pain, and probably won't happen too soon, but we can change the lang="" pretty easily.  I'll bug someone about it.  Is validator.nu up-to-date?  It looks like there are still a couple of warnings raised as errors.

[08:04:00.0000] <AryehGregor>
Apparently there are some Wikipedias in languages so obscure that the English Wikipedia doesn't have an article on them.

[08:04:01.0000] <AryehGregor>
http://cbk-zam.wikipedia.org/wiki/El_Primero_Pagina

[08:34:00.0000] <JonathanNeal>
Hello!

[09:01:00.0000] <MikeSmithX>
AryehGregor: I looked at a few of those invalid language tags case-by-case

[09:02:00.0000] <MikeSmithX>
e.g., for the "bat-smg" case, they should be using just "sgs"

[09:02:01.0000] <MikeSmithX>
which is not in the IANA registry yet

[09:02:02.0000] <MikeSmithX>
but http://www.sil.org/iso639-3/cr_files/2009-050_sgs.pdf seems to indicate there's already been a decision about it

[09:02:03.0000] <AryehGregor>
I concluded "bat" or "lit" would be best for now, for that one.

[09:03:00.0000] <AryehGregor>
Is there a code for "no code assigned yet"?

[09:03:01.0000] <MikeSmithX>
I dunno

[09:03:02.0000] <AryehGregor>
Or extensions for languages with no assigned code?

[09:03:03.0000] <MikeSmithX>
yeah there is that, sorta

[09:03:04.0000] <MikeSmithX>
can do this: "bat-x-smg"

[09:04:00.0000] <MikeSmithX>
the "x" means it's for "private use"

[09:04:01.0000] <MikeSmithX>
which is not exactly the right semantic for this case

[09:04:02.0000] <MikeSmithX>
but close enough

[09:05:00.0000] <MikeSmithX>
anyway, Richard Ishida says that any new tags for the form "aaa-bbb" are always wrong

[09:05:01.0000] <MikeSmithX>
if anything, they should be "aaa-bbbbb"

[09:05:02.0000] <MikeSmithX>
that is, the subtag after hyper should be at least 5 characters long

[09:05:03.0000] <MikeSmithX>
if it is indicating a dialect

[09:06:00.0000] <MikeSmithX>
that form indicates it's a "variant"

[09:06:01.0000] <AryehGregor>
Interesting.

[09:07:00.0000] <MikeSmithX>
s/hyper/hyphen/

[09:12:00.0000] <MikeSmithX>
AryehGregor: anyway, in general I would suggest that they should go through the registration procedure for registering actual new tags for all those cases

[09:13:00.0000] <MikeSmithX>
http://www.inter-locale.com/ID/rfc5646.html#registrationProc

[09:13:01.0000] <AryehGregor>
They = Wikipedia?

[09:13:02.0000] <MikeSmithX>
yeah

[09:13:03.0000] <AryehGregor>
The language approval committee is a few random volunteers, I guess someone should suggest it to them.

[09:13:04.0000] <AryehGregor>
Probably a non-starter if it requires significant work.

[09:13:05.0000] <AryehGregor>
I doubt Wikipedia's standards for a new wiki are as high as the IETF or ISO or whoever for a new language code.

[09:14:00.0000] <MikeSmithX>
well, the fact that a Wikipedia exists for the language/dialect is a good data point that it's important enough to a not-insignificant number of people

[09:16:00.0000] <MikeSmithX>
anyway, I would suggest that in the mean time at the language approval committee adopt a policy of using "foo-x-bar" tags for any languages that don't have standard tags

[09:16:01.0000] <AryehGregor>
No it's not, one person can approve a new wiki if no one actually objects.

[09:16:02.0000] <MikeSmithX>
oh, OK

[09:16:03.0000] <AryehGregor>
That person being some random dude who's doing it because no one else cares.

[09:16:04.0000] <MikeSmithX>
well then, they really should be doing the "foo-x-bar" thing for those cases

[09:17:00.0000] <AryehGregor>
I think it's being done in some cases.

[09:17:01.0000] <MikeSmithX>
OK

[09:17:02.0000] <AryehGregor>
Most of these wikis were set up quite a while ago.  I don't really know so much about language approval, I just ran into this because of the validation errors.

[09:18:00.0000] <MikeSmithX>
well, if there's anything more I can do to help with it, lemme know. it would be nice to have the wikipedia.org home page actually validate

[09:18:01.0000] <AryehGregor>
So could we just change abc-def to abc-x-def and that would be legitimate in all cases?

[09:18:02.0000] <AryehGregor>
Yes, I agree.

[09:18:03.0000] <AryehGregor>
I need to 1) figure out what to change the language codes to, 2) ask someone to do it.

[09:18:04.0000] <AryehGregor>
(2) is probably trivial.

[09:19:00.0000] <AryehGregor>
My conclusions after quickly looking up some stuff on Wikipedia were: bat-smg -> bat or lit; roa-rup -> rup; map-bms -> jav; roa-nrm -> ???; fiu-vro -> vro; cbk-zam -> ??? (doesn't even have Wikipedia article); eml -> egl or rgn; mo -> ro (?).

[09:19:01.0000] <MikeSmithX>
AryehGregor: yeah, I think if you change to abc-x-defgh, that would be legitimate in all cases

[09:19:02.0000] <MikeSmithX>
AryehGregor: yeah, that summary looks like what I had found too for the ones I looked at

[09:20:00.0000] <MikeSmithX>
and I suggest defgh because of the fact that real variant subtags must be at least 5 letters long

[09:20:01.0000] <MikeSmithX>
http://www.inter-locale.com/ID/rfc5646.html#variant

[09:21:00.0000] <Hixie>
wow, you know you're doing well when someone in a w3c mailing list balks at you taking several WEEKS to do something

[09:21:01.0000] <AryehGregor>
MikeSmithX, is eml still supposed to be an error, or only a warning?

[09:22:00.0000] <AryehGregor>
Apparently it's been split into egl and rgn, and I have no idea which is correct here, if either.

[09:22:01.0000] <MikeSmithX>
AryehGregor: I need to talk with i18n folks more about that one

[09:22:02.0000] <AryehGregor>
Also not sure about mo.

[09:22:03.0000] <MikeSmithX>
eml is "retired"

[09:22:04.0000] <AryehGregor>
Well, it only says deprecated for mo, so that's not an error, right?

[09:22:05.0000] <AryehGregor>
I take it validator.nu isn't running trunk of whatever code this is?

[09:22:06.0000] <MikeSmithX>
AryehGregor: yeah, that one is just a warning now, in the source

[09:23:00.0000] <AryehGregor>
How long until that goes live at validator.nu?

[09:23:01.0000] <MikeSmithX>
AryehGregor: probably a couple more weeks. Henri has some parser changes to check in

[09:23:02.0000] <AryehGregor>
k.

[09:23:03.0000] <hsivonen>
Sorry about the silence. I've been first sick and now traveling.

[09:24:00.0000] <hsivonen>
Past experience suggests I shouldn't try to redeploy V.nu while traveling or just before.

[09:24:01.0000] <AryehGregor>
:)

[09:24:02.0000] <MikeSmithX>
hsivonen: I got some other patches I want to bug you about anyway

[09:24:03.0000] <hsivonen>
MikeSmithX: ok

[09:25:00.0000] <hsivonen>
MikeSmithX: do you have them on pastebin?

[09:25:01.0000] <MikeSmithX>
AryehGregor: I have no idea why they picked "bat-" for that anyway, given that it's clearly a dialect of Lithuanian

[09:25:02.0000] <AryehGregor>
Heck if I know.

[09:25:03.0000] <MikeSmithX>
hsivonen: not yet anywhere except in my workspace

[09:26:00.0000] <AryehGregor>
This isn't a very professional operation, for all I know they picked it out of a hat.

[09:26:01.0000] <MikeSmithX>
I will try to get them to you in the next couple days

[09:27:00.0000] <hsivonen>
MikeSmithX: I've pushed the major parser changes. The ones that are still queued up are C++ translation-specific except for one astral NCR fix

[09:27:01.0000] <MikeSmithX>
OK

[09:28:00.0000] <MikeSmithX>
actually, stuff that I have queued up does not need to delay deployment

[09:28:01.0000] <MikeSmithX>
it can wait til after

[09:28:02.0000] <hsivonen>
(yay for test cases for the NCR thing. I did some bit masking that was right for a BMP code unit but wrong for a full code point)

[09:28:03.0000] <AryehGregor>
Okay, changes to the main page should be live within an hour.

[09:28:04.0000] <hsivonen>
my current plan is to redeploy on Monday-ish

[09:28:05.0000] <AryehGregor>
Only ones left are eml, mo, and the compatibility character thing.

[09:29:00.0000] <MikeSmithX>
hsivonen: OK, I will look through what I've got and e-mail you the diffs

[09:29:01.0000] <AryehGregor>
Is there a simple fix for that, like the exact same string without compatibility characters?

[09:29:02.0000] <hsivonen>
MikeSmithX: ok

[09:29:03.0000] <AryehGregor>
My Unicode-fu is weak.

[09:30:00.0000] <MikeSmithX>
mostly simple stuff but want to get your review. e.g., making the PrudentHttpEntityResolves connection timeout and socket timeout configurable through system properties

[09:31:00.0000] <AryehGregor>
There we go: http://validator.nu/?doc=http://www.wikipedia.org

[09:31:01.0000] <AryehGregor>
So eml will remain an error for now, the other two will be demoted to warnings on redeployment?

[09:32:00.0000] <MikeSmithX>
eml will become a warning

[09:32:01.0000] <AryehGregor>
Ah, okay.

[09:32:02.0000] <AryehGregor>
So it will validate.  Great.

[09:32:03.0000] <MikeSmithX>
oh no

[09:32:04.0000] <MikeSmithX>
sorry, no, it will remain an error

[09:32:05.0000] <MikeSmithX>
because it's not actually in the registry

[09:33:00.0000] <AryehGregor>
Right.

[09:33:01.0000] <hsivonen>
AryehGregor: is eml an actual ISO code?

[09:33:02.0000] <AryehGregor>
It's apparently "retired".

[09:34:00.0000] <MikeSmithX>
but it should be. I need to talk to the ltru people about getting it added. as far as I understand it, it is "retired" but still valid. so it should be listed in the registry as "grandfathered", I think, with a "deprecated" field and a "preferred-value" field. but the problem there is that the "preferred-value" field for that case is actually two tags, not one

[09:35:00.0000] <MikeSmithX>
s/should be/should be in the registry/

[09:35:01.0000] <AryehGregor>
Yes, that's why I didn't convert it.

[09:35:02.0000] <AryehGregor>
Not sure which is right here, if either.

[09:35:03.0000] <MikeSmithX>
actually, one of the patches I have here is the same one I e-mailed George about

[09:35:04.0000] <MikeSmithX>
which is a change to jing

[09:35:05.0000] <MikeSmithX>
not to v.nu itself

[09:36:00.0000] <MikeSmithX>
change to enabled the jing schematron backend to emit warnings

[09:36:01.0000] <MikeSmithX>
by recognizing assert/@role=warning

[09:37:00.0000] <MikeSmithX>
and that change wouldn't have any effect on v.nu anyway

[09:37:01.0000] <MikeSmithX>
hmm, or I guess it would for the bring-your-own-schema case

[09:38:00.0000] <zcorpan_>
MikeSmithX: do you fancy fixing bug 14? :)

[09:39:00.0000] <MikeSmithX>
/me looks

[09:40:00.0000] <MikeSmithX>
that would seem to be a parser change

[09:40:01.0000] <zcorpan_>
why?

[09:41:00.0000] <zcorpan_>
xml-stylesheet processing is layered on top of the infoset

[09:41:01.0000] <MikeSmithX>
hmm, yeah.. just that the schema-validation code is not doing any PI checking now

[09:41:02.0000] <MikeSmithX>
I guess it could

[09:44:00.0000] <hsivonen>
I think the implementation for bug 14 should be a SAX consumer that appears as a URI-identified pseudo-schema to the UI

[09:47:00.0000] <zcorpan_>
maybe you could create a new document for each potential xml-stylesheet pi, and feed the document "<xml-stylesheet "+pi.content+"/>"

[09:47:01.0000] <zcorpan_>
and have a schema for that

[09:47:02.0000] <MikeSmithX>
OK, I can take a crack at that one after I get some of these other changes in

[09:48:00.0000] <MikeSmithX>
is <?xml-stylesheet href="<"?> not simply an XML well-formedness error?

[09:48:01.0000] <zcorpan_>
no

[09:49:00.0000] <Hixie>
why would it be a well-formedness error?

[09:49:01.0000] <AryehGregor>
< isn't allowed in XML attributes unescaped, AFAIK.

[09:49:02.0000] <AryehGregor>
I have no idea why, but . . .

[09:49:03.0000] <gsnedders>
That's not an attribute.

[09:49:04.0000] <Hixie>
PIs don't have attributes

[09:49:05.0000] <AryehGregor>
Well, okay, then.

[09:49:06.0000] <Hixie>
:-)

[09:49:07.0000] <gsnedders>
That's a PI, which is a random bunch of characters between <? and ?>

[09:49:08.0000] <AryehGregor>
Great.

[09:50:00.0000] <gsnedders>
AryehGregor: The reason for < being disallowed in SGML compatibility

[09:50:01.0000] <AryehGregor>
Wasn't XML supposed to be reasonably simple?  I guess it is compared to SGML . . .

[09:50:02.0000] <gsnedders>
(As XML is meant to be an SGML subset)

[09:50:03.0000] <hsivonen>
AryehGregor: I think the DPH got forgotten pretty early 

[09:50:04.0000] <Hixie>
SGML certainly puts XML in perspective, in terms of complexity

[09:51:00.0000] <Philip`>
HTML does too

[09:51:01.0000] <AryehGregor>
So does JSON.  :)

[09:51:02.0000] <Hixie>
Philip`: indeed

[10:03:00.0000] <JonathanNeal>
So, I have multiple navigations on a page, should I title some of them differently?  "Navigation" then "Sub navigation" ?

[10:05:00.0000] <JonathanNeal>
Another option is to have the name of the page in the heading of the navigation, like "About Us Navigation", or "Our Leadership Navigation".  The other option I can think of is to just give them their title, like "Our Leadership".

[10:12:00.0000] <TabAtkins>
The latter is what I do for the section-nav, as differentiated from the main-nav.

[10:13:00.0000] <Hixie>
how convenient for shelley to suddenly decide that she's too good to reply to my e-mail, when if i did that she'd accuse me of ignoring her

[10:13:01.0000] <TabAtkins>
Hixie: She said she's leaving the email threads alone.  Don't question it.

[10:15:00.0000] <Hixie>
she said that _in an e-mail thread_

[10:16:00.0000] <Hixie>
it's not like she's actually leaving e-mail threads alone, just the ones that would inconveniently expose her inconsistencies

[11:29:00.0000] <AryehGregor>
whatwg.org is down?

[11:29:01.0000] <Hixie>
well that ain't good

[11:29:02.0000] <Hixie>
up again

[11:30:00.0000] <AryehGregor>
Up again?

[11:30:01.0000] <Hixie>
it's up

[11:30:02.0000] <AryehGregor>
Yeah.

[11:30:03.0000] <AryehGregor>
Hixie, what's supposed to happen if you have an infinite loop of iframes?

[11:30:04.0000] <Hixie>
the UA uses an infinite amount of RAM, i guess

[11:30:05.0000] <AryehGregor>
That's not what browsers do, so shall I file a bug?

[11:31:00.0000] <Hixie>
the browsers are probably applying the "hardware limitations" clause

[11:31:01.0000] <AryehGregor>
They stop after one or two nesting levels, though.

[11:31:02.0000] <AryehGregor>
So no hardware limitations in sight.

[11:32:00.0000] <AryehGregor>
Worth speccing, or better to just forget about it?

[11:32:01.0000] <Hixie>
the hardware limitations clause allows you to protect yourself from hitting the limits

[11:32:02.0000] <Hixie>
doesn't say you have to be near them

[11:33:00.0000] <AryehGregor>
That's pretty generous of it.  Okay, then, I won't bother filing a bug.  Probably not important to interop.

[11:35:00.0000] <zcorpan_>
is a UA that does nothing a conforming UA if it claims to do nothing as a protection against hardware limitations?

[11:37:00.0000] <zcorpan_>
/me has a kitchen table that is a conforming HTML5 UA running with very limited hardware

[11:38:00.0000] <Hixie>
zcorpan_: the exact line is "User agents may impose implementation-specific limits on otherwise unconstrained inputs, e.g. to prevent denial of service attacks, to guard against running out of memory, or to work around platform-specific limitations."

[11:38:01.0000] <Hixie>
i suppose you could say your table is constrained to 0-byte documents

[11:38:02.0000] <Hixie>
and 0-byte URLs

[11:39:00.0000] <zcorpan_>
yep

[11:39:01.0000] <zcorpan_>
i call it the Zero Table

[11:39:02.0000] <gsnedders>
Can you put anything on it? Or eat off it?

[11:39:03.0000] <zcorpan_>
sure

[11:42:00.0000] <Lachy>
good morning everyone

[11:43:00.0000] <gsnedders>
Good evening Australia.

[11:43:01.0000] <gsnedders>
zcorpan_: But that means putting something on nothing, and as zero is the additive identity (for real numbers) then it is no longer zero!

[11:44:00.0000] <zcorpan_>
gsnedders: i didn't say the table is nothing

[11:45:00.0000] <Lachy>
zcorpan_, does the table have a network connection, or does it only accept input printed on cards or paper?

[11:47:00.0000] <zcorpan_>
Lachy: it has a network connection, but it has a limitation of 0 connections at a time

[12:20:00.0000] <zcorpan_>
Hixie: what's the problem with defining it in terms of the infoset?

[12:29:00.0000] <Hixie>
zcorpan_: ?

[12:30:00.0000] <zcorpan_>
Hixie: what's the problem with defining xml-stylesheet in terms of the infoset as opposed to the dom?

[12:30:01.0000] <Hixie>
i meant CSSOM DOM stuff

[12:31:00.0000] <Hixie>
not DOM DOM stuff

[12:31:01.0000] <Hixie>
sorry for the confusion

[12:31:02.0000] <zcorpan_>
ah

[12:31:03.0000] <zcorpan_>
ok

[12:35:00.0000] <zcorpan_>
"The xml-stylesheet processor will say nothing about CSSOM integration." is the wg consensus

[12:55:00.0000] <Hixie>
lovely

[13:34:00.0000] <hsivonen>
can anyone remember why the "already executed" flag on scripts was changed to "already started"?

[13:34:01.0000] <hsivonen>
what problem was addressed?

[13:35:00.0000] <zcorpan_>
/me wasn't aware of the change

[13:38:00.0000] <hsivonen>
I wonder if I should freeze the presence of a for attribute at the same time src, async and defer get frozen...

[13:39:00.0000] <hsivonen>
hmm. problematic.

[13:39:01.0000] <hsivonen>
the already started flag doesn't get set for all non-defer, non-async parser-inserted scripts upon </script>

[13:40:00.0000] <hsivonen>
ouch

[13:40:01.0000] <hsivonen>
sigh

[13:41:00.0000] <hsivonen>
Hixie: I take it that it's intentional that scripts can cause a parser-inserted script to run later if the element didn't have non-whitespace content and didn't have the src attribute at the time the parser tried to run the script

[13:41:01.0000] <hsivonen>
Hixie: if that's the case, src, async and defer might need to be unfrozen at that point

[13:41:02.0000] <hsivonen>
/me tries to figure out what really needs to happen here

[13:42:00.0000] <zcorpan_>
hsivonen: opera seems to insert an empty text node for <script></script>

[13:43:00.0000] <hsivonen>
software.hixie.ch seems to be down :-(

[13:43:01.0000] <zcorpan_>
http://canvex.lazyilluminati.com/misc/dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cbody%3E%3Cscript%3E%3C%2Fscript%3E%3Cscript%3Edocument.body.firstChild.textContent%3D'w(1)'%3C%2Fscript%3E

[13:44:00.0000] <hsivonen>
zcorpan_: that's sad, but I can imagine the Web depending on it

[13:44:01.0000] <hsivonen>
we have a test case for it already in mochitest

[13:46:00.0000] <zcorpan_>
chrome seems to have weird behavior

[13:46:01.0000] <hsivonen>
how do I kill Opera on OS X when it goes crazy upon launch and keeps respawning itself with a new pid faster than I can kill it?

[13:47:00.0000] <zcorpan_>
sometimes the first script has no text node

[13:47:01.0000] <zcorpan_>
doesn't force quit work?

[13:48:00.0000] <hsivonen>
zcorpan_: by the time I can force quit, it has respawned itself under a different pid

[13:48:01.0000] <zcorpan_>
:/

[13:49:00.0000] <hsivonen>
ok. now it managed to exhaust the process limit and killed my Gecko build, too

[13:49:01.0000] <zcorpan_>
that's pretty bad, haven't heard of such crazyness

[13:50:00.0000] <hsivonen>
this has happend to me multiple times with Opera

[13:50:01.0000] <hsivonen>
after Opera got a self-updater on Mac

[13:50:02.0000] <hsivonen>
anyway, I can't load the test case in Opera right now

[13:51:00.0000] <zcorpan_>
opera runs the script

[13:51:01.0000] <hsivonen>
ok

[13:51:02.0000] <hsivonen>
thanks

[13:51:03.0000] <hsivonen>
so I need to implement src/async/defer thawing in that case

[13:51:04.0000] <hsivonen>
I guess I should file a spec bug, too.

[13:53:00.0000] <hsivonen>
I wonder if I can reclaim the number of permitted processes without rebooting

[14:02:00.0000] <nessy>
killall ?

[14:02:01.0000] <hsivonen>
sigh. this script stuff became much more complicated than I expected

[14:02:02.0000] <hsivonen>
nessy: now the problem is that I can't start any more processes

[14:03:00.0000] <hsivonen>
/me reboots

[14:03:01.0000] <nessy>
I was going to say ... :-)

[14:22:00.0000] <JonathanNeal>
I notice that when I set the codec html5 video stopped working in safari.

[14:23:00.0000] <JonathanNeal>
Is there a way to detect the proper codec to set for the video?  Does Safari not support the codec field set on the type attribute?

[14:24:00.0000] <zcorpan_>
do you mean the 'codecs' parameter?

[14:25:00.0000] <JonathanNeal>
yes

[14:26:00.0000] <zcorpan_>
i think it's supported but might have some bugs

[14:26:01.0000] <zcorpan_>
i'd point to the whatwg wiki but it's down atm

[14:27:00.0000] <JonathanNeal>
hancock, nice.

[14:27:01.0000] <JonathanNeal>
yea, i noticed that.

[14:36:00.0000] <JonathanNeal>
Would Content-Type: video/ogg;charset=UTF-8 throw off a browser?

[14:41:00.0000] <JonathanNeal>
You put in the "Please Update Your Code" song?

[14:45:00.0000] <hsivonen>
I hope my spec bug report was articulated well enough

[14:45:01.0000] <MindstormsKid>
Why is whatwg.org down? :/

[14:47:00.0000] <Philip`>
MindstormsKid: Because it's broken

[14:48:00.0000] <Philip`>
Blame Hixie :-)

[14:48:01.0000] <MindstormsKid>
Is there a mirror?

[14:50:00.0000] <Philip`>
Of the spec?

[14:50:01.0000] <hsivonen>
MindstormsKid: http://dev.w3.org/html5/spec/Overview.html

[14:51:00.0000] <MindstormsKid>
hsivonen: ok, though I like the other one better :P

[14:51:01.0000] <TabAtkins>
Everything from the table of contents down is identical.

[14:53:00.0000] <Philip`>
No it's not

[14:53:01.0000] <Philip`>
It's all different colours

[14:53:02.0000] <MindstormsKid>
yes :P

[14:55:00.0000] <TabAtkins>
Identical content.  >_<

[14:55:01.0000] <hsivonen>
colors of the spec are important

[14:59:00.0000] <jgraham>
hsivonen: Did you file a bug on the opera respawning thing?

[15:00:00.0000] <hsivonen>
jgraham: no. I whined about it here.

[15:00:01.0000] <hsivonen>
s/I whined/I just whined/

[15:01:00.0000] <jgraham>
hsivonen: Would you mind filing a bug? Especially if you have some idea of when it reproduces e.g. after update

[15:01:01.0000] <hsivonen>
jgraham: I don't have an idea of the conditions required to reproduce

[15:01:02.0000] <hsivonen>
jgraham: which makes it hard to file a useful bug report

[15:03:00.0000] <jgraham>
hsivonen: Oh. Well perhaps you could file a useless one instead (at least with the version number and the OSX version)

[15:09:00.0000] <MikeSmith>
very nice to see that WebGL draft made publicly available, plus a public mailing list

[15:15:00.0000] <JonathanNeal>
I'm just leaving the codec out for now with mp4 video @ http://vm-20.liferay.com/documentation/5.2/installing-with-tomcat-6

[15:24:00.0000] <aho>
https://www.khronos.org/webgl/

[15:24:01.0000] <aho>
woo webgl specs :D

[15:24:02.0000] <aho>
(draft specs, mind you) :>


2009-12-11
[16:04:00.0000] <ttepasse>
http://24ways.org/2009/incite-a-riot

[16:13:00.0000] <JonathanNeal>
ttepasse, fight for your right to cite!

[16:16:00.0000] <ttepasse>
:)

[17:43:00.0000] <hsivonen>
Does Kaazing have a Web Socket server already?

[17:48:00.0000] <MikeSmith>
hsivonen: yes

[17:49:00.0000] <MikeSmith>
not a separate server, but integrated into their gateway product

[17:49:01.0000] <MikeSmith>
iirc

[17:57:00.0000] <franksalim>
hsivonen, yes

[18:00:00.0000] <MikeSmith>
/me didn't know franksalim was on this channel

[18:01:00.0000] <MikeSmith>
hi franksalim 

[18:01:01.0000] <MikeSmith>
have you had time yet to look at SPDY much?

[18:01:02.0000] <franksalim>
MikeSmith, hello

[18:03:00.0000] <franksalim>
MikeSmith, not in great detail

[18:05:00.0000] <MikeSmith>
franksalim: I'm just wondering how overlap there will end up being in practice between use of SPDY protocol and use of Web Sockets protocol

[18:06:00.0000] <MikeSmith>
wondering whether the Web Sockets client API could also be used with SPDY as the network protocol

[18:07:00.0000] <franksalim>
the WebSocket api could certainly wrap the SPDY protocol in some way

[18:09:00.0000] <franksalim>
so far SPDY hasn't been exposed to JavaScript in any way beyond how HTTP has been exposed, AFAIK

[18:09:01.0000] <franksalim>
XHR and requests for resources (that may be cached)

[21:04:00.0000] <MikeSmith>
AryehGregor: http://www.alvestrand.no/pipermail/ietf-languages/2009-December/009754.html

[21:05:00.0000] <MikeSmith>
about the "eml" tag

[21:05:01.0000] <MikeSmith>
we'll see what response that gets

[21:05:02.0000] <MikeSmith>
but in the mean time, I reckon we can just add special-case handling for eml in the language-tag code

[21:06:00.0000] <MikeSmith>
and remove it later if/when "eml" does actually get added to the registry

[01:56:00.0000] <Hixie>
sorry about sporadic downtime overnight, dreamhost were downing some surgery on my web host config to fix some problems on some of my other domains

[03:24:00.0000] <annevk2>
"If getContext() is subsequently called with any other string, the WebGLRenderingContext shall be detached from its Canvas object (the canvas attribute will be null). API calls may still be made on the WebGLRenderingContext object and readPixels() may be used to obtain the contents of the drawing buffer. But the contents shall no longer be shown on the page."

[03:24:01.0000] <annevk2>
hmm

[03:25:00.0000] <annevk2>
it doesn't say it if can become attached again afterwards

[03:25:01.0000] <Philip`>
Also doesn't say what happens if you have a reference to a 2D context then get a WebGL context then use the 2D context

[03:28:00.0000] <Philip`>
Maybe HTML5 ought to define a mechanism for new contexts to be defined that interact with other contexts in various ways (e.g. one context being a vendor-specific extension of another context, or one being exclusive to all other contexts, etc)

[03:28:01.0000] <annevk2>
or it should say that exclusive contexts are verboten

[03:28:02.0000] <Philip`>
because otherwise it's quite possible that multiple specs will define contradictory behaviour

[03:29:00.0000] <annevk2>
yuhhuh

[03:30:00.0000] <Philip`>
How would it work with non-exclusive contents?

[03:30:01.0000] <Philip`>
If you draw something with WebGL then draw something else with 2D on the same canvas, what happens?

[03:30:02.0000] <Philip`>
and then when you call toDataURL, what happens?

[03:31:00.0000] <annevk2>
it should just draw on top

[03:31:01.0000] <annevk2>
i'd guess

[03:31:02.0000] <annevk2>
and share the backing store

[03:31:03.0000] <Philip`>
HTML5 seems to assume there is a single bitmap that's part of the canvas state, but I think that's incompatible with efficient WebGL implementations

[03:31:04.0000] <annevk2>
so everything draws on the same bitmap

[03:32:00.0000] <Philip`>
OpenGL doesn't draw onto bitmaps in system memory, unless you manually copy it back from the video card (which is pretty slow)

[03:32:01.0000] <annevk2>
you could implement 2d in WebGL

[03:34:00.0000] <Philip`>
I think that's kind of what's happening with some Direct2D browser experiments, but that's Windows-only and it's apparently really hard if you just have the raw 3D APIs

[03:34:01.0000] <jgraham>
/me seems to recall we had this conversation before :)

[03:35:00.0000] <Philip`>
jgraham: I don't recall coming to any conclusion before, sadly

[03:35:01.0000] <jgraham>
No, we didn't

[03:35:02.0000] <Philip`>
It would be nice to have one at some point

[03:35:03.0000] <Philip`>
before it gets shipped too much

[03:35:04.0000] <jgraham>
IIRC we decided it would be nice if everything could happen on the graphics card but probably impractical in the short term

[03:36:00.0000] <annevk2>
Hixie is always big on cracks between the specs

[03:36:01.0000] <annevk2>
he should take this on!

[03:36:02.0000] <Hixie>
file a bug

[03:36:03.0000] <annevk2>
mkay

[03:36:04.0000] <Hixie>
i wonder how webgl implementations handle the canvas being animated with a 3d css transform

[03:37:00.0000] <Hixie>
while a <video> plays on top...

[03:37:01.0000] <annevk2>
on a phone with hardware accelerated video

[03:37:02.0000] <annevk2>
...boom...

[03:39:00.0000] <annevk2>
bugs

[03:39:01.0000] <annevk2>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=8476

[03:40:00.0000] <jgraham>
In the future people will carry their phone in their pocket and a huge backpack containing a graphics card and a water cooling system

[03:41:00.0000] <Philip`>
/me wonders if jgraham has missed the trend for electronic miniaturisation over the past half a century

[03:44:00.0000] <jgraham>
Philip`: :) But presumably power consumption and heat generation are still rather large problems even with minature components

[03:52:00.0000] <Philip`>
jgraham: Things like PSPs and iPhones seem to do decent 3D quite happily, with similar capabilities to full-size consoles from ten years ago

[03:53:00.0000] <Philip`>
and I assume they must cope with the power and heat requirements adequately

[03:53:01.0000] <Philip`>
and I don't see any reason for the trend to stop

[03:54:00.0000] <workmad3>
most high-end phones have mobile versions of reasonable graphics chips nowadays

[03:55:00.0000] <workmad3>
and I'd say the capabilities are closer to pc graphics from about 5 years ago, but with less available memory

[03:56:00.0000] <Philip`>
workmad3: So they could run Half Life 2? :-)

[03:57:00.0000] <workmad3>
Philip`: potentially, if you reduced the graphics quality and map size considerably to fit within the memory available ;)

[03:59:00.0000] <workmad3>
at least in 3d graphics capabilities... there's a lot more to HL2 than that though :)

[04:00:00.0000] <Philip`>
I probably shouldn't have said "capabilities" - it seems like they support much more modern features and APIs, but performance is further behind

[04:00:01.0000] <Philip`>
so you have to use the capabilities more sparingly

[04:00:02.0000] <workmad3>
Philip`: that's probably about right yeah :)

[04:01:00.0000] <Philip`>
(I know basically nothing about this area though, so mostly I'm just guessing :-) )

[04:01:01.0000] <workmad3>
they don't run quite as fast, and have nowhere near as much memory, but the feature-set of the chips in use is pretty modern and allows for more sophisticated stuff than consoles from 10 years ago did

[04:01:02.0000] <Philip`>
At least they don't have to bother with HD resolution output

[04:01:03.0000] <workmad3>
that said, you don't need to render in anywhere near as high resolution or store as much geometry :)

[04:02:00.0000] <workmad3>
I think phone screens nowadays are 'QHD' (quarter HD) nowadays and clock in at 640x360 pixels :)

[04:03:00.0000] <workmad3>
but raw graphics power isn't really the problem on phones... it's control and all the other stuff that's needed

[04:04:00.0000] <Philip`>
That sounds more like ninth HD, unless you cheat and consider lower-than-standard-cheap-laptop-display to be HD :-)

[04:05:00.0000] <Philip`>
They should call it "middling definition", perhaps

[04:05:01.0000] <workmad3>
Philip`: a standard HD res is 1280 * 720

[04:06:00.0000] <workmad3>
(I think that's the res of 1080p)

[04:06:01.0000] <Philip`>
1080p is 1920x1080, hence the "1080" in the name, unless I'm horribly confused

[04:06:02.0000] <workmad3>
so taking the HD resolutions as those supported by HD tvs, then it is a quarter of that :)

[04:07:00.0000] <Philip`>
(and 1280x720 is 720p)

[04:07:01.0000] <workmad3>
hmm... I may be thinking of 720p then

[04:07:02.0000] <workmad3>
still, it's classed as a HD resolution, so the phone manufacturers went with a quarter of it as 'QHD' :)

[04:08:00.0000] <peroo>
actually qhd is 960x540, one quarter of 1920x1080 :)

[04:09:00.0000] <peroo>
640x360 is nHD apparantly, whatever that means

[04:10:00.0000] <Philip`>
/me guesses it must have been a marketing person who decided an acronym containing the word "high" was appropriate for sub-VGA resolutions :-p

[04:10:01.0000] <jgraham>
"Not High Definition", perhaps

[04:11:00.0000] <workmad3>
heh, I'm going by what the phones I worked on that had that resolution called it :)

[04:13:00.0000] <workmad3>
still, AFAIK, the graphics chips available to phones are capable of the raw power required to do all the effects that modern consoles are capable of, the restricting points are things like the physics engines and other processing that's required to get the data ready for the screen in a game

[05:33:00.0000] <Hixie>
i wonder how to add features to html5 without pissing off the htmlwg

[05:37:00.0000] <daedb>
Hixie: I don't think that's possible (unless it's XMLy or includes namespaces), but do it anyway :p

[05:38:00.0000] <Hixie>
well there are options like making the new features only appear in the whatwg version

[05:38:01.0000] <annevk2>
Hixie, what feature?

[05:38:02.0000] <Hixie>
which would piss off the people who get pissed off regardless of what i do, but wouldn't piss off the more sane people

[05:39:00.0000] <jgraham>
I would be pissed off if we introduced namespaces

[05:39:01.0000] <Hixie>
annevk2: e.g. a widget to trigger the display of UA UI to allow the user to select a video source

[05:39:02.0000] <Philip`>
What about the people who think it'll be really confusing if there's two substantially different HTML5 specs?

[05:39:03.0000] <Hixie>
Philip`: are they likely to cause me more or less trouble than the people who would complain at me adding features as we are trying to get to LC?

[05:40:00.0000] <annevk2>
Hixie, at what point do we declare HTML5 feature complete?

[05:40:01.0000] <Hixie>
we declared HTML_5_ feature complete about 2 years ago

[05:40:02.0000] <Hixie>
well that's not quite accurate

[05:41:00.0000] <Hixie>
we announced that HTML5 feedback shouldn't ask for new features about 2 years ago, and then announced LC (feature freeze) about 2 months ago

[05:41:01.0000] <Hixie>
but I'm talking about post-HTML5

[05:41:02.0000] <Hixie>
HTML6, or HTMLCurrent, or whatever

[05:41:03.0000] <annevk2>
ah ok

[05:41:04.0000] <annevk2>
so are you going to version the differences somehow?

