2007-12-01
[16:53:00.0000] <Dashiva>
Love that mail, "how about expanding it to the full thing like the HTML 4.01 DOCTYPE?"

[16:53:01.0000] <Dashiva>
No reason given, just let's make it longer

[17:09:00.0000] <jruderman>
http://www.webdevout.net/david-hammond/works/we-didnt-write-the-browsers

[17:11:00.0000] <Dashiva>
The revolution will not be namespaced

[18:07:00.0000] <Hixie>
Lachy: hey, i used to sleep in the room opposite yours

[18:42:00.0000] <MikeSmith>
Hixie - /me wonders where Lachy is staying. Alan's current place?

[18:44:00.0000] <Hixie>
no

[18:44:01.0000] <Hixie>
though i lived there too

[18:44:02.0000] <Hixie>
see his blog for link to pictures on flickr

[18:44:03.0000] <Hixie>
anywya gotta go

[18:44:04.0000] <MikeSmith>
see yah

[22:44:00.0000] <Hixie>
i hate how when you define a function inside a function of an object the 'this' of the inner function isn't the same as the outer function's

[22:47:00.0000] <othermaciej>
ES4 fixes that

[22:47:01.0000] <Hixie>
really?

[22:47:02.0000] <Hixie>
without breaking existing scripts?

[22:47:03.0000] <Hixie>
how?

[22:47:04.0000] <othermaciej>
but the canonical way to deal is to do this in the outer function:

[22:47:05.0000] <othermaciej>
var self = this;

[22:47:06.0000] <othermaciej>
and then use self in the nested func

[22:47:07.0000] <Hixie>
yeah

[22:47:08.0000] <othermaciej>
I don't know if it's without breaking existing scripts

[22:48:00.0000] <Hixie>
(i tend to do  var func = function (self) { return function (real args) { ... self ... }; }(this);  so as to not polute the outer scope with a 'self')

[22:48:01.0000] <Hixie>
(also known as currying)

[00:39:00.0000] <othermaciej>
sure, also an acceptable way to do things

[00:39:01.0000] <othermaciej>
although if the outer scope is a function scope that's captured by the continuation it's not much cleaner and may make things a tiny bit slower depending on implementation details

[00:40:00.0000] <Hixie>
yeah

[03:47:00.0000] <Lachy>
I don't understand Doug's response about Ogg Theora and Dirac not being *necessarily* patent free.

[03:47:01.0000] <Lachy>
We know they have patents, so it doesn't make sense

[03:53:00.0000] <Philip`>
Is Dirac affected by known patents?

[03:54:00.0000] <Philip`>
I know Theora has the ones released by On2, but haven't heard of Dirac relying on royalty-free patents

[03:55:00.0000] <Philip`>
Oh, apparently the BBC has some patents on Dirac

[04:23:00.0000] <Ben`>
will HTML 5 really include SQL execution?

[04:24:00.0000] <Philip`>
That is the current plan

[04:24:01.0000] <Ben`>
ok, thanks

[04:25:00.0000] <Philip`>
(like http://www.whatwg.org/specs/web-apps/current-work/multipage/section-sql.html#sql etc)

[04:34:00.0000] <othermaciej>
Lachy: yes, Doug seemed confused

[04:34:01.0000] <othermaciej>
Ben`: it will include local database storage with SQL-based access

[04:34:02.0000] <othermaciej>
or rather, a DOM API for it that embedded scripts can use

[04:35:00.0000] <othermaciej>
I'm saying it more precisely because "SQL execution in HTML" sounds like a scary security hole when phrased that way

[04:35:01.0000] <othermaciej>
but it's really the same kind of script-based local storage feature provided by extensions like Google Gears or Adobe Flash

[04:36:00.0000] <othermaciej>
Ben`: also, if you want to play around with it, WebKit has an experimental implementation: http://webkit.org/blog/126/webkit-does-html5-client-side-database-storage/

[07:33:00.0000] <gsnedders>
Hixie: ping

[07:41:00.0000] <Hixie>
gsnedders: pong, but i'm going offline in about 30 seconds

[07:41:01.0000] <gsnedders>
Hixie: then I'll ping you later

[07:41:02.0000] <gsnedders>
:)

[07:42:00.0000] <Hixie>
k :-)

[07:42:01.0000] <Hixie>
ttyl

[07:42:02.0000] <gsnedders>
(or just pong me)

[11:57:00.0000] <gavin>
what's the URI for hixie's list of issues he needs to address that was posted to public-html a while ago?

