00:11
<othermaciej>
AryehGregor: I agree, Hixie should probably take that out (I think he is out today though)
00:35
<AryehGregor>
"Microdata brings torment and dismay."
00:35
AryehGregor
suddenly realizes that the channel is +t, but there are no ops . . . do you get a freenode staffer in every time you want to change the topic?
00:38
<Dashiva>
Change the topic? What's that
00:39
<Dashiva>
I recall it was -t before, I'm guessing the mode got set during a netsplit or something
00:41
<GarethAdams|Home>
this channel isn't even registered?
00:43
<AryehGregor>
Nope.
01:27
<boblet>
I want “Microdata brings torment and dismay” on a t-shirt
01:29
<boblet>
actually “brings torment and dismay” would work even better. I could then change it to HTML5 or whatever is bringing the most torment at the time
02:34
<foolip>
boblet: please let me know when you have the t-shirts for sale, I will buy!
02:35
<boblet>
foolip: we are both naive in the ways of the internet; http://twitter.com/EvilandLazy/status/9656120722
02:37
<boblet>
someone has a real time search set for “t-shirt” and has a nice little business going on Zazzle
02:37
<boblet>
I didn’t have the heart to reply “yeah, but I’d want one that didn’t look like you made it in 5 min” ;-)
02:39
<boblet>
(at least it wasn’t Comic Sans)
02:51
<boblet>
so is @spellcheck only for forms? seems like it’d be for textarea and input, but it’s applicable to “HTML elements”…
03:01
<foolip>
boblet: isn't that just the DOM attribute? is <p spellcheck> really valid?
03:03
<boblet>
in section-index.html#attributes-0 it says elements = HTML elements, whereas eg placeholder has input; textarea
03:04
<foolip>
huh, ok
03:05
<boblet>
might file a bug on it just in case
03:05
<boblet>
or maybe a FORMAL OBJECTION!! (cue ominous da-da-daaaah!)
03:05
<foolip>
no no, it's Formal Objection
03:06
<boblet>
No one ever expects the Span… err Formal Objection!
03:07
<foolip>
no need for mockery though, I assume that's the way some WGs work, even if it seems unnecessarily formal
03:07
<boblet>
instead of the comfy chair, we can have the huge narcoleptic email thread
03:08
<boblet>
foolip: heh, yes I merely jest. I love process as much as the next guy
03:08
<boblet>
well, I see the need for it
03:09
<foolip>
I wonder if there are other WGs as strange as the HTML WG
03:12
<boblet>
this one has to be the best
03:37
<boblet>
anyone know where in the spec the @id value is precisely defined? it seems from /elements.html#the-id-attribute it now supports non-ascii characters like kanji, and I wanna confirm that
03:41
<miketaylr>
boblet: http://www.whatwg.org/specs/web-apps/current-work/complete.html#concept-id ?
03:42
<boblet>
miketaylr: that’s actually the same text I referred to ;-) it doesn’t say eg what unicode character range(s) are allowed, or what class of content is allowed
03:42
<miketaylr>
oops!
03:42
<miketaylr>
i don't recall reading that (yet?)
03:43
<boblet>
heh, np
03:43
<boblet>
I’d like to know what “at least one character” actually refers to
03:43
<miketaylr>
emoticons, hopefully
03:43
<boblet>
well, yeah, numbers are ok in ids now, so what else? can we use things like @ or £?
03:44
<boblet>
or is it just straight alphanumeric?
05:21
<zcorpan__>
boblet: anything goes basically in id
05:22
<boblet>
zcorpan__: so any unicode character except space or non-unicode characters (whatever they’re called) huh
05:22
<boblet>
rock
06:03
<nessy>
Hixie: the vote on the media a11y stuff will be over the next week, IIUC - then the emails will go to the WGs
08:40
<Philip`>
boblet: I'm not sure if it's explicitly defined anywhere, but I'm fairly sure "character" in HTML5 just means any arbitrary Unicode codepoint
08:40
<Philip`>
Hmm, it does define "The term Unicode character is used to mean a Unicode scalar value (i.e. any Unicode code point that is not a surrogate code point)"
08:40
<Philip`>
though I'm not sure if "character" is meant to be precisely that definition
08:41
<hsivonen>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=9001#c7 is interesting
09:58
<Lachy>
http://arstechnica.com/open-source/news/2010/02/ogg-theora-vs-h264-head-to-head-comparisons.ars
10:26
<Dashiva>
http://www.w3.org/2010/02/rdfa/meetings/2010-02-25#resolution_1
10:27
<Dashiva>
So is that yet another case of non-safe CURIE and URI overlap?
