00:04
<ttepasse>
http://24ways.org/2009/incite-a-riot
00:13
<JonathanNeal>
ttepasse, fight for your right to cite!
00:16
<ttepasse>
:)
01:43
<hsivonen>
Does Kaazing have a Web Socket server already?
01:48
<MikeSmith>
hsivonen: yes
01:49
<MikeSmith>
not a separate server, but integrated into their gateway product
01:49
<MikeSmith>
iirc
01:57
<franksalim>
hsivonen, yes
02:00
MikeSmith
didn't know franksalim was on this channel
02:01
<MikeSmith>
hi franksalim
02:01
<MikeSmith>
have you had time yet to look at SPDY much?
02:01
<franksalim>
MikeSmith, hello
02:03
<franksalim>
MikeSmith, not in great detail
02:05
<MikeSmith>
franksalim: I'm just wondering how overlap there will end up being in practice between use of SPDY protocol and use of Web Sockets protocol
02:06
<MikeSmith>
wondering whether the Web Sockets client API could also be used with SPDY as the network protocol
02:07
<franksalim>
the WebSocket api could certainly wrap the SPDY protocol in some way
02:09
<franksalim>
so far SPDY hasn't been exposed to JavaScript in any way beyond how HTTP has been exposed, AFAIK
02:09
<franksalim>
XHR and requests for resources (that may be cached)
05:04
<MikeSmith>
AryehGregor: http://www.alvestrand.no/pipermail/ietf-languages/2009-December/009754.html
05:05
<MikeSmith>
about the "eml" tag
05:05
<MikeSmith>
we'll see what response that gets
05:05
<MikeSmith>
but in the mean time, I reckon we can just add special-case handling for eml in the language-tag code
05:06
<MikeSmith>
and remove it later if/when "eml" does actually get added to the registry
09:56
<Hixie>
sorry about sporadic downtime overnight, dreamhost were downing some surgery on my web host config to fix some problems on some of my other domains
11:24
<annevk2>
"If getContext() is subsequently called with any other string, the WebGLRenderingContext shall be detached from its Canvas object (the canvas attribute will be null). API calls may still be made on the WebGLRenderingContext object and readPixels() may be used to obtain the contents of the drawing buffer. But the contents shall no longer be shown on the page."
11:24
<annevk2>
hmm
11:25
<annevk2>
it doesn't say it if can become attached again afterwards
11:25
<Philip`>
Also doesn't say what happens if you have a reference to a 2D context then get a WebGL context then use the 2D context
11:28
<Philip`>
Maybe HTML5 ought to define a mechanism for new contexts to be defined that interact with other contexts in various ways (e.g. one context being a vendor-specific extension of another context, or one being exclusive to all other contexts, etc)
11:28
<annevk2>
or it should say that exclusive contexts are verboten
11:28
<Philip`>
because otherwise it's quite possible that multiple specs will define contradictory behaviour
11:29
<annevk2>
yuhhuh
11:30
<Philip`>
How would it work with non-exclusive contents?
11:30
<Philip`>
If you draw something with WebGL then draw something else with 2D on the same canvas, what happens?
11:30
<Philip`>
and then when you call toDataURL, what happens?
11:31
<annevk2>
it should just draw on top
11:31
<annevk2>
i'd guess
11:31
<annevk2>
and share the backing store
11:31
<Philip`>
HTML5 seems to assume there is a single bitmap that's part of the canvas state, but I think that's incompatible with efficient WebGL implementations
11:31
<annevk2>
so everything draws on the same bitmap
11:32
<Philip`>
OpenGL doesn't draw onto bitmaps in system memory, unless you manually copy it back from the video card (which is pretty slow)
11:32
<annevk2>
you could implement 2d in WebGL
11:34
<Philip`>
I think that's kind of what's happening with some Direct2D browser experiments, but that's Windows-only and it's apparently really hard if you just have the raw 3D APIs
11:34
jgraham
seems to recall we had this conversation before :)
11:35
<Philip`>
jgraham: I don't recall coming to any conclusion before, sadly
11:35
<jgraham>
No, we didn't
11:35
<Philip`>
It would be nice to have one at some point
11:35
<Philip`>
before it gets shipped too much
11:35
<jgraham>
IIRC we decided it would be nice if everything could happen on the graphics card but probably impractical in the short term
11:36
<annevk2>
Hixie is always big on cracks between the specs
11:36
<annevk2>
he should take this on!
