00:11
<Hixie>
othermaciej: i think we need a state that is neither "Rejected" nor "Accepted" but which is "I don't know what you want"
00:12
<Hixie>
othermaciej: because i have the feeling that some of these bugs that i'm marking "Rejected" due to unclear requests are going to just be raised as issues, and will waste a lot of the WG's time, where if the request had just been made clear in the bug, I maybe would have been fine with making the change
00:16
<othermaciej>
Hixie: I suppose marking the bug NEEDSINFO and asking for more info might be too subtle - I think it's fine to have a "Did Not Understand Request" state or whatever
00:17
<Hixie>
ok, i'll change to "Did Not Understand Request" instead of "Rejected" when marking the bug NEEDSINFO
00:18
<Hixie>
thanks
00:18
<othermaciej>
you could even directly encourage the originator to reopen the bug with additional info in that case if they can expand on their request
00:19
<Hixie>
i do do that, in the Rationale field
00:19
<othermaciej>
sounds good
00:51
<Hixie>
othermaciej: http://www.w3.org/Bugs/Public/show_bug.cgi?id=8210 - do you think that should be a drawFocusRect(x,y,w,h) call or a drawFocusPath() call that uses the current path?
02:12
<roc>
If we start offering drawFocusRing, aren't we going to get requests for all kinds of native look-and-feel drawing in canvas?
02:20
<Hixie>
roc: possibly
02:20
<Hixie>
roc: if that's a problem, i'm certainly open to other suggestions to solve the focus ring problem
02:21
<roc>
it's not clear what problem the bug reporter is trying to solve
02:28
<Hixie>
http://lists.w3.org/Archives/Public/public-canvas-api/2009OctDec/thread.html#msg27
02:28
<Hixie>
basically how do you handle users who have magnifiers that follow the caret and system focus
02:28
<Hixie>
or who have the need for high-contrast focus rings
02:53
<gsnedders>
Since when was it 4am?
02:54
<gsnedders>
Oh, yeah. My computer is still on Europe/Stockholm, and I'm in Europe/London.
03:17
<annevk>
I cannot believe I'm awake
03:32
<annevk>
Hixie, "in optional canDrawCustom" does not specify a type
04:34
jwalden
snickers at the example in http://html5.org/tools/web-apps-tracker?from=4461&to=4462
05:04
<annevk>
so does Python in CGI mode suck because CGI sucks so much?
05:04
<annevk>
if so, why hasn't that been replaced over the years?
05:12
<Hixie>
annevk: oops
05:21
<Hixie>
gsnedders: http://www.w3.org/Bugs/Public/show_bug.cgi?id=8232 .... what?
05:52
<annevk>
I like how bz pointed out that browsers do byte-inflation rather than fiddling with ISO-8859-1
05:52
<annevk>
It is more or less the same of course, but it makes it seem much more simple :)
06:11
<Hixie>
gsnedders: nm
06:33
<Hixie>
annevk: it wouldn't be byte inflation when it's not inflating :-P
06:33
<Hixie>
it'd be byte stripping or something
06:33
<Hixie>
pretty bogus to send just the lower byte of a two byte codepoint, though
06:41
<annevk>
Hixie, ah, doh, byte-deflation then
06:41
<annevk>
I'd personally be happy with switching HTTP to use UTF-8 everywhere, break a few servers, and patch it all up
06:42
<annevk>
but that's not going to happen
06:43
<annevk>
and fiddling with the lower byte (thanks for that, didn't know the term I had to use yet) seems the easiest solution...
06:43
<annevk>
at some point we'll overload this method with ByteArray or whatever we end up calling it and encourage people to use that I guess
06:49
<Hixie>
i agree that just switching to UTF-8 is the right solution
06:50
<annevk>
do you know anything about Chrome's HTTP implementation?
06:50
<Hixie>
not off-hand
06:51
<annevk>
it seems to actually use UTF-8 at least here
06:51
<annevk>
if they use it everywhere...
06:51
<annevk>
then maybe there is a way
07:53
<annevk>
hmm, from looking at http://src.chromium.org/viewvc/chrome/trunk/src/net/http/ I cannot find code that deals with the ISO-8859-1 thingie...
07:54
<annevk>
though for parsing location headers it does call EscapeNonASCII saying something about a compat issue...