10:29
<Philip`>
Maybe they should be called Dangerous CURIEs
10:31
<Dashiva>
Imagine how much trouble the world would be saved if they had been sensible enough to scrap non-safe CURIEs from the outset
10:38
<jgraham>
A rather small amount, given how many people care?
10:40
<Dashiva>
Right now, yes, but if for some reason the tech takes off...
10:41
<Dashiva>
boblet: I propose "Spanish Objection" as the term for an objection nobody expects (possibly because it was made on a private list)
10:47
<Dashiva>
http://lists.w3.org/Archives/Public/public-html/2010Feb/0861.html
10:47
<Dashiva>
Supposing a browser let you change or override @lang, what would actually happen?
11:03
<Philip`>
Dashiva: Maybe it'd help screen-readers that use lang to guide pronunciation
11:04
<Philip`>
or the browser's spellchecker
11:07
<Dashiva>
Spellchecker is useful, but changing @lang is little different from changing the spellchecker's dictionary directly
11:15
<hsivonen>
Dashiva: it could affect font selection
11:24
Philip`
can't decide whether he likes FontForge or not
11:25
<Philip`>
It seems to work, and has all the functionality I've needed, but its UI has a 90s UNIX feeling and appearance and it keeps popping up weird error messages
11:39
<hsivonen>
https://twitter.com/Malarkey/status/9673991819 seems like HTML5
11:45
<Dashiva>
What is HTML5's equivalent of flex box layout? :)
11:48
<annevk>
Dashiva, <canvas>, doh
11:49
<hsivonen>
sounds like Crockford doesn't write tokenizers
11:49
<hsivonen>
re: goto
11:50
<Dashiva>
annevk: canvas seems to be working fine, the spec is moving along fast, implementations are happening
11:51
<annevk>
oh, in terms of that
11:51
<annevk>
<datagrid>
12:40
<boblet>
Dashiva: heh. but we still need a comfy chair…
13:33
<annevk>
so ISSUE-56 is about process too, in a way?
13:34
<annevk>
i.e. when we change references to IRIbis?
13:34
<annevk>
so weird
13:50
<Philip`>
http://msdn.microsoft.com/en-us/library/ms537512(VS.85).aspx - "<!--[if WindowsVersion 1]><p>You are using Windows Ultimate Edition.</p><![endif]-->"
13:50
<Philip`>
I wonder if any analytics scripts use that yet
13:51
<Philip`>
(though that page can't decide whether it's WindowsVersion or WindowsEdition)
14:03
hsivonen
notes "Detecting Firefox" in community content
14:10
<Philip`>
http://www.giorgiosardo.com/ie8/standards/demo1/7_Selectors.htm - "HTML 5"?
14:10
<Philip`>
(from http://blogs.msdn.com/giorgio/archive/2009/11/29/ie8-and-html-5.aspx)
14:11
<Philip`>
XDR and JSON and Selectors API are now all HTML5 too, it seems
14:12
<zcorpan>
everything cool is html5
14:12
<annevk>
XDR is not cool
14:13
<zcorpan>
ms presumably think it's cool
14:13
<hsivonen>
What's XDR?
14:13
<hsivonen>
ah. Cross-Domain Request
14:14
<hsivonen>
there's also XRD whose coolness is questionable
14:15
<Philip`>
http://tools.ietf.org/html/rfc4506 - XDR is already an IETF standard
14:16
<Philip`>
There ought to be a TLA registry to avoid naming conflicts
14:19
<Dashiva>
There needs to be a registry regtsiry too, to prevent people from creating overlapping registries
14:28
jgraham
wants a registry of all registries that don't include themselves as a member
14:31
<Philip`>
That's easy - just make a registry X such that if someone asks it whether X is registered in it, their request gets bogged down in bureaucratic processes for years and they never get a definitive answer back
14:32
<Philip`>
That way the registry will never have to give an incorrect answer
14:37
<jgraham>
But then how will I destroy the belief that all standards can be derived from the axiomatic proof?