[05:42:00.0000] <Hixie>
how do you mean

[05:42:01.0000] <annevk2>
<!--HTML6--> or some such

[05:42:02.0000] <Hixie>
you mean "are the new features going to appear in the w3c html5 draft"?

[05:43:00.0000] <annevk2>
just wondering about the WHATWG Last Call draft for the moment

[05:43:01.0000] <jgraham>
any draft that says HTML5 at the top I would have thought

[05:43:02.0000] <Hixie>
yeah that's the question, how to do it

[05:43:03.0000] <Hixie>
we could rename the whatwg draft to just "HTML", i guess

[05:43:04.0000] <jgraham>
/me thinks it would be fine to make a HTML.next draft for unstable stuff

[05:43:05.0000] <Hixie>
or have a separate view of it somehow

[05:44:00.0000] <annevk2>
I think we should maintain a HTML5 copy

[05:44:01.0000] <annevk2>
preferably on the same URL

[05:44:02.0000] <Hixie>
maybe have specs/web-apps/html5/ include just the html5 stuff, and specs/web-apps/current-work/ for the new stuff

[05:44:03.0000] <Hixie>
(and specs/web-apps/current-work/complete.html for everything but sql)

[05:44:04.0000] <annevk2>
current-work/ does suggest latest, so yeah, that seems ok

[05:45:00.0000] <Hixie>
btw if any of you want your specs merged into complete.html, we can do that

[05:45:01.0000] <Hixie>
e.g. to get the XHR cross-refs working

[05:45:02.0000] <annevk2>
people will find it somewhat confusing if we call it "HTML" though I suspect

[05:45:03.0000] <Hixie>
we could call it HTML6

[05:45:04.0000] <Hixie>
or 5.1

[05:45:05.0000] <Hixie>
or something

[05:45:06.0000] <annevk2>
Hixie, hmm yeah, I hope someone fixes anolis

[05:46:00.0000] <Hixie>
(though i really want to drop this versioning crap)

[05:46:01.0000] <annevk2>
HTML (formerly HTML5)

[05:46:02.0000] <annevk2>
for at least half a year or so

[05:46:03.0000] <annevk2>
and then we drop (formerly HTML5)

[05:47:00.0000] <Hixie>
or "HTML (including HTML5)", to indicate it's more than HTML5

[05:47:01.0000] <annevk2>
yeah, works too

[05:47:02.0000] <Hixie>
and then drop the paranthetical

[05:47:03.0000] <gsnedders>
annevk2: patches welcome.

[05:48:00.0000] <jgraham>
Call it HTML unstable or something

[05:48:01.0000] <Hixie>
(s/a/e/magic)

[05:48:02.0000] <gsnedders>
HTML Beta

[05:48:03.0000] <jcranmer>
HTML 5 rc i

[05:48:04.0000] <jgraham>
HTML Labs

[05:48:05.0000] <Hixie>
well the idea is to just have a single name that is always the top of the tree

[05:48:06.0000] <Hixie>
so "unstable" or "beta" is kinda wrong

[05:48:07.0000] <gsnedders>
jcranmer: So it's not a rc?

[05:48:08.0000] <Hixie>
i mean, <p> isn't going anywhere

[05:48:09.0000] <jcranmer>
gsnedders: it's an imaginary one

[05:49:00.0000] <Hixie>
i wish othermaciej was here, i'm curious as to his opinion on this

[05:49:01.0000] <jcranmer>
more complex than the real HTMLs

[05:49:02.0000] <jgraham>
Hixie: Well you always have an unstable and periodically make numbered stable releases, Debian style

[05:49:03.0000] <jgraham>
(or call it HTML Mainline which sounds cooler, like a VU song)

[05:49:04.0000] <Hixie>
that's what i want to move away from

[05:49:05.0000] <annevk2>
I'd rather be forever in release-state

[05:49:06.0000] <Hixie>
yeah

[05:50:00.0000] <Hixie>
bbiab, gotta get ready.

[05:50:01.0000] <jgraham>
Well so I would I but good luck getting the rest of the world to agree

[05:50:02.0000] <annevk2>
but until we convince the W3C it's prolly safer to have markers for the new stuff

[05:52:00.0000] <annevk2>
no need to convince the whole world

[05:53:00.0000] <annevk2>
just the patent lawyers

[05:53:01.0000] <annevk2>
which is prolly harder

[06:19:00.0000] <Hixie>
not really any need to convince those either so long as the w3c is publishing snapshots

[06:19:01.0000] <Hixie>
which i doubt they plan to stop doing

[06:21:00.0000] <annevk2>
legal effect of the patent policy only applies to W3C Recommendations

[06:21:01.0000] <annevk2>
I believe

[06:26:00.0000] <Hixie>
right, so so long as they keep publishing everything the whatwg does, everything is covered

[06:26:01.0000] <Hixie>
doesn't really matter when the publication happens, since the policy doesn't really kick in until REC anyway, as I understand it

[06:26:02.0000] <annevk2>
the issue I'm thinking about is that W3C Recommendation status is incompatible with forever-trunk

[06:27:00.0000] <Hixie>
i'm saying that what we do is forever-trunk at the whatwg, and continue working with the w3c to publish RECs every now and then

[06:27:01.0000] <Hixie>
as we've basically been doing

[06:27:02.0000] <Hixie>
the only problem is that the htmlwg is kinda stopping progress on html because of it trying to go to LC

[06:28:00.0000] <Philip`>
If the patent policy only kicks in at REC, I guess that means everyone is free to sue each other until 2022?

[06:28:01.0000] <annevk2>
that works though that also means you need to keep the REC markers in the document (or trunk markers)

[06:28:02.0000] <annevk2>
if you want to get rid of those too (I like!) then the convincing thingie above applies

[06:29:00.0000] <annevk2>
Philip`, pretty much

[06:29:01.0000] <annevk2>
Philip`, but I'm not a lawyer

[06:29:02.0000] <Hixie>
REC markers?

[06:31:00.0000] <annevk2>
<!--REC-ends-->

[06:31:01.0000] <annevk2>
or <!--HTML6--> or whatever

[06:31:02.0000] <jgraham>
That sounds... bad

[06:32:00.0000] <Hixie>
so long as one source document is generating more than one published document, i don't see how we could stop having some kind of markers

[06:33:00.0000] <jgraham>
It seems nightmarish to generate multiple versions of the same spec from the same source file

[06:34:00.0000] <Hixie>
i take it you haven't read http://wiki.whatwg.org/wiki/FAQ#What_are_the_various_versions_of_the_spec.3F recently

[06:34:01.0000] <annevk2>
Hixie, the difference is that this is a version marker, but maybe that's not much of a difference

[06:34:02.0000] <MattCampbell>
What version of Python does html5lib require?

[06:34:03.0000] <Hixie>
annevk2: i don't see how it's any different than html5 vs complete.html, or the whatwg microdata vocab spec vs w3c's three microdata vocab specs

[06:34:04.0000] <jgraham>
Hixie: This is quite different

[06:34:05.0000] <annevk2>
Hixie, and hopefully we do not let it scale to IE8 proportions

[06:35:00.0000] <jgraham>
I would have thought

[06:35:01.0000] <Hixie>
jgraham: how so?

[06:35:02.0000] <jgraham>
MattCampbell: 2.4 I guess

[06:35:03.0000] <Hixie>
annevk2: well these are just supersets of each other right, i mean, it's not like we're changing anything

[06:35:04.0000] <jgraham>
Hixie: Because I would expect changes from new versions to be much more invasive than changes that just involve splitting along broad feature lines

[06:36:00.0000] <MattCampbell>
darn, guess I'll have to compile Python from source to use html5lib on RHEL5 (which has Python 2.4)

[06:36:01.0000] <Hixie>
jgraham: i would expect editorial changes to just be made to all versions

[06:36:02.0000] <jgraham>
MattCampbell: (it might work under 2.3 but no promises)

[06:36:03.0000] <annevk2>
Hixie, sure, but it could get tricky nonetheless

[06:36:04.0000] <Hixie>
annevk2: well if it is too tricky we can always give up and do something else :-)

[06:37:00.0000] <annevk2>
Hixie, e.g. further overloading of existing methods, etc.

[06:37:01.0000] <gsnedders>
Hixie: Web DOM Core?

[06:37:02.0000] <annevk2>
diving and hiking

[06:38:00.0000] <jgraham>
Hixie: Yes but say you make a change that adds new steps to an existing algorithm, makes a few changes to conformance criteria and adds a few paragraphs of new material

[06:38:01.0000] <jgraham>
Then you end up with a bunch of ifdefs all over the place to try and seperate out the HTML6 stuff from the HTML5 stuff

[06:38:02.0000] <Hixie>
jgraham: we have a bunch of ifdefs already to do complete.html

[06:39:00.0000] <Hixie>
jgraham: it's not that big a problem

[06:39:01.0000] <jgraham>
Hixie: On such a fine-grained level? It seems like it should be a problem :)

[06:40:00.0000] <Hixie>
search for "END complete" in the source file

[06:41:00.0000] <annevk2>
Hixie, anyway, I don't have a better solution other than changing the way the W3C REC-track works and what you have works much better than what I do for XMLHttpRequest I think

[06:41:01.0000] <annevk2>
(which is trying to keep two documents in sync)

[06:41:02.0000] <Hixie>
on the long term i definitely think we should change the way the w3c REC track works, but that needs far more effort than i am willing to put out

[07:55:00.0000] <pesla\work>
sigh, firefox always crashes on the html5 specs -.-

[07:56:00.0000] <Hixie>
pesla\work: use http://whatwg.org/html5

[07:56:01.0000] <Hixie>
pesla\work: that works in firefox

[07:56:02.0000] <annevk2>
pesla\work, wrong channel to complain

[07:56:03.0000] <pesla\work>
Hixie: sweet :)

[07:57:00.0000] <Hixie>
pesla\work: if you need the one-page version, try http://www.whatwg.org/specs/web-apps/current-work/?slow-browser

[08:00:00.0000] <pesla\work>
annevk2: Man, I just needed someone to listen to my grief! ;)

[08:00:01.0000] <pesla\work>
Hixie: the /html5 method works perfectly :)

[08:05:00.0000] <Philip`>
http://www.whatwg.org/specs/web-apps/current-work/multipage/urls.html#htmlcollection-0

[08:05:01.0000] <Philip`>
"nthnthnth(&*@%WN@%HM)%25"

[08:06:00.0000] <Philip`>
Hixie: I'm not sure that's valid IDL

[08:06:01.0000] <Hixie>
yeah, trying to fix that. somehow my script thought i'd fixed it (which i had in the source) and so was refusing to regen the files

[08:06:02.0000] <Hixie>
ironically, it was there to test that the new webidl check in the pipeline was working

[08:06:03.0000] <Hixie>
(i've integrated dom's service into my update script)

[08:07:00.0000] <Philip`>
Ah

[08:09:00.0000] <annevk2>
pesla\work, just trying to find you a long term solution :p

[10:14:00.0000] <Hixie>
ok well i tried doing what we were talking about earlier

[10:14:01.0000] <Hixie>
we'll see how it goes

[10:22:00.0000] <hober>
Hixie: it seems like whatwg.org/html5 should redirect to the HTML5 draft

[10:22:01.0000] <hober>
maybe we could mint a whatwg.og/html redirect to current-work

[10:23:00.0000] <Hixie>
added /html

[10:23:01.0000] <Hixie>
i think people would find the full html spec more useful than the html5 spec

[11:05:00.0000] <jwalden>
annevk2: actually, I was thinking old-school, Sinatra; never even heard of MGS before :-)

[13:18:00.0000] <smaug>
"Your message to whatwg awaits moderator approval"

[13:18:01.0000] <smaug>
um

[13:19:00.0000] <smaug>
who is the moderator?

[13:34:00.0000] <Philip`>
smaug: There is no moderator

[13:34:01.0000] <Philip`>
smaug: so your message will never be approved

[13:34:02.0000] <Philip`>
smaug: You have to be subscribed before posting

[13:36:00.0000] <smaug>
Philip`: I'm subscribed

[13:36:01.0000] <smaug>
Philip`: the problem is that the email has apparently too many people CC'ed

[13:36:02.0000] <smaug>
(not that many, really)

[13:37:00.0000] <Philip`>
smaug: Oh, okay

[13:37:01.0000] <Philip`>
smaug: CC less :-)

[13:39:00.0000] <smaug>
Philip`: actually someone else CC'ed already so many people that couldn't really add new ones :)


2009-12-12
[19:04:00.0000] <MikeSmith>
AryehGregor: is there a way to contact the Wikimedia Language Committee by e-mail?

[19:06:00.0000] <MikeSmith>
ah, found it

[19:30:00.0000] <MikeSmith>
AryehGregor: see http://www.alvestrand.no/pipermail/ietf-languages/2006-November/005199.html

[19:33:00.0000] <MikeSmith>
Amorphous: from Doug Ewell, who's closely involved with the language-tag registry

[19:34:00.0000] <MikeSmith>
in reply to Gerard Meijssen asking about what WMF should do about language tags

[19:34:01.0000] <MikeSmith>
Doug replied that they should do the "x-" thing

[19:35:00.0000] <MikeSmith>
so I guess the change you made to the home page aligns with that

[22:30:00.0000] <hsivonen>
I've landed the </script> tokenization change on mozilla-central. 

[22:30:01.0000] <hsivonen>
should be in the next batch of nightlies

[22:36:00.0000] <hsivonen>
in fact, trunk should now be in sync with the spec except for <spacer>

[02:22:00.0000] <Philip`>
<span sandbox="random_token"><plaintext>...user-generated content...</span sandbox="same_token">

[02:23:00.0000] <Philip`>
That should solve Michal Zalewski's point 2.2, assuming the sandbox parser is properly sandboxing

[02:23:01.0000] <Philip`>
/me doesn't really want to imagine how such a parser would be implemented

[06:58:00.0000] <theMadness>
How is the attribute "required" supposed to work in html5 when applied to only one radio input if the others don't have it?

[06:59:00.0000] <Hixie>
you only need to set it on one of the radio buttons

[06:59:01.0000] <Hixie>
basically

[06:59:02.0000] <theMadness>
And it doesn't matter which. Any suggestion as per what would constitute "elegant coding"?

[07:00:00.0000] <theMadness>
AKA What Would Xhtml2 Do?

[07:01:00.0000] <Hixie>
xhtml2 would use xforms, so i don't know if that's what i'd do :-)

[07:01:01.0000] <Hixie>
i'd recommend putting required on all the radio buttons if they're supposed to be required

[07:01:02.0000] <theMadness>
I see.

[07:02:00.0000] <Hixie>
but really a required radio button set is a bit dubious -- you're only need it if you hadn't checked a radio button to start with

[07:02:01.0000] <Hixie>
and generally speaking it's better to check a radio button first

[07:02:02.0000] <Hixie>
instead of having none et

[07:02:03.0000] <Hixie>
set, even

[07:03:00.0000] <theMadness>
By the way, how does that compare to <select>? Does required fail in a <select> dropdown? (the first element is always selected, at least in most UA)

[07:06:00.0000] <theMadness>
Ah, there's no such thing as required for select.

[07:06:01.0000] <theMadness>
Funny.

[07:07:00.0000] <Hixie>
yeah there's no way to not have something selected

[07:08:00.0000] <Hixie>
gotta go, reffing an fll regional

[07:08:01.0000] <Hixie>
back later

[07:08:02.0000] <theMadness>
Thanks Ian.

[07:08:03.0000] <theMadness>
Have fun.

[07:54:00.0000] <Dashiva>
http://www.codinghorror.com/blog/

[07:54:01.0000] <Dashiva>
"We're overloading the class attribute with two meanings. Is "adr" the name of a class that contains styling information, or the name of a data field? (...) the semantics of the class attribute have been permanently altered."

[07:55:00.0000] <Dashiva>
Class is for CSS, it seems

[08:01:00.0000] <theMadness>
Yeah I'm puzzled too, "css class" means nothing.

[08:02:00.0000] <theMadness>
This attribute assigns a class name or set of class names to an element. Any number of elements may be assigned the same class name or names. Multiple class names must be separated by white space characters.

[08:02:01.0000] <theMadness>
Nothing about css in http://www.w3.org/TR/html4/struct/global.html#h-7.5.2

[14:32:00.0000] <Dashiva>
>Mozilla’s new offices have conference rooms named after memes! Had to take a shot of All Your Base room.


2009-12-13
[21:09:00.0000] <aho>
why is it getContext(String) instead of... like... getContext2d() and getContext3d()?

[21:09:01.0000] <aho>
the latter would make autocomplete voodoo a lot easier :>

[21:13:00.0000] <aho>
i also can't find another build-in js functions which returns different types depending on the passed parameter(s)

[21:13:01.0000] <aho>
-s

[04:10:00.0000] <AryehGregor>
"I for one would rather not go to such trouble. Can you imagine going

[04:10:01.0000] <AryehGregor>
to every site you visit and specifying that you want XHTML instead of

[04:10:02.0000] <AryehGregor>
HTML, rather than just specifying

[04:10:03.0000] <AryehGregor>
"application/xhtml+xml;q=1.0,text/html;q=0.0," in your request

[04:10:04.0000] <AryehGregor>
headers?"

[04:10:05.0000] <AryehGregor>
. . .

[05:35:00.0000] <daedb>
AryehGregor: Where is that quote from?

[07:57:00.0000] <AryehGregor>
daedb, the WHATWG list.


2009-12-14
[16:07:00.0000] <AryehGregor>
/me thinks he had a clever idea, but isn't sure

[16:07:01.0000] <AryehGregor>
The last time I thought I had a clever idea was onlyreplace, and that turned out to be fairly pointless.

[16:08:00.0000] <AryehGregor>
/me lets the list decide

[22:26:00.0000] <MikeSmith>
anybody know of any ABNF testing tools?

[22:27:00.0000] <MikeSmith>
I want to take a string and check it against an ABNF production to see it matches

[22:35:00.0000] <othermaciej>
you could try writing it as a yacc grammar

[22:35:01.0000] <othermaciej>
or as a regexp if it's a regular language

[22:38:00.0000] <MikeSmith>
I guess I could just generate a parser from it

[22:38:01.0000] <MikeSmith>
but that would kind of seem like overkill

[22:39:00.0000] <MikeSmith>
or I suppose I could even just read the ABNF spec to figure out what the production means

[03:42:00.0000] <zcorpan_>
text/x-html

[03:42:01.0000] <zcorpan_>
so that's where they got the name from

[03:44:00.0000] <virtuelv>
whatwg spec is quite an achievement

[03:44:01.0000] <virtuelv>
we have it freezing four browsers

[03:46:00.0000] <nessy>
lol

[03:46:01.0000] <nessy>
that's why I use the multipage version

[03:49:00.0000] <zcorpan_>
browsers suck

[03:50:00.0000] <Philip`>
virtuelv: Browser developers just need to implement better support for the HTML5 spec

[03:51:00.0000] <virtuelv>
Philip`: somehow I don't think that would've made the spec less freezy :-)

[03:52:00.0000] <jgraham>
/me wonders which browsers and how low a "freeze" threshold is for virtualv

[03:53:00.0000] <Philip`>
virtuelv: Supporting the spec involves not freezing on it, by (my) definition :-)

[04:13:00.0000] <gsnedders>
Hixie: Can you look at http://lists.w3.org/Archives/Public/public-html/2009Nov/0351.html ASAP?

[04:29:00.0000] <zcorpan_>
hmm... is it intended that this won't throw? window.postMessage('foo', '/');

[04:30:00.0000] <zcorpan_>
and whether window.postMessage('foo', ''); throws depends on the first script's base url

[04:49:00.0000] <zcorpan_>
Philip`: i guess the multipage script shouldn't use "HTML5" for current-work

[04:50:00.0000] <zcorpan_>
Philip`: can it use the original title instead of hard-coding the title?

[04:59:00.0000] <Hixie>
gsnedders: assume i'll do what you suggested

[05:00:00.0000] <Hixie>
zcorpan_: why would '/' not throw but '' throw?

[05:01:00.0000] <zcorpan_>
Hixie: because  resolving it relative to the first script's base URL ... results in a URL with a <host-specific> component that is neither empty nor a single U+002F SOLIDUS character (/)

[05:02:00.0000] <zcorpan_>
Hixie: ...for '' when the first script's base URL has a host-specific part other than '/'

[05:03:00.0000] <Hixie>
oh it's an origin? i forget.

[05:03:01.0000] <Hixie>
i thought it was the target url for some reason.

[05:03:02.0000] <zcorpan_>
yeah, the argument is targetOrigin

[05:03:03.0000] <gsnedders>
Hixie: k, thx

[05:06:00.0000] <zcorpan_>
Hixie: i filed http://www.w3.org/Bugs/Public/show_bug.cgi?id=8490 about that

[05:13:00.0000] <Hixie>
k

[05:16:00.0000] <zcorpan_>
step 4 of postMessage is broken because of the above-mentioned bug (since it passes an unresolved string to "same origin", which assumes a resolved url, if i understand things correctly)

[05:17:00.0000] <Hixie>
k

[06:56:00.0000] <Hixie>
any opera people here?

[06:56:01.0000] <zcorpan_>
yep

[07:05:00.0000] <jgraham>
+1

[07:10:00.0000] <gsnedders>
yep

[07:10:01.0000] <gsnedders>
/me is blatantly slow

[07:10:02.0000] <gsnedders>
But not as slow as payman.

[07:15:00.0000] <payman>
ok u proved ur point! 

[07:15:01.0000] <payman>
+1

[07:15:02.0000] <payman>
:p

[07:16:00.0000] <Hixie>
zcorpan_ won the prize

[07:16:01.0000] <Hixie>
since he replied first

[07:16:02.0000] <jgraham>
What, he quit first

[07:16:03.0000] <jgraham>
Oh

[07:16:04.0000] <jgraham>
Well he quit first too

[07:21:00.0000] <gsnedders>
Hixie: So, now you've established there are Opera people here…

[07:21:01.0000] <zcorpan_>
/me has established that Hixie causes zcorpan's opera build to crash :P

[07:26:00.0000] <zcorpan_>
test

[07:26:01.0000] <zcorpan_>
yay, survived

[08:00:00.0000] <Dashiva>
AryehGregor: <span sandbox> seems like it would be foiled by the attacker just starting his content with '</span>', am I missing something?

[08:00:01.0000] <Philip`>
zcorpan_: Fixed <title>

[08:01:00.0000] <Philip`>
Dashiva: I thought the idea was you'd escape the user's text before outputting it

[08:01:01.0000] <Philip`>
so you'd get <span sandbox>&lt;/span&gt;...

[08:02:00.0000] <Dashiva>
That's the idea, but then that's the idea without @sandbox as well

[08:02:01.0000] <Dashiva>
And we all know how well that works

[08:03:00.0000] <Philip`>
Without @sandbox the browser wouldn't interpret the tag's content as (restricted) HTML that it should parse and render

[08:04:00.0000] <Dashiva>
Assume an author who doesn't actually understand sandbox, he's just heard about it and wants to use it

[08:05:00.0000] <Philip`>
If the author wants to display 'safe' HTML from users, today they'd have to find a large complex library that they don't understand, like html5lib's sanitizer or a zillion other libraries, and then pass the content through it and print it out

[08:06:00.0000] <Philip`>
With <span sandbox>, it'd be basically the same except they don't need a sanitizer library, they just need a good old plain HTML escape function, and the browser is responsible for proper sanitization

[08:09:00.0000] <Philip`>
(Most XSS holes seem to come from people forgetting about escaping entirely, and letting unrestricted user-generated content leak out, and we can't stop that; but the idea is to make it easier when people do want to properly display content safely)

[08:09:01.0000] <Dashiva>
I'm wondering if the apparently simplicity will lead to more people trying to do it (and thus more people doing it incorrectly, instead of not doing it at all)

[08:10:00.0000] <Dashiva>
The idea here seems to be "Author tests with markup, notices nothing renderering, goes to figure out how to do it properly, does it properly"

[08:11:00.0000] <Philip`>
I think someone suggested making the browser delete the content if it sees an unescaped " or ', so they should very quickly discover that they need to escape the content

[08:11:01.0000] <Dashiva>
That looks very similar to how well-formed XHTML is supposed to work, yet we get YSODs

[08:11:02.0000] <Philip`>
and the escaping is easy

[08:15:00.0000] <Philip`>
YSODs triggered by user-generated content are usually not due to the author forgetting to use their serialize-to-well-formed-XML function, they're due to the serialize-to-well-formed-XML function being really really hard to implement perfectly correctly

[08:16:00.0000] <Philip`>
Correctly escaping 5 ASCII characters is far easier

[08:17:00.0000] <Hixie>
doc="" only requires you to escape one (two for correctness, three (four?) for XML), which is why i kinda like that idea

[08:18:00.0000] <Philip`>
(I think I don't really mind XML draconianness as a concept, I just mind it when it's coupled with very complex rules (like forbidden ranges of characters and forbidden attribute values and forbidden strings of characters that are otherwise plain text etc))

[08:19:00.0000] <Philip`>
<span sandbox> only requires you to escape one (<) too, I think

[08:19:01.0000] <Philip`>
(...in order to avoid attacks)

[08:39:00.0000] <Dashiva>
"omg interwebs collapsed" is an interesting bug tag

[09:08:00.0000] <hsivonen>
hmm. 3 Adobe reps have responded to the ISSUE 76 poll even though the chairs said they'd discount multiple replies from a single W3C Member

[09:10:00.0000] <Hixie>
they didn't say they'd discount multiple replies, they said they'd discount multiple identical replies. I haven't checked, but I assume the three reps each had different novel points that hadn't been raised before.

[09:10:01.0000] <hsivonen>
good point

[09:14:00.0000] <othermaciej>
what we said was if we count, all the reps from one company count as one vote, but we will also endeavor not to count that way at all

[09:15:00.0000] <jgraham>
othermaciej: Should we take Sam's statement that you should only vote if you intend to raise a FO should the vote not go your way seriously?

[09:19:00.0000] <jgraham>
(the specific quote is "What we are looking for is input from people within the group who seriously intend to take raise a formal objection, and the content of such objections.")

[09:26:00.0000] <jgraham>
/me guesses that later emails in that thread suggest that is not really the policy

[09:30:00.0000] <Dashiva>
It's an ideal

[09:34:00.0000] <jgraham>
It seems like the total opposite of an ideal

[09:34:01.0000] <jgraham>
Whatever that is. 

[09:35:00.0000] <jgraham>
(because decisions should be made on the basis of the weight of evidence, not the willingness of people to flight through Process to get their way)

[09:36:00.0000] <jgraham>
*fight

[09:36:01.0000] <Dashiva>
Well, that's just according to your personal ideals

[09:39:00.0000] <Hixie>
what i find interesting so far is that the poll seems to suggest manu's change proposal didn't do a good job of conveying why people felt it should be split out, while tab's did convey why people felt it should not be split out

[09:39:01.0000] <Hixie>
it's not clear to me what this means about the process per se

[09:39:02.0000] <gsnedders>
It's a writing competition!

[09:40:00.0000] <jgraham>
Dashiva: Well it doesn't seem hard to make an argument that using weight-of-evidence will tend to lead to a good specification whereas basing decisions on the degree to which various people have sociopathic tendencies will typically not

[09:41:00.0000] <othermaciej>
jgraham: it would be good for those *would* raise a Formal Objection to state their intent, but per the rules of the poll itself, if you have any objection you feel is strong, you can state it even if you would not raise a Formal Objection after but instead would choose to move on

[09:41:01.0000] <jgraham>
(note: I am not suggesting any real people have sociopathic tendencies. I am just using the term to indicate a willingness to continue to escalate their objections in the face of overwhelming evidence to the contrary. I don't think it is quite the right term to use but there we go)

[09:42:00.0000] <jgraham>
s/their//

[09:43:00.0000] <Hixie>
othermaciej: wait so if we _di_ want

[09:43:01.0000] <Hixie>
er

[09:43:02.0000] <Hixie>
othermaciej: wait so if we _would_ raise an FO we _should_ comment even if our argument has already been made?

[09:45:00.0000] <othermaciej>
it hurts my brain a little to explain this, because the chairs did not see 100% eye-to-eye on this so we ended up with a slightly squishy policy

[09:45:01.0000] <jgraham>
(and s/evidence to the contrary/evidence that their viewpoint failed to be convincing to others/)

[09:46:00.0000] <Philip`>
othermaciej: The chairs should have had a poll to decide on a non-squishy policy

[09:46:01.0000] <othermaciej>
it's not forbidden to make repeat arguments, just not (generally) necessary; however, intent to raise a formal objection can be useful info

[09:46:02.0000] <othermaciej>
Philip`: turtles all the way down, eh?

[09:47:00.0000] <jgraham>
We should have a squishy chair. Then we could all sit on it and be comfortable

[09:47:01.0000] <Philip`>
It's fewer and fewer turtles as you go down, and it's less painful to get agreement between a handful of turtles than between several hundred turtles

[09:48:00.0000] <Philip`>
and that's less painful that getting agreement between ten million turtles

[09:48:01.0000] <Philip`>
*than

[09:48:02.0000] <Hixie>
personally i imagine that whether or not google would raise a formal objection would depend on the reasoning behind the resulting decision, whichever way it goes, as with everything else

[09:48:03.0000] <Hixie>
so it's hard to say ahead of time

[09:48:04.0000] <Philip`>
So, to cause maximum pain, we should avoid making a decision and let authors choose between RDFa and microdata

[10:01:00.0000] <Hixie>
/me wonders how to (or whether to) do <device>

[10:02:00.0000] <jgraham>
How would it work?

[10:02:01.0000] <jgraham>
/me doesn't follow DAP 

[10:03:00.0000] <Hixie>
that's the question

[10:03:01.0000] <Hixie>
<device type=camera oncamera="myVideo.src = event.stream.URL"> is one option

[10:03:02.0000] <Hixie>
it shows a well

[10:03:03.0000] <Hixie>
that you click to get a window that lets you select a device

[10:04:00.0000] <jgraham>
So the point of <device> would be to give you browser-native UI?

[10:04:01.0000] <Hixie>
effectively "plugging" that device into the well

[10:04:02.0000] <Hixie>
a point, yes

[10:04:03.0000] <Hixie>
btu not click-jackable

[10:04:04.0000] <Hixie>
similar to <input type=file>

[10:05:00.0000] <Hixie>
(but not submittable)

[10:08:00.0000] <daedb>
Would <device> be able to handle joysticks/joypads? For web-based games and such...

[10:08:01.0000] <Hixie>
seems like those should just work, no?

[10:08:02.0000] <Hixie>
there's no security reason to make them not work as far as i can tell

[10:09:00.0000] <jgraham>
Hmm. I guess I would expect to have a device activated by an author-supplied UI element and a one of those dropdown bars that are fashionable these days

[10:09:01.0000] <Hixie>
doesn't scale

[10:09:02.0000] <jgraham>
Why?

[10:09:03.0000] <Hixie>
see public-device-apis for more discussion on that

[10:10:00.0000] <daedb>
I'm just curious because I don't know of any way to handle joypad input on a web page (unless you remap the buttons to keyboard buttons using Joytokey or other programs).

[10:10:01.0000] <jgraham>
Do you have a one irc message summary?

[10:10:02.0000] <Hixie>
http://lists.w3.org/Archives/Public/public-device-apis/2009Dec/0149.html

[10:10:03.0000] <othermaciej>
Hixie: how persistent is the permission grant?

[10:10:04.0000] <Hixie>
othermaciej: one of many questions

[10:10:05.0000] <othermaciej>
Hixie: also what types of devices are you imagining besides camera and microphone (if any)? wondering how much this generalizes

[10:10:06.0000] <Hixie>
further good questions :-)

[10:11:00.0000] <othermaciej>
if it was only camera and microphone I might just have <avinput> or <mediainput> or <videoinput>/<audioinput> instead of an apparently wildly general element

[10:11:01.0000] <Hixie>
jgraham: (also http://lists.w3.org/Archives/Public/public-device-apis/2009Dec/0194.html)

[10:11:02.0000] <othermaciej>
but that depends on whether there are other kinds of devices and if they offer similar semantics, API and user experience

[10:11:03.0000] <Hixie>
indeed

[10:12:00.0000] <othermaciej>
for example - does <device type=disk> make sense? probably not

[10:12:01.0000] <othermaciej>
nor <device type=ethernet>

[10:16:00.0000] <Hixie>
<device type=media-player> might make sense, to enable syncing-like apps

[10:16:01.0000] <othermaciej>
what would that offer access to?

[10:17:00.0000] <Hixie>
e.g. to enable an iTunes-like app to sync music to a music player

[10:17:01.0000] <Hixie>
maybe it would expose a file system

[10:17:02.0000] <othermaciej>
I see, so it would expose an attached external media player?

[10:17:03.0000] <Hixie>
or maybe it would expose a websocket-like interface

[10:18:00.0000] <Hixie>
<device type=rs232> would enable many legacy apps to be ported to the web, too

[10:18:01.0000] <othermaciej>
it seems like by the time this is popular, your most likely media player choice would be something capable enough that you could sync through The Cloud(tm)

[10:19:00.0000] <othermaciej>
but I'm not sure Apple would fully expose iPhone/iPod media collections either to the browser on the device or to the browser on a machine tethered to the device

[10:19:01.0000] <Hixie>
no, indeed

[10:20:00.0000] <othermaciej>
(though the cloud syncing approach could work just by running your web app on both your computer and your phone/ipod/media player thingie)

[10:50:00.0000] <Hixie>
ms2ger is awesome

[10:50:01.0000] <Hixie>
if we do go with this <Device> idea

[10:51:00.0000] <Hixie>
i wonder whether to use a different event for each type

[10:51:01.0000] <Hixie>
or the same event with a complex event interface

[10:51:02.0000] <bumblebird>
hey Hixie, I wrote the post about WebSocket vs xhr last wk.

[10:51:03.0000] <Hixie>
hi

[10:51:04.0000] <bumblebird>
just saw your reply. Cool to know compression will be looked at later on, and sorry if the 'spec is insane' comment grated :)

[10:52:00.0000] <Hixie>
no worries :-) several people have said that, you're not alone. it's just annoying that most people who say that point to specs that i think are insane as examples of the right way to do things :-)

[10:53:00.0000] <bumblebird>
sure. I can't really remember a spec that I didn't think was overly verbose etc TBH,

[10:54:00.0000] <bumblebird>
do you happen to know of any other browsers with WebSocket support? I've seen some references to it being in firefox trunk, but can't quite find it yet

[10:54:01.0000] <Hixie>
firefox devs are working on it, don't think it's ready yet though

[10:54:02.0000] <Hixie>
webkit trunk has it

[10:54:03.0000] <bumblebird>
ah ok, cool

[10:54:04.0000] <bumblebird>
so it should make it into safari etc in due course I expect?

[10:54:05.0000] <Hixie>
depends on what happens to the spec, probably

[10:55:00.0000] <bumblebird>
there was a load of discussion on hackernews+irc about it. People complaining that it should just provide raw socket support

[10:56:00.0000] <bumblebird>
FWIW I think the spec is pretty solid. Compression is the only thing missing for me

[10:56:01.0000] <Hixie>
i think this is as close to raw sockets we'll ever get

[10:56:02.0000] <Hixie>
given the security issues

[10:56:03.0000] <bumblebird>
sure. I don't think raw sockets would be good. You'd end up with 50 js libs to do data framing etc

[10:57:00.0000] <bumblebird>
and this solves the use case of having a bidirectional pipe to a webserver just fine

[10:57:01.0000] <bumblebird>
some people complain whatever :)

[10:57:02.0000] <Hixie>
hah

[10:57:03.0000] <bumblebird>
and agreed @ security concerns

[11:04:00.0000] <Philip`>
Hmm... Possible <canvas> information leakage: Find a browser with optimised alpha=255 drawing routines. Draw a one-pixel section of a different-origin image lots of times and time it. Then you can work out which pixels are transparent and which are solid.

[11:05:00.0000] <Philip`>
and if the image is the text "Today's intranet password is: PAfqGK3Rc0PwhApe" encoded as a PNG with black text on a transparent background then that could be bad

[11:15:00.0000] <gsnedders>
Philip`: Man, you come up with some obscure edge cases

[11:19:00.0000] <Philip`>
It's probably less obscure with WebGL, where you could (I think) write a shader that runs an expensive loop iff the texture pixel is black

[11:31:00.0000] <Hixie>
can't you get data out of a shader other than timing effects?