07:54
<annevk>
makes me wonder what other browsers do there
07:59
<annevk>
Hixie, progressBar.getElementsByTagName('span').textContent is also buggy
07:59
<annevk>
Hixie, maybe progressBar.firstChild.textContent
07:59
<annevk>
or [0] of course
08:00
<Hixie>
oops
08:04
<annevk>
arguably <code title="">-//W3C//DTD&nbsp;HTML&nbsp;4.0//EN</code> is somewhat problematic as it does not work if I copy and paste that
08:05
<annevk>
might be better to use actual spaces and have a class="token" with .token { white-space:pre } or some such
08:06
<Hixie>
browsers copy U+0020, actually
08:07
<othermaciej>
bug list is down to 178 I see
08:07
<othermaciej>
that's good progress
08:08
<Hixie>
it would be going faster if people didn't decide today was the day to file a bazillion bugs :-P
08:08
<annevk>
mkay
08:08
<annevk>
are public ids case-sensitive in SGML?
08:08
<annevk>
surprises me
08:11
<annevk>
looks like XMLHttpRequest will need at least one more Last Call
08:12
<nessy>
OMG - tell me I'm not the only person on this planet that has 13 different browsers installed on one computer
08:12
<nessy>
(admittedly in 3 different OSes)
08:13
<hsivonen>
annevk: public IDs are opaque strings in SGML
08:13
<hsivonen>
annevk: I'm pretty sure they don't get case-folded in SGML
08:14
<hsivonen>
nessy: I have that many on just one OS plus the ones on the other OSs
08:15
<nessy>
uff, good - I thought I was going mental
08:15
<nessy>
I'm just preparing demos for a html5 video tutorial
08:16
<hsivonen>
nessy: what's the status of the Theora ActiveX project?
08:16
<othermaciej>
do I get to count Safari 4.0.4, my WebKit nightly, my release build of Safari trunk and my debug build of Safari trunk as 4 different browsers?
08:16
<hsivonen>
othermaciej: I counted different builds from one codebase as separate
08:16
<nessy>
hsivonen: unfortunately not much progress yet - needs somebody to fund the poor soul - he has a main job, but would be prepared to reduce to do the thing
08:17
<hsivonen>
nessy: I see
08:17
<nessy>
othermaciej: I did count 3 different FF builds, so I guess, yes
08:17
<othermaciej>
Hixie: indeed I see 55 bugs resolved so far this month, but there has been a fair bit of incoming
08:17
<nessy>
actually, no, I only counted different browsers, not different versions - otherwise that'd be 3 more
08:18
<nessy>
hsivonen: we have FOMS coming up in 2 weeks and he will be there, so we'll see how far we can push it
08:18
<annevk>
hsivonen, kk
08:21
<othermaciej>
I have Firefox, Chrome, Opera and Safari, counting only distinct browsers
08:21
<hsivonen>
nessy: is the current recommended thing for IE support to use cortado as a fallback for <video>?
08:21
<othermaciej>
I have not bothered lately with Camino, OmniWeb, iCab or anything else
08:21
<hsivonen>
othermaciej: no IEs on VMs?
08:22
<nessy>
othermaciej: me, too, but 3 times, because of 3 different OSes
08:22
<nessy>
hsivonen: yes, cortado fallback right now
08:22
<hsivonen>
nessy: ok thanks
08:22
<nessy>
requires java installation, but there's a fair few of those around
08:22
<othermaciej>
I do also have IE on my Windows partition
08:22
<nessy>
go to http://theora.org/cortado
08:23
<othermaciej>
I'm not sure if IE/Mac is still runnable, but I don't have a copy here
08:23
<nessy>
IE/Mac died a long time ago, I think
08:23
<nessy>
never went back to it
08:23
<nessy>
but yeah, that on Windoze
08:23
<nessy>
bbiab
08:24
<hsivonen>
IE/Mac runs on Snow Leopard
08:26
<hsivonen>
nice, the progress/meter i18n can of worms is gone
08:49
<annevk>
hmm
08:49
<annevk>
getPosition takes an object
08:50
<annevk>
we discussed using object for APIs once in the WebApps WG and decided not to go there...