11:36
<Hixie>
file a bug
11:36
<annevk2>
mkay
11:36
<Hixie>
i wonder how webgl implementations handle the canvas being animated with a 3d css transform
11:37
<Hixie>
while a <video> plays on top...
11:37
<annevk2>
on a phone with hardware accelerated video
11:37
<annevk2>
...boom...
11:39
<annevk2>
bugs
11:39
<annevk2>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=8476
11:40
<jgraham>
In the future people will carry their phone in their pocket and a huge backpack containing a graphics card and a water cooling system
11:41
Philip`
wonders if jgraham has missed the trend for electronic miniaturisation over the past half a century
11:44
<jgraham>
Philip`: :) But presumably power consumption and heat generation are still rather large problems even with minature components
11:52
<Philip`>
jgraham: Things like PSPs and iPhones seem to do decent 3D quite happily, with similar capabilities to full-size consoles from ten years ago
11:53
<Philip`>
and I assume they must cope with the power and heat requirements adequately
11:53
<Philip`>
and I don't see any reason for the trend to stop
11:54
<workmad3>
most high-end phones have mobile versions of reasonable graphics chips nowadays
11:55
<workmad3>
and I'd say the capabilities are closer to pc graphics from about 5 years ago, but with less available memory
11:56
<Philip`>
workmad3: So they could run Half Life 2? :-)
11:57
<workmad3>
Philip`: potentially, if you reduced the graphics quality and map size considerably to fit within the memory available ;)
11:59
<workmad3>
at least in 3d graphics capabilities... there's a lot more to HL2 than that though :)
12:00
<Philip`>
I probably shouldn't have said "capabilities" - it seems like they support much more modern features and APIs, but performance is further behind
12:00
<Philip`>
so you have to use the capabilities more sparingly
12:00
<workmad3>
Philip`: that's probably about right yeah :)
12:01
<Philip`>
(I know basically nothing about this area though, so mostly I'm just guessing :-) )
12:01
<workmad3>
they don't run quite as fast, and have nowhere near as much memory, but the feature-set of the chips in use is pretty modern and allows for more sophisticated stuff than consoles from 10 years ago did
12:01
<Philip`>
At least they don't have to bother with HD resolution output
12:01
<workmad3>
that said, you don't need to render in anywhere near as high resolution or store as much geometry :)
12:02
<workmad3>
I think phone screens nowadays are 'QHD' (quarter HD) nowadays and clock in at 640x360 pixels :)
12:03
<workmad3>
but raw graphics power isn't really the problem on phones... it's control and all the other stuff that's needed
12:04
<Philip`>
That sounds more like ninth HD, unless you cheat and consider lower-than-standard-cheap-laptop-display to be HD :-)
12:05
<Philip`>
They should call it "middling definition", perhaps
12:05
<workmad3>
Philip`: a standard HD res is 1280 * 720
12:06
<workmad3>
(I think that's the res of 1080p)
12:06
<Philip`>
1080p is 1920x1080, hence the "1080" in the name, unless I'm horribly confused
12:06
<workmad3>
so taking the HD resolutions as those supported by HD tvs, then it is a quarter of that :)
12:07
<Philip`>
(and 1280x720 is 720p)
12:07
<workmad3>
hmm... I may be thinking of 720p then
12:07
<workmad3>
still, it's classed as a HD resolution, so the phone manufacturers went with a quarter of it as 'QHD' :)
12:08
<peroo>
actually qhd is 960x540, one quarter of 1920x1080 :)
12:09
<peroo>
640x360 is nHD apparantly, whatever that means
12:10
Philip`
guesses it must have been a marketing person who decided an acronym containing the word "high" was appropriate for sub-VGA resolutions :-p
12:10
<jgraham>
"Not High Definition", perhaps
12:11
<workmad3>
heh, I'm going by what the phones I worked on that had that resolution called it :)
12:13
<workmad3>
still, AFAIK, the graphics chips available to phones are capable of the raw power required to do all the effects that modern consoles are capable of, the restricting points are things like the physics engines and other processing that's required to get the data ready for the screen in a game
13:33
<Hixie>
i wonder how to add features to html5 without pissing off the htmlwg
13:37
<daedb>
Hixie: I don't think that's possible (unless it's XMLy or includes namespaces), but do it anyway :p
13:38
<Hixie>
well there are options like making the new features only appear in the whatwg version
13:38
<annevk2>
Hixie, what feature?