[12:22:00.0000] <Philip`>
Hixie: No, as far as I'm aware

[12:23:00.0000] <Philip`>
The ability to extract data (other than via the output buffers) would be incompatible with the notion of automatic highly-parallel execution

[12:25:00.0000] <Philip`>
(I think things like CUDA let you write to something like system memory directly from the GPU, but normal shaders are just side-effect-free functions from input pixels and textures and constants to output pixels, I think)

[12:25:01.0000] <Philip`>
(I could think wrongly, of course)


2009-12-15
[04:10:00.0000] <hsivonen_>
Does anyone happen to have a demo showing how document.write from a timeout during the parse behaves in IE?

[04:10:01.0000] <hsivonen_>
it blows away the document, right?

[04:13:00.0000] <zcorpan_>
i'm pretty sure it does

[04:13:01.0000] <hsivonen_>
thanks

[04:17:00.0000] <zcorpan_>
hsivonen_: did you see the script bug?

[04:17:01.0000] <hsivonen_>
zcorpan_: I didn't. What's the bug?

[04:18:00.0000] <zcorpan_>
hsivonen_: i sent an email. <script><!--<script></script>x</script>y

[04:18:01.0000] <zcorpan_>
the </script>x does not appear in the dom

[04:18:02.0000] <hsivonen_>
zcorpan_: thanks

[04:19:00.0000] <hsivonen_>
/me sees the PFWG has sent out responses to ARIA review comments

[04:19:01.0000] <hsivonen_>
/me needs to review the comments

[04:19:02.0000] <hsivonen_>
s/comment/responses/

[04:23:00.0000] <annevk2>
yeah me too

[04:23:01.0000] <annevk2>
at some point

[04:23:02.0000] <annevk2>
too much work :/

[04:23:03.0000] <hsivonen_>
zcorpan_: I think the bug you found is the same as https://bugzilla.mozilla.org/show_bug.cgi?id=533111

[04:24:00.0000] <hsivonen_>
zcorpan_: your test case is much nicer :-)

[04:26:00.0000] <zcorpan_>
hsivonen_: looking at comment 3, it appears to be different

[04:33:00.0000] <hsivonen_>
/me wonders what possesses Web authors to call document.write() from timeouts

[05:30:00.0000] <hsivonen>
gsnedders: is html5lib missing test coverage for the new script states?

[05:30:01.0000] <gsnedders>
hsivonen: No

[05:30:02.0000] <gsnedders>
hsivonen: Or at least, not much

[05:30:03.0000] <hsivonen>
gsnedders: hmm. I don't see script tests in escapeFlag.test or in contentModelFlags.test

[05:31:00.0000] <gsnedders>
hsivonen: test16.dat

[05:31:01.0000] <hsivonen>
gsnedders: ah. excellent. thanks

[05:32:00.0000] <gsnedders>
hsivonen: Complains about missing tests should be directed at zcorpan

[05:33:00.0000] <hsivonen>
eek. I see lots of test failures

[05:33:01.0000] <gsnedders>
In html5lib?

[05:34:00.0000] <hsivonen>
in V.nu

[05:34:01.0000] <gsnedders>
Or in validator.nu?

[05:34:02.0000] <gsnedders>
(html5lib test failures are caused by a lot of tests not having the correct expected parse errors)

[05:34:03.0000] <hsivonen>
so I'm noticing

[05:35:00.0000] <gsnedders>
/me turned that on as it turned up several parsing bugs

[05:35:01.0000] <hsivonen>
is it ok to push test fixes without fixing html5lib itself?

[05:35:02.0000] <gsnedders>
Yeah sure

[05:35:03.0000] <hsivonen>
ok

[05:51:00.0000] <zcorpan>
shouldn't a loop in structured clone throw INVALID_STATE_ERR instead of NOT_SUPPORTED_ERR?

[05:52:00.0000] <Philip`>
Loops aren't invalid, they just aren't yet supported by the algorithm

[05:55:00.0000] <zcorpan>
Hixie: ^

[05:56:00.0000] <annevk2>
what's wrong with loops?

[05:57:00.0000] <zcorpan>
it's hard to get a clone

[05:58:00.0000] <annevk2>
but technically you can clone them I think

[05:59:00.0000] <Philip`>
It's not that hard to implement, you just change the 'memory' list so that for each object you store a reference to its clone, and if you're cloning something that's already in the memory then you copy the stored reference

[06:19:00.0000] <zcorpan>
should we change the algorithm to support loops?

[06:22:00.0000] <Philip`>
(If it supports loops, it should also support x=[]; y=[x,x]; clone(y) without breaking the y[0]==y[1] equality)

[06:22:01.0000] <Philip`>
(which is different to the current behaviour)

[06:34:00.0000] <boblet>
hey Mike

[06:43:00.0000] <MikeSmith>
boblet: you're definitely welcome to stay at my place

[06:43:01.0000] <zcorpan>
seems like that's the current behavior in our impl

[06:43:02.0000] <boblet>
MikeSmith: thanks. sorry for the late notice (!)

[06:44:00.0000] <MikeSmith>
np

[06:44:01.0000] <MikeSmith>
don't need to bring anything

[06:45:00.0000] <boblet>
ok. see you at the event then

[06:47:00.0000] <MikeSmith>
boblet: yep, see you there

[06:50:00.0000] <MikeSmith>
zcorpan: I've been looking at implementing xml-sheetsheet PI checking (bug 14)

[06:51:00.0000] <MikeSmith>
talking with Yudai about it too

[06:52:00.0000] <MikeSmith>
Yudai is spending some time part-time as an intern and sysadmin at W3C/Keio

[06:52:01.0000] <zcorpan>
MikeSmith: cool

[06:53:00.0000] <MikeSmith>
zcorpan: one question, for cases like <?xml-stylesheet foo bar?>

[06:53:01.0000] <MikeSmith>
what error should we emit?

[06:54:00.0000] <zcorpan>
MikeSmith: an error with a message that says the pi failed to parse

[06:56:00.0000] <zcorpan>
MikeSmith: from the spec's perspective, it's a potential xml-stylesheet pi whose parsing results in an error

[06:57:00.0000] <MikeSmith>
zcorpan: understood. but I'm not sure that's the most helpful thing to say to end uses

[06:57:01.0000] <MikeSmith>
*users

[06:57:02.0000] <zcorpan>
MikeSmith: or rather, it's a pi with target xml-stylesheet that's not an xml-stylesheet pi (because it failed to parse)

[06:57:03.0000] <MikeSmith>
zcorpan: what I have it emitting now is, "Potential xml-stylesheet processing instruction is ill-formed."

[06:57:04.0000] <MikeSmith>
I guess I could make that "Potential xml-stylesheet processing instruction failed to parse."

[06:58:00.0000] <zcorpan>
if you could say why it failed to parse, that's probably be more helpful

[07:00:00.0000] <zcorpan>
MikeSmith: the message from the xml parser in v.nu is "required character (found b) (expected =)"

[07:01:00.0000] <zcorpan>
for <html foo bar>

[07:09:00.0000] <hsivonen>
Hixie: is there any chance of getting dreamhost to fix the gzip use on lists.whatwg.org for range requests?

[07:23:00.0000] <Philip`>
"Can sites host their own copy of foaf.html to avoid failure in communication when example.org is down?" - hmm, sadly http://uptime.netcraft.com/up/graph?site=example.org doesn't seem to give uptime data for it

[07:31:00.0000] <TabAtkins>
I am the luckiest man alive.  I won a TV on Friday, and a pie this morning.  I'm looking for a lotto ticket next.

[07:32:00.0000] <jgraham>
You might be hit by a bus. That is also a rather unlikely event so presumably indicates luck

[07:32:01.0000] <TabAtkins>
Nonono, good luck and bad luck are clearly separate things.

[07:33:00.0000] <zcorpan>
but both are still luck, no?

[07:33:01.0000] <TabAtkins>
Only in the abstract sense.  They're not commensurable unless you believe in karma.

[07:35:00.0000] <Philip`>
Regardless of whether it counts as luck, you would no longer be the luckiest man alive

[07:35:01.0000] <jgraham>
You have to hope that people who believe in karma don't get lucky and get you as a textbook example of karma in action

[07:35:02.0000] <TabAtkins>
Philip`: No jinxing.  >_<

[07:35:03.0000] <jgraham>
Philip`: Depends on the effects of the bus

[07:37:00.0000] <Philip`>
Well, I suppose it could be a magic bus which transforms into a field full of sunflowers when it hits somebody

[07:38:00.0000] <annevk2>
not all the ARIA replies seem satisfactory

[07:38:01.0000] <annevk2>
e.g. defining hidden element as done here http://www.w3.org/WAI/PF/aria/20091214/terms#def_hidden does not really seem implementable

[07:41:00.0000] <zcorpan>
hmm, definition reference to wcag2

[07:42:00.0000] <annevk2>
and with respect to my datatype comment the reply says it depends on the host language...

[07:43:00.0000] <annevk2>
so you need a separate ARIA attribute parser for SVG and HTML?

[07:43:01.0000] <annevk2>
and for any new language?

[07:49:00.0000] <gsnedders>
/me writes nonsense in his email to public-html

[07:52:00.0000] <MikeSmithX>
zcorpan: so the initial implementation of xml-stylesheet PI checking that I tried out involved internally transforming the PI into a document/element and then parsing that

[07:52:01.0000] <zcorpan>
MikeSmithX: ok

[07:52:02.0000] <MikeSmithX>
that was kind of what you suggested

[07:53:00.0000] <zcorpan>
yeah

[07:53:01.0000] <MikeSmithX>
and for that in the ill-formed pseudo-attribute case, Xerces says: 'Attribute name "foo" associated with an element type "xml-stylesheet" must be followed by the ' = ' character.'

[07:53:02.0000] <MikeSmithX>
which I guess is roughly same as your "required character (found b) (expected =)" message

[07:54:00.0000] <MikeSmithX>
zcorpan: anyway, for the actual end-user message, I think I will just go with "...failed to parse."

[07:54:01.0000] <MikeSmithX>
for that case

[07:55:00.0000] <zcorpan>
yeah, that's probably good enough

[07:56:00.0000] <MikeSmithX>
btw, I thought about the idea that you suggested of having an xml-stylesheet PI mini-schema and then validating the instances against that

[07:56:01.0000] <MikeSmithX>
but

[07:56:02.0000] <MikeSmithX>
the problem with that would be that the error messages would call the pseudo-attributes "attributes"

[07:57:00.0000] <MikeSmithX>
which they aren't of course

[07:57:01.0000] <zcorpan>
can't you post-process the message?

[07:57:02.0000] <MikeSmithX>
no, not easily

[07:58:00.0000] <MikeSmithX>
if we could there would be a lot of other places where it'd already be doing it

[07:59:00.0000] <MikeSmithX>
it's one of the tradeoffs in using an off-the-shelf general validation mechanism

[07:59:01.0000] <MikeSmithX>
as opposed to doing all the checking with custom code

[07:59:02.0000] <MikeSmithX>
Henri talks about it in his thesis

[07:59:03.0000] <MikeSmithX>
anyway, it's not a big problem in this case

[08:00:00.0000] <MikeSmithX>
I'm mostly just being lazy

[08:01:00.0000] <MikeSmithX>
and trying to figure out the easiest way to parse the set of pseudo-attributes without needing to write specific parsing code to do it

[08:01:01.0000] <zcorpan>
i guess it's not so bad if the messages say "attribute" instead of "pseudo-attribute"

[08:02:00.0000] <MikeSmithX>
zcorpan: maybe, but not sure hsivonen would agree about that

[08:04:00.0000] <zcorpan>
MikeSmithX: what happens currently for pis in the internal subset?

[08:04:01.0000] <MikeSmithX>
zcorpan: I don't know because I've not looked at the xmlparser code much

[08:05:00.0000] <MikeSmithX>
zcorpan: but I think the xmlparser code is mostly just doing whatever aelfred does

[08:05:01.0000] <zcorpan>
MikeSmithX: do you have a running prototype?

[08:06:00.0000] <MikeSmithX>
zcorpan: I can have it up on qa-dev shortly

[08:06:01.0000] <MikeSmithX>
well, relatively shortly

[08:06:02.0000] <MikeSmithX>
I need a break first

[08:07:00.0000] <zcorpan>
ok

[08:07:01.0000] <MikeSmithX>
after two hours of phone conferences

[08:07:02.0000] <MikeSmithX>
but I will ping you again when I have it up

[09:56:00.0000] <MikeSmithX>
if zcorpan comes back around, please let him know that http://qa-dev.w3.org:8888/ now has a minimal xml-stylesheet PI checker

[09:58:00.0000] <MikeSmithX>
it doesn't report anything yet except the simplest case, which is if code fails to parse the PI at all

[10:32:00.0000] <gsnedders>
'The "rows' and "cols" attributes aren't presentational, they give the wrapping edge of the textarea, which affects submission' — how?

[10:33:00.0000] <gsnedders>
Based upon the wrap attribute?

[10:33:01.0000] <gsnedders>
So it's only something new in HTML 5 that has any affect?

[10:58:00.0000] <Philip`>
gsnedders: wrap isn't new in HTML5

[10:58:01.0000] <Philip`>
It's existed at least since IE4

[10:59:00.0000] <gsnedders>
Ah

[11:46:00.0000] <AryehGregor>
gsnedders, well, cols is used for wrap=hard.  Doesn't seem rows is used for anything.

[11:47:00.0000] <AryehGregor>
Does CSS have a good way to emulate them?  Would you have to manually multiply the font-size times the line-height or something?

[11:48:00.0000] <AryehGregor>
That's just for rows, cols I have no idea on (but it's maybe less important, since textareas are more often wide than tall, and in a wide textarea it's not so noticeable if you have some extra empty pixels at the side)

[11:48:01.0000] <TabAtkins>
font-size times line-height?  ?_?  Just line-height times the number of rows you want.

[11:49:00.0000] <TabAtkins>
(Unless you are expressing your line height relative to your font-size, of course.

[11:49:01.0000] <TabAtkins>
cols are much less important, but you can typically estimate it with ems.

[11:50:00.0000] <AryehGregor>
Hmm, true, but you still have to manually check what that's supposed to be, right?

[11:50:01.0000] <AryehGregor>
The line-height, I mean.

[11:50:02.0000] <AryehGregor>
You can't just use ems or something, in other words.

[11:51:00.0000] <TabAtkins>
Yes.

[11:51:01.0000] <TabAtkins>
If your line-height is the standard 1.2, you'll have to do height:(1.2 * number of rows)em;

[11:52:00.0000] <TabAtkins>
At least, that should approximately work.

[11:52:01.0000] <AryehGregor>
Wow, Mozilla has a heck of a lot of bugs.  211 results for "textarea".

[11:53:00.0000] <AryehGregor>
Hmm, this looks relevant: https://bugzilla.mozilla.org/show_bug.cgi?id=502462

[11:53:01.0000] <AryehGregor>
/me has noted that <textarea maxlength> doesn't work in Firefox

[11:54:00.0000] <TabAtkins>
Nah, that bug's explanation is "<textbox multiline> doesn't work with maxlength because it's an html:textarea, which doesn't work with maxlength".

[11:55:00.0000] <AryehGregor>
That implies something to do with html:textarea not supporting maxlength, which is promising.

[12:05:00.0000] <zcorpan_>
MikeSmithX: i don't get any messages regarding for <?xml-stylesheet foo bar?>

[12:18:00.0000] <csarven>
Are localStorage keys sandboxed to the hosts in which they are created at or can they be shared?

[12:19:00.0000] <csarven>
Apparently they are

[12:37:00.0000] <TabAtkins>
Heh, there's nothing like demonstrating a working prototype of a proposal to make it really obvious why it's needed.

[12:40:00.0000] <AryehGregor>
/me is somewhat surprised that he was able to hack a trivial feature into Firefox in, like, an hour

[12:41:00.0000] <AryehGregor>
/me should try hacking on other projects more often

[12:49:00.0000] <mpt>
AryehGregor, what OS(es) do you use?

[12:49:01.0000] <AryehGregor>
mpt, Linux mainly, but I have Windows lying around too.

[12:50:00.0000] <AryehGregor>
(Ubuntu specifically)

[12:50:01.0000] <mpt>
ah

[12:50:02.0000] <AryehGregor>
Why do you ask?

[12:51:00.0000] <mpt>
We could always use some help with the Ubuntu Software Center. :-) Especially if you know your way around WebKit

[12:51:01.0000] <AryehGregor>
I don't, I'm a MediaWiki developer.

[12:51:02.0000] <AryehGregor>
If I were to branch out it would probably be into browser engines, though.

[12:53:00.0000] <mpt>
ok

[13:29:00.0000] <TabAtkins>
AryehGregor: What did you hack?

[13:37:00.0000] <virtuelv>
<offtopic>

[13:37:01.0000] <virtuelv>
I don't understand the software center

[13:38:00.0000] <virtuelv>
I install stuff once, and mostly won't touch the installation after that

[13:55:00.0000] <AryehGregor>
TabAtkins, I tried adding maxlength support to textarea.  The patch is five lines and seems to work.

[13:56:00.0000] <Philip`>
AryehGregor: You can't fit test cases inside five lines

[13:56:01.0000] <Philip`>
so you need to do more patching :-p

[13:56:02.0000] <AryehGregor>
Yes, I know that part is missing.

[13:56:03.0000] <AryehGregor>
I have to figure that out.

[13:56:04.0000] <AryehGregor>
I'll have to look at what kinds of tests there are for maxlength on inputs.

[13:57:00.0000] <Philip`>
/me would not be entirely surprised if there weren't any

[13:57:01.0000] <Philip`>
(but nor would I be surprised if there were)

[13:57:02.0000] <AryehGregor>
I looked at some recent commits and not all of them seemed to have tests attached.

[13:59:00.0000] <Philip`>
I think there weren't really any tests at all until a very small number of years ago

[13:59:01.0000] <Philip`>
and it's unlikely anyone would have bothered adding maxlength tests since then, unless as part of recent bug fixes to it

[14:04:00.0000] <AryehGregor>
The code ends up being just "enable the input maxlength implementation for textareas", as far as I can tell, so maybe it doesn't really need independent tests.  Perhaps I should just submit the patch and ask if it needs tests.

[14:06:00.0000] <Philip`>
AryehGregor: Everything should have tests :-)

[14:07:00.0000] <Philip`>
(That's probably not Mozilla policy, though)

[14:07:01.0000] <AryehGregor>
MediaWiki has an excellent test suite.  However, it's not automated.  You see, we use a wiki-based approach to testing: enable it on Wikipedia and wait for people to complain.

[14:07:02.0000] <TabAtkins>
!_!  What the crap, Firefox?  It's for some reason wrapping my <thead> across three rows in the table.

[14:07:03.0000] <Philip`>
(but it's common sense :-p )

[14:08:00.0000] <Philip`>
(Otherwise someone will break the feature and nobody will notice)

[14:08:01.0000] <AryehGregor>
Obviously then it wasn't an important feature anyway.

[14:08:02.0000] <AryehGregor>
Hardly worth the effort.

[14:09:00.0000] <Philip`>
Nobody noticed when Firefox completely broke canvas drawImage, until after it was released

[14:09:01.0000] <AryehGregor>
I know a programmer who works on bond-trading software.  He said he's had integer overflow when putting a sum of money in a 32-bit int.  And he doesn't use automated test cases.  So clearly some stupid *wiki* doesn't have to.

[14:10:00.0000] <AryehGregor>
:)

[14:10:01.0000] <AryehGregor>
See, that's one nice thing about developing MediaWiki, we barely have releases and don't care much about them.

[14:10:02.0000] <AryehGregor>
Break something?  Bug a sysadmin and get the fix synced to Wikipedia, no need to worry about anyone else unless it's a security problem.

[14:11:00.0000] <Philip`>
I guess it's easier to deploy updates when the software is on a handful of servers rather than on a hundred million users' PCs, so you don't need to be as careful

[14:12:00.0000] <AryehGregor>
Well, Mozilla can push out updates too if they feel like it.

[14:13:00.0000] <AryehGregor>
Things like layout engines are probably a lot more complicated and fragile than most code, though.

[14:13:01.0000] <AryehGregor>
And of course they actually have to worry about competition.

[14:13:02.0000] <AryehGregor>
:)

[14:14:00.0000] <Philip`>
I think it's not just a case of "feel like it", given the time and effort and bandwidth and user irritation involved in pushing updates :-)

[14:27:00.0000] <TabAtkins>
Wow.  Check out http://www.zachstronaut.com/posts/2009/02/17/animate-css-transforms-firefox-webkit.html in Firefox.  The text in the middle box constantly dances.

[14:30:00.0000] <AryehGregor>
Firefox bookmarklet from blog.johnath.com: javascript:document.body.style.MozTransform="rotate(90deg)";void(0);

[14:30:01.0000] <AryehGregor>
Completely pointless, but amusing!

[14:30:02.0000] <AryehGregor>
javascript:document.body.style.WebkitTransform="rotate(90deg)";void(0); works for WebKit.

[14:35:00.0000] <othermaciej>
all the browsing gangstas hold their web pages sideways

[14:36:00.0000] <roc>
AryehGregor: for something that's easy to test automatically, like <textarea maxlength>, the reviewer would demand tests, I think

[14:37:00.0000] <AryehGregor>
roc, okay, but what should I test?  The existence of .maxLength is easy to do in just JS, I guess.  What sort of test would I have to write to test "it doesn't let me type more than X characters", a Mochitest?

[14:37:01.0000] <TabAtkins>
othermaciej: Hahaha.

[14:38:00.0000] <AryehGregor>
roc, also, should I request review from you, or is there anyone else specific you'd recommend?

[14:38:01.0000] <roc>
patches that land without tests are generally either not supposed to be changing behavior (although some of those have tests anyway), or are changing something that's very hard to test in a robust manner (e.g. exactly what pixels end up on the screen due to some bizarro Win32 windowing API call), or are security bugs and the test is an exploit

[14:38:02.0000] <AryehGregor>
Okay.

[14:39:00.0000] <roc>
I can probably review it

[14:39:01.0000] <roc>
although it might be better to get jst⊙mo

[14:40:00.0000] <roc>
you probably want to write a mochitest and use EventUtils.js with waitForFocus and synthesizeKeyEvent

[14:40:01.0000] <roc>
(and thanks!!!)

[14:40:02.0000] <AryehGregor>
:)

[14:44:00.0000] <TabAtkins>
Argh why is it so impossible to create vertical table headers. >_<

[14:49:00.0000] <Hixie>
http://www.w3.org/mid/OF3B70DB4D.2F7BF1C9-ON8625768D.0078BE03-8625768D.007A8C21⊙uic may be interesting to implementors and authors interested in <canvas>

[14:49:01.0000] <Hixie>
(please send feedback straight to that list)

[14:57:00.0000] <zcorpan>
Hixie: doesn't the device example need <video autoplay>?

[15:01:00.0000] <Hixie>
zcorpan: probably wouldn't hurt, indeed

[15:14:00.0000] <fantasai>
Anyone know why html5lib's documentation page returns 404?

[15:28:00.0000] <Hixie>
i just checked in r4444

[15:28:01.0000] <Hixie>
four fours! it's gonna be a long time until r55555


2009-12-16
[16:32:00.0000] <Hixie>
hrm

[16:32:01.0000] <Hixie>
it's not clear how to proceed on the <device> thing until we have a codec

[17:38:00.0000] <AryehGregor>
Oh, I wanted to link roc to <https://bugzilla.mozilla.org/show_bug.cgi?id=535043>.  Oh well.

[00:57:00.0000] <jgraham>
fantasai: Interesting. Al the documentation seems to have vanished

[00:59:00.0000] <jgraham>
Still in svn though: http://html5lib.googlecode.com/svn/wiki/UserDocumentation.wiki

[01:00:00.0000] <jgraham>
I wonder if we are supposed to migrate it to mercurial somehow

[01:29:00.0000] <hsivonen>
gsnedders: do we really need contentModelFlags.test to support initializing the tokenizer to an arbitrary?

[01:29:01.0000] <hsivonen>
gsnedders: also, what's spec for initialStates?

[01:32:00.0000] <hsivonen>
whew. the failures went away when I fixed the test harness to use the new test case syntax

[01:32:01.0000] <hsivonen>
except a couple of real failures I had introduces in the tokenizer

[01:35:00.0000] <hsivonen>
hg outgoing says

[01:35:01.0000] <hsivonen>
(deprecated '%' in path default-push=https://hsivonen%%40iki.fi⊙hgc/hg/ from /Users/Shared/Projects/html5lib/.hg/hgrc:3)

[01:35:02.0000] <hsivonen>
should I take one of the % signs away?

[01:35:03.0000] <hsivonen>
apparently, yes

[02:10:00.0000] <mia^^>
hi

[02:10:01.0000] <jgraham>
The W3C should really do whatever is needed to update WebIDL under /TR/

[02:11:00.0000] <mia^^>
I am here to learn

[02:12:00.0000] <annevk2>
it's not really up to the W3C

[02:12:01.0000] <annevk2>
more up to weinig

[02:13:00.0000] <Philip`>
I thought the whole purpose of /TR/ was for hosting outdated documents

[02:13:01.0000] <jgraham>
Doesn't there have to be some sort of group decision or so?

[02:13:02.0000] <hsivonen>
mia^^: welcome

[02:14:00.0000] <jgraham>
Anyway having a draft that is greater then 1 year out of date under TR seems unusually poor (and would certainly faila per-document rather than per-group heartbeat requirement)

[02:14:01.0000] <hsivonen>
/me wonders what crazy UA sniffing WebKit and Opera are going to face when they try to implement <script defer>

[02:14:02.0000] <mia^^>
hsivone thank you 

[02:16:00.0000] <mia^^>
bbs phone

[02:16:01.0000] <jgraham>
mia^^: Anything in particular you are here to learn, or just trying to soak up the ambiance?

[02:16:02.0000] <annevk2>
jgraham, the reason is that weinig wanted to update some more things before pushing out a new version

[02:16:03.0000] <annevk2>
jgraham, based on TC39 comments or something

[02:16:04.0000] <hsivonen>
does ES5 standardize the IE @cc_on stuff/

[02:16:05.0000] <hsivonen>
?

[02:17:00.0000] <hsivonen>
/me wonders why the document.write in http://www.mnot.net/templates/hinclude.js runs in Gecko

[02:17:01.0000] <jgraham>
annevk2: I would prefer "push early puish often" as a policy. The current situation confuses the hell out of e.g. Google

[02:17:02.0000] <jgraham>
hsivonen: @cc_on?

[02:18:00.0000] <hsivonen>
jgraham: JScript conditional comments

[02:18:01.0000] <jgraham>
No

[02:18:02.0000] <hsivonen>
hmm.

[02:19:00.0000] <annevk2>
jgraham, sure, I'm just explaining the situation

[02:21:00.0000] <annevk2>
Hixie, what if you invoke StreamRecorder.stop() several times? seems you get ever-larger file objects with the current text

[02:21:01.0000] <annevk2>
Hixie, you might want file objects per fragment

[02:24:00.0000] <annevk2>
http://dev.w3.org/2009/dap/camera/ looks quite ugly

[03:24:00.0000] <gsnedders>
Philip`: yt?

[03:37:00.0000] <Philip`>
gsnedders: y

[03:42:00.0000] <hsivonen>
oops. sorry. It seems I broke one test I thought I was fixing. I'll push a fix shortly.

[04:11:00.0000] <hsivonen>
did we agree that we want foster-parented text nodes to coalesce?

[04:16:00.0000] <jgraham>
hsivonen: I think we did but I'm not sure Hixie did

[04:17:00.0000] <hsivonen>
hmmkay. should I change the "excite me" test case now or later?

[04:17:01.0000] <gsnedders>
/me didn't think anyone apart from him and hsivonen was in that thread

[04:17:02.0000] <gsnedders>
_We_ did. Dunno about anyone else.

[04:17:03.0000] <jgraham>
Well I did too

[04:25:00.0000] <hsivonen>
I'll go ahead and change the test

[04:29:00.0000] <hsivonen>
is it intentional that we don't have a tree builder test file numbered 13?

[04:31:00.0000] <hsivonen>
lovely. I see test failures in script escapes and in the AAA

[04:38:00.0000] <gsnedders>
It used to exist, IIRC

[04:40:00.0000] <Philip`>
hg log says it never did

[04:40:01.0000] <hsivonen>
gsnedders: if you wrote it, did you forget to hg add?

[04:41:00.0000] <gsnedders>
No

[04:41:01.0000] <gsnedders>
I didn't write it

[04:41:02.0000] <gsnedders>
I thought it had existed once

[04:41:03.0000] <hsivonen>
whoever wrote 14 must have assumed 13 existed

[04:41:04.0000] <hsivonen>
/me goes see where 14 came from

[04:42:00.0000] <hsivonen>
jgraham: blame points to you

[04:43:00.0000] <jgraham>
I thought we once did and then it got moved to tests99 or so

[04:43:01.0000] <jgraham>
But it could just have been a mistake :)

[04:43:02.0000] <hsivonen>
99?

[04:44:00.0000] <jgraham>
There were some tests that you made for not-yet-specified behaviour that got moved to a magic test file

[04:44:01.0000] <annevk>
airplanes skip row 13 too

[04:44:02.0000] <annevk>
it confused me recently

[04:44:03.0000] <jgraham>
By "magic" I mean "numbered 99"

[04:45:00.0000] <hsivonen>
/me doesn't see a file numbered 99 in the repo

[04:46:00.0000] <hsivonen>
/me wonders if the list of bad numbers in the airplane row range has much more than 13 on it or if only skipping 13 is a culturally-biased thing and proper i18n would require skipping many more rows

[04:46:01.0000] <gsnedders>
Then that was spec'd and then it was moved

[04:57:00.0000] <hsivonen>
my script escape states are shamefully buggy

[04:57:01.0000] <hsivonen>
it's a small wonder they work for dogfooding in practice

[05:01:00.0000] <jgraham>
I think if you skipped all the numbers that some culture considers unlucky you would have to go to a totally different system like named rows or something

[05:34:00.0000] <hsivonen>
ok if I change tests to anticipate Hixie fixing http://www.w3.org/Bugs/Public/show_bug.cgi?id=8373 ?

[05:36:00.0000] <annevk>
jgraham, not sure that'd go down better :)

[06:07:00.0000] <hsivonen>
now I have only two classes of test case failures left:

[06:07:01.0000] <hsivonen>
1) surrogates (wontfix)

[06:07:02.0000] <hsivonen>
2) space characters in the AAA

[06:07:03.0000] <hsivonen>
/me doesn't like touching the AAA

[06:07:04.0000] <hsivonen>
also, we don't have enough test coverage for frameset-ok

[06:07:05.0000] <annevk>
surrogates?

[06:08:00.0000] <hsivonen>
annevk: the test cases test lone surrogates in JSON input

[06:08:01.0000] <hsivonen>
annevk: I've decided checking for those doesn't belong in the tokenizer

[06:08:02.0000] <hsivonen>
but belongs on a lower layer that doesn't participate with JSON

[08:34:00.0000] <Hixie>
annevk3: i intend to make stop() make the object dead so a second call won't do anything

[08:34:01.0000] <Hixie>
or might throw an exception

[08:36:00.0000] <annevk3>
k

[08:57:00.0000] <annevk3>
whoa, what's up with all the new accounts on wiki.whatwg.org?

[09:02:00.0000] <drry_>
/me changes topic to 'WHATWG: http://www.whatwg.org/ -- logs: http://krijnhoetmer.nl/irc-logs/ -- stats: http://gavinsharp.com/irc/whatwg.html -- Please leave your sense of logic at the door, thanks!'

[11:30:00.0000] <Hixie>
i don't understand why people have such an adverse reaction to the websocket protocol spec

[11:31:00.0000] <Hixie>
when probed, people always say it should be more like other specs, e.g. HTTP or the BWTP proposal, but that was exactly the kind of mistake I set out to avoid...

[11:31:01.0000] <Hixie>
(mistake because those specs describe the wire protocol, but generally don't define client or server requirements, and don't define error handling)

[11:51:00.0000] <franksalim>
Hixie, I think they are looking for a formal grammar describing WebSocket frames

[11:53:00.0000] <franksalim>
Hixie, which, unless I am missing something, is not in the latest draft

[11:53:01.0000] <Hixie>
what for?

[11:55:00.0000] <franksalim>
Hixie, to understand and discuss the frame structure. FWIW, I also think you are right to precisely describe client and server requirements and error handling

[11:55:01.0000] <franksalim>
without a grammar, you have to do a lot of reading to understand what the protocol would typically look like on the wire

[11:56:00.0000] <Hixie>
well, there's a diagram in section 1.2 showing the wire format

[11:56:01.0000] <Hixie>
but i guess i could add a non-normative abnf blob if that's what people want

[11:58:00.0000] <franksalim>
That is only my best guess of what people want

[12:05:00.0000] <franksalim>
judging from comments on hybi

[12:10:00.0000] <Hixie>
huh. i don't think you can describe the wire protocol as ABNF.

[12:11:00.0000] <Hixie>
/me uses the <> escape hatch

[12:18:00.0000] <annevk3>
oh man

[12:18:01.0000] <annevk3>
Forms WG comments are generating all the email

[12:21:00.0000] <annevk3>
I guess we should just go the W3C way about this; make a WG decision to not address their comment and move on...

[12:21:01.0000] <annevk3>
I mean sure, ideally everything would be separate and completely unreadable and even less understandable, but that's not the case

[12:37:00.0000] <Hixie>
franksalim: well i tried adding some ABNF for the wire protocol, we'll see how that goes

[12:54:00.0000] <annevk3>
Hixie, so to prevent the race condition I remove all currently queued tasks as part of the steps for abort()?

[13:09:00.0000] <annevk3>
MacBook fail

[13:09:01.0000] <annevk3>
I was planning on buying a new one but now this one is freezing/crashing on me all the time I'm not so sure anymore...

[14:58:00.0000] <AryehGregor>
Have implementers said they don't want to implement any kind of fullscreen() API for video?

[14:59:00.0000] <AryehGregor>
You'd think they could apply whatever heuristics they use for window.open().

[14:59:01.0000] <AryehGregor>
For that matter, does Flash already allow scripts to fullscreen video?  If so, it doesn't seem to be abused much.

[14:59:02.0000] <AryehGregor>
It at least allows them to make buttons that will trigger fullscreen.

[15:00:00.0000] <annevk3>
the request was not for fullscreen video

[15:00:01.0000] <annevk3>
at least not this time around

[15:00:02.0000] <AryehGregor>
Oh, dear, I was totally misreading it.

[15:00:03.0000] <AryehGregor>
Wait, he really does want fullscreen video.

[15:00:04.0000] <AryehGregor>
Just he thinks you should be able to arbitrarily fullscreen pages.

[15:00:05.0000] <AryehGregor>
That seems unnecessary.

[15:01:00.0000] <AryehGregor>
Michael Dale is the one who works on video for Wikimedia, so I figured he was asking something about video.

[15:01:01.0000] <annevk3>
anyway, bedtimes

[15:01:02.0000] <AryehGregor>
Night.

[15:27:00.0000] <Hixie>
annevk3: yeah i guess that'd be one way (so long as you only pull the tasks for this xhr)


2009-12-17
[18:34:00.0000] <Hixie>
"World Wide Web Consortium has reported that the upcoming version of HTML a.k.a. HTML 5 will support external devices, directly from your browser." - http://www.livbit.com/article/2009/12/16/html-5-will-feature-support-for-external-devices-and-file-system/

[18:34:01.0000] <Hixie>
that's news to me!

[19:08:00.0000] <Templarian>
I'm a little unsure on something http://www.whatwg.org/specs/web-apps/current-work/multipage/the-map-element.html#svg-0 In SVG when it's implemented for HTML5 will it use the svg namespace "<svg:"?

[19:10:00.0000] <Templarian>
(for inline SVG that is)

[19:11:00.0000] <Templarian>
/me wonders if there is a cleric around that could help in translating these scriptures. hmm.

[19:24:00.0000] <blog_novak>
behind the scenes, are <video> and <canvas> tags similar?  Would it be possible to use WebGL shaders on <video> data?

[19:36:00.0000] <Templarian>
I just submitted one of those comment bugs to clear up that section since it sounds confusing. Will check back to it later and see if I can understand it then.

[22:09:00.0000] <hiptobecubic>
Looking at html5 and the <video> tag, would it be possible to stream realtime video? For example from a local source such as a digital video camera?