08:50
<annevk>
oh well
08:50
<annevk>
no consistency in the past, no consistency in the future either
08:51
<annevk>
one of the many flawed particles of the Web platform
09:11
<annevk>
the point Kai made has been addressed before
09:11
<annevk>
but I can't find the email :/
09:12
annevk
lets it be someone else's problem
09:12
<othermaciej>
well I just took a fresh shot at replying
09:12
<othermaciej>
if you have an older email to cite that would be nice too
09:14
<annevk>
cannot find it
09:16
<hsivonen>
the id numbers for saved docs on the live DOM viewer are surprisingly small
09:16
<hsivonen>
I guess the user base isn't particularly large
09:20
hsivonen
wonders if there are test cases for load/unload/pagehide/pageshow events for about:blank loading into an iframe
09:28
<hsivonen>
is onpageshow supposed to work as a content attribute?
09:29
<annevk>
per HTML5, yes
09:29
<hsivonen>
what about real world?
09:29
<hsivonen>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/342
09:29
<hsivonen>
is iframe not firing it or the handler not supported?
09:29
<annevk>
but only for body and frameset
09:30
<hsivonen>
oh
09:30
<annevk>
per html5 anyway
09:30
<hsivonen>
why frameset but not iframe?
09:30
<Hixie>
body and frameset are equivalent
09:30
<Hixie>
iframe and frame are equilavent
09:30
<annevk>
presumably because frameset takes the role of body
09:30
<hsivonen>
ok
09:33
<hsivonen>
is it even possible to register a pageshow event handler for an iframe that loads about:blank before it's too late?
09:33
<hsivonen>
parser-created iframe that is
09:33
<annevk>
prolly not
09:33
<annevk>
but you could load it again
09:33
<annevk>
after you add a listener to the window object
09:57
hsivonen
wonders what history behind about:blank loading synchronously is
09:57
<hsivonen>
not that it matters at this point to know who the guilty party is
09:59
<annevk>
i guess so that you can directly manipulate the document
09:59
<annevk>
makes sense to me
10:00
<hsivonen>
It's quite an annoyance
10:01
<annevk>
speaking as an author, i trump you :p
10:01
jgraham
was about to say the same thing as annevk
10:01
<hsivonen>
you shouldn't be using about:blank-created DOMs, because they aren in the quirks mode
10:02
<annevk>
thankfully jgraham loses with me
10:02
<annevk>
:p
10:03
<othermaciej>
is there any way to put an about:blank frame in standards mode?
10:03
<annevk>
about:standards-blank
10:03
<hsivonen>
othermaciej: I thinks there isn't
10:04
<othermaciej>
even if you document.write a standards-mode doctype into it?
10:04
<hsivonen>
othermaciej: document.writing into it implies document.open, which blows away html, head and body
10:05
<hsivonen>
and the "feature" here is that you get html, head and body synchronously and for free
10:06
<Lachy>
annevk, <iframe src="data:text/html,&lt;!DOCTYPE html&gt;"></iframe> will give you a blank document in standards mode
10:07
<hsivonen>
Lachy: but asynchronously
10:07
<annevk>
and in a different origin in some browsers
10:07
<Lachy>
oh, yeah.
10:07
<annevk>
though that's a bug per HTML5
10:18
<zcorpan>
about:blank is standards mode in opera
10:19
<hsivonen>
zcorpan: does it have a doctype?
10:19
<hsivonen>
zcorpan: bug or deliberate feature?
10:19
<annevk>
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN">
10:19
<zcorpan>
i've only seen one minor bug about it, so maybe we can change it to standards mode in other browsers too
10:19
<annevk>
has a <title> too
10:19
<hsivonen>
that's interesting
10:19
<annevk>
"Blank page"
10:20
<annevk>
and the root element has dir set to rtl
10:20
<hsivonen>
annevk: is the string localized?
10:20
<annevk>
euh ltr
10:20
<annevk>
no idea
10:20
<hsivonen>
about:blank in Gecko has a doctype, too
10:20
<hsivonen>
but a quirky one
10:21
<zcorpan>
i think it would be nice if about:blank were standards mode
10:21
Philip`
wonders how OS X interprets the encoding of non-ASCII filenames on 8-bit-filename-filesystems (like FAT)
10:22
<workmad3>
about:blank checks as valid in FF as well
10:22
<workmad3>
or at least, it reckons it's valid in the w3c validator
10:22
<workmad3>
web dev toolbar in FF reckons it's quirky :/
10:23
<annevk>
the validator is out of touch with reality
10:25
<hsivonen>
hmm. if I'm reading the code right, about:blank in Gecko has or doesn't have a doctype depending on how it got created
10:25
<hsivonen>
fun
10:26
<workmad3>
heh :)
10:29
<zcorpan>
hsivonen: can about:blank be changed to standards mode in gecko?
