2010-06-01
[17:00:00.0000] <gsnedders>
(I know the harness is a horrible hack, but it was done as such to get it working somewhat in all browsers, inc. Firefox just after the HTML5 Parser merge which had some odd, fun bug)

[17:01:00.0000] <roc>
it's great

[17:03:00.0000] <gsnedders>
roc: Try running a single test case in it :)

[17:04:00.0000] <gsnedders>
/me can't remember half his gripes with it, but he has plenty

[18:42:00.0000] <nessy>
foolip: does Opera fire the progress event on video elements?

[21:07:00.0000] <foolip>
nessy: yes, as per spec about the timing

[21:07:01.0000] <foolip>
last I checked anyway :)

[21:08:00.0000] <nessy>
yeah - about to write you a bug report :)

[21:08:01.0000] <nessy>
I think onprogress works, but not addEventListener

[21:08:02.0000] <nessy>
but then, it's really hard to test because after every test I have to change the file name, because Opera has buffered it and a shift-reload doesn't reload the video resource

[21:09:00.0000] <nessy>
I have the same problem with Chrome, incidentally

[21:09:01.0000] <nessy>
that's at least one bug ;)

[21:10:00.0000] <nessy>
actually - ditch that - it's only the reloading that doesn't work

[21:10:01.0000] <nessy>
progress events work

[21:13:00.0000] <foolip>
nessy: use a HTTP server and set the source to URL+'?'+something random

[21:14:00.0000] <nessy>
good idea

[21:14:01.0000] <foolip>
is it Math.random() perhaps?

[21:14:02.0000] <nessy>
still, it would be nice if shift-reload on the page ditched the buffered media resource

[21:14:03.0000] <foolip>
anyway, that's what zcorpan does

[21:14:04.0000] <nessy>
Firefox does that

[21:14:05.0000] <foolip>
Opera doesn't have a special reload I think

[21:14:06.0000] <foolip>
at least I never found one

[21:15:00.0000] <nessy>
well, I can reload the resource with shift-click on the reload button

[21:15:01.0000] <nessy>
also, I can remove all other elements from cache in "Tools"->"Delete private data…"

[21:15:02.0000] <nessy>
except for media resources

[21:20:00.0000] <foolip>
oh, that *is* a bug

[21:20:01.0000] <foolip>
can you file one at http://bugs.opera.com ?

[21:21:00.0000] <nessy>
ok

[21:22:00.0000] <nessy>
and webkit is next …. :(

[21:24:00.0000] <nessy>
gah, seems impossible for me to get an account there :(

[21:24:01.0000] <nessy>
will try the wizard ...

[21:27:00.0000] <jwm>
hehe

[21:28:00.0000] <jwm>
amazing how many opera people are in whatwg

[21:32:00.0000] <nessy>
jwm: I am using all browsers :)

[21:39:00.0000] <MikeSmithX>
jwm: it's not amazing if you consider how much market capital Opera has

[21:40:00.0000] <foolip>
we like standards, etc etc :)

[21:45:00.0000] <jwm>
MikeSmith: less than all other browsers?

[21:45:01.0000] <jwm>
hehe

[21:54:00.0000] <MikeSmith>
jwm: inversely proportional

[21:54:01.0000] <MikeSmith>
little-known axiom

[22:17:00.0000] <MikeSmith>
what are some synonyms for the word "encode"?

[22:17:01.0000] <MikeSmith>
in the general sense

[22:19:00.0000] <MikeSmith>
ok, encrypt is close I guess

[01:02:00.0000] <zcorpan_>
foolip: when are you moving?

[01:02:01.0000] <hsivonen>
how up-to-date is html5lib these days?

[01:26:00.0000] <jgraham>
hsivonen: According to the existing testcases it's not that bad

[01:26:01.0000] <jgraham>
A couple of patches haven't been pushed to google code

[01:31:00.0000] <hsivonen>
jgraham: OK. I'm going to advertise it to the SVG WG then

[01:32:00.0000] <hsivonen>
jgraham: it has SVG support, right?

[01:32:01.0000] <jgraham>
hsivonen: Right

[01:33:00.0000] <jgraham>
From memory the missing things is the recent change to <button> or something

[01:33:01.0000] <jgraham>
s/things/thing/

[01:34:00.0000] <hsivonen>
ok

[01:35:00.0000] <annevk>
the SVG minutes didn't consider html5lib to be a real implementation

[01:36:00.0000] <annevk>
if you're talking about that

[01:49:00.0000] <jgraham>
/me wonders what was being discussed

[01:50:00.0000] <annevk>
read it and weep: http://www.w3.org/2010/05/31-svg-minutes.html

[01:53:00.0000] <jgraham>
Sigh

[01:54:00.0000] <jgraham>
No one noticed that webkit are implementing then?

[02:02:00.0000] <annevk>
dunno, roc pointed that out on the list this morning

[02:03:00.0000] <annevk>
btw, I don't think anything will actually happen

[02:06:00.0000] <hsivonen>
/me just followed up to roc's email

[02:08:00.0000] <hsivonen>
the "spec X is not done" mentality *within* the W3C is so bizarre

[02:26:00.0000] <foolip>
zcorpan_: as soon as I'm back in sweden, tillträde is July 5 I think

[02:31:00.0000] <zcorpan_>
foolip: ok. fun :)

[02:32:00.0000] <zcorpan_>
foolip: if i go back to video work, i still have a reason to go to göteborg then :P

[02:33:00.0000] <jgraham>
foolip: You are moving? That's sad :(

[02:33:01.0000] <jgraham>
Well I mean it is hopefully nice for you

[02:34:00.0000] <foolip>
I'm quite thrilled

[02:35:00.0000] <foolip>
zcorpan_: you're always welcome to work on video

[02:35:01.0000] <jgraham>
To be leaving Linköping?

[02:37:00.0000] <foolip>
jgraham: Honestly, yes. Outside of work there's nothing I enjoy about Linköping.

[02:38:00.0000] <jgraham>
Well yeah I would hardly recommend it as a holiday destination

[02:40:00.0000] <foolip>
Most importantly, I think my wife will like Göteborg better.

[02:40:01.0000] <foolip>
More people. People are rare in Sweden...

[02:41:00.0000] <jgraham>
Not stockholm then?

[02:42:00.0000] <foolip>
No, Stockholm is the capital and thus clearly sucks, as I am not from the capital.

[02:42:01.0000] <foolip>
That's the theory anyway :)

[02:42:02.0000] <foolip>
Also, they talk funny.

[02:43:00.0000] <jgraham>
That makes no sense

[02:43:01.0000] <annevk>
foolip, that goes for all of Sweden, I have to say

[02:43:02.0000] <foolip>
Haha

[02:43:03.0000] <zcorpan_>
they talk funny in göteborg too

[02:43:04.0000] <hsivonen>
Is there an Opera office in Stockholm?

[02:43:05.0000] <jgraham>
But sure Göteborg is nice

[02:43:06.0000] <jgraham>
hsivonen: A tiny one. Two engineers and some sales guys

[02:43:07.0000] <zcorpan_>
hsivonen: yes

[02:44:00.0000] <hsivonen>
ok

[02:44:01.0000] <zcorpan_>
there's an office in örebro also

[02:44:02.0000] <zcorpan_>
a tinier one

[02:44:03.0000] <jgraham>
zcorpan_: You don't count as "an office"

[02:44:04.0000] <zcorpan_>
jgraham: :(

[02:44:05.0000] <jgraham>
One person is "working from home" You need two people to make an office

[02:45:00.0000] <jgraham>
Unless you are actually not working from home I guess

[02:46:00.0000] <annevk>
if Rijk is actually in wijk bij duurstede maybe we should petition for a Dutch office here in Utrecht

[03:02:00.0000] <hsivonen>
I wonder what version number I should give to the next release of the V.nu parser...

[03:02:01.0000] <hsivonen>
It should be 1.3 at minimum and I don't want to go all the way to 2.0 yet

[03:03:00.0000] <hsivonen>
possible choices: 1.3, 1.5 and 1.9

[03:07:00.0000] <annevk>
just do 1.3 then

[03:07:01.0000] <annevk>
you're not in competition (yet) so there's no reason to inflate your numbers

[03:11:00.0000] <hsivonen>
annevk: ok. makes sense

[03:11:01.0000] <hsivonen>
I'm thinking of going to 2.0 when I add a sanitizer

[03:16:00.0000] <jgraham>
Make it converge on the golden ratio

[03:17:00.0000] <annevk>
http://www.youtube.com/watch?v=zaT7thTxyq8 is really funny

[03:18:00.0000] <annevk>
for sure our marketing department has their highlights

[03:18:01.0000] <hsivonen>
New thing I learned today: Nautilus caches the file listings on samba mounts for days and doesn't show new files without clicking the reload button

[03:18:02.0000] <hsivonen>
FAIL

[03:24:00.0000] <annevk>
what happened to publishing new drafts?

[03:24:01.0000] <annevk>
did someone object in private or so or are all the chairs on holiday break?

[03:31:00.0000] <annevk>
http://lists.w3.org/Archives/Member/w3c-ac-forum/2010AprJun/0101.html (W3C Member-only) (doesn't provide an answer to the above question but is interesting)

[03:35:00.0000] <Dashiva>
annevk, you tease

[03:40:00.0000] <MikeSmith>
"Potato slow. Opera fast." is great

[03:40:01.0000] <MikeSmith>
who are the dudes in the video, I wonder?

[03:40:02.0000] <zcorpan_>
Odd & Even

[03:41:00.0000] <MikeSmith>
annevk: nobody privately objected to publishing new drafts, btw

[03:41:01.0000] <MikeSmith>
zcorpan_: hey!

[03:41:02.0000] <MikeSmith>
I wanted to bug you about something

[03:41:03.0000] <MikeSmith>
IRI checking

[03:41:04.0000] <zcorpan_>
MikeSmith: i know

[03:41:05.0000] <MikeSmith>
ah, OK

[03:42:00.0000] <zcorpan_>
MikeSmith: the messages look nice, but i don't have time right now to write tests

[03:43:00.0000] <zcorpan_>
/me is eating websocket cookies

[03:43:01.0000] <MikeSmith>
ok, no oproblem

[03:43:02.0000] <MikeSmith>
*problem (no o)

[03:43:03.0000] <MikeSmith>
zcorpan_: yeah, I figured you were still in Websocket Wonderland for a while

[03:43:04.0000] <annevk>
/me is sort of curious what Hixie is up to

[03:44:00.0000] <jgraham>
/me is sort of curious what the chairs are up to

[03:45:00.0000] <hsivonen>
/me wonders if the chairs noticed that Gecko, WebKit and Dreamweaver made progress on implementing elements proposed for removal since the start of the survey

[03:46:00.0000] <annevk>
at least one of them should have noticed stuff landing in WebKit :)

[03:46:01.0000] <hsivonen>
not to suggest that that's in any way relevant to Decisions, of course. ;-)

[03:47:00.0000] <annevk>
depending on the outcome of the decision, we may have came up with a few more <embed>-like elements

[03:48:00.0000] <annevk>
with as difference that these have a proper specification

[03:48:01.0000] <hsivonen>
annevk: do you mean elements supported by browsers and Adobe's authoring tools but not the spec?

[03:49:00.0000] <zcorpan_>
damned browsers making up their own tags

[03:49:01.0000] <annevk>
hsivonen, yeah

[03:49:02.0000] <annevk>
with "the spec" being the "the spec the W3C publishes"

[03:49:03.0000] <jgraham>
If these decisions don't keep the elements then the HTMLWG deserves to become a laughing stock

[03:52:00.0000] <Dashiva>
You can't really blame the WG for chair decisions, can you? Well, and the people who create the issues.

[03:55:00.0000] <Philip`>
You can blame the WG for putting up with it

[03:55:01.0000] <jgraham>
We could have a revolution

[03:55:02.0000] <Philip`>
The W3C and the chairs only have authority because people choose to give it to them

[03:56:00.0000] <jgraham>
With pitchforks and torches and things

[03:56:01.0000] <annevk>
Revolution is XHTML2

[03:56:02.0000] <jgraham>
Although those typically end with the revolutionary leader dyning in an unpleasant way, so if anyone asks it was Philip`'s idea

[03:56:03.0000] <annevk>
Though from a W3C perspective maybe HTML5 is indeed revolution... Strange world.

[03:56:04.0000] <jgraham>
*dying

[03:56:05.0000] <Dashiva>
annevk: That depends on what you revolt _against_

[03:57:00.0000] <Dashiva>
You could say XHTML2 was a revolution against HTML, and HTML5 is a revolution against XHTML2

[03:57:01.0000] <annevk>
What if you just wanna evolve?

[03:58:00.0000] <Philip`>
The WG members could form a new working group that's entirely independent of the W3C with its own non-consensus model for making decisions and that develops its own specification called HTML5

[03:58:01.0000] <Dashiva>
That's a question of authority, I suppose

[03:58:02.0000] <Philip`>
and so it's their fault if they choose to work within the existing system instead

[03:58:03.0000] <Dashiva>
From the W3C viewpoint, I'm sure XHTML2 _was_ evolution

[03:59:00.0000] <annevk>
From miles high it may indeed seem that way

[04:00:00.0000] <aho>
i liked xhtml2's "hx"

[04:00:01.0000] <aho>
using h1 (etc) in conjunction with nested sections is sorta weird

[04:00:02.0000] <annevk>
AlexD keeps it up: "HTML5 is vapourware _until_ the spec. reaches recommendation status."

[04:01:00.0000] <jgraham>
aho: Yeah but it is kinda broken because you can still use <h> and <h1> in the same document

[04:01:01.0000] <annevk>
vapourware: "New software that has been announced or marketed but has not been produced."

[04:01:02.0000] <jgraham>
This is like the opposite of vapourware

[04:01:03.0000] <jgraham>
We haven't announce it is done but it is still beiong produced

[04:01:04.0000] <aho>
jgraham, still... i'd like to point out that this is a section-nesting heading and not a regular one

[04:01:05.0000] <Dashiva>
solidware?

[04:02:00.0000] <jgraham>
aho: They are all just headings

[04:02:01.0000] <aho>
using a weight-free heading would do the trick

[04:02:02.0000] <annevk>
Guess I should stop... http://xkcd.com/386/

[04:02:03.0000] <aho>
jgraham, yea... they are all just tags (?) :>

[04:04:00.0000] <aho>
i just liked that there was some heading which indicated that its intended weight depends on the section nesting

[04:04:01.0000] <Dashiva>
aho: Isn't that obvious, though? Since h1 would only be at the very top of the document normally

[04:04:02.0000] <aho>
and tables are normally only used for tabular data? :D

[04:06:00.0000] <annevk>
aho, by that argument we should have <table2> which will be only used correctly, which of course is nonsense

[04:06:01.0000] <jgraham>
Dashiva: I think something like LiquidCrystalWare would be better since it appears all over the place but the theory is still not complete

[04:06:02.0000] <aho>
bad analogy :>

[04:07:00.0000] <jgraham>
(of course the theory of any state of matter is not complete but...)

[04:07:01.0000] <annevk>
aho, works for me :)

[04:09:00.0000] <Hixie>
you know, if the svgwg thinks html5 is vapourware, maybe we should just use the same process as svg 1.2 did

[04:10:00.0000] <Hixie>
and just stamp html5 as a REC and ignore 10% of the feedback we've gotten, just like they did

[04:10:01.0000] <Hixie>
http://ln.hixie.ch/?start=1155235213&count=1

[04:10:02.0000] <Dashiva>
Hixie: The rules don't apply to XML languages, you should know that by now

[04:11:00.0000] <Hixie>
html5 has an xml serialisation!

[04:11:01.0000] <Dashiva>
A fig leaf, nothing more

[04:11:02.0000] <aho>
annevk, "Sections may contain headings of any rank, but authors are strongly encouraged to either use only h1  elements, or to use elements of the appropriate rank  for the section's nesting level."

[04:12:00.0000] <aho>
if there were an element for headings without any weight, it would be clear which route you picked

[04:12:01.0000] <Dashiva>
It would be clear, but for what purpose?

[04:13:00.0000] <aho>
say there is a doc... and 5 years later some different person writes a new stylesheet

[04:14:00.0000] <Dashiva>
No problem at all

[04:15:00.0000] <aho>
there is also no problem with using single letter variables exclusively

[04:15:01.0000] <aho>
it just takes a bit longer to understand than necessary

[04:15:02.0000] <Dashiva>
There's a rather significant backwards compatibility problem, for no actual benefit

[04:15:03.0000] <annevk>
aho, just use <h1>

[04:16:00.0000] <annevk>
<h2-6> is only if you care about backwards compatibility

[04:16:01.0000] <aho>
if you use canvas and all those new apis... why would you care about such a small detail?

[04:16:02.0000] <annevk>
indeed

[04:16:03.0000] <annevk>
why do you?

[04:17:00.0000] <aho>
breaking backward copatibility of weight-free headings, that is

[04:17:01.0000] <Dashiva>
Because not every page uses canvas?

[04:18:00.0000] <aho>
article, aside, section, nav... they "break" backwards compatibility the same way

[04:18:01.0000] <aho>
not really much harm done there

[04:18:02.0000] <Dashiva>
Those are different. They _add_ semantics to what is considered meaningless by older browsers

[04:18:03.0000] <aho>
well, i did like h/hx. that's all :P

[04:19:00.0000] <aho>
being able to differenciate between weighted headings and weight-free/section-nesting headings doesn't add anything?

[04:20:00.0000] <annevk>
nope

[04:20:01.0000] <aho>
wouldn't that make the structure clearer?

[04:20:02.0000] <Dashiva>
You're the one claiming it does, shouldn't you answer?

[04:20:03.0000] <Dashiva>
<Dashiva> aho: Isn't that obvious, though? Since h1 would only be at the very top of the document normally

[04:24:00.0000] <Dashiva>
Hixie: Last link in that blog post is dead, by the way

[04:25:00.0000] <aho>
well, imo h/hx:h1-h6 is sorta like span:em/b/strong/i

[04:25:01.0000] <aho>
i really liked that

[04:25:02.0000] <aho>
but it's not like this will be added to any spec any time soon

[04:25:03.0000] <Hixie>
Dashiva: that's even better

[04:25:04.0000] <Hixie>
Dashiva: so much for stable links at the w3c!

[04:25:05.0000] <aho>
so, i don't really see why you think it's necessary to discuss this in depth

[04:25:06.0000] <Dashiva>
Maybe it wasn't cool enough to not change

[04:28:00.0000] <Dashiva>
aho: Span has no semantics, the proposed hx does.

[04:28:01.0000] <aho>
yes

[04:28:02.0000] <aho>
h/hx on its own doesn't have specific weight

[04:28:03.0000] <aho>
sorta like a td on its own doesn't have much use :>

[04:32:00.0000] <Dashiva>
I guess you're just an aho, after all...

[04:32:01.0000] <aho>
*clap* *clap*

[04:34:00.0000] <annevk>
Dashiva, Hixie, I guess it is http://www.w3.org/TR/2006/CR-SVGMobile12-20060810/lc3-replies-public.html

[04:38:00.0000] <aho>
Dashiva, if you see some hx somewhere, you know it's a heading which relies on section-wrapping. if you see h1, it's either the old kind of heading *or* one which relies on section-wrapping. the hx element doesn't have this kind of ambiguity - and that's something i generally like

[04:39:00.0000] <aho>
e.g. i like code conventions, because they help getting rid of that stuff

[04:39:01.0000] <aho>
saves time

[04:40:00.0000] <jgraham>
aho: I have no idea how you envison that working

[04:41:00.0000] <jgraham>
What happens with <body><hx><section><hx><section<h1>

[04:41:01.0000] <jgraham>
(ignore the typo)

[04:41:02.0000] <jgraham>
Is the <h1> a top level heading or a third level heading

[04:43:00.0000] <jgraham>
(exactly this use case happens any time a mashup embeds content that uses <h1> in a container that uses <section>)

[04:45:00.0000] <aho>
what happens in a mashup that embeds content that uses <h2-6> in a container that uses section? ;>

[04:47:00.0000] <annevk>
it would just work

[04:47:01.0000] <aho>
so, that section h4 thing you just embedded will look like what exactly?

[04:48:00.0000] <jgraham>
aho: The logical outline is well defined

[04:49:00.0000] <jgraham>
aho: The style will (by default) just depend on the tag name, but that is a CSS limitation taht would be worse if we also had <hx>

[04:49:01.0000] <jgraham>
Hopefully we will get ::section(n) and so on

[04:49:02.0000] <aho>
-moz-any() is nice for this stuff .)

[04:50:00.0000] <aho>
e.g. section any(h1,h2,h3,h4,h5,h6,hx) { first level heading}

[04:51:00.0000] <aho>
if it all were hx you wouldn't have to do that though

[04:51:01.0000] <jgraham>
Right, but make the section a pseudo class that matches all level one sections

[04:51:02.0000] <jgraham>
aho: It can't all be <hx>

[04:52:00.0000] <aho>
why not? (if we ignore mashups)

[04:52:01.0000] <jgraham>
We don't ignore mashups

[04:52:02.0000] <jgraham>
If we make invalid assumptions we can prove anything

[04:52:03.0000] <aho>
i never used any mashup, i never made one, i never will make one

[04:52:04.0000] <aho>
i don't even know one popular mashup

[04:53:00.0000] <Dashiva>
If you aren't making mashups, it's pretty much a non-issue

[04:53:01.0000] <jgraham>
You never used an online feed reader, for example?

[04:53:02.0000] <Dashiva>
Mashups are the important case, not something to ignore

[04:55:00.0000] <aho>
aggregators typically process the feeds quite a bit, but they generally don't really work all that well

[04:58:00.0000] <jgraham>
"process the feeds quite a bit" typically means "do some sanitization". Having to rewrite the heading tags to fit in with the output template is rather different

[04:59:00.0000] <aho>
well, currently you just have random headings there

[04:59:01.0000] <aho>
i don't really see how that's super awesome

[04:59:02.0000] <aho>
<:

[05:09:00.0000] <aho>
in retrospect i think it would have been better if i hadn't mentioned xhtml2's hx. i hadn't expected this kind of annoying discussion... giving that it's a dead spec n all... oh well :>

[05:09:01.0000] <aho>
it's nice that html5 picked up section tho (there! i said it) :>

[06:11:00.0000] <hsivonen>
/me notes "WebM will be supported by all three major browsers: Google Chrome, Mozilla Firefox and Opera" on OMG! UBUNTU!

[06:11:01.0000] <Peter`>
I'm still wondering what Apple will do with WebM

[06:12:00.0000] <Peter`>
no word on that I'm assuming?

[06:12:01.0000] <Peter`>
(Next to Steve Jobs saying he doesn't like it)

[06:17:00.0000] <Dashiva>
Of course he doesn't like it, it wasn't invited by him :)

[06:53:00.0000] <Lachy>
Peter`, Apple won't support WebM while there is still a percieved risk of patent trolls

[06:54:00.0000] <Peter`>
Clearity on the legal patent subjects might take months, if not years

[06:54:01.0000] <Lachy>
they will at least be waiting to see what comes out of the FUD from the MPEG LA, if anything at all. But, they'll likely just maintain the threat for ages, like they have with Theora.

[06:55:00.0000] <Lachy>
right, so I wouldn't recommend holding your breath for Apple to support it.

[06:55:01.0000] <hsivonen>
Lachy: citation needed

[06:55:02.0000] <Peter`>
I'm quite curious about the future and implementation of WebM, specifically now Microsoft is more or less supporting it as well

[06:56:00.0000] <Lachy>
but I believe there will at least be 3rd party quick time components made available one day, which will enable support in Safari.

[06:57:00.0000] <daedb>
Apple doesn't even support Vorbis yet, so any official support for VP8 seems incredibly unlikely.

[06:57:01.0000] <Peter`>
But following Theora and the lack of a formal stance of the company on WebM, different from Steve Jobs' own thoughts, it's still somewhat undefined right now. As hsivonen already said, while it's unlikely, nothing is clear

[06:57:02.0000] <hsivonen>
It's so frustrating that XiphQT still doesn't have Thusnelda. it would be nice to have WebM support rolled into XiphQT

[06:58:00.0000] <hsivonen>
Peter`: what have I said?

[06:58:01.0000] <Peter`>
"citation needed"

[06:58:02.0000] <Lachy>
hsivonen, http://www.theregister.co.uk/2010/05/20/jobs_on_vp8/

[06:59:00.0000] <Lachy>
http://www.theregister.co.uk/2010/05/21/mpegla_mulls_patent_license_for_webm/

[06:59:01.0000] <hsivonen>
Peter`: I meant I wanted to see Lachy's source

[06:59:02.0000] <hsivonen>
Lachy: thanks

[07:00:00.0000] <hsivonen>
Lachy: I think the Register is extrapolating quite a bit from Steve dropping one URL without commentary

[07:01:00.0000] <Lachy>
yes, that's true. But it's not too much of a stretch given what Jobs previously said about forming a patent pool around open video codecs

[07:02:00.0000] <Lachy>
There was another article I read that was a set of email exchanges between a reporter, Larry Horn from the MPEG LA and, I believe, someone from Xiph, which talked about the FUD that the MPEG LA have maintained for years over theora

[07:02:01.0000] <Peter`>
hsivonen: alright, sorry :)

[07:04:00.0000] <hsivonen>
Lachy: well, yeah, anyone can be "looking into" whatever

[07:05:00.0000] <hsivonen>
(As I understand it, the FUD Monty was referring to wasn't MPEG-LA FUD per se but FUD from the MPEG camp. IIRC, MP3 licenses haven't been managed by MPEG-LA.)

[07:06:00.0000] <Peter`>
I'm assuming Google knows what they're doing though, also considering the list of companies who already committed to VP8, there has to be a certain degree of certainty. A chance of submarine patents arising always remains, with any product that's entering an already-existing market

[07:08:00.0000] <hsivonen>
when creating a Ubuntu USB installation stick with a large (over 1 GB) persistence file, is the creation of the persistence file supposed to be insanely slow?

[07:08:01.0000] <hsivonen>
I'm wondering if I have bought a bad USB stick

[07:09:00.0000] <hsivonen>
also, it seem the progress meter was programmed by the person who came up with the Windows file copy progress meter

[07:14:00.0000] <Lachy>
Peter`, IANAL, but I think the number of other companies supporting WebM, many of whom I assume hold numerous patents of their own, give some level of protection because they may hold patents that they can use defensively against others who aggressively attack VP8

[07:17:00.0000] <zcorpan_>
http://twitter.com/microft/statuses/15176790401

[07:26:00.0000] <Peter`>
Lachy: That's a good point, specifically given that the license states that you cannot file patent claims against Google if you're using the specification. Widely interpreted, one could say that they're using open-source to minimize the chance of patent claims

[07:26:01.0000] <Peter`>
Then again, IANAL either

[07:30:00.0000] <MikeSmith>
can anybody remind me what happened the the "Copy and paste" section of the HTML5 spec?

[07:31:00.0000] <MikeSmith>
it seems to be commented out in the source now

[07:34:00.0000] <hsivonen>
MikeSmith: it used to define copy and paste in terms of drag and drop

[07:34:01.0000] <MikeSmith>
ok

[07:34:02.0000] <hsivonen>
MikeSmith: then it was pointed out to Hixie that some browser(s?) exposed copy and paste in a distinct way already

[07:35:00.0000] <MikeSmith>
ah

[07:35:01.0000] <MikeSmith>
I wonder if there is open bug on it

[07:36:00.0000] <hsivonen>
What kind of write speed (MB/s) should one expect from a USB stick?

[07:36:01.0000] <gsnedders>
hsivonen: Varies massively

[07:38:00.0000] <hsivonen>
OK, I bought USB sticks that suck

[07:38:01.0000] <hsivonen>
sigh

[07:40:00.0000] <hsivonen>
lesson learned: shop at a computer store. don't rely on office supply stores

[07:41:00.0000] <hsivonen>
aaargh. now after spending over 2 hours creating the persistence file, it's just starting to create a file system inside the persistence file

[07:42:00.0000] <MikeSmith>
hsivonen: found http://www.w3.org/Bugs/Public/show_bug.cgi?id=8800

[09:24:00.0000] <Peter`>
MikeSmith (?)/W3C: The link in the latest Twitter message gives a 404 error http://twitter.com/w3c/status/15185050915

[09:41:00.0000] <boblet>
anyone into diff and on a Mac should check out www.kaleidoscopeapp.com

[09:42:00.0000] <boblet>
/me is sad that he prolly fits this category

[10:18:00.0000] <jwalden>
/me is surprised he can't find an Opera bug-reporting menu item or UI in Opera itself (not for site bugs, but for Opera bugs)

[10:19:00.0000] <Philip`>
jwalden: Help -> Support -> Report a bug?

[10:19:01.0000] <jwalden>
Philip`: more indirect than I'd expected

[10:19:02.0000] <jwalden>
/me should note this is a beta version

[10:20:00.0000] <jwalden>
I guess Firefox doesn't have a similar item, so I guess it's not too odd

[10:22:00.0000] <jwalden>
/me finally got annoyed enough to report http://krijnhoetmer.nl/irc-logs/whatwg/20100405#l-486 now that it's breaking his site's intended rendering

[10:23:00.0000] <jwalden>
although, big blue background is probably a better rendering than the non-scaled background you get in a browser that doesn't support background-size, if your screen resolution/browser window are large enough

[10:23:01.0000] <jwalden>
or at least less failtastic

[12:35:00.0000] <variable>
.

[12:35:01.0000] <TabAtkins>
.?

[12:37:00.0000] <jgraham>
.*

[12:38:00.0000] <TabAtkins>
.{0,5}

[12:38:01.0000] <TabAtkins>
Hello

[12:38:02.0000] <jgraham>
True

[13:18:00.0000] <TabAtkins>
Today's surprising spellcheck list omission: "combinations".

[13:18:01.0000] <variable>
TabAtkins: which speller checker?

[13:19:00.0000] <TabAtkins>
Firefox on ubuntu.

[13:21:00.0000] <nielsle>
Are webgl and notifications likely to become official html?

[13:21:01.0000] <TabAtkins>
I know that Chrome either has them or is implementing them.  Can't speak for other browsers.

[13:23:00.0000] <gsnedders>
WebGL will remain a separate spec, and so will notifications

[13:23:01.0000] <gsnedders>
It makes little sense to make them part of the HTML spec, IMO

[13:23:02.0000] <gsnedders>
But I guess it's a question of what you mean by "official HTML"

[13:25:00.0000] <nielsle>
I meant html as defined by w3c, but I think that I got my answer.

[14:00:00.0000] <jgraham>
nielsle: WebGL is not being done at W3C. Noifications I'm not so sure about but only Google seem really interested at the moment

[14:01:00.0000] <jgraham>
(WebGL is nevertheless becoming a standard; other standards organisations are avaliable, as they would say on the BBC)

[14:05:00.0000] <nielsle>
Thank you for the answers. (I am not an expert on all this, but I am not fully convinced that non-standard html is good thing.)

[14:08:00.0000] <gsnedders>
/me gets caught out by the fact that null does not equal anything in SQL

[14:09:00.0000] <TabAtkins>
NULL IS NULL

[14:09:01.0000] <ment>
badam bam tssss...

[14:10:00.0000] <gsnedders>
TabAtkins: Yeah, found that now

[14:10:01.0000] <Dashiva>
Just like NaN isNaN()

[14:12:00.0000] <gsnedders>
How do GROUP BY and joins interact?

[14:12:01.0000] <TabAtkins>
GROUP BY applies to the joined table.

[14:12:02.0000] <TabAtkins>
(Unless you've grouped in a subquery, of course.)

[14:13:00.0000] <gsnedders>
Okay, so it is logical

[14:37:00.0000] <AryehGregor>
Does Steve Jobs actually respond to a large fraction of the random mail he receives from the public?

[14:38:00.0000] <AryehGregor>
Maybe it's just a press corps, but everyone seems to assume that the responses are actually from him.

[14:40:00.0000] <Philip`>
AryehGregor: If Santa can reply to all the messages I send him, I'm sure Steve Jobs can

[14:40:01.0000] <TabAtkins>
You think Santa doesn't have an elven press corps?

[14:41:00.0000] <Philip`>
Of course not

[14:41:01.0000] <Philip`>
Elves don't even exist

[14:41:02.0000] <TabAtkins>
Oh ye of little Christmas faith...

[14:41:03.0000] <Philip`>
It's just him and Rudolph

[14:42:00.0000] <TabAtkins>
I have a folk song that clearly contradicts your assertion.


2010-06-02
[17:05:00.0000] <annevk>
oh man

[17:05:01.0000] <annevk>
http://daringfireball.net/linked/2010/06/01/thurrott-understanding is brilliant

[17:06:00.0000] <annevk>
I really dig the whole quote what the expert said before the fact after the fact thing Gruber has going on

[17:09:00.0000] <annevk>
"First Draft of SPARQL 1.1 Federation Extensions Published; Five SPARQL 1.1 Drafts Updated" federation extensions?

[17:09:01.0000] <annevk>
gotta admit that this sounds pretty serious

[17:09:02.0000] <annevk>
serious and boring

[18:17:00.0000] <mikl0>
http://pastebin.com/iLAHWMAR can someone explain this DOM behavior to me?

[18:20:00.0000] <AryehGregor>
mikl0, browsers do not support the self-closing /> syntax in text/html.

[18:20:01.0000] <AryehGregor>
You have to use an explicit closing tag like </firstChild>.

[18:21:00.0000] <mikl0>
so i must use <firstChild></firstChild>. however if i do XHTML it will work as that does support self closing right?

[18:22:00.0000] <Philip`>
mikl0: Yes, if by "do XHTML" you mean serving the file as application/xhtml+xml

[18:22:01.0000] <Philip`>
(not just changing the doctype or the <meta>)

[18:22:02.0000] <mikl0>
AryehGregor: thanks for always asnwering i learn so much.

[18:23:00.0000] <mikl0>
Philip`: yes doctype, meta, http header, and .xhtml extension. thanks too Philip` 

[18:23:01.0000] <tabatkins>
Dammit, my computer's broken.

[18:23:02.0000] <Philip`>
TabAtkins: Buy a new one

[18:24:00.0000] <TabAtkins>
"computer" is defined very loosely here.  Something between me and the rest of the internet is broken, and causing all of my requests to take a retarded amount of time.

[18:25:00.0000] <TabAtkins>
And... now it's back.

[18:25:01.0000] <AryehGregor>
TabAtkins, make sure no one on your network is using P2P or other bandwidth-heavy stuff.

[18:25:02.0000] <AryehGregor>
I saw 600 ms ping recently to my ISP when another computer in the house was making an initial backup to Mozy.

[18:26:00.0000] <TabAtkins>
I'm on Google's network, so I have no idea how I would check, nor what I would do if they were.

[18:26:01.0000] <AryehGregor>
Oh, right.

[18:26:02.0000] <AryehGregor>
I thought you were at home.

[18:26:03.0000] <Philip`>
I would have thought Google would have enough bandwidth to cope with someone using P2P software

[18:27:00.0000] <TabAtkins>
AryehGregor: You're 3 hours later than me.  I'm at work for another half hour.

[18:27:01.0000] <TabAtkins>
Philip`: One would think so, yes.  I suspect it was something crazy related to startup stuff, since I just restarted my computer.

[18:27:02.0000] <AryehGregor>
Time is meaningless to me.  I've gotten up and have not yet gone to bed, that's all I notice.

[18:28:00.0000] <AryehGregor>
That's a slight exaggeration, but things like "working hours" are certainly meaningless to me.

[18:28:01.0000] <TabAtkins>
Somehow I was using 6 gigs of RAM between a handful of browser windows doing non-intensive things, and a few idle consoles.

[18:28:02.0000] <AryehGregor>
o_O

[18:28:03.0000] <AryehGregor>
Not counting buffers and cache?

[18:28:04.0000] <gsnedders>
TabAtkins: Have you been looking at memory usage bugs?

[18:29:00.0000] <TabAtkins>
gsnedders: Not on purpose.  ^_^

[18:29:01.0000] <TabAtkins>
Anyway, 20 days of uptime, so shrug.  Same thing happened after roughly 20 days of uptime previously.

[18:29:02.0000] <AryehGregor>
What OS?

[18:29:03.0000] <TabAtkins>
GUbuntu.

[18:30:00.0000] <AryehGregor>
I had XOrg using 1.5G of memory a while back, but that was after more like 90d of uptime.

[18:30:01.0000] <gsnedders>
Yeah, I've had that happen too

[18:34:00.0000] <AryehGregor>
This is why non-memory-managed languages stink.  But I guess we have to live with them.

[18:35:00.0000] <gsnedders>
memory-managed languages stink because their GC tends to have bugs too :)

[18:39:00.0000] <TabAtkins>
Single source of bugs in an area where the bugs are relevant to the very operation of the component is better than bugs distributed throughout every program that aren't critical to the operation of the programs.

[18:40:00.0000] <AryehGregor>
Well, I mostly write in PHP.  Despite the PHP developers' most strenuous efforts to introduce bugs and outright insanity into all parts of the reference language implementation, it's hard to get memory to leak past the end of a request.

[18:40:01.0000] <Philip`>
Memory-managed languages don't protect you from poor caching strategies, which is sometimes what causes painful memory usage

[18:40:02.0000] <AryehGregor>
Well, it's an improvement.

[18:47:00.0000] <Philip`>
AryehGregor: It's better than nothing, there's just a danger that people think GC means they no longer have to worry about the lifetimes of allocated objects

[18:48:00.0000] <AryehGregor>
From experience, it seems like people don't even notice the GC once they're used to it.  It doesn't make them less able to troubleshoot memory usage.

[18:48:01.0000] <Philip`>
when really you still ought to worry and the GC just means you don't have any explicit end-of-lifetime markers in your code

[18:48:02.0000] <AryehGregor>
It means that you almost never have to worry about explicitly deallocating things, not never.

[18:48:03.0000] <AryehGregor>
I've cut memory usage using unset() in PHP before.

[18:50:00.0000] <Philip`>
/me supposes he's just mildly unhappy with how Eclipse seems to be using 1GB of memory, when he's simply been using it as a fancy text editor for a day

[19:04:00.0000] <AryehGregor>
/me has never noticed vim using an appreciable amount of memory.

[00:06:00.0000] <MikeSmith>
hsivonen: I have a question about if/how to use the v.nu HTML5 parser in conjunction with existing command-line Java apps

[00:06:01.0000] <MikeSmith>
background is this  -

[00:06:02.0000] <MikeSmith>
if I want to use a different parser with, e.g., Saxon

[00:07:00.0000] <MikeSmith>
instead of the built-in parser that Saxon uses

[00:07:01.0000] <MikeSmith>
Aelfred I guess

[00:07:02.0000] <MikeSmith>
then I can do this:

[00:07:03.0000] <MikeSmith>
java \ 

[00:07:04.0000] <MikeSmith>
  -Djavax.xml.parsers.DocumentBuilderFactory=\

[00:07:05.0000] <MikeSmith>
     org.apache.xerces.jaxp.DocumentBuilderFactoryImpl \

[00:07:06.0000] <MikeSmith>
  -Djavax.xml.parsers.SAXParserFactory=\

[00:07:07.0000] <MikeSmith>
     org.apache.xerces.jaxp.SAXParserFactoryImpl \

[00:07:08.0000] <MikeSmith>
  com.icl.saxon.StyleSheet  \

[00:07:09.0000] <MikeSmith>
...

[00:08:00.0000] <MikeSmith>
and that'll cause Saxon to use Xerces instead

[00:08:01.0000] <MikeSmith>
as far as I can tell, I can't currently do that with the v.nu HTML parser, right?

[00:09:00.0000] <MikeSmith>
so is there some other way I could cause Saxon to use the v.nu parser instead?

[00:09:01.0000] <MikeSmith>
or could we add support for those properties to the v.nu parser?

[00:09:02.0000] <MikeSmith>
that's some JAXP thing, right?

[00:14:00.0000] <MikeSmith>
/me takes a look at the XSLT4HTML code for the time being

[00:27:00.0000] <fantasai>
What's the default output encoding for html5lib, utf-8?

[00:28:00.0000] <MikeSmith>
fantasai: yeah

[00:28:01.0000] <fantasai>
MikeSmith: or is it the system default encoding?

[00:29:00.0000] <MikeSmith>
I believe it is utf-8

[00:29:01.0000] <MikeSmith>
not whatever the default system encoding is

[00:32:00.0000] <MikeSmith>
/me looks at some code

[00:33:00.0000] <MikeSmith>
hmm, I find "serializer.HTMLSerializer(**kwargs).serialize(tokens, encoding='utf-8')" in code I'm using

[00:34:00.0000] <MikeSmith>
so maybe it's not actually utf-8 by default

[00:34:01.0000] <MikeSmith>
/me glances around for help

[00:38:00.0000] <MikeSmith>
checking the html5lib source, it seems that html5lib/serializer code does not itself set encoding to utf-8 

[00:39:00.0000] <MikeSmith>
fantasai: so I think I was wrong

[00:39:01.0000] <MikeSmith>
I can't see anywhere in the html5lib code itself where the output encoding is set to utf-8

[01:07:00.0000] <Mrmil>
Anyone has an idea why wrap="off" for textarea is not present in the spec?

[01:08:00.0000] <zcorpan_>
Mrmil: it's semantically equivalent to wrap="soft"

[01:09:00.0000] <annevk>
rendering is defined

[01:19:00.0000] <Mrmil>
wrap soft/hard doesn't seem to work (or I'm not getting it right), maybe browsers don't support it?

[02:10:00.0000] <MikeSmith>
listening to introductory presentation related to container-based virtualization

[02:10:01.0000] <MikeSmith>
network namespace isolation

[02:12:00.0000] <kennyluck>
It looks like a hardware-related talk to me, so I wasn't interested.

[02:16:00.0000] <MikeSmith>
kennyluck: it's not about software

[02:17:00.0000] <kennyluck>
OK

[02:17:01.0000] <MikeSmith>
kernel, socket and routing library code

[02:17:02.0000] <MikeSmith>
dude is from Tsinghua university

[02:18:00.0000] <kennyluck>
Ah, it still sounds interesting.

[02:18:01.0000] <kennyluck>
Yeah, I noticed that, FYI ,there's a Tsuinghua uni. in Taiwan and a Tsuinghua uni. in Mainland China.

[02:18:02.0000] <kennyluck>
I was wondering which one does he come from.

[02:19:00.0000] <MikeSmith>
solution slide: "associate each uid with an IP address"

[02:19:01.0000] <MikeSmith>
kennyluck: from accent, I guess mainland

[02:19:02.0000] <kennyluck>
Oh, that sounds quite crazy.

[02:20:00.0000] <kennyluck>
ref. "associate each uid with an IP address"

[02:20:01.0000] <MikeSmith>
http://www.nrc.tsinghua.edu.cn/7_english/profile/chenmaoke.htm

[02:32:00.0000] <jgraham>
So the Decisions seemed to mainly take into account what people said on the poll even though it said not to say anything in the poll unless it wasw not already in a change proposal

[02:32:01.0000] <jgraham>
Oh, TabAtkins said the same thing on the list

[02:35:00.0000] <jgraham>
and if I understand what Sam said it seems like the weight given to change proposals is asymmetric so only arguments in favour of change are considered

[02:35:01.0000] <jgraham>
/me is confused

[02:37:00.0000] <hsivonen>
jgraham: yeah, I think the instructions given by the chairs pre-poll and the explanations post-poll don't go well together

[02:38:00.0000] <hsivonen>
frankly, I'm not at all surprised

[02:41:00.0000] <hsivonen>
TabAtkins: maybe next time in addition to submitting a no-change Change Proposal, you should also object to the deletionist Change Proposal by saying that something like "I repeat and raise as objections against this Proposal all the points in favor of no change raised in the no-change Proposal as if fully restated herein"

[02:43:00.0000] <hsivonen>
jgraham: which is why I took the time to file an objection that was slightly different from the points made in the no-change Proposal

[02:46:00.0000] <jgraham>
hsivonen: It appears to have been worthwhile. But should have been unnecessary

[02:48:00.0000] <hsivonen>
In retrospect, I probably should have copied and pasted my objection to other polls too

[02:48:01.0000] <hsivonen>
just in case

[03:16:00.0000] <Hixie>
hsivonen, jgraham: you could decode sam's e-mails and figure out what the decisions were? I wasn't even able to penetrate his wishy washy language to work out what to change

[03:17:00.0000] <Hixie>
btw hsivonen you might want to comment on https://bugs.webkit.org/show_bug.cgi?id=40047

[03:34:00.0000] <daedb_>
Oh yay, there's fun stuff on public-html again. Nice conspiracy theories :D

[03:36:00.0000] <Hixie>
is there some equivalent of svn blame that will tell me what revisions subsequently _removed_ lines that were checked in at a particular revision?

[03:46:00.0000] <annevk>
Hixie, scan the commit messages? :)

[03:47:00.0000] <Hixie>
there are 1000s :-)

[03:48:00.0000] <Philip`>
grep and a bisect tool?

[03:49:00.0000] <Philip`>
like http://search.cpan.org/dist/App-SVN-Bisect/bin/svn-bisect

[03:49:01.0000] <annevk>
you can do an search through them

[03:50:00.0000] <Hixie>
so "no"? :-)

[03:50:01.0000] <jgraham>
So if I have <iframe src="javascript:something"> in the source of a document, what happens? How many load events do I get?

[03:51:00.0000] <jgraham>
(like is there an implicit load of about:blank first?)

[03:59:00.0000] <zcorpan_>
jgraham: seems browsers differ

[03:59:01.0000] <zcorpan_>
javascript:void() gives 1 load in opera and chrome, 0 in firefox

[03:59:02.0000] <zcorpan_>
javascript:'' gives 1 in opera and firefox, 2 in chrome

[04:00:00.0000] <jgraham>
zcorpan_: Any idea what the spec says, or whether it is defensible?

[04:00:01.0000] <zcorpan_>
haven't looked at the spec

[05:15:00.0000] <hsivonen>
Hixie: regarding https://bugs.webkit.org/show_bug.cgi?id=40047 , I think Gecko may still be blocking on inline scripts if there are pending sheets

[05:16:00.0000] <RiPPER>
hi

[05:27:00.0000] <hsivonen>
Hixie: actually, I think the spec in now racy in the blocking behavior of inline scripts

[05:27:01.0000] <hsivonen>
sigh

[05:28:00.0000] <Dashiva>
The smell of Process in the morning...

[05:28:01.0000] <RiPPER>
I read somewhere that in html 5 I can use xhtml syntax, is this right?

[05:29:00.0000] <Dashiva>
No, but some xhtml-isms are conforming

[05:29:01.0000] <Dashiva>
Like <br/> instead of <br>

[05:29:02.0000] <hsivonen>
RiPPER: you can 1) use an XML serialization of HTML5 when serving content as application/xhtml+xml or you can 2) use certain XHTMLisms in text/html

[05:30:00.0000] <hsivonen>
RiPPER: you can't use the whole of XHTML syntax in text/html, though

[05:30:01.0000] <Hixie>
hsivonen: really? i thought we'd fixed all the racy behaviour...

[05:32:00.0000] <hsivonen>
Hixie: by inspection, Gecko's implementation sure looks racy...

[05:33:00.0000] <Hixie>
hsivonen: well if the spec is racy, file a bug

[05:35:00.0000] <hsivonen>
Hixie: the spec isn't racy, it seems. But the spec looks unacceptable for implementation

[05:36:00.0000] <Hixie>
well, file a bug for that too

[05:36:01.0000] <hsivonen>
both to me and, it seems, to Eric Seidel per comment #2 in the WebKit bug

[05:37:00.0000] <hsivonen>
since it would be pretty crazy to spin a nested event loop from within document.write() in order to wait for a style sheet

[05:37:01.0000] <hsivonen>
I'll file a spec bug

[05:37:02.0000] <Hixie>
well how else are you going to wait for a style sheet?

[05:38:00.0000] <hsivonen>
Hixie: do we know that document.written inline scripts need to wait for sheet for Web compat?

[05:38:01.0000] <Hixie>
iirc that's what browsers did when i tested it

[05:39:00.0000] <Hixie>
but who knows

[05:39:01.0000] <hsivonen>
by inspection, it seems to me that Gecko doesn't do that

[05:39:02.0000] <annevk>
Gecko used to have synchronous style sheet loading

[05:39:03.0000] <Hixie>
making things dependent on whether they'd inserted via document.write() or not seems pretty wacky

[05:40:00.0000] <hsivonen>
my axiom is that nested event loops are wacky

[05:40:01.0000] <Hixie>
well on a side note, technically the html5 spec doesn't have any nested event loops

[05:40:02.0000] <Hixie>
but that's neither here nor there

[05:41:00.0000] <Hixie>
my point is that since we need to spin the event loop for regular <script>s, that boat has sailed

[05:41:01.0000] <hsivonen>
that doesn't help much, since we are dealing with C++ method call semantics

[05:41:02.0000] <hsivonen>
so there aren't fancy yields and continuations available

[05:41:03.0000] <Hixie>
so why make an exception for document.written ones?

[05:41:04.0000] <annevk>
I think we had the ability for some time to stop script execution when a layout DOM attribute was asked for to be able to fetch the necessary style sheets and compute something... Mostly to work with sites that were depending on the synchronous style sheet loading. I suspect things have changed though. For one Gecko no longer does the synchronous loading...

[05:42:00.0000] <hsivonen>
Hixie: because for regular scripts you don't need a nested event loop in order to be able to wait for the sheet

[05:42:01.0000] <Hixie>
without the event loop, the sheet can never load

[05:42:02.0000] <Hixie>
since the load happens on the event loop

[05:42:03.0000] <hsivonen>
Hixie: when you tested Gecko, did you make sure to document.write the <link rel=stylesheet> and the inline script together?

[05:42:04.0000] <Hixie>
no idea

[05:43:00.0000] <hsivonen>
if the <link rel=stylesheet> is in the network stream, you block whatever script would call document.write before you get to the document.write

[05:43:01.0000] <jgraham>
hsivonen: (you can do fancy yield and continuation in C++ if you design for it, of course)

[05:43:02.0000] <jgraham>
(I'm not suggesting that doesn't make it too high an implementation burden to require it)

[05:44:00.0000] <hsivonen>
jgraham: do you mean with arbitrary methods on the stack and without breaking the abstraction that the program isn't supposed to mess with its own stack as raw memory?

[05:45:00.0000] <Hixie>
you can do things with setjmp/longjmp

[05:45:01.0000] <Hixie>
but i wouldn't recommend it

[05:45:02.0000] <Hixie>
you can also simply write your code so that the continuation is a separate method

[05:46:00.0000] <Hixie>
it's not like the spinning is in an arbitrary place, it's in a well-defined place

[05:46:01.0000] <Hixie>
but in practice it doesn't matter much

[05:46:02.0000] <jgraham>
hsivonen: I am not familiar with the range of implementation approaches, I just know it is possible

[05:46:03.0000] <Hixie>
i was just saying the spec doesn't technically use nested event loops

[05:47:00.0000] <hsivonen>
Hixie: that's not much of a consolation

[05:47:01.0000] <Hixie>
i don't think anybody intended it to be a consolation :-)

[06:59:00.0000] <jgraham>
Hmm, is it generally true that the list of possibilities in the tree construction algorithm can be treated as unordered? It is at least untrue in the foreignContent case, but it seems like it would be nice if it was always true

[07:00:00.0000] <jgraham>
(apart from the "anything else" type clauses of course)

[07:06:00.0000] <hsivonen>
jgraham: IIRC, it's unordered

[07:07:00.0000] <jgraham>
Hmm, what am I missing about EOF handling in foreign content? It seems like it doesn't switch insertion modes but reprocesses the token

[07:07:01.0000] <jgraham>
So it will just pop everything off the stack of open elements

[07:09:00.0000] <jgraham>
hsivonen: It isn't in the case of 'A start tag whose tag name is one of: "b", "big", [...]' because that rule only applies if you don't match the rule 'A start tag, if the current node is an element in the HTML namespace.'

[07:10:00.0000] <hsivonen>
jgraham: oh ok.

[07:11:00.0000] <jgraham>
hsivonen: Any clues about what I am missing with the EOF thing?

[07:17:00.0000] <zcorpan_>
so what happened to binary arrays?

[07:17:01.0000] <hsivonen>
jgraham: I think you found a spec bug.

[07:18:00.0000] <hsivonen>
jgraham: before reprocessing the token, the tree builder should switch to the original mode

[07:18:01.0000] <hsivonen>
or at least that's what I have implemented

[07:19:00.0000] <jgraham>
hsivonen: For everything in that clause or just the EOF case?

[07:34:00.0000] <jgraham>
hsivonen: Filed a bug

[07:35:00.0000] <hsivonen>
jgraham: I meant the EOF case, but there may be other bugs

[07:35:01.0000] <jgraham>
I think the other case works either way

[07:36:00.0000] <jgraham>
But I may be wrong

[07:51:00.0000] <mikekelly>
hi fans

[07:52:00.0000] <mikekelly>
I've come for my bi-annual discussion about the type attribute

[07:53:00.0000] <Dashiva>
Please make sure you have new information before reopening old issues

[07:54:00.0000] <mikekelly>
I'm joking

[07:56:00.0000] <mikekelly>
why is it not possible for javascript to kick off a page transition and have control over the headers in the same as you do with XHR ?

[07:57:00.0000] <annevk>
so you were not joking?

[07:58:00.0000] <annevk>
yawn

[07:58:01.0000] <m_ali>
Hi

[07:59:00.0000] <m_ali>
i have a project that is using html5lib for sanitization of user input

[08:00:00.0000] <m_ali>
it is using the tokenizer and code like this:

[08:00:01.0000] <m_ali>
 p = html5lib.HTMLParser(tokenizer=HTMLSanitizer,

[08:00:02.0000] <m_ali>
                            tree=treebuilders.getTreeBuilder("dom"))

[08:00:03.0000] <m_ali>
    dom_tree = p.parseFragment(html)

[08:00:04.0000] <m_ali>
    walker = treewalkers.getTreeWalker("dom")

[08:00:05.0000] <m_ali>
    stream = walker(dom_tree)

[08:00:06.0000] <m_ali>
    s = serializer.HTMLSerializer(omit_optional_tags=False,

[08:00:07.0000] <m_ali>
                                  quote_attr_values=True)

[08:00:08.0000] <m_ali>
    output_generator = s.serialize(stream)

[08:00:09.0000] <m_ali>
    #logging.debug(''.join(output_generator))

[08:00:10.0000] <m_ali>
    return u''.join(output_generator)

[08:00:11.0000] <m_ali>
I am looking to allow the user to embed videos which are in <object> tags and the sanitization changes that to text.

[08:00:12.0000] <m_ali>
Any ideas what I need to do?

[08:01:00.0000] <zcorpan_>
if you allow <object>, there's no point in sanitizing at all

[08:02:00.0000] <mikekelly>
annevk: I was joking.. do you know the answer to that question?

[08:03:00.0000] <zcorpan_>
unless i guess you require data='' to be something that is trusted

[08:03:01.0000] <m_ali>
I can check the object tag that contains the allowed video sites

[08:06:00.0000] <TabAtkins_>
dammit, lachy.  why you gotta beat me to posting things?

[08:07:00.0000] <m_ali>
@zcorpan_: would I need to look at each token in the tokenizer or is there any other way?

[08:08:00.0000] <jgraham>
m_ali: YOu can allow object in the sanitizer. Restricting the allowed attribute values is harder

[08:09:00.0000] <jgraham>
There is a concept of filters for treewalkers so you might be able to use the built-in sanitizer as the first stage and then write something that implements the filter interface for your own custom sanitization

[08:09:01.0000] <jgraham>
Not really documented though so you will need to look at the code

[08:10:00.0000] <jgraham>
m_ali: (as an aside I guess using cElementTree or lxml is better than using DOM as the intermediate tree format, especially if you con't plan to manipulate the tree at all and so don't care about the API)

[08:12:00.0000] <jgraham>
(and you can do tree = html5lib.parse(input, treebulder="lxml") rather than constructing a intermediate HTMLParser object by hand)

[08:12:01.0000] <jgraham>
(I think parseFragment should work as well, if not file a bug)

[08:14:00.0000] <m_ali>
Can't see object as an accepted element here: http://wiki.whatwg.org/wiki/Sanitization_rules or am i missing something?

[08:15:00.0000] <m_ali>
Also, I don't understand why I need the two stage approach?

[08:15:01.0000] <m_ali>
*accepted=acceptable

[08:18:00.0000] <jgraham>
m_ali: You can customise what elements the sanitizer accepts by changing an attribute

[08:18:01.0000] <mikekelly>
did someone say accept?

[08:18:02.0000] <jgraham>
It is harder to add extra checks on allowed attribute values

[08:18:03.0000] <jgraham>
That's what I am suggesting the two-stage approach for

[08:19:00.0000] <jgraham>
(I vaugely considered redesigning the sanitizer as a series of plugabble filters, but apart from anything else it would kill perf)

[08:21:00.0000] <m_ali>
hmm... ok, so do you mean that the first stage sanitize would allow <object> and my custom sanitize would then only allow those <objects> that I want?

[08:24:00.0000] <jgraham>
m_ali: Yes

[08:25:00.0000] <zcorpan_>
http://www.google.se/search?hl=sv&safe=off&client=opera&hs=yuU&rls=en&tbs=prv%3A1&q=html5test&aq=f&aqi=&aql=&oq=&gs_rfai= - google scores 62 on html5test.com

[08:25:01.0000] <m_ali>
thanks jgraham

[08:27:00.0000] <jgraham>
zcorpan_: Hah!

[08:28:00.0000] <jgraham>
zcorpan_: We should get them to encode the passing tests in a barcode in the area presented by google so we can determine what they pass

[08:28:01.0000] <zcorpan_>
94 on beta.html5test.com

[08:29:00.0000] <zcorpan_>
'no bonus points' so no more than one video and one audio codec

[08:35:00.0000] <zcorpan_>
http://github.com/NielsLeenheer/html5test/issues#issue/41

[08:41:00.0000] <jgraham>
(also works with Acid 3)

[08:41:01.0000] <jgraham>
zcorpan_: Is it me or does your bug report attribute all lines to you? Not that it matters.

[08:43:00.0000] <Philip`>
Presumably Google is just rendering with some version of Chrome?

[08:43:01.0000] <jgraham>
Yeah, seems to be a reasonably old version though

[08:44:00.0000] <jgraham>
Is the fourth bucket on acid3 SVG? That doesn't seem to be working at all

[08:47:00.0000] <gsnedders>
SVG is a compile-time option for WebKit

[08:51:00.0000] <zcorpan_>
jgraham: github tampers with my text

[08:52:00.0000] <jgraham>
zcorpan_: OK

[09:31:00.0000] <gsnedders>
/me wonders what to do this evening

[09:31:01.0000] <gsnedders>
Doing something outside would be nice given the weather, but watching football seems, uh, annoying.

[09:33:00.0000] <jgraham>
gsnedders: Really it won't be that bad

[09:33:01.0000] <jgraham>
You can always bring a book

[09:33:02.0000] <jgraham>
and a mp3 player

[09:33:03.0000] <jgraham>
and headphones

[09:34:00.0000] <gsnedders>
Well, if I take my closed headphones, I can avoid all of the attempts to get me a girl.

[09:34:01.0000] <gsnedders>
But it doesn't really seem worthwhile going to a football match for that.

[09:34:02.0000] <Dashiva>
Get girls are a football match?

[09:35:00.0000] <Dashiva>
*at

[09:35:01.0000] <gsnedders>
Dashiva: Well, it's a women's football match, so presumably

[09:35:02.0000] <gsnedders>
/me doesn't have much experience of this

[09:35:03.0000] <Dashiva>
No field invasions

[09:35:04.0000] <jgraham>
gsnedders: We can make sure you don't sit with kilsmo

[09:35:05.0000] <gsnedders>
Dashiva: That's what my colleagues are probably planning for me.

[09:36:00.0000] <jgraham>
gsnedders: I'm really not, at least

[09:36:01.0000] <Dashiva>
If you do, make sure you streak

[09:36:02.0000] <jgraham>
I doubt Stina or Marlin are either 

[09:36:03.0000] <jgraham>
honestly

[09:36:04.0000] <jgraham>
/me has no idea who else, if anyone, is going

[09:36:05.0000] <gsnedders>
jgraham: Now I'm just realizing how little clue I have about going

[09:37:00.0000] <gsnedders>
jgraham: Lars I thought was

[09:37:01.0000] <gsnedders>
*who is going

[09:37:02.0000] <jgraham>
(actually I'm not sure about Stina or Marlin but I know they were at least encouraged to go)

[09:37:03.0000] <gsnedders>
Dashiva: Just make sure to inact the second rule of streaking?

[09:37:04.0000] <gsnedders>
jgraham: Half the office was, as far as I can tell :P

[09:38:00.0000] <Dashiva>
gsnedders: The first rule of football is, kick the ball

[09:38:01.0000] <gsnedders>
Gah. Sit indoors and try and catch up on the last month in my diary, or go and watch football

[09:38:02.0000] <gsnedders>
Neither seems particuarly great.

[09:38:03.0000] <gsnedders>
But the latter at least takes adavantage of the fact that Sweden has been weather than Scotland.

[09:39:00.0000] <gsnedders>
s/been/better/

[09:40:00.0000] <Dashiva>
Is it possible to not have better weather than Scotland?

[09:40:01.0000] <gsnedders>
Dashiva: The Polar regions.

[09:41:00.0000] <Dashiva>
The weather is less important when it's night half the year :)

[09:44:00.0000] <gsnedders>
jgraham: When are you going from the office, anyway?

[09:44:01.0000] <jgraham>
gsnedders: 18:30

[09:45:00.0000] <gsnedders>
/me might come, and will wear less black than earlier if he's to sit around in the sun

[12:31:00.0000] <hsivonen>
gsnedders: Did you see Avenue Q in London?

[13:06:00.0000] <gsnedders>
hsivonen: No, I never actually managed to see it as I, well, ended up with too much else going on with friends there

[15:10:00.0000] <AryehGregor>
Why are people complaining about the chairs' decisions, when the decision policy clearly states that the only recourse is filing a formal objection?

[15:11:00.0000] <TabAtkins>
Because they don't understand the process, or don't care about it and just want things to go their way.

[15:11:01.0000] <TabAtkins>
(I, on the other hand, am bitching about the process itself.)

[15:12:00.0000] <AryehGregor>
Who decided that it would be a good idea to allow random people off the Internet to file change proposals that the working group has to deal with?  Does any other WG work that way?

[15:12:01.0000] <othermaciej>
oh no, is there a giant flamewar?

[15:13:00.0000] <hober>
I bet Larry's new bug will go over really well with Sam...

[15:13:01.0000] <hober>
othermaciej: 

[15:13:02.0000] <hober>
err, http://www.w3.org/Bugs/Public/show_bug.cgi?id=9836

[15:13:03.0000] <TabAtkins>
heh, awesome.

[15:14:00.0000] <AryehGregor>
I'm pretty sure I remember the chairs explicitly saying something to the effect of "only strongly-held objections will be considered".

[15:17:00.0000] <Rik`>
This guy is paid by Adobe to contribute to HTML5 ? really ?

[15:18:00.0000] <othermaciej>
the exact text (still visible on the poll results page) is: "If you have strong objections to adopting this Change Proposal, please state your objections below."

[15:19:00.0000] <othermaciej>
I must admit that a response starting with "I'm not *strongly* opposed" sounds like the author is declaring it not to be a strong objection to me

[15:19:01.0000] <AryehGregor>
That does sound like a relatively reasonable inference.

[15:20:00.0000] <othermaciej>
and while certainly the chairs or WG could second guess them, it seems very unlikely something could be a strong objection if the person raising it says it isn't, and second guessing in such circumstances would amount to a fishing expedition

[15:22:00.0000] <mbrubeck>
Note that the reply read (my emphasis) "I'm not strongly opposed to the *concepts*.... *but*...."

[15:25:00.0000] <Hixie>
could someone explain to me why microdata and a suggestion to make UAs more accessible get split out, but <figure> and <aside> do not?

[15:25:01.0000] <Hixie>
i've been reading and rereading the respective "working group decisions" but can't for the life of me work out a consistent policy

[15:26:00.0000] <Hixie>
(that is assuming i'm even understanding the decisions correctly, which is hard given how utterly vague and wishy-washy they are)

[15:26:01.0000] <annevk>
they seem to be primarily based on the comments the change proposals receive

[15:27:00.0000] <annevk>
the one with the weakest comments against, wins

[15:27:01.0000] <Hixie>
so there's no consistent policy for me to apply?

[15:27:02.0000] <AryehGregor>
Sure there is, object very strongly to everything.

[15:27:03.0000] <AryehGregor>
I mean, any changes.

[15:27:04.0000] <Hixie>
no i mean as editor

[15:28:00.0000] <Hixie>
there's no policy that i can apply as editor before things get objected to

[15:28:01.0000] <annevk>
it depends on how well you can predict what the WG members are up to

[15:29:00.0000] <Hixie>
i can't even predict what the chairs are up to

[15:29:01.0000] <AryehGregor>
Well . . . make sure that your opinion happens to agree with the chairs'?  I mean, basically what we've got is that the chairs can overrule you if people convince them that you're wrong.

[15:29:02.0000] <annevk>
clearly you need more data :p

[15:29:03.0000] <AryehGregor>
So unless you can read their minds, or pay them off or something, I don't that you'd be able to avoid ever being overruled.

[15:29:04.0000] <Hixie>
AryehGregor: the chairs' opinions don't agree with the chairs'

[15:29:05.0000] <AryehGregor>
That's a problem with most people.

[15:30:00.0000] <TabAtkins>
AryehGregor: Not "people convince them that you're wrong".  It's "people convince them that they'll object more strenuously than people on the other side".

[15:30:01.0000] <TabAtkins>
It is *literally* a race to exhaustion.

[15:30:02.0000] <annevk>
he is not worried about being overruled, he is interested in applying rationale for approved change proposals to the rest of the specification

[15:30:03.0000] <AryehGregor>
Why?

[15:31:00.0000] <AryehGregor>
It seems like it would be easier to do what he thinks is best and then fix it if he's asked to.

[15:31:01.0000] <AryehGregor>
Given that there are no clear criteria being applied.

[15:31:02.0000] <Hixie>
i am not going to have my name on a spec that is internally inconsistent

[15:32:00.0000] <AryehGregor>
TabAtkins, no, they're supposed to be judging the technical merits of objections.  I don't win if I say "I HATE THIS FEATURE AND WILL COMPLAIN TO THE DIRECTOR AND PICKET OUTSIDE W3C OFFICES".  I have to convince them that my reasons are sound.

[15:32:01.0000] <Hixie>
if we're going to do something different than what i think is logical, that's fine, i just need to know what the logic is so that i can apply it consistently and make the spec internally consistent

[15:33:00.0000] <AryehGregor>
Hixie, are you suggesting that any form of HTML is internally consistent?  Surely it's less inconsistent to change a few minor things because the chairs disagree with you, than to use entire horribly-designed APIs because IE happens to support them already.

[15:33:01.0000] <Hixie>
i don't mean the technology being defined, i mean the spec defining it

[15:33:02.0000] <TabAtkins>
AryehGregor: No, that's what you *want* to happen.  That is definitely not what Sam specifically said they actually judge on.

[15:34:00.0000] <AryehGregor>
TabAtkins, I explicitly asked during microdata and was told something like what I just said.

[15:34:01.0000] <annevk>
while you guys figure this out I'm gonna read my next book

[15:34:02.0000] <Hixie>
i've been told they use the value of arguments too

[15:34:03.0000] <Hixie>
though clearly they don't

[15:35:00.0000] <TabAtkins>
AryehGregor: And now he's saying something different.

[15:35:01.0000] <othermaciej>
it seems like inclusion/exclusion is (a) a judgment call based on many factors and (b) relevant only to the snapshot of the language  that will be labeled "HTML5" by the W3C (therefore some of the factors being the situation at the present moment), and therefore I am not sure what it would mean to have "internal consistency" on the set of things included

[15:36:00.0000] <Hixie>
othermaciej: why is <aside> in but microdata out? the arguments to both seem to apply equally

[15:36:01.0000] <AryehGregor>
Consistency would be an issue if the chairs were dictating exact wording, but saying a feature or paragraph here or there should be removed doesn't seem like it meaningfully increases inconsistency.

[15:36:02.0000] <TabAtkins>
The change proposal part of the process turns out to be a farce, unfortunately.  Even when I "win" I'm angry at the results.  I think I just won't contribute in the future.

[15:37:00.0000] <Hixie>
AryehGregor: why is one paragraph of advice in but another not?

[15:37:01.0000] <AryehGregor>
Hixie, the major reason for leaving out Microdata was that it should compete with RDFa on an equal footing, wasn't it?

[15:37:02.0000] <mbrubeck>
Hixie: That's an inconsistency in process - it doesn't necessarily lead to an inconsistent spec.

[15:37:03.0000] <mbrubeck>
(any more than a consistent process would necessarily lead to a consistent spec)

[15:38:00.0000] <othermaciej>
I don't want to get too much into it, since I'm technically on vacation still, but Tab, I am curious what you find upsetting about the results and I'll ask you about it later maybe

[15:38:01.0000] <Hixie>
AryehGregor: shouldn't <aside> compete on an equal footing with other proposals?

[15:38:02.0000] <TabAtkins>
othermaciej: Sure, just ping me whenever.

[15:39:00.0000] <Hixie>
TabAtkins: i feel similarly, although in my guess i'm not angry at the results, i am just baffled

[15:39:01.0000] <AryehGregor>
Hixie, there are no other proposals that are written up in specs and actively edited, are there?  You can't say it should compete on an equal footing with something that's not in any spec and won't be put in one anytime soon.

[15:40:00.0000] <Hixie>
AryehGregor: so things that have active specs should be split out? i could apply that reasoning, if that's the reasoning

[15:40:01.0000] <AryehGregor>
It seemed fairly essential in the RDFa case.  But the chairs have declined to decide anything on a general basis, only case-by-case.

[15:40:02.0000] <Hixie>
AryehGregor: i'd have to take out the event handler attributes (compete with XML Events), <a> and <link> (competes with XLink), rel="" (competes with RDFa and Microdata)...

[15:40:03.0000] <Hixie>
AryehGregor: i don

[15:40:04.0000] <AryehGregor>
So maybe you should ask for a change in the decision policy.

[15:40:05.0000] <Hixie>
er

[15:41:00.0000] <Hixie>
AryehGregor: i don't mind them deciding on a case-by-case basis so long as they apply their reasoning consistently to everything in the spec

[15:41:01.0000] <Hixie>
but they don't

[15:41:02.0000] <TabAtkins>
AryehGregor: Lief's <object> proposal is a competing idea.  ^_^

[15:43:00.0000] <jgraham>
"""The decision about which objections are strong or are not strong should be

[15:43:01.0000] <jgraham>
working group consensus based"""

[15:43:02.0000] <jgraham>
We should totally have a poll for each objection

[15:44:00.0000] <Hixie>
consensus is a dumb way to do language design

[15:44:01.0000] <othermaciej>
jgraham: that suggestion reminds me of this: http://volokh.com/2010/06/01/sabotage-or-how-dilbert-won-the-war/

[15:46:00.0000] <jgraham>
Yes, I think it is rather clear Larry is seeking to slow down progress

[15:48:00.0000] <AryehGregor>
It doesn't seem that way to me.  His actions seem explainable by good faith.

[15:48:01.0000] <othermaciej>
To be clear, I meant your comical suggestion ("have a poll for each objection"), not Larry's

[15:48:02.0000] <AryehGregor>
I tend to always assume good faith.  It works pretty well.

[15:48:03.0000] <jgraham>
AryehGregor: It also means that you are wrong most of the time

[15:48:04.0000] <jgraham>
Hopefully that doesn't matter

[15:48:05.0000] <AryehGregor>
jgraham, I beg to differ.

[15:49:00.0000] <jgraham>
Well I can't really demonstrate "most"

[15:49:01.0000] <othermaciej>
I am not sure what Larry's motivations are, and ideally I can do my job without having to make such a determination

[15:49:02.0000] <mbrubeck>
I'm really confused about Larry's "objection" - is it an objection to the change proposal at all, or is it really a statement of support for the change proposal?

[15:49:03.0000] <jgraham>
So "a non-negligible fraction"

[15:49:04.0000] <othermaciej>
I think most people are acting in what they themselves consider to be good faith most of the time

[15:49:05.0000] <mbrubeck>
oh, I see

[15:50:00.0000] <mbrubeck>
It objects to the "Change Proposal to Keep" not the "Change Proposal to Remove"

[15:50:01.0000] <jgraham>
FWIW many people consider slowing down progress to be a reasonable goal

[15:50:02.0000] <jgraham>
for my definition of progress

[15:50:03.0000] <othermaciej>
sometimes people behave in ways that are consciously insincere, but that is the exception

[15:52:00.0000] <othermaciej>
If anyone really had a conscious goal of "delay progress on HTML5", then participating in the HTML WG on that basis would arguably be bad faith

[15:52:01.0000] <jgraham>
(fwiw I don't see how you could say "ecision about which objections are strong [...] should be consensus based" and not realise that this would have the effect of slowing things down)

[15:53:00.0000] <jgraham>
(so it is at least true that he must believe that slower progress is acceptable)

[15:53:01.0000] <jgraham>
(although he could conceviably believe that it would lead to faster progress at a later time for some reason I don't see)

[15:53:02.0000] <mbrubeck>
He might believe both that it would slow the process and that the slower process would lead to a better ultimate outcome.  That wouldn't be bad faith.

[15:54:00.0000] <mbrubeck>
Change is not always the same as "progress" - faster change does not magically make things better

[15:54:01.0000] <othermaciej>
on the other hand, seeking to improve things along dimensions that others do not care much about is good faith behavior

[15:54:02.0000] <jgraham>
Yes, I don't claim that he is necessarily acting in bad faith

[15:54:03.0000] <jgraham>
I really have no idea what his motivations are

[15:55:00.0000] <jgraham>
(so it makes little sense to speculate about good faith/bad faith)

[15:55:01.0000] <othermaciej>
even if it can be hard to distinguish at times from deliberate delay

[15:56:00.0000] <jgraham>
Anyway, I seem to have accidentially created a tangent from the more relevent point that the Process is driving good people away from the WG

[16:00:00.0000] <TabAtkins>
And thus it needs to be destroyed.

[16:00:01.0000] <AryehGregor>
The Process, the WG, or the W3C?  :)

[16:02:00.0000] <othermaciej>
if the process does drive good people away then we may have to revise it

[16:02:01.0000] <othermaciej>
I don't know what concerns have been raised of late but I'll certainly look into them when back stateside

[16:02:02.0000] <TabAtkins>
DESTROY THEM ALL

[16:04:00.0000] <crankharder>
hey all, i've been playing around with cache manifests today and am kinda unsure about somethin

[16:04:01.0000] <AryehGregor>
The problem with the HTMLWG is that, unlike the WHATWG, it operates under the conceit that everyone's voice needs to be heard and it's not acceptable to just dismiss people's objections out of hand.  This inevitably leads to people raising all sorts of issues that no one else cares much about, because others are forced to deal with them.

[16:05:00.0000] <AryehGregor>
It doesn't seem like the CSSWG operates this way, so I don't see why the HTMLWG should have to.

[16:05:01.0000] <AryehGregor>
As far as I can tell, in the CSSWG, random people from the Internet have no right to raise objections that anyone has to pay attention to.

[16:05:02.0000] <othermaciej>
it's harder to become a CSS WG member

[16:05:03.0000] <crankharder>
is there a way to specify a wildcard URL for the CACHE directive?

[16:06:00.0000] <othermaciej>
however, the CSS WG does have to at least respond to all Last Call comments in some meaningful way, and those could come from random people from the internet

[16:06:01.0000] <AryehGregor>
Yes, that's the problem.  Everyone is theoretically equal in the HTMLWG, more or less, and that inevitably requires mounds of procedure and bickering as everyone has to have their say.

[16:06:02.0000] <mbrubeck>
crankharder: What CACHE directive?

[16:06:03.0000] <AryehGregor>
They could be responded to by one group member in a few sentences.  You don't need to write lengthy counterproposals with several required sections and have a survey and so on.

[16:07:00.0000] <crankharder>
mbrubeck: in a cache manifest file

[16:07:01.0000] <Hixie>
crankharder: you mean matching everything?

[16:07:02.0000] <TabAtkins>
othermaciej: Harder to become a member, but not to talk on the list.

[16:07:03.0000] <mbrubeck>
crankharder: No, no wildcard.

[16:07:04.0000] <crankharder>
the NETWORK directive allows a wildcard, but not the CACHE

[16:07:05.0000] <crankharder>
NETWORK urls must only start with what is defined, afaik the CACHE ones must match up perfectly

[16:07:06.0000] <Hixie>
crankharder: the URLs in the CACHE section are fetched one by one by the browser, so i don't see how the CACHE could have a wildcard

[16:08:00.0000] <othermaciej>
TabAtkins: true, but random people talking on the list do not have to be taken as seriously as CSS WG members, since they do not have a direct say in what WG resolutions pass

[16:08:01.0000] <Hixie>
crankharder: the browser can't just make up the URLs :-)

[16:08:02.0000] <crankharder>
hrm

[16:08:03.0000] <TabAtkins>
othermaciej: Indeed.  And that's somewhat more useful.  I was a useful and productive contributor to the CSSWG long before I become a member.

[16:09:00.0000] <TabAtkins>
But had I turned out crazy, shrug.  I could just have been killfiled and that's that.

[16:09:01.0000] <AryehGregor>
The WHATWG and CSSWG both rely on a small group of qualified people making all decisions, and merely require them to address public comments (however perfunctorily) according to no particular procedure.

[16:09:02.0000] <AryehGregor>
(one person, in the WHATWG's case)

[16:10:00.0000] <jgraham>
I'm not sure the CSS WG is the best example to follow

[16:10:01.0000] <othermaciej>
well, one could certainly look at whether the CSS WG is on the whole a more effective organization than the HTML WG

[16:10:02.0000] <TabAtkins>
Hmm, for some reason my python script is now outputting things in a bad encoding.  It didn't do that a second ago, though I had basically equivalent characters.

[16:10:03.0000] <crankharder>
so the issue i'm having boils down to my framework (rails) appending timestamps to css/js/image files -- these timestamps are used for far-future expiration caching by the web server - it's therefore impossible to add the correct css/js/image filename to the manifest since the full name is dynamically generated

[16:10:04.0000] <AryehGregor>
It's the only W3C WG I've followed other than the HTMLWG.

[16:10:05.0000] <AryehGregor>
That's the only reason I use it for examples.

[16:10:06.0000] <jgraham>
It's not like it is known for being hyper efficient

[16:10:07.0000] <othermaciej>
I think the general consensus would be that it's not

[16:11:00.0000] <mbrubeck>
crankharder: Can you generate the manifest dynamically with the same timestamps?

[16:11:01.0000] <AryehGregor>
Perhaps it's also too big.

[16:11:02.0000] <AryehGregor>
It's hard to argue with the WHATWG's efficiency, and that seems to stem from the fact that only one person is responsible for making all decisions.

[16:11:03.0000] <crankharder>
mbrubeck: probably, but I was hoping for the wildcard solution ;)

[16:11:04.0000] <TabAtkins>
othermaciej: Of course, one must separate the problem of editors letting specs languish, and process making specs languish.  The CSSWG suffers from the former, but the latter is what we're complaining about here.

[16:12:00.0000] <jgraham>
It's not clear that the CSSWG problems are due to its size

[16:12:01.0000] <mbrubeck>
crankharder: In fact, you should be able to generate the manifest as a static file at deployment time with appropriate timestamps - that's basically what I did for an offline app project.

[16:12:02.0000] <jgraham>
Although they could be I guess

[16:12:03.0000] <othermaciej>
in the W3C, in theory the buck stops at the Director, however he does not make any of the front-line decisions

[16:13:00.0000] <othermaciej>
I don't think there is evidence that HTML5 is languishing; I believe the complaints cited were about bad decision process and/or incensistent basis for decisions and/or resulting inconsistency in the spec

[16:13:01.0000] <othermaciej>
not about the spec failing to progress

[16:13:02.0000] <jgraham>
othermaciej: From a real world point of view it is doing fine

[16:14:00.0000] <jgraham>
othermaciej: From a W3C process POV it is nearing the end of Q2 2010 and we don't seem to be much closer to LC

[16:14:01.0000] <TabAtkins>
othermaciej: Bad choice of words on my part.  Indeed, the problem is not anything languishing.

[16:14:02.0000] <jgraham>
Which was supposed to happen in Nov 1009

[16:14:03.0000] <jgraham>
*2009

[16:14:04.0000] <mbrubeck>
It's 1001 years late!

[16:14:05.0000] <jgraham>
(we're not quite that late...)

[16:14:06.0000] <othermaciej>
jgraham: I do wish we were further along than we are, but I also believe we are in the endgame of moving towards LC

[16:15:00.0000] <Hixie>
html5 is doing fine so far, the concern as i see it is that a decisions process is being applied to the spec that has only avoided causing massive damage so far by luck

[16:15:01.0000] <Hixie>
and by my interpreting the decisions in the least damaging way each time

[16:16:00.0000] <othermaciej>
I expect at least Microdata will get to LC pretty soon (I plan to make it a priority to work on that when I get back from vacation)

[16:16:01.0000] <mbrubeck>
Hixie: Or the process is inconsistent because it involves judgment and takes inherently unpredictable human factors into account, and can't be replaced by an algorithm as you seem to wish.

[16:16:02.0000] <jgraham>
othermaciej: The process that is taking us there seems to be causing attrition of people willing to contribute

[16:16:03.0000] <othermaciej>
and once we get one draft to LC I expect the path will be smoother for subsequent ones

[16:16:04.0000] <Hixie>
mbrubeck: judgement is fine, so long as it is consistently applied

[16:16:05.0000] <Hixie>
mbrubeck: (judgement is all that's ever been used)

[16:17:00.0000] <othermaciej>
jgraham: that's certainly something I would be concerned about if that is in fact the case

[16:18:00.0000] <mbrubeck>
Hixie: I mean, above someone suggested that microdata was split out because it has a separate spec, and you said that to be "consistent" with that "rule" you would also have to split out <a> and @rel.

[16:19:00.0000] <mbrubeck>
but it seems obvious that instead of a rule ("do not compete with any other specs, ever") there has to be a judgment call (which other specs are likely to matter / compete / be used?).

[16:19:01.0000] <Hixie>
mbrubeck: so RDFa is not likely to matter / compete / be used?

[16:19:02.0000] <mbrubeck>
NO

[16:19:03.0000] <Hixie>
rel="" and itemprop="" both compete with RDFa

[16:19:04.0000] <Hixie>
heck, RDFa even _uses_ rel=""

[16:20:00.0000] <mbrubeck>
You're still trying to turn my example of *factors to be considered* into some sort of binary rule that will always produce the same outcome.

[16:20:01.0000] <Hixie>
(in an incompatible manner, but that's another story)

[16:20:02.0000] <Hixie>
yes

[16:20:03.0000] <Hixie>
well not necessarily binary

[16:20:04.0000] <Hixie>
i'm trying to apply a rule that can predict wg decision outcomes

[16:20:05.0000] <mbrubeck>
And I think that's the wrong problem to try to solve.

[16:20:06.0000] <Hixie>
so that i can just make the spec what the wg wants without us having to waste months going through this insanely heavy-weight process for each paragraph

[16:21:00.0000] <mbrubeck>
heh.

[16:21:01.0000] <Hixie>
(and yes, we're literally doing it at the paragraph level at the moment)

[16:23:00.0000] <TabAtkins>
Yay, encoding issue was an encoding issue!  (That is, there was no problem, the file just thought it was the wrong encoding and so was displaying incorrectly.)

[16:53:00.0000] <MikeSmith>
Hixie: in the W3C copy of the spec, can you please change the "This specification is available in the following formats: single page HTML, multipage HTML." to "...available in the following formats: single page HTML, multipage HTML, Author Edition."

[16:54:00.0000] <MikeSmith>
...with the Author Edition text being a link to "author/" subdir


2010-06-03
[19:05:00.0000] <Hixie>
MikeSmith: done

[19:06:00.0000] <MikeSmith>
thanks

[19:08:00.0000] <MikeSmith>
I'll get the author-edition published in TR space along with the multipage

[19:08:01.0000] <MikeSmith>
for next round of WDs

[21:58:00.0000] <karlcow>
looking at "Proposing New Features" at the bottom of this page http://www.w3.org/MarkUp/HTML-WG/

[22:19:00.0000] <MikeSmithX>
"We value your opinion, really we do, but we're going to put a stiff tax on its expression, so that we won't have to hear it very often."

[22:23:00.0000] <MikeSmith>
+1

[00:01:00.0000] <hsivonen>
gsnedders: OK. I was hoping to get advice on which seats I should get...

[00:15:00.0000] <hsivonen>
AryehGregor: It's a bit of a myth that Hixie decides everything alone. Here's an example of influencing the decision process: http://krijnhoetmer.nl/irc-logs/whatwg/20090212#l-27

[00:18:00.0000] <hsivonen>
MikeSmith: maybe the Author Edition should be called something like an Edition for Web Authors

[00:18:01.0000] <hsivonen>
MikeSmith: otherwise it looks like the Editor's Cut

[00:21:00.0000] <hsivonen>
I still have trouble deciding if I should get the Avenue Q tickets from the Stalls or the Royal Circle (the first balcony level)...

[00:22:00.0000] <MikeSmith>
hsivonen: good point

[00:44:00.0000] <MikeSmith>
hsivonen: changed now to "Edition for Web Authors"

[00:44:01.0000] <MikeSmith>
http://dev.w3.org/html5/spec-author-view/

[00:45:00.0000] <hsivonen>
MikeSmith: thanks

[00:45:01.0000] <MikeSmith>
hsivonen: btw, did you see my question yesterday about how to make the v.nu HTML5 parser usable by other Java apps from the command line?

[00:45:02.0000] <MikeSmith>
/me goes to get pointer to logs

[00:46:00.0000] <MikeSmith>
http://krijnhoetmer.nl/irc-logs/whatwg/20100602#l-181

[01:54:00.0000] <hsivonen>
http://html4all.org/mailman/archives/list_html4all.org/2010-June/001066.html

[02:01:00.0000] <zcorpan_>
hsivonen: surely the reason is that Hixie wrote the text instead of the Experts

[02:04:00.0000] <Dashiva>
zcorpan_: A mutation of NIH?

[02:07:00.0000] <jgraham>
hsivonen: You are going to London?

[02:08:00.0000] <annevk>
bleh

[02:09:00.0000] <hsivonen>
jgraham: yes. In July.

[02:09:01.0000] <jgraham>
hsivonen: Nice. Holiday?

[02:10:00.0000] <hsivonen>
jgraham: yeah, a short one, since I'd have a layover at LHR anyway

[02:10:01.0000] <annevk>
http://blog.chromium.org/2010/06/websocket-protocol-updated.html

[02:11:00.0000] <hsivonen>
I booked tickets for Avenue Q. gsnedders is effective at this social media marketing thing.

[02:11:01.0000] <jgraham>
What you mean he convinced you to go without ever actually having gone himself or having any idea of product quality?

[02:12:00.0000] <hsivonen>
jgraham: right. very effective indeed.

[02:13:00.0000] <annevk>
haha

[02:14:00.0000] <hsivonen>
great to see that Chromium is tracking the spec changes for Web Socket

[02:19:00.0000] <hsivonen>
working with hg on a hard disk feels very slow after getting used to an SSD

[02:26:00.0000] <annevk>
http://twitter.com/mikeplate/statuses/15247455331

[02:26:01.0000] <annevk>
'"The most unreadable specification I've read in 20 years of standardization" Daniel Glazman on WhatWG #html5 doc spec #swdc'

[02:27:00.0000] <hsivonen>
/me wonders if the other specs that Daniel has read have been as precise

[02:27:01.0000] <Dashiva>
/me wonders if he has read any other specs at all

[02:28:00.0000] <jgraham>
/me wonders if the other specs will lead to such interoperable behaviour in such a complex area

[02:30:00.0000] <annevk>
/me wonders how carefully he read CSS 2.1 :-P

[02:31:00.0000] <workmad3>
annevk: surely you're not suggesting that the CSS 2.1 spec was unreadable! :)

[02:31:01.0000] <annevk>
I'm suggesting it is pretty hard to read

[02:32:00.0000] <annevk>
And incomplete unfortunately on certain important topics...

[02:32:01.0000] <Dashiva>
I can never remember where to find half the layout properties

[02:33:00.0000] <jgraham>
I think my serious response would be http://dbaron.org/log/20100531-specs

[03:14:00.0000] <annevk>
someone emailed me again the other day about href on any element... gave an explanation plus pointer to the FAQ entry

[03:18:00.0000] <aho>
easy way: you should have just pointed out that it's now fine to put whatever you want into a-elements :>

[03:22:00.0000] <annevk>
he was actually aware of that

[03:25:00.0000] <mikekelly>
I'm interested in understand why it isn't possible for javascript to kick off a page transition and have granular control over the request in the same as you do with XHR - anyone know?

[03:27:00.0000] <zcorpan_>
mikekelly: because there's only one person in the world who wants to do that

[03:35:00.0000] <gsnedders>
hsivonen: And Hixie was good at convincing me I should go (though mainly F2F, so no IRC logs of this)

[03:38:00.0000] <zcorpan_>
jgraham: who's marlin? did you mean malin?

[03:39:00.0000] <jgraham>
zcorpan_: Probably

[03:39:01.0000] <Dashiva>
He's implying she's fishy

[03:40:00.0000] <jgraham>
I consider it a good day if I manage to spell my own name right

[03:51:00.0000] <annevk>
jgraham, low standards much?

[03:52:00.0000] <gsnedders>
jgraham: Your spelling it better than mine. :P

[03:52:01.0000] <annevk>
though I guess it's a good thing it doesn't depend on externalities, like someone else spelling your name right for a change...

[03:52:02.0000] <gsnedders>
(There again, that is a really low standard.)

[03:52:03.0000] <jgraham>
annevk: I like to use the word "realistic"

[05:30:00.0000] <hsivonen>
Does anyone happen to have an URL for a style sheet that's guarenteed to load artificially slowly?

[05:30:01.0000] <Rik`>
hsivonen: steve souders should have that with cuzillion

[05:33:00.0000] <jgraham>
hsivonen: I guess OPera internal urls aren't that useful to you ;)

[05:33:01.0000] <Rik`>
hsivonen: http://stevesouders.com/cuzillion/

[05:34:00.0000] <jgraham>
Rik`: Neat

[05:36:00.0000] <hsivonen>
Rik`: thanks

[05:36:01.0000] <Rik`>
yeah, pretty useful

[06:09:00.0000] <hsivonen>
what's a cross-browser (incl. IE) way to ask an element for its color?

[06:09:01.0000] <Dashiva>
None that I know of, you have to combine getComputedStyle and currentStyle

[06:10:00.0000] <Rik`>
and don't forget the please argument

[06:11:00.0000] <Dashiva>
brb, implementing intercal support in webkit

[06:12:00.0000] <hsivonen>
ok. maybe I'll try to see if Opera and WebKit support offsetHeight

[06:12:01.0000] <Dashiva>
You can't just detect getComputedStyle and fallback to currentStyle?

[06:13:00.0000] <hsivonen>
Dashiva: I just want to check if a test style sheet is applied at a given point

[06:13:01.0000] <hsivonen>
I don't need to check arbitrary styles.

[06:30:00.0000] <zcorpan_>
offsetHeight should work

[06:34:00.0000] <hsivonen>
as I expected, none of the top 4 engines behaves like the spec says when a document.written internal script occurs right after an external style sheet in the same document.write

[06:35:00.0000] <hsivonen>
http://hsivonen.iki.fi/test/moz/sheet-blocking-script.html

[06:35:01.0000] <hsivonen>
time to file a spec bug

[06:36:00.0000] <hsivonen>
it took a lot of trying to come up with a test that explained the behavior of all 4

[06:54:00.0000] <Dashiva>
And there are _still_ people who object to the design principles, wonderful

[07:02:00.0000] <hsivonen>
whew. http://www.w3.org/Bugs/Public/show_bug.cgi?id=9843 filed.

[07:23:00.0000] <hsivonen>
is there a service that lets me search Hixie's commit messages?

[07:23:01.0000] <hsivonen>
or browse blame for the spec?

[07:24:00.0000] <Dashiva>
If there isn't, you could search the archives for commit-watchers

[07:32:00.0000] <jgraham>
hsivonen: I assume you don't count the avaliability of svn for checkout as "a service"

[07:33:00.0000] <hsivonen>
jgraham: isn't it slow? and no Web UI?

[07:34:00.0000] <hsivonen>
/me needs to find the SVG URL

[07:34:01.0000] <hsivonen>
/me guesses correctly

[07:38:00.0000] <jgraham>
hsivonen: Yes

[07:38:01.0000] <hsivonen>
OK, now I need to find the list discussion that motivated http://html5.org/tools/web-apps-tracker?from=13&to=14

[07:47:00.0000] <hsivonen>
I was unable to find a public discussion leading to that change.

[07:48:00.0000] <gsnedders>
r14!? Man, that's old sk00l.

[09:30:00.0000] <crankharder>
this manifest should hit /offline/update_all_events regardless of cache/network status right?

[09:30:01.0000] <crankharder>
http://pastie.org/990656

[10:34:00.0000] <AryehGregor>
Hmm.  <input required> in Chrome dev channel seemingly doesn't submit the form, but gives no error.

[10:34:01.0000] <AryehGregor>
It just focuses the password field.

[10:35:00.0000] <AryehGregor>
I mean, the required field.

[10:35:01.0000] <remysharp>
AryehGregor: it's flagged as invalid on the dom node - you have to go all JS on the element

[10:35:02.0000] <AryehGregor>
That seems like a really bad feature to deploy.  If you don't have UI, you shouldn't enforce the requirements either.

[10:36:00.0000] <remysharp>
AryehGregor: completely agree

[10:36:01.0000] <AryehGregor>
The form validation API is not meant to be JS-only.  It's meant to work even if the author provides no JS.

[10:37:00.0000] <AryehGregor>
Are there minimal UI requirements in the spec here?

[10:37:01.0000] <AryehGregor>
/me is looking

[10:37:02.0000] <gsnedders>
No

[10:37:03.0000] <gsnedders>
/me would hope there aren't any UI requirements

[10:37:04.0000] <AryehGregor>
Not even "you have to tell the user and not just silently fail"?

[10:37:05.0000] <AryehGregor>
That seems like a pretty reasonable UI requirement to me.

[10:37:06.0000] <gsnedders>
What if it's Googlebot?

[10:38:00.0000] <AryehGregor>
I'm talking about browsers here.  Googlebot is not a browser.

[10:39:00.0000] <AryehGregor>
A Googlebot does not fall into the "Web browsers and other interactive user agents" category.

[10:39:01.0000] <AryehGregor>
It's "Data mining tools", I suppose.

[10:39:02.0000] <AryehGregor>
So it only has to follow the semantic requirements, not UI requirements.

[10:43:00.0000] <AryehGregor>
http://www.whatwg.org/specs/web-apps/current-work/multipage/association-of-controls-and-forms.html#constraint-validation "Report the problems with the constraints of at least one of the elements given in unhandled invalid controls to the user."

[10:43:01.0000] <AryehGregor>
It explicitly suggests focusing the element, but that's just not enough.  There needs to be some text somehow.

[10:45:00.0000] <AryehGregor>
https://bugs.webkit.org/show_bug.cgi?id=34945

[11:36:00.0000] <zcorpan_>
hmm, so <link rel=next> is being spread on twitter as being 'html5 prefetch'

[11:37:00.0000] <zcorpan_>
http://keyboardy.com/programming/html5-link-prefetching/

[11:38:00.0000] <AryehGregor>
Didn't it way predate -- oh, never mind, buzzword.

[11:45:00.0000] <zcorpan_>
/me comments

[11:48:00.0000] <TabAtkins>
Haha, I love <canvas>+<video> *so* much - http://betaweb.csug.rochester.edu/~jparish/public/canvas/video.html

[11:49:00.0000] <miketaylr>
that's awesome

[11:50:00.0000] <TabAtkins>
I think it would be even more awesome if it didn't draw into the canvas, but just used it for calculation, and instead did the display in real text.

[11:50:01.0000] <miketaylr>
dhtml 4 life

[11:55:00.0000] <miketaylr>
somewhat unrelated, but did you see kangax's svg -> canvas parser thingy? http://yura.thinkweb2.com/canvas_demo/

[12:00:00.0000] <TabAtkins>
No, I had not previously seen that!

[12:00:01.0000] <TabAtkins>
Random note - anyone know who the dude in that image is?

[12:01:00.0000] <miketaylr>
the tongue sticking out guy?

[12:01:01.0000] <TabAtkins>
Yeah.

[12:01:02.0000] <miketaylr>
ha. no clue. i can ping kangax on IM and ask though

[12:02:00.0000] <TabAtkins>
kk.  I'm just wondering because it looks sorta like jimmy urine.

[12:04:00.0000] <miketaylr>
quoth the kangax, "just some svg i found on the web"

[12:04:01.0000] <TabAtkins>
Kk.

[12:19:00.0000] <AryehGregor>
Awesome, Ubuntu versions are now tracking latest Firefox instead of freezing on a stable branch forever.

[12:23:00.0000] <gratz|home>
hm

[12:23:01.0000] <gratz|home>
AryehGregor, link to info?

[12:24:00.0000] <AryehGregor>
gratz|home, http://fridge.ubuntu.com/node/2051

[12:26:00.0000] <gratz|home>
cool

[12:26:01.0000] <gratz|home>
was a bit surprised since they freeze everything for standardisation / support

[12:26:02.0000] <gratz|home>
good to see though :D

[12:28:00.0000] <gratz|home>
brb

[12:43:00.0000] <TabAtkins>
AryehGregor: You around?  Question about Hebrew numbering.

[12:43:01.0000] <AryehGregor>
TabAtkins, I'm here.

[12:43:02.0000] <TabAtkins>
Do you know how accurate the current css3-lists description of hebrew numbering is?  It looks like it's pretty easy to describe with a simple greedy additive algorithm.

[12:44:00.0000] <AryehGregor>
I commented about that some time ago, didn't I?

[12:44:01.0000] <TabAtkins>
Yeah, but I can't follow that discussion well enough to be sure.

[12:44:02.0000] <AryehGregor>
The conclusion was that it should just stop at 999,999.

[12:44:03.0000] <TabAtkins>
Okay, that's fine.

[12:44:04.0000] <AryehGregor>
If you do that, then it's very simple.

[12:44:05.0000] <TabAtkins>
And the appears if and only if the number is 1k or greater?

[12:45:00.0000] <TabAtkins>
the geresh

[12:45:01.0000] <AryehGregor>
Well, there are still a couple weird things, actually.

[12:45:02.0000] <TabAtkins>
Well, plus the 15/16 exceptions.

[12:45:03.0000] <AryehGregor>
I'd ditch all the places where it uses a Hebrew word, like 0, 1000, 2000, etc.

[12:45:04.0000] <TabAtkins>
The current algo doesn't use a word anywhere.

[12:46:00.0000] <AryehGregor>
Okay, what URL are you looking at?

[12:46:01.0000] <AryehGregor>
I'm looking at a WD, I think.

[12:46:02.0000] <TabAtkins>
http://dev.w3.org/csswg/css3-lists/#algorithmic

[12:46:03.0000] <TabAtkins>
It's just "choose a value from this list for the 100k digit, 10k digit, 1k digit, add a geresh if necessary, 100 digit, 10 digit, 1 digit".

[12:47:00.0000] <AryehGregor>
This looks fine to me, yes.

[12:47:01.0000] <TabAtkins>
Kk, cool.

[12:47:02.0000] <AryehGregor>
I don't know how standard it is for stuff above 1000.  I've seen that convention used, but people also sometimes just continue on without a geresh.

[12:47:03.0000] <AryehGregor>
E.g., תתר for 1000.

[12:47:04.0000] <TabAtkins>
I'm developing a syntax to allow authors to specify list-styles themselves.  All the non-algorithmics are trivial, and I'm trying to solve as many algorithmics as I can do simply.

[12:48:00.0000] <AryehGregor>
That gets unwieldy above about 1200, though.

[12:48:01.0000] <TabAtkins>
Most of them fall into either "cjk" or "additive" camps.

[12:48:02.0000] <AryehGregor>
The Hebrew one is quite simple.  The only oddity is the 15/16 thing, I think.

[12:49:00.0000] <TabAtkins>
Yeah, but that's easy enough to do with a greedy additive algorithm.  Just specify 19/18/17/16/15 as possible tokens.

[12:49:01.0000] <TabAtkins>
Then it'll take those in preference to building 15/16 the "normal" way with a 10 and a 5/6.

[12:49:02.0000] <AryehGregor>
Ah, I see.

[12:49:03.0000] <AryehGregor>
Yeah, that should work fine.

[12:50:00.0000] <TabAtkins>
Good to know.

[12:51:00.0000] <TabAtkins>
Stupid ethiopic and tamil numbering systems being all complicated.

[12:52:00.0000] <zcorpan_>
does css support roman numbers?

[12:52:01.0000] <TabAtkins>
I can handle tamil with an addition to my additive system, but ethiopic is impossible without creating something that might as well be general purpose.

[12:52:02.0000] <TabAtkins>
zcorpan_: CSS3 Lists describes it (under the algorithmic types).

[12:53:00.0000] <TabAtkins>
My method of defining list-styles also supports it.

[12:53:01.0000] <zcorpan_>
nice

[12:54:00.0000] <TabAtkins>
I currently have an example UA stylesheet that defines every non-glyph and non-algorithmic list type.  I'll be adding the algorithmics that I can handle soon.

[12:55:00.0000] <TabAtkins>
I wonder where I could find someone who knows the tamil numbering system?

[12:55:01.0000] <TabAtkins>
I want to confirm something before I give up on it.

[12:55:02.0000] <TabAtkins>
Random fun fact - in my syntax, it's pretty easy to describe a "one, two, three" list-style.

[12:56:00.0000] <AryehGregor>
You should support Roman numerals the way the romans actually used them most of the time, with IIII instead of IV and so on.  Then it's really trivial.

[12:57:00.0000] <AryehGregor>
(fun fact: watches with Roman numerals typically have IIII instead of IV, but IX instead of VIIII)

[12:57:01.0000] <TabAtkins>
Indeed, both of those systems are trivial.

[12:58:00.0000] <TabAtkins>
The subtractive roman numerals just require a handful more entries in the token list.

[12:58:01.0000] <TabAtkins>
(So that it'll take "XC" for 90 if it can, frex.)

[12:58:02.0000] <TabAtkins>
The only thing I can't do in my system is support overbars, but honestly who cares.

[12:59:00.0000] <TabAtkins>
I'm fine with the range just being 1-39999.

[12:59:01.0000] <zcorpan_>
/me didn't know about overbars in roman numbers

[13:00:00.0000] <AryehGregor>
There's no Roman-numeral-with-overbar Unicode character?

[13:00:01.0000] <AryehGregor>
characters.

[13:00:02.0000] <TabAtkins>
I dunno.  Maybe?

[13:02:00.0000] <AryehGregor>
I don't see it.

[13:02:01.0000] <AryehGregor>
There are Roman numerals around U+2160 ff., but none with overbars.

[13:03:00.0000] <AryehGregor>
At least not in my font.

[13:03:01.0000] <TabAtkins>
Yeah, though apparently unicode recommends against using them.  Apparently they're mainly for cjk usage.

[13:03:02.0000] <TabAtkins>
They recommend just using latin letters normally.

[13:03:03.0000] <AryehGregor>
Right, I remember that.

[13:09:00.0000] <TabAtkins>
Hmm, does it make sense to include the 5k and 10k characters?  They're marked as "archaic" roman numerals in the unicode listing.

[13:14:00.0000] <Philip`>
I thought most Roman numerals were pretty archaic

[13:15:00.0000] <TabAtkins>
They prefer the term "classical".

[13:34:00.0000] <AryehGregor>
Only Roman letters are usually used for Roman numerals.

[13:34:01.0000] <AryehGregor>
Up to M for 1000.

[13:34:02.0000] <AryehGregor>
Today, I mean.

[13:38:00.0000] <zcorpan_>
chrome 6.0.422.0 crashes on my websocket testsuite

[13:39:00.0000] <zcorpan_>
hmm, one test not only killed the tab but killed all of chrome

[13:41:00.0000] <zcorpan_>
now it seems to be any test - maybe chrome got fed up and didn't like websockets anymore

[13:41:01.0000] <AryehGregor>
You know Chrome changed the WebSocket implementation incompatibly in 6.0, right?

[13:41:02.0000] <AryehGregor>
/me assumes so, but is just checking

[13:42:00.0000] <zcorpan_>
yes, that's why my tests work at all

[13:42:01.0000] <zcorpan_>
i should probably look into making my tests public

[13:44:00.0000] <zcorpan_>
seems it crashes if the server sends a binary frame with length 2^32

[14:05:00.0000] <zcorpan_>
ok out of 158 tests, chrome passes 98 and crashes on 2, and sometimes seemingly crashes hard on any test

[14:05:01.0000] <gsnedders>
zcorpan_: On OS X?

[14:05:02.0000] <zcorpan_>
yes

[14:14:00.0000] <zcorpan_>
seems chrome happily opens the connection without checking the upgrade and connection fields

[15:21:00.0000] <TabAtkins>
Argh, damn you hebrew.  I'm just going to skip providing the hebrew list-type in actual characters and only do it with escapes, because bidi mixing is super-frustrating.  >_<

[15:30:00.0000] <AryehGregor>
It's super-frustrating even if you can actually read it.

[15:31:00.0000] <AryehGregor>
Especially when it keeps switching direction as you type.

[15:36:00.0000] <TabAtkins>
For real.  I was about to start cussing when it was swapping the direction of backspace and delete based on which direction I approached the comma from.

[15:38:00.0000] <AryehGregor>
Yep.


2010-06-04
[18:28:00.0000] <[bjoern]>
http://www.websitedev.de/temp/canvas-heatmap-2.html -- Should this be part of a circle with a solid blue color in the upper left corner (QtWebkit), a colorful circle somewhere on the canvas (Firefox nightly), or a large colorful rectangle (Opera)?

[19:54:00.0000] <MikeSmith>
[bjoern]: Safari and Chrome on OSX vote for making it the large colorful rectangle

[19:56:00.0000] <[bjoern]>
That suits me, I'll claim that's the correct behavior then. Thanks.

[20:32:00.0000] <MikeSmith>
http://twitter.com/azaaza/status/15383718153

[20:32:01.0000] <MikeSmith>
"Did you know Apple has a Web Apps directory? You have to be an ADC Premier member to submit. http://www.apple.com/webapps/"

[20:33:00.0000] <paul_irish>
it's the web app store that gets no traffic!

[20:33:01.0000] <paul_irish>
\o/

[20:35:00.0000] <MikeSmith>
:)

[20:36:00.0000] <MikeSmith>
http://www.apple.com/webapps/utilities/html5reference.html

[20:39:00.0000] <miketaylr>
link to: http://www.biocourseware.com/iphone/html5/index.htm#1 ?

[20:40:00.0000] <miketaylr>
bogus, totally isn't working in any browsers on this mac :/

[20:41:00.0000] <miketaylr>
/me failed to realize its an iphone webapp :(

[20:41:01.0000] <miketaylr>
/me goes away

[21:03:00.0000] <MikeSmithX>
http://twitter.com/emilbjorklund/status/15328003664

[21:03:01.0000] <MikeSmithX>
"Second annoyance of the day: the html5 spec links you get from the validator crashes firefox."

[21:03:02.0000] <MikeSmithX>
should have v.nu link to the multipage version of the spec

[23:51:00.0000] <zcorpan_>
awesome how http://www.apple.com/html5/ refuses to show the demos in any other browser than safari

[23:52:00.0000] <zcorpan_>
including chrome

[23:56:00.0000] <Peter`>
A few months from now they'll have a fair reason to do that with <video>, being the only major browser which doesn't support WebM ;-)

[23:56:01.0000] <zcorpan_>
i assume chrome is still going to support h.264

[23:57:00.0000] <Peter`>
As will IE 9 :) Loading the full URL of a test in Chrome doesn't work either ( http://www.apple.com/html5/showcase/typography/ ), redirects to the overview page

[23:57:01.0000] <Peter`>
Ironically the sliders are broken in Safari

[00:48:00.0000] <MikeSmith>
wtf

[00:48:01.0000] <MikeSmith>
http://www.w3.org/TR/MathML/chapter3.html#id.3.5.1.2

[00:48:02.0000] <MikeSmith>
allowed values are (top | bottom | center | baseline | axis) [ rownumber ]

[00:49:00.0000] <MikeSmith>
where [ rownumber ] is an optional number that can be appended to "top", "bottom", etc.?

[00:49:01.0000] <MikeSmith>
e.g., top3 ?

[00:49:02.0000] <MikeSmith>
can that really be?

[00:50:00.0000] <MikeSmith>
attribute align {xsd:string {

[00:50:01.0000] <MikeSmith>
    pattern ='\s*(top|bottom|center|baseline|axis)\s*[0-9]*'}}?

[00:51:00.0000] <MikeSmith>
wow

[01:03:00.0000] <boblet>
sounds funky

[01:03:01.0000] <boblet>
gsnedders’ outliner doesn’t show implicit sectioning does it

[01:06:00.0000] <boblet>
hrm… still not sure if I understand implied sections :/

[01:07:00.0000] <zcorpan_>
<h1>foo</h1><h2>hi, i'm in an implied section</h1>

[01:08:00.0000] <boblet>
d’oh! gsnedders’ outliner does show implied sections, but only those without typos

[01:08:01.0000] <boblet>
obviously the mind-reading update hasn’t happened yet

[01:09:00.0000] <boblet>
zcorpan_: I kinda understand the theory, but I’m not *feelin’ it* ye

[01:09:01.0000] <boblet>
yet

[01:10:00.0000] <boblet>
it’s basically an error correction mechanism, right? for UGE

[01:10:01.0000] <boblet>
(user-generated errors)

[01:13:00.0000] <hsivonen>
MikeSmith: curious that they allow whitespace before but not after

[01:13:01.0000] <hsivonen>
not after the number

[01:13:02.0000] <hsivonen>
but before the text

[01:15:00.0000] <MikeSmith>
hsivonen: the more I see of that vocabulary the more it scares me

[01:16:00.0000] <boblet>
another outlining q: using <body><article><h1>title</h1>… gives a root “Untitled section”. wassup with that? we’re meant to add a heading element outside sectioning content? I don’t get where it’s coming from

[01:16:01.0000] <MikeSmith>
anyway, it seems clear there are some more deficiencies in that MathML 2.0 rng schema

[01:17:00.0000] <MikeSmith>
I'm wondering if it would be easier to just start with the MathML 3.0 one and remove stuff

[01:17:01.0000] <MikeSmith>
maybe I should ping David Carlisle and ask

[01:17:02.0000] <jgraham>
boblet: Implied sections are basically HTML4 

[01:18:00.0000] <jgraham>
boblet: The <body> element is also a sectioning element

[01:18:01.0000] <boblet>
jgraham: aah! I forgot about sectioning root. doh

[01:18:02.0000] <boblet>
thanks

[01:18:03.0000] <jgraham>
So if you have <body><h1>My funky site</h1><article><h1>My article everything works

[01:19:00.0000] <boblet>
outlining algorithm needs to be communicated by mind-meld. explaining it in writing is brutal

[01:20:00.0000] <hsivonen>
MikeSmith: have you filed the scary stuff as spec bugs against MathML 3?

[01:20:01.0000] <MikeSmith>
no

[01:20:02.0000] <MikeSmith>
I guess I should

[01:20:03.0000] <MikeSmith>
I just wonder if any of this stuff is actually supported

[01:21:00.0000] <MikeSmith>
I mean, what does align=baseline3 even mean?

[01:21:01.0000] <hsivonen>
/me has no idea

[01:25:00.0000] <MikeSmith>
hsivonen: btw, you got my question from the other day about hooking v.nu HTML5 parser into other java apps?

[01:25:01.0000] <MikeSmith>
http://krijnhoetmer.nl/irc-logs/whatwg/20100602#l-181

[01:26:00.0000] <hsivonen>
MikeSmith: I've missed that bit in the logs

[01:27:00.0000] <hsivonen>
MikeSmith: you need a DocumentBuilderFactory that instantiates HtmlDocumentBuilder

[01:27:01.0000] <MikeSmith>
OK

[01:27:02.0000] <hsivonen>
MikeSmith: I haven't supplied one, because I wasn't aware of people actually needing one

[01:28:00.0000] <MikeSmith>
hsivonen: I was thinking it would be useful for anybody who wants to just use the HTML5 parser in place of whatever existing XML one

[01:28:01.0000] <MikeSmith>
e.g., with Saxon

[01:28:02.0000] <hsivonen>
I thought they'd just do new HtmlDocumentBuilder

[01:28:03.0000] <MikeSmith>
or jing or whatever

[01:28:04.0000] <hsivonen>
I didn't realize people actually want the factory stuff

[01:28:05.0000] <MikeSmith>
I mean using it from the command-line

[01:28:06.0000] <hsivonen>
but yeah, that would be doable

[01:29:00.0000] <MikeSmith>
OK

[01:29:01.0000] <MikeSmith>
should I open a bug for it?

[01:29:02.0000] <hsivonen>
MikeSmith: yes, please

[01:29:03.0000] <MikeSmith>
ok

[01:31:00.0000] <jgraham>
TabAtkins: Surely there is a combining overbar character somewhere in unicode?

[01:33:00.0000] <zcorpan_>
jgraham: with multiple overbars?

[01:34:00.0000] <MikeSmith>
hsivonen: what sax, what would the corresponding SAXParserFactory need to instantiate?

[01:35:00.0000] <MikeSmith>
sax/HtmlParser ?

[01:40:00.0000] <hsivonen>
yeah

[01:45:00.0000] <MikeSmith>
ok

[01:47:00.0000] <boblet>
does “sectioning element” mean anything specific in the spec? I’d like to use it as a reference for sectioning content and sectioning root elements, but wonder if it’s already being used to mean something else…

[01:49:00.0000] <jgraham>
zcorpan_: Maybe not

[01:49:01.0000] <jgraham>
(but you would have thought there would be if you need it for roman numerals)

[01:49:02.0000] <boblet>
can’t find it in the spec, so maybe I’m safe

[01:50:00.0000] <boblet>
roman numerals?

[01:51:00.0000] <boblet>
oh, you mean I should have said “as shorthand for…”?

[01:52:00.0000] <jgraham>
boblet: I was replying to zcorpan_ 

[01:52:01.0000] <boblet>
jgraham: haha, d’oh

[01:53:00.0000] <boblet>
orz

[03:56:00.0000] <zcorpan_>
'interact with web standards' just arrived to my door

[04:13:00.0000] <sid123>
hi I wanted to know about the storage limit of web data bases... There is a size parameter that is provided to the Database initilaization... But is that followed strictly or are there limits on it in any way? 

[04:32:00.0000] <Lachy>
wow. What a great way to promote HTML5 and web standards, by requiring the use of Safari to view any of their demos http://www.apple.com/html5/

[04:32:01.0000] <Lachy>
</sarcasm>

[04:32:02.0000] <zcorpan_>
old news

[04:32:03.0000] <Lachy>
well, new to me. I only just read my news feeds

[04:33:00.0000] <boblet>
adactio: thanks for the email (was starting to wonder if I had the right addy :P )

[04:33:01.0000] <adactio>
boblet: yeah, really sorry it took me so long to respond. :-(

[04:33:02.0000] <boblet>
heh

[04:33:03.0000] <boblet>
adactio: not like you’ve been up to much either :|

[04:34:00.0000] <adactio>
boblet: ;-)

[04:34:01.0000] <boblet>
thanks for the feedback tho, really appreciated (just finished a big screed on sectioning root today)

[04:37:00.0000] <hsivonen>
Lachy: I'm rather amazed that whoever put the browser sniffing in place didn't notice the mismatch with the message and readjust

[04:38:00.0000] <hsivonen>
or then, the message isn't really about standards but about Safari vs. Flash

[04:38:01.0000] <hsivonen>
not even about WebKit vs. Flash

[04:38:02.0000] <hsivonen>
/me hasn't yet bothered to examine the demos in Safari to see how the VR demo works

[04:42:00.0000] <hsivonen>
ah. CSS 3D Transforms

[04:43:00.0000] <hsivonen>
at least their developer site has the decency to call them "Safari Technology Demos"

[04:43:01.0000] <Rik`>
in the video example, they're using a CSS mask. It's not yet in any draft, is it ?

[04:44:00.0000] <zcorpan_>
it's web standards!1

[04:45:00.0000] <boblet>
bbl

[04:46:00.0000] <hsivonen>
more clever than dealing with a spherical mapping

[05:22:00.0000] <hsivonen>
where can I find a signed Chrome extension?

[05:23:00.0000] <Peter`>
Signed how?

[05:24:00.0000] <hsivonen>
Peter`: that's what I want to find out

[05:24:01.0000] <Peter`>
they're just zip files with a certificate to generate an unique hash

[05:24:02.0000] <hsivonen>
Peter`: where's the cert?

[05:25:00.0000] <Peter`>
hsivonen: On the computer of the person who packed it, as a .pem file

[05:25:01.0000] <Peter`>
You can rename .crx to .zip and it'll work as well

[05:25:02.0000] <hsivonen>
Peter`: where's the signature?

[05:26:00.0000] <Peter`>
I think it's only used to generate the extension hash, used for auto-updating and such. It's not in the archive either

[05:27:00.0000] <hsivonen>
Peter`: ah. that doesn't match what the intro video on youtube said

[05:28:00.0000] <Peter`>
I could be wrong, I'm not a Chrome developer, just somewhat familiar with the code. Best place to ask probably would be #chromium

[06:00:00.0000] <hsivonen>
well, I just asked on the whatwg list

[06:25:00.0000] <boblet>
hey all, more questions on outlining algorithm — what’s it’s raison d'être? does it assist UAs to provide direct access? is it to assist browsers in DOM reflow? something else?

[06:25:01.0000] <boblet>
I feel like I understand it without really getting it

[06:27:00.0000] <hsivonen>
boblet: the primary point is allowing browsers to provide UI for navigating directly a section or getting an overview of teh document structure

[06:27:01.0000] <hsivonen>
boblet: mainly for screen readers

[06:27:02.0000] <hsivonen>
boblet: the secondary point is styling based on outline depth

[06:28:00.0000] <hsivonen>
boblet: which depends on whether browser vendors and the CSS WG add a way of selecting on outline depth

[06:28:01.0000] <hsivonen>
if such a selector proves to be infeasible, well...

[06:41:00.0000] <boblet>
hsivonen: sorry — story time. thanks for your explanation

[06:43:00.0000] <boblet>
so for sectioning root elements like figure and blockquote, is it expected they’ll contain large chunks of content with headers, or that UAs will provide access to sectioning root outlines?

[06:44:00.0000] <boblet>
I’m not sure why sectioning root outlines don’t contribute to the main outline — perhaps because they normally lack a heading?

[06:49:00.0000] <hsivonen>
boblet: the idea is to hide quoted stuff from the main outline

[06:52:00.0000] <boblet>
hsivonen: quoted stuff (blockquote/figure) and generally non-outline stuff (details, fieldset, td) that sometimes contains headings, right?

[06:55:00.0000] <hsivonen>
boblet: right

[06:58:00.0000] <boblet>
hsivonen: ok, thanks for your explanation. I was worried there might have been something else that I wasn’t getting, but it appears not

[08:58:00.0000] <TabAtkins>
/me is sad whenever he sees attributes which accept true/false.

[08:58:01.0000] <TabAtkins>
Rather than either being binary or giving the values better names for future extensibility.

[09:00:00.0000] <Philip`>
true/false allows plenty of room for extensibility

[09:00:01.0000] <Philip`>
http://thedailywtf.com/Articles/What_Is_Truth_0x3f_.aspx

[09:00:02.0000] <TabAtkins>
true/false/file-not-found

[10:14:00.0000] <asmodai>
lol

[10:14:01.0000] <asmodai>
http://i.imgur.com/cT08B.png

[10:18:00.0000] <AryehGregor>
Is html5test.com actually a competent testing service?

[10:18:01.0000] <AryehGregor>
Some of the tests look bogus to me.

[10:19:00.0000] <AryehGregor>
Like testing for specific audio and video codecs.

[10:20:00.0000] <AryehGregor>
And it reports that Chrome dev channel doesn't support the autocomplete attribute, which is nonsense -- browsers have all supported that for years, haven't they?

[10:20:01.0000] <AryehGregor>
/me doesn't trust it

[10:20:02.0000] <AryehGregor>
Safari and Chrome have practically identical HTML5 conformance if you compare comparable versions, because they both use WebKit.

[10:20:03.0000] <TabAtkins>
We do different things on top of that base, though.

[10:21:00.0000] <TabAtkins>
Some things show up in Chrome before we push them all the way back to Webkit, frex.

[10:21:01.0000] <AryehGregor>
Not ones that are relevant to almost any of the features that html5test.com supposedly test.

[10:21:02.0000] <TabAtkins>
And presumably the same sometimes happens with Safari.

[10:21:03.0000] <AryehGregor>
A few, yes, but not many.

[10:21:04.0000] <TabAtkins>
Yeah, wouldn't know about that.

[10:37:00.0000] <Philip`>
It seems the popularity of a test suite is very highly correlated with how easy it is for a random user to run it in their own browser and see a big clear interestingly-designed page indicating the results

[10:38:00.0000] <Philip`>
and very loosely correlated with the quality or range of its test cases

[10:38:01.0000] <TabAtkins>
That makes sense.  Random users have no idea how to evaluate the quality of tests or their coverage.

[10:39:00.0000] <TabAtkins>
They *can* evaluate whether or not it runs, and how pretty a results page is.

[10:42:00.0000] <Philip`>
(Popularity isn't a particularly important metric for a test suite, though)

[11:00:00.0000] <boblet>
is there a section on design principles in WHATWG spec? similar to http://www.w3.org/TR/html-design-principles/

[11:08:00.0000] <Philip`>
boblet: Those principles were mostly just writing down the principles that had been used in the WHATWG process

[11:08:01.0000] <Philip`>
so there's no difference

[11:09:00.0000] <boblet>
Philip`: thanks

[11:09:01.0000] <Philip`>
(They weren't written down explicitly before then)

[11:09:02.0000] <Philip`>
(mostly since they were already obvious to most WHATWG members)

[11:34:00.0000] <dglazkov>
TabAtkins: btw, we don't have a fork of WebKit. All development for Chromium features happens on WebKit ToT. No pushing is involved :)

[11:37:00.0000] <TabAtkins>
Ah, I am mistaken then.

[11:42:00.0000] <boblet>
regarding the accessibility design principle, would you say that “Design features to be accessible to users with disabilities” could be accurately summarised in “accessibility by default”?

[11:43:00.0000] <Dashiva>
I would say so, but it's not entirely uncontroversial

[11:44:00.0000] <boblet>
Dashiva: nothing in a11y ever is :)

[11:44:01.0000] <TabAtkins>
boblet: Subtly different.

[11:44:02.0000] <Dashiva>
There's the "bolt-on accessibility is a sign of design failure" and "bolt-on accessibility is the way to great accessibility"

[11:45:00.0000] <TabAtkins>
The former could mean doing something that *can* be accessible if authors take some effort.  The latter would avoid doing that in favor of something that is, well, accessible by default.

[11:45:01.0000] <boblet>
TabAtkins: oh? uhoh

[11:46:00.0000] <TabAtkins>
Ideally we want to avoid requiring any author interaction whatsoever to acheive good accessibility.

[11:46:01.0000] <TabAtkins>
That isn't quite possible, so instead we just minimize the amount of author interaction necessary.

[11:46:02.0000] <boblet>
well I sense a trend towards things being more accessible by default, eg moving attribute-based meta info into main content. but I don’t see anything so nebulous in the design principles :)

[11:46:03.0000] <boblet>
aah

[11:47:00.0000] <Dashiva>
Where possible, trick authors into writing accessible content without realizing it

[11:47:01.0000] <TabAtkins>
Got it, Dashiva.  ^_^

[11:47:02.0000] <boblet>
:/

[11:48:00.0000] <boblet>
actually… Dashiva, can I quote you on that?

[11:48:01.0000] <boblet>
;)

[11:48:02.0000] <TabAtkins>
Simple, widely-attested truth - the amount of accessible content decreases exponentially with the amount of effort required.

[11:48:03.0000] <TabAtkins>
(Well, maybe not exponentially - I don't have any numbers to attest that.  But super-linear at least.)

[11:49:00.0000] <boblet>
no doubt

[11:49:01.0000] <Dashiva>
Re: moving attribute-based meta info into main content, that's not primarily for accessibility, that's because hidden metadata has problems in itself

[11:50:00.0000] <Dashiva>
But accessibility-targeted hidden metadata is has even more problems

[11:50:01.0000] <Dashiva>
-is

[11:50:02.0000] <boblet>
I was referring to previously a11y-focussed metadata like whatever-it-is on table

[11:50:03.0000] <Dashiva>
Since it's not only hidden, but it's basically non-existent to users who don't need the extra accessibility

[11:52:00.0000] <TabAtkins>
Also: people don't know how to write good @summary text.  And auto-generated @summary is useless - if it can be autogenerated by the producer, it can be auto-generated by the consumer equally well.

[11:52:01.0000] <TabAtkins>
And then the consumer *knows* that it's auto-generated and can treat it differently than custom-crafted text.

[11:53:00.0000] <boblet>
yeah, I completely agree with the idea

[11:53:01.0000] <boblet>
did the issue about that get resolved? I should check…

[11:56:00.0000] <boblet>
review huh

[11:58:00.0000] <AryehGregor>
http://beta.html5test.com/ actually looks pretty good.

[11:59:00.0000] <AryehGregor>
Blech, it tests <device>, is that even specced yet?

[11:59:01.0000] <TabAtkins>
It *exists* in a spec.

[11:59:02.0000] <boblet>
I hope someone starts graphing score over time in one of these sometime

[11:59:03.0000] <TabAtkins>
It's not "specced" though.

[12:16:00.0000] <AryehGregor>
http://damowmow.com/playground/htmlwg/chart.html "Last updated NaN days ago."

[12:21:00.0000] <TabAtkins>
var updateStr = ' Last updated ' + (age < 1 ? 'today' : age.toFixed(0) + ' day' + (age.toFixed(0) == 1 ? '' : 's') + ' ago') + '.';

[12:23:00.0000] <Dashiva>
Good-ole' NaN

[12:29:00.0000] <TabAtkins>
The problem is the toFixed(0) call.  It appears to give "NaN" as the result for any number.

[12:30:00.0000] <TabAtkins>
At least, here in Chrome.

[12:31:00.0000] <TabAtkins>
Firefox does it right.

[12:31:01.0000] <TabAtkins>
So, looks like a bug on our end.

[12:35:00.0000] <AryehGregor>
The display doesn't work for me either.

[12:36:00.0000] <TabAtkins>
Yeah, me neither.

[12:36:01.0000] <TabAtkins>
What do you want to bet it's toFixed again?

[12:36:02.0000] <Dashiva>
It errors in Opera on a lineTo call

[12:37:00.0000] <AryehGregor>
Is this test completely broken, or is it just me?  http://samples.msdn.microsoft.com/ietestcenter/html5/selection/select.htm

[12:38:00.0000] <Dashiva>
There's a NaN object in there

[12:38:01.0000] <Dashiva>
Hum

[12:39:00.0000] <AryehGregor>
Looking at the source code, it looks as though it selects the contents of the input, then checks whether the contents are *not* selected, or something similarly crazy.

[12:41:00.0000] <boblet>
I’m wondering what the big a11y changes in HTML5 are. I know alt text, longdesc, table summary, plus the current video/canvas and interactive element (progress, details etc) issues. anything else I’m missing?

[12:41:01.0000] <Dashiva>
It looks fine to me, where in the code is the problem?

[12:42:00.0000] <AryehGregor>
Dashiva, it does input1.select(), then checkDefaultSelectionAttributes().  The latter checks things like null != selection.anchorNode.  Why should selection.anchorNode be null if you just selected the input?

[12:45:00.0000] <TabAtkins>
boblet: Not an expert, but I think that's it.

[12:45:01.0000] <boblet>
TabAtkins: thanks. more sanity-checking than anything :)

[12:47:00.0000] <Dashiva>
AryehGregor: Those don't apply to selection in input fields

[13:13:00.0000] <AryehGregor>
Dashiva, where does the spec say that?  It says that text fields have their own selections, but not that selecting text in text fields doesn't also select it in the browsing context.  Am I missing something?

[13:14:00.0000] <AryehGregor>
It just says that the UA should allow the user to change the selection.

[13:14:01.0000] <AryehGregor>
"This one selection must be shared by all the content of the browsing context (though not by nested browsing contexts), including any editing hosts in the document."\

[13:14:02.0000] <AryehGregor>
s/\\//

[13:15:00.0000] <AryehGregor>
Firefox and WebKit seem to have the more plausible interpretation to me.

[13:27:00.0000] <Dashiva>
AryehGregor: Firefox seems weird to me

[13:34:00.0000] <AryehGregor>
Ugh, faked browser UI in Chrome: http://www.huffingtonpost.com/stewart-abercrombie-baker/smallpox-in-the-garage_b_596340.html

[13:34:01.0000] <AryehGregor>
Evil.

[13:34:02.0000] <AryehGregor>
They could have done better by using position: fixed.

[13:34:03.0000] <Dashiva>
AryehGregor: Looks to me like Firefox claims there's a zero-length selection at a random location in the document...

[13:35:00.0000] <Dashiva>
When there's no selection at all

[13:35:01.0000] <AryehGregor>
Dashiva, well, the spec says that the UA should provide a means for the user to change the selection, not that it has to make sense.  Clearly, this is merely a quality-of-implementation issue, and everyone passes.

[13:36:00.0000] <Dashiva>
The spec also says "Returns null if there's no selection."

[13:36:01.0000] <Dashiva>
Which firefox does not

[13:37:00.0000] <AryehGregor>
Well, perhaps Firefox has simply chosen to interpret page load as a user choice of selection.

[13:37:01.0000] <AryehGregor>
This is what happens when you leave UI totally undefined, am I right?  :)

[13:38:00.0000] <Dashiva>
If at any point you (or a script) have selected anything, it never forgets

[13:38:01.0000] <Dashiva>
E.g. select some random text, remove the selection, and it claims there's a zero-length selection at the start of the ex-selection

[13:39:00.0000] <Dashiva>
If there is no ex-selection (such as in that test), it picks body for some reason

[13:40:00.0000] <AryehGregor>
Whee.

[13:40:01.0000] <Dashiva>
Never mind

[13:40:02.0000] <Dashiva>
It picks the parent of the input the script selected

[13:41:00.0000] <Dashiva>
But the offsets are 1, not the offset of the input

[13:42:00.0000] <Dashiva>
http://dashiva.net/test/selection.html

[14:21:00.0000] <dglazkov>
nobody wants to stay in touch on LinkedIn. Well said.

[14:23:00.0000] <TabAtkins>
Haha, yes.

[14:28:00.0000] <AryehGregor>
What?

[14:28:01.0000] <TabAtkins>
On the whatwg list.

[14:28:02.0000] <AryehGregor>
Ohl

[14:28:03.0000] <AryehGregor>
Oh.

[14:59:00.0000] <TabAtkins>
Oh man, using unicode box-drawing characters when doing ascii illustrations is 1000% sexier than using +-|.

[15:00:00.0000] <TabAtkins>
Though I suppose it no longer qualifies as "ascii".  ^_^

[15:00:01.0000] <Dashiva>
Real men use 1x1 pixel divs

[15:01:00.0000] <TabAtkins>
Real men use tables with background-colors on the cells.

[15:01:01.0000] <Dashiva>
But that isn't _semantic_

[15:01:02.0000] <TabAtkins>
Box-drawing characters are the answer, then.  Just wrap it in a <pre> and you're golden.

[15:02:00.0000] <TabAtkins>
(Or just send plain-text emails like I do and hope everyone uses a fixed-width font to view emails.)

[15:02:01.0000] <Philip`>
TabAtkins: Tables like in http://blogs.msdn.com/b/oldnewthing/archive/2010/05/26/10015253.aspx ?

[15:02:02.0000] <TabAtkins>
Yes, exactly like that.

[15:02:03.0000] <Philip`>
Gmail ought to automatically detect ASCII art and switch to a fixed-width font

[15:04:00.0000] <TabAtkins>
That wouldn't be overly hard, I bet.

[15:05:00.0000] <Dashiva>
Your mission, should you choose to accept it, is to produce a non-edge case that would trigger a false positive

[15:05:01.0000] <Philip`>
TabAtkins: I will rely on you and await the feature eagerly

[15:05:02.0000] <TabAtkins>
What's the easiest way to add new key combos on linux?  I know on Windows you can just produce a new keyboard layout file.

[15:05:03.0000] <TabAtkins>
Philip`: I'll send the suggestion upstream.

[15:05:04.0000] <TabAtkins>
/me wants to put the box-drawing characters on his keyboard.

[15:05:05.0000] <Dashiva>
Probably depends on the WM

[15:06:00.0000] <Dashiva>
There are like four layers of stuff between the app and the keyboard

[15:06:01.0000] <TabAtkins>
That's dumb.

[15:06:02.0000] <TabAtkins>
The WM shouldn't be taking part in this.  >_<

[15:06:03.0000] <Dashiva>
I'm trying to stay blissfully unaware, but I stumbled across it via a discussion about some laptop key used to switch displays or something

[15:07:00.0000] <TabAtkins>
Maybe XKB will work?

[15:09:00.0000] <AryehGregor>
/me finally gets the LinkedIn joke

[15:14:00.0000] <AryehGregor>
It's probably not the WM, it's probably the desktop environment.

[15:14:01.0000] <AryehGregor>
Which certainly has to take part in things like shortcut keys, although it'd be nice if that sort of thing were standardized.

[15:17:00.0000] <jgraham>
Philip`: Popularity is an important metric for a testsuite if it causes pressure on implementors to fix specific issues

[15:17:01.0000] <jgraham>
e.g. Acid 3

[15:17:02.0000] <jgraham>
(for better or worse)

[15:25:00.0000] <Philip`>
jgraham: True

[15:31:00.0000] <jcranmer>
TabAtkins: xkb; if you're willing to do compose, a .XCompose file lets you add even more compose keys

[15:36:00.0000] <TabAtkins>
Ooh, interesting.

[16:09:00.0000] <AryehGregor>
"Interesting" is an excellent description for how Linux works in so many cases.  :)

[16:13:00.0000] <tabatkins>
Given <label><input type=checkbox><a href=#>foo</a></label>, I *think* that the 3.2.5.1 algo says to activate the link (and not toggle the checkbox) when you click the link.

[16:13:01.0000] <tabatkins>
Yes/no?

[16:14:00.0000] <Dashiva>
3.2.5.1 Kinds of content?

[16:14:01.0000] <TabAtkins>
Sorry, 3.2.5.1.7

[16:14:02.0000] <Dashiva>
ah

[16:15:00.0000] <TabAtkins>
It looks like you click on the <a>, the <a> is the nearest activatable element, then you run the click steps on it.

[16:15:01.0000] <Dashiva>
Yeah

[16:33:00.0000] <TabAtkins>
AryehGregor: Any way to trigger a monospace font in mediawiki?

[16:33:01.0000] <AryehGregor>
<span style="font-family:monospace">, <tt>, <code>, . . .

[16:33:02.0000] <AryehGregor>
You can also begin a line with a space.

[16:33:03.0000] <TabAtkins>
Danke.  

[16:34:00.0000] <TabAtkins>
Ah, cool.

[16:34:01.0000] <TabAtkins>
Someone stupid did the examples for the box-drawing characters using a table for some reason.

[16:35:00.0000] <TabAtkins>
Apparently it's for some sort of compat?  Whatever.  <pre><code> will work as expected?

[16:36:00.0000] <AryehGregor>
<pre> is magic in MediaWiki.

[16:36:01.0000] <AryehGregor>
It means the same as <pre><nowiki>, so the contents are escaped.

[16:37:00.0000] <AryehGregor>
So don't use that.  An initial space on a line will create a <pre>.

[16:37:01.0000] <TabAtkins>
kk

[16:37:02.0000] <AryehGregor>
That's the only way to create a normal <pre>, annoyingly.

[16:48:00.0000] <TabAtkins>
AryehGregor: You considered putting in a few -moz-linear-gradient() rules in the places where mediawiki uses gradients?

[16:48:01.0000] <AryehGregor>
TabAtkins, I didn't know it does use gradients.  :)

[16:49:00.0000] <TabAtkins>
Well, like, the bar on top of each page uses gradients for each item.

[16:49:01.0000] <AryehGregor>
In the new skin (Vector) or the old one (Monobook)?

[16:49:02.0000] <TabAtkins>
Of course, it's just a 1x100 px png, but still, one less network request on ff.

[16:49:03.0000] <TabAtkins>
new one

[16:51:00.0000] <AryehGregor>
Well, got to go.

[16:52:00.0000] <AryehGregor>
Maybe I'll do it later if I remember, or you can remind me if you like.

[16:53:00.0000] <TabAtkins>
kk

[16:53:01.0000] <TabAtkins>
+


2010-06-05
[17:04:00.0000] <Dashiva>
These web standards are open, reliable, highly secure, and only supported by Safari.

[17:06:00.0000] <Dashiva>
I accidentally opened the chatlogs from 2008 instead of 2010, and I didn't notice for a while

[18:16:00.0000] <TabAtkins>
Man, wth.  4 freaking "google promo" emails have gotten through the spamblocker today.

[18:36:00.0000] <jwm>
google knows your spamblocker

[18:36:01.0000] <jwm>
be afraid

[21:21:00.0000] <MikeSmith>
I'm trying to understand what the validation errors at http://validator.nu/?doc=https%3A%2F%2Feyeasme.com%2FJoe%2FMathML%2FHTML5%2Fextras.html

[21:22:00.0000] <MikeSmith>
the initial error is because you can't have an HTML img element in SVG content served as text/html

[21:22:01.0000] <MikeSmith>
right?

[21:24:00.0000] <MikeSmith>
er, make that MathML content

[21:27:00.0000] <MikeSmith>
I thought for this case the parser is supposed to just ignore the img, then pop the open MathML annotation-xml element from the stack and proceed from there

[21:48:00.0000] <Hixie>
/me tries to decode http://lists.w3.org/Archives/Public/public-html/2010Jun/0001.html to work out what the reason is for removing the text

[21:48:01.0000] <Hixie>
does anybody know what the actual reason behind the "widespread belief" that the paragraph is harmful might be?

[21:48:02.0000] <Hixie>
nothing seems to actually explain what the reason is...

[21:49:00.0000] <aho>
huh? :)

[21:55:00.0000] <jwm>
Hixie: ?

[21:55:01.0000] <jwm>
you there?

[21:56:00.0000] <Hixie>
here

[21:56:01.0000] <jwm>
cool

[21:56:02.0000] <jwm>
is there anyone wanting to do websocket serving in browsers?

[21:56:03.0000] <jwm>
I know the peer to peer provisions got removed

[21:56:04.0000] <jwm>
but I have some big plans for doing web p2p

[21:56:05.0000] <Hixie>
the peer to peer stuff got moved to a differnet part of the spec

[21:56:06.0000] <Hixie>
but it's still there

[21:56:07.0000] <jwm>
ohh it is?

[21:56:08.0000] <Hixie>
search for PeerToPeer

[21:56:09.0000] <jwm>
I didn't see it anymore

[21:56:10.0000] <Hixie>
it's very immature so far

[21:57:00.0000] <jwm>
yeah anyway I could help?

[21:57:01.0000] <Hixie>
sorry, "peer-to-peer", not "PeerToPeer"

[21:57:02.0000] <jwm>
I'm really serious about p2p on web as a platform

[21:57:03.0000] <Hixie>
interface name is ConnectionPeer

[21:57:04.0000] <jwm>
I think it's the only way to attempt to avoid firewalls/packet inspection properly heh

[21:58:00.0000] <Hixie>
i envisage some sort of mechanism that would let peers connect to each other

[21:58:01.0000] <jwm>
what about doing that coding in javascript

[21:59:00.0000] <Hixie>
but with the protocol defined in a spearate working group

[21:59:01.0000] <Hixie>
coding?

[21:59:02.0000] <jwm>
I mean relying on client coding to handle peer connections

[21:59:03.0000] <jwm>
just keep the spec real simple

[21:59:04.0000] <jwm>
maybe even utilize websockets?

[21:59:05.0000] <Hixie>
websockets isn't really applicable here

[21:59:06.0000] <Hixie>
there's nothing about web sockets that makes sense for peer to peer

[22:00:00.0000] <jwm>
well just being about to serve websockets from a browser should be real simple to do

[22:00:01.0000] <jwm>
about/able

[22:00:02.0000] <Hixie>
it would also not solve like 90% of the use cases :-)

[22:00:03.0000] <jwm>
it's like 5 lines in nodejs

[22:00:04.0000] <Hixie>
like, you wouldn't be able to connect to it

[22:00:05.0000] <Hixie>
i think we need a much better solution than that

[22:01:00.0000] <jwm>
can't we build on websockets?

[22:01:01.0000] <jwm>
I know we need a firewall breaker utilizing upnp or some other methods

[22:01:02.0000] <jwm>
but that can be subject to browser implementation

[22:02:00.0000] <Hixie>
i see nothing in websockets that would be even remotely useful in doing this

[22:03:00.0000] <jwm>
sad to hear that

[22:03:01.0000] <jwm>
I think websockets is a nice simple protocol

[22:03:02.0000] <jwm>
it'd be nice to have a simple low overhead protocol even one already implemented in some form to use and get p2p jump started

[22:04:00.0000] <jwm>
I guess it can be coded as a plugin for now

[22:04:01.0000] <jwm>
see if anyone finds it useful

[22:04:02.0000] <Hixie>
websockets is a very simple protocol, yes -- it has like three features: a handshake with origin protection, a client/server architecture, and framing for text

[22:05:00.0000] <Hixie>
for peer-to-peer we don't have origins, the architecture is client/client, and the framing needs to support video.

[22:05:01.0000] <Hixie>
it's like completely different.

[22:05:02.0000] <jwm>
websockets 2.0 :)

[22:05:03.0000] <Hixie>
i don't think trying to use websockets is a helpful way forward here

[22:06:00.0000] <jwm>
it's just a simple way forward

[22:06:01.0000] <jwm>
I don't think it's bad to be able to share text/base64 coded images 

[22:07:00.0000] <Hixie>
if people start sending base64 encoded data over 8-bit-safe network infrastructure, we've failed.

[22:08:00.0000] <jwm>
well could websockets be expanded upon and be made binary safe

[22:09:00.0000] <Hixie>
man, you're like someone with a hammer who sees everything as a nail :-)

[22:10:00.0000] <Hixie>
we should design solutions to fit the problems we have

[22:10:01.0000] <Hixie>
not try to take our existing solutions and bend them to all similar problems

[22:10:02.0000] <jwm>
I like extending solutions we have already

[22:10:03.0000] <jwm>
websockets is pretty new and it'd be nice to boost it up :)

[22:10:04.0000] <Hixie>
no, it really wouldn't :-)

[22:10:05.0000] <jwm>
I'm actually conforming to the web since I see it as the main future of transporting most content

[22:10:06.0000] <Hixie>
it would just make it be more complex

[22:11:00.0000] <Hixie>
why can't we just make a more appropriate protocol for this use case?

[22:11:01.0000] <jwm>
really I say screw the web even and let's make more and more efficient protocols tasked for the problem at hand

[22:11:02.0000] <Hixie>
(or use one of the many existing ones?)

[22:11:03.0000] <Hixie>
anyway

[22:11:04.0000] <jwm>
well I'd like the traffic to go across regular protocols/ports

[22:15:00.0000] <jwm>
ok then I see you don't believe in the same direction

[22:15:01.0000] <jwm>
could you point me to the new section with the peer to peer area?

[22:17:00.0000] <jwm>
http://www.mail-archive.com/whatwg⊙lwo/msg19815.html

[22:17:01.0000] <jwm>
I guess there have already been discussions on this

[22:17:02.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/#peer-to-peer-connections

[22:20:00.0000] <jwm>
really big on video

[22:21:00.0000] <Hixie>
yeah it was originally intended for video conferencing

[22:21:01.0000] <jwm>
ahh ok

[22:21:02.0000] <jwm>
I don't like the third party server concept

[22:22:00.0000] <Hixie>
we need it, no other way to do double NAT traversal

[22:24:00.0000] <jwm>
hybrid p2p

[22:25:00.0000] <jwm>
nat traversal can be done with upnp right?

[22:28:00.0000] <Hixie>
jwm: not on my network

[22:28:01.0000] <Hixie>
upnp is disabled on my router (because it's a HUGE security hole)

[22:31:00.0000] <jwm>
secure when your computers can be trusted :)

[22:31:01.0000] <jwm>
(which is never)

[22:31:02.0000] <jwm>
heh

[01:16:00.0000] <scotfl>
raham: aah! I forgot about sectioning root. doh

[05:51:00.0000] <boblet>
scotfl: what was your comment on sectioning root in reference to?

[09:26:00.0000] <henrikl>
When dealing with legacy systems that depend on RTMP for live streaming of events, is there any support for showing these streams in HTML5?

[11:35:00.0000] <Lachy>
Hixie, Requiem 1.9.4 is out. This should work with iTunes 9.1.1


2010-06-06
[23:28:00.0000] <boblet>
does anyone have an opinion on @aria-describedby and @aria-labelledby in regard to image fallback content (@alt text) and @longdesc?

[23:29:00.0000] <boblet>
I’m guessing that if the image performs a function @aria-labelledby is more appropriate, but otherwise @aria-describedby can be used for content that would have ended up in @alt or @longdesc…?

[23:33:00.0000] <Hixie>
it's hard to give a general guideline

[23:34:00.0000] <Hixie>
depends on the exact situation

[23:34:01.0000] <Hixie>
generally speaking if you write your markup correctly you don't need to use ARIA -- it's only really to be used when writing custom widgets that HTML doesn't support

[23:47:00.0000] <boblet>
Hixie: I’m thinking about using these ARIA features as fallback until <figure> is supported by AT, the ‘belt and suspenders’ approach

[23:48:00.0000] <boblet>
(which sadly might take a while :/)

[23:49:00.0000] <Hixie>
a figure's caption is a "labelledby", but you still need alt text

[23:49:01.0000] <Hixie>
they're separate issues

[23:49:02.0000] <boblet>
oh, just remembered img in figure must have alt text huh

[23:50:00.0000] <boblet>
thanks for the sanity check

[00:24:00.0000] <boblet>
just came across a perfect description of a11y “Hic etiam homines magna cornua habentes longitudine quatuor pedum, et sunt etiam serpentes tante magnitudinis, ut unum bovem comedant integrum” on http://en.wikipedia.org/wiki/Here_be_dragons

[00:24:01.0000] <boblet>
"Here, indeed, are men who have large horns of the length of four feet, and there are even serpents so large, that they could eat an ox whole."

[00:24:02.0000] <boblet>
eloquent *and* accurate

[01:11:00.0000] <boblet>
is the difference between obsolete and ‘obsolete but conforming’ that one generates a validation error and the other a warning? Or is there something else?

[01:31:00.0000] <boblet>
hixie: shouldn’t @headers be in http://www.w3.org/TR/html5/obsolete.html as obsolete, along with axis? want me to file a bug report?

[01:34:00.0000] <boblet>
also, no issue for axis/headers? wow

[01:59:00.0000] <boblet>
Hixie: aah, woops. @headers were specced under td and th, and I was looking at table. my bad

[09:12:00.0000] <sgfgdf>
hello, can someone tell me where should i start if i want to try learning HTML5?

[09:13:00.0000] <micheil>
sgfgdf: there is a good smashingmagazine article on it.

[09:13:01.0000] <micheil>
(as much as smashing mag is normally cruddy, some articles are really good.)

[09:18:00.0000] <sgfgdf>
micheil: i found something, hope it is what you are talking about. one more question - when you think html5 will be usable, i mean is it worth to start learning now or its to early? do you know when it should be released officialy?

[09:18:01.0000] <micheil>
it's usable now, with some work

[09:18:02.0000] <micheil>
html5shiv from what I've heard is good for backwards compat.

[09:18:03.0000] <AryehGregor>
sgfgdf, some features are usable now, some features will probably never be usable (they'll be cut from the spec and never re-added).

[09:19:00.0000] <AryehGregor>
It's a case-by-case thing, it's not going to become usable all at once.

[09:20:00.0000] <sgfgdf>
what are the major supported features far by now, that are usable?

[09:20:01.0000] <micheil>
depends on your definition of supported

[09:21:00.0000] <AryehGregor>
sgfgdf, caniuse.com

[09:21:01.0000] <sgfgdf>
micheil: i mean the last versions of the most popular browsers support.

[09:21:02.0000] <micheil>
sgfgdf: no idea.

[09:25:00.0000] <sgfgdf>
so if someone have a website and want to use html5 is it possible for him to detect if the user's browser supports it and send him the html5 code, and if not send him alternative, or less feature old way html?

[09:25:01.0000] <micheil>
no.

[09:26:00.0000] <micheil>
well, rather, it is possible, but it's the wrong approach.

[09:26:01.0000] <sgfgdf>
so what they do if the website is ery popular, but they want to use the new technologies? give the user a switch if he is sure that he owns capable browser?

[09:27:00.0000] <sgfgdf>
*very popular, sorry

[09:27:01.0000] <AryehGregor>
sgfgdf, you do graceful fallback.  Basically as you said.

[09:27:02.0000] <AryehGregor>
Most HTML5 features support it.

[09:27:03.0000] <AryehGregor>
E.g., <video> with fallback to Flash or Java in non-supporting browsers.

[09:29:00.0000] <sgfgdf>
understand.

[09:29:01.0000] <sgfgdf>
thank you very much for the help guys.

[09:31:00.0000] <flox>
hi

[09:32:00.0000] <flox>
I've some pain installing the html5 validator on Debian

[09:32:01.0000] <flox>
http://paste.pocoo.org/show/222569/

[10:09:00.0000] <AryehGregor>
TabAtkins, what do you think of this?  tkent seems to disagree with me . . . http://code.google.com/p/chromium/issues/detail?id=45804

[10:12:00.0000] <boblet>
sgfgdf: you might also want to try http://html5doctor.com and http://diveintohtml5.org

[10:13:00.0000] <boblet>
also, this interactive presentation is really good for getting an overview: http://apirocks.com/html5/html5.html

[10:15:00.0000] <boblet>
sgfgdf: finally here’s an overview of getting started I wrote: http://oli.jp/2009/html5-structure4/

[10:24:00.0000] <sgfgdf>
boblet: thank you, appreciate it.

[10:24:01.0000] <boblet>
np yo

[10:24:02.0000] <boblet>
if you have questions go for it. there’s normally someone around with time to answer

[10:26:00.0000] <sgfgdf>
boblet: yes, thanks again, btw are you from Japan?

[10:27:00.0000] <boblet>
I’m from New Zealand, but I live in Japan

[10:27:01.0000] <sgfgdf>
okay, i was just curious :).

[10:29:00.0000] <boblet>
np yo :)

[10:32:00.0000] <sgfgdf>
i just didn't even thought how much novelty can bring just one digit - 5. this is amazing.

[10:34:00.0000] <flox>
any idea about the HTML5 validator on Debian Lenny?  http://paste.pocoo.org/show/222569/

[10:35:00.0000] <flox>
I followed the instructions, http://about.validator.nu/#src but no chance

[10:35:01.0000] <boblet>
flox: you prolly wanna talk to MikeSmith or hsivonen 

[10:36:00.0000] <boblet>
dunno about hsivonen but mikesmith won’t be on for at least another 8 hours

[10:36:01.0000] <boblet>
and with that adieu

[10:36:02.0000] <flox>
ok, will try tomorrow :)

[10:36:03.0000] <flox>
good evening

[10:37:00.0000] <boblet>
coming on 2am for me ＞＿＜

[10:37:01.0000] <boblet>
nn

[13:14:00.0000] <drclue>
Howdy all. The other day in my updates came  Chrome 5.0.375.70 beta , and now none of my WebSocket server code works anymore. I've been trying to sort out what changes I need to make , but I'm a wee bit confused. Anyone got a moment to help me sort it out?

[13:23:00.0000] <AryehGregor>
drclue, the Chrome WebSocket implementation recently changed incompatibly to match a new draft.

[13:25:00.0000] <drclue>
I've been fiddling about trying different combinations to see if I could get the WebSeockets working again, and for this conversation I've posted a copy of the request and response so folks can help me figure out what I need to change http://drclue.net/ws.html

[13:25:01.0000] <AryehGregor>
I'm not a WS expert, but I think the entire handshake changed.

[13:26:00.0000] <AryehGregor>
I could be wrong, though.

[13:27:00.0000] <drclue>
@AryehGregor   Ya, I know it changed. I know that it's a spec in flux and I need to respond to it but everything is really vauge in the documentation I've found so far , so I'm just having a problem knowing what I'm supposed to be coding

[13:29:00.0000] <drclue>
Actually the bigger project I'm working is an HTML5 interface to the Asterisk VOIP PBX.  Basically it goes HTML5 => WebSockets => Asterisk AMI => Asterisk.  Everything was working great until the webSockets shifted out from under me, so all my work is stopped until I sort this thing out

[13:30:00.0000] <Hixie>
drclue: the documentation describing the new protocol is http://www.whatwg.org/specs/web-socket-protocol/

[13:30:01.0000] <Hixie>
specifically you want section 5

[13:36:00.0000] <drclue>
@Hixie   I've been looking at that , and I'm still scratching my head a bit , as the request from the browser seems to be somewhat at odds with the documentation , or perhaps I'm just mis-reading something. I do have the request/response sample http://drclue.net/ws.html  that simply reflects my banging my head , but is that request what I'm supposed to be seeing?

[13:44:00.0000] <drclue>
The host and origin documentation definitions seem to be reversed as relates to the header fields the browser is sending me 

[13:49:00.0000] <drclue>
The "host" field in the client request is including the port number , while the "origin" field is not including the port number.    The documentation says that the "origin" field of the client request should have the port number

[13:54:00.0000] <drclue>
@hixie   Am I hallucinating , or does the dump from the chrome client's request not match the spec? hrrp://drclue.net/ws.html 

[13:54:01.0000] <Hixie>
which port?

[13:55:00.0000] <Hixie>
both the stuff from the client and the stuff from the server look wrong on http://drclue.net/ws.html

[13:55:01.0000] <Hixie>
they're both missing the new handshake stuff

[13:55:02.0000] <Hixie>
like the key

[13:56:00.0000] <Hixie>
it should look more like: http://junkyard.damowmow.com/433

[13:56:01.0000] <drclue>
@hixi , I expect the server side to be wrong right now , as I've been messing with it trying to get a clue, but the request comes from chrome

[13:57:00.0000] <Hixie>
if that's what chrome is sending then either you have the wrong version of chrome or chrome is way wrong

[13:58:00.0000] <drclue>
@hixie  Chrome 5.0.375.70 beta on Linux

[14:00:00.0000] <drclue>
I suppose the weird and obviously defective request as dumped from the socket chrome is calling on ,perhaps explains my confusion  

[14:02:00.0000] <Hixie>
well chrome 6.0.422.0 dev on mac sends the right handshake for me

[14:04:00.0000] <Hixie>
and it works with my test server

[14:04:01.0000] <drclue>
I wonder why the request I'm getting is so out of whack?  I still have my original handshake function that worked up until the upgrade , but my understanding is that when they did the upgrade they blew away the old handshake. 

[14:06:00.0000] <Hixie>
maybe 5.0 is before the upgrade?

[14:06:01.0000] <Hixie>
does your old function still work?

[14:07:00.0000] <drclue>
@Hixi Nope , the old handshake died the second after the latest chrome update, so I'm in limbo

[14:08:00.0000] <Hixie>
weird

[14:08:01.0000] <Hixie>
switch to the dev builds instead of beta, i guess :-)

[14:08:02.0000] <drclue>
I have a few of these WebSocket servers , just so I can verify that changes to code are not breaking things and so I can roll back changes

[14:10:00.0000] <Hixie>
bbiab

[14:11:00.0000] <drclue>
Dev builds , ouch! :)   I'll have to think about that one.

[14:12:00.0000] <drclue>
At least I know that I have not gone insane (least not more than I already am)

[14:13:00.0000] <drclue>
I had invented at least 30 new cuss words in the past day or so

[14:32:00.0000] <AryehGregor>
Well, it's an experimental protocol.  :)

[14:32:01.0000] <ment>
drclue: such as? :)

[14:34:00.0000] <drclue>
@AryehGregor   I know , and I'm not bitching that much, but the situation really had me wondering if I had totally lost it. @ment In order for you to properly read the cuss words I would have to cut your eyes out :)

[14:36:00.0000] <drclue>
It's not so bad. I have some other code to work on that has been sitting on the back burner, and I'm sure that a new cut of the browser will have things back on the road soon

[14:39:00.0000] <drclue>
I've been getting a kick out of the HTML5 context , and actually making money with it , which is a pretty good deal

[14:41:00.0000] <ment>
well i've invented few cuss words myself, but they can only be written in runes (ancient, not utf-8) and when spoken, everyone who heard starts caughing blood and spiders

[14:43:00.0000] <drclue>
The last big cuss seems to have resulted in a three story building sinking down into a void in the earth

[14:44:00.0000] <drclue>
Then somebody married Rush Limbaugh   , I was cussin pretty good then too

[14:46:00.0000] <drclue>
A cuss that causes someone to marry Rush, gotta be far worse than relative benign blood and spiders

[14:58:00.0000] <drclue>
One thing that would be sorta neat for WebSockets is a server mode with of course all the "Are you sure" boxes and perhaps a requirement that the code using those sockets be located on the local machine

[14:59:00.0000] <Hixie>
what do you mean?

[14:59:01.0000] <drclue>
@hixie  It would be sorta neat to have other clients connect to my browser

[15:00:00.0000] <Hixie>
search for PeerConnection in the HTML spec on the WHATWG site

[15:00:01.0000] <Hixie>
still extremely early days, but the idea is there

[15:06:00.0000] <drclue>
@hixie Sorta interesting. Have to see how that pans out

[15:09:00.0000] <drclue>
Could end up being some wild applications if millions of browsers were collectively participating

[15:11:00.0000] <Dashiva>
PeerConnection + File system access = Bittorrent? :)

[15:11:01.0000] <drclue>
Please , no file system acess

[15:11:02.0000] <Dashiva>
It's already in the works

[15:12:00.0000] <Dashiva>
Restricted, obviously

[15:13:00.0000] <drclue>
While I like the idea of a WebTorrent , those restrictions would really need to be well thought out to keep the HTML5 context from becoming the bad side of Flash.  

[15:13:01.0000] <Dashiva>
The time to get involved and make sure that doesn't happen is right now :)

[15:15:00.0000] <drclue>
Well , I have registered over at the w3.org , so we'll just have to see how long that takes to go through.  I do like the idea of exploiting torrent clients for design concepts in communications

[15:42:00.0000] <drclue>
@hixie   In 6.3.4. TCP connections is there a logic typo in the port range definition for TCPConnection()?

[15:58:00.0000] <Hixie>
drclue: TCPConnection?

[15:59:00.0000] <drclue>
@hixie http://www.whatwg.org/specs/web-apps/2007-10-26/multipage/section-network.html

[16:01:00.0000] <drclue>
Maybe that page is way out of date or something, but it was what came up in a quick search

[16:01:01.0000] <Hixie>
um, that's from 2007

[16:01:02.0000] <Hixie>
you want http://www.whatwg.org/specs/web-apps/current-work/multipage/


2010-06-07
[00:04:00.0000] <theMadness>
Did a size attribute get dropped from the input type=file?

[01:12:00.0000] <hsivonen>
wow. running svn annotate on the spec is *slow*

[01:13:00.0000] <hsivonen>
/me wishes someone were caching a blame-annotated copy

[01:26:00.0000] <flox>
hello

[01:27:00.0000] <flox>
I retried to install html5 validator locally (Debian Lenny)

[01:27:01.0000] <flox>
still the same error http://paste.pocoo.org/show/222740/

[01:30:00.0000] <MikeSmith>
flox: run the build script one more time

[01:30:01.0000] <MikeSmith>
that error is a know issue

[01:30:02.0000] <MikeSmith>
it always happens the first time you run the build

[01:30:03.0000] <MikeSmith>
you can fix it by running "build/build.py localent"

[01:31:00.0000] <MikeSmith>
or just be re-running "build/build.py all"

[01:31:01.0000] <MikeSmith>
or maybe even just re-running "build/build.py build"

[01:32:00.0000] <flox>
wow ...

[01:32:01.0000] <flox>
yesterday I retried it repeatedly, without success

[01:33:00.0000] <flox>
today it works on 2nd run

[01:34:00.0000] <hsivonen>
/me decides to fix the documentation right now

[01:35:00.0000] <MikeSmith>
flox: dunno what you were doing yesterday, but re-running the build twice with "all" should always work

[01:35:01.0000] <MikeSmith>
flox: but anyway, you should never need to do that again unless/until you do a fresh checkout of the sources

[01:35:02.0000] <Hixie>
hsivonen: there's an out of date one at http://www.whatwg.org/specs/web-apps/current-work/blame

[01:35:03.0000] <Hixie>
hsivonen: i regen it every now and then

[01:36:00.0000] <hsivonen>
Hixie: thanks

[01:36:01.0000] <flox>
MikeSmith: thank you

[01:39:00.0000] <hsivonen>
hmm. the commit message of http://html5.org/tools/web-apps-tracker?from=3194&to=3195&context= doesn't really give hints about why -- > commenent closing was deemed necessary

[01:39:01.0000] <MikeSmith>
flox: yw

[01:40:00.0000] <hsivonen>
flox, MikeSmith: edited http://about.validator.nu/#src

[01:40:01.0000] <Hixie>
not only is the group discussing the html5 license at the w3c a secret group, I can't even subscribe to their mailing list without being appointed a rep by google's AC

[01:40:02.0000] <Hixie>
anyone want to lay bets on the odds of a group that secretive coming up with an appropriate open license for the spec?

[01:41:00.0000] <hsivonen>
Hixie: do you remember the story behind -- > ?

[01:41:01.0000] <Hixie>
i think zcorpan was behind it, but that's just a guess

[01:41:02.0000] <MikeSmith>
hsivonen: doc update looks good. That makes it clear, anyway

[01:43:00.0000] <hsivonen>
Hixie: I'm willing to believe that without reparsing, we'd be worse off if -- > didn't close, but the bad news is that it breaks CNN Money

[01:43:01.0000] <MikeSmith>
hsivonen: would be nice to find somebody with ant or maven chops to rework the build

[01:44:00.0000] <hsivonen>
Whoever edits the source of CNN Money seems to think that putting spaces around -- makes them no longer act as commment delimites

[01:44:01.0000] <Hixie>
hsivonen: and that works in all other browsers?