[22:09:01.0000] <hiptobecubic>
I'm not suggesting that the <video> tag be able to access the hardware, just would be possible for some program to write the video to disk AS it's being served?

[22:10:00.0000] <Templarian>
hiptobecubic: you mean like a right click download video option?

[22:11:00.0000] <hiptobecubic>
Templarian, well no. I mean have video playing in the browser which is not yet a complete file, and keep adding to that file on the server

[22:12:00.0000] <Templarian>
Yea, should be possible if you're using any type of media server that supports one of the formats.

[22:13:00.0000] <Templarian>
(you would want a media server, not sure how writing to the video file while having the server stream it would go)

[22:13:01.0000] <hiptobecubic>
Templarian, when you say 'should be possible' what are you envisioning? I have apache running and a test page which loads an .ogv fine. I tried replacing the file with a named pipe and catting the video into the pipe, but it fails to load in the page.

[22:13:02.0000] <hiptobecubic>
Templarian, ah, that's the problem.

[22:13:03.0000] <hiptobecubic>
Templarian, i don't know much of anything about media servers,

[22:13:04.0000] <nessy>
how about trying icecast?

[22:14:00.0000] <nessy>
you need a particular server that does live streaming - I think the video element/browser should already support progressively growing files

[22:14:01.0000] <nessy>
you could also try flumotion

[22:15:00.0000] <Templarian>
Yea, just use your webcam to test should work.

[22:15:01.0000] <Templarian>
(once you have the media portion setup to output the feed)

[22:15:02.0000] <hiptobecubic>
flumotion... hmm

[00:11:00.0000] <MikeSmith>
Yudai: http://html5.org/tools/web-apps-tracker

[00:53:00.0000] <annevk3>
http://berjon.com/blog/2009/12/consensus.html is nice

[00:58:00.0000] <hsivonen>
/me doesn't get the use of a grave accent over the e in archived.

[01:09:00.0000] <annevk3>
ah great

[01:09:01.0000] <annevk3>
Hixie guesses what I suggested is okay

[01:09:02.0000] <annevk3>
guess I'm starting to understand what I'm talking about

[01:11:00.0000] <annevk3>
hsivonen, did you see http://krijnhoetmer.nl/irc-logs/whatwg/20091217#l-215 ?

[01:13:00.0000] <hsivonen>
annevk3: I skimmed past that line.

[01:13:01.0000] <hsivonen>
hmm. Templarian already left.

[01:13:02.0000] <annevk3>
if that's not namespace confusion :)

[01:14:00.0000] <annevk3>
FYI: another 20 hours or so on http://www.w3.org/2002/09/wbs/40318/issue-76-objection-poll/

[01:24:00.0000] <zcorpan>
Philip`: do you have data about pages with empty attribute values for these attributes? http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-December/024562.html

[02:03:00.0000] <annevk3>
why is Web Database not part of complete but <device> is?

[02:04:00.0000] <Philip`>
zcorpan: I have many gigabytes of data about it, and if I don't forget later today then I should be able to summarise it

[02:06:00.0000] <hsivonen>
annevk3: maybe because Web Database isn't Part of the Platform™ because it has been vendor vetoed?

[02:08:00.0000] <annevk3>
it seems only parts of it are somewhat vetoed

[02:08:01.0000] <annevk3>
and it's mostly Oracle actually that's making the noise

[02:09:00.0000] <annevk3>
which is somewhat scary

[02:10:00.0000] <annevk3>
but I guess that is it

[02:10:01.0000] <hsivonen>
annevk3: well, there has been only one query language on the table, so vendor vetoing the query language doesn't leave a working spec

[02:10:02.0000] <hsivonen>
annevk3: and I meant Mozilla and Microsoft--not Oracle

[02:11:00.0000] <annevk3>
it's not the language that has been vetoed, it's the lack of a language that can be implemented without implementing SQLite

[02:11:01.0000] <annevk3>
without using SQLite, I should say

[02:11:02.0000] <annevk3>
Hixie has offered to spec that language if someone is interested in implementing it

[02:46:00.0000] <annevk>
Does anyone know what Perl tricks Hixie uses to process the source file to get to multiple files?

[02:47:00.0000] <annevk>
I want to change the XHR2 source file to output both XHR1 and XHR2 documents

[02:47:01.0000] <annevk>
Gives me centralized errors rather than distributed :)

[02:48:00.0000] <annevk>
<!--XHR1-start--> and <!--XHR2-start--> etc. labels work, but how do I deal with them?

[02:48:01.0000] <annevk>
just run a regular expression?

[02:48:02.0000] <annevk>
and does it make sense that the source document itself no longer represents anything if it's not processed?

[03:30:00.0000] <hendry>
i think the Bubbles at work example is wrong on http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-1.html#the-figure-element

[03:31:00.0000] <hendry>
http://b.dabase.com/figure/

[03:32:00.0000] <hendry>
or it could be a validation bug for hsivonen 

[03:32:01.0000] <Philip`>
Looks like a validator bug to me

[03:47:00.0000] <hsivonen>
hendry: yeah, it's a validation bug

[03:47:01.0000] <hsivonen>
pending the resolution of the <dt>/<dd> thing

[03:54:00.0000] <zcorpan>
/me creates http://bit.ly/livedom

[04:19:00.0000] <zcorpan>
"If V is null or undefined, then return a sequence of length zero." - http://dev.w3.org/2006/webapi/WebIDL/#es-sequence

[04:20:00.0000] <zcorpan>
"If the ports argument is null, then act as if the method had just been called with two arguments, message and targetOrigin." - http://www.whatwg.org/specs/web-apps/current-work/multipage/comms.html#posting-messages-with-message-ports

[04:20:01.0000] <zcorpan>
can the ports argument be null?

[06:02:00.0000] <zcorpan>
hmm, shouldn't new MessageChannel() start the ports automatically?

[06:45:00.0000] <fabiand>
Hello. Does someone know, if it is possible to create a canvas from some rendered html content? (like convertig some h2 to a canvas ..)

[06:49:00.0000] <Philip`>
fabiand: That's not currently possible

[06:51:00.0000] <fabiand>
Philip`: ok. thanks

[06:52:00.0000] <hsivonen>
(IIRC, it's possible in Gecko if the script runs with chrome privileges)

[06:52:01.0000] <hsivonen>
(which doesn't help Web authoring)

[06:53:00.0000] <Philip`>
(Indeed, with drawWindow)

[06:54:00.0000] <fabiand>
hsivonen: true, the snapshot utility i'm using is using this function ...

[06:58:00.0000] <hsivonen>
zcorpan: the fix for the script escape tokenization stuff should be in the Minefield nightlies. Thanks for finding the bug.

[07:36:00.0000] <annevk>
aah

[07:37:00.0000] <annevk>
in Python a . does not match newlines

[07:37:01.0000] <annevk>
so you need re.compile("regexp", re.DOTALL) ...

[07:37:02.0000] <annevk>
rather than just "regexp"

[07:37:03.0000] <annevk>
grmbl

[07:38:00.0000] <Philip`>
Isn't that true of (almost) all languages' regexps?

[07:38:01.0000] <Philip`>
e.g. you need /regexp/s in Perl

[07:40:00.0000] <gsnedders>
/me thinks (almost) all languages' regexps are based on Perl

[07:41:00.0000] <Philip`>
and you need to use (?:.|\n) in JS instead of . as far as I can tell

[07:42:00.0000] <annevk>
oh

[07:42:01.0000] <annevk>
well, I guess I'm a regexp noob

[07:42:02.0000] <Philip`>
I think Perl wins in cleanest regexp syntax :-)

[07:46:00.0000] <annevk>
I tried learning Perl once, but switched to Python

[07:46:01.0000] <Philip`>
/me finds it convenient to know both

[07:47:00.0000] <annevk>
I guess I should invest some more time in Perl yeah

[07:49:00.0000] <Philip`>
/me thinks Perl probably requires quite a bit more investment than Python

[07:49:01.0000] <Philip`>
e.g. if you see some code doing re.compile("regexp", re.DOTALL) then you can do help(re) to find out what it is

[07:49:02.0000] <Philip`>
but if you see some code with /regexp/s then you're kind of stuck unless you already know what it is and where to find the documentation

[07:50:00.0000] <jgraham>
/me wonders if it is really worthwhile investing time in perl compared to investing the same time in a more dissimilar language

[07:50:01.0000] <Philip`>
but the latter is far easier to write, and somewhat easier to remember

[07:51:00.0000] <mikekelly>
hi fans

[07:51:01.0000] <Philip`>
jgraham: Depends on what kind of code you want to write

[07:51:02.0000] <mikekelly>
un petite question

[07:52:00.0000] <mikekelly>
seal voo play

[07:52:01.0000] <jgraham>
Philip`: Really? I would expect that learning a greater variety of things would make you better in general than learning several closely related things

[07:52:02.0000] <mikekelly>
you know forms are getting new actions for PUT/DELETE ..?

[07:52:03.0000] <jgraham>
/me is not really a shining example of this philosophy though

[07:53:00.0000] <mikekelly>
what's the mofokin' dizzle with PATCH yo?!

[07:55:00.0000] <Philip`>
jgraham: I'd expect that if you only write code for quick scripts and glue code and some one-off CGIs and whatever then you would not get a great deal of value out of learning Haskell or assembly or Prolog etc, and it would be much harder to learn since you wouldn't be learning by solving problems you care about

[07:56:00.0000] <Philip`>
so it'd be better to spend the same amount of time learning a more directly relevant language that you would actually use sometimes

[07:56:01.0000] <gsnedders>
jgraham: You should learn Ruby.

[07:56:02.0000] <mikekelly>
yes

[07:56:03.0000] <mikekelly>
sinatra is the poop

[07:56:04.0000] <jgraham>
gsnedders: Why?

[07:56:05.0000] <Philip`>
(But once you know enough languages to do everything you want, then you should spend more time learning more languages :-) )

[07:57:00.0000] <gsnedders>
jgraham: Because you're not a shining example of the philosophy of different langauges.

[07:57:01.0000] <mikekelly>
jgraham: http://www.sinatrarb.com

[07:57:02.0000] <mikekelly>
that is the coolest hello world I've ever seen.

[07:57:03.0000] <jgraham>
Philip`: Typically one language is enough to do everything you want if it is one sutiable language for the tasks you care about

[07:57:04.0000] <jgraham>
gsnedders: I can't tell if you are being ironic.

[07:57:05.0000] <mikekelly>
which is why you want a language with an HTTP DSL! ))

[07:58:00.0000] <gsnedders>
jgraham: Aren't I always?

[07:58:01.0000] <jgraham>
gsnedders: no

[07:58:02.0000] <mikekelly>
please can someone explain what is going on with HTML5 and HTTP PATCH

[07:58:03.0000] <jgraham>
Philip`: (and you only care about a small range of tasks)

[07:59:00.0000] <gsnedders>
mikekelly: AFAIK nobody has made any proposal for it, and nor is PATCH actually standardized yet

[08:00:00.0000] <mikekelly>
er

[08:00:01.0000] <mikekelly>
I think it did recently no?

[08:00:02.0000] <gsnedders>
Not that I saw. AFAIK it's still an I-D.

[08:00:03.0000] <mikekelly>
ahh poop-wagons

[08:00:04.0000] <Philip`>
jgraham: I'd imagine most people do a sufficiently wide range of tasks that one language is not the optimum one for all of them

[08:01:00.0000] <mikekelly>
ok well let's add that to next century's roadmap anyway

[08:01:01.0000] <Philip`>
where by "imagine" I just mean I'm making things up to support anything I said earlier

[08:01:02.0000] <mikekelly>
are you a Physicist by any chance?

[08:02:00.0000] <jgraham>
I doubt one langauge is the optimum one for all of them but it may be a sufficiently good one

[08:02:01.0000] <gsnedders>
mikekelly: http://tools.ietf.org/html/draft-dusseault-http-patch-16 — that's the latest I-D I can see, and I don't see any RFC after that for it

[08:02:02.0000] <gsnedders>
mikekelly: Who? Me?

[08:03:00.0000] <workmad3>
http://www.musicademy.com/2009/12/worship-gaffe-2-christmas-carol-nightmare/ <-- who can listen to this the longest

[08:03:01.0000] <workmad3>
I just got to 2.30 ish

[08:03:02.0000] <gsnedders>
Philip`: I used PHP for everything for years because I couldn't be bothered learning another language. Then I learnt Python, and realized learning languages was normally a _lot_ less painful than learning PHP.

[08:03:03.0000] <workmad3>
oops, wrong channel

[08:04:00.0000] <workmad3>
sorry guys :)

[08:04:01.0000] <workmad3>
gsnedders: I found learning python almost painful... but then the language I'd learnt prior to python was ruby :)

[08:04:02.0000] <mikekelly>
gsnedders: apparently PATCH was never actually taken out or something

[08:05:00.0000] <mikekelly>
just not implemented

[08:05:01.0000] <gsnedders>
workmad3: I basically learnt Ruby in a day, already knowing PYthon

[08:05:02.0000] <gsnedders>
*Python

[08:05:03.0000] <jgraham>
I spent many years trying to convince people I knew doing PhDs that they should write trivial glue code in Python rather than C

[08:06:00.0000] <Philip`>
/me used to use QuickBASIC for everything, then tried learning C, and realised learning C was really painful, and then gave up and went back to QuickBASIC (until being forced to learn Perl to maintain/write some primitive CGI scripts, and after that it was easy to learn C)

[08:06:01.0000] <gsnedders>
So, if you want to have it easy when learning C, learn Perl?

[08:06:02.0000] <jgraham>
s/C/any other language/

[08:06:03.0000] <gsnedders>
Because that fucks up your mind sufficently?

[08:07:00.0000] <jgraham>
s/fucks up your mind sufficiently/gives you unreasonably low expectations/

[08:07:01.0000] <mikekelly>
gsnedders: The IESG has approved the following document:

[08:07:02.0000] <mikekelly>
- 'PATCH Method for HTTP ' <draft-dusseault-http-patch-16.txt> as a Proposed Standard 

[08:07:03.0000] <Philip`>
I had a dozen books on C/C++ but couldn't understand pointers after reading any of them, but it made sense years later when I realised they were like Perl references

[08:08:00.0000] <workmad3>
I thankfully managed to get pointers fairly easily... it was template metaprogramming in C++ that screwed with my mind

[08:09:00.0000] <workmad3>
but I know what you mean... I read through half a dozen or so books on C++ all based around learning the low level mechanics

[08:09:01.0000] <Philip`>
Learning a sane functional programming first helps with template metaprogramming :-)

[08:09:02.0000] <gsnedders>
/me can't read programming books

[08:09:03.0000] <workmad3>
I've now read a similar amount of text on Ruby and Rails, and gotten a lot more distance out of them :)

[08:09:04.0000] <gsnedders>
I just lose interest so quickly.

[08:10:00.0000] <workmad3>
/me reads on the bus to and from work

[08:10:01.0000] <gsnedders>
/me walks to work…

[08:10:02.0000] <gsnedders>
/me needs to get a scarf

[08:10:03.0000] <Philip`>
gsnedders: You need audio books

[08:10:04.0000] <workmad3>
aha, untapped market... audio books on programming :)

[08:10:05.0000] <jgraham>
I think my problem with learning C (which I occasionally try to do properly) is that it is so tedious to get anything done that I wind up doing something less boring instead

[08:10:06.0000] <gsnedders>
Philip`: So then I don't hear the bus coming and get killed?

[08:10:07.0000] <workmad3>
/me imagines Stephen Fry reading out code samples...

[08:11:00.0000] <gsnedders>
/me bursts out laughing

[08:11:01.0000] <Philip`>
"First write the following into your text editor: hash, exclamation mark, forward slash, user, that is yoo, ess, arr, forward slash, bin, ..."

[08:11:02.0000] <Philip`>
It'd be great

[08:12:00.0000] <Philip`>
gsnedders: You could turn the volume down a bit

[08:12:01.0000] <workmad3>
I think an audio book on lisp would be awesome :)

[08:13:00.0000] <Philip`>
gsnedders: Also, deaf people seem to avoid buses alright, so you should be fine

[08:13:01.0000] <jgraham>
"... right parenthesis, right parenthesis, right parenthesis, right parenthesis, right parenthesis..."

[08:13:02.0000] <Philip`>
jgraham: Maybe your problem with learning C is that you're trying to do things in it that are tedious to get done

[08:13:03.0000] <Philip`>
like anything involving text

[08:14:00.0000] <jgraham>
Philip`: Could be

[08:14:01.0000] <Philip`>
Write a PNG decoder or something :-)

[08:14:02.0000] <jgraham>
Philip`: But I don't really want to decode PNGs

[08:14:03.0000] <workmad3>
C isn't about doing stuff that you *want* to do :)

[08:15:00.0000] <jgraham>
OK, I don't *have* to decode PNGs

[08:15:01.0000] <Philip`>
Learning a language isn't about doing stuff that you *want* to do :-)

[08:15:02.0000] <gsnedders>
GIFs, instead?

[08:15:03.0000] <gsnedders>
/me wrote Anolis to learn Python

[08:15:04.0000] <gsnedders>
/me wanted to do that

[08:15:05.0000] <Philip`>
Well, it's not about doing stuff you *don't* want to do either

[08:16:00.0000] <Philip`>
It's about doing stuff that you can manage given your current skill level, and that will make you more comfortable with what you know and will teach you new things

[08:16:01.0000] <jgraham>
Philip`: Those things are not easy to find

[08:16:02.0000] <Philip`>
and then you just repeat until you know everything

[08:17:00.0000] <workmad3>
Philip`: until you know everything? or until you're an expert? :)

[08:17:01.0000] <workmad3>
/me is reminded of the cynical definition of an expert - 'A person who knows more and more about less and less until they know everything about nothing'

[08:20:00.0000] <jgraham>
Philip`: (or rather they can be if the language itself presents an interesting challenge -- thus making simpole excercises interesting -- or has a very shallow learning curve -- thus making intersting projects possible -- but these conditions do not always hold)

[08:23:00.0000] <Philip`>
Maybe it's better to find simple exercises within interesting projects

[08:24:00.0000] <Philip`>
i.e. patching someone else's project

[08:24:01.0000] <Philip`>
(as long as the project itself is reasonably clean and doesn't add a huge amount of extra complexity)

[08:24:02.0000] <Philip`>
because then you can do something that's both easy and useful

[08:26:00.0000] <Hixie>
annevk: oracle's comments had no effect on whether complete.html contained Web SQL Database, only Microsoft's and Mozilla's.

[08:27:00.0000] <Hixie>
annevk: and for the preprocessing, my script does various things but for the <!--START--> markers what i do is i have a script that runs through the file and keeps track of what is "on" and what is "off"

[08:27:01.0000] <Hixie>
and then outputs one file per marker, containing just the bits that are "on" and "off"

[08:27:02.0000] <Hixie>
i also have a separate output file which contains everything, which i use for validation

[08:29:00.0000] <Philip`>
print $f1 $_ if /<!--START1-->/ .. /<!--END1-->/; # this doesn't quite work since it'll print the first comment too :-(

[08:37:00.0000] <annevk>
I have a regular expression search that works quite well

[08:38:00.0000] <annevk>
however, annotating the XHR2 source file is a bitch

[08:38:01.0000] <annevk>
hopefully it is worth all the hours that it is going to cost me

[08:38:02.0000] <annevk>
I'm guessing it will help

[10:03:00.0000] <annevk>
and back from dinner

[10:03:01.0000] <annevk>
still at Schiphol

[10:04:00.0000] <annevk>
I think this is a personal low

[10:04:01.0000] <annevk>
and also a personal record

[10:54:00.0000] <annevk>
weird

[10:54:01.0000] <annevk>
http://googleblog.blogspot.com/2009/12/transliteration-goes-global.html

[10:54:02.0000] <annevk>
they don't trust Unicode / font support enough yet?

[11:06:00.0000] <Philip`>
Their alt text is hardly ideal

[11:14:00.0000] <AryehGregor>
Surely font support for Cyrillic is reliable.

[11:23:00.0000] <Hixie>
what's the selector api method name again?

[11:23:01.0000] <Hixie>
oh _query_ selector

[11:23:02.0000] <Hixie>
nm

[11:26:00.0000] <zcorpan__>
Philip`: got a summary yet? :)

[13:36:00.0000] <Philip`>
zcorpan: http://philip.html5.org/data/empty-url-attributes.txt

[13:37:00.0000] <Philip`>
zcorpan: http://philip.html5.org/data/empty-url-link-attributes.txt

[13:37:01.0000] <Philip`>
Are those the kinds of things you want?

[13:37:02.0000] <Philip`>
(I assume you can figure out the weird undocumented inconsistent XML formats)

[13:39:00.0000] <zcorpan>
Philip`: yep, thanks!

[13:59:00.0000] <zcorpan>
hsivonen: http://visionwow.com/ has lots of nul bytes which show up as u+fffd with html5.enable but don't render anything without html5.enable and in other browsers

[15:48:00.0000] <zcorpan>
Philip`: what's up with <attribute uri="gg.shenso.com/s/?q=ç element="img" name="src" value=""/> (missing quote)?

[15:56:00.0000] <Philip`>
zcorpan: Your browser is eating the quote

[15:57:00.0000] <zcorpan>
snap


2009-12-18
[16:29:00.0000] <zcorpan>
Philip`: is element-href-absolute the resolved url? does it take <base href> into account?

[16:39:00.0000] <Philip`>
zcorpan: It's just href resolved against element-uri

[16:39:01.0000] <Philip`>
and element-uri is just the page that was downloaded

[16:40:00.0000] <Philip`>
so it ignores <base> entirely

[16:40:01.0000] <zcorpan>
ok

[18:25:00.0000] <drhodes>
does anyone know if websockets will allow peer2peer connections?

[18:30:00.0000] <drhodes>
ok, well found this: http://stackoverflow.com/questions/1032006/will-html5-allow-web-apps-to-make-peer-to-peer-http-connections if by chance anyone is also curious.

[19:42:00.0000] <Hixie>
drhodes: the short answer is "not in this version"

[19:43:00.0000] <Hixie>
drhodes: however, it's definitely something that's come up in the past, and i wouldn't be surprised to see it come up again in the future

[19:43:01.0000] <Hixie>
drhodes: in particular, if the current exploration of the video conferencing space goes anywhere, i wouldn't at all be surprised to see something for peer-to-peer streaming

[19:56:00.0000] <drhodes>
ok, thanks Hixie.

[22:58:00.0000] <boblet>
hey all. Has any browser fully supported HTML4.01 yet? What about CSS2.1—still only IE8?

[23:01:00.0000] <boblet>
s/supported/implemented/

[23:01:01.0000] <hsivonen>
boblet: I'm pretty sure a browser that can browse the Web can't fully implement HTML 4.01

[23:02:00.0000] <boblet>
hsivonen: heh

[23:02:01.0000] <hsivonen>
boblet: and for CSS 2.1 there isn't a sufficient test suite yet to be able to tell

[23:02:02.0000] <JoePeck>
I heard people were going to skip "4" and go straight to "5" like ECMAScript =P

[23:02:03.0000] <hsivonen>
(I am nt suggesting that there'd be a sufficient test suite for HTML 4.01)

[23:03:00.0000] <boblet>
hsivonen: ohrly? I thought the test suite was pretty complete now after MS’s contributions. interesting

[23:03:01.0000] <boblet>
JoePeck: ++

[23:03:02.0000] <zcorpan>
JoePeck: you mean css5?

[23:04:00.0000] <hsivonen>
boblet: as I understand it, the MS test suite test features statically in isolation--not feature interactions or dynamic changes

[23:04:01.0000] <boblet>
hsivonen: aah. thanks for the explanation

[23:04:02.0000] <JoePeck>
I'm too tired to joke =), g'nite guys ;)

[23:06:00.0000] <zcorpan>
http://jhop.me/ie8-bugs suggests ie8 doesn't have a perfect css2.1 impl

[23:10:00.0000] <zcorpan>
(not surprisingly many bugs seem to be about feature interactions and dynamic changes)

[23:11:00.0000] <zcorpan>
though a surprising (to me) number of bugs are marked as regressions since ie7

[23:11:01.0000] <zcorpan>
maybe regressions are easier to find

[23:12:00.0000] <hsivonen>
hmm. otoh, maybe html 4.01 has so few MUSTs that apply to UAs that there are impls of all mandatory requirements

[23:13:00.0000] <zcorpan>
hsivonen: browsers assume a default encoding, which is a violation of html4 must

[23:14:00.0000] <hsivonen>
ok. that settles it then

[23:18:00.0000] <zcorpan>
html4 also has a must requirement about parsing media="" which is incompatible with css3-mq

[23:19:00.0000] <zcorpan>
and "user agents must always make the content

[23:19:01.0000] <zcorpan>
   of the TITLE element available to users (including TITLE elements that

[23:19:02.0000] <zcorpan>
   occur in frames)"

[23:20:00.0000] <zcorpan>
and that q has quotes

[23:21:00.0000] <zcorpan>
and some rendering requirements about the soft hyphen

[23:24:00.0000] <zcorpan>
i guess i could count the remaining testable requirements on my other hand's fingers, but it's too boring

[23:34:00.0000] <zcorpan>
MikeSmith: yo

[23:35:00.0000] <hsivonen>
MikeSmith: I re deployed V.nu yesterday

[23:37:00.0000] <MikeSmith>
hsivonen: cool

[23:37:01.0000] <MikeSmith>
zcorpan: hey

[23:37:02.0000] <zcorpan>
MikeSmith: i don't see any xml-stylesheet support in qa-dev from testing it

[23:37:03.0000] <MikeSmith>
zcorpan: hmm, OK

[23:38:00.0000] <MikeSmith>
I must have muffed something up

[23:38:01.0000] <MikeSmith>
I'll take a look later

[23:38:02.0000] <MikeSmith>
zcorpan: I plan to work on the xml-stylesheet checker this weekend

[23:38:03.0000] <zcorpan>
cool

[23:39:00.0000] <MikeSmith>
hopefully will have something fairly complete by Monday

[23:39:01.0000] <MikeSmith>
ready for testing

[23:39:02.0000] <zcorpan>
ok

[23:45:00.0000] <zcorpan>
MikeSmith: what does the internally generated document look like?

[23:46:00.0000] <MikeSmith>
zcorpan: what internally generated document?

[23:46:01.0000] <MikeSmith>
oh

[23:46:02.0000] <MikeSmith>
it looks like this:

[23:46:03.0000] <MikeSmith>
<xml-stylesheet href="">

[23:46:04.0000] <MikeSmith>
etc.

[23:46:05.0000] <MikeSmith>
<xml-stylesheet href=""/>

[23:46:06.0000] <MikeSmith>
actually

[23:47:00.0000] <MikeSmith>
it's just a single element

[23:47:01.0000] <zcorpan>
ok

[23:47:02.0000] <MikeSmith>
exactly the same as the PI, just with < and /> instead of <? and ?>

[23:48:00.0000] <MikeSmith>
uses XOM to parse that and then examine the attributes

[23:49:00.0000] <MikeSmith>
dunno if XOM is the best choice or if there's some better way of doing it

[23:49:01.0000] <MikeSmith>
but this way seems to work OK

[23:50:00.0000] <zcorpan>
i wonder if there's a way to break out of the element but still end up with a well-formed document

[23:50:01.0000] <zcorpan>
considering that > is allowed in pi content

[23:50:02.0000] <MikeSmith>
hmm, dunno

[23:50:03.0000] <zcorpan>
it would be possible with xml5's short end tag

[23:52:00.0000] <hsivonen>
MikeSmith: are you using code from the HTML5 tokenizer?

[23:52:01.0000] <MikeSmith>
hsivonen: no, using standard XOM directly

[23:53:00.0000] <MikeSmith>
zcorpan: that does remind me about one downside of doing it the way I am, which is that if the pseudo-element/document is not well-formed, all it really does right now is report it as "The syntax of the xml-stylesheet PI is incorrect."

[23:53:01.0000] <MikeSmith>
e.g., for a case like <?xml-stylesheet href="<"?>

[23:53:02.0000] <MikeSmith>
it doesn't really tell the user where the error is

[23:53:03.0000] <zcorpan>
yeah

[23:54:00.0000] <MikeSmith>
I could have it get the message from the exception that XOM throws

[23:54:01.0000] <MikeSmith>
but the problem with that is XOM messages use the word "attribute"

[23:54:02.0000] <zcorpan>
but that message would be confusing for <?xml-stylesheet >?>

[23:54:03.0000] <hsivonen>
MikeSmith: I don't see how XOM would be applicable

[23:55:00.0000] <MikeSmith>
hsivonen: for creating a document based on taking <?xml-stylesheet href=""?> and turning it into <xml-stylesheet href=""/>

[23:56:00.0000] <hsivonen>
oh

[23:57:00.0000] <MikeSmith>
I'm using a Builder the creating an Element then using getAttribute()

[23:57:01.0000] <zcorpan>
MikeSmith: what happens for namespace declarations?

[23:57:02.0000] <hsivonen>
MikeSmith: involving a tree model here seems over-complicated to me

[23:59:00.0000] <zcorpan>
i think you need a namespace-unaware xml parser

[23:59:01.0000] <MikeSmith>
hsivonen: if you have another way that'd work better, I can implement it that way

[00:00:00.0000] <hsivonen>
I'd take the data from the PI in a SAX handler and run it through a state machine that implements the pseudo-attribute syntax

[00:01:00.0000] <hsivonen>
and I'd create the state macine by copying and pasting from the HTML5 tokenizer

[00:01:01.0000] <MikeSmith>
hsivonen: OK

[00:01:02.0000] <zcorpan>
at least Char checking is already done by the xml parser

[00:02:00.0000] <MikeSmith>
hsivonen: OK, I can make that change this weekend

[00:02:01.0000] <MikeSmith>
zcorpan: Char checking?

[00:02:02.0000] <MikeSmith>
oh

[00:02:03.0000] <zcorpan>
(but not LegalChar for NCRs)

[00:03:00.0000] <MikeSmith>
zcorpan: you mean the Char checking is done before we get to the xml-stylesheet PI checking

[00:03:01.0000] <zcorpan>
yes

[00:03:02.0000] <MikeSmith>
zcorpan: why do we need a namespace-unaware parser?

[00:04:00.0000] <MikeSmith>
if somebody has <?xml-stylesheet xlink:href="foo"?> or whatever, that's just an invalid attribute name, right?

[00:04:01.0000] <zcorpan>
MikeSmith: because otherwise <?xml-stylesheet href='' xmlns:xml='http://www.w3.org/XML/1998/namespace'?><x/> would probably pass

[00:04:02.0000] <MikeSmith>
oh, I see

[00:05:00.0000] <MikeSmith>
hmm, yeah

[00:05:01.0000] <zcorpan>
but if you implement your own tokenizer, that won't be a problem

[00:05:02.0000] <MikeSmith>
right

[00:06:00.0000] <MikeSmith>
so Henri's suggestion is definitely not just a better way of doing it, but basically the right way

[00:07:00.0000] <MikeSmith>
there are probably other bad side effects of turning into a element

[00:07:01.0000] <MikeSmith>
anyway, I have to step out for a bit

[00:07:02.0000] <MikeSmith>
back later

[00:08:00.0000] <zcorpan>
hsivonen: how would you validate the pseudo-attributes?

[00:10:00.0000] <zcorpan>
with custom code, or with a schema?

[00:11:00.0000] <hsivonen>
zcorpan: I was thinking custom code

[00:11:01.0000] <zcorpan>
ok

[00:19:00.0000] <hsivonen>
/me wonders how "as well as non-browser environments that make use of Web technologies" got into the Web Apps WG charter

[00:19:01.0000] <zcorpan>
widgets?

[00:20:00.0000] <hsivonen>
maybe

[00:21:00.0000] <hsivonen>
too bad the bit is there. not having it there would make it easy to form a charter-based argument for why it's not worthwhile use of resources for the group to facilitate non-browser reuse of XHR

[00:22:00.0000] <hsivonen>
fwiw, I think such reuse is fine

[00:22:01.0000] <hsivonen>
but the group shouldn't use its finite resources to facilitate it

[00:23:00.0000] <hsivonen>
it should be the reusers' problem

[00:26:00.0000] <hsivonen>
(conceptually, I prefer Web apps pinned to HTML5 app cache over widgets)

[04:00:00.0000] <hsivonen>
sent comments

[04:00:01.0000] <hsivonen>
the e-book part is particularly odd

[06:57:00.0000] <annevk>
oh yes

[06:57:01.0000] <annevk>
win win win

[06:57:02.0000] <annevk>
I successfully merged XHR1 and XHR2 into a single document

[06:57:03.0000] <annevk>
now all I need to do is update the script with new file names and I can start editing again

[10:01:00.0000] <Hixie>
so i guess we have to wait til wednesday now for the chairs to come to a decision?

[12:24:00.0000] <hsivonen>
" <oedipus> could set up a google code group for the shadow DOM for CANVAS and use development of shadow DOM as a follow-up activity to the RWAB XG"

[12:25:00.0000] <hsivonen>
what does the rich web app backplane have to do with canvas subtree?

[12:29:00.0000] <miketaylr>
back

[14:59:00.0000] <TabAtkins>
Argh I am once again stymied by <dl>'s inconsiderate lack of <di>.

[14:59:01.0000] <TabAtkins>
Hixie: Will it do me any good to throw a request for <di> at you?  I can provide use-cases, all styling-related.

[15:00:00.0000] <Hixie>
can you make a convincing argument that the bug shouldn't be fixed in css rather than html? :-)

[15:00:01.0000] <TabAtkins>
…probably not.

[15:03:00.0000] <TabAtkins>
Damn your logic!

[15:11:00.0000] <othermaciej>
I wonder if anyone is going to propose an alternative to shelley's figure/dialog/dt/dd thing

[15:12:00.0000] <TabAtkins>
I have it on my calendar.

[15:12:01.0000] <othermaciej>
I can understand her aesthetic distaste for using <dt> and <dd> inside <figure>, but I can't say <fltcap> is much better

[15:12:02.0000] <othermaciej>
I'm considering making a proposal to use <label> instead

[15:12:03.0000] <TabAtkins>
Despite <label>'s problems with embedded form elements?

[15:12:04.0000] <othermaciej>
I can't say I love the status quo enough to ersonally advocate that 

[15:13:00.0000] <TabAtkins>
I won't be advocating the status quo either - I find the use of <dt>/<dd> to personally be ugly.

[15:13:01.0000] <othermaciej>
yes, despite <label>'s potential problems for some use cases in current browsers

[15:13:02.0000] <othermaciej>
I guess I would probably have to suggest how future browsers could finesse the issue

[15:14:00.0000] <othermaciej>
I could go even more crazy and propose <label> for <details> and <caption> for <figure>, which is semantically the best match, but would require a parsing change for <figure> and could render it unusable in tables in the short term

[15:14:01.0000] <othermaciej>
er, a parsing change for <caption>

[15:14:02.0000] <othermaciej>
I also wonder if there is a technical reason none of the <hn> elements would not do

[15:15:00.0000] <TabAtkins>
One nice advantage of using <label> is that it suggests the possibility of putting the <label> outside of the <details> and linking it with an idref.

[15:15:01.0000] <TabAtkins>
Though perhaps that's bad for accessibility of the markup?

[15:16:00.0000] <TabAtkins>
I suggested using an <h1> for it on the list some time ago.

[15:16:01.0000] <othermaciej>
I think a label where the details pop up somewhere else is actually useful in some cases

[15:16:02.0000] <Hixie>
<label> makes it impossible to have form controls in a template-like page with a <figure>, so i think it's a non-starter

[15:17:00.0000] <Hixie>
<h1> would interfere with the <figure> content model

[15:17:01.0000] <TabAtkins>
What's the use-case for putting form controls in a figure caption?  I can't recall one right now.

[15:17:02.0000] <Dashiva>
Editing, e.g. a checkbox in each figure for inclusion or exclusion of some sort

[15:17:03.0000] <othermaciej>
<label> would only have to create problems like that in the short term, which is still IMO better than <legend> was but to be fair is more objectively problematic than <dt>/<dd>

[15:18:00.0000] <othermaciej>
TabAtkins: Hot or Not

[15:18:01.0000] <Hixie>
<figure><legend><label>Caption: <input type=text></label></legend> ... </figure>

[15:20:00.0000] <TabAtkins>
I see thee template use, then.  The hot-or-not use is weird, though.  We don't really want the form being outside of the page outline, do we?