10:30
<hsivonen>
zcorpan: as in can authors change its mode or could a Gecko developer land code for changing it in the engine?
10:38
<othermaciej>
document.write()ing a doctype seems simple enough as a way to get standards mode
10:38
<othermaciej>
it doesn't depend on about:blank being synchronous in any way but it's convenient enough
10:38
<othermaciej>
I think there are compat reasons that about:blank has to be in quirks mode
10:38
<jgraham>
othermaciej: You have a somewhat unusual view of "simple enough"
10:38
<zcorpan>
hsivonen: the latter
10:39
<othermaciej>
jgraham: if you're about to poke at the DOM of a frame. doing someFrame.contentDocument.write("<!DOCTYPE html>") first seems prety simple to me
10:39
<zcorpan>
othermaciej: can you cite the compat reasons?
10:39
<othermaciej>
zcorpan: not offhand
10:40
<jgraham>
othermaciej: The incantation isn't hard once you know it; it is realising that that is the right incantation that is the non-simple part
10:40
<othermaciej>
sure
10:40
<othermaciej>
that would be true of any special incantation, other than changing the default
10:40
<othermaciej>
I wonder if about:blank in a standards mode document could be in standards mode
10:41
<othermaciej>
but I would not be surprised if there are standards-mode pages with about:blank iframes that they depend on being quirky
10:41
<hsivonen>
zcorpan: probably not, but you might a get a better answer from Boris.
10:42
<zcorpan>
othermaciej: it has not caused much of a problem for opera to always use standards mode
10:44
<annevk>
much of or no problem?
10:44
annevk
hasn't seen a thing
10:48
<zcorpan>
i've seen one bug where it caused unwanted top margin because we insert a <p> when pressing enter in contenteditable
10:48
<zcorpan>
which could either be solved by using quirks mode or by not using <p>
10:49
<zcorpan>
i think we might move away from <p> anyway
10:52
<hsivonen>
hmm. so is pageshow supposed to fire on initial page load, too, and not only when navigating back to a page in the back-forward cache?
10:53
<hsivonen>
apparently yes
10:53
<hsivonen>
and Gecko and WebKit either differ in handler support or in event firing
10:53
<zcorpan>
does pageshow use the Event interface?
10:53
<hsivonen>
dunno
10:54
<othermaciej>
I don't like how pageshow is defined
10:54
<othermaciej>
it makes it needlessly awkward to have code that *only* runs when restoring from b/f cache
10:55
<othermaciej>
if only one of "load" or "pageshow" fired, you could easily put code in one handler or both depending on whether it's one-time, every-time or restore-only initialization
10:55
<othermaciej>
not only that but "load" has to do complete initialization anyway, to support browsers that don't have "pageshow"
10:56
<othermaciej>
I believe I gave this complaint to the folks who first implemented these events for Gecko but they were unconvinced
11:00
<zcorpan>
not too late to change it
11:00
<hsivonen>
hasn't Gecko had a b/f cache for years already?
11:10
<othermaciej>
yes it has
11:10
<othermaciej>
I discussed it with the engineers working on it before it shipped, years ago
11:11
<othermaciej>
I'd suspect at this point a new event ("pagerestore"?) would be required
11:11
<othermaciej>
although I think it's hard to depend on "pageshow" always firing, and yet still be compatible with browsers that never fire it at all
11:28
<Lachy>
wtf? http://rebuildingtheweb.com/en/invitation-html5-team/
11:29
<hsivonen>
Lachy: I guess he missed the design principle debates
11:29
<Lachy>
"The first phase in these debates is a challenge, from me to Ian Hickson [...] to debate me on the following topic: Error messages should be displayed for corrupt HTML5."