13:38
<Hixie>
which would piss off the people who get pissed off regardless of what i do, but wouldn't piss off the more sane people
13:39
<jgraham>
I would be pissed off if we introduced namespaces
13:39
<Hixie>
annevk2: e.g. a widget to trigger the display of UA UI to allow the user to select a video source
13:39
<Philip`>
What about the people who think it'll be really confusing if there's two substantially different HTML5 specs?
13:39
<Hixie>
Philip`: are they likely to cause me more or less trouble than the people who would complain at me adding features as we are trying to get to LC?
13:40
<annevk2>
Hixie, at what point do we declare HTML5 feature complete?
13:40
<Hixie>
we declared HTML_5_ feature complete about 2 years ago
13:40
<Hixie>
well that's not quite accurate
13:41
<Hixie>
we announced that HTML5 feedback shouldn't ask for new features about 2 years ago, and then announced LC (feature freeze) about 2 months ago
13:41
<Hixie>
but I'm talking about post-HTML5
13:41
<Hixie>
HTML6, or HTMLCurrent, or whatever
13:41
<annevk2>
ah ok
13:41
<annevk2>
so are you going to version the differences somehow?
13:42
<Hixie>
how do you mean
13:42
<annevk2>
<!--HTML6--> or some such
13:42
<Hixie>
you mean "are the new features going to appear in the w3c html5 draft"?
13:43
<annevk2>
just wondering about the WHATWG Last Call draft for the moment
13:43
<jgraham>
any draft that says HTML5 at the top I would have thought
13:43
<Hixie>
yeah that's the question, how to do it
13:43
<Hixie>
we could rename the whatwg draft to just "HTML", i guess
13:43
jgraham
thinks it would be fine to make a HTML.next draft for unstable stuff
13:43
<Hixie>
or have a separate view of it somehow
13:44
<annevk2>
I think we should maintain a HTML5 copy
13:44
<annevk2>
preferably on the same URL
13:44
<Hixie>
maybe have specs/web-apps/html5/ include just the html5 stuff, and specs/web-apps/current-work/ for the new stuff
13:44
<Hixie>
(and specs/web-apps/current-work/complete.html for everything but sql)
13:44
<annevk2>
current-work/ does suggest latest, so yeah, that seems ok
13:45
<Hixie>
btw if any of you want your specs merged into complete.html, we can do that
13:45
<Hixie>
e.g. to get the XHR cross-refs working
13:45
<annevk2>
people will find it somewhat confusing if we call it "HTML" though I suspect
13:45
<Hixie>
we could call it HTML6
13:45
<Hixie>
or 5.1
13:45
<Hixie>
or something
13:45
<annevk2>
Hixie, hmm yeah, I hope someone fixes anolis
13:46
<Hixie>
(though i really want to drop this versioning crap)
13:46
<annevk2>
HTML (formerly HTML5)
13:46
<annevk2>
for at least half a year or so
13:46
<annevk2>
and then we drop (formerly HTML5)
13:47
<Hixie>
or "HTML (including HTML5)", to indicate it's more than HTML5
13:47
<annevk2>
yeah, works too
13:47
<Hixie>
and then drop the paranthetical
13:47
<gsnedders>
annevk2: patches welcome.
13:48
<jgraham>
Call it HTML unstable or something
13:48
<Hixie>
(s/a/e/magic)
13:48
<gsnedders>
HTML Beta
13:48
<jcranmer>
HTML 5 rc i
13:48
<jgraham>
HTML Labs
13:48
<Hixie>
well the idea is to just have a single name that is always the top of the tree
13:48
<Hixie>
so "unstable" or "beta" is kinda wrong
13:48
<gsnedders>
jcranmer: So it's not a rc?