[01:45:00.0000] <Hixie>
hsivonen: i think this is one of those cases where the three options are 1. x sites break, 2. x other sites break, or 3. we have a security vulnerability

[01:45:01.0000] <hsivonen>
Hixie: it seems that -- > doesn't terminate a comment in Opera 10.5x, IE8 or Chrome beta channel

[01:45:02.0000] <hsivonen>
MikeSmith: https://twitter.com/technomancy/status/10994115673

[01:46:00.0000] <Hixie>
hsivonen: did you test safari?

[01:46:01.0000] <hsivonen>
Hixie: I didn't

[01:46:02.0000] <Hixie>
prolly the same as chrome then

[01:46:03.0000] <hsivonen>
these days, I tend to test WebKit as Chrome, since Chrome runs on Linux but Safari doesn't

[01:46:04.0000] <Hixie>
the webkit guys are implementing the spec too, if they get there soon enough maybe we can just bully our way through under the banner of HTML5 or something

[01:48:00.0000] <Matjas>
Is using the solidus ('/>') required when using the XHTML serialization of HTML5 (by sending the document with Content-Type: application/xhtml+xml)?

[01:48:01.0000] <Hixie>
Matjas: only if you omit the end tag

[01:48:02.0000] <Hixie>
Matjas: in xml you can do <foo></foo> or <foo/>, they mean the same

[01:48:03.0000] <Hixie>
e.g. <img></img> or <img/>

[01:48:04.0000] <MikeSmith>
hsivonen: :) I guess I had assumed ant and maven were actually preferable to however else to put together a build for java apps.. but perhaps now

[01:48:05.0000] <MikeSmith>
*not

[01:48:06.0000] <Matjas>
Hixie: Thanks.

[01:49:00.0000] <Matjas>
Hixie: So <img></img> would be valid HTML5 (in XML mode)?

[01:49:01.0000] <hsivonen>
MikeSmith: well, obviously build.py isn't so nice, either

[01:50:00.0000] <Hixie>
Matjas: well you need src="" and alt="" also, but yeah, from a syntax perspective it's correct in XML

[01:50:01.0000] <hsivonen>
Matjas: <img alt='...' src='...'></img> is valid XHTML5, yes

[01:50:02.0000] <Hixie>
it's just like XHTML1 :-)

[01:51:00.0000] <Matjas>
I knew it was syntactically equivalent, but I didn't know it was actually confirming/valid. Mind = blown. Thanks guys!

[01:55:00.0000] <hsivonen>
Python md5 module deprecation warning is another piece of gratuitous change that humanity could have done without

[01:57:00.0000] <hsivonen>
MikeSmith: speaking of Maven, I need to learn how to make Maven run some Java code supplied by me as part of the build process...

[01:57:01.0000] <hsivonen>
also, the piece of code I have modified a .java file in place instead of having distinct input and output

[01:57:02.0000] <hsivonen>
which is probably considered wrong

[01:58:00.0000] <hsivonen>
/me gestures angrily in HotSpot's general direction

[01:59:00.0000] <MikeSmith>
I seem to run into the "considered wrong" cases often when I write anything at all in Java

[01:59:01.0000] <MikeSmith>
Java ecosystem seems to consider a lot of things wrong

[01:59:02.0000] <hsivonen>
hah

[02:00:00.0000] <Matjas>
Why is it that html5.validator.nu chokes on </img>? http://html5.validator.nu/?doc=http://mathiasbynens.be/edu/test.html&showsource=yes Note that this is in HTML serialization mode.

[02:01:00.0000] <Hixie>
HTML != XHTML

[02:01:01.0000] <Hixie>
</img> is only valid in XHTML

[02:01:02.0000] <hsivonen>
(The reason why I made it modify the file in place is that I don't want to rewrite the identity of the class just to apply a workaround for release binaries

[02:01:03.0000] <hsivonen>
)

[02:01:04.0000] <MikeSmith>
hsivonen: ah

[02:02:00.0000] <MikeSmith>
yeah, that makes sense

[02:02:01.0000] <Matjas>
Fair enough. Now, to trigger XML mode, I just need to serve the file with a header like Content-Type: application/xhtml+xml;charset=UTF-8, right?

[02:03:00.0000] <Matjas>
Or do I need to make other changes to the document as well?

[02:03:01.0000] <Matjas>
http://mathiasbynens.be/edu/test.html (content-type header added) seems to break, I must be doing something wrong.

[02:03:02.0000] <Hixie>
you need xmlns="http://www.w3.org/1999/xhtml" on the <html>

[02:04:00.0000] <Hixie>
you don't need the doctype in xhtml mode

[02:04:01.0000] <Hixie>
but if you have it it has to be uppercase

[02:04:02.0000] <Hixie>
<!DOCTYPE html>

[02:04:03.0000] <Hixie>
frankly though i wouldn't bother using XHTML unless you have a really good reason to :-)

[02:05:00.0000] <hsivonen>
/me wishes the last argument of addEventListener were optional

[02:05:01.0000] <Matjas>
Don't worry, I'm never gonna use XHTML. Just trying to understand how the XML serialization works :)

[02:05:02.0000] <Hixie>
:-)

[02:31:00.0000] <jgraham>
I think the -- > thing caused a bug in Opera or so, but zcorpan will remember

[02:33:00.0000] <hsivonen>
jgraham: caused a bug which way?

[02:33:01.0000] <jgraham>
Not supporting it

[02:33:02.0000] <hsivonen>
jgraham: in an internal build that no longer reparses?

[02:33:03.0000] <jgraham>
Umm, I might be thinking of the wrong thing

[02:34:00.0000] <jgraham>
zcorpan really has a better memory than me

[02:34:01.0000] <hsivonen>
Well, I CCed zcorpan on the Mozilla bug.

[02:45:00.0000] <henrikl>
Quick question: HTML5 + Live streaming video. What options do I have?

[02:53:00.0000] <doublec>
henrikl, browsers that support theora will play live streams using icecast

[02:57:00.0000] <henrikl>
doublec: Ok. Do you know how the lay of the land is with regards to H.264?

[02:57:01.0000] <doublec>
henrikl, unfortuntately not. I believe safari can do it via quicktime but I don't know the details.

[02:58:00.0000] <henrikl>
doublec: Ok, thanks! I`ll just have to set up some tests then. :)

[03:11:00.0000] <nessy>
darwin streaming server for quicktime

[03:11:01.0000] <nessy>
flumotion for theora and webm

[03:13:00.0000] <nessy>
several companies also offer h.264 streaming services using Adobe's Server

[06:53:00.0000] <hsivonen>
http://stackoverflow.com/questions/2972817/what-reasons-could-justify-apples-html5-demos-use-of-html4-doctypes

[06:59:00.0000] <Lachy>
It's good that people are recognising those "HTML5 demos" for what they are as just a marketing gimmick, rather than any serious attempt to showcasing standards.

[07:00:00.0000] <Phae>
the apple ones?

[07:00:01.0000] <Lachy>
yes

[07:00:02.0000] <Phae>
mm.. agreed.

[07:05:00.0000] <hsivonen>
Lachy: some people are, some people aren't

[07:07:00.0000] <Philip`>
Marketing gimmicks don't need to convince everybody in order to be successful

[07:07:01.0000] <Philip`>
I guess they just need to convince more people than the number who react adversely to the realisation that it's a gimmick

[07:08:00.0000] <hsivonen>
Lachy: e.g. Gruber's initial post seemed to take them at face value

[07:09:00.0000] <Lachy>
sure, we can't expect everyone to realise these things.

[07:32:00.0000] <jgraham>
You need to weight by the value of the person e.g. there is no point in convincing lots of people who can't afforrd Apple products that Apple are awesome. But there may be disproportionate weight in convincing percieved tech leaders

[11:25:00.0000] <Dashiva>
"In common usage, HTML5 may also refer to the additional use of CSS3, as both technologies are under development in parallel."

[11:25:01.0000] <AryehGregor>
In that case, presumably it can also refer to the use of SSDs or unmanned attack drones, since those are also under development in parallel.

[11:26:00.0000] <Dashiva>
Well, are you saying wikipedia is imperfect?

[11:26:01.0000] <AryehGregor>
No, that sentence accurately reflects popular use of "HTML5" as a buzzword.

[11:27:00.0000] <Dashiva>
"HTML5 is also used for any web technology that entered the public consciousness after 2005"

[11:27:01.0000] <Philip`>
HTML5 makes my computer faster and kills terrorists?

[11:27:02.0000] <Philip`>
Sounds like good marketing to me

[11:27:03.0000] <AryehGregor>
Well, maybe some unmanned attack drones have a web interface.  I don't know, it's probably classified.

[11:50:00.0000] <zcorpan_>
Philip`: yo

[11:51:00.0000] <zcorpan_>
Philip`: do you have newer dotbot data than 2009-04?

[11:53:00.0000] <zcorpan_>
seems dotbot doesn't have newer data

[11:54:00.0000] <Philip`>
zcorpan_: I'm not aware of any newer data than the year-old release

[11:55:00.0000] <zcorpan_>
i was considering the possibility of running a new research for -- > for https://bugzilla.mozilla.org/show_bug.cgi?id=570309

[11:58:00.0000] <gsnedders>
Philip`: Can you take a look in a similar manner to what you did for script for style with comments?

[12:00:00.0000] <Philip`>
What did I do for script for style with comments?

[12:00:01.0000] <gsnedders>
Parse Error.

[12:03:00.0000] <gsnedders>
/me tries to find that wiki page about comments in script

[12:05:00.0000] <flox>
/me was just reading this part :) http://is.gd/cGaXS

[12:40:00.0000] <jwalden>
"the first shipped implementation may have been developed in stealth mode without peer review during the development process" hah

[12:41:00.0000] <svl>
Hrm, H.264 for this Apple "FaceTime" video calling thing which they say will become "an open industry standard".

[12:43:00.0000] <jgraham>
Y'know I really wish the status markers in the W3C version of the spec weren't quite so ugly

[12:44:00.0000] <jgraham>
I feel somehow responsible

[12:44:01.0000] <jgraham>
flox: No need for URL shorteners here, this isn't twitter

[12:45:00.0000] <flox>
jgraham: it's my IRC client which is configured to shorten URLs

[12:48:00.0000] <AryehGregor>
What a broken IRC client.

[12:48:01.0000] <AryehGregor>
/me is wary of following shortened URLs from people he doesn't know

[12:49:00.0000] <jgraham>
Indeed

[12:50:00.0000] <Philip`>
Why are unshortened URLs any safer than shortened URLs?

[12:50:01.0000] <Philip`>
Unless you know and trust the domain name, it could be anything

[12:50:02.0000] <jgraham>
Philip`: Well that seems like a big win already

[12:51:00.0000] <jgraham>
You might also know and mistrust the domain name

[12:51:01.0000] <jwalden>
twitter will auto-shorten some URLs, in some cases; not sure why they don't just always do it, given the arbitrary length limits and all

[12:53:00.0000] <jgraham>
It also seems to be possible to totally break twitter posts by retweeting them so that the link at the end (i.e. the whole useful contents) gets totally stripped off

[12:53:01.0000] <jgraham>
The success of such a broken service is some kind of lesson

[12:53:02.0000] <AryehGregor>
Philip`, 1) Often I do know and trust the domain name.  2) Even if I'm not worried about a malicious URL, the name often gives a hint to the content.  3) Trolls are more likely to use a stock URL shortener for goatse than to register a whole domain.

[12:53:03.0000] <jgraham>
But I don't know what

[12:54:00.0000] <jgraham>
On the subject of which;

[12:54:01.0000] <jgraham>
gsnedders: You can't

[12:54:02.0000] <jwalden>
e.g. https://twitter.com/jswalden/status/5179465673 and https://twitter.com/jswalden/status/5179505187 demonstrating such a forced-shortening

[12:55:00.0000] <gsnedders>
jgraham: What? Number three?

[12:55:01.0000] <gsnedders>
jgraham: How evil do you think I am, bitch.

[12:55:02.0000] <gsnedders>
Oh, wait, that joke was too late on Saturday.

[12:55:03.0000] <gsnedders>
s/, bitch//

[12:55:04.0000] <jgraham>
gsnedders: Counting from the top down? Yes

[12:56:00.0000] <jgraham>
(one based)

[12:56:01.0000] <jgraham>
And no, I have no idea why you are calling me a bitch

[12:56:02.0000] <gsnedders>
jgraham: I mean the third item in AryehGregor's list.

[12:56:03.0000] <jgraham>
/me is confused

[12:56:04.0000] <jgraham>
But that is OK

[12:56:05.0000] <jgraham>
It's fun

[12:56:06.0000] <AryehGregor>
/me is too.

[12:57:00.0000] <gsnedders>
Uh, so Miss Squeaky was in car giving directions to their flat as if a GPS. I started appending "bitch" on to the end of all the instructions. This may have remained a joke for the rest of the day.

[12:57:01.0000] <gsnedders>
jgraham: I don't entirely understand what I'm being told I can't do either

[12:58:00.0000] <jgraham>
Well that's all good

[13:00:00.0000] <AryehGregor>
Are spaces prohibited in href?  If so, why?  Do browsers not encode them consistently when making the request?

[13:04:00.0000] <Philip`>
AryehGregor: They're prohibited in the URL syntax, I think

[13:04:01.0000] <AryehGregor>
The validator says so.  Do you know why?

[13:04:02.0000] <Philip`>
and it's not HTML's job to redefine the valid URL syntax

[13:06:00.0000] <AryehGregor>
Well, it looks like the browser encodes the spaces as %20.  Something must tell it to do so, unless that's unspecced.

[13:07:00.0000] <AryehGregor>
Anyway, it's HTML's job to say what markup is valid.  It could provide a preprocessing algorithm and then say the URL is valid HTML if the result of the preprocessing is a valid URL.

[13:09:00.0000] <jgraham>
AryehGregor: It does

[13:09:01.0000] <jgraham>
http://www.whatwg.org/specs/web-apps/current-work/multipage/urls.html#valid-url

[13:10:00.0000] <jgraham>
The %20 thing comes from the "resolve an address" algorithm

[13:10:01.0000] <AryehGregor>
Okay, so can anyone actually answer my original question rather than giving me procedural reasons for why the question doesn't make sense?

[13:11:00.0000] <jgraham>
Which is, I think, part of the spec-hot-potato that is the URL specification

[13:11:01.0000] <AryehGregor>
Is there some reason to ban spaces in href/src/etc.?

[13:12:00.0000] <jgraham>
AryehGregor: From HTML's point of view it is because RFC3987 says so

[13:12:01.0000] <jgraham>
I imagine the RFC says tht so they people can split lists of URLs on spaces to get distinct URLs

[13:12:02.0000] <AryehGregor>
Why does HTML have to require that URLs in HTML match RFC3987?

[13:12:03.0000] <jgraham>
for example

[13:12:04.0000] <AryehGregor>
That's incompatible with URL processing in HTML already.

[13:12:05.0000] <gsnedders>
It doesn't

[13:13:00.0000] <gsnedders>
It requires RFCs match RFC3986 or if the document is encoded in UTF-8 or UTF-16 it matches RFC3987

[13:13:01.0000] <gsnedders>
(IIRC)

[13:13:02.0000] <AryehGregor>
s/RFCs/URLs/ I assume?

[13:13:03.0000] <gsnedders>
Yeah

[13:15:00.0000] <jgraham>
gsnedders: That is mising the point somewhat

[13:16:00.0000] <gsnedders>
Of course, it's me.

[13:16:01.0000] <gsnedders>
What do you expect?

[13:16:02.0000] <jgraham>
AryehGregor: I assume because changing the definition of valid HTML would be a huge fight for no particular purpose

[13:16:03.0000] <jgraham>
You don't really gain much by allowing literal spaces

[13:16:04.0000] <gsnedders>
/me doesn't think it was a very common validation error

[13:16:05.0000] <Philip`>
Also you lose compatibility with tools that expect syntactically valid URL

[13:16:06.0000] <Philip`>
s

[13:17:00.0000] <AryehGregor>
I was asking because I noticed the error when validating a site (not one I control).

[13:17:01.0000] <Philip`>
...when they're hooked into an HTML parser and a dumb attribute-value extractor

[13:18:00.0000] <Philip`>
although actually you get the same problem even when not involving HTML, because some sites send spaces in URLs in HTTP redirect headers

[13:18:01.0000] <Philip`>
so I guess those tools should just be fixed

[13:18:02.0000] <gsnedders>
(within the URL, leading/trailing whitepace is allowed there)

[13:19:00.0000] <gsnedders>
(I presume Philip` means)

[13:20:00.0000] <Philip`>
I don't mean leading/trailing whitespace

[13:55:00.0000] <gsnedders>
(Really, how I felt didn't change my behaviour around her that much. The fact that both of us are in general really physically has far more to do with it than anything else.)

[13:56:00.0000] <Philip`>
(You are really physically?)

[13:57:00.0000] <AryehGregor>
I'm not sure if I'd prefer to be physically or mentally.  Maybe spiritually?  Virtually?

[13:57:01.0000] <gsnedders>
Definately mentally.

[13:58:00.0000] <gsnedders>
And where the hell did that come from?

[14:00:00.0000] <gsnedders>
Oh, there.

[14:01:00.0000] <gsnedders>
Heh. IRC script searching through logs going back five years, one sent to the wrong window. Interesting.

[14:01:01.0000] <gsnedders>
/me wonders where that even comes fro

[14:01:02.0000] <gsnedders>
*from

[14:01:03.0000] <gsnedders>
/me thinks his Perl is wrong

[14:02:00.0000] <vlad_>
hey folks, WebIDL question -- if an interface 'operation void foo(in long a);' is called as x.foo() from JS

[14:02:01.0000] <vlad_>
should that throw an exception? since 'a' does not have [Optional], or will the argument be 'undefined' in JS, and then converted to a long (0?)?

[14:03:00.0000] <gsnedders>
Treating it as undefined doesn't apply because it's a host object

[14:03:01.0000] <gsnedders>
I think throw TypeError

[14:05:00.0000] <vlad_>
hmm, ok

[14:05:01.0000] <vlad_>
the WebIDL draft doesn't seem to mention this case at all

[14:06:00.0000] <gsnedders>
It does

[14:06:01.0000] <gsnedders>
http://dev.w3.org/2006/webapi/WebIDL/#call

[14:06:02.0000] <gsnedders>
It falls into step 3 of that

[14:07:00.0000] <vlad_>
well, that depends on what happens in the overload resolution

[14:07:01.0000] <vlad_>
there's even an ed note that says to test how implementations actually behave with too few or too many args :)

[14:08:00.0000] <vlad_>
the question is I guess whether the type list is "undefined" in that case, or is the wrong length

[14:08:01.0000] <gsnedders>
I think it is

[14:11:00.0000] <jgraham>
iirc (whithout reading any specs or anyhting) this ase is underdefined

[14:12:00.0000] <jgraham>
I think the simplest reading is that it throws TypeError, but I don't really trust the spec to be right about that

[14:15:00.0000] <jgraham>
gsnedders is really physically in the sense that he had corporeal existence

[14:15:01.0000] <jgraham>
*has

[14:15:02.0000] <jgraham>
Well I guess had last time I saw him

[14:16:00.0000] <jgraham>
I couldn't say what happened after that

[14:17:00.0000] <gsnedders>
You have evidence of corporeal existence then, proof I wasn't just a ghost?

[14:17:01.0000] <gsnedders>
s/have/had/

[14:22:00.0000] <jgraham>
gsnedders: """Limits of the diaphane. But he adds: in bodies. Then he was aware of them bodies before of them coloured. How? By knocking his sconce against them, sure."""

[14:23:00.0000] <gsnedders>
Hey, I've never finished that book. There again, nor have you. :P

[14:24:00.0000] <jgraham>
Indeed not

[14:24:01.0000] <jwalden>
/me wonders what book that is, and assumes it's some sort of "high-class" literature that is also thoroughly unreadable

[14:24:02.0000] <gsnedders>
jwalden: Ulysses, Joyce

[14:24:03.0000] <jgraham>
jwalden: Yes

[14:24:04.0000] <jwalden>
ah

[14:25:00.0000] <jwalden>
I read some of dubliners, it was interesting but not remarkable unless you were particularly omphaloskeptic in your reading

[14:25:01.0000] <gsnedders>
Do you really expect people to understand omphaloskeptic?

[14:26:00.0000] <jwalden>
c'mon, it's a fun word to use!  :-D

[14:26:01.0000] <gsnedders>
True.

[14:26:02.0000] <jwalden>
/me demands an interesting story first and foremost; if the author wants to layer symbolism underneath that, that's fine, but it shouldn't be necessary to enjoy the book

[14:26:03.0000] <jgraham>
A Portrait of the Artist as a Young Man is excellent and accessible

[14:26:04.0000] <Philip`>
/me demands aliens and lasers and spaceships first and foremost

[14:27:00.0000] <gsnedders>
jwalden: Better yet is something where the symbolism is underneath a good story, and where the implied story is as good as the main one

[14:27:01.0000] <Philip`>
(or dragons)

[14:27:02.0000] <jgraham>
Philip`: But what about qualities yoou like in a book?

[14:27:03.0000] <jwalden>
gsnedders: true

[14:27:04.0000] <Philip`>
Actually, mostly I look for thickness

[14:27:05.0000] <Philip`>
All books seem to be about the same price so I might as well get one with twice as many pages

[14:28:00.0000] <gsnedders>
jgraham: Wow, that photo you uploaded of me on Flcikr makes me look really tired and otherwise terrible

[14:28:01.0000] <gsnedders>
*Flickr

[14:28:02.0000] <gsnedders>
<http://www.flickr.com/photos/jgraham/4679370417/> is the photo in question.

[14:29:00.0000] <jgraham>
gsnedders: Slep more and you won't look so tired 

[14:29:01.0000] <jgraham>
</helpful-advice>

[14:29:02.0000] <jgraham>
*Sleep

[14:29:03.0000] <gsnedders>
jgraham: "In the zoo, the mighty zoo/Geoffrey sleeps today"

[14:30:00.0000] <gsnedders>
There again, songs like that sung by people like me are probably not what you want mentioned :P

[14:30:01.0000] <gsnedders>
jgraham: Also, I slept 19 hours last nigh.t

[14:30:02.0000] <gsnedders>
*night

[14:30:03.0000] <gsnedders>
(and yesterday evening, and this morning)

[14:31:00.0000] <jgraham>
gsnedders: Well I guess you look less tired now

[14:31:01.0000] <gsnedders>
I looked less tired having slept in the car on the way back

[14:31:02.0000] <gsnedders>
Or at least was half asleep with my eyes shut for most of the time

[14:31:03.0000] <jgraham>
Although you should try to reduce the standard deviation

[14:31:04.0000] <gsnedders>
Yeah, totally :(

[14:32:00.0000] <jgraham>
A nice gaussian distribution centred on 8 or so

[14:32:01.0000] <jgraham>
/me realises this doesn't work in the limit of small t

[14:33:00.0000] <jgraham>
(since you can't slep for a negative amount of time)

[14:33:01.0000] <gsnedders>
jgraham: Also, damn you for taking way better photos than me

[14:33:02.0000] <jgraham>
*sleep

[14:33:03.0000] <jgraham>
dammit

[14:34:00.0000] <jgraham>
Really not very good

[14:34:01.0000] <gsnedders>
I didn't say they were very good, I just said way better than mine :P

[14:35:00.0000] <jgraham>
I know, that was just for the benefit of those following along at home

[14:38:00.0000] <AryehGregor>
Where does it say that CSS prefixes are dropped at CR?

[14:39:00.0000] <gsnedders>
Unwritten CSS WG policy, AFAIK

[14:40:00.0000] <AryehGregor>
I'm sure I've seen it written down somewhere.

[14:43:00.0000] <AryehGregor>
http://www.w3.org/TR/css-beijing/

[14:44:00.0000] <gsnedders>
/me gets confused by documents like that being a TR

[14:45:00.0000] <jgraham>
/me just gets confused by documents like that

[14:45:01.0000] <gsnedders>
jgraham: You're always confused.

[14:46:00.0000] <gsnedders>
/me wonders whether it's reasonable to reply to emails saying, "At least you didn't top-quote."

[14:46:01.0000] <gsnedders>
Or even s/At/I might not care, but at/

[14:47:00.0000] <jgraham>
gsnedders: Do you get many emails saying "At least you didn't top-quote"?

[14:47:01.0000] <gsnedders>
jgraham: No.

[14:47:02.0000] <gsnedders>
But I'm unconventional.

[14:47:03.0000] <jgraham>
Well it is quite unreasonable to reply to emails you don't get

[14:48:00.0000] <gsnedders>
I find it far more interesting to do that.

[14:49:00.0000] <jgraham>
I suppose you could create a fictional email to reply to

[14:49:01.0000] <jgraham>
People might be offended hough

[14:49:02.0000] <gsnedders>
I often do.

[14:49:03.0000] <jgraham>
+t

[14:50:00.0000] <gsnedders>
It's fun to see how many people believe the fictional email's authorship.

[14:50:01.0000] <jgraham>
gsnedders: I think you need to get out more

[14:50:02.0000] <gsnedders>
jgraham: You got any plans for Sunday?

[14:51:00.0000] <jgraham>
gsnedders: Not yet

[14:51:01.0000] <jgraham>
(afaik, usual disclimers apply)

[14:51:02.0000] <gsnedders>
/me will send email about Sunday tomorrow/Wednesday, unless beaten to it

[14:53:00.0000] <gsnedders>
(in which case I likely will send an email about it, just a reply)

[14:55:00.0000] <jgraham>
gsnedders: You should spend the time you spend talking about sending email sending email and the time you save doing something fun

[14:55:01.0000] <jgraham>
like making cupcakes

[14:56:00.0000] <gsnedders>
Not bagels, as suggested to me earlier?

[14:56:01.0000] <jgraham>
hmm?

[14:57:00.0000] <gsnedders>
(No, I haven't been online all day.)

[14:57:01.0000] <gsnedders>
jgraham: Steal you partner's Facebook account.

[14:57:02.0000] <gsnedders>
*your

[14:57:03.0000] <gsnedders>
(or, alternatively, get your own)

[14:59:00.0000] <gsnedders>
/me expects he was expected to put burgers inside the bagels

[14:59:01.0000] <gsnedders>
(Just because it alliterates)

[15:03:00.0000] <jgraham>
cupcakes can cunningly conceal cyanaide, causing curtains for consumers

[15:13:00.0000] <AryehGregor>
You couldn't figure out a way around "for"?

[15:14:00.0000] <AryehGregor>
Cupcakes can cunningly conceal cyanide, causing certain consumer cessation.

[16:33:00.0000] <Hixie>
christ

[16:33:01.0000] <Hixie>
the a11y community sure write bad alt text

[16:41:00.0000] <TabAtkins_>
Hixie: I find that too.  Which is very surprising.  Good alt text is ridiculously easy if you know wtf you're doing.

[16:42:00.0000] <TabAtkins_>
/me doesn't get why anyone ever has issues with it, but they do for some reason, so shrug.

[16:43:00.0000] <AryehGregor>
Because normal people have no reason to learn how to write good alt text, because they have so few blind users that it makes no practical difference to them.

[16:43:01.0000] <AryehGregor>
HTML cannot mandate solutions contrary to economics.

[16:43:02.0000] <TabAtkins_>
I've found that good alt text has been helpful for me personally.  I spent quite a while with a broken image once, frex, on a page that I read fairly regularly, because the alt-text was so good that I didn't notice it was broken.

[16:44:00.0000] <TabAtkins_>
/me supposes it's kind of weird that he considers it helpful that it hid the existence of a broken image.

[16:44:01.0000] <TabAtkins_>
I only found out about it when I made my 404 page start emailing me.

[16:45:00.0000] <AryehGregor>
Just because there are isolated cases where it worked well doesn't mean that it would have been worth the effort to learn how, if you weren't involved in standards anyway.

[16:45:01.0000] <AryehGregor>
(so that you need the knowledge for non-authoring purposes)

[16:46:00.0000] <AryehGregor>
TabAtkins_, oh, right, I meant to ask you -- actually, I did ask you -- could you look at this, since you're a Chromium person and all?  http://code.google.com/p/chromium/issues/detail?id=45804

[16:46:01.0000] <TabAtkins_>
True, I acquired the knowledge because of involvement in standards, but my point is that it's *very easy* knowledge.  It takes less than 5 minutes to learn.  This makes it even worse that some of the a11y people can't do it right in their examples, at least.

[16:49:00.0000] <AryehGregor>
It might take less than five minutes to learn how to do it, but people won't learn it anyway unless there's obvious and immediate benefit.

[16:49:01.0000] <AryehGregor>
Most people don't formally learn HTML at all, they just imitate existing use and maybe read some tutorials or skim a book.

[16:50:00.0000] <AryehGregor>
If it looks right to them, then it's good enough.

[16:51:00.0000] <TabAtkins_>
Yeah, I know that.  I'm specifically complaining about a11y people, though.

[16:51:01.0000] <TabAtkins_>
And their bad examples.

[16:52:00.0000] <AryehGregor>
Probably people who do use alt don't bother to learn the correct way because they think they already know it.

[16:52:01.0000] <AryehGregor>
Then it doesn't matter how easy it is to learn.

[16:53:00.0000] <AryehGregor>
I'm wondering who the a11y people at the W3C are.  Are they mostly people who have professional expertise, like working with blind people on a regular basis (or being blind)?

[16:53:01.0000] <AryehGregor>
I suppose I could try Google-stalking a bunch of them, but I'm too lazy.

[16:53:02.0000] <AryehGregor>
Also, I don't really care.

[16:55:00.0000] <AryehGregor>
Hi, dglazkov!  Do you have any thoughts on <http://code.google.com/p/chromium/issues/detail?id=45804>?


2010-06-08
[17:00:00.0000] <TabAtkins_>
AryehGregor: I think that a lot of them are disabled or work closely with the disabled.

[17:10:00.0000] <Hixie>
AryehGregor: i use text browsers often enough that for me it's actually an issue when alt text isn't good

[17:11:00.0000] <volkmar>
Hixie: ping

[17:14:00.0000] <Hixie>
pong

[17:14:01.0000] <Hixie>
anyone know what the mimetype for truetype fonts is?

[17:15:00.0000] <Philip`>
My /etc/mime.types says application/x-font-ttf

[17:21:00.0000] <hober>
there's no officially defined ones, IIRC

[17:21:01.0000] <hober>
though I've seen people pushing font/ttf, font/otf, font/eot, and font/woff lately

[17:22:00.0000] <volkmar>
Hixie: could i have your opinion about @autocomplete related bugs ?

[17:23:00.0000] <volkmar>
just to know if i should go on the direction of bug 9637 or you have a big reason to refuse

[17:23:01.0000] <Hixie>
given how much people push mime you'd think they would maintain the registry so that this kind of thing was easy to answer

[17:23:02.0000] <Hixie>
it's not like fonts are rare or something

[17:24:00.0000] <Hixie>
volkmar: which bug system? mozilla, webkit, w3c...?

[17:24:01.0000] <volkmar>
Hixie: w3c

[17:25:00.0000] <Hixie>
volkmar: i expect we can't change the autocomplete stuff, it's been implemented for more than a decade now

[17:27:00.0000] <volkmar>
Hixie: i thought autocomplete IDL attribute has been introduced with html5

[17:28:00.0000] <Hixie>
volkmar: IE6 didn't distinguish between IDL and content attributes

[17:33:00.0000] <volkmar>
Hixie: hmmm, is that really a blocker ?

[17:33:01.0000] <Hixie>
i dunno

[17:33:02.0000] <Hixie>
if we want to change it though we have to prove it's not a problem to change it, because with something that old, the safest guess is that it's impossible to change

[17:34:00.0000] <volkmar>
the only real difference that would happen between @autocomplete and .autocomplete is when @autocomplete is not set or set to an invalid value

[17:35:00.0000] <volkmar>
i doesn't sound really critic, does it ?

[17:35:01.0000] <Hixie>
/me points to the /topic

[17:35:02.0000] <volkmar>
Hixie: ok

[17:36:00.0000] <Hixie>
the web is a weird palce

[17:36:01.0000] <Hixie>
place

[17:36:02.0000] <Hixie>
people do all kinds of things we don't expect

[17:36:03.0000] <Hixie>
seriously though, unless there's a really important reason to worry about it, i would recommend not touching it

[17:37:00.0000] <volkmar>
Hixie: btw, who/what should use autocomplete except UAs ?

[17:37:01.0000] <Hixie>
how do you mean?

[17:37:02.0000] <Hixie>
it's most commonly used by banks

[17:37:03.0000] <Hixie>
google.com uses it too

[17:38:00.0000] <volkmar>
Hixie: sorry, i wasn't clear

[17:38:01.0000] <volkmar>
i mean, websites _sets_ autocomplete, the value is for the UA, right ?

[17:39:00.0000] <Hixie>
yes

[17:39:01.0000] <volkmar>
then the web shouldn't care about what .autocomplete returns

[17:39:02.0000] <volkmar>
... or that's still too much thinking about the wild web

[17:40:00.0000] <Hixie>
who knows

[17:41:00.0000] <Hixie>
people do crazy things

[17:41:01.0000] <Hixie>
often without good reason

[17:56:00.0000] <ttepasse>
Hm. The new reader feature in HTML 5 rearranges the hX in hgroup.

[17:59:00.0000] <TabAtkins_>
reader feature?

[18:00:00.0000] <ttepasse>
TabAtkins, like the readability-Bookmarklet: http://www.apple.com/safari/whats-new.html#reader

[18:01:00.0000] <TabAtkins_>
Ah, I see.

[18:32:00.0000] <karlcow>
you will have to love the marketing machine

[18:32:01.0000] <karlcow>
HTML 4.01 Support

[18:32:02.0000] <karlcow>
"Safari supports HTML 4.01, the authoring language that defines the structure and layout of web documents."

[18:32:03.0000] <karlcow>
-- http://www.apple.com/safari/features.html

[18:34:00.0000] <Rik`>
karlcow: the whole release feel rushed

[18:34:01.0000] <Rik`>
this is basically the same page as yesterday

[18:35:00.0000] <karlcow>
and they still didn't fix the page about html5 and web standards

[18:35:01.0000] <karlcow>
maybe maciej is fencing with marketing dept in Apple corridors

[18:38:00.0000] <karlcow>
safari 5 install 140,2 Mo of "things" and I would have to restart the computer… hmmm then later.

[18:40:00.0000] <karlcow>
wooot!

[18:40:01.0000] <AryehGregor>
Programs that make you restart the computer are lame.  What are they doing, installing drivers?

[18:40:02.0000] <karlcow>
http://www.w3.org/TR/html5/semantics.html#the-address-element

[18:41:00.0000] <karlcow>
"The address  element represents  the contact information for its nearest article  or body  element ancestor."

[18:41:01.0000] <Rik`>
AryehGregor: updating the webkit framework

[18:41:02.0000] <karlcow>
http://developer.apple.com/safari/library/documentation/AppleApplications/Reference/SafariHTMLRef/Articles/HTMLTags.html#//apple_ref/doc/uid/30001262-SW2

[18:41:03.0000] <karlcow>
address

[18:41:04.0000] <karlcow>
Specifies a street address.

[18:41:05.0000] <karlcow>
????

[18:42:00.0000] <karlcow>
top of the document: "Safari and WebKit implement a large subset of the HTML 5 Specification defined by the World Wide Web Consortium (W3C)."

[18:42:01.0000] <AryehGregor>
I can do stuff like that on Linux without restarting.  Just unlink the files, make the new ones, and hope that if everything gets seriously messed up the user intelligently restarts anyway.

[18:42:02.0000] <AryehGregor>
At least, I get the impression that's what apt-get upgrade does.  Seems to work pretty well, generally.

[18:42:03.0000] <AryehGregor>
Of course, you've got to restart whatever programs were updated.

[18:43:00.0000] <karlcow>
the HTML reference of Apple is quite funky and will do a lot of bad 

[18:43:01.0000] <karlcow>
cite

[18:43:02.0000] <karlcow>
Specifies a citation.

[18:43:03.0000] <AryehGregor>
:/

[18:43:04.0000] <AryehGregor>
They should just cite HTML5 verbatim.

[18:46:00.0000] <karlcow>
at least put a link to the spec for each element

[18:47:00.0000] <GPHemsley>
what does Safari 5's support for the HTML5 elements

[18:47:01.0000] <GPHemsley>
...consist of?

[18:47:02.0000] <GPHemsley>
err, lemme try that again

[18:47:03.0000] <GPHemsley>
what does Safari 5's support for the new HTML5 elements consist of?

[18:48:00.0000] <karlcow>
GPHemsley: see http://krijnhoetmer.nl/irc-logs/whatwg/20100608#l-147

[18:52:00.0000] <karlcow>
Apple-Specific Meta Tag Keys

[18:52:01.0000] <karlcow>
http://developer.apple.com/safari/library/documentation/AppleApplications/Reference/SafariHTMLRef/Articles/MetaTags.html#//apple_ref/doc/uid/TP40008193-SW1

[19:41:00.0000] <MikeSmith>
I'm trying to write up an explanation of the motivation behind the various proposals to add a binary data type to Javascript

[23:00:00.0000] <boblet>
anyone know how Safari5 is implementing Reader? seems given their HTML5 sectioning element support it should use <article> if present, but it’s only using first <section> inside <article> for me

[23:03:00.0000] <boblet>
wonder if the article needs role="main" or something, or if it’s just broken

[23:06:00.0000] <ttepasse>
It uses some strange heuristics. On one site it regroups the hX inside an hgroup.

[23:21:00.0000] <boblet>
so much for their new! support for these elements :/ 

[23:23:00.0000] <boblet>
http://www.focus.com/images/view/11905/ “WTF is HTML5?”

[23:31:00.0000] <aboodman>
MikeSmith: what seems bad about magic iframe?

[23:32:00.0000] <aboodman>
MikeSmith: the fact that you have the same name as a PM on chromium is weirding me out.

[23:33:00.0000] <aboodman>
just fyi

[23:39:00.0000] <hsivonen>
boblet: where did they pull the readiness percentages? seems questionable that Chrome's readiness is so much higher than Safari's in particular.

[23:40:00.0000] <boblet>
hsivonen: small text at the bottom lists caniuse and wikipedia. I wonder if the landscape has shifted since they made it

[23:41:00.0000] <boblet>
bug filed for Safari Radar. I hear through the twitters that it does something wack with titles in hgroup too (yet to verify)

[23:41:01.0000] <boblet>
s/Safari Radar/Safari Reader/

[23:42:00.0000] <ttepasse>
Boblet, an example for the hgroup thingy: http://jeena.net/brother

[23:44:00.0000] <boblet>
ttepasse: thanks. woah, that’s bizarro

[00:17:00.0000] <boblet>
hsivonen: I enjoyed your http://hsivonen.iki.fi/-webkit-html5/ article. another problem is it’s hard for authors to know what vendor prefixes to add. is there a resource listing what CSS things have vendor prefixes? if not can we set a wiki up somewhere?

[00:21:00.0000] <boblet>
hsivonen: also you might want to mention in the conclusion that for bits of the spec that are implemented the same way using vendor prefixes by more than one browser it’s good to add the non-prefixed version too, for future support.

[00:21:01.0000] <boblet>
bbl

[00:40:00.0000] <hsivonen>
boblet: adding the non-prefixed version (which I do on my site) defeats the point of prefixes, too 

[00:41:00.0000] <hsivonen>
boblet: since it creates a body of legacy content that assumes that standardized syntax works the same as the initial prefixed syntax

[00:42:00.0000] <hsivonen>
boblet: as I see it, the only real solution is to speed up the process from having two implementations to the point where the prefix is removed and the CSS WG agrees not to bikeshed the feature anymore

[00:43:00.0000] <hsivonen>
boblet: and to make that happen, vendors should be required to contribute sufficient editor time to the CSS WG when they come up with a CSS extension

[00:43:01.0000] <hsivonen>
boblet: but there's no way to enforce such a person-time contribution rule

[01:23:00.0000] <jgraham>
TabAtkins_: The a11y people seem to have rather different ideas about what constitues ideal alt text. For example they seem to assume that the user should be informed that there is an image, and get a description of the image. Having the page work as if there were no images in the first place doesn't seem to be a goal

[01:23:01.0000] <jgraham>
It would be interesting to test what is most helpful to actual users

[01:30:00.0000] <hsivonen>
jgraham: the "a11y people" aren't a unified "they" on this point

[01:31:00.0000] <jgraham>
hsivonen: Fiar enough

[01:31:01.0000] <jgraham>
*Fair

[01:32:00.0000] <MikeSmith>
hsivonen: I seem to be running into a problem with the v.nu servlet output... Is there some character limit on the length of error messages it will emit?

[01:33:00.0000] <hsivonen>
MikeSmith: of an individual error message? I can't recall.

[01:33:01.0000] <hsivonen>
MikeSmith: is this in Jetty or in Jigsaw?

[01:33:02.0000] <MikeSmith>
Jetty

[01:34:00.0000] <MikeSmith>
running locally from my workspace

[01:34:01.0000] <hsivonen>
MikeSmith: how does the problem manifest itself?

[01:35:00.0000] <MikeSmith>
hsivonen: If I validate a certain file directly from the command-line with Jing, I get the following error message:

[01:35:01.0000] <MikeSmith>
http://paste.plurk.com/show/267869/

[01:36:00.0000] <jgraham>
TabAtkins_: I'm not sure I agree it's easy either. For example I don't see the value of Hixie's alt text for the Blammo Corp example

[01:36:01.0000] <MikeSmith>
but if I try it with my local v.nu install, I get "error: " but with no message text

[01:36:02.0000] <MikeSmith>
but I do get the line number info and markup fragment

[01:36:03.0000] <hsivonen>
MikeSmith: and does the rest of the result page show up?

[01:36:04.0000] <hsivonen>
ok

[01:36:05.0000] <MikeSmith>
hsivonen: yeah

[01:37:00.0000] <Hixie>
jgraham: it's not clear to me that there's much value in making that kind of image accessible either

[01:37:01.0000] <hsivonen>
I'm suspecting a bug in the code that formats the message

[01:37:02.0000] <jgraham>
TabAtkins_: Whatever the photograph is trying to communicate, it is not "Our CEO is tall and has brown hair and a moustache"

[01:37:03.0000] <Hixie>
jgraham: my argument was not that my suggestion made it useful, but that it made it accessible

[01:37:04.0000] <MikeSmith>
hsivonen: OK, I can play around it some more and see if I can figure anything out on my own

[01:37:05.0000] <hsivonen>
jgraham: is that an actual alt text?

[01:38:00.0000] <jgraham>
hsivonen: No, it is a hypothetical one

[01:38:01.0000] <MikeSmith>
hsivonen: fwiw, I get no indication of a problem on the console log from v.nu

[01:38:02.0000] <jgraham>
Hixie: I think that is too reductionist a viewpoint

[01:38:03.0000] <jgraham>
I don't think you can say "is this image accessible"

[01:38:04.0000] <jgraham>
Only "is this content accessible"

[01:38:05.0000] <jgraham>
where content === whole page, typically

[01:39:00.0000] <Hixie>
jgraham: well that's as may be, but bruce was applying it to the figure, which is what i applied it to

[01:39:01.0000] <hsivonen>
jgraham: http://krijnhoetmer.nl/irc-logs/html-wg/20070927#l-332

[01:40:00.0000] <jgraham>
Hixie: I think my point is that the thread is trying to look at <figure>s in isolation, which is a silly thing to do

[01:40:01.0000] <Hixie>
jgraham: personally i'd be fine with not making the photograph accessible, since it's almost certainly purely decorative in context, but we shouldn't pretend that that means it's accessible

[01:40:02.0000] <Hixie>
jgraham: couldn't agree more

[01:40:03.0000] <jgraham>
hsivonen: Indeed

[01:40:04.0000] <jgraham>
:)

[01:46:00.0000] <Hixie>
(i'd omit the alt attribute in the case)

[01:54:00.0000] <boblet>
translation q: can anyone guess at what this means in English? “the pashto number in list taq” I’ve got list taq prolly being <li>, but have no idea what pashto might be

[01:55:00.0000] <Hixie>
Pashto is another name for the Afghani language

[01:57:00.0000] <boblet>
Hixie: you’re a scholar and a gentleman. will check CSS language support for list counters

[01:58:00.0000] <boblet>
aah GIYF (although it’s GIMF this time)

[02:02:00.0000] <Hixie>
boblet: wish i could claim credit... i just typed "pashto" into google and read the first hit (wikipedia) :-)

[02:02:01.0000] <Andros>
Semantic question: what sets a heading apart from a dt? The dt is defined as a term part of a term-description group. But isn't also a heading in a way a term, describing the content that follows it?

[02:03:00.0000] <Hixie>
Andros: you wouldn't put the terms of a term-description group into a book's table of contents

[02:03:01.0000] <Hixie>
Andros: and you wouldn't put the headings of a book into its index

[02:03:02.0000] <hsivonen>
Andros: the expectation is that headings appear in an outline but dts don't

[02:03:03.0000] <boblet>
Hixie: still drinking coffee so am apparently a little slow on the Google, but I’m now reading CSS3 Lists which you edited :)

[02:03:04.0000] <Hixie>
boblet: :-)

[02:09:00.0000] <Andros>
hsivonen & hixie: I think I see what you're meaning. 

[02:15:00.0000] <jgraham>
/me wonders if he is the only person trying not to respond to "what is not possible in HTML5" with "faster than light travel"

[02:16:00.0000] <gsnedders>
But what if the bits used by HTML5 have no mass?

[02:16:01.0000] <Hixie>
jgraham: has that thread gotten in need of some gardening?

[02:16:02.0000] <Hixie>
i haven't been paying close attention...

[02:16:03.0000] <Hixie>
gsnedders: you'd need negative mass to get faster than light

[02:16:04.0000] <gsnedders>
/me is around a month behind on email

[02:16:05.0000] <Hixie>
(which doesn't make much sense)

[02:16:06.0000] <gsnedders>
Hixie: Ah, true. Duh.

[02:16:07.0000] <Hixie>
/me hopes he's remembering his relativity maths well enough

[02:17:00.0000] <gsnedders>
You are.

[02:17:01.0000] <gsnedders>
(I think. jgraham will now say we're both wrong.)

[02:18:00.0000] <jgraham>
No, once you get down to zero mass things get "interesting" but you can't travel faster than light

[02:19:00.0000] <jgraham>
(light itself has zero rest mass for example)

[02:19:01.0000] <gsnedders>
Surely it therefore also has zero relativisty mass for all values of v?

[02:20:00.0000] <hsivonen>
and antimatter has still positive mass?

[02:20:01.0000] <jgraham>
Yes

[02:20:02.0000] <jgraham>
/me should point out the implied "in a vacuum")

[02:20:03.0000] <jgraham>
(since you can travel faster than light in a non-vacuum)

[02:20:04.0000] <Hixie>
negative mass is basically a meaningless concept like time before the big bang, if i'm not mistaken

[02:21:00.0000] <gsnedders>
/me doesn't care about places apart from in a vacnuum on a frictionless bench

[02:21:01.0000] <hsivonen>
/me is reminded of xkdc with physicists in a frictionless vacuum

[02:22:00.0000] <Hixie>
that xkcd is such a lie

[02:22:01.0000] <Hixie>
the professor and the laptop would both more or less disintegrate if friction was zero

[02:22:02.0000] <gsnedders>
http://xkcd.com/669/

[02:23:00.0000] <Hixie>
/me notes it's basically the first hit on google for "xkdc with physicists in a frictionless vacuum" [sic], and thus not hard to find :-P

[02:23:01.0000] <gsnedders>
/me may have done that

[02:25:00.0000] <jgraham>
Hixie: I think you are assuming a particular way of changing the laws of physics

[02:26:00.0000] <Hixie>
a way that doesn't involve changing the laws of biology and engineering? :-)

[02:27:00.0000] <gsnedders>
/me notes his sister and niece have vanished

[02:27:01.0000] <gsnedders>
/me wonders when they'll return home

[02:30:00.0000] <gsnedders>
And I was expecting, the moment I said that, they come back.

[02:30:01.0000] <gsnedders>
It worked!

[02:39:00.0000] <gsnedders>
I tell my niece off for reaching for a wine bottle and now she looks scared of me.

[02:46:00.0000] <MikeSmith>
hsivonen: http://www.w3.org/html/check?doc=https%3A%2F%2Feyeasme.com%2FJoe%2FMathML%2FHTML5%2Fextras.xhtml&showsource=yes

[02:46:01.0000] <MikeSmith>
hsivonen: above shows the error-message-missing problem I was describing earlier

[02:47:00.0000] <hsivonen>
MikeSmith: I suggest stepping through the message formatter in a debugger

[02:47:01.0000] <MikeSmith>
ok

[02:48:00.0000] <MikeSmith>
hsivonen: bt, I notice also that the error is caused by the fact that he has an <annotation-xml> element as the first/only child of <semantics> element

[02:48:01.0000] <MikeSmith>
the MathML 2.0 DTD and the existing MathML 2.0 schema that v.nu is using seem to consider that valid

[02:49:00.0000] <MikeSmith>
but the MathML 3.0 schema and the MathML 2.0 one that I'm testing with, which is generated from that, consider it invalid

[02:50:00.0000] <hsivonen>
MikeSmith: parser bugs or spec bugs are possible, too, as causes for the messages

[02:50:01.0000] <MikeSmith>
OK

[02:50:02.0000] <hsivonen>
/me steps away to perform accounting rituals

[02:50:03.0000] <MikeSmith>
hsivonen: btw, what I have on http://www.w3.org/html/check right now is the generated MathML 2.0 schema that David Carlisle provided

[02:51:00.0000] <gsnedders>
/me wants a good feed reader for OS X that actually manages to resolve URLs properly

[02:51:01.0000] <gsnedders>
and has a simple UI.

[03:20:00.0000] <jgraham>
(FWIW I think that SteveF's point about the definition of <figure> is reasonable)

[03:20:01.0000] <jgraham>
(a figure is just a piece of content with a caption)

[03:20:02.0000] <Hixie>
file a bug

[03:21:00.0000] <Hixie>
if there's a bug

[03:21:01.0000] <gsnedders>
If there's a bug, file one?

[03:21:02.0000] <gsnedders>
Huh?

[03:21:03.0000] <gsnedders>
/me thinks there's some missing negative here

[03:21:04.0000] <Hixie>
if there's a problem, file an entry into the database

[03:21:05.0000] <jgraham>
gsnedders: It made sense to me

[03:22:00.0000] <gsnedders>
Oh, I thought you meant if there was a bug, file a duplicate.

[03:22:01.0000] <gsnedders>
Not if a bug exists

[04:40:00.0000] <annevk>
/me would have expected at least some discussion on http://www.w3.org/QA/2010/06/xhtml_modularization_a_markup.html

[04:43:00.0000] <jgraham>
annevk: Not sure what you would say really

[04:44:00.0000] <hsivonen>
annevk: It's interesting that one of the most important technologies of 2001 is only a way of writing schemas

[04:44:01.0000] <annevk>
anything from "joy" to "who cares" I suppose

[04:45:00.0000] <annevk>
but I think that blog post is a justification for them continuing to publish XHTML Modularization crap without testing it in actual user agents

[04:45:01.0000] <hsivonen>
annevk: sure

[04:46:00.0000] <annevk>
despite the problems pointed out

[04:46:01.0000] <hsivonen>
the timing may have something to do with the PER still being in the limbo: not publicly transitioned to REC nor rejected as a Note.

[04:47:00.0000] <hsivonen>
If it were clearly REC-worthy, there wouldn't be a need to post justifications like that, and the Director probably would already have approved the transition to REC.

[04:52:00.0000] <hsivonen>
http://lists.w3.org/Archives/Member/w3c-ac-members/2010AprJun/0049.html

[04:52:01.0000] <hsivonen>
(sorry about the link to secret stuff)

[04:55:00.0000] <annevk>
anyone here planning to go to Maastricht end of July?

[04:55:01.0000] <annevk>
IETF has a meeting

[04:55:02.0000] <annevk>
should not be too hard to plan my as-of-yet non-existing vacation plans around the meeting and attend

[04:56:00.0000] <hsivonen>
http://www.apple.com/safari/whats-new.html#html5

[04:56:01.0000] <annevk>
though I'm not too thrilled about going either (apart from visiting Maastricht, it's quite nice) as the IETF meetings I attended over the phone have been nothing but utter boredom

[04:56:02.0000] <hsivonen>
is there documentation on how Safari supports closed captions?

[04:57:00.0000] <nessy>
do they?

[04:57:01.0000] <Lachy>
I love how they blatantly admit that "XHTML Modularization is not [...] implemented in browsers"

[04:57:02.0000] <hsivonen>
nessy: claimed at the URL I just pasted

[04:57:03.0000] <Lachy>
(I realise that's quote mining, but still funny)

[04:57:04.0000] <annevk>
I suppose they support whatever closed captions H264 supports

[04:57:05.0000] <nessy>
I would assume it's only those that come from within MPEG-4

[04:58:00.0000] <nessy>
as annevk says :)

[04:58:01.0000] <hsivonen>
/me notes that the URL also links to the "HTML5" demos that actually demo WebKit-specific CSS features

[04:58:02.0000] <annevk>
nessy, I guess you're more correct, as H264 is just the video stream :)

[04:59:00.0000] <Lachy>
My guess is that Safari will support whatever caption formats that QuickTime supports

[04:59:01.0000] <nessy>
IIRC that would then be decoded by quicktime and probably rendered by quicktime, too

[04:59:02.0000] <Lachy>
it's kind of a shame how they're going ahead with their own caption formats now, rather than waiting for WebSRT

[04:59:03.0000] <hsivonen>
/me was interested in actual docs/demos

[04:59:04.0000] <nessy>
we should just ask Eric

[05:00:00.0000] <nessy>
Lachy: I don't think it covers external caption formats, so it doesn't mean anything for WebSRT

[05:00:01.0000] <hsivonen>
It seems that the Decision Process is useless for actually putting ISSUEs behind us and moving on.

[05:02:00.0000] <nessy>
what Decision Process?

[05:02:01.0000] <Lachy>
nessy, the HTML WG formal decision process

[05:03:00.0000] <nessy>
ah, I thought we were talking WHATWG

[05:03:01.0000] <nessy>
 I do miss the email discussion of WebSRT a bit - might still be ahead of us

[05:03:02.0000] <Lachy>
the one that is seemingly only good at letting people drag issues on forever, despite supposedly coming to WG decisions

[05:05:00.0000] <annevk>
nessy, it's not finished yet

[05:05:01.0000] <nessy>
ah cool - I've totally lost touch in the last weeks - too busy with other stuff

[05:05:02.0000] <annevk>
(as in, the first draft)

[05:06:00.0000] <jgraham>
/me notes that the whole everything-is-HTML5 thing is the unexpectedly sour taste of victory (to quote Green Wing), so it's probably not worth getting to upset about

[05:18:00.0000] <Lachy>
jgraham, can it really be considered a victory when what people are referring to as being HTML5, pretty much includes everything but what acutally is HTML5, with few exceptions?

[05:20:00.0000] <Lachy>
HTML5 as just become the new term that replaces terms like DHTML and AJAX

[05:20:01.0000] <hsivonen>
Lachy: Safari 5 does better by supporting "HTML5 AJAX history" :-)

[05:20:02.0000] <Lachy>
haha

[05:24:00.0000] <Lachy>
hmm. They also claim to support article, section, etc. But they don't support the proper heading level algorithm, so that's really only a half truth.

[05:24:01.0000] <jgraham>
Lachy: That's not really true. We have the two biggest tech brands on the planet fighting to leverage the mindshare that HTML5 has. Compare to http://www.456bereastreet.com/archive/200705/help_keep_accessibility_and_semantics_in_html/

[05:25:00.0000] <jgraham>
If going from "you might be better off not using HTML5" to Apple trying to make anything and everything "HTML5" because it has such positive assosiations isn't victory I don't know what you were expecting

[05:26:00.0000] <jgraham>
s/Apple/Apple and Google/

[05:28:00.0000] <Lachy>
Sure, it's great that they're promoting HTML5, but it would be better if they would use the proper terminology for things that aren't in HTML5 instead of lumping it all together

[05:28:01.0000] <Philip`>
jgraham: Marketing isn't victory

[05:28:02.0000] <jgraham>
Lachy: Indeed. Like I said, the unepectedly sour taste of victory.

[05:28:03.0000] <Philip`>
I guess victory is having good technology and having lots of people use it (and marketing just helps with the latter)

[05:29:00.0000] <jgraham>
Philip`: On the classic scale of "First they ignore you, then they fight you, then you win", we are past "then they fight you"

[05:29:01.0000] <hsivonen>
jgraham: do you count XHTML-as-text/html as a victory for XML?

[05:31:00.0000] <jgraham>
hsivonen: Maybe a pyrrhic victory

[05:31:01.0000] <Philip`>
jgraham: I don't think everyone would agree with that scale - e.g. everyone who's been in a fight and then lost

[05:32:00.0000] <Lachy>
jgraham, I would be more comfortable calling it a victory if the things that apple were promoting as HTML5 were actually widely interoperable, rather than including things like h.264/AAC video/audio, CSS 3D Transforms, etc.

[05:34:00.0000] <zcorpan_>
i wonder how safari decides that a page is an 'article' and enables the Reader button

[05:34:01.0000] <Lachy>
and if their demos for features that are more interoperable, like WebFonts and CSS transitiosn, actually functioned in other browsers

[05:35:00.0000] <hsivonen>
Is Apple supporting WOFF yet?

[05:35:01.0000] <Rik`>
zcorpan_: apparently : http://idzr.org/ah242k

[05:36:00.0000] <annevk>
not many changes anymore to HTML5 based on what I had to write down for html5-diff

[05:36:01.0000] <hsivonen>
Rik`: is it only <div> or <article>, too?

[05:37:00.0000] <hsivonen>
it would be a bit sad to advertise support for <article> but not to support that feature for <article>

[05:37:01.0000] <Lachy>
I'm not sure. I just assumed their typography demo was using webfonts http://www.apple.com/html5/showcase/typography/

[05:37:02.0000] <Rik`>
hsivonen: don't know

[05:37:03.0000] <Lachy>
I didn't look at the source yet

[05:37:04.0000] <hsivonen>
Lachy: always look at the source

[05:37:05.0000] <Lachy>
hsivonen, that takes time

[05:37:06.0000] <Lachy>
I was lazy

[05:37:07.0000] <hsivonen>
Lachy: they have the fonts in multiple formats

[05:39:00.0000] <hsivonen>
/me wishes Apple added WOFF and TTF support to iOS to put make the need to use SVG fonts moot

[05:40:00.0000] <Lachy>
oh, interesting. They've added EOT, TTF, WOFF and SVG fonts, and yet still decided to make the demo Safari only.

[05:41:00.0000] <hsivonen>
/me suspects there are at least two different theys at work at Apple

[05:42:00.0000] <annevk>
they support EOT? oh god

[05:42:01.0000] <hsivonen>
annevk: in the demo

[05:42:02.0000] <hsivonen>
annevk: but they block IE from accessing it

[05:42:03.0000] <annevk>
oh, I thought to Safari

[05:46:00.0000] <Lachy>
have Microsoft announced which formats they will support in IE9? I assume they'll keep EOT. Will they add WOFF, TTF or OTF?

[05:50:00.0000] <zcorpan_>
Rik`: interesting. they claim to also support multipage articles

[05:50:01.0000] <Rik`>
zcorpan_: and they do. it's really well done

[05:51:00.0000] <Rik`>
that reminds me the fast forward feature of Opera

[05:51:01.0000] <Rik`>
zcorpan_: http://arstechnica.com/apple/news/2010/06/safari-5-faster-less-clutter-secure-browser-extensions.ars

[05:53:00.0000] <hsivonen>
whoa? you need to join a developer program to make extensions?

[05:53:01.0000] <hsivonen>
how different are Safari extensions from Chrome extensions?

[05:54:00.0000] <Rik`>
hsivonen: I'm not sure you need to join the developer program if you distribute them yourself

[05:54:01.0000] <hsivonen>
I guess Tiger is now clearly EOLed

[05:55:00.0000] <Rik`>
there is Safari 4.1 for Tiger, so not that dead

[05:55:01.0000] <hsivonen>
It still annoys me that Apple never says when a Mac OS X release is EOLed

[05:55:02.0000] <Lachy>
hsivonen, the update in that article says you can get extensions from 3rd party sites too. So no AppStore-like lock in

[05:55:03.0000] <hsivonen>
Rik`: whoa!

[05:55:04.0000] <hsivonen>
Rik`: so they now update the browser but no longer provide security patches for the OS?

[05:56:00.0000] <hsivonen>
is it documented anywhere that the advertised Safari features don't work on Windows unless the user chooses the Safari+QuickTime download option?

[05:56:01.0000] <hsivonen>
is providing the non-QuickTime download an antitrust thing?

[05:57:00.0000] <Lachy>
presumably, if you already have QuickTime installed, you don't need to get it again

[05:57:01.0000] <Rik`>
hsivonen: Safari 4.1 is apparently just an updated webkit version plus some security fixes afaict

[05:58:00.0000] <Rik`>
hsivonen: for Safari vs Chrome extensions http://developer.apple.com/safari/library/documentation/UserExperience/Conceptual/SafariExtensionsConversionGuide/

[05:59:00.0000] <hsivonen>
Rik`: thanks

[06:00:00.0000] <annevk>
MikeSmith, yt?

[06:00:01.0000] <MikeSmith>
annevk: yep

[06:00:02.0000] <annevk>
MikeSmith, I'm removing http://dev.w3.org/html5/html4-differences/#webarch

[06:00:03.0000] <annevk>
MikeSmith, the source code says you are using it, so you may want to copy it and put it somewhere safe

[06:00:04.0000] <MikeSmith>
/me looks

[06:00:05.0000] <MikeSmith>
I'm not using it anymore

[06:00:06.0000] <MikeSmith>
so fine by me

[06:01:00.0000] <annevk>
kk

[06:10:00.0000] <MikeSmith>
annevk: have you run the diff doc through pubrules?

[06:49:00.0000] <annevk>
MikeSmith, no

[06:49:01.0000] <MikeSmith>
ok

[06:49:02.0000] <annevk>
MikeSmith, I was addressing Marcos' comments

[06:49:03.0000] <MikeSmith>
ok

[06:50:00.0000] <annevk>
MikeSmith, it's not yet marked-as-WD; trying to find if I missed any comments but it seems I'm done

[06:51:00.0000] <MikeSmith>
when you do get to pubrules, let me know if you have any problems

[06:51:01.0000] <annevk>
are there any new rules or something?

[06:51:02.0000] <annevk>
because normally I'm automatically compliant minus a few things  that are bugs in pubrules

[06:54:00.0000] <annevk>
ah, copyright date was wrong

[06:55:00.0000] <MikeSmith>
we have been having some problems with pubrules lately

[06:55:01.0000] <annevk>
seems someone magically fixed that while publishing

[06:55:02.0000] <annevk>
last time

[06:55:03.0000] <MikeSmith>
I might have changed that

[06:55:04.0000] <MikeSmith>
dunno

[06:55:05.0000] <annevk>
anyway, it's all good now I think

[06:55:06.0000] <annevk>
at least in the same state it was last time, minus the copyright bug

[06:55:07.0000] <annevk>
http://dev.w3.org/html5/html4-differences/

[06:55:08.0000] <annevk>
if people find more issues  today I can probably fix them

[06:56:00.0000] <annevk>
and depending on when everything is moved over I can also fix things tomorrow or the date of publication

[06:57:00.0000] <MikeSmith>
ok

[06:57:01.0000] <MikeSmith>
hsivonen: I found the cause of the missing-error-message bug

[06:58:00.0000] <MikeSmith>
it was due to the code I added to the message emitter recently

[06:58:01.0000] <MikeSmith>
surprise surprise

[07:01:00.0000] <hsivonen>
MikeSmith: :-)

[07:02:00.0000] <hsivonen>
MikeSmith: I guess I suck at code review

[07:02:01.0000] <MikeSmith>
I think it was easy to miss

[07:12:00.0000] <hsivonen>
so which version of Web Sockets did Safari 5 ship?

[07:18:00.0000] <boblet>
hey all, re: <dl> I’ve had two ppl say today that dt/dd associations need an explicit wrapping element. one said it’s annoying to style and another said it’s hard to manipulate in Javascript. would the solution just to be make each name-value association an individual <dl>? so a description list becomes a list of multiple <dl> elements, each wrapping one name-value association

[07:18:01.0000] <hsivonen>
boblet: in theory that breaks semantics

[07:19:00.0000] <hsivonen>
boblet: Not sure whether it's a Real Problem, though.

[07:19:01.0000] <boblet>
hsivonen: in the styling case you’d need to apply classes to both dt and dd to style a name-value association

[07:22:00.0000] <boblet>
I’m guessing that we’re not able to put in eg a wrapping div to style a name-vale association (I think that’s what this means: “If a dl element contains non-whitespace text nodes, or elements other than dt and dd, then those elements or text nodes do not form part of any groups in that dl.”)

[07:35:00.0000] <annevk>
hmm, actually, IETF meetings are expensive

[07:35:01.0000] <annevk>
unless lots of friends are going to hang out in Maastricht I'm not going I think

[07:41:00.0000] <zcorpan_>
hsivonen: -75

[07:43:00.0000] <annevk>
hopefully they fix that in a .1

[07:43:01.0000] <jgraham>
kinda ironic given that othermaciej was instrumental in designing -76

[07:50:00.0000] <hsivonen>
zcorpan_: sigh. that sucks big time.

[07:52:00.0000] <zcorpan_>
yes

[07:52:01.0000] <annevk>
really?

[07:53:00.0000] <annevk>
Chromium dev already changed, nobody else has shipped, ...

[07:54:00.0000] <annevk>
and it's not like it's clear that post -75 is in a fixed state...

[07:54:01.0000] <annevk>
see e.g. the whole SSL handshake discussion

[07:54:02.0000] <hsivonen>
/me isn't reading hybi

[07:54:03.0000] <annevk>
prolly a smart thing

[07:55:00.0000] <annevk>
that was the first sensible discussion since a long time

[07:55:01.0000] <jgraham>
annevk: Depends if this is the same as whatever will ship on the iPhone / iPad

[07:55:02.0000] <jgraham>
People will make sites that sepend on those

[07:55:03.0000] <jgraham>
*depend

[07:55:04.0000] <jgraham>
Desktop Safari not si much

[07:55:05.0000] <jgraham>
*so

[07:57:00.0000] <Rik`>
I'm really concerned about the QA done with Chrome and Safari

[07:58:00.0000] <Rik`>
they don't mention the UI for <input type="number/date"> in the release notes which let me think it wasn't tested

[08:01:00.0000] <jgraham>
Rik`: It's not clar that that follows

[08:01:01.0000] <jgraham>
clear

[08:01:02.0000] <jgraham>
Sigh. I should try pressing the right keys

[08:02:00.0000] <Rik`>
jgraham: on http://www.apple.com/safari/whats-new.html, there are 7 lines for sectionning that adds nothing except parsing rules and display: block

[08:02:01.0000] <Rik`>
this was added to WebKit by Apple

[08:03:00.0000] <Rik`>
the new UI for forms was added by Google and there is no mention

[08:03:01.0000] <jgraham>
Rik`: Often QA and Marketing are not working that closely together

[08:04:00.0000] <Rik`>
sure

[08:05:00.0000] <Rik`>
but mentioning form validation and not the UI feels strange to me

[08:07:00.0000] <jgraham>
Rik`: Indeed. but better to test than to speculate

[09:27:00.0000] <TabAtkins>
jgraham: The purpose of the photograph is to provide a physical depiction of the CEO. Thus in this case a description is appropriate, as you need the caption to be referring to something (in most other cases it is purely decorative and so should receive the empty string).

[09:31:00.0000] <jgraham>
TabAtkins: That seems unlikely. I mean the purpose of the photograph is probably to make the company seem "more human" so stimulating a positive emotional response, or to make the CEO feel impoartant, or something

[09:32:00.0000] <jgraham>
it is very unlikely to be "to help the users identify the CEO in an identity parade"

[09:32:01.0000] <gsnedders>
Oh, is that the point of those photos?

[09:32:02.0000] <gsnedders>
/me have never got any point of those photos

[10:34:00.0000] <TabAtkins_>
Who's defining DOM 0 right now?

[10:34:01.0000] <jgraham>
TabAtkins: Hahahaha

[10:35:00.0000] <jgraham>
(that's the short answer)

[10:35:01.0000] <jgraham>
(the long answer is zcorpan and gsnedders worked on Web DOM Core but don't have time to pursue it, so no one)

[11:50:00.0000] <volkmar>
is there a reason why valueAsNumber is a double instead of a float ?

[11:55:00.0000] <Dashiva>
volkmar: I'd assume it's because javascript numbers are doubles

[11:59:00.0000] <volkmar>
Dashiva: why not but then there is a lake of consistency as progress element has only float attributes

[12:00:00.0000] <volkmar>
and the specs say we should apply "the rules for parsing floating point number values"

[12:05:00.0000] <Dashiva>
Hmm, so valueAsNumber is the only use of double?

[12:05:01.0000] <volkmar>
afaict

[12:07:00.0000] <volkmar>
Dashiva: even if it could be great to have double for progress.position

[12:09:00.0000] <Dashiva>
It seems valueAsNumber is used for date-based inputs, using float there wouldn't be future-proof I guess?

[12:10:00.0000] <Dashiva>
It used to be float, but was increased to double on request

[12:13:00.0000] <volkmar>
Dashiva: then, it explains why the spec say we should parse it with the float rule

[12:14:00.0000] <volkmar>
i suppose it should be changed if we keep float

[12:14:01.0000] <volkmar>
s/we keep float/you keep double/

[12:53:00.0000] <drclue>
OK, I switched over to Chrome 6.0.422.0 dev (Linux) to work out the new WebSocket handshake.  By using injected inputs of both the link Hixie gave me http://junkyard.damowmow.com/433 and the examples in the spec , I get the expected results , but when I use the request from Chrome , Key 3 looks odd as does the result and of course it does not work.     I've posted examples of the input and resulting output for all three cases at h

[12:55:00.0000] <AryehGregor>
drclue, cut off after "all three cases at".

[12:55:01.0000] <annevk>
http://hg.diveintohtml5.org/hgweb.cgi/rev/31e07449843a7982c119bc7fe2c69b595a7e46f5 -- did someone tell hsivonen? :)

[12:55:02.0000] <drclue>
...all three cases at http://www.drclue.net/ws.html     The only thing I'm modifying in my code is the handshake , and the code did work before, and does work with the example requests, just not in Chrome.   The only thing that strikes me as odd are the values for the fields "host" and "origin" , which sorta seem transposed in the request as relates to what I think I'm reading in the spec

[13:03:00.0000] <drclue>
In the spec it says that the "origin" request field should contain the scheme, hostname, and port  , but the chrome request only has the scheme and hostname , and not the port.  The "host" request field has the port tacked on the end though, and I'm not even sure that should be there

[13:05:00.0000] <drclue>
The only time the "origin" request field is supposed to exclude the port reference is if it is the default port, but I'm using port 10812

[13:06:00.0000] <drclue>
So am I just missing something obvious , or is something besides me broken here?

[13:17:00.0000] <jgraham>
drclue: afaict both host and origin should include non-defaul ports

[13:17:01.0000] <kpx>
Hey I am new to JS and I just wrote a small blog on Webstorage and HTML5 that I had studied in the past few days... Now I wanted to know some good place where I can post it so I could have some discussion and FB.... 

[13:18:00.0000] <jgraham>
The random bytes from Chrome, as written on your website look screwy

[13:18:01.0000] <jgraham>
kpx: I don't really understand the question

[13:19:00.0000] <drclue>
@jgraham  I most certainly agree they look way wrong.   Check out the "host" and "origin" fields in the request. Do those jive with your understanding of the spec?

[13:21:00.0000] <jgraham>
drclue: origin should have a port, at least

[13:21:01.0000] <zcorpan_>
drclue: Origin is the origin of the page which runs the script that opens the websocket

[13:21:02.0000] <kpx>
As in I used to submit blog entries to hacker news and all... Is there some place like taht for HTML5 and all?

[13:21:03.0000] <jgraham>
Ah, that's a good point

[13:21:04.0000] <jgraham>
If you are running the page on port 80 it shoudln't

[13:22:00.0000] <drclue>
@jgraham so what ya figure gives with the odd looking key 3?

[13:22:01.0000] <jgraham>
kpx: Not really.

[13:23:00.0000] <annevk>
drclue, if it's the default port of the scheme the origin production will not include the port

[13:23:01.0000] <annevk>
so http,80 -> no port; https,443 -> no port; http,81 -> port; etc.

[13:23:02.0000] <drclue>
OK, , I'm with ya on the origin

[13:24:00.0000] <drclue>
I still don't get this thing with the odd looking key 3.   

[13:24:01.0000] <zcorpan_>
drclue: looks like you're converting key3 to characters

[13:24:02.0000] <zcorpan_>
drclue: key3 is random bytes

[13:25:00.0000] <zcorpan_>
drclue: don't decode them to a string

[13:25:01.0000] <jgraham>
zcorpan_: It looks surprisingly non-random though

[13:25:02.0000] <jgraham>
What am I missing

[13:26:00.0000] <drclue>
I understand that they are random bytes.  I had only copy-n-pasted them from the shell output to the text editor where in the process they got mangled , but they were mangled looking when they came in

[13:26:01.0000] <jgraham>
Define "mangled looking"

[13:26:02.0000] <jgraham>
Were there 8 bytes?

[13:27:00.0000] <annevk>
opening _bytes_ in a _text_ editor doesn't really work ;)

[13:27:01.0000] <jgraham>
annevk: That really depends

[13:27:02.0000] <jgraham>
I mean they all map to something

[13:27:03.0000] <drclue>
Well, I'm not sure if the examples in the spec and the example in hixies example were rigged to be printable , but both of those were printable and did work in my handshake code

[13:28:00.0000] <annevk>
yeah, like the UTF-8 BOM followed by some stuff :p

[13:28:01.0000] <Hixie>
the spec examples were rigged 

[13:28:02.0000] <Hixie>
to avoid non-printable characters

[13:28:03.0000] <drclue>
Yes there were the proper number of bytes

[13:28:04.0000] <jgraham>
drclue: So I don't understand the problem

[13:29:00.0000] <drclue>
The problem is that chrome won't accept the handshake

[13:29:01.0000] <zcorpan_>
chrome's handshake looks ok to me when i look in wireshark, and it accepts the handshake to a pywebsocket server

[13:30:00.0000] <Hixie>
chrome's handshake works with http://software.hixie.ch/utilities/js/websocket/ connectring to ws://damowmow.com:11111

[13:30:01.0000] <Hixie>
chrome 6, that is

[13:30:02.0000] <Hixie>
dev channel

[13:30:03.0000] <Hixie>
haven't tested 5.x betas

[13:31:00.0000] <jgraham>
drclue: You have your server code anywhere?

[13:31:01.0000] <jgraham>
(public)

[13:31:02.0000] <drclue>
It has been my assumption that the only thing I should have to change from before is the handshake and that the javascript on the browser side should go unchanged.   I'm currently testing with Linux 6.0.422.0 dev

[13:32:00.0000] <drclue>
No , I don't have the server code up anywhere

[13:33:00.0000] <Hixie>
i'm pretty sure the JS should not need changing

[13:33:01.0000] <Hixie>
can you show us a dump of what the client sends and what the server sends?

[13:34:00.0000] <jgraham>
Hixie: should that page for for me? I pasted in the ws URL and hit connect but it always just disconnects

[13:34:01.0000] <drclue>
The only thing I'm messing with is the handshake , and for the most part I ripped off the code from http://webreflection.blogspot.com/

[13:36:00.0000] <Hixie>
sorry i meant ws://damowmow.com:11111/demo

[13:37:00.0000] <drclue>
@Hixie the closest I have to a dump is on http://www.drclue.net/ws.html   examples (A) and (C) are injected , while example (B) is from chrome

[13:37:01.0000] <zcorpan_>
Hixie: your demo needs s/URL/url/ :)

[13:37:02.0000] <jgraham>
Hixie: Ah, that works :)

[13:38:00.0000] <Hixie>
zcorpan_: hm, yeah

[13:38:01.0000] <Hixie>
drclue: i would guess your problem is in handling the bytes -- you're probably converting them to UTF-8 or something

[13:40:00.0000] <annevk>
if we break from HTTP we should just make it a UTF-8 stream

[13:44:00.0000] <epeus>
anyone know if Apple did fulfil Jobs's pledge to post open specs on their new FaceTime stuff yet?

[13:50:00.0000] <jgraham>
epeus: I thought he promised "open standards" which is a bit different to just "open specs"

[13:51:00.0000] <kpx>
jgraham: Hey I wrote this small blog entry on webdatabases wanted to ask if you could just have a look and suggest some improvements... http://mnesia.wikispaces.com/HTML5+and+webstorage... 

[13:51:01.0000] <kpx>
jgraham: thanks.. 

[13:51:02.0000] <epeus>
I saw a list fo acronyms I recognised...

[14:08:00.0000] <jgraham>
kpx: I haven't done a detailed technical review, but it is worth noting that the databases stuff is only really sutiable for extensions or other use cases where the target browser is known in advance

[14:09:00.0000] <jgraham>
because only Opera + Webkit have support and the others have rejected the idea

[14:09:01.0000] <kpx>
jgraham: Well i did all this while writing a chrome extension so its ok.... :)

[14:11:00.0000] <kpx>
jgraham: But thanks for looking, I'd be happy if you could give any pointers 

[14:19:00.0000] <jgraham>
kpx: Yeah, I saw you were developing a chrome extension. But the second sentence of the article is somewhat misleading

[14:34:00.0000] <zcorpan_>
anyone wanna place bets on whether the spec will be published tomorrow?

[14:34:01.0000] <gsnedders>
/me realzies he can't use being under 18 as an excuse for not betting now

[15:49:00.0000] <nessy>
Hixie: question of process - are any of the discussions on WHATWG still useful to include new stuff into the spec or are we beyond that and changes have to be made through the W3C process?

[15:52:00.0000] <Hixie>
nessy: i ignore where the feedback comes from when editing the spec, so discussions on the whatwg have as much sway as discussions on the w3c side

[15:53:00.0000] <Hixie>
more, maybe, since the discussions on the w3c side tend to end up mired in process crap

[15:53:01.0000] <nessy>
excellent - was just curious

[15:53:02.0000] <nessy>
yeah, and politics ;)

[15:53:03.0000] <Dashiva>
Much the same

[15:53:04.0000] <Hixie>
well there's plenty of politics on the whatwg side, it's just not discussed :-)

[15:53:05.0000] <nessy>
which is the nice bit

[15:53:06.0000] <Hixie>
:-)

[15:56:00.0000] <volkmar>
Hixie: could we have your advice on mozilla bug 568515 (beginning to comment 3) ?

[15:56:01.0000] <Hixie>
looking...

[15:56:02.0000] <volkmar>
thanks :)

[15:57:00.0000] <Hixie>
wow i've no idea what i meant by "User agents must treat plaintext elements in a manner equivalent to pre elements.

[15:57:01.0000] <Hixie>
"

[15:58:00.0000] <volkmar>
oh...

[15:58:01.0000] <volkmar>
if it's not for the interface, what could that be ?

[16:00:00.0000] <zcorpan_>
i think that text was added as a result of me suggesting the spec should say what the semantics of a plaintext element are, for the purpose of e.g. speech browsers

[16:01:00.0000] <Hixie>
volkmar: i think it just means for semantics

[16:01:01.0000] <volkmar>
Hixie: then which interface should they have ?

[16:01:02.0000] <Hixie>
volkmar: there's a similar statement for <acronym>, saying it must match <abbr>, but <abbr> doesn't have its own interface

[16:01:03.0000] <Hixie>
volkmar: HTMLElement

[16:01:04.0000] <Hixie>
i'll file a bug on the spec to make that clearer

[16:01:05.0000] <volkmar>
Hixie: ok, indeed, a clarification might be good

[16:02:00.0000] <Hixie>
volkmar: i filed http://www.w3.org/Bugs/Public/show_bug.cgi?id=9885

[16:03:00.0000] <volkmar>
Hixie: thanks :)

[16:06:00.0000] <volkmar>
Hixie: btw, i asked earlier about valueAsNumber being a double instead of a float and Dashiva told me it was upon request

[16:07:00.0000] <volkmar>
but the text about the algorithm to convert a string to a number tells it has to follow the rule of parsing a floating point number which is about float and not double

[16:07:01.0000] <volkmar>
i suppose a clarification is needed

[16:08:00.0000] <Hixie>
volkmar: file another bug :-)

[16:08:01.0000] <Hixie>
volkmar: you can file bugs using the text box at the bottom right of the spec

[16:08:02.0000] <volkmar>
Hixie: i know

[16:09:00.0000] <volkmar>
i wanted to be sure valueAsNumber is going to stay as double

[16:09:01.0000] <Hixie>
no idea

[16:09:02.0000] <volkmar>
ok :)

[16:09:03.0000] <Hixie>
since JS only has one Number type, i don't know that i really care either way

[16:10:00.0000] <volkmar>
for the implementation, that's quite different

[16:10:01.0000] <volkmar>
anyway, i will file the bug

[16:10:02.0000] <Hixie>
thanks

[16:43:00.0000] <Dashiva>
volkmar: I checked the discussion from when it was changed, and it turns out float won't work at all since the timestamp is milliseconds and not seconds

[16:47:00.0000] <Philip`>
Dashiva: Why does the scale matter? Doesn't that just mean the maximum is 10^34 seconds, not 10^37 seconds?

[16:48:00.0000] <Philip`>
(so it's still plenty of ages of the universe)

[16:49:00.0000] <Dashiva>
Philip`: Timestamps need to be precise, so a float is pretty much a 24-bit integer in this regard

[16:50:00.0000] <Dashiva>
But I suppose you're right, and float would be too small even for seconds

[16:51:00.0000] <Philip`>
Oh, so the issue is that it has a precision of milliseconds and a range of >2^24 larger than that, rather than the issue being that it's scaled to milliseconds

[16:52:00.0000] <Philip`>
Things would be much easier if time didn't change so frequently

[16:53:00.0000] <Dashiva>
Sounds like you're onto something

[16:54:00.0000] <Philip`>
24fps is enough for visually smooth animation in films, so we could just quantise the universe to ~100Hz and nobody would notice

[16:54:01.0000] <AryehGregor>
I'm pretty sure we'd notice when chemistry stopped working.  Well, actually, we wouldn't, but for entirely the wrong reason.

[16:57:00.0000] <Dashiva>
It seems today's date requires 31 bits for second precision, 41 for millisecond precision

[16:58:00.0000] <AryehGregor>
It requires 0 seconds for arbitrary precision, if you pick the start date correctly.

[16:58:01.0000] <AryehGregor>
It requires 0 bits for arbitrary precision, if you pick the start date correctly.

[16:59:00.0000] <Dashiva>
Sadly, we have a bit of a legacy environment to consider

[16:59:01.0000] <annevk>
oh look

[16:59:02.0000] <annevk>
seems we found a new issue to get all worked up about

[16:59:03.0000] <annevk>
the WHATWG reference


2010-06-09
[17:00:00.0000] <AryehGregor>
/me started to comment in that discussion, but thought better of it.

[17:00:01.0000] <annevk>
yeah, I failed

[17:01:00.0000] <annevk>
it's probably the short vacation

[17:01:01.0000] <annevk>
makes me naive

[17:10:00.0000] <hober>
I sometimes find myself contributing to such threads, thinking "this way, people with more important things to do won't have to."

[17:10:01.0000] <hober>
Afterwards, I wonder, "why did I feed the trolls again?"

[17:12:00.0000] <annevk>
yeah...

[17:12:01.0000] <annevk>
time for my awesome book

[17:12:02.0000] <annevk>
and then some sleep

[17:12:03.0000] <annevk>
nn

[17:16:00.0000] <TabAtkins_>
Philip`: 24fps is only visually acceptable because we're so used to it.  It's visually distinguishable from the 60fps that camcorders get.

[17:16:01.0000] <TabAtkins_>
(That's why "motion blur" is a visual option for Mass Effect - to make it blurrier so it looks more like a movie rather than a game.)

[19:05:00.0000] <skm>
need a bit of help on shared workers & workers

[19:05:01.0000] <skm>
i thought what i wanted to use was nested workers but this isnt implemented anywhere and im thinnking sharedworkers might do what i need

[19:06:00.0000] <skm>
i want one worker that selects from websql (200,000) rows, then multiple (aybe 5) other workers handling the processing

[19:06:01.0000] <skm>
is this something sharedworkers can do/are intended for

[19:06:02.0000] <skm>
or mainly for sharing tasks between pages/tabs

[19:46:00.0000] <boblet>
just tried <dt>…<dd>…</dd></dt> but it’s unhelpfully corrected by browsers. I can select a sibling dd with .class + dd-style selectors, but not once there’s more than one dd. basically forced to apply classes to every dt & dd in a description list ;(

[19:47:00.0000] <boblet>
lack of name-value association wrapper element also means any script that wants to manipulate more than single associations needs to implement the association algorithm

[20:13:00.0000] <boblet>
mikesmith Hixie ping me if you have a moment. would like to discuss if it’s possible to add an exception for <div> to “If a dl element contains non-whitespace text nodes, or elements other than dt and dd, then those elements or text nodes do not form part of any groups in that dl.”

[20:22:00.0000] <Hixie>
boblet: not sure what you mean

[20:24:00.0000] <boblet>
Hixie: have had a couple of ppl mention <dl> is hard to use with CSS or JS because name-value associations don’t have a wrapper element.

[20:24:01.0000] <Hixie>
yeah, that's a bug in css

[20:24:02.0000] <Hixie>
we should fix it in css

[20:25:00.0000] <Hixie>
we could provide an array of DocumentFragment or something to make it easier in JS

[20:25:01.0000] <Hixie>
not clear what the use case is in JS

[20:25:02.0000] <boblet>
Hixie: aah. what about for JS? just implement the dl parsing algorithm in scripts?

[20:25:03.0000] <Hixie>
parsing?

[20:25:04.0000] <Hixie>
finding groups in JS is trivial

[20:25:05.0000] <Hixie>
you just collect nodes until you hit a dd then collect nodes until you hit a dt

[20:26:00.0000] <Hixie>
and that's one group

[20:26:01.0000] <boblet>
well, the logic of dl (how multiple <dt> and <dd> are associated)

[20:26:02.0000] <Hixie>
then you repeat

[20:26:03.0000] <Hixie>
but depending on what the use cases are for manipulating <dl>s in script, we might be able to provide much better APIs for it

[20:26:04.0000] <boblet>
ok. need to follow up with that person coz not sure I understood fully (Japanese tweet, so brevity+language working against me :)

[20:27:00.0000] <boblet>
will find out use case and report back

[20:27:01.0000] <Hixie>
k

[20:28:00.0000] <boblet>
CSS selector won’t be any time soon tho huh, with Selectors Level 3 already proposed recommendation

[20:28:01.0000] <Hixie>
no idea, i'm not up on what the csswg is doing

[20:28:02.0000] <boblet>
k, will email www-style. thanks

[20:31:00.0000] <boblet>
Hixie: oh re: what I meant, wrapping name-vale associations in <div> lets me style enclosed <dt> and <dd>, but I understand that quote to mean that anything inside <div> would then not be part of the <dl> — correct?

[20:32:00.0000] <boblet>
so I was wondering if “or elements other than dt and dd” could become “or elements other than dt, dd and div” until CSS selectors catches up

[20:33:00.0000] <Hixie>
http://wiki.whatwg.org/wiki/FAQ#HTML5_should_group_.3Cdt.3Es_and_.3Cdd.3Es_together_in_.3Cdi.3Es.21

[20:33:01.0000] <Hixie>
it comes up quite often :-)

[20:33:02.0000] <Hixie>
it's clearly not _that_ much of a problem since nobody has done anything in the csswg about it

[20:34:00.0000] <boblet>
hah :) so I’m guessing that using <div> in the meantime is not recommended right?

[20:35:00.0000] <boblet>
(note to self, check FAQ before asking q)

[20:37:00.0000] <Hixie>
well it's non-conforming

[20:37:01.0000] <Hixie>
so not so much "not recommended" as "not allowed" :-)

[20:38:00.0000] <boblet>
heh. puts the kibosh on that idea then

[20:44:00.0000] <Hixie>
i updated the faq with a bit more on that question, fyi

[21:04:00.0000] <Hixie>
wow, fielding really lives in his own world

[21:32:00.0000] <wirepair>
does anyone have any links to browser layout/security testcases?

[21:32:01.0000] <wirepair>
i think chromiums src tree comes with a bunch

[21:33:00.0000] <wirepair>
but i'd be interested to see more 

[21:33:01.0000] <MikeSmith>
wirepair: I think it depends on what you are looking for

[21:33:02.0000] <MikeSmith>
but Adam Barth would be a good person to chat with

[21:34:00.0000] <MikeSmith>
abarth on #webkit

[21:34:01.0000] <MikeSmith>
/me re-reads wirepair question

[21:34:02.0000] <wirepair>
hehe

[21:34:03.0000] <MikeSmith>
what are "layout/security" testcases?

[21:35:00.0000] <wirepair>
well they call them LayoutTests in chromium, even though they clearly include testcases for security

[21:35:01.0000] <wirepair>
such as same origin policy checks

[21:35:02.0000] <MikeSmith>
ah, that

[21:35:03.0000] <wirepair>
make sure xhr doesn't follow redirects, that kind of thing

[21:35:04.0000] <MikeSmith>
yeah, that's just a name

[21:36:00.0000] <wirepair>
so adam would be good then huh?

[21:36:01.0000] <MikeSmith>
yep

[21:36:02.0000] <MikeSmith>
definitely

[21:36:03.0000] <wirepair>
he on irc a lot?

[21:36:04.0000] <MikeSmith>
he's around on #webkit usually

[21:37:00.0000] <MikeSmith>
I guess he's there mostly to talk with other webkit devs about build issues and such

[21:37:01.0000] <wirepair>
ah cool i'll try to ping him next time he pops in

[21:37:02.0000] <wirepair>
i just finished the infrastructure for my browser testing system and starting on building testcases

[21:38:00.0000] <wirepair>
figured i should see how other people are doing tests

[21:38:01.0000] <MikeSmith>
what kind of infrastructure?

[21:38:02.0000] <wirepair>
http://sh0dan.org/wbts/

[21:38:03.0000] <wirepair>
fake dns server, vhosts, ssl

[21:38:04.0000] <wirepair>
all for testing security

[21:38:05.0000] <MikeSmith>
automated? 

[21:38:06.0000] <wirepair>
thats the plan

[21:38:07.0000] <MikeSmith>
cross-browser?

[21:38:08.0000] <wirepair>
;>

[21:38:09.0000] <wirepair>
thats the plan

[21:38:10.0000] <wirepair>
hehe

[21:38:11.0000] <MikeSmith>
do you know about browsertests?

[21:38:12.0000] <wirepair>
nope

[21:39:00.0000] <MikeSmith>
/me goes to find a URL

[21:39:01.0000] <wirepair>
browsertests.com?

[21:39:02.0000] <wirepair>
.org? heh

[21:40:00.0000] <MikeSmith>
http://code.google.com/p/browsertests/

[21:40:01.0000] <MikeSmith>
it's worth taking a look at

[21:40:02.0000] <wirepair>
thanks, looking now

[21:40:03.0000] <wirepair>
ah yeah looks like www.browsertests.org is an instance of it

[21:41:00.0000] <MikeSmith>
http://code.google.com/p/browsertests/wiki/Installation

[21:41:01.0000] <MikeSmith>
it's pretty easy to install

[21:41:02.0000] <MikeSmith>
we also have a #testing channel on irc.w3.org where you might want to hang out

[21:41:03.0000] <MikeSmith>
port 6665 or port 80 on irc.w3.org

[21:42:00.0000] <MikeSmith>
the guy who developed browsertests is usually there

[21:42:01.0000] <MikeSmith>
syp

[21:42:02.0000] <MikeSmith>
Sylvain Pasche

[21:42:03.0000] <wirepair>
great will do

[21:50:00.0000] <wirepair>
yeah ok these look the same as the ones in the chromium source tree (http://www.browsertests.org/tests/search/?file=security)

[22:06:00.0000] <MikeSmith>
the prose of the MathML specs don't ever seem to actually define the content models for many elements

[22:06:01.0000] <MikeSmith>
e.g., http://www.w3.org/TR/MathML3/chapter3.html#presm.mi

[22:20:00.0000] <Dashiva>
MikeSmith: It seems RelaxNG is the conformance basis, the prose expands on that

[22:21:00.0000] <Dashiva>
"A valid MathML expression is an XML construct determined by the MathML RelaxNG Schema together with the additional requirements given in this specification."

[22:22:00.0000] <MikeSmith>
Dashiva: OK

[22:22:01.0000] <Dashiva>
Seems like an inconvenient arrangement for authors

[22:22:02.0000] <MikeSmith>
yeah

[22:23:00.0000] <MikeSmith>
but specs really aren't for authors anyway

[22:23:01.0000] <MikeSmith>
I can see from reading back to http://www.w3.org/TR/MathML3/chapter3.html#presm.tokel that there is something of a prose definition there

[22:24:00.0000] <MikeSmith>
but that really seems like and suboptimal way to do it

[22:25:00.0000] <MikeSmith>
"With the exception of the empty mspace element, token elements can contain any sequence of zero or more Unicode characters, or mglyph or malignmark elements"

[22:26:00.0000] <MikeSmith>
but then it never even explicitly defines what "token elements" are

[22:26:01.0000] <MikeSmith>
oh wait

[22:26:02.0000] <MikeSmith>
it does in the next paragraph

[22:26:03.0000] <MikeSmith>
"Token elements represent identifiers (mi), numbers (mn), operators (mo), text (mtext), strings (ms) and spacing (mspace)."

[22:26:04.0000] <MikeSmith>
sorta

[22:27:00.0000] <MikeSmith>
though no hyperlink for "token elements" or dfn markup or anything there

[22:28:00.0000] <Dashiva>
A bit weird that mspace is a token element when it isn't allowed to contain tokens

[22:29:00.0000] <MikeSmith>
yeah

[22:31:00.0000] <MikeSmith>
anyway, I'm just trying to figure out of having validator.nu allow HTML markup in mi and mn and mo violates MathML semantics

[22:32:00.0000] <MikeSmith>
I think I'll just try having it allow HTML markup within any "token element"

[22:33:00.0000] <MikeSmith>
if hsivonen is OK with that

[22:34:00.0000] <MikeSmith>
but that means also allowing it within mtext

[22:34:01.0000] <MikeSmith>
and within the ms element

[22:34:02.0000] <MikeSmith>
whatever that is

[22:34:03.0000] <MikeSmith>
"An mtext element is used to represent arbitrary text that should be rendered as itself."

[22:35:00.0000] <MikeSmith>
I guess an HTML <img> of text could be considered to "represent" text

[22:36:00.0000] <MikeSmith>
「The ms element is used to represent "string literals" in expressions meant to be interpreted by computer algebra systems or other systems containing "programming languages".」

[22:37:00.0000] <MikeSmith>
but 「Note that the string literals encoded by ms are made up of characters, mglyphs and malignmarks rather than "ASCII strings".」

[22:37:01.0000] <MikeSmith>
so I say we are OK there, too

[22:39:00.0000] <MikeSmith>
anyway, restricting it to allowing HTML phrasing content seems reasonable

[22:39:01.0000] <MikeSmith>
that is, allowing HTML phrasing content within any MathML "token element" 

[22:39:02.0000] <MikeSmith>
as far as validation goes

[22:40:00.0000] <MikeSmith>
hmm, yeah, and that seems to align with the HTML5 parsing algorithm too

[23:44:00.0000] <annevk>
remind me to take an afternoon nap

[23:44:01.0000] <annevk>
falling asleep at 2 and waking up at 7 is not fair

[23:46:00.0000] <annevk>
oh look

[23:46:01.0000] <annevk>
we're not publishing after all

[23:46:02.0000] <annevk>
quelle surprise

[00:04:00.0000] <Hixie>
i love the way some people think the w3c version of the html spec is somehow more normative

[00:08:00.0000] <annevk>
in the end what matters is what gets implemented and used by authors

[00:08:01.0000] <annevk>
that has not really changed in the past fifteen years

[00:09:00.0000] <Hixie>
yup

[00:39:00.0000] <annevk>
American political ads are great

[00:46:00.0000] <annevk>
"Regardless, if you're thinking that the WhatWG will pull out of the HTML5 effort, and doom it by their lack of participation, think again: the WhatWG organization is not a legal entity. It is an informal group of a handful of individuals, half of whom became disillusioned with the effort years ago and have not participated since."

[00:47:00.0000] <annevk>
1) She clearly is missing something. 2) What is with the "WhatWG" spelling?

[00:49:00.0000] <Hixie>
the whatwg has more participants now than at any time in its history, and more than 3 times more people than the htmlwg... who are the people who left?

[00:49:01.0000] <Hixie>
we should send them flowers or something, asking them to come back

[00:52:00.0000] <Hixie>
mpt and Matthew Raymond haven't posted in a while

[00:52:01.0000] <Hixie>
maybe she was referring to them

[01:25:00.0000] <annevk>
oh hey, Firefox landed WebM

[01:25:01.0000] <annevk>
way to go

[02:13:00.0000] <MikeSmith>
hsivonen: please ping me if/when you have time ... I want to ask about a change to the (X)HTML+SVG+MathML v.nu schemas to make it valid to have HTML phrasing content in MathML "token elements" (mi, mo, mn, ms, mtext)

[02:14:00.0000] <MikeSmith>
I made the changes experimentally in my workspace and pushed to http://www.w3.org/html/check for testing

[02:25:00.0000] <zcorpan_>
hidden="" is widely implemented?

[02:45:00.0000] <MikeSmith>
zcorpan_: who said it is?

[02:49:00.0000] <zcorpan_>
http://www.w3.org/mid/4C0F4FCE.6000304⊙kc

[02:57:00.0000] <Hixie>
hidden="" is newer than ping="" and implemented in fewer browsers, if i'm not mistaken

[02:59:00.0000] <hsivonen>
I think Jirka is confusing hidden="" with type=hidden

[03:04:00.0000] <hsivonen>
w00t. Youtube in WebM works great in Minefield trunk

[03:04:01.0000] <hsivonen>
however, the context menu doesn't work right

[03:05:00.0000] <hsivonen>
it worked right after Google I/O

[03:06:00.0000] <hsivonen>
is something broken in Minefield or did YouTube deliberately break  the context menu to discourage people from saving video

[03:07:00.0000] <zcorpan_>
i thought google had a <div> overlay to break the context menu

[03:07:01.0000] <zcorpan_>
s/google/youtube/

[03:07:02.0000] <hsivonen>
zcorpan_: ok

[03:08:00.0000] <hsivonen>
what a great way to give every browser an incentive to work on optimizing overlay compositing

[03:09:00.0000] <hsivonen>
/me wonders who is going to be the first to write an independent implementation of VP8

[03:10:00.0000] <Philip`>
Is it independent if you copy-and-paste the chunks of C that the specification is (apparently) largely made of?

[03:10:01.0000] <hsivonen>
Philip`: not really

[03:11:00.0000] <hsivonen>
Philip`: it's independent if Hixie expresses those chunks in English and then someone carefully implements the English spec in C again

[03:12:00.0000] <Hixie>
not gonna be me!

[03:20:00.0000] <MikeSmith>
hsivonen: I made an experimental change to the (X)HTML+SVG+MathML v.nu schemas to make it valid to have HTML phrasing content in MathML "token elements" (mi, mo, mn, ms, mtext)

[03:20:01.0000] <MikeSmith>
made the changes in my workspace and pushed to http://www.w3.org/html/check for testing

[03:21:00.0000] <MikeSmith>
good change? bad change?

[03:21:01.0000] <hsivonen>
MikeSmith: sorry, I had a bit of a situation with the tinderbox earlier

[03:21:02.0000] <MikeSmith>
no problem man

[03:21:03.0000] <hsivonen>
MikeSmith: is the change supported by any spec?

[03:21:04.0000] <MikeSmith>
well, the HTML spec...

[03:22:00.0000] <MikeSmith>
the parsing algorithm allows HTML in those elements

[03:22:01.0000] <hsivonen>
MikeSmith: that's parser behavior only, right?

[03:22:02.0000] <MikeSmith>
right

[03:22:03.0000] <MikeSmith>
only that

[03:22:04.0000] <MikeSmith>
the MathML spec says nothing about this

[03:22:05.0000] <MikeSmith>
not as far as I can tell at least

[03:22:06.0000] <hsivonen>
Hixie: when you made the parser allow HTML in MathML tokens, what was your plan about speccing?

[03:22:07.0000] <MikeSmith>
reading the MathML specs is like reading a foreign language

[03:22:08.0000] <hsivonen>
MikeSmith: does layout work as expected in Gecko?

[03:22:09.0000] <Hixie>
hsivonen: i provided hooks so that MathML could say what was allowed

[03:22:10.0000] <zcorpan_>
MikeSmith: that's why it's called 'foreign content'

[03:23:00.0000] <hsivonen>
zcorpan_: lol

[03:23:01.0000] <Hixie>
hsivonen: (and used hooks that SVG provided to make it defined for SVG)

[03:24:00.0000] <hsivonen>
Hixie: oh was the motivation for allowing HTML actually allowing SVG to support Jacques Distler's use cases without annotation-xml?

[03:24:01.0000] <hsivonen>
Hixie: oops. sorry, I read your lines in reverse order. scratch that

[03:24:02.0000] <Hixie>
not sure what that question meant :-)

[03:24:03.0000] <hsivonen>
Hixie: has anyone pinged the Math WG pointing out that they could now allow this?

[03:25:00.0000] <Hixie>
not to my knowledge

[03:25:01.0000] <Hixie>
they've been busy with mml3 and i've been busy with html5

[03:25:02.0000] <Hixie>
so it hasn't been a priority for anyone to get together on this and hammer it out

[03:25:03.0000] <MikeSmith>
hsivonen: layout works as expected in Gecko as far as I can tell

[03:26:00.0000] <MikeSmith>
for this case

[03:26:01.0000] <zcorpan_>
using annotation-xml to embed svg or html in mathml seems like an ugly hack, maybe we shouldn't support it at all and require use of <mn> et al instead

[03:26:02.0000] <MikeSmith>
zcorpan_: that'd be fine by me

[03:26:03.0000] <MikeSmith>
though I did already raise a spec bug about it

[03:26:04.0000] <MikeSmith>
but I raised that just to get the question answered

[03:27:00.0000] <MikeSmith>
not because I think necessarily that answer should be to allow html in annotation-xml

[03:27:01.0000] <zcorpan_>
commented on the bug

[03:28:00.0000] <MikeSmith>
and true what Hixie says about MathML WG being focused on getting MathML3 out

[03:28:01.0000] <MikeSmith>
as far as I can tell

[03:28:02.0000] <MikeSmith>
anything related to MathML 2.0 is a low- or non-priority for them at this point, I think

[03:29:00.0000] <zcorpan_>
so does mathml 3 allow html and svg in <mn>?

[03:29:01.0000] <MikeSmith>
but David Carlisle did help much recently by providing a good MathML 2.0 schema

[03:30:00.0000] <MikeSmith>
zcorpan_: mathml 3 doesn't say anything more about it that 2 did, afaict

[03:30:01.0000] <MikeSmith>
not the spec prose

[03:30:02.0000] <MikeSmith>
nor the RelaxNG schema for 3

[03:31:00.0000] <hsivonen>
MikeSmith: my take is that I like the idea, but it would be nice to be able to be able to point to a draft of some kind as justification

[03:32:00.0000] <hsivonen>
MikeSmith: that is, this isn't the kind of thing where validator developers just have to overrule the specs for sanity

[03:32:01.0000] <MikeSmith>
hsivonen: OK, I can try to get the attention of the Math group and see if I have any success

[03:32:02.0000] <hsivonen>
MikeSmith: I guess I'd be OK with it anyway if it's accompanied by an email to the Math WG

[03:32:03.0000] <hsivonen>
MikeSmith: ok

[03:33:00.0000] <Hixie>
it's a very easy fix for them -- they just need to say that the content model of their <mn> element is whatever it is plus phrasing content and reference the HTML spec.

[03:34:00.0000] <Hixie>
and so on for the elements that accept text

[03:35:00.0000] <MikeSmith>
yeah, I would hope so at least

[03:35:01.0000] <MikeSmith>
I wonder if they use W3C bugzilla

[03:35:02.0000] <MikeSmith>
/me goes to check

[03:36:00.0000] <MikeSmith>
hmm, looks like not

[03:36:01.0000] <MikeSmith>
anyway, I will get an e-mail message out to the group at least

[03:36:02.0000] <MikeSmith>
or to their comments list

[03:38:00.0000] <MikeSmith>
but for now I have to drop off

[03:38:01.0000] <MikeSmith>
back on later

[05:22:00.0000] <jgraham>
/me notes that the polyglot spec is a set of inferences rather than a definition

[05:29:00.0000] <Lachy>
TimBL suggestion that the polyglot spec be made normative is misguided. It must stay entirely non-normative because it doesn't specify anything beyond what HTML5 itself specifies, it just describes the requirements in a way targetted towards authors wanting to write polyglot documents

[05:31:00.0000] <jgraham>
Lachy: I just said something similar on the list

[05:31:01.0000] <Lachy>
also, I don't see the point of it. It just duplicates much of what will go into my HTML5 Reference and HTML5 Guide. (I'm hoping I will be allocated some time to work on that soon too)

[05:32:00.0000] <Philip`>
It has the advantage of not always being in the future tense

[05:32:01.0000] <Lachy>
I don't have a problem with it being published though, I just think the work should have gone into contributing to the HTML5 Reference, rather than being separate

[05:33:00.0000] <Lachy>
Philip`, yeah, that's true. It's just that I've been allocated to so much other QA tasks rather than spec work.

[05:34:00.0000] <MikeSmith>
it normatively specifies what an "XHTML/HTML polygot document" is

[05:35:00.0000] <jgraham>
MikeSmith: The only normative text needed to do that is "an XHTML/HTML polygot document is a document that conforms to all the authoring requirements of HTML and XHTML"

[05:36:00.0000] <Philip`>
jgraham: That's not a useful definition, since you want polyglot documents to act similarly when processed as XHTML and HTML

[05:37:00.0000] <Philip`>
which means you need more restrictions, based on subjective views of what is similar enough behaviour

[05:38:00.0000] <jgraham>
Philip`: documenting things that cannot be strictly inferred from either spec seems fine

[05:38:01.0000] <Lachy>
Philip`, there's no need for there to be a normative definition of what's considered to be polyglot anyway.  There only needs to be guidelines.

[05:39:00.0000] <jgraham>
So maybe "can't be normative" is wrong

[05:39:01.0000] <jgraham>
er

[05:39:02.0000] <jgraham>
Not that I said that

[05:40:00.0000] <jgraham>
So it is OK to have the parts that can't be infered from other specs considered normative

[05:40:01.0000] <jgraham>
I wouldn't bother myself, but there we go

[05:40:02.0000] <Lachy>
Although there are certainly some things authors need to do for strict polyglot documents that can't necessarily be implied purely from the conformance requirements of either HTML or XHTML, there is no reason for those to be normatively specified.

[05:41:00.0000] <Philip`>
jgraham: Do you think most of the HTML5 spec's syntax definition should be non-normative, because it can be inferred from the parsing definition?

[05:41:01.0000] <jgraham>
Philip`: It can't

[05:41:02.0000] <Lachy>
besides, most authors aren't really interested in writing strict polyglot documents anyway. They just want to use certain talismans for what they perceive as cleaner markup.

[05:41:03.0000] <jgraham>
Strictly speaking

[05:43:00.0000] <jgraham>
(they are independent in the sense that one applies to authors and one applies to UAs. Of course there has to be some common ground to make the definitions useful, but you can't infer what the common ground is without normative author requirements)

[05:44:00.0000] <Philip`>
jgraham: You could derive e.g. the syntax for start tags from the relevant fragment of the parser definition (avoiding cases that are parse errors), I think, so the spec doesn't need to define the syntax again in terms of all the angle brackets and spaces

[05:44:01.0000] <Philip`>
(It would just need to add a few extra requirements about not including ` in unquoted attribute values etc)

[05:46:00.0000] <Philip`>
The list of omittable tags could be omitted, and derived from examining the parser for tags that can be omitted without errors and without changing the DOM, etc

[05:48:00.0000] <jgraham>
Philip`: Yeah, I suppose that could theoretically work. Making seperate normative author/UA requirements still feels rather different to expressing the same author requirements in multiple places

[05:48:01.0000] <jgraham>
all with the same normative status

[06:13:00.0000] <hsivonen>
I think a the polyglot doc should be a set of inferences and, therefore, non-normative

[07:03:00.0000] <Lachy>
hsivonen, what do you mean by "ignoring the CDATA section exposure domain modeling error in the DOM itself"?

[07:03:01.0000] <Lachy>
are you referring to the fact that CDATA sections are exposed through the DOM API?

[07:05:00.0000] <Lachy>
or to the fact that things like <script><![CDATA[ ... //]]></script> inherently parse differently in HTML and XHTML, but which should be ignored for polyglot checking purposes?

[07:21:00.0000] <hsivonen>
Lachy: I mean that CDATA nodes should be considered equivalent to being next nodes and then adjacent text nodes should be coalesced before comparison

[07:21:01.0000] <hsivonen>
Lachy: since the HTML parser in foreign content generated text nodes, not CDATA nodes for <![CDATA[...]]>

[07:24:00.0000] <Dashiva>
http://twitter.com/jirkakosek/status/15764226697

[07:24:01.0000] <Dashiva>
Isn't the answer to this obviously yes?

[07:24:02.0000] <Dashiva>
The problem with XML isn't the use cases, it's the XML.

[07:28:00.0000] <micheil>
what's with the two data framing formats for websockets?

[07:37:00.0000] <jgraham>
micheil: Hixie claims that using a sentinal marker makes it less likely that server authors will screw up by counting characters rather than encoded bytes, and makes content of unknown length simpler

[07:37:01.0000] <jgraham>
+ sending

[07:37:02.0000] <micheil>
Hixie: has the data framing for client-sent messages changed between 75 and 76?

[09:04:00.0000] <micheil>
Hixie: also, when is the high byte variant likely to occur?

[09:16:00.0000] <jgraham>
micheil: Binary websockets will happen just as soon as the web platform supports binary

[09:17:00.0000] <jgraham>
i.e. has a bytearray type or so

[09:17:01.0000] <micheil>
okay, so, for now, probably best not to wrory about them?

[09:17:02.0000] <jgraham>
Yeah

[09:17:03.0000] <jgraham>
Though hopefully there will be progress

[09:17:04.0000] <jgraham>
soon

[09:17:05.0000] <jgraham>
the WebGL people want it, the ES people want it, the File people want it and the WebSockets people want it

[09:18:00.0000] <jgraham>
so the demand is there

[09:18:01.0000] <jgraham>
They just need to agree on something (or let implementations make their decisions for them)

[09:19:00.0000] <micheil>
I'm just finding some parts of the spec confusing, as I don't yet know how to fully read the spec's.. like the lingo and such

[09:23:00.0000] <micheil>
jgraham: under high-order use case: http://www.whatwg.org/specs/web-apps/current-work/complete.html?slow-browser=1#data-framing

[09:23:01.0000] <micheil>
after step 4, b doesn't change.

[09:24:00.0000] <micheil>
oh. fuck. wait.

[09:24:01.0000] <micheil>
missed an important point. my bad.

[09:35:00.0000] <annevk>
lol

[09:35:01.0000] <annevk>
Shelley is now claiming there was no agreement between Opera, Mozilla, and Apple

[09:36:00.0000] <annevk>
guess she missed the deadlock way early on and the recent announcement on the WHATWG list

[09:38:00.0000] <TabAtkins>
boblet: Selectors 4 will be started at some point this year.

[09:39:00.0000] <boblet>
TabAtkins: oh ho! good news

[09:39:01.0000] <boblet>
TabAtkins: if you’d like to see what the style complaints were: http://oli.jp/temp/dl-for-meta.html

[09:41:00.0000] <boblet>
basically floating dt and dd right reverses order, and you have to know width to display them inline. first 2 from Niels, the rest me trying to get something that did what he wanted. can send you the email background if you’re interested

[09:41:01.0000] <daedb_>
TabAtkins: Is there a list of expected new selectors in Selectors 4 available somewhere?

[09:41:02.0000] <Philip`>
annevk: Baseless claims of legal uncertainty are great because you can say anything you want and the best case is that people believe you and the worse case is that someone points out facts and you can say "IANAL" and then make more claims

[09:41:03.0000] <Philip`>
s/worse/worst/

[09:46:00.0000] <TabAtkins>
boblet: Yeah, I understand all the cases for <di> or ::di or ::wrap(dt,dd+dt) - I've agitated for them too.  ^_^

[09:46:01.0000] <TabAtkins>
And no, no list.

[09:47:00.0000] <neitcho>
hi, is this channel about html5?

[09:47:01.0000] <annevk>
::di would prolly make the most sense even though Hixie claims it's a generic problem...

[09:48:00.0000] <annevk>
or maybe some simple predefined XBL templates

[09:48:01.0000] <Philip`>
neitcho: Yes

[09:48:02.0000] <annevk>
neitcho, sure is

[09:48:03.0000] <boblet>
TabAtkins: heh, as long as someone is agitating :) to tell the truth once CSS3 support catches up I think we’re gonna take quite a while to work out how to use everything together anyhow

[09:48:04.0000] <TabAtkins>
Yeah, I think the generic solution should exist too, but HTML can, itself, specify a ::di to make this particular common case very easy.

[09:49:00.0000] <neitcho>
Do you know any method or program to make dynamic tutorials? maby somethin that uses html5 and video tag

[09:49:01.0000] <annevk>
Philip`, yeah, fortunately it is easily discovered; now I don't have to read the rest of it :)

[09:49:02.0000] <annevk>
neitcho, JavaScript?

[09:49:03.0000] <boblet>
hixie was saying the generic solution would also address implied sections — would that mean outlining algorithm implicit sections?

[09:50:00.0000] <neitcho>
<annevk> maybe I don't know where to start, I know I dont want flash

[09:50:01.0000] <TabAtkins>
boblet: Yeah.

[09:51:00.0000] <neitcho>
want like a clickable video

[09:52:00.0000] <boblet>
TabAtkins: cool, thought so

[09:54:00.0000] <annevk>
neitcho, if you don't want to learn JavaScript you could take a look at http://jquery.com/ I suppose

[09:54:01.0000] <annevk>
neitcho, there's prolly tutorials for slideshow / interactive like things

[09:55:00.0000] <neitcho>
<annevk> I'll have a look. Thanks!

[09:56:00.0000] <boblet>
nn

[09:58:00.0000] <mikekelly>
I'm behind you Ian.

[10:18:00.0000] <drclue>
OK folks , I saw this question float by the other day, but I had not gotten so far as to observe it.   Chrome now reaches an "open" state after my handshake response , but than immediately closes. Any idea what thats about?

[10:24:00.0000] <drclue>
Is this IRC channel logged somewhere that I might go look at yesterday's conversations?

[10:26:00.0000] <drclue>
Never mind that last , I see the link at the top of the screen

[10:31:00.0000] <Hixie>
drclue: you might be sending an extra byte after the handshake that chrome doesn't recognise

[10:32:00.0000] <drclue>
@hixie Thanks. That is the theory I'm working with now. Not sure where I might have picked it up , but none the less it seems a valifd thought

[10:41:00.0000] <micheil>
Hixie: do you know of any working draft76 & 75 servers for websockets other then the Go one?

[10:45:00.0000] <Dashiva>
http://www.flickr.com/photos/pixer/4681732186/sizes/o/

[10:45:01.0000] <Dashiva>
Geolocation might be an even stronger contender to HTML5-but-not-really than CSS3 :)

[10:54:00.0000] <jgraham>
It isn't in the colour palette things at the bottom though

[10:54:01.0000] <jgraham>
So the author is confused

[10:56:00.0000] <jgraham>
I love how the bar charts correspond not-at-all to the colour swatches

[10:57:00.0000] <miketaylr>
that's a feature

[10:58:00.0000] <jgraham>
miketaylr: ?

[10:59:00.0000] <miketaylr>
(sry, just being facetious)

[10:59:01.0000] <jgraham>
Ah

[11:00:00.0000] <jgraham>
Maybe it is designed to highlight the intellectual vacuousness of statements like "Opera is 72% HTML5 ready"

[11:01:00.0000] <Dashiva>
I didn't know all current browsers had messaging

[11:03:00.0000] <miketaylr>
yeah i didn't realize ie8 had postMessage either

[11:04:00.0000] <AryehGregor>
Heh, so true: http://en.wikipedia.org/w/index.php?title=Portable_Network_Graphics&diff=prev&oldid=366885637

[11:04:01.0000] <Dashiva>
Is there a "Best of Wikipedia Diffs" community, I wonder

[11:06:00.0000] <AryehGregor>
http://en.wikipedia.org/wiki/WP:BJAODN

[11:20:00.0000] <Hixie>
micheil: there's the one i wrote in perl: http://damowmow.com/playground/demos/websocket/

[11:20:01.0000] <Hixie>
micheil: is the new handshake causing you issues?

[11:21:00.0000] <micheil>
some, but I'm trying to work out how to correctly handle the messages right now

[11:24:00.0000] <Hixie>
ah

[11:24:01.0000] <Hixie>
that hasn't changed

[11:24:02.0000] <Hixie>
but the demos above should give you an idea if the spec isn't clear enough

[11:27:00.0000] <TabAtkins>
Hm, this is the first time I've ever actually seen <nobr> in the wild.

[11:27:01.0000] <TabAtkins>
But this site is a morass of bad design.

[11:27:02.0000] <volkmar>
There is a reason why input.min and input.max are DOMString and not float ?

[11:28:00.0000] <Hixie>
volkmar: they take dates when type=date

[11:28:01.0000] <volkmar>
Hixie: ok, thanks

[11:29:00.0000] <micheil>
Hixie: my main problem is that the data I get in isn't just a socket I read() off, but rather an unmutable buffer.

[11:32:00.0000] <Hixie>
ah

[11:35:00.0000] <jgraham>
/me thinks caniuse.com would be more useful if you could select n browsers that you care about and get a list of features supported by all n

[11:35:01.0000] <TabAtkins>
Go yell @Fyrd.

[11:36:00.0000] <gsnedders>
Loudly.

[11:36:01.0000] <micheil>
Hixie: I think the way my parser has to work is by using two nested while() loops.

[11:37:00.0000] <jgraham>
TabAtkins: On twitter? I don't really do twitter

[11:37:01.0000] <TabAtkins>
Well, do twitter.

[11:37:02.0000] <jgraham>
No

[11:38:00.0000] <gsnedders>
jgraham: You're just a grumpy old man.

[11:38:01.0000] <TabAtkins>
I sent him the suggestion for you.

[11:39:00.0000] <jgraham>
gsnedders: No, I just don't want to end up sounding like you

[11:39:01.0000] <jgraham>
It's bad enough that people confuse us in real life

[11:39:02.0000] <jgraham>
If it started happening on the internet...

[11:39:03.0000] <gsnedders>
I sound old? Cool. I mean, all teenagers want to be older. ;P

[11:40:00.0000] <gsnedders>
jgraham: It's already happened, I think. Though by people who confuse us in real life.

[11:40:01.0000] <jgraham>
gsnedders: No, I would sound like a teenager :p

[11:41:00.0000] <jgraham>
TabAtkins: Thanks

[11:41:01.0000] <gsnedders>
jgraham: You do anyway.

[11:43:00.0000] <jgraham>
gsnedders: Well no need to do so in a more multimedia way then

[11:43:01.0000] <drclue>
@hixie - Maybe I'm staring at things too hard and missing something obvious in constructing my handshake response. I've been looking for a stray byte to explain the "open" followed by "disconnect" problem , but I just don't see it. Heres a link to my handshake class (PHP) if anyone feels like taking a peek and pointing out my folly.  http://www.drclue.net/class.webSocketHandshake.inc

[11:43:02.0000] <Hixie>
/me looks (though i don't know php, so...)

[11:43:03.0000] <Hixie>
i get an error message

[11:44:00.0000] <jgraham>
/me to

[11:44:01.0000] <jgraham>
o

[11:46:00.0000] <erlehmann>
is there some RDFa -> microdata converter one of you can recommend ? my gsoc project went faster than expected, and why not add microdata.

[11:47:00.0000] <jgraham>
erlehmann: Not that I know of

[11:47:01.0000] <jgraham>
(what was the project?)

[11:49:00.0000] <erlehmann>
i am generating RDFa-enriched license markup. demo here: http://gsoc2010.dieweltistgarnichtso.net/?p=1

[11:49:01.0000] <erlehmann>
note the awesome CSS i wrote :D

[11:50:00.0000] <erlehmann>
you can extract the RDF using http://www.w3.org/2007/08/pyRdfa/#distill_by_uri

[11:52:00.0000] <Hixie>
rdfa to microdata is not a trivial conversion

[11:52:01.0000] <Hixie>
it's probably only usefully possible on a vocabulary-by-vocabulary basis

[11:53:00.0000] <erlehmann>
Hixie, i will then have to write it by hand, my use case is limited anyway.

[11:55:00.0000] <gsnedders>
Gah. Why did I start looking at lenses again? So bloody expensive.

[11:55:01.0000] <erlehmann>
Is there a way to have both RDFa and Microdata in the same markup ? (I may have asked this question months ago, but forgot the answer.)

[11:56:00.0000] <gsnedders>
As tempted as I am by the Canon EF 70-200 f/2.8 Mark 2, it's a bit expensive

[11:59:00.0000] <micheil>
Hixie: how does your perl server handle when someone does something like: for(var i=0; i<10000; i++){ conn.send("t"+i); }

[11:59:01.0000] <micheil>
conn being a client instance of WebSocket

[12:00:00.0000] <Hixie>
erlehmann: yeah, they don't overlap at all so you can easily include both

[12:00:01.0000] <Hixie>
micheil: why would that be a problem?

[12:01:00.0000] <micheil>
well.. currently I notice that it gets sent in multiple blocks

[12:02:00.0000] <micheil>
so, I get something on my server like: 0xFFtest0x000xFFtest0x00

[12:02:01.0000] <micheil>
although, it seems to arbitrarily get split

[12:03:00.0000] <Hixie>
you should get [0x00]test0[0xFF][0x00]test1[0xFF][0x00]test2[0xFF][0x00]test3[0xFF][0x00]test4[0xFF]...

[12:04:00.0000] <micheil>
rather, yeah

[12:04:01.0000] <micheil>
but something keeps killing my parser.. 

[12:04:02.0000] <micheil>
I guessing while(true) loops are not the way to go.

[12:05:00.0000] <Hixie>
seems to work fine: open http://software.hixie.ch/utilities/js/websocket/ then put in ws://damowmow.com:11111/demo and press the "Spam 100 packets" button

[12:05:01.0000] <Hixie>
(if you're using chrome, ignored the "undefined", which is due to a bug in chrome)

[12:05:02.0000] <jgraham>
gsnedders: IS?

[12:06:00.0000] <gsnedders>
jgraham: yeah

[12:06:01.0000] <jgraham>
That's insanely expensive

[12:06:02.0000] <gsnedders>
jgraham: OTW, I'm thinking of the 70-200 f/4L IS

[12:06:03.0000] <Hixie>
/me wonders what julian would consider adding something not "silently"

[12:06:04.0000] <gsnedders>
jgraham: But I really want a telephoto that quick

[12:06:05.0000] <jgraham>
Like way more than is worth paying unlerss you atre a pro

[12:07:00.0000] <jgraham>
gsnedders: for what?

[12:07:01.0000] <jgraham>
I think yagni applies

[12:07:02.0000] <gsnedders>
/me reminds jgraham that photography is effective at wasting money

[12:07:03.0000] <jgraham>
I know :)

[12:08:00.0000] <jgraham>
but pragmatism gets you more nice stuff for the same amount of money

[12:09:00.0000] <gsnedders>
I'm not entirely sure the f/4 has a field of depth small enough

[12:09:01.0000] <micheil>
okay.. got the parser working.. not to figure out wtf to do with broken packets.

[12:10:00.0000] <jgraham>
gsnedders: If you typo depth of field as field of depth yagni^2 applies

[12:10:01.0000] <micheil>
Hixie: this is what I mean about broken packets: http://gist.github.com/431901

[12:10:02.0000] <jgraham>
gsnedders: More to the point: for what?

[12:10:03.0000] <gsnedders>
But well, theoretical concern. I'm not buying the f/2.8

[12:10:04.0000] <micheil>
Hixie: that was generated from google chrome, 5.0.375.70, mac os x 6.02

[12:10:05.0000] <gsnedders>
jgraham: Hah. I really haven't been speaking English properly all week.

[12:11:00.0000] <micheil>
erm, 10.6.3, not 6.02

[12:11:01.0000] <drclue>
Since there might be somebody here now that has had this issue before , has anyone run into a situation where chrome accepts the handshake , goes to "open" and then immediately closes?

[12:11:02.0000] <gsnedders>
jgraham: (Like, how much of the time when you last saw me did I make sense?)

[12:14:00.0000] <gsnedders>
/me digs around and finds the f/4 lens is nowhere near as deep as he was expecting

[12:18:00.0000] <tom547>
Hi all. I am reading about HTML5/XHTML5 and it mentions that either html or xhtml should be used where appropriate, but can someone kindly explain to me how one decides which is more appropriate to learn and use. I am not comprehending the faq.

[12:21:00.0000] <AryehGregor>
tom547, use text/html for normal web pages.  XHTML5 is not very interesting to typical web authors.

[12:25:00.0000] <tom547>
out of curiosity what does xhtml5 turn pages into?

[12:26:00.0000] <AryehGregor>
Giant man-eating cougars.

[12:26:01.0000] <AryehGregor>
. . . No, really, what does that question mean?  They're XHTML5 pages.

[12:27:00.0000] <tom547>
lol. I am trying to figure out what the difference is, I keep on reading abuot it yet don't understand what the heck the fuss is about.

[12:27:01.0000] <AryehGregor>
XHTML is XML.  text/html is its own idiosyncratic format.

[12:27:02.0000] <Dashiva>
Just pretend XHTML5 doesn't exist

[12:27:03.0000] <AryehGregor>
Reasons to use text/html: IE doesn't support XHTML before IE9, and XHTML fails fatally on syntax errors.

[12:27:04.0000] <erlehmann>
tom547, xml has very few and clear rules. it is also easy to get wrong.

[12:27:05.0000] <AryehGregor>
Reasons to use XHTML: ???

[12:28:00.0000] <AryehGregor>
There are a couple of reasons to use XHTML instead of text/html, but they're mostly theoretical and massively outweighed by the disadvantages for normal authors.

[12:28:01.0000] <Dashiva>
XHTML and lojban, freinds for life

[12:28:02.0000] <erlehmann>
tom547, in contrast, html has an arcane, convoluted and not intuitive synthax, but recovers from pretty much all errors.

[12:28:03.0000] <AryehGregor>
Also works in all browsers.

[12:28:04.0000] <AryehGregor>
(interoperably if you stick to valid markup, mostly)

[12:28:05.0000] <Dashiva>
That's not exactly true, erlehmann

[12:29:00.0000] <Dashiva>
The arcane syntax is SGML, not HTML.

[12:29:01.0000] <erlehmann>
Dashiva, not ?

[12:29:02.0000] <tom547>
ok so it's just a different format of coding?

[12:30:00.0000] <erlehmann>
tom547, i'd suggest you write according xml rules (close every element, the final slash should not be a problem for html5 parsers) and send that as text/html

[12:30:01.0000] <erlehmann>
why ? because it is easy to remember.

[12:31:00.0000] <jgraham>
gsnedders: Let's try science http://www.cambridgeincolour.com/tutorials/depth-of-field.htm

[12:31:01.0000] <jgraham>
gsnedders: the calculator in particular

[12:31:02.0000] <AryehGregor>
Hixie, do you expect anything positive to come of starting a flame war with the HTMLWG chairs?

[12:34:00.0000] <AryehGregor>
I mean, truth and justice are nice and all, but I don't see them as being good for the web in this case.

[12:35:00.0000] <Dashiva>
If nobody speaks up, nothing will change

[12:35:01.0000] <micheil>
Hixie: if binary messages over websockets always send their lengths, then why not do the same for utf8 messages?

[12:35:02.0000] <AryehGregor>
And if someone speaks up, something will change, maybe.  Will that change be for the better?

[12:36:00.0000] <AryehGregor>
I don't see any likely outcome except the WHATWG being polite to the W3C, or the two of them splitting apart completely and no longer collaborating on HTML5.

[12:37:00.0000] <AryehGregor>
The latter would mean that implementers would have to choose which spec to follow, and I doubt that they'd all choose the same one.

[12:37:01.0000] <AryehGregor>
Well, or the WHATWG could be impolite to the W3C, the W3C could ignore the WHATWG and remove all references to it from its drafts, and the status quo otherwise remain unchanged.  That doesn't sound like a big improvement either.

[12:39:00.0000] <Dashiva>
That seems like the old-fashioned view of specs dictating reality

[12:39:01.0000] <AryehGregor>
Which part?

[12:39:02.0000] <Dashiva>
"implementers would have to choose which spec to follow"

[12:40:00.0000] <Dashiva>
The spec is a tool for interoperability, it is not a goal in itself

[12:40:01.0000] <gsnedders>
jgraham: Oh, wait. Now I feel really dumb. My confusion was failing to account for the fact that although the depth of field is larger, the background is likely further away, so the fact the depth of field is bigger is irrelevant

[12:40:02.0000] <gsnedders>
/me fails at physics

[12:40:03.0000] <gsnedders>
Well, not really at physics. My physics was fine. Just applying the physics.

[12:40:04.0000] <AryehGregor>
It doesn't serve that goal if there's a conflicting spec, and some vendors follow one and some the other.

[12:41:00.0000] <Dashiva>
But that's an improbable outcome

[12:41:01.0000] <Dashiva>
The actors involved don't have the right motivations

[12:42:00.0000] <AryehGregor>
What's a probable outcome that's better than the status quo?

[12:42:01.0000] <Dashiva>
HTML5 exists in the first place because the vendors themselves wanted interoperability

[12:42:02.0000] <AryehGregor>
You mean, the vendors other than Microsoft, which never joined until the HTMLWG formed.

[12:42:03.0000] <AryehGregor>
And has basically never acknowledged the existence of the WHATWG as far as I know.

[12:43:00.0000] <Dashiva>
Yeah, and they aren't using acid3 in their progress reports or anything

[12:43:01.0000] <Dashiva>
It's not about specs and organizations, it's about actual results in the field

[12:44:00.0000] <drclue>
Finally got my handshake working again. It appears that last little twist of close on open was simply me being tired and including the wrong server harness code from my code piles. So anyone who needs a handshake class for PHP , please feel free to steal a copy. http://www.drclue.net/class.webSocketHandshake.inc

[12:44:01.0000] <AryehGregor>
Getting actual results in the field requires getting all the major implementers to agree on a spec to follow.

[12:45:00.0000] <Dashiva>
No, it requires them to implement the same behavior. The spec is a tool to make that easy (or at least easier).

[12:45:01.0000] <jgraham>
I'm not sure implementors are the biggest problem here

[12:45:02.0000] <jgraham>
They can work things out amongst themselves, and WHATWG promises to follow implementations

[12:45:03.0000] <jgraham>
(assuming two competing specs)

[12:46:00.0000] <AryehGregor>
Dashiva, Microsoft doesn't seem to treat it that way.  I've seen multiple cases where they implement clearly wrong behavior by following specs to the letter.

[12:46:01.0000] <AryehGregor>
I would predict that there's a good chance they'd follow the HTMLWG spec no matter what.  If the HTMLWG version is no longer edited by Hixie and is no longer clear or sane, this is bad for the web.

[12:47:00.0000] <jgraham>
But authors are a problem because two competing language definitions causes confusion

[12:48:00.0000] <AryehGregor>
That's true too, although not a huge deal, since most authors don't care about specs anyway.

[12:48:01.0000] <Dashiva>
Authors are going to use a (possibly non-conforming) subset anyway

[12:49:00.0000] <drclue>
What would be nice  is if we are going to be developing code against an evolving spec that chrome and other releases don't just cut off the old implementation but leave one old and one new so that work does not have to come to a screaming halt while folks adjust bottleneck code like WebSockets. 

[12:50:00.0000] <Dashiva>
We most certainly do _not_ want to entrench outdated support

[12:51:00.0000] <drclue>
No , not entrench outdated support , but rather allow that some folks might be working on other things , while someone is updating to the newest spec.  

[12:52:00.0000] <AryehGregor>
The code is considered experimental at this point, so people aren't supposed to be using it for production stuff unless they're willing to keep up.

[12:52:01.0000] <AryehGregor>
Maybe the Chrome developers could make this clearer.

[12:52:02.0000] <Dashiva>
Remember how much fun autoplay and autobuffer was

[12:53:00.0000] <jgraham>
Authors may not care about specs but they care about perceptions of specs

[12:53:01.0000] <jgraham>
If there are two competing versions of the "same" spec, the added complexity may cause them to avoid both

[12:55:00.0000] <drclue>
A willingness to keep up is not the point. We all know that sooner or later the spec will settle and there will be a stable version of  WebSockets , and so we are developing applications that will use it. It would simply be nice to have the last version available so that other folks can keep going on their bits , while those in the "keeping up" department do their thing.

[12:55:01.0000] <AryehGregor>
drclue, #chromium might be a better place to complain.

[12:56:00.0000] <drclue>
I'll pop over there and suggest it.

[12:57:00.0000] <Dashiva>
If you desire a frozen support set, you could always just freeze the browser version

[12:57:01.0000] <AryehGregor>
By the way, my original remark to Hixie was about his response to Sam's request to halt heartbeat publication of HTML5.  I didn't see the later thread at that point, which looks like it might actually be productive.

[12:58:00.0000] <AryehGregor>
Explicitly saying in the WHATWG draft that sections are missing from the W3C draft because of politics is not productive.

[12:58:01.0000] <jgraham>
AryehGregor: FWIW I tend to agree with that

[12:58:02.0000] <jgraham>
Even though I think what is said in the WHATWG draft is largely accurate

[12:58:03.0000] <AryehGregor>
I somewhat agree.

[12:58:04.0000] <AryehGregor>
But it doesn't help anything to put it there.

[12:58:05.0000] <jgraham>
Indeed

[12:59:00.0000] <Dashiva>
Does it help anything to complain about it being put there?

[12:59:01.0000] <AryehGregor>
No, but I stand less chance of convincing the W3C than I do of convincing Hixie, I think.

[12:59:02.0000] <Hixie>
AryehGregor: I'm not trying to start a flame war. I'm trying to get coherent editing guidance to edit the W3C draft and I'm trying to answer questions people have asked me in the WHATWG draft.

[13:00:00.0000] <Dashiva>
People keep complaining about the lack of rationale, but these chair decisions aren't exactly helping

[13:00:01.0000] <Philip`>
Remove the paragraph from the WHATWG draft for a few days, get the W3C snapshot published, then put it back into the WHATWG draft - easy solution

[13:00:02.0000] <AryehGregor>
Hixie, on the latter point, you could answer them without explicitly stating anything that's derogatory toward the W3C.

[13:01:00.0000] <AryehGregor>
The former point I'm fine with.

[13:01:01.0000] <Dashiva>
It's a shame the truth has to be such as dirty thing

[13:01:02.0000] <Hixie>
AryehGregor: if the truth is derogatory, the problem isn't with the statement saying the truth.

[13:03:00.0000] <AryehGregor>
Hixie, saying the truth directly instead of indirectly is often not productive or wise.  Just because something is correct doesn't mean it's pragmatically a good idea to say it.

[13:03:01.0000] <AryehGregor>
You can rephrase things so as to be still true but less hostile.

[13:03:02.0000] <drclue>
@Hixie I tend to both agree and disagree.  The truth is the truth , but packaging makes a lot of difference in how long it takes to resolve an issue

[13:04:00.0000] <Hixie>
AryehGregor: people are asking me why the drafts are different. Can you give me text that is (a) true, (b) accurately answers the question, and (c) isn't beating around the bush?

[13:04:01.0000] <Hixie>
s/accurately/accurately and usefully/

[13:04:02.0000] <Dashiva>
Besides, it's only derogatory if you agree politics is a bad thing

[13:04:03.0000] <Dashiva>
Shouldn't you then be more concerned about bad things running the show?

[13:05:00.0000] <Fyrd>
jgraham: Re: Your suggestion for caniuse.com - So like http://caniuse.com/#agents=gecko,webkit_chr but not showing any feature with red in it? If not, could you elaborate?

[13:05:01.0000] <AryehGregor>
Hixie, "I don't know.  Here's a link to the decision that made me remove the text, but I don't understand it."

[13:05:02.0000] <Hixie>
Dashiva: i am, hence my question about editing guidance

[13:05:03.0000] <drclue>
Politics sucks , and often has little to do with achieving the best results , but same is a fact of life

[13:05:04.0000] <Hixie>
AryehGregor: you think that would be less upsetting?

[13:06:00.0000] <AryehGregor>
Hixie, yes.  You aren't accusing them of anything, you're just saying you don't understand them.

[13:06:01.0000] <Hixie>
AryehGregor: the current text doesn't accuse anyone of anything either

[13:06:02.0000] <Hixie>
it just states a fact

[13:06:03.0000] <AryehGregor>
"The W3C version omits a paragraph of implementation advice for political reasons" accuses the W3C of removing that paragraph for political rather than technical reasons, which is not a characterization they'd agree with.

[13:07:00.0000] <AryehGregor>
Linking to the actual decision rationale rather than summarizing it in a way that its authors disagree with would not only be more politic, it would also be more accurate.

[13:07:01.0000] <jgraham>
Fyrd: Yes, and working for any number of UAs, not just 2

[13:07:02.0000] <Hixie>
AryehGregor: it's also not really true -- i understand the reason for the change, the reason is that the chairs made a political decision instead of a technical one and then tried to bury that fact in a pseudo-technical description of their decision

[13:07:03.0000] <Dashiva>
The decision process itself is pretty clear that it's a political process

[13:07:04.0000] <drclue>
That is some harsh text , not sure how much forward motion one would hope for out of that , no matter how true it was

[13:08:00.0000] <jgraham>
FrOh, wait, it is the comparison thing that only works for 2

[13:08:01.0000] <jgraham>
Fyrd: ^

[13:08:02.0000] <Fyrd>
Right.

[13:09:00.0000] <Fyrd>
So maybe if I added a "hide features with missing support" checkbox, would that work?

[13:09:01.0000] <AryehGregor>
Hixie, incorporating that interpretation into the actual text of the spec does not help anything substantial and does harm something.  Namely, the chairs will apparently require that the W3C version of the spec not contain links to the WHATWG version.

[13:09:02.0000] <jgraham>
Fyrd: Well I was imaging something more like the comparison thing, but with N > 2 browsers and no table, just a list of supported features

[13:10:00.0000] <Fyrd>
Hmm...okay.

[13:10:01.0000] <Fyrd>
I think I've wanted that before myself, actually.

[13:10:02.0000] <Fyrd>
Will look into it.

[13:10:03.0000] <jgraham>
Fyrd: The use case was someone who has to su[port specific browsers wondering what tech they can use

[13:11:00.0000] <Fyrd>
Ah, gotcha.

[13:11:01.0000] <Fyrd>
Specific browser versions then too, yeah?

[13:11:02.0000] <jgraham>
Yeah

[13:11:03.0000] <drclue>
@AryehGregor  That's why I have a drafts folder , so I can be as "truthful" as I want , and then once I cool a bit I can gift wrap 

[13:11:04.0000] <Fyrd>
Alright, I see now. Thanks for the suggestion, jgraham.

[13:12:00.0000] <Fyrd>
Anyone else have any caniuse.com suggestions while I'm here?

[13:13:00.0000] <Dashiva>
An option for what jgraham suggested where it also has a list of features you can't use

[13:14:00.0000] <Fyrd>
Dashiva: Hm, alright.

[13:14:01.0000] <Dashiva>
http://lists.w3.org/Archives/Public/public-html/2010Jun/0242.html

[13:15:00.0000] <Dashiva>
Funny how JF worries about the users when it's the W3C that refuses to accept design principles such as priority of constituencies

[13:15:01.0000] <drclue>
If there are features one cannot use , why not just have a CSS class for them so that they can stay structurally in the same place and when they evolve the CSS for them can be changed

[13:19:00.0000] <TabAtkins>
Hixie: Take sicking's advice and just rename Microdata to RDFa5.

[13:20:00.0000] <Hixie>
TabAtkins, meet AryehGregor. He's trying to get me to _reduce_ the political friction. :-P

[13:20:01.0000] <Dashiva>
TabAtkins: 5 is too controversial

[13:20:02.0000] <AryehGregor>
/me hits TabAtkins with a stick

[13:20:03.0000] <Dashiva>
Maybe RDFa 4, followed by 4.01

[13:22:00.0000] <drclue>
I would probably keep my powder dry with the w3.org types and save the debate for things that really matter. (Unless one really thinks this is a big deal)

[13:23:00.0000] <Hixie>
can anyone remember why we added a reference to second WCAG to the w3c html draft?

[13:23:01.0000] <Hixie>
a second reference to

[13:23:02.0000] <Hixie>
not a reference to second

[13:24:00.0000] <Hixie>
i can't find the relevant e-mail

[13:24:01.0000] <Hixie>
/me is trying AryehGregor's suggested approach of citing e-mails rather than referring to things as political

[13:25:00.0000] <AryehGregor>
Dashiva, see, I told you I have more chance of convincing Hixie.

[13:25:01.0000] <AryehGregor>
Now I just have to convince him that legacy presentational markup shouldn't be lumped into the same category as things like parse errors and <image>.

[13:26:00.0000] <TabAtkins>
AryehGregor: If you ever come down to the bay area, I visited a great kosher restaurant last night.  Slow service, but food was quite good.

[13:26:01.0000] <drclue>
The lost art of winning by not having your opponent loose 

[13:26:02.0000] <AryehGregor>
TabAtkins, what was it called?

[13:26:03.0000] <Hixie>
apparnetly the answer is http://www.w3.org/Bugs/Public/show_bug.cgi?id=9241

[13:26:04.0000] <TabAtkins>
The Kitchen Table, iirc.

[13:26:05.0000] <jgraham>
TabAtkins: That description got worse as it went on

[13:26:06.0000] <jgraham>
from great -> slow -> quite good

[13:27:00.0000] <jgraham>
well I guess slow is worse than "quite good"

[13:27:01.0000] <Dashiva>
AryehGregor: You haven't actually solved the problem, though

[13:27:02.0000] <TabAtkins>
Yes, it is.

[13:27:03.0000] <jgraham>
but neither really sums to "great"

[13:27:04.0000] <TabAtkins>
"quite good" can be equivalent to "great".

[13:27:05.0000] <TabAtkins>
/me doesn't impose a total ordering on his comparitives.

[13:28:00.0000] <AryehGregor>
Dashiva, not yet.

[13:28:01.0000] <jgraham>
No, great is strictly better than quite good

[13:28:02.0000] <TabAtkins>
Fine.

[13:28:03.0000] <Dashiva>
AryehGregor: If anything, it might get worse, since now more people will be exposed to the full insanity of the decisions

[13:28:04.0000] <TabAtkins>
s/food was quite good/food was OMGWTFBBQORGASMIC/

[13:29:00.0000] <jgraham>
You get orgasams at BBQs? That must be socially akward

[13:29:01.0000] <TabAtkins>
That's just how we texans roll.

[13:29:02.0000] <Philip`>
Dashiva: People can easily block out insanity when it's disguised with sufficient waffle

[13:29:03.0000] <TabAtkins>
We're quite serious about our bbq.

[13:29:04.0000] <jgraham>
Mmmm waffles

[13:30:00.0000] <Dashiva>
Philip`: Now why did you have to bring up waffles. Now there'll be a waffle war.

[13:30:01.0000] <Philip`>
Dashiva: so you just need to be careful not to summarise the decisions in a single sentence, because that will expose people to it unfairly

[13:30:02.0000] <jgraham>
Is that like a beef war?

[13:31:00.0000] <jgraham>
Or more like a food fight?

[13:31:01.0000] <Hixie>
AryehGregor: how is the new text at http://www.whatwg.org/specs/web-apps/current-work/#is-this-html5? ?

[13:31:02.0000] <Dashiva>
It's when I point out that American waffles are horrible dry things with no taste

[13:32:00.0000] <TabAtkins>
Dashiva: How can anyone disagree, though?

[13:32:01.0000] <Dashiva>
I don't know, but they still do

[13:34:00.0000] <AryehGregor>
Hixie, I suggest you remove "redundant and inconsistent".  (If it's redundant and inconsistent, why didn't you say that in the bug you linked to?)  Other than that, you'll have to ask the chairs, I guess.

[13:35:00.0000] <AryehGregor>
TabAtkins, a restaurant that claims to be kosher but doesn't mention on its website who certifies it as kosher strikes me as extremely suspicious.

[13:35:01.0000] <TabAtkins>
AryehGregor: They mentioned the certification on their menu, though I didn't pay enough attention to it to remember who it was.

[13:36:00.0000] <AryehGregor>
TabAtkins, it's very possibly some rabbi that no one's heard of and no one reasonable would trust.

[13:36:01.0000] <TabAtkins>
Hixie: CSS should probably be a part of that last section.

[13:36:02.0000] <TabAtkins>
AryehGregor: Sure, that's possible.

[13:36:03.0000] <TabAtkins>
There were plenty of jews there, though, so you can at least get some assurance from the crowd.

[13:37:00.0000] <Hixie>
AryehGregor: i didn't say it in the bug because it wasn't worth my time to whine about it

[13:37:01.0000] <AryehGregor>
Hixie, but it's worth your time to whine about it in the WHATWG draft?

[13:37:02.0000] <zcorpan_>
Hixie: then why whine about it in the spec

[13:40:00.0000] <drclue>
"The W3C version includes a second reference to the WCAG document, with the two having some inconsistancies" 

[13:41:00.0000] <AryehGregor>
That's poor grammar.

[13:41:01.0000] <drclue>
OK , fix the grammar , but the idea is to ditch the blame fest

[13:42:00.0000] <AryehGregor>
You're still saying they're inconsistent.  Just ignore that.  If it's worth pointing out, point it out in the W3C and make sure your link to the reasoning includes that objection.

[13:43:00.0000] <AryehGregor>
No one will complain if you link to a discussion that happens to include objections, they'll complain if the spec text actually endorses one of those objections.

[13:43:01.0000] <AryehGregor>
(I mean, some people might complain, but I don't think anyone important will complain.)

[13:43:02.0000] <Hixie>
AryehGregor: it's not whining, it's just explaining that they're different

[13:44:00.0000] <drclue>
Why even call it an objection? Perhaps a need to consolidate or something. 

[13:45:00.0000] <AryehGregor>
Hixie, I think it will needlessly aggravate people.  My advice is to not include that wording in the spec.  That section only serves to highlight the differences, it doesn't need to include all the reasoning inline.

[13:45:01.0000] <Hixie>
the section serves to answer the question "why are they different" which i'm getting asked a lot

[13:46:00.0000] <Hixie>
zcorpan_: what should i link to, re CSS?

[13:46:01.0000] <zcorpan_>
maybe it should be in a wiki page instead of in the spec

[13:46:02.0000] <AryehGregor>
Then say "because of a request by XXX", like with the decision link.  If you want to say why you didn't implement it in the WHATWG draft, maybe say it in the bug.  I don't know.

[13:46:03.0000] <zcorpan_>
Hixie: CSS?

[13:46:04.0000] <AryehGregor>
TabAtkins, most Jews aren't religious at all, and many of those who consider themselves religious wouldn't be considered religious by most Orthodox Jews, so it's not really helpful to say that there are other Jews there.

[13:47:00.0000] <Hixie>
zcorpan_: oops, that was tab's idea

[13:47:01.0000] <TabAtkins>
Okay.

[13:47:02.0000] <Hixie>
TabAtkins: what should i link to, re CSS?

[13:47:03.0000] <drclue>
I would think Hixie is right in pointing out disparities , it just needs to be in a neutral tone 

[13:47:04.0000] <TabAtkins>
http://www.w3.org/Style/CSS/current-work ?

[13:48:00.0000] <AryehGregor>
The proper place to criticize the W3C's decisions is in the places the W3C has set up for that purpose, not in the official spec text of an organization that's supposed to be cooperating with the W3C.

[13:48:01.0000] <TabAtkins>
That page describes itself as "the place to start" if "you want to follow the development of CSS3".

[13:48:02.0000] <Hixie>
AryehGregor: k, reload?

[13:48:03.0000] <Hixie>
TabAtkins: k

[13:49:00.0000] <drclue>
My point exactly!  The declaration needs to be phrased as other than criticism , but none the less it is worth mentioning. 

[13:50:00.0000] <AryehGregor>
Hixie, that looks much better to me.  I suggest you ask the chairs if they're okay with it.

[13:50:01.0000] <Hixie>
k

[13:50:02.0000] <AryehGregor>
(and by "the chairs" I mean Sam, since Maciej is reportedly on vacation and Paul participates about as much as any Microsoft employee)

[13:50:03.0000] <AryehGregor>
(marginally more, I guess)

[13:50:04.0000] <zcorpan_>
Hixie: 'WebSRT' didn't get xreffed

[13:53:00.0000] <Hixie>
zcorpan_: yeah, that'll get fixed in due course

[13:59:00.0000] <gsnedders>
/me wonders what it is about him using words incessently like "normally" last week (esp. Saturday) and "inevitably" this week

[14:04:00.0000] <TabAtkins>
/me just took the W3C copy of Microdata and did a simple search-and-replace for "microdata"->"RDFa5".  It reads astonishingly well.

[14:05:00.0000] <zcorpan_>
TabAtkins: you should publish a FPWD

[14:05:01.0000] <AryehGregor>
I'm pretty sure the RDFa WG would object.

[14:05:02.0000] <TabAtkins>
For real.

[14:06:00.0000] <AryehGregor>
Hixie has trouble understanding other people's points of view if they don't seem logical to him.  You just like causing trouble.

[14:07:00.0000] <TabAtkins>
That's not true.

[14:07:01.0000] <AryehGregor>
I wasn't really serious.  :P

[14:07:02.0000] <TabAtkins>
I genuinely think that many people's objections to Microdata are that it doesn't contain the string "RDFa" in its name.  This is an easy fix.

[14:07:03.0000] <AryehGregor>
Okay, so I take it back, you aren't good at understanding other people's points of view either.

[14:08:00.0000] <AryehGregor>
Come on, you really think that?

[14:08:01.0000] <AryehGregor>
People's objections are that they and all their work got sidestepped.

[14:08:02.0000] <AryehGregor>
It's not the name.  RDFa 2 is based on RDFa 1.1, and made by basically the same people.  That's the difference.

[14:08:03.0000] <drclue>
Hixie suffers from having an actual IQ, and the natural impatience that can bring on in certain situations

[14:09:00.0000] <TabAtkins>
AryehGregor: I'm not talking about the people who worked on RDFa.

[14:09:01.0000] <TabAtkins>
I'm referring to the sideline commenters.

[14:09:02.0000] <AryehGregor>
They objected for the same reasons, because they see it as one person engaging in a pattern of marginalizing those he disagrees with.

[14:09:03.0000] <AryehGregor>
Which in other cases includes them, typically.

[14:09:04.0000] <TabAtkins>
I think you're imputing far too much thought to most of the positions held.

[14:10:00.0000] <zcorpan_>
Hixie: 404 - http://www.whatwg.org/specs/web-apps/current-work/complete/Essays1743.ttf

[14:10:01.0000] <AryehGregor>
I think that if you refuse to accept that people you disagree with have actual motives and aren't just acting randomly out of stupidity, you're not going to get very far in any situation requiring compromise and discussion.

[14:11:00.0000] <TabAtkins>
Now you're imputing too little thought to my objection.  ^_^

[14:11:01.0000] <AryehGregor>
I didn't say this is what they consciously think, by the way.  I said it's the reason.

[14:11:02.0000] <AryehGregor>
I mean, the reason for their actions.

[14:11:03.0000] <AryehGregor>
Also, they have the idea that the web data community has expertise, and microdata ignores that expertise.

[14:12:00.0000] <Hixie>
zcorpan_: oops

[14:12:01.0000] <TabAtkins>
Yes, they have that idea.  Which is incorrect, of course, as Microdata is almost functionally equivalent to RDFa.

[14:13:00.0000] <AryehGregor>
The basic point is that the WHATWG supporters think Hixie is smart and does a good job of balancing different concerns against each other, and the W3C supporters think that it's better to get everyone to agree to compromises rather than having one well-situated person arbitrate.

[14:14:00.0000] <Dashiva>
You mean having three well-situated people arbitrate instead of one

[14:14:01.0000] <AryehGregor>
Yes, that's technically true, but everyone pretends it isn't.

[14:14:02.0000] <AryehGregor>
It's also not quite true in practice, because the chairs try to minimize objections, while Hixie tries to do whatever is best regardless of objections.

[14:14:03.0000] <zcorpan_>
four - first hixie arbitrates and then three people arbitrate if someone objects

[14:14:04.0000] <Hixie>
don't forget tim

[14:15:00.0000] <zcorpan_>
fine, so five

[14:15:01.0000] <Hixie>
(you gotta love a process where the more you escalate something, the less knowledgable in the subject is the person making the decision)

[14:15:02.0000] <AryehGregor>
Isn't that how most processes work?

[14:16:00.0000] <drclue>
I tend to at times suffer the same handicap of pushing for whats best, and it does get me into trouble some times.

[14:16:01.0000] <AryehGregor>
People with broader authority are automatically going to have less expertise in any particular thing.

[14:17:00.0000] <Hixie>
AryehGregor: it's not how the whatwg process works -- if you escalate from me, the people who decide are all even more involved in the web than i am

[14:17:01.0000] <AryehGregor>
There's no formal procedure to escalate to the steering committee, though, right?

[14:18:00.0000] <AryehGregor>
What it comes down to is that the steering committee is reps of major browsers, and if any of the major browsers object, then you can't put it in the spec anyway.

[14:18:01.0000] <drclue>
Sometimes even the folks heavily involved , sorta have blinders on.  While I dutifully switched from tables to CSS tables , I'm glad to see the old tables back, so I have a choice.

[14:19:00.0000] <AryehGregor>
What old tables?

[14:19:01.0000] <TabAtkins>
Indeed, what?

[14:19:02.0000] <TabAtkins>
/me suspects drclue fell prey to some of the more misguided anti-table evangelism.

[14:20:00.0000] <drclue>
The table elements that went away for a while , or at least were strongly discouraged. I myself like tables

[14:20:01.0000] <zcorpan_>
http://chestofbooks.com/food/household/Woman-Encyclopaedia-4/images/Chippendale-occasional-table-with-gallery-and-carved-legs-T.jpg

[14:20:02.0000] <zcorpan_>
that's a pretty old table

[14:20:03.0000] <TabAtkins>
There is not and has never been anything wrong with tables, nor have they every disappeared.

[14:20:04.0000] <TabAtkins>
The problem is and has always been using tables as a page-layout tool.

[14:22:00.0000] <TabAtkins>
Some people read "don't use tables for layout" and heard "don't use tables", and then went crazy trying to design tables without using <table>s.  Which was silly and ridiculous, and caused some harm to people who thought the standardistas must have known what they were talking about.

[14:22:01.0000] <drclue>
Well, yes , I agree that having long running open elements was bad on performance , but it actually got to a point where project specs strictly prohibited ANY tables.

[14:22:02.0000] <TabAtkins>
That's retarded.

[14:22:03.0000] <drclue>
I know it's retarded

[14:22:04.0000] <TabAtkins>
You use <table>s to mark up tables.  You don't use them to mark up non-tables.

[14:23:00.0000] <gsnedders>
/me gets the point of buying lens, then realizes, "wait, I don't have enough money in this account, damned Swedes ot giving me a card I can use online"

[14:23:01.0000] <drclue>
I use tables for tabular data, and not much more

[14:23:02.0000] <gsnedders>
/me is not bitter or anything

[14:25:00.0000] <drclue>
Occasionally I bend my table rules , but only a little and never to the grand scale of abuse that seemed to give tables their bad rep

[14:26:00.0000] <drclue>
Of course I run into folks who get so rigid in their particular design rules that they would rather break a page than break a rule :) 

[14:27:00.0000] <Hixie>
so btw do i win anything for being right about how shelley leaving the group would have no effect on her posting to the list? :-)

[14:27:01.0000] <AryehGregor>
That doesn't even count as a prediction.  Hasn't she done it like four times already?

[14:30:00.0000] <Hixie>
filed http://www.w3.org/Bugs/Public/show_bug.cgi?id=9894 on the process

[14:31:00.0000] <Dashiva>
Hixie: That bet was never valid in the first place, since she had "left" before

[14:35:00.0000] <AryehGregor>
Hixie, this is what banning legacy presentational markup results in.  http://www.mediawiki.org/wiki/Special:Code/MediaWiki/66500

[14:38:00.0000] <Hixie>
AryehGregor: what am i looking at?

[14:39:00.0000] <AryehGregor>
"Replace <tt> by <code>, as <tt> is deprecated in HTML5"

[14:39:01.0000] <AryehGregor>
Followed by a diff that does a search-and-replace of <tt> for <code>.

[14:39:02.0000] <hober>
Hixie: no one took you up on that bet, right?

[14:44:00.0000] <Hixie>
AryehGregor: well so long as you're trying to convert a presentational language (wiki) into a semantic one (html), you're doomed

[14:45:00.0000] <AryehGregor>
Hixie, 1) The commit had nothing to do with wikitext.  It affected only MediaWiki's wikitext-independent output; <tt> in wikitext still gives you <tt> in HTML output.  2) Wikitext is mostly just a macro language on top of HTML, and is no more or less semantic than HTML itself.

[14:46:00.0000] <Hixie>
oh, ok

[14:46:01.0000] <Hixie>
well

[14:46:02.0000] <Hixie>
r-

[14:46:03.0000] <Hixie>
"this is the wrong fix, please go and actually make sure you use semantics correctly"

[14:46:04.0000] <AryehGregor>
Yes, great, but that doesn't fix the fact that other authors will be doing the same thing for the same reasons.

[14:46:05.0000] <AryehGregor>
It's illustrative.

[14:47:00.0000] <AryehGregor>
Most authors don't understand semantic markup.  If you tell them that the new standard says you can only use semantic markup, they'll just misuse it horribly.  Or, at best, translate automatically to <span style="font-family:monospace">.

[14:47:01.0000] <AryehGregor>
Wordpress (AFAICT) gives you an italics button, which looks exactly like the italics button in Microsoft Word, which outputs <em>.

[14:48:00.0000] <AryehGregor>
And a bold button that outputs <strong>.

[14:48:01.0000] <Hixie>
yes, people misuse html

[14:48:02.0000] <Hixie>
news at 11

[14:49:00.0000] <AryehGregor>
If you aren't designing the semantics to minimize misuse by authors, what's the point?  If even the authors of really large web apps don't get it, who will?

[14:49:01.0000] <Hixie>
the problem is that authors write inaccessible pages

[14:49:02.0000] <Hixie>
whether they do it using <font> or misuse of <em>, the problem is the same

[14:50:00.0000] <AryehGregor>
No, because UAs will know that <font> is not semantic.  They won't know that for <em>.

[14:50:01.0000] <Hixie>
i don't see how that's ad advantage

[14:50:02.0000] <jgraham>
The general statement of the problem is "a general purpose UA can't make use of inconsistent semantics"

[14:50:03.0000] <jgraham>
(or at least can't make optimal use)

[14:50:04.0000] <AryehGregor>
If authors are encouraged to misuse semantic elements for presentation, you get a <table> situation, where you can't infer any semantics because it's usually used for presentation.

[14:50:05.0000] <AryehGregor>
(encouraged as in they have an incentive to, not as in they're told to)

[14:51:00.0000] <jgraham>
(but allowing <tt> doesn't help much because people will still pick the wrong choice an appreciable fractio of the time)

[14:51:01.0000] <Hixie>
ok now sam is just treating me like a child

[14:51:02.0000] <AryehGregor>
If you treat <tt> the same as a parse error, that's a strong reason for authors to replace it with <code> (which is shorter and easier to type than <span style="font-family:monospace">).

[14:52:00.0000] <AryehGregor>
If the validator just says "maybe you could consider using semantic markup instead if possible", then you'll increase the signal-to-noise ration in semantic markup.

[14:53:00.0000] <Hixie>
we tried that in html4

[14:53:01.0000] <AryehGregor>
And authors didn't use semantic markup.  Okay, well, they aren't going to anyway.

[14:53:02.0000] <AryehGregor>
So nothing lost.

[14:53:03.0000] <AryehGregor>
Forcing them to either mass-convert their pages to a syntax that's (at best) logically equivalent, or just give up on validation, is not a win.

[14:54:00.0000] <Hixie>
bb in 20

[14:54:01.0000] <AryehGregor>
<tt> will mostly either get converted to <code> (wrong) or <span style="font-family:monospace"> (no better than <tt>).

[14:54:02.0000] <AryehGregor>
Okay.

[14:54:03.0000] <AryehGregor>
I've been meaning to reopen the bug I filed about this, I'll argue more there.

[14:59:00.0000] <AryehGregor>
Hmm.  It seems like the Nexus One engraving thing has a limited font selection.

[14:59:01.0000] <AryehGregor>
I'll have to rethink what I want to put there.  Sigh.

[15:13:00.0000] <Hixie>
back

[15:15:00.0000] <Hixie>
AryehGregor: i'm certainly in favour of making sure validators give error messages that are more appropriate than just "not allowed"

[15:15:01.0000] <Hixie>
AryehGregor: hence the long section on obsolete elements and the advice for how to replace them

[15:15:02.0000] <AryehGregor>
It's the verdict that matters to most people, not the details.  "Is this a standards-compliant document or not?"

[15:15:03.0000] <AryehGregor>
It's something that people have on their checklist for web application quality.

[15:15:04.0000] <Hixie>
well if they use presentational markup, it shouldn't be compliant

[15:16:00.0000] <Hixie>
because they shouldn't use presentational markup

[15:16:01.0000] <AryehGregor>
style="" is every bit as presentational as <tt>.

[15:16:02.0000] <Hixie>
yes, and i wanted to make that non-conforming too

[15:16:03.0000] <Hixie>
unfortunately it has a couple of valid use cases

[15:17:00.0000] <Hixie>
(as a rapid prototyping tool, and as a special-case override)

[15:17:01.0000] <Hixie>
it's non-conforming to use style="" in the same way as <tt>

[15:17:02.0000] <Hixie>
(though of course a validator can't easily report that)

[15:18:00.0000] <AryehGregor>
You can't make things not validate based on an idealized notion of what authors should do.  Validation has to reflect reality.  If it makes unreasonable demands, authors will ignore it or evade it.  You have to look at whether a requirement will actually improve pages in practice, not whether it would improve pages if authors hypothetically followed it.

[15:18:01.0000] <AryehGregor>
In practice, by allowing <code> and <span style="font-family:monospace"> but not <tt>, you'll just encourage authors to move en masse from the latter to the former.

[15:18:02.0000] <Hixie>
it's not an unreasonable demand to make pages accessible

[15:19:00.0000] <Hixie>
seriously. semantic markup is not _that_ complicated.

[15:19:01.0000] <Hixie>
people do understand it once they try.