11:30
<hsivonen>
(as if showing errors were up to Hixie to decide)
11:30
<Lachy>
I guess I shouldn't be too surprised he's still pushing for draconian error handling in HTML, even though that debate has been over for years.
11:30
<annevk>
http://xkcd.com/386/
11:31
<othermaciej>
who is this person?
11:31
<Hixie>
lachy: http://lists.w3.org/Archives/Public/www-archive/2010Jan/0009.html
11:31
<hsivonen>
othermaciej: the XStandard XML editor guy
11:34
<zcorpan>
i thought almost all discussion in public-html is about design principles/process/direction
11:34
<hsivonen>
weird. I no longer see his byline on xhtml.com
11:34
<zcorpan>
as about in-depth, well the threads are well indented at least
11:35
<annevk>
for some people standards is fun
11:35
<annevk>
for others it's process
11:35
<Hixie>
standards are a tool
11:35
<Hixie>
for getting interop
11:35
<Hixie>
nothing more, imho
11:36
<Hixie>
(sometimes they're not even the best tool for that job)
11:36
<annevk>
i meant working on standards
11:36
<Hixie>
oh working on standards is a lot of fun :-)
11:37
<Hixie>
especially when it's like-minded people working on things together
11:37
<Hixie>
it's less fun when it's people who disagree on the fundamental goals
11:38
<zcorpan>
down with people who disagree
11:39
<othermaciej>
I replied to his post (perhaps foolishly)
11:39
<zcorpan>
validator.nu shows error messages for corrupt html5
11:42
<Lachy>
zcorpan, based on what he's previously said on the topic, he means for browsers to show error messages to users
11:42
<Lachy>
I responded to his post too.
11:43
<zcorpan>
Lachy: of course, but the post didn't say that :)
11:49
<jgraham>
http://vocamus.net/dave/?cat=25 seems interesting
11:56
<Hixie>
ok got it down to 157 bugs
11:56
<Hixie>
i was aiming for 150
11:56
<Hixie>
but i started about 20 higher than i expected to start
11:56
<Hixie>
so i'll call this a qualified success
11:56
<Hixie>
good night
12:03
<othermaciej>
good nigh
12:03
<othermaciej>
t
12:07
<annevk>
nn
13:38
<jgraham>
Heh
13:38
<jgraham>
I obviously shouldn't have replied
13:38
<annevk>
me neither I think
13:38
<annevk>
but I did
13:39
<annevk>
I'm already worried about flames
13:39
<hsivonen>
I have to wonder if were baited to reply accidentally or intentionally
13:40
<annevk>
quite surprising how much we're aligned on certain points
13:41
<hsivonen>
evidence of groupthink!
13:41
<annevk>
heh
13:41
<annevk>
might be
13:42
<annevk>
earlier today when I discussed that email with othermaciej neither of us mentioned a view point but we both knew it was the same
13:42
<TabAtkins>
Pretty sure it was an intentional baiting.
13:42
<annevk>
it struck me a little later how convenient that was
13:47
<othermaciej>
the "Browser implementations" thread?
13:48
<Lachy>
othermaciej, yeah, I think so
13:49
<othermaciej>
I don't see evidence of baiting (not really sure what we were baited into if so)
13:50
<jgraham>
I think the supposed baiting was the suggestion that only released specs contribute to the legacy
13:50
<Lachy>
re the autobuffer thread, I'm not entirely convinced that making autobuffer itself a tri-state attribute (on/off, or absent) is impossible, despite it being boolean in Firefox 3.5
13:51
<jgraham>
But I'm not really sure that it was baiting (I'm not sure what the point would be)
13:52
<jgraham>
Lachy: Agreed but I don't really see the value in taking the risk
13:52
<othermaciej>
Firefox has pretty high uptake on updates
13:52
<Lachy>
since autobuffer is, I believe, currently only implemented in Firefox, and its omission currently has the same effect as autobuffer=off should have, then authors can simply omit the attribute for the relatively short transition from FF 3.5 to 3.6
13:53
<othermaciej>
there are advantages and disadvantages to reusing autobuffer vs. minting a new buffer attribute
13:53
<Lachy>
so, if 3.6 supported autobuffer as tri-state, then the issue would not have serious long term consequnces
13:53
<othermaciej>
a new buffer attribute could be more clear and would have less risk of legacy conflict, but at the same time it might leave us stuck with two attributes
13:53
<othermaciej>
I'm planning to list both options when I file the bug
13:53
<jgraham>
I think we could obsolete autobuffer at this stage
13:54
<Lachy>
of course, using buffer="..." as tri-state, and obsoleting autobuffer would also be acceptable to me
13:55
<Lachy>
I don't think there is significant content out there relying on autobuffer for it to be impossible to remove once buffer="" is more widely supported
13:55
<annevk>
autobuffer can easily be removed
13:55
<annevk>
that was never the concern
13:55
<annevk>
afaict
13:56
<hsivonen>
Lachy: Gecko 1.9.2 is in RC in mobile. changes to the engine feature set are highly theoretical at this point
13:57
<Lachy>
also, I'm slighly less convinced about the need for tri-state buffering, until I've seen more evidence about what sort of heuristics browsers can realistically apply when the attribute is absent to make it significantly different from the no-buffering case.