[15:21:00.0000] <Hixie>
?

[15:21:01.0000] <TabAtkins>
<figure> is an outline root, right?

[15:21:02.0000] <TabAtkins>
/me may be misremembering.

[15:24:00.0000] <othermaciej>
<figure><dd><img src="random-photo-12345.jpg"></dd><dt><label><input type="radio" name="hotornot" value="hot">Hot</label> or <label><input type="radio" name="hotornot" value="not">Not</label></dt></figure>

[15:24:01.0000] <othermaciej>
using current syntax

[15:24:02.0000] <othermaciej>
makes sense to me

[15:25:00.0000] <othermaciej>
I don't understand how the document outline would be relevant to whether those controls are part of the caption - certainly feels to me like they are

[15:26:00.0000] <othermaciej>
the way I would work around this if <figure>'s caption were <label> is to only let form controls inside a <figure> caption bind to a label through explicit <label for> or aria-labeledby and not implicitly via containment

[15:26:01.0000] <othermaciej>
that would make this use case for <figure> non-viable until all browsers update

[15:28:00.0000] <Hixie>
i really think we should avoid making the platform that quirky

[15:28:01.0000] <Hixie>
it's bad enough when we add quirks accidentally

[15:29:00.0000] <Hixie>
no need to make it worse :-)

[15:33:00.0000] <othermaciej>
it's true, that would be an odd quirk

[15:33:01.0000] <othermaciej>
and <label> is not a good enough match for <figure> to be worth fighting for

[15:33:02.0000] <TabAtkins>
That would be way too strange.  It would confuse me.

[15:33:03.0000] <othermaciej>
<caption> would be the best match

[15:33:04.0000] <TabAtkins>
Let's all go back to <caption>.

[15:33:05.0000] <TabAtkins>
Yus.

[15:33:06.0000] <othermaciej>
but the short-term cost of being unusable in tables might be too high

[15:33:07.0000] <TabAtkins>
Man, who puts figures in tables?

[15:34:00.0000] <othermaciej>
presumably that is what Flickr would do if they used <figure>

[15:35:00.0000] <Hixie>
we couldn't use <caption> at all -- you couldn't nest a figure in a <caption> if we used <caption>

[15:35:01.0000] <Hixie>
if we reuse elements, only <dt>/<dd> and <legend> make any sense, i believe

[15:35:02.0000] <Hixie>
if we make a new element, well, we'd be adding to HTML's quite ridiculous set of "heading"-like elements

[15:36:00.0000] <Hixie>
we're at 18 already iirc

[15:36:01.0000] <Hixie>
surely it's time to draw the line :-)

[15:36:02.0000] <TabAtkins>
Twitter's use of <table> doesn't even make *sense*.  They're a fixed-width site.  They can handle this in other ways.

[15:36:03.0000] <Hixie>
this far, no farther!

[15:36:04.0000] <TabAtkins>
s/Twitter/flickr/

[15:37:00.0000] <daedb>
Let's draw the line at 20, just in case we eventually need one more... also, even numbers are better ^^

[15:37:01.0000] <TabAtkins>
Why can't we nest <caption><figure><caption>?

[15:38:00.0000] <Hixie>
TabAtkins: <caption> would close the parent <caption>

[15:38:01.0000] <TabAtkins>
Damn legacy parsing constraints!

[15:38:02.0000] <TabAtkins>
I suppose that's a necessity for some reason.

[15:42:00.0000] <TabAtkins>
/me needs to write all this down.

[15:43:00.0000] <TabAtkins>
/me …since he did offer to do that whole "Why Features Exist" thing.

[15:44:00.0000] <othermaciej>
Hixie: I don't see why using <caption> would prevent you from nesting a <figure> in a <caption>, but I also don't see a use case for doing so

[15:45:00.0000] <TabAtkins>
He said why while you were gone.  Legacy parsing constraint, apparently.

[15:45:01.0000] <Hixie>
<caption> has to imply </caption>

[15:46:00.0000] <Hixie>
and as to why you'd have one in a table caption, putting <details> in a <caption> is one of the proposed ways of addressing the complex table accessibility issue

[15:46:01.0000] <TabAtkins>
Is that necessary if the <caption> is nested inside a <figure>, though?  We *don't* have any legacy constraints surrounding people putting <figure> in table captions.

[15:46:02.0000] <othermaciej>
Hixie: does <caption> have to imply </caption> inside <figure>

[15:46:03.0000] <othermaciej>
?

[15:46:04.0000] <TabAtkins>
/me doens't know if that complicates parsing or what.

[15:46:05.0000] <othermaciej>
I don't see how that could be a legacy constraint

[15:47:00.0000] <Hixie>
othermaciej: i really don't think adding a new parsing mode just to handle <figure> is the right balance to strike :-)

[15:47:01.0000] <othermaciej>
Hixie: I think <label> would be better for <details> than <caption> and I think control nesting is not likely to be an issue there

[15:48:00.0000] <TabAtkins>
It's not really an issue because the entirety of a details caption has to be active already.

[15:49:00.0000] <TabAtkins>
The *idea* of nesting form controls in a <details> caption is about as troublesome as nesting <a> inside of <a>, or <label> inside of <label> currently.

[15:49:01.0000] <Hixie>
othermaciej: i think <legend> would be orders of magnitude better than either of those

[15:50:00.0000] <Hixie>
TabAtkins: why? it's quite common

[15:50:01.0000] <TabAtkins>
I've never done so on either, and it strikes me as crazy-talk.

[15:50:02.0000] <othermaciej>
<legend> is broken for all uses in current browsers, as opposed to just in tables, so I can't see how it's worse than <caption> other than the fact that you have already written the spec for the needed parser chnges

[15:50:03.0000] <othermaciej>
er, can't see how it's better than <caption>

[15:51:00.0000] <Hixie>
the <legend> parsing changes are a simplification. The <caption> parser changes are so much extra complexity they actually require at least one more parse mode.

[15:51:01.0000] <TabAtkins>
Firefox *can't* nest <label>s inside of <label>s.

[15:51:02.0000] <TabAtkins>
(possibly other's can't too)

[15:52:00.0000] <Hixie>
anyway, bbiab, gotta go to work

[15:52:01.0000] <othermaciej>
I think either possible change would be the about same amount of code change to WebKit's existing parser, regardless of whether it is a simplification or addition in the spec

[15:52:02.0000] <othermaciej>
gotta pay attention to meeting

[15:52:03.0000] <othermaciej>
back later


2009-12-19
[18:32:00.0000] <miketaylr>
may i ask for some clarification re: boolean attributes?

[18:32:01.0000] <miketaylr>
"The presence of a boolean attribute on an element represents the true value, and the absence of the attribute represents the false value."

[18:32:02.0000] <miketaylr>
so i get that <input required> is legit

[18:32:03.0000] <miketaylr>
and that <input required=true> is not legit

[18:32:04.0000] <miketaylr>
what about <input required="required">?

[18:34:00.0000] <miketaylr>
oh wait...answered my own question..in the next sentence

[00:29:00.0000] <hsivonen>
whoa http://www.on2.com/index.php?id=472&news_id=693


2009-12-20
[21:03:00.0000] <Hixie>
btw, i'm on vacation but will be checking mail and irc sporadically for the next two weeks

[21:03:01.0000] <Hixie>
so if you need me, just /msg me and hang around for a few hours

[02:26:00.0000] <MikeSmith>
hsivonen: if you're around -- I did a copy-paste of the stateLoop code from the htmlparser tokenizer code, and I'm wondering what I should set the returnState value to at the start of the loop

[02:26:01.0000] <MikeSmith>
given that I know I'm reading in the data part of PI

[02:29:00.0000] <MikeSmith>
it seems like I should initially just set it to the same value as the state value

[02:29:01.0000] <MikeSmith>
that is, BEFORE_ATTRIBUTE_NAME

[13:23:00.0000] <Dashiva>
I wonder if anyone's going to implement XHR on a platform that already has full HTTP access


2009-12-21
[22:19:00.0000] <Aleoss>
Can someone tell me how to fix/resolve the Tidy errors in my document? http://pastebin.org/67228

[22:26:00.0000] <Vito`>
Aleoss: looks like Tidy doesn't like the HTML5 doctype, so it's not parsing the HTML entities

[22:26:01.0000] <Vito`>
because your page validates through html5.validator.nu

[22:27:00.0000] <Aleoss>
Vito: So is there a way I can make the entities validate through Tidy?

[22:27:01.0000] <Aleoss>
Vito: Or is that out of my control?

[22:29:00.0000] <Vito`>
I only see people complaining on it in a google search

[22:29:01.0000] <Vito`>
I'd guess out of your control until Tidy becomes HTML5-aware

[22:30:00.0000] <Vito`>
You could put in an HTML4 doctype to make sure your entities validate, and then ignore all the errors about the HTML5 stuff

[22:32:00.0000] <Aleoss>
Vito: I'm designing my page for HTML5 using HTML5 standards. I'm not going to test it through a doctype that I don't intend on using. That would leave room for REAL errors.

[22:32:01.0000] <Aleoss>
Vito: Thanks for confirming my conserns in regards to Tidy.

[22:36:00.0000] <Aleoss>
Truly is a pity..

[22:37:00.0000] <Vito`>
Why are you using Tidy in particular?  Isn't the other validator enough?

[22:37:01.0000] <Vito`>
(I've never personally used Tidy, so I don't know its benefits for < HTML5)

[22:40:00.0000] <Aleoss>
Vito: I like to use Tidy (Serial) + Total Validator (Serial with WAI-AAA WCAG) + W3 HTML Validator.

[22:42:00.0000] <Aleoss>
Vito: And if I'm really bored, I'll validate with cynthiasays.com as well.

[22:44:00.0000] <Aleoss>
Vito: Call it over-kill if you wish, but I call it getting the opinions of various sources to ensure 100% accessibility and validity.

[00:13:00.0000] <a1>
Hello

[00:13:01.0000] <a1>
Is this the place to ask questions about HTML5?

[00:14:00.0000] <Vito`>
that's what the topic in #html says ;)

[00:15:00.0000] <a1>
My question: is it possible to add <audio> tag from javascript, make it play file from some javascript scalar; it contains actual OGG data and not link to it?

[00:16:00.0000] <a1>
var audioElement = document.createElement('audio'); audioElement.setAttribute('src','data:audio/ogg;'+data);audioElement.play();

[00:16:01.0000] <a1>
this doesn't work

[00:17:00.0000] <a1>
Converting to base64 with javascript seems to be pretty slow

[00:18:00.0000] <Vito`>
how interesting

[00:18:01.0000] <Vito`>
does it work with a data URL without using javascript to inject the whole thing?

[00:18:02.0000] <Vito`>
as a regular <audio> tag in the page?

[00:19:00.0000] <Vito`>
the notes on MDC around data URLs say you shouldn't expect long data URLs to work everywhere, and I imagine an OGG file would be huge

[00:20:00.0000] <a1>
It works in firefox; 2MB ogg file

[00:20:01.0000] <a1>
It doesn't have to be data:

[00:21:00.0000] <a1>
I just want to play file I have stored in javascript variable.

[00:22:00.0000] <a1>
I actually only need this to work on firefox, as I'm writing a greasemonkey script

[00:32:00.0000] <Vito`>
a1: are you actually adding the element to the page?  appending it to the body or some such?  and have you tried delaying the play to make sure it has time to attach to the DOM?

[00:33:00.0000] <Vito`>
I haven't tried this myself but if it works with a data URL as part of the HTML, I'd expect it to also work when it's injected

[00:33:01.0000] <a1>
I'll try

[00:53:00.0000] <roc>
a1: that shoudl work

[00:53:01.0000] <roc>
the element doesn't need to be in the DOM

[00:54:00.0000] <roc>
if you base64 encode the data, you need to add 'base64' to your data: URL

[00:54:01.0000] <a1>
I think one problem is that 'data:audio/ogg;'+data should be 'data:audio/ogg;'+encodeURI(data)

[00:55:00.0000] <erlehmann>
a1, roc, cn you repost the offending data URI ?

[00:55:01.0000] <erlehmann>
just joined

[00:55:02.0000] <a1>
no, it's nothing

[00:55:03.0000] <a1>
problem in me, not data

[00:56:00.0000] <zcorpan>
al, it should be a comma, not a semicolon

[00:56:01.0000] <zcorpan>
s/al/a1/

[00:57:00.0000] <erlehmann>
… anyone visiting Berlin for christmas and attending 26C3 ? (Won't ask again.)

[01:03:00.0000] <GPHemsley>
Does HTML5 define named entities? Or point to a reference of them somewhere else?

[01:04:00.0000] <zcorpan>
it defines them

[01:04:01.0000] <GPHemsley>
Oh, found it: Section 9.6

[01:05:00.0000] <GPHemsley>
Called them "Named character references"

[01:05:01.0000] <zcorpan>
yeah, they were called entities before

[01:05:02.0000] <zcorpan>
but people said it clashed with xml's terminology

[01:06:00.0000] <GPHemsley>
ah

[01:07:00.0000] <GPHemsley>
Hixie: It'd be nice if that table also had a column that showed what the symbols looked like (e.g. by using the &#x....; form and letting the browser render them)

[01:09:00.0000] <zcorpan>
GPHemsley: there are nicer interfaces than a big table if you want to find and use a character...

[01:10:00.0000] <GPHemsley>
zcorpan: True. But my mind always goes straight to source when wondering about something.

[01:12:00.0000] <GPHemsley>
Hixie: BTW, the definition of "Named character references" in section 9.1.4 is in conflict with the table in section 9.6, specifically with regard to the point about having to end in a semicolon. The table has duplicate entries for NCRs that do not end with a semicolon (presumably, for backwards compatibility).

[01:14:00.0000] <zcorpan>
GPHemsley: 9.1.4 excludes those without semicolon

[01:15:00.0000] <GPHemsley>
zcorpan: Is that not the only place that references the NCRs defined in 9.6?

[01:15:01.0000] <GPHemsley>
IOW, wouldn't that make the NCRs without a semicolon in 9.6 orphans?

[01:17:00.0000] <zcorpan>
the parser also references 9.6

[01:17:01.0000] <GPHemsley>
which section is that?

[01:17:02.0000] <zcorpan>
9.2

[01:19:00.0000] <GPHemsley>
hmm... can't find where; do you have a subsection?

[01:19:01.0000] <zcorpan>
9.2.4.70 Tokenizing character references

[01:20:00.0000] <GPHemsley>
ah, that's why... 9.2.4 is a separate page :P

[01:22:00.0000] <GPHemsley>
zcorpan: So, are you referring to the "Anything else"."Otherwise, ..." part?

[01:25:00.0000] <zcorpan>
i see 2 references to 9.6 in "anything else"

[01:26:00.0000] <GPHemsley>
right... I'm asking specifically about the NCRs that do not end in a semicolon

[01:26:01.0000] <zcorpan>
what about them?

[01:27:00.0000] <Dashiva>
"Otherwise, if the next character is a U+003B SEMICOLON, consume that too. If it isn't, there is a parse error."

[01:27:01.0000] <Dashiva>
That part?

[01:27:02.0000] <GPHemsley>
zcorpan: well, you said they weren't orphans and/or that they were excluded from 9.1.4

[01:28:00.0000] <GPHemsley>
Dashiva: No, "Otherwise, return a character token for the character ..."

[01:28:01.0000] <GPHemsley>
zcorpan: So I was asking where they were included.

[01:28:02.0000] <zcorpan>
if the input is foo&aumlbar then the parser will treat it as fooäbar

[01:28:03.0000] <zcorpan>
with a parse error

[01:29:00.0000] <zcorpan>
"Consume the maximum number of characters possible, with the consumed characters matching one of the identifiers in the first column of the named character references table (in a case-sensitive manner)."

[01:29:01.0000] <GPHemsley>
right, that's what I was asking about

[01:31:00.0000] <GPHemsley>
I see, so the ones without the semicolon are still processed, but they also raise a parse error?

[01:31:01.0000] <zcorpan>
yes

[01:31:02.0000] <GPHemsley>
If, however, you use an NCR that does not have a semicolon-less counterpart, then it's just a parse error and nothing is shown, or everything is shown?

[01:32:00.0000] <zcorpan>
then it's a parse error but the entity is not replaced

[01:32:01.0000] <zcorpan>
because it's not an entity

[01:32:02.0000] <GPHemsley>
so, it's displayed as literal?

[01:32:03.0000] <zcorpan>
yes

[01:32:04.0000] <GPHemsley>
gotcha, OK

[01:32:05.0000] <GPHemsley>
thanks

[01:35:00.0000] <a1>
I managed to get it to work

[01:35:01.0000] <a1>
The problem was that data I was getting from xmlhttprequest was interpreted as utf-8 string

[01:36:00.0000] <a1>
and was base64 encoded incorrectly

[01:36:01.0000] <a1>
Thanks

[01:39:00.0000] <zcorpan>
a1 wants xhr2's responseBody

[01:41:00.0000] <zcorpan>
or possibly <audio src>

[01:50:00.0000] <GPHemsley>
Does HTML5 make any requirements for the style of the bullet for <li> with regard to embedded lists?

[01:51:00.0000] <gsnedders>
GPHemsley: No, HTML 5 has no rendering requirements

[01:54:00.0000] <GPHemsley>
gsnedders: Hmm... that's not entirely true: http://www.whatwg.org/specs/web-apps/current-work/multipage/rendering.html#punctuation-and-decorations

[01:55:00.0000] <GPHemsley>
But for some reason it specifies differing/alternating styles for embedded uls, but not embedded ols

[01:56:00.0000] <gsnedders>
That's non-normative

[01:56:01.0000] <gsnedders>
So it doesn't require anything :P

[01:56:02.0000] <GPHemsley>
:P

[01:57:00.0000] <GPHemsley>
looks like I'll have to specify ol[type]

[01:57:01.0000] <zcorpan>
type is invalid

[01:57:02.0000] <GPHemsley>
ugh, was just looking that up

[02:10:00.0000] <a1>
It's me again with same question.

[02:10:01.0000] <a1>
I have a javascript scalar contatining OGG stream

[02:11:00.0000] <a1>
it starts with "OggS\0\2" and contains some music I'd like to be able to play with it

[02:11:01.0000] <a1>
with <audio> tag

[02:12:00.0000] <a1>
It's possible to play it by setting src attribute of audio to 'data:audio/ogg;base64,'+encode64(a)

[02:12:01.0000] <a1>
but encoding huge sound file with javascript is very slow

[02:12:02.0000] <a1>
my browser freezes

[02:12:03.0000] <a1>
So I'm asking

[02:12:04.0000] <a1>
is there a way to set data of <audio> tag, and not source of that data with javascript?

[02:16:00.0000] <zcorpan>
no

[02:16:01.0000] <zcorpan>
do you need to base64 it though?

[02:16:02.0000] <a1>
well, it's binary data

[02:17:00.0000] <zcorpan>
sure

[02:17:01.0000] <a1>
'data:audio/ogg,'+a doesn't work

[02:17:02.0000] <zcorpan>
does encodeURIComponent(a) work?

[02:17:03.0000] <a1>
'data:audio/ogg,'+encodeURIComponent(a)

[02:17:04.0000] <a1>
this?

[02:17:05.0000] <a1>
No..

[02:18:00.0000] <zcorpan>
hmm, wonder why not

[02:18:01.0000] <jgraham>
zcorpan: encodeURIComponent does something quite different

[02:18:02.0000] <jgraham>
Doesn't it?

[02:19:00.0000] <jgraham>
Oh wait

[02:19:01.0000] <a1>
Output of encodeURIComponent(a) seems to be okay...

[02:19:02.0000] <a1>
"OggS%00%02%00%00%00%00%00%00%00%00'%EF%9F%B7%EF%9E%83Z%00%

[02:19:03.0000] <jgraham>
/me was answering a different question

[02:20:00.0000] <jgraham>
a1: Have you tried using a2b or whatever it's called, in browsers that support it?

[02:20:01.0000] <zcorpan>
atob

[02:22:00.0000] <jgraham>
/me wonders why that isn't a part of html5

[02:22:01.0000] <zcorpan>
file a bug? :)

[02:26:00.0000] <a1>
It's data I get from XMLHttpRequest, and characters that >127 need to be and-ed with 0xff

[02:26:01.0000] <a1>
calling btoa on that results in exception

[02:29:00.0000] <zcorpan>
why do you get the data with xhr rather than with <audio src>?

[02:30:00.0000] <a1>
Basically, it's a jpg image catenated with ogg audio file. I can't just specify that as source of <audio>

[02:30:01.0000] <a1>
I get it with xhr, throw away jpeg data, and feed what's left to <audio>

[02:31:00.0000] <jgraham>
OK new question: Why do you have a JPEG image concated with an OGG audio file?

[02:31:01.0000] <a1>
some widely used discussion board allows uploading of images

[02:32:00.0000] <a1>
I want to make to make it possible to allow users to upload and listen to sounds

[02:32:01.0000] <a1>
with greasemonkey

[02:32:02.0000] <gsnedders>
/me notes you'll hit issues with some browsers with null bytes in the data, and with stuff that isn't UTF-8

[02:33:00.0000] <jgraham>
/me wonders if you could use some clever hack with <img> <canvas> and getImageData

[02:34:00.0000] <a1>
getImageData?

[02:35:00.0000] <a1>
That returns pixels, doesn't it?

[02:36:00.0000] <zcorpan>
yes

[03:16:00.0000] <zcorpan>
can we update http://www.w3.org/QA/2002/04/valid-dtd-list.html ?

[03:18:00.0000] <hsivonen>
zcorpan: maybe dom could if you can convince him.

[03:19:00.0000] <zcorpan>
and the referenced http://www.w3.org/QA/Tips/Doctype has factual errors

[03:30:00.0000] <zcorpan>
does firefox still look at content-style-type?

[03:33:00.0000] <karlcow>
zcorpan: for the qa tip it is easy to modify, there is a process

[03:33:01.0000] <zcorpan>
what's the process?

[03:33:02.0000] <karlcow>
for the dtd list it's easy to do. 

[03:34:00.0000] <karlcow>
zcorpan: http://www.w3.org/QA/Tips/

[03:35:00.0000] <karlcow>
if you want I can send some suggestions, I see indeed some factual errors

[03:36:00.0000] <karlcow>
Tools which process HTML documents, such as Web browsers, need to know which DTD an (X)HTML document

[03:36:01.0000] <karlcow>
s/such as Web browsers//

[03:43:00.0000] <zcorpan>
sent an email to dom

[03:47:00.0000] <karlcow>
zcorpan: hehe http://lists.w3.org/Archives/Public/public-evangelist/2009Dec/0000

[03:48:00.0000] <zcorpan>
karlcow: thanks. my email was about the other document

[03:48:01.0000] <karlcow>
excellent

[03:52:00.0000] <zcorpan>
karlcow: i'd also suggest replacing the first further reading link with http://hsivonen.iki.fi/doctype/

[03:55:00.0000] <karlcow>
http://www.alistapart.com/stories/doctype/ plus this one has went the way of the dodo. the one of hsivonen should be added too maybe a bit technical for certain people but good to have indeed.

[04:00:00.0000] <gsnedders>
karlcow: http://www.alistapart.com/articles/doctype/ I guess is it

[06:01:00.0000] <hsivonen>
the third Definition below says only the 5 XML ones

[06:02:00.0000] <hsivonen>
MikeSmith: so I suggest making a copy of NamedCharacters that only has 5 entries

[06:03:00.0000] <MikeSmith>
that'd certainly make things easier

[06:04:00.0000] <MikeSmith>
but are you sure that's what the third Definition is actually saying?

[06:04:01.0000] <MikeSmith>
I can't see that it's stating that restriction

[06:04:02.0000] <hsivonen>
zcorpan: ^ spec not clear enough :-)

[06:05:00.0000] <MikeSmith>
hsivonen: seems like there are two parts to what is says there

[06:05:01.0000] <MikeSmith>
"Each CharRef in PseudoAttValue is replaced with the character it represents according to XML. [XML]"

[06:05:02.0000] <MikeSmith>
which isn't stating any restriction

[06:06:00.0000] <MikeSmith>
and "Each PredefEntityRef in PseudoAttValue is replaced with with U+0026 (&) if it is "&amp;", U+003C (<) if it is "&lt;", U+003E (>) if it is "&gt;", U+0022 (") if it is "&quot;" and U+0027 (') if it is "&apos;"."

[06:06:01.0000] <hsivonen>
MikeSmith: CharRef is a numeric reference, isn't it?

[06:06:02.0000] <hsivonen>
and PredefEntityRef is a named ref

[06:06:03.0000] <MikeSmith>
/me looks at XML spec now

[06:07:00.0000] <MikeSmith>
hsivonen: yeah, OK, CharRef is numeric

[08:08:00.0000] <foolip>
is there even a validator for XHTML5+RDFa that includes the microdata attributes?

[08:09:00.0000] <erlehmann>
dunno lol :3

[08:10:00.0000] <foolip>
I doubt we'll see one a validator.nu

[08:10:01.0000] <foolip>
at

[08:12:00.0000] <erlehmann>
Hmmm, if both is not possible at once, I've no idea how to continue. The Creative Commons person says RDFa is the way to go, Hixie wants Microdata … should I just make it an option in the interface?

[08:12:01.0000] <erlehmann>
like: ( ) RDFa (x) Microdata

[08:13:00.0000] <erlehmann>
Pro: Not my problem anymore. Con: Users will likely not understand anything at all.

[08:14:00.0000] <foolip>
I doubt any consumers care about validity and since the attributes don't clash they won't interfere with each other

[08:14:01.0000] <foolip>
but it seems a bit odd to support both in the same document

[08:14:02.0000] <foolip>
is there any benefit to supporting RDFa, like Google giving preferential treatment or something?

[08:15:00.0000] <erlehmann>
Anyway, nice Trilemma … Choose two: Microdata, RDFa, Valid Markup. :D

[08:15:01.0000] <foolip>
hehe

[08:16:00.0000] <erlehmann>
Well, there are already RDFa consumers out there, I assume.

[08:16:01.0000] <foolip>
I know which one I'd sacrifice, but then I'm kind of biased.

[08:17:00.0000] <erlehmann>
Hehe.

[08:18:00.0000] <foolip>
why not also stuff some microformats in there

[08:18:01.0000] <foolip>
at least that won't affect validity :)

[08:21:00.0000] <erlehmann>
=_=

[08:22:00.0000] <jgraham>
Maybe you could add the data as RDF/XML hidden inside commnts and as JSON blocks inside <script type="application/json"> while you're at it :)

[08:22:01.0000] <erlehmann>
Got it. I'll stay with Microformats for now and fix the styling. Is there some wiki page with a list of Microformat consumers ?

[08:23:00.0000] <foolip>
right, wasn't CC doing RDF/XML in comments?

[08:24:00.0000] <erlehmann>
well, since i link to the CC license deed … you can have microformats in the generated code and RDF in CC territory ;)

[08:24:01.0000] <jgraham>
foolip: Not sure. Trackback use(es|d) it I believe

[08:24:02.0000] <Philip`>
Anyone who wants to extract licensing information from the web will have to support whatever markup CC currently tells people to use

[08:25:00.0000] <Philip`>
so you might as well just use that, rather than adding complexity by requiring consumers support more types of markup

[08:25:01.0000] <MikeSmith>
Philip`: OK - thanks (about EBNF question)

[08:25:02.0000] <jgraham>
/me wonders how similar the set of all people who want to extract licesnisng information is to the empty set

[08:25:03.0000] <jgraham>
Do Google or Yahoo or Microsoft or someone read this data?

[08:26:00.0000] <Philip`>
jgraham: Well, I guess they're not disjoint sets so there is some similarity

[08:26:01.0000] <jgraham>
(and do people then make use of it when they read it?)

[08:26:02.0000] <Philip`>
*dissimilarity

[08:26:03.0000] <jgraham>
Philip`: :p

[08:26:04.0000] <erlehmann>
Dunno, Google or Yahoo at least have this data on their pages.

[08:26:05.0000] <erlehmann>
Damn, now I have to learn Git branching.

[08:26:06.0000] <Philip`>
I'm fairly sure some search engines can use it

[08:26:07.0000] <foolip>
I know I often search for a specific license on google images or flickr, perhaps some of that information came from scraping

[08:26:08.0000] <Philip`>
http://search.yahoo.com/web/advanced?ei=UTF-8

[08:27:00.0000] <Philip`>
"Search only for Creative Commons licensed content"

[08:27:01.0000] <foolip>
I don't remember ever finding anything with google though, perhaps the feature doesn't exist

[08:28:00.0000] <jgraham>
foolip: I use it on flickr but that doesn't require scraping since flickr have access to their own db

[08:29:00.0000] <foolip>
I used to use CC:s search tool, but that's just a frontend for other things

[08:29:01.0000] <Philip`>
/me wonders how common it is for sites to scrape themselves

[08:29:02.0000] <TabAtkins>
That would be kinda weird.

[08:29:03.0000] <Philip`>
I suppose lots have generic search engines that basically do that

[08:30:00.0000] <TabAtkins>
But yeah, generic search engines would be the common use case.

[08:30:01.0000] <TabAtkins>
I suppose I do that?  I use Google Custom Search on my company's site.

[08:30:02.0000] <jgraham>
TabAtkins: So you just confessed to weirdness?

[08:30:03.0000] <TabAtkins>
Yes.

[08:31:00.0000] <Philip`>
I was thinking more of ones that are self-hosted, like Ultraseek or (to some extent) the Google Search Appliance

[08:31:01.0000] <Philip`>
where they could technically get access to the private databases, but for practical purposes they probably just read data via the site's normal web interface

[08:31:02.0000] <foolip>
Philip`: does "scraping" with JavaScript count?

[08:33:00.0000] <foolip>
in that case any site using the Microdata DOM API (none yet I assume) would be scraping themselves

[08:34:00.0000] <Philip`>
/me doesn't have a sufficiently coherent point to be able to work out whether JS counts or not

[08:34:01.0000] <karlushi>
jgraham, I think you can imagine tools like tumblr, posterous or any blogging tools included in aggregator.

[08:35:00.0000] <karlushi>
parsing the content for the license and adding the credits in the reblogged page.

[08:36:00.0000] <karlushi>
another cool feature would be an authoring creating a blockquote with the right credits, authors, and/or license information

[08:36:01.0000] <karlushi>
s/authoring/authoring tool/

[08:57:00.0000] <erlehmann>
foolip, Microdata DOM isn't implemented yet, is it?

[08:59:00.0000] <foolip>
erlehmann: not in any browser, no

[09:00:00.0000] <erlehmann>
well, one could fake it with javascript pretty easily.

[09:01:00.0000] <foolip>
yes, which is what I've been playing with

[12:46:00.0000] <a1>
Hello.

[12:46:01.0000] <a1>
Could you please have a look at this:

[12:46:02.0000] <a1>
http://no-info.no-ip.info:6224/sound/test.html

[12:47:00.0000] <a1>
It's an audio tag with data: URI as its src

[12:47:01.0000] <a1>
I'm using firefox, and play/pause/seek functions are not working untl the whole sound has finished playing

[12:47:02.0000] <a1>
Is this how it should be working?

[12:50:00.0000] <annevk>
nope

[12:51:00.0000] <a1>
Bug in firefox or in my web page?

[12:52:00.0000] <annevk>
if your code is correct it's a bug in Firefox

[12:52:01.0000] <annevk>
haven't looked at the code since my connection is crappy

[12:52:02.0000] <a1>
<html><head></head><body><audio src="data:audio/ogg;base64,ALL-DATA-HERE" controls /></body></html>

[12:54:00.0000] <annevk>
looks good to me

[12:54:01.0000] <annevk>
well

[12:55:00.0000] <annevk>
<audio> needs an explicit closing tag

[12:55:01.0000] <annevk>
in text/html documents

[12:57:00.0000] <a1>
Still same behavior

[13:04:00.0000] <annevk>
i'd file a bug then

[13:19:00.0000] <jgraham>
annevk: I don't really know why you think the question about contexts is rhetorical. It seems you might mean "obviously not" but clearly you can write portable code by just using the 2D context or by doing feature testing or...

[13:19:01.0000] <jgraham>
So if by "that is rhetorical, right?" you meant "yes" it wasn't very clear

[13:36:00.0000] <roc>
a1: works for me

[13:36:01.0000] <roc>
oh, I see, we don't get the correct duration

[13:37:00.0000] <roc>
the first time through

[13:37:01.0000] <a1>
and can

[13:37:02.0000] <a1>
't seek

[13:37:03.0000] <roc>
right, because we only support seeking using HTTP

[13:38:00.0000] <a1>
'We' firefox or 'We' html?

[13:38:01.0000] <a1>
I filed a firefox bug.

[13:38:02.0000] <roc>
We firefox

[13:41:00.0000] <roc>
the problem is that Ogg doesn't include the duration at the start of the file, so we have to seek to the end of the file to get it, and that requires a transport like HTTP that supports seeking. We'll need to add some code so that when the download finishes, if we've got everything in the media cache we look at the data at the end to get the duration

[13:42:00.0000] <a1>
But the whole thing is loaded when player appears.

[13:42:01.0000] <a1>
"controls" goes after "src=data:..."

[13:42:02.0000] <a1>
so shouldn't whole steam already be in memory?

[13:42:03.0000] <Vito`>
oh neat

[13:43:00.0000] <Vito`>
glad you sorted it out a1

[13:43:01.0000] <a1>
Yeah, thanks

[13:43:02.0000] <a1>
apparently base64 was relatvely fast and some regexp I used to throw away data from picture was making firefox freeze

[14:06:00.0000] <annevk>
jgraham, okay...

[14:06:01.0000] <annevk>
jgraham, I was just going from if things work differently it's bad for interop

[15:14:00.0000] <abarth>
i'm really tired of the CORS/UM discussion on webapps

[15:15:00.0000] <abarth>
it doesn't seem to be going anywhere

[15:16:00.0000] <othermaciej>
abarth: did my suggestion that people only post if they have new information to add not work?

[15:16:01.0000] <abarth>
tyler and ian are still talking about x-frame-options

[15:18:00.0000] <abarth>
(i should stop complaining and just ignore the threads)


2009-12-22
[17:29:00.0000] <othermaciej>
abarth, Hixie: the specs themselves do not say anything about deleting the other spec from the universe

[17:29:01.0000] <othermaciej>
abarth, Hixie: thus, by proposing to advance both, the chair in question has sort of decided the controversy

[17:30:00.0000] <abarth>
yes, although that might have been before tyler and mark make their "UM and no CORS" position clear

[17:30:01.0000] <abarth>
i failed to refrain from posting on that thread :(

[17:31:00.0000] <othermaciej>
I can't even bring myself to read the latest emails right now

[17:32:00.0000] <othermaciej>
but I think I will personally try not to post unless I have new information to add

[17:46:00.0000] <AryehGregor>
roc, BTW, <https://bugzilla.mozilla.org/show_bug.cgi?id=535043>.  Do you know what kind of ballpark response time I should expect?  There seem to be 53 bugs with review requested from jst, some years old.

[17:53:00.0000] <roc>
try asking for an ETA in the bug

[17:54:00.0000] <AryehGregor>
Is he automatically CCed?

[17:54:01.0000] <roc>
no, good point

[17:54:02.0000] <roc>
CC him

[17:56:00.0000] <AryehGregor>
k, thanks.

[00:02:00.0000] <GPHemsley>
Is there any semantic value to @class?

[00:02:01.0000] <GPHemsley>
Or is it purely a connection to style?

[00:20:00.0000] <MikeSmith>
GPHemsley: is that a trick question?

[00:20:01.0000] <GPHemsley>
um... no?

[00:21:00.0000] <MikeSmith>
OK, it seems clear it does have semantic value

[00:22:00.0000] <MikeSmith>
though I wouldn't be surprised to find somebody dispute that for whatever reason

[00:23:00.0000] <GPHemsley>
so then it can be used (semantically) in lieu of an element that doesn't exist in the standard?

[00:23:01.0000] <MikeSmith>
yeah, that's part of the intent

[00:23:02.0000] <GPHemsley>
right, OK

[00:23:03.0000] <GPHemsley>
thanks

[00:23:04.0000] <MikeSmith>
but the semantics and style often go hand-in-hand

[00:23:05.0000] <GPHemsley>
well, yes and no

[00:24:00.0000] <MikeSmith>
that's why I qualified it with "often"..