13:48
<Hixie>
i mean, <p> isn't going anywhere
13:48
<jcranmer>
gsnedders: it's an imaginary one
13:49
<Hixie>
i wish othermaciej was here, i'm curious as to his opinion on this
13:49
<jcranmer>
more complex than the real HTMLs
13:49
<jgraham>
Hixie: Well you always have an unstable and periodically make numbered stable releases, Debian style
13:49
<jgraham>
(or call it HTML Mainline which sounds cooler, like a VU song)
13:49
<Hixie>
that's what i want to move away from
13:49
<annevk2>
I'd rather be forever in release-state
13:49
<Hixie>
yeah
13:50
<Hixie>
bbiab, gotta get ready.
13:50
<jgraham>
Well so I would I but good luck getting the rest of the world to agree
13:50
<annevk2>
but until we convince the W3C it's prolly safer to have markers for the new stuff
13:52
<annevk2>
no need to convince the whole world
13:53
<annevk2>
just the patent lawyers
13:53
<annevk2>
which is prolly harder
14:19
<Hixie>
not really any need to convince those either so long as the w3c is publishing snapshots
14:19
<Hixie>
which i doubt they plan to stop doing
14:21
<annevk2>
legal effect of the patent policy only applies to W3C Recommendations
14:21
<annevk2>
I believe
14:26
<Hixie>
right, so so long as they keep publishing everything the whatwg does, everything is covered
14:26
<Hixie>
doesn't really matter when the publication happens, since the policy doesn't really kick in until REC anyway, as I understand it
14:26
<annevk2>
the issue I'm thinking about is that W3C Recommendation status is incompatible with forever-trunk
14:27
<Hixie>
i'm saying that what we do is forever-trunk at the whatwg, and continue working with the w3c to publish RECs every now and then
14:27
<Hixie>
as we've basically been doing
14:27
<Hixie>
the only problem is that the htmlwg is kinda stopping progress on html because of it trying to go to LC
14:28
<Philip`>
If the patent policy only kicks in at REC, I guess that means everyone is free to sue each other until 2022?
14:28
<annevk2>
that works though that also means you need to keep the REC markers in the document (or trunk markers)
14:28
<annevk2>
if you want to get rid of those too (I like!) then the convincing thingie above applies
14:29
<annevk2>
Philip`, pretty much
14:29
<annevk2>
Philip`, but I'm not a lawyer
14:29
<Hixie>
REC markers?
14:31
<annevk2>
<!--REC-ends-->
14:31
<annevk2>
or <!--HTML6--> or whatever
14:31
<jgraham>
That sounds... bad
14:32
<Hixie>
so long as one source document is generating more than one published document, i don't see how we could stop having some kind of markers
14:33
<jgraham>
It seems nightmarish to generate multiple versions of the same spec from the same source file
14:34
<Hixie>
i take it you haven't read http://wiki.whatwg.org/wiki/FAQ#What_are_the_various_versions_of_the_spec.3F recently
14:34
<annevk2>
Hixie, the difference is that this is a version marker, but maybe that's not much of a difference
14:34
<MattCampbell>
What version of Python does html5lib require?
14:34
<Hixie>
annevk2: i don't see how it's any different than html5 vs complete.html, or the whatwg microdata vocab spec vs w3c's three microdata vocab specs
14:34
<jgraham>
Hixie: This is quite different
14:34
<annevk2>
Hixie, and hopefully we do not let it scale to IE8 proportions
14:35
<jgraham>
I would have thought
14:35
<Hixie>
jgraham: how so?
14:35
<jgraham>
MattCampbell: 2.4 I guess
14:35
<Hixie>
annevk2: well these are just supersets of each other right, i mean, it's not like we're changing anything
14:35
<jgraham>
Hixie: Because I would expect changes from new versions to be much more invasive than changes that just involve splitting along broad feature lines
14:36
<MattCampbell>
darn, guess I'll have to compile Python from source to use html5lib on RHEL5 (which has Python 2.4)
14:36
<Hixie>
jgraham: i would expect editorial changes to just be made to all versions
14:36
<jgraham>
MattCampbell: (it might work under 2.3 but no promises)
14:36
<annevk2>
Hixie, sure, but it could get tricky nonetheless
14:36
<Hixie>
annevk2: well if it is too tricky we can always give up and do something else :-)
14:37
<annevk2>
Hixie, e.g. further overloading of existing methods, etc.