14:45
<Philip`>
I hope you haven't been tricked into believing Gödel's mere "theorem"
14:46
<gsnedders>
Philip`: He's a physicist: he does nothing but believe theorems.
17:02
AryehGregor
wonders why mpilgrim closes off public comments on his blog posts so quickly.
17:03
<Philip`>
To improve the signal-to-noise ratio?
17:08
<AryehGregor>
How does that work? Smart people respond quicker than stupid people?
17:10
<Philip`>
People who respond quicker are more likely to be feed subscribers, and therefore excellent people, rather than random stranges who noticed the post on Reddit
17:11
<Philip`>
Also, people who respond quicker are more likely to be responding to the post, and not responding to other comments and straying far from the original topic
17:11
AryehGregor
has his doubts.
17:12
Philip`
too
17:12
<AryehGregor>
I just got the TeXBook a day or two after his post on the CSS2.1 appendices being in alphabetical order.
17:12
<AryehGregor>
It seems that some others already pointed out that the TeXBook has appendices in alphabetical order.
17:13
<AryehGregor>
But I also wanted to remark, since the comparison had come up, how much more elegant TeX is than CSS. Granted, partly this is because it's Turing-complete, but I suspect it's also because it follows one person's vision rather than accreting over time.
17:13
<AryehGregor>
I've been wondering if anyone's tried implementing a TeX-like line-breaking/hyphenation algorithm in HTML, actually. I'm sure someone has. I wonder how it worked out.
17:14
<AryehGregor>
HTML is really unsuitable for high-quality typography, sadly.
17:19
<TabAtkins>
AryehGregor: That exact topic has come up recently. Yes, there is definite interest in implementing Tex-like line-breaking and hyphenation, and specifying such precisely in a CSS module.
17:19
<AryehGregor>
That would be neat.
17:19
<TabAtkins>
And yes, someone has done so already *in Javascript and canvas*.
17:19
<AryehGregor>
Awesome. Link?
17:19
<TabAtkins>
I"ll have to find it. one sec.
17:20
<JonathanNeal>
Hey all.
17:20
<TabAtkins>
AryehGregor: http://ajaxian.com/archives/tex-line-breaking-algorithm-in-javascript
17:20
<JonathanNeal>
TabAtkins, http://sandbox.thewikies.com/html5-video/
17:21
<AryehGregor>
Seems like no hyphenation, though. Sad. The TeX hyphenation algorithm looks efficient enough for web use.
17:22
<TabAtkins>
Oh, right, that's just line-breaking. Yeah, hyphenation shouldn't be too bad either, from what I understand.
17:22
<AryehGregor>
I assume it does French spacing, since you can't do "~" or "\ " or such.
17:22
<AryehGregor>
Er, well, I guess you can.
17:22
<AryehGregor>
&nbsp; is ~.
17:22
<TabAtkins>
JonathanNeal: Strange implementation of your html5 shim. Did that come from performance tests or something?
17:24
<JonathanNeal>
TabAtkins, you mean with the while loop? It's just the fastest, smallest way to do it.
17:24
<JonathanNeal>
In a while loop, you get the best performance by sticking the math before the variable, like --i
17:24
<TabAtkins>
That's not the smallest way to do it. May be fastest, shrug.
17:25
<JonathanNeal>
I bet it's one of the smallest ways to do it, minus all the generous spaces and variable names I gave it.
17:25
<TabAtkins>
"abbr,article,etc".split(',');for(var i=0;i<e.length;i++){document.createElement(e[i])} is what I use.
17:26
<TabAtkins>
Sorry, clipped that wrong. Assume a "var e=" beginning that line.
17:26
<TabAtkins>
And I'm certain you can do it smaller by using Array functions.
17:27
<AryehGregor>
Does JS have map()? :)
17:27
<TabAtkins>
AryehGregor: Yes.
17:28
<JonathanNeal>
TabAtkins, by one space... var e='abbr article etc'.split(' ');i=e.length+1;while(--i){document.createElement(elems[i])}
17:28
<AryehGregor>
Does this work? "abbr,article,etc".split(',').map(document.createElement)
17:28
<JonathanNeal>
But the while loop is much better for performance.
17:28
<AryehGregor>
The performance here is negligible anyway.
17:28
<AryehGregor>
Premature optimization is the root of all evil.