[15:19:02.0000] <AryehGregor>
That position does not reflect reality.  Even highly competent authors rarely produce consistently semantic markup.

[15:19:03.0000] <AryehGregor>
None of the top websites consistently uses semantic markup, for instance.

[15:19:04.0000] <Hixie>
even highly competent software engineers write software with bugs, but we still expect them to try to not include bugs

[15:20:00.0000] <Hixie>
even highly competent doctors rarely have perfect track records in healing people but we still expect them to try

[15:20:01.0000] <AryehGregor>
The record shows that highly competent authors do not even try to consistently use semantic markup.

[15:20:02.0000] <Hixie>
nonsense

[15:20:03.0000] <AryehGregor>
Really?

[15:21:00.0000] <Hixie>
sure

[15:21:01.0000] <AryehGregor>
Have you looked at Gmail's HTML?  Does that look like trying to use semantic markup?

[15:21:02.0000] <AryehGregor>
Or google.com's HTML?

[15:21:03.0000] <TabAtkins_>
That looks like machine-generated markup coming from a java application.

[15:21:04.0000] <AryehGregor>
Precisely.

[15:21:05.0000] <AryehGregor>
And it's purely presentational.

[15:21:06.0000] <Hixie>
the existence of counter-examples is not proof of a generalisation

[15:21:07.0000] <TabAtkins_>
AryehGregor: I assure you that we are trying our best to move to better markup.

[15:22:00.0000] <Hixie>
there are certainly large numbers of pages that are written poorly

[15:22:01.0000] <AryehGregor>
So then where are your examples in support of the generalization?

[15:22:02.0000] <Hixie>
look at how people have jumped on the new html5 elements

[15:22:03.0000] <TabAtkins_>
Though I doubt it matters to use exactly what the tags are called - we just want to reduce the massive numbers of wrapping <div>s and similar.

[15:22:04.0000] <TabAtkins_>
s/use/us/

[15:22:05.0000] <AryehGregor>
I predict that the new elements won't be used correctly unless they provide tangible effects that make them differ practically from <div>s.

[15:23:00.0000] <Hixie>
there are plenty of languages that already exist if what you want is just a formatting language

[15:23:01.0000] <AryehGregor>
Yes, like HTML.

[15:24:00.0000] <AryehGregor>
Which has the advantage of being implemented in web browsers.

[15:24:01.0000] <AryehGregor>
Unlike, well, everything els.

[15:24:02.0000] <AryehGregor>
else.

[15:24:03.0000] <Hixie>
i find that to be a short-sighted approach: it's media-dependent, inaccessible without annotations, hard to maintain, etc.

[15:24:04.0000] <jgraham>
AryehGregor: Agreed, but they should if people implement the outlining algoritm

[15:24:05.0000] <jgraham>
(about tangible effects being critical)

[15:24:06.0000] <Hixie>
i'm interested in a language that is media-independent, accessible by default, easy to maintain, etc.

[15:24:07.0000] <AryehGregor>
You can write media-independent, maintainable HTML that's totally presentational.  That's mostly a matter of CSS.

[15:25:00.0000] <Hixie>
CSS is an optional technology

[15:25:01.0000] <AryehGregor>
Not in practice.

[15:25:02.0000] <Hixie>
yes in practice

[15:25:03.0000] <Hixie>
i use browsers with no CSS support regularly

[15:25:04.0000] <Hixie>
and I rely on a user agent with no CSS support constantly (google search)

[15:26:00.0000] <AryehGregor>
It can still be media-independent and maintainable without using semantic elements significantly.

[15:26:01.0000] <Hixie>
not anywhere near as easily as if you just use semantic elements

[15:26:02.0000] <AryehGregor>
Sure, but the difference is clearly not enough that most authors care.  You're left with the choice of telling those authors that they don't get a useful validator, or rethinking your approach.

[15:27:00.0000] <AryehGregor>
I think semantic HTML is great, and I use it myself wherever possible, but you have to take into consideration the practicalities of the matter.

[15:27:01.0000] <Hixie>
the practicalities of the matter are that it's time for authors to get with the programme

[15:28:00.0000] <jgraham>
Typically education doesn't work

[15:28:01.0000] <AryehGregor>
Then come up with an effective way of doing that.  Saying their pages don't validate if they contain machine-detectably non-semantic markup just means they'll move to non-semantic markup that's not machine-detectable.

[15:28:02.0000] <Hixie>
jgraham: it's worked pretty well for banning tables-for-layout

[15:29:00.0000] <AryehGregor>
Tables-for-layout died because CSS usually works better, not because of evangelism.

[15:29:01.0000] <Hixie>
i don't buy that at _all_

[15:29:02.0000] <Hixie>
css can't even do basic table layouts

[15:29:03.0000] <Hixie>
it's one of the things TabAtkins_ is working on

[15:29:04.0000] <AryehGregor>
Really?  Table-based markup is a nightmare to maintain.

[15:29:05.0000] <Hixie>
yes!

[15:29:06.0000] <AryehGregor>
Mostly you don't *want* table layouts, is the thing.

[15:29:07.0000] <Hixie>
so is <tt>-based and <font>-based markup

[15:29:08.0000] <Hixie>
and <div>-based markup for that matter

[15:30:00.0000] <Hixie>
(i originally wanted to ban <div> too but sadly people found good use cases for it too)

[15:30:01.0000] <AryehGregor>
If that's really true, authors will move away from it by themselves.  Few have.

[15:30:02.0000] <jgraham>
<em> and <i> seem roughly equally hard

[15:31:00.0000] <jgraham>
Which seems like the problem AryehGregor is concerned with

[15:31:01.0000] <jgraham>
(I know <i> is allowed)

[15:31:02.0000] <AryehGregor>
Provide them with a superior technology, one that they can clearly see is superior, and they'll use it.

[15:31:03.0000] <AryehGregor>
Validators never banned tables, and they disappeared regardless, because they were a pain to use.

[15:31:04.0000] <Hixie>
i don't expect all authors to get the finest points of the phrasing element definitions (though i certainly hope that html5's much clearer definitions will increase the understanding by orders of magnitude over html4's)

[15:32:00.0000] <AryehGregor>
Almost no authors read the spec, so definitions don't help much.

[15:32:01.0000] <AryehGregor>
Even very savvy authors don't read specs.

[15:32:02.0000] <Hixie>
the definitions spread to tutorials

[15:32:03.0000] <Hixie>
furthermore, using them results in more maintainable code than inline formatting, so i think it's a win already

[15:33:00.0000] <jgraham>
AryehGregor: I think you are wrong about <table>

[15:33:01.0000] <Hixie>
i don't really think it's the end of the world if people use <code> instead of <kbd> or <samp>, at least not to the extent that it's a problem if they use <table> instead of <aside>

[15:33:02.0000] <AryehGregor>
The point is that if you tell people they can't use <tt>, they'll port everything blindly to <code>.  They won't move to semantic markup.  You're getting rid of machine-detectable non-semantic markup in favor of non-machine-detectable non-semantic markup.

[15:33:03.0000] <Hixie>
similarly it's not the end of the world if they use <article> instead of <section> or vice versa

[15:34:00.0000] <jgraham>
Also, requiring locally better solutions doesn't work because the problems that are fixed are all exterialities from the author's point of view

[15:34:01.0000] <Hixie>
the finer grained semantics provide tools that the more adventurous authors can use to their advantage

[15:34:02.0000] <AryehGregor>
jgraham, what do you mean?

[15:35:00.0000] <Hixie>
AryehGregor: i don't think it's the end of the world if they use <code> instead of <tt> for a start. When they mature as authors they might start using <samp> where it's more appropriate, e.g. if they decide they want to style something differently (rather than using classes)

[15:35:01.0000] <jgraham>
AryehGregor: The advantage of using <code> in the correct place is typically not for the authors themselves but for some third party e.g. people using screen readers

[15:35:02.0000] <Hixie>
AryehGregor: if i were designing the language from scratch it's quite possible i wouldn't even bother with many of the current phrasing elements, they are quite fine grained and esoteric in some cases

[15:35:03.0000] <AryehGregor>
Hixie, but that elides a useful distinction: <tt> for "I don't have any opinion on the semantics, I just want it monospace" vs. <code> for "I want this monospace because it's code".

[15:36:00.0000] <Hixie>
AryehGregor: *shrug*

[15:36:01.0000] <AryehGregor>
To the extent that the semantics are useful, that's a significant difference.

[15:36:02.0000] <jgraham>
sleep++

[15:36:03.0000] <AryehGregor>
Requiring authors to do this kind of porting will also add to the burden of emitting valid HTML5, which will encourage authors to just ignore validation.

[15:37:00.0000] <AryehGregor>
Hixie, well, okay.  Just remember, I'm pretty sure I'm the first author who actually tried to port a major, heavily-used web app to HTML5.  Consider this early feedback.

[15:37:01.0000] <AryehGregor>
(Valid HTML5, of course I mean, not just <!doctype html>)

[15:38:00.0000] <Hixie>
AryehGregor: this is not a new problem - people have been running into this porting from HTML4 Transitional to Strict for a decade or more now

[15:38:01.0000] <AryehGregor>
They had the option of not porting to Strict.

[15:38:02.0000] <Hixie>
AryehGregor: which everyone was _supposed_ to do years ago

[15:38:03.0000] <AryehGregor>
Which the large majority of web apps and pages took.

[15:38:04.0000] <Hixie>
AryehGregor: yeah, and look at how that went

[15:38:05.0000] <AryehGregor>
MediaWiki was explicitly XHTML 1.0 Transitional because converting to Strict would have been stupid and pointless.

[15:38:06.0000] <Hixie>
it's not stupid and pointless if done right

[15:39:00.0000] <AryehGregor>
I defy you to tell me one single appreciable advantage of rewriting <table cellpadding=""> in wikitext to use some kind of CSS.

[15:40:00.0000] <Hixie>
i thought we weren't talking about wikitext

[15:40:01.0000] <Hixie>
the main advantage is reduced file size

[15:40:02.0000] <Hixie>
assuming we're talking about actual data tables

[15:41:00.0000] <Hixie>
the second advantage is that once you do all the table styling in CSS, you can get _much_ prettier data tables.

[15:41:01.0000] <Hixie>
(the third advantage doesn't apply to wikitext, presumably: when editing a table by hand, it's much easier to maintain if there's no presentational cruft on every table)

[15:42:00.0000] <AryehGregor>
Well, okay.

[15:42:01.0000] <Hixie>
the fourth advantage is that it is less likely that people will use data tables if their tables can't have presentational markup

[15:43:00.0000] <Hixie>
(also presumably not applicable here)

[15:43:01.0000] <Hixie>
the fifth advantage is that it lets you change the padding on a media-by-media basis

[15:43:02.0000] <Hixie>
e.g. to make tables tighter on a mobile device than when printed

[15:43:03.0000] <AryehGregor>
Let's put it this way: MediaWiki's interface will probably become marginally more semantic to conform to HTML5.  The real problem is that we have loads of legacy wikitext that was valid XHTML 1.0 Transitional but isn't valid HTML5.  I guess this is idiosyncratic.

[15:43:04.0000] <AryehGregor>
Probably not that idiosyncratic.

[15:44:00.0000] <AryehGregor>
But you can make a GIGO argument here.

[15:44:01.0000] <Hixie>
sure, if you need to bring a legacy of bad code into the modern day, you'll have a lot of work

[15:44:02.0000] <Hixie>
that's certainly very true

[15:44:03.0000] <Hixie>
for presentational wikitext, it's work that would probably need to happen anyway to make the output more accessible

[15:45:00.0000] <AryehGregor>
An impossible amount, in this case.  There are some tens of millions of pages, at least.

[15:45:01.0000] <Hixie>
yup

[15:45:02.0000] <AryehGregor>
What changes could be made in wikitext that would make the output more accessible, given editors who are clueless about accessibility?

[15:45:03.0000] <Hixie>
make wikitext just be semantic html -- semantic html is accessible by default

[15:46:00.0000] <AryehGregor>
Not if people misuse it.

[15:47:00.0000] <Hixie>
instead of http://en.wikipedia.org/wiki/Wikipedia:Tutorial_(Formatting) saying how to do bold, italics, and bold-italics, have the page say how to mark up the defining instances of terms on the page, how to mark up stress emphasis, how to mark up titles of works, etc

[15:47:01.0000] <Hixie>
people who can get to grips with how to write an encyclopedia can get to grips with how to write basic semantic markup

[15:47:02.0000] <Hixie>
it's not that hard

[15:47:03.0000] <AryehGregor>
We don't control that page.  It was written by users based on what they think makes the best tutorial.

[15:48:00.0000] <Hixie>
yes, but if the markup language had been semantic, they'd think that the best tutorial would be to teach the semantics of the language

[15:48:01.0000] <Hixie>
we see that all the time

[15:49:00.0000] <AryehGregor>
I really think you're ignoring a lot of evidence that most people find semantic markup too much trouble to bother with, even if they actually understand it.  A proper response is to figure out why they think that and try to fix it, not just figure they're all wrong.

[15:49:01.0000] <Hixie>
i'm open to suggestions, but "just let them write inaccessible pages" doesn't cut it

[15:49:02.0000] <Hixie>
why do people prefer to write presentational markup than semantic markup?

[15:49:03.0000] <Hixie>
how do we fix that?

[15:50:00.0000] <hober>
Have your employer de-index w3schools, that might help :/

[15:50:01.0000] <AryehGregor>
For one thing, requests from non-programmer are almost always phrased in presentational terms.

[15:50:02.0000] <AryehGregor>
non-programmers

[15:50:03.0000] <Hixie>
hober: yeah, no kidding.

[15:50:04.0000] <AryehGregor>
The specifications for a web page are normally totally presentational.

[15:50:05.0000] <Hixie>
AryehGregor: indeed, but why? how do we fix that?

[15:50:06.0000] <AryehGregor>
Designers decide how they want it to look, then developers are asked to make it look that way.

[15:51:00.0000] <AryehGregor>
Because the overwhelming majority of (perceived?) users use web browsers that ignore the semantics of the documents.

[15:51:01.0000] <AryehGregor>
Normal web browsers don't distinguish between <tt> and <code>, so normal users don't care, so authors don't care.

[15:51:02.0000] <Hixie>
how do we teach to designers that a web page will be rendered on dozens of devices and could have a visual presentation, a braille presentation, a voice presentation, etc?

[15:51:03.0000] <Hixie>
with different device capabilities, etc?

[15:52:00.0000] <AryehGregor>
You can't get them to care about screen readers if they have a negligible number of blind users.

[15:52:01.0000] <AryehGregor>
You might be able to get them to care about, say, cell phones.  Maybe.

[15:52:02.0000] <Hixie>
so what, screw the blind?

[15:52:03.0000] <zcorpan_>
make dreamweaver have multiple views of the document

[15:52:04.0000] <zcorpan_>
i think they've done that now

[15:52:05.0000] <AryehGregor>
No, just acknowledge that trying to mandate things that most authors won't listen to doesn't help the blind.  Authoring requirements have to reflect reality just as much as implementer requirements.

[15:53:00.0000] <Hixie>
i strongly disagree with that later statement

[15:53:01.0000] <Hixie>
i think in the case of authoring requirements, conformance requirements have to _lead_ reality

[15:53:02.0000] <AryehGregor>
Why the difference?

[15:54:00.0000] <Hixie>
because authoring requirements don't affect interop

[15:54:01.0000] <Hixie>
they affect the success with which an author can convey meaning

[15:54:02.0000] <Hixie>
whereas UA requirements are intended to result in interop

[15:54:03.0000] <Hixie>
it's a huge assymetry

[15:55:00.0000] <Hixie>
it's the same reason why UA requirements have to cover every eventuality of author input, but author requirements don't have to cover every eventuality of UA output

[15:56:00.0000] <AryehGregor>
Reasonable.

[15:57:00.0000] <zcorpan_>
authors have to deal with actual UA output

[15:57:01.0000] <AryehGregor>
Nevertheless, you cannot preempt cost-benefit analyses by authors.  If a business is making a website to sell a product that has effectively no blind users, then you are not going to be able to persuade them to expend nontrivial resources on making things accessible to blind people.

[15:57:02.0000] <AryehGregor>
Not unless you have more power over authors than standards bodies actually do in practice.

[15:58:00.0000] <AryehGregor>
The best you can do is hope people want their pages to validate, but that power becomes weaker with every additional requirement, because stricter requirements mean fewer people bother to comply.

[15:58:01.0000] <AryehGregor>
Ideally, HTML would be crafted so that it's really accessible with no effort involved beyond what it takes to get the page working right in a standard visual browser.

[15:58:02.0000] <AryehGregor>
Unfortunately, that's not possible.

[15:59:00.0000] <Hixie>
that's certainly true. It's why with HTML5 we've made a significant effort to remove requirements that are pointless (e.g. "ids can't start with a number") and make the actual requirements all have good reasons (like "doing this will disenfranchise the blind")

[15:59:01.0000] <Hixie>
("that" being the thing about power becoming weaker with each requirement)

[15:59:02.0000] <Hixie>
i think with HTML5 we actually get pretty close to what you describe

[15:59:03.0000] <AryehGregor>
But the blind will be disenfranchised anyway.  If you require alt text, authors will put empty alt text or lousy alt text.

[15:59:04.0000] <Hixie>
but obviously there are limits to what we can do

[15:59:05.0000] <AryehGregor>
Unless they care about the blind to begin with.

[16:00:00.0000] <AryehGregor>
Wikipedia didn't allow alt text on images until like last year, when I added the feature for basically idealistic reasons.  I don't recall ever seeing a single complaint.

[16:00:01.0000] <AryehGregor>
If authors don't get complaints from actual users, they won't fix things.

[16:01:00.0000] <Hixie>
wikipedia images almost always just repeat what is already said on the page

[16:01:01.0000] <AryehGregor>
It's way easier to just trivially fix all the validator errors, than to actually test in a screen reader (which is what you really need to do to get a good website for the blind).

[16:01:02.0000] <Hixie>
so it's not a big deal for wikipedia images to not have alt text in many cases

[16:02:00.0000] <Hixie>
AryehGregor: for authors who don't care about making content accessible, the alternatives i see are them using purely presentational markup, or them using non-presentational markup that approximates their semantics by virtue of defaulting to relatively close to the presentation they want

[16:03:00.0000] <Hixie>
AryehGregor: neither is ideal, but the second option is mildly better.

[16:03:01.0000] <Hixie>
AryehGregor: and had the benefit that it might make them look things up. maybe.

[16:03:02.0000] <AryehGregor>
Shrug.

[16:03:03.0000] <AryehGregor>
I guess we'll see what happens.

[16:03:04.0000] <Hixie>
AryehGregor: for authors who _do_ care about making content accessible, but simply don't know that it's an option, the validator complaining about presentational markup will lead them to find out how to do it.

[16:04:00.0000] <AryehGregor>
Except when all the tutorials switch to using semantic-as-presentational markup instead of actually presentational markup.

[16:04:01.0000] <Hixie>
some might

[16:04:02.0000] <Hixie>
some won't

[16:06:00.0000] <AryehGregor>
I'd probably be more inclined to agree with you if I weren't faced with the fact that most Wikipedia pages will never validate again once we switch to HTML5.  I'll just have to view that as GIGO, I guess.

[16:06:01.0000] <AryehGregor>
Maybe add an option to not allow the newly-invalid markup.

[16:07:00.0000] <Hixie>
there's definitely a dichotomy between newly created pages and legacy pages that are going to be ported

[16:07:01.0000] <Hixie>
i wish i had a good solution to that problem

[16:07:02.0000] <Hixie>
it's one we've fought with a lot

[16:07:03.0000] <Hixie>
most of the old "downplayed errors" and the current "obsolete but conforming" stuff are related to that problem

[16:08:00.0000] <AryehGregor>
If mechanical translation were used, do you think it would be better to do <tt> -> <span style="font-family:monospace"> or <tt> -> <code>?

[16:08:01.0000] <Hixie>
we need different validator behaviour for different pages, but without running into the HTML4 Transitional problem (where everyone writes new pages that claim to be legacy pages)

[16:09:00.0000] <AryehGregor>
Wikitext is really hard to parse, though, so we can't feasibly do mechanical translation of any kind on the source wikitext.

[16:09:01.0000] <Hixie>
because we don't have a good solution, i've tended to favour the needs of new pages, because there will be more new pages in the future than there are pages that will be migrated

[16:09:02.0000] <AryehGregor>
We could force everyone to switch by shutting off support for old elements/attributes, but that would break . . . everything.

[16:09:03.0000] <AryehGregor>
It's a headache.

[16:09:04.0000] <Hixie>
AryehGregor: i don't think you can successfully do a mechanical translation of presentational markup to semantic markup today without an intimate knowledge of the source document's typographic conventions

[16:10:00.0000] <AryehGregor>
If you could do reliable automated translation from presentational to semantic markup, semantic markup would not be needed.

[16:14:00.0000] <Dashiva>
Aren't most wikipedia images actually <figure> form?

[16:14:01.0000] <AryehGregor>
I guess your decision is reasonable, but it's a huge pain in the neck for me.

[16:14:02.0000] <AryehGregor>
Dashiva, yes, but we can't use that until IE8 is dead.  Don't want to rely on JS hacks.

[16:15:00.0000] <Dashiva>
You don't have to use the element, but you can use its aria mappings

[16:15:01.0000] <AryehGregor>
Interesting idea.

[16:15:02.0000] <Hixie>
AryehGregor: i completely agree that it's a huge pain in the neck for anyone with legacy content

[16:15:03.0000] <Hixie>
AryehGregor: i have no good solution for that sadly

[16:20:00.0000] <zcorpan_>
http://forums.whatwg.org/viewtopic.php?t=4315 - feedback on rel=prefetch

[16:26:00.0000] <zcorpan_>
"BUT Is it CORRECT to put a NAV tag within a HEADER TAG? I have read in the specs that you can put non H or HGROUP tags within a header but it refers to those other tags as "metadata"... That makes me nervous." - http://www.sitepoint.com/forums/showthread.php?p=4613614#post4613614

[16:46:00.0000] <Hixie>
jesus, i just got 6 copies of the BoF announcement -- and that's after deduping due to corss-posting

[16:48:00.0000] <abarth>
i think they're worried about it being last-minute

[16:51:00.0000] <Hixie>
face to face meetings like that are such a waste of time

[16:51:01.0000] <Hixie>
/me mumbles

[16:51:02.0000] <AryehGregor>
It's nice to meet people sometimes.

[16:51:03.0000] <AryehGregor>
Even if not much gets done.

[16:53:00.0000] <Hixie>
i'm all for people getting together for fun

[16:53:01.0000] <Hixie>
that's not what these things are


2010-06-10
[17:01:00.0000] <cardona507>
I had a good time at the TPAC - granted the most fun was the night we all went out for mexican food

[17:02:00.0000] <Hixie>
i'm not arguing that they're not fun, especially the first few

[17:02:01.0000] <Hixie>
i'm arguing that they're a waste of time

[17:02:02.0000] <Hixie>
i do think we should have some get togethers more often

[17:02:03.0000] <Hixie>
it's hard with us spread around so much though

[17:08:00.0000] <cardona507>
I was watching the documentary Code Rush - about Mozilla and noticed that they all have a celebration dinner at the same mexican restaurant that we all ate at. http://www.youtube.com/watch?v=u404SLJj7ig - 

[17:11:00.0000] <Hixie>
yup

[17:11:01.0000] <Hixie>
la fiesta

[17:12:00.0000] <Hixie>
i first learnt of la fiesta while interning at netscape

[17:12:01.0000] <cardona507>
:)

[17:12:02.0000] <Hixie>
we went there all the time

[17:12:03.0000] <cardona507>
wow - you were at netscape - interesting

[17:12:04.0000] <cardona507>
bet that was a wild ride

[17:19:00.0000] <Hixie>
cardona507: i was there for a year as an intern, basically during the netscape 6.0 release

[17:19:01.0000] <Hixie>
summer 2000 to summer 2001

[17:19:02.0000] <Hixie>
just before the bubble burst

[17:19:03.0000] <Hixie>
or as the bubble was bursting

[17:19:04.0000] <Hixie>
depending on your outlook

[17:19:05.0000] <cardona507>
any obvious similarities between then and now?

[17:21:00.0000] <Hixie>
hard to say

[17:23:00.0000] <Hixie>
the french have a saying, le plus ca change, le plus c'est la meme chose (modulo my not having entered appropriate accents)

[17:23:01.0000] <Hixie>
which applies here

[17:23:02.0000] <Hixie>
the fundamentals are all pretty much the same

[17:23:03.0000] <Hixie>
even the players are pretty much the same

[17:23:04.0000] <Hixie>
though with some new faces, and some old ones gone

[17:24:00.0000] <cardona507>
well said

[17:25:00.0000] <TabAtkins_>
Hixie: The english speaking world has that saying too.  (I dunno where it originated, but it's common now.)

[17:25:01.0000] <Hixie>
comes from Jean-Baptiste Alphonse Karr

[17:26:00.0000] <Hixie>
ok bbl

[17:26:01.0000] <cardona507>
HTML5 it - http://www.fastcompany.com/1658276/html5-the-next-buzzword-the-next-battle?partner=rss&utm_source=feedburner&utm_medium=feed&utm_campaign=Feed:+fastcompany/headlines+(Fast+Company+Headlines)

[17:26:02.0000] <cardona507>
ugg - what an ugly url

[17:28:00.0000] <Dashiva>
You can leave out the stuff after the ?

[17:29:00.0000] <cardona507>
yes - that is better

[17:31:00.0000] <Dashiva>
Evil google making our URLs ugly :P

[17:42:00.0000] <til>
this is slightly OT, but does anyone know where apple was going to publish the FaceTime spec yesterday?

[17:42:01.0000] <til>
there are so many standards bodies.  i haven't seen anything posted on IETF

[18:40:00.0000] <paul_irish>
anyone know if a css spec indicates i can legally mix hex notation with rgba or not?

[18:47:00.0000] <Dashiva>
paul_irish: Mix how? In the same stylesheet, or the same value?

[18:49:00.0000] <paul_irish>
well my usecase here is a linear gradient..

[18:50:00.0000] <paul_irish>
i can't think of any spec'd things that accept multiple colors like gradients do...

[18:51:00.0000] <paul_irish>
the problem is webkit doesnt allow a mix of hex and rgba values in a linear gradient definition, and i wanted to see if a spec indicates it should.

[18:53:00.0000] <Dashiva>
I guess that would depend on the specification for the property

[18:53:01.0000] <Dashiva>
E.g. https://developer.mozilla.org/en/CSS/-moz-linear-gradient doesn't say anything about the different <color> instances having to use the same syntax, so it should work in firefox

[19:50:00.0000] <TabAtkins>
paul_irish: A <color> is a <color> is a <color>.  It doesn't matter what notation you use, even for multiple occurences in one "value" like a linear-gradient() function.

[19:50:01.0000] <TabAtkins>
They all just get passed to the color parser and turned into a 4-byte number anyway.

[19:53:00.0000] <TabAtkins>
If webkit seriously doesn't allow mixed notations in color-stops in their -webkit-gradient() notation, that's extremely odd.  It would be substantially more work to disallow it than to allow it.

[20:30:00.0000] <paul_irish>
TabAtkins: i am now having trouble reproducing the bug i found. if i can, i'll file it. thx

[21:14:00.0000] <annevk>
oh, only 5:20

[21:15:00.0000] <annevk>
D66 FTW

[21:19:00.0000] <annevk>
aah, we went from 11 to 10 again

[21:19:01.0000] <annevk>
lame

[23:10:00.0000] <kpx>
Hey I am making a scrapbook plugin for google chrome and nearing completion. I am saving the pages as a  innerHTML of the 'HTML' element in a WEBDATABASE in HTML5. At rendering time I am just replacing the innerHTML of a blank page. This is obviously not sufficient as the style etc are not applied properly and page is distorted.... The question is more how can i retrieve the style of the...

[23:10:01.0000] <kpx>
...page so I may store it onto the database?? .... 

[23:10:02.0000] <kpx>
and then apply it back

[23:27:00.0000] <kpx>
Never mind I think I got it....

[00:31:00.0000] <kpx>
Hey does someone know a pure JS function to convert Images to base64 so I can store it as text?

[00:34:00.0000] <slartsa>
hm, I can't trust w3schools anymore, found a typo from the site :)

[00:49:00.0000] <hsivonen>
Are the Adobe heart images at http://www.fastcompany.com/1658276/html5-the-next-buzzword-the-next-battle?partner=rss&utm_source=feedburner&utm_medium=feed&utm_campaign=Feed:+fastcompany/headlines+%28Fast+Company+Headlines actual ads?

[02:07:00.0000] <jgraham>
Wow, the name of the polyglot document is being bikeshedded

[02:30:00.0000] <abarth>
jgraham: shocking

[02:33:00.0000] <MikeSmith>
hey, it's abarth

[02:33:01.0000] <MikeSmith>
abarth: I was suggesting to wirepair yesterday that he should ping you

[02:33:02.0000] <abarth>
hi MikeSmith 

[02:33:03.0000] <MikeSmith>
hey man

[02:33:04.0000] <abarth>
who is wirepair?

[02:34:00.0000] <MikeSmith>
he's interested in helping improved Web-application/browser security

[02:34:01.0000] <abarth>
ah, awesome

[02:34:02.0000] <MikeSmith>
and was asking about where he could find test cases

[02:35:00.0000] <abarth>
tests cases for what?

[02:35:01.0000] <abarth>
http://trac.webkit.org/browser/trunk/LayoutTests/http/tests/security

[02:35:02.0000] <abarth>
has a ton of security test cases

[02:35:03.0000] <MikeSmith>
yeah, like that

[02:35:04.0000] <MikeSmith>
I don't know if he had found that yet or not

[02:35:05.0000] <MikeSmith>
anyway, I just wanted to give you a heads-up that he may be pinging you

[02:36:00.0000] <abarth>
okiedokes

[02:36:01.0000] <abarth>
/me goes to bed

[02:46:00.0000] <wirepair>
doh just missed him

[02:49:00.0000] <Dashiva>
"1 in 4 has a vision difficulty (25%)" makes me think they include people with regular glasses

[04:42:00.0000] <zcorpan_>
http://twitter.com/jaffathecake/status/15839704934

[04:44:00.0000] <volkmar>
the specs say input.value, on setting should throw INVALID_STATE_ERR if the type is 'file' but SECURITY_ERR isn't more appropriate ?

[05:44:00.0000] <jgraham>
Is it just me that is having problems with recent firefox trunk on lucid?

[05:44:01.0000] <jgraham>
Like it won't start at all

[05:45:00.0000] <micheil>
jgraham: got that websocket-server working.

[05:46:00.0000] <jgraham>
micheil: Nice

[05:46:01.0000] <micheil>
yeah. parses something like 100,000 packets / ~1.2

[05:46:02.0000] <jgraham>
/me wonders if WebM is to blame or if this failure predates that landing

[05:47:00.0000] <kbrosnan>
webm was yesterday morning

[05:48:00.0000] <jgraham>
Hmm, I think it stopped loading yesterday, but I hasn't updated for a few days

[05:48:01.0000] <jgraham>
*hadn't

[05:49:00.0000] <kbrosnan>
starts fine for me, ppa or build from mozilla?

[05:49:01.0000] <jgraham>
ppa

[05:49:02.0000] <kbrosnan>
/me is using the latter

[05:50:00.0000] <kbrosnan>
try the usual, safe mode, rename .mozilla?

[05:51:00.0000] <jgraham>
Yeah, will do

[07:12:00.0000] <henrikbjorn>
hey guys. a little question. If i have a widget in my sidebar that has a h1 should it then be a <article> or <section>

[07:13:00.0000] <zcorpan_>
henrikbjorn: is the widget 'standalone'?

[07:13:01.0000] <henrikbjorn>
its a ul with links to discussions

[07:14:00.0000] <henrikbjorn>
which dosent really have anything to do with the content areas content

[07:14:01.0000] <zcorpan_>
sounds like an <aside> then :)

[07:14:02.0000] <henrikbjorn>
okay :) another stupid question then. What if it is a box width text like "Help us out with feedback click here" aside aswell ?

[07:16:00.0000] <zcorpan_>
sure. or you could have an <aside> that wraps all such boxes

[07:16:01.0000] <boblet>
oh oh stupid question time? perfect! :)

[07:16:02.0000] <boblet>
one from me; the reason for the change from block/inline to flow/phrasing was the confusion once CSS display values are added to the mix, right?

[07:16:03.0000] <zcorpan_>
boblet: yes

[07:17:00.0000] <boblet>
henrikbjorn: I like to think of <aside> as something that’s not essential to the main content, i.e. something a mobile browser could choose to not download to save bandwidth

[07:18:00.0000] <boblet>
zcorpan_: thanks :) thought I remember reading that somewhere but wanted to check

[07:18:01.0000] <boblet>
henrikbjorn: you could also use Safari Reader or Instapaper as a mental model

[07:18:02.0000] <henrikbjorn>
zcorpan_: yeah but the question is what the markup for the box it self should be not the sidebar :)

[07:19:00.0000] <henrikbjorn>
like iphone specific sites

[07:19:01.0000] <zcorpan_>
henrikbjorn: i'd say <section>

[07:20:00.0000] <henrikbjorn>
i guess a box like this falls under "The section element represents a generic document or application section…"

[07:21:00.0000] <henrikbjorn>
thanks guys :D

[07:22:00.0000] <boblet>
henrikbjorn: if you need more feedback a code sample is always easy to advise on

[07:23:00.0000] <boblet>
(not required of course :) )

[07:23:01.0000] <henrikbjorn>
yeah was just initial thoughts before starting coding. Im sure ill come up with a few questions incl. code samples later on :)

[07:24:00.0000] <zcorpan_>
/me likes the fact that http://simon.html5.org/sandbox/html/w3c-home-in-html5 now renders correctly in opera, firefox and webkit

[07:24:01.0000] <zcorpan_>
when i wrote that page, it only rendered correctly in opera

[07:24:02.0000] <zcorpan_>
wonder if ie9 gets it right

[07:31:00.0000] <zcorpan_>
oh, html5test.com finally updated

[07:32:00.0000] <hsivonen>
hmm. SVG and MathML parsing count as bonus points...

[07:32:01.0000] <hsivonen>
support for parsing is actually required

[07:32:02.0000] <hsivonen>
though it's probably better if a browser doesn't support parsing if it doesn't support rendering

[07:38:00.0000] <zcorpan_>
chrome doesn't support WAVE?

[07:48:00.0000] <Lachy>
zcorpan_, no, it never has

[07:49:00.0000] <Lachy>
only aac and vorbis, I believe

[07:53:00.0000] <zcorpan_>
wonder why they don't support wave

[07:57:00.0000] <hsivonen>
so there isn't even PCM WAVE as cross-browser audio baseline :-(

[08:01:00.0000] <zcorpan_>
http://code.google.com/p/chromium/issues/detail?id=23916

[08:01:01.0000] <hsivonen>
I can no longer find any mention of PCM in the spec

[08:01:02.0000] <hsivonen>
I thought PCM WAVE used to be a required feature

[08:01:03.0000] <hsivonen>
did it get vendor vetoed?

[08:03:00.0000] <zcorpan_>
hixie removed it at the same time as the codec big-issue paragraph was removed, citing the fact that all browsers were going to support wave anyway

[08:03:01.0000] <hsivonen>
right...

[08:19:00.0000] <jgraham>
Hmm, doesn't the W3C testsuite license grant a strict subset of the rights granted by the 3 clause BSD license?

[08:20:00.0000] <jgraham>
Oh it's probably about whether you get to mention the W3C or not

[08:20:01.0000] <jgraham>
In one case you have to, in the other you can't

[08:20:02.0000] <hsivonen>
It's relatively hard to grant a superset of the right granted by the 3-clause BSD license without going into trademark territory

[08:20:03.0000] <hsivonen>
*rights

[08:22:00.0000] <jgraham>
Yeah, I think the trademark issue is the only difference

[08:23:00.0000] <jgraham>
That is, if you want to mention the W3C you must use the things under the W3C testsuite license which doesn't allow modifications

[08:23:01.0000] <Dashiva>
"Here is a testsuite completely unrelated to w3c"

[08:25:00.0000] <nessy>
google chrome doesn't support PCM WAVE

[08:45:00.0000] <annevk>
what is the point again of having two licenses for a test? http://www.w3.org/html/wg/wiki/Testing/Submission/

[08:45:01.0000] <annevk>
one being less permissive than the other

[08:45:02.0000] <hsivonen>
annevk: the less permissive one is more Official

[08:46:00.0000] <annevk>
so basically for people that would love to use it under that license so they are forced to meet the extra requirements?

[08:47:00.0000] <hsivonen>
for people who want to pass the Official Test Suite, I guess

[08:47:01.0000] <zcorpan_>
a whole lotta yellow in http://krijnhoetmer.nl/irc-logs/whatwg/20100609

[08:50:00.0000] <annevk>
maybe because it was election day

[08:57:00.0000] <hsivonen>
http://krijnhoetmer.nl/irc-logs/whatwg/20100609#l-1225 I would have thought Googlebot had CSS support already

[09:01:00.0000] <TabAtkins>
Googlebot understands enough CSS to tell in some circumstances that you're trying to be sneaky with SEO things.  That's about it.

[09:02:00.0000] <zcorpan_>
TabAtkins: but it doesn't fetch external style sheets, does it?

[09:02:01.0000] <hsivonen>
/me wonders if Googlebot support all the newest Selectors

[09:33:00.0000] <TabAtkins>
AryehGregor: Is kosher a proper superset of vegetarian?

[09:37:00.0000] <TabAtkins>
I know that it's not a superset of fish-eating vegetarianism, because of dairy+fish.

[10:46:00.0000] <Dashiva>
Who is going to buy all the bikes we'll need to justify these bikesheds?

[10:48:00.0000] <Philip`>
Dashiva: You're missing the point of bikesheds, if you want to put bikes in them

[10:48:01.0000] <Philip`>
The point of bikesheds is the bikeshed

[10:52:00.0000] <henrikbjorn>
Hey. How would you guys markup http://drp.ly/1as46T in html5

[10:53:00.0000] <MikeSmith>
Philip`: I object to the way you've chosen to make your point about the point about bikesheds.

[10:54:00.0000] <TabAtkins>
<div><ul><li><a href="#forum">Forum</a></li><li><a href="popular">Popular</a></li><li><a href="#nyeste">Nyeste</a></li></ul> <ul id=forum><li>...</li>...</ul> <ul id=popular>...</div>

[10:54:01.0000] <henrikbjorn>
but wuldnt a dl be a better option. And no section / article ? :)

[10:55:00.0000] <TabAtkins>
A <dl> is a possibilty, but styling it would be harder.

[10:55:01.0000] <TabAtkins>
And no, no need for sectioning elements in particular here.

[10:55:02.0000] <henrikbjorn>
even tho its meant for a sidebar ?

[10:56:00.0000] <TabAtkins>
In that case, probably replace the <div> with an <aside>.

[10:56:01.0000] <henrikbjorn>
okay :)

[10:56:02.0000] <Philip`>
TabAtkins: Wouldn't something like <h1>Forum</h1> (forum links) <h1>Populaere</h1> ... be more appropriate?

[10:56:03.0000] <Philip`>
since those heading things look like heading things to me

[10:56:04.0000] <TabAtkins>
Still too hard to style.

[10:57:00.0000] <TabAtkins>
Some consideration must be given to visual effects, at least until I get some better layout modules done in CSS.

[10:57:01.0000] <henrikbjorn>
the styling dosent matter :) more the markup im interested in.. my first atemp was a aside with a header and a section which hold the ul

[11:21:00.0000] <variable>
I just added textarea to the rationale page on the wiki: http://wiki.whatwg.org/wiki/Rationale#textarea --> any comments?

[11:58:00.0000] <AryehGregor>
TabAtkins, dairy+fish is fine, only dairy+meat/poultry is a problem.  Kosher food isn't quite a subset of vegan food; although all the ingredients in vegan food are kosher, there are some restrictions on the preparation and handling of food that could be a problem even if none of the ingredients are inherently non-kosher.

[11:59:00.0000] <AryehGregor>
For instance, milk is a problem according to some if not produced under Jewish supervision.  Grape juice/wine can be a problem.  Some cooked food is a problem.

[11:59:01.0000] <AryehGregor>
Something like raw fruit or vegetables should always be fine, at least if you ignore the problem of bugs.

[12:00:00.0000] <TabAtkins>
"Problem"?

[12:00:01.0000] <AryehGregor>
I mean, it can be non-kosher.

[12:00:02.0000] <TabAtkins>
Bugs can be non-kosher?

[12:00:03.0000] <AryehGregor>
All bugs are non-kosher, except theoretically some species of locust.

[12:01:00.0000] <TabAtkins>
Ah, kk.

[12:01:01.0000] <AryehGregor>
The only kosher animals are beasts with cloven hooves that chew their cud, fish with fins and scales, and most birds that people actually eat.

[12:02:00.0000] <TabAtkins>
Random awesome: http://pterosaurs.files.wordpress.com/2008/05/more_azhdarchids_with_clouds1.jpg

[12:04:00.0000] <AryehGregor>
/me was about to question the accuracy of that image, but then looked up "azhdarchid" on Wikipedia and found the same image illustrating the article

[12:04:01.0000] <TabAtkins>
Haha.

[12:04:02.0000] <AryehGregor>
Apparently that's meant to be a "juvenile titanosaur" there.

[12:04:03.0000] <TabAtkins>
One particular family of pterosaurs is, apparently, confectured to be a plains walker.

[12:05:00.0000] <AryehGregor>
I was suspicious of the scale, because I assumed that thing was like a full-grown brontosaurus or such.

[12:05:01.0000] <TabAtkins>
It could still fly, but would often just walk around like a terrifying ostrich-analogue.

[12:05:02.0000] <TabAtkins>
Oh, no.  Just tiny food animal.

[13:21:00.0000] <othermaciej>
wow, there has been a lot of public-html mail the past two weeks

[13:21:01.0000] <cardona507>
thats for sure

[13:23:00.0000] <annevk>
hey, wb maciej

[13:23:01.0000] <TabAtkins>
 yo, maciej.

[13:24:00.0000] <TabAtkins>
/me is temporarily binning public-html for a month or two.

[13:26:00.0000] <gsnedders>
/me will eventually catch up on public-html

[13:26:01.0000] <gsnedders>
/me is around a month behind

[13:45:00.0000] <cardona507>
installing windows XP onto my mac for some IE testing - I feel dirty

[13:50:00.0000] <Philip`>
othermaciej: Fortunately approximately none of it was technical discussion, so it can be safely ignored by almost everyone (except by e.g. chairs)

[13:51:00.0000] <jgraham>
gsnedders: Don't bother

[13:52:00.0000] <gsnedders>
jgraham: But what else will I do next Monday!?

[13:53:00.0000] <Philip`>
The same thing you do every Monday - try to take over the world

[13:55:00.0000] <gsnedders>
Shh, don't mention that.

[13:58:00.0000] <jgraham>
gsnedders: is not actually pinky

[13:58:01.0000] <jgraham>
or indeed the brain

[14:54:00.0000] <AryehGregor>
Does anyone know roughly what percentage of Mozilla development (for instance) is done by employees vs. volunteers?

[14:56:00.0000] <jgraham>
AryehGregor: How are you measuing "development"?

[14:56:01.0000] <AryehGregor>
I don't know, anything handy.

[14:56:02.0000] <jgraham>
(the short anser is "no" btw)

[14:56:03.0000] <AryehGregor>
Heh.

[14:56:04.0000] <AryehGregor>
Number of commits would be a good start.

[14:57:00.0000] <jcranmer>
l10n commits skew that statistic a fair amount

[14:57:01.0000] <jgraham>
You probably need to weight by complexity of change somehow

[14:58:00.0000] <jgraham>
and remember that employees are more likely to have commit access

[14:58:01.0000] <AryehGregor>
It says who wrote the patch, though, yes?

[14:58:02.0000] <jgraham>
(does Mercurial allow seperate author/committer like git?)

[14:58:03.0000] <AryehGregor>
Pretty sure, yes.

[14:58:04.0000] <AryehGregor>
Or something like that.

[14:59:00.0000] <jgraham>
/me wonders how many of JF's disability categories he fits into

[15:02:00.0000] <jgraham>
AryehGregor: The longer version of "no" is "I guess it is mostly employees these days"

[15:02:01.0000] <AryehGregor>
This has my name: http://hg.mozilla.org/mozilla-central/rev/511a92d06acf

[15:15:00.0000] <AryehGregor>
/me finds a google.com page on https with a little skull and crossbones in Chrome, because of insecure resources

[15:15:01.0000] <AryehGregor>
I didn't even try!

[15:16:00.0000] <AryehGregor>
Yay for HTTPS being broken.

[15:17:00.0000] <AryehGregor>
I also like how the Google Checkout credit card number field doesn't have autocomplete disabled.

[15:38:00.0000] <cardona507>
can we postpone the dev meeting on next wed until 11am so that I can attend this? http://pages.engineyard.com/Rails3-MobileAppswithYehuda-62010.html

[15:42:00.0000] <Philip`>
cardona507: Wrong channel?

[15:44:00.0000] <cardona507>
haha - oops my bad :)


2010-06-11
[17:33:00.0000] <paul_irish>
i was interested in what specs were broken out of the html5 spec and when.. got any leads for where i should look?

[17:33:01.0000] <TabAtkins>
The first section of the whatwg spec should detail that.

[18:30:00.0000] <othermaciej>
paul_irish: http://wiki.whatwg.org/wiki/FAQ#What_are_the_various_versions_of_the_spec.3F

[18:38:00.0000] <paul_irish>
thx tab and maciej

[21:32:00.0000] <MikeSmith>
http://www.guardian.co.uk/media/pda/2010/jun/10/antiweb

[21:33:00.0000] <MikeSmith>
breaking news: journalist finds evidence of supposedly growing "anti-Web" movement

[23:36:00.0000] <hsivonen>
what doess paidContent:UK mean on the anti-Web article?  who pays for what?

[00:20:00.0000] <hsivonen>
/me looks for the base URL processing rules again

[00:20:01.0000] <hsivonen>
/me fails to find them in the obvious place

[00:21:00.0000] <hsivonen>
found it

[00:31:00.0000] <hsivonen>
paul_irish: http://html5readiness.com/ needs updates for Firefox and Safari--to the version number at least

[00:54:00.0000] <hsivonen>
<basefont> creates a crazy DOM in IE...

[01:10:00.0000] <paul_irish>
hsivonen: true. ill try to hit that up next week

[01:12:00.0000] <hsivonen>
paul_irish: nice

[01:13:00.0000] <hsivonen>
looks like html5lib's test coverage for basefont and bgsound sucks

[01:14:00.0000] <paul_irish>
url?

[01:21:00.0000] <hsivonen>
paul_irish: http://code.google.com/p/html5lib/source/browse/#hg/testdata/ There basically are no interesting tests about <basefont> or <bgsound>

[01:22:00.0000] <hsivonen>
that is, even if I change the parser notably, no tests starts failing

[01:34:00.0000] <jgraham>
Do basefont or bgsound do anything unusually magic?

[01:34:01.0000] <jgraham>
/me can't remember and loading the spec is tedius

[01:34:02.0000] <jgraham>
*tedious

[01:35:00.0000] <Hixie>
you know even complete has a multipage version now :-P

[01:36:00.0000] <jgraham>
Well I have a very low tedium thereshold right now :)

[01:39:00.0000] <jgraham>
Ah, OK, nothing too special

[01:40:00.0000] <jgraham>
hsivonen: It seems like it should be possible to have tests for every tag that has special behaviour in each mode in which it is special

[01:40:01.0000] <jgraham>
but I can't work out how to generate the expected output sanely

[01:41:00.0000] <hsivonen>
jgraham: they are now sensitive to reconstructing the list of formatting elements and to flipping frameset-ok

[01:41:01.0000] <jgraham>
hsivonen: Yeah

[01:41:02.0000] <hsivonen>
jgraham: I was just about to accidentally break both those behaviors

[01:42:00.0000] <jgraham>
(generating the expected input is also not trivial because you need to know that you should have formatting tags before, or frameset after, the element, in order to test

[01:42:01.0000] <jgraham>
)

[01:42:02.0000] <jgraham>
s/expected/interesting/

[01:44:00.0000] <hsivonen>
/me hasn't written framesets in years and needs to look up documentation

[01:45:00.0000] <hsivonen>
w3schools to rescue!

[01:48:00.0000] <hsivonen>
Hixie: Based on IE behavior, I believe the spec is wrong to set frameset-ok to 'not ok' on <basefont> and <bgsound>

[01:48:01.0000] <Hixie>
k

[01:48:02.0000] <hsivonen>
ok. now I can 'break' stuff intentionally instead of breaking stuff accidentally

[01:51:00.0000] <hsivonen>
I also think that given the function of <basefont> (none) and <bgsound> (not visual), I think it's useless to reconstruct the active formatting elements for them in 'in body'

[01:51:01.0000] <hsivonen>
s/I think//

[01:51:02.0000] <Hixie>
no argument there

[01:54:00.0000] <jgraham>
So we could just remove all the special rules?

[01:54:01.0000] <jgraham>
Sounds good

[01:59:00.0000] <hsivonen>
aargh. I still forgot to test <head><noscript><bgsound> with scripting turned off

[02:00:00.0000] <jgraham>
wiki down?

[02:00:01.0000] <jgraham>
Oh, no it was just being very very slow

[04:00:00.0000] <hsivonen>
http://twitter.com/mattmay/status/15798681102

[04:02:00.0000] <othermaciej>
who has raised a Formal Objection?

[04:03:00.0000] <othermaciej>
/me is not fully caught up on email

[04:12:00.0000] <Lachy>
no-one has raised an FO yet. I don't know where Matt has gotten his information from.

[04:13:00.0000] <othermaciej>
did anyone (other than Shelley) indicate an intent to do so?

[04:13:01.0000] <othermaciej>
/me is rather puzzled

[04:17:00.0000] <Hixie>
looks like he tweeted that just a few hours after http://lists.w3.org/Archives/Public/public-html/2010Jun/0238.html

[04:27:00.0000] <zcorpan_>
http://www.w3.org/mid/OF4C735B56.D6A802E0-ON8625773E.006210C1-8625773E.006248A3⊙uic - is this how to make emails accessible?

[04:27:01.0000] <Peter`>
it's clear..

[08:02:00.0000] <annevk>
it's quite amazing how often Julian can make the same point

[08:02:01.0000] <annevk>
even more amazing that I still read it all

[08:45:00.0000] <zcorpan_>
hsivonen: <base> was parsed the same as <basefont> in ie6

[08:47:00.0000] <zcorpan_>
hsivonen: should basefont and bgsound break out of noscript in head?

[08:49:00.0000] <hsivonen>
zcorpan_: I didn't research noscript interaction yet. I got distracted by other stuff.

[08:50:00.0000] <hsivonen>
zcorpan_: but it's hard for me to believe that they'd need to break out of <noscript> in head for Web compat

[08:51:00.0000] <zcorpan_>
hsivonen: indeed, however breaking out is the default behavior, isn't it?

[08:51:01.0000] <zcorpan_>
and you can't tell the difference without roundtripping. and basefont and bgsound are invalid anyway

[08:56:00.0000] <hsivonen>
zcorpan_: my suggestion is to make bgsound and basefont parse just like <link>

[08:56:01.0000] <zcorpan_>
ok

[10:22:00.0000] <Philip`>
Did the behaviour of public-html for messages from non-subscribers change some time recentishly?

[10:22:01.0000] <Philip`>
It used to delay the first message (for manual moderation?) but then accept subsequent ones immediately

[10:23:00.0000] <Philip`>
but now it sends me a message saying my message was forwarded to the maintainer for manual processing

[10:31:00.0000] <TabAtkins>
Maybe you should be less spammy, Philip`.

[10:32:00.0000] <Philip`>
It's the first I've sent to public-html in >2 months

[10:33:00.0000] <Philip`>
though it's a vaguely technical discussion and not about process and therefore it's probably off-topic for public-html

[10:33:01.0000] <TabAtkins>
There's your problem, then.

[10:50:00.0000] <hsivonen>
Looks like Flash 10.1 got finally released

[11:54:00.0000] <enderandpeter_>
Hello, all you brilliant crafters of the new web

[11:56:00.0000] <AryehGregor>
Hello, brilliant craftee, or beneficiary thereof.

[11:57:00.0000] <enderandpeter_>
Oh, I'm quite the novice, believe you me. But I hope to help out where I can...

[11:58:00.0000] <AryehGregor>
To help out, read the spec and point out anything unclear or wrong.  And try using it and tell whatwg⊙wo about it.

[11:58:01.0000] <AryehGregor>
(or maybe it's @lists.whatwg.org, I forget)

[12:03:00.0000] <enderandpeter_>
Oh wow. I like the tool for submitting comments about sections.

[12:05:00.0000] <enderandpeter_>
To tell you the truth, I have not really seriously sat down with any of the W3C specifications. When I familiarize myself about so many web technologies, I find myself focusing on the literature browser makers produce, and general tutorials and guides. I do think understanding that specs themselves is important, to know where everyone's coming from.

[12:06:00.0000] <AryehGregor>
This channel isn't about a W3C specification.  :)

[12:06:01.0000] <AryehGregor>
But never mind, that's a detail.

[12:07:00.0000] <enderandpeter_>
Quite true, but I assume that one day HTML5 will because a specification of theirs, won't it?

[12:07:01.0000] <AryehGregor>
There's a W3C HTML5 specification and also a WHATWG one.  The latter is a superset of the former.

[12:07:02.0000] <AryehGregor>
There's no reason to believe that the WHATWG will go away anytime soon.

[12:08:00.0000] <AryehGregor>
Some people here would prefer it if the HTMLWG went away instead.

[12:09:00.0000] <enderandpeter_>
The WHATWG spec is a superset of the W3C one? So there is much the WHATWG one entails that W3C one does not or will not?

[12:09:01.0000] <TabAtkins>
At the moment, not "much".  The exact differences are detailed at the very beginning of the whatwg version.

[12:10:00.0000] <AryehGregor>
Well, the WHATWG spec is a combination of a bunch of W3C specs plus some extra minor things.

[12:10:01.0000] <AryehGregor>
The W3C split it up a lot, so the spec called "HTML5" there is a lot smaller, but other parts are in other specs.

[12:11:00.0000] <enderandpeter_>
I see, I see... Personally, I think the more heads working on conforming web design, the better off the world will be. Well, maybe I will finally sit down and read the darn thing... both I suppose.

[12:13:00.0000] <enderandpeter_>
Oh man, just to really hammer in my lack of experience, what does it mean that a section is "non-normative". Is there a place you know of where I can get a hold of so much of the jargon these specs use?

[12:16:00.0000] <AryehGregor>
Try a dictionary, in this case.

[12:16:01.0000] <AryehGregor>
"Normative" means "binding, required".  Non-normative text is explanatory, not binding.

[12:19:00.0000] <enderandpeter_>
I see. Wow, I didn't mean to sound like THAT much of novice. It's just that I find a lot of technical writing about programming get's very esoteric very easily, so you just never know...

[12:33:00.0000] <cardona507>
maryz - holding down the spot

[12:33:01.0000] <TabAtkins>
cardona507: Wrong window again.  ^_^

[12:33:02.0000] <cardona507>
doh

[12:33:03.0000] <cardona507>
wow - I gotta work on that :)

[12:34:00.0000] <cardona507>
sorry bout that

[12:37:00.0000] <Philip`>
enderandpeter_: Specifications have a special kind of language that most other technical writing doesn't, though to make it more confusing some specifications aren't very consistent at using that language

[12:38:00.0000] <Philip`>
So it's probably best to ask seemingly obvious questions, because sometimes the answers will be important but unobvious :-)

[13:03:00.0000] <AryehGregor>
This post is awesome for the inverted answers at the bottom: http://www.xn--8ws00zhy3a.com/blog/2010/06/html5-atom-gone-wrong

[13:35:00.0000] <TabAtkins>
Heh, that is pretty cool.  And luckily easy to reverse.

[13:38:00.0000] <Philip`>
TabAtkins: Indeed - just stand on your head

[13:38:01.0000] <Lachy>
AryehGregor, several of those issues that post complains about are not issues at all. The author just failed to understand the spec and blamed his own mistakes on HTML5

[13:38:02.0000] <AryehGregor>
Lachy, dunno about that, I was just figuring the upside-down thing was cool.

[13:39:00.0000] <Lachy>
a couple of them seem like trivial spec bugs that can be fixed in the spec, and not really significant issues to be complaining about.

[13:42:00.0000] <AryehGregor>
If there are serious problems with the algorithm, they should be fixed.  This is part of why two specs is bad, because people with legitimate complaints will say "Well, it's only in that nasty WHATWG spec, so I don't have to worry about it."

[13:42:01.0000] <AryehGregor>
Same thing with Microdata vs. RDFa.

[13:42:02.0000] <AryehGregor>
(There might not be serious problems, of course.)

[13:43:00.0000] <TabAtkins>
Hmm.  Looking at the atom format, it appears that <title>s are supposed to be text, not markup.

[13:45:00.0000] <Lachy>
From the list in that page, issues 1, 3, 4 and 8 are trivial to fix. Issues 5, 6 and 7 are bogus. Issue 2 I'm not sure about, but most likely bogus. Issue 9 is legitimate, but not much that can be done about it and not a serious problem.

[13:46:00.0000] <TabAtkins>
Issue 2 was discussed on the list.  It's bogus as he phrased it, but there can indeed be problems in some circumstances.

[13:49:00.0000] <TabAtkins>
But the id he claims would be constructed by the html algo is bogus unless the feed is broken (to get that value, the author would have to write <a rel=bookmark href=index>Permalink</a>, essentially).

[13:52:00.0000] <TabAtkins>
I see no way to comment on his entry, nor any contact info for him anywhere on his site.  Way to "blog".

[13:53:00.0000] <AryehGregor>
TabAtkins, I saw a blog like that recently.

[13:53:01.0000] <Lachy>
Doesn't matter. Just respond on public-html and state which of the issues Hixie should fix, and which can be ignored

[13:53:02.0000] <AryehGregor>
Very annoying, because it attributed my idea to someone else.  :(

[13:54:00.0000] <Philip`>
TabAtkins: Some people have better things to do than spend every day moderating comments on their blog

[13:54:01.0000] <Philip`>
If people want to respond they can just write their own blog post

[13:55:00.0000] <AryehGregor>
Hixie adopts that policy for his blog, but I'm pretty sure he also has contact info somewhere.

[13:55:01.0000] <TabAtkins>
That's not a "response".  That's a separate entity which happens to be talking about the original post.

[13:55:02.0000] <TabAtkins>
I'm okay with no comments, but *some* way to contact the author is a requirement.

[13:55:03.0000] <TabAtkins>
Or else you're just an ass, imo.

[13:57:00.0000] <Philip`>
You could use one of the IM contacts on http://www.詹姆斯.com/foaf

[13:58:00.0000] <TabAtkins>
How did you discover that?

[13:58:01.0000] <Philip`>
I leveraged the magic of RDF

[13:58:02.0000] <TabAtkins>
Oh, ffs.  That list of differences?  It's a series of "<sup>1</sup> <span>...</span><br>" in a div.

[13:59:00.0000] <Philip`>
Actually I just did view source on the index page

[14:04:00.0000] <hober>
TabAtkins: <atom:title> is an Atom Text Construct, which can be all sorts of things, including HTML.

[14:05:00.0000] <TabAtkins>
hober: Thanks - I didn't feel like diving into the rfc enough to find the definition of "Text Construct".

[14:05:01.0000] <TabAtkins>
/me is writing up the necessary fixes right now.

[14:25:00.0000] <KaOSoFt>
There still doesn't seem to be an off-line HTML5 validator, right? Pretty much just like this: http://users.skynet.be/mgueury/mozilla/

[14:25:01.0000] <KaOSoFt>
I know the spec is constantly changing and all, but I'd love to have something like it.

[14:25:02.0000] <KaOSoFt>
I, for once, do not have Internet at home, and such an extension would make miracles for me.

[14:26:00.0000] <Philip`>
KaOSoFt: You can (probably) run the Validator.nu code offline

[14:27:00.0000] <KaOSoFt>
I didn't know it's download-able.

[14:27:01.0000] <Philip`>
http://about.validator.nu/#src

[14:28:00.0000] <KaOSoFt>
Wow, thank you. I'll check it out right away!

[14:57:00.0000] <jgraham>
KaOSoFt: You have to run the build script twice or it will crash in a mysterious way

[14:58:00.0000] <jgraham>
(unless that got fixed)

[14:58:01.0000] <KaOSoFt>
Yeah, I read the note in that section.

[15:08:00.0000] <boaz>
QWE IKTHJGWO;LGKUIAJER5'PLT OGPZ5

[15:08:01.0000] <boaz>
;LGN

[15:08:02.0000] <boaz>
K 4{W}P"r[f kM"{8og ijBK

[15:08:03.0000] <boaz>
]PORTH80 1

[15:08:04.0000] <boaz>
hq E{)&Tgfn ]

[15:08:05.0000] <boaz>
bot fgiu

[15:08:06.0000] <boaz>
{HN0o5rib 8.on ojvceip'OU;G NWE

[15:08:07.0000] <boaz>
R


[15:08:08.0000] <boaz>
IGJH8ow/4nhigf ojnk

[15:08:09.0000] <boaz>
k

[15:08:10.0000] <boaz>
hl;W

[15:08:11.0000] <boaz>
$M{

[15:08:12.0000] <boaz>
bguw/n

[15:08:13.0000] <boaz>
 uywr

[15:08:14.0000] <jgraham>
You lost me at QWE

[15:23:00.0000] <gsnedders>
Oh hell yeah. Four hour wait at Skavsta for the bus tomorrow!

[15:43:00.0000] <jgraham>
gsnedders: I didn't even know that was possible

[15:48:00.0000] <zcorpan_>
does new XMLHttpRequest; make any sense?

[15:48:01.0000] <jgraham>
Yeah

[15:48:02.0000] <jgraham>
That always works in ES

[15:48:03.0000] <zcorpan_>
also for normal functions?

[15:49:00.0000] <jgraham>
I think so, unless my memory is faulty

[15:49:01.0000] <zcorpan_>
ok

[15:50:00.0000] <jgraham>
Yeah just checked by experiment

[15:50:01.0000] <jgraham>
Anyway I have a rely to that email

[15:50:02.0000] <jgraham>
*reply

[15:50:03.0000] <jgraham>
I will find a mail system that lets me send it...


2010-06-12
[18:34:00.0000] <til>
i have a page that i wrote in HTML5 forms, then wrote a javascript implementation of the validity API

[18:34:01.0000] <til>
the latest builds of WebKit are breaking my page, and i'm trying to figure out if they or i implemented wrong

[18:35:00.0000] <til>
when i hit submit, the page jumps to the first empty form element

[18:35:01.0000] <til>
before calling onsubmit

[18:35:02.0000] <til>
any hints as to what i should look into?

[18:35:03.0000] <til>
i haven't played with HTML5 forms in a long whule

[18:37:00.0000] <TabAtkins_>
Forms *should* validate before submitting, and not submit if anything's invalid.

[18:40:00.0000] <miketaylr>
til: do you have the required attribute set on those empty inputs?

[18:41:00.0000] <til>
miketaylr: nope

[18:41:01.0000] <til>
i've found a couple bugs in earlier chrome builds.  one was a bug in their implementation of checkValidity

[18:41:02.0000] <til>
the other was a disagreement between Hixie and myself on what it means for a checkbox group to be @required

[18:42:00.0000] <til>
but this has otherwise worked great until now

[18:44:00.0000] <til>
TabAtkins_: i don't disagree, but i have to figure out if the rejected form should be considered invalid or not

[18:45:00.0000] <til>
right now, i'm using form.@onsubmit="return checkSurveyValidity();" to validate the form and block it from submitting on pre-HTML5 browsers

[18:46:00.0000] <til>
i may need to use a different event handler.  webkit isn't running my validator at all (which means my smooth scrolling and message-displaying script isn't running)

[19:29:00.0000] <dbaron>
AryehGregor, I think Asa Dotzler has a bunch of statistics on how much of various types of Mozilla contributions are employees vs. volunteers

[22:52:00.0000] <dbaron>
/me reads the W3C process document and learns that PER status is basically PR + LC at the same time

[22:53:00.0000] <dbaron>
/me didn't realize it had so many aspects of Last Call

[02:20:00.0000] <GPHemsley>
/me looks up, then goes to bed

[02:52:00.0000] <gsnedders>
jgraham: Nor did I. Until recetnly, you only had to wait two hours, worse case, AFAIK

[05:07:00.0000] <annevk>
var x = new XMLHttpRequest;

[05:07:01.0000] <annevk>
you learn something new every day

[05:09:00.0000] <aho>
i wouldn't omit those invocation parens though

[05:09:01.0000] <aho>
it's a function call... don't make it look like something else

[05:30:00.0000] <nielsle>
Is it possible to get two separate contexts ctx1 and ctx2 that refer to the same canvas? I would like to use ctx1.lineTo(x,y) and ctx2.lineTo(x,y) to draw two curves in parallel.

[05:33:00.0000] <nielsle>
I can probably hack a solution together myself, but I would like html5 to solve the problem for me :)

[05:35:00.0000] <aho>
what do you mean with "draw in parallel"?

[05:36:00.0000] <aho>
you draw stuff to the canvas... and once you're done with that and the UI is back in control... the updated drawing is shown

[05:36:01.0000] <aho>
drawing onto one context or thousands doesn't change anything

[05:37:00.0000] <aho>
/me pokes nielsle 

[05:50:00.0000] <nielsle>
I am toying with a canvas based multi user paint program. Whenever the user move the mouse I use lineTo(x,y) to draw a line. If another user draws something then the client recieves a json object and I also use lineTo to draw that. This works nicely unless the client revieves a json object while the user is in the process of drawing a stroke. I can probably solve this by creating some locks.

[05:51:00.0000] <annevk>
Reading http://www.w3.org/mid/4C1359A1.7050403⊙in I'm not really sure what to say... I did not quite agree with the removal of the Atom mapping, but going as far as objecting goes way too far.

[05:52:00.0000] <annevk>
It seems e.g. Julian is far less constrained in that regard going as far as raising an issue over something as simple as a reference.

[05:52:01.0000] <annevk>
Does that mean the progress is tilted in the advantage of people that like to bicker over every detail?

[05:54:00.0000] <aho>
nielsle, well... you just have to update your canvas if an event occurs (input, network, whatever)

[05:54:01.0000] <aho>
that's all there is to it

[05:54:02.0000] <aho>
(it's single threaded - there are no locks)

[05:55:00.0000] <annevk>
he means a lock to indicate it cannot be updated right this instant because something else is going on

[05:55:01.0000] <annevk>
nielsle, we might get dedicated path objects in the future, but not for now

[05:57:00.0000] <annevk>
pretty good read: http://bitworking.org/news/2010/06/distributed-extensibility

[05:58:00.0000] <nielsle>
annevk: Yes you are right. I was asking for a path object. (Sorry if I was unclear).

[06:00:00.0000] <annevk>
I believe the thinking is that it should have been that way from the beginning, but in the beginning everything went fast, and so we're stuck with some not too great design decisions

[06:00:01.0000] <annevk>
but we can also still paper over them in the future, and had we not it would probably have taken all much longer to get universally deployed

[06:04:00.0000] <nielsle>
Ah I see. Thank you for the answers.

[09:07:00.0000] <annevk>
what is http://wiki.whatwg.org/index.php?title=Talk:Timed_tracks&curid=1928&diff=4923&oldid=4716 about?

[09:08:00.0000] <annevk>
ah

[09:08:01.0000] <annevk>
it seems to be about http://wiki.whatwg.org/index.php?title=Timed_tracks&action=historysubmit&diff=4647&oldid=4645

[16:49:00.0000] <boblet>
Normally <li>’s content is just phrasing content. if I have an <aside> or other flow element inside an <li>, should I wrap the phrasing content in <p>?

[16:50:00.0000] <boblet>
eg <li>phrasing content</li> → <li><aside>…</aside>phrasing content</li> or <li><aside>…</aside><p>phrasing content</p></li>?

[16:55:00.0000] <daedb>
Works either way, imo

[16:59:00.0000] <boblet>
thanks daedb. yeah I know it works, but I wondered if it was more ‘correct’ to explicitly add the <p>

[16:59:01.0000] <daedb>
nah, I doubt it makes any difference

[16:59:02.0000] <gsnedders>
I like fixing bugs and then finding out other code relies upon those bugs.


2010-06-13
[17:00:00.0000] <gsnedders>
I'd have an explicit p there, as otherwise it is implied and doesn't make rendering quite so obvious

[17:02:00.0000] <othermaciej>
down to 216 unread in my public-html folder....

[17:21:00.0000] <Hixie>
boblet: it technically makes no difference -- in both cases the "phrasing content" text is a paragraph, it's just that the <p> element makes it more obvious

[17:22:00.0000] <boblet>
Hixie: thanks for the clarification

[17:22:01.0000] <Hixie>
boblet: there's a section in the spec that talks about implied paragraphs (note that this is not implied <p>s, just implied "semantic" paragraphs, like the implied sections cause by <h2> etc)

[17:22:02.0000] <boblet>
aah yeah, I think that’s what I’m half-remembering

[17:22:03.0000] <boblet>
will re-read that now

[17:23:00.0000] <Hixie>
it's not an uber-clear section

[17:23:01.0000] <Hixie>
it's hard to really explain it because of the way HTML has implied elements

[17:24:00.0000] <boblet>
Hixie: oh speaking of half-remembering, I think you originally did some research looking at class/id names to inform new sectioning element naming. was that http://code.google.com/webstats/2005-12/classes.html ?

[17:24:01.0000] <Hixie>
yeah

[17:24:02.0000] <boblet>
(I’m writing a little aside on the history of the new elements)

[17:25:00.0000] <boblet>
Hixie: if you have editing powers on that page it refers to <date> btw ;-)

[17:25:01.0000] <Hixie>
it's a report from 2005, it's not supposed to be updated

[17:26:00.0000] <boblet>
cool :)

[17:26:01.0000] <daedb>
There was a <date> element in 2005?

[17:26:02.0000] <Hixie>
it became <time> iirc

[17:26:03.0000] <Hixie>
or was it <t>?

[17:26:04.0000] <Hixie>
i forget

[17:27:00.0000] <boblet>
daedb: yeah I was surprised too :)

[17:28:00.0000] <daedb>
I know I've seen that page before, but don't remember the date element link :)

[17:29:00.0000] <boblet>
still looking for the email that introduces those changes to WHATWG list, but knowing the report is the basis is a good start

[22:14:00.0000] <boblet>
Hixie: re: class="link" in the popular class name survey, I doubt this is the reason but I sometimes use <a class="img-link"><img></a>, as I tend to use border-bottom over text-decoration:underline, and need to turn that off when the link content is an image

[22:15:00.0000] <boblet>
hopefully this won’t be necessary once the CSS3 text-decoration:underline extra properties are implemented (distance from text, color, line style/thickness etc)

[00:58:00.0000] <Hixie>
gotta love leif's simple view of the world

[07:20:00.0000] <henrikbjorn>
is it allowed to have mutiple aside's for example in a div#sidebar ?

[08:07:00.0000] <boblet>
henrikbjorn: sure. but it might be better to put multiple <section>s in an <aside id="sidebar">…

[08:09:00.0000] <boblet>
Hixie: re: the 2005 Web Authoring Stats, I’d be interested in the data for the current web. It’s been almost 5 years, so the differences would be interesting

[08:11:00.0000] <boblet>
Hixie: Also it’d be interesting to specifically check for class names that might indicate a page made with best practices — things like clearfix, group, amp — and see if the code from that small fraction of pages is different to the norm

[08:12:00.0000] <boblet>
Hixie: finally if you really want to blue-sky it I bet www-style would love something similar for CSS

[08:14:00.0000] <boblet>
Hixie: oh, sitename as id on body (eg <body id="whatwg-org">) would be another indicator

[08:19:00.0000] <henrikbjorn>
boblet: but the content in the sidebar is blocks/widgets with fx a list of recent items an ad or a tip of some sort

[08:20:00.0000] <henrikbjorn>
would that still be valid in a aside ??

[08:20:01.0000] <henrikbjorn>
with multiple sections

[08:20:02.0000] <boblet>
henrikbjorn: <aside> can contain other sectioning elements, such as multiple <section> elements

[08:21:00.0000] <henrikbjorn>
yeah but i thought aside was for referencing stuff on the page

[08:21:01.0000] <henrikbjorn>
and theese things are not related to the content

[08:21:02.0000] <boblet>
the thing to ask yourself is whether the content in the <aside> is related to the section/article it’s in, and whether the content is not essential to the understanding of the main content

[08:22:00.0000] <boblet>
aah

[08:23:00.0000] <henrikbjorn>
hopefully what i just wrote makes sense >P

[08:23:01.0000] <boblet>
if it’s unrelated to the main content (which is presumably in <article), then in an <aside> that’s a sibling of <article> (ie an <aside> that’s related to <body>) may be the way to go

[08:24:00.0000] <henrikbjorn>
okay :)

[08:24:01.0000] <henrikbjorn>
i think i get it now

[08:24:02.0000] <henrikbjorn>
thanks a lot

[08:25:00.0000] <boblet>
<body>…<article></article><aside></aside></body> is a page-level aside (applies to the page)…

[08:25:01.0000] <boblet>
<body>…<article><aside></aside></article></body> the aside applies to the article content

[08:26:00.0000] <henrikbjorn>
ohh

[08:26:01.0000] <henrikbjorn>
ahh :D

[08:26:02.0000] <boblet>
you can see an example of a page aside in the second code sample here: http://oli.jp/2009/html5-structure4/#article-page

[08:26:03.0000] <henrikbjorn>
then a aside related to body is the way to go

[08:27:00.0000] <henrikbjorn>
all this new stuff is hard >P

[08:28:00.0000] <boblet>
at the top of that page the coding style pulldown is in an aside that’s in <article>, because it’s related to the article (not the page)

[08:28:01.0000] <boblet>
heh, it gets easier

[08:28:02.0000] <boblet>
to give a simple example, you know how in HTML 4 <address> was only for the contact info for the page right?

[08:29:00.0000] <boblet>
in HTML5 you can also use <address> in an <article>, and when you do it becomes contact info for the article only.

[08:29:01.0000] <boblet>
both elements are scoped by their context — what they apply to depends on what they’re inside

[08:30:00.0000] <boblet>
Think Russian nested dolls… in cute little angle brackets ;-)

[08:32:00.0000] <henrikbjorn>
:p

[08:32:01.0000] <henrikbjorn>
thanks againg :)

[08:34:00.0000] <boblet>
henrikbjorn: np. if you have probs or questions this is a pretty good place to ask. after that check www.html5doctor.com, www.diveintohtml5.com, or even www.oli.jp ;-)

[08:35:00.0000] <henrikbjorn>
bookmarked .D

[08:35:01.0000] <henrikbjorn>
:D

[08:41:00.0000] <boblet>
henrikbjorn: (conflict of interest statement: I’m writing for two of them :) )

[08:41:01.0000] <henrikbjorn>
:)

[09:09:00.0000] <henrikbjorn>
is this valid / makes sense ? https://gist.github.com/92e7383a585cc0e1523f

[09:35:00.0000] <gsnedders>
jgraham: You'll be pleased to know I'm now listening to death metal to put certain other songs out of my head

[14:05:00.0000] <paul_irish>
anyone remember that trailer thats being used instead of bigbuckbunny in html5 demos recently?

[14:09:00.0000] <paul_irish>
Sintel ! http://durian.blender.org/download/

[14:28:00.0000] <aliok_>
hello everyone

[14:29:00.0000] <gsnedders>
'ello

[14:29:01.0000] <aliok_>
I am working on HTML5 support for MyFaces2 (the open JavaServer Faces implemenatation), and I want to ask something

[14:30:00.0000] <aliok_>
I saw there is a term "boolean attribute" in HTML5 spec

[14:30:01.0000] <aliok_>
the problem is

[14:30:02.0000] <aliok_>
I mean, I want to ask if I understood correctly

[14:32:00.0000] <Hixie>
aliok_: you have to ask the question to get an answer :-)

[14:33:00.0000] <Hixie>
aliok_: there's a lot of people who just look in every few hours and answer anything they see that they can answer

[14:33:01.0000] <aliok_>
I must write draggable="true" if I need to make a component draggable, while  I must write controls="controls" to show the controls.

[14:33:02.0000] <aliok_>
I am sure these difference makes sense, but I couldn't find out.

[14:36:00.0000] <aliok_>
why don't HTML5 does not accept "true" and "false" as a value of a boolean attribute?

[14:37:00.0000] <Hixie>
historical reasons from sgml (1986)

[14:37:01.0000] <Hixie>
draggable="" isn't a boolean attribute

[14:38:00.0000] <Hixie>
we really should rename "Boolean attribute"  to something else

[14:38:01.0000] <aliok_>
:)

[14:43:00.0000] <aliok_>
yes, I've learnt that. I was very mad at Chrome when it didn't allow me to drag with draggable="draggable" :) But it was right!

[14:43:01.0000] <aliok_>
thank you, <Hixie>

[14:44:00.0000] <Hixie>
np


2010-06-14
[21:47:00.0000] <MikeSmith>
http://item.rakuten.co.jp/icefield/c/0000000218/

[21:49:00.0000] <MikeSmith>
http://item.rakuten.co.jp/icefield/10011112/

[21:51:00.0000] <MikeSmith>
http://item.rakuten.co.jp/icefield/10010977/

[21:52:00.0000] <wirepair>
ha

[21:52:01.0000] <wirepair>
there's now a brand called html huh

[21:52:02.0000] <wirepair>
http://item.rakuten.co.jp/icefield/10008676/ is good

[22:14:00.0000] <MikeSmith>
wirepair: I got to buy some of this stuff

[22:16:00.0000] <wirepair>
hehe

[22:16:01.0000] <MikeSmith>
dude must have a shop somewhere in town

[22:16:02.0000] <wirepair>
i swear though rakuten gives me a seizure

[22:16:03.0000] <wirepair>
worst website ever

[22:17:00.0000] <wirepair>
i'm guessing harajuku or shibuya if they have a shop hehe

[22:19:00.0000] <MikeSmith>
渋谷PARCO パート1 ６F html渋谷店

[22:24:00.0000] <wirepair>
my terms ascii only ;<

[22:25:00.0000] <wirepair>
btw spoke with Abarth

[22:25:01.0000] <MikeSmith>
great

[22:25:02.0000] <wirepair>
looks like the layoutTestController they use is somewhat webkit specific

[22:25:03.0000] <MikeSmith>
yeah

[22:25:04.0000] <wirepair>
so i'm just going to incorporate those testcases into my own system

[22:25:05.0000] <MikeSmith>
knowed that already

[22:26:00.0000] <wirepair>
which i have working on ie mobile, ie8, opera 9 something, chrome, and firefox

[22:26:01.0000] <MikeSmith>
that's one reason I suggested talking with Slyvain on #testing

[22:26:02.0000] <wirepair>
yeah

[22:26:03.0000] <MikeSmith>
syp

[22:26:04.0000] <MikeSmith>
he's thought this stuff through quite a bit already

[22:26:05.0000] <wirepair>
yeah that irc is kinda quiet ;>

[22:27:00.0000] <wirepair>
i'll ping him there and see when i get a response

[22:27:01.0000] <MikeSmith>
yeah, if you ping syp there during CET hours, he will likely responde

[22:27:02.0000] <MikeSmith>
oh man, check out this hipster: http://image.space.rakuten.co.jp/lg01/48/0000990248/97/img760e01c1zikazj.jpeg

[22:27:03.0000] <MikeSmith>
that is so going to be me after I stop by Parco tonight

[22:27:04.0000] <wirepair>
haha

[22:28:00.0000] <wirepair>
Awesome heh

[22:28:01.0000] <wirepair>
so html isn't copyright'd then

[22:28:02.0000] <wirepair>
?

[22:29:00.0000] <ako>
txt also isn't copyrighted

[22:29:01.0000] <ako>
*shrug*

[22:30:00.0000] <wirepair>
well java is isn't it?

[22:30:01.0000] <wirepair>
even though it could refer to coffee

[22:30:02.0000] <ako>
"java" the name is trademarked

[22:30:03.0000] <wirepair>
yeah sorry trademarked, not copyright

[22:30:04.0000] <ako>
(different thing)

[22:32:00.0000] <ako>
using "html" as brand name is about as stupid as using "the" as brand name

[22:32:01.0000] <ako>
it's ungoogleable

[22:32:02.0000] <ako>
and therefore completely useless

[22:32:03.0000] <ako>
:>

[22:32:04.0000] <wirepair>
it won MikeSmith over did it not? ;>

[22:35:00.0000] <ako>
well, whatever... buh bai :>

[22:43:00.0000] <MikeSmith>
man, I'm going to grow broke buying HTML goods

[00:57:00.0000] <MikeSmith>
the single-page version of the HTML5 spec is inching up on 5MB

[00:57:01.0000] <MikeSmith>
currently 4.8MB

[00:59:00.0000] <nessy>
ouch

[01:00:00.0000] <zcorpan_>
MikeSmith: what about complete.html?

[02:03:00.0000] <MikeSmith>
zcorpan_: dunno, haven't tried it yet

[02:04:00.0000] <MikeSmith>
I just noticed how big the HTML5 spec had gotten because I needed to sync up the v.nu copy to the latest upstream

[02:04:01.0000] <MikeSmith>
btw

[02:04:02.0000] <MikeSmith>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=9595

[02:04:03.0000] <MikeSmith>
[[

[02:04:04.0000] <MikeSmith>
I just bumped into a wierd problem in all IE browsers and it turned out to be

[02:04:05.0000] <MikeSmith>
an HTML comment just before the DOCTYPE. The validator did not recognize it as

[02:04:06.0000] <MikeSmith>
an error but all IE browsers failed to properly assign css proeprties etc.

[02:04:07.0000] <MikeSmith>
]]

[02:05:00.0000] <MikeSmith>
that's not a validator bug of course

[02:05:01.0000] <MikeSmith>
but is what that says about IE behavior true?

[02:05:02.0000] <zcorpan_>
yep, at least up to and including ie8

[02:05:03.0000] <zcorpan_>
dunno about ie9

[02:05:04.0000] <MikeSmith>
ah, OK

[02:05:05.0000] <MikeSmith>
/me didn't know

[02:10:00.0000] <MikeSmith>
yee hah

[02:10:01.0000] <MikeSmith>
complete spec is 5.1MB

[02:56:00.0000] <annevk>
not a huge difference the splitting out has made then...

[03:00:00.0000] <MikeSmith>
annevk: I guess not (as far as the file size at least)

[03:02:00.0000] <zcorpan_>
if removing features from the spec has the consequence of the removed features getting implemented in browsers, then it seems like a good thing

[03:12:00.0000] <MikeSmith>
zcorpan_: heh

[03:51:00.0000] <MikeSmith>
http://validator.nu/?doc=http%3A%2F%2Fbikramgreenvalley.com%2FMain_Content%2FFAQ.php

[03:51:01.0000] <MikeSmith>
error: ILLEGAL_CHARACTER in QUERY

[03:51:02.0000] <MikeSmith>
src="/Parts/GenerateFiles/MakeFiles.php?name=ToggleDL&amp;type=text/javascript&amp;path=Parts/ToggleDL&amp;js[]=ToggleDL"

[03:51:03.0000] <MikeSmith>
what's the illegal character?

[03:51:04.0000] <MikeSmith>
brackets?

[03:52:00.0000] <MikeSmith>
solidus?

[03:56:00.0000] <hasather>
MikeSmith: brackets

[03:58:00.0000] <MikeSmith>
hasather: thanks

[04:01:00.0000] <craiggles>
I'm playing around with HTML5 canvas element and I'm now thinking of making a game, for a game like super mario bros - http://ian-albert.com/misc/gamemaps/super_mario_bros/full/mario-1-1.gif - would you suggest me using a 2D array for the map?

[04:09:00.0000] <hsivonen>
MikeSmith: IIRC, if you copy a URL with brackets from the URL bar, the brackets don't get percent escaped

[04:09:01.0000] <hsivonen>
MikeSmith: I'm not sure if this should be treated as a Firefox bug or as an RFC bug

[04:45:00.0000] <erlehmann>
hsivonen, i seem to remember that brackets were perfectly legal in URIs, aren't they ?

[04:50:00.0000] <hasather>
erlehmann: maybe it's only allowed around IP literals even, dunno

[04:50:01.0000] <erlehmann>
thx.

[05:01:00.0000] <MikeSmithX>
hsivonen: I see

[06:07:00.0000] <annevk>
"needed" could maybe be an alternative, but it kind of sucks

[06:07:01.0000] <annevk>
and might very well have similar issues

[06:09:00.0000] <annevk>
funny

[06:09:01.0000] <annevk>
http://blogs.msdn.com/b/ie/archive/2010/06/11/ie-test-center-updates.aspx

[06:10:00.0000] <annevk>
they updated the tests in a way that IE still passes and can continue to have their silly no whitespace nodes in the DOM bug

[06:10:01.0000] <annevk>
(foreign content parsing tests)

[06:11:00.0000] <annevk>
without even acknowledging the problem

[06:20:00.0000] <MikeSmith>
https://lists.webkit.org/pipermail/webkit-dev/2010-June/013215.html

[06:20:01.0000] <MikeSmith>
from abarth

[06:20:02.0000] <MikeSmith>
"HTML5 tokenizer landing soon"

[06:24:00.0000] <MikeSmith>
and https://spreadsheets.google.com/ccc?key=0AppchfQ5mBrEdDFJUW5DOGNsdmtvZkN0ZmIzMjdaT0E&hl=en

[06:40:00.0000] <hsivonen>
Interesting. Either WebKit just has broader test suite coverage than Gecko or their test suite depends on more invalid stuff.

[06:40:01.0000] <hsivonen>
IIRC, the HTML5 *tokenizer* regressed fewer tests in Mozilla's test suites

[06:42:00.0000] <gsnedders>
Looks like just testing more invalid stuff, not random stuff relying upno it

[06:42:01.0000] <hsivonen>
Also, good for them that they have the luxury of landing the tokenizer and the tree builder independently

[06:43:00.0000] <hsivonen>
i.e., their pre-existing abstaction between the tokenizer and the tree builder has to be relatively proper

[06:50:00.0000] <craiggles>
http://pastie.org/1003752 .. I'm having some problems with my code, perhaps I've done something wrong. well basically what is happening is that one circle isn't being filled green (#008000).

[07:19:00.0000] <craiggles>
is it possible to check if a certain shape is in a specified area of my canvas? i'm looking through the spec atm 

[07:23:00.0000] <annevk>
craiggles, there's isPointInPath

[10:14:00.0000] <ap>
hi john! hi everyone!

[10:14:01.0000] <ap>
sorry, wrong channel

[10:17:00.0000] <AryehGregor>
Hi anyway!

[11:26:00.0000] <tabatkins>
Man, switching my monitors to vertical is an improvement, but it'll take some getting used to.  I'm not used to seeing this much of a website at once.

[11:30:00.0000] <miketaylr>
it's always kinda freaked me out

[11:43:00.0000] <cardona507>
vertical is great for one screen as long as I have another that is horizontal

[12:02:00.0000] <jgraham>
Yeah that seems like a good setup

[12:03:00.0000] <jgraham>
I need a second big screen :)

[12:03:01.0000] <jgraham>
/me has one big widescreen and one small 4:3

[12:03:02.0000] <TabAtkins>
Two large widescreens here.  I think I'd like a small 4:3 to keep horizontal too.

[12:04:00.0000] <gsnedders>
TabAtkins: That I always found to be the real advantage of a 30" over two 24": vetical space.

[12:08:00.0000] <jgraham>
gsnedders: With one vertical, 2x24 gives you more vertical space than 1x30

[12:08:01.0000] <gsnedders>
Yes, that's true

[12:08:02.0000] <jgraham>
Really big monitors are useful for some things e.g. image processing

[12:08:03.0000] <jgraham>
For coding not so much

[12:08:04.0000] <jgraham>
(the basic question is "do I want one application maximised or several side-by-side)

[12:09:00.0000] <jgraham>
s/)/")/

[12:09:01.0000] <gsnedders>
Really big monitors tend to have less good colour reproduction than some, e.g., 24" screens available, so it's a question of which you care about more

[12:09:02.0000] <jgraham>
Doesn't that just depend on what type of display technology they use?

[12:09:03.0000] <gsnedders>
Well, yes. But all the 30" screens available aren't as good as some 24".

[12:10:00.0000] <jgraham>
(which is _never_ever_ specified)

[12:10:01.0000] <gsnedders>
Also: why is everything so frickin' expensive here?

[12:10:02.0000] <jgraham>
/me wonders if that is true or if you can get insanely expensive 30" screens

[12:10:03.0000] <jgraham>
gsnedders: High tax rates, high wages, high expectation of cost?

[12:10:04.0000] <othermaciej>
doh, I forgot to implement the IDL attribute for @hidden

[12:12:00.0000] <jgraham>
gsnedders: http://www.necdisplay.com/Products/Product/?product=ec17518a-366e-4f09-8173-ca42d439b357

[12:13:00.0000] <TabAtkins>
othermaciej, what the hell.

[12:13:01.0000] <gsnedders>
/me hadn't seen NEC announce any 30" screens

[12:13:02.0000] <othermaciej>
TabAtkins: sorry man, I know you're disappointed

[12:13:03.0000] <TabAtkins>
;_; /wrist

[12:14:00.0000] <othermaciej>
now I won't be able to say it takes 30 minutes to implement

[12:14:01.0000] <jgraham>
othermaciej: How long does adding an IDL attribute take? I thought you just used some sort of IDL in the source tree?

[12:15:00.0000] <jgraham>
So it would be a copy+paste?

[12:15:01.0000] <othermaciej>
jgraham: it will take me probably another 30 minutes, including test case

[12:15:02.0000] <jgraham>
Well an hour isn't so bad, really :)

[12:15:03.0000] <othermaciej>
it's actually slightly easier than hooking up a markup attribute to CSS

[12:16:00.0000] <lifechamp>
Canvas (html5) + SVG libraries VS using SVG directly: theoretically, should there be a big performance difference or no?

[12:16:01.0000] <TabAtkins>
Depends entirely on how efficient canvas and svg are in the chosen browser.

[12:16:02.0000] <jgraham>
I guess that depends on the library and what you are doing

[12:19:00.0000] <lifechamp>
Ok, so answer is no, in some ways. I suppose the one underlying question was whether SVG is hardware supported (whereas a library building SVG atop Canvas) would not be.  It seems there is some HW support for SVG, but not widespread: http://en.wikipedia.org/wiki/Scalable_Vector_Graphics#Hardware_rendering

[12:19:01.0000] <othermaciej>
depends on your library and your use case

[12:20:00.0000] <lifechamp>
err, sorry, i meant the answer seems to be yes, in some ways, where HW support is used and helpful.

[12:30:00.0000] <lifechamp>
Although, since WebGL atop Canvas (?) is HW supported, I suppose there can be a HW supported SVG library atop Canvas as well. I'll look for it, but if anyone knows about this, please LMK.

[13:06:00.0000] <gsnedders>
jgraham: You had some idea of using numpy for computing the average run length, how so?

[14:07:00.0000] <jgraham>
gsnedders: I hope you were doing something fun earlier to be working now

[14:07:01.0000] <jgraham>
gsnedders: But I haven't thought about the numpy thing. What does the input look like?

[14:09:00.0000] <MattCampbell>
html5lib 0.90 isn't handling the </form> end-tag correctly, at least when using simpletree

[14:09:01.0000] <Hixie>
you sure?

[14:09:02.0000] <MattCampbell>
given this trivial test document: <form action="http://example.com/blah" method=post></form><p>test</p>

[14:09:03.0000] <MattCampbell>
parsing and serializing that moves </form> after </p>

[14:11:00.0000] <Hixie>
http://james.html5.org/cgi-bin/parsetree/parsetree.py?source=%3Cform+action%3D%22http://example.com/blah%22+method%3Dpost%3E%3C/form%3E%3Cp%3Etest%3C/p%3E

[14:11:01.0000] <Hixie>
says the <p> is after the <form>

[14:11:02.0000] <Hixie>
(that's an html5lib dumper)

[14:11:03.0000] <MattCampbell>
using which tree builder?

[14:11:04.0000] <jgraham>
Hixie: The code there needs to be updated

[14:12:00.0000] <jgraham>
MattCampbell: testing...

[14:12:01.0000] <Hixie>
k

[14:12:02.0000] <gsnedders>
jgraham: Gah, I'll just ask you tomorrow. I have other things to do now :P

[14:15:00.0000] <jgraham>
MattCampbell: WFM

[14:16:00.0000] <jgraham>
At least parsing does

[14:16:01.0000] <jgraham>
I suppose the serializer could be broken in interesting ways

[14:16:02.0000] <MattCampbell>
jgraham: Which tree builder are you using?

[14:17:00.0000] <jgraham>
MattCampbell: SimpleTree (not that it should matter or simpletree should be used for anything serious)

[14:17:01.0000] <jgraham>
(of course it *could* matter)

[14:17:02.0000] <MattCampbell>
jgraham: I'm using simpletree too

[14:18:00.0000] <jgraham>
MattCampbell: You are using the release version, not HEAD

[14:18:01.0000] <MattCampbell>
jgraham: And the problem is in the parser or the tree builder; I just traversed the parse tree myself

[14:18:02.0000] <MattCampbell>
0.90

[14:18:03.0000] <jgraham>
MattCampbell: can you run ./parse.py {your input} --tree from a command line and paste the resulst somewhere?

[14:19:00.0000] <jgraham>
gsnedders: So were you doing something fun? :p

[14:19:01.0000] <gsnedders>
jgraham: Yeah

[14:19:02.0000] <MattCampbell>
jgraham: In the examples directory, right?

[14:19:03.0000] <jgraham>
MattCampbell: Possibly, yeah

[14:20:00.0000] <gsnedders>
(and I only did around 30 mins of work)

[14:20:01.0000] <jgraham>
gsnedders: Sometyhing secret?

[14:20:02.0000] <gsnedders>
jgraham: Of course :)

[14:21:00.0000] <jgraham>
Wow, intrigue

[14:21:01.0000] <jgraham>
I will be sure to make up improbable rumors

[14:21:02.0000] <MattCampbell>
Any preferred pastebin here?

[14:21:03.0000] <gsnedders>
jgraham: What was it you made up before?

[14:21:04.0000] <gsnedders>
/me tries to remember

[14:21:05.0000] <gsnedders>
MattCampbell: not really

[14:22:00.0000] <MattCampbell>
jgraham: http://pastebin.com/dZi1k9qk

[14:23:00.0000] <gsnedders>
Odd.

[14:23:01.0000] <jgraham>
Yeah, odd

[14:23:02.0000] <MattCampbell>
Remember, in the input, the p element is outside of the form

[14:23:03.0000] <jgraham>
MattCampbell: Can you try and reproduce with the latest hg trunk?

[14:23:04.0000] <MattCampbell>
Python 2.6, btw

[14:24:00.0000] <MattCampbell>
OK

[14:24:01.0000] <MattCampbell>
What's the repo URL?

[14:24:02.0000] <MattCampbell>
not on google code, is it?

[14:24:03.0000] <MattCampbell>
oh ok, it is

[14:25:00.0000] <MattCampbell>
http://html5lib.googlecode.com/hg/python/

[14:25:01.0000] <MattCampbell>
right?

[14:25:02.0000] <gsnedders>
eah

[14:25:03.0000] <gsnedders>
*yeah

[14:26:00.0000] <jgraham>
gsnedders: This time at least one of them will involve sailor hats

[14:27:00.0000] <jgraham>
;)

[14:28:00.0000] <MattCampbell>
works in hg trunk

[14:28:01.0000] <MattCampbell>
I'll use that then

[14:29:00.0000] <MattCampbell>
Any particular reason for the switch to setuptools?

[14:31:00.0000] <jgraham>
There was a switch?

[14:31:01.0000] <gsnedders>
jgraham: hah

[14:31:02.0000] <jgraham>
I thought we had always been using setuptools?

[14:32:00.0000] <MattCampbell>
I think 0.90 used plain distutils

[14:32:01.0000] <jgraham>
(I think the answer is "python packaging makes me feel queasy every time I have to read about it, so we do not necessarily do the best thing")

[14:32:02.0000] <gsnedders>
I don't think 0.90 did

[14:32:03.0000] <jgraham>
If it did it was an accident

[14:32:04.0000] <MattCampbell>
ah, 0.90 used a try-except

[14:32:05.0000] <gsnedders>
But maybe it did. All I had to do with that release was nagging jgraham until he made it.

[14:32:06.0000] <MattCampbell>
so it fell back on distutils in the absence of setuptools

[14:33:00.0000] <jgraham>
I'm not sure we gain much from setuptools tbh so maybe it is enough to just use distutils

[14:33:01.0000] <jgraham>
But I don't really know what the tradeoffs are

[14:33:02.0000] <MattCampbell>
And why do I care? setuptools + py2exe is problematic

[14:33:03.0000] <MattCampbell>
yes, using html5lib in a Windows app

[14:33:04.0000] <jgraham>
Have fun :)

[14:34:00.0000] <jgraham>
Does distutils let you specify dependencies

[14:34:01.0000] <MattCampbell>
No... I guess that's the main advantage of setuptools

[14:34:02.0000] <MattCampbell>
0.90 is 5 months old, right?

[14:35:00.0000] <MattCampbell>
I just noticed the date on setup.py in that release

[14:35:01.0000] <jgraham>
Already? 

[14:35:02.0000] <MattCampbell>
January 17

[14:35:03.0000] <gsnedders>
jgraham: yes, already

[14:35:04.0000] <gsnedders>
jgraham: Obviously with your busy social life time flies

[14:36:00.0000] <jgraham>
gsnedders: My busy social life? I'm not th one having secret rendevousz ;)

[14:36:01.0000] <jgraham>
sp?

[14:36:02.0000] <jgraham>
Yeah I guess it is that old

[14:36:03.0000] <jgraham>
We should make a new release soon then

[14:36:04.0000] <gsnedders>
rendezvous

[14:38:00.0000] <MattCampbell>
Are there any current release-critical bugs?

[14:38:01.0000] <jgraham>
Well we need to fix some issues that prevent all the tests passing

[14:39:00.0000] <jgraham>
Nothing too serious, just little spec changes

[14:39:01.0000] <gsnedders>
Did you get all the table fostering tests passing?

[14:39:02.0000] <jgraham>
Yeah I think they pass

[14:39:03.0000] <gsnedders>
I had a look at that a while ago, and it wasn't easy with the current code which is quite different to the spec

[14:39:04.0000] <gsnedders>
(current when I looked)

[14:39:05.0000] <jgraham>
It is just a couple of things with <button> iirc

[14:39:06.0000] <gsnedders>
(this was last year)

[14:48:00.0000] <hober>
html5lib peeps, have you seen http://github.com/aredridel/html5 ?

[14:49:00.0000] <hober>
( impl. of html5 parsing algorithm in js for node.js )

[14:52:00.0000] <jgraham>
Seems to be using a way out of date testsuite

[14:52:01.0000] <jgraham>
Dunno quite how it managed that

[14:53:00.0000] <jgraham>
hober: Or, in long, no

[14:55:00.0000] <jgraham>
"""Excessive blink rates, such as those over 3 times a second can cause epileptic seizures in some users. To address this issue the user will configure the operating system blink rate for carets or selections to one which does not induce a seizure. User agents must support this method so that authors can determine the system blink rate and adjust their canvas blink rate to match that set in the operating system. """

[14:56:00.0000] <jgraham>
Doesn't it make more sense just to have a lower blink rate always

[14:56:01.0000] <jgraham>
?

[14:56:02.0000] <jgraham>
/me finds no evidence that the blink rate can be changed on OSX

[14:57:00.0000] <jgraham>
Oh it can be via the command line

[14:57:01.0000] <zcorpan_>
does a blinking cursor really cause epileptic seizures in some users?

[14:57:02.0000] <jgraham>
So I guess some a11y thing could expose that

[14:59:00.0000] <zcorpan_>
and yeah, why would you want the cursor blink over 3 times a second anyway

[14:59:01.0000] <TabAtkins>
So you can cause seizures.

[14:59:02.0000] <zcorpan_>
need to make sure the cursor is at least 200x200 pixels also

[15:00:00.0000] <TabAtkins>
Really, though, I'll just include a <video> of Porygon if I wanted to make people seizure.

[15:00:01.0000] <jgraham>
zcorpan_: There are at least reports on the internet of blinking cursors causing seziures

[15:00:02.0000] <jgraham>
although few seem to be first hand

[15:01:00.0000] <AryehGregor>
Maybe they're fatal.

[15:02:00.0000] <jgraham>
(but I agree it seems odd to specifically optimise for that case when there are so many other things that could cause seziures if you are photosensitive)

[15:07:00.0000] <jgraham>
http://www.amazon.com/Photosensitive-Epilepsy-Clinics-Developmental-Medicine/product-reviews/1898683026 appears to be an actual human claiming to have seziures triggered by blinking cursors. Although the reporter is a little zealous in his call to ban blinking in all forms

[15:09:00.0000] <jgraham>
http://www.justdamn.com/?p=1627

[15:09:01.0000] <TabAtkins>
No blinking rates between .1Hz and 1000Hz?

[15:09:02.0000] <TabAtkins>
If anyone's injured by something blinking on off for 5 seconds each, they're pretty much screwed anyway.

[15:10:00.0000] <jgraham>
Although that one could be hyperbole

[15:10:01.0000] <jgraham>
TabAtkins: That was part of the zealousness

[15:10:02.0000] <jgraham>
I have never seen anyone else worry about rates < 2Hz

[15:11:00.0000] <TabAtkins>
I believe the blog post you just linked to is hyperbole.

[15:11:01.0000] <zcorpan_>
the range .1Hz .. 1000Hz probably excludes all monitors

[15:12:00.0000] <TabAtkins>
I'm pretty certain it does.  Most monitors run between 50 and 100 Hz.

[15:14:00.0000] <AryehGregor>
/me is fairly sure 50 and 100 lie between .1 and 1000

[15:15:00.0000] <zcorpan_>
i read about a new type of monitor being used for read pads that don't blink at all

[15:15:01.0000] <jgraham>
3-60 says http://www.epilepsy.org.uk/info/photosensitive/intro

[15:15:02.0000] <jgraham>
which I tend to trust more than some random guy on amazon

[15:16:00.0000] <zcorpan_>
instead it moved around some ink-like stuff and didn't require any battery to show the image so long as it doesn't change

[15:16:01.0000] <AryehGregor>
It's funny how some people in the HTMLWG seem to all think that the WHATWG is some nonentity that will go away if the W3C ignores it.

[15:22:00.0000] <AryehGregor>
Random note: whoever decided that the Nexus One should make a noise by default every time they received an e-mail, obviously didn't subscribe to many mailing lists.

[15:22:01.0000] <jgraham>
Hmm, I don't think the drawFocusRing(x,y,w,h) works very well

[15:23:00.0000] <jgraham>
Although I might be wring

[15:23:01.0000] <jgraham>
*wrong

[15:23:02.0000] <jgraham>
Because the algorithm doesn't convert w,h accoring to the current transform matrix

[15:24:00.0000] <jgraham>
(which makes sense if the result has to be a rectangle)

[15:24:01.0000] <jgraham>
(but means that the area defined by the rectangle might not cover what you expect)

[15:26:00.0000] <jgraham>
But presumably the UA can calculate the best-fitting rectangle from the path anyway, so I don't see why the author has to supply it

[15:28:00.0000] <jgraham>
I assume there is some reason it would be bad (perf?) to make drawCustomRing a callback that is called if the UA needs to draw a custom focus ring?

[15:30:00.0000] <jgraham>
/me wonders what "editable fallback content" means

[15:32:00.0000] <zcorpan_>
<canvas contenteditable>?

[15:32:01.0000] <jgraham>
Maybe

[15:32:02.0000] <zcorpan_>
or <canvas><textarea>

[15:32:03.0000] <jgraham>
It feels like there is some design assumption hidden in that phrase

[15:33:00.0000] <jgraham>
that is not explicitly stated

[15:33:01.0000] <zcorpan_>
i guess it's 'bespin made accessible'

[15:33:02.0000] <zcorpan_>
though my guess is as good as anyone's

[15:35:00.0000] <jgraham>
Yeah, this whole set of changes reads like a reaction to bespin

[15:35:01.0000] <jgraham>
But bespin deliberately avoided also using the DOM for content, at least initially

[15:35:02.0000] <jgraham>
Did the reasons for that evaporate?

[15:37:00.0000] <lifechamp>
Flash HTML5 compiler coming? How about this for now? http://csharksgames.blogspot.com/2009/07/jangaroo-html-5-killer-as3-to-js.html

[15:37:01.0000] <jgraham>
In the nicest posible way, this is not camera ready spec text

[15:38:00.0000] <jgraham>
I am finding it quite hard to follow

[15:40:00.0000] <jgraham>
Is there enough information in the text API to make computing the height/width of characters reasonable?

[15:42:00.0000] <jgraham>
I think the setCaretSelectionRect stuff doesn't make much sense without measureText providing all the required information

[15:42:01.0000] <Hixie>
jgraham: it has more fundamental problems, like nobody will use the selection part of the API since it doesn't get them anything

[15:42:02.0000] <Hixie>
jgraham: (other than accessibility in ATs, which they can't test, so it'll always be wrong)

[15:42:03.0000] <jgraham>
Hixie: Yes, that is also a problem

[15:43:00.0000] <Hixie>
it's better than usemap="" though!

[15:44:00.0000] <jgraham>
I don't see that we need to block on this stuff; it addresses problems that only a small percentage of apps will encounter

[15:44:01.0000] <jgraham>
I expect violent disagreement on that point though

[15:46:00.0000] <jgraham>
Why does setCaretSelectionRect return true?

[15:48:00.0000] <jgraham>
Actually I'm lost

[15:48:01.0000] <jgraham>
How does one know the selection is a rectangle?

[15:48:02.0000] <Hixie>
it won't be

[15:48:03.0000] <Hixie>
i'm guessing the rectangle is the zoom rectangle

[15:49:00.0000] <jgraham>
Yeah that makes sense

[15:49:01.0000] <jgraham>
But I don't think the spec actually says that

[15:51:00.0000] <jgraham>
The caretBlinkRate thing specifically notes that the default is 2Hz

[15:52:00.0000] <jgraham>
(on typical operating systems)

[15:52:01.0000] <jgraham>
Given the figures quoted earlier, I wonder why it is needed

[15:53:00.0000] <jgraham>
(actually I still wonder why we care so much about the specific case of carets when so many other things could cause similar but worse problems)

[15:54:00.0000] <jgraham>
(since they won't typically be at 2Hz)

[15:55:00.0000] <jgraham>
I guess I will send this feedback to the list tommorow, so if I am totally wrong about anything it would be nice to let me know

[16:10:00.0000] <MikeSmith>
http://dev.opera.com/articles/view/css3-show-and-hide/ is nice

[16:10:01.0000] <MikeSmith>
http://devfiles.myopera.com/articles/1861/css-menu-ex4.html

[16:11:00.0000] <zcorpan_>
hmm, basing it on :focus seems weird

[16:12:00.0000] <zcorpan_>
:target seems better if you want css-only

[16:13:00.0000] <MikeSmith>
dunno about the details, but for one thing, it seems like this could be used to emulate support for the display element

[16:13:01.0000] <MikeSmith>
until such time as we get native implementations

[16:13:02.0000] <TabAtkins>
Yeah, you really don't want :focus - it'll rehide or reshow things as focus moves, which is weird.

[16:13:03.0000] <zcorpan_>
you mean details element?

[16:13:04.0000] <MikeSmith>
zcorpan_: erp, yeah

[16:13:05.0000] <MikeSmith>
details

[16:13:06.0000] <TabAtkins>
/me still likes using the hidden :checked hack.

[16:14:00.0000] <MikeSmith>
would make a good html5doctor article maybe

[16:14:01.0000] <MikeSmith>
/me looks around for some HTML5 doctors in the house

[16:14:02.0000] <zcorpan_>
as TabAtkins says, :focus doesn't work except for a minimal demo page

[16:14:03.0000] <zcorpan_>
the article acks this at the end

[16:15:00.0000] <MikeSmith>
http://dev.opera.com/forums/topic/600712?t=1276554067&page=1#comment5792922

[16:15:01.0000] <MikeSmith>
yeah

[16:15:02.0000] <zcorpan_>
an infobox instead of a menu would be more appropriate for showcasing :focus

[16:16:00.0000] <zcorpan_>
but maybe even that would be annoying

[16:38:00.0000] <zcorpan_>
does safari not fire any events for the arrow keys?


2010-06-15
[17:03:00.0000] <TabAtkins>
zcorpan_: Looks like not.  I was just trying to use your demo on Chrome, and it doesn't do anything for the arrow keys.

[17:06:00.0000] <zcorpan_>
TabAtkins: so that means canvas games don't work in safari/chrome?

[17:06:01.0000] <TabAtkins>
Presumably!

[17:06:02.0000] <zcorpan_>
awesome

[17:08:00.0000] <MikeSmith>
http://www.oipf.tv/docs/Release1/Release1_1/OIPF-T1-R1-Specification-Volume-5-Declarative-Application-Environment-V1_1-2009-10-08.pdf

[17:10:00.0000] <MikeSmith>
peruse the table of contents there

[17:11:00.0000] <TabAtkins>
zcorpan_: I'm hearing that we intentionally only support keyup and keydown, not keypress.

[17:11:01.0000] <TabAtkins>
And just confirmed that swapping to onkeydown works as expected.

[17:12:00.0000] <zcorpan_>
onkeydown doesn't seem to work in safari 5

[17:13:00.0000] <TabAtkins>
...

[17:13:01.0000] <TabAtkins>
Well, it works in Chrome5 at least.

[17:13:02.0000] <zcorpan_>
yep, got it working in chrome

[17:14:00.0000] <zcorpan_>
so why not keypress?

[17:14:01.0000] <zcorpan_>
ie compat thing?

[17:15:00.0000] <TabAtkins>
No clue.  I'm asking right now, but via a fairly async method, so I won't find out for a bit.

[17:15:01.0000] <zcorpan_>
k

[17:15:02.0000] <zcorpan_>
i'll get some sleep meanwhile, but do share when you know more :)

[17:16:00.0000] <TabAtkins>
Will do.

[17:16:01.0000] <zcorpan_>
cheers

[17:33:00.0000] <TabAtkins>
zcorpan: Got some explanation.  We try to do some union of IE and Gecko key events.  For keypress specifically, we only fire for "character generating" keys (this matches IE).

[20:04:00.0000] <MikeSmith>
http://lists.w3.org/Archives/Public/public-xg-audio/2010Jun/0010.html

[20:04:01.0000] <MikeSmith>
http://chromium.googlecode.com/svn/trunk/samples/audio/specification/specification.html

[20:04:02.0000] <MikeSmith>
demos: http://chromium.googlecode.com/svn/trunk/samples/audio/index.html

[20:05:00.0000] <MikeSmith>
joining the Audio Incubator Group would be a good way to get involved in discussion

[20:06:00.0000] <MikeSmith>
although I guess you can also just subscribe to the mailing list:

[20:06:01.0000] <MikeSmith>
http://lists.w3.org/Archives/Public/public-xg-audio/

[20:13:00.0000] <MikeSmith>
so the new HTML5 tokenizer in Webkit is 5% faster on benchmarks than the old HTML parser

[20:13:01.0000] <MikeSmith>
https://lists.webkit.org/pipermail/webkit-dev/2010-June/013244.html

[20:15:00.0000] <MikeSmith>
and from other discussions in that thread, it seems the new code is much more amenable to further optimizations than the old everybodys-afraid-to-touch-for-fear-of-regressions code

[20:16:00.0000] <MikeSmith>
so that eventually there will be further optimizations that could make it quite significantly faster than the old code

[20:25:00.0000] <othermaciej>
that's the old tree builder with the new HTML5 tokenizer

[20:25:01.0000] <othermaciej>
I am sure there is plenty more optimization juice in there

[20:31:00.0000] <roc>
oiks

[20:35:00.0000] <roc>
that audio API is somewhat ... complex

[20:52:00.0000] <MikeSmith>
othermaciej: fwiw, I personally wish we could put a complete moratorium on use of the word "objection"

[20:52:01.0000] <MikeSmith>
especially when it's preceded by the word "formal"

[20:52:02.0000] <othermaciej>
heh :-)

[20:56:00.0000] <MikeSmith>
I think it could be argued that the use of the word "objection", instead of ensuring that something doesn't get overlooked, instead just causes many people to write it off as an attempt to fast-track a disagreement to becoming an Issue

[20:57:00.0000] <MikeSmith>
instead of a good-faith attempt to actual defend a position on basis of technical merit

[20:57:01.0000] <MikeSmith>
*actually defend

[21:00:00.0000] <MikeSmith>
I think I the Advanced Dungeons and Dragons Dungeonmaster's Guide actually puts a limit on how many times a player can invoke the W3C formal-objection magic

[21:01:00.0000] <MikeSmith>
a quota

[21:01:01.0000] <roc>
each formal objection ages you by five years

[21:02:00.0000] <MikeSmith>
I think for this case the rules also require that the DM can't reveal to the player what the quota limit is

[21:02:01.0000] <MikeSmith>
instead, when the player reaches the limit, a big fricking lightning bolt shoots down and obliterates them completely

[21:02:02.0000] <MikeSmith>
and game play moves on from there

[21:03:00.0000] <MikeSmith>
with the other players learning a lesson from that

[21:25:00.0000] <MikeSmith>
I wonder how long it takes to implement an HTML5 tree builder

[21:25:01.0000] <MikeSmith>
relative to implementing a tokenizer

[21:26:00.0000] <MikeSmith>
seems like it only took 3 weeks or so from Adam and Eric to land the new tokenizer code

[21:26:01.0000] <MikeSmith>
or I guess maybe they had gotten started well before I noticed

[21:32:00.0000] <MikeSmith>
"Why do you persist in this game?"

[21:32:01.0000] <MikeSmith>
a wonderful existential question

[22:02:00.0000] <MikeSmith>
well, this is interesting

[22:02:01.0000] <MikeSmith>
I found a picture of the guy who's responsible for "scheduled maintenance" at Twitter

[22:03:00.0000] <MikeSmith>
http://a.gd/bad5df

[22:03:01.0000] <miketaylr>
:D

[22:03:02.0000] <MikeSmith>
guy looks of familiar

[22:04:00.0000] <MikeSmith>
*kind of

[22:05:00.0000] <MikeSmith>
perhaps that mobile handset he's sporting is symbolic of the cutting-edge nature of the technology they're using on the backend

[22:06:00.0000] <MikeSmith>
I asked somebody at Twitter why they hai

[22:06:01.0000] <MikeSmith>
I asked somebody at Twitter why they hired that guy

[22:06:02.0000] <MikeSmith>
and they said, "He interviewed well."

[22:06:03.0000] <MikeSmith>
not sure what they meant

[22:08:00.0000] <MikeSmith>
but I somehow get a sense that dude would be pretty good at putting forward a "formal objection" or two

[22:08:01.0000] <MikeSmith>
were he involved in standards discussions

[22:10:00.0000] <MikeSmith>
anyway, dunno why I brought this up

[22:10:01.0000] <MikeSmith>
or why I persist in it

[22:10:02.0000] <MikeSmith>
to paraphrase another great thinker:

[22:10:03.0000] <MikeSmith>
"To persist or not to persist... That is the question."

[22:11:00.0000] <othermaciej>
MikeSmith: I think they seriously have been working on it only a few weeks - they are both very fast programmers

[22:14:00.0000] <MikeSmith>
othermaciej: great to hear, then

[22:14:01.0000] <MikeSmith>
so I reckon they can make similar rapid progress on the tree builder

[22:15:00.0000] <MikeSmith>
nice to have stuff to look forward to, though

[22:15:01.0000] <MikeSmith>
it's going to be kind of less motivating once everything gets implemented

[22:15:02.0000] <MikeSmith>
as far as HTML5

[22:16:00.0000] <MikeSmith>
but I guess we will have other new stuff coming along by then

[22:16:01.0000] <Hixie>
HTML5 is old hat

[22:16:02.0000] <Hixie>
already

[22:16:03.0000] <Hixie>
e.g. <device>

[22:17:00.0000] <MikeSmith>
yeah, at least we don't have worry about anybody implementing that for a while ;)

[22:18:00.0000] <MikeSmith>
/me heads off to lunch

[23:58:00.0000] <hsivonen>
MikeSmith: the HTML5 parser in Gecko is also significantly faster than the old parser. However, the HTML parser is a tiny part of the total pageload time, so making the tiny part smaller doesn't make a big difference to the overall perf

[00:00:00.0000] <hsivonen>
(and there's still room for perf improvements in the tokenizer in Gecko)

[00:04:00.0000] <MikeSmith>
hsivonen: yeah, I knew it didn't make any big difference in total pageload times

[00:05:00.0000] <MikeSmith>
just that part of the sentiment expressed in the discussion on the webkit-dev list was that it at least should not be any slower than the older parser

[00:06:00.0000] <hsivonen>
Yeah, it would be surprising if a fresh implementation of the spec were slower than an implementation that evolved over time

[01:23:00.0000] <annevk>
"100% of Task Force members voting on this recommendation via WBS voted to support it." ... that it's only a little over 20% of the Task Force doesn't seem to matter

[01:24:00.0000] <annevk>
What I mostly don't get though is this caret proposal. Do they still consider Bespin a valid example for <canvas> usage?

[01:38:00.0000] <Slaanesh>
http://twitter.com/fantasai/status/16181517781 :)

[01:39:00.0000] <annevk>
Seems the CSS WG is getting even better at not defining CSS... :/

[01:54:00.0000] <Slaanesh>
The minutes from the 2008 GeoPriv BOF shouldn't be surprising to me at this point, but I'm still surprised how people don't understand users

[02:19:00.0000] <annevk>
fwiw, CORS now has http://dev.w3.org/2006/waf/access-control/#http-access-control-expose-headers

[02:20:00.0000] <annevk>
I think the only thing that is stopping it from moving on is some coherent explanation of how you should not run untrusted script in your own origin and then let it talk cross-origin

[02:26:00.0000] <MikeSmith>
cool

[02:27:00.0000] <MikeSmith>
the CORS adventure has been quite an odyssey 

[02:28:00.0000] <MikeSmith>
epic poems could be written about it

[02:29:00.0000] <annevk>
a meeting was had / a proposal was made / done it seemed / but three more years await

[02:29:01.0000] <annevk>
now the rest of it...

[02:29:02.0000] <MikeSmith>
heh

[02:29:03.0000] <MikeSmith>
brilliant

[02:30:00.0000] <MikeSmith>
take a break from spec writing and focus on the poetry for a while

[02:30:01.0000] <MikeSmith>
you may be on to something

[02:30:02.0000] <nessy1>
I hope the media fragment journey won't be such an odyssey

[02:30:03.0000] <MikeSmith>
tragic comedy

[02:31:00.0000] <MikeSmith>
is the main form of narrative we often seem to find ourselves in

[02:31:01.0000] <MikeSmith>
as far as work on technology standards

[02:31:02.0000] <jgraham>
Yeah, and I was told it would be erotic fiction

[02:32:00.0000] <jgraham>
Imagine my disappointment

[02:32:01.0000] <MikeSmith>
heh

[02:32:02.0000] <nessy1>
lol

[02:33:00.0000] <annevk>
/me is somewhat tempted to point out to Roy: "Didn't he just do that?"

[02:34:00.0000] <jgraham>
No point in getting in pissing matches with trolls

[02:34:01.0000] <annevk>
yeah, deleted the email

[02:34:02.0000] <annevk>
it's the thought that counts, I guess :p

[02:36:00.0000] <MikeSmith>
I sometimes find myself wanting to just put a plague on both the houses

[02:37:00.0000] <jgraham>
MikeSmith: Isn't that just your general God complex playing up again?

[02:38:00.0000] <MikeSmith>
there's clearly no god in this universe

[02:38:01.0000] <MikeSmith>
only a demiurge

[02:38:02.0000] <MikeSmith>
and I was thinking more in terms of what happened to Mercutio

[02:38:03.0000] <jgraham>
In any case I think a plauge of hotly-worded-emails is the modern day equivalent of frogs, at least

[02:39:00.0000] <annevk>
http://www.w3.org/2000/09/dbwg/details?group=44977&public=1 -- pretty much all W3C old-timers

[02:39:01.0000] <MikeSmith>
hotly worded e-mails are like the nurse in the play

[02:39:02.0000] <MikeSmith>
comic relief

[02:40:00.0000] <jgraham>
/me doesn't remember what happened to Mercutio, apart from the fact that he ended up dead 

[02:40:01.0000] <MikeSmith>
or like Quentin Compson

[02:41:00.0000] <MikeSmith>
er, I mean Benji

[02:41:01.0000] <jgraham>
I saw them as more like the priest. Well intentioned, but ultimatley responsible for the tragedy

[02:42:00.0000] <MikeSmith>
or make that Benjy

[02:42:01.0000] <MikeSmith>
not the dog

[02:42:02.0000] <MikeSmith>
although if the story were told by a dog that would be maybe even be better

[02:45:00.0000] <MikeSmith>
hsivonen: when you're back around, I wanted to ask about how to proceed with that MathML update for v.nu

[02:46:00.0000] <MikeSmith>
based on my understanding from the offlist discussion, I have http://qa-dev.w3.org:8888/ confoir

[02:46:01.0000] <MikeSmith>
*configured to only allow HTML markup in <mtext>

[02:46:02.0000] <MikeSmith>
for now at least

[02:47:00.0000] <MikeSmith>
unless/until whatever ends up being added to the HTML5 spec says otherwiser

[02:47:01.0000] <annevk>
MathML needs to be updated for this, not HTML5, I think

[02:47:02.0000] <Slaanesh>
So now Roy is claiming that thousands of CMSes use Content-Language?

[02:48:00.0000] <jgraham>
Hixie: yt?

[02:48:01.0000] <MikeSmith>
annevk: I have had some offlist discussion about that, and it was pointed out to me that the MathML3 spec says this: "The draft HTML5 Recommendation, [HTML5], contains details for how MathML is included in HTML5 (and XHTML5)."

[02:49:00.0000] <MikeSmith>
http://www.w3.org/TR/MathML3/chapter6.html#interf.html

[02:49:01.0000] <jgraham>
Hixie: never mind

[02:50:00.0000] <MikeSmith>
btw, everybody, MathML3 is back at LC, with a deadline for comments of July 1

[02:50:01.0000] <annevk>
MikeSmith, but not how HTML5 is included in MathML

[02:50:02.0000] <annevk>
(which is what the question is here)

[02:50:03.0000] <annevk>
Slaanesh, according to Roy browsers are hardly relevant

[02:51:00.0000] <Slaanesh>
I know that, but previously it's been vague suggestions that some unspecified CMS uses it

[02:51:01.0000] <Slaanesh>
Thousands is quite a step up

[02:52:00.0000] <Slaanesh>
I imagine you need a really loose definition of CMS for there to exist thousands of them

[02:52:01.0000] <MikeSmith>
annevk: true. perhaps I need to make an LC comment to ask for text to be amended to the that: "And how HTML content in included in MathML content within HTML documents."

[02:52:02.0000] <jgraham>
IS it actually only possible to change the selection in an element that has focus?

[02:53:00.0000] <Slaanesh>
jgraham: What if the element doesn't take focus?

[02:53:01.0000] <jgraham>
Slaanesh: My reason for asking is that http://lists.w3.org/Archives/Public/public-canvas-api/2010AprJun/att-0054/2dcontext10-June-7.html#focus-management has setCaretSelectionRect which assumes the element with a selection is also focused

[02:54:00.0000] <MikeSmith>
annevk: I do now notice that sentence, "In the lax schema profile, elements from non-MathML namespaces are allowed in token elements, but not in other elements."

[02:56:00.0000] <jgraham>
Oh wait, maybe the "or is not a document descendant of the element with whose context the method is associated" applies

[02:56:01.0000] <jgraham>
yeah

[02:58:00.0000] <jgraham>
Argh. Logic fail

[02:59:00.0000] <Slaanesh>
Hum

[03:00:00.0000] <annevk>
MikeSmith, hmm yeah, though note that if browsers are to follow the HTML5 rules it will also work in "MathML documents"

[03:00:01.0000] <annevk>
there's no real distinction between those things in code

[03:00:02.0000] <Slaanesh>
I'm not quite sure I understand the problem, but e.g. text can be selected and won't trigger focus events (but it does blur the previous focus owner, if any)

[03:01:00.0000] <annevk>
Slaanesh, that can happen

[03:01:01.0000] <jgraham>
Slaanesh: If I do "select all" does that necessarily have to focus the control where the text is being selected?

[03:01:02.0000] <annevk>
Slaanesh, well actually, the document would have focus in that case so something is triggered (iirc)

[03:01:03.0000] <jgraham>
(assuming I am implementing with DOM)

[03:02:00.0000] <jgraham>
s/do/implement/

[03:04:00.0000] <Slaanesh>
Besides, once you involve script they're definitely not connected, e.g. having a 'select text' button in the app UI

[03:04:01.0000] <jgraham>
Yes, exactly

[03:04:02.0000] <jgraham>
Nice example

[03:04:03.0000] <MikeSmith>
hmm, anybody been following work on this doc? - http://www.w3.org/TR/xml-proc-profiles/

[03:05:00.0000] <MikeSmith>
zcorpan_ maybe?

[03:05:01.0000] <MikeSmith>
see https://lists.webkit.org/pipermail/webkit-dev/2010-June/013250.html

[03:06:00.0000] <MikeSmith>
[[

[03:06:01.0000] <MikeSmith>
Basically, the profile would be much like what is described as the

[03:06:02.0000] <MikeSmith>
"minimum XML processor profile" except that reading the external

[03:06:03.0000] <MikeSmith>
subset would not be required.

[03:06:04.0000] <MikeSmith>
]]

[03:06:05.0000] <MikeSmith>
indeed

[03:07:00.0000] <MikeSmith>
on the face of it, it seems odd that the minimal profile would require reading external markup declarations

[03:08:00.0000] <MikeSmith>
hmm, there is actually also a "minimum XML processor profile" that's meant to be simpler than the "basic XML processor profile"

[03:09:00.0000] <MikeSmith>
but it still requires support for reading external markup declarations

[03:09:01.0000] <annevk>
really?

[03:09:02.0000] <annevk>
silly

[03:10:00.0000] <MikeSmith>
as far as I can see, those are the only profiles it permits

[03:10:01.0000] <MikeSmith>
so browsers currently can't conform to this spec at all

[03:11:00.0000] <zcorpan_>
MikeSmith: i haven't seen that

[03:11:01.0000] <MikeSmith>
OK

[03:11:02.0000] <MikeSmith>
well, seems like it would be worthwhile to comment on it

[03:11:03.0000] <MikeSmith>
comments to public-xml-processing-model-comments⊙wo

[03:12:00.0000] <zcorpan_>
can't find any mention of it in my public-xml-core-wg inbox

[03:12:01.0000] <MikeSmith>
I guess maybe Alex Milowski is in that group

[03:12:02.0000] <MikeSmith>
zcorpan_: because it's a different group

[03:12:03.0000] <MikeSmith>
http://www.w3.org/XML/Processing/

[03:13:00.0000] <MikeSmith>
not XML Core

[03:18:00.0000] <Lachy>
"I don't think the text should even appear in the HTML standard until it has been implemented with some consistency" -- Roy T. Fielding

[03:18:01.0000] <Lachy>
haha. How does he expect implementation to occur without a spec?

[03:19:00.0000] <zcorpan_>
it's certainly possible to implement something without a spec

[03:19:01.0000] <zcorpan_>
even with some consistency

[03:20:00.0000] <annevk>
if the spec is a mental image?

[03:20:01.0000] <Lachy>
well, yes, but it requires reverse engineering

[03:20:02.0000] <zcorpan_>
Lachy: yes. maybe that's what roy is advocating?

[03:32:00.0000] <annevk>
guess he missed the part of history where we were doing exactly that and felt miserable

[03:53:00.0000] <annevk>
oh god

[03:53:01.0000] <annevk>
confusion all over

[03:53:02.0000] <annevk>
http://lists.w3.org/Archives/Member/member-math/2010Jun/0069.html (W3C Member-only #$@)

[03:53:03.0000] <annevk>
not that it matters

[03:55:00.0000] <annevk>
http://www.w3.org/2010/06/09-xhtml-minutes.html -- 'sad news, steven VITAL (Visionary Technology in Library Solutions) is scrapping plans for integrating xforms so as to be "compatible" with HTML5' 'still trying to talk them into using XHTML M12n'

[03:57:00.0000] <jgraham>
The W3C should admonish people that post things to private lists that have no sensitive information in them

[03:57:01.0000] <annevk>
interesting news about XHTML M12n "please advocate amongst W3C members to vote and sign up for it -- need at least 10 votes in favor"

[03:59:00.0000] <annevk>
and WebApps has a new charter: http://www.w3.org/2010/webapps/charter/

[04:00:00.0000] <annevk>
there are some interesting details attached to it, but alas, all W3C Member-only #$@

[04:02:00.0000] <roc>
I wonder if I should bother digging out my password

[04:03:00.0000] <annevk>
roc, for WebApps WG you might want to, unless you're a member of the group, in which case you should gave gotten an email

[04:03:01.0000] <roc>
hmm

[04:03:02.0000] <roc>
is the charter public?

[04:04:00.0000] <annevk>
yes, should be

[04:04:01.0000] <annevk>
http://www.w3.org/2010/webapps/charter/,access says it is anyway

[04:04:02.0000] <annevk>
but considerations in drafting the charter are private

[04:06:00.0000] <annevk>
http://lists.w3.org/Archives/Member/member-webapps/2010AprJun/0022.html which forwards http://lists.w3.org/Archives/Member/w3c-ac-members/2010AprJun/0056.html (both W3C Member-only) has the details

[04:06:01.0000] <roc>
interesting

[04:06:02.0000] <roc>
thanks

[04:07:00.0000] <roc>
ooh, we should have helped to try to get Widgets kicked out

[04:07:01.0000] <roc>
oh well

[04:07:02.0000] <roc>
grr

[04:08:00.0000] <roc>
man, this stuff needs to be in the open

[04:09:00.0000] <roc>
it's not fair to have people secretly blocking stuff

[04:09:01.0000] <annevk>
uhuh

[04:09:02.0000] <roc>
thanks

[04:21:00.0000] <hsivonen>
is there a back story to go along with http://www.w3.org/2000/09/dbwg/details?group=44977&public=1 ?

[04:31:00.0000] <karlcow>
hsivonen: http://www.w3.org/2010/04/w3c-vision-public/wiki/Newstd

[04:31:01.0000] <karlcow>
http://www.w3.org/2010/Talks/ij-newstd-201005/#%281%29

[04:31:02.0000] <karlcow>
http://www.w3.org/2010/04/w3c-vision-public/wiki/Main_Page

[05:15:00.0000] <jgraham>
/me wonders if the usemap thing conflicts with the drawFocusRing changes

[05:27:00.0000] <hsivonen>
was the new vision group open to charter review at some point?

[05:28:00.0000] <annevk>
it's a special Task Force so I guess not

[05:28:01.0000] <hsivonen>
I see

[05:28:02.0000] <annevk>
I'm guessing it came out of the W3C getting a new CEO or some such

[05:29:00.0000] <annevk>
it's somewhat announced here: http://www.w3.org/QA/2010/06/the_mission_of_w3c.html

[05:29:01.0000] <annevk>
but not in a very concrete way

[05:29:02.0000] <annevk>
"we are building teams that will look at each of these topics"

[05:31:00.0000] <hsivonen>
I guess "Make W3C the best place for new standards work." semi-implies that it isn't already, which may not be an entirely wrong implication

[05:32:00.0000] <annevk>
right, but then putting mostly W3C old-timers in the "vision" Task Force does not give much hope for change

[05:33:00.0000] <hsivonen>
I find it interesting that Andy Updegrove is in the task force

[05:48:00.0000] <hsivonen>
I think the placement of "Impl." in http://www.w3.org/2010/Talks/ij-newstd-201005/#%288%29 doesn't reflect reality

[05:50:00.0000] <jgraham>
No, the problem with reality is that it doesn't make for convenient slides or slogans

[05:52:00.0000] <annevk>
yes, lets blame reality

[05:55:00.0000] <annevk>
hmm

[05:56:00.0000] <annevk>
maybe XHR can be moved to CR if the WebApps WG can make a decision to that effect

[05:57:00.0000] <annevk>
it would also mean people would have to agree with the decisions made so far, but given the lack of email so far I guess that will go okay (or not)

[05:58:00.0000] <jgraham>
I'm not blaming reality; I'm pointing out that complex things are often poorly understood because they are hard to communicate so people communicate simple fantasies instead

[06:03:00.0000] <zcorpan_>
Hixie: "Let origin be the ASCII serialization of the origin of the script that invoked the WebSocket() constructor, converted to ASCII lowercase." - was the 'of the script' part deliberate?

[06:13:00.0000] <annevk>
zcorpan_, I think so

[06:13:01.0000] <annevk>
zcorpan_, XMLHttpRequest does it differently

[06:14:00.0000] <annevk>
zcorpan_, but it matters in e.g. an iframe scenario where the iframes are of different domains but can communicate to document.domain

[06:15:00.0000] <zcorpan_>
why does it matter? if they can talk to each other, there's no security reason to get the origin from the script instead of from the constructor

[06:17:00.0000] <annevk>
what do you mean "from the constructor"?

[06:17:01.0000] <annevk>
like XMLHttpRequest does it?

[06:17:02.0000] <annevk>
iirc XMLHttpRequest is quite exceptional

[06:18:00.0000] <zcorpan_>
i mean from the window object of the constructor

[06:19:00.0000] <annevk>
it's not that simple, you can copy the interface object

[06:19:01.0000] <zcorpan_>
how?

[06:20:00.0000] <annevk>
win.X2 = win2.XMLHttpRequest; x = new X2;

[06:21:00.0000] <zcorpan_>
that's not a copy though, just a reference. it's the same object

[06:21:01.0000] <zcorpan_>
does that change which origin is used?

[06:21:02.0000] <annevk>
no

[06:22:00.0000] <jgraham>
Running eval in the context of othre windows is "fun"

[06:22:01.0000] <jgraham>
*other

[06:22:02.0000] <karlushi>
hsivonen, annevk, any suggestions are super welcome for the task force.

[06:22:03.0000] <annevk>
the other thing in case of XHR is that the base URL is also based on the original document

[06:23:00.0000] <annevk>
unless you use XHR in the context of workers

[06:23:01.0000] <karlushi>
I can also step down if you think it will be better for it. I can only explain my goals joining it under invitation.

[06:23:02.0000] <annevk>
karlushi, no worries :)

[06:23:03.0000] <zcorpan_>
annevk: ok, but let's ignore baseurl since websockets doesn't have it

[06:23:04.0000] <annevk>
I'd focus on seeing how the process can be changed to meet reality

[06:24:00.0000] <annevk>
i.e. implementations happen before CR (before REC even, see also patent policy)

[06:25:00.0000] <karlushi>
My goals are to push w3c toward a more flexible way of working. It's a bit hard right now for one to spontaneously start things at w3c. I have no idea if it will happen but at least there is a door open and a welcoming hand. So let's try.

[06:25:01.0000] <annevk>
zcorpan_, most other contexts like window.location and such come to the base URL by using the "first script" (or whatever we call it now)

[06:25:02.0000] <annevk>
zcorpan_, so it would make sense to also get the origin there

[06:25:03.0000] <annevk>
zcorpan_, rather than the different way XMLHttpRequest goes about things

[06:25:04.0000] <hsivonen>
karlushi: in addition to the point annevk made, a major bug I see in the Process, and a bug that is actively harming HTML5, is the way dissent is handled

[06:26:00.0000] <hsivonen>
karlushi: specifically, "can live with" and "weakest objections" are ways to get a race to the bottom in terms of technical excellence if non-implementors start objecting a lot

[06:28:00.0000] <karlushi>
hsivonen, here you express the issue. What process implementation, do you propose?

[06:29:00.0000] <hsivonen>
karlushi: that WG decisions should be technically the best that implementors agree to implement even if the decisions draws loud objections

[06:30:00.0000] <hsivonen>
karlushi: that is, framing decisions in terms of technical merit instead of framing them in terms of least objection

[06:33:00.0000] <karlushi>
hsivonen, hmm it doesn't seem testable either. "should", "technical merit", "agree to implement" push a bit further the issue. 

[06:33:01.0000] <karlushi>
Would it help if a controversial issue was raised to have test cases, and implemented code to finally make the balance (according to different class of products)? 

[06:33:02.0000] <karlushi>
s/was raised/when raised/

[06:34:00.0000] <jgraham>
I'm not convinced test cases help the class of problems we have had in the HTMLWG

[06:34:01.0000] <jgraham>
Except insofar as they can be used to demonstrate claims about existing behaviour

[06:35:00.0000] <karlushi>
jgraham, maybe that would be a good starting point. Taking the type of issues we had and see how practically we could solve them in a testable way. If possible. 

[06:35:01.0000] <karlushi>
test cases are half part of the story. Implemented code is  necessary too.

[06:36:00.0000] <jgraham>
For some issues (X would be better as a seperate document) it's hard to see what type of testcases you could even write

[06:36:01.0000] <hsivonen>
karlushi: like jgraham says, test cases only help establishing what the current state of implementation is

[06:37:00.0000] <jgraham>
(because the experiment you want to do is to have X in/out of the main document and see if the implementations are of higher quality in the long term, but you can't do that)

[06:37:01.0000] <hsivonen>
karlushi: basically, stuff like "remove X", "split out Y", "make a non-normative reference to Z" should not be taking a group's time

[06:37:02.0000] <hsivonen>
karlushi: and proposals to change technical substance should be decided on technical merit--not on who threates to complain the most

[06:38:00.0000] <karlushi>
jgraham, splitting out spec is an entire different issue which might be handled at the start and has nothing to do with technical merit that was the initial point of starting the discussion :)

[06:39:00.0000] <karlushi>
* technical merit of a feature

[06:39:01.0000] <karlushi>
* specs organization

[06:40:00.0000] <hsivonen>
karlushi: that could be partially solved by letting a participant make non-technical proposals (like splitting the spec) only after the participant has a track record of also contributing technically

[06:40:01.0000] <hsivonen>
(by sending in good review comments, implementing the spec, writing test cases, etc.)

[06:41:00.0000] <karlushi>
hsivonen, stackoverflow model. Good participants get more ways of participating. Getting credentials in a community. Could be interesting indeed.

[06:42:00.0000] <hsivonen>
karlushi: the basic problem is that working groups need to be open so that new people who contribute positively can come in

[06:42:01.0000] <hsivonen>
karlushi: but

[06:43:00.0000] <hsivonen>
karlushi: not so open that anyone can come in and start process trolling the group

[06:44:00.0000] <karlushi>
hsivonen, basically you are proposing: open read access, and progressive write access with positive contributions. Worth trying, would have to be modelled a bit to understand the possible issues. 

[06:45:00.0000] <jgraham>
karlushi: It does have technical ramifications if it makes implementing harder, or affects the accuracy of the spec (because split documents diverge, for example)

[06:45:01.0000] <hsivonen>
karlushi: yes. also, an important point: one should be able to suggest technical "write" things right away but shouldn't feel an entitlement of getting one's suggestions accepted right away

[06:47:00.0000] <hsivonen>
karlushi: in fact, a strong sense of entitlement is part of the problem

[06:47:01.0000] <hsivonen>
karlushi: when I got involved, I didn't get paid to work on a browser

[06:48:00.0000] <hsivonen>
karlushi: I made suggestions, but I didn't have a grandiose feeling of entitlement

[06:48:01.0000] <hsivonen>
karlushi: that is, I didn't think the WHATWG were required to do things my way

[06:50:00.0000] <zcorpan_>
/me has the same experience as hsivonen

[06:51:00.0000] <karlushi>
hsivonen, yep difficult issue. Even in free community (not paid and nothing to pay), I have seen people asking for things like if we were forced to do it. (example validators when the source code was available.)

[06:51:01.0000] <karlushi>
But I'm not sure this issue can be solved.

[06:52:00.0000] <karlushi>
I guess it is the "moral" stance of each individuals which drives them. Some people are respectful. Some less. :/ very hard issue.

[06:54:00.0000] <karlushi>
/me will point out Ian to this discussion and pushes further. Back to my daily work. Thanks hsivonen and jgraham 

[06:54:01.0000] <hsivonen>
karlushi: thanks

[06:55:00.0000] <hsivonen>
karlushi: and one solution is that chairs tell people who walk in with a strong sense of entitlement that they need to adjust their attitude 

[07:02:00.0000] <Slaanesh>
The "Public Accountability part of process" bullet point makes me wonder if that only applies to decision and not to rationale and discussion on private lists

[07:14:00.0000] <Marcosc>
/me notes roc "ooh, we should have helped to try to get Widgets kicked out", why would you do that? 

[07:16:00.0000] <Marcosc>
yeah, I thought so... 

[07:25:00.0000] <Slaanesh>
"For instance, it seems there may be resistance to finalizing CSS 2.1 out of concern it will not be maintained once a Rec."

[07:25:01.0000] <Slaanesh>
What is this about?

[07:30:00.0000] <hsivonen>
Slaanesh: where's the quote from?

[07:31:00.0000] <Slaanesh>
From the wiki page about the new vision thing

[07:35:00.0000] <MikeSmith>
somebody should please create a "Web platform" article at Wikipedia

[07:35:01.0000] <zcorpan_>
MikeSmith: just rename the "HTML5" article

[07:35:02.0000] <MikeSmith>
heh

[07:35:03.0000] <MikeSmith>
well, I was thinking something more along the lines of what hober has written up at http://edward.oconnor.cx/2009/05/what-the-web-platform-is

[07:38:00.0000] <MikeSmith>
and what hsivonen diagrammed at http://hsivonen.iki.fi/web-stack/

[07:39:00.0000] <karlushi>
http://en.wikipedia.org/wiki/World_Wide_Web

[07:39:01.0000] <karlushi>
maybe starts with a section here.

[07:40:00.0000] <MikeSmith>
yeah, that'd be one way to do it

[07:41:00.0000] <annevk>
seems there would be conflict of interest

[07:42:00.0000] <MikeSmith>
well, there would be if I were to do it at least

[07:43:00.0000] <MikeSmith>
annevk: or maybe I misunderstand what you meant

[07:43:01.0000] <annevk>
nope :)

[07:44:00.0000] <hsivonen>
annevk: the problem with wikipedia rules is that if you actually know a lot about a topic, you've probably crossed the conflict of interest line

[07:45:00.0000] <annevk>
no suggestion there

[07:50:00.0000] <Slaanesh>
hsivonen: You're also likely to run into the whole "need third party sources, preferably paper ones" issue :)

[07:51:00.0000] <hsivonen>
Slaanesh: my other writing is already authoritative enough to be cited on wikipedia, so it's worth trying to use my writing as a source for more :-)

[08:09:00.0000] <Lachy>
So it turns out that the iPhone 4 will be available unlocked in the UK, and will work on carriers worldwide. :-)

[08:10:00.0000] <Lachy>
I'm considering getting one. Though I want to evaluate some other alternatives before I commit.

[08:10:01.0000] <hsivonen>
Lachy: don't you want a phone that runs Opera Mobile 10.x?

[08:15:00.0000] <annevk>
so I finally got time to look into CSSOM again

[08:16:00.0000] <annevk>
the way getComputedStyle works in Gecko is indeed as Boris described (unsurprisingly)

[08:17:00.0000] <annevk>
it takes the style rules of the document of the window on which it is invoked and then applies those to the tree/fragment the element is from (irrespective of whether it is in fact the same document or not) and gives you an object with answerss

[08:17:01.0000] <Lachy>
hsivonen, I will look at the available Android phones first

[08:17:02.0000] <annevk>
whenever something changes that would affect the computations the object is updated

[08:18:00.0000] <Lachy>
I believe there are some android phones on the market with Opera 10 on them. I'm not sure which though

[08:18:01.0000] <hsivonen>
annevk: what do other browsers do if the owner doc and the window don't match?

[08:18:02.0000] <hsivonen>
Lachy: is Opera Mobile available for Android as a self-install download?

[08:18:03.0000] <annevk>
usually assume there are no style rules

[08:18:04.0000] <annevk>
but apparently developers like what Gecko does

[08:19:00.0000] <annevk>
some of them anyway

[08:19:01.0000] <annevk>
and I'm happy either way

[08:20:00.0000] <annevk>
I wanna discuss it with Rune who is mainly responsible for layout in Opera, but I doubt he feels strongly either way

[08:22:00.0000] <Lachy>
hsivonen, I don't know.

[08:22:01.0000] <Lachy>
I can find out

[08:24:00.0000] <Lachy>
hsivonen, gsnedders tells me that it's only avaialble on Android phones that are shipped with it pre-installed.

[08:24:01.0000] <Lachy>
I believe Mini is available though

[08:25:00.0000] <gsnedders>
Mini in beta

[08:35:00.0000] <Lachy>
hsivonen, apparently there aren't any android devices with Mobile either. Only mini.

[09:04:00.0000] <annevk>
hmm, does that new speech input proposal lack a link or is it just me?

[09:04:01.0000] <annevk>
or is the link in the old proposal email?

[10:47:00.0000] <TabAtkins>
I... have no idea what that vision document is going on about re: css 2.1.  If anyone is resisting finalizing 2.1, it's certainly news to me, and I expect to the rest of the CSSWG.

[10:51:00.0000] <TabAtkins>
annevk: That status from fantasai was a joke.  ^_^

[10:58:00.0000] <hsivonen>
I hadn't realized tere was an ISSUE about the title of the ARIA section

[10:59:00.0000] <TabAtkins>
Ffs, really?

[10:59:01.0000] <hsivonen>
having the bikesheds as ISSUEs is kinda sad

[10:59:02.0000] <TabAtkins>
/me really isn't sure of proper policy in starting a sentence with an acronym that would be written lowercase...

[10:59:03.0000] <TabAtkins>
It's all about entitlement.  Sigh.

[11:00:00.0000] <hsivonen>
TabAtkins: http://www.w3.org/html/wg/tracker/issues/109

[11:02:00.0000] <TabAtkins>
Double sigh.

[11:02:01.0000] <TabAtkins>
What's the w3c process for "this is frivolous bullshit, and no one cares"?

[11:02:02.0000] <AryehGregor>
I think W3C process isn't relevant, only HTMLWG process?

[11:03:00.0000] <TabAtkins>
Either one.

[11:03:01.0000] <AryehGregor>
Process in some WGs would be to say "shut up", or have a quick vote.  Process in the HTMLWG appears to be the usual Decision Policy.

[11:07:00.0000] <hsivonen>
AryehGregor: "usual"? did you meand *un*usual?

[11:07:01.0000] <hsivonen>
s/d//

[11:08:00.0000] <AryehGregor>
Usual as in the same policy used to handle all other complaints in the HTMLWG, i.e., no difference between frivolous and non-frivolous.

[11:09:00.0000] <hsivonen>
now I feel like writing a counter-proposal to the ASCII ref ISSUE

[13:04:00.0000] <TabAtkins>
Hixie: Is there anything in particular preventing the WHATWG from adopting a similar patent policy to the W3C?

[13:12:00.0000] <othermaciej>
TabAtkins: it's definitely possible, but since there is no real concept of joining the WHATWG (you just subscribe to the list), so you couldn't require people or organizations to agree to it to join

[13:13:00.0000] <othermaciej>
I think it would be good for WHATWG to have a patent policy

[13:13:01.0000] <karlushi>
TabAtkins, you need a backup [money] when shit hits the fan. It can be super costly. 

[13:15:00.0000] <AryehGregor>
othermaciej, you could require the employers of Steering Committee members to agree to it.  And ask other parties nicely if they'd like to, I guess.

[13:17:00.0000] <othermaciej>
AryehGregor: that's certainly one option - though that would be a smaller set of IPR commitments than the HTML WG has

[13:17:01.0000] <AryehGregor>
Yes, I think considerably smaller.

[13:17:02.0000] <AryehGregor>
Adobe and Microsoft come to mind.

[13:17:03.0000] <othermaciej>
IBM also has a large patent portfolio

[13:17:04.0000] <AryehGregor>
Right.

[13:18:00.0000] <othermaciej>
besides large corporation patent portfolios, there is also the theoretical vulnerability of a patent troll deliberately proposing a feature on which they have a submarine patent

[13:19:00.0000] <othermaciej>
but the W3C process is vulnerable to that too, if the patent troll simply fails to join the WG

[13:19:01.0000] <AryehGregor>
That could happen just as well in the HTMLWG.  Right.

[13:19:02.0000] <othermaciej>
so the main advantage is getting patent commitments from some of the relevant industry players

[13:19:03.0000] <Lachy>
getting all the lawyers and management from so many companies to agree to a whatwg patent policy would be a nightmare to deal with. Unfortunately, it wouldn't be as simple as just reusing the W3C patent policy either

[13:30:00.0000] <othermaciej>
I suspect you could get the three copyright holders plus google to agree, but someone would have to do the work

[13:30:01.0000] <othermaciej>
unfortunately, you probably have to be a lawyer to do anything useful on that front

[13:31:00.0000] <AryehGregor>
Are any HTMLWG members likely to sue anyone else for implementing an open standard anyway?

[13:31:01.0000] <AryehGregor>
I suppose so, if it came to it.

[13:32:00.0000] <AryehGregor>
/me reads the W3C patent policy

[13:32:01.0000] <AryehGregor>
s/reads/skims/

[13:33:00.0000] <TabAtkins>
Yeah, I'm sure that you could get at least that much commitment.  Hopefully one could get MS to agree too.

[13:35:00.0000] <AryehGregor>
Okay, so does this policy only require licensing when the spec hits REC, or before that too?  I'm not clear on that point.

[13:35:01.0000] <AryehGregor>
Also, what if a party holds a patent they didn't disclose?  Is it auto-licensed?

[13:37:00.0000] <TabAtkins>
I *think* so, yes.  You disclose a patent specifically so you *don't* want to license it.

[14:38:00.0000] <TabAtkins>
My hobby: annoying my wife with strange unicode smilies.

[14:41:00.0000] <othermaciej>
AryehGregor: the W3C Patent Policy is a little convoluted, but here is basically how it works:

[14:42:00.0000] <othermaciej>
(1) For any First Public Working Draft or Last Call Working Draft, there is a call for exclusion, which is an opportunity for WG Members to list specific patents or patent claims for which they will *not* agree to grant royalty-free licenses. Anything not excluded is assumed to be licensed.

[14:43:00.0000] <othermaciej>
(2) Technically, the license does not actually go into effect until the spec hits REC, but you do have to license anything you didn't exclude by LC. That's part of why specs have to go back to Working Draft if they have substantial changes made in CR - those changes could possibly entrain new IP.

[14:43:01.0000] <othermaciej>
(3) You can put certain limits on the RF license, such as basing it on the other party cross-licensing any essential claims they have on the spec.

[14:44:00.0000] <othermaciej>
You can read the whole thing if you want the gory details.

[14:44:01.0000] <AryehGregor>
So if someone actually wanted to sue someone over HTML5, they could do so anyway until 2022 or whatever.

[14:44:02.0000] <AryehGregor>
By which point most possible patents will probably be expired anyway.

[14:45:00.0000] <AryehGregor>
(since HTML5 qualifies as prior art for any essential claims, I'd think)

[14:46:00.0000] <jgraham>
Sigh, WebIDl is officially without an editor

[14:48:00.0000] <AryehGregor>
And people in the HTMLWG want to get rid of their existing highly active editor.  :)

[14:48:01.0000] <TabAtkins>
/me is very, very glad he started binning HTMLWG when he did...

[14:49:00.0000] <AryehGregor>
/me is mildly interested to see what happens.

[14:49:01.0000] <TabAtkins>
I'll let you track it and summarize for me later.

[14:51:00.0000] <AryehGregor>
In short, Ian complained in rather undiplomatic terms about the chairs' decision-making, and various arguments ensued, with some flames directed in various directions from the people you'd expect, and calls to remove Hixie from the people you'd expect.  Still ongoing.

[14:51:01.0000] <jgraham>
Yeah, the WebIDL thing is a real pain because WebIDL is a really important spec and it blocks other specs

[14:51:02.0000] <jgraham>
And of course almost no one is qualified to actually edit it

[14:51:03.0000] <TabAtkins>
Fun.

[14:52:00.0000] <othermaciej>
TabAtkins: I did reply to one of your direct questions, you may want to consider reading that post (though there were some flames in the follow-up)

[14:53:00.0000] <TabAtkins>
othermaciej: I'll go find it.  My HTMLWG email is just getting auto-read-and-archived.

[14:58:00.0000] <daedb>
The spec for ASCII actually costs money? :O

[14:58:01.0000] <TabAtkins>
Yup.

[14:58:02.0000] <AryehGregor>
There's actually a spec for ASCII?

[14:59:00.0000] <AryehGregor>
Does it need one?

[14:59:01.0000] <daedb>
Apparently

[14:59:02.0000] <AryehGregor>
Can't you just read the Wikipedia article for the bits you haven't bothered to memorize?

[14:59:03.0000] <AryehGregor>
Do other W3C specs reference anything for ASCII?

[14:59:04.0000] <TabAtkins>
othermaciej: What's the title of the thread you responded to me in?

[15:00:00.0000] <TabAtkins>
Oh, I see, it's a top-level thread.

[15:00:01.0000] <TabAtkins>
nm

[15:02:00.0000] <AryehGregor>
TabAtkins, you should have it go to your inbox if it contains your name.

[15:03:00.0000] <AryehGregor>
/me has some Wikimedia lists set like that, so he only gets the stuff that contains his name or the name of other interesting people

[15:03:01.0000] <TabAtkins>
AryehGregor: Shrug, I can depend on it getting mentioned to me at some point if I don't respond directly.

[15:06:00.0000] <TabAtkins>
In other words, I'm willing to put the cost on the HTMLWG participants to make sure I get relevant email, because I'm purposely not caring about that list at the moment.

[15:23:00.0000] <jgraham>
hsivonen: I don't really see why .pdf.zip is not considered acceptable

[15:27:00.0000] <Lachy>
for a spec that no-one actually needs to read, the inconvenience of a zipped PDF seems negligible

[15:28:00.0000] <Lachy>
but the fact that there are other non-zipped alternatives available, it's preferable to go with those anyway

[15:30:00.0000] <jgraham>
Well sure

[15:31:00.0000] <jgraham>
But I wouldn't exclude something just because it was zipped

[15:31:01.0000] <jgraham>
If it was the actual useful reference

[15:37:00.0000] <TabAtkins>
othermaciej: Sorry for my email.  :\

[15:39:00.0000] <TabAtkins>
Urgh, HTMLWG stresses me out.

[15:42:00.0000] <gsnedders>
Do asteroids have to come from within the solar-system?

[15:42:01.0000] <TabAtkins>
No.

[15:42:02.0000] <gsnedders>
That's what I thought.

[15:43:00.0000] <TabAtkins>
There aren't very many rocks floating in intersteller space anyway, but you can call them asteroids if you find them.

[15:43:01.0000] <gsnedders>
That's what I thought

[15:43:02.0000] <gsnedders>
Yay for signs being wrong is museums.

[15:43:03.0000] <TabAtkins>
Does it say what interstellar rocks are called?

[15:45:00.0000] <karlushi>
gsnedders, the space is mostly empty. It is just that the asteroids are most likely to come from the solar system. They do not have to. Pave the solar system cowpath ;)

[15:46:00.0000] <gsnedders>
karlushi: :)

[15:46:01.0000] <TabAtkins>
Presumably the pathmakers are spherical frictionless cows.

[15:46:02.0000] <gsnedders>
TabAtkins: No, it just categorized rocks hitting the earth into two categories: asteroids from within the solar system, and comets, basically.

[15:47:00.0000] <gsnedders>
(or rather entering Earth's atmosphere)

[15:47:01.0000] <TabAtkins>
That's totally wrong.

[15:47:02.0000] <gsnedders>
Maybe the German was better, but I think it said the same

[15:47:03.0000] <gsnedders>
TabAtkins: Indeed, that's what I thought.

[15:47:04.0000] <TabAtkins>
Comets are quite distinct from asteroids.

[15:47:05.0000] <gsnedders>
I find it sad when my physics is better than a sign in a museum

[15:48:00.0000] <karlushi>
TabAtkins, indeed. They are a lot cuter, with long eyelashes

[15:48:01.0000] <TabAtkins>
Comets are the females of the subplanetary ecosystem?

[15:48:02.0000] <jgraham>
This isn't physics is it just naming things

[15:49:00.0000] <gsnedders>
When, physics related terminology

[15:49:01.0000] <karlushi>
In French at least. Un astéroïde, une comète.

[15:49:02.0000] <gsnedders>
*Well

[15:49:03.0000] <gsnedders>
/me hasn't drunk anything

[15:49:04.0000] <TabAtkins>
Astronymics!

[15:51:00.0000] <karlushi>
anyway, there are both SSSB since 2006

[15:51:01.0000] <TabAtkins>
Solar System Smash Bros?

[15:52:00.0000] <karlushi>
Small Solar System Body

[15:53:00.0000] <jgraham>
karlushi: You got a reference for "asteroids do not have to come from the solar system"?

[15:55:00.0000] <jgraham>
I can't imagine that there are a significant number of bodies that make trans-stellar journies and velocity higher than the escape velocity of the solar system

[15:55:01.0000] <TabAtkins>
I doubt there are many either.  But that doesn't make them not asteroids.

[15:56:00.0000] <jgraham>
It depends if "not many" means "none"

[15:56:01.0000] <jgraham>
Also, wikipedia disagrees with you

[15:56:02.0000] <TabAtkins>
The universe is a big place, and orbital dynamics is complicated.

[15:56:03.0000] <TabAtkins>
No it doesn't.  Note the "traditionally".

[15:56:04.0000] <jgraham>
And I can't find the IAU definition

[15:57:00.0000] <TabAtkins>
Or rather, "historically".

[15:57:01.0000] <jgraham>
Sure, orbital dynamics is complicated and it is easy to believe that some material is ejected from solar systems

[15:59:00.0000] <jgraham>
TabAtkins: Note that the "historically" applies to "of the inner slar system"

[15:59:01.0000] <jgraham>
*solar

[16:00:00.0000] <TabAtkins>
True.

[16:00:01.0000] <TabAtkins>
Still, I don't know if IAU has a better word for interstellar rocks, so I'm going with "asteroid" for them.

[16:01:00.0000] <othermaciej>
TabAtkins: wow, that's long; no time to read now

[16:01:01.0000] <TabAtkins>
Yeah, sorry.  No problem.

[16:01:02.0000] <TabAtkins>
Tell me when you respond, because I won't see it.

[16:02:00.0000] <jgraham>
/me defers to Feynman: "You can know the name of a bird in all the languages of the world, but when you're finished, you'll know absolutely nothing whatever about the bird... So let's look at the bird and see what it's doing -- that's what counts."

[16:04:00.0000] <AryehGregor>
/me observes that cssquirrel.com's captcha is guessable with 50% probability, so doesn't seem very effective

[16:04:01.0000] <TabAtkins>
Just having *a* captcha of some kind is effective for preventing most spammers.

[16:05:00.0000] <TabAtkins>
Anyone know any django in the room?

[16:07:00.0000] <hober>
TabAtkins: a bit

[16:07:01.0000] <jgraham>
TabAtkins: No, but ask anyway

[16:08:00.0000] <TabAtkins>
Kk.  I have a Test and a Comment class with a ManyToMany relationship.  The Comment class has an Author field.  I'd like to link things up so that Test has a reviewers field containing all the authors of comments connected to the test.

[16:08:01.0000] <TabAtkins>
Any automagic way to specify that, or do I have to create a getter that does it manually?

[16:11:00.0000] <gsnedders>
/me looks at how many photos he has from tonight, and wonders how many photos other people took

[16:12:00.0000] <jgraham>
I took about 4

[16:12:01.0000] <jgraham>
Well more actually but about 4 with my camera

[16:13:00.0000] <gsnedders>
/me remembers drinking water would probably be a good idea now

[16:14:00.0000] <jgraham>
And sleep

[16:14:01.0000] <zcorpan_>
drinking sleep?

[16:14:02.0000] <gsnedders>
zcorpan_: You should've come!

[16:15:00.0000] <TabAtkins>
zcorpan_: You see my message to you from yesterday re: keypress reasoning in webkit?

[16:15:01.0000] <zcorpan_>
TabAtkins: yes, thanks

[16:15:02.0000] <TabAtkins>
kk

[16:15:03.0000] <gsnedders>
jgraham: Because I manage to get so much of that nowadays

[16:17:00.0000] <gsnedders>
There again, despite lack of sleep I manage to look amazing alert for the amount of sleep I do get

[16:20:00.0000] <gsnedders>
Hmm, I think I took around half of the 54 photos on my camera

[16:31:00.0000] <zcorpan_>
http://www.sitepoint.com/forums/showthread.php?t=683147 - people should write tutorials for converting common flash stuff to html5

[16:32:00.0000] <TabAtkins>
Argh, their page-tracking script killed the tab.

[16:32:01.0000] <TabAtkins>
Ah, better this time.


2010-06-16
[17:40:00.0000] <Hixie>
wow i had no idea that you could mail draft-abarth-origin⊙tio

[17:40:01.0000] <Hixie>
that's pretty neat

[17:51:00.0000] <JonathanNeal>
In HTML, when you have a label and an input, what would you call that pairing.  I ask because I want to wrap it in a div and give it a classname that describes the content.

[17:52:00.0000] <JonathanNeal>
I think it's a control, right?

[18:33:00.0000] <TabAtkins>
Labelled input?

[20:02:00.0000] <JonathanNeal>
So, am I keeping up with the latest WHATWG changes @ http://www.iecss.com/whatwg.css ?

[20:13:00.0000] <karlcow>
jgraham: definition of IAU http://en.wikipedia.org/wiki/IAU_definition_of_planet#Final_definition

[20:13:01.0000] <MikeSmith>
"Beaver Taste Great With Mustard"

[20:13:02.0000] <MikeSmith>
now that's good

[20:13:03.0000] <MikeSmith>
give us more of that kind of stuff please

[20:13:04.0000] <karlcow>
including http://www.iau.org/public_press/news/release/iau0601/q_answers/

[20:14:00.0000] <karlcow>
Hmm beaver!

[20:15:00.0000] <karlcow>
eating beaver like a mountain man

[20:25:00.0000] <othermaciej>
Hixie: your description of which <a> elements on reddit are buttons vs. links makes me doubt my sanity

[20:34:00.0000] <Hixie>
othermaciej: it's just a matter of what they look like

[20:34:01.0000] <othermaciej>
really?

[20:34:02.0000] <othermaciej>
are the semantics of all controls defined by their appearance rather than behavior, or only links and buttons?