[11:59:00.0000] <aroben>
gavin: http://www.whatwg.org/issues/

[11:59:01.0000] <gavin>
thanks

[12:01:00.0000] <gavin>
hmm, when I try to register, I get "E-mail is not valid or does not correspond to a user who has sent feedback that has not yet been dealt with."

[12:02:00.0000] <gavin>
but my email is valid, and I have sent some feedback that hasn't been addressed

[12:02:01.0000] <gavin>
I wonder if my message somehow got lost, or was deemed to be something that doesn't need to be dealt with

[12:03:00.0000] <gavin>
(lost by whatever mechanism is used to track feedback, I mean - it's in the list archives)

[12:04:00.0000] <Philip`>
gavin: You could see if your message is in the issues list

[12:04:01.0000] <anne-mac>
was it a message to whatwg⊙wo or public-html⊙wo ?

[12:04:02.0000] <gavin>
it doesn't appear to be

[12:04:03.0000] <Philip`>
and if it is, the corresponding email address should work

[12:04:04.0000] <gavin>
whatwg@

[12:05:00.0000] <anne-mac>
and you're using the same address? :)

[12:05:01.0000] <gavin>
http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2007-September/012459.html

[12:05:02.0000] <gavin>
yes

[12:06:00.0000] <anne-mac>
hmm, we need that in the IE-way for <canvas> to work properly...

[12:06:01.0000] <anne-mac>
/me likes the irony here

[12:07:00.0000] <virtuelv>
anne-mac: since when did you use a mac?

[12:07:01.0000] <anne-mac>
i bought one a week ago to fool around

[12:08:00.0000] <virtuelv>
I have stuff to teach you ... over a beer next week

[12:08:01.0000] <Philip`>
anne-mac: or we need to not define canvas features in terms of img.complete

[12:09:00.0000] <anne-mac>
sure, but the IE behavior makes sense, imo

[12:09:01.0000] <Philip`>
Has someone investigated what happens to .complete when you modify .src?

[12:10:00.0000] <gavin>
I haven't

[12:10:01.0000] <gavin>
I believe Mozilla's implementation will accureately reflect the current state of the image at all times

[12:10:02.0000] <gavin>
(modulo the behavior I describe in my post)

[12:13:00.0000] <Philip`>
gavin: Hmm, that message is definitely not in the issues list

[12:13:01.0000] <Philip`>
You should poke Hixie about it :-)

[12:14:00.0000] <Philip`>
(I've had at least one fall into his spam filter)

[12:14:01.0000] <anne-mac>
I had several of those

[12:14:02.0000] <anne-mac>
that we somehow discovered quick enough, not sure if there are more

[12:15:00.0000] <anne-mac>
then again, not sure how much I care about some of those two year old comments

[12:16:00.0000] <gsnedders>
/me probably should've asked Hixie when he intended to be back

[14:47:00.0000] <gsnedders>
Hixie: back yet?

[15:26:00.0000] <Philip`>
/me wonders who set up http://hixietracker.blogspot.com/

[15:26:01.0000] <gsnedders>
me! choose me!

[15:27:00.0000] <Philip`>
(They ought to fix the non-escaping of the title "Re: [whatwg] <BIG> Element" since it makes the rest of the page big)

[15:27:01.0000] <gsnedders>
Philip`: isn't Hixie awesome enough to deserve that, though?

[15:27:02.0000] <Hixie>
gavin: your e-mail is one of a small number of e-mails that i received during a small window of time where dreamhost was down and the e-mail is only in my gmail inbox, not my imap box

[15:28:00.0000] <Lachy>
Philip`, IIRC, I think the hixietracker was set up by Mark Pilgrim

[15:28:01.0000] <Lachy>
it was mentioned in one of these IRC channels a few months ago

[15:28:02.0000] <gsnedders>
/me waits to be quoted out of context and claimed that he goes along with anything that Hixie says

[15:28:03.0000] <gavin>
Hixie: ah

[15:28:04.0000] <gsnedders>
Hixie: I guess your back then?

[15:28:05.0000] <Hixie>
gsnedders: heya

[15:29:00.0000] <Hixie>
gsnedders: sure, for a few minutes :-)

[15:29:01.0000] <gavin>
Hixie: should I resend it or something?