[00:24:01.0000] <GPHemsley>
I mean, if something were fed a pure HTML5 document and was asked to extract the data, the styling wouldn't matter

[00:24:02.0000] <MikeSmith>
true that

[00:25:00.0000] <MikeSmith>
but on the other hand, if some part of you content means something different you often want to dislay it to users differently

[00:25:01.0000] <GPHemsley>
and, in that case, <var> and <span class="expression"> (for example) would hold the same weight, correct?

[00:25:02.0000] <MikeSmith>
no, I don't think so

[00:25:03.0000] <MikeSmith>
<var> has a standard, documented meaning

[00:26:00.0000] <MikeSmith>
class="expression" means whatever you choose it to mean, and nobody else can know for sure what you mean by it

[00:26:01.0000] <GPHemsley>
well, assume that the meaning of each class was documented and used standardly

[00:26:02.0000] <MikeSmith>
OK

[00:27:00.0000] <MikeSmith>
also, all that said, you could also misuse <var> in a way that's at odd with with its standard meaning

[00:27:01.0000] <GPHemsley>
right

[00:28:00.0000] <GPHemsley>
this situation, however, assumes that whatever is doing the processing knows what everything means and expects everything to be used as defined

[00:28:01.0000] <GPHemsley>
so it expects <var> to be used correctly

[00:29:00.0000] <GPHemsley>
(and also <span class="expression">)

[00:32:00.0000] <GPHemsley>
...you can't nest <span>s?

[00:34:00.0000] <GPHemsley>
or Firefox can't... o_0

[00:41:00.0000] <GPHemsley>
oh, nope, my mistake

[00:41:01.0000] <GPHemsley>
/me is an idiot

[00:45:00.0000] <GPHemsley>
(overzealous find and replace)

[01:49:00.0000] <gsnedders>
Weee… Opera 10.50 pre-alpha shipped.

[01:49:01.0000] <gsnedders>
</mindlessspam>

[01:55:00.0000] <erlehmann>
gsnedders, AV-enabled?

[02:01:00.0000] <gsnedders>
erlehmann: Coming soon

[02:02:00.0000] <erlehmann>
/me needs more theora. :3

[02:02:01.0000] <doublec>
gsnedders, new JS engine looks interesting :)

[02:02:02.0000] <gsnedders>
doublec: Of course, it's what I've been working on ;P

[02:03:00.0000] <doublec>
haha, nice. I'm hearing good stuff about it on reddit and places

[02:03:01.0000] <doublec>
i'll have to try my 8080 emulator on it

[02:36:00.0000] <hsivonen>
Does Opera give any hints on what sort of general approach or techniques they used in the new JS engine?

[02:38:00.0000] <Rik`>
hsivonen: http://my.opera.com/core/blog/2009/02/04/carakan ?

[02:38:01.0000] <hsivonen>
Rik`: thanks

[02:39:00.0000] <hsivonen>
/me wonders if Opera still intends to keep Futhark around for small devices

[02:40:00.0000] <Philip`>
hsivonen: http://my.opera.com/core/blog/2009/12/22/carakan-revisited

[02:40:01.0000] <hsivonen>
thanks

[02:42:00.0000] <jgraham>
doublec: iirc your 8080 emulator claims that we don't support canvas

[02:42:01.0000] <jgraham>
(so it would be nice if you would fix that :)

[02:47:00.0000] <jgraham>
BTW if anyone is playing with the labs build and finds a site that breaks in a likely-javascript related way and doesn't want to file a bug for some reason (e.g. laziness), please let gsnedders or I know

[03:17:00.0000] <Philip`>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/337 - Opera 10.50 seems quite nicely fast at that

[03:18:00.0000] <Philip`>
(I get about 14 million ints/sec, vs about 2 M/sec in Firefox 3.5, and approximately zero in Opera 10.10)

[03:21:00.0000] <gsnedders>
(Though with holidays I dunno how we'll be around)

[03:36:00.0000] <doublec>
jgraham: I detect for createImageData and display the message if it doesn't exist

[03:36:01.0000] <doublec>
jgraham: does Opera not support it?

[03:36:02.0000] <gsnedders>
doublec: It does

[03:36:03.0000] <doublec>
odd -I wonder why it fails

[03:47:00.0000] <Lachy>
The announcement for Selectors API being published as CR should come out some time later today

[03:47:01.0000] <doublec>
gsnedders: document.getElementById('screen').getContext('2d').createImageData is undefined 

[03:48:00.0000] <doublec>
gsnedders: that's why the 8080 emulator is not running

[03:49:00.0000] <doublec>
gsnedders: tested in 10.50 pre-alpha using the js command line 

[03:49:01.0000] <doublec>
gsnedders: 'screen' is a <canvas id='screen'>

[03:50:00.0000] <gsnedders>
doublec: Does it work in 10.10?

[03:50:01.0000] <doublec>
gsnedders: I'll check

[03:50:02.0000] <gsnedders>
doublec: Or does it work with JIT disabled?

[03:50:03.0000] <gsnedders>
That seems really like a normal JIT bug :)

[03:51:00.0000] <gsnedders>
(Namely, random things being undefined)

[03:51:01.0000] <doublec>
gsnedders: how do I disable the jit?

[03:51:02.0000] <gsnedders>
doublec: opera:config, search for JIT

[03:57:00.0000] <doublec>
gsnedders: doesn't work in 10.10 either

[03:57:01.0000] <doublec>
gsnedders: on 10.5 with jit disabled same issue

[03:57:02.0000] <gsnedders>
doublec: Link to the page?

[03:57:03.0000] <doublec>
gsnedders: http://www.bluishcoder.co.nz/js8080/

[03:58:00.0000] <gsnedders>
Oh, apparently we don't sup[port it.

[03:58:01.0000] <gsnedders>
*support

[03:59:00.0000] <doublec>
that'll be it then :)

[04:05:00.0000] <erlehmann>
gsnedders, what in-joke is that with the power plants mentioned in the caveat?

[04:06:00.0000] <gsnedders>
huh?

[05:15:00.0000] <gsnedders>
First reply from uni application! A rejection!

[05:15:01.0000] <gsnedders>
(I suck at CS.)

[11:03:00.0000] <nirvdrum>
Hi.  Is the validator.nu validator known to be broken in trunk?  I'm get a ClassCastException when running it and before I spend too much time tracking it down, wanted to make sure others weren't already aware of it.

[11:09:00.0000] <nirvdrum>
Hmm . . . I found IRC logs highlighting the same issue going back to May.

[11:23:00.0000] <nirvdrum>
Alright.  If I run "all" a second time, it works.

[11:23:01.0000] <nirvdrum>
Repeatable, too.

[11:44:00.0000] <jgraham>
gsnedders: s'ton?

[12:28:00.0000] <hsivonen>
nirvdrum: try rerunning python build.py localent

[12:30:00.0000] <nirvdrum>
hsivonen: I hadn't run localent explicitly at all.  I just ran the "all" target.  Re-running that fixes the problem.

[12:31:00.0000] <nirvdrum>
I'll see if "localent" after "all" fixes it, too, but it might be hard to tell.

[12:33:00.0000] <nirvdrum>
hsivonen: I don't suppose there's any way to build a WAR out of this, is there?

[12:33:01.0000] <hsivonen>
nirvdrum: there isn't, because I haven't needed a WAR

[12:34:00.0000] <nirvdrum>
I couldn't quite work out how you were serving out of jetty.  Where is the web context being set up?

[12:35:00.0000] <hsivonen>
Main.java

[12:35:01.0000] <hsivonen>
in the validator svn repo

[12:35:02.0000] <hsivonen>
afk

[12:36:00.0000] <nirvdrum>
So, I guess I could write something to unpack the validator JAR and set up the WAR structure, adding in the dependencies to WEB-INF/lib.

[12:38:00.0000] <nirvdrum>
It'd just be handy to have a WAR because I'm already running Jetty and the CSS validator on the same machine.

[12:39:00.0000] <nirvdrum>
hsivonen: My other question is if I run it with th "run" target, how do you recommend detaching from the tty and capturing the pid?  I could do it with nohup and some bash incantation, but I'd assume there's a cleaner way.

[12:49:00.0000] <nirvdrum>
hsivonen: all => localent => run still induces the CCE.


2009-12-23
[16:29:00.0000] <AryehGregor>
/me doesn't understand what annevk's latest blog post means

[16:38:00.0000] <jwalden>
it seems like the sort of post sayrer would make in its style (perhaps its content as well) (ohai, sayrer, if you get pointed at this in logs ;-) )

[17:21:00.0000] <AryehGregor>
That prompted me to start reading through sayrer's blog history.  Interesting how many of the features he lists here did actually get removed: http://blog.mozilla.com/rob-sayre/2008/02/19/bloaty-parts-of-the-whatwg-html5-specification-that-should-be-removed/

[18:26:00.0000] <othermaciej>
AryehGregor: 6/17 it looks like

[01:23:00.0000] <hsivonen>
MikeSmith: I blogged: http://hsivonen.iki.fi/html5-parser-improvements/

[01:29:00.0000] <MikeSmith>
hsivonen: cool

[01:29:01.0000] <MikeSmith>
/me reads

[01:30:00.0000] <MikeSmith>
hsivonen: I will try to write up a whatwg blog entry tomorrow about other recent changes (non-parser ones)

[01:30:01.0000] <MikeSmith>
the warnings stuff etc.

[01:31:00.0000] <MikeSmith>
I guess it's mostly minor stuff as far as the end-user experience goes

[01:31:01.0000] <hsivonen>
MikeSmith: I'll deploy the parser changes to V.nu and update the post with a mention that the changes are on V.nu, too

[01:31:02.0000] <MikeSmith>
great

[01:35:00.0000] <MikeSmith>
hsivonen: so the parser changes will potentially change some v.nu behavior a bit, right?

[01:36:00.0000] <MikeSmith>
I mean for <title> and <script>, etc.

[01:36:01.0000] <MikeSmith>
especially script

[01:37:00.0000] <hsivonen>
MikeSmith: yeah, script, title, <table><form>, the ol/ul stuff, <frameset>

[01:37:01.0000] <MikeSmith>
Ok

[01:39:00.0000] <MikeSmith>
hsivonen: no changes to attribute parsing?

[01:39:01.0000] <MikeSmith>
or rather tokenizing

[01:39:02.0000] <hsivonen>
MikeSmith: none that I could recall

[01:39:03.0000] <MikeSmith>
OK

[01:39:04.0000] <hsivonen>
(after actually looking at the change log messages)

[01:42:00.0000] <MikeSmith>
I think for my own edification I'll probably take a look at the change log and diffs, try to learn a bit more about the parser code

[01:42:01.0000] <MikeSmith>
though I guess it's a probably a pretty big set of diffs

[01:45:00.0000] <hsivonen>
MikeSmith: the tokenizer diffs are big

[01:45:01.0000] <hsivonen>
MikeSmith: the tree builder diff not so big

[01:46:00.0000] <MikeSmith>
OK

[01:46:01.0000] <hsivonen>
MikeSmith: deployed

[01:46:02.0000] <hsivonen>
MikeSmith: including your recent changes

[01:47:00.0000] <MikeSmith>
ah great

[01:48:00.0000] <MikeSmith>
hsivonen: btw, about the xml-stylesheet checking

[01:48:01.0000] <MikeSmith>
the <!DOCTYPE x[<?xml-stylesheet href="x"?>]><x/> case

[01:48:02.0000] <MikeSmith>
I did what you suggested

[01:48:03.0000] <MikeSmith>
in CheckerValidator

[01:49:00.0000] <hsivonen>
MikeSmith: did it work as expected?

[01:49:01.0000] <MikeSmith>
partly

[01:49:02.0000] <hsivonen>
what didn't work?

[01:49:03.0000] <MikeSmith>
I changed the return null on getDTDHandler to return (checker instanceof DTDHandler) ? (DTDHandler)checker : null

[01:50:00.0000] <MikeSmith>
and I can see that's working as expected

[01:51:00.0000] <MikeSmith>
but in my subclass of the checker, it's not yet having any effect

[01:51:01.0000] <hsivonen>
MikeSmith: no effect as in not getting any DTDHandler calls?

[01:51:02.0000] <MikeSmith>
I think there's something I'm missting

[01:51:03.0000] <MikeSmith>
*missing

[01:52:00.0000] <MikeSmith>
I added notationDecl

[01:52:01.0000] <MikeSmith>
and unparsedEntityDecl methods

[01:52:02.0000] <MikeSmith>
to my Checker subclass

[01:52:03.0000] <hsivonen>
I wonder if the stream splitters forward DTDHandler correctly

[01:53:00.0000] <MikeSmith>
I'm you can give my some guidance on where to troubleshoot, I will take a look

[01:53:01.0000] <MikeSmith>
but I dropped some System.err stuff into those message

[01:53:02.0000] <MikeSmith>
*methods

[01:53:03.0000] <hsivonen>
hmm. CombineValidator handles DTDHander OK

[01:54:00.0000] <hsivonen>
MikeSmith: oops. Sorry. DTDHandler is useless for this.

[01:54:01.0000] <MikeSmith>
ah, OK

[01:55:00.0000] <MikeSmith>
I was beginning to wonder.. running out of places to check :)

[01:55:01.0000] <hsivonen>
you need LexicalHandler

[01:55:02.0000] <MikeSmith>
aha

[01:55:03.0000] <hsivonen>
sorry about giving bogus advice and wasting your time

[01:55:04.0000] <MikeSmith>
OK

[01:55:05.0000] <MikeSmith>
no problem man

[01:55:06.0000] <MikeSmith>
everything is still a learning experience for me at this point

[01:55:07.0000] <MikeSmith>
so it's honestly not wasted time

[01:55:08.0000] <hsivonen>
now, the problem is the Jing stuff doesn't pass the LexicalHandler stuff along

[01:56:00.0000] <MikeSmith>
OK

[01:56:01.0000] <hsivonen>
so you can't do this as a class that pretends to be a Jing Validator instance

[01:56:02.0000] <hsivonen>
instead, you need to hook it to the pipeline as a special case

[01:56:03.0000] <MikeSmith>
oh

[01:56:04.0000] <MikeSmith>
is there an example of a class where you've already done that?

[01:57:00.0000] <hsivonen>
yes. I'll take a look

[01:57:01.0000] <MikeSmith>
cool

[01:57:02.0000] <hsivonen>
MikeSmith: look for 

[01:57:03.0000] <hsivonen>
reader = new IdFilter(xmlParser);

[01:58:00.0000] <hsivonen>
in VerifierServletTransaction

[01:58:01.0000] <hsivonen>
I think you need to add yours as a filter on top of that

[01:58:02.0000] <hsivonen>
/me checks the filter mechanism to see if it can deal with LexicalHandler

[01:58:03.0000] <MikeSmith>
OK, I'll take a look there

[01:59:00.0000] <hsivonen>
aargh

[01:59:01.0000] <hsivonen>
it doesn't

[01:59:02.0000] <hsivonen>
all this infrastructure assumes that the app wants to do things that don't violate layering

[01:59:03.0000] <hsivonen>
let's try again

[02:00:00.0000] <hsivonen>
/me checks if the parser support LexicalHandler

[02:01:00.0000] <hsivonen>
it does

[02:01:01.0000] <hsivonen>
good

[02:01:02.0000] <MikeSmith>
oh good

[02:01:03.0000] <MikeSmith>
yeah

[02:01:04.0000] <hsivonen>
so

[02:01:05.0000] <hsivonen>
as long as the validator has only one LexicalHandler, you could keep your ContentHandler part as a Checker subclass

[02:01:06.0000] <hsivonen>
and also make it implement LexicalHandler

[02:02:00.0000] <MikeSmith>
OK

[02:02:01.0000] <MikeSmith>
/me takes a look at LexicalHandler docs

[02:03:00.0000] <hsivonen>
and then in validatorByUrl in VerifierServletTransaction

[02:03:01.0000] <hsivonen>
have a special case for your Checker

[02:03:02.0000] <hsivonen>
that also sets your Checker as the LexicalHandler of the parser in addition to returning it

[02:04:00.0000] <hsivonen>
except it has to dig out the Checker object from the Validator

[02:04:01.0000] <hsivonen>
hmm.

[02:05:00.0000] <hsivonen>
hmm. this is annoying

[02:06:00.0000] <hsivonen>
ok. here's what to do

[02:06:01.0000] <hsivonen>
call getContentHandler() on the Validator

[02:06:02.0000] <hsivonen>
then check if it's an instanceof LexicalHandler

[02:06:03.0000] <hsivonen>
and if it is, set it as the lexicalHandler of the parser

[02:06:04.0000] <MikeSmith>
/me nods

[02:07:00.0000] <hsivonen>
which will break once more than one Validator have a contenthandler that is also a lexicalhandler

[02:07:01.0000] <hsivonen>
hmm. 

[02:07:02.0000] <hsivonen>
perhaps it's safer to do the instanceof against your particular class name

[02:07:03.0000] <MikeSmith>
OK, yeah

[02:07:04.0000] <hsivonen>
to make the setup less brittle than trusting that nothing else implements LexicalHandler as a stub

[02:09:00.0000] <MikeSmith>
understood

[02:10:00.0000] <MikeSmith>
the call to getContentHandler() the Validator, I need to do that within the validate() method, right?

[02:10:01.0000] <MikeSmith>
*on the Validator

[02:11:00.0000] <hsivonen>
MikeSmith: I suggest putting the hack in the validatorByUrl method

[02:11:01.0000] <hsivonen>
MikeSmith: assuming you are going to make the checker URL-identified

[02:12:00.0000] <MikeSmith>
yep, I've already got it URL-identified

[02:12:01.0000] <MikeSmith>
so that should work fine

[02:14:00.0000] <hsivonen>
in other news, it's scary to discover how many test cases rely on about:blank loading synchronously.

[02:15:00.0000] <hsivonen>
yet, the HTML5 parser works quite well for dogfooding even if about:blank doesn't always load synchronously

[02:15:01.0000] <MikeSmith>
test cases from which suite?

[02:15:02.0000] <hsivonen>
MikeSmith: Mozilla Mochitests

[02:15:03.0000] <MikeSmith>
ah

[02:15:04.0000] <hsivonen>
(the test cases don't need to say about:blank. Having an <iframe> with src means the same thing.)

[02:16:00.0000] <hsivonen>
s/with/without/

[02:16:01.0000] <MikeSmith>
so how come the HTML5 parser still works for those cases?

[02:16:02.0000] <hsivonen>
it doesn't

[02:16:03.0000] <hsivonen>
at least not reliably

[02:16:04.0000] <MikeSmith>
oh

[02:16:05.0000] <MikeSmith>
I see now

[02:18:00.0000] <MikeSmith>
hey, btw, about the URL for this xml-stylesheet checker.. any suggestions for the name? I've got http://c.validator.nu/xmlstylepi/ now

[02:18:01.0000] <MikeSmith>
maybe http://c.validator.nu/xml-style-pi/ is better?

[02:18:02.0000] <MikeSmith>
or http://c.validator.nu/xml-stylesheet-pi/ ?

[02:18:03.0000] <MikeSmith>
dunno if you have any conventions you prefer to use for those names

[02:19:00.0000] <MikeSmith>
dashes or no dashes or whatever

[02:21:00.0000] <hsivonen>
MikeSmith: I've used /text-content/ with a hyphen

[02:21:01.0000] <hsivonen>
but otherwise I've managed to use single words, so the issue hasn't come up

[02:21:02.0000] <hsivonen>
how about /xml-stylesheet/ ?

[02:23:00.0000] <MikeSmith>
hsivonen: sounds fine to me

[02:23:01.0000] <MikeSmith>
I will make it so

[02:23:02.0000] <hsivonen>
ok

[02:23:03.0000] <hsivonen>
/me goes away for lunch

[02:24:00.0000] <MikeSmith>
hsivonen: thanks extremely much for all your help

[02:24:01.0000] <MikeSmith>
/me heads off to catch a train back to Tokyo

[02:37:00.0000] <Philip`>
hsivonen: "<--…--> escapes in <script>" (in html5-parser-improvements)

[02:38:00.0000] <Dashiva>
I don't recall seeing public-html this quiet for years

[03:08:00.0000] <hsivonen>
Philip`: thanks. fixed

[05:12:00.0000] <hsivonen>
http://livedom.validator.nu/ now has the latest parser version deployed

[05:18:00.0000] <annevk3>
Philip`, see "URL-encoded form data"

[05:21:00.0000] <Philip`>
annevk3: Did you mean foolip?

[05:31:00.0000] <annevk3>
yes

[05:38:00.0000] <foolip>
annevk3: thanks, looking now

[05:45:00.0000] <foolip>
that's just the encoding end, as expected

[05:47:00.0000] <hsivonen>
http://www.guardian.co.uk/environment/2009/dec/22/copenhagen-climate-change-mark-lynas

[05:51:00.0000] <karlcow>
hsivonen: it is at these moments that we wish there was a public transcript of the discussions

[05:51:01.0000] <hsivonen>
oops. wrong window. sorry about the off-topic URL

[05:56:00.0000] <karlcow>
hsivonen: still interesting ;)

[06:02:00.0000] <karlcow>
http://rebuildingtheweb.com/en/html5-make-web-more-invalid/

[06:07:00.0000] <Philip`>
Sounds like it's arguing that HTML5 makes the concept of validity much less important (because it removes the link between validity and interoperability / forward-compatibility)

[06:07:01.0000] <Philip`>
and so people will focus less on validity

[06:07:02.0000] <Philip`>
since everything will work fine for them anyway

[06:07:03.0000] <Philip`>
and, uh, somehow that's a bad thing (?)

[06:07:04.0000] <annevk3>
foolip, maybe you should file a bug and say that it should define server requirements as well

[06:07:05.0000] <gsnedders>
We have pretty uniform behaviour already though

[06:07:06.0000] <annevk3>
foolip, would be somewhat logical

[06:08:00.0000] <hsivonen>
Philip`: I thought HTML5 doesn't move the link between validity and forward compat

[06:08:01.0000] <hsivonen>
everything invalid is reserved for future specs

[06:08:02.0000] <Philip`>
I don't see anything in that post saying why "invalid" content is actually bad

[06:09:00.0000] <hsivonen>
(of course, if you deploy a given piece of syntax enough, it'll show up as a legacy constraint on what future specs can say)

[06:09:01.0000] <hsivonen>
(so you only get burned if you make stuff up and aren't successful in deploying it widely)

[06:09:02.0000] <Philip`>
hsivonen: That's true, but it's not what the mentioned post is arguing

[06:32:00.0000] <Dashiva>
Philip`: There's a second, underlying assumption, that valid == best practice (and probably also XHTML :P)

[06:33:00.0000] <foolip>
annevk3: I'll do that if no one in the media fragments WG can find a spec which we can use

[06:37:00.0000] <annevk>
i think html5 should define it regardless

[06:38:00.0000] <annevk>
well, as long as it's the spec that defines how to serialize

[06:44:00.0000] <hsivonen>
Today I broke my rule of not redeploying V.nu before a vacation. I hope it stays up for the rest of the year...

[06:45:00.0000] <nirvdrum>
hsivonen: Happen to see my message about localents not fixing the CCE?

[06:45:01.0000] <hsivonen>
nirvdrum: hmm. that's odd. usually it fixes the problem

[06:46:00.0000] <hsivonen>
nirvdrum: it's a problem with the local entity cache being corrupt somehow

[06:46:01.0000] <hsivonen>
nirvdrum: I suggest trying to delete the local cache and then running localent

[06:46:02.0000] <nirvdrum>
hsivonen: It's very reproducible for me.  If I execute "all" twice in a row, it works fine.

[06:47:00.0000] <nirvdrum>
I've just taken to making a tarball of my distribution and throwing it up on S3.

[06:47:01.0000] <nirvdrum>
I found the only way I could detach from the TTY, too, was to add a stop port to the jetty command.

[06:48:00.0000] <nirvdrum>
I think it's because Main has a System.in.read();

[07:01:00.0000] <hsivonen>
nirvdrum: http://pastebin.mozilla.org/692476 works for me for detaching from TTY

[07:01:01.0000] <nirvdrum>
hsivonen: You added a stop port there, too.

[07:01:02.0000] <nirvdrum>
Without it, it tries to read from stdin, it seems.

[07:02:00.0000] <nirvdrum>
And then immediately stops the server.

[07:02:01.0000] <nirvdrum>
I spent a few hours trying to figure out why I could get as far as binding to the port and then the process would just kill itself.

[07:03:00.0000] <hsivonen>
clearly, I should do a better job with documenting my deployment procedure

[07:04:00.0000] <nirvdrum>
I'd really like to take a stab at building a web app directory structure at the very least.  Going to WAR from there is simple.  It'd simplify deployment considerably.

[07:04:01.0000] <hsivonen>
depends on how one is deploying

[07:05:00.0000] <hsivonen>
having worked on Java enterprise software and having written my of main() for Jetty, I vastly prefer writing my own main() for Jetty

[07:05:01.0000] <hsivonen>
s/of/own/

[07:06:00.0000] <nirvdrum>
Well, if I want to use an existing app server, I could just plop it in and use my existing init scripts and monitoring tools.  If I were to still the embedded jetty you have, I wouldn't need to add all those classpath entries.  It'd just be an easier command to execute.

[07:06:01.0000] <nirvdrum>
I don't think you lose anything.

[07:06:02.0000] <nirvdrum>
Setting up a WebAppContext in Jetty is pretty straightforward.

[07:07:00.0000] <hsivonen>
classpath hell is bad enough with plain Java apps, I don't really want to deploy with more run-time classloader tricks

[07:07:01.0000] <hsivonen>
like an app server running itself and the .war under different classloaders or whatever

[07:07:02.0000] <hsivonen>
I am willing to consider patches for building a .war, though

[07:08:00.0000] <nirvdrum>
Well, what I was initially proposing was just a more standard approach to running the app in the embedded container *shrug*

[07:08:01.0000] <nirvdrum>
But I agree.  You don't want to be debugging people JBoss issues, for example.

[07:09:00.0000] <hsivonen>
fwiw, V.nu at least at one point supported running inside W3C Jigsaw

[07:09:01.0000] <hsivonen>
which is yet different

[07:09:02.0000] <nirvdrum>
I run the CSS validator as a WAR in Jetty.  I think that was designed to run in Jigsaw.

[07:09:03.0000] <nirvdrum>
So, maybe they've cleaned that up a bit.

[07:13:00.0000] <hsivonen>
nirvdrum: anyway, I understand that others may prefer to deploy a .war and I'm willing to accept code to enable it, but I don't see a compelling benefit to the deployment of the instances I'm running

[07:14:00.0000] <hsivonen>
btw, one reason of having my own main() is making debugging in Eclipse super-simple

[07:14:01.0000] <nirvdrum>
Fair enough.  I'm just looking that I killed about 10 hours yesterday trying to get everything going.

[07:14:02.0000] <nirvdrum>
hsivonen: I normally handle that by add a TestMain class or whatever to my test classpath and execute from there.

[07:16:00.0000] <nirvdrum>
Another issue I had run into is that mercurial is a required dependency.  I can file an issue to add that to the docs.

[07:16:01.0000] <nirvdrum>
It's been years since I've used bugzilla, though.  Hopefully I remember my way around ;-)

[07:17:00.0000] <hsivonen>
sorry about the defective docs and the unusual deployment

[07:17:01.0000] <nirvdrum>
Eh, no worries.  I'm just trying to think of ways to make it easier for the next time.

[07:18:00.0000] <hsivonen>
nirvdrum: did Jing's datatype library loading code work ok under your app server?

[07:18:01.0000] <hsivonen>
nirvdrum: previously, when I tried to sandbox the code, I gave up because what I did disrupted Jing's classloader tricks

[07:19:00.0000] <nirvdrum>
I guess so?  I got the embedded Jetty server up and could validate an HMTL 5 site, both from the server directly and from the W3C validator frontended to the nu validator.

[07:19:01.0000] <hsivonen>
ok

[07:19:02.0000] <nirvdrum>
But, I really didn't try any other doc types.

[07:36:00.0000] <MikeSmith>
Philip`: fwiw, I figured out the cause of my problem with v.nu not actually accepting connections

[07:36:01.0000] <MikeSmith>
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=560056

[07:37:00.0000] <MikeSmith>
latest debian "netbase" package sets a new kernel param

[07:37:01.0000] <MikeSmith>
net.ipv6.bindv6only=1

[07:37:02.0000] <Philip`>
Ah

[07:38:00.0000] <MikeSmith>
that totally muffs up all java networking

[07:38:01.0000] <Philip`>
Would http://[::1]:8888/ work?

[07:39:00.0000] <MikeSmith>
dunno

[07:39:01.0000] <MikeSmith>
I can try it

[07:40:00.0000] <MikeSmith>
I don't know what that syntax means

[07:41:00.0000] <MikeSmith>
wow

[07:41:01.0000] <MikeSmith>
Philip`: yeah, that works even when  net.ipv6.bindv6only=1

[07:41:02.0000] <MikeSmith>
so what does "[::1]" mean?

[07:44:00.0000] <MikeSmith>
/me steps away for a bit

[07:50:00.0000] <gratz|home_>
MikeSmith, ipv6 equivalent of 127.0.0.1

[07:52:00.0000] <Mitsurugi>
hi all

[07:54:00.0000] <Philip`>
MikeSmith: What gratz|home_ said

[07:54:01.0000] <annevk>
http://rebuildingtheweb.com/en/ is pretty much like our other fan site in a slightly different tone...

[07:54:02.0000] <Philip`>
::1 is like 127.0.0.1, [::1] is how you write it in URLs to be unambiguous

[08:06:00.0000] <annevk>
hsivonen, feel free to post such off-topic links btw

[08:06:01.0000] <annevk>
hsivonen, I'd go as far as encouraging it :)

[08:07:00.0000] <karlcow>
annevk: just realized that it was written by Vlad Alexander http://rebuildingtheweb.com/en/about/

[08:18:00.0000] <annevk>
there's a lot less email these past few days

[08:18:01.0000] <annevk>
very nice

[08:18:02.0000] <Philip`>
Just wait until a microdata decision is made

[08:20:00.0000] <annevk>
when is that?

[08:20:01.0000] <annevk>
after the holidays somewhere?

[08:20:02.0000] <Philip`>
/me shrugs

[08:21:00.0000] <foolip>
/me looks forward to another cycle of the same nonsense

[08:23:00.0000] <gratz|home_>
what decision is there to be made?

[08:24:00.0000] <Philip`>
Whether it should be split out of the W3C's HTML5 document or not

[08:25:00.0000] <gratz|home_>
what should?

[08:25:01.0000] <Philip`>
All the bits to do with microdata

[09:40:00.0000] <MikeSmith>
gratz|home_, Philip` - I see, thanks (::1 thing)

[10:11:00.0000] <annevk>
task queues are fun

[11:25:00.0000] <Lachy>
Hixie, whatwg.org seems to be down

[11:49:00.0000] <smaug>
if some attribute should be parsed as non-negative integer, what is the default value for the DOM attribute if the attribute isn't there or if the parsing fails? 

[11:52:00.0000] <smaug>
and since whatwg.org is down it is even harder to find out if the draft specifies that situation properly

[12:00:00.0000] <annevk>
http://dev.w3.org/html5/spec/infrastructure.html#non-negative-integers

[12:01:00.0000] <annevk>
oh, and http://dev.w3.org/html5/spec/infrastructure.html#reflecting-content-attributes-in-idl-attributes

[12:01:01.0000] <annevk>
smaug, ^^

[12:02:00.0000] <annevk>
maybe we should host all specs on multiple servers in case stuff goes down

[12:04:00.0000] <smaug>
annevk: ok, thanks

[12:04:01.0000] <smaug>
though, there is this: "The maxLength IDL attribute must reflect the maxlength content attribute, limited to only non-negative numbers."

[12:04:02.0000] <smaug>
which sort-of says that maxLength is always non-negative

[12:05:00.0000] <annevk>
it depends on the type of maxLength in the IDL

[12:06:00.0000] <annevk>
if it's actually a signed integer it will return -1 if it's not present

[12:06:01.0000] <annevk>
sounds like that is the case

[12:06:02.0000] <smaug>
yeah, I think that is what the draft tries to say

[12:08:00.0000] <annevk>
tries? :)

[12:09:00.0000] <annevk>
having read it it seems that's what it says

[12:24:00.0000] <AryehGregor>
smaug, "limited to only non-negative numbers" is linked to something that explains pretty clearly, I think.

[12:26:00.0000] <smaug>
AryehGregor: I couldn't find it easily

[12:26:01.0000] <smaug>
it is explained in #reflecting-content-attributes-in-idl-attributes

[12:27:00.0000] <annevk>
you can just click on it

[12:28:00.0000] <annevk>
takes you right there

[12:28:01.0000] <AryehGregor>
http://dev.w3.org/html5/spec/infrastructure.html#limited-to-only-non-negative-numbers "If, on the other hand, it fails or returns an out of range value, or if the attribute is absent, the default value must be returned instead, or −1 if there is no default value."

[12:29:00.0000] <AryehGregor>
And if you follow the link to "rules for parsing non-negative integers", it tells you that the algorithm fails if there's a - before the first digit, among other things, if you read through the algorithm.

[12:29:01.0000] <smaug>
ok, my mistake then

[12:29:02.0000] <smaug>
/me wishes the draft would be reasonable easy to read

[12:30:00.0000] <AryehGregor>
Well, hard to be reasonably easy to read while being fully precise, I guess.

[12:30:01.0000] <smaug>
but since I don't have any real suggestions (other than splitting it to many parts), I shouldn't perhaps complain so much

[12:31:00.0000] <annevk>
you think splitting it would make it easier to read?

[12:31:01.0000] <smaug>
yes!

[12:31:02.0000] <annevk>
interesting; I think it makes things harder (for where we splitted it)

[12:31:03.0000] <smaug>
splitting things like XHR from it made both drafts somewhat easier to understand

[12:32:00.0000] <smaug>
same with storage etc

[12:32:01.0000] <AryehGregor>
Probably not possible to do sanely for things like maxlength, though.

[12:33:00.0000] <smaug>
even for maxLength

[12:33:01.0000] <Dashiva>
A split that can be accomplished by variable use of display:none on a single document isn't a real split :)

[12:34:00.0000] <smaug>
there could be some "core" spec, which tells how DOM attributes and content attributes are reflected and then HTML forms spec could have links to that

[12:34:01.0000] <smaug>
though, that is quite close to what we have now

[12:34:02.0000] <annevk>
it is almost exactly what we have now...

[12:35:00.0000] <annevk>
except it's in the same file...

[12:35:01.0000] <annevk>
(the same goes for all the other splitting btw)

[12:35:02.0000] <annevk>
(XHR has been rewritten a lot, I think that's what made it clearer)

[12:50:00.0000] <Mitsurugi>
hi all, is there any lists of compatible browsers for html5 ?

[12:51:00.0000] <AryehGregor>
Some parts of HTML5 are implemented by all browsers.  Some parts aren't implemented by any yet.

[12:51:01.0000] <AryehGregor>
You'll have to ask about particular features.

[12:51:02.0000] <Mitsurugi>
aha

[12:51:03.0000] <Mitsurugi>
so, im looking for the browsers that support the video tag :)

[12:52:00.0000] <annevk>
I'd point to the wiki, but it's down

[12:52:01.0000] <annevk>
Mitsurugi, Chrome, Firefox, and Safari

[12:52:02.0000] <AryehGregor>
Mitsurugi, currently <video> is supported by Firefox 3.5 and up with Ogg Theora, recent Safari (4, maybe?) with H.264, and Chrome 3 and up with both.

[12:52:03.0000] <Mitsurugi>
thx annevk 

[12:52:04.0000] <AryehGregor>
More precisely, Safari supports whatever QuickTime does.

[12:53:00.0000] <AryehGregor>
As far as I understand.  Which includes H.264 but not Theora by default.

[12:53:01.0000] <AryehGregor>
But you can install a Theora codec if you want.

[12:53:02.0000] <Mitsurugi>
i was sure that i read that opera supports too

[12:53:03.0000] <AryehGregor>
Experimentally only, IIRC.

[12:53:04.0000] <Mitsurugi>
aha

[12:53:05.0000] <AryehGregor>
If annevk says it doesn't, then it doesn't.  :)

[12:53:06.0000] <Mitsurugi>
gracias

[12:54:00.0000] <Mitsurugi>
sorry, thx

[12:54:01.0000] <Philip`>
Opera has said it will support Theora in <video>, at some point in the future, I believe

[12:54:02.0000] <annevk>
we will for Opera 10.5

[12:54:03.0000] <Mitsurugi>
we believe that all browesers support 

[12:54:04.0000] <AryehGregor>
Yes, that's my recollection too.