14:22
<annevk>
othermaciej, did you file a bug on <img>.x/y
14:23
<annevk>
or can they can be nuked from the browsers
14:24
<othermaciej>
annevk: I don't know what that is
14:24
<othermaciej>
annevk: is it something we're missing?
14:24
<annevk>
it is something that's not in Opera of which we have a old bug report from 2007 or so about a compat issue
14:24
<annevk>
it is apparently in Firefox and WebKit
14:25
<annevk>
according to the bug "These properties are respectively the left offset and top offset relative to the viewport borders."
14:25
<annevk>
that sounds somewhat wrong terminology wise, but I guess you get the idea
14:25
<Huvet>
have anyone had any success with sending the lxml tree from html5lib to lxml.html.clean.Cleaner?
14:26
<Huvet>
I get a crash that _Element is missing a rewrite_links method
14:30
<virtuelv>
Image.x = The X coordinate of the image within the client display area.
14:31
<jgraham>
Huvet: That sounds like a difference between lxml.etree._Element and lxml.html.)Element or something
14:31
<virtuelv>
Have browsers nuked Image.vspace/hspace yet?
14:31
<annevk>
no :/
14:32
<annevk>
and othermaciej is advocating lots of more cruft I found out today!
14:32
<Huvet>
jgraham: Ah, I assumed that an lxml.etree was always the same... I'll have a look
14:32
<virtuelv>
properties found in my ancient Wrox reference:
14:32
<virtuelv>
border, complete, constructor, defaultValue, height, hspace, lowsrc, name, size, src, vspace, width, x, y
14:32
<virtuelv>
and select()
14:33
<othermaciej>
annevk: Im not familiar with <img>.x/y, if you would like I can search through history and see if I can find why it was added
14:33
<othermaciej>
annevk: what cruft am I advocating?
14:33
<virtuelv>
<img>.x was first added in Netscape 3
14:33
<annevk>
just the bugs you filed on even more interfaces for legacy elements :)
14:35
<annevk>
like HTMLAnchorElement.text
14:35
<virtuelv>
<a>. accessKey, dataFld, dataSrc, hash, host, hostname, href, Methods, mimeType, nameProp, pathname, port, protocol, protocolLong, rel, search, tabIndex, target
14:35
<annevk>
hmm, doesn't IE have a .text attribute for all elements?
14:35
<virtuelv>
dataFld/Src?
14:37
<Philip`>
virtuelv: They're a sort of client-side templating thing in IE (and in the HTML4 DTD)
14:38
<Philip`>
Fortunately approximately nobody uses that feature
14:39
<virtuelv>
some data binding thing, according to MSDN
14:40
<virtuelv>
I wonder if there's other, long-forgotten stuff this reference lists
14:41
<Philip`>
(I think I found one site that used the feature for real (to generate a list of cars from an XML data source), out of about half a million pages)
14:41
<virtuelv>
IMG.fileModifiedDate (Property)
14:42
<virtuelv>
iccProfile, fileSize, start
14:43
<virtuelv>
loop and dynSrc as well
15:37
<MikeSmith>
zcorpan: fyi, initial attempt at <script src=""></script> checking is live on qa-dev
15:54
<zcorpan>
MikeSmith: cool
15:57
<zcorpan>
MikeSmith: seems to work from quick testing
16:03
<Lachy>
Hixie, why does the complete version of the spec still have a version number in its title "Web Applications 1.0", given that the other version is now just called "WHATWG HTML" without a specific version?