[15:29:02.0000] <Hixie>
gavin: if you want to vote on it and if you don't mind resending, sure

[15:29:03.0000] <Hixie>
but i won't lose it if you don't

[15:30:00.0000] <Hixie>
(lose the e-mail that is)

[15:30:01.0000] <Hixie>
(some might say i've already "lost it")

[15:30:02.0000] <gavin>
ok, that's all I care about :)

[15:30:03.0000] <Philip`>
Lachy: Okay - I don't remember that at all, and can't find any search results, and it looks like Hixie added the link to it on the wiki

[15:30:04.0000] <gsnedders>
Hixie: not as badly as me

[15:30:05.0000] <gsnedders>
/me points Hixie at PM

[15:31:00.0000] <Philip`>
It reminds me of the Santa Tracker except without the presents

[15:31:01.0000] <Lachy>
the discussion occurred around the time the link was added to the wiki, but I couldn't find it in the logs either.  Maybe Krijn was offline at the time

[15:32:00.0000] <gsnedders>
Philip`: Santa is real kthxbai

[15:33:00.0000] <Philip`>
Lachy: I guess I was offline too since it's not in my logs either :-(

[15:34:00.0000] <gavin>
I can't find it in my logs of this channel

[15:35:00.0000] <gavin>
but I think I only have logs here since december 2006

[15:35:01.0000] <Lachy>
strangley, not in mine either, AFAICT

[15:35:02.0000] <Lachy>
though I don't have a complete set of logs

[15:35:03.0000] <Philip`>
The wiki change was Sep 25, which sounds recent until I realise the year has passed far quicker than I expected

[15:36:00.0000] <Philip`>
Maybe it was a different channel somewhere?

[15:36:01.0000] <gsnedders>
wait… it's Dec?

[15:36:02.0000] <Lachy>
I'm sure it was either #whatwg or #html-wg

[15:36:03.0000] <gsnedders>
that's maddening.

[15:37:00.0000] <gsnedders>
/me realises what that means: mocks next month!

[15:37:01.0000] <Lachy>
yes, gsnedders, December usually begins on December 1st every year

[15:38:00.0000] <gsnedders>
oh. I thought it started on December 31st.

[15:38:01.0000] <Philip`>
I think your calendar must be upside down

[15:39:00.0000] <gsnedders>
Ah. Maybe I should get it out the drawer its in?

[15:39:01.0000] <Philip`>
You might also want to check that your clocks are turning clockwise, to rule out the possibility of a time reversal

[15:40:00.0000] <gsnedders>
My sister had a clock that went the wrong way once. I missed the start of school because of it once.

[15:41:00.0000] <Hixie>
wow

[15:41:01.0000] <Hixie>
w3c actually censored me

[15:41:02.0000] <Lachy>
what?

[15:41:03.0000] <Hixie>
well that's a first

[15:41:04.0000] <Lachy>
where?

[15:42:00.0000] <Hixie>
http://lists.w3.org/Archives/Public/www-archive/2007Nov/0104.html

[15:42:01.0000] <Hixie>
dean complained that i replied to his private message by cc'ing the archives

[15:42:02.0000] <Hixie>
and w3c removed my message from the archives

[15:43:00.0000] <Lachy>
woah!  Luckily, i still have a copy in my personal archive :-)

[15:43:01.0000] <gavin>
I noticed that earlier

[15:43:02.0000] <gavin>
I wish I'd kept a copy :)

[15:44:00.0000] <Philip`>
Lachy: www-archive is meant to be a "Mail-to-web gateway", so subscribing to the list and using it for mail-to-mail is cheating :-p

[15:46:00.0000] <MikeSmith>
Hixie - fwiw, I wasn't aware that message had been removed. They were supposed to get your OK to remove it.

[15:47:00.0000] <MikeSmith>
"The need for editing or excising material contained in an archived message is addressed only when the sender of the archived message requests it. Carefully documented exceptions may be considered."