[12:54:05.0000] <AryehGregor>
Ah.

[12:54:06.0000] <annevk>
but not in the pre-alpha that was released yesterday

[12:54:07.0000] <Mitsurugi>
hehe

[12:55:00.0000] <annevk>
we also had experimental releases of Opera with <video> in the past

[12:56:00.0000] <Philip`>
How much resemblance does the current spec bear to Opera's initial experimental version?

[12:57:00.0000] <annevk>
a little

[12:57:01.0000] <annevk>
I think stop() got killed and a whole lot got added

[13:01:00.0000] <Dashiva>
I bet the first major video library will do HTMLVideoElement.prototype.stop = function(){...} :)


2009-12-24
[17:14:00.0000] <AryehGregor>
How can you unset an IDL attribute like maxLength?

[22:36:00.0000] <Hixie>
anyone heard anything about the microdata thing?

[01:13:00.0000] <MikeSmith>
annevk3: is Simon on vacation already maybe?

[01:14:00.0000] <annevk3>
maybe

[01:15:00.0000] <MikeSmith>
hsivonen: I was thinking today that if the xml-stylesheet Checker subclass implements LexicalHandler then maybe I should rename it to something more general so that it can also be used later for any other lexical-checking stuff that might come up

[01:15:01.0000] <annevk3>
oh, his office is closed today

[01:15:02.0000] <MikeSmith>
ah, OK

[01:17:00.0000] <MikeSmith>
annevk3: if you care to try some xml-stylesheet checking, http://qa-dev.w3.org:8888/ has support for it now

[01:17:01.0000] <MikeSmith>
still needs to be reviewed by hsivonen so not committed yet

[01:17:02.0000] <MikeSmith>
but enough for testing at this point at least

[10:33:00.0000] <gsnedders>
Hixie: thx

[10:55:00.0000] <gsnedders>
jgraham: yt?

[10:58:00.0000] <gsnedders>
jgraham: 9314e70287 seems to have broken SimpleTree cloning with namespaces

[10:58:01.0000] <gsnedders>
jgraham: And we have loads of test failures because of it

[11:02:00.0000] <gsnedders>
jgraham: Also, we really ought to fix all the parse error failures, as I found quickly one case of the wrong tree being created causing one of them ot fail

[11:02:01.0000] <gsnedders>
*to


2009-12-25
[16:00:00.0000] <jgraham>
gsnedders: Oops. 

[16:00:01.0000] <jgraham>
I thought I had run the tests with the patch applied

[16:01:00.0000] <jgraham>
I agree that fixing the parse errors would be nice but I needed to know how much was really failing and how much was parse errors

[17:17:00.0000] <Hixie>
foolip: i'm on vacation as annevk said, but if there's something you want me to prioritise when i get back, remind me early next year

[17:17:01.0000] <Hixie>
still no news about the microdata issue?

[18:17:00.0000] <AryehGregor>
Who is "The Team" in Michael™ Smith's comment?  http://www.w3.org/2002/09/wbs/40318/issue-76-objection-poll/results

[18:36:00.0000] <Dashiva>
markp's latest tweet confuses me

[19:47:00.0000] <Hixie>
AryehGregor: the w3c members

[19:47:01.0000] <Hixie>
er

[19:47:02.0000] <Hixie>
not members

[19:47:03.0000] <Hixie>
staff

[19:47:04.0000] <Hixie>
sorry

[04:07:00.0000] <gsnedders>
jgraham: Wrong pare errors.

[04:07:01.0000] <gsnedders>
jgraham: Waiting for what was in the oven, now moving my laptop away from coffee table for later :)

[09:36:00.0000] <Dashiva>
Twitter says "retweeted by Bob and 5 others", but I can't find a non-brute-force way of figuring out who...

[10:56:00.0000] <Vito`>
Dashiva: there's a "Retweets" link on the right, where it has Mentions, Favorites and DMs.  That's just a list of retweets and the links to everyone who RT'd it

[11:12:00.0000] <Dashiva>
Vito`: That's better, although it shows everyone who retweeted and not just my followees


2009-12-26
[02:53:00.0000] <MikeSmith>
zcorpan: thanks for the latest set of xml-stylesheet test cases

[02:54:00.0000] <MikeSmith>
I think I've now got handling for all those behaving as expected

[02:54:01.0000] <MikeSmith>
when you have time, please test and let me know

[02:54:02.0000] <MikeSmith>
i'm going through the previous test cases now to make sure the latest changes didn't introduce any regressions

[02:55:00.0000] <MikeSmith>
oops

[02:55:01.0000] <MikeSmith>
get some build errors

[02:59:00.0000] <zcorpan>
MikeSmith: should the ...all/ checker enable the xml-stylesheet checker?

[02:59:01.0000] <zcorpan>
MikeSmith: iirc i had to type in the checker url manually when i tested

[03:00:00.0000] <MikeSmith>
zcorpan: I think the all/ checker enables it

[03:01:00.0000] <MikeSmith>
I'm not setting anything manually when I test

[03:01:01.0000] <MikeSmith>
I just set the preset to XHTML5

[03:01:02.0000] <zcorpan>
ok

[03:02:00.0000] <zcorpan>
i can't connect to qa-dev atm

[03:05:00.0000] <MikeSmith>
zcorpan: yep, I'm working on it .. had a build problem

[03:05:01.0000] <MikeSmith>
I should have it back in a 5 minutes or so

[03:11:00.0000] <MikeSmith>
zcorpan: OK, it's back

[03:19:00.0000] <MikeSmith>
zcorpan: I have checking for the charset pseudo-attribute in my workspace

[03:19:01.0000] <MikeSmith>
lemme know if you want me to push that up to qa-dev too

[03:30:00.0000] <zcorpan>
MikeSmith: <?xml-stylesheet href=?> gives Error: Duplicate xml-stylesheet pseudo-attribute href.

[03:31:00.0000] <MikeSmith>
hmm, OK

[03:31:01.0000] <MikeSmith>
will try to fix that now

[03:34:00.0000] <zcorpan>
MikeSmith: is it possible to have a look at the code?

[03:35:00.0000] <MikeSmith>
yeah, I can post a patch to bugzilla

[03:35:01.0000] <MikeSmith>
hang on a minute

[03:37:00.0000] <MikeSmith>
clearly I've somehow got it adding the attribute name twice for that case

[03:40:00.0000] <zcorpan>
MikeSmith: regarding multiple xslt pis, i think firefox ignores all but the *last* (but other browsers use the first)

[03:40:01.0000] <MikeSmith>
zcorpan: http://bugzilla.validator.nu/attachment.cgi?id=135

[03:40:02.0000] <zcorpan>
MikeSmith: thanks

[03:40:03.0000] <MikeSmith>
zcorpan: hmm, that's kind of bad

[03:41:00.0000] <zcorpan>
what's the windows1252 stuff?

[03:43:00.0000] <zcorpan>
<?xml-stylesheet href="" title="&#x0D;"?> should be valid

[03:43:01.0000] <zcorpan>
see XML LegalChar

[03:44:00.0000] <zcorpan>
(maybe xml 1.0 4ed LegalChar; ask hsivonen for advice)

[03:44:01.0000] <zcorpan>
if they're different

[03:45:00.0000] <zcorpan>
you might want to test "EOF" inside entity and ncr, too

[03:45:01.0000] <MikeSmith>
OK

[03:45:02.0000] <MikeSmith>
that stuff is all just from the htmlparser Tokenizer code

[03:46:00.0000] <MikeSmith>
so I probably need to tweak it further to align with XML spec

[03:47:00.0000] <MikeSmith>
man, I can see already I have some kruft around the adding of attributes that I need to clean up

[03:50:00.0000] <zcorpan>
<?xml-stylesheet href="" title="&"?> should be invalid

[03:50:01.0000] <MikeSmith>
hmm, yeah

[03:50:02.0000] <MikeSmith>
that one will be harder to fix

[03:51:00.0000] <MikeSmith>
I've got the duplicate-attribute thing fixed in my workspace

[03:52:00.0000] <MikeSmith>
just needed to remove conditionals around lines 482 and 518

[03:53:00.0000] <MikeSmith>
you want to me to push the change to qa-dev now?

[03:53:01.0000] <zcorpan>
naw it's ok

[03:53:02.0000] <zcorpan>
not sure it's good to single out checking for "<"

[03:57:00.0000] <zcorpan>
but maybe it is

[03:59:00.0000] <zcorpan>
yeah, it is, i'm just not thinking clearly

[03:59:01.0000] <MikeSmith>
OK

[04:03:00.0000] <zcorpan>
MikeSmith: type="TEXT/XML" ... should be case-insensitive

[04:03:01.0000] <MikeSmith>
OK

[04:03:02.0000] <MikeSmith>
I thought I had it that way already

[04:03:03.0000] <MikeSmith>
but I guess not

[04:04:00.0000] <zcorpan>
MikeSmith: and should maybe even recognize things like "text/xml; charset=utf-8"

[04:04:01.0000] <MikeSmith>
I think it does do that already, at least

[04:04:02.0000] <zcorpan>
but maybe that's hard and not worth the bother

[04:04:03.0000] <zcorpan>
oh

[04:04:04.0000] <MikeSmith>
I'm just using the existing HTML5 datatype library for checking that

[04:05:00.0000] <MikeSmith>
so if it doesn't work for the PI, it means it's broken elsewhere too

[04:08:00.0000] <zcorpan>
if ("application/xml".equals(attrValue)

[04:09:00.0000] <zcorpan>
this is for the multiple xslt whining

[04:09:01.0000] <MikeSmith>
yeah

[04:09:02.0000] <MikeSmith>
should I remove that?

[04:10:00.0000] <MikeSmith>
application/xml is the actual recommended mime type, isn't it?

[04:10:01.0000] <MikeSmith>
I do realize it doesn't work across browsers

[04:10:02.0000] <MikeSmith>
I will add the warning stuff after this

[04:10:03.0000] <MikeSmith>
with a link to the wiki page

[04:10:04.0000] <zcorpan>
my point is ideally it should be case-insensitive and support parameters

[04:11:00.0000] <zcorpan>
so <?xml-stylesheet href="" type="TEXT/XML"?><?xml-stylesheet href="" type="application/xml; charset=utf-8"?> should give a message about multiple xslt pi

[04:11:01.0000] <MikeSmith>
ah, OK

[04:15:00.0000] <MikeSmith>
so I could lowercase the string before comparing and then just use startsWith()

[04:15:01.0000] <MikeSmith>
I think

[04:19:00.0000] <zcorpan>
type="text/xslLOL"

[04:19:01.0000] <zcorpan>
just lowecasing is probably good enough

[04:20:00.0000] <zcorpan>
gotta go

[15:45:00.0000] <aho>
webgl behaves like flash with wmode=transparent. this is pretty slow and usually totally unnecessary. are there any plans for some kind of switch which would make it behave more like an opaque overlay?


2009-12-27
[05:15:00.0000] <gsnedders>
Heh. So Flyglobespan went bust earlier this month.

[05:15:01.0000] <gsnedders>
/me was a bit worried about whether they would go bust when he flew back from TPAC last year

[06:35:00.0000] <gsnedders>
So, Hixie's birthday has come.

[06:36:00.0000] <gsnedders>
And time for me to reveal what I've been working on for the past week.

[06:36:01.0000] <gsnedders>
http://stuff.gsnedders.com/whatwg/vocabs.html

[06:37:00.0000] <gsnedders>
Take a look at the xref there :)

[06:40:00.0000] <gsnedders>
Though there does appear to be some bugs.

[06:52:00.0000] <gsnedders>
/me fixes wha looks to be the biggest bug

[07:04:00.0000] <gsnedders>
(Tell me about bugs in xref, bugs such as lack of TOC and substitutions are known issue)

[07:48:00.0000] <gsnedders>
foolip: yt?

[08:31:00.0000] <foolip>
gsnedders: on and off

[08:32:00.0000] <gsnedders>
foolip: What do you mean by references to multipage?

[08:32:01.0000] <gsnedders>
foolip: From where? Within the spec?

[08:32:02.0000] <gsnedders>
foolip: Another document?

[08:47:00.0000] <foolip>
gsnedders: I mean references from e.g. the vocabs spec to the multipage HTML5 spec

[08:47:01.0000] <foolip>
as opening the single-page spec will kill a certain web browser

[08:48:00.0000] <gsnedders>
Web browsers should suck less. :)

[08:52:00.0000] <GPHemsley>
/me glares at Hixie for going to the extreme when demonstrating the variety of syntax HTML5 can have.

[08:55:00.0000] <AryehGregor>
<VIdeO autoBUFFER="AUTObuffer" autoplay CONTRols=''         ></VIDEO>

[08:55:01.0000] <GPHemsley>
yeah

[08:55:02.0000] <GPHemsley>
see the second example for <section>

[08:55:03.0000] <AryehGregor>
/me should have used some tabs and newlines in the whitespace, too

[08:55:04.0000] <AryehGregor>
Haha.

[08:56:00.0000] <AryehGregor>
That's painful.

[08:56:01.0000] <GPHemsley>
On another note, what is the point of <dfn>?

[08:57:00.0000] <AryehGregor>
Isn't it one of those elements that's pretty stupid but not quite stupid enough to remove given that it was in previous versions of the standard?

[08:57:01.0000] <GPHemsley>
Probably

[08:58:00.0000] <GPHemsley>
but I was hoping for a more specific answer :P

[09:03:00.0000] <gsnedders>
GPHemsley: Look at all the cases of dfn in the spec :)

[09:03:01.0000] <gsnedders>
GPHemsley: It can be used to build xrefs

[09:04:00.0000] <gsnedders>
/me tries regenning spec with the next revision of Anolis 2

[09:05:00.0000] <GPHemsley>
But it can't be used to substitute for an <abbr> @title

[09:07:00.0000] <gsnedders>
Why not?

[09:10:00.0000] <gsnedders>
Uploading the spec is slow.

[09:11:00.0000] <GPHemsley>
gsnedders: Because (based on how I'm reading the spec) you have specify the @title for each <abbr>, regardless of whether you've previously <dfn>'d it.

[09:11:01.0000] <gsnedders>
http://stuff.gsnedders.com/whatwg/vocabs.html --- it has the date (again)!

[09:12:00.0000] <gsnedders>
(Sorry, I just added a brand-new sub implementation into Anolis 2, which is a lot quicker than before)

[10:02:00.0000] <jgraham>
gsnedders: Do you have to specify all the source files at once to get x-doc x-refs?

[10:03:00.0000] <gsnedders>
jgraham: You can either do that or you can specific URLs to load dfns from

[10:03:01.0000] <jgraham>
Hixie: Happy Birthday, as I understand it

[10:03:02.0000] <gsnedders>
(But only those with @id, as it can't add them in that case)

[10:03:03.0000] <jgraham>
gsnedders: Right so if someone wants to reference HTML5 without having access to the source document, it should work

[10:03:04.0000] <gsnedders>
jgraham: Yeah.

[10:03:05.0000] <jgraham>
(for example)

[10:03:06.0000] <jgraham>
gsnedders: Nice :)

[10:04:00.0000] <gsnedders>
/me has the advantage of having the source documents

[10:04:01.0000] <gsnedders>
jgraham: It can't so well cope with docs that reference each other both ways without the source files, for obvious reasons

[10:04:02.0000] <gsnedders>
jgraham: Also, I guess you'll be pleased to know that Anolis 2 should be a lot cheaper than 1

[10:05:00.0000] <jgraham>
Boxing day sales?

[10:05:01.0000] <gsnedders>
jgraham: Also I want to discuss one or two API related things about it with you, but probably won't do that until I'm back in Sweden

[10:06:00.0000] <gsnedders>
jgraham: You know what I mean :)

[10:07:00.0000] <jgraham>
gsnedders: Sure. Probaably better to discuss then ince I am on dody hotel wifi at the mment

[10:07:01.0000] <jgraham>
(which I blame for all the mising characters)

[10:07:02.0000] <jgraham>
(because doing corrections is to painful)

[10:08:00.0000] <gsnedders>
Also I don't have to deal with silly textual mediums, but can instead just speak :P

[10:09:00.0000] <jgraham>
Indeed

[10:09:01.0000] <gsnedders>
(TOC I think is the main thing, apart from any actual UI, that remains to be done)

[10:10:00.0000] <Philip`>
jgraham: Write your lines in a local text editor then copy-and-paste into your IRC client

[10:21:00.0000] <GPHemsley>
Which validator (w3.org or validator.nu) is usually more up-to-date with the spec?

[10:39:00.0000] <jgraham>
Philip`: No

[10:47:00.0000] <quesada>
for html5lib: should I use the release version (.11) or the tip of the hg repo?

[10:47:01.0000] <quesada>
that means: is the head of the repo stable?

[11:05:00.0000] <gsnedders>
quesada: More so than 0.1

[11:05:01.0000] <gsnedders>
*0.11

[11:05:02.0000] <gsnedders>
(So use it)

[11:05:03.0000] <gsnedders>
GPHemsley: validator.nu

[11:06:00.0000] <gsnedders>
GPHemsley: It's the same impl, but v.nu is maintained by the guy who wrote the impl, so normally haschanges pushed to it more quickly

[11:09:00.0000] <gsnedders>
jgraham: does test_tests14_2_lxml fail for you?

[11:09:01.0000] <gsnedders>
jgraham: Leading to four failures?

[11:09:02.0000] <gsnedders>
(in total)

[11:15:00.0000] <GPHemsley>
gsnedders: K, thanks.

[12:44:00.0000] <annevk4>
actually, the qa-dev version of the W3C might sometimes be more up to date thanks to MikeSmith

[12:52:00.0000] <zcorpan>
hello whatwg

[12:52:01.0000] <Hixie>
gsnedders: wow, dude, that's fantastic

[12:54:00.0000] <Hixie>
gsnedders: i'd offer to marry you, but i'm already engaged

[12:55:00.0000] <annevk4>
happy b-day Hixie

[12:56:00.0000] <Hixie>
thanks

[12:56:01.0000] <zcorpan>
happy birthday

[12:56:02.0000] <annevk4>
the xref stuff looks pretty neat

[12:56:03.0000] <annevk4>
maybe I should start using that too

[12:58:00.0000] <Hixie>
i am so going to be all over that

[14:28:00.0000] <annevk4>
http://on2.com/index.php?id=472&news_id=696

[14:34:00.0000] <othermaciej>
is that standard for corporate bylaws?

[14:34:01.0000] <othermaciej>
(that merger approval requires majority of shares outstanding, not just voted)

[14:37:00.0000] <annevk4>
no idea

[14:37:01.0000] <othermaciej>
wow, On2 is practically a penny stock

[14:37:02.0000] <annevk4>
I've never seen anything like this before, but then I never pay much attention to mergers

[14:38:00.0000] <othermaciej>
still somewhat below what you'd get in the exchange if the merger is approved

[14:38:01.0000] <annevk4>
I think part of the issue is that the stock used to be worth USD 20

[14:38:02.0000] <othermaciej>
most mergers you hear about, a large proportion of the shares are held by institutional investors who always vote their proxy

[14:38:03.0000] <annevk4>
30 even, at its high in 2000

[14:39:00.0000] <othermaciej>
the highest price I see in the past 5 years is $3

[14:40:00.0000] <othermaciej>
they have been losing money for some time

[15:09:00.0000] <jgraham>
gsnedders: Haven't checked but if that is the test I think it is then yes

[15:11:00.0000] <jgraham>
(I guess the magic infoset coercion thing isn't quite working right)

[15:12:00.0000] <jgraham>
(so it doesn't see the right tag name on the stack)


2009-12-28
[16:12:00.0000] <cardona507>
jquery

[16:12:01.0000] <cardona507>
ooops - wrong room

[17:30:00.0000] <othermaciej>
do any Web Apps WG specs have an official test suite yet?

[17:33:00.0000] <othermaciej>
hmm I see ElementTraversal has one

[21:20:00.0000] <jtbandes>
Hi all... HTML5 semantics question. Should I do this: <html>...<body><header>...</header><nav>...</nav>...<section id="posts"><article><header><h1>Post title</h1></header><p>Post content</p></article></section>...</body></html>

[21:21:00.0000] <jtbandes>
Or this? ...<div id="posts"><article><header><h1>Title</h1></header><section class="post-body"><p>Content</p></section></article></div>

[21:21:01.0000] <Hixie>
both are fine, depending on what you want your document outline to look like

[21:22:00.0000] <jtbandes>
I'm just unclear on where <section> is to be used and where it should be <div>

[21:22:01.0000] <jtbandes>
Like, right now I have <section id="main"><aside>...</aside>..a bunch of <article>s

[21:23:00.0000] <jtbandes>
But I'm running into compatibility issues styling the <article>s' <p> tags, because not all browsers support CSS3 :first-of-type and :last-of-type at the moment

[21:23:01.0000] <jtbandes>
So I'm wondering if I should go with <article><header><h1>Title</h1></header><section class="post-body"><p>... instead

[21:23:02.0000] <jtbandes>
Or something entirely different.

[21:25:00.0000] <jtbandes>
The HTML5 draft spec seems a bit vague in its examples; it says "A general rule is that the section element is appropriate only if the element's contents would be listed explicitly in the document's outline." and "Examples of sections would be chapters, the various tabbed pages in a tabbed dialog box, or the numbered sections of a thesis. A Web site's home page could be split into sections for an introduction, news items, contact information."

[21:27:00.0000] <jtbandes>
Hixie: HTML5 may not be at the point for what's "right" or "wrong", but maybe just suggestions on semantics right now. I don't know

[21:34:00.0000] <jtbandes>
Any ideas?

[21:53:00.0000] <Hixie>
jtbandes: sorry, got distracted

[21:53:01.0000] <jtbandes>
np

[21:53:02.0000] <Hixie>
jtbandes: basically, use <section> for anything where you would feel ok giving a section title (<h1>) and including that section title in a table of contents

[21:54:00.0000] <Hixie>
jtbandes: and use <div> if the only reason you have the element is styling or scripting

[21:55:00.0000] <jtbandes>
Hixie: does <article>/<hgroup>/something start a new heading level like <section> does?

[21:57:00.0000] <Hixie>
<article>, <nav>, <section>, and <aside> are the same in that respect, yes

[21:59:00.0000] <jtbandes>
Would this make sense then? <section id="main"><h1>Posts</h1><article><header><h1>Title</h1></header><section class="post-body"><p>...

[22:00:00.0000] <Hixie>
no need for the post-body one

[22:00:01.0000] <Hixie>
other than that, yes

[22:01:00.0000] <jtbandes>
I feel like I should mark up the body of each <article>

[22:01:01.0000] <jtbandes>
If nothing else, to make styling the first <p> element easier

[22:07:00.0000] <Hixie>
jtbandes: use <div> for that

[22:07:01.0000] <jtbandes>
Hm

[22:07:02.0000] <Hixie>
(if tabatkins reads the archives... there's another example of why we should fix <di> in CSS/DOM rather than in markup)

[22:26:00.0000] <jtbandes>
Hixie: <di>?

[22:27:00.0000] <Hixie>
a long running discussion unrelated to what you're talking about, don't worry :-)

[22:27:01.0000] <jtbandes>
But I'm curious now :)

[22:28:00.0000] <Hixie>
it's hard to style grounps of <dt>/<dd>s

[22:28:01.0000] <Hixie>
(as well as groups of lots of other things, like the stuff after an <h1> in a <section>, as you mentioned)

[22:29:00.0000] <jtbandes>
Well, in that particular case it'd be p:first-of-type, but that's CSS3 and not widely supported

[22:29:01.0000] <Hixie>
so some people, instead of suggesting we fix CSS to make it possible to style these groups once and for all, think we should have authors who want to style those sections add <di> elements to their <dl>s to wrap the groups

[22:29:02.0000] <jtbandes>
o_O

[22:29:03.0000] <Hixie>
it's not as bad as i made it sound

[22:30:00.0000] <Hixie>
but these are the kinds of things we often talk about in #whatwg :-)

[22:30:01.0000] <Hixie>
afk, playing game

[22:30:02.0000] <jtbandes>
Is WhatWG responsible for CSS too?

[22:31:00.0000] <jtbandes>
Because I really want nested selectors... like #main { ...styles... p { line-height: 1.3em;... } } 

[04:38:00.0000] <GPHemsley>
Yeah, I'd like to be able style elements based on what children they have....

[04:39:00.0000] <GPHemsley>
(in response to jtbandes and Hixie from 6 hours ago)

[04:39:01.0000] <MikeSmith>
zcorpan: I think xml-stylesheet checking is now passing all test cases you've given so far

[04:40:00.0000] <zcorpan>
MikeSmith: cool

[04:40:01.0000] <zcorpan>
MikeSmith: did you remove the rest of the win1252 table?

[04:40:02.0000] <MikeSmith>
I'm sure there are some other bugs in there that we've not found so far

[04:41:00.0000] <MikeSmith>
zcorpan: no

[04:41:01.0000] <MikeSmith>
should I?

[04:41:02.0000] <MikeSmith>
I don't actually even know what that's for

[04:41:03.0000] <zcorpan>
MikeSmith: yes, xml charref has no fixups

[04:41:04.0000] <MikeSmith>
OK

[04:41:05.0000] <MikeSmith>
I'll remove it now

[04:43:00.0000] <zcorpan>
it's for &#x80; showing a eurosign instead of a control character in text/html

[04:43:01.0000] <MikeSmith>
OK

[04:44:00.0000] <MikeSmith>
so the whole of C1 controls range is allowed?

[04:44:01.0000] <MikeSmith>
or I should still emit an error for that

[04:44:02.0000] <MikeSmith>
?

[04:45:00.0000] <zcorpan>
you should emit an err if the expanded character is not:

[04:45:01.0000] <zcorpan>
[2]    Char    ::=    #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF] /* any Unicode character, excluding the surrogate blocks, FFFE, and FFFF. */

[04:45:02.0000] <zcorpan>
http://www.w3.org/TR/REC-xml/#NT-Char

[04:46:00.0000] <MikeSmith>
OK

[04:46:01.0000] <zcorpan>
(and possibly warn for control characters and undefined unicode characters)

[04:48:00.0000] <zcorpan>
(and "compatibility characters", whatever they are)

[04:52:00.0000] <MikeSmith>
zcorpan: OK, I ripped out the win1252 stuff and pushed an update to qa-dev

[04:52:01.0000] <MikeSmith>
I can add the other warnings later

[04:55:00.0000] <zcorpan>
great

[04:56:00.0000] <MikeSmith>
zcorpan: update patch is at http://bugzilla.validator.nu/attachment.cgi?id=136 

[04:56:01.0000] <MikeSmith>
that includes some other stuff I added recently

[04:56:02.0000] <MikeSmith>
charset checker

[04:58:00.0000] <MikeSmith>
but if you see anything remaining in the XmlStylesheetChecker code that looks wrong/suspicious, lemme know

[04:58:01.0000] <MikeSmith>
btw, I think will suggest to hsivonen that we change the name of that to LexicalChecker

[04:58:02.0000] <MikeSmith>
or something

[04:59:00.0000] <MikeSmith>
because due to the need to check if instancesd of xml-stylesheets PIs are in doctype declarations, it now has to be an implementation of SAX LexicalHandler

[05:00:00.0000] <MikeSmith>
so we could choose to turn it into a variety lexical checker that includes specific checking for xml-stylesheet PI content

[05:01:00.0000] <MikeSmith>
(though the xml-stylesheet checking is itself not lexical)

[05:03:00.0000] <MikeSmith>
but one other thing that comes to mind that it could potentially be used for is warning about CDATA secions

[05:03:01.0000] <MikeSmith>
*sections

[05:04:00.0000] <zcorpan>
why would you warn about cdata sections?

[05:04:01.0000] <MikeSmith>
for checking reverse-polygot document case -- people who have XHTML documents and want to test to make sure they can be parsed as text/html

[05:05:00.0000] <MikeSmith>
we can warn, "Document contains CDATA sections; can't be parsed as text/html."

[05:05:01.0000] <MikeSmith>
right?

[05:05:02.0000] <MikeSmith>
hmm

[05:05:03.0000] <MikeSmith>
I guess maybe the htmlparser already warns about that

[05:05:04.0000] <MikeSmith>
/me checks

[05:05:05.0000] <zcorpan>
ah, i thought polyglot checking would be code in the html parser checking for xml compat

[05:06:00.0000] <annevk>
you should prolly rename it also for XBL

[05:06:01.0000] <annevk>
though that depends on whether we'll still introduce <?xbl?> or not

[05:06:02.0000] <zcorpan>
you can have cdata sections in foreign lands

[05:07:00.0000] <MikeSmith>
well, we could check and report it only if it's not in foreign lands

[05:07:01.0000] <MikeSmith>
annevk: yeah

[05:08:00.0000] <MikeSmith>
maybe I can just rename the class VarietyChecker

[05:08:01.0000] <MikeSmith>
that makes it sound fun

[05:09:00.0000] <MikeSmith>
annevk: remind me, the idea of having a PI for CORS has been dropped?

[05:10:00.0000] <MikeSmith>
annevk, zcorpan : btw, I notice that v.nu checking for media-query values seems to fail for many examples in the current draft

[05:11:00.0000] <MikeSmith>
is it worth updating/fixing it at this point?

[05:12:00.0000] <zcorpan>
i'm not really up to speed with mq status, but i'm under the impressino that browsers are pretty aligned with the latest draft

[05:13:00.0000] <MikeSmith>
zcorpan: if that's the case, I guess I should look into updating it

[05:14:00.0000] <MikeSmith>
the fixes might be simple

[05:14:01.0000] <MikeSmith>
well, probably not simple

[05:14:02.0000] <MikeSmith>
but maybe not too difficult

[05:21:00.0000] <annevk>
MikeSmith, yes

[05:21:01.0000] <MikeSmith>
ok

[05:21:02.0000] <annevk>
MikeSmith, dunno about media queries

[05:21:03.0000] <annevk>
they'd require a full CSS parser to properly support

[05:22:00.0000] <MikeSmith>
hmm

[05:22:01.0000] <annevk>
you can use e.g. CSS escapes and all

[05:22:02.0000] <MikeSmith>
I wonder if there's a java CSS parsing library

[05:22:03.0000] <MikeSmith>
if so, we could just import and use that

[05:22:04.0000] <annevk>
well, there's the W3C CSS parser right?

[05:22:05.0000] <annevk>
isn't that in Java?

[05:23:00.0000] <annevk>
would be cool if it could be integrated so <style> style="" etc. can be checked too

[05:24:00.0000] <zcorpan>
you'd want to check external css too if you're checking css

[05:24:01.0000] <MikeSmith>
annevk: yeah, current w3c css validator is written in java

[05:24:02.0000] <annevk>
and feeds, etc.

[05:24:03.0000] <MikeSmith>
yeah

[05:24:04.0000] <annevk>
and images

[05:25:00.0000] <annevk>
should prolly have a checkbox for external resources

[05:26:00.0000] <MikeSmith>
BN

[05:26:01.0000] <annevk>
BN?

[05:26:02.0000] <MikeSmith>
if the css valididator exposes a rest interface, I guess I could maybe just use that

[05:27:00.0000] <zcorpan>
MikeSmith: <?xml-stylesheet HREF=""?> is invalid

[05:27:01.0000] <MikeSmith>
zcorpan: OK

[05:27:02.0000] <MikeSmith>
/me checks now

[05:28:00.0000] <MikeSmith>
ah crap

[05:28:01.0000] <MikeSmith>
css validator only has a soap interface, I guess

[05:30:00.0000] <MikeSmith>
zcorpan: I see now that's an side effect of the fact I'm using switch/case for checking the attribute names

[05:30:01.0000] <annevk>
switch/case is case-insensitive?

[05:31:00.0000] <zcorpan>
MikeSmith: i thought it was because of  c += 0x20; in various places

[05:31:01.0000] <MikeSmith>
annevk: in java you can only use constants in case statements

[05:32:00.0000] <MikeSmith>
and I was just following the convention of using uppercase for constants

[05:32:01.0000] <MikeSmith>
the lowercasing to compare

[05:32:02.0000] <annevk>
Java is weird

[05:33:00.0000] <MikeSmith>
zcorpan: probably that too

[05:33:01.0000] <annevk>
though Python does not have case/switch at all, which is also somewhat weird

[05:33:02.0000] <zcorpan>
MikeSmith: your tokenizer normalizes pseudo-attributes to lowercase

[05:33:03.0000] <zcorpan>
MikeSmith: the constants' names shouldn't matter, should they?

[05:35:00.0000] <MikeSmith>
zcorpan: it matters in this case because I convert the name of the attribute and uppercase it and compare that to the constant name

[05:36:00.0000] <MikeSmith>
I can just make the constant names lowercase

[05:36:01.0000] <MikeSmith>
or just change it all to if

[05:36:02.0000] <zcorpan>
oh

[05:37:00.0000] <MikeSmith>
zcorpan: yeah, I can see I need to remove all that c += 0x20 stuff

[05:40:00.0000] <MikeSmith>
anyway, I gotta drop off for now

[08:27:00.0000] <annevk>
gsnedders, did you ever do header encoding tests with XHR?

[08:27:01.0000] <annevk>
gsnedders, the spec should be more explicit on the manner, but isn't currently

[08:40:00.0000] <annevk>
implementation code of this seems to be somewhat messy as well

[08:44:00.0000] <annevk>
I cannot even find where things are going from string to byte sequence

[08:44:01.0000] <annevk>
hack

[11:24:00.0000] <gsnedders>
annevk: no, I didn't get that far (I was always more concerned with the basic syntax, never got to semantics)

[11:25:00.0000] <gsnedders>
jgraham: k, it's just you having said all tests pass somewhere (in a bug?)

[11:25:01.0000] <gsnedders>
Hixie: heh. I didn't expect so much appreciation :)

[11:25:02.0000] <gsnedders>
Hixie: And you're engaged? congrats on taht

[11:25:03.0000] <gsnedders>
*that

[11:28:00.0000] <gsnedders>
Hixie: As for how do you use it, first thing to do is to wait for it to be finished, seeming I guess you want a TOC :)

[12:43:00.0000] <annevk>
gsnedders, too bad

[13:16:00.0000] <whiteinge>
Anyone familiar with an IRC channel more specifically focused on Canvas?

[14:41:00.0000] <foolip>
Would the WHATWG or public-iri@w3c be the best place to ask about IRI bis?

[14:58:00.0000] <annevk>
public-iri

[14:58:01.0000] <annevk>
what do you wanna ask foolip?

[15:03:00.0000] <foolip>
annevk: I want to ask if query string parsing should be a part of IRI bis, another IETF spec or maybe something else

[15:05:00.0000] <foolip>
IRI bis says "Specifications that define the syntax of any of the above components MAY divide them further and define smaller parts to be components according to this document."

[15:06:00.0000] <foolip>
I'm not really sure what exactly that means though

[15:07:00.0000] <foolip>
I've reverse engineered PHP, ASP (VBScript) and JSP a bit and they seem to work mostly in compatible ways, only interesting differences are in how they treat malformed %XX entities

[15:09:00.0000] <AryehGregor>
Doesn't Perl do some totally different thing involving semicolons?

[15:09:01.0000] <foolip>
AryehGregor: aha, might that be the source of the mysterious W3C suggestion that implementations handle both & and ; as a separator?

[15:10:00.0000] <AryehGregor>
I dunno, I just know that Perl apps tend to use semicolons where I'd expect ampersands.

[15:10:01.0000] <AryehGregor>
E.g.: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=6b7b284958d47b77d06745b36bc7f36dab769d9b

[15:10:02.0000] <foolip>
I know very little about Perl, is there even a standard module used for this kind of thing?

[15:11:00.0000] <AryehGregor>
Perl seems to have standard modules for everything.

[15:11:01.0000] <AryehGregor>
I don't do Perl either, though.