[20:35:00.0000] <Hixie>
i don't understand the question

[20:35:01.0000] <Hixie>
the point of ATs is to expose what the controls look like to non-AT users

[20:35:02.0000] <othermaciej>
all right, let me give a concrete example

[20:35:03.0000] <othermaciej>
look at this control:

[20:35:04.0000] <othermaciej>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cstyle%3E%0A%23mystery%20%7B%0A%20%20%20%20background%3A%20white%20url(http%3A%2F%2Fwww.reddit.com%2Fstatic%2Fgradient-button.png)%20repeat-x%20scroll%200%25%2050%25%3B%0A%20%20%20%20border%3A%201px%20solid%20%23C4DBF1%3B%0A%20%20%20%20display%3A%20block%3B%0A%20%20%20%20font-size%3A%20150%25%3B%0A%20%20%20%20font-weight%3A%20bold%3B%0A%20%20%20%20height%3A%2029px%3B%0A%20%20%2

[20:35:05.0000] <othermaciej>
tter-spacing%3A%20-1px%3B%0A%20%20%20%20line-height%3A%2029px%3B%0A%20%20%20%20position%3A%20relative%3B%0A%20%20%20%20text-align%3A%20center%3B%0A%20%20%20%20margin%3A%090px%3B%0A%20%20%20%20padding%3A%200px%3B%0A%7D%0A%3C%2Fstyle%3E%0A%0A%3Cinput%20id%3Dmystery%20type%3Dtext%20value%3Dmystery%3E%0A

[20:35:06.0000] <othermaciej>
I guess I should shorten that, hold on a sec

[20:35:07.0000] <Hixie>
and the point of ARIA is to enable that to happen when the browser can't figure it out

[20:35:08.0000] <othermaciej>
http://bit.ly/cK5Cfo

[20:35:09.0000] <Hixie>
you can use hte "save" link

[20:36:00.0000] <Hixie>
to get a shorter URL

[20:36:01.0000] <othermaciej>
what are the semantics of that control?

[20:36:02.0000] <othermaciej>
is that control semantically a button?

[20:36:03.0000] <Hixie>
i don't know what you mean by "semantic" in this context

[20:37:00.0000] <othermaciej>
I am refering to whatever you meant when you wrote "Authors must not use elements, attributes, or attribute values for purposes other than their appropriate intended semantic purpose."

[20:37:01.0000] <Hixie>
the intended semantic purpose of <input type=text> is a text control

[20:37:02.0000] <Hixie>
is that the purpose to which you are putting it here?

[20:37:03.0000] <othermaciej>
am I correctly using it for the intended semantic purpose?

[20:38:00.0000] <othermaciej>
or am I abusing it for the semantic purpose of a button?

[20:38:01.0000] <Hixie>
you are using it to try to prove a point in an argument, so no, it's no the appropriate intended semantic purpose

[20:38:02.0000] <Hixie>
:-P

[20:38:03.0000] <othermaciej>
you seemed to argue that a link with that styling has the semantics of a button, even if it has the behavior of a link

[20:38:04.0000] <othermaciej>
whereas something that looks like a link but has the behavior of a button is a link

[20:38:05.0000] <othermaciej>
I am wondering how far that extends

[20:39:00.0000] <othermaciej>
my understanding of control semantics is that it's tied to the behavior, not the appearance

[20:39:01.0000] <Hixie>
not sure what you mean by "behaviour of a link" vs "behaviour of a button"

[20:39:02.0000] <othermaciej>
and I was mighty confused to see you argue otherwise

[20:39:03.0000] <othermaciej>
are you saying the only semantic difference between links and buttons is their visual appearance?

[20:39:04.0000] <Hixie>
isn't it?

[20:40:00.0000] <othermaciej>
my understanding is that the semantics of a link are that it links to something

[20:40:01.0000] <othermaciej>
and the semantics of a button are that it actuates a command

[20:40:02.0000] <othermaciej>
regardless of what either looks like

[20:40:03.0000] <othermaciej>
and that properly using an <a> element for a link has consequences such as appropriate cursor and status bar behavior on hover, being recognized as indexable by search engines, etc

[20:41:00.0000] <Hixie>
isn't "link to something" a command?

[20:41:01.0000] <othermaciej>
it's a specific subset of commands

[20:41:02.0000] <othermaciej>
there are clearly commands that are not a "link to something"

[20:41:03.0000] <othermaciej>
(such as the command performed by the reddit submit button)

[20:42:00.0000] <othermaciej>
er, the reddit share button

[20:42:01.0000] <othermaciej>
similarly, "toggle" is a form of command, but that doesn't mean buttons and checkboxes are interchangeable other than appearance

[20:42:02.0000] <Hixie>
i think i could convincingly argue this about 5 different ways if i tried

[20:42:03.0000] <Hixie>
i don't think any of them would help us in the context of ATs and ARIA

[20:43:00.0000] <othermaciej>
I'm just trying to understand what it means to forbid using a link as a button

[20:44:00.0000] <othermaciej>
your argument seems to be based on the premise that doing that should be disallowed, but your examples of what does or does not constitute using a link as a button did not help my comprehension

[20:45:00.0000] <Hixie>
ok forget what i said about reddit. here's a better way to think about it based on what you said earlier.

[20:45:01.0000] <othermaciej>
and made me doubt my understanding of the requirement, ""Authors must not use elements, attributes, or attribute values for purposes other than their appropriate intended semantic purpose.""

[20:45:02.0000] <Hixie>
a link is something that provides the user with a URL

[20:45:03.0000] <othermaciej>
I would agree with that definition

[20:45:04.0000] <othermaciej>
(by that definition, "Submit a link" on reddit would be a link, but "share" would not)

[20:45:05.0000] <Hixie>
a button is something that provides the user with the option to activate something defined by the author at a user-selected time

[20:46:00.0000] <othermaciej>
and by that definition, "share" would be a button but you could also argue that all links (at least in the control sense) are also buttons

[20:46:01.0000] <Hixie>
so on reddit, the arrows and "share" should be buttons, and "submit a link" should be a link

[20:46:02.0000] <othermaciej>
I agree

[20:47:00.0000] <Hixie>
so using <a> for "share" is wrong

[20:47:01.0000] <othermaciej>
also agree

[20:47:02.0000] <Hixie>
ok

[20:48:00.0000] <othermaciej>
so looking at the specific example of "share", I think the point Steve and Sam were calling attention to was:

[20:49:00.0000] <othermaciej>
-> the spec makes it a validator error to put rule="button" on that <a> element, because it would be making an <a> element act as a button, contrary to its semantics

[20:49:01.0000] <Hixie>
the spec makes using <a> a validator error too, albeit one that few validators will ever be able to report in the coming few years

[20:49:02.0000] <othermaciej>
-> the spec does not make it a validator error to use href="#" onclick="..." on that <a> element, even though almost the sole use of that markup is to make an <a> element act like a button

[20:50:00.0000] <Hixie>
my point is that it does

[20:50:01.0000] <othermaciej>
(and similarly for href="javascript:...", although that's not used in this particular case)

[20:50:02.0000] <Hixie>
the spec does disallow that

[20:50:03.0000] <othermaciej>
it disallows the concept with a non-machine-checkable requirement, but does not make the syntax an error

[20:50:04.0000] <othermaciej>
however, it does make the syntax <a role="button"> an error

[20:50:05.0000] <Hixie>
right, role="" is a godsend here

[20:50:06.0000] <othermaciej>
with a machine-checkable rule

[20:51:00.0000] <Hixie>
it finally allows us an opportunity to catch this using a syntax error

[20:51:01.0000] <Hixie>
there are lots of things that cannot be caught using syntax errors

[20:51:02.0000] <Hixie>
they are no less errors

[20:51:03.0000] <karlcow>
othermaciej: the issue is in part what is the link. Is it a POST or a GET. http://bitworking.org/news/I_m_sorry__I_can_t_kiss_it_and_make_it_better_

[20:51:04.0000] <othermaciej>
in principle, a link with href="#" onclick="..." or href="javascript:..." could be a syntax error, unless that syntax serves a valid use case that cannot be achieved in any other way

[20:51:05.0000] <Hixie>
and when we can catch them we try to do so

[20:51:06.0000] <Hixie>
indeed

[20:52:00.0000] <Hixie>
(unfortunately, there are such use cases)

[20:52:01.0000] <othermaciej>
ok, so explaining what those are would help resolve the apparent contradiction

[20:52:02.0000] <Hixie>
the use cases?

[20:52:03.0000] <othermaciej>
since a naiive observer such as myself reading the change proposal might be unaware of those use cases

[20:52:04.0000] <Hixie>
href="javascript:"'s use case is exposing bookmarklets

[20:53:00.0000] <karlcow>
http://natbat.net/2009/Jun/10/styling-buttons-as-links/

[20:53:01.0000] <othermaciej>
good point

[20:53:02.0000] <othermaciej>
(my suggestion would be to cite these in your Change Proposal and/or in resolving SteveF's bugs when you get around to that)

[20:53:03.0000] <Hixie>
onclick="..."'s use case is for such things as click-tracking or general user interaction, or making links that are both functional in "AJAX" and static scenarios

[20:54:00.0000] <Hixie>
i don't understand why they are relevant to the change proposal

[20:54:01.0000] <othermaciej>
right, but you would never use href="#" [sic] for a click-tracked link that is supposed to work in both static and AJAX scenarios

[20:54:02.0000] <karlcow>
http://jontangerine.com/silo/html/button/

[20:54:03.0000] <Hixie>
making href="#" non-conforming might make sense

[20:55:00.0000] <Hixie>
it's kind of a weirdly specific case to disallow

[20:55:01.0000] <Hixie>
i mean i don't see much point linking to the same page, but why would we disallow it?

[20:55:02.0000] <Hixie>
it's just a relative URL...

[20:55:03.0000] <othermaciej>
I think Steve's point was that href="#" onclick="..." is exclusively used to make <a> elements act like buttons and has no valid use case

[20:56:00.0000] <othermaciej>
I agree that it would be an awfully specific thing to disallow, though one could claim the same of <a role="button">

[20:56:01.0000] <othermaciej>
anyway

[20:56:02.0000] <othermaciej>
I think those were the questions being asked

[20:56:03.0000] <othermaciej>
feel free to respond as you see fit

[20:56:04.0000] <othermaciej>
I should be going

[20:56:05.0000] <Hixie>
role="button" isn't specifically disallowed

[20:56:06.0000] <Hixie>
it's just not included in the list of things that are allowed

[20:57:00.0000] <othermaciej>
rather, it's not specifically allowed

[20:57:01.0000] <othermaciej>
anyway

[20:57:02.0000] <Hixie>
that's the opposite argument :-)

[20:57:03.0000] <othermaciej>
I don't want to argue the point further, you have explained these matters enough that I feel I understand them, but you may want to explain the above points on the list for the benefit of others

[20:58:00.0000] <othermaciej>
I hope that my confusion was understandable and I expect others may share it

[20:58:01.0000] <othermaciej>
thanks for the explanation

[20:58:02.0000] <Hixie>
the only others i care about here are sam and paul, frankly

[20:59:00.0000] <Hixie>
and last time i included counter-arguments in a change proposal i was told it was a waste of time

[20:59:01.0000] <Hixie>
and indeed it had no effect

[20:59:02.0000] <Hixie>
so i'm at a loss as to how to include this in a useful manner

[21:00:00.0000] <othermaciej>
I think they (and perhaps Steve) would likely benefit from an explanation of these points, even if it is just in an ordinary email to the list

[21:00:01.0000] <othermaciej>
explaining to Steve is useful because if he chooses to withdraw his proposal or agrees to some compromise (e.g. disallowing href="#" onclick="..." and/or explaining in more detail that buttons shouldn't be abused as links and explaining the valid use cases for <a href="javascript:">), that ends the process without Sam or Paul's opinion being relevant

[21:34:00.0000] <MikeSmith>
strawman attempt at a concise definition of what the term "the Web platform" means:

[21:34:01.0000] <MikeSmith>
[[

[21:34:02.0000] <MikeSmith>
  "a client-side (browser-side) runtime/processing environment for

[21:34:03.0000] <MikeSmith>
  Web applications and Web documents that are built using HTML,

[21:34:04.0000] <MikeSmith>
  CSS, Javascript and a couple of other core-standard formats, and

[21:34:05.0000] <MikeSmith>
  that are addressed and transmitted using HTTP, URIs and a couple

[21:34:06.0000] <MikeSmith>
  of other core-standard protocols"

[21:34:07.0000] <MikeSmith>
]]

[21:34:08.0000] <MikeSmith>
comments, suggestions, refinements, please

[21:35:00.0000] <wirepair>
'couple of other' used 2x ...

[21:36:00.0000] <roc>
the list of relevant standards is much longer than that

[21:36:01.0000] <wirepair>
seems lil vaguish

[21:41:00.0000] <MikeSmith>
it's intended to be minimal

[21:42:00.0000] <MikeSmith>
not to be an exhaustive list

[21:43:00.0000] <roc>
then just say "... and other standard formats"

[21:45:00.0000] <MikeSmith>
OK

[23:39:00.0000] <MikeSmith>
oh, issue 56 is _my_ fault

[23:39:01.0000] <MikeSmith>
[[

[23:39:02.0000] <MikeSmith>
ISSUE-56 was raised in error by Michael(tm) Smith based on a                                                                                                                                        

[23:39:03.0000] <MikeSmith>
misunderstanding of Roy's messages to the working group.

[23:39:04.0000] <MikeSmith>
]]

[23:42:00.0000] <MikeSmith>
fwiw, there was never any misunderstanding

[23:45:00.0000] <othermaciej>
wait, what?

[23:45:01.0000] <MikeSmith>
not on my part at least

[23:45:02.0000] <MikeSmith>
I'm not even sure that I was the one who raised that issue initially, though I ended up becoming the owner for it

[23:46:00.0000] <MikeSmith>
Adam's change proposal

[23:46:01.0000] <MikeSmith>
which change proposal I personally happen completely agree with, fwiw

[23:47:00.0000] <MikeSmith>
http://lists.w3.org/Archives/Public/public-html/2010Jun/0394.html

[00:06:00.0000] <othermaciej>
MikeSmithX: it's starting to seem like a better approach than either of the other proposals

[00:11:00.0000] <boblet>
 “We might actually be able to process hyperlinks interoperably, leading to joy and happiness… purveyors of whisky might go out of business.”

[00:11:01.0000] <boblet>
lol

[00:15:00.0000] <MikeSmith>
I was going to say, the lesson for me here is to not raise issues on behalf of anybody else

[00:15:01.0000] <MikeSmith>
especially when I don't personally agree it's an issue

[00:16:00.0000] <MikeSmith>
but I actually learned that lesson quite a while ago

[00:16:01.0000] <MikeSmith>
which is why I don't do it any more

[00:16:02.0000] <othermaciej>
fortunately, the new process encourages people to raise their own issues, or at least provide issue text

[00:17:00.0000] <MikeSmith>
yeah

[00:18:00.0000] <othermaciej>
anyway, I am glad Adam posted that Change Proposal, even if he partly meant it as a reductio ad absurdum of the process

[00:18:01.0000] <MikeSmith>
yeah

[00:18:02.0000] <MikeSmith>
the summary is a very good and succinct statement of the fact that this is a non-problem

[00:19:00.0000] <MikeSmith>
and arguably always has been

[00:20:00.0000] <othermaciej>
well, he's actually arguing for a change relative to the way the spec is currently

[00:20:01.0000] <othermaciej>
I think once I read his proposal closer I might conclude it lacks sufficient detail

[00:21:00.0000] <othermaciej>
though that could be fixed by linking to http://www.w3.org/html/wg/href/draft for instance

[00:23:00.0000] <MikeSmith>
othermaciej: well, let me put it this way: I don't think there is any risk of implementor confusion about this part of the spec, nor has there ever been

[00:23:01.0000] <MikeSmith>
I agree with what Adam writes about adding text for clarification, but I don't think it's all that necessary to add it

[00:24:00.0000] <MikeSmith>
except to appease those who have objected to the terminology

[00:24:01.0000] <othermaciej>
I don't think implementors would be confused by the different use of the word URL

[00:25:00.0000] <othermaciej>
so far none of them have been

[00:25:01.0000] <othermaciej>
I don't know about other possible audiences for the spec

[00:26:00.0000] <MikeSmith>
I don't think even a casual reader would be confused by it

[00:26:01.0000] <othermaciej>
I think the factoring issue (where should the processing to convert what occurs in an href attribute to a URI/IRI/whatever be defined, in HTML5 or in IRI or in some third document) is somewhat orthogonal to the terminology issue

[00:26:02.0000] <MikeSmith>
I don't think _anybody_ would really be confused by it

[00:26:03.0000] <othermaciej>
however, so far no one else has stepped up to the plate to write a viable spec for this

[00:27:00.0000] <MikeSmith>
yeah, the factoring issue is certainly orthogonal

[00:27:01.0000] <MikeSmith>
and we did have come half-hearted previous attempts at writing a separate spec

[00:27:02.0000] <othermaciej>
ISSUE-56 is primarily about where the relevant spec text goes, not the terminology, that was ISSUE-72 which is already closed (by timeout)

[00:28:00.0000] <MikeSmith>
the other thing is, the term "URL" is long-deprecated by the purists anyway

[00:28:01.0000] <MikeSmith>
they have abandoned it

[00:29:00.0000] <MikeSmith>
so it's always seemed to me to be somewhat ironic to see an objection to how it's been used here

[00:29:01.0000] <MikeSmith>
anyway, I think this is one of those issues where you could count the number of people who feel strongly about it on one hand

[00:30:00.0000] <MikeSmith>
even if you were missing fingers

[00:30:01.0000] <MikeSmith>
not that the number of people objecting strongly is the only measure of whether it's important or not

[00:31:00.0000] <othermaciej>
I will say this - I personally care that the proper interoperable processing rules for browsers are spec'd *some*where

[00:32:00.0000] <othermaciej>
I don't care that much where that place is

[00:32:01.0000] <othermaciej>
I don't think the people with specific preferences in that regard should get unlimited time to produce something usable though

[00:33:00.0000] <MikeSmith>
yeah

[00:42:00.0000] <hsivonen>
jgraham: ISO makes some of its spec available as .pdf.zip, which is just unbelievable user-hostile compared to posting .pdf with flate streams inside

[00:44:00.0000] <hsivonen>
jgraham: and you can get a useful reference by linking to a know-good rev of the wikipedia article

[00:50:00.0000] <hsivonen>
MikeSmith: for defining the Open Web Platform, I suggest the following definition: 

[00:51:00.0000] <hsivonen>
MikeSmith: give a feature one point for support in each of Gecko, WebKit, Presto and Trident in their latest versions

[00:51:01.0000] <hsivonen>
(so each feature gets a score from 0 to 4)

[00:51:02.0000] <hsivonen>
features that score 3 or 4 are part of the Platform

[01:03:00.0000] <MikeSmith>
hsivonen: that is a great definition for us to use with each other

[01:03:01.0000] <MikeSmith>
but it's one that does not go over quite as well with others

[01:03:02.0000] <MikeSmith>
e.g., with non-implementors

[01:22:00.0000] <hsivonen>
hooray to abarth_ for the Change Proposal for ISSUE-56

[01:22:01.0000] <abarth_>
hi hsivonen 

[01:22:02.0000] <abarth_>
i should have been a spec lawyer

[01:23:00.0000] <othermaciej>
I appreciate that abarth_ posted it as well, even though part of it was vaguely trying to make fun of me

[01:23:01.0000] <othermaciej>
they should totally have spec law school

[01:23:02.0000] <othermaciej>
but then you'd have to pass the spec bar exam to be allowed to practice as a spec lawyer

[01:24:00.0000] <abarth_>
the bit about whisky purveyors?

[01:24:01.0000] <abarth_>
:)

[01:24:02.0000] <othermaciej>
no, my borderline alcoholism is fair game

[01:25:00.0000] <othermaciej>
and may well go beyond borderline before HTML5 is done

[01:30:00.0000] <zcorpan_>
if we disallow href="#", all we'll achieve is to annoy authors and make them use href="#i-hate-html5" instead

[01:32:00.0000] <othermaciej>
I personally think trying to enforce semantics via syntax errors is an ill-conceived project

[01:32:01.0000] <zcorpan_>
i agree

[01:33:00.0000] <zcorpan_>
also, the world doesn't fall apart if authors use <a role=button>

[01:34:00.0000] <MikeSmith>
abarth_: hey man

[01:34:01.0000] <MikeSmith>
sorry for any vehemence in my replies on public-html

[01:35:00.0000] <abarth_>
MikeSmith: hi

[01:35:01.0000] <zcorpan_>
Hixie: "The current text satisfies the above edit instructions (as I'm sure Tab intended)." i thought hsivonen wrote the counter proposal

[01:35:02.0000] <MikeSmith>
abarth_: I have to carefully preserve the fiction that I actually know what I'm doing

[01:35:03.0000] <abarth_>
haha :)

[01:35:04.0000] <MikeSmith>
:)

[01:36:00.0000] <abarth_>
MikeSmith: i'm trying to read public-html after filter out all the emotion

[01:36:01.0000] <MikeSmith>
I find that drugs help

[01:36:02.0000] <abarth_>
so, i think i missed any vehemence

[01:41:00.0000] <Hixie>
zcorpan_: oops

[01:41:01.0000] <Hixie>
zcorpan_: oh well it's what happens because i ignore the From: line when reading e-mails!

[01:45:00.0000] <Hixie>
"the current text implies that HTML is super-easy, while XHTML is so hard that even the Working Group can't explain it"

[01:45:01.0000] <Hixie>
someone should propose a change proposal that makes it explicit that HTML is that hard

[01:45:02.0000] <Hixie>
since apparently julian doesn't think we've made that clear enough

[01:49:00.0000] <asmodai>
Hixie: hard/easy? In the context of writing a spec?

[01:50:00.0000] <Hixie>
er, s/HTML/XHTML/ in my above comment

[02:06:00.0000] <othermaciej>
to be fair, HTML is also hard in many ways! though not escaping rules in particular

[02:10:00.0000] <zcorpan_>
Q: What are the risks?

[02:10:01.0000] <zcorpan_>
A: We might actually be able to process hyperlinks interoperably,

[02:10:02.0000] <zcorpan_>
leading to joy and happiness.  With so much joy in the work, purveyors

[02:10:03.0000] <zcorpan_>
of whisky might go out of business.

[02:10:04.0000] <zcorpan_>
LOL

[02:10:05.0000] <annevk>
call for consensus for XHR CR

[02:10:06.0000] <annevk>
never got this close before

[02:10:07.0000] <annevk>
jippiedieyay

[02:11:00.0000] <abarth>
zcorpan_: omg, typo...   s/work/world/

[02:11:01.0000] <annevk>
maybe it's more like jippydiyay

[02:12:00.0000] <annevk>
abarth, note that it's not just HTML that uses that kind of processing

[02:12:01.0000] <abarth>
ssh!

[02:12:02.0000] <annevk>
abarth, as you're probably aware, XMLHttpRequest, CSS, HTTP, etc. all use it

[02:12:03.0000] <annevk>
haha

[02:12:04.0000] <annevk>
but yeah, whatever

[02:16:00.0000] <othermaciej>
annevk: write some CR exit criteria please so I can support the CR transition in good consience

[02:20:00.0000] <Hixie>
annevk: you can crib the CR exit criteria from the whatwg charter if you like :-)

[02:32:00.0000] <MikeSmith>
what does the "from TFA" part in dudes message mean?

[02:42:00.0000] <annevk>
othermaciej, I plan to just use http://www.w3.org/TR/CSS21/#crec

[02:44:00.0000] <Hixie>
that's basicaly what the whatwg charter has

[02:44:01.0000] <Hixie>
(word for word)

[02:44:02.0000] <othermaciej>
annevk: it would be simpler IMO to just say there should be two implementations that pass every test case

[02:44:03.0000] <hsivonen>
/me hadn't realized CSS had a 6-month rule

[02:45:00.0000] <othermaciej>
then you don't have to define the scope of what is a "feature"

[02:45:01.0000] <Hixie>
/me agrees with othermaciej that that requirement is better -- require a comprehensive test suite and two fully compliant implementations

[02:45:02.0000] <slartsa>
hsivonen: what do you mean?

[02:46:00.0000] <othermaciej>
given the relatively small size of XHR, it would be silly to define different pieces of it as separate features, and a waste of time to debate which subsection levels constitute feature boundaries

[02:46:01.0000] <hsivonen>
http://www.w3.org/WAI/PF/HTML/track/actions/24 looks odd given PF's own statements

[02:46:02.0000] <othermaciej>
also it does not seem like a huge burden to expect two complete implementations

[02:46:03.0000] <hsivonen>
slartsa: I mean that 6 months from the start of CR has to pass even if the other criteria were met sooner

[02:47:00.0000] <slartsa>
now would that just be there to narrow possible errors?

[02:47:01.0000] <othermaciej>
annevk: those CR exit criteria also don't seem to say anything about the comprehensiveness of the test suite used to judge implementations of any given feature

[02:48:00.0000] <hsivonen>
slartsa: yeah, I just hadn't realized there was a 6-month rule

[02:48:01.0000] <annevk>
othermaciej, ok

[02:48:02.0000] <othermaciej>
also they effectively make all features potentially "at risk"

[02:48:03.0000] <annevk>
othermaciej, I wouldn't let it pass CR without two complete implementations and an extensive test suite, but maybe it is better to put that in writing indeed

[02:48:04.0000] <othermaciej>
which is explicitly disallowed by the W3C process

[02:49:00.0000] <othermaciej>
annevk: I would like to see that in writing, so there's no misunderstandings later with other WG members

[02:49:01.0000] <othermaciej>
you're not allowed to just say "any feature without two implementations will be dropped", you have to list specific features as being "at risk" if they don't meet the relevant criteria

[02:50:00.0000] <othermaciej>
at least that is my recollection

[02:50:01.0000] <othermaciej>
for XHR I am not sure any features should be "at risk"

[02:52:00.0000] <annevk>
no I agree

[02:53:00.0000] <annevk>
ok, I'll take it more serious :)

[02:54:00.0000] <Hixie>
othermaciej: the csswg discussed that long ago and basically decided that they would go back to LC often enough anyway since they were going to maintain the spec that if that became an issue, they'd just cycle again and list the at-risk parts (or just remove them)

[03:05:00.0000] <annevk>
TabAtkins, oh, it was a joke? you know, the CSS WG leaving px undefined would not be so surprising really :)

[03:12:00.0000] <hsivonen>
annevk: what was a joke?

[03:13:00.0000] <annevk>
http://twitter.com/fantasai/status/16181517781

[03:13:01.0000] <annevk>
this WebM stuff is problematic

[03:13:02.0000] <annevk>
YouTube now works in Opera

[03:13:03.0000] <hsivonen>
annevk: that was too subtle for me to realize it was a joke

[03:14:00.0000] <hsivonen>
annevk: so what's the problem with WebM?

[03:14:01.0000] <hsivonen>
annevk: youtube harms your productivity?

[03:14:02.0000] <jgraham>
hsivonen: What will people complain about?

[03:14:03.0000] <annevk>
hsivonen, exactly, I might end up watching YouTube all day

[03:15:00.0000] <annevk>
not having Flash helped a lot, but then they found this workaround

[03:15:01.0000] <jgraham>
I mean for 10.50 we got _so_many_ people saying "why doesn't HTML5 video work in Opera"

[03:15:02.0000] <jgraham>
because we didn't support H.264 Youtube

[03:16:00.0000] <hsivonen>
Am I supposed to get a WebM-enabled Opera build from the "beta-stable" apt source?

[03:16:01.0000] <annevk>
should I read the 31 emails on "RE: Transferring File* to WebApps - redux"?

[03:16:02.0000] <annevk>
(I don't know the public sources for Opera distribution... I'm using the latest internal nightlies...)

[03:17:00.0000] <jgraham>
hsivonen: I think you have to look at the desktop team blog

[03:17:01.0000] <annevk>
well, from an internal URL; there doesn't seem to be a much of difference anymore between what I use and someone who doesn't work for Opera; which is pretty great

[03:17:02.0000] <hsivonen>
jgraham: :-( the repo is offering 10.53.6330 to me

[03:18:00.0000] <annevk>
yeah, we should set up something like Chrome does

[03:18:01.0000] <jgraham>
hsivonen: Yeah, I know, our update situation for dev builds really sucks

[03:18:02.0000] <gsnedders>
hsivonen: AFAIK beta still pushes stable for some raeson

[03:19:00.0000] <annevk>
in that thread someone mentions bondi

[03:19:01.0000] <annevk>
/me ran away

[03:36:00.0000] <Lachy>
annevk, the speech input proposal was linked to in a previous email. http://docs.google.com/View?id=dcfg79pz_5dhnp23f5

[03:37:00.0000] <annevk>
oh wow that looks complex

[03:37:01.0000] <Lachy>
I'm not sure I get the whole purpose of it. It's been possible to use speech input with any form control for years, given a microphone and speech recognition software.

[03:40:00.0000] <annevk>
there must be something why that doesn't just work, but it's not explained very well

[03:41:00.0000] <Lachy>
ah, it describes utterrance="" and interpretation="" attributes that help with defining what speech commands are accepted

[03:49:00.0000] <Hixie>
/me tries to work out what faulkner's e-mail means... is he saying that when the WAI said we should have "strong semantics" or whatever it was that they were wrong?

[04:04:00.0000] <AryehGregor>
Am I the only one who dies a little inside whenever I see a line beginning with ">From" in an e-mail when it should be "From"?

[04:05:00.0000] <annevk>
http://www.w3.org/TR/xbl/ has no exit criteria?

[04:05:01.0000] <annevk>
/me is looking at prior art

[04:09:00.0000] <Hixie>
AryehGregor: you are not

[04:10:00.0000] <Hixie>
AryehGregor: i die even more when it's my own e-mails

[04:11:00.0000] <AryehGregor>
:(

[04:11:01.0000] <AryehGregor>
If only the guy who came up with mbox had heard of reversible escaping.

[04:12:00.0000] <AryehGregor>
"Specifying a grammar is useful for apps which have limited vocabulary, for e.g. commands, navigation within page, maps etc. Such applications would not work as well with free-form text input. @pattern can be used to restrict the allowed inputs, but regular expressions are less expressive than context-free grammars."  That sounds extremely scary.

[04:13:00.0000] <workmad3>
wow, it's been a while since I came across a discussion involving context-free grammars :)

[04:14:00.0000] <workmad3>
now, if only I could remember what they are and what characterized them...

[04:15:00.0000] <AryehGregor>
Are context-free grammars exactly those that can be described in BNF, by any chance?  Like regular expressions and regex?

[04:15:01.0000] <AryehGregor>
/me isn't sure.

[04:16:00.0000] <workmad3>
that's the problem :)

[04:18:00.0000] <AryehGregor>
Am I the only one who finds back/forward in Chrome dev channel is sometimes incredibly flaky?

[04:20:00.0000] <AryehGregor>
It will regularly forget items in the history, and just now it completely made up a history entry.  A page I visited once and bookmarked, months ago, replaced the entire history for the tab.

[04:20:01.0000] <AryehGregor>
I would report it, but haven't tried to reproduce it consistently . . .

[04:28:00.0000] <oal>
Is there a good article that describes WebSockets, and what it's all about?

[04:29:00.0000] <Hixie>
the spec has some intro material that might help

[04:30:00.0000] <oal>
Over at w3.org?

[04:30:01.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/complete/network.html#introduction-7

[04:30:02.0000] <oal>
Thank you

[04:40:00.0000] <zcorpan_>
speaking of websockets, does anyone know websocket demos/games/apps (that run in the browser)?

[04:54:00.0000] <zcorpan_>
lol http://www.youtube.com/watch?v=EozwYbMTtS0&feature=player_embedded

[04:56:00.0000] <AryehGregor>
/me is struck by how businesses actually use the word "proprietary" in a positive sense in their advertising, instead of in the witheringly negative sense that it has in standards circles

[04:56:01.0000] <AryehGregor>
"we use BGP best-path routing, combined with Mzima's proprietary route-analysis system to ensure your audience is always served fastest."

[05:09:00.0000] <jgraham>
zcorpan_: I guess most demos are coded against -75 so far

[05:09:01.0000] <jgraham>
Oh wait

[05:10:00.0000] <jgraham>
Well I guess I mean most servers

[05:10:01.0000] <jgraham>
for the demos

[05:13:00.0000] <annevk>
oh yes

[05:13:01.0000] <annevk>
donkey kong returns

[05:13:02.0000] <annevk>
win

[05:13:03.0000] <annevk>
specifically, donkey kong country, which was my favorite video game for like way long

[05:15:00.0000] <zcorpan_>
jgraham: yeah :(

[05:50:00.0000] <annevk>
can people take a quick look at http://dev.w3.org/2006/webapi/XMLHttpRequest/#crec ?

[05:50:01.0000] <annevk>
it defines the CR exit criteria based on the discussion of this morning

[05:51:00.0000] <annevk>
just saying "ok" is enough :) I'd like to know if it's understandable

[05:54:00.0000] <jgraham>
annevk: ok afaict

[05:54:01.0000] <annevk>
cool

[05:57:00.0000] <remysharp>
question: should I be able to do document.querySelector('video').ontimeupdate = dostuff; ?

[05:58:00.0000] <remysharp>
I'm only able to bind to that event using addEventListener - which is fine, I just don't understand why

[05:58:01.0000] <remysharp>
compared to doing something like worker.onmessage = dostuff - which works with both addEventListener and inline

[05:58:02.0000] <annevk>
remysharp, bug in the browser

[05:58:03.0000] <remysharp>
okay - thought so.

[05:58:04.0000] <remysharp>
cheers

[05:59:00.0000] <annevk>
ontimeupdate is defined on HTMLElement iirc so it should be available on all HTML elements

[05:59:01.0000] <annevk>
maybe these so-called HTML5 testing sites should take that into account

[05:59:02.0000] <remysharp>
really? that ... seems odd to me, why would the timeupdate event be available on all element?

[05:59:03.0000] <annevk>
that's the way all event listeners work

[06:00:00.0000] <remysharp>
ah, cool - I did not know that - but that's nothing new ;-)

[06:00:01.0000] <remysharp>
cheers again.

[06:01:00.0000] <annevk>
just a historical thing; though it does allow you to dispatch custom timeupdate events and listen for them

[06:01:01.0000] <annevk>
using the simpler on-eventname mechanism

[06:48:00.0000] <hsivonen>
/me finds it weird that Opera has Google in the toolbar but Bing in speed dial

[06:48:01.0000] <hsivonen>
(obviously, it has to be about how much each of them pays for placement, but it's still weird)

[06:49:00.0000] <hsivonen>
(yes, I did a fresh install on a fresh OS install)

[06:54:00.0000] <hsivonen>
looks like YouTube has solved HTML5 video captioning on the JS level

[07:56:00.0000] <annevk>
sometimes it feels like there is more regrets emails than actual progress

[07:56:01.0000] <TabAtkins>
On that note, it looks like I'll have to regret today.

[07:57:00.0000] <TabAtkins>
Didn't realize I couldn't use my phone without logging into it, and I was never told me extension/password for the phone.

[07:57:01.0000] <TabAtkins>
Also left my cellphone at home.  >_<

[07:58:00.0000] <zcorpan_>
TabAtkins: excuses!

[07:59:00.0000] <TabAtkins>
/me hones his telepathic connection into the cellular network.

[08:01:00.0000] <annevk>
unless specifically asked I'm no longer doing them

[08:01:01.0000] <annevk>
and even then I probably can't

[08:01:02.0000] <annevk>
I broke Skype

[08:06:00.0000] <TabAtkins>
Ooh, we the ability to upgrade our desktops to Lucid now.  That should push my gnome version up high enough to finally install the compiz grid plugin!

[08:15:00.0000] <TabAtkins>
Crap, I've had a Dr. Horrible song in my head since yesterday.  I need to go watch it again.

[08:28:00.0000] <johnst>
TabAtkins, I wish for less noise in this channel.

[08:29:00.0000] <TabAtkins>
johnst: ?_?

[08:30:00.0000] <annevk>
pony!

[08:30:01.0000] <jgraham>
johnst: Try #htmlwg on the w3c servers

[08:30:02.0000] <annevk>
if we're gonna do wishes and all

[08:31:00.0000] <jgraham>
er #html-wg

[08:31:01.0000] <johnst>
jgraham, :-)

[08:32:00.0000] <TabAtkins>
othermaciej: Worth my time to update my two pending change proposals?  I dunno if anything's happened to them since I stopped listening.

[08:35:00.0000] <zcorpan_>
should a document in an iframe be unloaded when the iframe is removed from the (parent) document?

[08:36:00.0000] <TabAtkins>
Related question: Does an <iframe> have to be in a parent document to receive it's own document?

[08:37:00.0000] <TabAtkins>
(I was looking at Dean Edward's hack yesterday for ripping builtins out of an iframe to get around IE's bugs with subclassing Array.)

[08:37:01.0000] <zcorpan_>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/506

[08:43:00.0000] <annevk>
gonna write a domintro box for XHR

[08:45:00.0000] <zcorpan_>
seems webkit and firefox unload but ie and opera don't

[08:45:01.0000] <zcorpan_>
ie8 is annoying with it's handling of data: urls

[08:46:00.0000] <zcorpan_>
(if you try to open a data: url in an iframe, it redirects you to an error page)

[08:47:00.0000] <TabAtkins>
That does seem annoying.

[08:54:00.0000] <AryehGregor>
Even more annoying, it doesn't support data: entered manually into the URL bar.

[08:55:00.0000] <AryehGregor>
That's like half the point of data: from my point of view.

[08:56:00.0000] <zcorpan_>
AryehGregor: indeed

[08:57:00.0000] <AryehGregor>
Although I guess normal web developers don't need to do minimal test cases much, which is what I use it for.

[08:57:01.0000] <AryehGregor>
So we're all probably atypical.

[09:16:00.0000] <Smylers>
TabAtkins: In case this was after you stopped listening: http://lists.w3.org/Archives/Public/public-html/2010Jun/0072.html

[09:17:00.0000] <Smylers>
TabAtkins: I can't remember what Sam's feedback was, or why my comments would address it.

[09:18:00.0000] <Smylers>
TabAtkins: I wasn't attempting to address any of Sam's points when I wrote that.

[09:19:00.0000] <zcorpan_>
Hixie: does ws://damowmow.com:11111/demo still work? seems to disconnect for me

[10:23:00.0000] <gsnedders>
http://www.w3.org/blog/systeam/2010/06/16/why_we_chose_mercurial_as_our_dvcs

[10:29:00.0000] <AryehGregor>
I fiddled with Mercurial a bit, but I've gotten really used to git.

[10:30:00.0000] <AryehGregor>
git is complicated and unpredictable, but I was really annoyed more than once at Mercurial by not being allowed to roll back more than one revision, and some other things in that vein.

[10:30:01.0000] <jgraham>
I think mq will save you there

[10:31:00.0000] <jgraham>
(well depending what you mean "rollback"_

[10:31:01.0000] <jgraham>
s/_/)/

[10:31:02.0000] <AryehGregor>
Probably.

[10:32:00.0000] <AryehGregor>
I can't be bothered to seriously attempt to learn all the popular VCSes, though.

[10:32:01.0000] <jgraham>
I mean you can revert to an earlier revision 

[10:32:02.0000] <AryehGregor>
I'm familiar with git and svn, that's good enough for me.

[10:32:03.0000] <AryehGregor>
I can figure out enough of the others to check stuff out and make patches to submit if necessary, that's about all I need.

[10:32:04.0000] <jgraham>
I have been using git a lot recently and I quite like it, but it is pretty baroque in places. Mercurial is really very simple once you have the general idea

[10:32:05.0000] <jgraham>
Although hard things are still hard

[10:32:06.0000] <jgraham>
just not as hard

[10:33:00.0000] <jgraham>
I don't think I'm making any sense

[10:33:01.0000] <gsnedders>
You never do

[10:33:02.0000] <AryehGregor>
Yes, "baroque" is a good word for git.

[10:34:00.0000] <AryehGregor>
I'm sure it makes perfect sense if you fully understand its implementation, but not otherwise.

[10:34:01.0000] <AryehGregor>
Otherwise you learn how to do common tasks by trial and error and hope Google will help you find out how to do something new.

[10:34:02.0000] <AryehGregor>
Because the man pages are often not helpful except as a reference if you already know exactly what you're doing.

[10:35:00.0000] <jgraham>
I think I mean that mercurial is easier for a large range of common use cases but might be harder for some less-common use cases

[10:37:00.0000] <jgraham>
AryehGregor: If you can use git, it is a few minutes of effort to get reasonably proficient in mercurial

[10:38:00.0000] <AryehGregor>
Yeah, I've used it.

[10:38:01.0000] <jgraham>
Not "expert" level 

[10:38:02.0000] <AryehGregor>
Both for fiddling and to submit a patch to Mozilla.

[10:38:03.0000] <jgraham>
But good enough for common things

[10:47:00.0000] <oal>
Has anyone here done any work with web sockets yet?

[11:03:00.0000] <jgraham>
oal: Yes

[11:03:01.0000] <jgraham>
people here have

[11:03:02.0000] <jgraham>
if you have a question I recommend asking it

[11:03:03.0000] <jgraham>
:)

[11:06:00.0000] <oal>
jgraham, just wondering if there are any good resources for learning how to use it? I've managed to send a handshake from a simple python server, but I can't send any messages

[11:07:00.0000] <oal>
I've looked at Amazon, but all the html5 related books are coming later this summer/fall

[11:13:00.0000] <jgraham>
oal: Using what client?

[11:13:01.0000] <oal>
Google Chrome/Chromium

[11:19:00.0000] <AryehGregor>
I hate it when discussion is split among multiple threads and I reply to one before I see other relevant replies.

[11:19:01.0000] <oal>
jgraham, any ideas?

[11:51:00.0000] <JonathanNeal>
What would a tokenizer do? Like the one mentioned @ https://bugs.webkit.org/show_bug.cgi?id=39259#c3 ?

[11:52:00.0000] <AryehGregor>
http://en.wikipedia.org/wiki/Tokenizer

[11:56:00.0000] <paul_irish>
Aha... so what advantages do we expect from the html5 tokenizer landing in webkit?

[11:56:01.0000] <TabAtkins>
Little by itself.  But it's the necessary first step to coding the HTML5 treebuilder.

[11:57:00.0000] <TabAtkins>
Which *does* have benefits, namely everyone convering on using the particular parsing/building algo specified in HTML5, so all our pages always turn out the same everywhere.

[12:00:00.0000] <paul_irish>
Ah. gotcha. thx tab

[12:05:00.0000] <JonathanNeal>
Thanks TabAtkins.

[13:08:00.0000] <oal>
jgraham, I just bought "Pro HTML5 Programming" from Apress. It's an "alpha book" but some chapters are finished. The Web Socket one is :)

[13:08:01.0000] <hsivonen>
paul_irish: with the HTML5 landing in WebKit, the top 4 engines tokenize <div<div> the same way

[13:09:00.0000] <hsivonen>
*the HTML5 tokenizer landing

[13:11:00.0000] <zcorpan_>
oal: i wonder if the book describes the server impl part, and if so, if it's -75 or -76

[13:13:00.0000] <oal>
zcorpan_, what chapter would that be in? These chapters are finished: Html 5 Canvas API, Working with HTML 5 audio and video, Geolocation, Communication API, Web Sockets API, Web Workers API and Web storage api

[13:13:01.0000] <oal>
brb

[13:13:02.0000] <zcorpan_>
oal: web sockets

[13:14:00.0000] <oal>
zcorpan_, ok, back in 20 min, poke me if I forget to look for it

[13:32:00.0000] <zcorpan_>
does the webkit tokenizer support the script states and never reparses?

[13:41:00.0000] <oal>
zcorpan_, back

[13:42:00.0000] <oal>
zcorpan_, I searched for impl, didn't find anything

[13:43:00.0000] <bros>
Is it ok to discuss HTML 5 here?

[13:43:01.0000] <zcorpan_>
oal: ok

[13:59:00.0000] <TabAtkins>
bros: yeah.

[13:59:01.0000] <jgraham>
oal: Sorry, went home

[14:00:00.0000] <oal>
jgraham, I'm reading the Web Sockets chapter now. Interesting :)

[14:01:00.0000] <jgraham>
oal: OK. It is woth knowing that the protocol changed a bit recently and so different versions of chrome connect to different servers

[14:01:01.0000] <oal>
I will try the code in the book and see how it works

[14:02:00.0000] <gsnedders>
/me wonders what to do in the absence of any secret rendezvous

[14:03:00.0000] <TabAtkins>
gsnedders: overt rendezvous instead.

[14:04:00.0000] <jgraham>
oal: http://software.hixie.ch/utilities/js/websocket/ is a working demo you can play with

[14:05:00.0000] <zcorpan_>
jgraham: not working for me today

[14:05:01.0000] <gsnedders>
TabAtkins: Totally. With you and fanastai having done most of the (generic W3C) testing stuff (as well as the CSS stuff), we should totally have some informal F2F around the CSS WG F2F in Oslo

[14:05:02.0000] <gsnedders>
(if both of you will bethere)

[14:05:03.0000] <gsnedders>
*be there

[14:05:04.0000] <oal>
jgraham, will check it out. Thanks

[14:05:05.0000] <zcorpan_>
jgraham: or ws://damowmow.com:11111/demo is not working

[14:05:06.0000] <jgraham>
gsnedders: Unsecrify the previous one

[14:05:07.0000] <jgraham>
zcorpan_: Ah, I knew that

[14:06:00.0000] <jgraham>
gsnedders: ;)

[14:06:01.0000] <jgraham>
When is the CSS F2F?

[14:06:02.0000] <TabAtkins>
gsnedders: I... have no idea what you're talking about.  But yes, we totally should.

[14:06:03.0000] <TabAtkins>
Augst 23-25.

[14:06:04.0000] <jgraham>
TabAtkins: You will get used to it

[14:06:05.0000] <TabAtkins>
I think both me and Fantasai arrive on saturday the 21st.

[14:07:00.0000] <gsnedders>
/me should blatantly just go there for the weekend to party ;P

[14:07:01.0000] <TabAtkins>
Sounds fun.

[14:07:02.0000] <TabAtkins>
I'll be using Sunday to un-jetlag myself.  Partying would presumably help.

[14:07:03.0000] <gsnedders>
hah

[14:07:04.0000] <TabAtkins>
We've got a big 7-person apartment that a bunch of us are sharing.

[14:08:00.0000] <jgraham>
I might tryto find an excuse to be in Oslo then

[14:08:01.0000] <gsnedders>
heh

[14:08:02.0000] <gsnedders>
Obviously you should be working on what I'm working on. It makes it so trivial. :P

[14:08:03.0000] <jgraham>
Just need to keep working on projects with people in Oslo and it shouldn't be too hard

[14:11:00.0000] <zcorpan_>
http://webreflection.blogspot.com/2010/06/websocket-handshake-76-simplified.html - hmm, i thought preg_match was not the way to parse the handshake

[14:11:01.0000] <jgraham>
zcorpan_: So find bugs in it :)

[14:12:00.0000] <jgraham>
Sigh. So much to do

[14:13:00.0000] <zcorpan_>
well for one it would fail (sometimes, depending on the order the fields end up in) if i set subprotocol to Sec-WebSocket-Key1: lol

[14:13:01.0000] <TabAtkins>
Hehe.

[14:15:00.0000] <Hixie>
zcorpan_: that would be a pretty silly thing to do though :-)

[14:16:00.0000] <zcorpan_>
i guess being a qa i spot such holes without thinking

[14:16:01.0000] <zcorpan_>
Hixie: so why is your websocket server not working?

[14:16:02.0000] <Hixie>
which?

[14:17:00.0000] <jgraham>
Hmm, I think it is more that being zcorpan_ you spot such holes without thinking

[14:17:01.0000] <Hixie>
/me is pretty happy to see that all the work he did to come up with a system that could be implemented naively and yet still be secure worked :-)

[14:17:02.0000] <zcorpan_>
ws://damowmow.com:11111/demo

[14:17:03.0000] <Hixie>
zcorpan_: server is down - want me to turn it back on?

[14:17:04.0000] <Hixie>
up now

[14:17:05.0000] <zcorpan_>
yes please :)

[14:17:06.0000] <zcorpan_>
thanks

[14:18:00.0000] <zcorpan_>
jgraham: not all qas are as good as a zcorpan_ :)

[14:19:00.0000] <jgraham>
zcorpan_: I think I am not as evil as you. This makes me sad

[14:19:01.0000] <zcorpan_>
heh

[14:20:00.0000] <jgraham>
So on an entirely different topic setCaretSelectionRect

[14:21:00.0000] <jgraham>
Why do they need a width and height?

[14:21:01.0000] <jgraham>
It seems like it should be sufficient to give a point an a direction for a magnifier

[14:21:02.0000] <jgraham>
since presumably it needs to know about the direction of the text and where it starts

[14:22:00.0000] <jgraham>
but the size of a single character seems less relevant

[14:22:01.0000] <gsnedders>
Well, I'm glad your less evil than zcorpan_, jgraham 

[14:23:00.0000] <gsnedders>
I mean, I actually (really) have to share a room with you, rather than just in theory

[14:24:00.0000] <gsnedders>
s/your/you're/

[14:40:00.0000] <Hixie>
well this is a pretty good explanation of why technical development shouldn't be done using the htmlwg issue process or the consensus-driven process the task force is using: http://lists.w3.org/Archives/Public/public-html/2010Jun/0427.html

[14:42:00.0000] <AryehGregor>
"ARIA does not intrude on the visual experience and has near ubiquitous adoption."

[14:43:00.0000] <AryehGregor>
. . . how are we counting "adoption" here?

[14:43:01.0000] <TabAtkins>
Damn, you beat me to it.

[14:43:02.0000] <TabAtkins>
/me was just pasting that line in.

[14:46:00.0000] <jgraham>
Hah, I hadn't even noticed the bit about adoption

[14:51:00.0000] <zcorpan_>
http://www.google.com/codesearch?q=%5Cbaria- 4000 results

[14:52:00.0000] <zcorpan_>
http://www.google.com/codesearch?q=%3Cplaintext 4000 results

[14:56:00.0000] <jgraham>
Well to be fair I guess they mean adoption amongst UAs

[14:57:00.0000] <jgraham>
although a11y is hard because you can get 100% adoption amongst UAs and it doesn't matter a whit if authors don't use it or use it incorrectly

[14:57:01.0000] <zcorpan_>
it was in response to "... That suggests that it will be widely ignored or misused."

[15:04:00.0000] <bros>
I would like to discuss the handshaking process of the WebSockets API. Is it open for discussion?

[15:05:00.0000] <AryehGregor>
a11y people seem to focus on making features that authors could possibly use correctly if they really tried, rather than making ones that typical authors are likely to use correctly in practice without trying hard (or at all).

[15:05:01.0000] <AryehGregor>
bros, speak and you'll be heard.

[15:05:02.0000] <AryehGregor>
Various people here are qualified to comment.

[15:05:03.0000] <TabAtkins>
AryehGregor: ?_?

[15:05:04.0000] <TabAtkins>
Yes, that statement is completely correct.

[15:05:05.0000] <AryehGregor>
To be fair, it's probably impossible to design a system where authors design accessible pages without trying.

[15:07:00.0000] <bros>
I think trying to formulate an argument solved my own problem. I was trying to make my own very mini "WebSocket server" in C when I realized I would have implement, or copy an already exisiting implementation of the MD5 hashing function. I thought this was crazy just to handshake but now I realized WebSocket servers are going to be implementing functionality such as caching, etc. which would use MD5 checksums.

[15:08:00.0000] <zcorpan_>
you could design a system where authors design pages that are accessible as a result of the author trying to achieve something else (seo, validation, or zeldman semantic kool-aid)

[15:09:00.0000] <AryehGregor>
Only to a limited degree, in practice.

[15:09:01.0000] <zcorpan_>
yeah

[15:09:02.0000] <bros>
Is the WebSocket client-server handshaking process thought of as acceptable? 2 keys needing to be decoded/converted to integers/divided then a separate key needed to be concatenated with the other 2 back in string form, then hashed?

[15:09:03.0000] <AryehGregor>
I haven't seen an effective solution to get typical authors to write decent alt text.

[15:09:04.0000] <AryehGregor>
Would it be in the spec if no one thought it was acceptable?

[15:09:05.0000] <bros>
No.

[15:09:06.0000] <bros>
What is the purpose of it?

[15:10:00.0000] <AryehGregor>
I'm not a WebSocket expert, but my impression is it's meant to prevent various types of security problems.

[15:10:01.0000] <jgraham>
bros: It is designed to prevent cross-protocol attacks

[15:10:02.0000] <jgraham>
using web sockets to spoof non-web-socket protocols

[15:11:00.0000] <jgraham>
which would be bad if the servers talking those protocols weren't expecting random web pages to connect

[15:12:00.0000] <jgraham>
(which they aren't, in general)

[15:12:01.0000] <bros>
It is there so current HTTP servers will not have to be re-written, and only servers that want to handle WebSockets need to implement it?

[15:13:00.0000] <jgraham>
Hixie: why does drawFocusRing take a canDrawCustom argument rather than a function that would be called to draw a custom focus ring if needed?

[15:14:00.0000] <jgraham>
bros: That is a seperate issue really

[15:14:01.0000] <zcorpan_>
bros: i think http://www.ietf.org/mail-archive/web/hybi/current/msg01198.html identified some security problems with the old protocol

[15:15:00.0000] <jgraham>
The fact that you can implement WebSockets on top of existing HTTP servers is considered an advantage though

[15:32:00.0000] <MikeSmith>
as far as HTML5 processing of MathML and SVG in text/html, is it intended that a UA should actually render content in <annotation-xml> elements?

[15:32:01.0000] <MikeSmith>
because as far as I can see from checking in Minefield, it doesn't seem to be rendering SVG content in <annotation-xml>

[15:34:00.0000] <zcorpan_>
sicking: rich said 90% of web sites, not users

[15:35:00.0000] <zcorpan_>
MikeSmith: i thought Jacques was using svg in annotation-xml

[15:35:01.0000] <MikeSmith>
I thought he was too

[15:36:00.0000] <MikeSmith>
/me goes back to look at his site

[15:36:01.0000] <MikeSmith>
http://golem.ph.utexas.edu/~distler/blog/archives/001475.html

[15:40:00.0000] <MikeSmith>
but he's not serving that as text/html

[15:41:00.0000] <MikeSmith>
and when I try to process it as text/html, it does not seen to render as expected

[15:41:01.0000] <MikeSmith>
the SVG-in-MathML part, I mean

[15:44:00.0000] <MikeSmith>
ah wait

[15:45:00.0000] <MikeSmith>
after removing the namespace prefixes, it does

[15:48:00.0000] <zcorpan_>
seems annotation-xml is display:none in opera

[15:50:00.0000] <MikeSmith>
zcorpan_: that seems odd

[15:52:00.0000] <zcorpan_>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/508

[15:52:01.0000] <zcorpan_>
/me files a bug

[15:56:00.0000] <zcorpan_>
annotation, annotation-xml {display:none;} -- http://www.w3.org/TR/mathml-for-css/

[16:02:00.0000] <Dashiva>
Wouldn't it be fun if someone made an issue for removing author conformance reqs from the w3c spec and they weren't removed from the whatwg version?

[16:04:00.0000] <AryehGregor>
Well, all implementer requirements are already removed from the author version of the spec, so shouldn't be too much harder.

[16:04:01.0000] <Dashiva>
Hmm... what is left if you hide both author and implementer requirements, I wonder

[16:05:00.0000] <AryehGregor>
Non-normative text?

[16:10:00.0000] <MikeSmith>
zcorpan_: I also realize now that if you put MathML content into <annotation-xml>, it has to have a <math> element wrapped around it, right?

[16:10:01.0000] <MikeSmith>
I mean in order for it to be rendered correctly

[16:11:00.0000] <boblet>
hey Mike — you’re up too eh? poor us

[16:12:00.0000] <MikeSmith>
boblet: telcon :(

[16:12:01.0000] <zcorpan_>
MikeSmith: no idea

[16:12:02.0000] <boblet>
input type=email failing for kanji (unknown browser) http://twitter.com/pukupi/status/16308745936 following up…

[16:13:00.0000] <AryehGregor>
Does the type=email support in the browser in question handle IDNs?

[16:13:01.0000] <AryehGregor>
Nothing in the spec requires it to, IIRC.

[16:13:02.0000] <AryehGregor>
If it does accept IDNs, it has to ASCIIfy them before submitting.

[16:14:00.0000] <zcorpan_>
Hixie: hmm. i'm trying to connect to your websocket server from origin http://simon.html5.org but your server responds with Sec-WebSocket-Origin: http://software.hixie.ch :)

[16:15:00.0000] <boblet>
re: input type="email", it’s broken in Safari. I’m assuming that “1*( atext / "." ) "@" ldh-str 1*( "." ldh-str” in states-of-the-type-attribute.html#e-mail-state means any unicode character right?

[16:15:01.0000] <MikeSmith>
AryehGregor: I think it does accept IDNs

[16:15:02.0000] <MikeSmith>
iirc

[16:16:00.0000] <AryehGregor>
boblet, why don't you look it up in the referenced spec rather than assuming?

[16:16:01.0000] <zcorpan_>
Hixie: in particular this page http://simon.html5.org/dump/ws-limit.html to find out if there's a limit on the number of connections on certain networks

[16:16:02.0000] <MikeSmith>
/me looks back at the spec text about valid e-mail address

[16:17:00.0000] <boblet>
AryehGregor: your suggestion would assume that I could understand what the referenced specs actually said :| but yes, looking up now…

[16:18:00.0000] <AryehGregor>
Nope, regular domain names only.

[16:18:01.0000] <AryehGregor>
<ldh-str> ::= <let-dig-hyp> | <let-dig-hyp> <ldh-str>

[16:18:02.0000] <AryehGregor>
<let-dig-hyp> ::= <let-dig> | "-"

[16:18:03.0000] <AryehGregor>
<let-dig> ::= <letter> | <digit>

[16:18:04.0000] <AryehGregor>
<letter> ::= any one of the 52 alphabetic characters A through Z in

[16:18:05.0000] <AryehGregor>
upper case and a through z in lower case

[16:18:06.0000] <AryehGregor>
<digit> ::= any one of the ten digits 0 through 9

[16:18:07.0000] <AryehGregor>
Safari is allowed to accept IDNs, but it has to convert them to punycode.

[16:18:08.0000] <AryehGregor>
No one seems to do this yet, although the spec suggests it.

[16:19:00.0000] <boblet>
heh, thanks for the translation :)

[16:21:00.0000] <boblet>
well surely that’s a massive i18n disaster waiting to happen. I own several kanji domains, and type="email" would completely prevent me from using them, no? (unless “convert them to punycode” means they’re automatically translated to that crazy alphanumeric+hyphens representation of unicode domains)

[16:21:01.0000] <Hixie>
zcorpan_: yeah it's only a demo server

[16:21:02.0000] <AryehGregor>
boblet, yes, that's what it means.

[16:21:03.0000] <AryehGregor>
All domain names are really ASCII, non-ASCII stuff only exists for display to users.

[16:21:04.0000] <Hixie>
bros: there are many open source implementations of MD5 for C that you can use

[16:22:00.0000] <AryehGregor>
What browsers should do is display your domain name as Unicode, then convert it to punycode before submitting.

[16:22:01.0000] <boblet>
AryehGregor: aah, so it’s a bug with Safari not doing the punycode conversion then eh? will check bugzilla…

[16:22:02.0000] <Hixie>
jgraham: booleans was either to spec, a callback would be fine too i guess

[16:22:03.0000] <AryehGregor>
It's not a bug per spec, it's a quality-of-implementation issue.

[16:22:04.0000] <boblet>
thanks for your translation assistance

[16:22:05.0000] <AryehGregor>
You can still use your domain names, you just have to convert to punycode by hand.

[16:23:00.0000] <boblet>
that’s a pretty big quality-of-implementation issue, that to an end user would be the same as broken

[16:24:00.0000] <AryehGregor>
Yes, to the small minority of users who use IDNs in their e-mail addresses.

[16:24:01.0000] <AryehGregor>
Not a bug per the spec, though.

[16:24:02.0000] <AryehGregor>
Maybe it should be . . . at least in an "expected behavior" kind of way like the rendering section.

[16:25:00.0000] <AryehGregor>
Authors need to be able to expect reasonably consistent functionality here from a user perspective to want to use the feature, even if technically it doesn't affect interop.

[16:25:01.0000] <boblet>
AryehGregor: small minority now, but I think the non-English web is gonna change that in a few years (well, prolly more like 5-10 years)

[16:25:02.0000] <AryehGregor>
Probably.

[16:26:00.0000] <MikeSmith>
if I had an e-mail address with an IDN in it, I would definitely be using it now

[16:26:01.0000] <MikeSmith>
just for fun

[16:26:02.0000] <TabAtkins>
Me too.

[16:26:03.0000] <MikeSmith>
and to impress the girls

[16:26:04.0000] <boblet>
kanji domains are being promoted by registrars in Japan but still haven’t taken off, prolly due to poor mobile browser support. however I expect that’ll change within 3-4 years (cellphone average life 2 years)

[16:27:00.0000] <boblet>
MikeSmith TabAtkins want me to make one for you on オリ.jp? :P

[16:27:01.0000] <TabAtkins>
Is that kanji somehow related to my name?

[16:28:00.0000] <TabAtkins>
I need a vanity IDN.

[16:28:01.0000] <boblet>
nope, it’s “ori” or “oli” (my IRL name)

[16:28:02.0000] <TabAtkins>
What's a cheap registrat that'll do idn domain names?

[16:29:00.0000] <MikeSmith>
+1 to vanity IDN

[16:29:01.0000] <TabAtkins>
/me requires http://xn--_-ncfb.com.

[16:29:02.0000] <boblet>
TabAtkins: normally they’re country-specific. I used muumuu-domain.com for Japanese IDN, but their interface is … all Japanese :)

[16:30:00.0000] <TabAtkins>
Eh, looks like there are plenty of english ones.

[16:30:01.0000] <boblet>
that’s good

[16:30:02.0000] <TabAtkins>
Once I get my hands on $10, I'm totally registering ಠ_ಠ.com (or .whatever).

[16:31:00.0000] <AryehGregor>
WebKit's implementation of form validation is completely broken anyway.  It has no UI for it.

[16:31:01.0000] <AryehGregor>
So the form refuses to submit but doesn't tell you why.

[16:32:00.0000] <AryehGregor>
I tried to complain, but tkent disagreed with me, and other Chromium devs I bugged didn't say anything.  http://code.google.com/p/chromium/issues/detail?id=45804

[16:32:01.0000] <AryehGregor>
Does this ship in Safari 5 too?  Maybe I could bug Safari devs.

[16:32:02.0000] <boblet>
AryehGregor: yep IDN failure also in Safari

[16:33:00.0000] <AryehGregor>
/me solicits comment from othermaciej on <http://code.google.com/p/chromium/issues/detail?id=45804>, on the theory that it probably affects WebKit/Safari too

[16:49:00.0000] <MikeSmith>
zcorpan_: btw, I also note that svg content in annotation-xml has to be wrapped in an svg element (in order for Minefield to process it)

[16:58:00.0000] <zcorpan_>
MikeSmith: svg elements without an ancestor <svg> never render anything

[16:58:01.0000] <MikeSmith>
..and after re-reading the parsing algorithm, I realize that's exactly as expected

[16:58:02.0000] <zcorpan_>
that too

[16:58:03.0000] <MikeSmith>
zcorpan_: yeah, realize that now.. was being "slow on the uptake"

[16:59:00.0000] <zcorpan_>
what i meant was <circle xmlns='svg'/> in xml doesn't render a circle unless there's an <svg> ancestor

[16:59:01.0000] <MikeSmith>
OK


2010-06-17
[17:01:00.0000] <MikeSmith>
zcorpan_: so the problems with putting HTML content within annotation-xml is that the user would similarly need to wrap it in something, right? e.g, a <div>

[17:02:00.0000] <MikeSmith>
zcorpan_: I think I'm starting to agree that dealing with processing of content in annotation-xml might not be worth the trouble..

[17:02:01.0000] <TabAtkins>
What's the point of <annotation-xml>?  I feel like I've missed something.

[17:03:00.0000] <MikeSmith>
I guess it's for annotating non-presentational "content" MathML

[17:03:01.0000] <TabAtkins>
So you use it inside of MathML to represent an annotation?

[17:04:00.0000] <MikeSmith>
yeah

[17:04:01.0000] <TabAtkins>
So it's somewhat similar to SVG's <foreignContent> or whatever?

[17:04:02.0000] <MikeSmith>
yeah

[17:04:03.0000] <MikeSmith>
sorta

[17:05:00.0000] <zcorpan_>
http://www.w3.org/TR/MathML2/chapter4.html#contm.annotation-xml

[17:05:01.0000] <MikeSmith>
but currently different in one really significant way as far as inclusion in HTML5 goes

[17:05:02.0000] <MikeSmith>
which is that HTML content in  <foreignContent>  will get processed as expected

[17:05:03.0000] <MikeSmith>
but HTML content in <annotation-xml> will now

[17:05:04.0000] <MikeSmith>
*not

[17:06:00.0000] <MikeSmith>
which I raised a bug for : http://www.w3.org/Bugs/Public/show_bug.cgi?id=9887

[17:06:01.0000] <TabAtkins>
Oh.  I would naively think it should just be treated the same.

[17:07:00.0000] <MikeSmith>
yeah, I think that's what others are likely to expect too

[17:07:01.0000] <zcorpan_>
MikeSmith: from spec example, the content-mathml would go directly in <semantics>, while presentational-mathml would go in <annotation-xml>

[17:07:02.0000] <MikeSmith>
zcorpan_: yeah, that's how I understand it too

[17:07:03.0000] <MikeSmith>
(though probably I didn't say that clearly)

[17:07:04.0000] <zcorpan_>
which makes me wonder why the css profile spec makes it display:none

[17:08:00.0000] <othermaciej>
AryehGregor: if you think it's a WebKit bug, then please file it in bugs.webkit.org

[17:08:01.0000] <othermaciej>
AryehGregor: I agree that respecting the constraints without UI for validation failure is a bad state

[17:08:02.0000] <MikeSmith>
TabAtkins: but I don't think that just because it's what people expect, we necessarily need to make it work the same

[17:08:03.0000] <TabAtkins>
I think we need a good reason to make it not work the same.

[17:09:00.0000] <MikeSmith>
TabAtkins: I don't personally feel strongly about how it ends up getting resolved, but I think it does need to get resolved ... and maybe made clear the document-conformance parts of spec, too

[17:09:01.0000] <MikeSmith>
TabAtkins: "not worth the trouble" isn't a good reason?

[17:10:00.0000] <MikeSmith>
TabAtkins: users can put HTML content in the MathML <mtext> element instead

[17:10:01.0000] <TabAtkins>
Not worth the trouble to save implementor effort and make two elements act identically?

[17:10:02.0000] <MikeSmith>
TabAtkins: the whole <semantics>...<annotation-xml> thing in MathML is a arguably an ugly language hack to begin with

[17:11:00.0000] <MikeSmith>
well, Jacques Distler is one person that has said that, at least

[17:11:01.0000] <TabAtkins>
I wouldn't disagree.  I'm just wondering what the benefit is of having a special element that is *almost* the same as SVG's foreignContent, but different in subtle ways.

[17:12:00.0000] <MikeSmith>
true

[17:12:01.0000] <zcorpan_>
TabAtkins: annotation-xml needs to support embedding mathml elements

[17:12:02.0000] <zcorpan_>
TabAtkins: that's incompatible with how foreignContent works

[17:12:03.0000] <MikeSmith>
ah yeah

[17:13:00.0000] <MikeSmith>
so it's not really "foreign content" ... it's more like "maybe foreign content, or maybe not"

[17:14:00.0000] <MikeSmith>
btw, Jacques comment: "I really had no idea how to properly include SVG fragments inside a MathML equation. It turns out that the “correct” answer is some horrible kludge, involving embedding the SVG in the MathML <semantics> element. Go figure …"

[17:14:01.0000] <TabAtkins>
Ah, so it needs to support mathml children that aren't wrapped in <math>?

[17:14:02.0000] <MikeSmith>
(from http://golem.ph.utexas.edu/~distler/blog/archives/001475.html)

[17:14:03.0000] <zcorpan_>
TabAtkins: yes

[17:14:04.0000] <TabAtkins>
That's weird.

[17:14:05.0000] <bros>
I am at a loss. I can not implement the handshaking element of WebSockets. Would anybody here be willing to review a paste of the code and see if my errors pop out?

[17:15:00.0000] <MikeSmith>
but fwiw, Minefield doesn't actually seem to render MathML content in <annotation-xml> unless it's wrapped in an <math> element

[17:16:00.0000] <MikeSmith>
afaict, at least

[17:17:00.0000] <zcorpan_>
/me might be wrong

[17:19:00.0000] <Hixie>
bros: ure

[17:19:01.0000] <Hixie>
sure even

[17:19:02.0000] <zcorpan_>
  <annotation-xml encoding="MathML-Content">

[17:19:03.0000] <zcorpan_>
    <apply>

[17:19:04.0000] <zcorpan_>
http://www.w3.org/TR/MathML2/chapter2.html#fund.mix

[17:20:00.0000] <TabAtkins>
Every single time someone directs a response to bros, I think they're addressing the room like a surfer dude.  "Bros, I submit that these waves are totally sweet."

[17:20:01.0000] <MikeSmith>
zcorpan_: as far as I can tell from the parsing algorithm, it should not be required.. but as far as I can tell from actually testing it in Minefield, Minefield seems to require it

[17:20:02.0000] <zcorpan_>
  <annotation-xml encoding="MathML-Presentation">

[17:20:03.0000] <zcorpan_>
    <mrow>

[17:20:04.0000] <MikeSmith>
TabAtkins: :)

[17:20:05.0000] <zcorpan_>
certainly the spec examples don't use a wrapping <math>

[17:21:00.0000] <TabAtkins>
So I guess the next thing to do is see if there's much mathml in the wild that uses <annotation-xml> with mathml children that aren't wrapped in <math>.

[17:21:01.0000] <TabAtkins>
The spec examples can just be fixed if not.

[17:21:02.0000] <MikeSmith>
zcorpan_: yeah, I know. I copied and pasted some of them.. but can't get them to render in text/html unless I put <math> around them

[17:22:00.0000] <zcorpan_>
TabAtkins: there isn't much mathml in the wild

[17:22:01.0000] <bros>
Hixie: Thank you. http://codepad.org/YBQD16oO The problem I face is no matter what, the WebSocket's readyState is always closed, which makes me think I am sending something wrong.

[17:22:02.0000] <TabAtkins>
zcorpan_: Then let's fix the spec and make the content model of <annotation-xml> more reasonable.

[17:23:00.0000] <zcorpan_>
TabAtkins: i don't see how it's not reasonable

[17:23:01.0000] <boblet>
seems that webkit is being smart and running type="email" validation before submit — getting failure for IDN email addys on non-submit ajax form

[17:24:00.0000] <Hixie>
bros: well to start with, read() might not return the whole handshake at once

[17:24:01.0000] <Hixie>
bros: read() just returns what the kernel has received so far, which could be on byte, two bytes, the whole handshake plus a few frames, anything really

[17:24:02.0000] <TabAtkins>
zcorpan_: It seems more reasonable that every direct child of an <annotation-xml> have a root element, rather than just "every child that isn't mathml".

[17:24:03.0000] <bros>
Hixie: Very true. I can confirm it does in this case. I removed all of my output debugging. It does though, all the way to the end of the 3rd part of the key.

[17:26:00.0000] <Hixie>
bros: is this code running on a public server anywhere?

[17:26:01.0000] <zcorpan_>
Hixie: a websocket client wouldn't send frames along with the handshake

[17:26:02.0000] <Hixie>
zcorpan_: not a browser one, anyway

[17:26:03.0000] <bros>
Hixie: No. I could make it.

[17:27:00.0000] <Hixie>
bros: that would be helpful

[17:27:01.0000] <zcorpan_>
Hixie: iirc, the protocol doesn't allow a conforming websocket client to do it

[17:27:02.0000] <Hixie>
bros: btw there's a security bug in your code if i'm not mistaken -- if i connect to the server and send only a few bytes, it'll crash with an out-of-bounds buffer read

[17:28:00.0000] <Hixie>
zcorpan_: ok, not a conforming one then :-)

[17:28:01.0000] <bros>
Hixie: I wouldn't doubt it. I have been trying very hard to just get this, working. What line specifically?

[17:28:02.0000] <bros>
Hixie: Permission to private message you?

[17:28:03.0000] <Hixie>
bros: the "keys[8] = buf[r - 8];" stuff doesn't check that r > 8

[17:28:04.0000] <Hixie>
bros: sure

[17:29:00.0000] <zcorpan_>
bros: why not discuss it here? people can learn from it

[17:29:01.0000] <Hixie>
i imagine he wants to send me his IP address :-)

[17:29:02.0000] <bros>
http://98.249.142.158/

[17:29:03.0000] <zcorpan_>
oh

[17:29:04.0000] <Hixie>
bros: which port is the ws server on?

[17:31:00.0000] <bros>
Hixie: 8080. The link I linked is a page with new WebSocket()

[17:33:00.0000] <Hixie>
bros: i don't have a good way to sniff the network here. I'll have to go home and do it. Will you still be around in ~ an hour?

[17:33:01.0000] <bros>
I have Sec-WebSocket-Origin hard-coded.

[17:33:02.0000] <bros>
Hixie: Most likely. Let me work on making the code "better".

[17:34:00.0000] <Hixie>
k. bbiab.

[17:34:01.0000] <zcorpan_>
bros: WebSocket handshake failure, expected field 'sec-websocket-location' to have value 'ws://98.249.142.158:8080/', found 'ws://98.249.142.158'.

[17:34:02.0000] <bros>
That is something. Thank you. How did you find that error? I was using Chrome developer tools and couldn't find anything.

[17:35:00.0000] <zcorpan_>
i used an internal build of opera :)

[17:36:00.0000] <bros>
very nice. :)

[17:37:00.0000] <MikeSmith>
zcorpan_: um, please ignore what I said about needing to wrap MathML content in <math> to get it work in annotation-xml in Minefield... you don't need to .. not sure what I was doing wrong before

[17:37:01.0000] <MikeSmith>
clearly a sign that I need a morning coffee break

[17:38:00.0000] <MikeSmith>
/me goes to make coffee

[17:44:00.0000] <AryehGregor>
othermaciej, I don't know if it's a WebKit bug, since I normally only use Linux, and there don't seem to be non-Chromium WebKit builds readily available for Linux.

[17:44:01.0000] <AryehGregor>
I could try using a Windows machine, but I bet there are lots of people here with copies of Safari or WebKit around.

[17:44:02.0000] <AryehGregor>
To those people: in the data URL data:text/html,<!doctype html><form><input required><input><button>Submit</button></form>

[17:45:00.0000] <AryehGregor>
Enter text into the second field, leaving the first blank, and try to submit.  Does it work?

[17:45:01.0000] <MikeSmith>
AryehGregor: you know about Arora? (QtWebkit-based Linux browser)

[17:45:02.0000] <zcorpan_>
AryehGregor: it focuses the first field

[17:45:03.0000] <AryehGregor>
Nope.  Should I try in that

[17:45:04.0000] <AryehGregor>
?

[17:45:05.0000] <AryehGregor>
zcorpan_, right, so that's the bug.

[17:45:06.0000] <AryehGregor>
I'll file at webkit.org, then.

[17:46:00.0000] <MikeSmith>
AryehGregor: yeah -- Arora is actually a pretty decent browser

[17:47:00.0000] <MikeSmith>
AryehGregor: and there's an #arora channel on freenode where you can bug the developers when needed

[17:48:00.0000] <MikeSmith>
http://code.google.com/p/arora/

[17:49:00.0000] <GPHemsley>
Am I right in interpreting that <input> and other form elements do not actually require a <form> parent?

[17:50:00.0000] <MikeSmith>
GPHemsley: depends on if you want them to actually do anything

[17:50:01.0000] <MikeSmith>
they do render without a <form> ancestor

[17:50:02.0000] <GPHemsley>
if I don't need them to submit with a form, then it's valid?

[17:51:00.0000] <MikeSmith>
yes

[17:51:01.0000] <MikeSmith>
as far as document conformance goes

[17:51:02.0000] <GPHemsley>
ok, cool

[17:55:00.0000] <AryehGregor>
othermaciej, https://bugs.webkit.org/show_bug.cgi?id=40747

[17:59:00.0000] <othermaciej>
AryehGregor: ty

[18:19:00.0000] <Hixie>
bros: did you get it working?

[18:23:00.0000] <karlcow>
/me wonders if there is a use case where an input element has to be in the document without a form and then if not why is it conformant.

[18:23:01.0000] <karlcow>
not that important maybe.

[18:24:00.0000] <Dashiva>
Anything that doesn't submit to the server is justified outside a form

[18:24:01.0000] <Dashiva>
E.g. a javascript calculator

[18:25:00.0000] <karlcow>
ah I see the use case, UI widget for client side application

[18:28:00.0000] <AryehGregor>
"FWIW I think Ian’s, er, disregard for social-niceties  is sub-optimal. But at this point, I’m so frustrated with the glacial progress of web markup since 1997, I don’t care if he tells the w3c to go fuck itself and publishes HTML5 in limited editions by personally carving it into the skins of dead baby seals. Or ex-XHTML2 WG members. Using Princess Diana’s hipbone. While stamping on the Dali Lama’s spectacles. On stage. On fire. In

[18:28:01.0000] <AryehGregor>
 front of the Queen. Just get it out."

[18:30:00.0000] <Dashiva>
Well, that's certain an opinion!

[18:30:01.0000] <Dashiva>
*certainly

[18:30:02.0000] <boblet>
I MIS

[18:31:00.0000] <boblet>
I misread it as “just get it out … (in) front of the Queen” to start with (reading from the bottom up

[18:31:01.0000] <AryehGregor>
Writing HTML5 on the skin of dead XHTML2 WG members would certainly provide a strong incentive for working groups to pay attention to user needs.

[18:33:00.0000] <Dashiva>
I liked this one too: http://twitter.com/SickingJ/status/16284113128

[19:01:00.0000] <MikeSmith>
AryehGregor: I sense a subtle hint of frustration in whoever wrote that

[19:01:01.0000] <MikeSmith>
but maybe that's just me

[19:01:02.0000] <MikeSmith>
I'm not a psychologist

[19:01:03.0000] <AryehGregor>
MikeSmith, that's one possible interpretation of that comment.

[19:01:04.0000] <MikeSmith>
heh

[19:01:05.0000] <AryehGregor>
Perhaps he was just obliquely expressing his antipathy toward baby seals.

[19:04:00.0000] <MikeSmith>
AryehGregor: did the CSS Squirrel write that? Or somebody from the Last Week in HTML5 collective? I ask because it seems to have the same kind of subtle hint of frustration that characterizes those works

[19:04:01.0000] <Dashiva>
I don't agree

[19:05:00.0000] <Dashiva>
It seems very different from last week, at least

[19:05:01.0000] <AryehGregor>
MikeSmith, both of those parties would tend to express the opposite viewpoint.  It was a commenter on CSS Squirrel's latest post.

[19:05:02.0000] <AryehGregor>
"Mattur".

[19:05:03.0000] <AryehGregor>
http://www.cssquirrel.com/2010/06/15/693comic-update-html5-unicorn-heuristics/#comment-31994

[19:09:00.0000] <Dashiva>
http://www.cssquirrel.com/comic/?comic=62

[19:09:01.0000] <Dashiva>
I'm pretty sure this isn't a real language

[19:09:02.0000] <othermaciej>
that is an amusing comment

[19:14:00.0000] <Dashiva>
Wow, the blog post itself is pretty ignorant

[19:15:00.0000] <AryehGregor>
Yes, most of Kyle's posts on HTML5 process are pretty ignorant.

[19:15:01.0000] <AryehGregor>
Kind of a shame.

[19:17:00.0000] <AryehGregor>
Good thing we don't have to give those ignorant non-implementers any kind of say in anything important, am I right?

[19:19:00.0000] <MikeSmith>
AryehGregor, Dashiva - I see. I guess it was the subtlety of his wording that threw me off

[19:19:01.0000] <MikeSmith>
the felicitous language

[19:20:00.0000] <MikeSmith>
He should state himself more strongly

[19:20:01.0000] <MikeSmith>
you know, tell us what he really thinks

[19:20:02.0000] <MikeSmith>
instead of beating around the bush

[19:22:00.0000] <AryehGregor>
Maybe you should comment on the blog to advise him.

[19:22:01.0000] <Dashiva>
If it wasn't for the point espoused, I could have been convinced it was squirrel writing the comment

[19:22:02.0000] <AryehGregor>
Yes, the tone is very much like him.

[19:22:03.0000] <Dashiva>
But I don't think he would sock puppet himself that way...

[19:22:04.0000] <AryehGregor>
People tend to pick up the tone of the sites they post on, I notice.

[19:23:00.0000] <AryehGregor>
If you go to the discussion forums for Dinosaur Comics, for example, you'll find everyone writing in run-on sentences and spouting non sequiturs.

[19:23:01.0000] <MikeSmith>
Mattur is definitely no sock puppet

[19:24:00.0000] <MikeSmith>
I think he does some self-sock-puppetry occasionally

[19:24:01.0000] <MikeSmith>
But I do that too

[19:24:02.0000] <MikeSmith>
in the privacy of my own home, of course

[19:24:03.0000] <MikeSmith>
but hey, we all do that, right?

[19:24:04.0000] <Dashiva>
I may have inadvertently sock puppeted myself with my second IRC client

[19:24:05.0000] <AryehGregor>
. . . what?

[19:25:00.0000] <Dashiva>
I don't think anyone in here know who it is

[19:25:01.0000] <MikeSmith>
I like the Squirrel too

[19:25:02.0000] <MikeSmith>
that Beavers Taste Great with Mustard stuff is solid gold

[19:25:03.0000] <karlcow>
Dali Lama is the extravaganza of buddhism

[19:25:04.0000] <MikeSmith>
he needs to give us more stuff like that

[19:26:00.0000] <MikeSmith>
and more violence, and more nudity

[19:26:01.0000] <MikeSmith>
if not outright intercourse

[19:26:02.0000] <Dashiva>
He makes good comics, and he serves a purpose, I just wish his blog posts would stick to facts and rational criticism

[19:27:00.0000] <MikeSmith>
as Karl Marx once said, there is nothing funnier in the entire world than depictions of cartoon characters have sex with each other

[19:27:01.0000] <Dashiva>
I don't think he said that

[19:27:02.0000] <MikeSmith>
oh, I am pretty sure he did

[19:27:03.0000] <MikeSmith>
it was either him or Albert Einstein

[19:27:04.0000] <Dashiva>
Unless by Karl Marx you mean Mark Twain and Oscar Wilde (co-authored)

[19:27:05.0000] <MikeSmith>
I think the Dalai Lama even quoted that one

[19:28:00.0000] <MikeSmith>
Karl Marx said Huckleberry Finn is his favorite book of all time

[19:28:01.0000] <MikeSmith>
and he also said Oscar Wilde is crap

[19:28:02.0000] <Dashiva>
Is the Dalai Lama allowed to find things funny, or does that ruin his chances for nirvana?

[19:29:00.0000] <MikeSmith>
and Marx said that Nabokov is "even more crap"

[19:29:01.0000] <karlcow>
Dashiva: rationalism… comics… hmmm I'm worried about mental health of Dashiva now. do you have Specification orgasm? :p

[19:29:02.0000] <MikeSmith>
Dalai Lama is a punk, seriously

[19:29:03.0000] <MikeSmith>
or a chump

[19:29:04.0000] <MikeSmith>
or a jackass

[19:30:00.0000] <Dashiva>
I think MikeSmith had a few too many beer breaks today

[19:30:01.0000] <othermaciej>
most exciting thing about those comments: reading Shelley's confident predictions of what Apple would do in particular situations

[19:30:02.0000] <Dashiva>
othermaciej: Or the comment about all the browser companies leaving WHATWG

[19:31:00.0000] <MikeSmith>
/me struggles to find a good other term that accurately captures the nature of the petty fraud that is the Dalai Lama's entire life

[19:31:01.0000] <karlcow>
hehe

[19:31:02.0000] <Dashiva>
MikeSmith: Sort of like every other holy man and woman after a little close examination?

[19:31:03.0000] <MikeSmith>
no, no

[19:32:00.0000] <karlcow>
MikeSmith: what about Dalai Lamarketing

[19:32:01.0000] <MikeSmith>
some holy men are genuine

[19:32:02.0000] <MikeSmith>
like Pope Ratzinger

[19:32:03.0000] <MikeSmith>
he will be a Saint one day

[19:32:04.0000] <Dashiva>
His excuses are full of holes, yes

[19:32:05.0000] <MikeSmith>
Saint Ratzinger

[19:32:06.0000] <MikeSmith>
Dashiva: beer bong

[19:32:07.0000] <MikeSmith>
not beer

[19:32:08.0000] <MikeSmith>
today

[19:32:09.0000] <MikeSmith>
http://www.hecklerspray.com/owen-wilson-vs-the-dalai-lama/2005686.php

[19:33:00.0000] <MikeSmith>
"corny purveyor of goofball ideas

[19:33:01.0000] <Dashiva>
karlcow: Also, I'm just acting as MikeSmith's foil, I don't think it's fair to make judgements about my own state

[19:33:02.0000] <MikeSmith>
that's another site filled with great subtlety of expression, btw

[19:33:03.0000] <karlcow>
Ratzinger could fit for a beer name

[19:34:00.0000] <MikeSmith>
Pope Ratzinger Beer

[19:34:01.0000] <MikeSmith>
I would buy it

[19:34:02.0000] <Dashiva>
Is there a beer you wouldn't buy?

[19:34:03.0000] <Dashiva>
Except, like, Asahi

[19:35:00.0000] <MikeSmith>
Ringnes

[19:36:00.0000] <Dashiva>
Nice google-fu

[19:36:01.0000] <karlcow>
There is a beer I would not buy. Budweiser, though it doesn't really qualify as a beer

[19:36:02.0000] <Dashiva>
Yeah, it was a trick question

[19:37:00.0000] <MikeSmith>
othermaciej: are you suggesting that Apple doesn't make critical technology decisions based on confidently stated third-party assertions found in blog comments?

[19:38:00.0000] <Dashiva>
Apple doesn't make forward-looking statements, so he can't answer your question

[19:38:01.0000] <othermaciej>
^ this

[19:39:00.0000] <karlcow>
Apple doesn't decide. Steve does.

[19:39:01.0000] <karlcow>
/me notes he has not drank, just ate chocolate :p

[19:40:00.0000] <MikeSmith>
Dashiva: no google-fu involved.. I have experienced the wonder that is Ringnes firsthand... had my life's fill of Ringnes, and then some

[19:41:00.0000] <MikeSmith>
Norway needs to do something about its affordable-beer situation

[19:41:01.0000] <bros>
Should /query Hixie Hello

[19:42:00.0000] <karlcow>
Ringnes NOR way

[19:46:00.0000] <myakura>
uh oh http://fantasai.inkedblade.net/style/specs/css2.1/px-unit

[19:48:00.0000] <myakura>
it doesn't look right in html5-enabled-Gecko, </ins> doesn't close the <p>.

[19:49:00.0000] <MikeSmith>
myakura: which part is rendering wrong?

[19:49:01.0000] <MikeSmith>
it seems to look correct in my Minefield

[19:50:00.0000] <myakura>
Chrome 6 (dev) doesn't close either, while Opera 10.60 beta and IE8 does.

[19:50:01.0000] <kbrosnan>
should have a lot less green

[19:50:02.0000] <MikeSmith>
or at least it looks the same as it does in other browsers

[19:50:03.0000] <MikeSmith>
oh

[19:50:04.0000] <myakura>
MikeSmith: lines 45-47. <ins><p>In cases where the <a href="cascade.html#usedValue">used</a>  length cannot be supported, user agents must approximate it in the <a href="cascade.html#actual-value">actual value.</a></ins> 

[19:50:05.0000] <myakura>
oops

[19:51:00.0000] <MikeSmith>
yeah, I see now

[19:56:00.0000] <AryehGregor>
/me needs to block tvtropes.org at the firewall or something to prevent himself from gratuitously staying up late and wasting huge amounts of time.

[19:56:01.0000] <AryehGregor>
(although the "Jews in Space" skit it linked to was possibly awesome enough to be worth it this time)

[19:56:02.0000] <kbrosnan>
livedom.validator.nu gives the same dom tree as Firefox trunk

[20:52:00.0000] <MikeSmith>
man, I really am slow on the uptake

[20:52:01.0000] <MikeSmith>
I am just now realizing that annotation-xml content is, well, annotations

[20:53:00.0000] <MikeSmith>
with the implication that it's not displayed in the actual body text of a document

[20:53:01.0000] <MikeSmith>
but made available in some other way

[20:53:02.0000] <MikeSmith>
I guess

[20:53:03.0000] <MikeSmith>
a pop-up

[20:53:04.0000] <MikeSmith>
or whatever

[20:54:00.0000] <MikeSmith>
at least I think that's the actual intent

[20:54:01.0000] <MikeSmith>
based on seeing the rule that Simon cited in http://www.w3.org/TR/2009/CR-mathml-for-css-20091215/#d1e2465

[20:55:00.0000] <MikeSmith>
default UA stylesheet

[20:55:01.0000] <MikeSmith>
annotation, annotation-xml

[20:55:02.0000] <MikeSmith>
 {display:none;}

[20:55:03.0000] <MikeSmith>
which Opera apparently conforms to

[20:55:04.0000] <MikeSmith>
but Firefox does not seem to

[21:06:00.0000] <MikeSmith>
ah geez

[21:07:00.0000] <MikeSmith>
Opera Dragonfly apparently does not let you edit CSS from "local user stylesheets"

[21:07:01.0000] <MikeSmith>
of which I guess the MathML stylesheet it uses is one

[21:09:00.0000] <MikeSmith>
/me finds ./Contents/Resources/styles/mathml.css

[21:13:00.0000] <boblet>
re: earlier conversation about IDNs silently failing <input type="email"> validation (blocking form submission with zero user feedback) http://pukupi.com/post/2070/ Ouchies

[21:13:01.0000] <boblet>
Opera bug filed

[21:19:00.0000] <MikeSmith>
after reading more, I find the whole <semantics><annotation-xml> markup in MathML to be seriously overloaded

[21:20:00.0000] <MikeSmith>
annotations can either be "content" annotations (=non-presentational) or presentational annotations

[21:20:01.0000] <MikeSmith>
and it is not clear at all from the spec under what circumstances (if any) there are meant to be presented to the user by default

[21:21:00.0000] <MikeSmith>
and it is not clear at all from the spec under what circumstances (if any) there are meant to be presented to the user by default."

[21:22:00.0000] <MikeSmith>
oops

[21:22:01.0000] <MikeSmith>
"A particularly important case is the use of a presentation MathML expression to indicate a preferred rendering for a content MathML expression."

[21:22:02.0000] <MikeSmith>
"Note that when a presentation MathML annotation is present in a semantics element, it may be used as the default rendering of the semantics element, instead of the default rendering of the first child."

[21:23:00.0000] <MikeSmith>
..."may be used as the default rendering"

[21:24:00.0000] <karlcow>
my house is a computer code. No tag soup there. - Sim City in Ixtapaluca, Mexico - http://hydeparkblvd.com/?p=1494

[21:26:00.0000] <MikeSmithX>
"may be used as the default rendering" = "might also not be used as the default rendering, it's up to you!"

[23:13:00.0000] <kennyluck>
Hmm... MikeSmith told me that the browser vendors are not that interested in W3C Device API (http://www.w3.org/2009/dap/ ) because of the policy framework stuff (XACML?). But are there any plans for the browser vendors on implementation things like the Contact API or Calendar API ?

[23:55:00.0000] <boblet>
is there anything in the spec on what UAs should do when an input fails type validation?

[23:56:00.0000] <Hixie>
yes

[23:57:00.0000] <boblet>
hey Hixie — following checkedness link now, I may be getting close :)

[00:03:00.0000] <MikeSmith>
kennyluck: Mozilla has some interest in the Contacts API

[00:04:00.0000] <MikeSmith>
iirc

[00:07:00.0000] <MikeSmith>
kennyluck: http://lists.w3.org/Archives/Public/public-device-apis/2010Jun/0147.html

[00:07:01.0000] <boblet>
Hixie: atm IDN values in type="email" cause silent submission failure in all supporting browsers. any suggestions apart from using type="text"? i.e. should I file a bug to request spec say if input fails validation this should be communicated to the user?

[00:07:02.0000] <MikeSmith>
[[

[00:07:03.0000] <MikeSmith>
Where 

[00:07:04.0000] <MikeSmith>
specifications worked on in the DAP WG lend themselves to implementation 

[00:07:05.0000] <MikeSmith>
plans, I think Mozilla participants interested in these can comment on 

[00:07:06.0000] <MikeSmith>
them (e.g. Contacts API, at least for now).

[00:07:07.0000] <MikeSmith>
]]

[00:08:00.0000] <boblet>
or is there a way to override per-input validation (or browser default chrome for eg date picker)?

[00:19:00.0000] <MikeSmith>
boblet: is there not an attribute for that?

[00:19:01.0000] <MikeSmith>
"novalidate" or whatever

[00:19:02.0000] <boblet>
MikeSmith: seems to be for form

[00:19:03.0000] <boblet>
let me doublecheck

[00:20:00.0000] <boblet>
oh nice, yeah it’s also an attribute on input

[00:20:01.0000] <boblet>
doh

[00:20:02.0000] <boblet>
/me must skim less

[00:21:00.0000] <MikeSmith>
only on input/type=submit, right?

[00:23:00.0000] <boblet>
aah yeah, checking the table we can only use formnovalidate on submit (and novalidate on form)

[00:23:01.0000] <boblet>
so all or nothing

[00:29:00.0000] <kennyluck>
Thanks, MikeSmith. I was trying to find a relevant bug in bugzilla@Mozilla but found nothing related.

[00:30:00.0000] <MikeSmith>
kennyluck: see also http://lists.w3.org/Archives/Public/public-webapps/2010AprJun/0909.html

[00:30:01.0000] <MikeSmith>
though that was qualified a bit further in http://lists.w3.org/Archives/Public/public-webapps/2010AprJun/0911.html

[00:31:00.0000] <MikeSmith>
not that I necessarily agree with any of those assessments

[00:31:01.0000] <MikeSmith>
but just to give you an sense of some of the discussion that's been taking place

[00:32:00.0000] <kennyluck>
/me is reading

[00:33:00.0000] <MikeSmith>
kennyluck: reading through the whole thread at http://lists.w3.org/Archives/Public/public-device-apis/2010Jun/thread.html#msg128 would not be a waste of time

[00:33:01.0000] <MikeSmith>
in particular the messages from Arun and Brian

[00:34:00.0000] <kennyluck>
Yeah, I am actually following those "in-reply-to" because I am a bit confused.

[00:56:00.0000] <Hixie>
boblet: do you mean non-ASCII text input or non-ASCII value=""s? and why is it silent?

[00:58:00.0000] <boblet>
non-ASCII text input (eg example@日本語.jp), and apparently there’s no indication to the user that this fails validation, and that in doing so the form can’t be submitted.

[00:59:00.0000] <boblet>
this has been a problem for a long time due to \w in regex treating double-byte characters as if they had a space on each side (again apparently)

[01:02:00.0000] <Hixie>
sounds like a lot of bugs

[01:02:01.0000] <Hixie>
file them :-)

[01:04:00.0000] <Hixie>
looks like the spec doesn't mention it, but really if the user types in a unicode e-mail address it should get converted to punycode

[01:06:00.0000] <boblet>
Hixie: heh http://code.google.com/p/chromium/issues/detail?id=45804 by AryehGregor and https://bugs.webkit.org/show_bug.cgi?id=40761 by me

[01:08:00.0000] <boblet>
the word is using formnovalidate gets around this, so I’m planning to use that with type="email" for iPhone keyboard (mobile-targetted site)

[01:09:00.0000] <boblet>
I just wonder given how broken this is if adding a note about error handling to the spec would be a good idea, ref: http://krijnhoetmer.nl/irc-logs/whatwg/20100617#l-61

[01:10:00.0000] <boblet>
focussing the form field of the first validation error is not equal to providing a user UI to this

[01:10:01.0000] <Hixie>
agreed

[01:11:00.0000] <Hixie>
feel free to file a bug on the spec or send mail to the mailing list if you think the spec needs updating a bit

[01:11:01.0000] <Hixie>
though i'm way behind on dealing with feedback

[01:12:00.0000] <boblet>
rgr, will do

[01:27:00.0000] <MikeSmith>
Hixie: would it be possible to have the parsing algorithm just read the contents of annotation-xml elements into the DOM but do any processing on them at all?

[01:27:01.0000] <MikeSmith>
that is, even if those contents are SVG

[01:27:02.0000] <MikeSmith>
or even if the contents are MathML

[01:28:00.0000] <MikeSmith>
*into the DOM but _not_ do any processing on them at all

[01:29:00.0000] <MikeSmith>
for one thing, those contents are, after all, annotations that are not intended to be rendered in the main text flow

[01:30:00.0000] <MikeSmith>
both Gecko and Presto have default UA stylesheets with "annotation-xml { display: none }"

[01:31:00.0000] <othermaciej>
what is annotation-xml?

[01:31:01.0000] <MikeSmith>
and the MathML spec says: "The default rendering of a semantics element is the default rendering of its first child"

[01:31:02.0000] <MikeSmith>
where the first child is <mn> or whatever

[01:32:00.0000] <MikeSmith>
and the second or subsequent children are annotation-xml elements

[01:32:01.0000] <MikeSmith>
othermaciej: it's a mechanism in MathML for providing an "alternate representation" of MathML content

[01:33:00.0000] <Hixie>
MikeSmith: "processing"?

[01:33:01.0000] <MikeSmith>
http://www.w3.org/TR/MathML3/chapter5.html#mixing.alternate.representations

[01:33:02.0000] <MikeSmith>
Hixie: not render it as whatever it would otherwise render it as

[01:33:03.0000] <MikeSmith>
even if it's SVG, don't render it as such

[01:33:04.0000] <Hixie>
aren't they display:none?

[01:33:05.0000] <MikeSmith>
yeah

[01:33:06.0000] <Hixie>
by default?

[01:33:07.0000] <MikeSmith>
by default

[01:33:08.0000] <MikeSmith>
yeah

[01:34:00.0000] <Hixie>
so...?

[01:34:01.0000] <othermaciej>
MikeSmith: does MathML call for it to be display: none, or is that convergent evolution among Gecko and Presto?

[01:34:02.0000] <MikeSmith>
so, that's overrideable

[01:34:03.0000] <Hixie>
isn't that enough?

[01:34:04.0000] <othermaciej>
if annotation-xml is itself display: none, nothing inside it will render by default

[01:35:00.0000] <MikeSmith>
Hixie: ewhat if a site stylesheet sets it to actually display?

[01:36:00.0000] <Hixie>
then we'd want it to display, no?

[01:36:01.0000] <Hixie>
why else would it be set to display?

[01:38:00.0000] <MikeSmith>
OK, the problem currently is that the parsing algorithm doesn't allow you to put HTML content into <annotation-xml> elements

[01:38:01.0000] <MikeSmith>
but I think users will expect to be able to put HTML content there

[01:39:00.0000] <MikeSmith>
just as they can put SVG and MathML

[01:39:01.0000] <Hixie>
not sure how to allow it without hard-coding a list of elements

[01:39:02.0000] <MikeSmith>
yeah

[01:40:00.0000] <MikeSmith>
it could be allowed by allowing it

[01:41:00.0000] <Hixie>
annotation-xml didn't seem important enough to hard-code a lot of elements there

[01:41:01.0000] <MikeSmith>
and instead treating SVG and MathML elements in <annotation-xml> as arbitrary unknown elements

[01:41:02.0000] <Hixie>
it'd be a lot of complexity in every parser for a very obscure case

[01:41:03.0000] <MikeSmith>
Hixie: I fully agree it's not important

[01:42:00.0000] <Hixie>
i supported svg since that seemed useful and easy to support

[01:42:01.0000] <Hixie>
some mathml examples do it

[01:42:02.0000] <MikeSmith>
yeah, I know the examples do

[01:43:00.0000] <MikeSmith>
but I don't know if there are any real-world use cases for it in Web content

[01:43:01.0000] <MikeSmith>
other than those examples

[02:10:00.0000] <MikeSmith>
Hixie: another (hopefully easy to answer question) -

[02:10:01.0000] <MikeSmith>
the MathML spec current says this: "The draft HTML5 Recommendation, [HTML5], contains details for how MathML is included in HTML5 (and XHTML5)."

[02:11:00.0000] <Hixie>
not fully true yet but it will be

[02:11:01.0000] <MikeSmith>
well, I think it could be refined

[02:12:00.0000] <MikeSmith>
to say this:

[02:12:01.0000] <MikeSmith>
> >   The draft HTML5 Recommendation, [HTML5], contains details for

[02:12:02.0000] <MikeSmith>
> >   how MathML is included in HTML5 (and XHTML5), and, in turn,

[02:12:03.0000] <MikeSmith>
> >   details on where HTML content is allowed to be included in

[02:12:04.0000] <MikeSmith>
> >   MathML content within HTML documents.

[02:12:05.0000] <MikeSmith>
to make it more clear that doesn't just say how MathML can be included in HTML5

[02:13:00.0000] <Hixie>
wfm

[02:13:01.0000] <MikeSmith>
OK

[02:13:02.0000] <MikeSmith>
no suggestions for tweaking it any further?

[02:13:03.0000] <Hixie>
looks good to me

[02:13:04.0000] <MikeSmith>
OK, thanks

[02:14:00.0000] <othermaciej>
MikeSmith: it seems inaccurate to refer to the HTML5 Working Draft as "The draft HTML5 Recommendation"

[02:14:01.0000] <othermaciej>
or at least, imprecise

[02:14:02.0000] <Hixie>
draft HTML5 standard? :-)

[02:15:00.0000] <othermaciej>
I forget, does the W3C make standards this week?

[02:15:01.0000] <MikeSmith>
the W3C makes technologies

[02:15:02.0000] <MikeSmith>
normal people don't care about standards

[02:16:00.0000] <MikeSmith>
they do care about new technologies

[02:16:01.0000] <MikeSmith>
shiny stuff

[02:16:02.0000] <asmodai>
One thing raised on the Unicode list is that right now it is hard for i18n work to know which character ranges a particular UA supports. That way if someone lacks a particular font you could use some way (JavaScript?) to supply missing support via font-face. There's no plans for any API or detection way for supported scripts I guess?

[02:16:03.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/ says "Draft Standard"

[02:17:00.0000] <MikeSmith>
othermaciej, Hixie - I don't see why they could not just refer to there simply as "HTML5" or as "the HTML5 specification"

[02:17:01.0000] <MikeSmith>
right?

[02:17:02.0000] <MikeSmith>
I can ask for a change to that as well

[02:17:03.0000] <MikeSmith>
they are in LC with the spec now

[02:18:00.0000] <othermaciej>
I think both of those would be accurate ways to refer to HTML5, and less likely to be invalidated over time

[02:18:01.0000] <MikeSmith>
ok

[02:18:02.0000] <Hixie>
yeah

[03:07:00.0000] <zcorpan_>
http://www.w3.org/mid/6797BD1D-BA7A-440D-B6D6-4D0162E66883⊙gc here, roy is making a whole lot of sense

[03:12:00.0000] <zcorpan_>
"and that an <img> without a @role attribute is assumed to be the equivalent of <IMG @role="img"> (and would follow the rules in #1 above)" - http://www.w3.org/2009/06/Text-Alternatives-in-HTML5

[03:12:01.0000] <zcorpan_>
this doesn't make any sense given the definition of role="img" (which is closer to <figure>)

[03:13:00.0000] <Hixie>
holy cow, roy agreed with something html5 says?

[03:14:00.0000] <zcorpan_>
annevk: http://daringfireball.net/2007/12/fastcompany has just one sentence in the heading afaict

[03:14:01.0000] <annevk>
lol

[03:14:02.0000] <annevk>
fail

[03:15:00.0000] <annevk>
fortunately james example had plenty

[03:16:00.0000] <zcorpan_>
TR/ versions should have a floating message about the snapshotness, not just a paragraph in the SotD which engineers don't read

[03:17:00.0000] <annevk>
yeah, but that would need more than HTML WG buy in

[03:18:00.0000] <zcorpan_>
or the image in the margin should say "THIS DOCUMENT IS OBSOLETE DON'T READ IT"

[03:18:01.0000] <Hixie>
"Is there some kind of multiple-sentence-header abuse you are trying to combat?" lol

[03:19:00.0000] <Hixie>
zcorpan_: the paragraph right now isn't even in the SotD, it's in the History section

[03:19:01.0000] <zcorpan_>
Hixie: engineers don't read that either

[03:19:02.0000] <Hixie>
i guess we could make it more obvious

[03:20:00.0000] <zcorpan_>
Hixie: the way they end up in TR is they google on what they're looking for

[03:20:01.0000] <Hixie>
true

[03:20:02.0000] <zcorpan_>
maybe the solution is to change google's algorithms

[03:21:00.0000] <zcorpan_>
or change the links in wikipedia from TR/ to dev

[03:27:00.0000] <hsivonen>
this bug is awesome: https://bugzilla.mozilla.org/show_bug.cgi?id=515171

[03:27:01.0000] <hsivonen>
see comment #6

[03:33:00.0000] <zcorpan_>
/me changes his language to 'DROP TABLE *;'

[03:33:01.0000] <Hixie>
zcorpan_: well i looked into what it would take to add a floating warning on TR/ drafts, and it's doable pretty easily, but if i do it just like that i'll have half the wg on my ass

[03:33:02.0000] <Hixie>
zcorpan_: so if you want it please try to shore up some support in the wg first

[03:38:00.0000] <annevk>
should I file a bug on adding a magic square example to the spec?

[03:38:01.0000] <annevk>
not sure whether it's worth it, but they're kind of cool

[03:38:02.0000] <Hixie>
sure

[03:39:00.0000] <Hixie>
worse that can happen is i reject it and someone raises an issue for it

[03:40:00.0000] <jgraham>
Or a sudoku puzzle

[03:40:01.0000] <jgraham>
I hate them but it would be kinda cool to have a puzzle in the spec :)

[03:41:00.0000] <Hixie>
better idea than a magic square, file a bug :-)

[03:41:01.0000] <annevk>
nice one

[03:41:02.0000] <annevk>
/me files

[03:43:00.0000] <annevk>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=9943

[03:45:00.0000] <zcorpan_>
the puzzle should be interactive

[03:45:01.0000] <zcorpan_>
let's also add pacman to the canvas section

[03:46:00.0000] <Hixie>
so long as you get permission from the pacman rights owners...

[03:46:01.0000] <annevk>
and was Google pacman even done using <canvas>?

[03:46:02.0000] <annevk>
it worked in IE, no?

[03:47:00.0000] <Hixie>
that was a <div> disaster

[03:47:01.0000] <Hixie>
:-)

[03:47:02.0000] <Hixie>
http://www.google.com/pacman/ btw

[03:48:00.0000] <annevk>
yeah I know, it's fun

[03:53:00.0000] <Hixie>
so did anyone ever get around to giving some URLs of example pages that are actually helped by this role=button stuff?

[03:55:00.0000] <MikeSmith>
hsivonen: "For modsecurity 2.1.7 with core rules ver 1.5.1, workaround is to remove string “|id” (yes, remove that pipe character also) on the regular expression of rule 950006 in file modsecurity_crs_40_generic_attacks.conf"

[03:55:01.0000] <MikeSmith>
I guess they got a little carried away there

[03:56:00.0000] <zcorpan_>
annevk: a script isn't run if it's inserted but doesn't have any childnodes and doesn't have src set, iirc

[03:57:00.0000] <Dashiva>
http://www.cssquirrel.com/2010/06/15/693comic-update-html5-unicorn-heuristics/#comment-31995

[03:58:00.0000] <zcorpan_>
annevk: step 5

[03:58:01.0000] <annevk>
ah

[03:59:00.0000] <annevk>
resolved

[04:02:00.0000] <hsivonen>
MikeSmith: at leaast it's not a default feature of apache

[04:03:00.0000] <zcorpan_>
http://twitter.com/opera/statuses/16372312957 lol

[04:06:00.0000] <annevk>
haha

[04:06:01.0000] <Lachy>
haha, isn't it great when April Fools jokes just keep fooling people long after April Fools? :-)

[04:07:00.0000] <Dashiva>
Unless they know something you don't!

[04:08:00.0000] <Dashiva>
I've been wondering why the latest Opera builds seem to activate my webcam at random times...

[04:09:00.0000] <annevk>
that's part of our world dom^H^H^H^H...

[04:31:00.0000] <hsivonen>
hmm. Opera still sends the U; token in the UA string. I wonder if that's really needed for Web compat

[04:31:01.0000] <hsivonen>
IE doesn't send it

[04:38:00.0000] <karlcow>
opera doesn't support face gestures… no. shame on you ;)

[04:39:00.0000] <annevk>
jaja

[04:46:00.0000] <Lachy>
what was the U; token for originally?

[05:01:00.0000] <hsivonen>
Lachy: distinguishing between U.S., French and other versions of Netscape back when the U.S. and France had worse crypto regulations than they have today

[05:02:00.0000] <hsivonen>
U is U.S. ("strong" crypto), N is France (no crypto) and I is international (weak crypto)

[05:08:00.0000] <hsivonen>
Would anyone care to load http://www.delorie.com:81/some/url.txt in a Japanese version of IE8 and pastebin the result?

[05:09:00.0000] <hsivonen>
any non-English version of IE would be interesting, too, but not quite as interesting as Japanese

[05:09:01.0000] <hsivonen>
/me only has English IEs around

[05:09:02.0000] <hsivonen>
and non-Ultimate Windows, so I can't switch languages

[05:10:00.0000] <daedb>
hsivonen: http://pastebin.com/YUtFfz2r <-- swedish IE8

[05:11:00.0000] <hsivonen>
daedb: thanks!

[05:12:00.0000] <hsivonen>
confirms my hypothesis that IE doesn't put the UI language in the UA string

[05:27:00.0000] <Lachy>
why doesn't windows let you switch languages if you don't have Windows Ultimate?!

[05:28:00.0000] <Lachy>
that seems like such a basic feature, and so unnecessary to force people to purchase the Ultimate package for.

[05:30:00.0000] <gsnedders>
It's of use to a small number of people, and you get more profit if you force people to do so

[05:36:00.0000] <Lachy>
But it would be cheaper for them if they only had to produce 1 master copy of the discs to distribute globally, instead of having to then produce individual masters for each language in which Windows is distributed.

[05:37:00.0000] <Lachy>
and it would make it easier for foreigners to buy a copy that supports their own native language while in a foreign country

[05:37:01.0000] <hsivonen>
/me expects MS is competent at calculating the cost of tiered versions against the profit extracted by exploiting different points on the demand curve

[05:38:00.0000] <gsnedders>
Lachy: AFAIK they produce a single master disk, it's just it's locked by license key at install time

[05:39:00.0000] <Lachy>
so then each disc actually has all of the langauges on there anyway, and it's just their greed that prevents people from taking advantage?

[05:39:01.0000] <hsivonen>
I specifically ordered my Windows 7 Professional from Amazon.co.uk to get more googlable error messages

[05:39:02.0000] <hsivonen>
for some reason, the English version was priced unusually high compared to Finnish and Swedish versions in the local stores I usually buy from

[05:40:00.0000] <Hixie>
i downloaded ubuntu to get fewer error messages overall

[05:40:01.0000] <hsivonen>
the UK package allowed me to install and activate en-US, though

[05:40:02.0000] <Lachy>
fair enough. I might have to do the same, since I definitely don't want norwegian if and when I get a copy. I had previously assumed I would just be able to select English as install time, like I can with OSX

[05:40:03.0000] <hsivonen>
Hixie: I run it on top of Ubuntu

[05:40:04.0000] <Hixie>
:-P

[05:42:00.0000] <hsivonen>
however, the only times Windows 7 has done an ungraceful shutdown so far have been when X on Ubuntu has died due to USB devices getting plugged in...

[05:53:00.0000] <Lachy>
hsivonen, other than the language issue, does windows 7 professional generally have all the necessary features? Is there anything in Ultimate you wish you had?

[05:56:00.0000] <Lachy>
the compaison only shows bitlocker as the other difference, and I probably don't need that.

[06:02:00.0000] <hsivonen>
Lachy: The only other difference I am aware of is Bitlocker, which I don't need, since I only use Windows 7 for browser testing

[06:02:01.0000] <hsivonen>
and I assume Windows is untrusted, so I don't put my site login credentials there

[06:10:00.0000] <Lachy>
ok, in that case, I might got for Professional instead of Ultimate too, when I decide to get a copy soon.

[06:33:00.0000] <hsivonen>
Lachy: note: I've had Windows 7 for a day now

[06:43:00.0000] <Lachy>
ah, ok.

[07:16:00.0000] <kamathln>
Is there anything in HTML5 that gives AJAX like behaviour without script

[07:16:01.0000] <kamathln>
scripting* 

[07:17:00.0000] <kamathln>
Something that allows the server side to tell the otherside "Replace  only X part of the page, with Y item"

[07:19:00.0000] <zcorpan_>
kamathln: i guess that would be <iframe> and <a target>

[07:19:01.0000] <kamathln>
:D

[07:20:00.0000] <kamathln>
I wanted the server side to decide on the target

[07:20:01.0000] <kamathln>
I think thats not possible with iframe and target .

[07:21:00.0000] <zcorpan_>
it can if you're ok with it deciding up front when the first page loads :)

[07:22:00.0000] <kamathln>
zcorpan_: didnt catch you

[07:22:01.0000] <annevk>
there was a feature like that but it was dropped

[07:22:02.0000] <annevk>
though it did not quite do what you want

[07:23:00.0000] <kamathln>
annevk: Oh! any hint where I can look at the discussion, and any particular term they used that I can search for?

[07:23:01.0000] <kamathln>
Ok

[07:23:02.0000] <zcorpan_>
why do you want to do it without js?

[07:24:00.0000] <kamathln>
zcorpan_: I want it to work on devices/browsers that dont depend on js 

[07:25:00.0000] <zcorpan_>
what is it you're trying to do?

[07:25:01.0000] <kamathln>
Some people also have Javascript disabled 

[07:26:00.0000] <kamathln>
zcorpan_: Nothing at the moment .. But I have always found this as a major hinderence. That just disabling scriipting Simply makes the whole Server side programs a lot heavier

[07:29:00.0000] <kamathln>
Disabling javascript makes browsing a tad bit secure. 

[07:29:01.0000] <kamathln>
Well, I think I will just keep quiet for a while till I gather the right words

[07:30:00.0000] <annevk>
there has been discussion on a feature that's closer to what you want on the WHATWG list

[07:31:00.0000] <kamathln>
again, any term I can search for 

[07:31:01.0000] <kamathln>
or a date range

[07:31:02.0000] <kamathln>
?

[07:31:03.0000] <annevk>
I forgot the terminology :/

[07:31:04.0000] <zcorpan_>
/me also wonders what annevk is talking about

[07:31:05.0000] <daedb>
onlyreplace?

[07:31:06.0000] <zcorpan_>
oh

[07:32:00.0000] <zcorpan_>
/me didn't read that thread so no wonder he doesn't recall it

[07:32:01.0000] <kamathln>
http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-October/023661.html

[07:33:00.0000] <annevk>
http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-April/010801.html

[07:33:01.0000] <kamathln>
whoa! back in 2007!

[07:34:00.0000] <annevk>
nothing is novel

[07:36:00.0000] <annevk>
the proposal that was in a spec at some point was form pre-filling

[07:36:01.0000] <annevk>
material from 2004 or so

[07:36:02.0000] <kamathln>
All you guys were right. This comes close enough. But my point makes one but big change .. In my idea target(s) are to be decided by the server 

[07:36:03.0000] <annevk>
which in turn is based on some XForms stuff

[07:38:00.0000] <hsivonen>
whoa. scary. I got an automated prompt to install Flash when going to msn.com in IE8.

[07:39:00.0000] <hsivonen>
I wonder how many people click Yes to random prompts like that.

[07:39:01.0000] <zcorpan_>
hsivonen: 50%

[07:40:00.0000] <hsivonen>
seriously, it would be better for the security behavior of users to bundle Flash than to make the default home page autoprompt the install

[07:42:00.0000] <hsivonen>
wow. IE is hard to use.

[07:45:00.0000] <hsivonen>
oh great. after a day of using Windows, Minefield already has NPAPI plug-ins installed into it by 3 other browser vendors without explicitly telling me

[07:46:00.0000] <hsivonen>
kudos to Opera for not injecting native code into my Gecko-based browser

[07:49:00.0000] <Lachy>
oh, I hate when other programs try to install firefox extensions without asking.

[07:50:00.0000] <Lachy>
I uninstall them immediately. May require registry hacking.

[07:50:01.0000] <Lachy>
but what extensions did Chrome and Safari install?

[07:51:00.0000] <hsivonen>
(for those wondering what I'm talking about: Installing Chrome also installs a Google Update plug-in for Firefox/Minefield, installing a fully functional version of Safari also installs QuickTime Plug-in for Firefox/Minefield and to install legacy IE as part of XP mode, you have to go through Genuine Validation which apparently sticks a Windows Activation Technologies plug-in into Firefox/Minefield)

[07:51:01.0000] <Lachy>
is that true if you get the standalone Safari installer, instead of the Safari+QuickTime installer?

[07:52:00.0000] <hsivonen>
Lachy: I assume not, but I said "fully functional"

[07:52:01.0000] <hsivonen>
Lachy: I gather Safari 4 didn't support <video> without QuickTime

[07:53:00.0000] <Lachy>
if I install quicktime, then I expect it to install the NPAPI compatible plugin in the appropriate location for use by other browsers

[07:53:01.0000] <hsivonen>
I didn't test if Safari 5 can fall back to MS-provided decoders

[07:54:00.0000] <hsivonen>
Lachy: well, the other plug-ins can sorta be explained to be legitimate, but what is Google Update doing in Firefox after I install Chrome?

[07:54:01.0000] <Lachy>
no idea.

[07:55:00.0000] <daedb>
Actually, <video> does work without installing quicktime

[07:55:01.0000] <Dashiva>
Oh look, another YSOD, good thing XHTML is so easy

[07:55:02.0000] <hsivonen>
daedb: ooh. interesting!

[07:56:00.0000] <daedb>
I think, anyway

[07:56:01.0000] <daedb>
wait, no, I was wrong... quicktime is required :-/

[07:57:00.0000] <daedb>
/me should remember to actually test things before speaking...

[08:02:00.0000] <zcorpan_>
hsivonen: it's there so google can investigate why people switch back to firefox

[08:04:00.0000] <hsivonen>
awesome. now Windows Activation Technologies crashed

[08:04:01.0000] <hsivonen>
thanks to out-of-process plug-ins, Minefield didn't crash

[08:04:02.0000] <fagan>
hsivonen: trying to use IE?

[08:05:00.0000] <hsivonen>
fagan: trying to install XP Mode via Minefield after failing to install it via IE8...

[08:05:01.0000] <fagan>
sounds hard 

[08:05:02.0000] <fagan>
/me is a happy linux user

[08:06:00.0000] <hsivonen>
I'm a Linux user, too, though an unhappy one

[08:06:01.0000] <fagan>
hsivonen: why unhappy? 

[08:06:02.0000] <fagan>
its getting better year on year 

[08:06:03.0000] <hsivonen>
/me filed two customer dissatisfaction surveys on Canonical support case handling today

[08:07:00.0000] <hsivonen>
fagan: 1) I have situations that need support. 2) Support hasn't fully solved a single problem yet.

[08:07:01.0000] <fagan>
hsivonen: wow paid for support ill never do that :)

[08:07:02.0000] <fagan>
hsivonen: the best way I found is to ask on the forums sometimes its better than canonical support 

[08:08:00.0000] <hsivonen>
support has partially solved one out of six so far.

[08:08:01.0000] <fagan>
im an Ubuntu community developer 

[08:08:02.0000] <hsivonen>
fagan: before I purchased support, I tried Launchpad Questions

[08:08:03.0000] <fagan>
hsivonen: the forums are the place or IRC 

[08:09:00.0000] <hsivonen>
fagan: I also tried the Finnish Ubuntu IRC channel first

[08:09:01.0000] <fagan>
launchpad answers is kinda dead 

[08:09:02.0000] <hsivonen>
I guess I need to try the forums, then

[08:09:03.0000] <fagan>
hsivonen: must be a complicated problem

[08:10:00.0000] <fagan>
I wonder why the whatwg site is so basic 

[08:11:00.0000] <fagan>
shouldnt you guys have an insane video/audio experience to show off what your doing? :D

[08:13:00.0000] <hsivonen>
I want to do radical stuff with Ubuntu, such as configure a button on my Wacom stylus (has GUI on Windows and Mac), make a bootable clone of the system (productized for Mac) or attach USB devices while the system is running

[08:13:01.0000] <hsivonen>
and distribution upgrade to completion

[08:14:00.0000] <Lachy>
fagan, because the WHATWG is more focussed on developing the spec, than on creating and publishing fancy demos. There are plenty of other sites doing that already

[08:15:00.0000] <Dashiva>
It's definitely a task worth doing, though

[08:15:01.0000] <Lachy>
sure, if there are volunteers

[08:19:00.0000] <hsivonen>
oh great. Windows 7 crashes when trying to setup the XP Mode

[08:19:01.0000] <hsivonen>
bah

[08:21:00.0000] <Lachy>
that's not a bug, that means XP mode is functioning normally ;-)

[08:21:01.0000] <hsivonen>
I wonder if VirtualBox kills the guest if the guest tries to run an VT-x instruction

[08:22:00.0000] <hsivonen>
so much broken stuff

[08:22:01.0000] <hsivonen>
/me heads to the gym instead of fighting with operating systems

[08:23:00.0000] <fagan>
Lachy: very true I suppose it takes some effort to do good tutorials and thats what w3schools is for 

[08:24:00.0000] <Lachy>
fagan, try http://html5doctor.com/

[08:25:00.0000] <Lachy>
I wouldn't trust W3 schools. They're tutorials are notoriously bad quality

[08:25:01.0000] <fagan>
Lachy: I actually have been reading the spec so I dont really need to look at tutorials :D

[08:25:02.0000] <fagan>
but I only subbed the mailing list today 

[08:26:00.0000] <fagan>
I just popped in and had a look at it as it was being built to see what was going on 

[08:26:01.0000] <zcorpan_>
/me wonders if there's a way to make the material on w3schools be good

[08:27:00.0000] <fagan>
zcorpan_: well I found the best way to develop websites was to just look at other sites you like and try to make it look the way you want it 

[08:28:00.0000] <fagan>
then use tidy to fix the mistakes 

[08:29:00.0000] <fagan>
so for the most recent site I made I copied the look of apple.com ish and then changed the color scheme and images..etc so then it looked good

[08:29:01.0000] <fagan>
if w3schools broke down a few good websites to show people how to do it like a pro it would be a lot more helpful I think 

[09:00:00.0000] <zcorpan_>
clearly shelley doesn't know safari 5 supports <aside>

[09:12:00.0000] <karlushi>
zcorpan_, since march 2010? https://bugs.webkit.org/show_bug.cgi?id=32943

[09:12:01.0000] <karlushi>
http://en.wikipedia.org/wiki/Comparison_of_layout_engines_%28HTML5%29#Elements

[09:14:00.0000] <fagan>
karlushi: thats out of date I think for opera's engine 

[09:14:01.0000] <fagan>
that has most of them done in the most recent version 

[09:14:02.0000] <fagan>
or it should from what I remember 

[09:14:03.0000] <gsnedders>
No, it doesn't

[09:15:00.0000] <gsnedders>
I think the only thing wrong in that section is canvas should be 2.4

[09:16:00.0000] <karlushi>
hmmm

[09:16:01.0000] <karlushi>
a common format for publishing release notes would be cool. That would help to dynamically create implementation status report across browsers engines

[09:20:00.0000] <karlushi>
that would require a feature id (maybe the uri of the feature), version number of the engine, status: [bug|partial|ok|no], comment (optional), uri of bug report (optional)

[09:20:01.0000] <karlushi>
then people could publish their own report if they disagree with the report published by the company.

[09:22:00.0000] <Rik`>
would be cool to have release notes for Safari :(

[09:23:00.0000] <gsnedders>
When is something buggy and when is something not?

[09:25:00.0000] <karlushi>
gsnedders, to be defined :) maybe when there is an open bug about this feature. Though people would open open bug report just to change the status :p

[09:25:01.0000] <karlushi>
looking at the changelog of Opera 10.6 beta http://www.opera.com/docs/changelogs/mac/1060b1/

[09:26:00.0000] <karlushi>
the level of granularity is a big issue too

[09:27:00.0000] <karlushi>
I see for example in opera 10.6 beta 1

[09:27:01.0000] <karlushi>
"# Removing support for div nowrap"

[09:28:00.0000] <karlushi>
so something which was implemented in the past is not anymore. But in an implementation report it doesn't say, if because it was too buggy, or if the feature was dropped from the technology. 

[10:35:00.0000] <jgraham>
I wonder what "officially objecting" is

[10:36:00.0000] <zcorpan_>
i wonder if daniel objects to all non-machine-checkable requirements

[10:36:01.0000] <jgraham>
Logically he has to

[10:37:00.0000] <jgraham>
Given his second sentence

[10:38:00.0000] <jgraham>
In other news, I loath the python unittest module

[10:38:01.0000] <jgraham>
*loathe

[10:42:00.0000] <hober>
zcorpan_: yeah, just asked him that on the list

[10:43:00.0000] <Dashiva>
Removing non-machine-checkable requirements, hwh

[12:28:00.0000] <gsnedders>
http://gsnedders.html5.org/outliner/process.py?url=http%3A%2F%2Fgsnedders.com

[12:28:01.0000] <gsnedders>
Looks like it's trying to use the wrong Python version's folder

[13:51:00.0000] <jgraham>
gsnedders: Seems to have charcter encoding issues

[13:52:00.0000] <gsnedders>
jgraham: I see

[14:36:00.0000] <jgraham>
Can a <select> token ever be processed as if In Body when the current node is a <colgroup> element?

[14:37:00.0000] <jgraham>
Or rather when in the In Column Group insertion mode

[14:41:00.0000] <zcorpan_>
jgraham: afaict, <select> in column group just implies </colgroup> (which switches to in table) and gets reprocessed

[14:41:01.0000] <zcorpan_>
jgraham: or ignored in the fragment case

[14:42:00.0000] <jgraham>
zcorpan_: That's what I think

[14:43:00.0000] <jgraham>
So the text about using In Select In Table if the insertion mode is In Column Group is redundant?

[14:44:00.0000] <zcorpan_>
which text?

[14:45:00.0000] <zcorpan_>
If the insertion mode is one of in table", "in caption", "in column group", "in table body", "in row", or "in cell", then switch the insertion mode to "in select in table". Otherwise, switch the insertion mode to "in select".

[14:46:00.0000] <zcorpan_>
yeah

[14:46:01.0000] <zcorpan_>
while you whine about that, there's a missing quote also :)

[14:52:00.0000] <jgraham>
Whined

[15:07:00.0000] <AryehGregor>
Now, how long until the YouTube HTML5 beta works on embedded videos?

[15:07:01.0000] <TabAtkins>
I'm counting the days. :\

[15:07:02.0000] <AryehGregor>
Does that imply you know how many days are left?

[15:07:03.0000] <TabAtkins>
I wish it did.

[15:08:00.0000] <AryehGregor>
They'd have to change the embed code they provide, of course.  Probably just make it an iframe or script tag.

[15:08:01.0000] <jgraham>
How high is your count?

[15:08:02.0000] <TabAtkins>
jgraham: Q.

[15:08:03.0000] <zcorpan_>
AryehGregor: would be cool if it could be made to work retroactively i.e. without changing the embed code

[15:09:00.0000] <Lachy>
there are some issues that I think need to be solved before before YouTube will start allowing 3rd party sites to embed HTML5 video

[15:09:01.0000] <AryehGregor>
The embed code is <object type="application/x-shockwave-flash">, so I don't see that working.

[15:09:02.0000] <AryehGregor>
Lachy, really?  Like what?

[15:09:03.0000] <Lachy>
in particular, full screen support, more widespread support in browsers

[15:09:04.0000] <Lachy>
for webm

[15:09:05.0000] <zcorpan_>
AryehGregor: hmm, yeah

[15:09:06.0000] <Lachy>
in released products, not just betas

[15:10:00.0000] <AryehGregor>
I'm saying they could support the beta for embeds.

[15:10:01.0000] <AryehGregor>
I.e., use HTML5 if you're logged into YouTube and have opted in.

[15:10:02.0000] <AryehGregor>
Not for general deployment, of course, except as a fallback if Flash isn't installed.

[15:13:00.0000] <Lachy>
they could make it easier to deploy HTML5 later, if they changed the embed code from <embed> to a <script> that then dynamically generates the markup.

[15:14:00.0000] <Lachy>
that way, even if such a script simply embedded the current flash player, they can update the script later to generate code for HTML5 when that's supported

[15:14:01.0000] <othermaciej>
or they could use an iframe embed

[15:14:02.0000] <othermaciej>
that would let them conditionally serve flash or html

[15:14:03.0000] <Lachy>
yeah, that too.

[15:15:00.0000] <othermaciej>
without having to convince sites to let them inject a <script?

[15:15:01.0000] <othermaciej>
er, <script>

[15:15:02.0000] <othermaciej>
we at Apple have been recommending <iframe> embedding to video content providers who ask us

[15:20:00.0000] <AryehGregor>
Why would sites be upset by <script> but not mind <object>?

[15:20:01.0000] <TabAtkins>
Because people are dumb?

[15:20:02.0000] <AryehGregor>
<iframe> is certainly more secure than either, if you don't trust Google's JS.

[15:20:03.0000] <othermaciej>
I don't know of sites would mind <script>

[15:20:04.0000] <othermaciej>
apparently the way Web video embedding works now is...

[15:20:05.0000] <AryehGregor>
They use <script> for ads and analytics.

[15:21:00.0000] <othermaciej>
many of the popular providers have their player SWFs on a whitelist for embedding into posts / comments / user generated content / etc

[15:21:01.0000] <othermaciej>
like, popular blogging platforms, things like facebook, etc

[15:21:02.0000] <othermaciej>
if they come up with a new mechanism, it will take some time for them to get those whitelists updated

[15:22:00.0000] <othermaciej>
of course, an iframe is safer than <object> and maybe does not even need whitelisting but that might be a hard sell

[15:23:00.0000] <Lachy>
it would be nice if they also provided the <video> markup directly for authors to copy and paste, so that sites can apply their own custom controls for videos.

[15:24:00.0000] <AryehGregor>
I doubt they'll want to do that.  It gives them much less control.

[15:24:01.0000] <jgraham>
So what am I missing:

[15:24:02.0000] <jgraham>
<!doctype html><body><rp><rp>

[15:24:03.0000] <jgraham>
er

[15:24:04.0000] <AryehGregor>
Actually, maybe part of the reason that they haven't done this is that the video URLs are obfuscated and they don't want to publicize the URL generation algorithm.

[15:24:05.0000] <jgraham>
<!doctype html><body><rp></rp>

[15:24:06.0000] <TabAtkins>
jgraham: <ruby>?

[15:24:07.0000] <jgraham>
stack of open elements is

[15:24:08.0000] <jgraham>
[html, body, rp]

[15:24:09.0000] <AryehGregor>
They probably don't want people hotlinking their videos without their scripts and so on being there as well.

[15:25:00.0000] <jgraham>
</rp> has no special processing

[15:25:01.0000] <AryehGregor>
Technically it's always been possible, but I guess <video> makes it a lot easier.

[15:25:02.0000] <jgraham>
So we do any other end tag

[15:25:03.0000] <jgraham>
The first step is to generate implied end tags, which removes the rp element from the stack of open elements

[15:26:00.0000] <jgraham>
this is bad

[15:26:01.0000] <jgraham>
Should it generate implied end tags excluding rp?

[15:26:02.0000] <jgraham>
(or whatever the current token name is)

[15:28:00.0000] <Lachy>
in theory, the optimal markup to offer to authors would be <video style="binding:url(http://example.com/player.xml)"><source ...>...</video>, where player.xml is an XBL template that dynamically creates the custom UI.

[15:28:01.0000] <Lachy>
that would also allow 3rd parties to create and distribute video player UIs that sites can easily apply to their own videos

[15:30:00.0000] <zcorpan_>
jgraham: isn't <!doctype html><body><option></option> the same issue?

[15:32:00.0000] <jgraham>
zcorpan_: Yeah, but <rp> is shorter to type

[15:34:00.0000] <zcorpan_>
jgraham: yeah. it should say "except for elements with the same tag name as the token." in any other end tag

[15:34:01.0000] <jgraham>
In step 2.1?

[15:35:00.0000] <zcorpan_>
yes

[15:35:01.0000] <jgraham>
Good, because that's the bug I field

[15:35:02.0000] <jgraham>
*filed

[15:36:00.0000] <zcorpan_>
i thought bugs like these had been ironed out by now :)

[15:40:00.0000] <zcorpan_>
hmm, why does <rt> outside <ruby> not generate implied end tags?

[15:41:00.0000] <jgraham>
zcorpan_: Me too :)

[15:41:01.0000] <jgraham>
/me just managed to crash emacs

[15:44:00.0000] <zcorpan_>
wonder how ie handles rt and rp outside ruby

[15:47:00.0000] <zcorpan_>
seems ie implicitly closes rt and rp for <rt> or <rp> even outside ruby

[15:47:01.0000] <zcorpan_>
and doesn't implicitly close <p> even inside ruby

[15:48:00.0000] <zcorpan_>
while html5 treats rt and rp as unknown elements outside ruby, and forcefully pops anything inside ruby

[15:48:01.0000] <Hixie>
IE's behaviour is nutto

[15:48:02.0000] <Hixie>
i based the parser on actual pages

[15:49:00.0000] <Hixie>
i literally scanned every page in the google index that used <ruby> and then tried to come up with something that worked with all of them

[15:49:01.0000] <zcorpan_>
Hixie: did you look at pages that use <rt> or <rp> outside <ruby>?

[15:51:00.0000] <Hixie>
i don't recall there being a significant number

[15:58:00.0000] <jgraham>
<b><button>foo</b> should reopen the <b> inside the button these days, right?

[16:00:00.0000] <fagan>
jgraham: well the button itself isnt going to be bold the text is so it should be inside the button tag

[16:00:01.0000] <jgraham>
fagan: Indeed, but not quite the point

[16:00:02.0000] <jgraham>
The point is how the error recovery is supposed to work

[16:01:00.0000] <jgraham>
button used to insert a marker on the list of active formatting elements so that the <b> would never apply inside the <button>. That seems to have changed

[16:01:01.0000] <fagan>
jgraham: just tested it in chrome and <b> outside the button tag doesnt work 

[16:02:00.0000] <jgraham>
fagan: That could be why it changed :)

[16:02:01.0000] <fagan>
true 

[16:02:02.0000] <jgraham>
Although I guess it is more likely to be IE compat

[16:02:03.0000] <jgraham>
Er, wait

[16:03:00.0000] <jgraham>
The chrome behaviour is the old behaviour, obviously

[16:03:01.0000] <fagan>
lol

[16:06:00.0000] <jgraham>
Hmm, I am clearly too tired for this

[16:06:01.0000] <jgraham>
Hixie: If you fancy looking at parser bugs relatively soon it would be appreciated

[16:08:00.0000] <fagan>
jgraham: which browser ?

[16:09:00.0000] <zcorpan_>
fagan: jgraham is tearing apart the html5 parsing algorithm

[16:09:01.0000] <fagan>
ah ok 

[16:10:00.0000] <fagan>
/me knows the engine side of things but not the algorithm stuff

[16:13:00.0000] <fagan>
I wonder how up to date the html checker is on the w3 site for html5 

[16:14:00.0000] <fagan>
Im doing a few websites in it at the moment and they are passing fine but id hate for them to not be valid later on 

[16:14:01.0000] <fagan>
no matter anyway 

[16:14:02.0000] <Hixie>
use validator.nu, that's probably the most up to date

[16:15:00.0000] <Hixie>
validators will catch more and more errors as they get better though

[16:15:01.0000] <Hixie>
so it's normal to see new errors be caught over time :-)

[16:15:02.0000] <zcorpan_>
not to mention that html5 can still change out from under you

[16:16:00.0000] <fagan>
isnt it last call for changes now?

[16:16:01.0000] <zcorpan_>
sure

[16:16:02.0000] <Hixie>
last call doesn't mean no changes :-)

[16:16:03.0000] <zcorpan_>
it can still change

[16:16:04.0000] <Hixie>
the spec will continue changing its whole life, as people find new problems

[16:16:05.0000] <Hixie>
when it stops changing is when it's dead

[16:17:00.0000] <fagan>
true but still I would have thought that the biggest changes have already happened 

[16:18:00.0000] <Hixie>
the biggest ones probably have, yeah

[16:20:00.0000] <fagan>
I just love the video and audio if nothing else was done but that I would have been very very very happy 

[16:20:01.0000] <fagan>
but thats not the only thing I know 

[16:33:00.0000] <AryehGregor>
Where's the big combined Web Apps 1.0 spec?

[16:33:01.0000] <AryehGregor>
/me is trying to determine how the size of Hixie-edited specs compares to the total output of the CSSWG across its entire existence

[16:34:00.0000] <AryehGregor>
Oh, drat, not all the W3C specs are single-page.

[16:34:01.0000] <zcorpan_>
http://www.whatwg.org/specs/web-apps/current-work/complete.html

[16:34:02.0000] <AryehGregor>
Thanks.

[16:35:00.0000] <AryehGregor>
I didn't see a link anywhere on the website.

[16:35:01.0000] <Hixie>
it's there, but i try not to scare people off too much

[16:35:02.0000] <fagan>
I would be nice if it was split up a little

[16:35:03.0000] <Hixie>
fagan: http://www.whatwg.org/specs/web-apps/current-work/complete/

[16:35:04.0000] <AryehGregor>
Web Apps 1.0: ~5.3 MB

[16:35:05.0000] <AryehGregor>
All of the stuff I could easily get from the CSS WG without looking at every spec: <3 MB

[16:36:00.0000] <Hixie>
AryehGregor: you really want to compare the source documents, the post-processor adds a ton of cross-reference stuff

[16:36:01.0000] <AryehGregor>
Are the source documents for all the CSS specs readily available?

[16:36:02.0000] <Hixie>
i use way more hyperlinks in my specs than most people

[16:36:03.0000] <Hixie>
AryehGregor: no idea

[16:36:04.0000] <fagan>
Hixie: that is a huge page :)

[16:36:05.0000] <AryehGregor>
The CSS specs have a lot of repeated boilerplate, though, from the headers on every one.

[16:37:00.0000] <AryehGregor>
Source is almost as large.

[16:37:01.0000] <AryehGregor>
>4 MB.

[16:37:02.0000] <Hixie>
fagan: complete/ is a split up version of complete.html

[16:42:00.0000] <fagan>
Hixie: but still its about a 2 day read start to finish

[16:42:01.0000] <fagan>
Oh and I love the joke in the how to read this document part :D

[16:42:02.0000] <Hixie>
splitting it up doesn't make it smaller, no :-)

[16:42:03.0000] <Hixie>
just makes it have multiple pages

[16:43:00.0000] <fagan>
true but most people would be interested in specific parts

[16:43:01.0000] <fagan>
but I suppose they could just look at the index and jump around 

[16:44:00.0000] <fagan>
/me understands

[16:46:00.0000] <zcorpan_>
fagan: what joke?

[16:46:01.0000] <fagan>
zcorpan_: read it the entire way through, read it again, read it backwards then jump around to the bits you are interested in 

[16:47:00.0000] <zcorpan_>
fagan: i think that's not a joke

[16:47:01.0000] <fagan>
has to be joking

[16:47:02.0000] <Hixie>
it's one of those "it's funny because it's true" things

[16:47:03.0000] <fagan>
reading stuff backwards makes my head hurt 

[16:48:00.0000] <fagan>
words stop making sense 

[16:48:01.0000] <fagan>
:)

[16:48:02.0000] <annevk>
it just means you have to read it a lot

[16:48:03.0000] <annevk>
which not many people do, so there may be some issue

[16:49:00.0000] <fagan>
annevk: I just like to keep up with it :)

[16:50:00.0000] <annevk>
for that all you need is http://html5.org/tools/web-apps-tracker

[16:50:01.0000] <annevk>
:)

[16:50:02.0000] <zcorpan_>
in my experience just reading a spec doesn't give you a useful understanding of it, no matter how many times you read it

[16:50:03.0000] <fagan>
annevk: true but I hate that page 

[16:51:00.0000] <zcorpan_>
you have to do something actively with it to properly understand it, like implement it or write a testsuite for it

[16:51:01.0000] <fagan>
dont ask me why 

[16:51:02.0000] <zcorpan_>
and you're better off focusing on one part at a time instead of reading cover to cover

[16:52:00.0000] <fagan>
zcorpan_: I focus on the bits im interested in and the bits I would use most 

[16:52:01.0000] <annevk>
zcorpan_, I can answer quite a lot of questions within seconds/minutes by having read most bits of the specification a few times

[16:52:02.0000] <fagan>
(and just skip the rest)

[16:52:03.0000] <fagan>
I read all the common stuff too though 

[16:52:04.0000] <zcorpan_>
annevk: sure, but you can't find fundamental spec bugs by just reading it, or can you?

[16:52:05.0000] <zcorpan_>
i can't

[16:54:00.0000] <zcorpan_>
bedtime, nn

[16:54:01.0000] <annevk>
oh yeah, finding the real big issues requires implementation experience


2010-06-18
[17:47:00.0000] <AryehGregor>
It's amazing how much smoother to use Gmail is on my tiny Nexus One than on my desktop.

[17:47:01.0000] <AryehGregor>
I wonder: is this some deficiency in the web platform, or just Gmail not caching things locally aggressively enough?

[17:48:00.0000] <AryehGregor>
/me bets the latter, wonders why

[17:50:00.0000] <AryehGregor>
Local Storage size on disk for mail.google.com: 3 KB for http, 3 KB for https.  :/

[18:32:00.0000] <karlcow>
hmm the discussions about editors drafts, working drafts, etc. is argueing if it's bad or not

[18:32:01.0000] <karlcow>
INSTEAD of trying to list the issues/needs and how we solve them.

[18:34:00.0000] <Dashiva>
Issue: working drafts are always out of date

[18:34:01.0000] <karlcow>
Dashiva: I would add a "for [here an audience]"

[18:34:02.0000] <karlcow>
for example in your sentence.

[18:35:00.0000] <AryehGregor>
Why would anyone want to use an out-of-date draft?

[18:35:01.0000] <AryehGregor>
For any purpose?

[18:35:02.0000] <AryehGregor>
It's obsolete.

[18:35:03.0000] <karlcow>
Issue: last drafts of a specification is useful for some implementers.

[18:35:04.0000] <Hixie>
the only audience for which out of date drafts is useful is spec archeologists, like when henri and i try to work out why something was done

[18:35:05.0000] <Hixie>
but with HTML5, the whatwg svn repo is far more useful for that now

[18:36:00.0000] <karlcow>
How do we solve it? What is needed in the document, etc. Editor drafts could be retitled Developers Edition

[18:36:01.0000] <Dashiva>
Resolution: Stop issuing working drafts

[18:36:02.0000] <Dashiva>
Editor's draft covers all the use cases

[18:37:00.0000] <karlcow>
AryehGregor: People who need stability: Book writers, documentation writers, etc. A lot easier to update your work big by big step than because on a moving target with micro-edits

[18:37:01.0000] <Hixie>
Dashiva: that's what i did at the whatwg when i realised this problem existed

[18:37:02.0000] <Dashiva>
Should we really be encouraging people to write outdated documentation?

[18:37:03.0000] <AryehGregor>
karlcow, why would you want to write a book based on a snapshot that's months out of date?  If you really want that, you can pick a specific revision from SVN and only use that.

[18:38:00.0000] <AryehGregor>
At least that will be up-to-date as of the time you start writing.

[18:38:01.0000] <AryehGregor>
But I'll stick to stuff written by people who actually follow spec changes, for my part.

[18:38:02.0000] <karlcow>
AryehGregor: because not everyone is a geek aka pick up a version in SVN

[18:38:03.0000] <AryehGregor>
Then why the heck are they writing documentation for a spec?

[18:38:04.0000] <Dashiva>
People who write books and documentation should be competent enough to do so

[18:39:00.0000] <karlcow>
Dashiva: why do you hate freedom? ;)

[18:39:01.0000] <Hixie>
nobody's stopping someone from taking a snapshot of the PDF or HTML versions of the spec

[18:39:02.0000] <Dashiva>
I don't hate freedom, I hate freedumb

[18:39:03.0000] <Hixie>
doesn't mean we need to make explicit copies

[18:40:00.0000] <AryehGregor>
/me just wrote a 1377-word blog comment about HTML5

[18:40:01.0000] <karlcow>
It is usually wiser for the peace of each community to give tools that these communities need. (if they need it :) )

[18:40:02.0000] <othermaciej>
AryehGregor: 40 words too many!

[18:40:03.0000] <othermaciej>
which blog?

[18:40:04.0000] <AryehGregor>
othermaciej, <http://www.cssquirrel.com/2010/06/15/693comic-update-html5-unicorn-heuristics/comment-page-1/#comment-32000>, but it has to pass moderation because of the links I put in for references.

[18:40:05.0000] <Dashiva>
(People are free to be dumb, but we are not obligated to help them)

[18:41:00.0000] <othermaciej>
there's been some interesting discussion in those comments

[18:41:01.0000] <AryehGregor>
karlcow, so people who actually want a snapshot can just save the page to their hard disk.  What's the problem with that?

[18:41:02.0000] <AryehGregor>
WD would make sense if there were some meaningful procedure for approving it, so it were actually more likely to be correct than an ED, but that's not the case.

[18:41:03.0000] <AryehGregor>
othermaciej, I have hopes of converting Kyle Weems to the Light Side.

[18:42:00.0000] <karlcow>
AryehGregor: I have no issue with people saving a copy, they can do it already :)

[18:42:01.0000] <AryehGregor>
I think he's just misinformed.

[18:43:00.0000] <karlcow>
/me wishes that people could accept diversity in opinions ;)

[18:43:01.0000] <Dashiva>
Diversity in opinions is fine

[18:43:02.0000] <Dashiva>
Diversity in facts is not

[18:43:03.0000] <karlcow>
religious statement

[18:44:00.0000] <karlcow>
too far for me

[18:44:01.0000] <karlcow>
:)

[18:44:02.0000] <othermaciej>
AryehGregor: I think your messages in those comments come across as fairly rational and balanced

[18:44:03.0000] <Dashiva>
I should probably cite properly

[18:44:04.0000] <AryehGregor>
othermaciej, hope so!

[18:44:05.0000] <othermaciej>
AryehGregor: even though you are kind of questioning the quality of my service as chair :-p

[18:44:06.0000] <AryehGregor>
The last sentence of my new one is "If the W3C is unwilling to accept that, it will become a victim of natural selection, like it or not", but I think I supported it fairly well.

[18:45:00.0000] <Dashiva>
"You’re entitled to your own opinions. You’re not entitled to your own facts." - Daniel Patrick Moynihan

[18:45:01.0000] <AryehGregor>
othermaciej, don't worry, I'm only comparing you unfavorably to Hixie.  That's not an insult in absolute terms.

[18:46:00.0000] <AryehGregor>
Dashiva, can I be entitled to my own facts if I write a suitably advanced virtual reality program that implements them?

[18:48:00.0000] <karlcow>
AryehGregor: that is a lot more reasonable and basically what every society does.

[18:48:01.0000] <AryehGregor>
What, writing suitably advanced virtual reality programs?

[18:48:02.0000] <Dashiva>
No, you have to pass a Trial of Philosophy to justify why a virtual reality should be considered factual first

[18:49:00.0000] <Dashiva>
How come these discussions always start hours past midnight

[18:49:01.0000] <Dashiva>
Time to sleep

[18:49:02.0000] <karlcow>
hehe

[18:49:03.0000] <AryehGregor>
Because you live in the wrong time zone?

[18:49:04.0000] <karlcow>
Dashiva: because you are living in your own reality 

[18:50:00.0000] <Dashiva>
CET is the best time zone ever

[18:51:00.0000] <AryehGregor>
Dashiva, reality is indistinguishable from a sufficiently good illusion.  Ergo, there is no difference according to any empirical worldview, and empirical worldviews are the only ones that count because all others are stupid.  QED.

[18:51:01.0000] <Dashiva>
Then you have to prove that your virtual reality is sufficiently good

[18:51:02.0000] <Dashiva>
A sort of a turing test, if you will

[18:51:03.0000] <AryehGregor>
That was postulated.

[18:51:04.0000] <AryehGregor>
"suitably advanced"

[18:52:00.0000] <karlcow>
http://en.wikipedia.org/wiki/Allegory_of_the_Cave

[18:52:01.0000] <Dashiva>
Besides, people outside would be able to notice those on the inside, and thereby be able to reject their statements about the warped reality

[18:53:00.0000] <AryehGregor>
Dashiva, yes, but that's exactly what we were talking about: everyone gets their own version of the facts.

[18:53:01.0000] <Dashiva>
No, some people get a warped reality and other people get facts

[18:54:00.0000] <Dashiva>
Surely you do not consider the worldview of a lunatic to be factual?

[18:54:01.0000] <Dashiva>
"There will have to be a couple of implementations before CR of HTML5, anyway."

[18:55:00.0000] <Dashiva>
It's like nobody even cares what the process really says anymore

[18:55:01.0000] <AryehGregor>
Dashiva, but if everyone uses VR adjusted to taste, then everyone's reality will be equally accurate.

[18:56:00.0000] <Dashiva>
You postulate a world in which a) everyone is in a virtual reality and b) nobody realizes they are in a virtual reality

[18:56:01.0000] <AryehGregor>
Sure.  Maybe we're in one already.  Can't say.

[18:56:02.0000] <AryehGregor>
Just as good as actual reality.

[18:56:03.0000] <karlcow>
hehe

[18:56:04.0000] <Dashiva>
Not quite

[18:56:05.0000] <Dashiva>
Even if we cannot say for sure if our reality is indeed reality, we can say for sure that virtual realities inside our own reality are not reality

[18:57:00.0000] <karlcow>
Dashiva: and it is not an issue

[18:57:01.0000] <AryehGregor>
Also, a lunatic's worldview is indistinguishable from a factual worldview, from the lunatic's perspective.  Other people think the opposite, but in principle, that can be a perfectly symmetric disagreement, and we only win by majority vote.

[18:57:02.0000] <karlcow>
as long as you have developed a coherent system for you and maybe your social groups. 

[18:57:03.0000] <AryehGregor>
We don't know that virtual realities inside our own reality are not reality if we make ourselves all forget.

[18:58:00.0000] <karlcow>
Physics (science) is all about *creating* a reality.

[18:58:01.0000] <AryehGregor>
I'm speaking from the viewpoint of us after we all go into the VR and forget everything.

[18:58:02.0000] <Dashiva>
If everyone in reality forgets about reality in favor of a virtual reality, then the reality is no longer part of their conceptual universe

[18:58:03.0000] <AryehGregor>
In that situation, it would be fair to say that everyone gets to have their own version of reality.

[18:59:00.0000] <AryehGregor>
Right, exactly.

[18:59:01.0000] <Hixie>
as a physicist by training, i have to take objection ot the idea that physics _creates_ reality

[18:59:02.0000] <AryehGregor>
Me too, as a physics minor!

[18:59:03.0000] <AryehGregor>
Although it would be cool if it did.

[18:59:04.0000] <Hixie>
unless you have a very odd definition of "reality"

[18:59:05.0000] <AryehGregor>
I always play wizards in fantasy games.  :(

[18:59:06.0000] <Dashiva>
That's not very physicsy of you

[19:00:00.0000] <karlcow>
http://www.w3.org/2010/09/web-on-tv/

[19:00:01.0000] <AryehGregor>
No, it's perfectly physicsy.  I'm just utilizing the laws of physics that are present in a hypothetical alternate reality where they allow me to, say, cause you to spontaneously catch fire when I get annoyed at you.

[19:00:02.0000] <karlcow>
:D

[19:00:03.0000] <Dashiva>
Also known as wish fulfillment

[19:01:00.0000] <AryehGregor>
Exactly!

[19:01:01.0000] <karlcow>
Dashiva: and you are still not in bed.

[19:01:02.0000] <Dashiva>
I actually have both feet in bed

[19:01:03.0000] <karlcow>
TMI

[19:01:04.0000] <Dashiva>
The bed is my foot stool

[19:02:00.0000] <Dashiva>
The bookshelf is my monitor stand

[19:07:00.0000] <AryehGregor>
Awesome, this is how web PDF support should have always been done: http://blog.chromium.org/2010/06/bringing-improved-pdf-support-to-google.html

[19:10:00.0000] <karlcow>
hmmm, if I can still download the pdf locally or save the generated html on my computer I'm ok.

[19:10:01.0000] <karlcow>
Browser is not the best tool for me to read a pdf.

[19:10:02.0000] <karlcow>
I have seen this today too http://simonwillison.net/2010/Jun/17/pdfjs/

[19:11:00.0000] <AryehGregor>
I'd way prefer to get a good PDF reader implementation in a browser than to have to use external programs.

[19:11:01.0000] <karlcow>
AryehGregor: and me the opposite. every taste in nature ;)

[19:12:00.0000] <AryehGregor>
Anyway, things I was supposed to do tonight: file health insurance claim, buy several things, study some religious texts, pack for trip, go to bed early.  Things I actually did: chat on IRC.

[19:12:01.0000] <AryehGregor>
Sigh.

[19:12:02.0000] <AryehGregor>
Good night.

[19:12:03.0000] <roc>
roc

[19:13:00.0000] <AryehGregor>
o_O

[19:14:00.0000] <karlcow>
AryehGregor: nite nite

[19:16:00.0000] <roc>
oops

[19:22:00.0000] <karlcow>
http://lists.w3.org/Archives/Public/public-web-perf/ Web Browser Performance mailing List

[20:59:00.0000] <abarth>
Hixie: no complaints from day 1 of the HTML5 tokenizer in WebKit nightly

[21:00:00.0000] <abarth>
Hixie: that's pretty good :)

[21:13:00.0000] <roc>
karlcow: what is that mailing list actually for?

[21:23:00.0000] <karlcow>
roc: I guess what it says, but we might see soon.

[21:23:01.0000] <roc>
it doesn't say anything

[21:24:00.0000] <roc>
why do we need a mailing list to discuss browser performance?

[21:25:00.0000] <karlcow>
roc: I think you should ask plh. I found out through twitter

[21:25:01.0000] <roc>
oh, I thought you W3C people had some sort of cabal

[21:25:02.0000] <karlcow>
/me notes to roc that he is not working at W3C

[21:25:03.0000] <roc>
hum

[21:26:00.0000] <roc>
oh well

[21:27:00.0000] <karlcow>
roc: http://twitter.com/w3c/statuses/16421966743

[21:27:01.0000] <karlcow>
and the page says "Discussion related to Web browser performance "

[21:41:00.0000] <MikeSmith>
I forget -- is there a way to express negation with CSS selectors?

[21:41:01.0000] <MikeSmith>
e.g., to express "the first child that is not [something]"?

[21:55:00.0000] <MikeSmith>
roc, karlcow - that list is essentially just for discussion of the Web Timing spec

[21:55:01.0000] <MikeSmith>
at this point

[21:56:00.0000] <karlcow>
ah

[21:56:01.0000] <karlcow>
MikeSmith: thanks. We always need a Mike in one's pocket… hmmm did I write that. I guess so ;)

[21:58:00.0000] <MikeSmith>
http://dev.w3.org/2006/webapi/WebTiming/

[22:01:00.0000] <MikeSmith>
riding on a bus from Tokyo to Nagano is not the best way to get stable Internet connectivity

[22:01:01.0000] <wirepair>
i love nagano

[22:02:00.0000] <MikeSmith>
karlcow: btw, have you been following the current problems with the W3C markup validator

[22:02:01.0000] <MikeSmith>
wirepair: I never been there

[22:02:02.0000] <MikeSmith>
this will be my first visit

[22:02:03.0000] <wirepair>
ah nice, it's very very clean air

[22:02:04.0000] <MikeSmith>
to Matsumoto, actually

[22:02:05.0000] <wirepair>
ahh... 

[22:02:06.0000] <wirepair>
i was out in the woods hehe

[22:02:07.0000] <karlcow>
MikeSmith: about LibXML? I agree with your suggestion

[22:05:00.0000] <karlcow>
checking the page served as text/html with an html validator. The issue is to give the possibility for people who are working with local document. Upload of an XHTML document. You do not know how it will be served. So in the UI it has to be super clear what is happening

[22:16:00.0000] <MikeSmith>
karlcow: I suppose I would not object to having it do XML WF-ness checking on file uploads by default

[22:17:00.0000] <karlcow>
/me replied

[22:17:01.0000] <karlcow>
MikeSmith: have you seen Iwate's Matsuo. Quite impressive. http://www.michaeljohngrist.com/2009/08/the-mist-wreathed-apartments-of-matsuo-mine/   

[22:17:02.0000] <karlcow>
never seen it myself

[22:17:03.0000] <MikeSmith>
but I really don't think it's productive for it to be doing XML WF checking on remote docs served as text/html

[22:18:00.0000] <MikeSmith>
and iirc the performance problem is for checking of remote docs

[22:18:01.0000] <karlcow>
yep agreed

[22:20:00.0000] <MikeSmith>
I seem to remembers somebody commenting on the validator list that they had no problems if they checked files by upload instead of by HTTP

[22:20:01.0000] <MikeSmith>
have not seen

[22:20:02.0000] <MikeSmith>
wow, looks great

[22:21:00.0000] <MikeSmith>
beautiful decay

[22:21:01.0000] <karlcow>
yep

[22:21:02.0000] <MikeSmith>
"dilapidation's progresses are organized decays"

[22:57:00.0000] <hsivonen>
what's the web on tv workshop trying too accomplish?

[22:57:01.0000] <othermaciej>
hello all

[22:57:02.0000] <hsivonen>
hi

[22:58:00.0000] <hsivonen>
s/too/to/

[22:58:01.0000] <othermaciej>
there's a "web on tv" workshop?

[22:59:00.0000] <hsivonen>
according  to the scrollback

[23:01:00.0000] <hsivonen>
what pdf implementation does chrome use?

[23:02:00.0000] <hsivonen>
and will it be in chromium, too?

[23:04:00.0000] <hsivonen>
I wish Gecko came with Poppler built-in, but probably not gonna happen due to licensing

[23:12:00.0000] <hsivonen>
AryehGregor: when you say "Firefox currently", do you mean 3.6 and not the trunk?

[23:12:01.0000] <hsivonen>
my notion of current is the trunk

[01:28:00.0000] <annevk>
sigh

[01:28:01.0000] <annevk>
PFWG persists in their thinking that having three value syntaxes for WAI-ARIA is better than one

[01:29:00.0000] <annevk>
that seems like something an overarching body like the TAG could tackle

[01:29:01.0000] <annevk>
instead of all this polyglot bullshit

[01:30:00.0000] <annevk>
(and with three I mean potentially infinite)

[02:16:00.0000] <Dashiva>
http://twitter.com/simonstl/status/16439272227

[02:19:00.0000] <annevk>
yeah, and some part doesn't work in Chrome, and some not in Opera, and some not in Firefox... ooh

[02:19:01.0000] <Dashiva>
And some parts do work in IE6 :P

[02:20:00.0000] <annevk>
heh

[02:24:00.0000] <jgraham>
/me is disturbed that he seems to have had a dream about breaking and entering Hixie's house

[02:25:00.0000] <jgraham>
Although I think I was looking for him at the time rather than wanting to steal all his stuff

[02:26:00.0000] <jgraham>
And I think (although this part is really hazy) I eventually found him but he looked a lot older than in real life

[02:36:00.0000] <annevk>
sounds like vacation time :)

[02:38:00.0000] <annevk>
also that the PFWG puts DAISY on equal footing with HTML and SVG ... oh well

[02:39:00.0000] <annevk>
maybe we should just define the "host blah blah" mapping in HTML5 and use that everywhere and fix the specification to match reality later

[02:39:01.0000] <annevk>
that's probably the easier way to fix this

[02:43:00.0000] <hsivonen>
annevk: I think the PF really should focus on Web languages

[02:45:00.0000] <annevk>
meanwhile the WOFF people have not read the CORS specification which clearly says the API (@font-face in their case) is in charge of the request semantics whereas they make it dependent on the format

[02:46:00.0000] <annevk>
it is becoming quite clear that the reason the web platform is so inconsistent is because nobody really knows how it works but is extending it in their own vision anyway without regards to precedence and design patterns that came before

[02:47:00.0000] <annevk>
and in this case not even with regards to things that came before, as CORS is being developed pretty much in line with @font-face / WOFF and is pretty clear on how CORS API specifications ought to work and all

[02:50:00.0000] <annevk>
(not that I think it is at all justified that something like @font-face would use CORS, but that is a different matter)

[03:06:00.0000] <hsivonen>
annevk: indeed.

[03:07:00.0000] <hsivonen>
annevk: There's some old stuff that came about by Netscape Netcenter or Bugzilla devs just asking for something and Gecko developers simply said "OK".

[03:08:00.0000] <hsivonen>
/me wonders if Apple and Opera are taking action on fingerprinting. (see https://panopticlick.eff.org/ )

[03:10:00.0000] <abarth>
hsivonen: that's a hard problem

[03:12:00.0000] <hsivonen>
abarth: I think there are easier parts and harder parts.

[03:12:01.0000] <hsivonen>
there are 4 zones: HTTP request, CSS, JS and plug-ins

[03:12:02.0000] <abarth>
does it help to do some but not all of them?

[03:13:00.0000] <hsivonen>
abarth: I think it does

[03:13:01.0000] <jgraham>
presumably that depends on how much information you need to make reliable guesses about identity

[03:13:02.0000] <hsivonen>
abarth: if you do HTTP, you force fingerprinters to expose evidence of their activities to the client

[03:13:03.0000] <abarth>
ic

[03:14:00.0000] <hsivonen>
abarth: which means that big players with the widest ability to track people across the Web would be exposed to regulatory solutions

[03:14:01.0000] <abarth>
i have a link to the Fingerprint moz wiki page

[03:14:02.0000] <abarth>
but i haven't read it yet

[03:15:00.0000] <hsivonen>
my wishful thinking is that the plug-in part can be mitigated by treating everything but Flash as obsolete and by having .swf as click-to-play by default

[03:15:01.0000] <hsivonen>
but I realize this is wishful thinking

[03:16:00.0000] <hsivonen>
also, I have a naive Mac/Linux user POV

[03:17:00.0000] <hsivonen>
I'm saddened to see what kind of NPAPI cruft gets installed on Windows machines

[03:22:00.0000] <hsivonen>
whoa! after a couple of days of having Windows 7 installed and only installing other browsers, security patches, virus scanner and trying to install XP Mode but failing, IE8 running in full screen is already unique

[03:23:00.0000] <hsivonen>
maybe that's because I haven't installed Flash or Java, and not having those is rare

[03:23:01.0000] <jochen__>
hsivonen: i heard you think about reducing fingerprinting possibilities with http?