[15:47:01.0000] <MikeSmith>
http://www.w3.org/Mail/ArchiveEditingPolicy

[15:48:00.0000] <Lachy>
Philip`, I subscribed to it since I noticed there was a lot of interesting off-list discussion going on there

[15:48:01.0000] <gavin>
I suppose you could argue that Dean was the sender of a message that was archived

[15:49:00.0000] <gavin>
(indirectly through hixie's reply)

[15:50:00.0000] <MikeSmith>
gavin - yeah, I suppose

[15:50:01.0000] <Hixie>
well it's now archived at http://damowmow.com/playground/archives/reply-to-dean-edridge.html

[15:51:00.0000] <Lachy>
LOL! I was just about to suggest that you archive it elsewhere

[15:56:00.0000] <gsnedders>
so, us follows of Hixie…

[15:56:01.0000] <gsnedders>
*followers

[15:56:02.0000] <Philip`>
Hixie: It's a shame that the change in month breaks the W3C mail archive's reply-linking feature


2007-12-02
[16:00:00.0000] <gsnedders>
g'nite y'all

[16:07:00.0000] <Hixie>
Philip`: yeah

[16:09:00.0000] <Dashiva>
Isn't there a three-month view?

[16:09:01.0000] <Hixie>
depends on the list

[16:12:00.0000] <gavin>
depends on the number of messages per month, I thought

[16:12:01.0000] <gavin>
i.e. I thought it was automatic based on number of messages, instead of configured manually

[16:12:02.0000] <Hixie>
no

[16:12:03.0000] <Hixie>
manual

[16:12:04.0000] <gavin>
ah, ok

[16:13:00.0000] <Hixie>
gotta go again

[16:37:00.0000] <othermaciej>
hey, I got reddit karma points for defending html5: http://programming.reddit.com/info/61vbd/comments/

[16:37:01.0000] <othermaciej>
(the thread was going much the other way before my post)

[16:45:00.0000] <Dashiva>
"I say let Adobe's Air & Google's Gears work out the kinks, then maybe a few years down the road (once the issues with the core languages are settled) go about implementing them."

[16:47:00.0000] <Dashiva>
Although this one was more convincing: "html5 is a joke anyways and will never be adopted or supported and if it is I will get an ak 47 and find a rooftop"

[16:47:01.0000] <othermaciej>
heh

[16:47:02.0000] <Philip`>
Don't AK-47s work just as effectively at ground level?

[16:47:03.0000] <Dashiva>
Why are people so unpatriotic? Can't they use proper western firearms?

[16:51:00.0000] <jwalden>
the day of the tommy gun has passed

[16:52:00.0000] <bradee-oh>
http://www.ak-47.us/USmade.php  <--- AK-47 varieties made in the United States

[16:58:00.0000] <othermaciej>
US military issue firearms are generally much harder to maintain than the AK-47

[16:58:01.0000] <othermaciej>
which is why they are not used as much by guerilla armies and criminal gangs and the like

[17:00:00.0000] <Philip`>
The US military should develop a really bad gun, then make it easy to obtain and put a huge PR campaign behind it, so the guerilla armies and criminal gangs think it's excellent and buy lots of it

[17:01:00.0000] <Dashiva>
Make a gun that calls the police when you fire it!

[17:05:00.0000] <Philip`>
Dashiva: The problem with that plan is that it would have to have something like a mobile phone embedded in the gun, and prolonged use might lead to excess levels of radiation, which would be a health hazard

[17:07:00.0000] <Dashiva>
It's three mile island all over again

[21:58:00.0000] <Hixie>
that's odd

[21:58:01.0000] <Hixie>
i posted a huge comment to sam ruby's blog the other day

[21:58:02.0000] <Hixie>
and it hasn't appeared

[23:23:00.0000] <othermaciej>
mine appeared

[23:30:00.0000] <othermaciej>
I just posted a slightly mean reply on his blog

[23:52:00.0000] <Hixie>
another?

[00:03:00.0000] <othermaciej>
I didn't think the first one was that mean

[00:08:00.0000] <Hixie>
indeed, hence my question

[00:10:00.0000] <othermaciej>
so, yes, another reply

[07:27:00.0000] <hendry>
hsivonen_: don't understand this transparent error http://validator.nu/?doc=http%3A%2F%2Fvideo.natalian.org%2Ftest%2Fhtml5-media-mp3.html

[07:28:00.0000] <Dashiva>
transparent?

[07:28:01.0000] <anne-mac>
/me gets that <audio> can't occur in <body>

[07:33:00.0000] <hendry>
anne-mac: really? not <body>? even with controls?