[15:11:02.0000] <foolip>
meh

[15:11:03.0000] <foolip>
sounds worthwhile investigating though

[15:17:00.0000] <Philip`>
foolip: The "CGI" module is the standard#

[15:17:01.0000] <Philip`>
s/#//

[15:18:00.0000] <Philip`>
The other standard is the regexps in whatever tutorials/examples people copy-and-paste argument-splitting code from

[15:18:01.0000] <foolip>
Philip`: learning about CGI feels like going back in time :)

[15:18:02.0000] <Philip`>
The other other standard is probably whatever mod_perl does

[15:19:00.0000] <Philip`>
foolip: Perl often uses the CGI module API even if the implementation is not actually based on CGI

[15:19:01.0000] <foolip>
Philip`: ok, thanks

[15:20:00.0000] <Philip`>
CGI.pm splits on /[&;]/

[15:20:01.0000] <Philip`>
then it splits key/value on the first =

[15:20:02.0000] <Philip`>
(or puts everything in key if no value)

[15:21:00.0000] <Philip`>
(and sets value to "")

[15:21:01.0000] <Philip`>
then it unescapes key, and unescapes value

[15:21:02.0000] <foolip>
sounds like everyone else appart from the ;

[15:22:00.0000] <Philip`>
where unescape finds /%(?:([0-9a-fA-F]{2})|u([0-9a-fA-F]{4}))/ and decodes as hex

[15:22:01.0000] <annevk>
foolip, it seems you can parse it however you like

[15:22:02.0000] <annevk>
much like you can parse /foo/bar however you like on the server

[15:22:03.0000] <Philip`>
unless you're running on EBCDIC in which case it finds /%([0-9a-fA-F]{2})/ and decodes as ASCII hex

[15:23:00.0000] <AryehGregor>
:/

[15:23:01.0000] <Philip`>
Oh, and unescape replaces '+' with ' ' before that

[15:23:02.0000] <foolip>
annevk: yes, but the problem is that the media fragments WG wants define something on top of something that looks like a query string

[15:24:00.0000] <annevk>
hmm

[15:24:01.0000] <foolip>
hmm, %u422A is actually a valid percent escape?

[15:24:02.0000] <foolip>
I've never seen that format actually

[15:24:03.0000] <annevk>
and they expect it to be processed client-side?

[15:25:00.0000] <AryehGregor>
Valid for Perl, I guess.

[15:25:01.0000] <foolip>
annevk: client side for #t=10 and server-side for ?t=10

[15:25:02.0000] <annevk>
it seems to me they should stick to #syntax and define some kind of HTTP header

[15:25:03.0000] <AryehGregor>
Isn't there some standard for encoding GET form data in query strings?

[15:25:04.0000] <foolip>
there is

[15:25:05.0000] <foolip>
in HTML5

[15:25:06.0000] <foolip>
but none for decoding

[15:25:07.0000] <AryehGregor>
Interesting.

[15:26:00.0000] <foolip>
that I have found anyway

[15:26:01.0000] <foolip>
     #!/usr/bin/perl -w

[15:26:02.0000] <foolip>
     print "Content-type: text/html\n\n";

[15:26:03.0000] <foolip>
sorry :/

[15:26:04.0000] <annevk>
it seems a very fragile approach to me

[15:26:05.0000] <AryehGregor>
If this is just for media fragments, can't you just require it to behave like HTML5 requires it gets output?  Or is there some reason you want compatibility with major scripting languages?

[15:26:06.0000] <annevk>
how many implementors reviewed this?

[15:26:07.0000] <annevk>
/me can't really believe such a thing would get past basic review

[15:27:00.0000] <foolip>
the reason is that I'm meeting resistance when saying that the WG should define the processing requirements of this syntax, because presumably it belongs in another spec

[15:28:00.0000] <foolip>
annevk: I haven't seen any browser vendors apart from Opera (me) make significant noise since I joined

[15:28:01.0000] <annevk>
are they part of the group though?

[15:29:00.0000] <annevk>
I agree that some group should define processing requirements

[15:29:01.0000] <foolip>
I think some of the Apple guys are

[15:29:02.0000] <foolip>
annevk: UAs will of course just ignore the query string, but we need to know what to do for fragment URIs

[15:29:03.0000] <annevk>
if HTML5 keeps the definition of encoding it should also define how to decode imo

[15:29:04.0000] <foolip>
which part in particular do you think is fragile?

[15:30:00.0000] <Hixie>
/me looks in

[15:30:01.0000] <Hixie>
wassup

[15:30:02.0000] <annevk>
that one part is for the client and one part for the server

[15:30:03.0000] <annevk>
and that the server part might use names already in use

[15:30:04.0000] <foolip>
well HTML5 defines how to encode form data, what we are decoding isn't really form data, but I wouldn't mind if it's in HTML5

[15:31:00.0000] <foolip>
annevk: it's not the intention that both be used at once, and in either case the server part is opaque to the client and vice versa

[15:32:00.0000] <foolip>
apparently the thinking is that in some cases doing the chopping on the server is faster, e.g. archive.org already implements this

[15:32:01.0000] <foolip>
annevk: what are "names already in use"?

[15:32:02.0000] <annevk>
what if I have a resource that's named /x?t=10

[15:32:03.0000] <annevk>
(not sure if t is the right param name)

[15:32:04.0000] <foolip>
that's right

[15:33:00.0000] <annevk>
Hixie, discussing media fragments

[15:33:01.0000] <foolip>
annevk: well, I doubt anyone ever considered that

[15:33:02.0000] <annevk>
Hixie, oh, and whether HTML5 should define how to decode <form> GET URLs besides defining how to encode them

[15:34:00.0000] <foolip>
in my feedback I have pretended the query string part doesn't exist, because it is irrelevant to UAs

[15:34:01.0000] <annevk>
foolip, also, apart from maybe the fragment URIs are supposed to be opaque

[15:35:00.0000] <annevk>
the server assigns meaning to it for sure, but with this the client and server have some shared understanding of a part of the URI

[15:35:01.0000] <foolip>
annevk: I mean that the server doesn't get to see the fragment part

[15:35:02.0000] <annevk>
it feels somewhat wrong

[15:35:03.0000] <foolip>
there is no shared part

[15:35:04.0000] <annevk>
foolip, sure, the query part

[15:35:05.0000] <foolip>
the UA doesn't (shouldn't) care about that

[15:36:00.0000] <annevk>
the UA generates it and the server is expected to respond in a particular way

[15:36:01.0000] <annevk>
if the UA doesn't, how does the server ever get it?

[15:36:02.0000] <foolip>
no, the UA doesn't generate a query string

[15:36:03.0000] <foolip>
the UA can either be clever and seek to the right place with simple heuristics and range requests, or use the HTTP headers that the WG is defining

[15:37:00.0000] <annevk>
so why do we have the query string?

[15:37:01.0000] <annevk>
and who will be using it?

[15:37:02.0000] <foolip>
e.g. archive.org uses it, or something that looks like it

[15:38:00.0000] <foolip>
however, I would be happy if it wasn't in the spec at all

[15:38:01.0000] <annevk>
I'm not sure why it would need to be defined

[15:38:02.0000] <foolip>
presumably because it is convenient to use the same syntax for both, or something like that

[15:39:00.0000] <foolip>
and because server-side implementations actually exist

[15:39:01.0000] <Hixie>
why would we need to define how to decode the query component? What's the interoperability concern?

[15:39:02.0000] <AryehGregor>
But you don't need standards if it's done purely on the server side . . .

[15:39:03.0000] <Hixie>
Surely each URL can only be read by one server

[15:40:00.0000] <Hixie>
so there's nothing to interoperate with

[15:40:01.0000] <annevk>
I guess the idea is to standardize a common API for media resources

[15:40:02.0000] <foolip>
Hixie: I'm not really interested in decoding a query string, I'm interested in decoding a fragment component which uses the same syntax as a query string

[15:40:03.0000] <annevk>
but I'm not sure why that needs to be standardized...

[15:41:00.0000] <AryehGregor>
The fragment component decoding can be done on UAs, so that needs to be standardized.

[15:41:01.0000] <AryehGregor>
In fact, it has to be done on UAs.

[15:41:02.0000] <foolip>
AryehGregor: exactly

[15:41:03.0000] <Hixie>
well surely the fragment component decoding would be defined by the spec that defines the fragment?

[15:41:04.0000] <foolip>
Hixie: that's what I've been trying to argue for

[15:41:05.0000] <Hixie>
sounds good to me

[15:41:06.0000] <annevk>
Hixie, well, the idea would be to define how servers can get the data out of the URL

[15:42:00.0000] <foolip>
however, I've been a bit thrown off by the query string issue it seems

[15:42:01.0000] <annevk>
Hixie, if you define how that works presumably you can swap the software that gets data out of the URL

[15:42:02.0000] <AryehGregor>
Just define some subset of generally-used query string encoding.

[15:42:03.0000] <Hixie>
annevk: presumably that would be defined by the spec for that software

[15:42:04.0000] <Hixie>
annevk: different servers have different needs for how to parse query components

[15:43:00.0000] <annevk>
Hixie, it seems to me it's sort of an integral part of <form> submission

[15:43:01.0000] <Hixie>
e.g. http://example.com/?foo+bar vs http://example.com/?a=b&c=d vs http://example.com/a=b;c=d

[15:43:02.0000] <Hixie>
these are all valid query components

[15:43:03.0000] <Hixie>
but they need different parsing rules

[15:43:04.0000] <foolip>
it seems that one should do either (1) define how to parse a query string (probably in another spec) and reuse that definition for the fragment part in MF or (2) remove any mention of query strings from MF and just define how to parse the fragment component

[15:43:05.0000] <AryehGregor>
Hixie, I think you're missing a ? in the last.

[15:43:06.0000] <Hixie>
er yes

[15:44:00.0000] <AryehGregor>
foolip, (2) sounds more sensible.  Just make up something simple.

[15:44:01.0000] <AryehGregor>
We don't define how to parse query strings for anything else, no reason to start for media fragments.

[15:45:00.0000] <foolip>
ok, I don't expect arguing this line is going to be easy though

[15:45:01.0000] <foolip>
I agree it makes the most sense however

[15:45:02.0000] <foolip>
will try writing an email suggesting as much tomorrow when my brain has slept

[15:48:00.0000] <foolip>
thanks all


2009-12-29
[20:39:00.0000] <Zerxer>
now that I'm in the right area.. messing around with the <audio> tag and I can't get ogg files to play nicely in firefox. keeps skipping. perhaps I am just using a bad encoder? suggestions?

[21:03:00.0000] <Hixie>
hard to tell without a sample page :-)

[21:04:00.0000] <Hixie>
(#developers on irc.mozilla.org may be able to help better with mozilla-specific issues, btw)

[01:13:00.0000] <MikeSmith>
annevk: about media queries, is it a conformance error for a media query to use a media feature with a media type that it does not apply to?

[01:14:00.0000] <MikeSmith>
the spec does not seem to explicitly say it is a conformance error

[01:15:00.0000] <MikeSmith>
e.g., @media print and (scan: progressive)

[02:03:00.0000] <annevk>
it's not

[02:03:01.0000] <annevk>
you could warn if you want

[02:06:00.0000] <annevk>
http://bitworking.org/news/2009/12/xrd

[02:07:00.0000] <MikeSmith>
annevk: OK, I'll ask hsivonen if he thinks it's worth warning about

[02:07:01.0000] <annevk>
you could look up an old thread on media queries on www-style

[02:07:02.0000] <annevk>
hsivonen asked about this before

[02:11:00.0000] <MikeSmith>
annevk: OK

[02:14:00.0000] <MikeSmith>
annevk: found the thread

[02:15:00.0000] <annevk>
cool

[04:29:00.0000] <Dashiva>
annevk: And now we have two ladders to paint, twice the fun

[04:30:00.0000] <annevk>
it made me think of Microdata / RDFa mostly

[04:30:01.0000] <annevk>
though maybe there the problem is somewhat more scoped

[04:32:00.0000] <Dashiva>
Microdata and RDFa are which material you build the ladder from. Still out there, but somewhat more relevant than the color.

[06:34:00.0000] <annevk>
anyone knows why Ubuntu 9.10 is no longer offered as installation option under updates?

[06:34:01.0000] <annevk>
pretty sure it was before

[06:34:02.0000] <annevk>
but I didn't do it because I thought there might be trouble

[06:36:00.0000] <annevk>
http://guides.macrumors.com/MacBook_Pro_Buyer's_Guide is most excellent

[06:37:00.0000] <Dashiva>
Would we be having this autobuffer thread at all without safari's bug?

[06:38:00.0000] <annevk>
lots of talk, little practical impact

[06:40:00.0000] <wavded>
Does anyone have any experience with the Firefox 3.6 File API implementation?

[06:41:00.0000] <wavded>
I'm trying to do a XHR POST with file read in with FileReader, but I want to add more params to the post.  xhr.sendAsBinary() doesn't seem to allow me to do that...

[06:41:01.0000] <wavded>
Any help would be appreciated

[06:42:00.0000] <annevk>
you prolly need to wait until they do an impl of some XHR2 extensions

[06:42:01.0000] <annevk>
i.e. the FormData object

[06:43:00.0000] <wavded>
The docs said you could use xhr.send() with a FileReader but my attempts don't seem to work..

[06:44:00.0000] <wavded>
Hmm... there docs for XHR2 somewhere?

[06:45:00.0000] <annevk>
http://dev.w3.org/2006/webapi/XMLHttpRequest-2/ has the latest

[06:45:01.0000] <wavded>
thx

[08:44:00.0000] <TabAtkins>
Dashiva: Nope, the entire problem is spawned from Safari's bug causing it to autobuffer everything.  If it was already acting like Firefox, then no one would have complained, and we wouldn't have people trying to draw ridiculous distinctions like autobuffer=auto being somehow different than autobuffer not being present at all (let alone autobuffer=no actually having a usefully different...

[08:44:01.0000] <TabAtkins>
...definition).

[11:32:00.0000] <annevk>
hmm, I only get it if I run update-manager -d

[11:32:01.0000] <annevk>
but then sites also advice not to install it...

[11:32:02.0000] <annevk>
/me is confused

[11:32:03.0000] <annevk>
(still about Ubuntu 9.100

[11:32:04.0000] <annevk>
)

[11:32:05.0000] <gratz|home>
get what?

[11:32:06.0000] <gratz|home>
oh right

[11:33:00.0000] <gratz|home>
what advises not to install it and why?

[11:35:00.0000] <annevk>
there's a series here: http://www.insidesocal.com/click/2009/11/pulling-the-trigger-on-ubuntu.html

[11:35:01.0000] <annevk>
since I think I have a similar card for video I'll wait

[11:35:02.0000] <annevk>
especially with my MacBook being in repairs and all :)

[11:37:00.0000] <gratz|home>
heh

[11:43:00.0000] <wavded>
annevk - seems as though you are right, I haven't been able to find anyway in FF 3.6 to push file data via XHR along with other POST data.  :-(

[12:03:00.0000] <AryehGregor>
annevk, you're on 9.04, not 8.04 or anything, right?  Maybe apt-get update will fix it, if it thinks it's a development release for some crazy reason.

[12:10:00.0000] <annevk>
AryehGregor, yeah, I'm on 9.04

[12:11:00.0000] <annevk>
and I found a way to make the upgrade option appear, but I decided to wait

[13:27:00.0000] <foolip>
hmm, which spec defines encodeURIComponent?

[13:36:00.0000] <Philip`>
foolip: ECMAScript

[13:50:00.0000] <foolip>
Philip`: found it, thanks

[14:00:00.0000] <zcorpan>
MikeSmith: yo

[14:01:00.0000] <MikeSmith>
zcorpan: hej

[14:01:01.0000] <zcorpan>
MikeSmith: maybe checking for valuelessness and lessthanness should be in the tokenizer instead of the switch block

[14:02:00.0000] <MikeSmith>
/me takes a look

[14:02:01.0000] <zcorpan>
MikeSmith: also, i think case INVALID: is redundant with default:

[14:03:00.0000] <MikeSmith>
about checking for valuelessness and lessthanness, I think I made a conscious choice to do it the current way but can't recall why

[14:03:01.0000] <MikeSmith>
but I will look again

[14:04:00.0000] <MikeSmith>
about INVALID, I guess that's true -- I'll remove it

[14:06:00.0000] <MikeSmith>
ah

[14:06:01.0000] <MikeSmith>
zcorpan: I remember

[14:07:00.0000] <MikeSmith>
the reason I don't do the valuelessness check in the tokenizer is that I was thinking, we don't want to report valuelessness for invalid attributes

[14:07:01.0000] <MikeSmith>
I would think we'd just want to report it only for the case of attributes that we already know are valid

[14:07:02.0000] <zcorpan>
why not?

[14:08:00.0000] <MikeSmith>
because it will cause two error messages to be generated for each invalid attribute, if it lacks a value

[14:08:01.0000] <zcorpan>
yeah

[14:09:00.0000] <MikeSmith>
I guess my rationale was, if the attribute is invalid, it doesn't matter to the user whether it lacks a value

[14:09:01.0000] <zcorpan>
fair enough

[14:10:00.0000] <MikeSmith>
about lessthanness, where is the code checking that now?

[14:10:01.0000] <MikeSmith>
oh

[14:10:02.0000] <MikeSmith>
you mean the literal "<" character

[14:11:00.0000] <zcorpan>
yes :)

[14:11:01.0000] <MikeSmith>
I guess I could move that check

[14:12:00.0000] <MikeSmith>
hmm, but again, that's the same kind of case

[14:13:00.0000] <MikeSmith>
if I move it into the tokenizer, it will mean that you'd get two error messages for an invalid attribute if it has a value with "<"

[14:13:01.0000] <zcorpan>
a < will cause browsers to ignore the whole pi

[14:13:02.0000] <zcorpan>
an unknown pseudoatt won't

[14:13:03.0000] <MikeSmith>
ah, OK

[14:13:04.0000] <zcorpan>
same with valueless

[14:14:00.0000] <MikeSmith>
I see

[14:15:00.0000] <MikeSmith>
that trumps the two-message issue

[14:15:01.0000] <MikeSmith>
I will move those now and push it up to qa-dev when I'm done

[14:16:00.0000] <MikeSmith>
btw, I also made some fixes in media-query handling

[14:16:01.0000] <zcorpan>
did you fix case-sensitiveness?

[14:17:00.0000] <MikeSmith>
for xml-stylesheet pseudo-attribute names? yeah

[14:19:00.0000] <MikeSmith>
and also handling of XSLT indicators that have parameters

[14:20:00.0000] <zcorpan>
yep

[14:20:01.0000] <MikeSmith>
and added a warning if type/subtype is something other than text/xsl or text/css

[14:20:02.0000] <zcorpan>
that's in the patch i'm looking at

[14:22:00.0000] <zcorpan>
MikeSmith: i think there are some things that could be removed in the tokenizer, given that some characters are not allowed in xml

[14:22:01.0000] <MikeSmith>
OK

[14:22:02.0000] <zcorpan>
like case '\u000C': and case '\u0000':

[14:22:03.0000] <zcorpan>
but maybe it's not worth it to fiddle with something that works :)

[14:23:00.0000] <MikeSmith>
nah, I think it's worth getting it right

[14:23:01.0000] <MikeSmith>
so I can remove those

[14:26:00.0000] <zcorpan>
what does silentLineFeed() do?

[14:27:00.0000] <zcorpan>
'\r' also can't appear in pi data

[14:29:00.0000] <MikeSmith>
I am not sure what the purpose of silentLineFeed() is.. I just inherited it from the original copy/paste

[14:29:01.0000] <MikeSmith>
but I think basically it's for handling the CRLF case for HTML parsing as the spec requires

[14:29:02.0000] <MikeSmith>
to drop the CR

[14:30:00.0000] <zcorpan>
likely

[14:31:00.0000] <MikeSmith>
so what should I do for the "\r" case? emit a specific error?

[14:31:01.0000] <zcorpan>
remove, since it can't happen

[14:31:02.0000] <zcorpan>
\r is normalized to \n

[14:31:03.0000] <zcorpan>
by the xml parser

[14:50:00.0000] <zcorpan>
+                  // TODO warn about apos (IE) and TRADE (Opera)

[14:51:00.0000] <zcorpan>
hsivonen: &TRADE; works in opera now

[14:51:01.0000] <MikeSmith>
OK, I can remove that comment

[14:52:00.0000] <MikeSmith>
zcorpan: so U+000C is explicitly disallowed in XML? or does it get normalized?

[14:52:01.0000] <zcorpan>
it's disallowed

[14:53:00.0000] <MikeSmith>
OK

[14:58:00.0000] <zcorpan>
MikeSmith: maybe you can remove the stuff in the tokenizer that checks for semicolonless entities

[14:58:01.0000] <MikeSmith>
ah yeah

[14:58:02.0000] <MikeSmith>
hmm

[14:58:03.0000] <MikeSmith>
I guess

[14:58:04.0000] <MikeSmith>
so

[14:59:00.0000] <MikeSmith>
the XML parser will catch those?

[14:59:01.0000] <MikeSmith>
oh, I guess not

[14:59:02.0000] <zcorpan>
the table doesn't contain semicolonless entities

[15:00:00.0000] <zcorpan>
so it's dead code i think

[15:04:00.0000] <MikeSmith>
which table?

[15:05:00.0000] <zcorpan>
the entity table

[15:06:00.0000] <MikeSmith>
OK

[15:06:01.0000] <MikeSmith>
but I should keep the checking for semicolonless character references, right?

[15:07:00.0000] <zcorpan>
yes

[15:10:00.0000] <zcorpan>
did you implement the CharRef Legal Character thing?

[15:14:00.0000] <MikeSmith>
zcorpan: you mean an explicit check for "Char    ::=    #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] | [#x10000-#x10FFFF]" ?

[15:14:01.0000] <zcorpan>
yes

[15:15:00.0000] <MikeSmith>
not yet

[15:15:01.0000] <MikeSmith>
but I can add it now

[15:15:02.0000] <zcorpan>
ok

[15:17:00.0000] <MikeSmith>
I guess I can add it in the checkChar method

[15:17:01.0000] <zcorpan>
not handleNcrValue ?

[15:18:00.0000] <MikeSmith>
ah, for character refs

[15:19:00.0000] <zcorpan>
yeah... the xml parser has already checked Char for the pi data

[15:20:00.0000] <MikeSmith>
OK

[15:33:00.0000] <zcorpan>
time to sleep

[15:33:01.0000] <zcorpan>
g'nite


2009-12-30
[16:55:00.0000] <cardona507>
hello - I want to use offline cache and db - but am unclear how the two relate in terms of code - local storage also - I am kind of unclear about use cases - right now I am trying to help some friends put their books onto the web and we are making them mobile friendly - but I just don't know how to make my cache manifest clearly defined in the db... any advice?

[17:14:00.0000] <cardona507>
for example with a book - the text and images are all in the markup - so do I need a db at all? or just an cache manifest?

[00:56:00.0000] <zcorpan>
MikeSmith: maybe you should skip validation for xml-stylesheet pis that are not in the prolog

[00:58:00.0000] <zcorpan>
MikeSmith: since they're not considered to be potential xml-stylesheet pis

[00:59:00.0000] <MikeSmith>
zcorpan: maybe, but a possible rationale for checking them is that they're just misplaced and the user would still like their content to be checked

[00:59:01.0000] <MikeSmith>
but I don't feel strongly about it

[01:00:00.0000] <zcorpan>
yeah maybe

[01:00:01.0000] <MikeSmith>
if you think it seems preferable to not check them, I can change it

[01:03:00.0000] <MikeSmith>
hmm

[01:03:01.0000] <MikeSmith>
but I see your point

[01:04:00.0000] <MikeSmith>
if the user sees error messages about the content of the data part, that might cause them to assume it will actually be processed, even though it's misplaced

[01:05:00.0000] <zcorpan>
yes

[01:05:01.0000] <MikeSmith>
OK

[01:05:02.0000] <MikeSmith>
it's trivial to change -- just need to add an early return

[01:06:00.0000] <MikeSmith>
should I do the same for the case where it's in the doctype declaration?

[01:06:01.0000] <MikeSmith>
ah no

[01:06:02.0000] <MikeSmith>
spec says only to warn about that

[01:06:03.0000] <MikeSmith>
or doesn't say to warn

[01:06:04.0000] <MikeSmith>
but that case is just a "should", right?

[01:07:00.0000] <zcorpan>
you're allowed to skip pis in the doctype

[01:07:01.0000] <zcorpan>
per spec

[01:07:02.0000] <zcorpan>
but then the warning should probably say that you've skipped it

[01:08:00.0000] <zcorpan>
"For potential xml-stylesheet processing instructions that are in the [children] property of a document type declaration information item, xml-stylesheet processors may report to the application the parsing result of invoking the rules for parsing pseudo-attributes from a string, using the processing instruction information item's [content] property as the string. [Definition: If it's not reported to the application, the processing inst

[01:08:01.0000] <zcorpan>
information item is said to be ignored.]"

[01:10:00.0000] <zcorpan>
and "Documents should not use xml-stylesheet processing instructions that are in the [children] property of a document type declaration information item."

[01:11:00.0000] <zcorpan>
technically, <!DOCTYPE x[<?xml-stylesheet lol?>]><x/> is invalid, but you need to process it to tell

[01:15:00.0000] <MikeSmith>
maybe for both the cases, I should append "To check the contents of the xml-stylesheet instruction, move it and then resubmit the document for validation."

[01:16:00.0000] <zcorpan>
maybe (Suppressing further errors from this processing instruction.)

[01:19:00.0000] <MikeSmith>
OK

[01:27:00.0000] <MikeSmith>
zcorpan: OK, I just made those changes and also added warnings for charset and for media, etc., in XSLT indicators 

[01:28:00.0000] <MikeSmith>
and pushed to qa-dev

[01:28:01.0000] <MikeSmith>
and also made other changes we talked about yesterday

[01:28:02.0000] <MikeSmith>
see comments at bug 14

[01:28:03.0000] <MikeSmith>
oh, that reminds me

[01:29:00.0000] <MikeSmith>
wanted to ask about a couple remaining errors that aren't XML conformant

[01:32:00.0000] <zcorpan>
are they errors according to charmod?

[01:33:00.0000] <MikeSmith>
zcorpan: 

[01:33:01.0000] <MikeSmith>
if (value >= 0xFDD0 && value <= 0xFDEF) {

[01:33:02.0000] <MikeSmith>
    errNcrUnassigned();

[01:33:03.0000] <MikeSmith>
} else if (value >= 0x007F && value <= 0x009F) {

[01:33:04.0000] <MikeSmith>
    errNcrControlChar();

[01:33:05.0000] <MikeSmith>
those ranges are legal, right?

[01:33:06.0000] <MikeSmith>
if so, I can change those errors to warnings

[01:35:00.0000] <zcorpan>
legal according to the xml-stylesheet spec yes, but are they legal according to the charmod spec?

[01:37:00.0000] <MikeSmith>
ah

[01:37:01.0000] <MikeSmith>
/me goes to check charmod

[01:44:00.0000] <zcorpan>
can't find anything in charmod

[01:48:00.0000] <MikeSmith>
it mentions something about not using code points that are for Unicode internal use

[01:48:01.0000] <MikeSmith>
I wonder if permanently unassigned code points are considered to be for internal use

[01:55:00.0000] <MikeSmith>
"value >= 0xFDD0 && value <= 0xFDEF" seem to definitely be non-characters, so it seems right to emit an error for those

[01:55:01.0000] <MikeSmith>
but can't find any must-not prohibition on "value >= 0x007F && value <= 0x009F", so will change that to a warning

[04:05:00.0000] <annevk>
/me has trouble remembering all the places he's been

[05:46:00.0000] <oskude>
heh, the most best fitting topic i've seen "Please leave your sense of logic at the door, thanks!" when it comes to web-development ;P

[05:53:00.0000] <annevk>
yeah, browsers...

[08:24:00.0000] <annevk>
hmm, EUR 530 to get my MacBook fixed

[08:24:01.0000] <annevk>
doesn't really seem worth it

[08:25:00.0000] <Philip`>
That's more than the cost of my current main laptop

[08:25:01.0000] <annevk>
I paid a little more for it, but still...

[08:27:00.0000] <annevk>
pretty bad that it breaks down in two years :/

[13:53:00.0000] <AryehGregor>
I hope the HTMLWG publishes a Last Call soon for the sole reason that then people will stop trying to move the Wikipedia article to "HTML 5" based on looking at the August Working Draft.

[13:57:00.0000] <Philip`>
/me suggests not fighting the people who want to move it

[13:59:00.0000] <AryehGregor>
Well, then someone else will fight them and move it back.

[13:59:01.0000] <AryehGregor>
Anyway, it's not much of a fight, took two minutes or something.

[14:07:00.0000] <GPHemsley>
AryehGregor: Your WerdaBot doesn't appear to be working anymore... :P

[14:07:01.0000] <annevk>
publishing an interim draft is prolly coming first, if anything

[14:07:02.0000] <AryehGregor>
GPHemsley, WerdnaBot is Werdna's, not mine.

[14:07:03.0000] <GPHemsley>
AryehGregor: Well, I meant, for your talk page

[14:07:04.0000] <AryehGregor>
GPHemsley, nobody talks to me anyway, so no big deal.

[14:07:05.0000] <GPHemsley>
heh

[14:17:00.0000] <Hixie>
AryehGregor: ain't nothing wrong with calling it "HTML 5", just like "HTML4" and "HTML 4" are both acceptable

[14:17:01.0000] <AryehGregor>
Hixie, well, Wikipedia might as well be consistent.

[14:17:02.0000] <Hixie>
what does it do for HTML4?

[14:20:00.0000] <annevk>
oh yes, made Ubuntu stream video to my Xbox 360

[14:20:01.0000] <annevk>
win win win

[14:25:00.0000] <AryehGregor>
"HTML 4.0"

[14:25:01.0000] <AryehGregor>
That's the name used in the spec.

[14:26:00.0000] <AryehGregor>
Whereas HTML5 uses "HTML5" in the spec.


2009-12-31
[16:25:00.0000] <websafestudio>
Greetings! Why have I been asked to leave my sense of logic at the door? *curious*

[16:27:00.0000] <websafestudio>
I ask, because I thought that logic was the sine qua non of computer work of any kind.

[16:27:01.0000] <gsnedders>
HTML is far from logical.

[16:27:02.0000] <websafestudio>
lol

[16:27:03.0000] <websafestudio>
*mulls this over*

[16:27:04.0000] <gsnedders>
(See: <image> becomes "img" in the DOM)

[16:28:00.0000] <websafestudio>
But that was a saving of two letters!

[16:28:01.0000] <websafestudio>
<p> is a lot shorter than <paragraph> would be.

[16:29:00.0000] <gsnedders>
That's not what I mean, put <image> in an HTML file, you get "img" in the DOM.

[16:29:01.0000] <gsnedders>
The tag name magically changes.

[16:29:02.0000] <gsnedders>
Oh, and then there's <isindex>

[16:29:03.0000] <gsnedders>
That's also nice and logical.

[16:30:00.0000] <websafestudio>
*scratches head to remember* Isn't <isindex> for makign image maps?

[16:30:01.0000] <AryehGregor>
websafestudio, HTML isn't logical because it accumulated over years by dint of vendors randomly adding stuff they thought would be neat.  HTML5 tries to reflect that reality by specifying all the craziness clearly.

[16:30:02.0000] <AryehGregor>
As opposed to initiatives like XHTML2 that have tried to do things in a more "logical" but less pragmatic way.

[16:30:03.0000] <AryehGregor>
At least that's how I interpret it.

[16:31:00.0000] <websafestudio>
Pragmatic = moneymaking? in this instance

[16:31:01.0000] <websafestudio>
*thinks she understands*

[16:31:02.0000] <daedb>
pragmatic = stuff that actually works on the web

[16:32:00.0000] <websafestudio>
But ... more logical, but less pragmatic? *confused*

[16:32:01.0000] <websafestudio>
Isn't using a logical approach a practical thing?

[16:32:02.0000] <daedb>
XHTML2 had some neat logical features, but it wasn't backwards compatible so actually using it would be highly impractical.

[16:33:00.0000] <websafestudio>
I see ... In a perfect Web world, what browser would you like to see used?

[16:34:00.0000] <annevk>
in a perfect world there would be no computers, doh

[16:34:01.0000] <websafestudio>
Hmmm ...

[16:34:02.0000] <websafestudio>
Would there be mental telepathy?

[16:34:03.0000] <annevk>
/me should get some sleep

[16:34:04.0000] <annevk>
nn

[16:35:00.0000] <daedb>
In a perfect world the web would just stream right into our brains.

[16:35:01.0000] <websafestudio>
Sleep is a form of mental telepathy -- perhaps -- or rather, dreaming may be

[16:35:02.0000] <websafestudio>
Right, daedb -- Will that happen in our lifetimes?

[16:36:00.0000] <daedb>
I dunno, depends on how long we live :)

[16:36:01.0000] <websafestudio>
On a pragmatic note: I have just finished reading "Designing with Web Standards," Edition 3 (2009) and am slowly absorbing its message

[16:37:00.0000] <websafestudio>
I am earnestly trying to become a better Web designer -- is it hopeless? Is the game worth the candle? (I ask myself this after wrestling with minutiae like margins of lists)

[16:39:00.0000] <websafestudio>
All right, thanks for chatting, must go eat supper.

[03:48:00.0000] <smaug>
Grr, session history. 

[10:23:00.0000] <gsnedders>
/me pushes Anolis 1.1

[10:25:00.0000] <gsnedders>
(There basically nothing interesting new in it)

[10:26:00.0000] <gsnedders>
(It supports html5lib tip, it's up-to-date with the spec for the outliner impl, and data-anolis-xref exists)

[10:26:01.0000] <gsnedders>
Wait, didn't I agree to change that to data-xref?

[10:26:02.0000] <gsnedders>
Gah.

[15:42:00.0000] <nessy>
foolip: congrats on the HTML5 video release of Opera!!

[15:44:00.0000] <foolip>
nessy: thank you :)

[15:45:00.0000] <foolip>
you can see why I haven't replied to the media fragments thread in a few days now

[15:45:01.0000] <nessy>
say, if I install gstreamer on the mac, will it work?

[15:45:02.0000] <nessy>
sure, you do what you have to do

[15:45:03.0000] <foolip>
no, the feature is disabled, so it won't even try looking for GStreamer

[15:45:04.0000] <nessy>
I feel really bad for not having had any time to work on a11y stuff since TPAC - only media fragments and other contracts

[15:46:00.0000] <foolip>
also, I don't believe there is a standard GStreamer distribution for Mac, is there?

[15:46:01.0000] <nessy>
there are always priorities :-)

[15:46:02.0000] <nessy>
I have just port installed gstreamer on the mac, so yes, there is

[15:47:00.0000] <nessy>
port is a standard distro for the mac

[15:47:01.0000] <nessy>
fink is the other one

[15:48:00.0000] <nessy>
and even fink has a gstreamer package http://pdb.finkproject.org/pdb/package.php/gstreamer-0.10

[15:49:00.0000] <nessy>
this is the port package: http://trac.macports.org/browser/trunk/dports/gnome/gstreamer/Portfile

[15:49:01.0000] <nessy>
you might want to consider falling back to these on the mac when installed

[15:51:00.0000] <foolip>
right, I installed those once

[15:52:00.0000] <foolip>
however, we can't ask users to install that, it has a ridiculous amount of dependencies

[15:52:01.0000] <foolip>
with the lack of a package manager, it can't be automated even

[15:52:02.0000] <nessy>
no, you cannot expect it

[15:53:00.0000] <nessy>
but just like on Linux, where you also cannot expect it to be installed, you can still fall back to it if available

[15:53:01.0000] <foolip>
but on linux it is "native", and we can let the packages depend on the right packages

[15:54:00.0000] <nessy>
it's as much "native" on macports as it is on ubuntu

[15:54:01.0000] <nessy>
it's just a package to be installed

[15:54:02.0000] <nessy>
I just did "port install gstreamer" and it got all the dependencies

[15:54:03.0000] <foolip>
macports isn't native to Mac though is it?

[15:54:04.0000] <nessy>
if you do any kind of development on the mac, it is

[15:55:00.0000] <nessy>
normal users probably won't have it installed

[15:55:01.0000] <nessy>
this is why I say: when it's available, it would make sense to fall back to it

[15:55:02.0000] <nessy>
otherwise you can do what you do on windows

[15:55:03.0000] <nessy>
btw: some ppl also have gstreamer "natively" installed on Windows - but probably a lot less than those who have macports or fink

[15:55:04.0000] <foolip>
I will make a minimal mac build, like on windows

[15:56:00.0000] <nessy>
yeah, that will be the best

[15:56:01.0000] <nessy>
but why not have that for all platforms and only use it where gstreamer is not installed natively?

[15:59:00.0000] <foolip>
on Unix there's no need - GStreamer will be there (once our package dependencies or installer scripts make it so)