14:37
<gsnedders>
Hixie: Web DOM Core?
14:37
<annevk2>
diving and hiking
14:38
<jgraham>
Hixie: Yes but say you make a change that adds new steps to an existing algorithm, makes a few changes to conformance criteria and adds a few paragraphs of new material
14:38
<jgraham>
Then you end up with a bunch of ifdefs all over the place to try and seperate out the HTML6 stuff from the HTML5 stuff
14:38
<Hixie>
jgraham: we have a bunch of ifdefs already to do complete.html
14:39
<Hixie>
jgraham: it's not that big a problem
14:39
<jgraham>
Hixie: On such a fine-grained level? It seems like it should be a problem :)
14:40
<Hixie>
search for "END complete" in the source file
14:41
<annevk2>
Hixie, anyway, I don't have a better solution other than changing the way the W3C REC-track works and what you have works much better than what I do for XMLHttpRequest I think
14:41
<annevk2>
(which is trying to keep two documents in sync)
14:41
<Hixie>
on the long term i definitely think we should change the way the w3c REC track works, but that needs far more effort than i am willing to put out
15:55
<pesla\work>
sigh, firefox always crashes on the html5 specs -.-
15:56
<Hixie>
pesla\work: use http://whatwg.org/html5
15:56
<Hixie>
pesla\work: that works in firefox
15:56
<annevk2>
pesla\work, wrong channel to complain
15:56
<pesla\work>
Hixie: sweet :)
15:57
<Hixie>
pesla\work: if you need the one-page version, try http://www.whatwg.org/specs/web-apps/current-work/?slow-browser
16:00
<pesla\work>
annevk2: Man, I just needed someone to listen to my grief! ;)
16:00
<pesla\work>
Hixie: the /html5 method works perfectly :)
16:05
<Philip`>
http://www.whatwg.org/specs/web-apps/current-work/multipage/urls.html#htmlcollection-0
16:05
<Philip`>
"nthnthnth(&*@%WN@%HM)%25"
16:06
<Philip`>
Hixie: I'm not sure that's valid IDL
16:06
<Hixie>
yeah, trying to fix that. somehow my script thought i'd fixed it (which i had in the source) and so was refusing to regen the files
16:06
<Hixie>
ironically, it was there to test that the new webidl check in the pipeline was working
16:06
<Hixie>
(i've integrated dom's service into my update script)
16:07
<Philip`>
Ah
16:09
<annevk2>
pesla\work, just trying to find you a long term solution :p
18:14
<Hixie>
ok well i tried doing what we were talking about earlier
18:14
<Hixie>
we'll see how it goes
18:22
<hober>
Hixie: it seems like whatwg.org/html5 should redirect to the HTML5 draft
18:22
<hober>
maybe we could mint a whatwg.og/html redirect to current-work
18:23
<Hixie>
added /html
18:23
<Hixie>
i think people would find the full html spec more useful than the html5 spec
19:05
<jwalden>
annevk2: actually, I was thinking old-school, Sinatra; never even heard of MGS before :-)
21:18
<smaug>
"Your message to whatwg awaits moderator approval"
21:18
<smaug>
um
21:19
<smaug>
who is the moderator?
21:34
<Philip`>
smaug: There is no moderator
21:34
<Philip`>
smaug: so your message will never be approved
21:34
<Philip`>
smaug: You have to be subscribed before posting
21:36
<smaug>
Philip`: I'm subscribed
21:36
<smaug>
Philip`: the problem is that the email has apparently too many people CC'ed
21:36
<smaug>
(not that many, really)
21:37
<Philip`>
smaug: Oh, okay
21:37
<Philip`>
smaug: CC less :-)
21:39
<smaug>
Philip`: actually someone else CC'ed already so many people that couldn't really add new ones :)