[03:23:02.0000] <jochen__>
hsivonen: any plan what to do there without making half of the webpages unusable? :)

[03:24:00.0000] <hsivonen>
jochen__: well, I filed all the obvious pieces of the HTTP part of the puzzle as bugs yesterday, but I'm not sure where it's going to go

[03:24:01.0000] <hsivonen>
jochen__: as far as I can tell, the only hard part about HTTP request is Accept-Language

[03:24:02.0000] <jochen__>
mind linking them from wiki.mozzila.org/Fingerprinting?

[03:25:00.0000] <hsivonen>
jochen__: the tracking bug is https://bugzilla.mozilla.org/show_bug.cgi?id=http-fingerprint

[03:25:01.0000] <jochen__>
ta

[03:25:02.0000] <hsivonen>
jochen__: I'll add it to teh wiki

[03:28:00.0000] <hsivonen>
oh yeah, OS/CPU is probably politically hard, even though is terms of actual Web compat you get far without them

[03:30:00.0000] <volkmar>
i see some css selectors are defined in html5 specs, does that mean a request/issues related to that should be expressed in whatwg ml?

[03:30:01.0000] <volkmar>
or i should use www-style

[03:33:00.0000] <Matjas>
tabindex=0 enables keyboard navigation for any element. However, I would expect the click event to fire for those elements when pressing Enter/Space while they’re focused. Any reason why that’s not the case?

[03:34:00.0000] <Matjas>
Enter/Space fires click for every element that’s keyboard-focusable by default

[03:36:00.0000] <annevk>
I argued that we should do that

[03:36:01.0000] <annevk>
and make onclick="" elements focusable

[03:36:02.0000] <annevk>
but not enough browsers did it to begin with so they would first have to be changed as the reasoning, I believe

[03:37:00.0000] <hsivonen>
annevk: Hixie's reasoning, PF reasoning or someone else's reasoning?

[03:43:00.0000] <annevk>
Hixie; he just wanted to match reality for now

[03:43:01.0000] <annevk>
(though Opera has the nicer behavior afaik)

[03:46:00.0000] <jgraham>
hsivonen: <b><button></b>. AFAICT that should reopen the <b> element inside the <button>, but it seems that gecko does not

[03:48:00.0000] <jgraham>
hsivonen: Am I missing something?

[03:57:00.0000] <Matjas>
annevk: Exactly. What Opera does feels like "the right thing"

[04:01:00.0000] <annevk>
hsivonen, interesting work on fingerprinting

[04:02:00.0000] <annevk>
hsivonen, I think I'll copy some of your bugs and file them on Opera; see what happens

[04:07:00.0000] <annevk>
nuking Accept-Charset makes a lot of sense

[04:17:00.0000] <hsivonen>
annevk: the sad part is that there's so much fear of making Firefox lead on this issue. The parts that are the easiest to argue are the ones where IE is already leading... Like Accept-Charset.

[04:18:00.0000] <hsivonen>
jgraham: the <button> situation sucks. I haven't properly though about <b><button></b>.

[04:18:01.0000] <hsivonen>
one option is to just revert it into scoping and let the German Figure Collector site break

[04:18:02.0000] <daedb>
woah, .Net framework adds itself to Firefox's ua string now too, not just IE :O

[04:20:00.0000] <hsivonen>
daedb: using what mechanism? Whatever ships with Windows 7 doesn't tamper with the UA string of Minefield

[04:20:01.0000] <hsivonen>
everyone wants to add their mark to the UA string :-(

[04:20:02.0000] <daedb>
I don't know, but I have "(.NET CLR 3.5.30729)" at the end of my Firefox ua string right now (and Firefox was installed long after .Net 3.5)

[04:21:00.0000] <hsivonen>
daedb: do you see a .Net-specific add-on?

[04:21:01.0000] <daedb>
The :net framework assistant, yes, but it's disabled.

[04:22:00.0000] <annevk>
hsivonen, the others are more tricky I guess, yes

[04:23:00.0000] <hsivonen>
speaking of HTTP request headers, Opera Mobile on Maemo sure has a long Accept header

[04:31:00.0000] <karlcow>
hsivonen: worse than everyone wants to add their mark, nobody wants to remove the old ones. the pile is going up, up, up. :) I wonder if there is a physical limit to the size of UA string (characters number)

[04:32:00.0000] <annevk>
we cleaned it up a little bit

[04:32:01.0000] <annevk>
but then it exploded again ffs

[04:33:00.0000] <hsivonen>
annevk: Opera has cruft in the various Accept headers that clearly isn't critical for Web compat

[04:33:01.0000] <hsivonen>
Accept-Encoding is longer than in IE and Gecko, for example

[04:33:02.0000] <hsivonen>
Accept-Charset advertises UTF-16!!!!

[04:34:00.0000] <hsivonen>
those aren't fingerprinting-relevant, but just saying

[04:34:01.0000] <karlcow>
hehe

[04:34:02.0000] <annevk>
hsivonen, bug on accept-charset is filed

[04:35:00.0000] <hsivonen>
Mozilla has a "guardian of the Accept header", but clearly, there should be "guardian of the UA string, too"

[04:35:01.0000] <hsivonen>
s/, too"/", too/

[04:35:02.0000] <annevk>
what is "identity" in A-E? or "sdch" (Chrome has that)

[04:36:00.0000] <annevk>
we even have spaces between the values

[04:36:01.0000] <annevk>
very generous

[04:37:00.0000] <hsivonen>
annevk: Shared Dictionary Compression over HTTP says the first hit on Google

[04:37:01.0000] <hsivonen>
IIRC, identity is just sending the plain bytes

[04:37:02.0000] <hsivonen>
without compression

[04:37:03.0000] <annevk>
http://groups.google.com/group/SDCH

[04:39:00.0000] <karlcow>
http://www.w3.org/TR/ct-guidelines/#sec-Mobile-DOCTYPES

[04:40:00.0000] <hsivonen>
karlcow: that's sad

[04:40:01.0000] <karlcow>
reality is sad ;) 

[04:41:00.0000] <annevk>
filed a bug on Accept-Encoding

[04:41:01.0000] <annevk>
"This list is not exhaustive and is likely to change." jaja

[04:41:02.0000] <hsivonen>
karlcow: fortunately, Fennec, Opera Mobile, Opera Mini and Mobile Safari work despite "reality"

[04:42:00.0000] <annevk>
2.0 of "Web Content Transformation Proxies" will say they realize now the whole thing is sort of dead

[04:42:01.0000] <karlcow>
hsivonen: do you mean they are happy? :)

[04:42:02.0000] <annevk>
(I'm making that up, but that appears to be reality)

[04:42:03.0000] <hsivonen>
karlcow: they seem happy to me :-)

[04:43:00.0000] <karlcow>
/me is making apologizes for his morning syllogism :) still only one eye opened. :)

[04:44:00.0000] <annevk>
another thing that is sad is that Accept: is still not */*

[04:45:00.0000] <hsivonen>
annevk: yeah, though it might be wise to wait for IE9 to have been in the market for a while before trying to change that

[04:45:01.0000] <annevk>
I thought IE had that since IE6 or so

[04:46:00.0000] <hsivonen>
annevk: I meant the reason not to have Accept: */* is negotiating application/xhtml+xml

[04:46:01.0000] <annevk>
but we don't want that :)

[04:47:00.0000] <hsivonen>
if IE9 says */* but support XHTML+SVG *maybe* it'll make Accept-based negotiation less popular

[04:47:01.0000] <annevk>
more often than not XML does harm to our end users

[04:49:00.0000] <karlcow>
what would be the benefit of Accept: */*, because I had a use case in the past where a precise Accept would have helped filtering a DDOS

[04:50:00.0000] <hsivonen>
karlcow: fewer bytes to send, fewer problems for non-IE browsers when webmasters goof

[04:50:01.0000] <annevk>
omitting it entirely breaks

[04:51:00.0000] <karlcow>
my issue was the Accept for GET /blabla resulting from an IMG 

[04:51:01.0000] <hsivonen>
karlcow: you should just fix your site :-)

[04:51:02.0000] <karlcow>
I had no control on it

[04:52:00.0000] <karlcow>
the ddos was generated by a site with a very high trafic and a bad link.

[04:52:01.0000] <hsivonen>
karlcow: 99.99999% of webmasters wouldn't apply your mitigation method anyway, so it's not useful to support your mitigation method on the browser side

[04:52:02.0000] <annevk>
karlcow, you could have done something based on Referer

[04:53:00.0000] <hsivonen>
karlcow: if it was cross-origin, the Origin header will save us

[04:54:00.0000] <karlcow>
annevk: nope. because it would have removed valid request too :) but I'm more sympathetic to the argument of hsivonen. In the end, we just block the query on that page until we found out where the problem was :)

[04:55:00.0000] <karlcow>
s/block/blocked/

[04:55:01.0000] <annevk>
no it wouldn't

[04:55:02.0000] <annevk>
if you just blocked that specific referer for that resource it would've been fine

[05:07:00.0000] <Dashiva>
karlcow: I don't see how it makes sense to talk about a stable snapshot or milestone of an incomplete in-development specification

[05:07:01.0000] <Dashiva>
You can make snapshots and you can assign milestones, but they are arbitrary, not stable

[05:08:00.0000] <jgraham>
My pet example of why /TR is evil is WebIDL

[05:09:00.0000] <jgraham>
I have eventually learnt that you can find the actual latest version by being careful about whether you insert a space between Web and IDL in google

[05:09:01.0000] <Dashiva>
karlcow: If you agree about the stable part, wouldn't a permalink in each editor's draft satisfy the snapshot use case?

[05:10:00.0000] <annevk>
great, Jonny was already on Accept-Encoding simplification

[05:23:00.0000] <zcorpan_>
shouldn't the web timing spec put its stuff on the navigator object?

[05:24:00.0000] <zcorpan_>
/me imagines window.performance is already taken by web authors

[05:26:00.0000] <annevk>
yeah, you should say that

[05:26:01.0000] <annevk>
putting new stuff on navigator when it is somewhat sensible is better than bloating window even more

[05:30:00.0000] <hsivonen>
there should be a standing memo for spec editors about that

[05:30:01.0000] <jgraham>
But everyone loves everything being in the global scope, right?

[05:34:00.0000] <Dashiva>
Maybe we need a new parent object for stuff that doesn't really fit on navigator :)

[05:36:00.0000] <jgraham>
Personally I want more properties on the window object that just refer to the window object

[05:37:00.0000] <jgraham>
window.Hofstadter for example

[05:38:00.0000] <annevk>
Dashiva, the bar for navigator is low enough

[06:13:00.0000] <annevk>
god

[06:13:01.0000] <annevk>
why would http://html5.org/tools/web-apps-tracker?from=5105&to=5106 fail but http://html5.org/tools/web-apps-tracker?from=5104&to=5105 would work?

[06:18:00.0000] <hsivonen>
/me wonders why Opera still ships with both the "Opera Standard" and "Windows Native" skins when the "Opera Standard" skin is *more* Windows 7 -like of the two very similar skins

[06:29:00.0000] <annevk>
hsivonen, we'll soonish no longer do Accept-Language

[06:30:00.0000] <annevk>
maybe I should file some more bugs on our network module; he's being pretty active :)

[06:30:01.0000] <annevk>
euh, Accept-Charset

[06:30:02.0000] <annevk>
my bad

[06:30:03.0000] <hsivonen>
annevk: cool anyway

[06:37:00.0000] <hsivonen>
aargh. Radiance and Ambiance are FAIL with Eclipse

[06:38:00.0000] <hsivonen>
still wondering how Canonical didn't have some kind of organizational function to stop the Lucid default theme

[06:38:01.0000] <annevk>
btw, I'm not really part of Opera QA, but then I don't really have a better description available either :)

[06:38:02.0000] <jgraham>
It is pretty awful

[06:38:03.0000] <hsivonen>
annevk: sorry

[06:38:04.0000] <jgraham>
Not anne's job, the Lucid theme

[06:39:00.0000] <annevk>
heh

[06:39:01.0000] <annevk>
how do you know? :p

[06:39:02.0000] <hsivonen>
jgraham: they should hire whoever is developing Elemental 

[06:40:00.0000] <hsivonen>
(but having him/her/they make a theme that doesn't feel like too much of an Apple rip-opp)

[06:40:01.0000] <hsivonen>
*off

[06:41:00.0000] <hsivonen>
also, Radiance could be tweaked not to suck

[06:41:01.0000] <hsivonen>
Ambiance is just massive FAIL

[06:42:00.0000] <jgraham>
It's hard to believe that the people who designed it actually use it

[06:48:00.0000] <hsivonen>
s/Elemental/Elementary/

[07:44:00.0000] <karlushi>
Dashiva, it could be. :) we just have to find a model which makes it usable for different types of communities and not too burdensome to manage. Often we know our current burden, it's rare we understand the issues we create when changing. But yes it could be a good idea.

[07:51:00.0000] <jgraham>
karlushi: I am really not sure who the community is that wants a version of the spec with fewer bugfixes in than the current version

[07:52:00.0000] <jgraham>
I think the stable/editor's draft model is designed around an idea of spec development that doesn't really match reality, at least for the way HTML5 is developed

[07:54:00.0000] <hsivonen>
/me observes that when document.write() implies a call to document.open() in Gecko, the call is always to the built-in open() and not to whatever the page has replaced open() with

[07:54:01.0000] <karlushi>
I do not claim reality. ;) Just trying to identify use cases.

[07:55:00.0000] <hsivonen>
karlushi: so what's the use case of reading a spec you know is an already-obsolete draft when you know you could read the latest draft?

[07:56:00.0000] <jgraham>
That is, if you have stable versions, N, N+1, etc. and unstable drafts at all times t, it only makes sense to promote version N if, on averge, version N is closer to version N+1 than unstable version at time t(N)<t<t(N+1)

[07:56:01.0000] <karlushi>
hsivonen, will there be at a point a stable version?

[07:57:00.0000] <hsivonen>
karlushi: I expect there will be, but not one that's useful to read

[07:57:01.0000] <karlushi>
hsivonen, for you and implementers, not useful indeed.

[07:58:00.0000] <Lachy>
hsivonen, the use case for snapshots is so you can compare what was written at the time of the snapshot with what is now written, which is easier than digging through CVS/SVN logs.

[07:59:00.0000] <jgraham>
karlushi: It is possible to be unstable, but to be convering on a stable point

[07:59:01.0000] <hsivonen>
Lachy: that's a reasonable use case

[07:59:02.0000] <karlushi>
translation is another use case. Translating a document which is changing all the time is hell. 

[08:00:00.0000] <hsivonen>
karlushi: you just pick the draft that's the most current when you start translating

[08:00:01.0000] <karlushi>
s/you/one/ who's not necessary a geek.

[08:02:00.0000] <gsnedders>
/me likes what jgraham is typing now

[08:02:01.0000] <jgraham>
Heh, putting <plaintext> in <table> so it gets foster parented is amusing

[08:02:02.0000] <jgraham>
gsnedders: tinfb

[08:02:03.0000] <gsnedders>
/me is psycic

[09:12:00.0000] <annevk>
hmm, guess I'll look into the html5.org issue tomorrow or so

[09:12:01.0000] <annevk>
not gonna happen now

[10:14:00.0000] <AryehGregor>
hsivonen, my notion of "current" for most software is "current version I'm using", which might be current stable or trunk or something in between or maybe slightly older than current stable.  In this case, 3.6; didn't bother to test on trunk because it didn't matter.  I should have specified, though, since it's ambiguous.

[12:51:00.0000] <Gurpartap>
HTML6 http://www.artisopensource.net/hacks/deliciouspoetry.php?cshow=5217210

[13:22:00.0000] <jrgarrison>
I'm confused about something mentioned in the last paragraph of the message at http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2010-June/026728.html

[13:22:01.0000] <jrgarrison>
"I should have written input type=file accept='audio/x-speex;quality=(0-10);bitrate=...;source=microphone' to reflect the newly-announced Android/Firefox placement of the source= parameter."

[13:23:00.0000] <jrgarrison>
does anybody know what announcement this refers to?

[13:38:00.0000] <AryehGregor>
/me has no idea.

[13:48:00.0000] <variable>
question: is there a slightly more offical method of tracking proposals than the mailing list? I know about the bug list for the W3C. Does the WHATWG have anything similar?

[14:08:00.0000] <Dashiva>
The issue tracker

[14:09:00.0000] <Dashiva>
http://www.whatwg.org/issues/

[14:11:00.0000] <variable>
Dashiva, thank

[14:33:00.0000] <AryehGregor>
My giant comment got accepted.  Everyone feel free to complain about how terribly wrong I am, particularly since I don't have such a good grasp on the history here. http://www.cssquirrel.com/2010/06/15/693comic-update-html5-unicorn-heuristics/comment-page-1/#comment-32000

[14:44:00.0000] <johnst>
tl;dr

[14:44:01.0000] <AryehGregor>
:)

[16:19:00.0000] <Hixie>
annevk: index didn't change in r5106 but did in the other two

[16:46:00.0000] <karlcow>
AryehGregor: only one wrong fact in the comment

[16:47:00.0000] <karlcow>
"this is directly due to completely wrong-headed thinking by the W3C administration." 

[16:47:01.0000] <karlcow>
s/W3C Administration/W3C Members/

[16:50:00.0000] <karlcow>
Also "The W3C has to this day contributed almost nothing to the spec for practical purposes." makes me huh?

[16:50:01.0000] <karlcow>
W3C = Members of W3C

[16:50:02.0000] <karlcow>
if you meant W3C employees have not contributed. Then 1) it's not their primary role. 2) They did contributed in terms of handling all the administrivia of WGs life, liaison, setting mailing lists, web spaces. I would be surprised to consider MikeSmith work = nothing


2010-06-19
[20:57:00.0000] <patx>
<video src="file:///home/patx/out.ogv"></video>  Thats does not seem to work

[20:57:01.0000] <patx>
why is this?

[21:08:00.0000] <jwalden>
patx: what browser, and is the file containing that directly in /, /home/, or /home/patx/?

[21:08:01.0000] <patx>
/home/patx

[21:08:02.0000] <patx>
chrome

[21:09:00.0000] <jwalden>
hm

[21:09:01.0000] <jwalden>
I don't know chrome's local-file security model, but I'm pretty sure it should permit that

[21:12:00.0000] <patx>
doesnt work on firefox either

[21:12:01.0000] <jwalden>
are you sure the file isn't malformed?

[21:12:02.0000] <jwalden>
I *know* Firefox's security model is fine with a local HTML file accessing a sibling file

[21:13:00.0000] <jwalden>
what does file properties in nautilus say is the type of the file?

[21:23:00.0000] <patx>
now its working.

[21:23:01.0000] <patx>
hmm

[00:15:00.0000] <patx>
So I am having a little problem... http://patx.me/paste/6183.html

[00:19:00.0000] <Hixie>
patx: we'll prolly need to see the code to be able to help

[00:19:01.0000] <patx>
Hixie, the python code or the html5?

[00:20:00.0000] <Hixie>
the code the browsers get

[00:20:01.0000] <Hixie>
so html

[00:21:00.0000] <hsivonen>
patx: when on the local file system, the html file and the video file need to be in the same directory

[00:21:01.0000] <hsivonen>
patx: when on server, .ogv needs to be mapped to video/ogg

[00:22:00.0000] <patx>
oh ok

[00:22:01.0000] <patx>
hsivonen, that explains it!

[00:22:02.0000] <patx>
thanks a bunch

[00:22:03.0000] <patx>
theyre in diff directorys

[00:22:04.0000] <Hixie>
ah yeah that'll do it

[00:23:00.0000] <Hixie>
that's to protect you from downloading an HTML file that then grabs your system passwords or something

[00:23:01.0000] <patx>
oh ok!

[00:23:02.0000] <patx>
thanks

[00:25:00.0000] <patx>
Hixie - hsivonen : <video controls="controls" preload="metadata" src="file:///home/patx/Desktop/5496.ogv" width="800"></video> 

[00:25:01.0000] <patx>
so i cant do that?

[00:26:00.0000] <Hixie>
generally speaking whether file:// works at all and how it works is really up to the browser -- but in most browsers, it'll only work if the video file is in the same directory as the HTML file, when you use file://

[00:26:01.0000] <Hixie>
easiest solution is to just say src="5496.ogv" instead of saying file://... etc.

[00:26:02.0000] <Hixie>
and just make sure they're in the same directory

[00:40:00.0000] <patx>
ok thanks

[04:22:00.0000] <jgraham>
Is there some special reason that web workers disallows workers in data: uris?

[04:30:00.0000] <Dashiva>
Seems to be r3857

[04:31:00.0000] <Dashiva>
Origin reasons, apparently

[04:43:00.0000] <jgraham>
Dashiva: Yeah, but what does banning data uris gain us?

[04:44:00.0000] <gsnedders>
/me wonders if jgraham ever stops working on HTML5-related stuff

[04:44:01.0000] <Dashiva>
Isn't it just that since data: URIs have unique origins, they just don't work? Disclaimer: I don't really know why.

[04:45:00.0000] <gsnedders>
Dashiva: pong, seeming you pinged me last week while I was away

[04:47:00.0000] <Dashiva>
I think the TTL expired long ago

[04:48:00.0000] <Dashiva>
I can't seem to find the original request, are you sure it was me?

[04:48:01.0000] <gsnedders>
Dashiva: Yes

[05:22:00.0000] <jgraham>
gsnedders: I really don't do that much useful stuff. I would really like to get more done

[05:22:01.0000] <jgraham>
OTOH, it is fun

[05:22:02.0000] <jgraham>
More fun than whining about  it anyway

[05:39:00.0000] <annevk>
I fixed the issue on html5.org

[05:39:01.0000] <annevk>
not in a very pretty way

[05:53:00.0000] <hsivonen>
annevk: what was the root cause?

[05:54:00.0000] <annevk>
before when a diff was empty it would return an empty string

[05:55:00.0000] <hsivonen>
oh. an empty diff

[05:55:01.0000] <annevk>
apparently SVN changed and now something is at the top at all times leading my script to assume there was a diff and not finding a number and therefore throwing an error

[05:56:00.0000] <annevk>
so now I changed to a try / except where I try to find the number given a diff and if that fails I assume there was no diff

[06:12:00.0000] <hsivonen>
aargh. https://answers.edge.launchpad.net/launchpad/+faq/1024 is *so* annoying. I wish the Origin header had full deployment already

[07:47:00.0000] <annevk>
oh hey, Chrome nightlies support SVG referenced from the 'content' property

[07:47:01.0000] <annevk>
so they finally render my blog

[08:46:00.0000] <patx>
Hixie, hsivonen - i was talking to u last night... i put the vid in the same dir as the html... however i get the same result...

[12:53:00.0000] <jgraham>
Philip`: You on public-html-testsuite? We should import your canvas tests into the official testsuite

[12:53:01.0000] <jgraham>
So you can do it if you like, or I can do it if it is more convenient

[12:57:00.0000] <Philip`>
jgraham: I'm on the list

[12:57:01.0000] <Philip`>
(though I've been away for a week so I've missed anything that might have been said on it or elsewhere)

[12:57:02.0000] <Philip`>
I prefer solutions that don't require me to do any work

[12:59:00.0000] <Philip`>
though I also like solutions that let me update the tests by editing the source files instead of manually touching the HTML files, and I don't know how that'll interact with any importing of the tests

[13:00:00.0000] <jgraham>
Philip`: I think that is OK. I think we can say that the source is the canoncial version and everything else is just generated

[13:00:01.0000] <jgraham>
I hope so because we will have to do the same with the html5lib tests


2010-06-20
[18:48:00.0000] <gsnedders>
jgraham, Philip`: was there not a discussion about importing them before?

[18:52:00.0000] <Philip`>
gsnedders: I think there have been many discussions

[18:52:01.0000] <Philip`>
but then nothing really happened

[18:53:00.0000] <Philip`>
(As far as I'm aware, Mozilla/WebKit/Opera have already got imported copies of the tests, so there's not much immediate practical benefit from importing it into anywhere else)

[18:54:00.0000] <Philip`>
(so I don't really care much about it myself at the moment)

[03:05:00.0000] <Dashiva>
shepazu seems to have the priority of constituencies principle backwards. The WHATWG spec is the one trying to accomodate authors and implementors by being useful and convenient, the W3C one splits and cuts for theoretical purity

[03:44:00.0000] <jgraham>
I should note that the W3C has never agreed to converge on the design principles adopted by the WHATWG, as we are unfailingly reminded whenever they come up

[04:18:00.0000] <Dashiva>
hsivonen: I would like to see evidence of these so-called webm lolcats

[05:55:00.0000] <MikeSmith>
"We don’t write firmware for missile guidance systems, we write applications that let you pretend to steal your friends’ cartoon cows."

[05:55:01.0000] <MikeSmith>
http://garann.wordpress.com/2010/06/19/defending-terrible-things/

[06:05:00.0000] <erlehmann>
MikeSmith, I HATE THIS THING WORDPRESS DOES

[06:05:01.0000] <erlehmann>
i even defined my own quicktags that actually insert <i>s and <b>s

[07:48:00.0000] <karlcow>
do not ask what others do wrong, ask yourself what others do right, and what you do wrong ;)

[10:11:00.0000] <karlcow>
http://www.aldebaran-robotics.com/en/nao_robot_interactif

[10:12:00.0000] <karlcow>
http://www.youtube.com/watch?v=4t1NWH6G1f0 

[10:12:01.0000] <karlcow>
programmable in python

[10:28:00.0000] <annevk>
requires Flash :/

[10:36:00.0000] <karlcow>
annevk: yep, when in fact there is nothing requiring flash in these specific pages.

[11:13:00.0000] <AryehGregor>
karlcow, my impression is that: 1) The browser implementers wanted to create a W3C working group for Web Applications 1.0 or such, but the W3C administration (whoever has to approve new WGs?) denied it because it conflicted with the direction they were trying to take the web.  2) While people have certainly done lots of useful stuff in the W3C, almost all of that could just as well have been done in the WHATWG, and most would have been if the

[11:13:01.0000] <AryehGregor>
 HTMLWG was never created.

[11:18:00.0000] <AryehGregor>
I freely admit to being a total ignoramus posting from the peanut gallery, though.

[11:19:00.0000] <karlcow>
AryehGregor: at a few occasions the W3C had workshop on the future of the Web (including html). Members decided to go towards XML. They had invested time and money into it. There is the 2004 workshop, but there is also the one before this. Anyway following the Process a majority of people agreed to continue going to the XML road. 

[11:19:01.0000] <karlcow>
Only a handful of members wanted to reopen the work on html. Logically this is the road which has been followed. (not judging about the pertinence specifically at that time). I would count myself at that time in the XHTML camp for many reasons. Put let's put that aside.

[11:19:02.0000] <karlcow>
Maybe the mistake is to not have given the possibilities of two Working groups at the same time. BUT at this time, the W3C (aka organization means members) was often targeted for lack of inconsistencies by the public with regards to competing technologies such as css and xsl.

[11:19:03.0000] <karlcow>
it seems a natural and easy decision to make now. It was not in the past.

[11:19:04.0000] <AryehGregor>
So you're saying the decision was made by the non-implementer members of the W3C, not by its administration.

[11:19:05.0000] <karlcow>
nope

[11:20:00.0000] <karlcow>
different implementers, not the same market segment. 

[11:20:01.0000] <AryehGregor>
Ah.

[11:20:02.0000] <AryehGregor>
. . . but not, like, implementers of mainstream web browsers.

[11:20:03.0000] <karlcow>
right.

[11:20:04.0000] <AryehGregor>
Okay.

[11:21:00.0000] <AryehGregor>
Thanks for the correction.

[11:24:00.0000] <karlcow>
no issue. It's difficult. because everything is happening in times with not necessary the same people either. People like annevk, hsivonen and Lachlan were not born (kidding), but let's say they are the 2nd if not the 3rd generation of implementers, and they have bitten the dust of web failures. There is often a generation conflict in the group. It is interesting to see.

[11:24:01.0000] <karlcow>
and I'm pretty sure when the html5 generation will be older, the new generation will be very tough on the decision we are making these days. :))) whatever we could say to justify them.

[11:25:00.0000] <gsnedders>
(and I get called a baby)

[11:25:01.0000] <gsnedders>
:P

[11:25:02.0000] <karlcow>
;)

[11:29:00.0000] <AryehGregor>
Does anyone know how hard it would be to rewrite, e.g., the infobox in this article using non-table-based layout (working in IE6 is required)?  http://en.wikipedia.org/wiki/United_States

[11:29:01.0000] <AryehGregor>
It seems like it would be pretty easy unless you insisted on the auto-width thing for the columns . . .

[11:30:00.0000] <AryehGregor>
You could probably just set a width of 50% on the columns and it would look fine.

[11:32:00.0000] <AryehGregor>
/me isn't so great at CSS.

[12:04:00.0000] <jgraham>
karlcow: One can make the reasonable claim that a Process in which implementors in one market segment can block work in a different market segment is pretty broken

[12:44:00.0000] <AryehGregor>
Phone numbers should be assigned so that the minimum pairwise Levenshtein distance is maximal.  In particular, it should always be strictly greater than one.

[12:44:01.0000] <AryehGregor>
/me just got two people calling his cell phone by mistake in like the last hour

[12:51:00.0000] <invariable>
AryehGregor, my phone number is the same one a pharmacy had 20 years ago. I still get people calling for prescriptions 

[12:52:00.0000] <AryehGregor>
:/

[13:00:00.0000] <invariable>
karlcow, what market segments were pushing for XML? Where exactly is the dichotomy?

[13:16:00.0000] <hsivonen>
one problem the W3C has is that it pretends to be one happy community when there really are several communities

[13:18:00.0000] <hsivonen>
the Semantic Web community and the browsable Web community are almost disjoint and there's the ex-SGML community turned XML community somewhere in between


2010-06-21
[17:12:00.0000] <karlcow>
http://blog.mozilla.com/faaborg/2010/04/22/dont-talk-about-users/

[17:34:00.0000] <roc>
karlcow: we fought against the results of the 2004 workshop (and earlier decisions) as much as we could

[17:35:00.0000] <roc>
I hope you're not suggesting that no-one had the foresight to see that the W3C's detour down the XML road was a mistake

[20:27:00.0000] <zcorpan_>
i guess flock 3 beta can't play h.264

[20:38:00.0000] <zcorpan_>
has anyone tested vp8 in flash 10.1?

[20:55:00.0000] <roc>
does flash 10.1 support vp*?

[20:55:01.0000] <roc>
vp8?

[20:55:02.0000] <roc>
I'd be surprised

[20:56:00.0000] <karlcow>
it was a "will" in May http://blogs.adobe.com/flashplatform/2010/05/adobe_support_for_vp8.html but I do not think it is the case now.

[21:08:00.0000] <zcorpan_>
yay, minefield has -76

[21:08:01.0000] <zcorpan_>
let's see how it fares on my tests

[21:11:00.0000] <zcorpan_>
hmm, they have .URL

[21:11:01.0000] <roc>
websockets?

[21:11:02.0000] <zcorpan_>
yes

[21:12:00.0000] <roc>
how does it fare on your tests?

[21:13:00.0000] <zcorpan_>
you should look over the handling of the opening handshake, most tests there fail

[21:13:01.0000] <zcorpan_>
broken utf-8 in messages fail

[21:15:00.0000] <zcorpan_>
87 passed out of 159

[21:18:00.0000] <zcorpan_>
at least it doesn't crash like chrome

[21:21:00.0000] <zcorpan_>
http://krijnhoetmer.nl/irc-logs/whatwg/20100603#l-706

[21:27:00.0000] <roc>
zcorpan_: hmm, so your tests aren't publish?

[21:27:01.0000] <roc>
public?

[21:28:00.0000] <zcorpan_>
not yet, no

[21:38:00.0000] <zcorpan_>
i don't get any close event for invalid handshakes

[21:44:00.0000] <franksalim>
zcorpan_, hey that's exciting. not the missing close events, but -76 support

[21:44:01.0000] <zcorpan_>
franksalim: yes

[03:10:00.0000] <gsnedders>
/me finds out about the lexical enviroment scope of event handlers

[03:10:01.0000] <gsnedders>
Wow, that's so much more screwed up than

[03:10:02.0000] <gsnedders>
I thought it was just the element's object.

[04:51:00.0000] <gsnedders>
http://www.whatwg.org/specs/web-apps/current-work/#event-handler-content-attributes  requires the form owner to be on the lexical scope

[04:51:01.0000] <gsnedders>
WebKit does this, Gecko and Presto don't

[04:51:02.0000] <gsnedders>
/me wonders why it's needed

[04:52:00.0000] <jgraham>
WDIED?

[04:52:01.0000] <gsnedders>
Dunno, don't have IE

[05:41:00.0000] <zcorpan_>
maybe we should allow xmlns:xlink on html elements?

[05:43:00.0000] <annevk>
SVG should just get native href/src already

[05:43:01.0000] <annevk>
that's in the planning for quite a while now, at which point everything else will be moot

[05:46:00.0000] <zcorpan_>
hmm. so now what, the hybi decides to ignore all feedback that's sent to whatwg?

[05:47:00.0000] <annevk>
I somewhat doubt Hixie agreed to all what that guy said

[05:47:01.0000] <jgraham>
HyBi are moving so slowly they are in real danger of being irrelevant

[05:48:00.0000] <zcorpan_>
maybe the whatwg should declare that the hybi version is unofficial :)

[05:51:00.0000] <jgraham>
The best part about that email was how they declared that everyone had to move from the WHATWG mailing list to the IETF one but didn't actually CC the email to WHATWG

[05:55:00.0000] <hsivonen>
what's the upside of having hybi involved in Web Socket?

[05:56:00.0000] <annevk>
it's somewhat unclear at this point

[05:56:01.0000] <annevk>
aiui

[06:02:00.0000] <zcorpan_>
jgraham: sending emails to whatwg is counter productive

[06:03:00.0000] <annevk>
jgraham, yeah, noticed that too; wtf

[06:08:00.0000] <hsivonen>
zcorpan_: depends on the content of the emails :-)

[06:10:00.0000] <jgraham>
hsivonen: AFAICT the value proposition is that IETF has been around a long time. I'm not sure exactly what value that is bringing though

[06:11:00.0000] <hsivonen>
/me fails to see how having been around is a value proposition

[06:11:01.0000] <jgraham>
Well that seems to be the major difference between IETF and WHATWG

[06:13:00.0000] <jgraham>
I presume the value proposition that implies is having a preexisting concentration of expertise

[06:13:01.0000] <jgraham>
Not sure how that is panning out with websockets though

[06:14:00.0000] <hsivonen>
jgraham: so far, I'm unimpressed by "expertise" value propositions around HTML5

[06:43:00.0000] <franksalim>
zcorpan_, the IETF version already declares itself unoffical: "NOTE!  THIS COPY OF THIS DOCUMENT IS OBSOLETE."

[06:45:00.0000] <franksalim>
which i now see is mentioned in the latest email to HyBi

[06:47:00.0000] <franksalim>
has any of the feedback from the HyBi working group been incorporated? some of it was quite good, in my opinion

[06:53:00.0000] <zcorpan_>
franksalim: i see some hybi on the list of pending emails in http://www.whatwg.org/issues/#websocket

[06:54:00.0000] <zcorpan_>
franksalim: i agree some feedback from hybi was good

[07:50:00.0000] <zcorpan_>
bufferedAmount doesn't seem to work so well in chrome and firefox -- if i use variants of the bufferedAmount example in the spec, both chrome and firefox send lots of messages in the beginning while almost no messages at all after a short while

[07:51:00.0000] <zcorpan_>
opera seems to more evenly send messages

[08:00:00.0000] <annevk>
I think it's a bit sad there's been more time to consider the handshake feedback a little better

[08:01:00.0000] <annevk>
not having an HTTP-based handshake and integrating the handshake for TLS connections in the TLS layer would have been nicer I think

[08:01:01.0000] <annevk>
but getting there now seems rather hard

[08:01:02.0000] <annevk>
though I suppose yet one more change could be ok, if it happens quickly enough

[08:08:00.0000] <franksalim>
annevk, are you talking about next protocol negotiation in TLS?

[08:09:00.0000] <annevk>
the email from Adam Barth

[08:09:01.0000] <jgraham>
annevk: I think I would hate to write a server with that handshake

[08:09:02.0000] <jgraham>
At least I would expect it to be much harder

[08:10:00.0000] <annevk>
jgraham, there would be a far simpler handshake for non-TLS connections in my vision of how it ought to be...

[08:10:01.0000] <annevk>
jgraham, without all the HTTP crap

[08:10:02.0000] <jgraham>
annevk: Different handshakes depending on the connection type? Sounds odd

[08:10:03.0000] <franksalim>
i think that is a good design, however http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00 is deployed almost precisely nowhere

[08:10:04.0000] <franksalim>
and is in draft stage

[08:11:00.0000] <annevk>
franksalim, the "it is in draft stage" is how some W3C groups end up designing formats on top of XSL:FO rather than CSS

[08:11:01.0000] <annevk>
franksalim, it's not really a compelling argument in my view :)

[08:11:02.0000] <franksalim>
fair enough

[08:12:00.0000] <jgraham>
"it's deployed almost precisely nowhere" is a good argument though

[08:12:01.0000] <franksalim>
but one of the appeals of websocket, imho, is that it works with a lot of what we already have

[08:12:02.0000] <franksalim>
we would need updated TLS libraries

[08:12:03.0000] <franksalim>
and that means you wouldn't be able to do the nice little python/perl/* websocket servers

[08:12:04.0000] <franksalim>
in a few dozen lines

[08:13:00.0000] <franksalim>
but when nextprotoneg is everywhere, we can do websocket2 very easily

[08:15:00.0000] <franksalim>
or not, since nextprotoneg everywhere probably implies SPDY

[08:17:00.0000] <annevk>
well, if TLS is going to be the dominant form as some people seem to indicate, it would be nice if it didn't require more roundtrips than needed

[08:18:00.0000] <annevk>
and for the real simple deployment scenario you can have a non-TLS handshake that is simpler than what we have now

[10:43:00.0000] <TabAtkins>
JonathanNeal: Where's your HTML5 default UA stylesheet?  Google isn't helpful right now.

[10:46:00.0000] <JonathanNeal>
Heyo

[10:46:01.0000] <JonathanNeal>
http://www.iecss.com/whatwg.css

[10:46:02.0000] <TabAtkins>
Excellent, thank you.

[10:46:03.0000] <JonathanNeal>
I'm sure it could use some revisioning, but that's the one you guys reviewed most recently, based on our last conversations.

[11:16:00.0000] <JonathanNeal>
Whatcha up to, TabAtkins?

[11:16:01.0000] <TabAtkins>
Someone on the CSS lists asked for an HTML5 UA stylesheet, is all.

[11:21:00.0000] <JonathanNeal>
Ah, I was getting excited for new critisms and requests.

[11:21:01.0000] <TabAtkins>
Heh, nah.

[11:21:02.0000] <JonathanNeal>
*criticism

[11:45:00.0000] <TabAtkins>
Yay, I'm glad the PDF announcement made its way to the chromium blog.  I hate not knowing what things are confidential and what things aren't.

[11:46:00.0000] <gsnedders>
linky?

[11:46:01.0000] <TabAtkins>
http://blog.chromium.org/2010/06/bringing-improved-pdf-support-to-google.html

[11:51:00.0000] <JonathanNeal>
That's awesome, TabAtkins.

[11:52:00.0000] <gsnedders>
Does it actually use WebKit for rendering?

[11:52:01.0000] <TabAtkins>
I dunno.

[11:53:00.0000] <TabAtkins>
I somewhat doubt it, reasoning from some of the things I've heard about it.

[11:54:00.0000] <Dashiva^2>
Yarr

[11:54:01.0000] <TabAtkins>
Sock puppet!

[11:54:02.0000] <TabAtkins>
/me can't see joins/quits/renames, and so doesn't know if this is a new Dashiva or the old one with a new name.

[11:55:00.0000] <gsnedders>
It's a new Dashiva

[11:56:00.0000] <Dashiva^2>
New and improved?

[11:56:01.0000] <TabAtkins>
You tell us.

[11:57:00.0000] <Dashiva^2>
I just thought I

[11:58:00.0000] <Dashiva^2>
would flag my new colors in the interest of openness and transparency

[12:00:00.0000] <JonathanNeal>
Does something like audio[controls] exist yet that would actually work, in Chrome or some other?

[12:00:01.0000] <TabAtkins>
In terms of a selector, or a replacement, or what?

[12:00:02.0000] <TabAtkins>
Or an audio with controls?

[12:01:00.0000] <TabAtkins>
I know <audio controls> works in Firefox, and I suspect in webkit and opera.

[12:12:00.0000] <erlehmann>
JonathanNeal, <audio controls="controls"/> works.

[12:13:00.0000] <erlehmann>
HA HA, I AM USING THE XML SYNTHAX

[12:13:01.0000] <JonathanNeal>
I meant in terms of the selector.

[12:13:02.0000] <erlehmann>
why shouldnt it ?

[12:13:03.0000] <TabAtkins>
Synthrax is like anthrax, only worse.

[12:13:04.0000] <erlehmann>
afaik you can even make up selectors on the fly

[12:13:05.0000] <erlehmann>
attribute selectors, that is

[12:14:00.0000] <TabAtkins>
Indeed, you can.  You can select based on data-* attrs, fre.

[12:14:01.0000] <erlehmann>
<span lol="internet">rofl</span> is matched by span[lol]

[12:14:02.0000] <annevk>
Dashiva^2, new colors meaning Google?

[12:14:03.0000] <erlehmann>
TabAtkins, i am selecting based on RDFa

[12:14:04.0000] <TabAtkins>
That works too.

[12:14:05.0000] <erlehmann>
which is nice :3

[12:16:00.0000] <JonathanNeal>
Okay, then our previous discussion this remains true, cool.

[12:16:01.0000] <Dashiva^2>
annevk, yes

[12:16:02.0000] <JonathanNeal>
In the whatwg css file I had that.

[12:16:03.0000] <annevk>
Dashiva^2, congrats

[12:18:00.0000] <Dashiva^2>
So if anything suddenly breaks, it

[12:18:01.0000] <Dashiva^2>
's not my fault

[12:18:02.0000] <Dashiva^2>
Although I could blame it on this new and strange keyboard

[12:25:00.0000] <TabAtkins>
Re: WYSIWYG editors using <em> instead of <i> at all times - "That made me think of adopted foreign words, too, and how completely douchey your computer would sound reading out je nais c’est quoi or something with added emphasis, all like “Ooooh, look at me, I’m a computer and I speak French.”"

[12:28:00.0000] <JonathanNeal>
haha

[12:29:00.0000] <Dashiva^2>
I would be impressed

[12:29:01.0000] <JonathanNeal>
I was also using something like :section and :heading in that, not sure if that will ever be valid either.

[12:32:00.0000] <erlehmann>
http://github.com/wordpress/

[12:32:01.0000] <erlehmann>
this

[12:32:02.0000] <erlehmann>
I WILL FORK NOW

[12:32:03.0000] <erlehmann>
and change all <em>s to <i>s.

[12:32:04.0000] <Dashiva^2>
If you liked it then you shoulda made a fork of it

[12:32:05.0000] <erlehmann>
THAT WILL TELL THEM.

[12:33:00.0000] <erlehmann>
Dashiva^2, forking is the new poking.

[12:34:00.0000] <erlehmann>
hardcore forking action in progress.

[12:34:01.0000] <kpx>
HI I wanted to know that is there a javascript function that can read HTML as a binary safe string.... and dont say innerHTML that is not binary safe....

[12:35:00.0000] <Dashiva^2>
Are you sure binary safe is the phrase you mean?

[12:35:01.0000] <erlehmann>
trinary safe !

[12:36:00.0000] <JonathanNeal>
hey erlehmann, how about innerHTML!

[12:36:01.0000] <JonathanNeal>
:D

[12:36:02.0000] <erlehmann>
i actually used that function.

[12:36:03.0000] <erlehmann>
once.

[12:36:04.0000] <TabAtkins>
I use it all the time, implicitly.

[12:37:00.0000] <gsnedders>
kpx: How is anything not binary safe? JS has no concept of binary data, only of Unicode.

[12:39:00.0000] <kpx>
gsnedders: Well you are right, I meant "Unicode" safe actually.... 

[12:39:01.0000] <gsnedders>
kpx: In what way?

[12:39:02.0000] <gsnedders>
kpx: Any string returned by DOM is nessisarily a Unicode string?

[12:39:03.0000] <TabAtkins>
AryehGregor: Yeah, I could totally write that America infobox without a table.  However, a table is the correct thing to use for the middle half of it.

[12:40:00.0000] <kpx>
gsnedders: I was trying to save pages as in a data base as innerHTML and examining some of it... Turned out there was some data missing.... 

[12:40:01.0000] <erlehmann>
Receiving objects:   6% (4871/81182), 1.19 MiB | 4 KiB/s

[12:40:02.0000] <gsnedders>
How did data get from innerHTML to the database?

[12:41:00.0000] <gsnedders>
erlehmann: Hell yeah.

[12:41:01.0000] <erlehmann>
this will take a while

[12:41:02.0000] <kpx>
gsnedders: web databases

[12:42:00.0000] <kpx>
gsnedders: Is that normal to actually loose some data? technically if you see it on the page and you save it it should be there....

[12:42:01.0000] <kpx>
I calculated a SHA1 of the data also

[12:43:00.0000] <Dashiva^2>
What exactly was missing?

[12:45:00.0000] <kpx>
Well some tags were entirely gone...Some chars were sometimes gone.... I think SQLite backend should store the Unicode data safely...

[12:45:01.0000] <kpx>
Weird...

[12:45:02.0000] <TabAtkins>
That doesn't sound like an issue caused by unicode-related failures.

[12:45:03.0000] <kpx>
TabAtkins: Well i am not processing the data at all

[12:46:00.0000] <kpx>
as in document.getElementsByTagName('HTML')[0].innerHTML and then insert that into a webdatabase BLOB column in a table...

[12:47:00.0000] <kpx>
Any ways if someone had such an issue and all then would appreciate if you let me know....

[12:47:01.0000] <TabAtkins>
Shrug.  I dunno what the problem is, I'm just saying what I think the problem *isn't*.

[12:47:02.0000] <kpx>
:)

[12:47:03.0000] <TabAtkins>
Try pushing the data into something else and seeing if the problem persists.

[12:48:00.0000] <gsnedders>
Try using TEXT and not BLOB?

[13:12:00.0000] <TabAtkins>
When defining classes in Python, do you really have to define a class before it can be used, even if it's defined later in the file?

[13:19:00.0000] <AryehGregor>
TabAtkins, could you de-tablefy something like this?  http://en.wikipedia.org/wiki/Template:Protected_Areas_of_Colorado

[13:19:01.0000] <AryehGregor>
Also, yes, of course you do.  In what language is that *not* the case?

[13:19:02.0000] <AryehGregor>
Same with functions.

[13:23:00.0000] <TabAtkins>
AryehGregor: You dont' need to in Lisp.  The object system is designed such that it will throw warning for unimplemented things, but then slot in the proper class references when you define them later.

[13:24:00.0000] <TabAtkins>
AryehGregor: That looks like it should be a table too.

[13:24:01.0000] <AryehGregor>
I don't get how that would work in a typical object-oriented language.  You might have a constructor, for instance, so how could you process $foo = new MyClass(); without knowing the definition for MyClass()?

[13:24:02.0000] <AryehGregor>
Lisp is partly functional, and functional languages are weird like this.

[13:24:03.0000] <TabAtkins>
If you try to *use* the class before it is defined, sure, it's a problem.

[13:24:04.0000] <AryehGregor>
What are you trying to do?

[13:25:00.0000] <TabAtkins>
I have a big batch of classes in a file.  I want to organize them in a particular logical order such that the more important classes are near the top, and the little helper classes are at the bottom.  But I can't do that, because the little classes are used by the important classes, and thus must come before them.

[13:26:00.0000] <AryehGregor>
Maybe it deliberately requires that earlier classes not use later classes, to avoid infinite loops?

[13:27:00.0000] <TabAtkins>
That's just as silly.  It's perfectly fine to have classes hold references to each other (other than the GC problems such causes).

[13:27:01.0000] <AryehGregor>
Do other totally-non-functional languages permit this kind of thing?

[13:28:00.0000] <TabAtkins>
I have no idea.  Lisp's object system *is* generally more advanced than most other languages'.

[13:28:01.0000] <AryehGregor>
/me discovers that he's supposed to spend like two hours going down to Courant and back to fill out some form . . . yay.

[13:28:02.0000] <AryehGregor>
Got to go, I guess.

[13:28:03.0000] <TabAtkins>
kk

[13:31:00.0000] <TabAtkins>
Anyway, Lisp's object system is totally normal and not functional-polluted.  Lisp is a multi-paradigm language, and is perfectly happy integrating things like OO into it without having to do backflips to fit it into a functional mold.

[13:31:01.0000] <TabAtkins>
Everyone else's object system is just weak and annoying, is all.

[13:31:02.0000] <TabAtkins>
/me grumps.

[13:54:00.0000] <gsnedders>
You mean you don't love JS? :o

[13:54:01.0000] <gsnedders>
:P

[14:06:00.0000] <JonathanNeal>
Someone tweeted about my http://www.iecss.com/whatwg.css so I'd like to append some notes to it, especially about imaginary selectors that do not yet exist.  Anyone wanna help, chime in?

[14:06:01.0000] <JonathanNeal>
http://twitter.com/PageRankSEO/status/16710927231 is that you TabAtkins?

[14:07:00.0000] <jgraham>
TabAtkins: Javascript hoists the function definitions to the top of the scope so you don't have to define things in order there

[14:07:01.0000] <jgraham>
But in python you can do all kinds of stuff like

[14:07:02.0000] <jgraham>
class Foo(Bar):

[14:08:00.0000] <jgraham>
  if baz:

[14:08:01.0000] <jgraham>
    def __init__(self):

[14:08:02.0000] <jgraham>
(and so on)

[14:08:03.0000] <jgraham>
  else:

[14:08:04.0000] <jgraham>
  def __init__(self):

[14:08:05.0000] <jgraham>
      (some other stuff)

[14:10:00.0000] <jgraham>
Not claiming that is really useful, but I have changed the metaclass based on whether logging is enabled by putting an if statement in the class body

[14:10:01.0000] <jgraham>
Dunno if this really interferes with the ability to lazily resolve symbols though

[14:11:00.0000] <broquaint>
There's also Joose, inspired by Moose, which was inspired by "The Art of the Meta Object Protocol" which in turn describes Lisp's CLOS :)

[14:18:00.0000] <TabAtkins>
JonathanNeal: No, not me.  Dunno where that came from.  ^_^

[14:19:00.0000] <TabAtkins>
jgraham: In Lisp all this talk of "methods" is silly.  You just have functions specialized on certain arguments.  You can certainly define different versions of a function in an if.

[14:20:00.0000] <TabAtkins>
(That said, Lisp certainly has "methods".  It reuses that name for functions specialized on their arguments.  ^_^)

[14:23:00.0000] <erlehmann>
JonathanNeal, what is "img:complete" ?

[14:23:01.0000] <erlehmann>
also "video:metadata" ?

[14:26:00.0000] <JonathanNeal>
erlehmann, well complete is a state of img

[14:26:01.0000] <erlehmann>
loaded ?

[14:26:02.0000] <JonathanNeal>
Yes, but someone told me it would be complete.

[14:28:00.0000] <TabAtkins>
"would be" = "has been proposed to be at least once on the CSS list".

[14:32:00.0000] <Hixie>
AryehGregor: you don't have to define classes in perl before you use them

[14:32:01.0000] <Hixie>
AryehGregor: it's all late-dispatch

[14:33:00.0000] <TabAtkins>
So Python's the stupid stepchild of dynamic languages here.

[14:33:01.0000] <Hixie>
AryehGregor: you can even make a method that takes as an argument the name of the method that was called and have it dynamically make up the method as it goes

[14:33:02.0000] <Hixie>
and even in typed languages you don't have to give the full definition -- e.g. in objectpascal so long as you declare that the identifier will be defined as a class later, you can use it as a class before it's defined

[14:34:00.0000] <Hixie>
(the compiler does two passes, one to get the types and then one to get the code, or something)

[14:37:00.0000] <LoneStar99>
what is the simplest way to draw a line from one mouse point to anohter?

[14:38:00.0000] <TabAtkins>
Use <canvas>?

[14:38:01.0000] <LoneStar99>
oh yeah meant using canvas...

[14:39:00.0000] <JonathanNeal>
TabAtkins, thank you for the correction.

[14:39:01.0000] <JonathanNeal>
These are the types of things I want to add in a notes section.

[14:39:02.0000] <JonathanNeal>
Now that people are distributing it.

[14:41:00.0000] <LoneStar99>
TabAtkins: have the following

[14:42:00.0000] <LoneStar99>
TabAtkins: the code seems to work, but on mouse up, and try to draw another line, the previous one is deleted

[14:43:00.0000] <TabAtkins>
Yes, every time you draw to <canvas>, it's completely fresh.  If you want to keep the old lines around, you have to remember them and redrew them in the next draw step.

[14:43:01.0000] <jgraham>
TabAtkins: You can bind any function to a class in python. Methods are just bound functions

[14:43:02.0000] <TabAtkins>
Or, um...

[14:43:03.0000] <TabAtkins>
That's wrong.  But anyway!  I suspect that's your issue.

[14:44:00.0000] <LoneStar99>
TabAtkins: here is the complete code

[14:44:01.0000] <LoneStar99>
http://pastebin.com/jqxZyqMV

[14:44:02.0000] <TabAtkins>
jgraham: Hmm, I wonder if the problem is with Django, actually.

[14:44:03.0000] <jgraham>
TabAtkins: What problem?

[14:44:04.0000] <jgraham>
In general it is a requirement that you declare things before you use them

[14:45:00.0000] <jgraham>
in python

[14:45:01.0000] <TabAtkins>
Whereby it reports an error because one of my models uses another model declared later in the file.

[14:45:02.0000] <LoneStar99>
TabAtkins: what is the best way to remember all the lines drawn?

[14:45:03.0000] <TabAtkins>
And so I have to order the classes in the file in a suboptimal way.

[14:45:04.0000] <jgraham>
TabAtkins: Possibly, yes

[14:46:00.0000] <jgraham>
TabAtkins: You could split it into several smaller files, maybe

[14:46:01.0000] <TabAtkins>
LoneStar99: The problem is that clearRect() call you have.  It... clears the canvas every time you move the mouse.

[14:46:02.0000] <TabAtkins>
jgraham: Could, but that's even more trouble.

[14:49:00.0000] <LoneStar99>
TabAtkins: yeah, with out "clearRect()"  the line draws like the following: http://admin.work.lonestarbeer.operaunite.com/file_sharing/content/funky.png

[14:49:01.0000] <TabAtkins>
I can't connect to your computer there.

[14:59:00.0000] <LoneStar99>
TabAtkins:  is it still not working the link above?

[15:00:00.0000] <erlehmann>
i should be considered a hero http://github.com/erlehmann/wordpress/commit/f1d3dec90eb2ec18e973aa1e374ce78944a2fd68

[15:00:01.0000] <TabAtkins>
LoneStar99: Yeah.  It keeps saying your computer is offline.

[15:01:00.0000] <TabAtkins>
LoneStar99: Are you drawing it in the mouseMove so that the user can see the line before clicking the second time?

[15:02:00.0000] <LoneStar99>
TabAtkins: ok try this link http://freagair.com/RON/funky.png

[15:02:01.0000] <TabAtkins>
Yeah, that's what I thought.

[15:03:00.0000] <TabAtkins>
Everything is acting normally, then.  You're just doing it wrong.  ^_^

[15:03:01.0000] <TabAtkins>
So, you're drawing it in the mouseMove so that people can see the line before releasing their mouse and committing to a second point.

[15:03:02.0000] <TabAtkins>
This means that, yes, you do indeed need to clear the canvas between draws and redraw the line, since you don't want the previous "preview" lines to persist.

[15:04:00.0000] <TabAtkins>
But that means that, obviously, you're clearing the canvas, and thus have to redraw *everything* you want to persist.

[15:04:01.0000] <TabAtkins>
So save the points of previous lines in an array or something and redraw them at every drawing step.

[15:05:00.0000] <TabAtkins>
This is the downside of immediate-mode graphics like <canvas>.  They have *no memory* of what's inside of them.  They're just a bag of dumb pixels.

[15:06:00.0000] <LoneStar99>
TabAtkins:  ok, makes sense....

[15:06:01.0000] <TabAtkins>
If you want to manipulate lines by themselves without worrying about previous lines you've drawn, you want a retained-mode graphics.  That is, SVG.

[15:06:02.0000] <erlehmann>
they kicked me out of #wordpress for this.

[15:06:03.0000] <erlehmann>
wtf

[15:07:00.0000] <TabAtkins>
In SVG, you'd just create a <line> or whatever element on mousedown and adjust its attributes on mouseMove.  Then when they mousedown again, you're creating a new element, which doesn't interact with the old one.

[15:07:01.0000] <TabAtkins>
erlehmann: Hahaha.

[15:07:02.0000] <aho>
you can of course also render your stuff onto another off-screen canvas, update that one, then draw it onto the whole thing onto the visible one

[15:07:03.0000] <aho>
(dirty rectangles etc)

[15:07:04.0000] <erlehmann>
TabAtkins, thats not funny.

[15:07:05.0000] <erlehmann>
;_;

[15:07:06.0000] <LoneStar99>
TabAtkins:  so the only way is to save previous points, and redraw them right?

[15:08:00.0000] <erlehmann>
i'll go join the habari cabal right after gsoc. hello gsnedders.

[15:08:01.0000] <TabAtkins>
True that, aho.  To clarify, you can use the visible canvas as "scratch", copying the hidden canvas at every redraw (rather than just clearing the canvas).  Then, when the user commits to a particular line, draw that final line to the hidden canvas.

[15:09:00.0000] <TabAtkins>
That way the hidden canvas is only manipulated in an additive way, and never needs to be cleared, while the visible scratch canvas can be manipulated as much as you want.

[15:09:01.0000] <aho>
http://kaioa.com/k/ct/image_combine/index.html <- that renderToCanvas function creates such an off-screen canvas for example

[15:09:02.0000] <erlehmann>
wait i do not even know, where to apply.

[15:10:00.0000] <erlehmann>
then i'll probably have to hack my own blog system, using tinyldap, C and libowfat.

[15:10:01.0000] <TabAtkins>
erlehmann: I recommend the latter.  Only build it on top of Markdown, like I'll do someday when I'm not lazy.

[15:10:02.0000] <TabAtkins>
<3 Markdown.

[15:11:00.0000] <LoneStar99>
TabAtkins and aho -> thanks

[15:11:01.0000] <erlehmann>
TabAtkins, i'd probably use python anyway.

[15:13:00.0000] <LoneStar99>
TabAtkins:  this is the orginal code sample http://devfiles.myopera.com/articles/649/example5.html

[15:14:00.0000] <TabAtkins>
That seems to work fine now.

[15:14:01.0000] <LoneStar99>
it works, fine on the browser, but not in "mobile app"

[15:16:00.0000] <paul_irish>
You think a data uri cache.manifest file would work?

[15:17:00.0000] <TabAtkins>
paul_irish: Do cache manifests have any sort of origin restrictions?

[15:17:01.0000] <TabAtkins>
(Remember that data uris are automatically unique-origin.)

[15:17:02.0000] <paul_irish>
yup

[15:17:03.0000] <TabAtkins>
Then, presumably, no, it wouldn't.

[15:18:00.0000] <paul_irish>
Hmmm true.. though data uri fonts are allowed in FF where cross origin fonts are not.

[15:18:01.0000] <TabAtkins>
/me thinks we need more exceptions for places where data uris are same-origin as the embedding page.

[15:18:02.0000] <TabAtkins>
paul_irish: Probably a bug.  ^_^

[15:18:03.0000] <paul_irish>
Def.

[15:32:00.0000] <LoneStar99>
TabAtkins: how would you recommend making the dynamic array to keep previous "x, y" values?

[15:33:00.0000] <TabAtkins>
LoneStar99: By... making an array?  Something on a long-lived object storing [x,y] pairs?  Just push the final [x,y] onto the array on every mouseUp.

[15:34:00.0000] <LoneStar99>
TabAtkins: ok thanks

[15:45:00.0000] <AryehGregor>
TabAtkins, Hixie: looks like PHP also doesn't mind "$x = new A(); class A {...}".

[15:45:01.0000] <AryehGregor>
It runs the constructor and everything.

[15:45:02.0000] <AryehGregor>
Oh well, Python is picky.

[15:46:00.0000] <TabAtkins>
Excellent.

[15:46:01.0000] <AryehGregor>
It seems PHP just reads the whole file before it actually executes anything.

[15:48:00.0000] <AryehGregor>
This delays three seconds, prints "1\n2\n", then delays another three seconds: (echo '<?php echo "1\n";'; sleep 3; echo 'echo "2\n"; sleep(3);') | php

[15:48:01.0000] <AryehGregor>
(second delay to ensure that it's not just buffering output)

[15:51:00.0000] <Philip`>
/me can't think of any serious languages that don't read the whole file before actually executing anything

[15:51:01.0000] <AryehGregor>
I assume "serious languages" is intended to exclude shells.  :)

[15:52:00.0000] <Philip`>
Yes, and BASICs

[15:53:00.0000] <AryehGregor>
I see the same behavior in Python.  Don't really know any other interpreted languages to test in.

[15:56:00.0000] <AryehGregor>
Why do they read the whole file first?

[15:56:01.0000] <TabAtkins>
For PHP, one reason is so they can run any require()s immediately.

[15:57:00.0000] <AryehGregor>
I don't see why that would be necessary, if you actually executed statements in order.

[15:57:01.0000] <AryehGregor>
Clearly PHP executes statements out of order, so of course it has to read the whole file first.

[15:57:02.0000] <TabAtkins>
require()s fire even if they're in a non-executed path.

[15:57:03.0000] <TabAtkins>
if(false) { require "foo" }, frex, still requires and runs "foo".

[15:57:04.0000] <AryehGregor>
But the code in the file you fetch doesn't actually execute, I assume.

[15:58:00.0000] <Philip`>
(I think the issue with Python is that class definitions are just code that's executed at runtime and bind to the class name, so if you have code that executes earlier then it hasn't been bound yet)

[15:58:01.0000] <TabAtkins>
It... might?

[15:59:00.0000] <TabAtkins>
Philip`: Correct, Python's just not being smart enough about the binding.  You don't actually *need* a class there until you actuall interact with the object, you just need a handle.  Good object systems let you attach classes to that handle later.

[15:59:01.0000] <Philip`>
(It's fine to refer to classes that aren't syntactically earlier in the file, they just have to be dynamically earlier in runtime else you get a NameError when trying to use them)

[15:59:02.0000] <AryehGregor>
Python has a weird object system.

[15:59:03.0000] <Philip`>
TabAtkins: How could it be smarter? Before it's executed the class definition, there is nothing it could reasonably bind to

[16:00:00.0000] <TabAtkins>
Philip`: You tell me.  Tons of other dynamic languages get it right.

[16:01:00.0000] <Philip`>
Python lets you do 'print 1; class C: print 2; if thing_with_side_effects(): def method(): etc'

[16:01:01.0000] <Philip`>
and the code inside the class C definition gets executed after the first print statement

[16:01:02.0000] <TabAtkins>
Yes.

[16:01:03.0000] <AryehGregor>
You can define classes conditionally in PHP, and it seems to work out.

[16:01:04.0000] <TabAtkins>
And in Lisp, and in Perl, and in...

[16:02:00.0000] <Philip`>
Other languages either work in effectively the same way (it's a runtime error if you use a class name that hasn't been bound), or are much more static (the class definitions can be extracted without executing the code)

[16:02:01.0000] <AryehGregor>
TabAtkins, by the way, <http://en.wikipedia.org/wiki/Template:Protected_Areas_of_Colorado> is semantically a table, but not semantically a table in a table in a table.  :)

[16:03:00.0000] <AryehGregor>
Philip`, I'm guessing that at least in PHP, it's not a runtime error if you use a class that hasn't been bound, as long as the class isn't defined conditionally.

[16:03:01.0000] <AryehGregor>
/me tests

[16:04:00.0000] <TabAtkins>
AryehGregor: Oh god, I didn't actually look at the markup.

[16:04:01.0000] <TabAtkins>
a;sldhfa;kljg;a

[16:04:02.0000] <TabAtkins>
/me dies from the bad markup.

[16:05:00.0000] <TabAtkins>
For real?  Entire *rows* sacrificed on the black alter of *creating cell spacing*?  What. The. Fuck.

[16:05:01.0000] <AryehGregor>
$ echo '<?php $x = new C(); class C { function __construct() { echo "success\n"; } }' | php

[16:05:02.0000] <AryehGregor>
success

[16:05:03.0000] <AryehGregor>
$ echo '<?php $x = new C(); if (true) { class C { function __construct() { echo "success\n"; } } }' | php

[16:05:04.0000] <AryehGregor>
PHP Fatal error:  Class 'C' not found in - on line 1

[16:05:05.0000] <AryehGregor>
Philip`, ^^

[16:06:00.0000] <AryehGregor>
TabAtkins, do you think it could be rewritten to not be insane and still look basically the same in all browsers?  There's a discussion right now on wikitech-l about this.

[16:06:01.0000] <TabAtkins>
Cool, so PHP does some analysis on the source to find classes that it knows it'll be able to define.

[16:06:02.0000] <TabAtkins>
AryehGregor: Yes, I'm quite certain of this.

[16:06:03.0000] <AryehGregor>
http://lists.wikimedia.org/pipermail/wikitech-l/2010-June/048147.html

[16:06:04.0000] <TabAtkins>
That's an absolutely basic use of colspan and <th>.

[16:07:00.0000] <TabAtkins>
AryehGregor: You want me to produce something for you?

[16:07:01.0000] <Philip`>
TabAtkins: By "Cool" I assume you mean "That's totally hideous"

[16:10:00.0000] <AryehGregor>
TabAtkins, I dunno if there's any point.  Some enwiki sysop would have to get interested if it's going to get adopted.

[16:10:01.0000] <AryehGregor>
Hakon already proposed a div-based infobox thing, but basically got ignored.

[16:10:02.0000] <AryehGregor>
(although his didn't degrade that well in IE6/7)

[16:10:03.0000] <TabAtkins>
AryehGregor: Well, just let me know.  It's just a few minutes to put together a non-disgusting version of that box.

[16:11:00.0000] <Philip`>
(...though maybe it's no worse than e.g. Java where class definitions can define any one of about six things depending on whether it's in the top-level scope or in a class or in an expression and whether it's static or not)

[16:12:00.0000] <AryehGregor>
TabAtkins, hmm.  Looks like it's actually generated by a horrifyingly complicated template.

[16:14:00.0000] <TabAtkins>
Well, it's done as a navbox in a navbox.

[16:17:00.0000] <TabAtkins>
At the very least, you could change the navbox template to use ordinary tables and @cellspacing or whatever, rather than empty <tr>.

[16:18:00.0000] <TabAtkins>
Either changing to a new template that allowed nested groups natively, or having the template generator look for nested navboxes and interpret it correctly, would let you do even better.

[16:19:00.0000] <TabAtkins>
As a bonus, all of the greens would line up, rather than the second column being as wide as each group like it is now, since each group is a separate table.

[16:23:00.0000] <TabAtkins>
Actually, I don't understand who the navbox ended up being a table-in-a-table.  The outer table only has a single cell, and it appears its only purpose is to put a border on the navbox.

[16:23:01.0000] <TabAtkins>
s/who/how/

[16:23:02.0000] <AryehGregor>
I think it's because it uses this: http://en.wikipedia.org/w/index.php?title=Template:Navbox&action=edit

[16:23:03.0000] <AryehGregor>
And probably anyone who actually understands how that works is so thoroughly insane that they can't help with it much.

[16:24:00.0000] <TabAtkins>
Oh god.

[16:24:01.0000] <TabAtkins>
Protip: wikitext is a horrible, horrible language.

[16:25:00.0000] <AryehGregor>
I wish I didn't have to know that.

[16:26:00.0000] <TabAtkins>
I'm going to go with "someone who didn't know what they were doing hacked it together first, and wikitext is writeonly enough that noone ever improved it".

[16:27:00.0000] <AryehGregor>
Plus there's the small fact that it's used on millions of pages, and if you edit templates like that very often, sysadmins tend to get very annoyed at you.

[16:27:01.0000] <TabAtkins>
Pfft.

[16:27:02.0000] <AryehGregor>
(and like, maybe, delete the template)

[16:27:03.0000] <AryehGregor>
(that happened once on the Italian Wikipedia, IIRC)

[16:27:04.0000] <AryehGregor>
Plus of course, since it's used on so many pages, it's protected so only sysops can edit it.

[16:33:00.0000] <TabAtkins>
Was wikipedia invented before the days of JSON?  At least that sort of kv pairing is easier to read.

[16:34:00.0000] <TabAtkins>
/me hates when data formats invent new ways to pair keys and values.

[16:37:00.0000] <aho>
json wasn't invented, it was discovered :>

[16:38:00.0000] <TabAtkins>
I have no conclusive evidence that JSON wasn't transcribed from an obelisk found on one of Saturn's moons, but that's irrelevant to whether or not wikipedia was created before or after the sacred event.

[16:40:00.0000] <AryehGregor>
Wikipedia was launched on January 15, 2001.

[16:40:01.0000] <aho>
JSON was used at State Software in 2001. The JSON.org website was launched in 2002.

[16:40:02.0000] <aho>
<:

[16:40:03.0000] <AryehGregor>
Although that used preexisting software.

[16:40:04.0000] <AryehGregor>
MediaWiki was a couple years later.

[16:41:00.0000] <aho>
(json also didn't became popular right away)

[16:42:00.0000] <AryehGregor>
The real problem is that the language is meant to be a simple declarative thing, not an actual programming language.

[16:43:00.0000] <AryehGregor>
We could make it an actual programming language, but then performance would be even more horrifyingly terrible, since Wikipedia editors have no idea how to code.

[16:43:01.0000] <AryehGregor>
Also, they'd probably write programs that are just as baroque and unmaintainable.

[16:44:00.0000] <TabAtkins>
Yes.  Filling in a template using wikitext is only slightly horrifying.  Actually *implementing* a template with wikitext is getting to r'lyeh levels.

[16:47:00.0000] <AryehGregor>
Well, one of the proposals that gets floated is to use Lua or something.  But then we'd lose compatibility with typical shared hosts, which we don't want to do.

[16:50:00.0000] <TabAtkins>
Lua seems an odd choice.  Python is much more widely supported, if you don't want to move to the ubiquitous choice of PHP.

[16:53:00.0000] <AryehGregor>
It would need to support robust sandboxing, since this would be untrusted code.  Apparently Lua does that natively.

[16:54:00.0000] <TabAtkins>
Ah, true.

[16:54:01.0000] <TabAtkins>
Yeah, I imagine that a language *designed* to be an embedded scripting language running untrusted code would do so.


2010-06-22
[17:16:00.0000] <MikeSmith>
AryehGregor: is there a way in MediaWiki to comment out some wikitext in the source for a page?

[17:17:00.0000] <MikeSmith>
nm

[17:17:01.0000] <MikeSmith>
found it

[17:32:00.0000] <Philip`>
AryehGregor: Surely JavaScript would be a better idea than Lua, given that everybody on the web knows how to write JS and approximately nobody knows Lua

[17:33:00.0000] <Philip`>
(and since JS implementations are designed (and extensively tested) to run untrusted code in a sandboxed environment)

[17:56:00.0000] <AryehGregor>
Philip`, yeah, that was another thought.  Probably makes the most sense.

[17:57:00.0000] <AryehGregor>
But the shared hosting thing is a blocker, unless someone has a JavaScript implementation in PHP that's not even more ridiculously inefficient than PHP itself.  ;)

[19:16:00.0000] <drclue>
Here is something I would like to suggest that would make my life and probably that of others easier.  That would be two attributes added to at least the DIV and SPAN elements.  Those two attributes would be XML and XSL. It certainly would make doing XML/XSL trasforms more straight forward.

[19:16:01.0000] <AryehGregor>
What?

[19:16:02.0000] <AryehGregor>
What are you trying to do and how would this help?

[19:18:00.0000] <drclue>
Well, I personally have always liked using XSL templates against XML data and being able to simply assign the URLs to those attributes would greatly simplify using templates

[19:18:01.0000] <AryehGregor>
So what are you asking for to help you with that?

[19:20:00.0000] <drclue>
Basically I would like to see the attributes XML and XSL added to the spec , and then the browser vendors could tuck their proprietary methods of doing transforms behind those attributes , and authors could simply use the setAttribute methods to perform the manipulations

[19:21:00.0000] <TabAtkins>
So, again, what are you actually trying to do?

[19:22:00.0000] <TabAtkins>
I'm hearing something about taking XML from somewhere, and specifying an XSL transform to apply to it.  Presumably the result will then render somewhere.

[19:24:00.0000] <drclue>
I'm not sure I understand the question, but what I'm trying to do is drum up some support for the thought of making XML/XSL more accessible in a standards based way , so that when driving dynamic displays , I can simply reference the url of the XML source and the XSL view and not always have to carry a bunch of browser object sniffing code to perform dynamic transforms.

[19:24:01.0000] <AryehGregor>
What do you mean "the attributes XML and XSL"?

[19:26:00.0000] <drclue>
<div XML="http://...." XSL="http://" />   

[19:27:00.0000] <AryehGregor>
Um, what are those supposed to do?

[19:30:00.0000] <drclue>
The XSL attribute specifies an external XSL stylesheet to apply to the XML data referenced by the XML attribute. In the absence of the XSL attribute , the browser would first look to the XML file itself for a style reference and then fall through to the native rendering of the XML tree.

[19:32:00.0000] <AryehGregor>
You can already do client-side XSL.

[19:34:00.0000] <drclue>
Yes I can already do client-side XML/XSL, but it suffers from vendor specific hacks that could be neatly tucked away by giving some attributes for that purpose

[19:42:00.0000] <drclue>
I really don't see any downside in affording a more effective way to integrate XML/XSL w3.org standards in a page. The bandwidth savings afforded by XSL are hard to deny. The promotion of MVC techniques associated with XML/XSL is a positive thing too. Affording browser vendors a cheap way to standardize is a big bonus too.

[19:43:00.0000] <variable>
if I want to propose a new CSS attribute on text "gender: male; or gender: female;" for spoken media - where is the appropriate place to do so?

[19:43:01.0000] <AryehGregor>
variable, www-style.

[19:43:02.0000] <AryehGregor>
drclue, if vendor-specific hacks are required, then that should be fixed by itself.  You don't need a totally new way of including XSL to do that.

[19:44:00.0000] <AryehGregor>
And there is a cost for implementing every feature.  "There's no downside" isn't sufficient reason to get anything added to the spec.  It's long enough already.

[19:45:00.0000] <variable>
#2: I am capable of voting on the issues page but I don't see any of my emails there. Am I missing something ?

[19:45:01.0000] <variable>
I thought only people who have issues on the list can vote :|

[19:46:00.0000] <AryehGregor>
What?

[19:46:01.0000] <variable>
"Only people who have sent feedback that has ended up on this list are able to vote." --> I can vote but I don't see any of my emails 

[19:48:00.0000] <drclue>
@AryehGregor  Just by itself , those XML XSL attributes would have significant value. The fact that providing  a standards based way of doing something also ditches the need to use proprietary methods is added gravey

[19:48:01.0000] <AryehGregor>
You can include XSL without proprietary methods.

[19:48:02.0000] <AryehGregor>
<link rel=stylesheet href=mystylesheet.xsl>

[19:48:03.0000] <AryehGregor>
If browsers don't all support that, that won't be fixed by making up new attributes that mean a similar thing.

[19:51:00.0000] <drclue>
That would be a page wide item.  Elements in pages appear dynamically and get included in a wide variety of ways. Being able to specify the XML data and the XSL structure at the element level allows for better compartmentalization and a more effect way to draw in units of information into a page

[19:52:00.0000] <AryehGregor>
Okay, well, send it to the list if you like.  I suggest you focus on use-cases rather than solutions.

[19:55:00.0000] <nessy>
Hixie:  check out this ass subtitled video http://www.animereactor.dk/kara/files/flcl-ed.mkv - awesome crack!

[20:01:00.0000] <drclue>
@AryehGregor   Use cases include tabular data browsing , charts ,info graphics , news feeds, multi-view data, calenders  and miles and miles of common every day activities. If one sees integration of data and presentation in an element of a page, those XML/XSL attributes could handle the vast majority of it. 

[20:01:01.0000] <AryehGregor>
Okay, well, I can't change the spec.  You'll have to post to the list.

[20:37:00.0000] <drclue>
@AryehGregor    Well, I've taken your advice and posted the thought to the list , and we'll have to see if I can get any traction in there amongst the hurricane force winds of conversation.

[21:04:00.0000] <karlcow>
http://latimesblogs.latimes.com/technology/2010/06/apple-location-privacy-iphone-ipad.html

[22:48:00.0000] <MikeSmith>
kennyluck: at t45 now

[23:49:00.0000] <hsivonen>
nessy: aren't those subtitles burned into the video frame? they scale as bitmaps in Totem.

[23:50:00.0000] <nessy>
that file has it all burnt in, but I am told they were authored as ASS

[23:50:01.0000] <nessy>
it's just that not all players support all the functionalities, so it's burnt in

[23:51:00.0000] <nessy>
I'm pretty sure it's not your 90% use case for subtitles/captions, but it's totally awesome :)

[23:53:00.0000] <othermaciej>
I can't believe someone had the gumption to name a file format "ASS"

[23:53:01.0000] <hsivonen>
nessy: I see

[23:54:00.0000] <nessy>
othermaciej: ;)

[23:54:01.0000] <MikeSmith>
whoever picked that name deserves an award of special merit

[23:54:02.0000] <nessy>
didn't WHATWG almost turn into a task force? ;)

[23:55:00.0000] <nessy>
ASS really is ASSA for Advanced SubStation Alpha - but ASS is nicer

[23:55:01.0000] <hsivonen>
turns out I need to point people to #AVNormalize so often that I have it in my AwesomeBar 

[23:59:00.0000] <nessy>
oh! what is #AVNormalize ?

[00:00:00.0000] <hsivonen>
nessy: http://www.w3.org/TR/xml/#AVNormalize

[00:00:01.0000] <hsivonen>
(not related to the subtitle discussion at all)

[00:00:02.0000] <nessy>
ah, and here I was thinking of AV=audio/video!

[00:01:00.0000] <nessy>
wrong filters on my eyes ;)

[00:01:01.0000] <hsivonen>
:-)

[00:21:00.0000] <MikeSmith>
hsivonen: do you have any thoughts of whether any HTML elements should be allowed in <annotation-xml> ?

[00:21:01.0000] <MikeSmith>
or, more importantly, if the parsing algorithm should be changed to allow them there

[00:21:02.0000] <MikeSmith>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=9887

[00:22:00.0000] <MikeSmith>
Simon suggested that maybe they could be dealt with by requiring they be wrapped in a <div>

[00:23:00.0000] <hsivonen>
MikeSmith: I'm OK with making <div> (maybe <span>, too, just in case?) magic there, but what's the use case?

[00:23:01.0000] <hsivonen>
hmm. actually, <div> is less dangerous, since it is scoping on its own right

[00:24:00.0000] <hsivonen>
MikeSmith: even without a use case now, I'd be OK with making <div> magic there in case a use case shows up later

[00:24:01.0000] <MikeSmith>
hsivonen: the use case is for using HTML to annotate MathML expressions in the same way that SVG or MathML itself can be used for such... but that said, I don't really understand the general use case for annotating MathML expressions to begin with

[00:25:00.0000] <hsivonen>
MikeSmith: the best use case I've seen for annotations is using SVG to draw notation that MathML itself doesn't support

[00:25:01.0000] <hsivonen>
dunno how HTML would help there

[00:27:00.0000] <MikeSmith>
hsivonen: and even for that use case, couldn't SVG just be used within an <mtext> element, with no need to use the whole <semtantics>/<annotation-xml> apparatus?

[00:32:00.0000] <hsivonen>
MikeSmith: if the validity definition is adjusted, sure

[00:34:00.0000] <MikeSmith>
hsivonen: we have the freedom to adjust the validity definition

[00:34:01.0000] <MikeSmith>
to define it in the HTML5 spec

[00:34:02.0000] <hsivonen>
MikeSmith: works for me

[00:37:00.0000] <MikeSmith>
but getting back to annotation-xml, I guess I'm just wondering whether it is worth the trouble to try to convince implementors and Hixie that changing the parsing algorithm to allow <div> (and/or <span>) within annotation-xml as a holder for HTML content is worth the effort

[00:37:01.0000] <MikeSmith>
it sounds like you're amenable at least

[00:38:00.0000] <MikeSmith>
I can ask the MathML WG if they have use cases for HTML content in annotation-xml

[00:39:00.0000] <Hixie>
i think allowing just <div> would just end up confusing authors

[00:39:01.0000] <Hixie>
if we allow only one HTML element it should be <body> (or <html>)

[00:40:00.0000] <Hixie>
nessy: actually with WebSRT and the CSS extensions to address individual bits of the text cues, that's more or less possible

[00:40:01.0000] <hsivonen>
Hixie: Degrade Gracefully...

[00:41:00.0000] <Hixie>
hsivonen: both of those would degrade fine if no attributes are present

[00:41:01.0000] <hsivonen>
Hixie: <div> is the new <body>. see Atom

[00:41:02.0000] <Hixie>
Atom is not an example of good language design

[00:41:03.0000] <Hixie>
(it's not especially terrible, but it's not what i would suggest we base our decisions on)

[00:42:00.0000] <othermaciej>
<body> doesn't seem like a good element for other languages to use when embedding a snippet of HTML

[00:42:01.0000] <othermaciej>
for one thing, it means that when you nest that element back in HTML, you get nested <body>, which is pretty confusing

[00:42:02.0000] <othermaciej>
second, there shouldn't be an expectation that a snippet is necessarily like a complete document

[00:43:00.0000] <othermaciej>
(though really you'd want <html> if you are claiming to provide a complete embedded HTML document)

[00:43:01.0000] <drclue>
I already embed HTML in XML for use with XSL transforms

[00:45:00.0000] <Hixie>
othermaciej: without a solid use case, it's not clear whether it's for snippets or documents or something else

[00:45:01.0000] <othermaciej>
fair enough

[00:46:00.0000] <MikeSmith>
add to this that the MathML spec doesn't really define the rendering behavior of annotation-xml content

[00:47:00.0000] <MikeSmith>
it leaves it up to other specs to do that

[00:47:01.0000] <MikeSmith>
e.g., HTML5

[00:47:02.0000] <drclue>
I'd actually like to see an XML and XSL attribute for HTML elements myself. I use so many XSL transforms in day to day work , it would seem worth while to have an easy way to integrate them at the element level as opposed to having to muck around with vendor specific javascript

[00:49:00.0000] <Yudai>
MikeSmith: do you have any ticket for me to write patches?

[00:50:00.0000] <MikeSmith>
Yudai: have not talked with hsivonen about it yet

[00:50:01.0000] <MikeSmith>
hsivonen: Yudai is looking to help with v.nu bug fixes

[00:50:02.0000] <MikeSmith>
so if you have suggestions of any specific bugs he could work on, that'd be great

[00:50:03.0000] <MikeSmith>
/me goes to look through the bug list himself

[00:51:00.0000] <MikeSmith>
Yudai: there's at least one open bug I already Cc'ed you on

[00:52:00.0000] <hsivonen>
Yudai: oops. sorry, my email app had filtered your email to a folder where I didn't expect it.

[00:52:01.0000] <MikeSmith>
Yudai: http://bugzilla.validator.nu/show_bug.cgi?id=720 "Validating post-processed files but reporting source line numbers and file names"

[00:53:00.0000] <MikeSmith>
which Hixie filed

[00:53:01.0000] <Yudai>
MikeSmith: okay, i try to work on this

[00:53:02.0000] <MikeSmith>
Yudai: it's not a particularly easy one, I guess

[00:54:00.0000] <MikeSmith>
but certainly useful

[00:54:01.0000] <hsivonen>
I'll land 721

[00:54:02.0000] <MikeSmith>
Yudai: and you could bug Hixie for test cases and such

[00:54:03.0000] <MikeSmith>
hsivonen: ah, cool

[00:55:00.0000] <Yudai>
hsivonen: thx

[00:55:01.0000] <Yudai>
MikeSmith: i got it

[01:05:00.0000] <webben>
drclue: By browser-specific hacks, do you mean: http://www.w3schools.com/xsl/xsl_client.asp ?

[01:08:00.0000] <drclue>
@webben Maybe not that one in particular , but often there are variations of that. That kind of thing aside though , I think having an XML and XSL attribute would prove very useful of it's own right. I actually posted something to the mailing lists about it today

[01:08:01.0000] <webben>
yeah i saw

[01:09:00.0000] <webben>
like AryehGregor I don't see how the attribute removes the need for such hacks

[01:10:00.0000] <webben>
a consistent xsltprocessor and xslt implementation would remove the need (eventually) and be more generally useful.

[01:10:01.0000] <MikeSmith>
hsivonen: I think I can now check in the jing required-but-missing attributes/elements patch directly

[01:11:00.0000] <MikeSmith>
OK if I do that

[01:11:01.0000] <drclue>
Perhaps in some cases , folks might need to resort to that kind of thing, but for most of the use cases I come across , simply having the attributes would deal them

[01:11:02.0000] <MikeSmith>
?

[01:11:03.0000] <MikeSmith>
hsivonen: I believe I had made all changes from your review

[01:11:04.0000] <hsivonen>
MikeSmith: cool. yes, OK.

[01:11:05.0000] <MikeSmith>
OK, great

[01:11:06.0000] <webben>
drclue: But there's no difference between the two solutions from that perspective.

[01:12:00.0000] <webben>
drclue: Both solutions would take time to become commonly implemented. Until that time, the hackery would still be required.

[01:12:01.0000] <hsivonen>
Yudai: landed and deployed. Thank you.

[01:12:02.0000] <Yudai>
hsivonen: roger. thanks

[01:13:00.0000] <hsivonen>
And sorry about the delay.

[01:16:00.0000] <drclue>
@webben the attributes would be trivial to implement , and with the promise of HTML5 , this is the time to do it. having 

[01:19:00.0000] <drclue>
It will take time for browsers to support all of HTML5 anyways , so I'm not swayed at all by the idea that such may be the case for the XML / XSL attributes 

[01:19:01.0000] <MikeSmith>
Opera Mobile (non-Mini) on an Android device - http://www.pcpro.co.uk/blogs/2010/06/21/toshiba-ac100-mobile-internet-device-first-look-review/

[01:20:00.0000] <MikeSmith>
though it's not a handset or smartphone

[01:20:01.0000] <MikeSmith>
netbook, I guess

[01:20:02.0000] <drclue>
I do know that I would put them to work almost immediately for my closed audience applications , and build a little crutch script for those cases where I needed to provide legacy support. Basicly the same kind of thing I've been doing forever

[01:21:00.0000] <hsivonen>
I wonder how Android/Meego usability works when the platforms are stretched across both touch screen devices and netbooks that have a keyboard and a trackpad

[01:22:00.0000] <MikeSmith>
the touchscreen difference definitely seems like the biggest one

[01:22:01.0000] <othermaciej>
some Android devices do have an auxiliary hard keyboard

[01:22:02.0000] <MikeSmith>
I've not actually seen Android running on non-touchscreens

[01:22:03.0000] <othermaciej>
and a number have a trackball

[01:22:04.0000] <hsivonen>
also, I wonder if Opera Mobile for Android is OEM-only or if it'll be user-installable

[01:22:05.0000] <othermaciej>
to me, having a touchscreen *and* a pointing device on a phone seems weird

[01:22:06.0000] <MikeSmith>
hsivonen: I would bet that at this point it's OEM-only

[01:22:07.0000] <othermaciej>
is Opera Mobile for Android not in the Android Market?

[01:23:00.0000] <hsivonen>
I'm rather disappointed that Nokia put a hardware keyboard in N900 instead of making the on-screen keyboard rock

[01:23:01.0000] <slartsa>
I disagree

[01:23:02.0000] <hsivonen>
I think the onscreen keyboard on N900 is already better than the hardware keyboard but not as good as on iPhone

[01:23:03.0000] <MikeSmith>
othermaciej: no Opera Mobile in Android Market as far as I know

[01:24:00.0000] <slartsa>
most people still need the actual feeling of a keyboard at their fingertips

[01:24:01.0000] <slartsa>
..even though on-screen would be more efficient

[01:24:02.0000] <hsivonen>
slartsa: N900 is heavier than e.g. HTC Desire

[01:24:03.0000] <MikeSmith>
I do think having both a touchscreen and pointing device is odd

[01:24:04.0000] <hsivonen>
slartsa: being heavy sucks

[01:24:05.0000] <slartsa>
hey! I am heavy!

[01:24:06.0000] <slartsa>
well it's not that bad

[01:24:07.0000] <hsivonen>
slartsa: sorry. I meant it sucks for phones

[01:24:08.0000] <slartsa>
:D

[01:25:00.0000] <MikeSmith>
slartsa: I guess most people also prefer the feeling of viewing things on a 21-inch+ monitor

[01:25:01.0000] <slartsa>
I'm glad the phone had sold so well

[01:25:02.0000] <slartsa>
Not sure it would have if it didn't have that keyboard

[01:25:03.0000] <drclue>
The only problem I have with soft keyboards is finger registration. I often type pretty much in the dark, sorta like using my TV remote. It gets a little rough when there is nothing but looking  as an option

[01:30:00.0000] <hsivonen>
anyway, it seems that keypad without mouse, keyboard with mouse, single-touch and multi-touch all call for different interaction design, so it's not clear that it's a good thing to strech the interaction layer of a platform across those different device categories

[01:30:01.0000] <MikeSmith>
what's the difference between a MID and a netbook?

[01:30:02.0000] <hsivonen>
although Opera Mobile is doing really well across the first three (I haven't tried Opera Mobile on a multi-touch device)

[01:31:00.0000] <MikeSmith>
/me finds http://en.wikipedia.org/wiki/Mobile_internet_device

[01:33:00.0000] <drclue>
Keystrokes with evt.x evt.y , sorta sounds weird, but then again maybe that is not so weird.

[01:38:00.0000] <webben>
drclue: Without implementor experience, I'd be wary of declaring anything trivial to implement.

[01:39:00.0000] <webben>
drclue: But my argument is not time would be required for your solution, but simply that other solutions would remove the need for hacks without introducing new features to the platform.

[01:40:00.0000] <webben>
drclue: One could always build a crutch script that uses a data-* attribute instead, again without introducing a new feature.

[01:40:01.0000] <drclue>
@webben  I've been implementing things since before Microsoft and Apple, so I would not worry too much about my assumptions regarding implementation difficulties. All the code is there in the browsers already

[01:41:00.0000] <webben>
drclue: Implementing things in JS... or implementing things in browser code?

[01:41:01.0000] <drclue>
Interesting assumption that there was javascript way back then :)

[01:42:00.0000] <webben>
drclue: I don't meant random implementation experience, I mean actual browser implementation experience.

[01:44:00.0000] <drclue>
Well, obviously there are only so many people who have had direct exposure to implementing current day browser code, although I did write a browser way back when. There sorta has been little need for me to do so these days

[01:44:01.0000] <drclue>
I've also written web servers from scratch, implemented my own TCP/IP stacks and many other things over the years

[01:46:00.0000] <drclue>
The fact is that the code to do each aspect of what I'm suggesting is already built into all the browsers, and the few bits of code to leverage whats there would not seem to be such a big deal as to try and make a point of it

[01:47:00.0000] <webben>
I'm not arguing it is or is not relatively easy ; I'll leave that to somone who's actually going to try to code it.

[01:48:00.0000] <MikeSmith>
othermaciej: do you recall a decision about the shortname for the polyglot author guide?

[01:49:00.0000] <MikeSmith>
I find the FPWD request that went to the chairs list

[01:49:01.0000] <MikeSmith>
but that does not have a specific request for a shortname

[01:49:02.0000] <webben>
drclue: If you're pointing to an XML source and an XSLT source, what would be the practical advantage over pointing to an seamless iframe with a link in it?

[01:50:00.0000] <othermaciej>
MikeSmith: I do not

[01:50:01.0000] <MikeSmith>
ok

[01:50:02.0000] <othermaciej>
MikeSmith: is that within the discretion of the chairs/team?

[01:50:03.0000] <webben>
drclue: Might the iframe not be better if you really want to build page-independent widgets?

[01:50:04.0000] <othermaciej>
if so, I can ask Sam tomorrow (my suggestion would be "polyglot" as a shortname)

[01:51:00.0000] <MikeSmith>
yeah, it is chairs discretion with PLH signoff -- it just does not seem to have been discussed yet

[01:51:01.0000] <drclue>
@webben I might want to change the XML or XSL attribute independently.  

[01:51:02.0000] <MikeSmith>
the FPWD cites http://dev.w3.org/html5/html-xhtml-author-guide/html-xhtml-authoring-guide.html but not a shortname

[01:51:03.0000] <webben>
drclue: You might. But with an iframe you have the ability to style your widget in a sandbox.

[01:52:00.0000] <MikeSmith>
so the implication I guess is that the request is for the shortname "html-xhtml-author-guide", which isn't particularly short..

[01:52:01.0000] <hsivonen>
/TR/polyglot/ ?

[01:52:02.0000] <webben>
drclue: And if you did change the value of the XSL attribute, wouldn't the original DOM have been replaced already?

[01:53:00.0000] <drclue>
XML is the data, XSL is the template the rendering is the transform of the two. They are three data items

[01:54:00.0000] <drclue>
If I have the XML , but decide I want to perhaps look at a calender in day/week/month views , I only have to change the XSL attribute

[01:55:00.0000] <hsivonen>
drclue: I think getting implementors interested in adding XSLT-related features is unlikely.

[01:56:00.0000] <drclue>
The XSL processor is already in there. This is for the implementor just a rephrasing of the call of existing functions

[01:56:01.0000] <webben>
drclue: yeah, good point, you can refetch the DOM via the XML source.

[01:57:00.0000] <MikeSmith>
hsivonen: maybe /TR/html-polygot

[01:58:00.0000] <MikeSmith>
drclue: maybe you could demonstrate the usefulness of it by implementing support for it in a Javascript library first

[01:59:00.0000] <MikeSmith>
drclue: or lobbying with developers of some existing JS library to

[02:00:00.0000] <drclue>
There are javascript libraries for that kind of thing out there all over the place, but there really is no reason that one should need to resort to javascript to perform this task

[02:03:00.0000] <MikeSmith>
drclue: I guess by that logic it could be argued that there's also not need for browser implementors to resort to implementing support for it natively

[02:04:00.0000] <MikeSmith>
part of the point of doing it through a JS library first would be to demonstrate that it's something that people would actually use

[02:05:00.0000] <MikeSmith>
if you implemented it in library, the JS part of it would not need to be exposed to developers

[02:05:01.0000] <drclue>
By and large the majority of my server side tools emit data as XML. Many many sites provide data via XML. The majority of that data simply needs to be formatted , and really has no need to be manipulated with javascript. 

[02:06:00.0000] <MikeSmith>
drclue: so all that developers would need to do is include your JS library, and things would work just like you describe

[02:06:01.0000] <MikeSmith>
they would not have to do any actual JS coding themselves

[02:06:02.0000] <MikeSmith>
you would do it all for them

[02:06:03.0000] <MikeSmith>
in the library you're going to develop

[02:07:00.0000] <drclue>
Since there is no XML or XSL attribute in HTML at this time , nobody would be able to create valid markup unless they perhaps wanted to write their own DTD, or perhaps do multiple transforms or other gymnastics

[02:08:00.0000] <jgraham>
drclue: What hsivonen said. XSL is rare on the web so sinking more resources into it when vendors could be working on making javascript performance more awesome or video use less CPU or implementing new form input types, or whatever, seems like a bad idea

[02:08:01.0000] <jgraham>
drclue: You can do what you want using data-foo attributes and javascript

[02:10:00.0000] <drclue>
Perhaps the reason that XSL is so rare on the web is that deploying it in a valid manner is generally a pain

[02:10:01.0000] <drclue>
given the current context. 

[02:12:00.0000] <jgraham>
Perhaps. Or perhaps it is because people don't like writing xslt. It doesn't make sense for vendors to sink resources into that experiment

[02:12:01.0000] <jgraham>
especially given that you can do it yourself rather easilly

[02:15:00.0000] <drclue>
I know that XML/XSL saves tons of server resources and bandwidth, and also fits in well with the MVC school of design. Some people think CSS is really hard to code, and maybe for them it is. I tend to hand code just about everything so it does not typically bother me. Some folks think that even a simple HTML page is too complicated to type, but I guess why for them there are products like dreamweaver

[02:17:00.0000] <drclue>
BRB

[02:21:00.0000] <hsivonen>
CSS hard compared to XSLT????

[02:22:00.0000] <drclue>
There are many aspects of web development that are not everyones cup of tea, but I for one really appreciate the many w3.org standards and encourage interoperability among those standards. 

[02:23:00.0000] <annevk>
finally someone with a clue has entered the channel; unfortunately we don't do logic

[02:23:01.0000] <annevk>
;p

[02:29:00.0000] <drclue>
XSLT is not hard at all. It like CSS has it's interesting nooks and crannies. Actually in some ways , XSL and CSS are but flip sides of the same coin

[02:29:01.0000] <hsivonen>
Hixie: when you decided to make cache manifests triggered on root instead of <link> did you happen to write down all the reasons why <link> would have sucked?

[02:30:00.0000] <hsivonen>
Hixie: because resource packages will have the same problems

[02:30:01.0000] <hsivonen>
resource packages formerly known as "Limi packages"

[02:30:02.0000] <Hixie>
i needed something that happened before the root element's style="" attribute was seen

[02:31:00.0000] <annevk>
hsivonen, what is a Limi package?

[02:31:01.0000] <hsivonen>
Hixie: ouch. I didn't think of that

[02:31:02.0000] <annevk>
hsivonen, oh, the stuff that won't be needed if we get SPDY?

[02:31:03.0000] <hsivonen>
annevk: http://stanford.edu/~jlebar/moz/respkg/

[02:32:00.0000] <annevk>
hsivonen, seen the commentary from mnot?

[02:32:01.0000] <hsivonen>
annevk: nope

[02:32:02.0000] <hsivonen>
Hixie: style='' is going to be a problem. sigh. thanks

[02:33:00.0000] <annevk>
http://www.mnot.net/blog/2010/02/18/resource_packages

[02:33:01.0000] <Hixie>
hsivonen: note that i failed to handle <?xml-stylesheet?> and <?binding?> in XML

[02:33:02.0000] <annevk>
<?xbl?> mister

[02:34:00.0000] <hsivonen>
Hixie: well, at least those need to be racy

[02:34:01.0000] <Hixie>
annevk: might as well be <?unicorn?> right now for all the practical effect on the world it's had :-P

[02:34:02.0000] <hsivonen>
Hixie: but consider <html style='...'> (pause in network stream that flushes everything) <link>

[02:34:03.0000] <annevk>
jaja :)

[02:35:00.0000] <hsivonen>
vs. <html style='...'><link> (flush speculations here)

[02:36:00.0000] <Hixie>
hsivonen: the network isn't hugely relevant for manifest="", it's just a matter of making sure you didn't pick the wrong appcache if you're offline

[02:36:01.0000] <hsivonen>
I wonder if we could defer the processing of style attributes until head has been popped off the stack

[02:36:02.0000] <Hixie>
that would suck

[02:37:00.0000] <hsivonen>
Hixie: there was a bug about speculations arriving without manifest

[02:37:01.0000] <hsivonen>
Hixie: which I fixed by putting manifests on the speculative load side channel

[02:37:02.0000] <hsivonen>
since manifest processing doesn't really need the DOM node 

[02:37:03.0000] <Hixie>
what the manifest loading can be sync

[02:37:04.0000] <Hixie>
it's not network or file bound

[02:38:00.0000] <Hixie>
assuming you cache the known manifests in ram

[02:38:01.0000] <hsivonen>
Hixie: I don't understand what you are saying

[02:39:00.0000] <Hixie>
i'm saying i don't understand why you would need to put manifests on a load channel of any kind to solve any bugs

[02:39:01.0000] <hsivonen>
Hixie: what I'm saying is that stuff broke when we had speculative script loads start before the app cache stuff knew it's going to be seeing a manifest

[02:39:02.0000] <Peter`>
Is there a reason (or just a lack of proposal?) for not including DOM methods for a dynamic manifest, giving JavaScript the ability to add/remove files, opposed to a single static file?

[02:39:03.0000] <hsivonen>
I don't know the detail, either, I just know stuff broke

[02:40:00.0000] <Hixie>
hsivonen: it doesn't need to see the manifest, it just needs to know which cache to load files from, which is keyed from the manifest string -- no network or disk I/O needs to happen to do the processing necessary to prepare the browser for handling loads due to a manifest="" attribute

[02:52:00.0000] <hsivonen>
ok. this style stuff is going to suck majorly for interop across browsers with different speculative parsing implementations and document.open() impl. details 

[02:53:00.0000] <hsivonen>
any reason why I shouldn't suggest changing resource packages to use <html packages='url1 url2 url3'> instead of <link>?

[02:58:00.0000] <annevk>
MikeSmith, how likely is it that we publish?

[03:00:00.0000] <MikeSmith>
/me checks with his bookies

[03:01:00.0000] <annevk>
:)

[03:02:00.0000] <hsivonen>
hmm. abarth is testing evil cases. good.

[03:08:00.0000] <MikeSmith>
abarth is good at exposing evil

[03:09:00.0000] <Hixie>
hsivonen: what's the use case for this?

[03:10:00.0000] <MikeSmith>
http://twitter.com/s0enke "@w3c should consider rfc1867 for xhr uploads"

[03:10:01.0000] <hsivonen>
Hixie: avoiding multiple HTTP requests in order to speed up sites / web apps

[03:10:02.0000] <MikeSmith>
/me looks ups RFC 1867

[03:11:00.0000] <hsivonen>
Hixie: I already sent my suggestion

[03:11:01.0000] <annevk>
MikeSmith, that guy should take a look at XHR2

[03:11:02.0000] <MikeSmith>
annevk: I'll tell him

[03:11:03.0000] <MikeSmith>
annevk: or somebody will

[03:12:00.0000] <annevk>
I could, but I'm getting twitter-lazy

[03:12:01.0000] <annevk>
IRC starts to becoming the only convenient message-gateway

[03:13:00.0000] <annevk>
pretty much always up, fast, archived...

[03:13:01.0000] <hsivonen>
email is full of spam and my XMPP server stopped working

[03:14:00.0000] <annevk>
TabAtkins, you around?

[03:15:00.0000] <Hixie>
MikeSmith: is the microdata spec publishing WD or LC?

[03:27:00.0000] <MikeSmith>
Hixie: I don't recall that you ever got a response to your request to transition it to LC, did you?

[03:28:00.0000] <MikeSmith>
/me pokes othermaciej 

[03:28:01.0000] <Hixie>
i think sam said it was in his court or something

[03:28:02.0000] <othermaciej>
WD still for this round

[03:28:03.0000] <Hixie>
(i have great trouble understanding what he's saying half the time)

[03:28:04.0000] <othermaciej>
though we hope to publish an LCWD of it soon

[03:28:05.0000] <Hixie>
not really sure why it's had to wait, what, two months so far?

[03:28:06.0000] <MikeSmith>
I guess that spec would be a good was to tests the waters on LC process

[03:32:00.0000] <hsivonen>
speaking of chair responses, I haven't seen a response to my request for permission to submit "high-level prose" in the ASCII ref Change Proposal

[03:33:00.0000] <othermaciej>
noted

[03:34:00.0000] <othermaciej>
I think given Sam's follow-up, we would likely interpret your proposal as a very roundabout way of saying "change nothing" (thought it also admits that other solutions satisfying its constraints would also be acceptable)

[03:34:01.0000] <othermaciej>
I would personally prefer if it were more explicit about the change being requested

[03:35:00.0000] <othermaciej>
I will ask Sam what he thinks

[03:35:01.0000] <hsivonen>
othermaciej: as drafted, it allows the Editor to use a better reference without reopening a Decision

[03:36:00.0000] <annevk>
instead of ASCII we could just use Windows-1252 maybe

[03:36:01.0000] <annevk>
for UAs they're pretty much the same anyway

[03:37:00.0000] <hsivonen>
annevk: there are many ways to draft the passage without ever referencing ASCII

[03:38:00.0000] <hsivonen>
annevk: the point of my CP isn't to suggest an ideal wording but to make the point that we shouldn't be using the Decision Process for something as trivial as this so the Decision should uphold editor discretion

[03:41:00.0000] <annevk>
agreed that this is a waste of time; though all small one-liner comments at this point are marginal compared to what went before... :/

[03:42:00.0000] <annevk>
arguably it should be spelled out, as references are typically not followed

[03:43:00.0000] <Hixie>
i can't believe how much of our time julian and the chairs have so far wasted with this asinine issue

[03:43:01.0000] <annevk>
but I guess we can only do that once we have a fixed set of encodings and the encoding proliferation battle has been won

[03:53:00.0000] <MikeSmith>
annevk: http://twitter.com/s0enke/status/16761371509 "i've read the spec and i'm missing a xhr.upload() method which uses rfc1867 (which every serverside language supports ootb)"

[03:53:01.0000] <annevk>
hmm whatever then

[03:54:00.0000] <annevk>
send() accepts a request entity body as argument

[03:54:01.0000] <annevk>
which can be a File, but also a FormData object

[03:55:00.0000] <annevk>
the only reason 1867 is not mentioned because I defined it in terms of HTML5 stuff, as to not duplicate much and encourage code reuse

[03:55:01.0000] <hsivonen>
the environment of validator.nu supports straight http request bodies ootb but not form-based uploads...

[03:56:00.0000] <hsivonen>
differemce between servlet and php philosophies I guess

[03:57:00.0000] <MikeSmith>
does php have a philosophy?

[03:57:01.0000] <MikeSmith>
besides anarchy, I mean

[03:58:00.0000] <MikeSmith>
or whatever name there might be for "absence of any philosophy"

[04:06:00.0000] <drclue>
My take on PHP is that it is more organic than philosophic. I do code a lot with it and even write servers with it. For decades I coded most everything in C/C++ but not so much these days. It's weird , I have versions of many of my projects written independently in both languages

[04:08:00.0000] <s0enke>
hi

[04:09:00.0000] <s0enke>
MikeSmith, thx for the hint

[04:11:00.0000] <s0enke>
proposal for xhr2: define a "upload" method which uses default rfc1867 upload semantics. this way server software (php etc) would not have to reinvent the wheel for file uploads. any plans / ideas / hints on this?

[04:13:00.0000] <MikeSmith>
annevk: ↑

[04:13:01.0000] <Hixie>
MikeSmith: done the pub update

[04:14:00.0000] <Hixie>
bed now

[04:14:01.0000] <MikeSmith>
Hixie: thanks

[04:14:02.0000] <Hixie>
nn

[04:15:00.0000] <annevk>
s0enke, see the FormData object

[04:15:01.0000] <MikeSmith>
s0enke: hsivonen also points out that Java servlet environment that validator.nu runs under does not support form-based uploads ootb

[04:15:02.0000] <annevk>
s0enke, send() already accepts a request entity body; there's no need for a new method

[04:15:03.0000] <MikeSmith>
s0enke: said servlet environment is Jetty-based (for validator.nu)

[04:17:00.0000] <s0enke>
MikeSmith, well ok, but I guess the rfc1867 way is the most spread and default way to do file uploads, isnt it? ;-)

[04:17:01.0000] <s0enke>
annevk, will have a look, thx

[04:18:00.0000] <annevk>
I thought you read the specification? ;)

[04:19:00.0000] <s0enke>
/me hides

[04:19:01.0000] <MikeSmith>
s0enke: I suppose. But I recommend being deeply suspicious of anything published by the IETF

[04:22:00.0000] <s0enke>
MikeSmith, :D I rather implement formdata right now than discussing the value of rfcs ;-)

[04:24:00.0000] <MikeSmith>
s0enke: you clearly have your priorities all messed up

[04:27:00.0000] <MikeSmith>
s0enke: anyway, seriously, this is the best place to ask questions about anything related to actually implementing Web technologies

[04:27:01.0000] <MikeSmith>
at least the client-side part

[04:28:00.0000] <MikeSmith>
browser technologies

[04:28:01.0000] <annevk>
and you can get laughed at by spec editors

[04:28:02.0000] <MikeSmith>
and you can laugh at the spec editors too

[04:28:03.0000] <s0enke>
MikeSmith, yeah. all the real nerds live in irc ;)

[04:29:00.0000] <annevk>
MikeSmith, it's a popular sport

[04:30:00.0000] <drclue>
heh heh heh

[04:30:01.0000] <hsivonen>
s0enke: what server side stuff doesn't give you access to the HTTP request body directly?

[04:33:00.0000] <s0enke>
hsivonen, they all do. but at least in PHP, rfc1867 gives you powerful features like validating the request, putting the data into a tmpfile, creating a $_FILES array with all the metadata. In a normal post, you'd have to stream the content into a file in userland context (and blow up your server memory). the c context is universes more efficient.

[04:35:00.0000] <hsivonen>
streaming to a file blowing up server memory seems like an indication of something being very wrong

[04:36:00.0000] <hsivonen>
s0enke: actually, RFC 1867 is more of a buffering problem than the HTTP request body itself

[04:36:01.0000] <hsivonen>
s0enke: in enviroments that expose the real buffer cost and don't hide it from the programmer

[04:37:00.0000] <s0enke>
hsivonen, yeah. but then the world has to write their web apps in C again ;-)

[04:38:00.0000] <s0enke>
I'm just a slave of available technologies

[04:43:00.0000] <s0enke>
damn, the formData will be implemented in firefox 4 :-/

[04:45:00.0000] <drclue>
@sOenke:  You don't have to go back to C/C++ to do what you want. If you want to process a request in a unique way, you could always run a PHP based server on some other port to do that. 

[04:48:00.0000] <drclue>
There are times when C/C++ is needed for some particular bit of performance , or to do something really low level, but most days of the week PHP will do just fine, and one gets to skip the compile/link coffee break.

[04:50:00.0000] <hsivonen>
s0enke: why is it a "damn" for it to get implemented?

[04:59:00.0000] <hsivonen>
I wonder if there are proper statistics on how often HTTP requests are satisfied from proxy cache

[05:00:00.0000] <hsivonen>
or what percentage of Web users is behind proxies

[05:02:00.0000] <drclue>
@s0enke I think you can find a Firefox 3.6 hack-a-round for the time being at http://hacks.mozilla.org/2010/05/formdata-interface-coming-to-firefox/  and then perhaps trade it out when Firefox 4 comes out.

[05:04:00.0000] <drclue>
I for one am hoping that between WebM and a good showing in FireFox 4 that they get a new lease on life. These days I'm traveling in chrome (dev version at that)

[05:08:00.0000] <drclue>
I've also ordered an extra dose of heart medication for when IE9 comes out , in the event that they actually shock me with a decent offering. Now that they want to be all friendly with cloud computing , they really don't have the luxury of continuing to use IE to keep the Internet impudent   

[05:10:00.0000] <drclue>
Of course , since I boot to a Linux desktop , I'll have to take other folks word for how well IE9 does

[05:14:00.0000] <s0enke>
hsivonen, i missed a "not until" ;)

[05:17:00.0000] <drclue>
@s0enke   That happens to me too sometimes, that is data going missing between the ball cap and the keyboard :)

[05:19:00.0000] <s0enke>
layer eight packet loss ;)

[05:19:01.0000] <drclue>
skull ->preferences->type ahead buffer

[05:20:00.0000] <hsivonen>
s0enke: well, even if your alternative suggestion were specced, there'd be no chance of it being in Firefox before 4.0

[05:20:01.0000] <s0enke>
hsivonen, indeed.

[05:21:00.0000] <drclue>
At least they will have WebSockets in the bag

[05:22:00.0000] <drclue>
I think FF4 will probably test out pretty darn well for the most part

[05:23:00.0000] <drclue>
HTML5 is probably the first HTML version that everyone is trying to take seriously at the same time

[05:31:00.0000] <drclue>
Just thinking about this stuff sends me down memory lane Gopher , FIDONET , CHI 9 track, Altair , moon landings ,  rotary phones ,300 pound portable black and white TV's with a handy handle on top.

[05:33:00.0000] <erlehmann>
i wasn't born than yet.

[05:33:01.0000] <drclue>
Was you born before or after "I'm not a crook"

[05:33:02.0000] <hsivonen>
gopher died last week or so

[05:34:00.0000] <erlehmann>
drclue, for me its commodore pc 10, DOS, black-green screen, trying out all of the commands in the DOS manual.

[05:34:01.0000] <erlehmann>
i stopped after trying FORMAT, for obvious reasons

[05:34:02.0000] <drclue>
Ya , gopher died , but someone has nearly the whole gopher net on a hard drive

[05:34:03.0000] <erlehmann>
but hey, i was 10 or 11 yro

[05:35:00.0000] <erlehmann>
in the future, someone will have the whole internet of 2010 on a quantum holo storage medium the size of my fingernail

[05:35:01.0000] <drclue>
So you were probably around for those MFM full height 5 meg hard drives

[05:35:02.0000] <erlehmann>
my father told me the hard drive was a whopping 20 MB

[05:35:03.0000] <erlehmann>
in my first pc

[05:36:00.0000] <drclue>
I often contemplate what archaeologists will be able to find of us so many thousands of years after we are gone, since all our technologies are based upon materials with incredibly short life spans  

[05:37:00.0000] <drclue>
My very first "personal" computer was an altair

[05:39:00.0000] <s0enke>
erlehmann, 40MB with diskdoubler�

[05:39:01.0000] <erlehmann>
wat

[05:39:02.0000] <hsivonen>
/me thought DiskDoubler was a Mac-only thing

[05:40:00.0000] <drclue>
What was that , "stacker"?   Turned hard drives into door stops at an incredible rate

[05:40:01.0000] <s0enke>
hsivonen, I used it with dos 5.0 or so

[05:41:00.0000] <drclue>
Then there was DOS's daddy CPM

[05:41:01.0000] <drclue>
I remember them old Molecular computers , and dictaphone made some tanks too.

[05:42:00.0000] <drclue>
Ya remember them osborn's with the itty bitty screens?

[05:44:00.0000] <drclue>
Those were really some nerdy old days. I remember stripping 150 baud modems

[05:46:00.0000] <drclue>
I forget how many times I had to dive behind the teletype terminal to keep the night guards from catching me using "borrowed" computer time.

[05:47:00.0000] <erlehmann>
höhö höhöhö

[05:49:00.0000] <drclue>
What's really sorta weird is that computers have not really changed that much in all these decades.

[05:49:01.0000] <gsnedders>
erlehmann: Pff, the Habari cabal is by invitation only ;P

[05:52:00.0000] <erlehmann>
gsnedders, i'm going to pass all your secret tests of character, songoku-style !

[05:54:00.0000] <drclue>
(: ǝɹǝɥ punoɹɐ ƃuıƃuɐɥ ɹnoʎ ɟı ɹǝʇɔɐɹɐɥɔ ou ǝʌɐɥ noʎ

[05:56:00.0000] <erlehmann>
i r-r-reject your assumption undt replace it mitt mein own

[05:56:01.0000] <erlehmann>
mister jones

[05:57:00.0000] <drclue>
It was just a myth anyways 

[05:59:00.0000] <drclue>
Last night I looked upon the stair, and saw a man who was not there.

[05:59:01.0000] <drclue>
He was not there again today, how I wish that man would go away.

[06:49:00.0000] <erlehmann>
TabAtkins, can you refer me to an RDFa checker besides the N3 bookmarklet and pyRDFa ?

[06:52:00.0000] <karlushi>
oooh http://www.google.com/search?q=define:html5&hl=en

[06:55:00.0000] <karlushi>
oh it doesn't do only wikipedia

[06:55:01.0000] <karlushi>
oooh http://www.google.com/search?q=define:anarchy&hl=en

[06:55:02.0000] <karlushi>
wordnet also, when the word is defined at many places

[06:59:00.0000] <karlushi>
erlehmann, are there things which you specifically need for a rdfa checker that pyRDFa doesn't address?

[07:00:00.0000] <karlushi>
http://check.rdfa.info/

[07:00:01.0000] <erlehmann>
karlushi, my mentor insists on something not being properly scoped and namespacing missing, but I don't see it.

[07:00:02.0000] <erlehmann>
the site in question is this http://gsoc2010.dieweltistgarnichtso.net/

[07:01:00.0000] <karlushi>
the good channel for this question would be #swig (Semantic Web Interest Group), I guess.

[07:09:00.0000] <erlehmann>
karlushi, thanks.

[07:10:00.0000] <MikeSmith>
ah, what a rich means for thoughtful discussion twitter is

[07:10:01.0000] <MikeSmith>
twitter should have a "kneejerk reaction" button

[07:11:00.0000] <MikeSmith>
and a "kneejerk retweet reaction to kneejerk" reaction button

[07:23:00.0000] <karlushi>
MikeSmith, maybe a "rejerk" button

[07:27:00.0000] <MikeSmith>
karlushi: perfect

[10:27:00.0000] <TabAtkins>
annevk: No, I'm not going to be here at 2am my time.  I'm here now.

[10:27:01.0000] <TabAtkins>
erlehmann: No, I can't.

[10:27:02.0000] <erlehmann>
TabAtkins, thanks. but i already got help from karlushi and the people in #swig.

[10:29:00.0000] <TabAtkins>
cool

[10:30:00.0000] <annevk>
but now it is over

[10:30:01.0000] <annevk>
maybe it will align next time

[10:30:02.0000] <TabAtkins>
?_?

[10:57:00.0000] <jgraham>
/me starts humming The Smiths

[10:59:00.0000] <gsnedders>
jgraham: Liar. (Thankfully.)

[11:05:00.0000] <jgraham>
Sounds like an obscure b-side

[12:53:00.0000] <ap>
Hixie: perhaps HTML5 needs to specify that both WebSocket.url and WebSocket.URL should be supported now. I'm not sure if we'll want to drop support for what we shipped already, and if we're keeping it indefinitely, it can as well be specced

[13:01:00.0000] <zcorpan_>
ap: i doubt there's any content using URL

[13:02:00.0000] <ap>
zcorpan_: we can't go back and change it in Safari 5 though

[13:03:00.0000] <zcorpan_>
safari 5 shipped -75, which is a much bigger breaking change

[13:04:00.0000] <ap>
zcorpan_: that's a server side change, different department for many sites

[13:04:01.0000] <ap>
zcorpan_: and many servers already support both 75 and 76

[13:05:00.0000] <zcorpan_>
i wouldn't want to support both URL and url in opera

[13:05:01.0000] <zcorpan_>
i'd rather just have URL if it's required for compat, but as i said, i highly doubt it is (or even will be in a year)

[13:07:00.0000] <ap>
as I said, I don't know if we'll need to support it, I'm not sure. I don't think such a nonsense change should have been made to the spec that has shipping implementations

[13:07:01.0000] <annevk>
Firefox will ship with just url

[13:08:00.0000] <ap>
awesome :-/

[13:08:01.0000] <zcorpan_>
i argued for having URL everywhere, but in the end i gave up

[13:08:02.0000] <annevk>
you could support both for one more release or so and then drop it

[13:09:00.0000] <annevk>
yeah, I didn't agree with the no-URL camp either

[13:13:00.0000] <zcorpan_>
ap: there are more API changes compared to what shipped in safari 5: CloseEvent and the CLOSING state

[13:14:00.0000] <ap>
zcorpan_: true, although I'm not sure if adding these can practically break compatibility

[13:18:00.0000] <zcorpan_>
ap: EventSource.URL also changed to url

[13:18:01.0000] <ap>
zcorpan_: argh, I didn't know that

[13:27:00.0000] <AryehGregor>
How cryptic: http://blog.mozilla.com/rob-sayre/2010/06/21/networking/

[15:09:00.0000] <Hixie>
ap: dude, just drop .URL, it's too early in the life of the API to worry about back compat

[15:09:01.0000] <Hixie>
ap: we're still discussing changing the entire protocol!

[15:09:02.0000] <ap>
Hixie: by the time we can ship the change, it might not be too early. and same for EventSource

[15:10:00.0000] <Hixie>
so ship it now

[15:10:01.0000] <gsnedders>
/me notes "by the time we can" implies not now

[15:11:00.0000] <Hixie>
ap: we're not adding both, that's just dumb. Either we should use .URL throughout or .url throughout (except Document), but either way you and mozilla need to decide which it is and stick to it.

[15:12:00.0000] <Hixie>
ap: you can ship a point update with this change if it's going to be that long until the next update

[15:12:01.0000] <ap>
Hixie: we shipped something that matched the draft spec, is there a good reason for it to change under us? this seems like a completely arbitrary change

[15:13:00.0000] <Hixie>
well first, drafts spec are subject to change at a whim, you know that. that's the risk of being on the cutting edge. Like I said, we're still considering a completely new protocol to replace the current websocket one.

[15:14:00.0000] <Hixie>
but the reason for this change was a discussion on public-webapps which had people from webkit, mozilla, and opera and which decided that we should consistently use lowercase attribute names so as to not make authors go crazy trying to work out what was lowercase and what was uppercase

[15:14:01.0000] <Hixie>
the minimum change was to change from URL to url

[15:14:02.0000] <Hixie>
going the other way would have required even more implementations to change

[15:16:00.0000] <ap>
Hixie: XMLHttpRequest is also a draft, let's maybe change it to XMLHTTPRequest for consistency?

[15:17:00.0000] <Hixie>
XMLHttpRequest isn't a draft, it was shipped in like 1999

[15:17:01.0000] <ap>
Hixie: I think that there should be a cost/benefit analysis be applied, even to bleeding edge drafts

[15:17:02.0000] <Hixie>
there is

[15:17:03.0000] <Hixie>
and here the cost is MASSIVELY more on the side of being inconsistent

[15:17:04.0000] <ap>
there is no consistency in "url" casing on the Web, is there?

[15:17:05.0000] <jgraham>
s/on the web/in DOM APIs/

[15:18:00.0000] <ap>
Hixie: what about document.URL?

[15:19:00.0000] <Hixie>
see the thread on public-webapps

[15:19:01.0000] <Hixie>
we discussed this to death

[15:19:02.0000] <Hixie>
I was on the side of going all uppercase, but that had a higher cost than going all lowercase except Document

[15:20:00.0000] <abarth>
Hixie: how would you like to handle implementation feedback on the parsing algorithm?

[15:20:01.0000] <ap>
so, evaluating this spec change. benefit: none, since there is still no consistency

[15:20:02.0000] <abarth>
Hixie: emails to public-html? bugs in the w3c bugzilla?

[15:20:03.0000] <ap>
cost: high, need to change shipping implementations

[15:21:00.0000] <Hixie>
abarth: bugs in bugzilla or e-mails to the whatwg list -- if you put bugs in bugzilla, mark them P1/critical so I'll deal with them at the same time as henri's

[15:21:01.0000] <Hixie>
ap: if you want to change the decision, convince opera and mozilla to go the other way.

[15:21:02.0000] <abarth>
Hixie: ok, will do.  thanks

[15:22:00.0000] <ap>
Hixie: zcorpan_ and annevk seemed to prefer URL per the above discussion

[15:22:01.0000] <ap>
Hixie: that leaves Mozilla as the odd one out

[15:22:02.0000] <jgraham>
abarth: There is also a page on the whatwg wiki documenting parser issues

[15:23:00.0000] <jgraham>
If you could add things to that it would be good

[15:23:01.0000] <abarth>
jgraham: link?

[15:23:02.0000] <Hixie>
ap: i want uppercase too, but it's not a popularity decision.

[15:23:03.0000] <Hixie>
ap: it's a matter of what involved the least changes to implementations.

[15:23:04.0000] <jgraham>
abarth: http://wiki.whatwg.org/wiki/ParserIssues

[15:23:05.0000] <abarth>
jgraham: thxs

[15:23:06.0000] <ap>
Hixie: Firefox didn't ship EventSource or WebSocket

[15:24:00.0000] <jgraham>
ap: Having less inconsistency is still better

[15:24:01.0000] <Hixie>
ap: i'm not the one you have to convince

[15:25:00.0000] <ap>
jgraham: that's probably the part that was discussed to death on the mailing list (although for me, document.URL is the most familiar API, so being consistent with it is most important)

[15:25:01.0000] <Hixie>
(for me too)

[15:26:00.0000] <ap>
Hixie: who should I convince?

[15:26:01.0000] <Hixie>
reopen the thread on public-webapps and get the participants to agree to implement the opposite

[15:26:02.0000] <TabAtkins>
ap: Check out the thread for those who argued for lowercase, I suppose.

[15:28:00.0000] <gsnedders>
Hixie: http://www.hixie.ch/tests/adhoc/http/content-type/sniffing/ seems bogus

[15:28:01.0000] <gsnedders>
Hixie: 016 isn't that MIME type at all

[15:28:02.0000] <Hixie>
gsnedders: quite likely

[15:28:03.0000] <gsnedders>
Content-Type: application/x-httpd-php

[15:29:00.0000] <Hixie>
gsnedders: send me a mail with bugs, and i'll add them to the pile of e-mails about test bugs that i've been keeping. One day I'll have an intern or something I can get to fix them.

[15:32:00.0000] <Hixie>
abarth: the </p</div> thing was for compat with IE

[15:33:00.0000] <abarth>
Hixie: i see.  some folks in the webkit community are worried about whether that will cause problems for mobile sites that are used to a webkit monoculture :(

[15:33:01.0000] <Hixie>
yup, it probably will.

[15:34:00.0000] <zcorpan_>
abarth: opera does the same thing as IE for < in tags. we've found that there are some pages that break if we follow mozilla, and other pages break if we follow ie

[15:34:01.0000] <zcorpan_>
abarth: however we've followed ie here for a long time and it doesn't come up very often

[15:42:00.0000] <abarth>
zcorpan_: i see

[15:42:01.0000] <abarth>
zcorpan_: maybe its a crapshoot either way

[15:43:00.0000] <abarth>
i'm all for aligning behavior so we don't have these problems in the future

[15:43:01.0000] <abarth>
i'm not sure how to estimate the risks in the mobile space

[15:44:00.0000] <abarth>
but there are also likely counterbalancing risks in the intranet that IE folks are worried about

[15:44:01.0000] <zcorpan_>
you could look at opera's market share on mobile :)

[15:44:02.0000] <zcorpan_>
and ie's

[15:44:03.0000] <abarth>
zcorpan_: my understanding is there's a bunch of mobile content that's authored specifically for webkit, but i could be wrong about that

[15:45:00.0000] <abarth>
we had a lot of random test cases in the test suite that relied on this quirk

[15:45:01.0000] <abarth>
including one from the official CSS test suite

[15:46:00.0000] <TabAtkins>
The clear solution is to break everyone's pages so the monoculture breaks up.

[15:48:00.0000] <erlehmann>
TabAtkins, that reminds me of an exchange two of my buddies hat. one was "my collegues broke the nightly build" and the response was "break them"

[15:49:00.0000] <TabAtkins>
Your buddies hats are rather eloquent for hats.

[15:50:00.0000] <erlehmann>
I use the neo keyboard you insensitive clod.

[15:51:00.0000] <TabAtkins>
neo keyboard?

[15:51:01.0000] <erlehmann>
T and D keys are right next to each other.

[15:51:02.0000] <erlehmann>
neo2 is a german layout optimized for speed. like dvorak, but for german and closely related languages.

[15:51:03.0000] <TabAtkins>
Ah, kk.

[15:51:04.0000] <erlehmann>
and it has all the funny special chars other germans never get right, like „these“

[15:52:00.0000] <erlehmann>
But I laughed like a little girl, thank you, hats of.

[15:52:01.0000] <TabAtkins>
Man, I wish English used asymmetric quotes like that.

[15:52:02.0000] <erlehmann>
Incidentally, I have two hats right in this room.

[15:52:03.0000] <TabAtkins>
/me has no hats, and thus is sad.

[15:52:04.0000] <zcorpan_>
erlehmann: are they talking to each other?

[15:52:05.0000] <TabAtkins>
Next time I visit my parents I'm buying a proper cowboy hat, for reals this time.

[15:52:06.0000] <TabAtkins>
/me has to show off his Texan pride.

[15:53:00.0000] <erlehmann>
zcorpan_, i believe those hats were already dead and stuffed when my dad bought them.

[15:53:01.0000] <zcorpan_>
/me wonders what he should play with next using websockets

[15:54:00.0000] <franksalim>
zcorpan_, what have you played with so far?

[15:54:01.0000] <zcorpan_>
chat room and multi-user painting

[15:55:00.0000] <TabAtkins>
Multi-user doom.

[15:55:01.0000] <TabAtkins>
Which means, I suppose, Quake.

[15:56:00.0000] <zcorpan_>
i want something i can hack in a few hours :)

[15:58:00.0000] <zcorpan_>
maybe i should find some single-player canvas game that i can easily implement multiplayer support

[15:58:01.0000] <Philip`>
Canvex!

[15:59:00.0000] <AryehGregor>
Is there any such thing as a single-player game where it's easy to implement multiplayer support?

[15:59:01.0000] <AryehGregor>
I'd think that would typically require a rewrite.

[15:59:02.0000] <TabAtkins>
The box-drawing game.  Where you draw lines on a grid and score a point (and get to go again) whenever you finish all four sides of a box.

[15:59:03.0000] <Philip`>
/me wishes he could remember where he put his experimental multiplayer version

[16:01:00.0000] <Philip`>
TabAtkins: Turn-based games are bad demos, since latency doesn't matter and you could just do an XHR per turn

[16:01:01.0000] <Philip`>
Need something real-time to demonstrate the benefit of having a persistent socket

[16:01:02.0000] <roc>
pong

[16:01:03.0000] <TabAtkins>
Philip`: True.

[16:02:00.0000] <TabAtkins>
2-player asteroids!

[16:02:01.0000] <TabAtkins>
Both ships up at once, explode on collision, winner is the one with the most points at the end of each round.

[16:02:02.0000] <Philip`>
Pong is probably too real-time - it'll look weird when the ball bounces even though it looked like it missed the opponent's bat

[16:02:03.0000] <Philip`>
and you can't do much decent lag compensation for Pong

[16:04:00.0000] <ment>
what about ssh client? that's a good game that could benefit from persistent connection where latency matters but not too much

[16:07:00.0000] <zcorpan_>
pong seems interesting to do

[16:07:01.0000] <Hixie>
zcorpan_: what would be really useful is a web page you can go to and give the client information (origin, ws: URL, etc) and have it tell you what the server is doing wrong

[16:07:02.0000] <Hixie>
zcorpan_: including checking for various obvious mistakes, like whether the challenge was constructed little-endian rather than big-endian, or whatnot.

[16:08:00.0000] <Hixie>
/me is writing a multiplayer websocket game here http://software.hixie.ch/fun/cuddlyworld/

[16:09:00.0000] <Hixie>
(very early days yet)

[16:09:01.0000] <zcorpan_>
Hixie: i use opera's error console to tell what the server is doing wrong ;)

[16:09:02.0000] <Hixie>
(and ignore how slow it is, that's just because i have all the debugging checks enabled)

[16:10:00.0000] <Philip`>
Tron is more interesting than Pong and not much harder to implement

[16:10:01.0000] <Philip`>
/me wrote a basic Tron on his calculator but it was a really very slow calculator and only got about 0.5fps :-(

[16:11:00.0000] <zcorpan_>
Hixie: what user/pass should i use to play?

[16:11:01.0000] <Hixie>
any

[16:11:02.0000] <Hixie>
it creates an account when you join

[16:11:03.0000] <Hixie>
the password is for rejoining

[16:11:04.0000] <Philip`>
(then I wrote an emulator (in Perl and C) for a ~1MHz 8-bit processor and then wrote Tron for that and it was much faster)

[16:12:00.0000] <Hixie>
zcorpan_ a utility for testing servers would be good for checking that they handle all the various keys correctly --just testing in a browser isn't going to hit all the edge cases each time

[16:13:00.0000] <zcorpan_>
hmm, clearly we need a longer timeout before giving up in the opening handshake

[16:13:01.0000] <Hixie>
er no

[16:13:02.0000] <Hixie>
my server crashed

[16:13:03.0000] <Hixie>
for unrelated reasons

[16:13:04.0000] <Hixie>
restarting it

[16:16:00.0000] <Hixie>
well, it seems i have some bugs to figure out before i can demo this!

[16:16:01.0000] <Hixie>
anyway, bbiab

[16:23:00.0000] <AryehGregor>
I remember recently seeing statistics on how long things tend to stay in browsers' caches, but I can't seem to find them.  Does anyone have any pointers?

[16:23:01.0000] <AryehGregor>
(My recollection is typically less than a day, but I want to double-check that.)


2010-06-23
[18:48:00.0000] <annevk>
hmm, will update html5-diff tomorrow

[18:48:01.0000] <annevk>
time for some water now, you know

[19:31:00.0000] <jwm>
does     QQaalalala

[19:31:01.0000] <jwm>
hmm

[19:31:02.0000] <jwm>
heh

[19:31:03.0000] <jwm>
time for dinner

[21:07:00.0000] <othermaciej>
hi all

[22:02:00.0000] <roc>
TabAtkins: have you tried inspecting the shadows with some kind of screen magnifier? Preferably one that can show the actual color values of pixels?

[22:14:00.0000] <Viper550>
Is it me, or is the forum scrpt that whatwg uses not html-compliant?

[22:35:00.0000] <Viper550>
hey?

[22:40:00.0000] <Hixie>
Viper550: it is almost certainly not particularly good code, indeed

[22:41:00.0000] <Hixie>
Viper550: we didn't use the quality of the code as a deciding factor in picking the software :-)

[22:41:01.0000] <Viper550>
you've seen fluxbb 1.4?

[22:43:00.0000] <Viper550>
In case you haven't heard, its a fork of punbb 1.2

[22:44:00.0000] <Hixie>
Viper550: i don't know much about forum software personally

[00:00:00.0000] <annevk>
you'd think Leif would have noticed I didn't reply to any comment on html5-diff since last time I made it ready and that last time I replied to age old comments

[00:01:00.0000] <annevk>
it is of course more interesting to assume the chairs and I ignored him, because then you can vent frustrations, even if not at all justified

[00:28:00.0000] <MikeSmith>
kennyluck: is Guo Fu a family name or a personal/given name?

[00:28:01.0000] <kennyluck>
MikeSmith: it's a word. It means "Country Father" :)

[00:29:00.0000] <kennyluck>
father of a country

[00:29:01.0000] <MikeSmith>
oh

[00:29:02.0000] <kennyluck>
(well, it's possible that someone has this given name, but this is kind of odd then)

[00:34:00.0000] <annevk>
MikeSmith, we're gonna publish tomorrow?

[00:34:01.0000] <MikeSmith>
annevk: yes

[00:34:02.0000] <annevk>
okay then

[00:36:00.0000] <MikeSmith>
annevk: as soon as you have your doc ready, let me know, and I will move it over to the dated TR location

[00:36:01.0000] <MikeSmith>
and will let the webmaster know it's ready

[00:37:00.0000] <annevk>
yeah, I'll go through the new comments now and see what needs changing

[00:59:00.0000] <abarth>
Hixie: the bit about spinning the event loop isn't who things work internally

[00:59:01.0000] <abarth>
Hixie: nested event loops are bad new bears

[01:05:00.0000] <MikeSmith>
interesting to see a Microsoft commenter on the whatwg mailing list

[01:06:00.0000] <MikeSmith>
/me wonders what "Live Labs / Seadragon" is

[01:07:00.0000] <MikeSmith>
abarth: I seem to remember hsivonen mentioning something about nested event loops recently

[01:07:01.0000] <MikeSmith>
in bugzilla or here

[01:07:02.0000] <MikeSmith>
I wonder if the context is the same

[01:07:03.0000] <abarth>
I think the behavior ends up being the same

[01:07:04.0000] <abarth>
it's just a difference in the processing model from the actual processing

[01:08:00.0000] <abarth>
the problem with nested event loops

[01:08:01.0000] <abarth>
is you get in trouble when you try to unwind them

[01:08:02.0000] <abarth>
because lots of stuff gets re-entered

[01:08:03.0000] <MikeSmith>
ah

[01:08:04.0000] <abarth>
that might not be expecting it

[01:08:05.0000] <MikeSmith>
yeah, I can imagine

[01:08:06.0000] <abarth>
the parser needs to be re-entrant, which makes it tricky

[01:09:00.0000] <abarth>
DOM mutation also needs to be re-entrant, which is really hard'

[01:09:01.0000] <abarth>
and the source of lots of bugs :(

[01:09:02.0000] <MikeSmith>
that I also can imagine

[01:10:00.0000] <MikeSmith>
or actually, remember

[01:10:01.0000] <MikeSmith>
from previous life in browser product development

[01:13:00.0000] <MikeSmith>
abarth: the context of your current discussion about this is the tokenizer spec?

[01:14:00.0000] <MikeSmith>
the closing-the-input-stream part?

[03:19:00.0000] <MikeSmith>
kennyluck: Murata-san is http://ja.wikipedia.org/wiki/村田真

[03:19:01.0000] <MikeSmith>
and http://en.wikipedia.org/wiki/Makoto_Murata

[03:43:00.0000] <kennyluck>
MikeSmith: "In Mainland China, Malaysia  and Singapore....vertical writing is now very rare, more so in print than in writing and signage." from http://en.wikipedia.org/wiki/Horizontal_and_vertical_writing_in_East_Asian_scripts

[03:43:01.0000] <kennyluck>
I guess I was wrong about novels in Mainland China.

[03:43:02.0000] <MikeSmith>
interesting

[03:48:00.0000] <annevk>
wtf is sam on

[04:18:00.0000] <jgraham>
annevk: See, this is why you shouldn't do things for people. They will only complain

[04:19:00.0000] <annevk>
yeah, before I settled on "I already had." I had a lot of other drafts

[04:20:00.0000] <annevk>
dealing with the HTML WG is painful

[04:20:01.0000] <annevk>
so I'm calling it a day for now as far as that is concerned

[04:21:00.0000] <annevk>
http://code.google.com/apis/maps/documentation/javascript/examples/streetview-map.html is indeed very neat hsivonen

[04:23:00.0000] <annevk>
hehe, Google visited my place while I was home

[05:00:00.0000] <gsnedders>
annevk: Unusual :P

[05:02:00.0000] <annevk>
http://www.w3.org/2010/06/14-ab-minutes (W3C Member-only) -- worth reading

[05:04:00.0000] <Omer>
Hi

[05:05:00.0000] <Omer>
Someone here?

[05:05:01.0000] <Guest45746>
Hi

[05:05:02.0000] <Guest45746>
Someone here?

[05:05:03.0000] <annevk>
better to just ask your question

[05:05:04.0000] <Guest45746>
ok

[05:06:00.0000] <Guest45746>
how i can send a message with postMessage between windows/tabs?

[05:06:01.0000] <annevk>
by getting a reference to the other window and invoke postMessage() on it with the correct arguments?

[05:07:00.0000] <Guest45746>
how i getting reference to other window?

[05:07:01.0000] <Guest45746>
with the same domain

[05:08:00.0000] <annevk>
I think you can only get a reference if it's a related browsing context

[05:08:01.0000] <annevk>
so if it's window.opener, or in window[x]

[05:09:00.0000] <Guest45746>
i can't to access to other window?

[05:11:00.0000] <annevk>
no

[05:11:01.0000] <Guest45746>
ok thanks you ill do that on the sever side

[05:12:00.0000] <annevk>
you can maybe sync some info using localStorage for same-origin scenarios

[05:12:01.0000] <annevk>
the storage event is supposed to dispatch in each tab of the same origin

[05:13:00.0000] <Guest45746>
what i wanted is to share the same websockets connection to all of the windows

[05:14:00.0000] <Guest45746>
but ill do that on the server side i dont think have anuther way

[05:14:01.0000] <annevk>
shared worker?

[05:14:02.0000] <Guest45746>
what you mean?

[05:15:00.0000] <annevk>
see http://www.whatwg.org/ww

[05:16:00.0000] <annevk>
you can share a JavaScript "thread" among several same-origin windows and then you can use sockets from that thread

[05:18:00.0000] <Guest45746>
so when i created one shared worker all the pages use him?

[05:19:00.0000] <annevk>
pretty much

[05:24:00.0000] <Guest45746>
thanks tou very much

[05:24:01.0000] <Guest45746>
:)

[05:24:02.0000] <Guest45746>
i understand that

[05:25:00.0000] <Guest45746>
ww is cross browsers?

[05:26:00.0000] <Guest45746>
i think its desnt but i wanna to be sure

[05:27:00.0000] <Guest45746>
you*

[05:28:00.0000] <annevk>
nothing of this is cross browser really

[05:30:00.0000] <Guest45746>
but its great idea :)

[05:32:00.0000] <jgraham>
Well it is intended to be cross browser

[05:32:01.0000] <jgraham>
Just need to wait for the browsers

[05:36:00.0000] <Guest45746>
so i can be sure its will be in the next??

[05:39:00.0000] <annevk>
there's no guarantees really

[05:39:01.0000] <annevk>
but it's shipping in some already

[05:44:00.0000] <Guest45746>
you know when websockets protocol ill finished?

[07:15:00.0000] <annevk>
Guest45746, no

[07:16:00.0000] <annevk>
Guest45746, when three implementations are shipping it's prolly finished for better or worse

[07:39:00.0000] <jgraham>
So who is going to write the change proposal for @summary?

[07:52:00.0000] <Guest45746>
element.parentElement its part of html5?

[07:54:00.0000] <Philip`>
Guest45746: That seems to be a non-standard IE-only feature

[07:54:01.0000] <MikeSmith>
annevk: it's been pointed out to me that "The autobuffer attribute on media elements was renamed to preload." is mentioned twice in the diffs doc

[07:55:00.0000] <MikeSmith>
in section 5.1 and also in 5.2

[07:55:01.0000] <MikeSmith>
/me goes to check when it actually changed

[07:56:00.0000] <Lachy>
jgraham, what change proposal is needed?

[07:56:01.0000] <Lachy>
Just a zero-edit proposal, or one that says to make it fully obsolete?

[07:56:02.0000] <Lachy>
or both?

[07:56:03.0000] <Guest45746>
so how can i get the parent

[07:57:00.0000] <Guest45746>
ans why chrome and safari use parentElement?

[07:57:01.0000] <Guest45746>
and

[07:57:02.0000] <Philip`>
Guest45746: Use parentNode

[08:00:00.0000] <jgraham>
Lachy: I would prefer both. But at least a zero-edits one that collects together all the evidence that summary is pointless or harmful

[08:00:01.0000] <Guest45746>
thanks now its work on firefox :)

[08:01:00.0000] <jgraham>
Lachy: I have no idea how one is supposed to write a change proposal in such a way that the chairs actually pay attention to the points made though

[08:01:01.0000] <Guest45746>
but why chrome and safari use it?

[08:17:00.0000] <Lachy>
jgraham, no idea either, given how they've merely counted arguments and counter arguments as balancing each other out, regardless of their technical merit

[08:18:00.0000] <Lachy>
but that's an issue I need to follow up with sam, as I'm still waiting for him to stop ignoring my email about the issue.

[08:26:00.0000] <MikeSmith>
annevk: seems like "The autobuffer attribute for media elements is now named preload" should be removed from the "Changes since 4 March 2010" list

[08:27:00.0000] <MikeSmith>
because that change was actually made in February

[08:27:01.0000] <MikeSmith>
and you already to have it listed in "Changes from 25 August 2009 to 4 March 2010"

[08:43:00.0000] <TabAtkins>
roc: Well, I'm screenshotting and then checking pixels in GIMP.  I'd think that's close enough.

[08:43:01.0000] <jgraham>
/me wonders what the context for that was

[08:43:02.0000] <TabAtkins>
There's a www-style convo about how the *-shadow properties do/should work.

[08:44:00.0000] <TabAtkins>
There's disagreement about how text-shadow works *right now*, which is weird.

[08:45:00.0000] <Philip`>
Reverse-engineering shadow algorithms is fun

[08:45:01.0000] <Philip`>
/me wonders how similar CSS shadows are to canvas shadows

[08:46:00.0000] <TabAtkins>
canvas shadows are an exact copy of webkit's bugs in css shadows.

[08:47:00.0000] <TabAtkins>
Including a substantial change in shadow algorithm when the shadow blur is > 8px.

[08:48:00.0000] <Philip`>
Does it do the slightly weird thing with sigma=sqrt(2*blur) if blur>8 ?

[08:48:01.0000] <TabAtkins>
So, yes.

[08:48:02.0000] <Philip`>
That's not a substantial change, it's just interpreting the argument a bit differently

[08:48:03.0000] <TabAtkins>
It makes things weird, at least.

[08:48:04.0000] <TabAtkins>
smfr was complaining about it.

[08:49:00.0000] <Philip`>
It makes the shadow appear to scale more linearly, without big visual differences between small integers, if I remember correctly

[08:50:00.0000] <TabAtkins>
I was told it was a performance optimization.

[08:50:01.0000] <Philip`>
I assume the behaviour was inherited from CG, rather than being an intentional WebKit feature

[08:50:02.0000] <TabAtkins>
Yeah, I assume so.

[08:51:00.0000] <Philip`>
(I assume by "webkit's bugs" you meant the CG port of WebKit)

[08:51:01.0000] <TabAtkins>
So anyway, roc and I are disagreeing about how far Firefox's shadow extends for a 100px blur.  I say 50px, based on screenshots + pixel comparisons.  He says 79px, based on a screen magnifier and pixel checking.

[08:51:02.0000] <TabAtkins>
I dunno, I'm just paraphrasing from smfr.

[08:52:00.0000] <Philip`>
How can you disagree on something as objective as that? :-)

[08:52:01.0000] <TabAtkins>
I don't know!

[08:53:00.0000] <Philip`>
Just post the screenshot and say what pixel value you see at the edge of the blur

[08:53:01.0000] <TabAtkins>
Yeah, doing that now.

[08:53:02.0000] <Philip`>
and then find out it varies between different platforms or something :-)

[08:55:00.0000] <TabAtkins>
Well, it appears to be the same between my linux desktop and my windows laptop.

[08:55:01.0000] <TabAtkins>
But roc is checking on win and linux as well.

[08:57:00.0000] <Philip`>
With FF3.6.3 on Linux, I see (254,254,254) at 80ish pixels from the edge of the shape

[08:57:01.0000] <TabAtkins>
wtf

[08:59:00.0000] <Philip`>
You're not using Gimp's magic wand with a threshold of 15 to determine the edge of the shadow?

[08:59:01.0000] <TabAtkins>
No, I'm just zooming in and using the dropper to check the color manually.

[08:59:02.0000] <Philip`>
Ah, good :-)

[08:59:03.0000] <TabAtkins>
And when I did use the wand, I specifically dropped the threshold to 0.

[09:04:00.0000] <TabAtkins>
...huh.

[09:05:00.0000] <TabAtkins>
When I do a white shadow on black, the shadow is 50px wide.  +-1.  When I do a black shadow on white, it's about 80px wide, like roc says.

[09:06:00.0000] <Philip`>
Oh

[09:06:01.0000] <Philip`>
(That doesn't seem surprising since it depends on the rounding)

[09:06:02.0000] <TabAtkins>
That's stupid.

[09:08:00.0000] <TabAtkins>
Oh, wait.  Huh.  Maybe I was just stupid?  Now I'm measuring 80px for white-on-black.

[09:09:00.0000] <TabAtkins>
roc: So, nevermind.  I think I just measured webkit's screenshot twice, rather than measuring webkit once and firefox once.

[09:11:00.0000] <Philip`>
/me just tried yellow on magenta, and it looks like actually it rounds up the same as it rounds down, so that's not an issue

[09:11:01.0000] <TabAtkins>
Yeah, rounding's not the issue.  I was juggling too many screenshots at once.

[09:19:00.0000] <Philip`>
Hmm, why does text-shadow indicate the blur is measured in pixels?

[09:20:00.0000] <TabAtkins>
Where does it do so?

[09:21:00.0000] <Philip`>
Isn't that slightly meaningless, since the blur is theoretically infinite in size and I assume the specified size is not equal to the Gaussian sigma

[09:21:01.0000] <Philip`>
"text-shadow: 0 0 100px black;"

[09:22:00.0000] <Philip`>
(s/$/?/ in earlier question)

[09:22:01.0000] <TabAtkins>
The blur length specifies how far you have to go before the blur has gotten to within 1% of the full color/full transparency.  Or at least, that's probably what we're specifying now.

[09:22:02.0000] <TabAtkins>
It is remarkably underspecified right now.

[09:23:00.0000] <TabAtkins>
So yeah, the blur is theoretically infinite, but the area within the threshold isn't.

[09:23:01.0000] <Philip`>
1% in what colour space?

[09:23:02.0000] <TabAtkins>
rgb, the default css color space

[09:24:00.0000] <TabAtkins>
/me goes to get breakfast, bbiab

[09:29:00.0000] <karlushi>
/me sighs at the difficulties to share Web files with people at the offices. No interoperability for packaging between the different browsers. :((

[09:30:00.0000] <Philip`>
TabAtkins: Doesn't "1% of the full color" mean you need to solve a cubic equation to work out what sigma to use for the Gaussian, or am I getting my maths all wrong?

[09:30:01.0000] <Philip`>
(I'm interpreting it as meaning G(x) = 0.01 as defined by http://upload.wikimedia.org/math/b/5/a/b5a3ab06a75210694218bcbd7370eea3.png)

[09:32:00.0000] <TabAtkins>
Philip`: I have no clue.

[09:32:01.0000] <TabAtkins>
But I'm certain there's a much faster approximation.

[09:35:00.0000] <Philip`>
So the text-shadow property will be giving an approximation of an arbitrary percentage threshold?

[09:36:00.0000] <Philip`>
That doesn't sound very mathematically sane :-(

[09:36:01.0000] <TabAtkins>
Who cares about math, we're doing webdesign.

[09:37:00.0000] <jgraham>
TabAtkins: Seems like a good idea to understand the maths before you specify something insane

[09:38:00.0000] <TabAtkins>
To be fair, it appears that the implementors who're paying attention are cool with all that anyway.

[09:39:00.0000] <TabAtkins>
The only question is whether a 100px blur length means that the full blur area (extending inside and outside of the unblurred shadow boundary) is 100px wide, or that the blur area must extend 100px inside and outside.

[09:39:01.0000] <TabAtkins>
And that's just a question of bikeshedding and convergence with sufficiently related technologies, luckily.

[09:40:00.0000] <Philip`>
Is canvas sufficiently related?

[09:40:01.0000] <Philip`>
(I think that has a mathematically reasonably sane definition of shadow rendering :-) )

[09:40:02.0000] <TabAtkins>
Only insofar as canvas is literally "whatever webkit did for shadows".

[09:40:03.0000] <TabAtkins>
So, since we're already considering webkit, canvas just weights that slightly.

[09:41:00.0000] <annevk>
MikeSmith, still around?

[09:41:01.0000] <jgraham>
/me would have assumed that the blur length would be defined as a sigma of a given number of pixels, or something

[09:41:02.0000] <jgraham>
/me has no idea how it actually works though

[09:41:03.0000] <Philip`>
It's not precisely what Safari did, since Safari had (has?) incorrect clipping for small shadow sizes

[09:41:04.0000] <TabAtkins>
jgraham: Your average web author has absolutely no idea what a sigma is.

[09:41:05.0000] <Philip`>
(It would cut off the leftmost pixel of the shadow, in a visible way)

[09:41:06.0000] <jgraham>
TabAtkins: So?

[09:42:00.0000] <jgraham>
TabAtkins: Presumably they just adjust until they get the result they want

[09:42:01.0000] <TabAtkins>
So we want the blur length to mean something reasonable?

[09:42:02.0000] <Philip`>
TabAtkins: They just need to know that it's correlated to the apparent width of the shadow

[09:42:03.0000] <jgraham>
It does mean something reasonable

[09:42:04.0000] <Philip`>
(It can never be equal to the width since the shadow doesn't have an edge)

[09:43:00.0000] <jgraham>
It means that the shadow is 1/e of the maximum value at that number of pixels

[09:43:01.0000] <TabAtkins>
That's putting implementor interests over author interests, for little benefit to implementors and measurable downside to authors.

[09:43:02.0000] <jgraham>
I challenge you to measure that downside

[09:43:03.0000] <TabAtkins>
I did.  It's -13 utilons.

[09:43:04.0000] <jgraham>
I don't see why an author cares if it is 1/e or 1/100

[09:44:00.0000] <annevk>
MikeSmith, I fixed that and some other issues

[09:44:01.0000] <TabAtkins>
An author cares that the language is predictable.  Saying "well, bigger numbers mean bigger shadows" is only *barely* acceptable.  Ideally, the author shoudl be able to predict how large the shadow will be for a given length.

[09:45:00.0000] <jgraham>
The shadows will be infinitely big, apart from hardware limitations

[09:45:01.0000] <TabAtkins>
Now you're being pedantic.

[09:46:00.0000] <jgraham>
Not really :)

[09:46:01.0000] <jgraham>
I mean, in theory, there is no edge of the shadow.

[09:46:02.0000] <TabAtkins>
The measure we're planning to use on defining the "edge" of the shadow is 1% from full color/full transparency.

[09:47:00.0000] <jgraham>
/me would really use n sigma for integer n

[09:47:01.0000] <TabAtkins>
The two are interchangeable.

[09:47:02.0000] <TabAtkins>
Also, we dont' specify the blurring algorithm, so we can't assume a gaussian blur.

[09:48:00.0000] <jgraham>
I assume no one will implement a non-gaussian blur

[09:48:01.0000] <TabAtkins>
If you just wanted to use 3 sigma, that's fine.

[09:48:02.0000] <jgraham>
If they did it would be even weirder

[09:48:03.0000] <Philip`>
How can you specify shadowing without specifying the blurring algorithm?

[09:48:04.0000] <TabAtkins>
Man, I dunno.  I just know there was pushback against us specifying a precise algo.

[09:49:00.0000] <Philip`>
That seems to be basically leaving the entire thing undefined

[09:49:01.0000] <TabAtkins>
Philip`: Barely.  We just define the boundaries, and then say there should be a gradient of some kind.

[09:51:00.0000] <Philip`>
So someone could e.g. do a 100%-to-0% radial gradient around the edge of the shape, so that a tiny dot would have as bright a shadow as a large box?

[09:51:01.0000] <Philip`>
That doesn't sound good for interoperability

[09:51:02.0000] <TabAtkins>
See: pushback against specifying a precise algo.

[09:52:00.0000] <jgraham>
TabAtkins: It needen't bee to precise

[09:52:01.0000] <Philip`>
(Either people will implement things very differently, which is bad for interop, or they'll implement it the same anyway, in which case it can be specced)

[09:52:02.0000] <jgraham>
It should say "implementation defined approximation to a gaussian blur" or something

[09:53:00.0000] <jgraham>
You could argue that anything is an approximation to a gaussian of course, but that would be deliberatly perverse

[09:53:01.0000] <TabAtkins>
Time for conference call!

[09:54:00.0000] <Philip`>
/me just thinks it seems crazy to specify a rendering behaviour without actually specifying the rendering behaviour

[09:57:00.0000] <TabAtkins>
I agree, actually.

[12:00:00.0000] <hober>
whoa, emails to the whatwg list from a @microsoft.com address

[12:00:01.0000] <Peter->
Apparently IE9 DP3 is pending to be released today

[12:01:00.0000] <Peter->
I have noticed more Microsoft activity around various lists in the past few days

[12:02:00.0000] <gsnedders>
WHY THE HELL WOULD YOU SAY DON'T GIVE YOUR PASSWORD TO ANYONE… THEN SEND IT IN PLAINTEXT IN AN EMAIL!?

[12:02:01.0000] <gsnedders>
Sorry.

[12:02:02.0000] <TabAtkins>
...?

[12:02:03.0000] <gsnedders>
I had a moment there.

[12:03:00.0000] <gsnedders>
Well, especially when dealing with organizations which have to have bank details, that makes me ueasy

[12:03:01.0000] <gsnedders>
*uneasy

[12:16:00.0000] <AryehGregor>
gsnedders, presumably they care about phishing but not MITM.

[12:30:00.0000] <MikeSmith>
annevk: back now & synched up the copy again

[12:30:01.0000] <MikeSmith>
thanks

[12:51:00.0000] <annevk>
/me was about to offer Mike a beer

[13:28:00.0000] <Philip`>
hober: Not just "a @microsoft.com address", but three in a day

[13:30:00.0000] <Peter->
They posted on Canvas accessibility, a WebGL related topic and some HTML5 parser bits yesterday

[13:38:00.0000] <AryehGregor>
In whatwg rather than public-html?  Remarkable.

[13:42:00.0000] <Philip`>
It's good that Microsoft is so secretive and quiet in the HTML standards groups, otherwise we couldn't have fun overanalysing their occasional utterances to try to extract patterns and trends

[13:46:00.0000] <MikeSmith>
Peter-: posted about Canvas accessibility?

[13:46:01.0000] <Dashiva^2>
/me starts selling umbrellas to protect against falling pieces of sky

[13:47:00.0000] <Peter->
MikeSmith: http://lists.w3.org/Archives/Public/public-html/2010Jun/0519.html

[13:47:01.0000] <Peter->
that's on public-html

[13:48:00.0000] <MikeSmith>
ah yeah

[13:48:01.0000] <Peter->
It would be rather odd to reply to <canvas> accessibility if they wouldn't be implementing the entire element themself..

[13:48:02.0000] <MikeSmith>
stranger things have happened

[13:49:00.0000] <Peter->
No argument there

[13:49:01.0000] <Peter->
He is posting on behalf of the company however, while other posts mostly are on a more personal base

[13:51:00.0000] <AryehGregor>
My guess is that they'd like to implement it, modulo lawyers.

[13:51:01.0000] <AryehGregor>
/me sees no other feasible explanation.

[13:52:00.0000] <zcorpan_>
w3 bugzilla reached bug 10000!

[13:52:01.0000] <Peter->
Four of the seven Apple disclosed patents related to canvas have yet to be published, according to http://www.w3.org/2004/01/pp-impl/40318/status#current-disclosures

[13:53:00.0000] <Peter->
But IANAL. According to some larger news sites DP3 is pending for today, so that might clear things up

[13:55:00.0000] <Philip`>
Peter-: I believe those patents are all the same one (which is basically a description of the whole canvas API), just in different jurisdictions

[13:58:00.0000] <Peter->
I did notice the US/HK/EP indicators, but the last four aren't exactly clear. It would indeed make sense

[14:04:00.0000] <AryehGregor>
Oh, yeah, Ben Vanik is @microsoft.com.  He also seemed to be posting totally personally, which is a first for Microsoft on HTML5 lists in my experience . . .

[14:05:00.0000] <Peter->
Sylvain Galineau seems to be posting a lot on subjects Microsoft isn't involved with, CSS mostly

[14:06:00.0000] <TabAtkins>
Well, Sylvain's been an active member of the csswg for some time.

[14:06:01.0000] <TabAtkins>
Microsoft is pretty active in the csswg.

[14:09:00.0000] <AryehGregor>
Yes, it always has been.

[14:09:01.0000] <TabAtkins>
Hmm, these posts by MS.  Are they crossposted to htmlwg or something?

[14:09:02.0000] <Philip`>
No

[14:10:00.0000] <TabAtkins>
Interesting.  I'm not seeing them for some reason, then.

[14:10:01.0000] <AryehGregor>
. . . why do the chairs seem to misannounce dates in public-html announcements at least 25% of the time?

[14:11:00.0000] <Philip`>
TabAtkins: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2010-June/026825.html, http://lists.whatwg.org/pipermail/implementors-whatwg.org/2010-June/000273.html, etc

[14:11:01.0000] <TabAtkins>
Oh, I'm not on the implementors list.  That's why.

[14:14:00.0000] <Peter->
Philip`: that last mail is interesting, "I encountered this bug in my own implementation."

[14:16:00.0000] <AryehGregor>
Why does the implementers list even exist?  It only gets like three posts a month.

[14:16:01.0000] <AryehGregor>
/me subscribes

[14:17:00.0000] <Philip`>
Peter-: It seems he's one of the select bunch of people who implement HTML5 parsers just for fun :-)

[14:17:01.0000] <Philip`>
(An earlier mail says "to avoid misinterpretation, I am not on the IE team and do not speak for the IE team's plans")

[14:18:00.0000] <Peter->
Aah, at least he's passionate about HTML then ^-^

[14:19:00.0000] <AryehGregor>
Then why is he using his official e-mail address?

[14:19:01.0000] <AryehGregor>
MS doesn't have company policies on that?

[14:35:00.0000] <gsnedders>
(But what about other MS teams?)

[14:35:01.0000] <AryehGregor>
Maybe he's doing it for the Bing team.

[14:36:00.0000] <AryehGregor>
Hmm, this page renders somewhat unreadably in Chrome but is fine in Firefox.  I wonder why?  http://www.pcworld.com/article/168658/seven_reasons_microsofts_profits_are_tanking.html

[14:37:00.0000] <AryehGregor>
(Opera matches FF)

[14:37:01.0000] <TabAtkins>
Oh god.

[14:39:00.0000] <AryehGregor>
Seems like simple floats.

[14:39:01.0000] <TabAtkins>
Yeah, it's the #mainWell element that gets big, but I can't tell why.

[14:39:02.0000] <TabAtkins>
It doesn't have an explicit width on it, and its parent, #leftColumn, is a correct width.

[14:40:00.0000] <AryehGregor>
#blogArticle #mainWell { width: auto }

[14:40:01.0000] <AryehGregor>
Removing that fixes Chrome.

[14:40:02.0000] <AryehGregor>
Well, sort of.

[14:41:00.0000] <AryehGregor>
It's too narrow.

[14:41:01.0000] <TabAtkins>
Yeah, that allows the #mainWell { width: 351px; } rule to apply.

[14:41:02.0000] <AryehGregor>
Oh.

[14:42:00.0000] <AryehGregor>
So still odd.

[14:42:01.0000] <TabAtkins>
Yes.

[14:42:02.0000] <AryehGregor>
Removing float: left fixes it?

[14:42:03.0000] <TabAtkins>
Wow, yeah.  That's... bizarre.

[14:43:00.0000] <AryehGregor>
Seems to change the padding or something in Firefox, but looks fine.

[14:43:01.0000] <AryehGregor>
Do you want to make a minimal test case, or shall I?

[14:43:02.0000] <AryehGregor>
/me wonders if this is a Chrome bug or a spec bug

[14:43:03.0000] <TabAtkins>
I'll do it, I suppose.  I've got the relevant people to show it to just a few feet away.

[14:44:00.0000] <AryehGregor>
Show me the test case when you're done so I can poke at it.

[14:44:01.0000] <TabAtkins>
kk

[14:48:00.0000] <TabAtkins>
Ah, found it.  Look in the last comment for an attempt at a bbcode url tag.

[14:48:01.0000] <TabAtkins>
It's long enough to stretch the element, and the BFC chain makes the article grow as well.

[14:49:00.0000] <TabAtkins>
So now to actually reduce and find the inconsistency.

[14:49:01.0000] <AryehGregor>
I hate it when content can stretch the whole page.