16:09
<zcorpan>
why is it "WHATWG HTML" instead of just "HTML"?
16:11
<Lachy>
I think that's just to help distingish it from the subset of HTML currently being developed in W3C. Though, I don't feel strongly either way about it.
16:50
<Philip`>
How does replacing the boolean autobuffer attribute with a new enumerated attribute that has two values help with the problem of how browsers will react to new values they haven't heard of before?
16:51
<workmad3>
Philip`: allows them to die a bit more gracefully? :)
16:51
<Philip`>
(Is it just that having new values defaulting to don't-buffer is better than defaulting to do-buffer? Doesn't that depend on what the new values are going to be, which we don't know yet?)
16:52
<Philip`>
workmad3: I'm not sure I'd ever consider death to be graceful
16:52
<jgraham>
Philip`: I assume defaulting to more conservative behaviour is better
16:53
<jgraham>
(iundependent of what the less conservative options are)
16:53
<Philip`>
jgraham: i.e. defaulting to behaviour that harms the typical user experience by not buffering and delaying their enjoyment when they want to watch the video?
16:53
<workmad3>
I'd figure it makes it easier to expand future versions in a backwards compatible manner though
16:54
<Philip`>
jgraham: Seems like neither option is really "better", they all have drawbacks and people could complain either way
16:54
<jgraham>
Philip`: I think lots of flash0using sites already do no autobuffering so it can't be that bad for user experience
16:54
<workmad3>
adding a new value to the values available is less disruptive than changing the attribute to an enumeration and also adding a new value
16:55
<jgraham>
s/0/-/
16:56
<workmad3>
imagine the complaints if the standard changed from 'autobuffer="autobuffer" ' enabling autobuffering to that being a non-valid value and therefore disabling it
16:56
<workmad3>
so betting that the attribute starts off with 'autobuffer="on" ' to enable it and never needing to change :)
16:57
<jgraham>
workmad3: I'm imagining the complaints and I can't really see them being that bad
16:57
<jgraham>
You simply make <video autobuffer> spec-wise identical to <video>
16:57
<workmad3>
true... not as bad as if it were the other way around
16:58
<jgraham>
and then people who really want autobuffering in FF3.5 just add the invalid attribute (in addition to the new attribute for other UAs)
16:59
<annevk>
Hixie, people are complaining about mozillaquestquest.com to me
17:27
<annevk>
foolip, if we have a new attribute that might support multiple states going forward but is currently just a flag there does not seem to be a need for more than one value
17:28
<annevk>
(the other value is simply omitting the attribute)
17:56
<TabAtkins>
Crap, maybe I shouldn't have gotten a computer upgrade. Now my dropdown menus are smooth as silk. At least before I could be reasonably assured that my customers had a computer equal to or better than mine, and so would have an equal or better experience.
18:53
<ehird>
which is correct: <article>...<section class="comments">...</section></article> or <article>...</article><section class="comments">...</section>?
18:56
<AryehGregor>
ehird, "For instance, a blog entry on a site that accepts user-submitted comments could represent the comments as article elements nested within the article element for the blog entry."
18:56
<AryehGregor>
http://www.whatwg.org/specs/web-apps/current-work/multipage/sections.html#the-article-element
18:56
<ehird>
OK. Just seemed weird to consider comments part of the article itself, to me.
18:56
<AryehGregor>
So the way the spec suggests is <article>...text of article... <article>comment 1</article> <article>comment 2</article>...</article>
18:56
<AryehGregor>
The idea is to suggest that they're related to it, I guess.
18:57
<AryehGregor>
It does seem awkward for CSS.
22:30
<gsnedders>
http://www.molly.com/2010/01/05/bob-dylan-meets-html5/
22:53
<virtuelv>
gsnedders: I'd rather have hoped Dylan himself would write of it
22:53
<virtuelv>
But I'll settle for the Dylan christmas songs I bought a few weeks ago
23:36
<Hixie>
anyone know anything about nested iframes andonbeforeunloads?
23:36
<Hixie>
er
23:36
<Hixie>
"and onbeforeunload?"
23:45
<cardona507>
dylan meets html5 - sheesh :-/