17:29
AryehGregor
doubts .map() will work in IE, though
17:29
<TabAtkins>
JonathanNeal: Actually, you can probably shave off more if you correctly reference e in the createElement call. ^_^
17:29
<AryehGregor>
Kind of defeats the point.
17:30
<JonathanNeal>
TabAtkins, yes.
17:30
<JonathanNeal>
var e="abbr,article,etc".split(',');for(var i=0;i<e.length;i++){document.createElement(e[i])} VS
17:30
<JonathanNeal>
var e='abbr article etc'.split(' ');i=e.length+1;while(--i){document.createElement(e[i])}
17:30
<TabAtkins>
Really, it was the storing of the createElement result that confused me.
17:31
<JonathanNeal>
TabAtkins, that was unnecessary, in a previous implementation I was nulling out that value after the loop (for memory purposes, but I think it's negliable.
17:31
<paul_irish>
JonathanNeal: /*@cc_on'DetailsCanvasMeterHeader {....}'.replace(/.[a-z]+/g,function(n){document.createElement(n)})@*/
17:31
<TabAtkins>
Anyone want to try and play http://igofigure.com/movie/dhsab_p1/ in IE8? I'm getting a seemingly-random subset of people not able to play videos off of my site in IE8.
17:32
<JonathanNeal>
paul_irish uses this in modernizr, and would know a lot about it.
17:32
<JonathanNeal>
TabAtkins ^^
17:33
<paul_irish>
TabAtkins: here's the most recent round of hacks on it: http://gist.github.com/296128
17:33
<JonathanNeal>
Back they placed the shiv in modernizr, I just use it as a reference whenever I need to use it, and that's where the var comes from. paul_irish why do you null it out?
17:34
<JonathanNeal>
paul_irish, what is the performance impact between these?
17:35
<paul_irish>
a reverse while loop vs a str.replace iterator? hmmm.. i dont think anyone has tested. but its all <20ms and you know.. happens once per page load
17:36
<JonathanNeal>
I'll run it through my tester.
17:36
<JonathanNeal>
it = them
18:24
<JonathanNeal>
TabAtkins, I never heard if you got what you needed about <video>
18:28
<TabAtkins>
Nah, not yet. For some reason all flash-based players I've tried are failing when trying to play videos off of either of the servers I control in IE8, for certain people.
18:32
<TabAtkins>
This isn't helped by the fact that the IE8 on my work comp is refusing to recognize that I even have a valid version of flash, while the IE8 on my laptop plays everything perfectly fine.
18:39
Dashiva
fights with zip encodings
18:40
<Philip`>
I thought zips don't have encodings
18:40
<Dashiva>
They just don't tell you which one it is
18:41
<Dashiva>
The extraction assumes windows-1252 (or similar) and the content is actually shift_jis. But the result is madness.
18:42
<Dashiva>
Most of the bytes in the raw filename are in the 0x80 - 0x9F range, but the result filename has characters in 0xC0 through 0xF9
18:50
<TabAtkins>
JonathanNeal: Could you try visiting http://www.igofigure.com/flowplayer/example ?
18:50
<TabAtkins>
In ie8?
18:50
<JonathanNeal>
Sure thing.
18:50
<TabAtkins>
And tell me if it breaks or not? By 'break', I mean play a second or less of video, then hang the IE process.
18:51
<JonathanNeal>
Looks good to me, loved the Sing a long blog
18:51
<TabAtkins>
Anyone else want to try it out in IE8?
18:51
<TabAtkins>
Now I'm wondering if it's something to do with our VPN. I dunno if any customers have reported problems yet.
18:58
<mpilgrim>
AryehGregor: i close comments (a) when the discussion dies down, (b) when a wave of low-quality comments hits, or (c) after 7 days, whichever comes first
19:11
<miketaylr>
TabAtkins: i also see doogie howser from IE8 from that link
19:11
<TabAtkins>
Thanks, miketaylr.
19:11
<miketaylr>
np
19:11
<Dashiva>
0x83 0x80 x83 becomes 0xE2 0xC7 0xE2... how on earth
19:27
<AryehGregor>
"When the user presses the key combination corresponding to the assigned access key for an element, if the element defines a command, and the command's Hidden State facet is false (visible), and the command's Disabled State facet is also false (enabled), then the user agent must trigger the Action of the command."