[14:56:00.0000] <TabAtkins>
Ah, and the issue is that Firefox is treating the /s in the url as break opportunities, while Chrome waits for the ? before finding one.

[14:56:01.0000] <AryehGregor>
Oh.

[14:56:02.0000] <AryehGregor>
I guess there's no good place to spec that.

[14:56:03.0000] <TabAtkins>
I... dunno.  Hm.

[14:56:04.0000] <AryehGregor>
Kind of annoying, though.  That should probably be standard.

[14:56:05.0000] <AryehGregor>
Since it can have major effects on layout like this.

[14:57:00.0000] <zcorpan_>
unicode has suggestions on breaking opportunities iirc

[14:57:01.0000] <zcorpan_>
http://unicode.org/reports/tr14/

[14:58:00.0000] <Philip`>
"The third Internet Explorer Platform Preview is available for download.  We're excited to announce the availability of several major features you've asked for, such as Canvas, HTML5 Video and Audio, SVG gradients, Web Fonts, and more."

[14:58:01.0000] <AryehGregor>
Woo-hoo!

[14:58:02.0000] <AryehGregor>
I guess still no WebM by default, though, huh?

[14:59:00.0000] <zcorpan_>
what do they mean by Web Fonts?

[14:59:01.0000] <AryehGregor>
/me assumes that's also because of lawyers, but is pretty sure he's right this time.

[14:59:02.0000] <Philip`>
zcorpan_: Maybe they mean WOFF

[14:59:03.0000] <zcorpan_>
WOFF?

[14:59:04.0000] <TabAtkins>
zcorpan_: Presumably proper support for @font-face.

[14:59:05.0000] <TabAtkins>
And/or WOFF, yeah.

[14:59:06.0000] <TabAtkins>
/me thinks IE8 had proper @font-face support.

[15:00:00.0000] <Philip`>
"This release includes support for the HTML5 <audio>, <video>, and <canvas> elements, the Web Open Font Format (WOFF), and many other new and improved capabilities."

[15:01:00.0000] <Philip`>
(Seemingly can't download it yet though - they're not very good at updating all their pages simultaneously)

[15:01:01.0000] <Philip`>
Ah, http://ie.microsoft.com/testdrive/info/ReleaseNotes/Default.html updated now

[15:02:00.0000] <Philip`>
"MP4 H.264 playback support, using hardware or software decoding. Support for WebM software is not included in this release"

[15:02:01.0000] <Philip`>
"Web font formats: EOT. WOFF. Raw installable fonts (TrueType, OpenType)"

[15:02:02.0000] <gsnedders>
msPerformance?

[15:02:03.0000] <AryehGregor>
What?

[15:02:04.0000] <AryehGregor>
TTF and OTF support?

[15:02:05.0000] <gsnedders>
window.msPerformance is what?

[15:03:00.0000] <gsnedders>
"HTML5-conformant whitespace handling" — childNodes?

[15:03:01.0000] <Philip`>
gsnedders: Maybe it's like a turbo button on old computers

[15:03:02.0000] <gsnedders>
Most of the new ES5 stuff except strict mode

[15:04:00.0000] <Philip`>
Before IE8 they wanted opt-in standards support, so maybe now they're doing opt-in performance

[15:08:00.0000] <Philip`>
http://ie.microsoft.com/testdrive/Graphics/DeepZoom/Default.html seems quite nice

[15:10:00.0000] <gsnedders>
With or without the turbo button pressed?

[15:12:00.0000] <Philip`>
I suppose turbo buttons couldn't work on modern computers because they have too many megahertz to display on the LED display, so users wouldn't know how fast their CPU was

[15:12:01.0000] <jgraham>
/me didn't realise that computers came with LCD displays until today

[15:13:00.0000] <Philip`>
It was bad enough when they exceeded 100MHz and had to add an extra two LEDs onto the left

[15:13:01.0000] <gsnedders>
jgraham: Then you tried to work out what you could do with it

[15:13:02.0000] <jgraham>
gsnedders: Well obviously

[15:13:03.0000] <jgraham>
Anyway I can't imagine his Steveness allowing such tasteless bling

[15:14:00.0000] <TabAtkins>
AryehGregor: Yeah, MS had been talking about TTF support.

[15:14:01.0000] <TabAtkins>
If it's what I've heard, you can only use TTF if none of the embedding bits are set.  That excludes most existing desktop fonts.

[15:14:02.0000] <AryehGregor>
I thought the only reason WOFF existed is because MS refused to implement TTF support.

[15:14:03.0000] <Philip`>
So if everyone supports TTF now, why are we bothering with WOFF?

[15:14:04.0000] <TabAtkins>
Wonderful question!

[15:14:05.0000] <jgraham>
I thought font foundaries liked WOFF

[15:15:00.0000] <TabAtkins>
Yeah, for some reason they do.

[15:15:01.0000] <TabAtkins>
I'm pretty sure they don't understand it.

[15:15:02.0000] <AryehGregor>
Well, WOFF has built-in compression or something, right?

[15:15:03.0000] <AryehGregor>
So there's not absolutely no reason to use it?

[15:15:04.0000] <AryehGregor>
/me should test how IE9 handles his website . . .

[15:15:05.0000] <TabAtkins>
Yes, it uses a per-table compression algo.

[15:16:00.0000] <jgraham>
YEah I'm not sure how a tiny technical barrier can turn a whole industry from "impossible" to "fine"

[15:16:01.0000] <TabAtkins>
Which gets decently better numbers than just gzipping the whole thing.

[15:16:02.0000] <jgraham>
How much better?

[15:16:03.0000] <AryehGregor>
It's good anyway, you don't have to worry about configuring your web server to gzip it.

[15:16:04.0000] <TabAtkins>
I forget.  There were comparisons with actual numbers on the font mailing list a year ago.

[15:16:05.0000] <TabAtkins>
Like 30% iirc.

[15:17:00.0000] <AryehGregor>
I'm all in favor of formats with built-in compression.

[15:17:01.0000] <AryehGregor>
Unless they're text formats to begin with, in which case it might be annoying.

[15:17:02.0000] <AryehGregor>
But fonts aren't text files, so there's basically no reason to not compress them.

[15:17:03.0000] <TabAtkins>
Yeah, but you're not losing anything with binary formats.

[15:17:04.0000] <AryehGregor>
(at the format level, I mean)

[15:17:05.0000] <jgraham>
I guess 30% might be non-negligible

[15:17:06.0000] <AryehGregor>
Even if the compression were just plain old gzip on the whole file.

[15:18:00.0000] <AryehGregor>
jgraham, you mean, when the font file blocks page rendering in some browsers?  I'd say so.

[15:18:01.0000] <jgraham>
AryehGregor: Blocking rendering seems silly

[15:18:02.0000] <AryehGregor>
Well, tell the WebKit developers that.

[15:18:03.0000] <jgraham>
AryehGregor: Sure

[15:18:04.0000] <AryehGregor>
You have some FOUC-style effect either way.

[15:19:00.0000] <jgraham>
webkit developers: blocking rendering seems silly

[15:19:01.0000] <AryehGregor>
Unless you actually block rendering of the whole page from that point on, which no one does.

[15:19:02.0000] <jgraham>
AryehGregor: Presumably fonts are highly amenable to being cached

[15:19:03.0000] <AryehGregor>
Except that a large fraction of hits have empty cache in practice.

[15:19:04.0000] <AryehGregor>
Especially if most of your traffic is from occasional visitors, which is true for a lot of sites.

[15:20:00.0000] <jgraham>
In general, but for sites that you visit with any regularity it should be fine

[15:20:01.0000] <jgraham>
But I am not disagreeing with you

[15:20:02.0000] <AryehGregor>
Do I get to add parity-ie to the whiteboard of <https://bugzilla.mozilla.org/show_bug.cgi?id=276431> now?

[15:21:00.0000] <jgraham>
(you can probably get similar savings just by using a better general-purpose compression algorithm on the whole file, since fonts tend to be rather static)

[15:22:00.0000] <jgraham>
(won't really work on the web though)

[15:22:01.0000] <AryehGregor>
/me needs to go use IE9 to check out his web page, and also the HTML5 beta on YouTube, just to see for himself that it's actually true

[15:22:02.0000] <AryehGregor>
jgraham, but that will be much slower to decompress.

[15:23:00.0000] <AryehGregor>
Oh, maybe YouTube does UA sniffing and blacklists IE9?

[15:23:01.0000] <jgraham>
AryehGregor: No idea how it compares to the format specific compression or other timescales

[15:29:00.0000] <JonathanNeal>
AryehGregor, probably not a blacklist, probably a whitelist "Supported Browsers"

[15:29:01.0000] <Philip`>
http://blogs.msdn.com/b/ie/archive/2010/06/23/html5-native-third-ie9-platform-preview-available-for-developers.aspx

[15:40:00.0000] <paul_irish>
thx Philip` for breaking the news. :D

[15:40:01.0000] <Peter->
paul_irish: didn't you see it coming? :P

[15:41:00.0000] <paul_irish>
well everyone was hoping and expecting... but official confirmation is shit hot sweetness.

[15:41:01.0000] <Philip`>
paul_irish: I didn't do anything worthy of thanks - I just saw elsewhere on IRC that it had apparently been released, and then reloaded all the pages until MS got around to uploading them :-)

[15:41:02.0000] <Peter->
I've spent quite some time last week writing a blog mock up about it, MS was just making it quite obvious

[15:42:00.0000] <Peter->
replying on the canvas accessibility thing made it real obvious imo

[15:42:01.0000] <TabAtkins>
AryehGregor: https://bugs.webkit.org/show_bug.cgi?id=41103

[15:43:00.0000] <TabAtkins>
Ironically, the bug itself demonstrates our linebreaking behavior.

[15:43:01.0000] <Philip`>
Peter-: I think they could reply on the canvas accessibility thing without having got into the details of actually implementing canvas

[15:43:02.0000] <Peter->
Philip`: In theory, yes, but I consider it unlikely as he replied on behalf of Microsoft, instead of just on a personal note

[15:43:03.0000] <Philip`>
Their questions about specific requirements in the canvas spec seem more obvious, because nobody would look in that much detail unless they were implementing it or writing tests

[15:45:00.0000] <Peter->
That too, indeed :

[15:48:00.0000] <Rik`>
83/100 on Acid3. That's not something I expected so early on IE

[15:50:00.0000] <AryehGregor>
/me confirms that HTML5 YouTube works in the Platform Preview, although it's choppy right now

[15:52:00.0000] <AryehGregor>
So, my reaction to IE9 is now even more strongly 1) awesome!, and 2) be very afraid.

[15:53:00.0000] <AryehGregor>
TabAtkins, now, where should this be specced?

[15:53:01.0000] <TabAtkins>
Apparently, in Unicode.

[15:53:02.0000] <AryehGregor>
/me doesn't think "it shouldn't be specced" is a good answer, given the large rendering discrepancy on that page

[15:53:03.0000] <AryehGregor>
/me reads the Unicode spec

[16:00:00.0000] <Hixie>
http://www.youtube.com/watch?v=91bNp7HJolE

[16:05:00.0000] <AryehGregor>
TabAtkins, looks like U+002F SOLIDUS is in the class "SY: Symbols Allowing Break After (A)", so breaks should occur after it except if it immediately precedes a digit.

[16:05:01.0000] <AryehGregor>
"URLs are now so common in regular plain text that they need to be taken into account when assigning general-purpose line breaking properties. Slash (solidus) is allowed as an additional, limited break opportunity to improve layout of Web addresses. As a side effect, some common abbreviations such as “w/o” or “A/S”, which normally would not be broken, acquire a line break opportunity. The recommendation in this case is for the layout 

[16:05:02.0000] <AryehGregor>
system not to utilize a line break opportunity allowed by SY unless the distance between it and the next line break opportunity exceeds an implementation-defined minimal distance.

[16:05:03.0000] <AryehGregor>
"

[16:05:04.0000] <AryehGregor>
Ack, too long.

[16:06:00.0000] <TabAtkins>
K, so we're not following unicode.  That's cool.

[16:06:01.0000] <TabAtkins>
FWIW, IE doesn't follow that either.  However, it puts a linebreak opportunity before [, so it still ends up wrapping that line sufficiently.

[16:06:02.0000] <AryehGregor>
The full algorithm looks quite complicated.  I suspect browsers don't implement it for some reason.

[16:07:00.0000] <AryehGregor>
For instance, it says that breaks are permitted before %, but not if it's preceded by a number, even if there's intervening whitespace.

[16:07:01.0000] <AryehGregor>
/me tests that

[16:08:00.0000] <Philip`>
http://canvex.lazyilluminati.com/83b/play.xhtml - yay, Canvex sort of works in IE9, if I disable the SVG <object>s (otherwise it doesn't seem to load the wall textures) and do s/const/var/ in the JS

[16:08:01.0000] <AryehGregor>
Firefox 3.7, Chrome dev, and Opera 10.60 all ignore that.

[16:10:00.0000] <Philip`>
Still drops down to ~15fps in Ultra size when moving, though :-(

[16:11:00.0000] <AryehGregor>
/me can't figure out how [ is supposed to be handled

[16:11:01.0000] <AryehGregor>
It looks like "(AL | NU) × OP" means "do not break between an alphanumeric character and an opening punctuation character".

[16:17:00.0000] <AryehGregor>
Hmm, it seems like the prose contradicts the actual algorithm?

[16:17:01.0000] <TabAtkins>
That's not surprising.

[16:17:02.0000] <AryehGregor>
Oh, no, I see.

[16:17:03.0000] <AryehGregor>
It's covered by the last rule, "Break everywhere else."

[16:18:00.0000] <AryehGregor>
It actually looks to be very thorough and well-considered.

[16:18:01.0000] <AryehGregor>
The algorithm, I mean.

[16:18:02.0000] <TabAtkins>
That's cool.

[16:18:03.0000] <AryehGregor>
Although as with some other Unicode algorithms, it concedes that you may have to tailor it for best results.

[16:19:00.0000] <zcorpan_>
ok now i have ie9 preview 3 up and running

[16:19:01.0000] <zcorpan_>
doesn't seem to support <video controls>

[16:20:00.0000] <zcorpan_>
but does support <audio controls>

[16:21:00.0000] <zcorpan_>
hmm. only respects width/height on video if both are supported, and stretches the video instead of letterboxing

[16:22:00.0000] <zcorpan_>
ah, does support <video controls>, but they don't appear if the video is too small

[16:22:01.0000] <zcorpan_>
it seems they did fix the fallback bug i reported

[16:30:00.0000] <zcorpan_>
Philip`: i've run it through your canvas tests. do you want me to click 'submit results'?

[16:31:00.0000] <Philip`>
zcorpan_: No - that'll just give a 404, I think

[16:31:01.0000] <TabAtkins>
a;lskjdf;aslkdh;akl I cant' install preview 3.  And i couldn't install preview 2 either.  I uninstall preview 1.9, install preview 3, and it just gives me 1.9 back again.

[16:31:02.0000] <f1lt3r>
IE9 get's Canvas huh? Is this the end of SVG?! ...just kidding ;)  http://bit.ly/aSIwRu

[16:32:00.0000] <Philip`>
zcorpan_: If you copy-and-paste the test output from the textarea thing then I could probably stick it into the results table fairly easily

[16:32:01.0000] <zcorpan_>
i wonder how i go back in ie9 preview

[16:33:00.0000] <TabAtkins>
zcorpan_: You use my computer, apparently.

[16:33:01.0000] <Philip`>
zcorpan_: I'm not sure that you can

[16:34:00.0000] <zcorpan_>
sigh

[16:34:01.0000] <zcorpan_>
/me runs the tests again

[16:39:00.0000] <Philip`>
Can it cope with lots of tests on a single page?

[16:40:00.0000] <zcorpan_>
yeah

[16:40:01.0000] <zcorpan_>
is the 'generate report' button supposed to navigate to a different page?

[16:40:02.0000] <zcorpan_>
i don't get a chance to copy the textarea

[16:40:03.0000] <Philip`>
No, it should just fill in the textarea

[16:41:00.0000] <zcorpan_>
ie9 navigates...

[16:41:01.0000] <Philip`>
IE8 with excanvas doesn't, unless I'm horribly mistaken

[16:41:02.0000] <zcorpan_>
why the ******* is there no back functionality??

[16:41:03.0000] <zcorpan_>
or stop?

[16:42:00.0000] <Philip`>
<button type="button" onclick="document.getElementById('report').value = genreport()">Generate test report</button>

[16:42:01.0000] <Philip`>
I assume I shouldn't need a return false in there, or anything like that?

[16:44:00.0000] <zcorpan_>
you shouldn't... but it seems ie9 wants to submit the form there

[16:47:00.0000] <zcorpan_>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/518 (ie9 also screwed up the save function and gave "http://software.hixie.ch:80utilities/js/live-dom-viewer/saved/518")

[16:47:01.0000] <AryehGregor>
zcorpan_, I think the idea is so no one uses it as their normal browser and then hates on them for its missing features.

[16:47:02.0000] <AryehGregor>
They seem to go a little overboard, though.

[16:48:00.0000] <Hixie>
does http://software.hixie.ch/utilities/js/live-dom-viewer/saved/519 show usemap="" working on <canvas> in any browser? I can't reproduce chaals' claim that all the browsers support it already.

[16:48:01.0000] <Hixie>
am I just doing it wrong?

[16:59:00.0000] <zcorpan_>
Philip`: ok i edited the DOM and could now copy the textarea


2010-06-24
[17:21:00.0000] <hober>
Hixie: works in chrome

[17:22:00.0000] <Hixie>
what do you see exactly?

[17:22:01.0000] <hober>
when I click on the cats I land on: You have reached this web page by typing "example.com", "example.net", or "example.org" into your web browser.\n\nThese domain names are reserved for use in documentation and are not available for registration. See RFC 2606, Section 3.

[17:23:00.0000] <Hixie>
sure but is the <canvas> also showing the image map?

[17:23:01.0000] <Hixie>
i know the <img> supports it

[17:23:02.0000] <hober>
oh, sorry, didn't notice the <img>

[17:24:00.0000] <hober>
the <canvas> is the blank area between the ellipses?

[17:24:01.0000] <Hixie>
yeah

[17:24:02.0000] <hober>
nothing happens when I click that.

[17:24:03.0000] <hober>
so yeah, I don't know what chaals was on about there

[17:28:00.0000] <Hixie>
ok

[17:28:01.0000] <Hixie>
/me goes back to his MUD for now

[17:28:02.0000] <sicking>
man, i can't believe no-one took the w3c bugzilla bug 10000 before I did

[17:29:00.0000] <sicking>
it was sitting there wide open for a long time

[17:29:01.0000] <TabAtkins>
I can't believe that Julian got bug #9001.  >_<

[17:30:00.0000] <sicking>
/me doesn't see the significance of #9001

[17:30:01.0000] <TabAtkins>
IT'S OVER NINE THOUSAAAAAAAAAAAAAAAAAAAAND

[17:31:00.0000] <sicking>
riiiiiiight

[17:31:01.0000] <sicking>
/me feels like he is still lacking context

[17:31:02.0000] <Hixie>
y'all have issues

[17:31:03.0000] <TabAtkins>
http://www.youtube.com/watch?v=SiMHTK15Pik

[17:31:04.0000] <TabAtkins>
^^^ context

[17:31:05.0000] <TabAtkins>
Well, hm.  No, that's not it.  That' sjust the source.  One sec.

[17:32:00.0000] <sicking>
urban dictionary is my friend

[17:32:01.0000] <sicking>
also google

[17:32:02.0000] <TabAtkins>
That works too.

[17:32:03.0000] <sicking>
fortunately i've been successful in avoiding Dragon Ball Z

[17:33:00.0000] <TabAtkins>
over9k doesn't really depend on dbz.  It's... a thing all its own.

[17:35:00.0000] <sicking>
/me also notes that 10000 also is over 9000

[17:36:00.0000] <sicking>
i know, i know

[17:58:00.0000] <MikeSmith>
sicking: did anybody tell you yet about the prize for filing bug but 10000?

[18:00:00.0000] <sicking>
MikeSmith: no?! What do I win? A brand new car? $10000000 from dead Nigerian prince? A years supply of viagra?

[18:01:00.0000] <MikeSmith>
you won editorship of the WebIDL spec

[18:01:01.0000] <MikeSmith>
so, congratulations!

[18:01:02.0000] <MikeSmith>
/me pats sicking on the back

[18:01:03.0000] <sicking>
hahaha

[18:01:04.0000] <sicking>
/me runs

[18:01:05.0000] <MikeSmith>
heh

[18:02:00.0000] <MikeSmith>
maybe we should rename that spec Hot Potato

[18:03:00.0000] <sicking>
Or FileAPI2

[18:05:00.0000] <MikeSmith>
or XBL3...

[19:17:00.0000] <boblet>
re: CSS I was told UAs can’t drop vendor prefixes until the relevant spec goes CR. Given border-radius on Opera even though CSS3 Backgrounds is still Working Draft, that’s wrong, no?

[19:18:00.0000] <TabAtkins>
Backgrounds went CR for a short time.

[19:18:01.0000] <TabAtkins>
And then we pulled it back to make some useful changes.

[19:19:00.0000] <TabAtkins>
It should be back to Last Call any time now, too.

[19:19:01.0000] <boblet>
TabAtkins: aaah, that explains it. so regarding adding non-prefixed versions to ‘future-proof’ (like for transitions), is that bad? (given the spec could change etc)

[19:22:00.0000] <TabAtkins>
Yes and no.  It's good, and it's bad, and it's good and bad to do the opposite.

[19:22:01.0000] <boblet>
lol

[19:22:02.0000] <boblet>
thanks for your definitive answer :)

[19:22:03.0000] <TabAtkins>
If you put in unprefixed versions, figuring that they won't change, you create some pressure to not change it, since you have deployed content that won't be compatible.

[19:22:04.0000] <boblet>
(it certainly covered all the bases)

[19:23:00.0000] <TabAtkins>
But if you don't, then you make it more painful to drop prefixed versions, since you might not update your site to use the unprefixed one.

[19:23:01.0000] <TabAtkins>
So, shrug?

[19:23:02.0000] <TabAtkins>
I use unprefixed, but only if I'm using the property in a simple enough manner that I'm confident it won't change.

[19:24:00.0000] <boblet>
yeah, I personally work on the expectation that most specs won’t change significantly, and if there is a major change it’ll be something I notice and then go fix

[19:25:00.0000] <erlehmann>
i use prefixes. after all, that is what they are for.

[19:25:01.0000] <boblet>
there’s no site (that I know of) tracking what browsers are or have been using prefixes, so it’s hard to keep on top of adding unprefixed versions in

[19:25:02.0000] <erlehmann>
boblet, progressive degradation or how it is called ;)

[19:25:03.0000] <erlehmann>
make your site work without fancy backgrounds, problem solved

[19:26:00.0000] <boblet>
erlehmann: graceful degredation ;-)

[19:26:01.0000] <boblet>
but the unprefixed versions could be seen as progressive enhancement (for the glorious future of unprefixed support)

[19:27:00.0000] <MikeSmithX>
erlehmann: what if the purpose of your site is two demonstrate fancy backgrounds?

[19:28:00.0000] <erlehmann>
i would always settle for one fancy background, comrade.

[19:44:00.0000] <MikeSmith>
you sound like a possible appreciator of the "wabi sabi" aesthetic

[19:45:00.0000] <MikeSmith>
which, btw and unfortunately, most Japanese websites demonstrate completely the opposite of the wabi-sabi aesthetic

[19:45:01.0000] <MikeSmith>
with some exceptions

[19:45:02.0000] <MikeSmith>
like the sites that boblet creates in Japan

[19:45:03.0000] <MikeSmith>
which are much more wabi and sabi

[19:46:00.0000] <erlehmann>
wat

[19:46:01.0000] <boblet>
MikeSmith: i like to think of them as more wasabi

[19:47:00.0000] <MikeSmith>
boblet: please make em more like yuzu koshou

[19:48:00.0000] <boblet>
MikeSmith: ooh man! I’m so addicted to that stuff atm. although it prolly should be called yuzu shio huh

[19:49:00.0000] <MikeSmith>
dunno

[19:50:00.0000] <MikeSmith>
it's got a lot of pepper in there

[19:50:01.0000] <MikeSmith>
of some kind

[19:50:02.0000] <boblet>
the one I’ve got is more salty than peppery

[19:50:03.0000] <boblet>
regardless it’s awesome

[19:55:00.0000] <MikeSmith>
/me found http://en.wikipedia.org/wiki/Yuzukoshō

[19:56:00.0000] <MikeSmith>
http://ja.wikipedia.org/wiki/柚子胡椒

[19:57:00.0000] <MikeSmith>
pathetic -- dinnet know kanji for 胡椒 til now

[19:59:00.0000] <MikeSmith>
Hixie: please add a Previous Versions link to source of http://dev.w3.org/html5/spec/

[19:59:01.0000] <MikeSmith>
for http://www.w3.org/TR/2010/WD-html5-20100304/

[20:12:00.0000] <boblet>
/me wants a yuzukosho kitkat!

[20:32:00.0000] <variable>
I'm looking for more stuff to add to the rationale document - if anyone has any ideas please let me know

[20:32:01.0000] <variable>
I'm following the mailing list but there isn't that much to go on atm

[20:43:00.0000] <MikeSmith>
variable: going back through the commit-watchers list and checking commit descriptions might help some

[20:43:01.0000] <MikeSmith>
http://lists.whatwg.org/pipermail/commit-watchers-whatwg.org/

[20:44:00.0000] <MikeSmith>
...or it might also not help so much

[20:45:00.0000] <MikeSmith>
most of the commit descriptions don't provide much rationale

[20:45:01.0000] <variable>
MikeSmith, I'm looking for recent contecious decisions that have some kind of trail on the mailing list

[20:45:02.0000] <MikeSmith>
just recent ones?

[20:46:00.0000] <variable>
*contentious

[20:46:01.0000] <variable>
MikeSmith, older ones are OK as well 

[20:46:02.0000] <variable>
anything that is contentious and needs to be documented ;)

[20:46:03.0000] <MikeSmith>
for some recent ones, reviewing W3C bugzilla might help

[20:47:00.0000] <MikeSmith>
http://www.w3.org/Bugs/Public/buglist.cgi?cmdtype=runnamed&namedcmd=HTML%20WG%20issues

[20:47:01.0000] <MikeSmith>
and for older ones, going back through the whatwg archives and checking Hixie messages

[20:48:00.0000] <MikeSmith>
to find ones where he indicates that he's made a change to the spec

[20:48:01.0000] <variable>
MikeSmith, anything specific ?

[20:48:02.0000] <MikeSmith>
variable: nothing specific

[20:48:03.0000] <MikeSmith>
http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/

[20:48:04.0000] <MikeSmith>
you may find there is some pattern to the messages from Hixie

[20:49:00.0000] <MikeSmith>
something in the subject line, I dunno

[20:49:01.0000] <variable>
MikeSmith, I've been following the mailing list - looking for questions where people ask or answer why certain things are the case. But I can't go over everything -  hence fishing for more things to add 

[20:49:02.0000] <MikeSmith>
going back through Mark Pilgrim's This Week in HTML5 postings might actually be the best way

[20:50:00.0000] <variable>
hrm ok

[20:50:01.0000] <MikeSmith>
and his "What's Next" ones

[20:50:02.0000] <MikeSmith>
http://blog.whatwg.org/category/whats-next

[20:50:03.0000] <MikeSmith>
http://blog.whatwg.org/category/weekly-review

[20:51:00.0000] <MikeSmith>
Mark already did a lot of archaeology on some changes

[20:51:01.0000] <variable>
ok - I'll read throught them soon

[20:51:02.0000] <MikeSmith>
and documented them in some of those postings

[20:51:03.0000] <variable>
I'm trying to update this page: http://wiki.whatwg.org/wiki/index.php?title=Rationale 

[20:51:04.0000] <MikeSmith>
I don't know how much of that might already be reflected in the Rationale page

[20:51:05.0000] <MikeSmith>
variable: yeah, I know which page you mean

[20:52:00.0000] <variable>
" I need a ***legitimate*** login and password to continue. "  - as if you wanted an invalid username as password ;)

[20:53:00.0000] <variable>
(from the W3C bugzilla)

[20:54:00.0000] <variable>
MikeSmith, I added those to my bookmarks - I'll get to them soon - thanks

[20:58:00.0000] <MikeSmith>
cheers

[00:05:00.0000] <Hixie>
Lachy: yt?

[01:07:00.0000] <GPHemsley>
What do we think of Cufón? http://cufon.shoqolate.com/

[01:08:00.0000] <Dashiva^2>
krijnh: the logs page seems to be down, but I'm sure you know

[01:19:00.0000] <krijnh>
Dashiva^2: yeah, my IP changed again

[01:51:00.0000] <roc>
Philip`: have you run your canvas tests against IE9 dev preview yet?

[01:51:01.0000] <roc>
or maybe you're the wrong Philip

[01:52:00.0000] <zcorpan_>
/me has and has emailed the results to Philip`

[01:54:00.0000] <othermaciej>
how does it do?

[01:56:00.0000] <hsivonen>
are all the tweets saying things about the performance of IE9 still violating the license?

[02:01:00.0000] <zcorpan_>
othermaciej: i don't know, waiting for Philip` to put it in the comparison table

[02:10:00.0000] <jgraham>
Must not scream

[02:10:01.0000] <jgraham>
Must be calm in the face of ASCII Reference spam

[02:10:02.0000] <annevk>
that plugin email is pretty daunting too

[02:11:00.0000] <zcorpan_>
html4 and xml allows > in attribute values. they even allow ]]> in attribute values.

[02:11:01.0000] <annevk>
Hixie explains a teapot is being rendered and Julian suggests he did not want enhanced there, lol

[02:13:00.0000] <jgraham>
In the spirit of actually doing useful things, getting Philip`'s canvas tests in the testsuite seems high priority 

[02:13:01.0000] <roc>
yeah

[02:13:02.0000] <annevk>
when are we getting a useful test framework?

[02:14:00.0000] <jgraham>
annevk: We are getting a useful test framework?

[02:14:01.0000] <annevk>
I'd be interested in making tests for a number of things if something is in place

[02:14:02.0000] <annevk>
jgraham, yeah, so maybe not

[02:14:03.0000] <jgraham>
I mean I wrote something for general DOM tests

[02:14:04.0000] <jgraham>
it needs more fleshing out but the basic structure is there

[02:14:05.0000] <jgraham>
I don't plan to port existing testsuites to use it though

[02:16:00.0000] <annevk>
writing a new dom2html test suite should not be too hard

[02:16:01.0000] <jgraham>
annevk: http://dvcs.w3.org/hg/html/file/94563fe91cb0/tests/submission/Microsoft/common/domtestcase.js is the test harness

[02:16:02.0000] <annevk>
though admittedly HTML5 has more complex (and more useful) interfaces

[02:17:00.0000] <jgraham>
annevk: I plan to add assertions that map to WebIDL keywords

[02:18:00.0000] <annevk>
why does the W3C have its own BSD license?

[02:18:01.0000] <annevk>
is that normal?

[02:19:00.0000] <roc>
for those of you wondering about Ben Lerner (Microsoft) posting on the WHATWG list ...

[02:19:01.0000] <roc>
he's an intern

[02:19:02.0000] <roc>
working on MSR's managed-code Web browser

[02:19:03.0000] <roc>
/me talked to him quite a bit at PLDI a couple of weeks ago

[02:40:00.0000] <a-ja>
annevk: ping

[02:40:01.0000] <annevk>
yo

[02:41:00.0000] <a-ja>
annevk: you support http://html5.org/tools/web-apps-tracker ?

[02:41:01.0000] <annevk>
in a way

[02:41:02.0000] <a-ja>
something appears to be busted with last few changes

[02:41:03.0000] <a-ja>
since 5105

[02:41:04.0000] <annevk>
yeah, I wonder what is up

[02:42:00.0000] <annevk>
maybe I should run svn manually to see what comes out

[02:46:00.0000] <annevk>
I wonder what is different

[02:48:00.0000] <annevk>
nothing at all it seems :/

[02:48:01.0000] <annevk>
oh, maybe it's the caching

[02:49:00.0000] <zcorpan_>
if nothing is different, it seems appropriate for the diff to say 'no result' :)

[02:51:00.0000] <annevk>
pretty sure it's the cache now

[02:51:01.0000] <annevk>
I'll flush it and see what happens

[02:51:02.0000] <annevk>
though I suppose deleting 30GiB might take a while

[02:51:03.0000] <zcorpan_>
is the 'next' button screwing things up?

[02:52:00.0000] <annevk>
no

[02:52:01.0000] <annevk>
well, I don't think so

[02:52:02.0000] <a-ja>
you think? 

[02:52:03.0000] <zcorpan_>
i figured maybe someone clicked 'next' and it gave 'no result' and then 'no result' was cached

[02:52:04.0000] <annevk>
hmm

[02:53:00.0000] <annevk>
I don't think so

[02:53:01.0000] <annevk>
otherwise how would you explain http://html5.org/tools/web-apps-tracker?from=5093&to=5095&context=10

[02:54:00.0000] <zcorpan_>
ok

[02:55:00.0000] <annevk>
this is gonna take ages

[02:56:00.0000] <annevk>
oh great

[02:56:01.0000] <annevk>
the cache was what made some of it still work

[02:56:02.0000] <annevk>
lol

[02:56:03.0000] <annevk>
god dammit

[02:57:00.0000] <a-ja>
/me hides

[03:04:00.0000] <a-ja>
hm....working now, it seems

[03:05:00.0000] <annevk>
it seems that I will have to remove the context option as the external diff program caused the whole thing to break down

[03:05:01.0000] <annevk>
sorry hsivonen

[03:05:02.0000] <annevk>
though when I looked through the cache it seems nobody was really using that anyway

[03:09:00.0000] <a-ja>
for those keeping score...ff trunk now has support for figure/figcaption and mark

[03:10:00.0000] <annevk>
you can keep score here: http://wiki.whatwg.org/wiki/Implementations_in_Web_browsers

[03:11:00.0000] <a-ja>
haven't checked to see whether special width stuff for figures works right

[03:12:00.0000] <annevk>
thanks a-ja, all trackers appear to be fully functioning again

[03:12:01.0000] <annevk>
and we got rid of a feature in the process

[03:12:02.0000] <annevk>
always a good thing

[03:12:03.0000] <a-ja>
yeah...looks good from here,too 

[03:21:00.0000] <hsivonen>
Nice IE9 PP3 improves the DOM namespace compliance with HTML and SVG nodes

[03:23:00.0000] <hsivonen>
/me notices "The new rgba() color notation works in quirks, 7, and 8 document modes. In future versions of IE9, this syntax will only apply in the browser's most standard mode." in the relnotes

[03:24:00.0000] <zcorpan_>
i wonder if ie10 will have yet another mode

[03:25:00.0000] <annevk>
when quirks mode gets standardized the house of cards will collapse

[03:25:01.0000] <zcorpan_>
annevk: why? they can just choose to ignore the quirks mode spec

[03:26:00.0000] <annevk>
from the IE blog

[03:26:01.0000] <annevk>
"We encourage the other browser vendors to help the W3C finish the HTML5 specification by providing additional tests to the official HTML5 testing task force."

[03:26:02.0000] <annevk>
zcorpan_, it seems they haven't been able to ignore most other specs

[03:27:00.0000] <jgraham>
Yeah, we need some more work on the testsuite

[03:27:01.0000] <zcorpan_>
i wonder if the tests they submit is all tests they have

[03:27:02.0000] <jgraham>
I have not been impressed with the quality or number of the tests submitted so far

[03:28:00.0000] <jgraham>
zcorpan_: I doubt it

[03:28:01.0000] <jgraham>
I assume they vet tests that they submit to make sure they don't fail too many

[03:28:02.0000] <annevk>
so on the one hand the quality of the tests Microsoft has submitted are quite poor but on the other hand they have actually submitted tests

[03:28:03.0000] <hsivonen>
it's a pretty big shift in the Web that now the latest preview releases of all top 5 browsers can play YouTube videos without Flash

[03:28:04.0000] <jgraham>
Yeah

[03:29:00.0000] <jgraham>
annevk: ^

[03:29:01.0000] <jgraham>
annevk: Don't really know how to encourage others to submit though

[03:29:02.0000] <zcorpan_>
jgraham: lead by example?

[03:30:00.0000] <jgraham>
zcorpan_: Yeah, Opera need to do better too

[03:30:01.0000] <annevk>
jgraham, we could submit the html5lib tests and surrounding framework I suppose

[03:30:02.0000] <annevk>
jgraham, is the i

[03:30:03.0000] <annevk>
licensing compatible?

[03:31:00.0000] <jgraham>
annevk: I think so

[03:31:01.0000] <annevk>
the thing is, each time we add new tests we would then have to update a number of repositories :/

[03:31:02.0000] <jgraham>
I am happy for the W3C copy to be the official home for html5lib tests

[03:31:03.0000] <jgraham>
Assuming that they don't abuse that position

[03:32:00.0000] <MikeSmith>
roc: thanks -- I had been wondering what he might be working on

[03:32:01.0000] <MikeSmith>
hey krijnh is back

[03:32:02.0000] <jgraham>
annevk: I have a better HTML harness for the html5lib tests

[03:33:00.0000] <jgraham>
also

[03:33:01.0000] <annevk>
jgraham, that moves it from MIT to BSD

[03:33:02.0000] <drclue>
OK , I've contributed to the list and such and the only responses I've received is that it is a good idea, and in other venues the only opposition has been that since we have not done it we should not.  

[03:33:03.0000] <annevk>
jgraham, so new tests only contributed to the W3C cannot be moved back?

[03:34:00.0000] <hsivonen>
https://twitter.com/mhausenblas/status/16912067471

[03:34:01.0000] <jgraham>
Hmm, it depends if you think the BSD license is a problem or not I guess

[03:34:02.0000] <mhausenblas>
yes, hsivonen? :)

[03:35:00.0000] <drclue>
We should IMHO integrate XSL with HTML5 with the element level attributes XML/XSL and perhaps include an XML attribute

[03:35:01.0000] <jgraham>
In the short term the next thing that should be submitted is the canvas testsuite

[03:36:00.0000] <drclue>
Include an XML element I meant

[03:37:00.0000] <jgraham>
On consideration, I think it would be easiest if Philip` did that, assuming he is willing

[03:38:00.0000] <drclue>
Be it twitter,fox news , or many other venues , XML is totally out there, Having an XML and XSL attributes simply makes sense , and an XML attribute a natural extension

[03:39:00.0000] <drclue>
I could certainly fake these things in javascript , but given the volume of traffic , why make what should be in the HTML5 dialect a vendor specific javascript happening?

[03:42:00.0000] <drclue>
I know that some consider CSS , or other corners of web development not worth their time, but XML/XSL is totally worth the effort  

[03:43:00.0000] <drclue>
Anything in HTML5 that integrates other w3.org specs is worth at least vigorous debate 

[03:45:00.0000] <drclue>
Be it the weather, stocks, ebay,twitter, the daily news , or an appointment calender or hundreds of other situations , ignoring the thought of XML/XSL attributes and possibly an XML element is simply foolish

[03:46:00.0000] <annevk>
you may have missed it, but we're not big believers in this XML thing

[03:46:01.0000] <annevk>
it's sort of the reason we started doing HTML5

[03:46:02.0000] <drclue>
I know , not believing in XML (HTML4, HTML5 etc) might be an issue

[03:47:00.0000] <annevk>
having said that, there's XML and XSLT APIs available, so if people want it they can easily use them together

[03:48:00.0000] <annevk>
compared to DOM manipulation though their usage is negligible

[03:49:00.0000] <drclue>
Your so full of excuses my friend. There has never been a truly valid recognition of any standard that was not already strongly in ones wallet. I'm championing something that might not be your thought of today but actually respects the standards that would be in your wallet tomarrow.   

[03:50:00.0000] <jgraham>
YOu know I have no idea what you just said

[03:51:00.0000] <annevk>
it's not about recognition, it's about adoption

[03:51:01.0000] <drclue>
Why should I ignore a standard that saves copious bandwidth, increases productivity and profit, just because somebody might not at first glance understand it

[03:51:02.0000] <annevk>
if something is clearly in use patterns emerge and simplifications can be made

[03:52:00.0000] <annevk>
e.g. lots of sites have a date picker, so we can assume <input type=date> will be useful to them

[03:52:01.0000] <drclue>
Slavery of black people was a standard , but did not reflect the future , but I guess I should use your claim to justify same

[03:52:02.0000] <annevk>
jaja and hitler was against smoking, so being against smoking is evil

[03:53:00.0000] <annevk>
geez

[03:53:01.0000] <drclue>
input type=date seems fine to me as the input is asked often enough and could benefit from standardization 

[03:54:00.0000] <drclue>
Smoking is a personal decision and not something related to standards , but given such arguments I can see the value in your position (NOT)

[03:55:00.0000] <a-ja>
/me was wondering when Godwin would appear :)

[03:55:01.0000] <drclue>
Give me your argument against XML/XSL options!

[03:57:00.0000] <drclue>
I can start the for arguments with the many many many services offering XML outputs and that XSL options allow all them XML outputs to be integrated into any presentation

[03:58:00.0000] <drclue>
Your against arguments start with ......?

[03:58:01.0000] <hsivonen>
krijnh: logs aren't responding. maybe DNS out of sync again?

[04:05:00.0000] <Dashiva^2>
hsivonen: Yeah, he said so earlier

[04:05:01.0000] <drclue>
As I suspected, either you have no valid argument against the proposal or your argument requires so much encapsulation as to both take to much time to respond and or not even be valid 

[04:05:02.0000] <hsivonen>
Dashiva^2: OK

[04:07:00.0000] <drclue>
XML is a very common method of interaction and XSL a very free form method of presentation and really should be as a matter of w3.org interoperability be given a fair hearing  

[04:07:01.0000] <drclue>
XSL is but the flip side of CSS

[04:07:02.0000] <Dashiva^2>
They have been given a fair hearing, look at actual usage

[04:07:03.0000] <Philip`>
zcorpan_: Ping

[04:07:04.0000] <zcorpan_>
Philip`: pong

[04:08:00.0000] <jgraham>
drclue: The valid argument is "very few people use XSLT. Implementors would likely not implement more XSLT features. The functionality you want can be trivially written in javascript. Ergo your proposal is not worth standardising until several of these things change"

[04:08:01.0000] <Philip`>
zcorpan_: Could you send the IE9 results as an attachment or something? I don't know how to easily fix the linewrapping in the email

[04:09:00.0000] <drclue>
The actual usage of XSL has not been held up by it's value but rather it's implementation and standard. I could most easily site the same excuses for / against the deployment of the video element

[04:10:00.0000] <drclue>
Can you cite an excuse against  the XSL that has any more validity than one might have against the video element?

[04:11:00.0000] <Dashiva^2>
People code non-standard and non-cross-browser-compatible features without restraint if they actually want them

[04:11:01.0000] <Dashiva^2>
Video, you may have noticed, was rather popular before <video> was added

[04:12:00.0000] <drclue>
Again , can you cite a reason against the support of the w3.org standard XSL that would not apply to the video element?

[04:13:00.0000] <jgraham>
drclue: (a clue: the words "w3.org standard" don't magically win arguments here)

[04:13:01.0000] <hsivonen>
drclue: lolcats with XSL is harder than with video

[04:13:02.0000] <hsivonen>
drclue: most W3C standards are not implemented in browsers

[04:13:03.0000] <drclue>
So your argument would be like those that used bgcolor="" as opposed to those that used the CSS equivalent 

[04:14:00.0000] <drclue>
XSL is implemented in all major and most minor browsers

[04:14:01.0000] <Dashiva^2>
So there's no problem

[04:14:02.0000] <Dashiva^2>
The feature already exists

[04:14:03.0000] <drclue>
The exploitation of the feature is not implemented , but the proposed standard would make it more mainstream

[04:15:00.0000] <annevk>
if it was desired it would be mainstream, through workarounds

[04:15:01.0000] <annevk>
see video

[04:15:02.0000] <annevk>
see <input type=date>

[04:15:03.0000] <hsivonen>
drclue: if the usage of the current JS APIs for XSLT become ridiculously mainstream, then, maybe, there's a reason to provide a declatative way

[04:16:00.0000] <annevk>
see what went before Web Sockets

[04:16:01.0000] <Philip`>
Someone ought to make an XSLT Compact Syntax

[04:17:00.0000] <drclue>
Server load would be greatly reduced , developer coding would be greatly reduced. To say that because one has not added something to the spec is a reason not to add it is so lame as to cause me to choke on the laughter 

[04:17:01.0000] <zcorpan_>
drclue: http://wiki.whatwg.org/wiki/FAQ#Is_there_a_process_for_adding_new_features_to_a_specification.3F

[04:18:00.0000] <drclue>
Yippie skippy , good try , not playing here

[04:18:01.0000] <Dashiva^2>
I get the feeling drclue isn't actually reading what anyone else is writing

[04:18:02.0000] <drclue>
I read lots of things , but perhaps not always what you want me to read or in the case of reading same not always agreeing 

[04:18:03.0000] <jgraham>
It's like having mookid back

[04:19:00.0000] <Dashiva^2>
I thought they were the same person?

[04:19:01.0000] <jgraham>
Oh well that's why then

[04:19:02.0000] <drclue>
So in lue of a position you write what you do. Obviously you have no validity to your position

[04:20:00.0000] <drclue>
One should argue their position , not play junior highschool kid games

[04:21:00.0000] <annevk>
we did plenty to explain our position

[04:21:01.0000] <drclue>
Plese cite your valid arguments

[04:21:02.0000] <annevk>
please read what we wrote

[04:21:03.0000] <MikeSmith>
I miss mookid

[04:22:00.0000] <annevk>
/me -> lunch

[04:22:01.0000] <drclue>
I've been reading what you have all written and it sounds like a bunch of kids at lunch

[04:22:02.0000] <MikeSmith>
drclue: I strongly recommend you look into content negotation

[04:22:03.0000] <zcorpan_>
lol

[04:22:04.0000] <AryehGregor>
annevk, when quirks mode gets standardized, I'm betting IE9 will support it in standards mode (ironically), but keep some or all of the old modes for IE-specific code paths that still aren't updated.

[04:23:00.0000] <Philip`>
zcorpan_: http://philip.html5.org/tests/canvas/suite/tests/results.html - thanks

[04:23:01.0000] <drclue>
So you think that citing "content negotiation" is supposed to scare me?

[04:23:02.0000] <Philip`>
roc: ^

[04:23:03.0000] <roc>
thanks

[04:23:04.0000] <AryehGregor>
drclue, do you think repeated claims that we're all ignoring you and not making any arguments, rather than actually addressing the arguments, is going to persuade us?

[04:23:05.0000] <Philip`>
/me also fixed a few test bugs but didn't update the results so they don't quite match any more

[04:24:00.0000] <zcorpan_>
AryehGregor: how could they implement quirks mode in standards mode? that doesn't make any sense

[04:24:01.0000] <AryehGregor>
You can't make browser implementers implement what you want just by putting it in a spec.  If you can't convince them of your argument, there's no point in proceeding.

[04:24:02.0000] <roc>
people on my blog keep annoying me by claiming Mozilla is evil for not implementing every W3C REC

[04:24:03.0000] <drclue>
I don't think that closed minds ignoring years of w3.org efforts are easily persuaded as those ignoring such efforts need to take responsibility for their actions  which just may be impossible for them to do

[04:25:00.0000] <hsivonen>
/me points to http://spreadsheets.google.com/ccc?key=0AtnDcoh7FXfAdEFuWVlVVDZTVkFWdnRqaWFGMzNYM3c&hl=en

[04:25:01.0000] <AryehGregor>
zcorpan_, I mean that someone will spec the behavior other browsers use for choosing quirks mode, and will spec quirks mode itself, but IE will still let you explicitly opt out of that and rely on blacklists and other heuristics.

[04:26:00.0000] <AryehGregor>
drclue, then I guess you're out of luck.  Try making your own web browser, I guess.

[04:26:01.0000] <AryehGregor>
Or get a law passed.

[04:26:02.0000] <AryehGregor>
Not much else for you to do about it.

[04:26:03.0000] <drclue>
@AryehGregor:   There indeed is not much to actually implement as all the code is already in the browsers , even old IE

[04:27:00.0000] <jgraham>
roc: You don't implement Web Services Choreography Description Language?! Clear evil.

[04:27:01.0000] <drclue>
The only fight here is should we leverage what is in all browsers to the benefit and advancement of the w3.org standards or simply talk shit for the sake of doing same

[04:27:02.0000] <roc>
We've actually been removing support for W3C RECs. that makes us extra evil

[04:28:00.0000] <AryehGregor>
Which ones?

[04:28:01.0000] <jgraham>
(I guess that one is actually CR)

[04:28:02.0000] <roc>
we had some support for SOAP, WSDL and XForms

[04:28:03.0000] <roc>
all gone

[04:28:04.0000] <drclue>
The spec I suggest simply leverages code already implemented in ALL browsers

[04:29:00.0000] <roc>
(I don't recall if they ever actually shipped in a browser)

[04:29:01.0000] <MikeSmith>
^-

[04:29:02.0000] <MikeSmith>
oops

[04:29:03.0000] <MikeSmith>
fat fingered

[04:29:04.0000] <roc>
I think we've removed some XLink and XPointer stuff too but I don't remember the details

[04:30:00.0000] <roc>
not sure if they're RECs

[04:30:01.0000] <hsivonen>
I *think* SOAP shipped in Netscape circa 2002 or 2003

[04:30:02.0000] <drclue>
Even IE6 (Gasp if I can even call that a browser) includes the support as does every piece of software we would actually call a browser

[04:31:00.0000] <drclue>
All I'm asking is to map a couple of att4ributes to the existing support in those browsers

[04:32:00.0000] <hsivonen>
drclue: you need to spec, implement and test all sorts of things starting with how this feature would integrate with the event loop and the progression of the parse of the document having those attributes

[04:32:01.0000] <drclue>
Of course there will always be those that would argue against light at sunrise, but asking to leverage what is already there is simply a no brainer

[04:33:00.0000] <drclue>
@hsivonen are you trying to make me laugh?

[04:33:01.0000] <drclue>
You are either trying to make me laugh or are trying to snow me

[04:34:00.0000] <drclue>
I've been at the keyboard probably longer than you have been alive

[04:34:01.0000] <hsivonen>
Chrome thinks software.hixie.ch is written in Irish Gaelic and offers to translate

[04:34:02.0000] <hsivonen>
source says lang="en"

[04:35:00.0000] <drclue>
I'm always bothered by spec that thinks in lang-en

[04:36:00.0000] <hsivonen>
and it offers to translate to Icelandic in a Finnish Chrome running on English Windows

[04:36:01.0000] <hsivonen>
FAIL

[04:36:02.0000] <drclue>
All machines , all languages and that is just a base line

[04:37:00.0000] <AryehGregor>
hsivonen, I found that feature annoyingly unreliable.  I turned it off.

[04:37:01.0000] <drclue>
Obviously those developing something are apt to look at things through colored lenses

[04:37:02.0000] <AryehGregor>
It would offer to translate all-ASCII Apache directory listings with English file names (I forget from what).

[04:38:00.0000] <drclue>
ASCII is biased too as some languages could not even fit in such a constrained space

[04:39:00.0000] <drclue>
It's a hard biscuit to think in broad terms , but it is required to do so as otherwise ones opinions are stupid

[04:39:01.0000] <drclue>
BRB

[04:41:00.0000] <hsivonen>
curiously, Google decides what UI language to install for Chrome from the IP address of the downloader

[04:42:00.0000] <hsivonen>
as opposed to making the installer stub sniff for the UI language of the OS, for example

[04:42:01.0000] <AryehGregor>
You'd think that the latter would be more reliable . . . do they have data to the contrary?

[04:43:00.0000] <AryehGregor>
Maybe it's less reliable in some places, but more reliable in others.

[04:44:00.0000] <Philip`>
Maybe the application UI languages aren't a subset of the OS UI languages

[04:44:01.0000] <Philip`>
i.e. the OS doesn't provide enough information to select one of the application languages

[04:45:00.0000] <hsivonen>
Philip`: that's possible but Finnish is a in the set of possible UI languages for Windows, Mac OS X and the top Linux distros

[04:45:01.0000] <drclue>
There are a lot of things in spec and implementation that bother me , like CSS trying to do display:table etc , but in the end I would rather allow a spec to over reach than to under reach

[04:48:00.0000] <drclue>
CSS was always to me style while HTML was structure but it is inevitable that any particular spec is going to reach as far as it can, which as far as I've seen has had more benefits than detractions as much as such detractions might at time disturb me

[04:51:00.0000] <drclue>
My basic hope is despite these excesses that each spec will in the course of things try to advance other specs in a mutual sort of way so that the entire spec ecosystem can grow together 

[04:52:00.0000] <roc>
admirable restraint, everyone

[04:54:00.0000] <drclue>
Of course I personally like all the pluses that XML/XSL has to offer an over strained system where the USA struggles to be even number 15 in the world. I try to advance the integration of this spec if for nothing else that it would allow number 15 in the world to function better than the current number one

[04:58:00.0000] <drclue>
@roc we were restrained?  Oh ya , I did not deploy my fusion weapon , shit , I guess your right :) 

[05:12:00.0000] <annevk>
"I've been at the keyboard probably longer than you have been alive" -- I'm waiting for claims of axiomatic proofs now. This stuff is great :-)

[05:35:00.0000] <drclue>
@annevk: You have already had plenty of axiomatic proofs that XSL is worthless

[05:45:00.0000] <Dashiva^2>
No, those proofs are mostly empirical

[05:45:01.0000] <drclue>
Well, each would seem about as worthy an argument against XSL

[05:48:00.0000] <drclue>
There are many instances where I don't care for` a particular aspect of a w3.org standard , but none where I would simply because it was not my favorite argue against it in the face of it's benefits

[05:50:00.0000] <annevk>
AryehGregor, that seems pretty hard to do, but maybe to some extent

[05:50:01.0000] <Dashiva^2>
Those benefits are irrelevant until people show they care about them enough to use them

[05:52:00.0000] <drclue>
Well, I for one care about them enough to use them, and in the course of learning them consulted with scores of others who had learned them, just as when I learned CSS I lerned from many others in the face of at the time many nay sayers

[05:53:00.0000] <drclue>
There are many religions that object to vaccination , but I would be happy to disagree with them too

[05:54:00.0000] <drclue>
One could get a room full of folks to tell me that blacks should but shine shoes and I would disagree with them as well

[05:55:00.0000] <drclue>
Now if you would actually like to discuss the merits of the propositions as opposed to taking a bar stool poll , I'm more than up for that

[05:58:00.0000] <Dashiva^2>
Scores of people is not very much on internet scale

[05:59:00.0000] <drclue>
Well , are we talking the might of stagnation or the right of advancement?

[06:00:00.0000] <Dashiva^2>
We're talking about priorities, doing work on features that people actually want

[06:01:00.0000] <drclue>
Did people really WANT display:table vs element TABLE?

[06:01:01.0000] <Philip`>
We're talking about spending time developing features that are likely to give the most benefit to the most people, based on evidence of how widespread are people demonstrating a need for that feature

[06:02:00.0000] <Philip`>
Oh, like what Dashiva^2 said

[06:03:00.0000] <annevk>
display:table is sunk cost and anyway this is not the CSS WG

[06:03:01.0000] <drclue>
Well, I'm talking about the minimal investment to exploit functionality already built into every browser, to the benefit of all be they server side developers , client side developer , those paying for bandwidth etc.

[06:03:02.0000] <Philip`>
drclue: Lots of people clearly wanted table-like layouts, and there were clear disadvantages to using <table> for layout (unmaintainable markup, can't adapt to widely varied display devices, etc), so there was a problem that needed to be solved

[06:04:00.0000] <Philip`>
(display:table may or may not be an adequate solution to that problem)

[06:04:01.0000] <Dashiva^2>
The only benefit seems to be convenience

[06:05:00.0000] <Dashiva^2>
The feature set remains unchanged, and they are features people clearly do not care much about

[06:06:00.0000] <drclue>
@Philip I'm not arguing against the CSS attributes , as obviously you think they have value, but there was not a huge public demand for that feature , but rather a thought be the spec folks that it had value. Here I'm just saying that what I suggest has extreme value despite a public outcry for it's implementation. Specifications are often an act of consderation , not mob rule

[06:07:00.0000] <drclue>
If we had put everything to a bar stool vote , blacks would still be picking cotton

[06:08:00.0000] <Philip`>
drclue: The intended approach for HTML5 has not been to look for demand expressed as public outcry, but to do research into what people do in practice

[06:09:00.0000] <drclue>
XSL saves server load , saves bandwidth , affords portability , scales between the ignorant and the savvy

[06:09:01.0000] <Dashiva^2>
drclue: You are the one who wants to force other people to follow your commands

[06:09:02.0000] <Dashiva^2>
Adjust your analogies accordingly

[06:09:03.0000] <Peter`>
.. or are likely to be practicing when the features become available

[06:09:04.0000] <Philip`>
which gives a more accurate indication of what problems people experience and are having to work around in more inefficient ways and are worth solving better

[06:10:00.0000] <drclue>
Well , maybe in practice one utilized selective questioning where they only aske the opinions of those on the bar stool next to them

[06:11:00.0000] <drclue>
I'm not forcing anyone to use XSL , but rather trying to make it available in a context beyond vendor specific javascript and http header mdness

[06:13:00.0000] <drclue>
XSL while supported in every browser, has been forced to contend with an implementation madness that has thwarted its value and for which I hope to provide a reasonable avenue of normalcy 

[06:13:01.0000] <Dashiva^2>
Authors have coped with all kinds of madness. They would not be stopped if they actually wanted XSL.

[06:15:00.0000] <hsivonen>
annevk: IE9 PP3 bloated their Accept header.

[06:16:00.0000] <annevk>
I noticed that

[06:16:01.0000] <drclue>
I would not put an effort behind XSL if I did not believe it to have a very significant merit. XSL is an amazing specification both server side and client side , but alas the various implementations have made it almost impossible for the authors to deploy it. It was not simply a matter of author choice , but having to mess with apache add-type and a maze of mumbo jumbo to get browsers to properly deploy.

[06:17:00.0000] <annevk>
hsivonen, guess that might increase the amount of complaints a little, but I think we have a pretty strong case

[06:17:01.0000] <drclue>
@annevk , the only case you;ve made is to say you have one

[06:19:00.0000] <drclue>
I can say that XSL saves server load , saves bandwidth , reduces developer load and cite all day use cases for it that leverage existing sources all over the net

[06:19:01.0000] <drclue>
By simply adding an XSL and XSL attribute , I can immediately leverage many thousands of data sources

[06:20:00.0000] <annevk>
you can implement your proposal in a few lines of JavaScript as we have said several times so the arguments about load and saving bandwidth are bullshit

[06:20:01.0000] <drclue>
Fox,CNN,Twitter,Flicker, NOAA, and almost any other place of note

[06:21:00.0000] <drclue>
We could implement <video> in a few lines of javascript leveraging Flash , if all you have is that kind of argument

[06:22:00.0000] <annevk>
yes we could, but we do not want Flash

[06:22:01.0000] <annevk>
so we made <video>

[06:22:02.0000] <drclue>
Of course we would both stand up on each of our points citing the go do it in javascript thing as a bogus argument

[06:22:03.0000] <annevk>
and you still need JavaScript for <video> most of the time

[06:22:04.0000] <hsivonen>
IE blog says: "Note this video uses the HTML5 video tag (with the H.264 codec) if your browser supports it, and falls back to other methods otherwise. It’s a good example of same markup in action." I see "Install Microsoft Silverlight" but no video.

[06:22:05.0000] <Peter`>
hsivonen: in Firefox?

[06:23:00.0000] <drclue>
But of course you need far less javascript , making things more standardized and robust. Same holds for XSL

[06:23:01.0000] <annevk>
hsivonen, yeah, "same markup" is such nonsense, but I lost the will to argue it long ago

[06:23:02.0000] <hsivonen>
Peter`: in Minefield with Flash Player installed

[06:23:03.0000] <Peter`>
hsivonen: ah, then they're only supplying Silverlight fallbacks... makes sense (not)

[06:24:00.0000] <drclue>
I hate Flash too, but I really dislike arguments that hide behind such similar thinking as well

[06:24:01.0000] <annevk>
drclue, it's pretty clear people are willing to use JavaScript; if they want XSLT on the client in the way you suggest they can do that too, using some JavaScript; if they don't maybe they're not interested in XSLT

[06:24:02.0000] <Peter`>
drclue, you could start implementing the "jQuery for XSLT" in Javascript and become famous!

[06:24:03.0000] <drclue>
I could use javascript to inflict flash video too , so lets ditch the lame arguments

[06:25:00.0000] <drclue>
XSL has true legitimate value , can you at least admit that , or do you have no honesty in this discussion at all?

[06:26:00.0000] <annevk>
lol

[06:26:01.0000] <MikeSmith>
http://flashlab.com/html5/ is fun

[06:27:00.0000] <drclue>
LOL , I can too LOL at the idea that a quality w3.org specification can be given the short sheet by a wannbe specification that I also support

[06:29:00.0000] <drclue>
Why is it that one specification should disrespect another specification when the goal should be interoperability betyween all quality specs?

[06:29:01.0000] <Dashiva^2>
I think I'm going to stop feeding drclue now

[06:29:02.0000] <drclue>
HTML is after all like SVG , just more XML

[06:29:03.0000] <jgraham>
hsivonen: YT?

[06:30:00.0000] <jgraham>
I was thinking of doing something to fix the html5lib tokenizer tests with unpaired surrogates

[06:30:01.0000] <jgraham>
But I want to consult on what

[06:30:02.0000] <drclue>
XSL is a transformation syntax that allows XML to become any other thing

[06:30:03.0000] <gsnedders>
drclue: How can HTML after XML when it predates XML?

[06:31:00.0000] <jgraham>
hsivonen: Do those tests actually work with your test harness? They are a problem in the python case because the built-in json module (possibly erronously) believes that JSON forbids unpaired surrogates

[06:32:00.0000] <hsivonen>
jgraham: the surrogate tests work with the Java harness, IIRC, but not with the tokenizer

[06:32:01.0000] <drclue>
Because HTML while originating earlier was realized to be in the end XML which is the basis of a number of  of things like SVG and XSL and many others. Sometimes it takes a bit to understand that something that came first is really a derivative of something more basic 

[06:33:00.0000] <hsivonen>
jgraham: the Java tokenizer assumes its input is valid UTF-16

[06:33:01.0000] <jgraham>
hsivonen: Ah. OK

[06:33:02.0000] <MikeSmith>
"HTML5 compared to Flash is like a hat compared to an umbrella. They can be used together, but a hat won't replace an umbrella." 

[06:34:00.0000] <jgraham>
hsivonen: So what do you do with those tests?

[06:34:01.0000] <hsivonen>
jgraham: fail them

[06:34:02.0000] <drclue>
If we were to go strictly by timelines and such multicol would still be in HTML and not trying to live in CSS

[06:36:00.0000] <MikeSmith>
HTML *is* a hat - in fact, it's a whole bunch of hats: http://item.rakuten.co.jp/icefield/10008496/ and http://item.rakuten.co.jp/icefield/10001179/ and http://item.rakuten.co.jp/icefield/10001141/ and http://item.rakuten.co.jp/icefield/10001180/

[06:36:01.0000] <drclue>
Although multicol in HTML was only ever a Netscape thing

[06:36:02.0000] <hsivonen>
MikeSmith: where's the hat quote from?

[06:37:00.0000] <annevk>
hsivonen, flashlab.com

[06:37:01.0000] <MikeSmith>
hsivonen: http://flashlab.com/html5/ 

[06:37:02.0000] <annevk>
but Flash is a Marvel character, not an umbrella

[06:37:03.0000] <jgraham>
hsivonen: So if I were to move them into a seperate file and do something like double escape them (and add an extra flag to indicate that they were double escaped), would you be able to skip that file without too many problems?

[06:37:04.0000] <drclue>
Anyway , my point is that despite your favorite approach whatever that might be , having ready access to to XSL transforms , be they for XML to HTML or XML to SVG or whatever , ther eis a major value here and almost any majoe site you could think of provides data to drive it

[06:38:00.0000] <hsivonen>
MikeSmith: why is the last one Checkers instead of VR?

[06:38:01.0000] <hsivonen>
jgraham: moving them to a separate file would be great. I'd just skip the file

[06:39:00.0000] <drclue>
It matters not if it is twitter , fox , cnn , flicker, many infarstructure and database tools etc , to ignore XSL simply because it aint your favorite is not the best expression of wisdom

[06:39:01.0000] <MikeSmith>
hsivonen: because I didn't have time to build a VR demo.. my Flash skillz are limited to Checkers-level stuff

[06:39:02.0000] <jgraham>
abarth: I hope that works for you too

[06:42:00.0000] <hsivonen>
MikeSmith: are those Flash-based images supposed to be clickable to get actual demos? they aren't clickable for me

[06:44:00.0000] <jgraham>
/me doesn't have an umbrella, but has a coat with an integrated hat to replace it

[06:44:01.0000] <hsivonen>
oh. the Flash stuff doesn't work with click events in my browser. I guess flashlab.com needs a "support matrix" to use jd's terminology

[06:44:02.0000] <jgraham>
Dunno if that says more about me or the mataphor though

[06:44:03.0000] <MikeSmith>
hsivonen: dunno.. I didn't actually try the demos. I just like the words

[06:46:00.0000] <hsivonen>
oh. now I see why the links don't behave right

[06:46:01.0000] <hsivonen>
the whole page is Flash

[06:47:00.0000] <jgraham>
hsivonen: So you want me to move all the tests with invalid unicode characters?

[06:47:01.0000] <hsivonen>
jgraham: only the surrogate tests and the tests that test U+0000 next to CR

[06:47:02.0000] <jgraham>
hsivonen: OK

[06:48:00.0000] <drclue>
Flash needs but the nuclear explosion that is supposed to precede it 

[06:49:00.0000] <MikeSmith>
hsivonen: did you open a bug for the "bug when emitting tokenizer start tags" bug?

[06:49:01.0000] <hsivonen>
MikeSmith: nope

[06:50:00.0000] <MikeSmith>
oh

[06:50:01.0000] <MikeSmith>
you waiting for dude to open it?

[06:50:02.0000] <MikeSmith>
or no bug is needed?

[06:50:03.0000] <MikeSmith>
or ?

[06:50:04.0000] <hsivonen>
MikeSmith: I figured he can file a bug and if not, the email probably makes it to Hixie's mailboxes

[06:50:05.0000] <MikeSmith>
ok

[06:55:00.0000] <drclue>
Really XSL is just the flip side of CSS if one thinks about it. 

[06:57:00.0000] <jgraham>
hsivonen: It seems that the only test that is problematic for the json module is the one with input "\uD800"

[06:57:01.0000] <jgraham>
(suggests there should be more cases covered)

[06:57:02.0000] <jgraham>
Do you have a list of the other ones do you fail for this reason?

[07:00:00.0000] <drclue>
with XSL and but perhaps an onclick handler that called but one setAttribute one could have a news reader for every RSS feed on the planet

[07:01:00.0000] <drclue>
With XSL one could have but an inline onclick handler to show month , week , and day views of a calender

[07:02:00.0000] <hsivonen>
jgraham: Invalid Unicode character U+D800 and the test that tests U+0000 and carriage return next to each other

[07:03:00.0000] <drclue>
With XSL , forum programs could output perhaps 10% of what they have to do now and every forum could be totally unique even though driven by a simpler server side program

[07:04:00.0000] <jgraham>
hsivonen: Where is the latter? for some reason I can't find it

[07:04:01.0000] <drclue>
Every tabular listing from a database could share but one XSL file no matter the table involved

[07:05:00.0000] <jgraham>
ah, seen it

[07:06:00.0000] <drclue>
Widgets like the weather could be deployed with zero javascript or for that matter any scripting at all

[07:09:00.0000] <drclue>
Charts and graphs could be displayed against existing real time data , again without ever writing a line of code. That same data could be words,graphs or infographics and never require any scripts

[07:11:00.0000] <drclue>
changes to database browses often would require nothing but the change in SQL statement and no changes to code as of course with XSL there would be no code required

[07:14:00.0000] <MikeSmith>
the channel seems to have sensed some disturbance in the Force, resulting in krijnh site being busted

[07:14:01.0000] <MikeSmith>
I wonder if we can imagine the disturbance back into the cornfield

[07:14:02.0000] <MikeSmith>
/me concentrates

[07:14:03.0000] <MikeSmith>
or make it into a jack-in-the-box

[07:15:00.0000] <drclue>
More corn grows as a result of the concentration , but the field does not go away :)

[07:16:00.0000] <Dashiva^2>
MikeSmith: Maybe you can imagine the DNS updating to the new IP faster instead ;)

[07:17:00.0000] <MikeSmith>
ommm ommm ommm ommm ommm

[07:17:01.0000] <drclue>
I code a lot of javascript, and have done so since the 90's , but I do know that XSL is in many cases a much better alternative with the only issue having been the weird things one has to go through to access it in the various browsers these days. That is why I am fighting hard to have a clean way to use it implemented in HTML5

[07:18:00.0000] <MikeSmith>
rama rama hare hare

[07:19:00.0000] <drclue>
Emm Rama, did not they do everything in threes?

[07:21:00.0000] <MikeSmith>
nam myo ho renge kyo

[07:22:00.0000] <drclue>
I had a date once that chanted like that. It was amazing , after four hours of heavy sweating they were still able to chant , although it sounded much more like a whimper at that point

[07:24:00.0000] <drclue>
Of course the next day the IRS revoked their 501.c3

[07:27:00.0000] <drclue>
Yup, I would much rather that HTML5 provide a decent hook to XSL than have to continue a voodo of server side and client side code to make XSL work with those complications used as an excuse for not fixing XSL access in HTML5

[07:30:00.0000] <drclue>
The only reason XSL sucks has in reality nothing to do with XSL itself, however all them things that currently make XSL hard to deploy is why XSL is not popular. If XSL were accessible in a consistent way people would not understand how they lived without it

[07:31:00.0000] <daedb>
...and after four hours, you're still able to sit on irc and chant about the magical XSL unicorns. Amazing!

[07:32:00.0000] <drclue>
You bet your arse , I can chant about it, because XSL is amazing

[07:32:01.0000] <zcorpan_>
drclue: don't you have anything better to do today?

[07:32:02.0000] <Philip`>
I've used XSLT about once, and one of the most annoying things is that it didn't have a way to serialise XML :-(

[07:33:00.0000] <Philip`>
so I had to write (by which I mean copy-and-paste from a blog post from 2002) a partial serialiser in XSLT

[07:33:01.0000] <drclue>
Exactly what did you want to do that you could not figure out how to do in but "1" time using it?

[07:34:00.0000] <drclue>
Dis you also give up on CSS after "1" time using it?

[07:34:01.0000] <drclue>
How did you manage to learn HTML after giving up on it after using it but "1" time?

[07:35:00.0000] <Philip`>
I was generating HTML documentation of a file format based on a RelaxNG grammar, and wanted to include fragments of the grammar inside <pre> blocks in the output

[07:35:01.0000] <Philip`>
I didn't give up, I checked various documentation and mailing lists and it appeared that the only solution was to manually write an XML serialiser

[07:36:00.0000] <drclue>
I use XSL to ease the development of documentation, and it works great for me. I'm able to embed HTML , XML , whatever I like

[07:37:00.0000] <drclue>
The real problem has been the odd ways in which XSL was deployed in the browsers. It made it really hard for folks to sort out the documentation

[07:38:00.0000] <Philip`>
I like having a pattern-matching language that works cleanly over XML, but I'd like it more if there was extensive library support and if the syntax was much less verbose

[07:38:01.0000] <drclue>
One had to screw around with content-type headers that varied by browser and browser version and do so much screwing around before one could actually get to just doing XSL

[07:39:00.0000] <drclue>
The verbosity of the syntax was indeed annoying at first , but once I got the hang of it , the verbosity was no longer an issue

[07:40:00.0000] <Peter`>
You must realize that most web developers will have given up before getting the hang of it at all, because of that

[07:40:01.0000] <drclue>
Well, that would be the excuse to give up everything you know about web development

[07:41:00.0000] <drclue>
HTML sucks , CSS sucks , javascript sucks , XSL sucks , at l;east until you know them

[07:42:00.0000] <drclue>
How many times have you written the code to show a tabular list of something?

[07:43:00.0000] <MikeSmith>
the disturbance in the Force seems to be persisting

[07:43:01.0000] <drclue>
If you had to type the words a little longer but only had to do it once , might not it be worth it?

[07:43:02.0000] <MikeSmith>
I wonder if disturbances in the Force ever sleep, and if so, at what time

[07:43:03.0000] <Peter`>
Just don't switch over to the dark side, MikeSmith

[07:43:04.0000] <drclue>
@MikeSmith well , when are you going to sleep?

[07:44:00.0000] <MikeSmith>
Philip`: real men always write their own XML serializers

[07:44:01.0000] <Philip`>
MikeSmith: Real men are bozos

[07:45:00.0000] <drclue>
Both my C++ and PHP tools output dual XML and JSONized XML which on the client side always end up being run through XSL anyways

[07:45:01.0000] <MikeSmith>
o

[07:46:00.0000] <drclue>
I used to do it via other means , but I actually learned that XSL was more than worth the trip

[07:46:01.0000] <MikeSmith>
I sense that the disturbance in the Force may have tried to reach out to me a little bit ago, to try to draw me into further communion

[07:47:00.0000] <MikeSmith>
I feel a kind of "fascination with the abomination"

[07:47:01.0000] <drclue>
Mind you , at first I hated XSL , cussed and screamed and stomped my feet , but like every other language I've learned , I made it through to the other side and appreciated the worth of the effort

[07:49:00.0000] <MikeSmith>
/me wonders if anybody may perhaps have a clue what timezone the DITF might be in, and if it possibly might be close to bedtime

[07:49:01.0000] <drclue>
If XSL were made uniformly accessible I tend to believe folks would find a liking for it, as it is really a cool spec

[07:50:00.0000] <Peter`>
MikeSmith: Unfortunately in minus seven, so it might be a while

[07:51:00.0000] <MikeSmith>
Peter`: ah

[07:51:01.0000] <MikeSmith>
understood, thanks

[07:52:00.0000] <MikeSmith>
so I think I might deal with it by medicating and coming back again after a while

[07:52:01.0000] <MikeSmith>
/me looks around for his prescription from Dr. Kush

[07:53:00.0000] <drclue>
Writes MikeSmith a prescription for medical marijuana

[07:58:00.0000] <drclue>
Hers a really simple XSL I used for documenting some XML commands to an Asterisk PBX middle ware I'm writing using HTML5 and WebSockets http://asterclick.drclue.net/Assets/XML/API.xml

[08:23:00.0000] <drclue>
It is interesting that each time I lobby for the support of XSL which aims to support the good programming practice of having a separation in data ,code and presentation , that real smart folks come out of the woodwork to say I should do that presentation work in javascript. Who would have thunk it. 

[08:24:00.0000] <drclue>
If to lobby for the support of w3.org standards is a DITF , then listen to me breath in your helmet

[08:34:00.0000] <jgraham>
"In practice, no one puts ">" in attribute values." -- that sounds like a foolishly testable assertion

[08:35:00.0000] <Philip`>
I'm not sure what his point was about > being invalid in rel="...", because surely &gt; would be equally invalid (since it's not part of the expected datatype)

[08:35:01.0000] <jgraham>
I don't think he is making any sense in general

[08:36:00.0000] <jgraham>
It's not like we could change parsers to make foo="abc>" break even if it was invalid

[08:36:01.0000] <Philip`>
I don't think he's not making sense, I think he's just wrong

[08:37:00.0000] <Philip`>
Probably he's assuming that parsers only have to parse valid documents

[08:37:01.0000] <jgraham>
Well I guess the need to parse actual documents is such a fundamental part of my world view that people who don't share it slip from "wrong" to "not making sense"

[08:39:00.0000] <drclue>
Hopefully though if a parser wants to blow up on a badly formed document , it will at least leave a decent complaint :)

[08:41:00.0000] <jgraham>
/me wishes hg log interacted with the console like git log

[08:43:00.0000] <Philip`>
How does git log interact with the console?

[08:43:01.0000] <Philip`>
(Is it something that can't be emulated with "hg log|less"?)

[08:43:02.0000] <jgraham>
no, it is just like hg log | less

[08:44:00.0000] <jgraham>
but without having to remember to type |less

[08:44:01.0000] <jgraham>
and it works sensibly when writing to files and so on, as you would expect

[08:46:00.0000] <Philip`>
I remember to type |less so frequently that I often write "svn commit|less" which causes the nano UI to be displayed via less, which causes quite a mess

[08:49:00.0000] <zcorpan_>
Philip`: did you get my second email with the attachment?

[08:50:00.0000] <Philip`>
zcorpan_: I did

[08:50:01.0000] <Philip`>
zcorpan_: Did you get my link to the results page that incorporated your results?

[08:50:02.0000] <Philip`>
(on IRC)

[08:51:00.0000] <zcorpan_>
hmm no

[08:52:00.0000] <Philip`>
11:28 < Philip`> zcorpan_: http://philip.html5.org/tests/canvas/suite/tests/results.html - thanks

[08:54:00.0000] <Philip`>
(IE9 seems to do quite well, particularly given that it hasn't got globalCompositeOperation yet)

[08:56:00.0000] <zcorpan_>
yeah

[08:58:00.0000] <Philip`>
Peculiar that it parses rgb(255 0 0) as red

[08:58:01.0000] <Philip`>
but rejects all the other invalid values

[08:59:00.0000] <jgraham>
Hmm, I think lots of those unicode tokenizer tests are wrong. The spec says to convert surrogates to U+FFFD

[09:00:00.0000] <jgraham>
I wonder when that changed...

[09:13:00.0000] <zcorpan_>
ie9 supports &apos;

[09:23:00.0000] <zcorpan_>
http://simon.html5.org/sandbox/html/w3c-home-in-html5 renders correctly in ie9 (but the <code>s are tiny)

[09:24:00.0000] <zcorpan_>
seems they don't support drawImage(video)

[09:30:00.0000] <Philip`>
/me has no tests for drawImage(video), sadly

[09:31:00.0000] <zcorpan_>
annevk: have you noticed that hovering the logo on your blog in opera is ... weird?

[09:32:00.0000] <Philip`>
Is http://www.w3.org/TR/CSS/fonts.html#font-shorthand more correct than http://www.w3.org/TR/css3-fonts/#font for defining syntax?

[09:34:00.0000] <Mike^ses>
jgraham: http://html5.org/tools/web-apps-tracker?from=3870&to=3871

[09:42:00.0000] <zcorpan_>
hmm, ie9 parses the "<b>" as text in <svg><title><b>

[09:43:00.0000] <zcorpan_>
but they break out correctly for <svg><b>

[09:43:01.0000] <zcorpan_>
<svg><style><b> is like title

[09:43:02.0000] <zcorpan_>
but <svg><script><b> inserts a B element

[09:54:00.0000] <taptapir>
hey Philip`

[09:54:01.0000] <Philip`>
Hello

[09:55:00.0000] <taptapir>
so what you are suggesting is that, everytime an image is loaded they should set a flag or somethign to true, and when all of the image flags are set to true the last one should start drawing and other thigns...

[09:56:00.0000] <Philip`>
If your aim is to wait until they've all downloaded, and you want to download them all in parallel, then you'll need to do something like that

[09:56:01.0000] <taptapir>
yep. it's a webgl game so I need my assets (images and audio) loaded before anythign happens

[09:57:00.0000] <Philip`>
You could keep an array of all the images and check the img.complete property to see if they've been downloaded yet

[09:57:01.0000] <taptapir>
oh that would be more elegant fo sure

[09:57:02.0000] <Philip`>
though that's probably slow if you've got hundreds of files being loaded

[09:57:03.0000] <taptapir>
or what about hidden <img> fileds and placing an onLoad event to the body tag of the page and doing everything after that?

[09:58:00.0000] <Philip`>
You could just increment a counter and see if it's reached number_of_images

[09:58:01.0000] <Philip`>
or have a timer that polls a few times a second to see if everything's loaded yet, etc

[09:59:00.0000] <taptapir>
why would the "img.complete" method be slower?

[09:59:01.0000] <Philip`>
Yeah, you could use <body onload> too if you know all the images before the page has first finished loading

[10:00:00.0000] <Philip`>
If you have 1000 images, and each time an image loads you loop through all 1000 images to check whether they're all loaded yet, you'll be doing 10^6 checks

[10:01:00.0000] <Philip`>
(though if you have 1000 images you should stick them in a texture atlas or something so that you don't have 1000 images :-) )

[10:01:01.0000] <zcorpan_>
waiting for <body onload> seems like the cleanest solution

[10:02:00.0000] <taptapir>
Philip`: probably :)

[10:02:01.0000] <taptapir>
zcorpan_: can hidden img tags slow down the browser

[10:03:00.0000] <zcorpan_>
taptapir: new Image() also delays the load event

[10:03:01.0000] <taptapir>
zcorpan_: because in that method I could end up with 1000 hidden img tags

[10:03:02.0000] <zcorpan_>
don't need to be in the document

[10:03:03.0000] <Philip`>
I like doing it more like http://canvex.lazyilluminati.com/83/play.xhtml where it doesn't force you to wait for everything to download first :-)

[10:04:00.0000] <Philip`>
(That just polls for img.complete every frame)

[10:04:01.0000] <taptapir>
Philip`: are textures loaded dynamically every frame?

[10:05:00.0000] <Philip`>
taptapir: Not quite sure what you mean

[10:08:00.0000] <taptapir>
zcorpan_: can you elaborate it a little. I'm too new to js and dom :/ what I know is that I can use an image from the dom by querying its id. how can I query it if its not in the document ?

[10:08:01.0000] <taptapir>
I would understand if you 2 would choose to ignore me hehe :)

[10:09:00.0000] <taptapir>
Philip`: you said it polls for im.complete every frame and I thought you load the textures when they are needed

[10:09:01.0000] <taptapir>
thats what I meant by dynamic

[10:09:02.0000] <zcorpan_>
taptapir: you don't query, you create the image from script

[10:09:03.0000] <zcorpan_>
taptapir: <script> var x = new Image(); x.src = 'foo'; </script> <body onload="alert(x.complete)">

[10:10:00.0000] <taptapir>
ohh that clarifies a lot of things. somehow I thought body onload would only work for html data

[10:10:01.0000] <Philip`>
taptapir: Ah - it starts loading all of the textures immediately when you start the game, not dynamically based on visibility, and then just renders each one as the 'loading' image until the real image has been downloaded

[10:11:00.0000] <taptapir>
ok, it was real fast so I didn't noticed the loading images :)

[10:11:01.0000] <taptapir>
well that's another way. not bad actually. like a webpage :)

[10:13:00.0000] <zcorpan_>
taptapir: are you making your game multiplayer with websockets? :)

[10:13:01.0000] <taptapir>
zcorpan_: not this one. it's a single player adventure game in the spirit of Machinarium that I originally wrote in C (not complete)

[10:14:00.0000] <zcorpan_>
ok

[10:14:01.0000] <taptapir>
but I have lots of ideas about websockets and multiplayer games :)

[10:14:02.0000] <zcorpan_>
awesome :)

[10:14:03.0000] <Peter`>
Porting OpenTTD to canvas/websocket would be awesome

[10:15:00.0000] <Peter`>
but highly unlikely

[10:15:01.0000] <taptapir>
somtehing like Transformice. it's a flash game where a shaman mouse leads other mice to cheese :) there is physics involved etc..

[10:15:02.0000] <Peter`>
(fyi: open source version of Transport Tycoon Deluxe)

[10:15:03.0000] <taptapir>
Peter I think it's likely, did you see freeciv 

[10:16:00.0000] <Peter`>
Oh it certainly can be done

[10:16:01.0000] <Peter`>
but it would take an insane amount of time

[10:17:00.0000] <taptapir>
hmm. if it uses opengl I think it should be pretty straight forward to port it to webgl

[10:17:01.0000] <taptapir>
input etc. would take time though

[10:19:00.0000] <Philip`>
http://philip.html5.org/demos/canvas/spritepick/example.html - that's like half of TTD already

[10:19:01.0000] <taptapir>
I should get the hang of this async. business first. Coming from C, I'm lost!

[10:19:02.0000] <Philip`>
It just needs the buildings to be arranged a bit better, and for gameplay to be added

[10:19:03.0000] <taptapir>
Philip`: wow nice sprites.

[10:20:00.0000] <taptapir>
oww you stole them :p

[10:20:01.0000] <Philip`>
Of course :-)

[10:22:00.0000] <Philip`>
Peter`: Would it be more complex than a straightforward (but tedious) line-by-line C++->JS port?

[10:23:00.0000] <Philip`>
I guess the biggest pain is when games use lots of giant external libraries (for graphics or GUI or physics etc), because you wouldn't want to port all of them too

[10:23:01.0000] <taptapir>
so from what I understand, async means: every function is run from top to down but the VM doesn't wait for them to complete their job?

[10:23:02.0000] <taptapir>
and everything in JS works that way?

[10:23:03.0000] <Philip`>
but otherwise it might not be too hard

[10:24:00.0000] <Philip`>
taptapir: Every function call always runs to completion

[10:25:00.0000] <Philip`>
The asynchronous bit is like in 'img.onload = function () { do something }' - that's just defining a function, not executing it, and then storing it somewhere (in the image) and carrying on

[10:25:01.0000] <Philip`>
and some time after that script has finished, the browser will decide to call the function you stored in onload

[10:26:00.0000] <Philip`>
JS itself doesn't do anything asynchronous - it just registers callbacks with the browser environment

[10:26:01.0000] <Philip`>
effectively

[10:26:02.0000] <taptapir>
hmm, I started some javascript books yesterday but none of them seems to mention the async nature of it. any good books/sources you can recommend?

[10:26:03.0000] <Philip`>
/me doesn't know of any

[10:27:00.0000] <Philip`>
Are you familiar with function pointers and callbacks in C?

[10:27:01.0000] <taptapir>
yep

[10:28:00.0000] <Philip`>
It's basically the same thing, except instead of using a function pointer like "&function_that_does_stuff" you can define it inline like "function() { do stuff }"

[10:30:00.0000] <taptapir>
finally! thanks I get it a little bit better now

[10:31:00.0000] <Philip`>
(You can also do "function do_stuff() { do stuff }; ... img.onload = do_stuff;" which is equivalent to "img.onload = function() { do stuff }")

[10:31:01.0000] <taptapir>
yeah the first one is more C like :)

[10:31:02.0000] <Philip`>
(except in the latter case the function can refer to variables in its container scope)

[10:31:03.0000] <taptapir>
inline callbacks, that made sense.

[10:32:00.0000] <Philip`>
JS books should talk about closures, which are the relevant thing here

[10:32:01.0000] <taptapir>
checking...

[10:33:00.0000] <taptapir>
yes it's mentioned, I'm not there yet though. I'll pay special attention to that chapter

[10:34:00.0000] <Philip`>
I think they're a slightly strange concept when coming from C, but pretty useful once you understand them properly :-)

[10:35:00.0000] <taptapir>
got it! thanks Philip`and zcorpan_

[12:16:00.0000] <taptapir>
in JS, functions are objects right?

[12:16:01.0000] <taptapir>
howcome "document.body.insertAdjacentElement(document.createElement("div").innerHTML = "ljkghljkg");" is illegal?

[12:18:00.0000] <TabAtkins>
The return value of 'document.createElement("div").innerHTML = "ljkghljkg"' isn't a valid arg for insertAdjacentElement().

[12:18:01.0000] <TabAtkins>
This isn't jQuery, where everything returns a reasonable value for chaining.  ^_^

[12:27:00.0000] <taptapir>
arghh, right sry

[12:30:00.0000] <TabAtkins>
Was your "functions are objects" question unrelated?

[12:31:00.0000] <TabAtkins>
Because yes, they are.  And most objects are functions, so you can get a constructor.

[12:32:00.0000] <ayapo>
woah, just checked out the site hosting the irc logs. The yellow highlight function is epic awesome.

[12:34:00.0000] <taptapir>
TabAtkins: i thought the createElement().innetHTML part was the illegal one so I was assuming that createElement was not an object.

[12:34:01.0000] <TabAtkins>
Ah, that's unrelated.  createElement() returns a Node object.

[12:37:00.0000] <taptapir>
what's a node object? html tags?

[12:37:01.0000] <TabAtkins>
Yeah, it's a class that represents an element in the DOM.  Also other things, like text or comments.

[12:46:00.0000] <paul_irish>
boblet proposed using css transitions and transforms without the prefix (in addition to the prefixed variants) as the spec isnt likely to change and end users are unlikely to update their code when the spec hits CR... 

[12:46:01.0000] <paul_irish>
i can totally see that POV but i've been holding off on going unprefixed for un-cr things so far..

[12:47:00.0000] <paul_irish>
anyone here take a side on that?

[12:47:01.0000] <TabAtkins>
As long as you're willing to go back and fix your code, there's no problem.

[12:47:02.0000] <hsivonen>
yes, there is

[12:47:03.0000] <hsivonen>
implementors don't know if your site is maintained by a person who is willing to adjust

[12:47:04.0000] <paul_irish>
i can handle either case, i just want to make the right recommendation to other developers.

[12:47:05.0000] <TabAtkins>
/me wants to write a mixin spec for CSS which would solve the prefix problem without the problems of removing them entirely.

[12:48:00.0000] <paul_irish>
who are likely a lot more lazy

[12:48:01.0000] <hsivonen>
(I use unprefixed stuff myself, though.)

[12:48:02.0000] <TabAtkins>
hsivonen: But going with or without prefixes both cause problems for implementors.

[12:48:03.0000] <taptapir>
let's say I create an image in JS and point it to a PNG file in the server with _.src property, when does the browser actually starts to download it? also when the image.onload event occurs, does it mean that the download is finished or that the image is in the memory? or do I make any kind of sense :)

[12:48:04.0000] <hsivonen>
TabAtkins: indeed, I use unprefixed versions partly in protest against long-term prefixing

[12:49:00.0000] <paul_irish>
webkit animation is one i'd certainly not leave unprefixed. that seems to be a bit more contentious, yes?

[12:49:01.0000] <hsivonen>
yeah

[12:49:02.0000] <TabAtkins>
Animation's complex enough that I wouldn't unprefix it.  Too much change of *something* changing underneath you.

[12:49:03.0000] <paul_irish>
For sure. thx guys

[12:50:00.0000] <TabAtkins>
And yeah, the convo's died down somewhat, but I'm agitating for some changes to animations.

[12:50:01.0000] <Philip`>
taptapir: It could start to download in the background as soon as you set .src, though I don't know whether browsers will wait for the current script to return first

[12:51:00.0000] <Philip`>
taptapir: onload will be called some time (probably immediately) after the image is downloaded and decoded and found to not have errors (else it'll call onerror instead, I think), and you can safely use the image for drawing in canvases etc after that

[12:52:00.0000] <taptapir>
Philip`: so in this phase image is still not read to the memory right?

[12:53:00.0000] <Philip`>
taptapir: Depends on what you mean by being in memory

[12:54:00.0000] <Philip`>
taptapir: Browsers are free to do whatever optimisations they want in terms of caching data in memory, lazy decoding, etc, so you can never really know what they're doing - you just know whether or not it's safe to perform certain actions on images (like drawing them)

[12:55:00.0000] <taptapir>
I mean, is the image data transferred to a buffer in the memory? or is it simply presented by something like file pointers in C?

[12:56:00.0000] <TabAtkins>
Again, you don't know, because browsers are allowed to do what they want.  All you know is that certain actions are allowed at specific points.

[12:56:01.0000] <Philip`>
It's been downloaded from the network (so you can use it immediately), but everything else is just implementation details that you shouldn't need to know about

[12:57:00.0000] <AryehGregor>
annevk, the entire IE compatibility list thing can be fairly described as "hard to do", but IE has to do it anyway, since they have to deal with IE-only sites and IE-only code paths, so if they just implemented standards without compatibility they'd break.  (See: IE7.)

[12:57:01.0000] <taptapir>
Philip`: you know I want to download my game assets before the game starts,  what I wonder is that if I do that will it fill up the memory?

[12:57:02.0000] <taptapir>
lets say for 1000 images

[12:58:00.0000] <AryehGregor>
Everyone else tends to get the same code paths in most cases, due to "develop your site for standards-compliant browsers, then fix for IE".  So they can converge on a common behavior without causing more bugs than they prevent, usually.

[12:58:01.0000] <AryehGregor>
IE9's goal is explicitly to work with the same exact code everyone else gets, so hopefully IE10 won't need an IE9 compatibility mode.

[13:00:00.0000] <Philip`>
taptapir: That's an implementation detail :-)

[13:01:00.0000] <Philip`>
taptapir: It seems plausible that some implementations would store all loaded images as decoded RGBA in RAM, so you don't want to load huge amounts of image data and just hope the browser will do something clever with it

[13:02:00.0000] <taptapir>
Philip`: I better load each levels' assets when the time to use comes just to be safe then.

[13:02:01.0000] <Philip`>
taptapir: Depends on whether you're talking about megabytes or hundreds of megabytes or gigabytes of decoded image data

[13:03:00.0000] <Philip`>
taptapir: (If you're using them with WebGL then you have to worry about VRAM too, which is much more limited than RAM+swap)

[13:03:01.0000] <taptapir>
exactly

[13:03:02.0000] <taptapir>
I'M talking about couple hundreds of MB image data

[13:03:03.0000] <taptapir>
and about 100 MB of audio

[13:05:00.0000] <taptapir>
hmm wait, I'm not sure I can use an Image() object directly with webgl

[13:05:01.0000] <taptapir>
maybe there is a loadtexture function like in opengl to read it to VRAM

[13:05:02.0000] <airox>
hi guys

[13:06:00.0000] <airox>
Would watchPosition in the Geolocation API still report positions when the browsers (safari on iphone for example) is not active ?

[13:07:00.0000] <Philip`>
The web environment is not designed to provide much control of resource usage (since it's meant to provide portability instead, which means taking resource control away from authors and leaving it to browser developers), so if you want to push the boundaries then you'll just have to do lots of experimentation and make browser-specific assumptions

[13:07:01.0000] <AryehGregor>
Or wait till everyone supports NaCl.

[13:07:02.0000] <Philip`>
which is an irritating concept when you're used to having complete control over your environment

[13:08:00.0000] <Philip`>
You can pass an Image to glTexImage2D or something like that, I think

[13:08:01.0000] <taptapir>
Philip`: yeah I'm already irritated but the concept of playing hardware accelerated games in browser is just too good to pass

[13:09:00.0000] <taptapir>
Philip`: yes I checked it. I need to explicitly pass it to a webgl texture so I don't have to worry about browser filling the VRAM. and I can live with a couple hundreds MB of RAM usage

[13:12:00.0000] <taptapir>
I think I'll start by writing a decent game resource manager in javascript before anything else.

[13:24:00.0000] <taptapir>
is Image() object inherited form a regular node object? if so why there is an Image() object when it can also be created with .createElement("img")?

[13:44:00.0000] <AryehGregor>
What format do HTML5 tests need to be in?  It occurs to me that I wrote some tests for my Mozilla patch to enable maxlength on textarea, and could probably submit those.

[13:46:00.0000] <Philip`>
I don't think any format has really been agreed yet

[13:46:01.0000] <AryehGregor>
Well, there must be some way to run all the tests and figure out how many passed, no?

[13:47:00.0000] <Philip`>
There ought to be

[13:48:00.0000] <Philip`>
Actually, there probably shouldn't be a way to run *all* the tests

[13:48:01.0000] <Philip`>
because nobody will really want ten thousand results all at once, they'll just focus on a single feature

[13:49:00.0000] <TabAtkins>
AryehGregor: That's the part that I'm working on.

[13:49:01.0000] <AryehGregor>
Ah.

[13:50:00.0000] <AryehGregor>
When there is a standard test format, hopefully all the browsers can write tests for new features using that format and wrap it in their own format for internal use.

[13:50:01.0000] <AryehGregor>
That would be cool.

[13:50:02.0000] <TabAtkins>
Exactly.

[13:50:03.0000] <Philip`>
/me will just submit his own idiosyncratic format for now

[13:50:04.0000] <AryehGregor>
Since it seems to me that WebKit and Gecko must be writing tons of HTML5 tests for internal use, but they aren't submitting them like Microsoft is.

[13:50:05.0000] <AryehGregor>
I mean, both of them have an "all new features must have tests" policy . . .

[13:51:00.0000] <TabAtkins>
Yeah, I know our format is very ideosyncratic.

[13:51:01.0000] <AryehGregor>
Everyone's is.

[13:56:00.0000] <othermaciej>
AryehGregor: WebKit definitely has tests for HTML5 features

[13:56:01.0000] <othermaciej>
I think there would be a few steps required to make them into appropriate tests for a standards test suite:

[13:57:00.0000] <othermaciej>
1) ensure that everything tested is associated with specific conformance requirements in the spec

[13:57:01.0000] <othermaciej>
2) annotate the tests with the relevant conformance requirements

[13:57:02.0000] <othermaciej>
3) convert to whatever format the w3c wants them in

[13:57:03.0000] <AryehGregor>
Come to think of it, my maxlength tests won't work at all cross-browser, because they synthesize keyboard input.  Oops.

[13:58:00.0000] <othermaciej>
we have a lot of tests that are script-driven, and would totally work cross-browser

[13:58:01.0000] <othermaciej>
we have some that are based on dumps of the render tree structure which would need substantial reworking

[13:58:02.0000] <othermaciej>
I don't believe Apple, Google, or any of the other big players in WebKit-land have anyone currently tasked with converting WebKit regression tests for test suite use

[14:00:00.0000] <Rik`>
I think Opera contributes a lot of test, no ?

[14:02:00.0000] <othermaciej>
I believe they do in general, I am not sure how many have been contributed for HTML5 in particular

[14:08:00.0000] <annevk>
zcorpan, yeah, haven't filed a bug yet

[15:34:00.0000] <MikeSmith>
Hixie: can you please add http://www.w3.org/TR/2010/WD-html5-20100304/ to the set of Previous Version links in the spec

[15:34:01.0000] <Rik`>
anyone know a very simple module to work with WebSockets ? (php, python, nodejs, whatever)

[15:37:00.0000] <Hixie>
MikeSmith: yeah, hold on, i have to check in what i'm doing at the moment first

[15:37:01.0000] <MikeSmith>
k

[15:37:02.0000] <AryehGregor>
Why does the "Negative effects" part of the change proposal thing even exist?  Half the proposals don't seem to even use it, and in the ones that do it's usually not very convincing.

[15:40:00.0000] <MikeSmith>
Rik`: dunno how they might compare

[15:41:00.0000] <MikeSmith>
I guess the python one that the chrome dev team made was the first one

[15:41:01.0000] <MikeSmith>
but might not be the simplest to start with

[15:41:02.0000] <TabAtkins>
AryehGregor: Shrug.  Yeah, someone writing a change proposal usually isn't the best person to write up negative effects against their proposal.

[15:42:00.0000] <MikeSmith>
Rik`: http://jwebsocket.org/ looks quite active at least

[15:43:00.0000] <Hixie>
Rik`: write your own :-)

[15:43:01.0000] <Rik`>
Hixie: well, I might end up doing that but I wanted a simple one to start and then see where I can go from there

[15:44:00.0000] <MikeSmith>
http://twitter.com/_romuald/status/16914358097 "Web standards my ass : http://blog.chromium.org/2010/06/websocket-protocol-updated.html #html5"

[15:44:01.0000] <Rik`>
MikeSmith: thanks but, hum, not Java :)

[15:44:02.0000] <MikeSmith>
Rik`: yeah, I figured that would probably not end up at the top of your list :)

[15:45:00.0000] <Hixie>
Rik`: well you can play with my perl demo: http://damowmow.com/playground/demos/websocket

[15:47:00.0000] <Rik`>
Hixie: I meant something to abstract the protocol for me so that I can just implement my application

[15:51:00.0000] <Rik`>
http://github.com/miksago/node-websocket-server might be what I'm looking for

[15:54:00.0000] <MikeSmith>
Rik`: that's micheil I guess

[15:54:01.0000] <MikeSmith>
he's around here on #whatwg sometimes

[15:58:00.0000] <Hixie>
Rik`: there's not much to abstract, but ok :-)

[16:01:00.0000] <othermaciej>
hello all

[16:01:01.0000] <MikeSmith>
hei

[16:04:00.0000] <Rik`>
MikeSmith: well, next time you see him, tell him congrats, works like a charm 

[16:06:00.0000] <MikeSmith>
Rik`: cool

[16:07:00.0000] <MikeSmith>
http://www.reuters.com/article/idUSLDE65N14720100624

[16:07:01.0000] <MikeSmith>
includes "<^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Graphic on Nokia N-series losing to iPhone: here ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^>"

[16:08:00.0000] <MikeSmith>
good technique

[16:08:01.0000] <MikeSmith>
much easier than taking the trouble to actually create a graphic that illustrates whatever point you are trying to make

[16:08:02.0000] <MikeSmith>
to be fair the "here" part is a link to an actual graphic, but let's pretend for a moment that's not really there

[16:09:00.0000] <paul_irish>
TabAtkins: i dont see this announced but looks like flexible box is going into ie9 http://paulirish.com/i/2520.png

[16:09:01.0000] <Hixie>
MikeSmith: ok, i'm at a stopping point -- what was the change you wanted?

[16:09:02.0000] <MikeSmith>
to add http://www.w3.org/TR/2010/WD-html5-20100304/ to the set of Previous Version links

[16:09:03.0000] <Hixie>
ah ok

[16:09:04.0000] <Rik`>
paul_irish: wow, seriously ?

[16:09:05.0000] <TabAtkins>
paul_irish: Interesting.  That would explain my AlexMog has been bugging about flexbox so much.

[16:10:00.0000] <TabAtkins>
That image gonna stick around for a bit, paul_irish?

[16:10:01.0000] <Hixie>
man, i wish they'd implement stuff that's stable like DOM Events rather than stuff that's still in development

[16:10:02.0000] <annevk>
it's already June 25 here; did we fail to publish?

[16:11:00.0000] <paul_irish>
TabAtkins: yup. 

[16:11:01.0000] <Rik`>
Hixie: haven't they done that with the first preview ?

[16:11:02.0000] <TabAtkins>
K, just asked AlexMog about it.

[16:11:03.0000] <MikeSmith>
yeah, Alex, implement DOM Events! forget about that CSS crap

[16:11:04.0000] <MikeSmith>
annevk: I failed to publish

[16:11:05.0000] <Hixie>
Rik`: there's tons of stuff they could implement that's more stable than unfinished css3 or html5 stuff

[16:12:00.0000] <MikeSmith>
we will publish on June 25 and pretend that it's June 24

[16:12:01.0000] <Hixie>
Rik`: do they really have a bug-free implementation of dom events yet?

[16:12:02.0000] <TabAtkins>
Especially Flexbox, which is definitely changing from the current draft.

[16:12:03.0000] <annevk>
MikeSmith, kk

[16:12:04.0000] <TabAtkins>
Not sure *how much* it'll change now, but change it will.

[16:12:05.0000] <annevk>
it's prefixed so that seems cool

[16:12:06.0000] <Hixie>
MikeSmith: k checked it in

[16:12:07.0000] <annevk>
having an experimental impl is nice

[16:12:08.0000] <Hixie>
MikeSmith: anything else?

[16:12:09.0000] <TabAtkins>
Yeah, that's good at least.

[16:12:10.0000] <MikeSmith>
Hixie: thanks

[16:12:11.0000] <annevk>
what do you mean "at least"?

[16:12:12.0000] <MikeSmith>
Hixie: that should be it I believe

[16:13:00.0000] <Hixie>
MikeSmith: cool

[16:13:01.0000] <annevk>
/me doesn't see the problem

[16:13:02.0000] <Rik`>
Hixie: don't know about the bug free

[16:13:03.0000] <annevk>
anyway, naptime

[16:13:04.0000] <TabAtkins>
I didn't really mean anything.  Habit.

[16:13:05.0000] <Rik`>
Hixie: it seems to me they've implemented pretty much all the stable stuff used on the web

[16:14:00.0000] <TabAtkins>
http://www.tehrantimes.com/index_View.asp?code=221957

[16:14:01.0000] <MikeSmith>
browser projects need to find more engineers who really, really prefer to implement stable boring old stuff rather than fun and interesting new stuff

[16:14:02.0000] <MikeSmith>
that'd solve the problem

[16:14:03.0000] <paul_irish>
annevk: can i ask you about elem.style.VendorProps in pm?

[16:16:00.0000] <Hixie>
Rik`: it's the stable stuff not used on the web that they're holding back

[16:17:00.0000] <Hixie>
Rik`: my concern is that they have a habit of implementing a great breadth of standards with a very shallow depth of quality

[16:17:01.0000] <Hixie>
Rik`: and we end up having to clean up the mess of all the bugs they introduce into the platform

[16:18:00.0000] <annevk>
paul_irish, email annevk⊙oc

[16:18:01.0000] <annevk>
paul_irish, I'm going to bed

[16:18:02.0000] <annevk>
nn

[16:18:03.0000] <Hixie>
nn

[16:22:00.0000] <AryehGregor>
Hixie, are there major bugs in IE8's CSS2.1 support?

[16:23:00.0000] <Rik`>
yeah, bugs are bad. but if there were more W3C tests…

[16:23:01.0000] <AryehGregor>
Has anyone reported big interoperability problems with IE9's stuff, either?  I imagine a major goal would be "get all the big canvas/video/etc. demos working flawlessly".

[16:23:02.0000] <TabAtkins>
Well, it passes the CSS2.1 test suite better than anyone else, iirc.

[16:23:03.0000] <AryehGregor>
It looks to me like Microsoft has been doing a good job of standards support since IE8.

[16:23:04.0000] <AryehGregor>
Arguably since IE7, if you account for the fact that that was crippled by not throwing out IE6's engine and writing a CSS engine from scratch.

[16:23:05.0000] <Rik`>
AryehGregor: http://jhop.me/ie8-bugs

[16:24:00.0000] <Hixie>
AryehGregor: honestly it's been so long since i've actually done real testing, i've no idea

[16:24:01.0000] <AryehGregor>
Rik`, well, there are probably a lot more than 68 CSS2.1 bugs filed in either Mozilla's or WebKit's tracker.

[16:25:00.0000] <AryehGregor>
In practice, it seems to be just fine.

[16:25:01.0000] <AryehGregor>
Also, "Only partial implementation of CSS3 writing-mode property" <-- that doesn't look like CSS2.1 to me.

[16:27:00.0000] <Hixie>
back when i was testing as a full-time job, most "lists of bugs" for browsers for which i knew the filed bug counts were off by a factor of 100 or more

[16:27:01.0000] <Hixie>
if that ratio is still true today, then 68 bugs is equivalent to about 7000 bugs if you did real thorough testing

[16:28:00.0000] <AryehGregor>
And the probably hundreds or thousands of CSS bugs filed in other UAs' bug trackers is how many?  :)

[16:28:01.0000] <Hixie>
hundreds or thousands

[16:29:00.0000] <AryehGregor>
I'm pretty sure I've seen bzbarsky or someone say IE8 has the best CSS2.1 implementation.

[16:29:01.0000] <Hixie>
well i guess we'll see when it comes time for me to write acid4 :-P

[16:29:02.0000] <Hixie>
since acid4 is likely to be a CSS test suite

[16:29:03.0000] <TabAtkins>
And I've seen tables listing CSS2.1 test suite failures across browsers produced by arronei.  Consider the source, of course, but still.  ^_^

[16:29:04.0000] <Hixie>
how are they doing with acid3 at the moment?

[16:30:00.0000] <AryehGregor>
83 or something.

[16:30:01.0000] <AryehGregor>
They got a lot less hostile to it with PP3.

[16:31:00.0000] <AryehGregor>
Previously they played it down a lot, but now they have it displayed on their landing page and everything.

[16:31:01.0000] <AryehGregor>
I wouldn't be at all surprised if they shoot for 100 on the final release.

[16:31:02.0000] <Hixie>
their rendering still has some way to go

[16:31:03.0000] <Hixie>
but it's getting there

[16:31:04.0000] <Hixie>
(based on the screenshot on wikipedia)

[16:32:00.0000] <Hixie>
wow they do @font-face with ttf fonts?

[16:32:01.0000] <TabAtkins>
Yup, that made it in.

[16:32:02.0000] <Rik`>
ttf, otf, woff

[16:33:00.0000] <Hixie>
that's quite a nice

[16:33:01.0000] <daedb>
IE9 PP3 has the funniest Acid3 bug though, it shows fail if png files are associated with a non-default program :)

[16:33:02.0000] <Hixie>
change

[16:33:03.0000] <Hixie>
daedb: hah

[16:33:04.0000] <daedb>
They noted it in the release notes and everything.

[16:33:05.0000] <Hixie>
do they show the right favicon?

[16:34:00.0000] <Hixie>
(which is to say, no favicon or a cat, but not a red square)

[16:34:01.0000] <daedb>
There's no adress bar, so can't see favicons...

[16:34:02.0000] <Hixie>
ah

[16:39:00.0000] <Hixie>
anyone got any opinions on how WebSRT cues should parse if they don't have a valid timestamp?

[16:40:00.0000] <TabAtkins>
Drop on the floor?

[16:41:00.0000] <Hixie>
the whole cue up to the next double blank line?

[16:41:01.0000] <Hixie>
that's kinda what i figured should happen

[16:41:02.0000] <Hixie>
other options are to just ignore the one line and try reading a timestamp from the next one

[16:41:03.0000] <Hixie>
or to drop the whole file from there on

[16:42:00.0000] <Hixie>
or to assume a timestamp of 0

[16:42:01.0000] <Hixie>
or of equal to the last timestamp seen

[16:43:00.0000] <TabAtkins>
I don't think anything that assumes a timestamp would be good.

[16:43:01.0000] <TabAtkins>
Dropping the whole file is no good either, if the rest is properly formatted.

[16:43:02.0000] <TabAtkins>
Ignoring just that line and trying the next sounds potentially cool.

[16:44:00.0000] <TabAtkins>
So, one of those two.  Either skip the line, or skip everything until the next double blank line.

[16:45:00.0000] <Hixie>
i guess skip the block is safest

[16:47:00.0000] <variable>
is there any undefined or illegal code here: http://codepad.org/Ghv5oeHJ ?? 

[16:48:00.0000] <AryehGregor>
#c++?

[16:48:01.0000] <variable>
woops

[16:48:02.0000] <variable>
wrong channel sorry


2010-06-25
[17:18:00.0000] <Hixie>
hsivonen: http://trac.webkit.org/export/LATEST/trunk/LayoutTests/fast/parser/residual-style-hang.html doesn't show text in gecko

[17:23:00.0000] <roc>
AryehGregor: there's no way IE9 will hit 100 on Acid3, because they've said they're not doing SVG Animation or SVG Fonts

[17:23:01.0000] <TabAtkins>
Man, for real?

[17:24:00.0000] <roc>
for IE9 anyway

[17:24:01.0000] <TabAtkins>
I like my animated SVG...

[17:24:02.0000] <Hixie>
we're going to need to get the browsers together and decide if SVG fonts is really worth keeping, i think

[17:24:03.0000] <roc>
of course, the fact that Acid3 tests those is a huge misfeature IMHO

[17:24:04.0000] <TabAtkins>
/me doesn't know how he feels about SMIL in general, he just likes having his SVG animate itself.

[17:24:05.0000] <TabAtkins>
I don't think anyone thinks SVG fonts are really worth keeping.

[17:25:00.0000] <roc>
Hixie: Doug Schepers has said the SVG group will split out SVG Fonts into their own spec

[17:25:01.0000] <Hixie>
TabAtkins: i mean in the codebases of those who have implemented it

[17:25:02.0000] <Hixie>
roc: interesting

[17:26:00.0000] <TabAtkins>
/me is using animated SVG in an emulator for a flash game.

[17:26:01.0000] <TabAtkins>
Just for a conveyor belt, but still.

[17:26:02.0000] <roc>
Hixie: http://lists.w3.org/Archives/Public/www-svg/2010Jun/0003.html near the bottom

[17:26:03.0000] <roc>
although the whole message is relevant

[17:28:00.0000] <roc>
Hixie: regardless of the merits of SVG Fonts, I really strongly feel that SVG Fonts should not become part of the Web platform just because they snuck into Acid3

[17:29:00.0000] <Hixie>
i agree

[17:30:00.0000] <Hixie>
but i don't want to remove them without opera and webkit agreeing to remove support (or at least agreeing that the test shouldn't have them and that they might remove support), because then it just looks like i'm trying to make it easy for firefox to pass after having "tricked" webkit and opera into implementing something, or whatever

[17:30:01.0000] <roc>
well, if you read the feedback we've been getting for years, including some of the comments on one of my recent blog posts --- http://weblogs.mozillazine.org/roc/archives/2010/06/not_implementin.html#comments --- there is tremendous pressure to implement SVG Fonts because they're in Acid3

[17:30:02.0000] <othermaciej>
WebKit's SVG font support is in active use by iTunes Extras

[17:31:00.0000] <othermaciej>
and perhaps other content

[17:31:01.0000] <Hixie>
roc: i understand

[17:31:02.0000] <roc>
Hixie: removing SVG Fonts from Acid3 isn't a decision that SVG Fonts shouldn't be part of the Web platform. It just means letting them stand on their own merits

[17:31:03.0000] <othermaciej>
(yes, even though they could embed a TTF font in theory just as easily)

[17:31:04.0000] <Hixie>
roc: that's the idea of the acid tests, after all :-P

[17:32:00.0000] <Hixie>
roc: you can use that argument for everything in the test

[17:32:01.0000] <roc>
only for the new-feature stuff

[17:32:02.0000] <roc>
not for the bug-fix stuff

[17:32:03.0000] <Hixie>
hard to determine which is which

[17:32:04.0000] <Hixie>
e.g. is dom events "new-feature stuff" or "bug-fix stuff"?

[17:32:05.0000] <othermaciej>
I think it would have been better if Acid3 covered things which there was rough consensus should be part of the Web platform

[17:33:00.0000] <othermaciej>
but it's hard to evaluate what that means in a truly objective way

[17:33:01.0000] <roc>
Hixie: there has always been consensus that DOM Events should be part of the Web platform

[17:33:02.0000] <Hixie>
not from microsoft

[17:33:03.0000] <roc>
they edited the spec

[17:33:04.0000] <Hixie>
that doesn't mean anything

[17:33:05.0000] <roc>
ok

[17:34:00.0000] <roc>
did anyone ask them?

[17:34:01.0000] <Hixie>
not to my knowledge

[17:34:02.0000] <roc>
I sort of had the impression they were always "planning" to implement DOM events

[17:34:03.0000] <roc>
I don't recall them ever pushing back on it

[17:34:04.0000] <othermaciej>
I don't think they ever disputed the value or propriety of DOM events, they just stopped development for a while after implementing an early draft

[17:34:05.0000] <othermaciej>
at least that is my impression

[17:34:06.0000] <roc>
mine also

[17:34:07.0000] <Hixie>
it's hard to tell, since in my experience they push back on everything until the day they ship it

[17:35:00.0000] <Hixie>
but anyway

[17:35:01.0000] <othermaciej>
anyway, since SVG fonts are in actual use, I don't think Apple would favor removing them from WebKit, even if Acid3 removed those tests

[17:35:02.0000] <roc>
right, I wouldn't suggest removing them from Webkit

[17:35:03.0000] <othermaciej>
that being said, I personally wouldn't have any objection to dropping/replacing those tests

[17:36:00.0000] <othermaciej>
I'm responding to Hixie's suggestion that they should be removed from the test if and only if Opera and WebKit-based browsers drop support

[17:36:01.0000] <Hixie>
my recommendation would be to co-write an e-mail with someone from three of mozilla, webkit, microsoft, and opera, and send it to me asking for the fonts stuff to be removed

[17:36:02.0000] <roc>
ok

[17:36:03.0000] <Hixie>
i've no personal objection to removing them, i just don't want a PR problem

[17:36:04.0000] <roc>
you can probably keep the tests, just use a TTF font instead of an SVG font

[17:37:00.0000] <Hixie>
there's already a TTF test in there

[17:37:01.0000] <roc>
right

[17:37:02.0000] <Hixie>
the SVG font tests are testing SVG font APIs

[17:37:03.0000] <roc>
right

[17:37:04.0000] <roc>
you can test those font APIs with a TTF font

[17:37:05.0000] <Hixie>
oh you mean make new tests?

[17:37:06.0000] <Hixie>
oh

[17:37:07.0000] <othermaciej>
Microsoft refuses to implement TTF

[17:37:08.0000] <roc>
no

[17:37:09.0000] <Hixie>
wait, you want to keep those APIs?

[17:37:10.0000] <roc>
they just did

[17:37:11.0000] <Hixie>
othermaciej: they've implemented it

[17:37:12.0000] <roc>
yes

[17:37:13.0000] <othermaciej>
plain TTF?

[17:37:14.0000] <roc>
yes

[17:37:15.0000] <TabAtkins>
othermaciej: They're in preview 3.

[17:37:16.0000] <roc>
yes, I'm stunned too

[17:37:17.0000] <Hixie>
othermaciej: they pass the TTF part of Acid3, at least

[17:37:18.0000] <othermaciej>
/me wonders what the point of WOFF is then

[17:37:19.0000] <Hixie>
othermaciej: so that means they support at least one character from Ahem

[17:38:00.0000] <hober>
plain OTF too

[17:38:01.0000] <Hixie>
WOFF does indeed seem rather pointless at this point

[17:38:02.0000] <roc>
I want to blog about that and refer to their blog posts from two years ago where they said the world would explode if they implemented raw TTF

[17:38:03.0000] <roc>
but I'll let them off

[17:38:04.0000] <Hixie>
roc: post that after they ship IE9 :-)

[17:38:05.0000] <Hixie>
roc: they haven't shipped yet :-)

[17:38:06.0000] <roc>
there's still the font vendors who will ship WOFF but not TTF

[17:39:00.0000] <roc>
so we shall see

[17:39:01.0000] <roc>
but if WOFF becomes pointless, that's OK

[17:39:02.0000] <othermaciej>
I would be highly surprised if there is any font that is available in WOFF format but not TTF at all

[17:39:03.0000] <othermaciej>
though I can imagine fonts that are only legally licensed for use on the Web in WOFF format only

[17:40:00.0000] <roc>
there are --- http://www.fontshop.com/freefonts/ for example

[17:40:01.0000] <Hixie>
roc: some of these svg font tests seem pretty <svg:font>-specific

[17:40:02.0000] <roc>
Hixie: ok

[17:40:03.0000] <Hixie>
roc: maybe in addition to an e-mail, mail me a diff of what you want the tests to be if it's not just removing them wholesale :-)

[17:40:04.0000] <TabAtkins>
roc: I'd need to check details, but they're *probably* doing what I was told at TPAC, and only allowing linking of TTFs with none of the embedding bits set.

[17:41:00.0000] <roc>
othermaciej: so, would you co-sign an email "with someone from three of mozilla, webkit, microsoft, and opera, and send it to me asking for the fonts stuff to be removed"?

[17:41:01.0000] <roc>
TabAtkins: ah yes, as I recall that is exactly what they are doing

[17:41:02.0000] <TabAtkins>
Which means that almost no commercially-produced desktop fonts will work.

[17:41:03.0000] <roc>
so only "free" TTFs

[17:41:04.0000] <roc>
right

[17:41:05.0000] <roc>
I forgot about that, sorry

[17:41:06.0000] <zcorpan_>
roc: i'd prefer if the test just made supporting svg fonts optional (like attribute nodes is optional in acid3) rather than just removing the test

[17:42:00.0000] <othermaciej>
roc: depends on whether it is the sort of thing that could be construed as an official statement on behalf of Apple and/or would draw significant press attention

[17:42:01.0000] <othermaciej>
in the latter case I would probably need to get sign-off from others at Apple

[17:42:02.0000] <zcorpan_>
roc: so if you don't support svg fonts at all, you can get full score, but if you do support svg fonts, you need to pass the svg fonts test

[17:42:03.0000] <roc>
zcorpan_: ok that makes sense

[17:42:04.0000] <Hixie>
my intention would be to post a blog post announcing the change

[17:42:05.0000] <Hixie>
dunno how much attention that would get

[17:42:06.0000] <roc>
othermaciej: that's up to Hixie

[17:42:07.0000] <Hixie>
probably some

[17:43:00.0000] <roc>
probably a lot, but can't you phrase it so it's clearly not an Apple position?

[17:43:01.0000] <Hixie>
yeah i can say "i got an e-mail from roc, maciej, and (whoever)"

[17:44:00.0000] <othermaciej>
I can certainly try, however, if there is any significant chance it would result in a C|Net story citing the remarks as "Apple says X", I would have to preflight it, or else people would get really upset at me

[17:45:00.0000] <roc>
Hixie: what if you don't name names, but just say "representatives from Mozilla, Webkit and ..."?

[17:45:01.0000] <Hixie>
roc: sure, i just assumed that would be considered more formal

[17:46:00.0000] <roc>
if you say "Webkit" the trail doesn't lead back to Apple

[17:46:01.0000] <roc>
unless an enterprising reporter crawls this channel, in which case maciej is already in trouble :-)

[17:46:02.0000] <Hixie>
roc: i really don't mind either way, whatever is most convenient for you -- i just want to have some cover so i can honestly tell people it was a request from a plurality of browser vendors

[17:46:03.0000] <roc>
sure ok

[17:47:00.0000] <roc>
zcorpan_: I wonder if Opera would co-sign ... who should I ask? Hakon?

[17:48:00.0000] <Hixie>
howcome would be good

[17:48:01.0000] <zcorpan_>
roc: it seems ed wrote the test

[17:48:02.0000] <Hixie>
ed would be good too

[17:48:03.0000] <roc>
ok

[17:48:04.0000] <roc>
I'll ask them

[17:48:05.0000] <roc>
thanks

[17:48:06.0000] <Hixie>
ed would be better since howcome and i do the tests together

[17:48:07.0000] <Hixie>
though cc howcome so he's in the loop

[17:48:08.0000] <roc>
ok

[17:48:09.0000] <roc>
will do

[17:48:10.0000] <Hixie>
thanks

[17:49:00.0000] <roc>
thank you

[17:54:00.0000] <doublec>
given the reddit thread and comment's on roc's weblog I suspect the change will result in a fair amount of publicity around the "acid 3 test changed so IE and Firefox can get 100" idea

[17:55:00.0000] <roc>
if we can get all browser vendors to agree, I hope it will be OK.

[17:55:01.0000] <Hixie>
doublec: yeah that's why i want the cover and webkit/opera's approval :-)

[17:56:00.0000] <doublec>
Hixie, yeah I thought I'd just mention it in case someone thought the change would go quietly :)

[17:57:00.0000] <doublec>
nothing ever seems to happen quietly in the web world

[17:57:01.0000] <Hixie>
all the attention is making the web better, i think

[17:57:02.0000] <Hixie>
i'd rather have all this attention than the crickets and tumbleweed of ten years ago

[17:58:00.0000] <doublec>
true

[18:00:00.0000] <roc>
yeah

[18:01:00.0000] <roc>
this is actually a great time

[19:03:00.0000] <MikeSmith>
a question

[19:04:00.0000] <TabAtkins>
an answer

[19:04:01.0000] <MikeSmith>
heh

[19:04:02.0000] <MikeSmith>
why is it that WebGL needs TypedArray but canvas 2D context does not?

[19:05:00.0000] <MikeSmith>
what's different about that 3D api that requires binary support?

[19:05:01.0000] <Philip`>
The only large arrays in the 2D context are pixel arrays, which are just arrays of octets and therefore easy

[19:05:02.0000] <Philip`>
3D needs large arrays of floats (vertex coordinates), shorts (vertex array indices), etc

[19:06:00.0000] <Philip`>
and large arrays of different types interleaved

[19:06:01.0000] <Philip`>
(so each array element can encode all the data for a vertex)

[19:08:00.0000] <othermaciej>
yeah, ImageData is in fact a typed array

[19:08:01.0000] <othermaciej>
it's just that you don't need more than that

[19:27:00.0000] <MikeSmith>
Philip`, othermaciej - I see, thanks

[19:28:00.0000] <MikeSmith>
I think it would be good for vlad to consider adding that as informative background to the Introduction section of the TypedArray spec

[19:31:00.0000] <Hixie>
how should i detect if a file is an SRT file: by a whitelist of MIME types, by searching for --> on the second non-blank line, by looking for a magic string (which would work for WebSRT but would mean plain SRT files need a trivial modification to be used), some combination thereof, or by some other mechanism?

[19:33:00.0000] <MikeSmith>
le

[19:33:01.0000] <MikeSmith>
it would seem like something that did not rely on sniffing would be preferable

[19:33:02.0000] <MikeSmith>
so, whitelist of MIME types

[19:33:03.0000] <Hixie>
that's still a kind of sniffing

[19:34:00.0000] <MikeSmith>
OK, so that type

[19:34:01.0000] <Hixie>
k

[19:34:02.0000] <Hixie>
(the problem with "sniffing" is not that it's in-band, but that it's not interoperable)

[19:34:03.0000] <Hixie>
(this would be interoperable)

[19:35:00.0000] <Hixie>
if there's no MIME type, do I assume WebSRT or do I bail?

[19:38:00.0000] <MikeSmith>
playing devil's advocate for architectural-purity/sanity advocates: I guess I would suggest that you bail

[19:38:01.0000] <Hixie>
well there are quite practical reasons to bail too

[19:38:02.0000] <Hixie>
i just don't know if they overwhelm the practical reasons not to bail

[19:38:03.0000] <Hixie>
in practice, most resources have a mime type

[19:38:04.0000] <Hixie>
it's rare that none be given

[19:39:00.0000] <Hixie>
(it's usually text/plain or text/html, if it's wrong)

[19:40:00.0000] <othermaciej>
I think any fallback behavior is ok as long as it's defined

[19:40:01.0000] <othermaciej>
whether that is hard fail or some kind of sniffing

[19:40:02.0000] <MikeSmith>
yeah but I guess they are not so rare enough to have not cause serious problems in the past

[19:40:03.0000] <othermaciej>
the interop problem is mainly if you don't define it

[19:40:04.0000] <Hixie>
indeed

[19:40:05.0000] <MikeSmith>
e.g., like that Apache default being to treat everything as text/plain even it if doesn't have a MIME type

[19:40:06.0000] <othermaciej>
which is why so much video still gets served as text/plain

[19:41:00.0000] <othermaciej>
the more kinds of resources get sniffed, the harder it is to verify that the full set of sniffing algorithms does not have holes

[19:41:01.0000] <othermaciej>
therefore I would advocate not doing sniffing for new kinds of formats if there isn't a deep need for it

[19:41:02.0000] <othermaciej>
(and by "holes" I mean opportunity to make things get sniffed multiple ways to create a security hole)

[19:42:00.0000] <Hixie>
makes sense

[19:42:01.0000] <Hixie>
i guess

[19:42:02.0000] <Hixie>
i dunno

[19:42:03.0000] <Hixie>
i kinda like unambiguously self-describing formats

[19:43:00.0000] <MikeSmith>
in general for new specs it seems like the most sane thing for interop is being as unforgiving as possible

[19:43:01.0000] <MikeSmith>
and keep the option open of adding forgiveness later

[19:47:00.0000] <MikeSmith>
I wonder if the Typed Arrays spec draft is the first to ever include a section explicitly titled with "Bikeshedding"

[19:47:01.0000] <MikeSmith>
regardless, I like it. So I hope it starts a trend

[19:54:00.0000] <MikeSmith>
so does it seem like browser implementors are going to be satisfied with the Typed Arrays spec as a binary-data API in general (rather than just specifically for the WebGL case)?

[19:55:00.0000] <Hixie>
bbiab

[21:00:00.0000] <othermaciej>
heh

[21:00:01.0000] <othermaciej>
(re "Bikeshedding

[21:00:02.0000] <othermaciej>
" section)

[21:31:00.0000] <MikeSmithX>
othermaciej: so does it seem to you like the Typed Arrays spec is sufficient as a binary-data API in general (rather than just specifically for the WebGL case)?

[21:31:01.0000] <othermaciej>
I have not reviewed it closely enough to know

[21:32:00.0000] <othermaciej>
(yet)

[21:32:01.0000] <MikeSmithX>
OK

[21:32:02.0000] <MikeSmith>
I know the other proposals at http://wiki.commonjs.org/wiki/Binary have already had a lot of discussion

[21:33:00.0000] <MikeSmith>
I'm kinda wondering if those will continue to be considered/necessary if browsers end up implementing the Typed Arrays spec

[21:34:00.0000] <othermaciej>
I would like to see typed arrays be the one thing but I have not had time to study close

[21:34:01.0000] <MikeSmith>
OK

[21:34:02.0000] <MikeSmith>
so get to reviewing it, man

[21:35:00.0000] <MikeSmith>
it's not like you have a bunch more important stuff to do

[01:12:00.0000] <annevk>
I do think it would be nice if we had a signature for SRT

[01:12:01.0000] <annevk>
But that would mean subtitle files cannot be exchanged in an interoperable way until players are updated, but given how fast WebM is adopted it might be easy to patch existing players...

[01:13:00.0000] <annevk>
Especially given how trivial SRT is

[01:18:00.0000] <annevk>
http://people.opera.com/danield/html5/vuvuzela/ lol

[01:48:00.0000] <MikeSmith>
hey, I find some signs of good news in my mail for a change

[01:49:00.0000] <MikeSmith>
Jonas message about Notifications

[04:09:00.0000] <annevk>
hsivonen, filing bugs on removing crypto strength and language from the UA strenth on Opera now

[04:09:01.0000] <annevk>
hsivonen, will let you know what the outcome is

[05:24:00.0000] <annevk>
rhetoric, rhetoric, rhetoric, blah, blah, blah

[05:25:00.0000] <MikeSmith>
/me thinks he knows which message annevk is referring to, but will keep his mouth shut

[06:21:00.0000] <Philip`>
Lachy: You ought to fix your mail client since it thinks you're called Lachln

[06:30:00.0000] <Lachy>
thanks

[06:48:00.0000] <gsnedders>
hsivonen: yt?

[06:49:00.0000] <gsnedders>
hsivonen: In the "In foreign content" insertion mode, under the "A start tag whose tag name is one: [Long list of HTML element]" step, it seems to reprocess the token in the same state and go into an infinite loop

[06:49:01.0000] <gsnedders>
hsivonen: What did you do? Change to the secondary insertion mode?

[12:48:00.0000] <TabAtkins>
Yay, IE9 pp3 *almost* correctly displays my SVG-and-javascript emulator for a flash game.

[12:48:01.0000] <TabAtkins>
It just messes up one of the SVGs for some reason, and won't animate them yet.

[12:49:00.0000] <TabAtkins>
Those are possibly related - the messed up one is the animated one.

[12:56:00.0000] <cardona507>
good ole IE9

[13:20:00.0000] <MikeSmith>
for those interested in e-books, and/or specifically in the ePub format -

[13:20:01.0000] <MikeSmith>
http://www.idpf.org/idpf_groups/IDPF-EPUB-WG-Charter-4-27-2010.html

[13:20:02.0000] <MikeSmith>
see the Preliminary Solutions section

[13:21:00.0000] <MikeSmith>
where it lists "Adopt HTML5 Features"

[13:22:00.0000] <MikeSmith>
also see "Current Industry Problems" the list of problems that they are trying to address with this 2.1 update to ePub

[13:22:01.0000] <MikeSmith>
e.g., "Need for enhanced global language support"

[13:23:00.0000] <MikeSmith>
and specific mention of "including but not limited to special line-breaking rules and vertical writing direction"

[13:23:01.0000] <Hixie>
"it is acknowledged that a number of significant issues would need to be resolved" i wonder what those are

[13:24:00.0000] <MikeSmith>
dunno

[13:24:01.0000] <MikeSmith>
but we can try to find out

[13:25:00.0000] <MikeSmith>
they seem to be taking a fairly pragmatic approach to this update

[13:25:01.0000] <MikeSmith>
so far

[13:25:02.0000] <MikeSmith>
e.g., see the statement, "Since most distributed Reading Systems utilize Web browsers to present content to end users, and many standalone Reading Systems utilize Web browser rendering technologies in their implementations, it is desirable to improve alignment with Web standards as implemented by modern browsers."

[13:26:00.0000] <MikeSmith>
"as implemented"

[13:26:01.0000] <MikeSmith>
as opposed to, say, "as we would have preferred them to have been implemented"

[13:27:00.0000] <MikeSmith>
or "as not implemented yet but we will spec things out on the basis of blind faith"

[13:28:00.0000] <MikeSmith>
keep in mind that ePub 2.0 is normatively based on XHTML 1.1

[13:28:01.0000] <Hixie>
they seem to have done a reasonable job overall of being pragmatic within the constraints of not being experts in web tech

[13:28:02.0000] <MikeSmith>
yeah

[13:29:00.0000] <MikeSmith>
and switching away from XHTML 1.1 to "as implemented in modern browsers" is a great move to start with, at least

[13:30:00.0000] <MikeSmith>
anyway, vertical text support is extremely important for e-book-reading use cases in Japan

[13:30:01.0000] <MikeSmith>
I am told it is for Taiwan also

[13:30:02.0000] <MikeSmith>
though less so for mainland China

[13:31:00.0000] <Hixie>
yup

[13:31:01.0000] <othermaciej>
Apple's version of ePub will (as I'm sure has been previously mentioned) allow arbitrary HTML5 constructs

[13:31:02.0000] <othermaciej>
and people have in fact used it to embed video and audio and such in books

[13:31:03.0000] <othermaciej>
which is neat

[13:32:00.0000] <othermaciej>
from what I hear, there are a number of language features important for books in east asian locales

[13:32:01.0000] <othermaciej>
and people are pickier about some of them for books than for web sites

[13:32:02.0000] <othermaciej>
if Apple ends up doing any of them of course we'll feed the work into WebKit so it will be there for websites too

[13:33:00.0000] <othermaciej>
but I think most of the east asian language issues are mostly CSS level questions

[13:34:00.0000] <Hixie>
the only two vertical text things i'm aware of for html5 are to do with canvas drawText and captions

[13:34:01.0000] <MikeSmith>
othermaciej: yeah, I think for east-asian language/layout issues, it is mostly about CSS-level questions

[13:34:02.0000] <Hixie>
in canvas' case it's actually already specced, i just commented it out since it depends on CSS spec text that isn't yet written

[13:34:03.0000] <MikeSmith>
othermaciej: and I will say that from what I see, people in Japan are definitely much pickier about language features for books than they are for websites

[13:35:00.0000] <TabAtkins>
We're arguing about vertical-text stuff in the CSSWG right now anyway, spurred mostly by epub.

[13:35:01.0000] <Hixie>
in the case of subtitles it'll be similarly commented out once i'm done unless css has caught up, but it's already designed in

[13:35:02.0000] <MikeSmith>
e.g., almost all books -- paperback books, mass-market everyday reading stuff -- uses vertical text

[13:36:00.0000] <MikeSmith>
TabAtkins: good to hear that the discussion is going on in the CSS WG

[13:36:01.0000] <MikeSmith>
TabAtkins: is Murakami-san from Antenna House in the gropu?

[13:36:02.0000] <TabAtkins>
Yeah.

[13:36:03.0000] <MikeSmith>
ah

[13:36:04.0000] <Hixie>
the vertical text stuff was being discussed in the csswg 2000, too

[13:36:05.0000] <Hixie>
when i joined the group

[13:36:06.0000] <TabAtkins>
He's an editor of one of the relevant specs, actually.

[13:36:07.0000] <MikeSmith>
ok

[13:36:08.0000] <Hixie>
we even went to CR once!

[13:36:09.0000] <TabAtkins>
Yeah, the consensus is still sorta *shrug*.

[13:36:10.0000] <Hixie>
(and then took it back when fantasai didn't like it)

[13:38:00.0000] <MikeSmith>
well, Murakami-san can argue with some vehemence, for sure. But his an implementor. Not a browser implementor but anyway an implementor of a fairly complex and powerful application

[13:39:00.0000] <MikeSmith>
TabAtkins: I guess it's not reasonable to expect everybody in the group to see vertical text as a hugely higher priority

[13:40:00.0000] <MikeSmith>
maybe a lot of the work on driving it will end up coming from developers on the Chrome team in Japan

[13:40:01.0000] <TabAtkins>
Well, we all want it.  The issue is *how*, because CSS in general contains a lot of assumptions that depend on horizontal text layout.

[13:40:02.0000] <MikeSmith>
ah

[13:40:03.0000] <MikeSmith>
I see

[13:40:04.0000] <TabAtkins>
There's the IE approach, frex, where switching the writing mode to vertical just "rotates" everything, so that, for example, padding-top now adds padding to one of the horizontal sides.

[13:41:00.0000] <TabAtkins>
But most people in the group think that's stupid.

[13:42:00.0000] <MikeSmith>
"frex" is an odd choice of a name for that

[13:42:01.0000] <TabAtkins>
frex = for example

[13:42:02.0000] <MikeSmith>
ah

[13:42:03.0000] <MikeSmith>
heh

[13:42:04.0000] <MikeSmith>
OK

[13:42:05.0000] <TabAtkins>
I'm inconsistent in my usage of that word.

[13:42:06.0000] <MikeSmith>
didn't grok the abbreviation at first

[13:42:07.0000] <TabAtkins>
I picked it up from some message board years ago and it's stuck in my head.

[13:43:00.0000] <MikeSmith>
just struck me in this case because in Japanese it's pronounced same as "flex"

[13:43:01.0000] <TabAtkins>
Ah.

[13:43:02.0000] <MikeSmith>
no difference in translitering "r" and "l"

[13:44:00.0000] <TabAtkins>
Yeah.

[13:44:01.0000] <MikeSmith>
which is why I thought those freaking horn things were called "bubuzera"

[13:44:02.0000] <MikeSmith>
because that's what Japanese TV calls them

[13:44:03.0000] <TabAtkins>
Hahaha.

[13:44:04.0000] <MikeSmith>
oh, no difference often between "b" and "v" too

[13:44:05.0000] <TabAtkins>
The correct term is "vulvazookeepers".

[13:45:00.0000] <MikeSmith>
heh

[13:45:01.0000] <TabAtkins>
So, back to technical talk.

[13:46:00.0000] <TabAtkins>
Another possible solution is to introduce logical-direction equivalents of every physical-direction property, so you can do padding-before, frex, and have it mean padding-top in horizontal writing and padding-left or whatever in vertical writing.

[13:47:00.0000] <TabAtkins>
This would work "automatically", but only in relatively limited circumstances where you're not adjusting the layout based on writing mode.

[13:47:01.0000] <TabAtkins>
It also doesn't solve related problems, like the fact that in japanese vertical writing an underline goes on the opposite side of what you'd expect.

[13:48:00.0000] <TabAtkins>
So a third possibility is to have some sort of selector or media query or whatnot that distinguishes between when a document/element is in vertical or horizontal mode and lets you apply arbitrary CSS to it.

[14:06:00.0000] <MikeSmith>
TabAtkins: would seem like a good use for media queries

[14:07:00.0000] <TabAtkins>
There are complications, but yeah, probably.

[14:43:00.0000] <jgraham>
Random fact of the day: whenever I hear "Let Down" by Radiohead I hear the start skipping and jumping in my head, just like it does on my CD. It is quite disorienting when it plays normally

[14:44:00.0000] <gsnedders>
doh

[14:44:01.0000] <gsnedders>
(a deer)

[14:44:02.0000] <TabAtkins>
a female deer?

[14:44:03.0000] <jgraham>
gsnedders: You are supposed to be out partying

[14:45:00.0000] <gsnedders>
TabAtkins: Sorry, but the fact that myself and a colleague have had that in our heads for the past two months (almost) and sung it numerous times around jgraham is obviously missed on you.

[14:46:00.0000] <gsnedders>
jgraham: Sorry, I chose sanity.

[14:46:01.0000] <TabAtkins>
I was continuing the song!

[14:46:02.0000] <jgraham>
I wonder if we could convince the ePub people to support microdata

[14:46:03.0000] <gsnedders>
TabAtkins: You were questioning it, though.

[14:47:00.0000] <gsnedders>
(I may have been punched by jgraham for singing that song.)

[14:47:01.0000] <gsnedders>
(Multiple times.)

[14:47:02.0000] <gsnedders>
(He's a big bully.)

[14:47:03.0000] <jgraham>
gsnedders: At leat I don't hit girls

[14:47:04.0000] <jgraham>
*least

[14:48:00.0000] <gsnedders>
jgraham: …normally.

[14:48:01.0000] <jgraham>
Umm

[14:48:02.0000] <jgraham>
Ever, really

[14:48:03.0000] <gsnedders>
(There again, I don't normally either. I just poke them.)

[14:48:04.0000] <jgraham>
The innuendo levels are rising

[14:49:00.0000] <gsnedders>
It turns out poking girls makes them jump and scream.

[14:50:00.0000] <MikeSmith>
poke

[14:50:01.0000] <gsnedders>
/me jumps and screams

[14:50:02.0000] <jgraham>
I think my innuendo meter just exploded

[14:50:03.0000] <MikeSmith>
heh

[14:56:00.0000] <MikeSmith>
so a read an article that says "HTML5 is to HTML4 as Werner Herzog's 'The Bad Lieutenant: Port of Call - New Orleans' is to Abel Ferrara's 'Bad Lieutenant' -- and HTML5 is to Flash as Herzog's 'The Bad Lieutenant: Port of Call - New Orleans' is to 'Sex and the City 2' and its underlying franchise.'

[14:56:01.0000] <MikeSmith>
(whatever all that means)

[14:57:00.0000] <TabAtkins>
That's the most densely pretentious statement I've ever heard referring to HTML5.

[14:57:01.0000] <jgraham>
TabAtkins: You should get out more

[14:58:00.0000] <TabAtkins>
/me has no clue where he could get out to where he could hear more densely pretentious things, nor does he understand why he would *want* to.

[14:59:00.0000] <MikeSmith>
it could be that I just made up that entire line on my own

[14:59:01.0000] <MikeSmith>
or it could also be that I did actually read it somewhere

[14:59:02.0000] <jgraham>
TabAtkins: It is not that pretentious really. I mean they are all films

[14:59:03.0000] <jgraham>
and one of them is Sex and the City

[15:00:00.0000] <Hixie>
is it bad that i didn't understand that metaphor at all

[15:00:01.0000] <TabAtkins>
I suppose I have just never heard of the first two films, nor their directors(?).

[15:00:02.0000] <jgraham>
I don't understand it at all

[15:00:03.0000] <gsnedders>
/me estimates that almost nobody here has seen Sex and the City

[15:00:04.0000] <jgraham>
But wikipedia helps

[15:02:00.0000] <AryehGregor>
/me doesn't understand it at all.

[15:02:01.0000] <gsnedders>
/me doesn't ether

[15:02:02.0000] <gsnedders>
*either

[15:04:00.0000] <jgraham>
The later "Bad Lieutenant" film apparently borrows from the earlier one, whilst the director claims never to have seen the earlier one. The director of the earlier one is apparently very angry about the existence of the later one

[15:05:00.0000] <jgraham>
It you believe wikipedia, the earlier one has much more sex than the later one

[15:06:00.0000] <jgraham>
So I guess the point of the metaphor is that HTML5 is less likely to get you laid

[15:07:00.0000] <Hixie>
less likely than HTML4?

[15:07:01.0000] <Hixie>
man if HTML4 was supposed to get me laid i've been doign it wrong

[15:07:02.0000] <jgraham>
Anyway, it is clearly all wrong

[15:08:00.0000] <jgraham>
HTML4 is like a Kafka novel

[15:08:01.0000] <jgraham>
HTML5 is lik Joyce

[15:09:00.0000] <jgraham>
Flash is like the Very Hungry Catepillar

[15:10:00.0000] <MikeSmith>
jgraham: that description of the relationship between the two Bad Lieutenant films sounds somewhat apt in terms of comparison

[15:10:01.0000] <MikeSmith>
I guess

[15:10:02.0000] <MikeSmith>
I did see the first Bad Lieutenant film

[15:10:03.0000] <MikeSmith>
it is a really awful movie

[15:12:00.0000] <MikeSmith>
so at least it's nice to hear that Abel Ferrara is angry about Herzog making a film with the same name without telling him ahead of time

[15:13:00.0000] <MikeSmith>
Ferrara deserves to be made angry in return for foisting on the world the crappiness which is that film

[15:14:00.0000] <MikeSmith>
and even better to hear that Herzog never actually even watched his movie

[15:16:00.0000] <MikeSmith>
cool to see that abarth is starting work on the webkit html5 tree builder

[15:16:01.0000] <abarth>
MikeSmith: it's just a sketch at this point

[15:16:02.0000] <abarth>
MikeSmith: i'm trying a new control structure now

[15:16:03.0000] <MikeSmith>
I see

[15:16:04.0000] <abarth>
the problem is the way the spec is written is hard to implement directly

[15:17:00.0000] <abarth>
lots of "process blah as if foo"

[15:17:01.0000] <abarth>
also, as written, you need to make a ton of jump tables and branches for each character

[15:17:02.0000] <abarth>
which is really slow

[15:18:00.0000] <abarth>
instead of having the first branch be the insertion mode, i'm trying with the first branch being what token we're dealing with

[15:18:01.0000] <abarth>
which should push the character loop below all the jumping/branching

[15:20:00.0000] <MikeSmith>
interesting

[15:21:00.0000] <MikeSmith>
abarth: is that also in part because it'll make it easier for you to reuse the existing code?

[15:22:00.0000] <abarth>
MikeSmith: that design is closer to the way the old code works

[15:22:01.0000] <MikeSmith>
ok

[15:22:02.0000] <abarth>
the old code has a hash table keyed off the kind of token

[15:22:03.0000] <MikeSmith>
ah

[15:22:04.0000] <abarth>
and then a jump through a function pointer

[15:22:05.0000] <abarth>
to figure out what to do with the token

[15:23:00.0000] <TabAtkins>
Where's the html5 test suite?  (Or at least as much of it as exists?)

[15:24:00.0000] <abarth>
TabAtkins: we have a bunch of parser tests

[15:24:01.0000] <abarth>
TabAtkins: that we stole from henri and then added to

[15:25:00.0000] <MikeSmith>
TabAtkins: did you mean the parser test suite specifically? or HTML5 in general?

[15:26:00.0000] <paul_irish>
HTML5 in general

[15:29:00.0000] <paul_irish>
(I was just asking Tab, which prompted his question).. basically looking for whatever test suites exist that cover what's in HTML5 and friends

[15:29:01.0000] <AryehGregor>
/me discovers that his spam folder has like a billion non-spam things, wtf

[15:29:02.0000] <TabAtkins>
That's why I skim and clear my spam regularly.

[15:29:03.0000] <TabAtkins>
once or twice a day keeps the number below 50 at a time, so I can see it all on one page.

[15:29:04.0000] <AryehGregor>
I guess I need to do that too.

[15:30:00.0000] <AryehGregor>
Most of mine all goes to one address that I don't use for much, and I have those labeled, so the legit stuff is pretty easy to spot.

[15:30:01.0000] <AryehGregor>
I got suspicious when the first post of Doug Scheper's post to the WHATWG was missing.

[15:31:00.0000] <TabAtkins>
Set your filters so that they never spam anything you've labelled.

[15:31:01.0000] <TabAtkins>
That eliminated 99% of my false positives.

[15:31:02.0000] <AryehGregor>
I thought it was funny that I didn't get any commit mail the last week or so.

[15:31:03.0000] <AryehGregor>
Like 90% of my spam is labeled.

[15:31:04.0000] <TabAtkins>
Oh, heh.

[15:32:00.0000] <AryehGregor>
It mostly comes from an address I use that has no spam filtering.

[15:32:01.0000] <TabAtkins>
Well, set *specific labels* to be never spammed.

[15:32:02.0000] <AryehGregor>
But I should set most of my filters to do that, yeah.

[15:32:03.0000] <Philip`>
TabAtkins: There's loads of wiki pages with lists of existing test suites

[15:32:04.0000] <TabAtkins>
Philip`: I have no clue what any of them are.

[15:33:00.0000] <Philip`>
TabAtkins: http://wiki.whatwg.org/wiki/Testsuite#Existing_tests

[15:33:01.0000] <Hixie>
TabAtkins: my spam folder has 9540 e-mails in it, and that's just over the last 30 days. no way i can check that twice a day. :-P

[15:33:02.0000] <TabAtkins>
paul_irish: ^^^

[15:33:03.0000] <Philip`>
TabAtkins: http://www.w3.org/html/wg/wiki/Existing_Html5_Tests

[15:33:04.0000] <Philip`>
TabAtkins: etc

[15:34:00.0000] <TabAtkins>
Excellent, thanks.

[15:34:01.0000] <paul_irish>
Thx Philip`

[15:40:00.0000] <AryehGregor>
/me doesn't understand why it's so hard to reliably tell apart legitimate mailing lists from spam.

[15:41:00.0000] <TabAtkins>
For the longest time any mailing-list email from a google employee was getting spammed.

[15:41:01.0000] <TabAtkins>
(That's when I set all my label filters to never spam.)

[15:41:02.0000] <AryehGregor>
I was getting that too.

[15:41:03.0000] <AryehGregor>
Maybe because of SPF or something?

[15:42:00.0000] <TabAtkins>
Or the lack of it.  I just always found it funny that google.com addresses, presumably sent from inside the company, were spammed by gmail's filter.

[15:42:01.0000] <Hixie>
it wasn't sent from inside the company

[15:42:02.0000] <Hixie>
that's the problem

[15:42:03.0000] <AryehGregor>
I'm glad I found out about this mass spammification less than 30 days after it started.

[15:42:04.0000] <Hixie>
it was sent by the mailing list software

[15:43:00.0000] <AryehGregor>
Wait, if you have SPF, does that mean you can't send to mailing lists?  :/

[15:43:01.0000] <Hixie>
i don't know the details

[15:44:00.0000] <AryehGregor>
Actually, I think it'd only be if the Return-To was set to your address.

[15:44:01.0000] <Hixie>
i think it's actually Return-Path that matters

[15:44:02.0000] <Hixie>
and presumably the people who wrote SPF thought of this and the google issue was just misconfiguration

[15:45:00.0000] <Hixie>
i dunno

[15:45:01.0000] <AryehGregor>
That's what I meant.

[15:45:02.0000] <jgraham>
The HTML5 official tests are at http://dvcs.w3.org/hg/html/

[15:45:03.0000] <jgraham>
So far we are doing pretty badly

[15:46:00.0000] <jgraham>
An insignificant number of tests from Microsoft and a half-finished test harness

[15:51:00.0000] <AryehGregor>
1500 e-mails incorrectly marked as spam.  Ugh.

[15:52:00.0000] <Hixie>
wow

[15:52:01.0000] <Hixie>
that sucks

[15:55:00.0000] <MikeSmith>
hey, speaking of mail, has there been any agreement/resolution on the list discussion about renaming the @required attribute?

[15:55:01.0000] <AryehGregor>
I thought the resolution was "keep it, not enough sites to matter".

[15:55:02.0000] <annevk>
not renaming it

[15:56:00.0000] <Hixie>
i haven't looked at it yet

[15:56:01.0000] <Hixie>
depends on what browser vendors are willing to do

[15:56:02.0000] <Hixie>
opera and chrome don't seem to be willing to implement it as is without hacks

[15:57:00.0000] <Hixie>
(targetted hacks in opera's case, way over-broad hacks in chrome's case)

[15:57:01.0000] <Hixie>
so something will probably have to give

[15:57:02.0000] <Hixie>
i'll have to look more closely at the actual sites to see what we can do

[15:57:03.0000] <AryehGregor>
Maybe they would be okay with dropping the hacks if Mozilla implements it and evangelizes the sites in question?

[15:57:04.0000] <AryehGregor>
Mozilla has bigger market share, more likely to convince them.

[15:57:05.0000] <Hixie>
that might help

[16:03:00.0000] <annevk>
Jonas seemed to prefer that

[16:03:01.0000] <annevk>
me too

[16:03:02.0000] <annevk>
renaming things is no fun

[16:07:00.0000] <othermaciej>
Hixie: We are considering removing @required from WebKit entirely

[16:07:01.0000] <TabAtkins>
And replacing it with something?

[16:07:02.0000] <othermaciej>
at least until (a) it has a name with little enough conflict (for whatever reason) and (b) we actually have UI to tell you what happened if you can't submit a form because of it

[16:08:00.0000] <othermaciej>
I think the hack of supporting it but only in strict mode is ridiculous and broken

[16:08:01.0000] <Hixie>
othermaciej: well implementing it without hte UI was a terrible idea

[16:08:02.0000] <Hixie>
othermaciej: since the whole point of the feature is the UI

[16:08:03.0000] <TabAtkins>
Yeah, Aryeh has a bug asking webkit to turn off validation until there's UI that I can +1 on.

[16:08:04.0000] <othermaciej>
Hixie: agreed

[16:09:00.0000] <AryehGregor>
Someone else from WebKit also agreed on my bug.

[16:09:01.0000] <Hixie>
othermaciej: i told the people who did it as much, but they kinda blew me off, so... :-)

[16:09:02.0000] <othermaciej>
anyway, if Mozilla wants to go on a crusade to make the Web safe for an attribute named "required", I'm fine with waiting for the results, but I think the more practical choice would be to rename it

[16:09:03.0000] <AryehGregor>
They blew me off too.

[16:09:04.0000] <AryehGregor>
othermaciej, I thought there were only a limited number of sites with compatibility problems.

[16:09:05.0000] <Hixie>
i'm fine with renaming it too, i just need to look at the facts before i can work out what the right course of action is

[16:09:06.0000] <Hixie>
and subtitles come first :-P

[16:09:07.0000] <AryehGregor>
Also, good UI would really help a lot.

[16:10:00.0000] <annevk>
WebKit shipped without UI? blargh

[16:10:01.0000] <annevk>
I was told in a bug it would all be okay

[16:11:00.0000] <othermaciej>
AryehGregor: I believe there are JavaScript libraries that use an attribute named "required" for their own purposes

[16:11:01.0000] <AryehGregor>
annevk, yep.  MediaWiki had to disable form validation completely because of that.

[16:11:02.0000] <AryehGregor>
othermaciej, oh, bleh.

[16:11:03.0000] <othermaciej>
AryehGregor: that makes it likely the problem will be widespread

[16:11:04.0000] <AryehGregor>
I thought Opera only had to blacklist a few smallish sites?

[16:11:05.0000] <othermaciej>
and it is widespread enough that we actually got bugs

[16:12:00.0000] <othermaciej>
I'm fine with just disabling the feature til we get this straightened out

[16:17:00.0000] <annevk>
does seem good to disable those that require UI if the UI is not working

[16:31:00.0000] <paul_irish>
AryehGregor: was mediawiki using a javascript library in particular that used a required attribute?/

[16:32:00.0000] <AryehGregor>
paul_irish, no, it was using required correctly for the most part.  The problem was the lack of UI in WebKit.

[16:32:01.0000] <paul_irish>
ok

[16:32:02.0000] <AryehGregor>
People were trying to submit forms and they'd refuse to submit, with no error message.

[16:32:03.0000] <AryehGregor>
Which makes the feature worse than useless.

[16:32:04.0000] <AryehGregor>
Also, Opera's UI for it is pretty ugly, and that's the only other shipping implementation.

[16:39:00.0000] <Lachy>
AryehGregor, we know. we have our dev relations guys kicking us internally about the state of our HTML5 form controls.

[16:39:01.0000] <AryehGregor>
I know you know.  :)

[16:40:00.0000] <AryehGregor>
I just hope we don't have zillions of Safari 5s hanging around forever that we have to sniff around for this.

[16:40:01.0000] <othermaciej>
some aspects of WebKit's version are ugly too

[16:40:02.0000] <othermaciej>
Safari update uptake is pretty quick and our release cycles tend to be a year max for major releases

[16:40:03.0000] <Lachy>
what form HTML5 form controls has webkit implemented these days?

[16:41:00.0000] <othermaciej>
and we fix significant compat bugs in minor releases

[16:41:01.0000] <othermaciej>
I think we have at least half-assed implementations of all the new <input> types, but most are not good enough (IMO) to whole-heartedly recommend to developers

[16:41:02.0000] <Lachy>
oh, type="date" looks like an number control

[16:42:00.0000] <TabAtkins>
type=color looks like a text input, unfortunately.

[16:42:01.0000] <Lachy>
and pressing up or down on an <input type="number"> control without a value gives +/- 1.7976931348623157e+308

[16:42:02.0000] <othermaciej>
like I said

[16:42:03.0000] <TabAtkins>
Haha, wow.

[16:42:04.0000] <othermaciej>
not yet done to good quality

[16:43:00.0000] <othermaciej>
if anyone would like to be paid by Apple to work in this area, send me a resume

[16:43:01.0000] <othermaciej>
I will add that many of the controls are even worse on MobileSafari than Safari or Chrome on Mac/Win/Linux

[16:44:00.0000] <Lachy>
and <input type="week"> starts at the year 1583-W01 :-)


2010-06-26
[17:13:00.0000] <paul_irish>
othermaciej: do you have a recommendation on how to feature detect the new input types in safari?

[17:14:00.0000] <paul_irish>
assuming the tests would fail now..

[17:16:00.0000] <othermaciej>
paul_irish: I hadn't thought about it - I don't know if the spec has any mechanism for feature detection

[17:17:00.0000] <othermaciej>
in any case feature tests would currently pass, even though our implementations are pretty crufty

[17:17:01.0000] <Hixie>
checking the value of input.type should work iirc

[17:17:02.0000] <Hixie>
iirc it'll be "text" if the type isn't supported at all

[17:17:03.0000] <paul_irish>
yeah.. basic approachis to check value of input.type and see if it isn't 'text' .. chrome loves to false positive with that

[17:17:04.0000] <Hixie>
yeah that's the problem with feature tests :-)

[17:17:05.0000] <othermaciej>
I believe both Chrome and Safari will claim to support almost all the types by that standard

[17:18:00.0000] <othermaciej>
which in a sense we do, but perhaps not in a way you would actually want to use

[17:18:01.0000] <Hixie>
using cutting edge feature is fraught with such difficulties

[17:18:02.0000] <paul_irish>
so i've been throwing text like ':)' into the value of input type="date" and reading it out.. assuming that a proper date input would only allow proper date input..  that's seemed to work so far

[17:18:03.0000] <Hixie>
othermaciej: i was sad to see the incomplete implementations ship, didn't expect that

[17:39:00.0000] <MikeSmith>
latest HTML WG heartbeat WDs and FPWDs are now alive at http://www.w3.org/TR/html5/ etc.

[17:40:00.0000] <MikeSmith>
for convenient access, see reverse-chronological index at http://www.w3.org/TR/tr-date-drafts

[17:41:00.0000] <MikeSmith>
/me also notices http://www.w3.org/TR/2010/WD-media-frags-20100624/

[17:41:01.0000] <MikeSmith>
whatever that is, it's at LC

[17:41:02.0000] <MikeSmith>
so perhaps worth reviewing

[17:42:00.0000] <MikeSmith>
/me sees nessy1 listed there among the editors

[17:42:01.0000] <MikeSmith>
"specifies the syntax for constructing media fragment URIs and explains how to handle them when used over the HTTP protocol."

[17:44:00.0000] <MikeSmith>
/me notices LC of SVG 1.1 "Second Edition"

[17:44:01.0000] <MikeSmith>
http://www.w3.org/TR/2010/WD-SVG11-20100622/

[17:45:00.0000] <MikeSmith>
which appears to have a pretty big list of substantive changes

[17:45:01.0000] <MikeSmith>
http://www.w3.org/TR/2010/WD-SVG11-20100622/changes.html#Substantial

[17:46:00.0000] <MikeSmith>
seems the precedent set by XML 1.0 doing the "Fifth Edition" thing to incorporate substantive changes and not just errata is now firmly rooted

[17:47:00.0000] <MikeSmith>
that seems to perhaps be a really unfortunate trend

[17:47:01.0000] <MikeSmith>
anyway, the SVG 1.1 draft is at LC also

[17:48:00.0000] <MikeSmith>
reminds me that I want to send LC comments on MathML3 before July 1 deadline

[17:50:00.0000] <othermaciej>
MikeSmith: woot!

[17:50:01.0000] <MikeSmith>
othermaciej: yeah, sorry I didn't get them out yesterday

[17:50:02.0000] <MikeSmith>
I blame it on too much reefer

[17:50:03.0000] <othermaciej>
do you know if the w3c plans to post any kind of announcement?

[17:51:00.0000] <othermaciej>
or should it just be announced on the list?

[17:51:01.0000] <MikeSmith>
we will do an announcement soon

[17:51:02.0000] <othermaciej>
it would be nice if there was a single thing to link to for the set of all the new drafts

[17:51:03.0000] <MikeSmith>
yes

[17:51:04.0000] <MikeSmith>
I think we will have that by end of day US/West

[17:52:00.0000] <MikeSmith>
they will all be linked to in the announcement

[17:52:01.0000] <othermaciej>
cool beans

[17:52:02.0000] <MikeSmith>
Home Page News

[17:56:00.0000] <MikeSmith>
Hixie: I realize you are probably long ago bored with my mentioning the processing issues of the annotation-xml element 

[17:56:01.0000] <MikeSmith>
but I remembered something I didn't mention before

[17:57:00.0000] <MikeSmith>
the annotation-xml element has an "encoding" attribute

[17:57:01.0000] <MikeSmith>
which specifies a MIME type

[17:58:00.0000] <MikeSmith>
so if that is present and it specifies text/html, then seems like the parsing algorithm could use that info to do the similar behavior it does with SVG in there now

[17:58:01.0000] <Hixie>
that might make sense, yeah

[17:58:02.0000] <Hixie>
file a bug?

[17:58:03.0000] <MikeSmith>
so, we would not need a <div> wrapper or <body> or <html>

[17:58:04.0000] <MikeSmith>
yeah, I will update the bug I already opened

[17:59:00.0000] <MikeSmith>
after I get some sleep

[17:59:01.0000] <Hixie>
k

[18:59:00.0000] <boblet>
Anyone know what Microformats wiki means by “Avoid the "microdata vCard" vocabulary as it is an out-of-date fork/snapshot of hCard”? http://microformats.org/wiki/html5

[19:00:00.0000] <boblet>
I thought Microdata vCard was based on vCard

[19:00:01.0000] <boblet>
are there some issues regarding this I’m not aware of?

[19:13:00.0000] <MikeSmith>
boblet: I remember hearing some details from tantek about that

[19:13:01.0000] <MikeSmith>
maybe ping him on twitter

[19:13:02.0000] <boblet>
MikeSmith: ok will do, thanks

[19:14:00.0000] <boblet>
also have yet to check µF mail archives

[19:14:01.0000] <MikeSmith>
but iirc the problem is that the RFC that the microdata vCard vocab was based on is superseded now

[19:14:02.0000] <MikeSmith>
by an updated RFC

[19:14:03.0000] <MikeSmith>
which the hCard vocab is in sync with

[19:15:00.0000] <boblet>
woah holy shit. They *update* RFCs?

[19:15:01.0000] <MikeSmith>
replace them

[19:15:02.0000] <MikeSmith>
new ones obsolete the old ones

[19:15:03.0000] <boblet>
I thought those things got carved into stone back in the mesozoic era

[19:15:04.0000] <TabAtkins>
Um.  I'm not authorized to comment on the Issue-30 poll.