19:27
<AryehGregor>
What does this imply if you have duplicate accesskeys? They all get triggered?
19:27
<AryehGregor>
That doesn't seem to be what anyone implements (for good reason).
19:36
<Dashiva>
AryehGregor: That's probably meant to be covered by the optional behaviors in steps 3.3 and 4
19:44
<Philip`>
Hmm, seems nobody has bothered implementing textBaseline properly yet
19:45
<Philip`>
Opera and Chromium use top/bottom of the bounding box, not top/bottom of the em square
19:45
<Philip`>
(Firefox gets that right)
19:46
<Philip`>
None bother handling hanging and ideographic baselines
19:47
<AryehGregor>
Dashiva, interop would be nice here. Firefox 3.5 or so changed behavior so that if there are multiple elements with an accesskey, hitting the accesskey toggles between them instead of activating the first.
19:48
<AryehGregor>
It broke some stuff. Rather inconvenient from an authoring perspective.
19:48
<AryehGregor>
(it's still not fixed in vBulletin)
19:48
<Dashiva>
Woo, finally cracked it
19:49
<Dashiva>
It wasn't reading the shift_jis as windows-1252 or anything like it, it was using IBM437
19:49
<Dashiva>
(Why?)
19:53
<Philip`>
Dashiva: Looks like that's the one that lets you draw cools borders and gradients in plain text files
19:53
<Philip`>
I used it all the time for graphics in my QuickBASIC programs
19:59
<Dashiva>
Well, now I can fix my Japanese files at long last
20:00
<Philip`>
"// spec says to not throw error for invalid arg, but do it anyway"
20:00
<Philip`>
Good to see they're striving for conformance there
20:04
AryehGregor
tries to figure out the hg equivalent of git show for revisions
20:05
<AryehGregor>
hg log seems to work.
20:06
<Philip`>
If you want to see the revision as a diff, then I think it's hg diff -c 1234
20:06
Philip`
has no idea what git show is
20:07
<AryehGregor>
git show displays a git object. For revisions, it typically gives you the commit message, date, diff, etc.
20:07
<AryehGregor>
https://bugzilla.mozilla.org/show_bug.cgi?id=436904#c4
20:08
<Philip`>
Ah, hg log -pr 1234 sounds like it, then
20:21
<AryehGregor>
That's what I wound up with.
20:28
<Dashiva>
"Can we please add to the Web Sockets API methods getResponseHeader() and getAllResponseHeaders() as the per the XMLHttpRequest API."
20:29
<Dashiva>
On one hand, I can see how this might be useful, but on the other hand you can just as easily send that metadata as the first message from the server
20:57
<Philip`>
I thought Web Sockets wasn't HTTP
20:57
<Philip`>
so it doesn't have headers
20:59
<Dashiva>
Yes. But there are plenty of people over at IETF who want to turn web sockets into HTTP 2; also, there would be HTTP headers if a connection via HTTP upgrade failed
21:01
<AryehGregor>
Why don't they make HTTP 2 if they want HTTP 2?
21:04
<Dashiva>
I think that's what the hybi group is for
21:48
<JonathanNeal>
HTML5 shiv further tests @ http://pastie.org/844872
21:51
<JonathanNeal>
I didn't even remember to mention that these are tests done in IE6.
21:54
<Philip`>
JonathanNeal: low is higher than high, and avg is higher than both?
21:54
<JonathanNeal>
I must have written things wrong then.
21:55
<JonathanNeal>
I guess I can throw out these tests.
21:55
<JonathanNeal>
My last tests were about the same though.
21:56
<JonathanNeal>
Actually ... I flipped the highs and lows, that's all.
22:04
<JonathanNeal>
Philip`, updated @ http://pastie.org/844911
22:06
<Philip`>
JonathanNeal: Your average is still higher than your high
22:06
<JonathanNeal>
**** it, I'll redo all the tests.
23:00
<JonathanNeal>
I've redone the tests, Philip`
23:03
<JonathanNeal>
http://pastie.org/844999
23:05
<JonathanNeal>
These tests still are ****ed, I have no idea why shortest time can be longer than the average time.
23:06
<JonathanNeal>
I give up, crying, taking my ball and going home :-O
23:08
<JonathanNeal>
This is what I have @ http://pastie.org/845011
23:08
<JonathanNeal>
That's the test.