[19:16:00.0000] <boblet>
(which is why they’re fixed width)

[19:16:01.0000] <boblet>
thanks for the info MikeSmith. bbl

[19:16:02.0000] <MikeSmith>
TabAtkins: lemme check now

[19:17:00.0000] <TabAtkins>
It's possible that I'm logged in as tabatkins, rather than tabatkins2

[19:18:00.0000] <MikeSmith>
ah

[19:18:01.0000] <MikeSmith>
that might be it

[19:18:02.0000] <TabAtkins>
I don't know how to log out of Basic Auth.

[19:19:00.0000] <MikeSmith>
arr, me neither

[19:19:01.0000] <MikeSmith>
lemme check

[19:19:02.0000] <TabAtkins>
/me heads home, and will respond on Monday he guesses.

[19:21:00.0000] <MikeSmith>
ok

[20:54:00.0000] <Hixie>
boblet: i think that's out of date -- the microdata vcard spec has been changed to just work directly from vCard rather than hCard

[21:38:00.0000] <jmcneese>
howdy canvasers

[21:40:00.0000] <jmcneese>
so, can anyone answer if it is per spec/functionality that if i stroke a line in a canvas, it does not clear if if reset the canvas via changing the width or clearRect()?

[21:41:00.0000] <Hixie>
how do you mean?

[21:41:01.0000] <Hixie>
do you have a sample page showing the problem?

[21:41:02.0000] <jmcneese>
not on a reachable server.

[21:42:00.0000] <jmcneese>
it's simple to replicate.  start with an empty canvas.  draw a line down the center.  do a clearRect() on the canvas.  the line stays

[21:43:00.0000] <jmcneese>
curious if i'm missing a step

[21:45:00.0000] <Hixie>
what arguments are you passing clearRect()?

[21:48:00.0000] <jmcneese>
0,0,CANVASWIDTH,CANVASHEIGHT

[21:48:01.0000] <jmcneese>
it's working for any canvas i draw actual shapes on (rects, arcs, etc)

[21:49:00.0000] <Hixie>
so like:

[21:49:01.0000] <Hixie>
c.beginPath();

[21:49:02.0000] <Hixie>
c.moveTo(100,100);

[21:49:03.0000] <Hixie>
c.lineTo(200,300);

[21:49:04.0000] <Hixie>
c.stroke();

[21:49:05.0000] <Hixie>
c.clearRect(0, 0, 640, 480);

[21:49:06.0000] <Hixie>
...doesn't work?

[21:49:07.0000] <Hixie>
(where c is the context of course)

[21:50:00.0000] <jmcneese>
aha... is beginPath required?

[21:50:01.0000] <Hixie>
unless you do beginPath(), you're just adding to the last line each time

[21:50:02.0000] <Hixie>
stroke() doesn't reset the path

[21:51:00.0000] <jmcneese>
ok, noted.  but i don't see how that prevents the line from clearing when calling clearRect

[21:51:01.0000] <Hixie>
e.g. try:

[21:51:02.0000] <Hixie>
c.lineWidth=100; c.strokeStyle = 'red'; c.stroke();

[21:51:03.0000] <Hixie>
c.lineWidth=50; c.strokeStyle = 'orange'; c.stroke();

[21:51:04.0000] <Hixie>
c.lineWidth=10; c.strokeStyle = 'yellow'; c.stroke();

[21:52:00.0000] <Hixie>
well it's not that it prevents the line from being cleared, so much as you're probably calling stroke() again after you clear

[21:52:01.0000] <Hixie>
e.g. on a timer or something

[21:52:02.0000] <Hixie>
right?

[21:52:03.0000] <Hixie>
hard to say without seeing the code

[21:52:04.0000] <jmcneese>
i am calling redraw on mousedrag

[21:53:00.0000] <jmcneese>
now, since i couldn't get it to clear the lines for a grid, i ended up drawing rectangles (with three sides off canvas), which clear just fine

[21:54:00.0000] <Hixie>
yeah the rect methods don't affect the path

[21:55:00.0000] <Hixie>
so they wouldn't get added each time

[21:55:01.0000] <Hixie>
the key is just that all the liens you add to the path don't get reset until you call beginPath() again

[21:55:02.0000] <jmcneese>
so, paths must be aha

[21:55:03.0000] <jmcneese>
that makes sense

[21:58:00.0000] <jmcneese>
thanks mate

[21:58:01.0000] <Hixie>
np

[23:54:00.0000] <boblet>
Hixie: re: vcard & hcard, I’m guessing that the microdata vcard and vevent specs have been updated to the latest RFCs right? (apparently the reason for the microformats.org comment)

[23:54:01.0000] <boblet>
if so I’ll edit the wiki or contact Tantek about it

[23:54:02.0000] <Hixie>
have the RFCs changed recently?

[23:55:00.0000] <Hixie>
like in the last year?

[00:00:00.0000] <boblet>
Hixie: Mike seemed to think that was the reason. Will do some digging…

[00:06:00.0000] <Hixie>
i know there was some work to update the rfcs, but i haven't updated the vocabs to the new ones if that's it

[00:06:01.0000] <Hixie>
i didn't think hcard had been updated to them either though

[00:08:00.0000] <boblet>
Hixie: yeah it seems vcard 1.1 is still draft anyhow: http://datatracker.ietf.org/doc/draft-ietf-vcarddav-vcardrev/

[00:09:00.0000] <Hixie>
k

[00:09:01.0000] <boblet>
guess I’ll have to ping tantek and see what was meant by those statements. seems a bit strange

[00:54:00.0000] <annevk>
oh look 

[00:54:01.0000] <annevk>
http://www.w3.org/News/2010#entry-8843

[00:54:02.0000] <annevk>
yay for Mike

[00:58:00.0000] <MikeSmithX>
annevk: yay for editors getting their work done

[00:58:01.0000] <MikeSmith>
annevk: btw, thanks for all the time you spent on responding to comments about the diffs doc

[00:58:02.0000] <MikeSmith>
I know it took a lot of time

[00:59:00.0000] <MikeSmith>
and most people have no idea how much time and work goes into

[00:59:01.0000] <MikeSmith>
but it's a very useful document to have available

[00:59:02.0000] <MikeSmith>
for people who are trying to follow the changes

[00:59:03.0000] <MikeSmith>
and especially for people who are just starting to read up on HTML5

[01:00:00.0000] <MikeSmith>
when people ask me where to start with learning about HTML5, I always tell them to start with reading that document

[01:02:00.0000] <annevk>
cool man

[01:04:00.0000] <MikeSmith>
so along with getting updated WDs published and despite all the drama, all the actually important stuff still keeps on getting done

[01:05:00.0000] <MikeSmith>
e.g., giantly newsworthy stuff like canvas support finally being added to IE

[01:06:00.0000] <MikeSmith>
which the IE tech people deserve a lot of serious congrats for

[01:06:01.0000] <MikeSmith>
despite it taking some time to make it happen

[01:07:00.0000] <MikeSmith>
two years from now nobody will even remember that it took so long

[01:07:01.0000] <MikeSmith>
well I guess we may remember

[01:07:02.0000] <MikeSmith>
but nobody else will

[01:08:00.0000] <MikeSmith>
and then all kinds of other bits of important work are steadily and mostly quietly getting done

[01:08:01.0000] <MikeSmith>
e.g., moving ahead with getting agreement on Notifications

[01:09:00.0000] <annevk>
the way WebKit reverses what to switch first on in the tree builder is interesting

[01:09:01.0000] <annevk>
what you pointed out in a blog post

[01:09:02.0000] <MikeSmith>
heh

[01:09:03.0000] <MikeSmith>
ah yeah

[01:09:04.0000] <annevk>
euh tweet

[01:09:05.0000] <MikeSmith>
that's a good example

[01:10:00.0000] <MikeSmith>
abarth was talking about that a little bit here a few hours back

[01:10:01.0000] <MikeSmith>
example of "there's more than one way to skin a cat"

[01:10:02.0000] <MikeSmith>
dunno what that equivalent idiom is elsewhere

[01:10:03.0000] <MikeSmith>
I don't know if that began as a UK idiom or a US one

[01:11:00.0000] <MikeSmith>
seems like a US one

[01:11:01.0000] <MikeSmith>
/me tries to think of a less opaque idiom to capture that same meaning

[01:13:00.0000] <MikeSmith>
annevk: dunno how many people in this world would judge that to be interesting

[01:13:01.0000] <MikeSmith>
I mean, I did not because I know a whole lot but mostly because I had spent time looking at hsivonen htmlparser code

[01:13:02.0000] <MikeSmith>
and htmllib

[01:14:00.0000] <MikeSmith>
and not just looking at in the case of hsivonen code

[01:14:01.0000] <MikeSmith>
but actually copying and tweaking it

[01:14:02.0000] <MikeSmith>
for the xml-stylesheet checking feature that I added to validator.nu a while back

[01:15:00.0000] <MikeSmith>
needed it for pseudo-attribute-value parsing

[01:15:01.0000] <MikeSmith>
so it is really interesting to compare, anyway

[01:17:00.0000] <boblet>
who’s in charge of the W3 CSS validator? and will they ever change vendor prefixes from error to warning? ;-)

[01:17:01.0000] <MikeSmith>
boblet: not sure who is currently

[01:17:02.0000] <MikeSmith>
DanC rewrote it recently

[01:17:03.0000] <MikeSmith>
or wrote a new one

[01:17:04.0000] <MikeSmith>
hmm, not sure if that is public info

[01:17:05.0000] <boblet>
it is now :)

[01:18:00.0000] <MikeSmith>
well, maybe there was an announcement about it

[01:18:01.0000] <MikeSmith>
I don't keep up with it so much

[01:18:02.0000] <MikeSmith>
anyway, DanC wrote something in Scala

[01:18:03.0000] <boblet>
I might ask on wwwstyle

[01:18:04.0000] <MikeSmith>
part of the problem with getting people interested in working on the current CSS validator code is that it's written in Java

[01:19:00.0000] <MikeSmith>
which must people immediately just say "ick"

[01:19:01.0000] <MikeSmith>
Philippe le Hegaret wrote the CSS validator originally, I think

[01:20:00.0000] <MikeSmith>
boblet: lemme go look at the source repo know and see who's working on it currently

[01:20:01.0000] <boblet>
thanks MikeSmith 

[01:20:02.0000] <MikeSmith>
boblet: btw, did you figure out what actual difference there are between vcard microformat and hcard one?

[01:21:00.0000] <MikeSmith>
I remember tantek telling me what the differences were and I remember those differences between important

[01:21:01.0000] <MikeSmith>
but I can't remember now that the differences actually are :(

[01:22:00.0000] <boblet>
MikeSmith: no. am planning to ping Tantek after checking µF email archives. the notes were added Oct 2009 though, so I’m guessing it’s something to do with the old hcard-based Microdata vcard (although currently it’s RFC2426-based so theoretically no worries)

[01:22:01.0000] <MikeSmith>
OK

[01:24:00.0000] <MikeSmith>
boblet: I find http://dev.w3.org/cvsweb/2002/css-validator/org/w3c/css/css/

[01:24:01.0000] <MikeSmith>
etc.

[01:25:00.0000] <MikeSmith>
and I remember that my W3C sempai Yves Lafon is the main one making CSS validator maintenance changes these days

[01:25:01.0000] <boblet>
heh, searching http://lists.w3.org/Archives/Public/www-validator-css/ atm

[01:25:02.0000] <MikeSmith>
most recent change seems to be from 7 days ago

[01:25:03.0000] <MikeSmith>
boblet: anyway, it is very much one of those "squeaky wheel gets the grease" kind of things

[01:26:00.0000] <MikeSmith>
if you really want some particular changes made, file a bugzilla bug

[01:26:01.0000] <boblet>
MikeSmith: thanks. will squeak a bit then :)

[01:26:02.0000] <MikeSmith>
and feel free to Cc me on the bug, at mike⊙wo

[01:27:00.0000] <MikeSmith>
and I can then harass Yves about it until he his tired of me mentioning it

[01:27:01.0000] <MikeSmith>
boblet: http://www.w3.org/Bugs/Public/enter_bug.cgi?product=CSSValidator

[01:27:02.0000] <boblet>
cheers — will do!

[01:28:00.0000] <MikeSmith>
yeah, please do

[01:28:01.0000] <MikeSmith>
also consider mentioning it on the www-validator list, and including a link to the bug

[01:29:00.0000] <MikeSmith>
and others who are interested can add themselves to the Cc for the bug

[01:29:01.0000] <boblet>
har, paul_irish bet me by 10 days http://www.w3.org/Bugs/Public/show_bug.cgi?id=9932

[01:29:02.0000] <MikeSmith>
well, there you go

[01:29:03.0000] <MikeSmith>
God is clearly on your side

[01:30:00.0000] <MikeSmith>
and paul_irish is one of his secret agents

[01:30:01.0000] <MikeSmith>
/me adds himself to the Cc for that bug

[01:30:02.0000] <boblet>
Thank science! … uh thank God!

[01:31:00.0000] <MikeSmith>
I suspect I might probably be able to make that change relatively quickly and easily myself

[01:31:01.0000] <MikeSmith>
but I'm not going to try

[01:32:00.0000] <MikeSmith>
because if I start fixing css validator bugs, I will then become at least partially responsible for maintaining it too

[01:32:01.0000] <boblet>
ref http://lists.w3.org/Archives/Public/www-validator-css/2009Jul/0002.html

[01:33:00.0000] <MikeSmith>
thanks

[01:33:01.0000] <MikeSmith>
yipes

[01:33:02.0000] <MikeSmith>
[[

[01:33:03.0000] <MikeSmith>
> /* Do not Validate BEGIN */

[01:33:04.0000] <MikeSmith>
>  -moz-....

[01:33:05.0000] <MikeSmith>
>  -webkit-...

[01:33:06.0000] <MikeSmith>
> /* Do not Validate END */

[01:33:07.0000] <MikeSmith>
]]

[01:33:08.0000] <MikeSmith>
yuck

[01:34:00.0000] <boblet>
yeah I agree with Yann. but if the reason they generate an error is the starting hyphen, that’s bs

[01:34:01.0000] <MikeSmith>
well, I hope some people can step up to the plate to help Yves out with maintenance

[01:34:02.0000] <boblet>
if the validator doesn’t validate them (and assuming hyphen is the correct way to add a vendor prefix) they should be a warning not error

[01:34:03.0000] <MikeSmith>
true I guess

[01:35:00.0000] <MikeSmith>
but there are many other things that the CSS validator is not handling

[01:35:01.0000] <MikeSmith>
CSS3 stuff

[01:36:00.0000] <MikeSmith>
it is to the point that pretty much every working-draft we go to publish these days, when I check it with the W3C "pubrules" checker, which includes a CSS check, I get errors about the CSS

[01:36:01.0000] <MikeSmith>
which I and pubs team now pretty much just have to ignore

[01:36:02.0000] <MikeSmith>
because most of them are spurious

[01:37:00.0000] <MikeSmith>
and so that defeats the purpose of having a CSS check to begin with

[01:37:01.0000] <MikeSmith>
so we really should fix it

[01:37:02.0000] <MikeSmith>
I am pretty sure that Yves would be very happy to hand over the reins to anybody who could commit to helping with it

[01:38:00.0000] <MikeSmith>
and who had the coding chops

[01:38:01.0000] <MikeSmith>
or would at least be willing to spend some time learning

[01:39:00.0000] <MikeSmith>
I wish I could make time to help out with it myself

[01:39:01.0000] <MikeSmith>
but I would be the wrong person to do it anyway

[01:39:02.0000] <MikeSmith>
since I don't know jack about CSS

[01:40:00.0000] <MikeSmith>
boblet: learn some little Java, and you can get commit access and make the changes yourself directly

[01:40:01.0000] <boblet>
will add it to my list of things to do ＞＿＜

[01:40:02.0000] <MikeSmith>
or maybe you know some Java already

[01:40:03.0000] <MikeSmith>
but you are keeping it quiet

[01:41:00.0000] <boblet>
har!

[01:41:01.0000] <MikeSmith>
so that you don't get stuck with maintaining the code

[01:41:02.0000] <MikeSmith>
I'm on to you, man

[01:41:03.0000] <MikeSmith>
boblet: btw, DanC's new code is at http://dvcs.w3.org/hg/css-val/

[01:42:00.0000] <MikeSmith>
http://dvcs.w3.org/hg/css-val/file/f5931981d7dd/

[01:42:01.0000] <MikeSmith>
http://dvcs.w3.org/hg/css-val/raw-file/f5931981d7dd/README.html

[01:42:02.0000] <MikeSmith>
Scala seems to be relatively easier and less painful to learn than Java

[01:42:03.0000] <MikeSmith>
and more forgiving, maybe

[01:44:00.0000] <MikeSmith>
"CSS syntax evolves in a backward- and forward-compatible way; the CSS 2.1 specification includes a uniform grammar, applicable to all CSS versions. The original (2002) W3C CSS Validator predates this uniform grammar and maintenance is becoming an undue burden as a result. This is an attempt to build a validator using the uniform syntax."

[01:44:01.0000] <MikeSmith>
hmm, so sounds like we are in pretty much the same situation with the CSS validator that we are with the current W3C Markup Validator

[01:45:00.0000] <boblet>
hrm, hopefully DanC gets some volunteers then eh

[01:45:01.0000] <boblet>
ok bbl, the little princess demands I play with her

[01:46:00.0000] <MikeSmith>
from what Dan says there, I can see that the existing CSS validator code, like the existing W3C Markup Validator code, is based on an obsolete backend, and we are not getting a good return on investment for the time we might put into maintaining it any further

[01:46:01.0000] <MikeSmith>
we just need to replace it

[01:46:02.0000] <MikeSmith>
boblet: k

[01:50:00.0000] <MikeSmith>
annevk: (if you are around) do you know, is there a formal grammar for CSS3

[01:50:01.0000] <MikeSmith>
?

[01:50:02.0000] <MikeSmith>
yet

[01:51:00.0000] <MikeSmith>
I don't know what "uniform grammar" means

[01:53:00.0000] <MikeSmith>
I have serious misgivings about basing checking in a validator on a formal grammar to begin with

[01:53:01.0000] <MikeSmith>
but maybe the CSS case is different than others

[01:54:00.0000] <MikeSmith>
/me finds http://www.w3.org/TR/css3-syntax/

[01:54:01.0000] <annevk>
hmm not really

[01:54:02.0000] <MikeSmith>
http://www.w3.org/TR/css3-syntax/#detailed-grammar

[01:54:03.0000] <annevk>
we just patch the CSS 2.1 grammar basically

[01:54:04.0000] <MikeSmith>
annevk: seems like dbaron is working on that

[01:55:00.0000] <MikeSmith>
oh

[01:55:01.0000] <MikeSmith>
well, that's better than nothing

[01:55:02.0000] <annevk>
css3-syntax is hopelessly out of date

[01:55:03.0000] <annevk>
it's from 2003

[01:55:04.0000] <MikeSmith>
http://www.w3.org/TR/css3-syntax/#lexical

[01:55:05.0000] <MikeSmith>
ah

[01:55:06.0000] <annevk>
many many grammar changes have been made since then

[01:56:00.0000] <MikeSmith>
yet another spec in need of an active editor

[01:58:00.0000] <MikeSmith>
I wish the people who aren't otherwise actually yet doing editing or implementing or testing or documenting but instead seem to have endless amounts of time to repeatedly engage in process discussions ad nauseam might actually consider trying to really help out

[01:59:00.0000] <MikeSmith>
help out in other ways than suggesting ways to fix decision process, etc.

[02:00:00.0000] <MikeSmith>
which is not to say that is not useful as well

[02:00:01.0000] <MikeSmith>
but damn

[02:00:02.0000] <annevk>
everyone their talent I guess o_O

[02:00:03.0000] <MikeSmith>
well, everybody benefits from us having an up-to-date CSS validator

[02:01:00.0000] <MikeSmith>
that groks CSS3

[02:01:01.0000] <MikeSmith>
and we need some sound modern backend to base it on

[02:01:02.0000] <MikeSmith>
rather than throwing good money after bad on the old one

[02:01:03.0000] <annevk>
yeah, done in such a way that Validator.nu can invoke it, too

[02:02:00.0000] <annevk>
Validator.nu code, that is

[02:02:01.0000] <MikeSmith>
yeah

[02:02:02.0000] <MikeSmith>
and we need a good grammar to use for that -- to base the CSS checker on .. and a we need a good spec to hold that.. and we need an active editor for the spec

[02:03:00.0000] <MikeSmith>
everyone their talent, yeah

[02:03:01.0000] <MikeSmith>
but christ

[02:04:00.0000] <MikeSmith>
it is not a massive undertaking for somebody who's already smart to learn a little bit about how to write a grammar

[02:04:01.0000] <MikeSmith>
and to do a little bit of spec editing maybe

[02:04:02.0000] <MikeSmith>
we ain't asking for the world here

[02:07:00.0000] <MikeSmith>
/me notices a typo in home-page news for announcement of updated HTML WG drafts

[02:07:01.0000] <MikeSmith>
s/machine-readable date/machine-readable data/

[02:07:02.0000] <MikeSmith>
dam

[02:07:03.0000] <MikeSmith>
mnit

[02:07:04.0000] <MikeSmith>
my fault, too

[02:08:00.0000] <MikeSmith>
but I don't have perms to change it now

[02:08:01.0000] <MikeSmith>
ah well

[03:30:00.0000] <boblet>
foolip: yt?

[04:17:00.0000] <Philip`>
"why do we care about 1%? Who decided where we draw the line?" - I thought it was 1% because there's lots of Opera people in the group and Opera isn't much more than 1%

[04:35:00.0000] <gsnedders>
boblet: Probably flying now.

[04:40:00.0000] <MikeSmith>
annevk: btw, update of diffs doc has already been added to the Japanese translation at html5.jp

[04:40:01.0000] <MikeSmith>
http://twitter.com/futomi/status/17073669143

[04:40:02.0000] <MikeSmith>
http://www.html5.jp/trans/w3c_differences.html

[06:28:00.0000] <Philip`>
The W3C Test Suite License seems to have not been designed with the idea of including its text in the files you're going to distribute under the license

[06:29:00.0000] <Philip`>
since it uses hyperlinks and bold and italic and non-ASCII

[06:35:00.0000] <boblet>
gsnedders: ok thanks

[07:18:00.0000] <erlehmann>
can anyone advise me on this issue ? wordpress is using <em> and <strong> by default, but the buttons are marked <i> and <b> http://core.trac.wordpress.org/ticket/10465

[07:19:00.0000] <erlehmann>
it's just too dumb and i don't know how to tell them to change this.

[07:19:01.0000] <erlehmann>
last time i tried, they kicked me out of #wordpress :/

[07:24:00.0000] <Philip`>
erlehmann: Since i/em and b/strong are synonyms in practice, it seems a waste of time to try to get them to change

[07:26:00.0000] <erlehmann>
Philip`, but emphazised text is different from italiced text (like names that need to be set apart from the regular flow of text), especially in quotes.

[07:26:01.0000] <erlehmann>
also, these buttons confuse me.

[07:26:02.0000] <erlehmann>
all other buttons do what's on the label :(

[07:27:00.0000] <Philip`>
erlehmann: Few people care about the distinction, and the people who do care are usually on the wrong side, so it doesn't seem a good fight to get into :-)

[07:28:00.0000] <erlehmann>
fine. i will call my next helper script „bake_cake.zsh“ then. but make it burn the users home directory. huehuehue

[07:33:00.0000] <Philip`>
I've already learned not to trust anyone who promises to bake me a cake, so that's fine

[07:55:00.0000] <gsnedders>
Oooo! Cake nom nom!

[08:52:00.0000] <jgraham>
gsnedders: No! You can't om nom ... oh wait what? Ah. I've just ben told that om noming cake is fine. As you were.

[09:09:00.0000] <erlehmann>
gsnedders, jgraham, the WHATWG wishes to remind you that upon completion of the spec, there will be cake.

[09:14:00.0000] <boblet>
erlehmann: cake that deletes completed specs?

[09:17:00.0000] <boblet>
in foolip’s excellent http://foolip.org/microdatajs/live/ the hcard in <span itemprop="location" itemscope itemtype="http://microformats.org/profile/hcard"> isn’t shown unless the itemprop="location" is removed. looks like a bug

[09:40:00.0000] <boblet>
nimbupani inna house!

[09:40:01.0000] <nimbupani>
yo boblet !!

[09:40:02.0000] <nimbupani>
emoticon koan guru :)

[09:40:03.0000] <boblet>
so was it the 微妙 or the … aah, the emoticons huh

[09:40:04.0000] <boblet>
:|

[09:40:05.0000] <boblet>
didn’t realise I was being so deep ;-)

[09:42:00.0000] <nimbupani>
boblet: you were very deep coz I did not understand 微妙 :)

[09:43:00.0000] <boblet>
bimyou = Japanese for dodgy or iffy

[09:43:01.0000] <nimbupani>
ohh okie :)

[10:24:00.0000] <boblet>
hixie: yt?

[10:24:01.0000] <boblet>
or anyone into microdata?

[10:25:00.0000] <boblet>
when I add itemscope to a property, that property now represents the contained item’s name-value pairs, but does it also have a value of all contained text as well?

[10:28:00.0000] <boblet>
eg <p itemprop="members" itemscope>The band is <b itemprop="name">Bob</b> and <b itemprop="name>Sally</b></p>; members = contained name-value pairs (name:Bob, name:Sally) only, or also the value “The band is Bob and Sally.”?

[14:13:00.0000] <Hixie>
has hober been around recently?


2010-06-27
[17:01:00.0000] <MikeSmith>
Hixie: hober just tweeted, so if you need to ask him something, maybe DM him on twitter

[17:05:00.0000] <Hixie>
k

[17:05:01.0000] <Hixie>
thx

[01:51:00.0000] <foolip>
boblet: found a bug?

[01:52:00.0000] <foolip>
boblet: in your example there's no top-level item, nothing should show up, unless I'm misunderstanding

[03:09:00.0000] <boblet>
foolip: welcome back

[03:09:01.0000] <boblet>
foolip: that was inside a top level item (a vevent, hence itemprop="location")

[09:15:00.0000] <AryehGregor>
/me finds out about "cat /proc/sys/kernel/random/uuid", cute

[10:24:00.0000] <foolip>
LS

[15:17:00.0000] <Lachy>
Best optical illusion ever :-D http://www.reddit.com/r/pics/comments/cjkdj/four_stunning_optical_illusions_that_mess_with/

[16:45:00.0000] <taptapir>
how can I achieve alternative sources for audio tag as in this: http://pastebin.com/hZ6r4eCm with javascript:http://en.saturngod.net/blogpost/html5-audio-with-javascript?

[16:49:00.0000] <aho>
var audio = new Audio();

[16:49:01.0000] <aho>
if (audio.canPlayType('audio/ogg; codecs=vorbis') === 'probably') {... go with ogg...} else if etc

[16:52:00.0000] <taptapir>
aho: thanks, I guess there is no way to just specify alternative sources and let the browser figure it out as in HTML.

[16:53:00.0000] <aho>
well, you can write a small function for that

[16:59:00.0000] <roc>
you can create an anonymous audio element, give it anonymous <source> children, etc

[16:59:01.0000] <roc>
well not anonymous ,just not in the DOM

[16:59:02.0000] <roc>
note that "new Audio()" is equivalent to document.createElement("audio")


2010-06-28
[17:03:00.0000] <taptapir>
roc, can you give little example?

[17:03:01.0000] <roc>
var a = new Audio();

[17:05:00.0000] <roc>
a.innerHTML = "<source type='video/webm' src='potatoes.webm'><source type="video/ogg' src='potatoes.ogg'";

[17:05:01.0000] <roc>
a.play();

[17:05:02.0000] <roc>
that should work I think

[17:05:03.0000] <aho>
if you add some escape backslashes

[17:05:04.0000] <aho>
:>

[17:05:05.0000] <roc>
yeah?

[17:05:06.0000] <roc>
depends on the contxt I guess

[17:05:07.0000] <roc>
I also missed a closing >

[17:05:08.0000] <aho>
ye

[17:06:00.0000] <taptapir>
thanks

[17:06:01.0000] <taptapir>
will try

[17:06:02.0000] <aho>
i liked my approach better though .>

[17:07:00.0000] <aho>
i.e. not touching the dom and not invoking the html parser as well

[17:07:01.0000] <aho>
<:

[17:07:02.0000] <roc>
either way's fine, but my way's less code to write

[17:09:00.0000] <aho>
well, my personal use-case is somewhat different... i check codec is supported in order to download the right resource-blob (which contains everything)

[17:09:01.0000] <aho>
+which

[19:38:00.0000] <TheFuzzball>
Hi, I am trying to use <ul> and <li> recursively to make a list of lists, but it's not validating, can someone suggest an alternative?

[19:38:01.0000] <annevk>
yes, put the nested <ul> inside the <li> as per the specification

[19:40:00.0000] <TheFuzzball>
Ah, that makes sense, I was putting it outside of </li>

[19:40:01.0000] <TheFuzzball>
Thanks for that

[21:00:00.0000] <MikeSmith>
http://code.google.com/p/epub-revision/wiki/F2FDayTwo#EPUB_&_the_Web

[21:01:00.0000] <MikeSmith>
http://code.google.com/p/epub-revision/wiki/F2FDayTwo#CSS_2.1_/_CSS_3

[21:02:00.0000] <MikeSmith>
rather: http://code.google.com/p/epub-revision/wiki/F2FDayTwo#relationship_to_CSS_2.1_and_CSS_3

[21:02:01.0000] <MikeSmith>
oops

[21:02:02.0000] <MikeSmith>
ah un-oops

[21:02:03.0000] <MikeSmith>
that's the right links

[21:02:04.0000] <MikeSmith>
*link

[21:03:00.0000] <MikeSmith>
see this part:

[21:03:01.0000] <MikeSmith>
"Markus: trial close - we as WG want to have a preferred/core grammar which is compatible with a subset of HTML5"

[21:03:02.0000] <MikeSmith>
and this:

[21:03:03.0000] <MikeSmith>
http://code.google.com/p/epub-revision/wiki/F2FDayTwo#STRAW_POLL_=_overwhelming_support_for_Markus's_proposition

[21:03:04.0000] <MikeSmith>
"STRAW POLL = overwhelming support for Markus's proposition"

[21:05:00.0000] <annevk>
also contains commentary from Norman Welsh on HTML5 -- rare

[21:06:00.0000] <MikeSmith>
yeah

[21:06:01.0000] <MikeSmith>
I believe Norm joined the ePub 2.1 WG recently

[21:06:02.0000] <MikeSmith>
I'm trying to find a public mailing list for the group

[21:06:03.0000] <MikeSmith>
I think there isn't one

[21:06:04.0000] <MikeSmith>
but they do seem to at least publish their telcon and f2f minutes

[21:07:00.0000] <MikeSmith>
e.g., most recent telcon: http://code.google.com/p/epub-revision/wiki/Telcon20100624

[21:08:00.0000] <MikeSmith>
the recent huge growth in the e-book market would seem to make this ePub 2.1 update especially important

[21:08:01.0000] <MikeSmith>
draft requirements for global language support -

[21:08:02.0000] <MikeSmith>
http://code.google.com/p/epub-revision/wiki/EGLS_requirement_list

[21:11:00.0000] <roc>
that transcript pretty annoying

[21:15:00.0000] <MikeSmithX>
roc: what particular part of it?

[21:15:01.0000] <roc>
the Webkit == HTML5 stuff

[21:15:02.0000] <MikeSmithX>
the fact that they seem to think Webkit is the only HTML5 browser engine on the market?

[21:15:03.0000] <MikeSmithX>
ah

[21:15:04.0000] <MikeSmithX>
yeah, we need to educate them about that

[21:16:00.0000] <MikeSmithX>
roc: would behoove us to try to get some browser people in that group

[21:16:01.0000] <MikeSmithX>
I think most of their f2f meetings are in NYC

[21:16:02.0000] <roc>
get Rob Sayrer in their group, that'd shake them up

[21:16:03.0000] <MikeSmithX>
heh

[21:16:04.0000] <MikeSmithX>
that'd shake any group up

[21:17:00.0000] <roc>
the other annoying thing is their complaining about CSS 2.1 being unstable

[21:17:01.0000] <MikeSmithX>
yeah, there clearly seem to be some misperceptions there about a few things

[21:18:00.0000] <roc>
there are all these W3C specs that made it to REC because they didn't have good test suites, or wide implementation, or real-world content to deal with

[21:18:01.0000] <MikeSmithX>
annevk: I suggest you might want to bring the ePub 2.1 stuff to that attention of Lars-Erik and Chaals

[21:18:02.0000] <MikeSmithX>
annevk: and bug them about getting some people into that group

[21:18:03.0000] <roc>
so a lot of people think that stuff like XSL-FO is somehow "more mature" than CSS 2.1

[21:18:04.0000] <roc>
which is laughable

[21:19:00.0000] <MikeSmithX>
yeah

[21:19:01.0000] <MikeSmithX>
but in that transcript there does at least seem to be a hint of pragmatism and understanding about market realities

[21:20:00.0000] <roc>
yeah

[21:20:01.0000] <roc>
that's good

[21:21:00.0000] <MikeSmithX>
e.g., "Markus: rational for supporting XML with DAISY was that web was evolving to XML vocabularies as first-class, but in practice this hasn't happened"

[21:21:01.0000] <MikeSmithX>
nattokirai: good timing

[21:21:02.0000] <annevk>
"Brady: multi-column layout is eligible for exiting CR in next few days, maybe it's time to go to W3C and use muscle" -- they're in for a surprise :)

[21:21:03.0000] <nattokirai>
MikeSmithX: what's up?

[21:21:04.0000] <MikeSmithX>
nattokirai: we were just talking about the new ePub 2.1 work

[21:22:00.0000] <nattokirai>
oh lordy

[21:22:01.0000] <MikeSmithX>
nattokirai: ePub 2.1 seems likely to end up being based on what's actually being implemented in current browsers

[21:22:02.0000] <MikeSmithX>
nattokirai: btw, I had lunch with Makoto Murata the other day

[21:23:00.0000] <MikeSmithX>
part of what we talked about was... vertical text

[21:23:01.0000] <MikeSmithX>
see http://code.google.com/p/epub-revision/wiki/EGLS_requirement_list

[21:24:00.0000] <MikeSmithX>
nattokirai: I think it would be really helpful to get technical people from some browser projects involved in that group

[21:24:01.0000] <nattokirai>
so, the problem is this

[21:24:02.0000] <nattokirai>
ePub has a very specific set of requirements

[21:24:03.0000] <MikeSmithX>
nattokirai: Murata-san tells me there is a related meeting going on in August in Sapparo

[21:24:04.0000] <nattokirai>
that don't quite match HTML in the large

[21:24:05.0000] <MikeSmithX>
/me nods

[21:24:06.0000] <nattokirai>
for example

[21:25:00.0000] <nattokirai>
wtf is to be done with form controls?

[21:25:01.0000] <MikeSmithX>
nattokirai: I believe that part of what they are doing with the ePub 2.1 is some course correction to address that problem

[21:25:02.0000] <nattokirai>
vertical versions?

[21:25:03.0000] <MikeSmithX>
nattokirai: yeah, understood

[21:25:04.0000] <nattokirai>
/me shudders

[21:25:05.0000] <nattokirai>
and the dark corners of css3 are all impacted

[21:26:00.0000] <MikeSmithX>
that's why we should get some people in there to help them understand

[21:26:01.0000] <annevk>
ah, /whois nattokirai clears things up, hey John :)

[21:26:02.0000] <nattokirai>
howdy

[21:26:03.0000] <nattokirai>
i could be wrong but i think vertical text is really going to mean a full rewrite 

[21:26:04.0000] <MikeSmith>
nattokirai seems to have a magic way of showing up any time we start to talk about stuff related to layout

[21:26:05.0000] <nattokirai>
of a lot of layout engines

[21:26:06.0000] <MikeSmith>
wow

[21:27:00.0000] <MikeSmith>
well, that'd suck

[21:27:01.0000] <MikeSmith>
or actually, it would not happen

[21:27:02.0000] <MikeSmith>
not for a long time at least

[21:27:03.0000] <roc>
right, that won't happen

[21:27:04.0000] <nattokirai>
well, i think the right time to address it is as part of some form of grid layout

[21:27:05.0000] <boblet>
ouch

[21:27:06.0000] <nattokirai>
in whatever form that takes

[21:28:00.0000] <roc>
where do things come unstuck if we define vertical layout as a rotation transformation plus setting some kind of 'glyph-orientation' property? (That can also affect certain replaced elements such as images)

[21:28:01.0000] <MikeSmith>
which means that people in Japan and Taiwan who want to read e-books that more closely match the user experience of real books will instead have to read PDFs or something else not based on HTML+CSS

[21:28:02.0000] <nattokirai>
roc: so you need to flow across columns

[21:28:03.0000] <annevk>
or Internet Explorer starts dominating their market :p

[21:29:00.0000] <nattokirai>
so transformations aren't really right

[21:29:01.0000] <roc>
I'm not sure what you mean

[21:29:02.0000] <roc>
some Chinese newspapers have horizontal columns, but that would work OK

[21:29:03.0000] <MikeSmith>
annevk: yeah, no joke -- seriously market opportunity for IE9 in the eBook market .. let's try to keep it a secret

[21:30:00.0000] <nattokirai>
well, like mirroring in bidi you have different vertical punctuation forms

[21:30:01.0000] <boblet>
heh

[21:30:02.0000] <roc>
well, except "IE9" and "mobile" can't really be used in the same sentence

[21:30:03.0000] <nattokirai>
heh

[21:30:04.0000] <nattokirai>
MikeSmith: i actually think xsl 1.1 is closer to what ePub wants

[21:31:00.0000] <MikeSmith>
roc: btw, another note of pragmatism from the minutes: "Markus: rational for supporting XML with DAISY was that web was evolving to XML vocabularies as first-class, but in practice this hasn't happened" (Markus is CTO of DAISY and chair of the WG)

[21:31:01.0000] <boblet>
ruby on vertical text would no doubt be an extra layer of fun to implement too

[21:31:02.0000] <annevk>
given how complex CSS layout is nowadays it sort of dazzles me when I try to think of vertical

[21:31:03.0000] <nattokirai>
well, ruby is easy compared to things like warichu

[21:31:04.0000] <MikeSmith>
nattokirai: xsl 1.1 as opposed to what? CSS?

[21:31:05.0000] <annevk>
especially how certain complicated parts of CSS layout nowadays are not defined (e.g. tables)

[21:31:06.0000] <nattokirai>
which is two columns of text displayed within a larger column

[21:31:07.0000] <roc>
nattokirai: selecting correct character forms could be handled without much hassle

[21:31:08.0000] <nattokirai>
yup

[21:32:00.0000] <nattokirai>
roc: i think it's completely doable but there are still lots of dependencies that need to be worked out

[21:32:01.0000] <roc>
seems like the way to go then

[21:33:00.0000] <roc>
since it can be done incrementally

[21:33:01.0000] <nattokirai>
so you define vertical versions of pop-up menus?

[21:33:02.0000] <roc>
popups already have interesting interactions with CSS transforms

[21:33:03.0000] <roc>
this would be no different

[21:33:04.0000] <nattokirai>
i can imagine

[21:34:00.0000] <roc>
making popups transform correctly isn't actually all that hard

[21:34:01.0000] <boblet>
damn, can’t afford to make it to Tokyo this week. will be sorry to miss your talk nattokirai 

[21:34:02.0000] <roc>
make the widget the right size, make it transparent, carry in the transform when rendering, and Bob's your uncle

[21:34:03.0000] <nattokirai>
boblet: i'll post slides...

[21:34:04.0000] <roc>
if you actually want to do that. I think it would freak people out.

[21:35:00.0000] <nattokirai>
heh

[21:35:01.0000] <roc>
I think having an untransformed popup appear in roughly the right place would be satisfactory, wouldn't it?

[21:35:02.0000] <nattokirai>
dunno, might work

[21:36:00.0000] <nattokirai>
but it relates back to the whole top is right debate

[21:36:01.0000] <roc>
I mean, it's not as if we're striving to faithfully render Qin dynasty popups

[21:36:02.0000] <nattokirai>
well, it's not totally unreasonable

[21:36:03.0000] <nattokirai>
tests in japan are given in vertical form

[21:37:00.0000] <nattokirai>
so having radio buttons flow vertically should be natural

[21:37:01.0000] <nattokirai>
and fairly simple

[21:38:00.0000] <nattokirai>
i think this really requires a vertical layout module

[21:38:01.0000] <nattokirai>
that tries to cover the minimal set needed to make things work

[21:38:02.0000] <nattokirai>
otherwise you'll be playing spec whack-a-mole

[21:39:00.0000] <nattokirai>
MikeSmith: yeah, so i think xsl 1.1 instead of css makes sense for the ePub requirements

[21:42:00.0000] <nattokirai>
roc: btw, did you follow the whole physical vs. logical dimensions debate?

[21:42:01.0000] <roc>
I have in the past

[21:42:02.0000] <roc>
in fact, I instigated the latest round of it

[21:42:03.0000] <roc>
I may have not kept up recently

[21:42:04.0000] <nattokirai>
so you think it makes sense to add margin-before/after/start/end?

[21:43:00.0000] <roc>
ohhhh

[21:43:01.0000] <roc>
sorry

[21:43:02.0000] <roc>
I was thinking about units

[21:43:03.0000] <roc>
nattokirai: vertical-flowing radio buttons would work fine with a transforms-based approach

[21:44:00.0000] <MikeSmith>
nattokirai: where are you going to be presenting?

[21:44:01.0000] <roc>
the question you raised was whether we need to be able to have a single stylesheet that can intelligently style content without knowing what the orientation is

[21:44:02.0000] <roc>
right?

[21:44:03.0000] <nattokirai>
yes

[21:44:04.0000] <nattokirai>
MikeSmith: WDE ex, Apple Store Ginza, wed eve

[21:45:00.0000] <MikeSmith>
ah good

[21:45:01.0000] <nattokirai>
roc: not sure about transforms

[21:46:00.0000] <nattokirai>
roc: you need to use vertical metrics, use vertical glyphs, and some layout features will be vertical specific

[21:46:01.0000] <roc>
right so some layout features need to use vertical forms

[21:47:00.0000] <nattokirai>
so the natural thing is to key off writing-mode

[21:47:01.0000] <nattokirai>
plus you have to deal with full-width/half-width issues to do it right

[21:48:00.0000] <nattokirai>
i.e. in horizontal mode use proportional glyphs but in vertical use full-width

[21:48:01.0000] <roc>
but at least you start from a position where everything is well-defined, and then you add features or specialize behaviour in some places to get better results

[21:49:00.0000] <nattokirai>
you can force that down into the content but it's really more styling

[21:50:00.0000] <nattokirai>
yeah, i think the best way forward is to use writing-mode

[21:50:01.0000] <nattokirai>
and add some property *values* like start, end to text-align

[21:50:02.0000] <nattokirai>
but stay away from "logical" properties like margin-before/after/start/end

[21:50:03.0000] <roc>
yeah

[21:51:00.0000] <roc>
did anyone come up with good use-cases for being able to style content independent of the orientation?

[21:51:01.0000] <nattokirai>
it's an ePub thing

[21:51:02.0000] <nattokirai>
ePub readers allow a user to flip orientation

[21:52:00.0000] <nattokirai>
horizontal vs. vertical

[21:52:01.0000] <roc>
they could have two stylesheets?

[21:52:02.0000] <nattokirai>
yes but the claim is that's too much of a burden

[21:53:00.0000] <roc>
how about a pseudo-class on the root element for the "master" writing-mode?

[21:53:01.0000] <nattokirai>
yes, i think that's what murata-san is thinking about now

[21:53:02.0000] <nattokirai>
a 'principal-writing-mode' or something like that

[21:53:03.0000] <roc>
really need some examples

[21:54:00.0000] <annevk>
then you have a selector directly depending on a property

[21:54:01.0000] <roc>
of content that can be viewed in either mode and that's burdensome to style with multiple rules or stylesheets

[21:54:02.0000] <roc>
annevk: not if you have the UA set that psuedo-class externally

[21:54:03.0000] <roc>
so the same thing that sets the initial writing-mode also sets the selector state

[21:55:00.0000] <annevk>
so if you then overwrite it it's okay?

[21:55:01.0000] <annevk>
seems kind of odd, but I guess you only need it for such a flip scenario

[21:56:00.0000] <nattokirai>
roc: http://lists.w3.org/Archives/Public/www-style/2010Jun/0133.html

[21:56:01.0000] <nattokirai>
roc: examples of Latin text and numerals in Japanese text layout

[21:57:00.0000] <roc>
yeah

[21:57:01.0000] <roc>
I've seen similar stuff in some Chinese papers

[21:57:02.0000] <roc>
note though

[21:57:03.0000] <roc>
some of what they do is probably "we don't know what we're doing" rather than "it really needs to be done this way"

[21:57:04.0000] <nattokirai>
heh

[21:58:00.0000] <roc>
especially where they're being inconsistent!

[21:58:01.0000] <nattokirai>
well, there are absolute rules and fuzzy rules, as with anything

[21:58:02.0000] <nattokirai>
distinguishing between the two is a subtle art i think

[22:03:00.0000] <nattokirai>
MikeSmith: so what's the event in Sapporo?

[22:09:00.0000] <MikeSmith>
nattokirai: gimme a few minutes, I'll get you the details

[23:25:00.0000] <hsivonen>
interesting change of attitude at epub/DAISY

[23:54:00.0000] <MikeSmith>
hsivonen: yeah

[23:54:01.0000] <MikeSmith>
and encouraging

[00:16:00.0000] <hsivonen>
MikeSmith: whether it's encouraging depends on whether it results in requirements that are off-focus for the Web being brought into the HTML WG

[00:17:00.0000] <hsivonen>
but how can a group be concerned that CSS 2.1 might not be stable enough and at the same time be wanting extensibility

[00:17:01.0000] <hsivonen>
?

[00:17:02.0000] <hsivonen>
isn't extensibility a gaping loophole of instability?

[00:23:00.0000] <MikeSmith>
I'm not even sure what they mean by "extensibility"

[00:24:00.0000] <MikeSmith>
or what anybody does

[00:25:00.0000] <othermaciej>
who thinks CSS 2.1 is not stable enough?

[00:25:01.0000] <MikeSmith>
Murata-san did mention something to me about a particular vocabulary from another vocab that there is some use case for including in ePub files

[00:26:00.0000] <MikeSmith>
to which I said I'd like to know the particular use case

[00:26:01.0000] <MikeSmith>
and that it may well be that addressing that use case does not require namespace support in order to include some existing vocab

[00:26:02.0000] <MikeSmith>
and may not even require that particular vocab at all

[00:29:00.0000] <MikeSmith>
hsivonen: and yeah, one would think that for a format like this especially, where you have particularly constrained devices that it's meant to be deployed on, a mechanism for arbitrary extensibility would not be a great way to ensure interoperability and stability

[00:29:01.0000] <MikeSmith>
but as always, it'd help for everybody to keep an open mind

[00:30:00.0000] <MikeSmith>
othermaciej: I am told there are some Apple folk who are in the ePub 2.1 WG, but in looking through the names of the people who attended the f2f, I don't see names of anybody familiar from Webkit team

[00:31:00.0000] <othermaciej>
I can try to find out who if anyone is involved from Apple

[00:31:01.0000] <MikeSmith>
othermaciej: there's a long list of attendees here:

[00:31:02.0000] <MikeSmith>
http://code.google.com/p/epub-revision/wiki/F2F201006Agenda

[00:32:00.0000] <MikeSmith>
damn, that really is a long list, too

[00:32:01.0000] <MikeSmith>
60 people

[00:32:02.0000] <othermaciej>
I didn't recognize any of the names

[00:32:03.0000] <othermaciej>
that is indeed a long list

[00:33:00.0000] <hsivonen>
othermaciej: Gregory's comment seemed to imply instability of CSS 2.1, but maybe I was reading too much into it

[00:34:00.0000] <hsivonen>
then in the other end of the spectrum there's the participant who'd just go with what WebKit does

[00:35:00.0000] <hsivonen>
MikeSmith: is the Michael Smith on the participant list someone else who doesn't have (tm)?

[00:35:01.0000] <MikeSmith>
hsivonen: it's someone who's not me :)

[00:36:00.0000] <hsivonen>
ok

[00:36:01.0000] <MikeSmith>
and also someone who's probably not the same Michael Smith who is a Google Chrome product manager

[00:36:02.0000] <MikeSmith>
hsivonen: about the Webkit mentions, roc commented on that earlier

[00:36:03.0000] <MikeSmith>
/me notices rolandsteiner 

[00:37:00.0000] <MikeSmith>
rolandsteiner: we been having some discussion here today about the ePub 2.1 work

[00:37:01.0000] <MikeSmith>
which includes some work to try to address "global language" needs, and vertical text in particular

[00:37:02.0000] <othermaciej>
someone seemed to think MathML was a reason for extensibility, because WebKit doesn't support it yet (at least not complete or on by default)

[00:38:00.0000] <MikeSmith>
rolandsteiner: if you're interested, see scrollback and/or http://krijnhoetmer.nl/irc-logs/whatwg/20100628

[00:38:01.0000] <MikeSmith>
hamaji: ↑ also

[00:39:00.0000] <othermaciej>
it is quite surprising how many times webkit is mentioned by name in the minutes there

[00:39:01.0000] <MikeSmith>
othermaciej: hmm, OK.. well that just seems like confusion about what extensibility is about

[00:39:02.0000] <MikeSmith>
othermaciej: one guy who mentions it a few time seems to be at Adobe

[00:39:03.0000] <MikeSmith>
"Ric" in the minutes

[00:40:00.0000] <hsivonen>
is Bill McCoy the same Bill McCoy as the one who participated briefly on the WHATWG list in 2005? http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2005-January/002784.html

[00:40:01.0000] <MikeSmith>
Ric = Ric Wright, Adobe

[00:41:00.0000] <othermaciej>
someone in those minutes indirectly says the HTML5 standards process is "highly politicized"

[00:41:01.0000] <othermaciej>
"CSS3 != HTML5. The specification process for CSS3 modules is not highly-politicized and many modules have seen renewed interest in recent weeks."

[00:41:02.0000] <MikeSmith>
hsivonen: I would guess so, yes - http://blogs.adobe.com/billmccoy/

[00:42:00.0000] <MikeSmith>
http://billmccoybooks2.blogspot.com/

[00:42:01.0000] <othermaciej>
someone else says: "to all intents and purposes HTML5 is whatever is in Webkit nightlies, you are deluding yourself if you are waiting for W3C"

[00:42:02.0000] <othermaciej>
(that's Ric W I guess?)

[00:42:03.0000] <MikeSmith>
othermaciej: is the HTML5 standards process not highly politicized?

[00:42:04.0000] <hsivonen>
/me notes that the WHATWG email was before the Macromedia merger when Adobe has the Adobe Intelligent Document Platform thing going on with PDF and Reader

[00:43:00.0000] <hsivonen>
s/has/had/

[00:43:01.0000] <MikeSmith>
and anybody who thinks that there's not some politics in the CSS WG as well obviously doesn't know much about that group

[00:43:02.0000] <othermaciej>
MikeSmith: I decline to comment on the merits of that statement one way or the other, for obvious political reasons

[00:43:03.0000] <MikeSmith>
heh

[00:43:04.0000] <othermaciej>
standards are politics

[00:43:05.0000] <MikeSmith>
amen

[00:44:00.0000] <MikeSmith>
anybody who believe otherwise is fooling themselves

[00:44:01.0000] <othermaciej>
the moment independent parties need to agree on something, you are fundamentally engaged in a political exercise

[00:44:02.0000] <MikeSmith>
yep

[00:45:00.0000] <MikeSmith>
and if you just instead sit around agreeing with each other, that's a possible sign that you are not talking about anything that anybody else cares about

[00:45:01.0000] <othermaciej>
or as wikipedia would put it, "Politics is a process by which groups of people make collective decisions."

[00:45:02.0000] <MikeSmith>
othermaciej: I disagree with you about that

[00:45:03.0000] <MikeSmith>
(sorry, I just figured I needed to throw in some token disagreement to support my point)

[00:45:04.0000] <othermaciej>
MikeSmith: no, you!

[00:46:00.0000] <MikeSmith>
OK, clearly I will only be victimized and ridiculed here -- I will now go off to start a different channel

[00:46:01.0000] <MikeSmith>
or actually, I think I will instead just do all my commenting on twitter

[00:47:00.0000] <MikeSmith>
or as comments on other people's blogs

[00:49:00.0000] <MikeSmith>
for some reason that seems like a better way than the strategy that has so for resulted in seeing actual features getting implemented all over the place in shipping browsers

[00:49:01.0000] <MikeSmith>
but hey, who cares about stuff being implemented

[00:49:02.0000] <MikeSmith>
anyway, I digress

[00:51:00.0000] <MikeSmith>
meanwhile, abarth|InBody wakes up, parachutes in, and hits the ground running

[00:52:00.0000] <MikeSmith>
if I don't have a conformant HTML5 tree builder in Webkit by the end of this week, I'm going to withdraw my ambassador

[00:53:00.0000] <othermaciej>
heh

[00:53:01.0000] <nessy>
lol

[00:53:02.0000] <MikeSmith>
or file a formal diplomatic protest with Ban Ki Moon

[00:59:00.0000] <MikeSmith>
hsivonen: http://www.webpaper.net/blog seems to be Bill McCoy's current venture

[01:01:00.0000] <MikeSmith>
"And that Adobe would be passing around back-slaps for an Apple-forced port from ActionScript to Objective C, calling that "innnovation" is, well, kind of sad."

[01:01:01.0000] <MikeSmith>
he has a way with words, that's for sure

[01:10:00.0000] <abarth|InBody>
MikeSmith: we only recently where able to parse the document "<html>"

[01:11:00.0000] <abarth|InBody>
MikeSmith: having a fully conformant tree builder by the end of the week is a bit ... optimistic :)

[01:15:00.0000] <hsivonen>
MikeSmith: it's kinda sad that so many people want to put paper-like stuff on the Web instead of putting Web-like stuff on the Web

[01:19:00.0000] <hsivonen>
wow. if Adobe's Reader Mobile SDK doesn't support the "enterprise" features of PDF, it seems like an admission that PDF peaked at 1.4

[01:20:00.0000] <hsivonen>
PDF 1.4 seems to be what everyone outside Adobe was targeting anyway

[01:24:00.0000] <Dashiva^2>
What are the enterprise features?

[01:25:00.0000] <othermaciej>
what version of PDF has those?

[01:25:01.0000] <hsivonen>
Dashiva^2: "interactive forms, security of the Livecycle PolicyServer variety, JavaScript, and digital signatures"

[01:25:02.0000] <hsivonen>
quoting from http://billmccoybooks2.blogspot.com/2010/01/q-what-pdf-version-for-ebooks.html

[01:25:03.0000] <MikeSmithX>
abarth|InBody: well, I'm an optimist.. if I weren't, I guess I'd be doing something else for a living

[01:26:00.0000] <hsivonen>
also, the later versions of PDF have *totally* off-focus CAD features

[01:26:01.0000] <MikeSmith>
abarth|InBody: anyway, I'm just trying to giving you some grief.. I realize it'll take a while 

[01:26:02.0000] <MikeSmith>
e.g., maybe two weeks instead of one

[01:27:00.0000] <hsivonen>
too bad Adobe's PDF went bad after 1.4

[01:28:00.0000] <hsivonen>
now if you want some good post-1.4 features, there's no easy shorthand for talking about 1.4 + the good parts of later PDF versions

[01:29:00.0000] <hsivonen>
(I count the ability to apply the deflate algorithm to everything a good part albeit a marginal one)

[01:32:00.0000] <Dashiva^2>
Javascript seems more like a bug than a feature in PDF...

[01:32:01.0000] <hsivonen>
Dashiva^2: indeed

[01:33:00.0000] <hsivonen>
btw, what's for Windows what Preview is for Mac and Evince is for Gnome?

[01:34:00.0000] <Dashiva^2>
Foxit? Not sure what Preview and Evince "are"

[01:34:01.0000] <hsivonen>
Dashiva^2: Preview and Evince are faster than Adobe Reader and (I believe) have fewer serious exploitable security holes or are at least less actively being exploited

[01:35:00.0000] <abarth|InBody>
hsivonen: notepad?

[01:35:01.0000] <Dashiva^2>
Then probably Foxit

[01:36:00.0000] <hsivonen>
abarth|InBody: I meant for users who aren't comfortable with only viewing source in the PDF case

[01:36:01.0000] <MikeSmith>
hsivonen: about the wanting paper-like stuff on the Web problem -- yeah, it seems like some people maybe "can't see the forest for the trees" (or whatever metaphor) -- that is, they seem overly focused on trying to replicate the print-reading experience on the Web, and not nearly enough focused on trying to see how to best exploit the richer features of the Web to deliver and even better reading experience

[01:36:02.0000] <hsivonen>
Dashiva^2: thanks

[01:36:03.0000] <othermaciej>
MikeSmith: making good flexible layouts is hard

[01:37:00.0000] <hsivonen>
is the business model of Foxit Reader to be a gateway drug for their PDF creation stuff?

[01:37:01.0000] <MikeSmith>
well, there are some features that we currently have only in print that are very important

[01:37:02.0000] <othermaciej>
it's also hard to get truly top quality typography about of browser-grade text layout engines

[01:37:03.0000] <hsivonen>
othermaciej: that's partly because raw speed is overrated

[01:38:00.0000] <hsivonen>
/me gestures at the direction of WebKit on Mac not doing ligatures for Latin text where Cocoa text views and Gecko do

[01:38:01.0000] <MikeSmith>
othermaciej, hsivonen - e.g., vertical text is very important part of the user experience for reading in Japan .. lack of it for normal reading of e-books is a pretty big deal

[01:38:02.0000] <othermaciej>
raw speed of text layout?

[01:38:03.0000] <MikeSmith>
othermaciej: top-grade typography is not a high priority.. but being able to read vertical text is

[01:39:00.0000] <hsivonen>
othermaciej: why else would you not have the usual Mac ligature experience in WebKit?

[01:39:01.0000] <othermaciej>
I'm not even thinking specifically of ligatures, necessarily, I'm thinking of things like good quality justification (for which hyphenation is one prerequisite)

[01:39:02.0000] <hsivonen>
/me wants hyphenation in Web engines

[01:40:00.0000] <othermaciej>
actually we just landed some hyphenation support in webkit

[01:40:01.0000] <othermaciej>
not sure which ports support it right now though

[01:40:02.0000] <othermaciej>
(it's opt-in with a CSS property)

[01:40:03.0000] <hsivonen>
othermaciej: sure, but isn't the lack of ligatures an indication that typography is sacrificed for speed benchmarks?

[01:40:04.0000] <othermaciej>
I can imagine supporting contextual forms and kerning pairs in latin scripts eventually, either through an opt-in CSS property or because we find a way to make the common case very fast

[01:41:00.0000] <othermaciej>
it's true that we care more about page load speed than ligatures in latin scripts (or at least have historically)

[01:41:01.0000] <hsivonen>
I think at least at some point there was a hidden pref in Safari for turning on better typography

[01:41:02.0000] <othermaciej>
WebKit's text rendering (in the fast-path case) is much much much much faster than any of the native APIs on Mac OS X

[01:42:00.0000] <othermaciej>
you can force everything to go through the slow path

[01:42:01.0000] <hsivonen>
othermaciej: but also uglier

[01:42:02.0000] <othermaciej>
I think our eventual strategy has to be to extend the fast-path code to handle things like kerning and ligatures

[01:42:03.0000] <hsivonen>
"Qu" on my site looks worse in WebKit than it does in Gecko on Mac

[01:43:00.0000] <othermaciej>
a lot of very visually picky people use Safari every day

[01:43:01.0000] <othermaciej>
I am inclined to say our choice of tradeoffs is reasonable

[01:43:02.0000] <hsivonen>
in Gecko, the 'Q' extends nicely under the 'u'

[01:44:00.0000] <hsivonen>
othermaciej: too bad there aren't many ligatures in Helvetica, so we don't got Gruber to call you on this :-(

[01:44:01.0000] <othermaciej>
Qu is of course a kerning pair not a ligature

[01:45:00.0000] <othermaciej>
(as far as I can tell anyway, and in the default font

[01:45:01.0000] <hsivonen>
I thought the Q was a contextual alternative glyph

[01:45:02.0000] <hsivonen>
it's not just kerning

[01:45:03.0000] <hsivonen>
(I'm not suggesting it were a ligature)

[01:46:00.0000] <othermaciej>
I'm looking in Times 16 point and I don't see indications of an alternate glyph

[01:46:01.0000] <othermaciej>
but then, I'm not the most visually acute person in the world

[01:46:02.0000] <hsivonen>
othermaciej: the text is supposed to be in @font-face Linux Libertine

[01:46:03.0000] <hsivonen>
/me wonders why WebKit isn't applying the font 

[01:46:04.0000] <othermaciej>
oh, on your site in particular

[01:47:00.0000] <othermaciej>
I thought you were referring to Qu in general

[01:47:01.0000] <othermaciej>
mea culpa

[01:47:02.0000] <hsivonen>
othermaciej: I meant "Qu" on my site in particular

[01:48:00.0000] <othermaciej>
it does however look like Safari is not applying the font (at least for me) on your site

[01:48:01.0000] <hsivonen>
othermaciej: I guess that's a bug in WebKit/Safari, then

[01:48:02.0000] <hsivonen>
othermaciej: I've had people point out to me that the font isn't getting applied

[01:49:00.0000] <hsivonen>
othermaciej: but debugging it is hard, because Apple doesn't allow me to virtualize a vanilla copy of OS X

[01:49:01.0000] <hsivonen>
and if the font exists on any mounted volume, Safari finds it

[01:49:02.0000] <hsivonen>
which is kinda creepy

[01:50:00.0000] <hsivonen>
and makes it hard to investigate the problem on a system I actually use

[01:50:01.0000] <othermaciej>
I believe the font server is going to get less aggressive with hunting for fonts

[01:51:00.0000] <othermaciej>
anyway, I can see the alternate glyph looking at your page in Firefox (that might actually be a ligature)

[01:51:01.0000] <hsivonen>
othermaciej: does WebKit require a particular MIME type for fonts?

[01:51:02.0000] <othermaciej>
it's also true that Firefox kerns Qu tighter than Safari does in Times

[01:51:03.0000] <othermaciej>
probably, but I don't know for sure

[01:51:04.0000] <zcorpan_>
hmm, the Q doesn't extend under the u in opera

[01:52:00.0000] <othermaciej>
I notice all your fonts have a .gz extension which makes me wonder if you end up serving gzip files rather than truetype files with gzip transfer-encoding

[01:52:01.0000] <hsivonen>
othermaciej: IIRC, they are served as text/plain with gzip encoding

[01:52:02.0000] <othermaciej>
if I download one of the fonts, FontBook opens it, so something in the system is sniffing it

[01:52:03.0000] <othermaciej>
but that might not happen in the font loading path

[01:52:04.0000] <hsivonen>
maybe I should try font/ttf or application/octet-stream

[01:53:00.0000] <othermaciej>
I can definitely see that Safari loads all the font files

[01:53:01.0000] <hsivonen>
othermaciej: surely safari should ungzip those resources before the font subsystem gets to sniff?

[01:53:02.0000] <othermaciej>
if it's a transfer-encoding then it will

[01:54:00.0000] <othermaciej>
if it's just gzipped, then Safari will still decompress while downloading and sniff when done and/or based on extension

[01:54:01.0000] <Dashiva^2>
I find myself wanting a native hash function in JS

[01:54:02.0000] <Dashiva^2>
Wasn't that supposed to be in es5?

[01:54:03.0000] <othermaciej>
I believe the ES committee has been persuaded to add such a thing

[01:54:04.0000] <othermaciej>
but I fear the next ES version may take forever

[01:54:05.0000] <hsivonen>
Content-Encoding: gzip

[01:57:00.0000] <othermaciej>
I guess I will debug it sometime, but not tonight

[01:57:01.0000] <othermaciej>
from reading the code it seems like it should ignore the MIME type

[01:58:00.0000] <abarth|InBody>
othermaciej: ignoring mime type = sadness in the future

[01:59:00.0000] <othermaciej>
abarth|InBody: probably! - might be too late for this one though

[02:00:00.0000] <othermaciej>
I'm not sure there even is a correct mime type for opentype fonts

[02:00:01.0000] <hsivonen>
I'm pretty sure there isn't

[02:00:02.0000] <hsivonen>
all hooray for IANA

[02:00:03.0000] <zcorpan_>
iirc annevk tried to get font/ttf but failed

[02:00:04.0000] <hsivonen>
it's possible that WebKit has been ignoring those fonts since at least February 2009

[02:01:00.0000] <hsivonen>
but I haven't gotten around to investigating and filing a bug, because I was unable to investigate on the instances of Mac OS X I actually use myself

[02:01:01.0000] <othermaciej>
I don't think I have ever seen your site in the correct font before

[02:01:02.0000] <othermaciej>
zcorpan_: why did he fail?

[02:02:00.0000] <zcorpan_>
othermaciej: not sure. i think he was the only one who cared, or something

[02:02:01.0000] <othermaciej>
I'm surprised there is not even a "font" top-level content type

[02:38:00.0000] <roc>
hsivonen: raw performance is not overrated

[02:38:01.0000] <roc>
raw performance is everything

[02:39:00.0000] <roc>
the CAD stuff in Acroread is actually licensed from a New Zealand company

[02:40:00.0000] <roc>
I think we could get completely adequate book-reading typography in browsers with a combination of optimizations and CSS opt-in

[02:42:00.0000] <hsivonen>
roc: if raw perf is everything, why does Gecko do nice things for Latin text?

[02:42:01.0000] <roc>
on Windows and Linux, we actually don't

[02:43:00.0000] <hsivonen>
oh. :-(

[02:43:01.0000] <roc>
we have fast paths that bypass the shaper

[02:43:02.0000] <roc>
on Mac, well, we never got around to implementing the fast paths

[02:43:03.0000] <roc>
also the performance difference was never measured to be stunningly huge on Windows and Linux

[02:44:00.0000] <hsivonen>
roc: will the feature be available for opt-in once harfbuzz is used on all platforms?

[02:44:01.0000] <roc>
it's available for opt-in right now

[02:44:02.0000] <roc>
users can opt in via about:config, authors can opt in via text-rendering:optimizeLegibility

[02:44:03.0000] <roc>
(the latter also works on some Webkit ports, as I understand it)

[02:45:00.0000] <hsivonen>
nice.

[02:45:01.0000] <roc>
with Harfbuzz we're going to carefully measure the performance difference. Right now, Harfbuzz has no fast path

[02:45:02.0000] <hsivonen>
it sucks, though, that the default is optimizeBenchmarks

[02:45:03.0000] <roc>
it's not just benchmarks

[02:45:04.0000] <hsivonen>
what's the deal with camelCase in CSS?

[02:46:00.0000] <roc>
SVG screwed up

[02:46:01.0000] <hsivonen>
:-(

[02:46:02.0000] <roc>
Harfbuzz is considerably more optimizable of course, so I'm quite hopeful we can get away without any fast paths

[02:47:00.0000] <roc>
and we cache shaped text at the word level quite aggressively

[02:48:00.0000] <kling>
roc: harfbuzz performance is on the ToDo whiteboard behind me :]

[02:48:01.0000] <roc>
also, with Harfbuzz we can move most text shaping off the main thread if it's still showing up in profiles

[02:48:02.0000] <roc>
kling: get it off the whiteboard and check it in :-)

[02:48:03.0000] <annevk>
othermaciej, nobody seemed to care enough and then browsers shipped and all was lost

[02:49:00.0000] <kling>
roc: getting there eventually. memory usage is the larger problem ATM

[02:49:01.0000] <othermaciej>
annevk: still sorta seems worth doing but it seems like the MIME registry operates too slowly for Web reality

[02:50:00.0000] <roc>
interesting. if there are any regressions there we'll see them when Jonathan flips the switch

[02:52:00.0000] <hsivonen>
/me wonders what happens when a benchmark starst using optimizeLegibility...

[02:52:01.0000] <roc>
then we optimize it!

[02:53:00.0000] <hsivonen>
there should be an acid test measuring legibility and giving browsers points for typographic nice things

[02:55:00.0000] <hsivonen>
wow. Gecko has special UTF-8 decoder code for arm

[02:55:01.0000] <roc>
yes

[02:55:02.0000] <roc>
shaping text runs across element boundaries should be worth five million points

[02:56:00.0000] <roc>
but since that's not a "CSS feature", I don't expect to see it

[03:00:00.0000] <zcorpan_>
MikeSmith: hmm. i'm not sure it's a good idea to make <annotation-xml encoding=...> affect the tree building

[03:01:00.0000] <annevk>
http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2005-January/002852.html is somewhat interesting (via that link from hsivonen earlier) on an attempted definition of the platform

[03:02:00.0000] <hsivonen>
/me notes that Hixie had xml:id on the list. boo

[03:03:00.0000] <hsivonen>
whoa. XML Events is on Hixie's list too

[03:03:01.0000] <hsivonen>
why?

[03:07:00.0000] <annevk>
maybe because of SVG and Opera

[03:07:01.0000] <annevk>
I was kind of surprised by those too

[03:10:00.0000] <MikeSmith>
zcorpan_: why would it not be a good idea to make <annotation-xml encoding=...> affect the tree building?

[03:10:01.0000] <zcorpan_>
MikeSmith: seems complicated and i still don't know the use case :)

[03:10:02.0000] <zcorpan_>
MikeSmith: how do you parse the encoding="" attribute? parsing mime types is not defined

[03:11:00.0000] <hsivonen>
I hope we aren't going to parse attribute values beyond comparing with a constant

[03:11:01.0000] <MikeSmith>
zcorpan_: what hsivonen said

[03:12:00.0000] <MikeSmith>
we have existing HTML attributes that are defined in HTML5 as taking an MIME type

[03:12:01.0000] <annevk>
why is <annotation-xml> back in the picture?

[03:12:02.0000] <MikeSmith>
or at least one such attribute

[03:12:03.0000] <zcorpan_>
MikeSmith: they don't affect parsing

[03:12:04.0000] <MikeSmith>
annevk: because there is existing content that has HTML in annotation-xml

[03:13:00.0000] <MikeSmith>
and the parsing algorithm currently barfs on HTML in annotation-xml

[03:13:01.0000] <MikeSmith>
in a particularly ungraceful way

[03:13:02.0000] <zcorpan_>
where's the existing content?

[03:15:00.0000] <MikeSmith>
zcorpan_: some examples at Mozilla MDC, at least

[03:15:01.0000] <MikeSmith>
and I suspect that there are a few people who have created content base on those

[03:16:00.0000] <zcorpan_>
mozilla mdc is a wiki, it's easier to fix than changing the spec :)

[03:16:01.0000] <MikeSmith>
heh

[03:16:02.0000] <MikeSmith>
anyway, the reason those were created to begin with is that apparently people believe that is the way to include foreign content in MathML

[03:16:03.0000] <MikeSmith>
they have come to believe that for whatever reason

[03:17:00.0000] <MikeSmith>
perhaps based on a misreading of the MathML spec

[03:18:00.0000] <zcorpan_>
http://www.google.se/search?q=%22annotation-xml%22+site%3Adeveloper.mozilla.org

[03:19:00.0000] <zcorpan_>
http://www.google.com/codesearch?q=%3Cannotation-xml+encoding%5Cs%2A%3D%5Cs%2A%28%22%7C%27%29text%2Fhtml&hl=en&btnG=Search+Code

[03:20:00.0000] <MikeSmith>
zcorpan_: point taken

[03:21:00.0000] <MikeSmith>
I will personally not be very unhappy if the spec were changed to just drop annotation-xml content on the floor

[03:22:00.0000] <MikeSmith>
but if we keep the current SVG-in-annotation-xml support, I can understand that users might find it frustrating that the can include SVG but not HTML

[03:23:00.0000] <MikeSmith>
hey, there is even an example in the MathML3 spec of HTML in annotation-xml

[03:23:01.0000] <zcorpan_>
google code searching for just "<annotation-xml" i see mostly mathml in there, but also openmath and "<tt><code>" but that's without any encoding attribute

[03:23:02.0000] <MikeSmith>
/me goes to find it

[03:23:03.0000] <MikeSmith>
at this point, the case of HTML without the encoding attribute is not one I'd suggest trying to support

[03:25:00.0000] <Lachy>
Philip`, do you have any data analysing the use of <meta http-equiv="Content-Language"> compared with HTTP Content-Language?

[03:25:01.0000] <zcorpan_>
http://www.w3.org/TR/MathML3/chapter5.html#mixing.elements.annotation.xml

[03:25:02.0000] <hsivonen>
oh right. there are polls that need objecting to :-(

[03:26:00.0000] <MikeSmith>
zcorpan_: http://www.w3.org/TR/MathML3/chapter5.html#id.5.2.3.2

[03:26:01.0000] <Lachy>
Specifically, I'm interested in data that shows whether or not sites that use the <meta> element for this also use Content-Language in HTTP. (My assumption is that they don't, and thus the meta is indeed not widely used for server side processing, as Roy claimed)

[03:27:00.0000] <hsivonen>
zcorpan_: I'm not fond of the idea of supporting <html>, <head> or <body> there...

[03:27:01.0000] <zcorpan_>
hsivonen: i agree

[03:28:00.0000] <MikeSmith>
zcorpan_: that is admittedly not a great example. but because it's in the MathML spec, I think it could also lead users to assume it should be supported in UAs

[03:28:01.0000] <MikeSmith>
anyway, I will send a message today to the public MathML list asking for use cases

[03:28:02.0000] <annevk>
removing annotation-xml entirely wfm

[03:29:00.0000] <MikeSmith>
wfm too personally

[03:29:01.0000] <MikeSmith>
not sure it will be a popular choice among end users, though

[03:30:00.0000] <Philip`>
Lachy: All I can see is http://philip.html5.org/data/meta-http-equiv-content-language-values.txt

[03:31:00.0000] <Philip`>
Lachy: (and http://philip.html5.org/data/http-headers.txt for header counts)

[03:31:01.0000] <hsivonen>
Dashiva^2: I can now see what the business model for the free Foxit reader is

[03:32:00.0000] <hsivonen>
Dashiva^2: seems to be installware

[03:32:01.0000] <Philip`>
Lachy: (Those are from different sets of pages, I think)

[03:32:02.0000] <hsivonen>
Dashiva^2: it defaults to changing the default browser search engine to Ask.com, installing a toolbar of somekind somewhere and adding an eBay shortcut to somewhere

[03:33:00.0000] <hsivonen>
Dashiva^2: not to suggest that Adobe Reader were any better

[03:35:00.0000] <Lachy>
Philip`, ok. No worries then. I probably don't need it, but would might have been a nice addition to my argument for ISSUE-88

[03:37:00.0000] <Philip`>
Lachy: I could extract the raw content-language header/meta data from the latest page set pretty easily, and you could find some way to analyse it (count the number of pages in one set or the other or both)

[03:40:00.0000] <Lachy>
Philip`, could you extract it in a way that associates the <meta> elements used on the page with the value of the Content-Language used, if any?

[03:42:00.0000] <Lachy>
so, e.g. I can see if example.com uses "Content-Language: en, fr" in the HTTP, but decalares <meta http-equiv="Content-Language" content="en"> in the page.

[03:44:00.0000] <Lachy>
my assumption is that if, as I suspect, servers aren't using server side processing to extract the meta Content-Language to and set the HTTP headers, then we'll see a significant portion of pages using the the meta element, either without the HTTP header or with different values.

[03:47:00.0000] <Philip`>
Lachy: I think I could do that later today, if I don't forget

[03:49:00.0000] <Lachy>
thanks.

[04:00:00.0000] <Dashiva^2>
hsivonen: Yeah, it's become a lot worse recently

[04:00:01.0000] <Dashiva^2>
Like three different screens you have to disable stuff on

[04:09:00.0000] <boblet>
anyone have thoughts about RDFa’s @datatype? I’m trying to think of a situation where it’d be applicable to have more than one kind of data type for an @itemprop that wouldn’t be specced in advance (similar to datetime’s conformance rules)

[04:10:00.0000] <boblet>
yoshiaki: hello :)

[04:10:01.0000] <yoshiaki>
boblet, hello!!

[04:16:00.0000] <MikeSmith>
yoshiaki: hey!

[04:16:01.0000] <yoshiaki>
Hey! MikeSmith!!

[04:17:00.0000] <MikeSmith>
yoshiaki is an intern at W3C Japan and a doctoral student at Keio U. here

[04:17:01.0000] <MikeSmith>
and a great guy

[04:17:02.0000] <MikeSmith>
so be nice to him!

[04:18:00.0000] <MikeSmith>
or he will kill you

[04:18:01.0000] <yoshiaki>
Nice to meet you, all. I am Yoshiaki Fukami.

[04:18:02.0000] <kennyluck>
Yoshiaki's famous quote, "you need different protocol to talk to different people" (ref. old-style Japanese companies)

[04:22:00.0000] <jgraham>
MikeSmith: Is on-on-one combat a requirement for W3C internships these days?

[04:25:00.0000] <MikeSmith>
jgraham: yeah, the test is that all interns have to be able to beat me up

[04:25:01.0000] <yoshiaki>
I will never kill person who loves world wide web :) I am very happy to join this community.

[04:25:02.0000] <hsivonen>
boblet: as I understand it, @datatype isn't there to support a use case but it's there for RDF completeness

[04:26:00.0000] <jgraham>
MikeSmith: Yeah, but I was asking about combat

[04:26:01.0000] <hsivonen>
boblet: RDFa puts the ability to encode arbitrary RDF above use cases

[04:26:02.0000] <jgraham>
boblet: I thought the usecase was something like being able to have a distance and specify a unit of miles or a unit of picometres

[04:27:00.0000] <jgraham>
But that might have been something else

[04:27:01.0000] <MikeSmith>
jgraham: you probably don't really want me to go into details about the combat.. it's not pretty

[04:28:00.0000] <MikeSmith>
hsivonen: I recall that Manu recently said the HTML+RDFa spec is a strict superset of the Microdata spec

[04:28:01.0000] <hsivonen>
MikeSmith: yeah, he said so on twitter. I haven't verified the claim.

[04:29:00.0000] <jgraham>
boblet: (if that is the use case it is a bad idea; mixing units like that is a really bad idea)

[04:29:01.0000] <Dashiva^2>
Yeah, data types should be handled by the ontology, IMO

[04:29:02.0000] <hsivonen>
/me wonders how datatypes ended up on the RDF level in the first place

[04:30:00.0000] <jgraham>
hsivonen: Did you look at the API? I let my eyes slide across it for a few fractions of a second and it looked somewhat complex, but I didn't look hard enough to form a proper opinion

[04:30:01.0000] <hsivonen>
jgraham: I haven't reviewed the latest API draft properly

[04:30:02.0000] <MikeSmith>
well, it sounds like it is an intentional design goal, at least. given that, it would seem like it might be useful for the HTML+RDFa spec to include a section explaining to implementors and users which features to drop in order can to create Microdata-friendly processors and documents

[04:30:03.0000] <hsivonen>
jgraham: I searched it for the string "Microdata" and concluded that the document doesn't spell it out to the reader that the API is supposed to support Microdata

[04:31:00.0000] <jgraham>
Yeah, iirc it just said something like "this could also be used for microdata"

[04:31:01.0000] <MikeSmith>
perhaps that HTML+RDFa spec could define a Microdata-compatible processor as one of its conformance classes

[04:32:00.0000] <MikeSmith>
and then the spec could specify that Microdata-compatible processors are not required to implement support for particular features

[04:34:00.0000] <hsivonen>
MikeSmith: if done properly, wouldn't the result be a restatement of the Microdata spec?

[04:35:00.0000] <hsivonen>
is Object a magic type in WebIDL?

[04:35:01.0000] <jgraham>
hsivonen: What do you mean "magic type"?

[04:35:02.0000] <hsivonen>
jgraham: I mean something special

[04:36:00.0000] <hsivonen>
/me sees http://dev.w3.org/2006/webapi/WebIDL/#idl-object

[04:36:01.0000] <hsivonen>
but that's lower case

[04:36:02.0000] <hsivonen>
but the last sentence of that section says Object

[04:36:03.0000] <hsivonen>
hmm

[04:37:00.0000] <jgraham>
hsivonen: What is the context?

[04:38:00.0000] <hsivonen>
jgraham: the RDFa API spec has WebIDL stuff where type is Object

[04:41:00.0000] <hsivonen>
the RDFa API spec has some non-normative advertising content

[04:41:01.0000] <hsivonen>
"RDFa 1.0 [RDFA-SYNTAX] has seen substantial growth since it became an official W3C Recommendation in October 2008. It has seen wide adoption among search companies, e-commerce sites, governments, and content management systems. There are numerous interoperable implementations and growth is expected to continue to rise with the latest releases of RDFa 1.1 [RDFA-CORE], XHTML+RDFa 1.1 [XHTML-RDFA], and HTML+RDFa 1.1 [HTML-RDFA]."

[04:42:00.0000] <hsivonen>
/me wonders if those who oppose to Hixie's political editorialization have already filed bugs about the advertising in the RDFa API spec

[04:44:00.0000] <othermaciej>
I think self-puffery is a different matter than deprecation of others

[04:44:01.0000] <othermaciej>
(though in this case it seems a little over the top for a technical specification)

[04:45:00.0000] <othermaciej>
MikeSmith: is it claimed to be a feature-wise superset, or does it literally include all the exact markup constructs of Microdata?

[04:48:00.0000] <othermaciej>
I cannot find evidence of "itemprop" or "itemscope" in any of RDFa 1.1 Core, XHTML+RDFa 1.1 or HTML+RDFa 1.1

[04:51:00.0000] <MikeSmith>
othermaciej: feature-wise, I assume. I don't think the intention it to use the exact markup constructs

[04:53:00.0000] <kennyluck>
I had the same impression as othermaciej got, and was very surprised that "itemprop" or "itemsope" is in RDFa. :)

[04:53:01.0000] <kennyluck>
s/is/was/

[04:54:00.0000] <MikeSmith>
hsivonen: yeah, I suppose it could be seen as being a restatement, if it does in fact end up being that.. I guess the difference would be that it defines additional features for those implementations that choose to support them, and those authors who choose to use them

[04:54:01.0000] <othermaciej>
if there's a subset of RDFa 1.1 that has a trivial one-to-one mapping to Microdata, that might be interesting to identify

[04:55:00.0000] <othermaciej>
it might be complicated to do that if you want to make the microdata generate the same RDF triples when converted to RDF that RDFa does

[04:56:00.0000] <kennyluck>
'''

[05:24:00.0000] <Lachy>
Does anyone know if there is any reason to avoid the WIndows 7 64bit edition in favour of 32bit? Does Windows 7 x64 still suffer from the same compatibility problems that XP x64 did?

[05:25:00.0000] <Lachy>
I suppose, I should check driver compatibility for my hardware and availability of 64bit anti-virus software, etc.

[05:26:00.0000] <Philip`>
Are you going to run any applications that need >4GB memory each?

[05:27:00.0000] <gsnedders>
Lachy: Yes, there's still no support for Win16.

[05:27:01.0000] <gsnedders>
Oh, wait, you don't care about that?

[05:27:02.0000] <hsivonen>
Philip`: I gather perf benefits from more general purpose registers are a more significant consideration than RAM

[05:27:03.0000] <jgraham>
Philip`: Used any modern web browsers recently? :p

[05:27:04.0000] <gsnedders>
Then Vista and Win7 have been fine in 64-bit mode

[05:29:00.0000] <Lachy>
are there any real benefits, given that my current hardware doesn't support more than 4GB of RAM anyway?

[05:30:00.0000] <gsnedders>
More general purpose registers

[05:30:01.0000] <Philip`>
hsivonen: There's also perf loss from using up more cache storing pointers, so I'm not sure there's a clear benefit except in a few specific applications (ones that do lots of computation but don't do it all with SSE)

[05:30:02.0000] <Lachy>
does that help much with widely deployed, older software that I have?

[05:30:03.0000] <hsivonen>
Philip`: I have been lead to believe the difference is significant on SunSpider

[05:31:00.0000] <hsivonen>
Philip`: and perf on SunSpider matter more for perception than real-world perf :-/

[05:32:00.0000] <Philip`>
Lachy: It makes no difference unless the software is recompiled as a 64-bit binary

[05:32:01.0000] <hsivonen>
/me notes "HyBI (BiDirectional or Server-Initiated HTTP)" in the IETF 78 announcement/invitation

[05:33:00.0000] <Lachy>
Philip`, that's what I thought

[05:34:00.0000] <Lachy>
although, I suppose, since my main use case is running it on my Macs, in a virtual machine, and if and when I upgrade my Mac, I'll likely get 8GB of RAM. That might be a good excuse to go for 64 bit now.

[05:35:00.0000] <Lachy>
my other use case, will be for running it on the HTPC that I'm considering building soon, but I don't know the exact hardware for that yet.

[05:35:01.0000] <Philip`>
Lachy: With 32-bit Windows and 4GB RAM and suitable kernel options, each application can only use 3GB but the whole system can use all 4GB

[05:36:00.0000] <Philip`>
Actually that works for any value of "4GB" that is <= 64GB

[05:36:01.0000] <hsivonen>
Philip`: why 3 GB?

[05:36:02.0000] <Philip`>
hsivonen: http://blogs.msdn.com/b/oldnewthing/archive/2004/08/05/208908.aspx

[05:38:00.0000] <Dashiva^2>
Not all 4 GB, surely, since some of the address space is also lost to device mapping (graphics card)

[05:39:00.0000] <hsivonen>
Philip`: interesting that they implemented 3 GB instead of doing the simpler 2 GB user / 2 GB kernel thing only

[05:39:01.0000] <Lachy>
hmm, is Windows available with a multiple-machine licence, like OS X's family pack, or will I have to buy a new full copy for every machine I have?

[05:41:00.0000] <jgraham>
So is it deliberate that html breakout elements in foreign lands only break out of one level of foreignness?

[05:42:00.0000] <Philip`>
Dashiva^2: I think what happens is that if you don't have PAE then Windows can only see 3.5GB physical address space (the rest is BIOS stuff), but if you do have PAE then it can see 64GB physical address space and can then access all 4GB of RAM plus the BIOS stuff

[05:43:00.0000] <hsivonen>
Lachy: Microsoft doesn't even allow you to install 32-bit and 64-bit in dual boot without purchasing 2 licenses

[05:43:01.0000] <Philip`>
(You probably can't have 64GB RAM then, but at that point you ought to get a 64-bit CPU)

[05:43:02.0000] <MikeSmith>
for those who've not gotten their quota of drama for the day -

[05:43:03.0000] <MikeSmith>
http://news.cnet.com/8301-30685_3-20008935-264.html

[05:43:04.0000] <hsivonen>
jgraham: I don't know

[05:43:05.0000] <Dashiva^2>
I didn't know there was a quota

[05:44:00.0000] <Philip`>
hsivonen: Why wouldn't they implement the 3GB switch, since it's presumably not terribly complex and it lets customers easily use 50% more RAM?

[05:44:01.0000] <MikeSmith>
though that article's less dramatic than some might have hoped for I suppose

[05:44:02.0000] <MikeSmith>
"oversight of the future Web isn't collapsing, it's just rebalancing"

[05:45:00.0000] <hsivonen>
Philip`: I admit that I don't know how the 3 GB switch is implemented on the CPU memory protection level, so I'll shut up now

[05:50:00.0000] <Dashiva^2>
/me notes that MikeSmith's article includes mention of the awesome HTML5 feature geolocation

[05:53:00.0000] <MikeSmith>
all good new features that have gone into the Web platform recently are by definition HTML5 features

[05:54:00.0000] <MikeSmith>
and any bad features that may have also gone in are by definition not HTML5 features

[06:01:00.0000] <hsivonen>
speaking of drama: http://lists.w3.org/Archives/Public/public-html-comments/2010Jun/0035.html

[06:08:00.0000] <boblet>
hsivonen, jgraham thanks for your input

[06:08:01.0000] <boblet>
MikeSmith: “it might be useful for the HTML+RDFa spec to include a section explaining to implementors and users which features to drop in order can to create Microdata-friendly processors and documents” I’d like that too

[06:09:00.0000] <annevk>
hsivonen, will be interesting to see how that goes

[06:11:00.0000] <boblet>
and an HTML5 version of the RDFa primer http://www.w3.org/TR/xhtml-rdfa-primer/ would be nice too

[06:38:00.0000] <roc>
now I  don't feel so bad about duking it out with other Mozilla people in public

[07:27:00.0000] <MikeSmith>
"Needs more vuvuzela."

[07:30:00.0000] <annevk>
argh

[07:30:01.0000] <annevk>
security considerations for XMLHttpRequest are a pain

[07:31:00.0000] <annevk>
basically, I don't think it needs any, but trl wants some weasel wording in there that repeats what's been said and adds some fluff that's really a way more generic problem than XMLHttpRequest

[07:31:01.0000] <annevk>
like DNS rebinding

[07:32:00.0000] <zcorpan_>
fluff should be easy

[07:32:01.0000] <annevk>
it's all easy in a way, but it doesn't make any sense to me

[07:33:00.0000] <zcorpan_>
just copy stuff from abarth|InBody's cookie spec, it has lots of stuff on security that surely applies to xhr also

[07:33:01.0000] <annevk>
I believe I can certainly write something that will make the "problem" go away, but I don't believe it will actually improve anything

[07:40:00.0000] <boblet>
the phenny bot in #swik is pretty cool

[07:40:01.0000] <boblet>
s/swik/swig/

[07:42:00.0000] <boblet>
/me is resisting the urge to be all “thanks for the interwebs” fanboi-ish to timbl who’s there

[07:44:00.0000] <mhausenblas>
/me thinks we owe timbl a lot, boblet  ... for example an open and free WWW ;)

[07:45:00.0000] <boblet>
mhausenblas: oh noes! you’ve discovered I’m part of the cabal ;-)

[07:45:01.0000] <hsivonen>
how long should it take for a response to show up on http://www.w3.org/2002/09/wbs/40318/issue-88-objection-poll/results ?

[07:45:02.0000] <annevk>
yours shows for me

[07:45:03.0000] <boblet>
mhausenblas: and yeah I definitely feel so, but I bet timbl has heard it enough times. don’t wanna be annoying

[07:46:00.0000] <hsivonen>
annevk: odd. it shows to me in a browser I didn't use for responding

[07:46:01.0000] <mhausenblas>
sure thing, boblet ;)

[07:46:02.0000] <hsivonen>
but it doesn't show in the browser I used after reloading several times

[07:46:03.0000] <annevk>
maybe some wrong caching directives

[07:46:04.0000] <annevk>
HTTP isn't easy

[08:03:00.0000] <jgraham>
Sigh. I just found out one of our developers had been unwittingly reading the /TR/ version of the spec

[08:04:00.0000] <annevk>
nothing new there :/

[08:04:01.0000] <Dashiva^2>
Oh dear

[08:04:02.0000] <MikeSmith>
jgraham, annevk - time for you guys to add some stuff to browser.js

[08:05:00.0000] <MikeSmith>
have it redirect all TR requests to dev.w3.org

[08:05:01.0000] <MikeSmith>
that's assuming this Opera developer is actually using Opera as her/his browser of course

[08:05:02.0000] <Dashiva^2>
Or have it redirect HTML5 to whatwg.org :)

[08:05:03.0000] <annevk>
jgraham, http://lists.w3.org/Archives/Public/public-html/2010Jun/0451.html

[08:07:00.0000] <Philip`>
jgraham: Someone from Mozilla (I forget who) was unwittingly reading the author version of the spec some time ago, and got confused about there not being any behaviour specified

[08:07:01.0000] <Dashiva^2>
Philip`: Sounds like HTML4, eh

[08:08:00.0000] <jgraham>
Philip`: That is sad

[08:08:01.0000] <jgraham>
Hmm

[08:08:02.0000] <jgraham>
Sad isn't quite what I mean

[08:08:03.0000] <jgraham>
Anyway... I at least believe that the author view has real utility

[08:09:00.0000] <jgraham>
I think the group of people that want stable but wrong snapshots is purely hypothetical

[08:09:01.0000] <annevk>
I've yet to hear something coherent in favor of that, indeed

[08:09:02.0000] <jgraham>
So... what was the last publication date of the HTML5 spec before the one that just happened?

[08:10:00.0000] <annevk>
March 4

[08:10:01.0000] <jgraham>
Thanks

[08:11:00.0000] <jgraham>
annevk: It would be nice if the html5 tracker let you specify a date in the UI :)

[08:13:00.0000] <annevk>
I should prolly sync up the latest version of that so you can patch it :)

[08:13:01.0000] <jgraham>
Is the version in svn out of date? If you put the latest version there I might as well make a patch

[08:13:02.0000] <annevk>
html5-diff has SVN revision dates in the source of the HTML5 changelogs section by the way

[08:14:00.0000] <annevk>
working on it now

[08:20:00.0000] <annevk>
argh, password troubles

[08:21:00.0000] <annevk>
ok, should be good now

[08:21:01.0000] <annevk>
r175

[08:33:00.0000] <MikeSmith>
kennyluck: you read http://www.w3.org/DesignIssues/RDF-Future.html already?

[08:33:01.0000] <kennyluck>
Yeah. This shows that timbl is a logician. :)

[08:34:00.0000] <Philip`>
"RDF/MXL" - that sounds intriguing

[08:34:01.0000] <kennyluck>
hahhah

[08:42:00.0000] <MikeSmith>
Philip`: that's one of the many little-known features of RDF

[08:43:00.0000] <MikeSmith>
which is, there's a rule that when the string "RDF" precedes another three-letter string, the first two letters of the second string must then be transposed

[08:43:01.0000] <MikeSmith>
a lot of people think RDF is complicated, but I think that's a good example of a case where the rule is pretty simple

[08:44:00.0000] <MikeSmith>
I'm pretty kennyluck agrees with me about that

[08:44:01.0000] <annevk>
sounds like you're in a bar :p

[08:44:02.0000] <MikeSmith>
RDF TFW!

[08:47:00.0000] <MikeSmith>
kennyluck: anyway, about what we were discussing earlier, I guess N3 still remains relevant

[08:47:01.0000] <MikeSmith>
judging from that at least

[08:49:00.0000] <kennyluck>
I would encourage you guys to write your personal profile in N3, namely FOAF.

[08:50:00.0000] <kennyluck>
RDF/XML is a very complicated serialization of RDF that semantic web folks complain about for years.

[08:51:00.0000] <kennyluck>
I know RDF/XML might is the origin of XML namespace. I guess it's very unfortune. 

[08:51:01.0000] <Philip`>
Hopefully they've now learned the lesson that using XML for human-readable file formats is often not a good idea

[08:52:00.0000] <MikeSmith>
kennyluck: I guess I would make my personal profile more usable, if I wanted more people to contact me

[08:52:01.0000] <MikeSmith>
but I have too many people contacting me already

[08:52:02.0000] <MikeSmith>
so for me, maybe RDF/XML is a good choice

[08:53:00.0000] <kennyluck>
Why not RDF/N3? I have actually only written RDF in RDF/XML for only once or twice.

[08:53:01.0000] <kennyluck>
Almost all RDF library parses RDF/N3 or a subset of it, called Turtle.

[08:55:00.0000] <kennyluck>
I totally agree with you Philip`.

[08:56:00.0000] <MikeSmith>
I think the future of RDF should maybe involve renaming it to something else

[08:56:01.0000] <kennyluck>
One of the topic of the RDF Next step workshop is to think about a way to standardize RDF/JSON, IIRC.

[08:56:02.0000] <MikeSmith>
a funner name would help

[08:56:03.0000] <MikeSmith>
like ARF

[08:57:00.0000] <MikeSmith>
or like the subtitle format called "ASS"

[08:57:01.0000] <Philip`>
kennyluck: I don't think a format without e.g. comments would be a good basis for human-readable file formats either

[08:57:02.0000] <Philip`>
(where by "human-readable" I mean more "human-editable")

[08:58:00.0000] <kennyluck>
JSON does not allow /* this is a comment */ ?

[08:58:01.0000] <Philip`>
People shouldn't be scared of custom formats

[08:58:02.0000] <Philip`>
kennyluck: No

[08:59:00.0000] <kennyluck>
Hmm... sorry, this is new to me.

[08:59:01.0000] <Philip`>
The JSON grammar is just what http://json.org/ shows

[09:14:00.0000] <MikeSmith>
btw, that MXL typo corrected itself

[09:14:01.0000] <MikeSmith>
that's another little-known feature of RDF: self-correcting typos

[09:55:00.0000] <boblet>
/me is trying not to hyperventilate from laughing at a 27bslash6 link via MikeSmith 

[10:15:00.0000] <MikeSmith>
boblet: blame John Allsopp for that one

[10:16:00.0000] <boblet>
yeah I saw

[10:16:01.0000] <boblet>
classic

[10:55:00.0000] <MikeSmith>
http://blogs.gnome.org/rbultje/2010/06/27/googles-vp8-video-codec/

[10:56:00.0000] <MikeSmith>
"we wrote a video decoder that heavily reuses existing components in FFmpeg, leading to a vp8.c file that is a mere 1400 lines of code (including whitespace, comments and headers) and another 450 for the DSP functions"

[10:57:00.0000] <MikeSmith>
"Google’s VP8 specs are not always equally useful. They only describe the baseline profile (0). Other profiles (including those part of the vector testsuite, i.e. 1-3) use features not described in the specifications"

[11:01:00.0000] <Philip`>
"in the end, it’s much quicker to just read libvpx source code rather than depend on the spec" - that's not really the "independent implementations" that help make a good standard

[11:07:00.0000] <MikeSmith>
Philip`: silly you

[11:08:00.0000] <MikeSmith>
in this case, by "independent implementations", they mean "independent of any specification"

[11:08:01.0000] <MikeSmith>
and dependent on one another

[11:08:02.0000] <MikeSmith>
so is more of an "independent community of implementations"

[11:08:03.0000] <MikeSmith>
kind of like a hippy commune

[11:09:00.0000] <MikeSmith>
with free love

[11:09:01.0000] <MikeSmith>
and LSD

[11:10:00.0000] <gsnedders>
MikeSmith: What have you been doing _now_!?

[11:10:01.0000] <TabAtkins>
gsnedders: Free love and LSD.

[11:10:02.0000] <annevk>
MikeSmith, and a test suite, apparently

[11:11:00.0000] <MikeSmith>
morninglory seeds

[11:11:01.0000] <gsnedders>
Free love, LSD, and a test suite? Okay…

[11:11:02.0000] <annevk>
you're not seeing the pattern?

[11:12:00.0000] <gsnedders>
Not quite.

[11:12:01.0000] <krijnh>
-_-

[11:38:00.0000] <variable>
does anyone know of any bugs or offical responses from Hixie about the "advert" tag /

[11:38:01.0000] <variable>
*about an advert tag

[11:38:02.0000] <TabAtkins>
That we'd never do it, because it would immediately defeat itself?

[11:39:00.0000] <variable>
TabAtkins, agreed - I'm looking for an offical response from hixie

[11:40:00.0000] <variable>
or some W3C bug or something like that says exactly what you are saying

[11:40:01.0000] <TabAtkins>
Hmm, I'll check.

[11:40:02.0000] <Dashiva^2>
Was there ever a formal suggestion to add it?

[11:40:03.0000] <variable>
Dashiva^2, I don't know. 

[11:42:00.0000] <variable>
TabAtkins, failing that a mailing list post from some influential WHATWG contributer

[11:42:01.0000] <variable>
I'm searching my archives but I can't find one

[11:44:00.0000] <TabAtkins>
variable: http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2008-February/013939.html

[11:47:00.0000] <TabAtkins>
variable: And, two messages down that reply chain, Hixie points back to jgraham's response.

[14:35:00.0000] <annevk>
gsnedders, postMessage is also on window so no

[14:36:00.0000] <annevk>
gsnedders, hack: (self != window)

[14:37:00.0000] <annevk>
gsnedders, shared workers do not have postMessage on the global object by the way, it seems

[14:46:00.0000] <jgraham>
annevk: You mean this !== window I guess

[14:47:00.0000] <annevk>
jgraham, self exists

[14:48:00.0000] <jgraham>
Ah

[14:48:01.0000] <jgraham>
Indeed it does

[14:48:02.0000] <jgraham>
That's confusing

[14:49:00.0000] <jgraham>
It should have been named workerGlobal or something

[14:49:01.0000] <jgraham>
Anyway to late to paint that bikeshed

[14:49:02.0000] <jgraham>
*too

[14:49:03.0000] <annevk>
self is much nicer

[14:49:04.0000] <annevk>
there

[14:49:05.0000] <jgraham>
Not if you have a python background

[14:50:00.0000] <annevk>
kids of the future will have an ECMAScript background

[14:50:01.0000] <jgraham>
boy will they be screwed up

[14:50:02.0000] <annevk>
ECMAScript is prolly one of the lesser problems of the future :)

[14:51:00.0000] <jgraham>
Well I meant DOM really

[14:51:01.0000] <jgraham>
Or "the union of all web apis"

[14:52:00.0000] <jgraham>
Since I guess there will not be that many people with a pure es background

[14:52:01.0000] <jgraham>
unless it breaks out of the browser in a really big way

[14:52:02.0000] <jgraham>
much bigger than node.js


2010-06-29
[18:31:00.0000] <j0hng4lt>
I'm wondering what will happen to XHTML5 in strict mode. My previous attempts to use XHTML1 resulted in two problems.

[18:32:00.0000] <j0hng4lt>
Problem #1: Due to a rather extreme interpretation of Postel's law in early browser designs, most developers don't even know what constitutes valid markup and how to escape content. Juniors on the team were clued out when it came to ampersands

[18:33:00.0000] <j0hng4lt>
Problem #2: External libraries and other "foreign" code (e.g. https seals, ads, etc) always assume you're in text/html and use document.write()

[18:33:01.0000] <j0hng4lt>
this blows up in Firefox

[18:33:02.0000] <TabAtkins>
Yes?

[18:33:03.0000] <j0hng4lt>
(when served as application/xhtml+xml)

[18:34:00.0000] <webben>
j0hng4lt: I don't see how the new spec can make a dent in either problem.

[18:34:01.0000] <j0hng4lt>
me neither...

[18:34:02.0000] <TabAtkins>
XML is always going to be strict, and suffer from the problems you've mentioned.  That's why XML isn't very useful for the web.

[18:34:03.0000] <j0hng4lt>
What I'm afraid of is that everyone will start "switching back to tag soup".

[18:34:04.0000] <webben>
j0hng4lt: But does it matter? Do you need XML for anything?

[18:35:00.0000] <j0hng4lt>
The problem is not with XML... No one suffers from, say, a C# compiler not accepting invalid C#. Even VB programmers know how to close their while loops

[18:35:01.0000] <j0hng4lt>
It's purely historical

[18:36:00.0000] <webben>
j0hng4lt: Ah but there's a big difference between closely controlled syntaxes (e.g. python) and loose syntaxes (e.g. perl, php)

[18:36:01.0000] <webben>
j0hng4lt: defined tag soup parsing is pretty much a loose syntax for the web.

[18:36:02.0000] <j0hng4lt>
I'm just afraid that everyone will start switching back to tag soup and start generating even more libraries that rely on Postel's Law Version 2 (be liberal in what you produce and in what you accept)

[18:37:00.0000] <webben>
j0hng4lt: And the "solution" for development teams seeking to regulate that syntax is the same as for such languages - use a linter.

[18:37:01.0000] <webben>
j0hng4lt: So long as the output is defined - does it matter? (Anymore than the looseness of Perl and PHP matter?)

[18:38:00.0000] <TabAtkins>
j0hng4lt: C# compilers can afford to reject invalid C#, because the programmer is guaranteed to see the failure before the user does.  On the web, the opposite is usually the case - users usually suffer from an XML failure before an author sees it.

[18:38:01.0000] <TabAtkins>
In simple terms, the user is the one holding the compiler on the web, and the author just ships source code.  This is obviously completely different from regular applications.

[18:39:00.0000] <j0hng4lt>
Well... it's a bit of a "chicken-or-the-egg" question... Had the syntax been strict from the start, the authors would see the errors before the users did. The tools would have been different.

[18:39:01.0000] <j0hng4lt>
But we are where we are...

[18:40:00.0000] <TabAtkins>
No, that doesn't fix anything at all.  People who know what they're doing and are *trying* to produce valid XML fail *all the time*.

[18:40:01.0000] <j0hng4lt>
and it is impractical / infeasible to convert everything to strict syntax

[18:40:02.0000] <webben>
j0hng4lt: "the authors would see the errors before the users did". Not necessarily. HTML isn't a programming language - it's often (typically) /generated/ by code.

[18:40:03.0000] <j0hng4lt>
it's not that hard to produce valid xml if it's done by means other than string concatenation

[18:40:04.0000] <webben>
So there's no error for the programmer to see.

[18:41:00.0000] <j0hng4lt>
(templates, linq-to-xml, XmlWriter, etc.)

[18:41:01.0000] <TabAtkins>
j0hng4lt: Yes, it is.  And still, nearly all programmers use string concatenation.

[18:41:02.0000] <j0hng4lt>
and in environments that escape by default

[18:42:00.0000] <TabAtkins>
Virtually every framework ever invented uses string concat, and those that do try to do some escaping of the data often still miss things (like forgetting to remove U+0000 from input before shoving it into an element, which'll crash strict XML).

[18:43:00.0000] <webben>
e.g. the php dom implementation happily outputs characters that can't be legally placed in xml documents.

[18:45:00.0000] <j0hng4lt>
again, this is a chicken-or-the-egg question... Most people who want to output valid XHTML fail - either because they can easily get away with it or because the tools mess it up for them...

[18:45:01.0000] <j0hng4lt>
They don't normally fail to output valid XML

[18:45:02.0000] <j0hng4lt>
(e.g. valid RSS, valid sitemaps, etc.)

[18:45:03.0000] <j0hng4lt>
non-HTML applications of xml

[18:45:04.0000] <TabAtkins>
By "most" you mean "almost every single person on the web, given sufficient time or attackers".

[18:45:05.0000] <TabAtkins>
RSS is quite often invalid.

[18:47:00.0000] <webben>
locus classicus: http://www.xml.com/lpt/a/1450

[18:50:00.0000] <TabAtkins>
Note as well:  "tag soup" is bad because everyone interpreted it differently, so you could never be sure what pops out.  HTML5, on the other hand, defines with absolute precision how to produce an HTML document from any random bag of bytes you throw at it.  As long as everyone follows the HTML5 parsing algorithm, you can know with absolute precision how your document will be interpreted.

[18:50:01.0000] <j0hng4lt>
In my opinion, loose handling of semantics (e.g. a new element) is fine and dandy... What tends to get frustrating is loose handling of basic syntax... If something does break because of bad nesting (e.g. <form><table><tr><td>...</tr></td></form></table> - and most of the page looks like this) - you can't easily find the error because the validator will show an error on every line...

[18:51:00.0000] <j0hng4lt>
I don't object to tag soup being handled given the situation

[18:51:01.0000] <j0hng4lt>
I would have objected to it in the early 90s

[18:51:02.0000] <j0hng4lt>
but now you have to handle it whether you like it or not

[18:51:03.0000] <TabAtkins>
It would have failed back then as well, for the same reasons it fails today for people who are actually trying to produce valid XML.

[18:52:00.0000] <j0hng4lt>
In environments that actually use xml parsers, failing to produce valid xml isn't that common

[18:53:00.0000] <TabAtkins>
My anecdotes contradict your assertion.

[18:53:01.0000] <webben>
in fairness, it would be easy to make a "strict" markup syntax that's simpler to produce than xml

[18:53:02.0000] <j0hng4lt>
And Postel's law was meant for forward and backward compatibility - not for the handling of badly written code

[18:53:03.0000] <TabAtkins>
In carefully controlled environments that aren't doing anything complicated, sure, valid XML might get produced by accident.

[18:56:00.0000] <j0hng4lt>
Most applications of XML don't use string concat... They might generate semantic errors (e.g. throwing in an element that's not in the schema), sometimes by design, but you won't see nesting errors, naked ampersands, etc (i.e. well-formedness errors)

[18:57:00.0000] <TabAtkins>
Actually, a significant number of XML-producing applications do produce wellformedness errors.  Usually they get nesting right, but they'll often happily output characters that aren't allowed unescaped in XML.

[18:57:01.0000] <j0hng4lt>
Standard xml processing libraries in .net, java, and even in php are pretty good at well-formedness

[18:57:02.0000] <j0hng4lt>
the same goes for other strict formats

[18:57:03.0000] <j0hng4lt>
(json, .ini files, Apache config files)

[18:58:00.0000] <TabAtkins>
JSON is strict but also *very simple*, and has a single valid encoding.  It's trivial to produce valid JSON, unlike XML.

[18:59:00.0000] <Hixie_>
i dunno about _that_

[18:59:01.0000] <Hixie_>
it's simpler, sure, but not trivial

[18:59:02.0000] <othermaciej>
JSON has some quirky details

[18:59:03.0000] <othermaciej>
but indeed, it is much simpler than XML

[19:00:00.0000] <Hixie_>
JSON also still lacks error handling rules

[19:00:01.0000] <j0hng4lt>
oh well... I've been seeing a lot of flame wars on the Interwebs along the lines of "xml sucks because it's complicated" or "xml sucks because it's sucks because it's xml"... In my opinion, XML is good for some tasks, bad for others. And generating any language using string concat is error-prone without a liberal parser

[19:00:02.0000] <TabAtkins>
Hixie_: Right, JSON doesn't define what invalid JSON means.  It's "strict", in that invalid stuff just isn't JSON at all.

[19:00:03.0000] <j0hng4lt>
The main reason for the backlash against XML is its marketing... It was marketed as a panacea for everything

[19:01:00.0000] <TabAtkins>
But how can you get any simpler than the diagram at JSON.org?

[19:01:01.0000] <j0hng4lt>
.ini files are simpler than json - but also more limited

[19:01:02.0000] <TabAtkins>
True.  But JSON isn't more limited than XML, so that comparison doesn't seem relevant.

[19:02:00.0000] <j0hng4lt>
fixed-width records are still simpler than ini files

[19:03:00.0000] <j0hng4lt>
The popularity of JSON is explained by the fact that it's easy to use in JavaScript.

[19:03:01.0000] <j0hng4lt>
and very few people construct it manually

[19:04:00.0000] <j0hng4lt>
most of the time, it is produced by a tool that cannot possibly generate invalid JSON

[19:04:01.0000] <j0hng4lt>
e.g.

[19:04:02.0000] <j0hng4lt>
asp.net mvc: return Json(myObject);

[19:04:03.0000] <j0hng4lt>
php: echo json_encode($array);

[19:05:00.0000] <j0hng4lt>
Constructing JSON with string concatenation would result in the same road blocks

[19:05:01.0000] <j0hng4lt>
(e.g. having to escape quotes within strings)

[19:06:00.0000] <j0hng4lt>
Also, JSON is more limited than xml. It lacks namespaces.

[19:07:00.0000] <j0hng4lt>
But in the environment where it's normally used (AJAX calls), this limitation is irrelevant

[19:07:01.0000] <TabAtkins>
No it doesn't.  It lacks XML Namespaces.  That's not a weakness either.

[19:10:00.0000] <j0hng4lt>
Without the tooling (XPath, XSLT, namespaces, schemas, etc.), it's a bare-bones format that does the same thing as XML (i.e. it's a text representation of a tree structure). If you reconstruct the tooling, it will have the same strengths and weaknesses as XML except for this:

[19:10:01.0000] <j0hng4lt>
It will somewhat more succint (less space/bandwidth per payload if it is not compressed)

[19:11:00.0000] <j0hng4lt>
And it will be slightly less readable / editable

[19:11:01.0000] <TabAtkins>
I think most people find JSON much easier to read and edit than XML (me included).

[19:12:00.0000] <j0hng4lt>
in notepad, yes... In an IDE with syntax completion... not really...

[19:12:01.0000] <j0hng4lt>
But other than that, it reminds me of some of those good old flame wars

[19:12:02.0000] <j0hng4lt>
vi versus emacs

[19:12:03.0000] <j0hng4lt>
bash verus csh

[19:12:04.0000] <j0hng4lt>
*versus

[19:12:05.0000] <j0hng4lt>
tabs versus spaces

[19:13:00.0000] <TabAtkins>
The difference being that the web has shown pretty conclusively that XML is not the right solution for webpages, unlike your choice of editor, shell, or indentation character.

[19:15:00.0000] <j0hng4lt>
I think the reasons for that are purely historical...

[19:16:00.0000] <TabAtkins>
Most people appear to disagree, and have expressed their reasoning in multiple blog posts and articles across the web.

[19:16:01.0000] <j0hng4lt>
it tends to go back and forth...

[19:17:00.0000] <j0hng4lt>
I think the biggest problem with standards in general

[19:17:01.0000] <j0hng4lt>
(not just strict versus loose syntax)

[19:17:02.0000] <j0hng4lt>
is that on the web, standardization was a bit of an afterthought

[19:17:03.0000] <j0hng4lt>
(like with sql)

[19:18:00.0000] <j0hng4lt>
most of the folks who actually develop web pages know anything beyond what gets displayed in the browser and what doesn't

[19:19:00.0000] <j0hng4lt>
But I don't see why incorrect syntax couldn't be handled further upstream

[19:19:01.0000] <j0hng4lt>
(e.g. by development tools)

[19:19:02.0000] <j0hng4lt>
but theory has to deal with reality...

[19:20:00.0000] <j0hng4lt>
we have to deal with tag-soup parsing because we have a ton of tag soup... and wishing it disappeared wouldn't make it go away.

[19:20:01.0000] <j0hng4lt>
What could've handled it back in the day is a strict-mode switch that gets ignored by old browsers and is honored by newer ones

[19:21:00.0000] <j0hng4lt>
e.g. X-Parse-Mode=xml

[19:22:00.0000] <j0hng4lt>
the switch that was chosen would break everything in IE

[19:22:01.0000] <j0hng4lt>
so everyone would serve it as text/html

[19:23:00.0000] <j0hng4lt>
since most people used ie anyway

[19:24:00.0000] <j0hng4lt>
many third-party tools would continue with document.write(), direct output of html to stdout (<a href="<?=$foo->getLink() ?>">...</a> resulting in <a href="<b>Error:</b> File not found>...</a>)

[19:24:01.0000] <j0hng4lt>
this made strict mode next to impossible

[19:26:00.0000] <roc>
awww I was just about to comment

[19:27:00.0000] <roc>
I will anyway

[19:27:01.0000] <roc>
Postel's world was a world where most people producing wire data knew what the standard was and were trying to hit it

[19:28:00.0000] <roc>
the world of Web standards is one where most producers of wire data have know little about the standards, don't care, and largely wouldn't be able to get it right if they did

[19:29:00.0000] <roc>
optimal behaviour in these two worlds is completely different

[19:29:01.0000] <roc>
</end>

[19:32:00.0000] <MikeSmith>
roc: don't try to cloud the discussion with irrelevant details about the realities of the environment

[19:34:00.0000] <j0hng4lt>
In my opinion, the phenomenon that is the World Wide Web defies the logic of development / software engineering... It's a huge mess, and I find it rather surprising that it works at all...

[19:34:01.0000] <j0hng4lt>
But the job of the Web developer has consisted of pushing square pegs through round holes from day 1

[19:35:00.0000] <roc>
ha

[19:35:01.0000] <j0hng4lt>
(remember spacer gifs, two-layered tables to create a one-pixel border, javascript branching, etc...

[19:35:02.0000] <j0hng4lt>
)

[19:35:03.0000] <j0hng4lt>
And it's not anyone's fault... 

[19:35:04.0000] <fecalpatties>
shit, i remember flaming horizontal rules and animated email box icons

[19:36:00.0000] <j0hng4lt>
The installation base and the heterogeniety of the systems involved is unprecedented.

[19:37:00.0000] <j0hng4lt>
The most common anti-pattern found on the Web is Abstraction Inversion

[19:37:01.0000] <j0hng4lt>
http://en.wikipedia.org/wiki/Abstraction_inversion

[19:37:02.0000] <j0hng4lt>
Here are a few examples:

[19:37:03.0000] <MikeSmith>
j0hng4lt: I think you mean Inversion Abstraction

[19:37:04.0000] <MikeSmith>
(common mistake)

[19:38:00.0000] <j0hng4lt>
Well... I learned the mistake from Wikipedia ;)

[19:38:01.0000] <j0hng4lt>
here are a few examples:

[19:38:02.0000] <j0hng4lt>
- Using DIVs to render pixels in plotting libraries

[19:39:00.0000] <j0hng4lt>
- Using a hidden iframe to communicate with the server from a client-side script

[19:39:01.0000] <MikeSmith>
j0hng4lt: Wikipedia is yet another example of Inversion Abstraction

[19:39:02.0000] <j0hng4lt>
(before the invention of ajax)

[19:40:00.0000] <j0hng4lt>
- using background positioning to load multiple images at once

[19:41:00.0000] <j0hng4lt>
Another big challenge on the Web is the fuzzy border between users and programmers

[19:44:00.0000] <j0hng4lt>
(hence the reverse engineering of malformed code)

[19:47:00.0000] <MikeSmith>
indeed, hence

[19:48:00.0000] <MikeSmith>
j0hng4lt: you should consider writing about some of these theories

[19:50:00.0000] <MikeSmith>
we sometimes tend to spend too much time focusing on trying to find pragmatic ways to solve existing problems

[19:52:00.0000] <j0hng4lt>
I probably will... I'm thinking of starting a small tech blog...

[19:53:00.0000] <j0hng4lt>
Some people have pointed out that many programmers have failed to produce valid (XHTML|XML|HTML5|SGML|insert your language here).

[19:53:01.0000] <j0hng4lt>
In my opinion, most people, including myself, will fail to consistently produce anything valid by hand

[19:53:02.0000] <j0hng4lt>
unless it's a simple example you can literally "run in your head"

[19:54:00.0000] <MikeSmith>
yes, fail

[19:54:01.0000] <MikeSmith>
we are all failures and admitting that is Step 1 in getting better

[19:54:02.0000] <MikeSmith>
j0hng4lt: I suggest titling that blog either "Hence" or "Inversion Abstraction"

[19:54:03.0000] <j0hng4lt>
lol

[19:57:00.0000] <j0hng4lt>
I personally believe no matter what syntax you choose, human nature being what it is, there needs to be a tool somewhere between the keyboard and the user's screen that catches the most obvious errors.

[19:57:01.0000] <MikeSmith>
j0hng4lt: that solution sounds to practical

[19:57:02.0000] <MikeSmith>
it needs... more theory, more Wikipedia citations, or something

[19:58:00.0000] <MikeSmith>
1. We admitted we were powerless over the Web — that the Web has become unmanageable.

[19:58:01.0000] <MikeSmith>
2. Came to believe that a Power greater than ourselves could restore sanity to the Web.

[19:58:02.0000] <j0hng4lt>
Why can C# programmers produce valid C#? Actually, they can't. The compiler cactches the errors for them... And Visual Studio catches most errors before they even hit Compile.

[19:59:00.0000] <MikeSmith>
that's an interesting theory too

[20:01:00.0000] <j0hng4lt>
One of the reasons why most of the mark-up on the Web is syntactically invalid, whereas most C# code is valid (despite HTML and XHTML constructs being far simpler than C#) is that the mark-up goes to the browser before anyone or anything sees it, and nothing sees it until it's FUBAR (i.e. until it's so messed up that the HTML parser would need an Artificial Neural Network to figure it out).

[20:02:00.0000] <j0hng4lt>
"Postel-enabled" browsers aren't going away anytime soon, love them or hate them...

[20:02:01.0000] <MikeSmith>
yeah, it's clearly something in need of government regulation

[20:03:00.0000] <j0hng4lt>
Hell no...

[20:03:01.0000] <MikeSmith>
http://blogs.msdn.com/b/ie/archive/2010/06/28/measuring-web-page-performance.aspx

[20:03:02.0000] <j0hng4lt>
Would you give the Internet to the people who are running the post office?

[20:04:00.0000] <j0hng4lt>
I'm thinking of a validator that gets attached to all commonly used development environments (ASP.Net with and without MVC, PHP, Python, Ruby, etc)

[20:04:01.0000] <j0hng4lt>
as a post-processor

[20:04:02.0000] <j0hng4lt>
which can be disabled in production mode

[20:04:03.0000] <j0hng4lt>
(in case you're using third-party tools that could mess up your mark-up)

[20:04:04.0000] <j0hng4lt>
but catches everything during development

[20:05:00.0000] <j0hng4lt>
i.e. before it goes to the browser

[20:06:00.0000] <othermaciej>
Lachy: sadly, I don't think the actual terms of the Bilski decision are good news

[20:07:00.0000] <MikeSmith>
j0hng4lt: http://about.validator.nu/#src

[20:08:00.0000] <MikeSmith>
othermaciej: aza raskin had a good tweet along those lines

[20:08:01.0000] <MikeSmith>
http://twitter.com/azaaza/status/17270383318

[20:08:02.0000] <MikeSmith>
"Historic day in patent law. Supreme court says business as usual for business and software patents."

[20:09:00.0000] <MikeSmith>
j0hng4lt: do you do any Java programming?

[20:09:01.0000] <j0hng4lt>
Not at the moment, but I've done some

[20:10:00.0000] <MikeSmith>
othermaciej: would be good if you or others on your team might could take a look at the Web Timings draft

[20:10:01.0000] <MikeSmith>
http://dev.w3.org/2006/webapi/WebTiming/

[20:10:02.0000] <MikeSmith>
if you've not already

[20:10:03.0000] <j0hng4lt>
I'm familiar with the Java APIs for XML

[20:11:00.0000] <j0hng4lt>
as well as JSF, Struts (1.x only) and "classic" JSP

[20:12:00.0000] <j0hng4lt>
Right now, I code in php (my day job) and C#/ASP.net with jQuery + jQuery-UI (on my own time)

[20:16:00.0000] <j0hng4lt>
MikeSmith: Why were you asking me about Java? Was it in relation to the validator?

[20:17:00.0000] <MikeSmith>
j0hng4lt: validator.nu is a Java app that runs as a RESTful Web service and has an API that allows you to hook it into any dev environment you want

[20:18:00.0000] <MikeSmith>
you can check out the source and run it locally

[20:18:01.0000] <MikeSmith>
or you can set up a instance of it for shared use over a LAN

[20:18:02.0000] <j0hng4lt>
thanks for the link

[20:18:03.0000] <MikeSmith>
or whatever

[20:19:00.0000] <MikeSmith>
http://wiki.whatwg.org/wiki/Validator.nu_Web_Service_Interface

[20:20:00.0000] <MikeSmith>
if you try it and find that it's missing something you'd like for it to have, the backend Java is well-designed code that is pretty easy to hack on 

[20:20:01.0000] <MikeSmith>
using SAX to hook various parts together

[20:21:00.0000] <j0hng4lt>
REST is another interesting phenomenon... I do understand the initial arguments (i.e. using a GET request when you're not modifying stuff), but in a way, it's another case of going "back to the 90s"

[20:22:00.0000] <othermaciej>
MikeSmith: we plan to look at it, and also the associated charter for the potential WG around this

[20:24:00.0000] <j0hng4lt>
i.e. In practice, you end up having to simplify / flatten the data structures you send to the service (i.e. you go from XML-serialized objects to key/value pairs), and the response could be just about anything as well - xml, json or even key-value pairs...

[20:24:01.0000] <j0hng4lt>
I remember coding Web services before they were called Web services

[20:25:00.0000] <j0hng4lt>
That was exactly what you would end up doing... You would send a GET or a POSt with URL-encoded key/value pairs and you would get the same thing back

[20:25:01.0000] <MikeSmith>
nothing wrong with the 90s

[20:25:02.0000] <MikeSmith>
we could use some more of the 90s around here

[20:25:03.0000] <MikeSmith>
the Web has too much of the 00s

[20:26:00.0000] <MikeSmith>
"The World Wide Web: The 00 Years"

[20:26:01.0000] <MikeSmith>
in terms of standards at least, not the best years

[20:27:00.0000] <j0hng4lt>
My biggest frustration with the Web has been that, unlike with desktop development, you never get adequate tooling... The moment tools get built up, either they get obsolete or the Web goes up in arms with torches and pitchforks, calling for simplification

[20:27:01.0000] <j0hng4lt>
then you get a conspiracy between Mort and Einstein to drive out Elvis

[20:28:00.0000] <j0hng4lt>
So in the end, you're always coding in notepad (or do the equivalent thereof)... 

[20:29:00.0000] <j0hng4lt>
The worst thing of the 00 years had to do with standards designed by committee

[20:29:01.0000] <j0hng4lt>
that never got implemented

[20:29:02.0000] <j0hng4lt>
(e.g. XHTML2, CSS2, the beginnings of CSS3)

[20:30:00.0000] <j0hng4lt>
the positive part of the OOs was the drive for structure (stricter syntax, schemas, etc.)

[20:30:01.0000] <j0hng4lt>
+ no more if (ie) { doThis() } else if (netscape) { doThat() } else { goToHell() }

[20:35:00.0000] <j0hng4lt>
But in a way, we're seeing the Revenge of Mort in full swing... Loose syntax, no data structures, more copy-and-paste development, less widget-building

[20:36:00.0000] <j0hng4lt>
Ideally, I would like to be in a situation where Mort and Elvis can use the same Web without competing for space...

[21:13:00.0000] <karlcow>
/me noticed that less and less people do ascii art in their explanation of features (found cute ascii art for FIG element in 1995)

[21:13:01.0000] <karlcow>
example: http://lists.w3.org/Archives/Public/www-html/1995Jul/0105.html

[21:13:02.0000] <MikeSmith>
well, that's a sad change

[21:13:03.0000] <MikeSmith>
charter for proposed Web Performance WG has been updated -

[21:13:04.0000] <MikeSmith>
http://www.w3.org/2010/06/webperf.html

[21:14:00.0000] <MikeSmith>
proposed chair is Arvind Jain, Google

[21:42:00.0000] <MikeSmith>
my strategy of trying to generate as much animi

[21:42:01.0000] <MikeSmith>
oops

[21:43:00.0000] <MikeSmith>
wrong channel

[21:43:01.0000] <MikeSmith>
or maybe not

[21:44:00.0000] <MikeSmith>
fwiw, I was saying that my strategy of trying to generate as much animosity toward the W3C as possible seems to be failing

[21:47:00.0000] <MikeSmith>
I had been led to believe that the recent Stephen Shankland CNET article would engender quite a bit of righteous indignation and some robust W3C-bashing among the widest number of communities possible

[21:47:01.0000] <MikeSmith>
but I've been disappointed

[21:47:02.0000] <karlcow>
:)

[21:48:00.0000] <MikeSmith>
maybe if that article could instead have been written by someone less insightful, less prone to actually considering things carefully

[21:48:01.0000] <karlcow>
I had not even noticed this article and person before you mentioning it :)

[21:50:00.0000] <MikeSmith>
karlcow: well, you're clearly not doing enough to try to help whip things into a proper frenzy

[21:50:01.0000] <karlcow>
hmm whips…

[21:52:00.0000] <MikeSmith>
anyway, lacking the sufficient quantity of W3C-bashing that I had been expecting to get from that this week, I guess I have to fall back to again mostly relying on some folks here to regularly supply some knee-jerk W3C bashing now and then

[21:53:00.0000] <MikeSmith>
so I hope people won't disappoint me here like others elsewhere have

[21:54:00.0000] <MikeSmith>
I'm depending on it

[21:55:00.0000] <karlcow>
it is a looooooooooong article

[21:56:00.0000] <MikeSmith>
karlcow: it's designed to be savored

[21:56:01.0000] <MikeSmith>
enjoy every moment

[21:56:02.0000] <MikeSmith>
there's only one "first time" for everything

[22:08:00.0000] <karlcow>
http://infomesh.net/html/history/early/

[22:08:01.0000] <karlcow>
The Early History of HTML

[22:09:00.0000] <karlcow>
Full HTML Test Suite - http://www.w3.org/History/1991-WWW-NeXT/Implementation/Test/backup_of_test.html

[22:09:01.0000] <karlcow>
:)

[22:38:00.0000] <hsivonen>
wow. http://www.w3.org/DesignIssues/RDF-Future.html doesn't have a year in the URI

[22:41:00.0000] <karlcow>
hsivonen: none of DesignIssues have year in URIs

[22:58:00.0000] <hsivonen>
karlcow: ah. so the directory is so *old* it doesn't have a year. I was hoping *new* articles no longer have a year.

[23:05:00.0000] <karlcow>
hsivonen: this directory is just the directory of the personal notebook of the guy who created this web site you pointed to ;)

[23:05:01.0000] <MikeSmith>
" I would in fact prefer, instead of <H1>, <H2> etc for headings [those come from the AAP DTD] to have a nestable <SECTION>..</SECTION> element, and a generic <H>..</H> which at any level within the sections would produce the required level of heading."

[23:06:00.0000] <MikeSmith>
/me had never read that particular bit before

[23:07:00.0000] <MikeSmith>
ah, I get to the end of page and see that Sean Palmer put this together

[23:07:01.0000] <MikeSmith>
bravo Sean

[23:07:02.0000] <MikeSmith>
/me wonders what stuff Sean is working on these days

[23:08:00.0000] <MikeSmith>
http://inamidst.com/sbp/

[23:09:00.0000] <MikeSmith>
"As of April 2010, I've been working on automatic table of contents generation using jQuery,"..

[23:09:01.0000] <karlcow>
sean is one of my heroes :) 

[23:12:00.0000] <MikeSmith>
karlcow: http://inamidst.com/whits/2008/technobunkum is still relevant

[23:16:00.0000] <MikeSmith>
as are http://inamidst.com/whits/2008/ditching and http://inamidst.com/whits/2008/ambient

[23:20:00.0000] <karlcow>
hehe but but but MikeSmith if there is no more technobunkum, there is no more Gargantua, Olympus, Valhalla, …

[23:20:01.0000] <karlcow>
suddenly we would be forced to solve real issues, I mean plain boring stuff

[23:21:00.0000] <karlcow>
no no no no… this is a nightmare ;)

[23:23:00.0000] <karlcow>
ambient information is cool. Not always easy to put in place, specifically in a non intrusive way (aka GPS = cool, Celltower loc = bad)

[23:23:01.0000] <MikeSmith>
I guess a lot of real problems are often not particularly fun puzzles

[23:23:02.0000] <karlcow>
or you really have to be a geek to really do cool stuff with ambient

[23:24:00.0000] <MikeSmith>
the problems that are more mundane but are nevertheless difficult to solve

[23:24:01.0000] <karlcow>
ooops talking about ambient… 1:30am

[23:24:02.0000] <MikeSmith>
the problems that don't lead to particularly elegant solutions, but instead to kind of messy ones

[23:24:03.0000] <karlcow>
need to do boring puzzles tomorrow morning for a steak in my plate

[23:24:04.0000] <MikeSmith>
karlcow: that's early, man

[23:24:05.0000] <MikeSmith>
hai hai

[23:25:00.0000] <karlcow>
oyasumi nasai

[23:27:00.0000] <MikeSmith>
oyasumi

[00:04:00.0000] <MikeSmith>
dbaron: I'm not sure I understand your latest public-html message correctly

[00:04:01.0000] <MikeSmith>
so let me give an example and ask if it's the same as what you have in mind

[00:05:00.0000] <MikeSmith>
/me trots out a recent hobbyhorse

[00:05:01.0000] <dbaron>
ok

[00:06:00.0000] <MikeSmith>
the example is related to parsing of MathML content

[00:07:00.0000] <MikeSmith>
if the parsing algorithm encounters a MathML <annotation-xml> element, it essentially expects the content of that to be either and <svg> element or more MathML content

[00:08:00.0000] <MikeSmith>
but if the byte stream it parsing instead has HTML content in there, it goes back and pops elements off the stack until it encounters either a <math> or <svg> element

[00:09:00.0000] <MikeSmith>
the end

[00:11:00.0000] <hsivonen>
dbaron: I believe it is implied that there must be no other data structures hiding in the DOM tree than a Level 2 DOM and the references from form controls to forms

[00:11:01.0000] <hsivonen>
dbaron: but it's possible that it's not actually stated anywhere

[00:12:00.0000] <hsivonen>
it probably should be stated to make it clear that the kind of hidden data Opera and IE have is non-conforming

[00:13:00.0000] <hsivonen>
oh, and there are also the HTMLness flag and the quirkiness enum

[00:14:00.0000] <dbaron>
"No other data structures" might be too strong; it might make more sense to limit certain classes of behavior to not depending on other data structures

[00:15:00.0000] <dbaron>
in any case, I looked for a bit, and couldn't find any such requirements

[00:15:01.0000] <dbaron>
MikeSmith, I don't see how your example is related to what I was saying, although I could be missing something

[00:22:00.0000] <MikeSmith>
dbaron: no, I don't think you're missing anything -- after re-reading I see that you're referring to something other that what I meant by that example

[00:22:01.0000] <MikeSmith>
the "differences that must not matter" part in your subject line should have clued me in

[00:30:00.0000] <othermaciej>
good evening folks

[00:31:00.0000] <othermaciej>
hsivonen, dbaron: I think it would be a valid statement of fact that no conformance requirements in the spec would depend on the input document in any way other than as it affects the DOM

[00:31:01.0000] <othermaciej>
but I do not believe there is a requirement that browsers may not invent their own such behaviors

[00:31:02.0000] <othermaciej>
(so long as those behaviors are consistent with the existing UA conformance requirements)

[00:31:03.0000] <othermaciej>
I think it would be hard to scope it appropriately

[00:32:00.0000] <othermaciej>
for example, it's clear to me that a syntax-highlighted source view should be able to depend on the exact text input

[00:32:01.0000] <othermaciej>
or XHR.responseText

[00:39:00.0000] <dbaron>
but could one say, for example, that anything that depends on the element type (other than a display of the tag name) must not depend on things other than X?

[00:51:00.0000] <hsivonen>
othermaciej: source view isn't really part of the interoperable platform

[00:53:00.0000] <othermaciej>
it's hard to construct conformance criterion that's a universal negative

[00:54:00.0000] <othermaciej>
maybe it would be preferable to just cite specific things that MUST NOT be affected by anything in the source document besides the output of the parsing algorithm

[01:08:00.0000] <MikeSmithX>
dbaron: btw, I wanted to ask if you might be planning an updates to the CSS3 syntax spec

[01:09:00.0000] <MikeSmith>
in particular, to the grammar and tokenizer parts

[01:09:01.0000] <MikeSmith>
I ask because Dan Connolly recently wrote a new CSS Validator (in Scala)

[01:10:00.0000] <MikeSmith>
with is designed to take a CSS grammar as an input

[01:10:01.0000] <MikeSmith>
http://dvcs.w3.org/hg/css-val/raw-file/f5931981d7dd/README.html

[01:11:00.0000] <MikeSmith>
" the CSS 2.1 specification includes a uniform grammar, applicable to all CSS versions. The original (2002) W3C CSS Validator predates this uniform grammar and maintenance is becoming an undue burden as a result. This is an attempt to build a validator using the uniform syntax."

[01:11:01.0000] <dbaron>
I'm not planning an update; it's currently basically unowned.

[01:12:00.0000] <dbaron>
somebody writing a validator might make a good owner, though :-)

[01:12:01.0000] <dbaron>
anyway, g'night

[01:13:00.0000] <annevk>
oh god -- even the Microsoft canvas tests assume there's no such thing as a high-resolution canvas backing bitmap

[01:14:00.0000] <annevk>
nobody is gonna get it right

[01:16:00.0000] <hsivonen>
did Apple ship high-res backing bitmap by default in Mobile Safari on iPhone 4?

[01:16:01.0000] <hsivonen>
'cause if they didn't, who is going to?

[01:17:00.0000] <annevk>
well yeah, the iPhone 4 would certainly be interesting to test on this

[01:22:00.0000] <annevk>
I don't get the Google China thing

[01:22:01.0000] <annevk>
they cannot redirect to google.com.hk but they can point to it and users can access it either way but the former is forbidden but the latter is ok?

[01:22:02.0000] <annevk>
it just doesn't make much sense

[01:26:00.0000] <MikeSmith>
karlcow: http://inamidst.com/whits/2008/technobunkum is still relevant

[01:26:01.0000] <MikeSmith>
oops

[01:26:02.0000] <MikeSmith>
kennyluck: http://inamidst.com/whits/2008/technobunkum

[01:26:03.0000] <MikeSmith>
kennyluck: http://inamidst.com/whits/2008/ditching and http://inamidst.com/whits/2008/ambient

[01:29:00.0000] <kennyluck>
/me frowned a little bit when he saw words such as "psychological" and "technobunkum".

[01:38:00.0000] <foolip>
annevk: they just redirected, at least when I left Beijing a few weeks ago

[01:39:00.0000] <foolip>
annevk: also, expecting the mainland "regulation" of the internet is... funny :)

[01:40:00.0000] <annevk>
foolip, they're switching to linking, see latest blog post

[01:47:00.0000] <foolip>
hehe, they even used an image of a search form that links to .com.hk

[01:47:01.0000] <foolip>
this ICP thing they're talking about is actually quite funny

[01:48:00.0000] <foolip>
in http://www.miibeian.gov.cn/share/liuchengtu2.jsp there's a flowchart for how to register

[01:49:00.0000] <foolip>
the first step is "Use Windows IE6 browser to log on to www.miibeian.gov.cn"

[01:49:01.0000] <foolip>
that's Chinese e-government for you :)

[02:36:00.0000] <asmodai>
Is ff 3.7 nightly what will become 4.x? And does anyone know if it can coexist with a stable ff (without screwing up any existing profiles).

[02:37:00.0000] <hsivonen>
asmodai: the version number of nightlies got changed less than 6 hours ago. 

[02:37:01.0000] <hsivonen>
asmodai: so the next round of nightlies isn't "3.7" anymore

[02:38:00.0000] <hsivonen>
asmodai: I wouldn't run the nightlies with the same profile as 3.6.x

[02:38:01.0000] <hsivonen>
asmodai: I did, and Firefox Sync and password manager got confused

[02:40:00.0000] <asmodai>
hsivonen: I will run it on a Windows box. So I should use the profile manager, create a 4.x profile, but can I automate something to have 4.x use that profile (or at least ask which one to use prior to starting)?

[02:41:00.0000] <hsivonen>
asmodai: the profile manager has a checkbox for making it ask before starting

[02:42:00.0000] <annevk>
so when is 4 shipping?

[02:47:00.0000] <asmodai>
annevk: at the end of the year I think I just read

[02:47:01.0000] <asmodai>
/me looks for link

[02:47:02.0000] <hsivonen>
annevk: https://wiki.mozilla.org/Releases has a tentative lower bound data point

[02:47:03.0000] <asmodai>
October/November

[02:47:04.0000] <asmodai>
http://geektech.geektech.netdna-cdn.com/wp-content/uploads/2010/06/Firefox4ReleaseDate.png

[02:48:00.0000] <hsivonen>
3.6.6 shipped already

[02:48:01.0000] <hsivonen>
and there was no 3.6.5

[02:49:00.0000] <asmodai>
hsivonen: I guess the 3.7 nightly will update itself to whatever the new one has as version number?

[02:49:01.0000] <hsivonen>
asmodai: it should

[02:50:00.0000] <asmodai>
/me tries ;)

[02:54:00.0000] <asmodai>
I forgot, can't run regular and new side by side

[02:54:01.0000] <hsivonen>
asmodai: it looks like the nightlies haven't cycled yet after the version number change

[02:57:00.0000] <asmodai>
hsivonen: That's ok -- got a 3.7apre6 now

[02:57:01.0000] <asmodai>
a6pre even

[03:10:00.0000] <gsnedders>
annevk: thx

[03:12:00.0000] <annevk>
gsnedders, timely

[03:12:01.0000] <gsnedders>
annevk: 13 hours? Nothing.

[03:20:00.0000] <annevk>
gsnedders, at some point you have to start living in the present :p

[03:20:01.0000] <gsnedders>
annevk: It's hard to leave the past behind

[03:23:00.0000] <jgraham>
gsnedders: There is hardly any past behind you

[03:24:00.0000] <annevk>
I was gonna say "drink more" but it would have just gone down from there and it was pretty bad already :)

[03:25:00.0000] <gsnedders>
annevk: Hey, I had a beer or two when you told me I should have a beer for you in your abscence :P

[04:28:00.0000] <asmodai>
hsivonen: Well

[04:28:01.0000] <asmodai>
hsivonen: Do people depend heavily on those litmus tests?

[04:40:00.0000] <hsivonen>
asmodai: which litmus tests?

[04:40:01.0000] <asmodai>
litmus.mozilla.org 

[04:40:02.0000] <hsivonen>
asmodai: I have no idea

[04:41:00.0000] <asmodai>
Grrrr *slaps the test runner*

[04:41:01.0000] <asmodai>
Cute, press a button to submit a single test result, instead of sends all >_<

[04:43:00.0000] <hsivonen>
hmm. bulletproofing new elements by using <div>s also pretty much defeats any convenience benefits from the new elements

[04:43:01.0000] <hsivonen>
see http://tantek.com/presentations/2010/06/html5-vtm/

[04:47:00.0000] <jgraham>
and yeah, that nested div thing seems pointless

[04:49:00.0000] <jgraham>
the microdata vs microformats opinion seems biased

[04:50:00.0000] <jgraham>
web storage already has wide support, no?

[04:50:01.0000] <jgraham>
and postMessage + friends

[05:38:00.0000] <annevk>
hmm, why do Minefield releases "installed" in a separate directory use plug-ins from the version of Firefox that ships with the distribution?

[05:38:01.0000] <annevk>
specifically, the Ubuntu theming nonsense

[05:39:00.0000] <crash\>
there's a seperate plugin dir

[05:39:01.0000] <annevk>
okay, just curious

[05:39:02.0000] <annevk>
i wonder why that confusing bookmark menu is there

[05:39:03.0000] <annevk>
has that always been there?

[05:40:00.0000] <crash\>
seems all Firefox version looking for that directory

[05:40:01.0000] <annevk>
you'd think people stopped using bookmarks the moment search engines started to work

[05:41:00.0000] <annevk>
Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:2.0b2pre) Gecko/20100629 Minefield/4.0b2pre

[05:41:01.0000] <annevk>
Opera/9.80 (X11; Linux x86_64; U; en) Presto/2.6.30 Version/10.60

[05:42:00.0000] <annevk>
and I thought ours was bloated :/

[05:43:00.0000] <hsivonen>
annevk: you can disable the Ubuntu addon via the add-on manager if you don't like it

[05:43:01.0000] <crash\>
It bloated since old script use bad detection for browsers :(

[05:43:02.0000] <annevk>
it just disabled itself automatically it seems

[05:43:03.0000] <annevk>
yeah, that's why we have that Version/10.60 crap

[05:44:00.0000] <annevk>
we shouldn't have done that I think

[05:44:01.0000] <annevk>
Presto/... is enough imo

[05:44:02.0000] <crash\>
you should see the Chromes UA string

[05:44:03.0000] <annevk>
rather not :)

[05:45:00.0000] <crash\>
Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10_6_4; en-US) AppleWebKit/534.2 (KHTML, like Gecko) Chrome/6.0.451.0 Safari/534.2

[05:45:01.0000] <annevk>
lol

[05:45:02.0000] <hsivonen>
annevk: if you put Opera there, trying to get people to sniff for Presto is about as futile (or more futile) than getting people to sniff for Gecko instead of Firefox 

[05:46:00.0000] <hsivonen>
the same goes for WebKit and Safari

[05:46:01.0000] <crash\>
UA string sniffing is bad at all

[05:46:02.0000] <crash\>
I never needed it

[05:46:03.0000] <hsivonen>
what's the deal with the underscores in the OS X version in Chrome's UA string?

[05:47:00.0000] <crash\>
donno

[05:47:01.0000] <annevk>
the idea was that you could sniff for the engine common across Opera deliveries; obviously not a very good idea

[05:47:02.0000] <hsivonen>
crash\: well, if a browser vendor doesn't want to be sniffed, the vendor should kill the User-Agent header or freeze it to a constant

[05:47:03.0000] <annevk>
though in theory it sounds nice

[05:47:04.0000] <gsnedders>
hsivonen: IIRC it's how it is represented in some system define

[05:47:05.0000] <gsnedders>
But that would be compile time

[05:47:06.0000] <gsnedders>
Hmm

[05:47:07.0000] <gsnedders>
Dunno

[05:47:08.0000] <hsivonen>
annevk: "normal people" don't think in terms of engines. they think in terms of browser names

[05:48:00.0000] <crash\>
hsivonen: they use it themself, for example at addons.mozilla.org

[05:49:00.0000] <hsivonen>
crash\: unfortunately, I'm very aware of Mozilla's own sites using the UA string

[05:49:01.0000] <crash\>
to detect if an addon is avaiable for the curent Firefox version

[05:49:02.0000] <annevk>
I guess at the time this discussion went on I should have weighed in and pointed at what a failure Firefox/Gecko is for the Mozilla community but I did not really care back then :/

[05:49:03.0000] <hsivonen>
Dear lazy IRC, how do I tell pulseaudio on Lucid to use a USB headset for Skype but built-in audio output for everything else?

[05:50:00.0000] <annevk>
Well, it's not like I really care now, but trimming HTTP weight and fingerprinting variables is something that's somewhat fun

[05:53:00.0000] <roc>
Gecko clearly isn't a complete failure, since Webkit adopted it

[05:53:01.0000] <jgraham>
roc: I plan to quote that out of context :)

[05:54:00.0000] <hsivonen>
roc: they adopted it before Firefox was as well known as it is today

[05:59:00.0000] <roc>
jgraham: yeah, I realized after I said it that that could get me in trouble :-)

[06:00:00.0000] <roc>
hsivonen: maybe so

[06:00:01.0000] <roc>
the whole engine vs browser distinction is a quandry

[06:01:00.0000] <roc>
it's good in principle but it confuses people

[06:17:00.0000] <Dashiva^2>
hsivonen: lazy IRC is too lazy to reply

[06:19:00.0000] <Dashiva^2>
Re: nested div, I thought the IE shiv fixed all the problems with new block elements?

[06:29:00.0000] <boblet>
anyone know about HTML5 support  in .net?

[06:31:00.0000] <hsivonen>
boblet: someone from the .net side of Microsoft was implementing the parsing algorithm

[06:32:00.0000] <boblet>
hsivonen: trying to answer a query about <input type="number"> getting an invalid type error, and Google/Bing can’t help. did you hear an eta?

[06:32:01.0000] <boblet>
(what idiot approved using such a generic name. jeez)

[06:33:00.0000] <hsivonen>
boblet: I don't know of an eta. 

[06:33:01.0000] <hsivonen>
volkmar: ^

[06:34:00.0000] <boblet>
hsivonen: thanks anyhow. good to be able to pass on that it’s in progress

[06:38:00.0000] <hsivonen>
boblet: I don't know of progress, either. In fact, I don't know what the issue is.

[06:38:01.0000] <boblet>
hsivonen: “Our dev team use .net 3.5. We are attempting to use HTML5 form input fields. When using <input type=”number” /> we get a “Parser Error Message: 'number' is not a valid type for an input tag".”

[06:40:00.0000] <hsivonen>
boblet: oh. that's probably something on the ASP.Net layer

[06:42:00.0000] <boblet>
hsivonen: oho! that’s improved my search results no end. thanks!

[06:47:00.0000] <boblet>
woah, forum.asp.net is pretty scary

[06:50:00.0000] <hsivonen>
http://inamidst.com/whits/2008/ditching is interesting

[06:51:00.0000] <MikeSmith>
hsivonen: I figured you had read that previously.. I thought I was the only one who hadn't

[06:51:01.0000] <MikeSmith>
kennyluck and I were talking about that earlier today

[06:52:00.0000] <MikeSmith>
boblet: learning asp.net stuff is fun.. there's a whole series of books and courses about it.. you can spend as much time as you want and there's always more to learn

[06:53:00.0000] <boblet>
MikeSmith: you had a typo there s/as much time as you want/as much money as you want/

[06:53:01.0000] <MikeSmith>
hey man

[06:53:02.0000] <MikeSmith>
don't disrespect technologies

[06:53:03.0000] <MikeSmith>
"to each his own"
 as they say

[06:54:00.0000] <boblet>
no disrespect intended

[06:54:01.0000] <hsivonen>
MikeSmith: I hadn't read it before

[06:54:02.0000] <boblet>
(at least not of that part)

[06:54:03.0000] <MikeSmith>
boblet: I remember the introductory book for asp.net is called "How to distinguish your ass from a hole in the ground"

[06:54:04.0000] <MikeSmith>
I found that book really helpful

[06:54:05.0000] <hsivonen>
why is Modernizr dual-licensed under BSD and MIT?

[06:54:06.0000] <kennyluck>
I think the document is just too old.

[06:55:00.0000] <MikeSmith>
boblet: and then of course there is a natural progression of the various asp.net up from that fundamental level. it's all very scientific almost

[06:55:01.0000] <hsivonen>
kennyluck: Sean Palmer's blog post?

[06:55:02.0000] <boblet>
hsivonen: woah this is some strong stuff

[06:55:03.0000] <MikeSmith>
paul_irish: <hsivonen> why is Modernizr dual-licensed under BSD and MIT?

[06:55:04.0000] <kennyluck>
Yes, hsivonen.

[06:56:00.0000] <hsivonen>
kennyluck: even though it's 2 years old, the content seems on-point and fresh to me

[06:57:00.0000] <annevk>
hsivonen, hmm, that was a couple of years ago

[06:57:01.0000] <annevk>
hsivonen, weird that I haven't seen it then

[06:57:02.0000] <kennyluck>
He indeed shows the point about RDF -> you can easily write a RDF client. no documentView or whatever that's needed to implement a HTML client.

[06:58:00.0000] <kennyluck>
His client failed to attract folks, and I think this might be what his frustration came from.

[06:59:00.0000] <kennyluck>
s/what/where/

[07:01:00.0000] <annevk>
heh, that's so vastly different from what I read into it

[07:01:01.0000] <boblet>
 use cases and practical examples are a big help for anyone trying to read a spec too

[07:01:02.0000] <hsivonen>
kennyluck: I didn't read it that way

[07:02:00.0000] <kennyluck>
Well, that's my personal explanation.

[07:02:01.0000] <hsivonen>
kennyluck: I think the point about use cases and solutions looking for problems makes sense regardless of the popularity of his software

[07:02:02.0000] <kennyluck>
Yes, I agree.

[07:03:00.0000] <kennyluck>
But I believe we all know about this. It shouldn't be associated with Semantic Web.

[07:04:00.0000] <annevk>
Why not? It's been repeatedly cited as a problem of RDF...

[07:05:00.0000] <kennyluck>
We can blame those professors. :) And those who mention the word "ontology" again and again.

[07:06:00.0000] <boblet>
lol: http://iscss3partofhtml5.com/

[07:07:00.0000] <boblet>
(sorry if that’s old news)

[07:07:01.0000] <hsivonen>
kennyluck: you might be in denial, since the URL in your /wii points to an N3 resource :-)

[07:07:02.0000] <zcorpan_>
boblet: next up is ishtml5partofhtml5.com

[07:08:00.0000] <annevk>
also http://isgeolocationpartofhtml5.com/

[07:08:01.0000] <boblet>
zcorpan_: stop it, you’re killing me here!

[07:08:02.0000] <boblet>
/me needs to stop laughing so he can breathe

[07:11:00.0000] <Dashiva^2>
annevk: Missing a "yet" in the domain name :P

[07:20:00.0000] <annevk>
someone should get partofhtml5.com

[07:23:00.0000] <kennyluck>
By the way, Sean Palmer post gives us a hint on a use case of Semantic Web/RDF - decentralized issue tracking systems, so that we can make sure what we are working on (an issue) has a purpose (an issue that blocks this issue).

[07:24:00.0000] <kennyluck>
The reason why we need "decentralization" is the following use case. MikeSmith might want to create an issue "(super master bug) HTML5 implematations".

[07:24:01.0000] <kennyluck>
And this is blocked by the Mozilla HTML5 implemenation master bug and Webkit HTML5 implementation master bug.

[07:26:00.0000] <kennyluck>
As long as the Mozilla/Webkit HTML5 implementaion bugs are closed. We can close the "HTML5 implemetations" bug and release the W3C CR (correct me if I am wrong)

[07:26:01.0000] <Philip`>
He can do that without any RDF or semanticity, just by adding himself to the CC list on the Mozilla and WebKit bugzillas and noticing updates

[07:28:00.0000] <kennyluck>
Well, what I suggested was the feature -> "Allow an issue in a issue tracking system to block other issue in another domain". RDF is not required, indeed.

[07:28:01.0000] <kennyluck>
Any format that allows Linking *externally* is good enough. Microdata could be a choice.

[07:30:00.0000] <Philip`>
Why is <p>This bug is blocked by <a href="http://bugzilla.example.org/12345">this other bug</a></p> insufficient?

[07:30:01.0000] <kennyluck>
It is good, but it might not be machine readable.

[07:30:02.0000] <Philip`>
Why does it need to be machine readable?

[07:31:00.0000] <kennyluck>
micrioformat, microdata, JSON, RDF, CSV, XML

[07:31:01.0000] <kennyluck>
Oh, sorry. I thought you asked about the formats.

[07:32:00.0000] <kennyluck>
Because then we can do more things with the data. Microdata is machine readable.

[07:32:01.0000] <Philip`>
What "more things"?

[07:32:02.0000] <mahound_>
well, the only advantage i see is that it can be closed automatically or something...

[07:32:03.0000] <kennyluck>
Like rendering dependency graph, etc.

[07:33:00.0000] <mahound_>
i mean... the blocked issue can crawl the other two and close itself when they're closed

[07:33:01.0000] <kennyluck>
mahound_, It's a good advantage, isn't it. Just think about why do we the the field "Blocking"

[07:33:02.0000] <Philip`>
Seems like working backwards from the solution (providing machine-readable data) to come up with hypothetical problems that it could help with

[07:34:00.0000] <kennyluck>
You don't like Microdata, Philip` ?

[07:34:01.0000] <mahound_>
Philip`, well, it's still useful :)

[07:35:00.0000] <Philip`>
In the concrete example that was given (moving the HTML5 spec to CR when implementors have closed their HTML5-implementation bugs) you wouldn't actually want to tie the closed status, because browsers say closed when fixed in source control but the spec needs to wait for them to ship

[07:35:01.0000] <annevk>
except it's not magically solved by adding some RDF

[07:36:00.0000] <mahound_>
of course it's discussible whether it would be worth implementing such a thing...

[07:36:01.0000] <Philip`>
so automatically responding to status changes won't help that use case

[07:36:02.0000] <kennyluck>
annevk, no it's not. Lots of hard work needs to be done to Bugzilla, Redmine, etc.

[07:37:00.0000] <kennyluck>
I can imagine,as long as we have such system, "Google Todos".

[07:37:01.0000] <mahound_>
i think microformats would be more than enough

[07:37:02.0000] <kennyluck>
Then you can ask "Oh God Google, could you give me something worth doing. Please no sudoko, which does not have a dependency".

[07:38:00.0000] <kennyluck>
mahound_, maybe.

[07:38:01.0000] <mahound_>
anyway, i think a simple link or CC is still the best solution ;)

[07:39:00.0000] <mahound_>
as Philip` says, the workflow is sometimes too complex

[07:39:01.0000] <mahound_>
and requires human intervention

[07:40:00.0000] <kennyluck>
Well, the goal is to make issue trackers go global.

[07:40:01.0000] <kennyluck>
Of course issue trackers don't understand human.

[07:42:00.0000] <mahound_>
speaking of microformats... when will i see native hCalendar support in major web browsers?

[07:43:00.0000] <kennyluck>
:) that I am interested in, too.

[07:43:01.0000] <mahound_>
i really don't get it, it's such an interesting, useful use case

[07:43:02.0000] <mahound_>
how many times do i have to copy-paste dates and times, and places to my google calendar...

[07:53:00.0000] <variable>
mahound_, Firefox has an addon that does that

[07:54:00.0000] <variable>
I duno about chrome 

[07:54:01.0000] <kennyluck>
Operator you mean?

[07:55:00.0000] <variable>
kennyluck, I think so

[07:57:00.0000] <variable>
kennyluck, http://microformats.org/wiki/hcalendar-implementations#Firefox_Greasemonkey_Plugins

[08:01:00.0000] <boblet>
mahound_: I think FireFox was planning to release native support a while back, but decided to pull it and work on interface more. check azzazza’s blog

[08:03:00.0000] <variable>
boblet, link?

[08:04:00.0000] <kennyluck>
Hmm... I have to thank them then. I can't live without the Firefox smart bar.

[08:04:01.0000] <zcorpan_>
<mo><xsl:text disable-output-escaping="yes"><![CDATA[&InvisibleTimes;]]></xsl:text></mo>

[08:05:00.0000] <boblet>
variable: wetware only orz

[08:05:01.0000] <zcorpan_>
http://www.google.com/codesearch/p?hl=en#FQzGEBTCSY4/soft/OMathML/releases/omathml-0.1.tar.gz%7CDuTpMw-4EDI/xsl/mathmlc2p.xsl&q=%3C(%5B%5E%5Cs/%5D+:)?math%20%3C(%5B%5E%5Cs/%5D+:)?mo%3E%5B%5E%3C%5D*%3C%5B%5E/%5D

[08:05:02.0000] <variable>
wetware ?

[08:06:00.0000] <boblet>
the “I think…” part

[08:06:01.0000] <variable>
boblet, I meant link to azzazza's blog

[08:06:02.0000] <boblet>
looking… i spelled it wrong. wrong number of zs?

[08:06:03.0000] <zcorpan_>
Hixie: i see <mo><mrow> here http://www.google.com/codesearch/p?hl=en#KGV8vs9gsGI/content/abacus/content/XNS/docs/BigDecimal.html&q=%3C(%5B%5E%5Cs/%5D%2B:)%3Fmath%20%3C(%5B%5E%5Cs/%5D%2B:)%3Fmo%3E%5B%5E%3C%5D*%3C%5B%5E/%5D&sa=N&cd=25&ct=rc

[08:07:00.0000] <zcorpan_>
Hixie: what made you special-case "mglyph" and "malignmark"?

[08:07:01.0000] <boblet>
http://www.azarask.in/blog/

[08:07:02.0000] <variable>
boblet, thanks

[08:08:00.0000] <zcorpan_>
Hixie: that also has <mi><msub>

[08:09:00.0000] <boblet>
variable: sorry if it’s not Aza, it was a Firefox UI designer’s blog tho. I remember thinking just do it like Apple Mail

[08:10:00.0000] <variable>
boblet, if you find the link let me know please

[08:10:01.0000] <variable>
I'm goolging as well

[08:10:02.0000] <variable>
but my google fu isn't as good as it used to be. I think google is getting worse :\

[08:11:00.0000] <boblet>
heh, the web is getting larger

[08:11:01.0000] <boblet>
will do

[08:13:00.0000] <boblet>
oh dear, it seems the English defeat has had an effect on my fellow doctors: http://isbrucepartofhtml5.com/

[08:13:01.0000] <boblet>
they’re obviously hitting the gin still a little too hard

[08:14:00.0000] <Dashiva^2>
All these domain purchases are helping rejuvenate the world economy, if nothing else

[08:15:00.0000] <jgraham>
Also, knob has a k in it

[08:15:01.0000] <variable>
Everyone says domain purchases are cheap. but if you have 50 of them thats $500/yr :\

[08:16:00.0000] <Rik`>
if you're going to do a lot jokes about html5, just buy partofhtml5.com

[08:22:00.0000] <Lachy>
Rik`, annevk already suggested that

[08:22:01.0000] <annevk>
variable, but with all the banner income you'll be rich!

[08:23:00.0000] <Lachy>
haha. That only works if people actually visit your site, and the best way to to do that is by domain squatting

[08:23:01.0000] <MikeSmith>
haha http://isbrucepartofhtml5.com/ is great

[08:24:00.0000] <variable>
annevk, the owner of com.com (IIRC its cnet) should be rich. Everyone else has to hope that people visit the websites. That is why all the domain parking sites take a monthly fee instead of a percentage

[08:26:00.0000] <Lachy>
anyone know who set up the bruce one?

[08:26:01.0000] <variable>
I don't get it.

[08:27:00.0000] <Philip`>
Lachy: View source

[08:27:01.0000] <Lachy>
from a comment in the source "All of these elements should really be <bruce>'s but those damn standards folks just won't let it happen! Give the man some love: @brucel"

[08:27:02.0000] <Lachy>
:-)

[08:27:03.0000] <MikeSmith>
zcorpan_: mathml schema has "token.content = mglyph|malignmark|text"

[08:28:00.0000] <variable>
who is bruce ?

[08:28:01.0000] <MikeSmith>
zcorpan_: "With the exception of the empty mspace element, token elements can contain any sequence of zero or more Unicode characters, or mglyph or malignmark elements."

[08:28:02.0000] <MikeSmith>
http://www.w3.org/TR/MathML3/chapter3.html#presm.tokenchars

[08:29:00.0000] <annevk>
variable, that sounds somewhat naive

[08:29:01.0000] <MikeSmith>
zcorpan_: so in standard mathml, those examples you found are not valid, fwiw

[08:30:00.0000] <annevk>
variable, I made quite a bit of money for a while, but I stopped doing advertising

[08:30:01.0000] <variable>
annevk, with content? or with regular parking?

[08:31:00.0000] <annevk>
variable, also, http://www.brucelawson.co.uk/

[08:31:01.0000] <variable>
annevk, oh - I've been reading that blog for a while. I need to remember names better :\

[08:33:00.0000] <MikeSmith>
zcorpan_: I don't know what to make of the one that has <xsl:text> .. but the standard mathml schema does not allow foreign-namespace content like that anywhere except with (effing) annotation-xml

[08:38:00.0000] <davidc>
mike: wasn't that a fragment of xslt generating mathml, rather than mathml?

[08:38:01.0000] <MikeSmith>
davidc!

[08:38:02.0000] <zcorpan_>
MikeSmith: ok

[08:38:03.0000] <zcorpan_>
i just filed http://www.w3.org/Bugs/Public/show_bug.cgi?id=10047

[08:39:00.0000] <MikeSmith>
davidc: you mean as part of an XSL stylesheet?

[08:39:01.0000] <MikeSmith>
I guess it must be

[08:39:02.0000] <MikeSmith>
if so, that'd certainly make a lot more sense

[08:39:03.0000] <zcorpan_>
MikeSmith: the xsl one was just a WTF (how many levels of escaping can you get for a single character?)

[08:39:04.0000] <davidc>
yes, although I missed the initial conext but <mo><xsl:text disable-output-escaping="yes"><![CDATA[&InvisibleTimes;]]></xsl:text></mo> is xslt not mathml

[08:40:00.0000] <MikeSmith>
zcorpan_: ah

[08:41:00.0000] <MikeSmith>
davidc: yeah, I just misunderstood zcorpan_ reason for mentioning it

[08:42:00.0000] <MikeSmith>
davidc: the spec for annotation-xml is giving me heartburn

[08:43:00.0000] <MikeSmith>
I would be curious to hear details about how non-browser mathml UAs implement support for that

[08:43:01.0000] <jgraham>
Argh. How the f— did Foreign Content mode get so complex 

[08:47:00.0000] <davidc>
MikeSmith: on the heartburn, sorry about that, but there must be something that I'm missing, really the only thing the browser needs to do is to correctly step past the content.  Many MathML editing tools stick on all sorts of annotations (eg original source formats, or fallback display formats or whatever) and once we start telling people they can use mathml in html, it would be good if you...

[08:47:01.0000] <davidc>
...could just drop these iannotated expressions in and have the annotations not mess things up. having more support is a bonus, but just ignoring the annotations for rendering is a good start.

[08:49:00.0000] <zcorpan_>
davidc: "there are essentially no real world cases of html in math in a text/html document as this (as you know) is all new." - this is not the case, there *are* text/html pages with bogus <math> elements in them (and expect it to not affect anything)

[08:49:01.0000] <zcorpan_>
davidc: which existed before the html5 parser had any of this

[08:49:02.0000] <davidc>
zcorpan_: 

[08:51:00.0000] <davidc>
zcorpan_: yes but that's my point, if some weird pages misused some syntax for some other use,  too bad.  they could never have worked cross browser, they were always in error, how can you consider disabling mathml forever because of these pages?

[08:52:00.0000] <boblet>
oh dear. we appear to have started … something : http://ishtml5goingtokillflash.com/ 

[08:52:01.0000] <zcorpan_>
davidc: why would they not work cross-browser? all browsers do pretty much the same for unknown tags

[08:54:00.0000] <davidc>
zcorpan: yes well it depends what's inside the elements,  i was thinking you meant some pseudo-mathml, but i guess you mean <math><i>x</i>=3</math> or something?

[08:54:01.0000] <zcorpan_>
davidc: my question about html in annotation-xml was supposed to be read as xhtml in annotation-xml (mozilla has supported xhtml and mathml for a long time)

[08:54:02.0000] <zcorpan_>
davidc: yep. or just <math><!doctype html> (rest of page)

[08:55:00.0000] <davidc>
zcorpan:  but if you found <video><i>x</i>=3</video> on some page somewhere would you have not added video?

[08:56:00.0000] <annevk>
XMLHttpRequest now has class=domintro boxes too

[08:56:01.0000] <annevk>
so you can quickly tell what each method does and whether it throws (and what)

[08:57:00.0000] <zcorpan_>
davidc: we would have called video something else, yes, if it was considered too much breakage

[08:58:00.0000] <davidc>
zcorpan:  on xhtml in annotation-xm, I really don't know we have a lot of mathml here (at nag.co.uk) but i strip it down as much as I can so it works as cross browser as possible, other people are in same boat. putting  mathml into text/html is supposed to make it easier, having undocumentable parsing rules for annotations rather works against that

[08:58:01.0000] <zcorpan_>
davidc: (for instance we renamed <input action> to <input formaction> for web compat that had bogus <input action>, and we're now considering renaming <input required> since pages have bogus <input required>)

[08:58:02.0000] <annevk>
davidb, undocumentable?

[08:59:00.0000] <davidb_>
heh too many davids

[08:59:01.0000] <annevk>
oops

[09:00:00.0000] <davidc>
zcorpan:  yes I would guess that there's orders of magnitude more pages with <input... than  <math in current text/html, and you were naming new features, not features specified a decade ago which are used by the main authoring tools

[09:00:01.0000] <zcorpan_>
davidc: i guess we could make <annotation-xml> kind of scoping so that it doesn't break out so easily

[09:01:00.0000] <MikeSmith>
yeah

[09:01:01.0000] <MikeSmith>
the problem is determining whether that is worth doing

[09:02:00.0000] <davidc>
MikeSmith: making me happy must be worth doing

[09:03:00.0000] <MikeSmith>
heh :)

[09:03:01.0000] <MikeSmith>
yeah, it's worth something

[09:03:02.0000] <zcorpan_>
davidc: there might be other ways to make you happy :)

[09:03:03.0000] <davidc>
zcorpan_: sadly not (speaks the xml geek at heart)

[09:04:00.0000] <MikeSmith>
davidc: about the spec and HTML content within MathML content: the current handling is such that a conformant parser does what I described in my message, except within token elements

[09:05:00.0000] <MikeSmith>
that is, it basically ends up moving the HTML content out of document order to follow the mathml content

[09:05:01.0000] <MikeSmith>
so it does not do that only for the annotation-xml case

[09:06:00.0000] <jgraham>
/me has not been following but 

[09:06:01.0000] <davidc>
MikeSmith: well in the other cases the input is in error so I don't really care what fixup the html parser does.,

[09:06:02.0000] <jgraham>
erm, no

[09:07:00.0000] <jgraham>
can someone give me an example of the problem markup

[09:07:01.0000] <MikeSmith>
davidc: we can make it an error in annotation-xml also, within HTML documents

[09:07:02.0000] <MikeSmith>
jgraham: <annotation-xml><img src=foo>

[09:08:00.0000] <MikeSmith>
jgraham: <mathml><semantics><mi></mi><annotation-xml><img src=foo>

[09:08:01.0000] <jgraham>
Oh, I thought that was fine

[09:08:02.0000] <davidc>
MikeSmith: You could except it will look a bit odd to say that annotation-xml can have as child any markup from any vocabulary except in html it cant have html. Wouldn't it? 

[09:08:03.0000] <jgraham>
but I was being dumb

[09:09:00.0000] <jgraham>
we should just s/whose tag name is "svg"// in the spec

[09:09:01.0000] <jgraham>
I think that would solve the problem

[09:09:02.0000] <zcorpan_>
jgraham: you need to support <annotation-xml><somemathmlelement> also

[09:09:03.0000] <MikeSmith>
davidc: well, we could say that annotation-xml is not allowed at all in text/html documents

[09:09:04.0000] <MikeSmith>
jgraham: what zcorpan_ said

[09:10:00.0000] <MikeSmith>
that is the big complication

[09:10:01.0000] <jgraham>
Hmm?

[09:10:02.0000] <jgraham>
In what way support?

[09:10:03.0000] <MikeSmith>
that is what makes it different from the mathml token content elements

[09:10:04.0000] <MikeSmith>
*token elements

[09:10:05.0000] <MikeSmith>
the token elements can't contain mathml

[09:11:00.0000] <zcorpan_>
jgraham: <annotation-xml cd="mathmlkeys" name="contentequiv" encoding="MathML-Content">  <apply>

[09:11:01.0000] <davidc>
MikeSmith: yes but as I said somewhere if the mathml output from editors and CA systems etc is anotated then saying you can't use annotations in html means the editors have to change or users have to strip out the content or both

[09:11:02.0000] <MikeSmith>
jgraham: mathml within annotation-xml needs to be parsed as mathml, while html within mathml would need to be parsed as html

[09:12:00.0000] <MikeSmith>
davidc: well, somebody has to deal with costs here

[09:12:01.0000] <jgraham>
MikeSmith: Why? I thought it was fine to just ignore anything in annotation-xml?

[09:13:00.0000] <MikeSmith>
browser projects should not be expected to incur costs of implementing something with additional complexity unless the use cases for it are clear

[09:13:01.0000] <MikeSmith>
jgraham: it is display:none by *default*

[09:14:00.0000] <MikeSmith>
but what if an author sets its display property to something other than "none"?

[09:14:01.0000] <jgraham>
MikeSmith: It seems like we could just change the rules to put everything in annotation-xml in the html namespace. What use case requires something else?

[09:14:02.0000] <MikeSmith>
svg

[09:14:03.0000] <MikeSmith>
and mathml child content

[09:14:04.0000] <jgraham>
What does it mean if you do those things?

[09:14:05.0000] <jgraham>
What's the use case?

[09:15:00.0000] <davidc>
MikeSmith: yes but in my mental model just getting to the end of /annotation-xml without falling over is the simplest of options, not the most complicated, but perhaps 10 minutes on this irc isn't long enough to get my mental model aligned with the collective...

[09:15:01.0000] <MikeSmith>
jgraham: see http://golem.ph.utexas.edu/~distler/blog/archives/001475.html for some examples

[09:15:02.0000] <zcorpan_>
jgraham: you can annotate presentation-mathml with content-mathml in annotation-xml and vice versa

[09:15:03.0000] <MikeSmith>
jgraham: we are trying to determine what the use cases are

[09:16:00.0000] <MikeSmith>
davidc: it is not the simplest of options, due to the fact that the spec chooses to allow both mathml and non-mathml content in annotation-xml

[09:16:01.0000] <jgraham>
MikeSmith: I have seen Jacques example before, but if the svg is actually a representation of an operator or something, it is hardly an annotation

[09:17:00.0000] <jgraham>
isn't it possible to put svg in other places in MathML?

[09:17:01.0000] <jgraham>
/me guesses he should read the spec

[09:17:02.0000] <MikeSmith>
davidc: the reasons is because the parsing algorithm necessarily must handle HTML content within MathML in the way I've described

[09:17:03.0000] <davidc>
MikeSmith: well it was written for a simpler universe where every language used the same parsing model to avoid these problems.

[09:17:04.0000] <MikeSmith>
davidc: except for the special case of the token elements

[09:18:00.0000] <MikeSmith>
davidc: that is not the world we seem to find ourselves in

[09:18:01.0000] <davidc>
MikeSmith: oh? :-)

[09:18:02.0000] <jgraham>
zcorpan_: I think I would like to pretend that content mathml doesn't exist for the purpose of the web

[09:19:00.0000] <MikeSmith>
jgraham: it is not possible if you limit yourself to the standard/default MathML schema

[09:19:01.0000] <MikeSmith>
the standard schema only allows other namespaced content in annotation-xml

[09:19:02.0000] <jgraham>
Hmm

[09:19:03.0000] <MikeSmith>
that is why Jacques did it that way

[09:19:04.0000] <jgraham>
This seems like a lot of trouble

[09:20:00.0000] <MikeSmith>
but HTML5 parser allows it in the mathml "token elements" -- mi, mn, mn, mo, mtext

[09:20:01.0000] <MikeSmith>
so if Jacques were to use HTML5, he could/should just put it in mtext

[09:20:02.0000] <davidc>
jgraham: It's no trouble at all in the xhtml+mathml world for which it was designed, you can reliably parse past any content in annotation-xml, and by default you can always ignore it.

[09:20:03.0000] <jgraham>
I think I would prefer that approach in the long term

[09:20:04.0000] <MikeSmith>
what Jacques has in his examples is not even valid

[09:21:00.0000] <MikeSmith>
Jacques examples have <semantics><annotation-xml> (annotation-xml element as 1st child of semantics, which is not valid)

[09:22:00.0000] <davidc>
MikeSmith: if the use case is to make the svg appear in the math, putting in a token element rather than the annotation makes perfect sense

[09:22:01.0000] <MikeSmith>
Jacques it just relying there on some rendering behavior that just happens to work in Firefox but that is not really conformant to the MathML spec

[09:23:00.0000] <MikeSmith>
davidc: that is exactly what Jacques is using it for

[09:23:01.0000] <davidc>
If annotation-xml is proving difficult in text/html then just parsing it and leaving it in the dom unrendered would do, or if that's not possible, even just throwing the content away would do.

[09:23:02.0000] <jgraham>
MikeSmith: Based on what davidc is saying it seems like we really don't want to make annotation-xml insanely complex because it is mostly just for tool-specific junk that is not relevant to end users

[09:23:03.0000] <MikeSmith>
it is exactly what every page I have seen on the Web is using <annotation-xml> for

[09:23:04.0000] <MikeSmith>
davidc: understood

[09:23:05.0000] <davidc>
Yes but like the rest of us he tries to tread a line between what's specified and what works. He would change if what was specified also worked.

[09:25:00.0000] <MikeSmith>
davidc: I suggested earlier that we have UAs rely on the @encoding attribute of annotation-xml to address this issue

[09:25:01.0000] <MikeSmith>
but zcorpan_ pooh-poohed that for some reason

[09:25:02.0000] <MikeSmith>
I can't remember why

[09:26:00.0000] <davidc>
jgraham: well it might be relevant to end users, but it isn't relevant to the browser except for just being a data format. If I cut and paste an expression out of the browser, it would be nice (and works by the way in IE) if the annotations went along with it, then if I paste into say (maple0 it may well be able to evaluate the content mathml, even if teh presentation mathml is showing some...

[09:26:01.0000] <davidc>
...strange notation only used in one textbook

[09:26:02.0000] <MikeSmith>
e.g, in order to have some annotation-xml content parsed as HTML, you'd need to do annotation-xml/@encoding=text/html

[09:26:03.0000] <zcorpan_>
MikeSmith: because it complicates the parser and we've generally tried to avoid inspecting attributes in the parser

[09:27:00.0000] <MikeSmith>
zcorpan_: OK

[09:27:01.0000] <MikeSmith>
ah yeah

[09:27:02.0000] <MikeSmith>
that makes sense, yeah

[09:28:00.0000] <MikeSmith>
davidc: what zcorpan_ said .. inspecting the values of particular attributes during DOM tree-building is something to be avoided

[09:28:01.0000] <davidc>
MikeSmith: agreed

[09:30:00.0000] <MikeSmith>
davidc: stepping back, let me make clear: as currently specced, everything essentially works as users might expect -- with the one exception of the case of HTML content in annotation-xml

[09:31:00.0000] <davidc>
All I really care about is that the parser 9using whatever parsing model is convenient for you to specify) gets as far as </annotation-xml> and then carries on the rendering as if it wasn't there. It would be nice if it stuffed the annotation somewhere usable by script but that isn't absolutely essential.

[09:31:01.0000] <davidc>
MikeSmith (how do you folks keep up with teh bandwidth on irc conversations:-) perhaps that is not as bad as it originally seemed

[09:32:00.0000] <MikeSmith>
davidc: so it will essentially to that for everything except HTML in annotation-xml

[09:32:01.0000] <davidc>
It would be good if there was some way to get an html annotation someone mentioned using a dov containere as a possibility?

[09:32:02.0000] <MikeSmith>
davidc: yeah, that is possibility

[09:32:03.0000] <MikeSmith>
but that is why I asked for use cases on www-html

[09:32:04.0000] <davidc>
modulo my spelling

[09:33:00.0000] <MikeSmith>
I can't personally help to make a compelling case to implementors for changing the spec unless I understand myself what the use cases are

[09:33:01.0000] <zcorpan_>
davidc: do you want <annotation> to similarly survive-but-not-render?

[09:34:00.0000] <davidc>
yes

[09:35:00.0000] <davidc>
zcorpan: basically (in my mental model of the universe) they should be treated pretty much as unknown attributes, carrying data but by default not doing anything

[09:35:01.0000] <jgraham>
We could treat annotation-xml as RCDATA or something. I wonder if that would break sites

[09:36:00.0000] <zcorpan_>
CDATA in that case

[09:37:00.0000] <davidc>
jgraham; I would parse the annotation using normal text parsing: If people put elements in there do whatever html fixup you find convenient, but don't go to trouble to make it "work"

[09:38:00.0000] <davidc>
jgraham: sorry scratch that you were talking about annotation-xml, I read it as ,annotation> sorry

[09:39:00.0000] <davidc>
apparently, i have a day job, got to go

[09:40:00.0000] <MikeSmith>
a possible use case for HTML in annotation-xml that I can think of have been but reluctant to bring up is audio content

[09:40:01.0000] <MikeSmith>
I don't know if the math wg has thought of that one yet

[09:40:02.0000] <MikeSmith>
but I am sure somebody else will eventually

[09:42:00.0000] <MikeSmith>
I would think that one possible use case for annotation-xml is to provide alternative representations for accessibility purposes

[09:42:01.0000] <MikeSmith>
e.g., a recording of an actual human being reading out a mathml expression

[09:43:00.0000] <jgraham>
MikeSmith: Hmm, that would work with the "just parse it all as HTML" model but not the CDATA model

[09:43:01.0000] <MikeSmith>
jgraham: yeah

[09:44:00.0000] <MikeSmith>
which is why I think the HTML case might be important

[09:44:01.0000] <MikeSmith>
but I'd rather not do their use-case documenting for them

[09:45:00.0000] <MikeSmith>
I don't know what the current accessibility story is with MathML

[09:45:01.0000] <MikeSmith>
but I would like to figure it out

[09:46:00.0000] <MikeSmith>
preemptively

[09:46:01.0000] <MikeSmith>
before it blossoms into a contentious Issue for the HTML WG 

[09:46:02.0000] <MikeSmith>
maybe they already have all kinds of accessibility hooks in there, for all I know

[09:48:00.0000] <zcorpan_>
did Hixie publish a list of pages with bogus <math> and <svg> from his research?

[09:51:00.0000] <zcorpan_>
http://twitter.com/ibmkhd/statuses/17341570393 - disappointing his link is a 404

[09:55:00.0000] <zcorpan_>
http://lazeroids.com/

[09:55:01.0000] <zcorpan_>
get an error in firefox 'console not defined'

[09:56:00.0000] <jgraham>
http://www.bennadel.com/blog/1955-Pushing-Base64-Encoded-Images-Over-HTML5-WebSockets-With-Pusher-And-ColdFusion.htm

[09:56:01.0000] <zcorpan_>
thanks

[09:56:02.0000] <jgraham>
zcorpan_: Open firebug :)

[09:56:03.0000] <jgraham>
(for the "console not defined" thing)

[10:01:00.0000] <zcorpan_>
don't have firebug

[10:03:00.0000] <MikeSmith>
zcorpan_: he meant Web Inspector

[10:04:00.0000] <zcorpan_>
in opera i get TypeError: 'console.dir' is not a function

[10:05:00.0000] <jgraham>
zcorpan_: It does indeed work with firebug open

[10:05:01.0000] <jgraham>
report a bug?

[10:07:00.0000] <gsnedders>
http://iscss3partofhtml5.com/

[10:07:01.0000] <zcorpan_>
reported

[10:07:02.0000] <zcorpan_>
gsnedders: old news

[10:07:03.0000] <gsnedders>
Meh.

[10:36:00.0000] <hsivonen>
whoa. a bug about the axis attribute!

[10:38:00.0000] <jgraham>
oh no not again

[10:39:00.0000] <zcorpan_>
where?

[10:39:01.0000] <jgraham>
or do you mean a Gecko bug?

[10:42:00.0000] <hsivonen>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=10045

[10:46:00.0000] <jgraham>
Seems easy to close: "no axis is not going to be supported"

[10:50:00.0000] <Rik`>
video::before {content : "a"} should work, right ?

[11:00:00.0000] <zcorpan_>
Rik`: seems to work in opera (same with img)

[11:11:00.0000] <Rik`>
zcorpan_: well, Opera was the only one I haven't tested

[11:37:00.0000] <TabAtkins>
Oh wow, that lazeroids game *really* needs to put a cap on how fast your ship can go.

[11:38:00.0000] <TabAtkins>
It's too easy to just jam on the up key and quickly be moving fast enough that you advance past each screen before it can scroll to show you.

[11:43:00.0000] <zcorpan_>
TabAtkins: file a bug :)

[11:45:00.0000] <TabAtkins>
zcorpan_: Too late, already bored!

[11:47:00.0000] <TabAtkins>
Also, heh, hitting a bullet with another bullet destroys both for 0 points.

[11:49:00.0000] <paul_irish>
hsivonen: yahoo wanted to distribute modernizer as a YUI module and apparently needed it licensed BSD to satify something something. 

[11:50:00.0000] <paul_irish>
it didn't make much sense to me.

[11:51:00.0000] <AryehGregor>
TabAtkins, it's physically unrealistic for spaceships to be unable to accelerate without limit.

[11:51:01.0000] <AryehGregor>
(although once they reach relativistic velocities, they'll start seeing diminishing returns from the same fuel investment)

[11:52:00.0000] <TabAtkins>
AryehGregor: In space, constant thrust = constant velocity.  TV teaches us this.

[11:52:01.0000] <AryehGregor>
I said "physically unrealistic", not "televisionally unrealistic".

[11:53:00.0000] <TabAtkins>
Computers are like TVs, so clearly they should behave similarly.

[11:53:01.0000] <Philip`>
AryehGregor: Acceleration requires fuel, and fuel is limited

[11:53:02.0000] <AryehGregor>
Whoa, massively-multiplayer Asteroids?

[11:53:03.0000] <AryehGregor>
Philip`, yes, but that's not the constraint we're discussing here.

[11:54:00.0000] <AryehGregor>
If your fuel ran out and your acceleration was thereby limited, that would be different.

[11:54:01.0000] <AryehGregor>
That would be entirely realistic, since fuel is likely to be a serious limiting factor even for advanced spacecraft.

[11:54:02.0000] <AryehGregor>
Depending on how fast you want to travel.

[11:54:03.0000] <Philip`>
Perhaps you could say the spaceship has very limited fuel(/energy) storage (to save mass) and so it has to generate energy as it's required, and it can use that energy for either acceleration or shielding against micrometeorites

[11:54:04.0000] <TabAtkins>
In any case, I'm talking game-fun.  If you go fast enough, you literally hover off the edge of the screen.

[11:55:00.0000] <Philip`>
and you're in a region of space where there's a lot of very slow-moving micrometeorites

[11:55:01.0000] <Philip`>
so if you're stationary relative to them, then you don't need much shielding and can accelerate fast

[11:55:02.0000] <TabAtkins>
You're just reintroducing drag with more technobabble.

[11:55:03.0000] <Philip`>
but as you move faster you have to slow down acceleration, until you reach a top speed where all your energy is shielding you from collisions

[11:55:04.0000] <TabAtkins>
Just say that space is filled with luminiferous ether and be done with it.

[11:56:00.0000] <Philip`>
It's not drag since you're not transferring momentum to your surroundings

[11:56:01.0000] <TabAtkins>
The effect is the same - as you go faster, an outside force gradually saps more of your acceleration.

[11:57:00.0000] <AryehGregor>
Philip`, your technobabble is fairly impressive.

[11:57:01.0000] <AryehGregor>
Now just tell me what the heck these "shields" are that can deflect micrometeorites if you put energy into them.

[11:57:02.0000] <AryehGregor>
(also, we aren't talking about light, what does the luminiferous ether have to do with it?)

[11:57:03.0000] <Philip`>
TabAtkins: Sure - the point is to try to have a more physically plausible explanation for the same effects, so you don't have to implement stupid impossible gameplay

[11:58:00.0000] <TabAtkins>
AryehGregor: The ether also produces a drag on mass, clearly.

[11:58:01.0000] <Philip`>
(Realistically you'd attack a planet by accelerating an asteroid into an orbit that will collide with it a few years later, but it's much more fun to have an explanation for why you'd want to build a Death Star)

[11:59:00.0000] <TabAtkins>
AryehGregor: The shields are magnetically-contained plasma envelope that vaporize micrometeorites before they touch the hull.

[11:59:01.0000] <AryehGregor>
Philip`, then your target could decelerate it.  They'd be able to spot you accelerating such a large object that much from billions of miles away.

[12:00:00.0000] <Philip`>
AryehGregor: The 'shields' are just lasers that vapourise micrometeorites, so that the collision energy is dispersed and easily handled by the hull

[12:00:01.0000] <AryehGregor>
TabAtkins, if they're magnetically contained, why do you need any energy input?  Also, vaporizing them doesn't get rid of the kinetic energy.

[12:00:02.0000] <TabAtkins>
AryehGregor: You start the acceleration out pretty far, preferably off the ecliptic as well so it's harder to notice.

[12:00:03.0000] <AryehGregor>
Philip`, okay, I'm beginning to buy this.

[12:01:00.0000] <AryehGregor>
TabAtkins, the further away you accelerate it from, the longer it takes to get there.  Unless you accelerate it more, but then that cancels out the greater distance, so it doesn't help you.

[12:01:01.0000] <TabAtkins>
AryehGregor: Electromagnets.  Also, vaporizing them allows the plasma to absorb their momentum, thus spreading the impact over the entire magnetic field rather than focused on a point.

[12:01:02.0000] <Philip`>
AryehGregor: You find a rock that's going to go close to the sun, secretly attach some rockets to it, then activate them when it's got close so the light from the sun masks the rocket activity

[12:01:03.0000] <TabAtkins>
AryehGregor: The point is that the further you start it out, the bigger head start you have.

[12:01:04.0000] <AryehGregor>
Electromagnets don't consume energy to stay on, if you're smart and use a superconductor.

[12:02:00.0000] <AryehGregor>
Philip`, but it will be clearly visible from a different angle.

[12:02:01.0000] <TabAtkins>
AryehGregor: These people aren't smart enough to realize that the ether isn't real, you think they'd use a superconductor?

[12:02:02.0000] <Philip`>
or you just stick a giant screen around the rock before you boost it so that nobody sees

[12:02:03.0000] <AryehGregor>
TabAtkins, well, since space is only 4 K anyway, it seems like a pretty obvious idea.

[12:02:04.0000] <AryehGregor>
Philip`, the heat will leak out through the screen, you can't prevent that.

[12:02:05.0000] <Philip`>
AryehGregor: Space is big and you can't watch all of it all the time from all directions with highly accurate instruments

[12:03:00.0000] <TabAtkins>
Thus why I suggested launching it from out of the ecliptic.

[12:03:01.0000] <AryehGregor>
You don't need highly accurate instruments to spot that much acceleration being applied.  The thermal radiation would be greater than your typical star.

[12:03:02.0000] <TabAtkins>
Nobody looks out there.

[12:03:03.0000] <AryehGregor>
/me goes to look up his website

[12:03:04.0000] <Philip`>
and there's stars and rocks and ice and supernovas and spaceships everywhere, so it's easy to hide

[12:03:05.0000] <Philip`>
AryehGregor: You apply the acceleration very very slowly

[12:04:00.0000] <TabAtkins>
True - you need very little acceleration to make it work.

[12:04:01.0000] <AryehGregor>
http://www.projectrho.com/rocket/index.html

[12:04:02.0000] <TabAtkins>
Hell, use a magdriver rather than an engine.

[12:05:00.0000] <TabAtkins>
A couple of robots, a railgun attached to a nuclear pile throwing off mass from the rock, the whole thing burrowed in so the rock itself acts as thermal shielding...

[12:05:01.0000] <AryehGregor>
http://www.projectrho.com/rocket/rocket3w.html#nostealth

[12:05:02.0000] <TabAtkins>
No stealth for rocket engines, sure.  Those things are crazy.

[12:05:03.0000] <AryehGregor>
The rock won't act as thermal shielding over long periods of time.  The heat has to escape somehow.

[12:05:04.0000] <AryehGregor>
No, no stealth for anything reasonably warm.

[12:06:00.0000] <TabAtkins>
Hm, true.  Probably better to just put it on the surface and keep it low-power so it can radiate as much as possible while facing away from detection.

[12:06:01.0000] <Philip`>
It doesn't need to work for long periods of time, just long enough that your target doesn't have time to reach it and accelerate it in a different direction

[12:06:02.0000] <AryehGregor>
Anything that's, say, human body temperature will be visible from insane distances.  Although maybe if it were a robot, with a very cool engine . . . how cool can engines get?

[12:06:03.0000] <TabAtkins>
Very cool?

[12:06:04.0000] <AryehGregor>
You've inevitably got thermodynamic inefficiencies here, no?

[12:06:05.0000] <AryehGregor>
Railguns get hot.

[12:07:00.0000] <Philip`>
and you can stick a load of laser defence towers on the rock to discourage your target's population from trying to touch it

[12:07:01.0000] <TabAtkins>
Fire it once every week, or whatever you need to dissipate heat sufficiently.

[12:07:02.0000] <AryehGregor>
"even a puny ship using ion drive to thrust at a measly 1/1000 of a g could be spotted at one astronomical unit."

[12:07:03.0000] <TabAtkins>
1/1000 of a g?  Such luxury!

[12:07:04.0000] <AryehGregor>
Philip`, then they accelerate a rock to blow up your rock from out of laser range?

[12:07:05.0000] <hsivonen>
paul_irish: I see.

[12:07:06.0000] <AryehGregor>
TabAtkins, then it will never actually get anywhere until the people you're annoyed at have been dead for millennia anyway . . .

[12:08:00.0000] <TabAtkins>
When you're planning to kill a planet, a century or ten here or there isn't that important.

[12:08:01.0000] <Philip`>
AryehGregor: It's much harder for them to hit your little rock than for your little rock to hit their giant planet, particularly when you have years longer to prepare for it

[12:08:02.0000] <TabAtkins>
Philip`: To be fair, planets are pretty freaking tiny.

[12:08:03.0000] <AryehGregor>
Do we need to start doing back-of-the-envelope calculations now?

[12:09:00.0000] <TabAtkins>
No.

[12:09:01.0000] <AryehGregor>
Okay.

[12:09:02.0000] <AryehGregor>
Good.

[12:09:03.0000] <TabAtkins>
I'll just pretend that warp drives are possible and the whole thing is done in hyperspace.

[12:11:00.0000] <AryehGregor>
That's one approach.

[12:11:01.0000] <AryehGregor>
Rather the most common one, I'd think.

[12:12:00.0000] <zcorpan_>
ok that's the last time i link to a space ship game

[12:13:00.0000] <TabAtkins>
Other solutions involve doing all the acceleration in a nearby star system so you *can* crank it up nice and fast, and then just waiting a century or two for it to hit at .01c or something.

[12:13:01.0000] <AryehGregor>
TabAtkins, would be clearly visible.

[12:14:00.0000] <AryehGregor>
If you accelerate it more, at greater distance so that it hits in the same time, they cancel out exactly.

[12:14:01.0000] <AryehGregor>
My link said something like that.

[12:15:00.0000] <TabAtkins>
Just use an engine with power low enough to be hidden in the star's signature.  Or even start it *behind* the star, timing things so that it just passes near the star on the way to the target.

[12:15:01.0000] <AryehGregor>
Then all the locals will see it, and spies will report back.  Anyone on the other side of the star.

[12:16:00.0000] <TabAtkins>
That's fine.

[12:16:01.0000] <AryehGregor>
The general strategy is clearly plausible, but the details would need a lot of thought.

[12:16:02.0000] <TabAtkins>
Once it gets going, it's going to be hard to stop.

[12:16:03.0000] <AryehGregor>
Very easy to deflect it enough to miss, if you're sending it from light-years away.

[12:16:04.0000] <TabAtkins>
Most plausible "asteroid deflection" plans involve putting something on the asteroid and accelerating it to the side.

[12:16:05.0000] <AryehGregor>
You only need to deflect it a minute amount.

[12:16:06.0000] <AryehGregor>
Yep.

[12:17:00.0000] <TabAtkins>
If it's already going quite fast, you'd need to match velocities to land on it, which may be difficult.

[12:17:01.0000] <AryehGregor>
So it's an attack that's far easier to deflect than to conduct.

[12:17:02.0000] <Philip`>
TabAtkins: It's no good if your attack takes so long that the political situation has changed and you're no longer enemies with the people you're sending an unstoppable asteroid to

[12:17:03.0000] <AryehGregor>
Still much, much easier than accelerating the thing to start with.

[12:17:04.0000] <TabAtkins>
Philip`: It's good enough if you're sure that your enemies are infidels whos descendants will be forever hated by your alien god.

[12:18:00.0000] <AryehGregor>
You don't need to match velocities, either.

[12:18:01.0000] <AryehGregor>
Get a few really big nukes, that's enough.

[12:18:02.0000] <AryehGregor>
Especially at a large distance.

[12:18:03.0000] <TabAtkins>
Well, maybe.  You dont' want to just fracture it.

[12:18:04.0000] <TabAtkins>
Multiple small nukes is better.

[12:18:05.0000] <AryehGregor>
Possibly.

[12:18:06.0000] <AryehGregor>
Also less chance of failure.

[12:19:00.0000] <AryehGregor>
This attack would be more plausible if you found an asteroid that was passing very close to the planet anyway, and just deflected it slightly.

[12:19:01.0000] <AryehGregor>
But then your enemy could do the same.  They'd want to know about random asteroids passing nearby anyway.

[12:20:00.0000] <AryehGregor>
So even if they don't know you did it, their asteroid defense team would notice it and go to stop it.

[12:20:01.0000] <AryehGregor>
0.01c is not very fast, plenty of time to spot it if it's large enough to blow up the world at that speed.

[12:20:02.0000] <TabAtkins>
Assuming you haven't already blown up their asteroid defense, and aren't just dropping a rock on them to finish them off.

[12:21:00.0000] <AryehGregor>
If you're just finishing them off, you don't need to get fancy.  Nuke 'em from orbit.

[12:21:01.0000] <TabAtkins>
I assume putting a cheap thruster on a nearby asteroid is less expensive than that.

[12:22:00.0000] <AryehGregor>
I really doubt that.

[12:22:01.0000] <TabAtkins>
I'd assume that "drop a big rock on them" is the least fancy you can get.  ^_^

[12:22:02.0000] <AryehGregor>
Nukes are not very expensive at all.

[12:22:03.0000] <AryehGregor>
(relative to space travel)

[12:22:04.0000] <Philip`>
/me notes that when he last read about the asteroid idea, the aim wasn't to hit the target but to pass by close enough to pull it out of its stable Lagrange point orbit

[12:22:05.0000] <AryehGregor>
Also much faster.

[12:22:06.0000] <TabAtkins>
Nukes are expensive when you have to carry them in space, though.

[12:22:07.0000] <AryehGregor>
Maybe.  It depends.

[12:23:00.0000] <TabAtkins>
Just use a couple of those nukes to deflect a rock rather than decimate a single city, and the rest of the work is done by Newton himself.

[12:26:00.0000] <AryehGregor>
After a very long time.

[12:27:00.0000] <TabAtkins>
That might be enough.  Bashing apart their industrial capacity is easy, and won't take much actual destructive power.  The rock is for sterilizing purposes.

[12:28:00.0000] <AryehGregor>
/me notes that he happened to be reading: "It is also true that if the attacker's objective is sheer devastation, nuclear bombardment from high orbit will do the job. (So will the ever popular asteroid toss, but this is just a Rube Goldberg way to get the same effect.)" http://www.rocketpunk-manifesto.com/2009/06/space-warfare-i-gravity-well.html

[12:28:01.0000] <AryehGregor>
So it's two against one now.  You're outvoted.

[12:28:02.0000] <TabAtkins>
Who cares if it takes 50 years for it to hit the planet, if they can't scrounge up enough industrial capacity in that time to do anything about it, and so you can immediately leave for the next planet in your life-killing spree.

[12:29:00.0000] <TabAtkins>
Rube Goldberg devices can sometimes be more energy-efficient.  ^_^]

[12:29:01.0000] <zcorpan_>
AryehGregor: i thought whatwg didn't do voting

[12:29:02.0000] <AryehGregor>
zcorpan_, no, but we're discussing physics, and physical reality is decided by vote.  Didn't you know?

[12:29:03.0000] <Philip`>
In Spore I usually just terraform the planet until all life is extinguished, and then terraform it back again

[12:30:00.0000] <Philip`>
and then paint the seas purple

[12:30:01.0000] <TabAtkins>
Insult to injury, I see.

[12:46:00.0000] <AryehGregor>
Am I the only one who would gladly take a cell phone that's three times as thick and heavy if it means twice the battery life?

[12:46:01.0000] <AryehGregor>
(since I clearly haven't spammed enough off-topic conversation here today)

[12:46:02.0000] <Philip`>
The markets suggests you are

[12:46:03.0000] <Philip`>
*market

[12:47:00.0000] <AryehGregor>
Maybe that's because you can immediately tell how thin and light a device is, but you don't know how long its battery lasts until you've actually bought it.

[12:47:01.0000] <AryehGregor>
First impressions and all.

[12:47:02.0000] <AryehGregor>
(since obviously no one in their right mind believes manufacturers' battery life claims)

[12:48:00.0000] <Philip`>
Do people buy phones without reading any independent reviews?

[12:48:01.0000] <Philip`>
and/or do reviews not discuss battery life?

[12:48:02.0000] <AryehGregor>
Yes in both cases.  :)

[12:49:00.0000] <Philip`>
/me would be happy with any battery life longer than the time between waking and sleeping, since he can recharge it overnight

[12:49:01.0000] <AryehGregor>
I could do that, but isn't it supposed to be bad for the battery if you recharge it before it gets low?

[12:50:00.0000] <Philip`>
(My phone seems to last at least several days before the battery indicator goes non-green, anyway)

[12:51:00.0000] <Philip`>
I don't think it's bad enough that you'll notice, particularly if you buy a new phone every couple of years, and if it's a problem you can just buy a new battery

[12:51:01.0000] <AryehGregor>
Reasonable points.

[12:51:02.0000] <AryehGregor>
What phone do you have?

[12:52:00.0000] <Philip`>
It appears to be a K530i

[12:52:01.0000] <zcorpan_>
aren't phone batteries pretty expensive?

[12:52:02.0000] <AryehGregor>
See, I got such a fancy phone (Nexus One) that its battery barely lasts 24 hours even if I don't actually use it.

[12:53:00.0000] <AryehGregor>
Google suggests $49.99 new.

[12:54:00.0000] <Philip`>
/me has never seen any references to any actual experiments on battery life and the effects of recharging, just a load of (often contradictory) apocryphal claims about the dangers of charging too much or too little or too fast or too slow

[12:54:01.0000] <Philip`>
(though maybe I just haven't looked in the right places)

[12:54:02.0000] <AryehGregor>
Philip`, well, your phone weighs significantly less than mine but its battery lasts longer, so there goes my hypothesis.  It's also shorter and narrower, also a few millimeters thicker.

[12:55:00.0000] <Philip`>
AryehGregor: Your problem is that you have a phone with features

[12:55:01.0000] <AryehGregor>
But it doesn't last as long as yours even when it's *off*.

[12:56:00.0000] <Hixie>
ok i've finished describing the websrt format authoring rules

[12:56:01.0000] <Hixie>
now i just have the parser to finish, and the renderer, and the DOM API, and the examples, and the....

[12:57:00.0000] <Philip`>
Then I guess you merely need to convince the world to adopt the format

[13:02:00.0000] <Hixie>
details

[13:02:01.0000] <zcorpan_>
Hixie: surely the cue identifier can't contain linebreaks?

[13:02:02.0000] <zcorpan_>
A WebSRT cue identifier is any sequence of one or more characters not containing the substring "-->" (U+002D HYPHEN-MINUS, U+002D HYPHEN-MINUS, U+003E GREATER-THAN SIGN).

[13:03:00.0000] <Hixie>
good catch

[13:04:00.0000] <zcorpan_>
also, shouldn't there be a linebreak separating the cue identifier and the cue timings?

[13:09:00.0000] <jgraham>
AryehGregor: I belive with lithium ion batteries you are supposed to recharge little and often rather than running them right down

[13:09:01.0000] <jgraham>
That was some other type of battery

[13:09:02.0000] <jgraham>
Hixie: You are working on the parser now? Awesome :)

[13:10:00.0000] <Hixie>
been working on the websrt parser for a while now

[13:10:01.0000] <Hixie>
it's mostly done except for the cue text

[13:10:02.0000] <Hixie>
(which of course is the hard part)

[13:10:03.0000] <jgraham>
(I meant the HTML parser)

[13:10:04.0000] <AryehGregor>
jgraham, Wikipedia says they have no memory effect, and that sounds like what I was remembering, so you're probably right.  Thanks.

[13:10:05.0000] <jgraham>
(but I misunderstood you)

[13:10:06.0000] <jgraham>
(so ignore me)

[13:13:00.0000] <gsnedders>
(I often ignore him, it seems quite effective)

[13:13:01.0000] <jgraham>
(I know)

[13:13:02.0000] <gsnedders>
(Then he throws things at me, or gets people to throw things at me)

[13:13:03.0000] <gsnedders>
(that kinda sucks)

[13:14:00.0000] <jgraham>
(that hardly ever happens)

[13:15:00.0000] <jgraham>
(mostly other people throw thingat you of their own volition)

[13:18:00.0000] <TabAtkins>
(I'm whispering in parentheses)

[13:19:00.0000] <jgraham>
(I'm shouting in parenthesis)

[13:19:01.0000] <jgraham>
(but you can' tell)

[13:19:02.0000] <TabAtkins>
(NO YOU'RE NOT, I AM)

[13:19:03.0000] <jgraham>
(*can't)

[13:23:00.0000] <dglazkov>
()

[13:23:01.0000] <jgraham>
Hmm surely we have tests for the &pound vs &prod thing

[13:24:00.0000] <gsnedders>
That's what you think.

[13:24:01.0000] <jgraham>
I would think that. I wonder if I would be right

[13:24:02.0000] <jgraham>
I also wonder if webkit are using those tests

[13:24:03.0000] <jgraham>
assuming they exit

[13:24:04.0000] <jgraham>
*exist

[13:25:00.0000] <zcorpan_>
+  <p>Each <span>WebSRT timestamp</span> in a <span>WebSRT file</span>

[13:25:01.0000] <zcorpan_>
+  must represent a time greater than than all the times represented by

[13:25:02.0000] <zcorpan_>
+  <span title="WebSRT timestamp">WebSRT timestamps</span> earlier in

[13:25:03.0000] <zcorpan_>
+  the file.</p>

[13:25:04.0000] <zcorpan_>
Hixie: i thought we wanted to support overlapping cues?

[13:26:00.0000] <Hixie>
oh, right

[13:26:01.0000] <Hixie>
good point

[13:26:02.0000] <Hixie>
hmm

[13:26:03.0000] <Hixie>
i need a way to say that the karaoke timestamps have to be in order and in the timeframe of the cue

[13:26:04.0000] <Hixie>
and that the end stamp has to be later than the start stamp

[13:26:05.0000] <zcorpan_>
start > earlier starts; end > current start

[13:26:06.0000] <variable>
Hixie, do you recall any bugs or mailing list statments where you said that we won't create a advert tag.....

[13:26:07.0000] <Hixie>
and that the start stamp has to be later than the previous start stamp

[13:26:08.0000] <TabAtkins>
variable, I sent you links yesterday about that

[13:27:00.0000] <jgraham>
variable: TabAtkins answered that yeserday, no?

[13:27:01.0000] <jgraham>
heh

[13:27:02.0000] <variable>
I recall TabAtkins saying he will but I don't recall seeing the links

[13:27:03.0000] <variable>
/me greps logs

[13:27:04.0000] <Hixie>
variable: not off-hand, but i can tell you the rationale is just that such an element wouldn't work, since as soon as people used it, other people would use it to remove their ads, and then the people would stop using it

[13:27:05.0000] <variable>
Hixie, yep - I just want to add a source ;)

[13:28:00.0000] <TabAtkins>
One sec, I'm looking up the line in the logs.

[13:28:01.0000] <TabAtkins>
http://krijnhoetmer.nl/irc-logs/whatwg/20100628#l-1013

[13:28:02.0000] <Hixie>
:-)

[13:28:03.0000] <variable>
TabAtkins, thanks. I don't see that in my logs :\

[13:29:00.0000] <variable>
% grep -R "TabAtkins" \#whatwg.log|grep "two me"|wc -l

[13:29:01.0000] <variable>
       0

[13:29:02.0000] <variable>
:\

[13:30:00.0000] <variable>
but thanks for the link

[13:31:00.0000] <annevk>
zcorpan_, I filed a bug on that missing line break separator

[13:32:00.0000] <jgraham>
/me wonders if Julian is just expecting us to return the originally quoted text, verbatim

[13:33:00.0000] <annevk>
what do you mean?

[13:34:00.0000] <Hixie>
can julian not read english?

[13:36:00.0000] <annevk>
oh, I didn't realize he had a reply in that thread

[13:37:00.0000] <annevk>
guess it makes sense for his anti-non-BNF campaign

[13:38:00.0000] <annevk>
he missed out on a "for the record" note there

[13:38:01.0000] <jgraham>
Is anti-non- the same as pro-?

[13:39:00.0000] <Hixie>
julian isn't pro anything

[13:39:01.0000] <jgraham>
It doesn't really make sense because BNF is even less like English than English is

[13:39:02.0000] <jgraham>
Most things are in fact

[13:39:03.0000] <annevk>
he never really appears pro, he just complaints

[13:39:04.0000] <AryehGregor>
jgraham, an anti-non-BNF person might accept having no grammar at all, since that wouldn't include anything non-BNF.

[13:40:00.0000] <AryehGregor>
Well, why should he bother endorsing the parts of the spec he agrees with?  They're unlikely to change unless someone happens to object.

[13:41:00.0000] <AryehGregor>
(I guess he could oppose someone else's objection, and thereby be pro- something, but then he'd be anti-the objection)

[13:41:01.0000] <jgraham>
AryehGregor: Bits of the spec I agree with kep getting changed

[13:41:02.0000] <jgraham>
Although that typically happens via people objecting

[13:42:00.0000] <jgraham>
often with the word "Formally" dangling just out of sight

[13:42:01.0000] <annevk>
saying formally is just a formality

[13:43:00.0000] <jgraham>
So, as a distraction from one of the several useful things I should be doing, does anyone know how one deals with random numbers in Haskell in a sane way

[13:44:00.0000] <TabAtkins>
"sane" meaning "without monads"?

[13:44:01.0000] <annevk>
I was planning on replacing cookies with localStorage on html5.org earlier today, but I guess I'll leave that for tomorrow

[13:44:02.0000] <jgraham>
TabAtkins: I think sane might mean "with monads but without making the whole program non-pure"

[13:44:03.0000] <Hixie>
AryehGregor: he doesn't seem to endorse anything, he just says things are bad and doesn't say what he'd prefer

[13:44:04.0000] <TabAtkins>
jgraham: http://en.wikibooks.org/wiki/Haskell/Hierarchical_libraries/Randoms#The_Standard_Random_Number_Generator

[13:45:00.0000] <TabAtkins>
That not good enough for you?

[13:45:01.0000] <jgraham>
TabAtkins: I am a total n00b though, this is just a learning project

[13:45:02.0000] <TabAtkins>
Essentially doing random numbers as a sequence.

[13:45:03.0000] <AryehGregor>
He wrote the pro-multiple-languages-in-Content-Language Change Proposal.  That counts as saying what he'd prefer.

[13:45:04.0000] <annevk>
in theory it should simplify matters, and reduce traffic I suppose

[13:45:05.0000] <hsivonen>
speaking of objections, the objector turnout is pretty low even though the polls close tomorrow

[13:45:06.0000] <TabAtkins>
I did my objecting.

[13:46:00.0000] <jgraham>
The polls close tomorrow?

[13:46:01.0000] <jgraham>
Oh

[13:48:00.0000] <zcorpan_>
Hixie: have you seen http://www.storiesinflight.com/js_videosub/ ?

[13:48:01.0000] <AryehGregor>
Apparently some people don't understand what "objections" means.  Two of the five people objecting to making the Content-Language pragma non-conforming actually support it.

[13:48:02.0000] <hsivonen>
/me wonders if metacity and compiz allow apps to draw in the title bar while the WM draws the buttons

[13:48:03.0000] <hsivonen>
as opposed to getting rid of the title bar altogether as in Chrome and Wine

[13:49:00.0000] <hsivonen>
or if apps can set a per-window synthetic theme

[13:52:00.0000] <Philip`>
jgraham: I'm sure we have tokeniser tests for character references, but they're probably not trivial to run in a browser-hosted parser implementation

[13:52:01.0000] <jgraham>
Philip`: Good point

[13:52:02.0000] <Philip`>
and I wouldn't be surprised if the html5lib tree constructor tests didn't attempt to test that part of the tokeniser

[13:52:03.0000] <jgraham>
Philip`: I am pretty sure they don't

[13:54:00.0000] <hsivonen>
Philip`: Jonathan Griffin made a harness that turned tokenizer tests into tree builder tests, but the harness hasn't landed and is rotting as bugzilla attachment :-(

[13:54:01.0000] <jgraham>
hsivonen: Bug number?

[13:54:02.0000] <jgraham>
I might resurrect that

[13:55:00.0000] <hsivonen>
jgraham: https://bugzilla.mozilla.org/show_bug.cgi?id=html5-parsing

[13:55:01.0000] <hsivonen>
jgraham: you'll need a freash build of the JS version of the parser

[13:56:00.0000] <hsivonen>
*fresh

[13:56:01.0000] <jgraham>
hsivonen: Thanks

[13:56:02.0000] <hsivonen>
/me tries to remember to recompile it tomorrow

[14:02:00.0000] <annevk>
Hixie, http://www.w3.org/Bugs/Public/show_bug.cgi?id=10024 is another WebSRT syntax issue

[14:03:00.0000] <zcorpan_>
annevk: wasn't it fixed in the last checkin?

[14:11:00.0000] <hober>
hsivonen: by "the JS version of the parser," do you mean http://github.com/aredridel/html5 ?

[14:12:00.0000] <gsnedders>
hober: I presume he means the GWT compiled version of the Java one

[14:12:01.0000] <gsnedders>
(Instead of the C++ compiled version of the Java one)

[14:12:02.0000] <gsnedders>
(Or the Java one itself)

[14:12:03.0000] <hober>
ahh, indeed.

[14:14:00.0000] <gsnedders>
(Because all ES QA love GWT)

[14:15:00.0000] <Philip`>
gsnedders: You should consider the ability to view the original source code of everything you're debugging as a privilege, not a right

[14:28:00.0000] <hsivonen>
hober: I mean the parser from http://livedom.validator.nu/

[14:29:00.0000] <TabAtkins>
Man, fuck clearance.  Just defining it as "partially-collapsed margins" would have been simpler.

[14:37:00.0000] <zcorpan_>
Hixie: "WebSRT cue text line" is a confusing term since it's not actually a line

[14:38:00.0000] <annevk>
zcorpan_, oh, missed that first change

[14:39:00.0000] <zcorpan_>
Hixie: maybe "WebSRT cue markup"

[14:39:01.0000] <TabAtkins>
cue item?

[14:40:00.0000] <zcorpan_>
is "&amp;" an item?

[14:41:00.0000] <annevk>
cue body?

[14:41:01.0000] <TabAtkins>
Ooh, I like that one.

[14:41:02.0000] <zcorpan_>
seems weird to nest bodies :)

[14:41:03.0000] <TabAtkins>
"A WebSRT cue body contains one or more of the following components, in any order:"

[14:44:00.0000] <annevk>
zcorpan_, well, you'd have to change the other bits to match I suppose

[14:45:00.0000] <zcorpan_>
annevk: how would you change http://www.whatwg.org/specs/web-apps/current-work/complete/video.html#websrt-cue-italics-span ?

[14:45:01.0000] <TabAtkins>
Actually, I think I'd be okay with it saying that you nest cue bodies.

[14:48:00.0000] <annevk>
zcorpan_, so how is it not a line?

[14:49:00.0000] <zcorpan_>
annevk: a line to me implies a line break after it

[14:49:01.0000] <annevk>
zcorpan_, the way it is defined is that you have a lines separated by line terminators; makes sense to me

[14:49:02.0000] <annevk>
zcorpan_, and line terminators cannot be part of the line

[14:50:00.0000] <zcorpan_>
annevk: see <i>, <b> and <ruby>

[14:51:00.0000] <annevk>
zcorpan_, you mean that it refers back to cue text line?

[14:51:01.0000] <zcorpan_>
yeah

[14:51:02.0000] <TabAtkins>
lines embedded in lines without line separators involved does sound kinda weird.

[14:52:00.0000] <annevk>
fair enough

[14:52:01.0000] <annevk>
I wonder about the parsing details

[14:53:00.0000] <annevk>
and also why we do not allow all entities

[14:54:00.0000] <annevk>
for instance, VLC supports &copy;

[14:56:00.0000] <annevk>
also &trade; etc.

[14:56:01.0000] <annevk>
GStreamer doesn't by default though

[14:56:02.0000] <annevk>
but GStreamer is pretty crappy with text/srt in general

[15:11:00.0000] <zcorpan_>
i wonder if it's possible to download a whole bunch of srt subtitles from somewhere for research puposes

[15:12:00.0000] <zcorpan_>
other than crawling a site and downloading them one-by-one

[15:24:00.0000] <zcorpan_>
yay lazeroids works in opera now :)

[15:24:01.0000] <zcorpan_>
(and firefox without firebug)

[15:27:00.0000] <zcorpan_>
except it doesn't seem to work so good in opera :/

[15:29:00.0000] <AryehGregor>
Lazeroids is buggy for me in Chrome.  Movement doesn't work right.

[15:29:01.0000] <AryehGregor>
/me tries again

[15:30:00.0000] <AryehGregor>
Also, I thought you said you'd 

[15:30:01.0000] <AryehGregor>
Also, I thought you said you wouldn't mention space ship games here again.

[15:33:00.0000] <AryehGregor>
I think the problem is that holding down arrow keys doesn't work for moving.

[15:33:01.0000] <AryehGregor>
Or something.

[15:34:00.0000] <zcorpan_>
well i lied

[15:36:00.0000] <annevk>
haha -- http://twitter.com/jreschke/status/17364260158

[15:37:00.0000] <annevk>
that email did only make sense in context of the rambling earlier

[15:45:00.0000] <Philip`>
On the subject of test result formats: I used YAML for canvas tests, which is nice for humans to write and hard for computers to parse

[15:45:01.0000] <Philip`>
which is completely not what's needed for test results

[15:45:02.0000] <Philip`>
so don't use that

[15:47:00.0000] <AryehGregor>
How about JSON?

[15:49:00.0000] <Philip`>
That was suggested already

[15:49:01.0000] <Philip`>
(according to the minutes)

[15:59:00.0000] <jgraham>
I suggested JSON

[16:00:00.0000] <jgraham>
Since XML is massive amounts of overkill and json is only small amounts of overkill

[16:01:00.0000] <jgraham>
Oh I just thought of another issue that wasn't discussed

[16:01:01.0000] <jgraham>
We need a way to deal with multiple tests per file

[16:01:02.0000] <jgraham>
Hmm

[16:02:00.0000] <jgraham>
Oh well, bedtime

[16:13:00.0000] <Hixie>
zcorpan_: sweet

[16:13:01.0000] <Hixie>
i haven't even half finished the spec yet and we're already got implementations

[16:14:00.0000] <TabAtkins>
Haha, I just made several people look through multiple rulebooks before they realized I was making a rule-34 joke.

[16:15:00.0000] <TabAtkins>
http://forums.xkcd.com/viewtopic.php?p=2215558#p2215558

[16:16:00.0000] <TabAtkins>
/me is a dork.

[16:19:00.0000] <Philip`>
zcorpan_: http://urd.let.rug.nl/tiedeman/OPUS/OpenSubtitles.php ?

[16:20:00.0000] <zcorpan_>
Philip`: thanks

[16:21:00.0000] <Philip`>
zcorpan_: It's not clear how many are SRT but it sounds like some are

[16:22:00.0000] <MikeSmith>
TabAtkins: nice

[16:22:01.0000] <MikeSmith>
"I don't see it. What edition?"

[16:22:02.0000] <TabAtkins>
I love being a D&D nerd.  The thought of D&D defining that is actually plausible.

[16:23:00.0000] <MikeSmith>
D&D defining anything is plausible

[16:23:01.0000] <Philip`>
zcorpan_: ...unless they've only got converted-to-XML versions in there

[16:23:02.0000] <gsnedders>
TabAtkins: Rule 34?

[16:23:03.0000] <Philip`>
zcorpan_: but I don't fancy downloading 1.6GB just to check

[16:24:00.0000] <TabAtkins>
gsnedders: http://lmgtfy.com/?q=rule+34

[16:24:01.0000] <gsnedders>
Oh.

[16:24:02.0000] <AryehGregor>
TabAtkins, they would only define that in the BoVD or something.

[16:24:03.0000] <AryehGregor>
No way would that be in the DMG.

[16:24:04.0000] <TabAtkins>
BoEF, more than likely, but that's not an actual wotc book.

[16:25:00.0000] <AryehGregor>
I know of BoVD and BoED, what's BoEF?

[16:25:01.0000] <AryehGregor>
I probably don't want to know about it.

[16:25:02.0000] <TabAtkins>
Erotic Fantasy.

[16:25:03.0000] <AryehGregor>
Yeah.

[16:26:00.0000] <MikeSmith>
reminds me of the way to give annoying people driving directions: Give them such that they'll head off in a completely wrong direction, in the boondocks, and when you are done explaining most of it, say, "...once you get around there, you might feel like you've gone too far in that direction, but don't worry, just keep going a little further and you'll find it"

[16:27:00.0000] <TabAtkins>
I actually had to always append that to directions to my old house, because otherwise I'd get calls from people driving to my place wondering where they missed their turn.

[16:28:00.0000] <AryehGregor>
Why?

[16:28:01.0000] <AryehGregor>
Also, I guess this was in the pre-Google Maps/GPS era.

[16:28:02.0000] <TabAtkins>
No, it was just a few months ago.

[16:28:03.0000] <AryehGregor>
A Luddite, then?

[16:28:04.0000] <AryehGregor>
Or do people not use those things as much as I'd think?

[16:29:00.0000] <AryehGregor>
(I don't drive, of course, since I live in Manhattan.)

[16:29:01.0000] <TabAtkins>
Houston highways are almost completely developed, so you never go any distance without seeing stores or strip malls or whatnot.  My house was just far enough out from the city that there was a stretch of forest and a creek before the strip malls started again.

[16:29:02.0000] <TabAtkins>
It confused people, because usually seeing that means you're on your way to Dallas or Austin.

[16:31:00.0000] <AryehGregor>
I see.

[16:34:00.0000] <MikeSmith>
http://www.w3.org/TR/1998/NOTE-xml-names-0119 "error on line 2 at column 30: Encoding error"

[16:34:01.0000] <MikeSmith>
/me resorts to lynx

[16:35:00.0000] <MikeSmith>
let's hope the lynx developers add an XML parser

[16:35:01.0000] <MikeSmith>
that'd be fun

[16:37:00.0000] <TabAtkins>
lynx developers just need to add <video> support via one of the video-to-ascii converters.

[16:38:00.0000] <TabAtkins>
Then I will finally be able to fulfill my dream: browsing youtube in lynx.

[16:41:00.0000] <TabAtkins>
re: xml error, hoisted by their own doctype, it appears.

[16:45:00.0000] <MikeSmith>
Murata-san is 1998 - http://lists.xml.org/archives/xml-dev/199808/msg00164.html

[16:46:00.0000] <MikeSmith>
[[

[16:46:01.0000] <MikeSmith>
As a member of the WG, I have been involved.  I have agreed on colonization, 

[16:46:02.0000] <MikeSmith>
and have always believed that colonization provides a good basis for 

[16:46:03.0000] <MikeSmith>
the namespace extension.  Now that we have local scoping and declaration by 

[16:46:04.0000] <MikeSmith>
attributes, I start to wonder.

[16:46:05.0000] <MikeSmith>
Since the same prefix can be bound to 

[16:46:06.0000] <MikeSmith>
different namespaces, it is no longer possible to construct an equivalent 

[16:46:07.0000] <MikeSmith>
XML 1.0 DTD from a collection of namespace-schema pairs.  Then, what is 

[16:46:08.0000] <MikeSmith>
the point of using prefixes?

[16:46:09.0000] <MikeSmith>
 It would have 

[16:46:10.0000] <MikeSmith>
been a lot simpler if we had introduced a reserved attribute for specifying 

[16:46:11.0000] <MikeSmith>
the namespace of the element.

[16:46:12.0000] <MikeSmith>
]]

[16:46:13.0000] <Hixie>
man that would have been hella verbose

[16:47:00.0000] <TabAtkins>
<svg namespace="foobar">?

[16:47:01.0000] <MikeSmith>
yeah, sure

[16:47:02.0000] <MikeSmith>
more <svg xmlns="foobar> I guess

[16:48:00.0000] <MikeSmith>
oops

[16:48:01.0000] <Hixie>
if it was scoping that might work for svg, html, mathml, etc; wouldn't be very pretty for rdf

[16:49:00.0000] <zcorpan_>
Philip`: seems it was xml with sentence-for-sentence translations

[16:49:01.0000] <TabAtkins>
No, rdf would have sucked with that.

[16:49:02.0000] <TabAtkins>
But it feels like it would have been saner for normal elements.

[16:50:00.0000] <MikeSmith>
I just point it out as a reminder that there was nothing close to real consensus in the WG at that time

[16:50:01.0000] <MikeSmith>
no consensus that prefix mechanism was the right way

[16:51:00.0000] <MikeSmith>
just as there was also no consensus then about draconian error handling being the right way

[16:51:01.0000] <Hixie>
just as there's no consensus in the htmlwg that html5 is the right way :-)

[16:51:02.0000] <Hixie>
consensus isn't a good language design mechanism

[16:53:00.0000] <MikeSmith>
well it's especially not a good design mechanism when claims are made that it exists for a particular issue when it really does not

[16:53:01.0000] <MikeSmith>
decision about draconian error handling was made by simple majority vote

[16:53:02.0000] <MikeSmith>
which is not consensus anyway


2010-06-30
[17:00:00.0000] <Philip`>
zcorpan_: Maybe those people or the people they got the original data from would be willing to share the original data if asked

[17:00:01.0000] <Philip`>
/me was unable to easily find easily-downloadable subtitle collections

[17:01:00.0000] <Philip`>
(which I guess is because the people who collect the subtitles want to show loads of adverts to visitors, and don't want other sites copying all the data they've collected from volunteers)

[17:03:00.0000] <zcorpan_>
it says source: http://www.opensubtitles.org

[17:10:00.0000] <zcorpan_>
Hixie: the first random srt i download has <i> wrapping multiple lines

[17:10:01.0000] <zcorpan_>
(well, the second random. the first didn't have any markup)

[17:11:00.0000] <zcorpan_>
http://www.opensubtitles.org/en/download/sub/3695049

[17:11:01.0000] <zcorpan_>
i have no idea which encoding that one is using

[17:13:00.0000] <Hixie>
zcorpan_: ah, excellent, good to know, thanks

[17:18:00.0000] <zcorpan_>
it seems existing srts use different legacy encodings and don't declare it :(

[17:19:00.0000] <variable>
Has there been any discussion on client side includes for HTML files? 

[17:19:01.0000] <roc>
yes

[17:19:02.0000] <roc>
see if <iframe seamless> is what you want

[17:22:00.0000] <variable>
the spec is the only document that has more links to itself than wikipedia ;)

[17:24:00.0000] <zcorpan_>
i wonder what to do about the encoding issue

[17:25:00.0000] <variable>
roc, as I read the spec "seamless" means that <html>___stuff___</html> is the same as <html><iframe srcdoc="____stuff____" seamless></html> - am I correct ?

[17:25:01.0000] <zcorpan_>
hopefully people will reencode their legacy subtitles to utf-8 for web use

[17:25:02.0000] <variable>
(or src="document" where document has ___stuff___ in it)

[17:25:03.0000] <AryehGregor>
variable, no, it's radically different in lots of ways.

[17:25:04.0000] <AryehGregor>
However, it should be similar enough for most purposes.

[17:26:00.0000] <roc>
what Aryeh said

[17:26:01.0000] <variable>
AryehGregor, can you explain? I'm not /that/ knolegeable. 

[17:26:02.0000] <AryehGregor>
Well, it's an iframe.

[17:26:03.0000] <variable>
*not that familier with how it works

[17:26:04.0000] <AryehGregor>
It integrates better than a regular iframe, but still an iframe.

[17:26:05.0000] <AryehGregor>
So there's a whole separate document inside.

[17:27:00.0000] <AryehGregor>
But you can make it blend in to the bigger document visually, etc.

[17:27:01.0000] <roc>
it's a separate document, it's in a rectangle

[17:27:02.0000] <variable>
AryehGregor, in a seamless iframe is top.location == location ?

[17:27:03.0000] <roc>
for example you can't expect to put inline content in it and wrap that content at line breaks

[17:27:04.0000] <roc>
it has its own DOM

[17:27:05.0000] <roc>
its top.location != location in general

[17:28:00.0000] <variable>
so what does "When specified, it indicates that the iframe  element's browsing context is to be rendered in a manner that makes it appear to be part of the containing document (seamlessly included in the parent document). "

[17:28:01.0000] <variable>
mean? or can you point me to some "newbie" docs the issue?

[17:29:00.0000] <TabAtkins>
It means that it should look like the <iframe> tag wasn't there at all.

[17:30:00.0000] <TabAtkins>
(Replaced by the contents of the contained document.)

[17:30:01.0000] <variable>
so to the enclosing page there is no iframe but to the iframe content it is on its own

[17:31:00.0000] <variable>
/me has to learn how to read the spec better

[17:31:01.0000] <TabAtkins>
Visually, yes.  In all other respects, the <iframe> is still there and acts normally.

[17:31:02.0000] <variable>
Ah. I see. 

[17:31:03.0000] <variable>
There is no "client side include" but you can get the visual effect. OK

[17:32:00.0000] <TabAtkins>
Yeah.

[17:32:01.0000] <TabAtkins>
Actualy client-side includes have to be faked with javascript and xhr.

[17:32:02.0000] <TabAtkins>
s/Actualy/Actual/

[17:32:03.0000] <variable>
ok. good to know

[17:35:00.0000] <zcorpan_>
/me finds an srt with Traducerea ∫i adaptarea: <font color=#ff99cc>Kprice</font>

[17:35:01.0000] <zcorpan_>
/me <font color=#ffffff>Subtitr„ri-Noi Team</font>

[17:36:00.0000] <TabAtkins>
Is that an integral sign?

[17:36:01.0000] <variable>
TabAtkins, yes

[17:36:02.0000] <zcorpan_>
wrong encoding

[18:50:00.0000] <MikeSmith>
um

[18:51:00.0000] <MikeSmith>
how is the word "activateable" spelled?

[18:51:01.0000] <MikeSmith>
is that not a word?

[18:54:00.0000] <MikeSmith>
activatable

[18:54:01.0000] <MikeSmith>
only ~100,000 ocurrances

[18:55:00.0000] <MikeSmith>
/me wonders what other ways people use to express the "activatable"

[18:57:00.0000] <Philip`>
"can be activated"

[18:58:00.0000] <MikeSmith>
Philip`: my context is, "Make drag-and-drop events keyboard-activatable"

[18:59:00.0000] <MikeSmith>
is "Make drag-and-drop events fireable by keyboard" any better?

[18:59:01.0000] <MikeSmith>
you don't really "activate" an event, anyway

[19:00:00.0000] <Philip`>
"Allow keyboard activation of drag-and-drop events"

[19:02:00.0000] <MikeSmith>
thanks

[19:07:00.0000] <MikeSmith>
Hixie: when you have a chance, please flip the spec back to ED

[19:08:00.0000] <Hixie>
k

[19:08:01.0000] <MikeSmith>
thanks

[19:17:00.0000] <Hixie>
remind me if i forget to do it in the next few hours

[19:30:00.0000] <MikeSmith>
man, David Carlisle is a model of how inter-WG problem-resolving should be conducted 

[19:30:01.0000] <MikeSmith>
despite me talking trash about my frustrations with MathML stuff, he always remains gracious

[19:31:00.0000] <MikeSmith>
and focuses on trying to figure out where the problems are and how we can try to fix them

[19:32:00.0000] <nessy>
it can be hard - keeping feelings under control is always hard - but it's definitely more productive ;)

[19:32:01.0000] <MikeSmith>
yep

[19:37:00.0000] <MikeSmith>
Hixie: if/when you might find some real-time discussion with David to be useful, he says feel free to e-mail or skype him and he can join the channel for the discussion

[19:39:00.0000] <MikeSmith>
hmm, I find a David Carlisle whose Skype handle is "abadassmother"

[19:39:01.0000] <MikeSmith>
somehow I think that's not likely him

[19:40:00.0000] <AryehGregor>
See, that's the nice thing about having a distinctive name.  No confusion necessary.

[19:40:01.0000] <AryehGregor>
(I assume you could tell us the opposite end of that story.)

[19:41:00.0000] <MikeSmith>
heh

[19:41:01.0000] <AryehGregor>
No other Mike Smith had the foresight to trademark their name, though.  That's how I was able to find you on Twitter or something one time.

[19:42:00.0000] <MikeSmith>
I'm trademarking the fact that I have a trademark in my name

[19:42:01.0000] <MikeSmith>
/me looks for Aryeh on twitter

[19:42:02.0000] <AryehGregor>
Good luck.

[19:43:00.0000] <AryehGregor>
/me uses no social networking site or anything vaguely similar.

[19:43:01.0000] <AryehGregor>
Well, I guess I technically use Google Buzz, but only because I wasn't given a choice.

[19:43:02.0000] <MikeSmith>
heh

[19:43:03.0000] <AryehGregor>
(I don't post to it, in my defense)

[19:43:04.0000] <MikeSmith>
that's the next step in their plan

[19:43:05.0000] <AryehGregor>
(Nor does anyone else in the universe, though, except TabAtkins)

[19:44:00.0000] <AryehGregor>
(and possibly a single-digit number of other Google employees)

[19:44:01.0000] <MikeSmith>
I hear Buzz2 will actually generate posts without your consent, and send them out to all your contacts automatically, with them all Cc'ed so that they know who each other are (of course)

[19:45:00.0000] <MikeSmith>
so that'll be a nice feature to have

[19:45:01.0000] <AryehGregor>
Even better than what I've heard about Facebook!

[19:45:02.0000] <MikeSmith>
they will call it a feature for "lazy bloggers who have anxiety about not having posted for 3 months"

[19:46:00.0000] <MikeSmith>
Facebook is the market leader in privacy-infringing special features

[19:46:01.0000] <MikeSmith>
they are setting the de facto standards

[19:46:02.0000] <AryehGregor>
Yes, but Google is in a strong position to overtake them if they put their mind to it.

[19:46:03.0000] <MikeSmith>
yes

[19:46:04.0000] <AryehGregor>
I mean, where was Facebook when Google was out collecting people's unsecured Wi-Fi data without telling them?

[19:47:00.0000] <MikeSmith>
they just need to try harder, find more novel ways to expose people's private info

[19:47:01.0000] <MikeSmith>
AryehGregor: good point

[19:47:02.0000] <MikeSmith>
Google should blog about that

[19:47:03.0000] <MikeSmith>
"Nobody can infringe on privacy at the scale we can!"

[19:48:00.0000] <AryehGregor>
They do have an effective opt-out program, to be fair.  http://www.theonion.com/video/google-opt-out-feature-lets-users-protect-privacy,14358/

[19:49:00.0000] <AryehGregor>
The video doesn't work for me.  :(

[20:01:00.0000] <MikeSmith>
"The Opt-Out Village"

[20:02:00.0000] <MikeSmith>
beautiful

[20:02:01.0000] <MikeSmith>
"Christian Groups: Biblical Armageddon Must Be Taught Alongside Global Warming" is nice too

[20:06:00.0000] <othermaciej>
MikeSmith: that's good to hear, about David Carlisle

[20:27:00.0000] <MikeSmith>
othermaciej: yeah, kind of leading by example

[01:00:00.0000] <MikeSmith>
I wonder if Chris Grigg in the Audio XG is the same as the one who was a member of Negativland

[01:01:00.0000] <MikeSmith>
...and who also maybe the same who wrote the music for Maniac Mansion

[01:12:00.0000] <MikeSmith>
is there any way in MediaWiki to add an ID to a paragraph or term/span of text or whatever?

[01:14:00.0000] <annevk>
http://www.w3.org/QA/2010/06/an_update_on_css_21.html

[01:18:00.0000] <MikeSmith>
somebody posted a comment to that blog entry but it has not appeared yet for some reason

[01:18:01.0000] <MikeSmith>
though I got an feed notification about it

[01:44:00.0000] <MikeSmith>
https://bugs.webkit.org/show_bug.cgi?id=41363 is interesting

[01:44:01.0000] <MikeSmith>
what to do if author puts "body  { text-rendering: optimizeLegibility; }"

[01:45:00.0000] <MikeSmith>
paul_irish: btw, I don't think putting fragment IDs onto shortened URLs works the way you might have intended :)

[01:46:00.0000] <MikeSmith>
e.g., http://webk.it/6136#c3 

[01:46:01.0000] <Hixie>
/me mumbles something about it being bad to give authors control over how UAs optimise

[01:46:02.0000] <MikeSmith>
what to do if author puts "body  { text-rendering: optimizeEntireBrowsingExperience; }"

[01:47:00.0000] <Peter`>
MikeSmith: Chrome forwards me to https://bugs.webkit.org/show_bug.cgi?id=6136#c3, also scrolls the page to the intended position

[01:47:01.0000] <MikeSmith>
Peter`: well lah dee dah

[01:47:02.0000] <MikeSmith>
good for you

[01:47:03.0000] <MikeSmith>
seriously?

[01:47:04.0000] <Peter`>
Uh..?

[01:47:05.0000] <MikeSmith>
does Chrome cook breakfast for people now too?

[01:48:00.0000] <MikeSmith>
it seems like chrome should not do that

[01:48:01.0000] <hsivonen>
I'm concerned of a situation where a perf test suite turns on all these typographic nice things but doesn't test that they are taking effect and then Brand X browser scores better than Gecko simply by not implementing any of that stuff

[01:49:00.0000] <Peter`>
Whatever. I'm just saying it works on Chrome, I'm unaware of what the spec says on it, or of what other browsers do.

[01:49:01.0000] <Lachy>
MikeSmith, I just tried to update my answer on the issue 88 objetion poll, but it doesn't seem to be accepting my changes. It's still showing my original answers.

[01:49:02.0000] <MikeSmith>
Lachy: I'll check now

[01:50:00.0000] <hsivonen>
Lachy: do you see the new answer in another browser?

[01:50:01.0000] <MikeSmith>
Peter`: sorry, I was just trying to be funny, but I guess it sounded like I was trying to be a dick instead.. 

[01:51:00.0000] <Lachy>
ah, yeah, it works now. Must have been just a cache issue.

[01:51:01.0000] <Peter`>
Ok, I misinterpreted your message in that case, thank you for clearing it up

[01:52:00.0000] <MikeSmith>
Peter`: hmm, Minefield and Opera also do as Chrome does for http://webk.it/6136#c3 .. so egg on my face

[01:52:01.0000] <MikeSmith>
Safari does not, though

[01:52:02.0000] <MikeSmith>
or at least Webkit does not

[01:52:03.0000] <MikeSmith>
nightly

[01:52:04.0000] <Peter`>
IE doesn't do it either

[01:52:05.0000] <MikeSmith>
interesting

[01:53:00.0000] <MikeSmith>
I wonder what the specs say

[01:53:01.0000] <Hixie>
20 hours left on the polls, Philip`, zcorpan, MikeSmith, and anyone else i missed :-)

[01:54:00.0000] <Hixie>
foolip: ^

[01:54:01.0000] <MikeSmith>
I payed at the office.

[01:55:00.0000] <MikeSmith>
Lachy: OK. I have noticed some other unusualities lately that seem to indicate borked caching somewhere

[01:55:01.0000] <Peter`>
MikeSmith: http://trac.tools.ietf.org/wg/httpbis/trac/ticket/43 seems to be related

[01:55:02.0000] <annevk>
MikeSmith, fragids ought to be forwarded

[01:55:03.0000] <MikeSmith>
Peter`: thanks

[01:55:04.0000] <MikeSmith>
/me reads

[01:55:05.0000] <annevk>
MikeSmith, though client-side

[01:55:06.0000] <Peter`>
while it more specifically relates to overriding fragment ids

[01:56:00.0000] <MikeSmith>
annevk: sez who?

[01:56:01.0000] <annevk>
MikeSmith, it's like <image> needs to be parsed as <img>; it wasn't always defined, but it's certainly true :p

[01:56:02.0000] <annevk>
MikeSmith, I believe the HTTP WG might catch up though

[01:57:00.0000] <annevk>
MikeSmith, actually, I think it has always been defined for the simple case

[01:59:00.0000] <annevk>
MikeSmith, maybe not, "fragment" is mentioned only once in 2616 (haven't looked at errata though)

[02:00:00.0000] <Peter`>
http://www.w3.org/Protocols/HTTP/Fragment/draft-bos-http-redirect-00.txt

[02:00:01.0000] <Peter`>
That seems to be spot-on

[02:01:00.0000] <annevk>
no official standing whatsoever though

[02:01:01.0000] <annevk>
so I guess it's like <image> after all

[02:02:00.0000] <MikeSmith>
good times

[02:05:00.0000] <hsivonen>
Hixie: I'm debugging an interesting crash with document.documentElement.innerHTML = "<math></html>"

[02:06:00.0000] <hsivonen>
Hixie: I think the spec doesn't take this into account, and my ad hoc fixes to the spec don't take it into account, either

[02:08:00.0000] <foolip>
ddfdf

[02:08:01.0000] <annevk>
one more and you have a color

[02:08:02.0000] <foolip>
hehe

[02:10:00.0000] <foolip>
just trying to shut things down, this computer i moving to Gothenburg

[02:11:00.0000] <hsivonen>
Hixie: I assume we wouldn't want to switch out of 'in foreign' and into 'after body' with <math> on the top of the stack

[02:12:00.0000] <annevk>
the spec doesn't crash though, does it?

[02:13:00.0000] <zcorpan_>
https://connect.microsoft.com/IE/feedback/ViewFeedback.aspx?FeedbackID=559531 - IE9 fails html5lib html parser tests - Closed 

[02:13:01.0000] <zcorpan_>
as Fixed

[02:14:00.0000] <hsivonen>
Hixie: OTOH, if we don't (or only change the secondary mode to 'after body'), the end tag handling in 'in foreign content' ends up popping <html> off the stack

[02:15:00.0000] <hsivonen>
which should never happen and, therefore, crashes in EOF handler

[02:17:00.0000] <jgraham>
hsivonen: Would it be possible for you to document your ad-hoc fixes to the spec in the relevant bug reports?

[02:17:01.0000] <hsivonen>
jgraham: I suppose it would be possible

[02:18:00.0000] <jgraham>
hsivonen: Prose is easier to read than java :)

[02:18:01.0000] <zcorpan_>
seems they still fail many html5lib tests

[02:18:02.0000] <annevk>
hsivonen, so it doesn't crash for <math></head> or <math></body> or <math></td> ?

[02:19:00.0000] <hsivonen>
annevk: let's check

[02:19:01.0000] <hsivonen>
annevk: no crash

[02:20:00.0000] <hsivonen>
annevk: (with html as the context node)

[02:20:01.0000] <hsivonen>
so it seems wrong to change the primary mode to 'after body' without popping the foreign stuff off the stack

[02:21:00.0000] <hsivonen>
I think this leaves three options:

[02:21:01.0000] <hsivonen>
 1) Ignoring </html> (and maybe </body>) in 'in foreign content'

[02:21:02.0000] <hsivonen>
 2) popping all foreign stuff off the stack when transitioning to 'after body'

[02:22:00.0000] <hsivonen>
3) making an exception in the end tag handling in 'in foreign' so that the first item on the stack never gets popped no matter what

[02:22:01.0000] <jgraham>
hsivonen: It feels like </html> in innderHTML should be ignored

[02:22:02.0000] <annevk>
is this specific to in foreign content or would a document <math></html> do it too?

[02:22:03.0000] <jgraham>
Although I don't know what happens in other cases

[02:22:04.0000] <MikeSmith>
zcorpan_: are there bugs open for the other tests they fail?

[02:23:00.0000] <MikeSmith>
zcorpan_: or is their any distinguishable pattern to the bugs they fail?

[02:23:01.0000] <zcorpan_>
ie breaks out of <mtext>

[02:23:02.0000] <hsivonen>
annevk: <math></html> crashes even without fragment mode

[02:23:03.0000] <MikeSmith>
zcorpan_: hmm, that especially ain't good

[02:23:04.0000] <zcorpan_>
MikeSmith: no (afaik) and no (they just fail lots of tests)

[02:23:05.0000] <MikeSmith>
k

[02:24:00.0000] <hsivonen>
jgraham: so ignoring </html> in the fragment case is not the right fix

[02:24:01.0000] <zcorpan_>
oh they don't support mathml parsing at all

[02:24:02.0000] <zcorpan_>
and i was using <svg>, heh

[02:25:00.0000] <zcorpan_>
hmm they don't do case fixup for <foreigncontent>

[02:25:01.0000] <jgraham>
hsivonen: Well, it might be part of the right fix :)

[02:26:00.0000] <zcorpan_>
but they break out of <foreignContent>

[02:26:01.0000] <annevk>
hsivonen, and it's due to the spec? still not quite clear to me where it goes wrong; but I guess I should do something else

[02:26:02.0000] <jgraham>
annevk: The spec for end tag handling in foreign content mode is broken on its own

[02:26:03.0000] <hsivonen>
annevk: it is possible that my ad hoc fixes to spec crashes introduced this crash

[02:27:00.0000] <jgraham>
annevk: hsivonen has added some fixes

[02:27:01.0000] <hsivonen>
annevk: Hixie's fixes to SVG </a> and </font> handling caused the spec to crash

[02:28:00.0000] <jgraham>
hsivonen: This case crashes my local copy of html5lib which I think implements the spec as written

[02:28:01.0000] <zcorpan_>
at least they parse <foobar/> outside foreign content correctly now :)

[02:28:02.0000] <abarth>
are they trying to implement the HTML5 parsing algorithm?

[02:29:00.0000] <jgraham>
abarth: They seem to be trying to implement foreign content in HTML

[02:29:01.0000] <annevk>
I think the rest of it too

[02:29:02.0000] <zcorpan_>
abarth: half-assed, yes. they have a proper tree now (they said they implemented AAA) and svg in html

[02:29:03.0000] <jgraham>
it's not clear whether they are implementing the rest of the algorithm or trying to cherry pick

[02:30:00.0000] <hsivonen>
jgraham: I'm thinking of either ignoring </html> in 'in foreign content' or modifying step 3 under "An end tag, if the current node is not an element in the HTML namespace." stop right before it pops the last node off the stack

[02:30:01.0000] <zcorpan_>
they don't implement foster parenting

[02:30:02.0000] <abarth>
cherry picking seems silly

[02:30:03.0000] <jgraham>
abarth: Indeed

[02:30:04.0000] <abarth>
the main benefit is from going "all in" (to borrow a phrase from my favorite blog)

[02:31:00.0000] <hsivonen>
zcorpan_: interesting. I have thought that implementing parts of the parsing algorithm in an existing parser would be more painful than implementing the whole thing from scratch

[02:31:01.0000] <jgraham>
hsivonen: Have you made other changes to that algorithm?

[02:32:00.0000] <hsivonen>
jgraham: I fixed http://www.w3.org/Bugs/Public/show_bug.cgi?id=9582 , http://www.w3.org/Bugs/Public/show_bug.cgi?id=9581 and http://www.w3.org/Bugs/Public/show_bug.cgi?id=9580

[02:32:01.0000] <zcorpan_>
hsivonen: maybe their goal for ie9 was just proper tree and svg support

[02:33:00.0000] <hsivonen>
jgraham: I'll look up my code changes and try to figure out how to map those back to English

[02:33:01.0000] <hsivonen>
jgraham: the diff is http://hg.mozilla.org/projects/htmlparser/rev/b08b9993cd5d

[02:34:00.0000] <hsivonen>
jgraham: don't pay attention to the "changes" inside switch (mode)

[02:35:00.0000] <jgraham>
hsivonen: thanks

[02:40:00.0000] <hsivonen>
jgraham: eltPosForeign represents /node/ from the spec (by index)

[02:40:01.0000] <hsivonen>
jgraham: the three lines at http://hg.mozilla.org/projects/htmlparser/rev/b08b9993cd5d#l1.66 are one addition

[02:41:00.0000] <hsivonen>
fixing http://www.w3.org/Bugs/Public/show_bug.cgi?id=9580

[02:42:00.0000] <hsivonen>
jgraham: http://hg.mozilla.org/projects/htmlparser/rev/b08b9993cd5d#l1.64 fixes http://www.w3.org/Bugs/Public/show_bug.cgi?id=9582

[02:42:01.0000] <hsivonen>
jgraham: currentPtr is the index of the top aka. bottom of the stack

[02:44:00.0000] <hsivonen>
jgraham: it it is smaller than the just-decremented eltPosForeign, the secondary insertion mode has popped so much stuff that it would be bad to continue per spec before /node/ is in range for the current stack again

[02:45:00.0000] <hsivonen>
jgraham: the left-hand part of the |if| condition on line http://hg.mozilla.org/projects/htmlparser/rev/b08b9993cd5d#l1.57 fixes http://www.w3.org/Bugs/Public/show_bug.cgi?id=9581

[02:45:01.0000] <hsivonen>
jgraham: that's all, I think

[02:46:00.0000] <jgraham>
hsivonen: OK

[02:50:00.0000] <hsivonen>
krijnh: logs are broken

[02:55:00.0000] <hsivonen>
zcorpan_: they aren't supposed to case-correct <foreigncontent> but <foreignobject>

[02:57:00.0000] <MikeSmith>
zcorpan_: what's AAA?

[02:58:00.0000] <annevk>
adoption-agency-algorithm

[02:58:01.0000] <zcorpan_>
hsivonen: oops

[02:58:02.0000] <MikeSmith>
ah

[02:58:03.0000] <hsivonen>
jgraham: currently, </body> in 'in foreign' pops more stuff than </body> with <div> but no foreign content on the stack

[02:59:00.0000] <hsivonen>
jgraham: I'd expect this to be a bug, too

[02:59:01.0000] <zcorpan_>
hsivonen: still, they didn't turn <fooBarBaz> to lowercase

[02:59:02.0000] <hsivonen>
what if we ignored </body> and </html> in 'in foreign content'?

[02:59:03.0000] <hsivonen>
would that be too blunt a solution?

[03:01:00.0000] <zcorpan_>
/me wouldn't mind always ignoring </body> and </html>

[03:02:00.0000] <jgraham>
wfm I think

[03:04:00.0000] <hsivonen>
hmm. I think the new end tag handling is probably more aggressive that Hixie intended

[03:04:01.0000] <hsivonen>
consider <dl><svg><foreignContent><div><math></dl><foo>

[03:04:02.0000] <hsivonen>
</dl> pops stuff until <dl> has been popped

[03:04:03.0000] <hsivonen>
in particular, it doesn't stop at the <div>

[03:05:00.0000] <hsivonen>
aaagh.

[03:05:01.0000] <hsivonen>
zcorpan_'s foreignContent is messing with my head

[03:05:02.0000] <zcorpan_>
heh

[03:06:00.0000] <hsivonen>
anyway, <foo> becomes the next sibling of <dl>

[03:07:00.0000] <zcorpan_>
why would it stop at the div? that doesn't make any sense

[03:08:00.0000] <hsivonen>
in the <dl><svg><foreignObject><div><math></dl><foo> case, both <foreignObject> and <div> are scoping

[03:08:01.0000] <hsivonen>
normally, </dl> doesn't match a <dl> if there's a scoping element in between

[03:08:02.0000] <jgraham>
I agree that it should stop at scopiung elements

[03:09:00.0000] <jgraham>
*scoping

[03:09:01.0000] <zcorpan_>
if there's a scoping element in between, shouldn't the end tag just be treated as if the <dl> start tag wasn't there?

[03:10:00.0000] <hsivonen>
zcorpan_: possibly, but that's not what the spec does

[03:10:01.0000] <hsivonen>
zcorpan_: which looks like a spec bug to me

[03:10:02.0000] <hsivonen>
it's a bug of such magnitude that I'm not fully comfortable with improvising as solution

[03:11:00.0000] <hsivonen>
but I should land something to get rid of the crasher :-(

[03:11:01.0000] <zcorpan_>
<dl><div></dl>x

[03:11:02.0000] <zcorpan_>
closes the dl

[03:12:00.0000] <hsivonen>
zcorpan_: good point

[03:12:01.0000] <hsivonen>
<dl><div></body>x 

[03:13:00.0000] <hsivonen>
doesn't close dl

[03:13:01.0000] <hsivonen>
but if </body> happens in 'in foreign' it pops stuff all the way to <body>

[03:14:00.0000] <hsivonen>
well, I guess ignoring </body> and </html> would good enough to address the worst cases

[03:16:00.0000] <MikeSmith>
.

[03:16:01.0000] <MikeSmith>
(sorry, fat-finger)

[03:17:00.0000] <hsivonen>
whoa. <dl><svg><foreignObject><div><math></body><foo> is worse than I thought

[03:17:01.0000] <hsivonen>
<foo> becomes the next sibling on <body>!

[03:19:00.0000] <MikeSmith>
sweet

[03:20:00.0000] <MikeSmith>
huh?

[03:20:01.0000] <MikeSmith>
isn't already the next sibling of body in your example?

[03:21:00.0000] <MikeSmith>
oh

[03:21:01.0000] <MikeSmith>
/me puts down his meth pipe

[03:21:02.0000] <MikeSmith>
(ignore me)

[03:26:00.0000] <zcorpan_>
awesome that websrt has <comment> and ie also has <comment> in html

[03:28:00.0000] <jgraham>
hsivonen: I think the end tag handling in In Foreign Content needs to be totally rewritten...

[03:29:00.0000] <hsivonen>
jgraham: :-(

[03:29:01.0000] <hsivonen>
jgraham: how?

[03:30:00.0000] <hsivonen>
is the whatwg wiki broken or my browser? some operations that are supposed to redirect me don't

[03:30:01.0000] <hsivonen>
and I'm left with a blank page

[03:30:02.0000] <jgraham>
hsivonen: I don't know

[03:30:03.0000] <jgraham>
hsivonen: That's the problem

[03:31:00.0000] <jgraham>
But it feels like we're trying to patch around the wrong approach

[03:31:01.0000] <jgraham>
I could be wrong of course

[03:33:00.0000] <hsivonen>
jgraham: it's also possible that my fix for http://www.w3.org/Bugs/Public/show_bug.cgi?id=9582 is wrong

[03:33:01.0000] <annevk>
http://html5.org/tools/web-apps-tracker now remembers the show editorial flag

[03:34:00.0000] <hsivonen>
jgraham: and the right fix would be setting the insertion mode to the secondary mode and returning instead of continuing the loop

[03:35:00.0000] <annevk>
only browser it goes wrong in is Firefox

[03:35:01.0000] <annevk>
it does not seem to reflow the page after a class name change to the body element

[03:36:00.0000] <zcorpan_>
annevk: nice

[03:41:00.0000] <roc>
looks to me like the body class is not being set in firefox

[03:42:00.0000] <zcorpan_>
http://forums.whatwg.org/viewtopic.php?t=4334 - how to produce HTML5 printable pdf

[03:43:00.0000] <annevk>
roc, weird, everyone gets the same code

[03:43:01.0000] <annevk>
roc, and invoking updateEditorial() after page load does work...

[03:44:00.0000] <annevk>
roc, maybe a bug due to parser changes?

[03:44:01.0000] <roc>
checking

[03:45:00.0000] <roc>
so I think it's to do with form history restoration

[03:46:00.0000] <hsivonen>
does it involve innerHTML?

[03:46:01.0000] <roc>
no

[03:46:02.0000] <hsivonen>
good

[03:46:03.0000] <roc>
if you load the page the first time, the checkbox is checked. You uncheck the checkbox and reload. We remember that the checkbox was unchecked, so in the reloaded page, we uncheck it

[03:46:04.0000] <roc>
then "if(showEdits() && readCookie("editorial") == "editorial") {" is false because showEdits returned false because the editorial checkbox is not checked

[03:47:00.0000] <roc>
so you don't call updateEditorial

[03:48:00.0000] <jgraham>
hsivonen: So am I reading your fixes right? If I am they boil down to:

[03:48:01.0000] <jgraham>
In step 4 abort if node is the root node

[03:49:00.0000] <jgraham>
After step 5, if node is no longer on the stack of open elements, set node to the bottommost element on the stack of open elements

[03:49:01.0000] <annevk>
roc, ah, you remember the value of the checkbox, but not the class name on the body

[03:49:02.0000] <roc>
yeah

[03:49:03.0000] <annevk>
hmm

[03:49:04.0000] <roc>
why would we remember the class name on the body?

[03:51:00.0000] <annevk>
mkay, removed the showEdits() check there

[03:51:01.0000] <annevk>
went wrong in Chrome too with history navigation

[03:51:02.0000] <annevk>
Opera works fine

[03:52:00.0000] <roc>
I'm not really sure what the status of form history restoration is

[03:52:01.0000] <roc>
spec-wise

[03:58:00.0000] <annevk>
I think Hixie might have specced it out, but I'm not familiar with the details

[03:59:00.0000] <roc>
annevk: I'm confused about one thing in your page

[03:59:01.0000] <roc>
   function updateEditorial() {

[03:59:02.0000] <roc>
     var editorial = showEdits() ? "" : "editorial"

[03:59:03.0000] <roc>
shouldn't that be the other way around?

[04:00:00.0000] <annevk>
the naming is all confusing :/

[04:00:01.0000] <annevk>
but editorial means that stuff will be hidden

[04:00:02.0000] <annevk>
well "editorial"

[04:01:00.0000] <hsivonen>
yeah, the YouTube blog *so* makes me want to support Content Protection to enable them to serve "This rental is currently unavailable in your country."

[04:01:01.0000] <annevk>
this code needs a few more passes before it's logical again

[04:01:02.0000] <annevk>
it also has createCookie and readCookie for non-cookie code

[04:02:00.0000] <roc>
hsivonen: heh, we got that too

[04:02:01.0000] <roc>
annevk, OK, I'm going to stop trying to make it work then :-)

[04:02:02.0000] <annevk>
sorry, the code is logical, it's just not logically named

[04:04:00.0000] <hsivonen>
umm. how do sites that embed the YouTube swf player ensure that the swf doesn't get at private information on the embedding page?

[04:04:01.0000] <annevk>
hsivonen, yeah, blargh

[04:04:02.0000] <annevk>
content protection is teh evil

[04:11:00.0000] <zcorpan_>
hsivonen: they trust youtube?

[04:12:00.0000] <annevk>
most of the other issues they mention will be addressed in due course though

[04:13:00.0000] <corey__>
is this the proper channel for html5 discussions?

[04:14:00.0000] <zcorpan_>
yes

[04:14:01.0000] <roc>
/me hacks on fullscreen to make that one go away

[04:15:00.0000] <hsivonen>
zcorpan_: as far as I can tell, the embedding model is already "they trust YouTube", yes

[04:15:01.0000] <zcorpan_>
roc: api? for video only or anything?

[04:15:02.0000] <annevk>
roc, someone hacking on a spec too?

[04:15:03.0000] <roc>
yep

[04:15:04.0000] <roc>
I already proposed it in the whatwg list a while ago

[04:15:05.0000] <roc>
zcorpan_: everything

[04:16:00.0000] <hsivonen>
zcorpan_: so doing <script src="http://www.youtube.com/do/whatever/you/do.js"></script> should work fine

[04:16:01.0000] <roc>
zcorpan_, annevk: https://wiki.mozilla.org/Gecko:FullScreenAPI#Gecko_Implementation

[04:16:02.0000] <roc>
erp

[04:16:03.0000] <roc>
ignore the implementation bit

[04:17:00.0000] <roc>
that's just a slight refinement of what was discussed here: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2010-January/024872.html

[04:19:00.0000] <corey__>
are websockets a dependable technology?

[04:20:00.0000] <annevk>
corey__, what does that mean?

[04:20:01.0000] <annevk>
roc, looks pretty good, though it would be nice if we could avoid putting more stuff on Window

[04:20:02.0000] <corey__>
I'm asking if it's doing to be adopted by major browsers in the near future. seems to be only on chrome at the moment.

[04:20:03.0000] <roc>
yeah

[04:20:04.0000] <zcorpan_>
maybe fullscreen stuff should be on window.screen :)

[04:21:00.0000] <roc>
no it shouldn't

[04:21:01.0000] <roc>
you don't make the screen fullscreen

[04:21:02.0000] <roc>
I could go with "navigator is the new window"

[04:21:03.0000] <annevk>
corey__, other browsers are implementing it, yes

[04:21:04.0000] <roc>
corey__: it's on Firefox trunk

[04:21:05.0000] <roc>
so the next Firefox will have it

[04:22:00.0000] <corey__>
I was lead to believe Firefox 4 would have it?

[04:22:01.0000] <roc>
yeah

[04:22:02.0000] <roc>
annevk: thing is, Gecko already has a "fullscreen" attribute on Window

[04:22:03.0000] <roc>
er, fullScreen

[04:23:00.0000] <hsivonen>
roc: why is there already fullScreen on window?

[04:23:01.0000] <hsivonen>
roc: XUL stuff leaking to the Web?

[04:23:02.0000] <roc>
maybe, I don't know without doing the archaeology

[04:24:00.0000] <annevk>
seems to be unique to Gecko

[04:24:01.0000] <annevk>
well, at least Opera/Chrome don't have it

[04:25:00.0000] <roc>
I could add the new APIs elsewhere, and even a new 'fullScreen' attribute

[04:26:00.0000] <roc>
it's just that Window is the most logical place for it, apart from the global namespace pollution problem

[04:37:00.0000] <othermaciej>
roc: it seems odd that requestFullScreen is on Window but operates on an Element, instead of just being an Element method

[04:37:01.0000] <roc>
you might not have an element

[04:37:02.0000] <roc>
we could have two no-arg methods, one on Element and one on Window

[04:38:00.0000] <othermaciej>
would the one on Window do anything different than applying full-screen to the root element?

[04:38:01.0000] <hsivonen>
roc: what about non-arg methods on Element and Document?

[04:38:02.0000] <hsivonen>
roc: to avoid window pollution

[04:38:03.0000] <roc>
although there's the variation for disabling keys, so that's two methods on Element and two on Window

[04:38:04.0000] <annevk>
is it likely we would have other options than keys in the future?

[04:39:00.0000] <othermaciej>
adding to Window pollutes the global namespace, which is bad

[04:39:01.0000] <roc>
annevk: I don't know

[04:39:02.0000] <roc>
yes, I know

[04:39:03.0000] <hsivonen>
roc: has Document been considered?

[04:39:04.0000] <roc>
othermaciej: yes, making the root element fullscreen has side effects ... the default :full-screen style rule would kick in 

[04:39:05.0000] <annevk>
othermaciej, root can be positioned within the initial containing block

[04:40:00.0000] <roc>
hsivonen: I'm considering it right now :-)

[04:40:01.0000] <othermaciej>
I think WebKit's enterFullscreen() and exitFullscreen() extensions on HTMLVideoElement could just be applied to all elements

[04:40:02.0000] <MikeSmith>
corey__: https://bugzilla.mozilla.org/show_bug.cgi?id=472529

[04:41:00.0000] <zcorpan_>
/me wonders what happens if you fullsceen an element that's display:none

[04:41:01.0000] <roc>
othermaciej: I think they behave differently from what I'm working on

[04:41:02.0000] <MikeSmith>
corey__: I think most of the code has already landed in gecko trunk

[04:41:03.0000] <othermaciej>
also it doesn't make sense to me for the shorter simpler name to be the one with keyboard disabled

[04:41:04.0000] <roc>
it's not

[04:41:05.0000] <othermaciej>
er

[04:41:06.0000] <othermaciej>
with keyboard enabled I mean

[04:41:07.0000] <corey__>
I'm using a recent minefield build and not seeing websockets enabled =/

[04:41:08.0000] <hsivonen>
hmm. I guess the whatwg wiki isn't broken but instead my snapshot of Gecko doesn't support HTTP redirects

[04:41:09.0000] <othermaciej>
having keyboard enabled is a big security risk, so I am not sure why it should be allowed at all

[04:42:00.0000] <zcorpan_>
corey__: works for me

[04:42:01.0000] <roc>
WriteRoom, stuff like that

[04:42:02.0000] <roc>
lots of games need the keybaord

[04:42:03.0000] <othermaciej>
have you thought of a sufficient way to mitigate the security risk of a full-screen OS simulacrum?

[04:42:04.0000] <corey__>
zcorpan_, your version?

[04:42:05.0000] <hsivonen>
othermaciej: shouldn't full-screen first-person WebGL shooters support movement by letter keys?

[04:42:06.0000] <annevk>
it does make sense to reverse the logic I think

[04:42:07.0000] <zcorpan_>
corey__: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:2.0b2pre) Gecko/20100629 Minefield/4.0b2pre

[04:42:08.0000] <annevk>
requestFullScreenEnableKeyboard or some such

[04:42:09.0000] <roc>
hsivonen: well, one thing I'm thinking of is suppressing keypress, not keydown or keyup

[04:42:10.0000] <othermaciej>
Flash deals with this by (a) having an ugly text warning when you first enter full-screen and (b) only allowing a very limited set of keys in fullscreen mode

[04:43:00.0000] <roc>
annevk: ok

[04:43:01.0000] <MikeSmith>
kennyluck: http://blog.jclark.com/2010/01/xml-namespaces.html

[04:43:02.0000] <othermaciej>
hsivonen: that's a valid use case, but how do you allow it without creating the security risk of an OS simulacrum that steals your password?

[04:44:00.0000] <roc>
passive opt-in UI

[04:44:01.0000] <MikeSmith>
kennyluck: http://www.xml.com/pub/a/2005/04/13/namespace-uris.html is worth reading too (from 2005 and that article was not the first to point out the problems but it's a good summary)

[04:44:02.0000] <roc>
othermaciej: so if I understand correctly, webkitEnterFullScreen creates a new window showing just the video, right?

[04:44:03.0000] <zcorpan_>
speaking of FPS games, i wonder how to solve the problem of the mouse reaching the edge of the window

[04:45:00.0000] <kennyluck>
thanks, reading.

[04:45:01.0000] <othermaciej>
roc: are you asking about the implementation, or logically what it does?

[04:45:02.0000] <MikeSmith>
kennyluck: "XML Namespaces Don't Need URIs" 

[04:45:03.0000] <roc>
the latter

[04:45:04.0000] <hsivonen>
/me notes that e.g. OS X "Application foo requires you to type your password" dialog could be faked without fullscreen

[04:45:05.0000] <othermaciej>
roc: there is no second <video> element

[04:45:06.0000] <hsivonen>
but Ubuntu's can't

[04:45:07.0000] <othermaciej>
roc: it displays the existing <video> element fullscreen

[04:45:08.0000] <othermaciej>
it happens to do this by creating another fullscreen window to render to, but that's an implementation detail

[04:46:00.0000] <corey__>
MikeSmith, every xml namespace has a uri

[04:46:01.0000] <othermaciej>
if keyboard access has to be granted special permission, and regular fullscreen doesn't, then it seems like poor form to have the simpler method name be the one that enables the keyboard

[04:46:02.0000] <MikeSmith>
corey__: yeah, too bad

[04:46:03.0000] <roc>
is there documentation for webkitEnterFullScreen? http://developer.apple.com/safari/library/documentation/appleapplications/reference/WebKitDOMRef/HTMLVideoElement_idl/Classes/HTMLVideoElement/index.html is a bit sparse

[04:47:00.0000] <othermaciej>
anyway, we would like to generalize enterFullscreen / exitFullscreen to all elements, and we were thinking of writing a proposal, but we may just comment on yours, since it is close to what we want

[04:47:01.0000] <hsivonen>
I agree that the way for requesting the keyboard should look more special than the way of going fullscreen without keyboard

[04:47:02.0000] <roc>
othermaciej: you already did comment on mine

[04:47:03.0000] <roc>
well, Simon did

[04:47:04.0000] <othermaciej>
my only concern would be that you seem to have features that are not obviously necessary for the common use case

[04:47:05.0000] <roc>
we had a thread about almost this exact proposal on the whatwg list in January

[04:47:06.0000] <MikeSmith>
corey__: sorry I meant yeah, that's the beauty of it

[04:48:00.0000] <othermaciej>
well, one comment I'd give is that I think enter / exit are fine verbs, and while I have no huge problem with changing them, request/cancel don't seem obviously better

[04:48:01.0000] <Philip`>
zcorpan_: If it's fullscreen, that doesn't seem much of a problem - there's no browser chrome for the user to click on so there's probably no security risk in disabling their cursor, so you just disable their cursor and send mouse events with dx/dy instead of x/y

[04:48:02.0000] <roc>
the thing is that in my model, fullscreen transitions are asynchronous

[04:48:03.0000] <othermaciej>
having methods on Window instead of on elements seems pretty clearly worse (to me)

[04:48:04.0000] <roc>
when you call requestFullScreen, you don't necessarily get it immediately, or at all

[04:48:05.0000] <MikeSmith>
corey__: or, I meant No, not actually

[04:49:00.0000] <roc>
yes, I agree moving the methods to Element and Document make sense

[04:49:01.0000] <MikeSmith>
corey__: take your pick

[04:49:02.0000] <othermaciej>
"might not get it at all" seems weird

[04:49:03.0000] <roc>
it might be denied

[04:49:04.0000] <corey__>
MikeSmith, you are very indecisive

[04:49:05.0000] <othermaciej>
so the user gets prompted whether to allow fullscreen?

[04:49:06.0000] <othermaciej>
or are you imagining a blacklist?

[04:50:00.0000] <roc>
I'm imagining a passive prompt

[04:50:01.0000] <roc>
like we do for popups

[04:50:02.0000] <othermaciej>
I would say that for the common limited-keyboard case, no prompting is required

[04:50:03.0000] <roc>
I at least want to *allow* the UA to use a passive prompt

[04:50:04.0000] <kennyluck>
MikeSmith, I read the latter one before. And I think it's about the centralized vs decentralized debate.

[04:50:05.0000] <annevk>
if it's like popups, openFullScreen() / closeFullScreen() ?

[04:50:06.0000] <othermaciej>
for the full-keyboard case, I don't think a prompt would be an adequate security measure

[04:50:07.0000] <annevk>
/me is just brainstorming a bit; don't pay too much attention

[04:51:00.0000] <roc>
it's not like popups other than that

[04:51:01.0000] <roc>
othermaciej: I don't think it's a good idea to design an API that's incompatible with any kind of prompting

[04:51:02.0000] <othermaciej>
I am dubious about implementing a version with full keyboard support (as opposed to, say, limited to space, enter, arrow keys)

[04:51:03.0000] <jgraham>
Is this supposed to be bound to explicit user actions?

[04:51:04.0000] <roc>
it can be

[04:51:05.0000] <annevk>
maybe full keyboard should only be for this new "installed apps" concept, if that is going to happen

[04:52:00.0000] <roc>
othermaciej: what about devices that don't have a keyboard? I mean, if you only have a touch screen, disabling the keyboard isn't a big security win

[04:52:01.0000] <othermaciej>
that makes more sense to me, if there is such a concept

[04:52:02.0000] <othermaciej>
yes, I was just thinking today about how to defend against OS simulacrum on touchscreen-only devices

[04:52:03.0000] <jgraham>
If it's not then the ability to have a prompt is essential. If it is then it still seems like a useful feature

[04:53:00.0000] <Philip`>
othermaciej: If FPS-style games are a use case, then you need to support at least the WASD keys

[04:53:01.0000] <othermaciej>
I think one of the most common use cases for this will be video that wants to go full-screen with custom controls

[04:53:02.0000] <roc>
I don't want to constrain the security design space here by making the API assume synchronous transitions

[04:53:03.0000] <othermaciej>
for that case, prompting of any form would be obnoxious

[04:53:04.0000] <roc>
is that so crazy?

[04:54:00.0000] <roc>
especially given that I want fullscreen features that work when the user explicitly commands fullscreen (F11 in Firefox), so asynchronous transitions happen anyway

[04:54:01.0000] <othermaciej>
I'm not saying it should assume synchronous transitions, because in any case it would be nice for the transition to animate and for the content to remain live during that

[04:54:02.0000] <hsivonen>
Philip`: and the WASD-equivalent Dvorak, French, etc. locations

[04:54:03.0000] <roc>
othermaciej: I agree about the common use case

[04:55:00.0000] <othermaciej>
Philip`: it might be interesting to investigate what subset of keys Flash allows

[04:55:01.0000] <roc>
othermaciej: is webkitEnterFullScreen documented anywhere?

[04:55:02.0000] <Philip`>
http://www.adobe.com/devnet/flashplayer/articles/fplayer10_security_changes_03.html#head5

[04:55:03.0000] <Philip`>
Not sure if that's a comprehensive list

[04:56:00.0000] <othermaciej>
roc: I don't really know, but I can get people to write a rough description of it (and related events) or write it myself

[04:57:00.0000] <othermaciej>
our HTMLVideoElement.idl documents the entry points pretty clearly, if not the details of behavior

[04:57:01.0000] <othermaciej>
there are also some events

[04:57:02.0000] <roc>
what my proposal does is just make the window fullscreen and apply a default style to the fullscreen element, if any

[04:57:03.0000] <othermaciej>
http://trac.webkit.org/browser/trunk/WebCore/html/HTMLVideoElement.idl

[04:58:00.0000] <roc>
that means that cases like fullscreening a display:none element are perfectly well defined

[04:58:01.0000] <annevk>
what is the default style?

[04:58:02.0000] <roc>
it can also integrate nicely with the fullscreen features Firefox already has

[04:58:03.0000] <othermaciej>
do you do anything to make the full-screened element grow to the size of the Window?

[04:59:00.0000] <roc>
*:full-screen { position:fixed;  top:0;  left:0;  right:0;  bottom:0;  z-index:2147483647;  background:black; }

[04:59:01.0000] <annevk>
interesting

[05:00:00.0000] <zcorpan_>
so by default you get black text on black background?

[05:00:01.0000] <othermaciej>
doing it that way is clever

[05:00:02.0000] <roc>
zcorpan_: only if you decide to fullscreen a DIV full of default-style text ... I don't think that's the common case

[05:01:00.0000] <othermaciej>
what I would imagine is that the element's bounds become the viewport, so you can't use CSS to make stuff surrounding the element appear in view through weird CSS rules

[05:01:01.0000] <roc>
the common case is <video>, which generally wants black

[05:01:02.0000] <othermaciej>
that makes it more obvious how to do a nice transition

[05:01:03.0000] <hsivonen>
can CSS animations provide nice transitions from an in-flow box into a position:fixed box?

[05:01:04.0000] <roc>
no

[05:01:05.0000] <annevk>
you could add color:white

[05:02:00.0000] <roc>
sure, but I honestly don't think it would matter :-)

[05:02:01.0000] <roc>
most of the transitions you might want to do require OS integration to make a window zoom out in some manner

[05:02:02.0000] <Philip`>
annevk: If somebody wants to fullscreen a page of text, they probably want it to stay black-on-white, so you should minimise the number of CSS properties they need to override

[05:03:00.0000] <zcorpan_>
i agree that full screen video should be black, but for anything else it's less clear it should be black

[05:03:01.0000] <roc>
with my approach you can make the style change happen immediately, snap the window to the old bounds of the element, and zoom it out to the full screen

[05:05:00.0000] <roc>
that transition would look good in most cases

[05:05:01.0000] <roc>
almost all cases, I suspect

[05:05:02.0000] <Philip`>
*:full-screen { position: ...; } video:full-screen { background: black; } ?

[05:05:03.0000] <roc>
yes

[05:05:04.0000] <roc>
that would work

[05:05:05.0000] <othermaciej>
if you are using this with custom controls, you are likely to be using an ancestor of the <video>

[05:05:06.0000] <roc>
although we don't want UA style sheets to get *too* clever

[05:05:07.0000] <roc>
othermaciej: yep

[05:06:00.0000] <annevk>
the flexibility it provides for is pretty great

[05:07:00.0000] <roc>
I'll make some changes, thanks for the feedback

[05:07:01.0000] <othermaciej>
it's not clear to me that adding a black background in fullscreen is helpful on net 

[05:07:02.0000] <roc>
I dunno

[05:07:03.0000] <roc>
that's a minor detail anyway

[05:07:04.0000] <annevk>
yeah, me neither, but we can fiddle with those details later I suppose

[05:07:05.0000] <annevk>
once there's some developer feedback

[05:08:00.0000] <othermaciej>
I will say that our main point of feedback for enterFullscreen on <video> has been a desire for custom controls

[05:09:00.0000] <roc>
yes, that's essential

[05:09:01.0000] <othermaciej>
we have not had much direct interest in taking non-video content full-screen, but it makes sense to be fully general if you are general enough to allow custom JS-driven video controls

[05:10:00.0000] <othermaciej>
also one can imagine games and perhaps distraction-free apps as future use cases

[05:10:01.0000] <othermaciej>
or slideshows

[05:10:02.0000] <roc>
once you allow custom controls you have to deal with most of the security issues anyway

[05:11:00.0000] <othermaciej>
well, for keyboard&mouse devices at least, keyboard access is the main security risk

[05:12:00.0000] <othermaciej>
for touch devices, most of them have a status display at the top, so perhaps an always-on distinctive status display would be good enough

[05:12:01.0000] <roc>
so what would you do for touch devices?

[05:12:02.0000] <roc>
heh

[05:12:03.0000] <othermaciej>
that's the best I was able to think of

[05:12:04.0000] <roc>
ok

[05:12:05.0000] <roc>
thanks

[05:14:00.0000] <jgraham>
hsivonen: I am clearly still missing something. How do you cope with <table><caption><svg></caption>?

[05:17:00.0000] <jgraham>
It feels like it should break after the caption element is popped, but afaict it does not

[05:18:00.0000] <roc>
I don't really like going to requestFullScreen and requestFullScreenEnableKeys

[05:18:01.0000] <roc>
in the modern world it feels a bit wrong for the simple API to disable one particular input device that isn't even ubiquitous anymore

[05:19:00.0000] <othermaciej>
the simple API should be the one that likely does not require permission grants or pre-approval by the user

[05:19:01.0000] <roc>
in that case we should disable all input except the ESC key :-)

[05:19:02.0000] <othermaciej>
if I were designing it I'd just leave out the EnableKeys variant for v1, since Flash gets along fine without it

[05:20:00.0000] <othermaciej>
I think the Flash set of keys, and ESC hardcoded to always exit, is likely a pretty safe mode, at least for devices with a keyboard

[05:20:01.0000] <roc>
that's a big caveat

[05:21:00.0000] <roc>
but I'll go with this and see what people think

[05:22:00.0000] <roc>
oh, now I remember why I decided to have methods that take an element parameter

[05:23:00.0000] <roc>
it's because then I can have requestFullScreen(Element); cancelFullScreen() where cancelFullScreen takes no parameter

[05:23:01.0000] <roc>
so there's no confusion about what should happen if you do elem1.requestFullScreen(); elem2.cancelFullScreen();

[05:23:02.0000] <roc>
I suppose I can have cancelFullScreen on the Document only

[05:24:00.0000] <zcorpan_>
/me was just going to suggest that

[05:24:01.0000] <othermaciej>
elem1.requestFullScreen(); elem2.cancelFullScreen(); is something that is likely to arise only by mistake or in artificial test cases

[05:25:00.0000] <othermaciej>
or maybe I am too optimistic

[05:25:01.0000] <othermaciej>
anyway

[05:25:02.0000] <othermaciej>
I'm way overdue for some sleep, so good night folks

[05:26:00.0000] <roc>
heh, it must be 5:30am there

[05:26:01.0000] <jgraham>
hsivonen: It seems to work OK with the html5lib test cases if you break when node is no longer in the stack of open elements after step 5

[05:26:02.0000] <jgraham>
hsivonen: And for the <math></html> case

[05:27:00.0000] <jgraham>
hsivonen: I havne't really thought about the implications of that fix though

[05:46:00.0000] <karlcow>
"<video> tag does not currently meet all the needs of a site like YouTube:" - http://apiblog.youtube.com/2010/06/flash-and-html5-tag.html  implementer feedback

[06:01:00.0000] <annevk>
karlcow, you might be interested in the logs

[06:08:00.0000] <hsivonen>
karlcow: I believe that's "author feedback" from the usual point of view

[06:08:01.0000] <hsivonen>
usual on #whatwg that is

[06:16:00.0000] <Amadiro>
Good evening. Can someone recommend a small, handy webserver or proxy that supports websockets, both the new standard (with all the fancy encryption and hanshake stuff) as well as the older formats (as older versions of chrome implement them, for instance)? I'm currently using misultin (a tiny erlang webserver w/ websocket support) as reverse proxy, but it does not support the newer revisions of the websocket protocol.

[06:16:01.0000] <Amadiro>
Basically, what I want to do is connect my websocket application to my server, which uses a plain TCP connection. So the gateway has to accept the websocket, and relay the connection.

[06:25:00.0000] <karlcow>
hsivonen: thanks

[07:01:00.0000] <annevk>
why does https://wiki.mozilla.org/index.php?title=Gecko:FullScreenAPI&limit=500&action=history only show changes from today?

[07:03:00.0000] <jgraham>
annevk: Because it was created today?

[07:05:00.0000] <annevk>
ooh

[07:05:01.0000] <annevk>
I thought this was going on for some time

[07:33:00.0000] <kamathln>
<a href="http://appspot.com" href="http://appengine.google.com"> Appspot</a> 

[07:34:00.0000] <kamathln>
<a href="svn://svn.something.com/something"  href="http://something.com/svn"> Appspot</a> 

[07:36:00.0000] <kamathln>
<a href="htt://mylinux.dyndns.org"  href="http://myblog.com/LapppyNotReachable"> Me :-)</a> 

[07:36:01.0000] <kamathln>
what say

[07:36:02.0000] <kamathln>
?

[07:36:03.0000] <kamathln>
alternate hrefs when the orig hrefs go bad

[07:38:00.0000] <annevk>
unlikely I'd say

[07:38:01.0000] <kamathln>
annevk: in what way

[07:38:02.0000] <annevk>
on a basic level it's incompatible with the HTML parser and the DOM

[07:39:00.0000] <annevk>
and it's not really clear there's a real need for it

[07:39:01.0000] <annevk>
I suggest reading through http://wiki.whatwg.org/wiki/FAQ#Is_there_a_process_for_adding_new_features_to_a_specification.3F

[07:40:00.0000] <kamathln>
I am only discussing here .. more than just for when things become unreachable, I want the browsers to be able to handle when they dont know how to handle a rotocol

[07:40:01.0000] <kamathln>
protocol*

[07:40:02.0000] <kamathln>
eg: svn protocol there

[07:40:03.0000] <kamathln>
<a href="htt://mylinux.dyndns.org"  href="http://myblog.com/LapppyNotReachable"> Me :-)</a> 

[07:41:00.0000] <kamathln>
whups not that

[07:41:01.0000] <kamathln>
<a href="svn://svn.something.com/something"  href="http://something.com/svn"> Appspot</a> 

[07:41:02.0000] <kamathln>
that one

[07:41:03.0000] <Amadiro>
kamathln, so if the browser doesn't understand the svn:// protocol, it can use the alternative link?

[07:41:04.0000] <kamathln>
Amadiro: yes, thats the idea

[07:41:05.0000] <Amadiro>
(Does any browser understand the svn protocol?)

[07:41:06.0000] <Peter`>
wondering; there's registerProtocolHandler now, is there some kind of isProtocolKnown method?

[07:42:00.0000] <Amadiro>
kamathln, I kinda like the idea, I have to say

[07:42:01.0000] <kamathln>
Peter`: that would require Javascript right?

[07:42:02.0000] <Peter`>
Yes

[07:42:03.0000] <kamathln>
or at least some dynamic scripting 

[07:42:04.0000] <annevk>
Peter`, there's no infrastructure around it, no

[07:42:05.0000] <annevk>
Peter`, not really clear that's needed

[07:42:06.0000] <kamathln>
with alternate linking, it need not even be dynamic

[07:43:00.0000] <Peter`>
annevk: I haven't thought of any use-cases or whatsoever, was mostly wondering

[07:45:00.0000] <Amadiro>
kamathln, I'd imagine the browser would display a message ala "This location is not available because $reason, the website suggests to visit $alternative_link instead"

[07:47:00.0000] <kamathln>
Amadiro: hmm, that would be possible as well. I Think I should have used the wordings  "it *need not* be dynamic"

[07:48:00.0000] <kamathln>
non-dynamic fallback usefull when you are using resources from alternative sites 

[07:49:00.0000] <kamathln>
from mysite1 :   <img src="http://mysite2.com/kamathln/1.jpg" alt_src="http://mysite3..com/kamathln/1.jpg"> 

[07:52:00.0000] <kamathln>
programmers can think of this as a minimalistic try(){}catch(){}

[07:52:01.0000] <kamathln>
:)

[07:57:00.0000] <TabAtkins>
That only gives you two sources, though - even better would be to leverage CSS and the image() function (when it's supported), like <img src=foo style="content: contents, image(bar, baz, qux);">.  That'll try each of the four images in turn.

[07:58:00.0000] <kamathln>
TabAtkins: I have been thinking about the problem.. There must be a way to specify any number of URLs 

[07:58:01.0000] <TabAtkins>
There is.  I just gave one.

[07:59:00.0000] <TabAtkins>
It just so happens that nobody implements that part of CSS yet.

[07:59:01.0000] <kamathln>
TabAtkins: that would work only for the IMG problem .. what about a href , js  and css srcs ?

[07:59:02.0000] <TabAtkins>
Then it indeed would not work.

[07:59:03.0000] <kamathln>
and now, audio and video .. 

[07:59:04.0000] <kamathln>
exactly ..

[08:00:00.0000] <zcorpan_>
kamathln: audio and video support multiple fallbacks already

[08:01:00.0000] <kamathln>
zcorpan_: Oh yeah! then it would be (sort of) shamefull to not have them for anywhere we need to specify hrefs and srcs 

[08:01:01.0000] <kamathln>
/me starts groping for audio and video documentation to find how its implemented

[08:02:00.0000] <zcorpan_>
i don't see why we'd want to support fallbacks everywhere

[08:05:00.0000] <kamathln>
zcorpan_: did you login after I started to chat ? I already gave 3 usecases .. 1)general unaivalability of the page, 2)unaivalability of a domain providing  a resource, and 3) inability of a browser to handle a special protocol (like svn, I guess)

[08:05:01.0000] <annevk>
they're very minor

[08:06:00.0000] <annevk>
and most can be checked beforehand by the page provider providing an even better user experience

[08:06:01.0000] <zcorpan_>
what would a browser do with an svn: link?

[08:06:02.0000] <annevk>
TabAtkins, not really convinced that image() proposal is a good idea to be honest

[08:06:03.0000] <kamathln>
launch tortoissvn?

[08:06:04.0000] <kamathln>
tortoisesvn?

[08:07:00.0000] <zcorpan_>
what's the use case for that?

[08:08:00.0000] <zcorpan_>
i'd rather have svn: urls written in plain text so i can copy it and paste it into my terminal or a script

[08:10:00.0000] <zcorpan_>
for 1), if there's reason to believe a page will become unavailable, and you have a backup page that is not unavailable, why not just link to the backup page directly?

[08:11:00.0000] <kamathln>
irc:// xmpp:// lastfm:// webcal:// .. etc..

[08:11:01.0000] <kamathln>
svn was just one example

[08:11:02.0000] <zcorpan_>
what would you fall back to for those?

[08:11:03.0000] <kamathln>
webclients

[08:12:00.0000] <annevk>
(besides, only 0.0000000001% or so of all links fall in that category)

[08:12:01.0000] <Philip`>
zcorpan_: You may believe the backup page will become unavailable too, and more frequently than the original page, but rarely at the same time

[08:12:02.0000] <zcorpan_>
why not provide two links? some users might prefer the web client over the native app

[08:13:00.0000] <kamathln>
/me wears a second thinking cap

[08:14:00.0000] <zcorpan_>
Philip`: fair enough. however i haven't seen anyone work around this limitation so i'm not convinced it's a problem that needs solving

[08:14:01.0000] <doublec>
zcorpan_, video and audio don't have fallback in the case of the resource being unavailable - which is what I think kamathln was wanting

[08:15:00.0000] <zcorpan_>
doublec: sure it does

[08:15:01.0000] <zcorpan_>
doublec: if a video is not available, the next <source> is tried

[08:16:00.0000] <kamathln>
why would you provide it for video, and not img ? I still dont get.

[08:16:01.0000] <zcorpan_>
that's not the problem <source> was intended to solve, but it does as a side effect

[08:16:02.0000] <kamathln>
zcorpan_: oh okies

[08:16:03.0000] <Philip`>
zcorpan_: I guess the common case is that you own both the linker and linkee and run them on the same server, so if it's down then it doesn't matter what links you write; or the linker and linkee are different people, and the linker doesn't know enough about the linkee's backup/mirroring/update strategy to sensibly link to multiple URLs

[08:17:00.0000] <doublec>
zcorpan_, erm, I don't think it does

[08:17:01.0000] <doublec>
zcorpan_, if the resource fails to load then an error is raised

[08:18:00.0000] <TabAtkins>
annevk: image() in general, or just for this use-case?  image() is just wrapping up the behavior of the "content" property in a more general fashion for images themselves.

[08:18:01.0000] <zcorpan_>
doublec: could you point to the relevant part of the spec that says so?

[08:18:02.0000] <doublec>
zcorpan_, 4.8.10.5

[08:19:00.0000] <doublec>
although it appears to have changed a lot since I last read it

[08:19:01.0000] <doublec>
/me double checks

[08:19:02.0000] <zcorpan_>
"⌛ If absolute URL was not obtained successfully, then end the synchronous section, and jump down to the failed step below."

[08:20:00.0000] <zcorpan_>
"⌛ If the node after pointer is a source element, let candidate be that element."

[08:21:00.0000] <zcorpan_>
"⌛ If candidate is null, jump back to the search loop step. Otherwise, jump back to the process candidate step."

[08:22:00.0000] <paul_irish>
miketaylr: really? shorturl+fragment was working fine for me. it broke something? :(

[08:23:00.0000] <doublec>
what does "If absolute URL was not obtained successfully" mean?

[08:23:01.0000] <miketaylr>
paul_irish: ?

[08:24:00.0000] <doublec>
It's really this bit we're talking about, right? "Run the resource fetch algorithm with absolute URL. If that algorithm returns without aborting this one, then the load failed. "

[08:26:00.0000] <doublec>
yeah it looks like you're right zcorpan_ 

[08:26:01.0000] <zcorpan_>
doublec: what you quoted is about <video src="">, not <source>

[08:27:00.0000] <doublec>
zcorpan_, no what I quoted is also in the "Otherwise, the source  elements will be used; run these substeps:" section

[08:27:01.0000] <zcorpan_>
oh sorry

[08:27:02.0000] <doublec>
it seems hard to refer to parts of the spec in this area

[08:28:00.0000] <doublec>
lots of numbered steps and parts without subheadings

[08:29:00.0000] <doublec>
the resource fetch algorithm lumps 4XX errors in the same class as unsupported media which is where it confirms what you are saying I think

[08:29:01.0000] <doublec>
http://www.whatwg.org/specs/web-apps/current-work/multipage/video.html#concept-media-load-resource

[08:29:02.0000] <zcorpan_>
doublec: so if it isn't aborted (because the load failed), then the next step is run which will process the next <source>

[08:29:03.0000] <doublec>
yep

[08:30:00.0000] <doublec>
lucky I didn't work on the resource selection/loading part of our implementation

[08:30:01.0000] <doublec>
:)

[08:32:00.0000] <paul_irish>
miketaylr: wrong mike. :)

[08:33:00.0000] <miketaylr>
so many mikes!

[08:33:01.0000] <zcorpan_>
firefox fails 28 out of my 35 resource selection tests

[08:35:00.0000] <zcorpan_>
opera fails 4, chrome fails 28

[08:37:00.0000] <no_mind>
is there any attempt or demo or displaying streaming video under html5 <video> tag ?

[08:40:00.0000] <doublec>
zcorpan_, where are your tests?

[08:41:00.0000] <doublec>
zcorpan_, no doubt bug 485288 https://bugzilla.mozilla.org/show_bug.cgi?id=485288

[08:42:00.0000] <zcorpan_>
doublec: not public yet :(

[08:45:00.0000] <annevk>
TabAtkins, image() in general I think; I don't really believe in fallback too much

[08:45:01.0000] <TabAtkins>
Ah, ok.

[08:45:02.0000] <annevk>
that it works and is used in practice, that is

[08:46:00.0000] <annevk>
it's definitely way too small a thing to hit 80/20 in general

[08:47:00.0000] <annevk>
(media might be slightly different, but even there it is mostly because the format situation is fucked up, not because fallback is needed in general)

[08:50:00.0000] <TabAtkins>
Hm, maybe.

[08:51:00.0000] <TabAtkins>
I do like the ability to fallback to a color, though.

[08:52:00.0000] <annevk>
how often do you see images not load?

[08:52:01.0000] <TabAtkins>
When I specifically know that my text won't be legible on the underlying background and am relying on an image to provide the proper color contrast, but the image is partially transparent so I can't just pair it with a background-color.

[08:52:02.0000] <annevk>
when we get SPDY it's even less likely you'll notice delay by images

[08:53:00.0000] <annevk>
it's just way too much into theoretical space

[08:53:01.0000] <annevk>
there's a lot of other things we could tackle instead

[08:53:02.0000] <TabAtkins>
Certainly.

[08:53:03.0000] <zcorpan_>
TabAtkins: text-shadow? :)

[08:54:00.0000] <TabAtkins>
zcorpan_: That only works if you *want* a shadow.

[08:54:01.0000] <annevk>
that images are not loading is not something that happens nowadays though

[08:54:02.0000] <annevk>
maybe in the nineties

[08:55:00.0000] <TabAtkins>
I've had broken images sit around on my sites before.  ^_^

[08:55:01.0000] <TabAtkins>
Though, to be fair, that stopped happening as soon as I hooked up an email script to my 404 page.

[08:55:02.0000] <TabAtkins>
And if an author is smart enough to know that they need fallback, they're smart enough to make their 404 page email them too.

[08:55:03.0000] <annevk>
cnn.com would be a fair example (especially if their developers want to tackle the issue)

[08:56:00.0000] <annevk>
your own site not so much

[08:56:01.0000] <annevk>
lets be at least a little pragmatic here :)

[08:56:02.0000] <TabAtkins>
I meant "a company site that I managed as the webmaster".

[09:01:00.0000] <annevk>
things that seem more relevant for the CSS WG to tackle: <meta name=viewport>, fullscreen API, gradients, form control styling, ...

[09:01:01.0000] <annevk>
also a lot more important

[09:02:00.0000] <hsivonen>
hmm. Nokia announces an XForms client for J2ME

[09:02:01.0000] <annevk>
we're now implementing this border-image thing which has become quite the complexity while form controls are still not figured out, while the table model is still undefined...

[09:02:02.0000] <annevk>
the priorities are way wrong imo

[09:03:00.0000] <hsivonen>
https://projects.forum.nokia.com/XFormsJ2ME

[09:05:00.0000] <annevk>
Symbian, J2ME, Maemo, ... where are they going? 

[09:07:00.0000] <hsivonen>
annevk: looks like they are trying to keep all three for different market segments

[09:07:01.0000] <hsivonen>
but why XForms for any of the market segments? and why now?

[09:07:02.0000] <TabAtkins>
annevk: Dude, start kicking some asses then.  You never talk about any of this stuff on the list or in meetings.  I get easily distracted by shiny new things.

[09:10:00.0000] <annevk>
TabAtkins, heh, I certainly try, but it's not really working

[09:10:01.0000] <annevk>
TabAtkins, also, I haven't found much time to work on CSS stuff other than CSSOM

[09:10:02.0000] <TabAtkins>
I haven't seen much trying.  ^_^

[09:11:00.0000] <TabAtkins>
Yeah, granted.

[09:16:00.0000] <TabAtkins>
Um, does anyone implement a box-shadow-style property?

[09:16:01.0000] <TabAtkins>
/me doesn't think so.

[09:22:00.0000] <annevk>
TabAtkins, I haven't done it much recently, but it's also quite hard

[09:23:00.0000] <Peter`>
TabAtkins: sorry, meant box-shadow by itself

[09:25:00.0000] <TabAtkins>
Ah, that Peter is you?

[09:25:01.0000] <Peter`>
Yes

[09:25:02.0000] <Peter`>
I overlooked the fact that border-radius is in fact a shorthand method

[09:26:00.0000] <TabAtkins>
Heh, I actually forgot about that too.  I only ever use the shorthand.

[09:26:01.0000] <TabAtkins>
(Though that's partially because of the mismatch between subproperty names between webkit and gecko.)

[09:26:02.0000] <TabAtkins>
/me doesn't know if that's been fixed yet.

[09:27:00.0000] <Peter`>
Same here, indeed :-)

[09:27:01.0000] <TabAtkins>
Also, remember that you can do multiple shadows right now by just specifying multiple shadows comma-separated.

[09:28:00.0000] <Peter`>
I didn't know that, thanks! I'll do some experimenting with it. I do remember seeing some fancy coloured shadow-examples which didn't immediately make sense -they do now-.

[10:05:00.0000] <TabAtkins>
Hixie: Webkit is implementing dataset right now, and uncovered a spec bug.  I've emailed whatwg about it, and would appreciate a quick response.

[10:07:00.0000] <Philip`>
TabAtkins: Why do you think there's a problem?

[10:08:00.0000] <Philip`>
The part you cut out of the quote says the attribute name needs to be XML-compatible too

[10:08:01.0000] <Philip`>
and the processing algorithm ignores that too

[10:09:00.0000] <Philip`>
(i.e. empty and/or XML-incompatible names aren't allowed but will be extracted by the algorithm)

[10:10:00.0000] <TabAtkins>
Well, it doesn't gain us anything to process <div data-=foo>, since there isn't anything on the web that depends on it.  Might as well align author and impl requirements here.

[10:11:00.0000] <Philip`>
Do you want to align to the XML-compatible requirement too?

[10:14:00.0000] <TabAtkins>
Maybe, though I'd be equally happy to drop that requirement.

[10:19:00.0000] <jgraham>
The empty string case is confusing

[10:19:01.0000] <jgraham>
So that shouldn't work, I think

[10:21:00.0000] <TabAtkins>
We're already verifying that the name doesn't include any capital letters, so I suspect it's not overly a burden to verify that it only contains characters in the xml-compatible range.

[10:21:01.0000] <TabAtkins>
(http://www.w3.org/TR/REC-xml/#NT-NameStartChar and NameChar)

[10:21:02.0000] <annevk>
that seems like overkill

[10:22:00.0000] <TabAtkins>
Possibly, yeah, so I'm not against ignoring that.

[10:23:00.0000] <annevk>
I don't really see why author and user agent requirements need to be aligned here

[10:24:00.0000] <Philip`>
TabAtkins: Do you want http://www.w3.org/TR/REC-xml/#NT-NameChar or http://www.w3.org/TR/2006/REC-xml-20060816/#NT-Name ?

[10:24:01.0000] <Philip`>
TabAtkins: by which I mean, I'd personally be happy to not have to care and instead to allow anything

[10:24:02.0000] <TabAtkins>
I linked to the one that HTML5 does, which is the extent of my caring.

[10:24:03.0000] <jgraham>
I think it only makes sense to allow anything, not to only process XML-compatible names

[10:25:00.0000] <jgraham>
Down that path lies madness

[10:26:00.0000] <Philip`>
Down that path lies a lot of work writing test cases and filing bugs, for no benefit to anybody

[10:26:01.0000] <TabAtkins>
That's cool with me.

[10:26:02.0000] <jgraham>
But dropping the processing of empty names seems fine

[10:26:03.0000] <TabAtkins>
I'm also cool with changing the author conformance to say that "data-" is valid.

[10:26:04.0000] <jgraham>
I see no benefit to processing them

[10:26:05.0000] <jgraham>
and allowing them would be weird

[10:28:00.0000] <Philip`>
If you don't process "data-", then should the name-setting algorithm throw an exception on dataset[""] = "foo"?

[10:28:01.0000] <Philip`>
to avoid losing the (I think) guaranteed roundtripping

[10:28:02.0000] <TabAtkins>
If we did restrict it, then yes.

[10:29:00.0000] <jgraham>
It could just silently ignore it

[10:29:01.0000] <jgraham>
That would be The ECMAScript Way (TM)

[10:30:00.0000] <jgraham>
(but sure, throwing INVALID_ARGUMENTS_ERROR or whatever it's called seems fine)

[10:30:01.0000] <Philip`>
It should silently ignore dataset["-A"] too, instead of throwing like that?

[10:30:02.0000] <Philip`>
(in your proposed Way (TM))

[10:31:00.0000] <Philip`>
s/like that/like it does now/

[10:31:01.0000] <TabAtkins>
Obviously we should match everything one way or another.  ^_^

[10:31:02.0000] <jgraham>
Philip`: I have no idea what it does now, and would mildly prefer throwing to ignoring

[10:32:00.0000] <TabAtkins>
/me sighs.

[10:33:00.0000] <TabAtkins>
I wish XML weren't so stupid.

[10:33:01.0000] <annevk>
http://lists.w3.org/Archives/Public/public-device-apis/2010Jun/0334.html -- I guess with some people "URIs" is some kind of gut reaction to problems :/

[10:33:02.0000] <jgraham>
TabAtkins: Blame Tim Bray

[10:33:03.0000] <TabAtkins>
I will!

[10:33:04.0000] <jgraham>
Since everyone has the (somewhat inaccurate) idea he is the Father of XML

[10:34:00.0000] <TabAtkins>
Argh, he's work-at-home.  So I can't bug him on campus.

[10:34:01.0000] <annevk>
implement XML5 in Chrome

[10:34:02.0000] <annevk>
it's much saner

[10:36:00.0000] <jgraham>
TabAtkins: You could wait 'till he next visits and follow him home

[10:36:01.0000] <jgraham>
and then ring on his doorbeel at three in the morning

[10:36:02.0000] <jgraham>
and hide out in his bins

[10:36:03.0000] <TabAtkins>
jgraham: Brilliant!

[10:37:00.0000] <TabAtkins>
annevk: XML5 being backwards compatible means that a substantial bit of craziness is still present, right?

[10:37:01.0000] <TabAtkins>
Like the namechar restrictions?

[10:38:00.0000] <annevk>
those are gone, for instance

[10:38:01.0000] <TabAtkins>
So it's somewhat backwards compatible?

[10:38:02.0000] <annevk>
there's no need for arbitrary restrictions to stay compatible with well-formed content

[10:38:03.0000] <TabAtkins>
kk

[10:39:00.0000] <TabAtkins>
Ah, so XML5 processors read XML1 content, but not the other way round.  Got it.

[10:39:01.0000] <jgraham>
The doctype madness is still there iirc

[10:39:02.0000] <jgraham>
(I mean it has to be)

[10:39:03.0000] <annevk>
TabAtkins, depends on how you craft your XML5 content, yup

[10:39:04.0000] <TabAtkins>
Yeah, of course.

[10:39:05.0000] <TabAtkins>
/me wants something even simpler.  Screw backwards compatibility.

[10:39:06.0000] <annevk>
jgraham, yes, it's about half the complexity of the tokenizer (or the whole parser, really)

[10:40:00.0000] <annevk>
TabAtkins, I don't think you can get much simpler (apart from taking out all the DOCTYPE states) while still having the expressiveness of the DOM

[10:41:00.0000] <TabAtkins>
I should actually look at XML5 before declaring it not-simple-enough.

[10:42:00.0000] <Philip`>
It seems quite common to implement XML parsers that don't support internal subsets

[10:43:00.0000] <Philip`>
so there'd probably be demand for a spec that isn't entirely backward-compatible in that way

[10:50:00.0000] <TabAtkins>
Oof, the PI and DOCTYPE stuff really is the majority of the parser.

[10:51:00.0000] <TabAtkins>
Ignoring those two groups, there are only 22 states.

[10:51:01.0000] <TabAtkins>
(And about half of *those* are just dealing with comments and CDATA.)

[10:52:00.0000] <TabAtkins>
Well, a third.  Looks like 7 or 8 states devoted to those two.

[10:53:00.0000] <zcorpan_>
TabAtkins: you want XML5 but without PI, doctype, comment and CDATA?

[10:54:00.0000] <TabAtkins>
That'd be nice, yeah.

[10:54:01.0000] <TabAtkins>
Though comments are useful enough that I can accept those in there.

[10:55:00.0000] <jgraham>
CDATA seems pretty useful

[10:55:01.0000] <jgraham>
PI not so much

[10:55:02.0000] <jgraham>
(pity about the CDATA syntax, really)

[10:58:00.0000] <TabAtkins>
annevk: If I'm reading the comment states correctly, the intention is that <!-- and --> are actual comment start/end, and -- inside of them is just treated as characters?

[10:58:01.0000] <annevk>
yeah

[10:58:02.0000] <TabAtkins>
Cool.

[11:02:00.0000] <TabAtkins>
And all the doctype and pi states are solely to consume doctypes, not do anything with them?

[11:03:00.0000] <TabAtkins>
jgraham: is CDATA very useful?  I really don't know - I just use <style> and <script> without worrying about the fact that they might contain things that look disturbingly like markup.

[11:03:01.0000] <annevk>
DOCTYPEs are processed actually

[11:03:02.0000] <TabAtkins>
I just assume the rest is magic.

[11:03:03.0000] <annevk>
to support entity references

[11:03:04.0000] <annevk>
and such nonsense

[11:03:05.0000] <annevk>
PIs are put in the DOM

[11:03:06.0000] <TabAtkins>
Ah, yeah, I see the PI part.

[11:03:07.0000] <TabAtkins>
I just skimmed the doctype states, so must have missed where they were actually parsed.

[11:04:00.0000] <annevk>
large parts of it are ignored

[11:04:01.0000] <annevk>
and are just about correctly processing them

[11:04:02.0000] <annevk>
but some parts are taken into consideration

[11:04:03.0000] <jgraham>
TabAtkins: In XML is guess it is useful

[11:04:04.0000] <jgraham>
I mean it is confusing if you can't do if (1<b) in your script

[11:04:05.0000] <TabAtkins>
jgraham: Presumably so you can do the script/style magic without having to know that the elements are magic beforehand?

[11:05:00.0000] <jgraham>
Yeah

[11:05:01.0000] <TabAtkins>
Makes sense.

[11:06:00.0000] <TabAtkins>
Though it should be even more explicit - it's too easy to get a nested array-reference in a numeric comparison and fake out the cdata-end state.  ^_^

[11:09:00.0000] <annevk>
you can just do &lt;

[11:09:01.0000] <annevk>
works fine

[11:10:00.0000] <TabAtkins>
There is no way I'm writing |if(a[b[c]]&lt;42) {...}|

[11:10:01.0000] <annevk>
you rather write <![CDATA[ crap?

[11:11:00.0000] <annevk>
to each his own I suppose

[11:11:01.0000] <annevk>
but my way requires less states :p

[11:11:02.0000] <TabAtkins>
If I'm handwriting?  Yeah.

[11:11:03.0000] <TabAtkins>
/me supposes he could just avoid directly embedding non-XML languages in an xml document.

[11:14:00.0000] <TabAtkins>
/me would enjoy a language with only &lt; and &quot; as entities.

[11:15:00.0000] <zcorpan_>
and &amp;?

[11:16:00.0000] <TabAtkins>
Right, yes.

[11:17:00.0000] <TabAtkins>
/me would like to find a way to avoid that, too.

[11:19:00.0000] <zcorpan_>
/me finds an srt with <b><font color="#00afad">Join us on Facebook !

[11:19:01.0000] <zcorpan_>
/me Squadra Dell'Ombra</font></b>

[11:19:02.0000] <zcorpan_>
Hixie: ^

[11:25:00.0000] <zcorpan_>
Hixie: http://www.tvsubtitles.net/subtitle-132566.html has <i>Et maintenant "Les transcroyables

[11:25:01.0000] <zcorpan_>
exploits de Zapp Brannigan"

[11:25:02.0000] <zcorpan_>
(i.e. unclosed <i>

[11:25:03.0000] <zcorpan_>
)

[11:26:00.0000] <Hixie>
zcorpan_: i'm intentionally not going to be supporting attributes at this point

[11:26:01.0000] <Hixie>
zcorpan_: and will support unclosed <i>s

[11:28:00.0000] <zcorpan_>
Hixie: it was just the first srt i found with <b>

[11:28:01.0000] <Hixie>
ah

[11:28:02.0000] <zcorpan_>
{\pos(192,240)}On a des photos.

[11:28:03.0000] <zcorpan_>
http://www.tvsubtitles.net/subtitle-132551.html

[11:29:00.0000] <zcorpan_>
<i>{\a6}TY,

[11:29:01.0000] <zcorpan_>
L'ASSISTANT DU CORONER</i>

[11:31:00.0000] <Hixie>
wow, i wonder what UA supports that

[11:38:00.0000] <zcorpan_>
VLC doesn't support it

[11:42:00.0000] <annevk>
hmm, VLC is pretty good usually

[11:42:01.0000] <Hixie>
TabAtkins: you're e-mail confuses authoring criteria with implementation requirements, there's no ambiguity

[11:52:00.0000] <zcorpan_>
MPlayer seems to support {\pos(192,240)}

[11:53:00.0000] <zcorpan_>
and ignores {\a6}, or replaces it with a space or something

[11:56:00.0000] <Hixie>
(um, s/you're/your/)

[12:02:00.0000] <zcorpan_>
Hixie: all subtitles from tvsubtitles.net seem to have 9999

[12:02:01.0000] <zcorpan_>
00:00:0,500 --> 00:00:2,00

[12:02:02.0000] <zcorpan_>
<font color="#ffff00" size=14>www.tvsubtitles.net</font>

[12:02:03.0000] <zcorpan_>
at the *end* of the file

[12:08:00.0000] <Hixie>
yeah, i think the parser will likely support out-of-order cues

[12:09:00.0000] <Hixie>
(in fact it already does)

[12:10:00.0000] <Hixie>
it doesn't support times with only one digit for the seconds or two digits for the thousandths, though

[12:10:01.0000] <Hixie>
do UAs support that?

[12:10:02.0000] <Hixie>
it's trivial for me to add support if necessary

[12:12:00.0000] <Hixie>
zcorpan_: btw so i don't miss them it'd really useful if you could note your observations of what you found in the wild on the iki

[12:12:01.0000] <Hixie>
wiki

[12:12:02.0000] <zcorpan_>
VLC supports single-digit seconds

[12:12:03.0000] <zcorpan_>
ok

[12:12:04.0000] <Hixie>
thanks for doing this btw, it's really helpful

[12:13:00.0000] <zcorpan_>
MPlayer too

[12:16:00.0000] <zcorpan_>
vlc seems to drop the thousands if they're 2 digits

[12:17:00.0000] <zcorpan_>
heh, vlc interprets 00:00:00,5000 as if it were 00:00:05,000

[12:18:00.0000] <zcorpan_>
it's possible that it takes the two digits but it's too close to 0 for me to tell

[12:20:00.0000] <Hixie>
100ms should distinguishable from 0ms, try having multiple overlapping titles

[12:20:01.0000] <Hixie>
oh wait doesn't vlc do automatic overlay protection?

[12:21:00.0000] <Hixie>
if so you can just have some that would overlap and some that would not based on how it interprets two digit times

[12:21:01.0000] <Hixie>
and see how it positions them

[12:21:02.0000] <Hixie>
bbiab, going to the office

[12:21:03.0000] <zcorpan_>
mplayer also interprets 00:00:00,5000 as if it were 00:00:05,000

[12:25:00.0000] <zcorpan_>
ok vlc interprets 00:00:01,99 as 00:00:01,099

[12:28:00.0000] <zcorpan_>
vlc seems to just overlap without changing position

[12:32:00.0000] <zcorpan_>
for the pay-per-minute or per-byte use case, can't the server just throttle the download as it sees fit?

[12:33:00.0000] <zcorpan_>
if it's on the client side it seems easy to bypass anyway

[13:03:00.0000] <Hixie>
zcorpan_: thanks for the notes on the wiki

[13:03:01.0000] <zcorpan_>
np

[13:25:00.0000] <Hixie>
zcorpan_: we can go several ways on the timestamp parsing thing

[13:25:01.0000] <Hixie>
zcorpan_: we could just support arbitrary values in each field, and then add them together without checking ranges

[13:26:00.0000] <Hixie>
zcorpan_: so e.g. 1:60:60,60000 would turn into 1 hour + 60 minutes + 60 seconds + 60000 thousandths, i.e. 2h2m, same as 02:02:00.000

[13:27:00.0000] <Hixie>
zcorpan_: we could check ranges, but support missing zero padding, so e.g. 1:1:1,1 => 01:01:01.001

[13:28:00.0000] <Hixie>
zcorpan_: or we could be strict, and not support this contentz

[13:28:01.0000] <Hixie>
zcorpan_: any opinions?

[13:34:00.0000] <Hixie>
i went with the first option

[13:35:00.0000] <zcorpan_>
that's what i was going to suggest (for compat with vlc and mplayer; i have no idea if there's any content that relies on this)

[13:37:00.0000] <TabAtkins>
Adding them all together is the best.  Makes so many things simpler when you don't have to futz around with doing base-60 arithmetic yourself.

[13:37:01.0000] <Hixie>
well there's the content you pasted which relies on us doing either the first two and not the third

[13:38:00.0000] <Hixie>
TabAtkins: well they're still not valid at the moment

[13:38:01.0000] <Hixie>
this is just error handling behaviour

[13:38:02.0000] <TabAtkins>
d'oh.

[13:38:03.0000] <TabAtkins>
I like it when that sort of thing is expicitly allowed.

[13:38:04.0000] <Hixie>
we can talk about what should be valid too, but that's a discussion for another time

[13:38:05.0000] <Hixie>
i'm starting the format as strict as possible in terms of what's allowed, e.g. it even wants the cues in order

[13:38:06.0000] <Hixie>
which i imagine we'll later relax

[13:38:07.0000] <Hixie>
but we'll see

[13:39:00.0000] <TabAtkins>
/me thinks specifically of date libraries that are perfectly happy being asked to create a datetime for March -3 or whatever.

[13:40:00.0000] <zcorpan_>
Hixie: ah, i didn't even notice the lack of zeros in 00:00:0,500 --> 00:00:2,00 when i pasted it

[13:41:00.0000] <Hixie>
heh

[13:41:01.0000] <Hixie>
TabAtkins: there are arguments both ways on this -- in this particular case, given how easy it is to write a script that normalises the files, the benefit might not outweigh the cost

[13:42:00.0000] <Hixie>
(the cost is that it makes understanding what's going on kinda confusing)

[13:42:01.0000] <Hixie>
(e.g. you can't tell at a glance which cue is higher than which cue)

[13:43:00.0000] <TabAtkins>
I blame human-based timestamps.

[13:46:00.0000] <annevk>
http://html5.org/tools/web-apps-tracker?from=5123&to=5124 -- I hope we don't get these

[13:46:01.0000] <hsivonen>
annevk: congrats on the chairship

[13:48:00.0000] <annevk>
thanks, though the charter still needs to get approved by the AC

[13:48:01.0000] <Hixie>
chairship?

[13:49:00.0000] <annevk>
http://www.w3.org/2010/06/notification-charter.html

[13:50:00.0000] <Hixie>
oh, i've been ignoring that thread

[13:51:00.0000] <Hixie>
glad you're chairing that

[13:51:01.0000] <Hixie>
grats

[13:51:02.0000] <Hixie>
keep 'em sane!

[13:53:00.0000] <Hixie>
/me wonders how to structure the output of the cue text parser

[13:53:01.0000] <Hixie>
i could just output DOM nodes

[13:53:02.0000] <Hixie>
but that seems a bit heavy-weight if implementations decide to actually take me to my word

[13:57:00.0000] <zcorpan_>
Hixie: i think you can assume that implementors are going to cut corners

[14:18:00.0000] <hsivonen>
Hixie: if the cue text parser outputs DOM nodes, why not use the HTML fragment parsing algorithm?

[14:21:00.0000] <karlushi>
trying to imagine the concrete use cases from the report http://www.w3.org/2010/02/mbui/report.html and have difficulties to see them. I guess I should read the papers

[14:21:01.0000] <othermaciej>
annevk: I offer you a virtual beer of sympathy

[14:24:00.0000] <zcorpan_>
hsivonen: it's interleaved with karaoke timestamps and voices

[14:25:00.0000] <hsivonen>
zcorpan_: I didn't notice voices inside the cue text

[14:25:01.0000] <hsivonen>
/me looks

[14:25:02.0000] <zcorpan_>
hsivonen: and maybe VLC doesn't want to bundle a whole HTML parser just for websrt

[14:26:00.0000] <zcorpan_>
hsivonen: oh maybe not voices

[14:27:00.0000] <hsivonen>
zcorpan_: VLC already bundles the kitchen sink

[14:27:01.0000] <hsivonen>
zcorpan_: I thought we'd have learned that avoiding HTML parsing is the wrong optimization

[14:28:00.0000] <hsivonen>
we could mint an element for karaoke timing within cue text

[14:28:01.0000] <karlushi>
http://www.google.com/search?q=websrt+site%3Avideolan.org

[14:29:00.0000] <karlushi>
[nil]

[14:31:00.0000] <hsivonen>
/me still thinks karaoke stuff is a distraction as far as solving an accessibility problem goes

[14:32:00.0000] <karlushi>
http://trac.videolan.org/vlc/search?q=Subtitles nothing related to websrt

[14:33:00.0000] <zcorpan_>
hsivonen: if browsers support any elements in captions then soon enough all media players need to bundle a browser engine for compat with content

[14:34:00.0000] <hsivonen>
zcorpan_: maybe. one could argue that if browsers support any elements in prose, word processors need to bundle browser engines

[14:34:01.0000] <hsivonen>
when designing a format for the Web, needing a browser engine isn't a bug, IMO

[14:37:00.0000] <karlushi>
hsivonen, maybe, but it doesn't necessary help with the format having success in other environments. (thinking). If the format has no strong ties with the browser engine, I guess it doesn't matter that much.  

[14:39:00.0000] <hsivonen>
karlushi: SRT already has success in other environments. WebSRT is leveraging that. Not the other way round.

[14:54:00.0000] <annevk>
both sides make sense

[14:55:00.0000] <annevk>
though I suspect browsers will at some point be a low-level piece of software, like unix, so having them as a requirement is probably not too bad

[14:56:00.0000] <TabAtkins>
"at some point"?

[14:57:00.0000] <annevk>
it's not quite there yet

[15:00:00.0000] <TabAtkins>
AryehGregor: Plus, the only pay-per-minute video I've ever seen on the web is porn.

[15:03:00.0000] <karlushi>
/me suddenly think that TabAtkins has spent money for nothing, there is so much free stuff out there.

[15:04:00.0000] <TabAtkins>
Oh, I've never paid for it.  You kidding?  Giving your credit card to a porn company is just *asking* for identity theft.

[15:04:01.0000] <TabAtkins>
I'm just saying, that's the only place I've ever seen pay-per-minute used.

[15:05:00.0000] <Hixie>
hsivonen: because i don't want to require that VLC embed an HTML parser when it's not necessary

[15:05:01.0000] <karlushi>
/me was just quoting TabAtkins ;) "I've ever seen" and being playful.

[15:08:00.0000] <annevk>
Hixie, the counter argument is that VLC will be eaten by the browser Godzilla in the end and that therefore complicating the browser more for the sake of VLC is not worth it long term

[15:10:00.0000] <Philip`>
/me kind of likes having several small independent applications, rather than one giant monolithic one

[15:10:01.0000] <annevk>
(and maybe/hopefully, long term HTML parsing will be like number parsing in programming languages)

[15:11:00.0000] <karlushi>
/me agrees with Philip` 

[15:11:01.0000] <karlushi>
/me fears we are geeks though

[15:12:00.0000] <annevk>
I don't know

[15:12:01.0000] <annevk>
applications are on top of the platform

[15:12:02.0000] <annevk>
we're discussing the platform here

[15:15:00.0000] <annevk>
I don't mind yet another parser so much though

[15:16:00.0000] <annevk>
compared to layout, dynamic updates, etc., parsers are cheap

[15:17:00.0000] <annevk>
(though admittedly the HTML parser with script injection and all can get pretty darn complex)

[15:17:01.0000] <hsivonen>
Hixie: I think WebSRT is repeating the mistake that Pingback made (speccing a regexp instead of requiring real parsing)

[15:18:00.0000] <hsivonen>
(this time it's not a regexp but still the "can't require HTML parsing" mentality)

[15:18:01.0000] <hsivonen>
and we are the group that has finally solved HTML parsing so that it could now be required!

[15:19:00.0000] <Hixie>
hsivonen: who's speccing a regexp?

[15:19:01.0000] <hsivonen>
Hixie: you did in Pingback

[15:19:02.0000] <Hixie>
i was ten years younger then :-)

[15:20:00.0000] <Hixie>
i can't use the html fragment parser because it doesn't support things like timestamps, and i don't want to use the html parser because that's about 1000 times the needed complexity 

[15:20:01.0000] <annevk>
Pingback was about finding something inside a text/html resource. That seems quite different from finding something in a text/srt resource.

[15:20:02.0000] <Hixie>
but i have no intention of speccing a fake parser, it'll be a real one

[15:20:03.0000] <hsivonen>
my point is that HTML parsing shouldn't be considered to be the bogeyman it was 10 years ago

[15:21:00.0000] <hsivonen>
Hixie: I'm suggesting that the HTML parser be used only for cue text

[15:21:01.0000] <Hixie>
sure, but the cue text has timestamps (for karaoke)

[15:22:00.0000] <hsivonen>
Hixie: I'm ok with a dedicated parser for the WebSRT encapsulation

[15:22:01.0000] <hsivonen>
Hixie: an element could be minted

[15:22:02.0000] <Hixie>
i don't consider the html parser a bogeyman, i consider it a massive piece of engineering and a pretty big part of my spec editing career... but it's still orders of magnitude more than we need here

[15:22:03.0000] <hsivonen>
Hixie: (and karaoke is on the wrong side of 80/20 anyway, IMO)

[15:23:00.0000] <Hixie>
i beg to differ on that last one, karaoke occurs in many (most?) anime videos, e.g.

[15:24:00.0000] <annevk>
most Ghibli movies I've seen so far need it

[15:24:01.0000] <TabAtkins>
Agreed.  It's either part of most movies, or at the least part of the opening/closing credits.

[15:25:00.0000] <annevk>
yeah, usually at the start/ending

[15:25:01.0000] <Hixie>
anyway, i really don't see why we'd reuse the html parser other than just because we already have one, which is a pretty poor reason given that media players aren't going to be the same as html browsers in many cases

[15:25:02.0000] <annevk>
I should say Totoro again; it's pretty great

[15:25:03.0000] <hsivonen>
would the creators of those movies be satisfied with text tracks anyway, or would they burn their art to the pixel data?

[15:25:04.0000] <annevk>
s/say/see/ doh

[15:26:00.0000] <boazsender>
There any mention in any spec of file writing with data uri's?

[15:26:01.0000] <boazsender>
like if I wanted to be able to do window.location='data:dataType,' + URIEncodedData;

[15:26:02.0000] <hsivonen>
Hixie: I think the WHATWG would be designing a captioning format for use on the Web. if it gets picked up by standalone media players, that would be a bonus, but IMO shouldn't be a design goal

[15:27:00.0000] <boazsender>
... and have that cause a file download. 

[15:27:01.0000] <boazsender>
the above works in firefox 3.6

[15:27:02.0000] <annevk>
hsivonen, you don't want to create captions twice

[15:28:00.0000] <hsivonen>
annevk: for the vast majority of cases, the WebSRT captions would be just like today's SRT

[15:28:01.0000] <hsivonen>
without fancy stuff

[15:28:02.0000] <annevk>
I wonder more how long the standalone player with no browser argument holds true...

[15:29:00.0000] <annevk>
hsivonen, presumably they would like to have conforming implementations too

[15:30:00.0000] <hsivonen>
annevk: we could use the "CSS is optional" pixie dust :-)

[15:31:00.0000] <zcorpan_>
annevk: legacy encodings is a bigger problem with existing srt

[15:33:00.0000] <annevk>
zcorpan_, I don't think it's about how much existing content works with WebSRT, it's more how easy WebSRT can be adopted by various vendors

[15:33:01.0000] <annevk>
hsivonen, as in HTML parsing is optional?

[15:33:02.0000] <hsivonen>
annevk: no, as in complex rendering of the data structure you get out of the HTML parser would be optional

[15:34:00.0000] <hsivonen>
HTML parsing is / will be off-the-shelf stuff

[15:34:01.0000] <annevk>
I think Hixie's point was that implementing an HTML parser was too much already

[15:35:00.0000] <annevk>
yeah, I think that will be true in due course too

[15:35:01.0000] <hsivonen>
annevk: my point is that you'll be able to use an HTML parser that's already been written

[15:35:02.0000] <annevk>
on the other hand, by that argument we should have used HTML for EventSource or manifest=""

[15:36:00.0000] <annevk>
if the part of WebSRT that is markup is significantly simpler than a full HTML parser it might be worth to have it actually be simpler

[15:36:01.0000] <hsivonen>
(I realize that what I say would be more credible if I was able to point to a C++ version of the V.nu/Gecko parser that didn't need Gecko stuuf around it)

[15:37:00.0000] <annevk>
your argument also kind of sounds like "lets use XML; it's great"

[15:39:00.0000] <annevk>
and we have rejected that too because of the sheer complexity

[15:39:01.0000] <annevk>
even though standard libraries are all over the place

[15:42:00.0000] <annevk>
anyway, I'm about 45min late for an appointment with a novel; nn :)

[15:43:00.0000] <roc>
WebSRT is renderable markup, EventSource and manifest are not

[15:43:01.0000] <roc>
the argument that "right now we only need a simple subset, so let's spec our own" is why SVG started encroaching on HTML and CSS

[15:44:00.0000] <roc>
it's troublesome because of course the requirements grow over time, and at each step the easiest thing to do for implementers, spec authors and Web authors is to add one more feature to the stand-alone subset

[15:45:00.0000] <roc>
however, argument by analogy is weak so that doesn't settle the issue

[15:46:00.0000] <AryehGregor>
I don't suppose there's any way to have, say, <input type="number"> that doesn't get validated, without adding attributes to the form itself or submit buttons?

[15:48:00.0000] <AryehGregor>
/me doesn't see one

[15:48:01.0000] <zcorpan_>
AryehGregor: i think you can override the validation of specific elements with setCustomValidity or something

[15:48:02.0000] <AryehGregor>
I was thinking without JS.  I could always add oninvalid="return false", yeah.

[15:49:00.0000] <zcorpan_>
then no, afaik

[16:11:00.0000] <othermaciej>
for those of you who have criticized the information considered, quality of reasoning, or clarity of writing of WG decisions, I would appreciate private feedback on http://lists.w3.org/Archives/Public/public-html/2010Jun/att-0659/issue-93-decision.html

[16:12:00.0000] <othermaciej>
(I believe this includes Lachy, TabAtkins, sicking, Hixie and probably others)

[16:13:00.0000] <othermaciej>
I'll also be asking some non-WHATWG folks who complained in the past for their private feedback

[16:13:01.0000] <TabAtkins>
othermaciej: Will do.

[16:14:00.0000] <othermaciej>
(would rather not have a huge public debate about it but I'd like to know if people who were concerned in the past feel their concerns were addressed)

[16:17:00.0000] <hober>
othermaciej: as one of the zero-edit CP authors, I think that the points we raised were adequately addressed in the decision's stated rationale.

[16:35:00.0000] <MikeSmith>
paul_irish: about the frag ID thing, I was wrong .. seems it works in most UAs

[16:35:01.0000] <MikeSmith>
just wasn't working in the first one I had tried it in

[16:53:00.0000] <JonathanNeal>
If you were putting an image next to a word that it represents, then the alt of the image is usually the word, but in this case a screen reader will just read the word twice.

[16:53:01.0000] <JonathanNeal>
How should this be handled?

[16:55:00.0000] <JonathanNeal>
Sorry about that.  Whoops.

[16:56:00.0000] <TabAtkins>
JonathanNeal: Description of the image, in that case.

[16:58:00.0000] <JonathanNeal>
But the description of the image is the content next to it.

[16:59:00.0000] <JonathanNeal>
Is there any way of tying these together?