[07:34:00.0000] <hendry>
same goes for video I expect http://validator.nu/?doc=http%3A%2F%2Fvideo.natalian.org%2Ftest%2Fhtml5-media-ogg.html

[07:35:00.0000] <anne-mac>
yes

[07:36:00.0000] <anne-mac>
needs to be inside a paragraph like element, such as <p> or <figure>

[07:38:00.0000] <hendry>
anne-mac: ok, fine

[07:38:01.0000] <hendry>
though http://validator.nu/?doc=http%3A%2F%2Fvideo.natalian.org%2Ftest%2Fhtml5-media-ogg.html

[07:38:02.0000] <hendry>
seems to expect children on the video element

[07:40:00.0000] <hendry>
i don't understnad this 'transparent' business

[07:42:00.0000] <Dashiva>
Looks like fallback

[07:43:00.0000] <Dashiva>
Basically says "If you replace the element with its transparent children" which is essentially what is supposed to happen with fallback, as I understand it

[07:43:01.0000] <hsivonen_>
hendry: transparent means that the content model is block if the parent's content model was block. et.

[07:43:02.0000] <hsivonen_>
etc.

[07:43:03.0000] <hendry>
so it want's me to put something in there, like "You failed to playback the media"

[07:44:00.0000] <hsivonen_>
hendry: also, IIRC, there's a spec bug near here and I've implemented the spec pedentically with the bug

[07:44:01.0000] <hsivonen_>
hendry: the bug is that if the parent is figure, the transparent content model sucks and you have to put an <img> there

[07:45:00.0000] <Dashiva>
"When a transparent or semi-transparent element has no parent" <-- How does that work, unless documentElement is transparent?

[07:45:01.0000] <hsivonen_>
Dashiva: document fragments I guess

[07:46:00.0000] <hsivonen_>
hendry: hmm. interesting. I don't know why you are getting the error you are getting

[07:47:00.0000] <Dashiva>
Maybe you need a container (div,p) inside it to hold the (non-existent) content

[07:48:00.0000] <hendry>
oh gosh. i hope this block, inline stuff becomes a whole lot easier

[07:48:01.0000] <Philip`>
/me too :-)

[07:49:00.0000] <hsivonen_>
hendry: there's a chance you are seeing a validator bug. I'm not sure. I checked the schema and it isn't obvious to me why you get the error you are getting

[07:49:01.0000] <hsivonen_>
oohh!

[07:49:02.0000] <hsivonen_>
I have a bug!

[07:51:00.0000] <hsivonen_>
hendry: I'm fixing it now

[07:52:00.0000] <Dashiva>
Now that's customer service

[07:52:01.0000] <hendry>
hsivonen_: :)

[07:56:00.0000] <hsivonen>
and I should remove significant inline checks as well...

[08:01:00.0000] <hsivonen>
hendry: fixed

[08:01:01.0000] <hsivonen>
hendry: I'll remove the significant inline thing in due course

[08:12:00.0000] <hsivonen>
hendry: now fixed

[08:12:01.0000] <hsivonen>
hendry: thanks for finding this

[08:21:00.0000] <hendry>
hsivonen: thanks for the fix :)

[09:48:00.0000] <gsnedders>
Hixie: I'm wondering whether to integrate Content-Type sniffing into my spec, as it really does have an effect on what is effectively in Content-Type

[09:49:00.0000] <gsnedders>
Hixie: you got any thoughts on whether it should or should not be there?

[14:24:00.0000] <hendry>
i've written a web app using WF2, http://letter2.dabase.com/


2007-12-03
[16:42:00.0000] <othermaciej>
/me wonders why it is not grossly offensive when Dr. Olaf suggests smell-o-vision

[16:43:00.0000] <Philip`>
Probably because of the context of the suggestion

[17:06:00.0000] <othermaciej>
a long email that no one wants to read all the way through?

[17:07:00.0000] <Dashiva>
othermaciej: Obviously because the formal complaint has already been made, no need to dupe it :P

[20:32:00.0000] <Hixie>
wow, whatwg⊙wo passed 800 members

[00:47:00.0000] <Hixie>
i'm still getting e-mails from thursday

[00:47:01.0000] <Hixie>
poor ailed mailbox

[01:16:00.0000] <Hixie>
right well the body of my annotation script is done

[01:16:01.0000] <Hixie>
now i just need the serverside

[01:16:02.0000] <Hixie>
MikeSmith: yt?

[01:17:00.0000] <annevk2>
where is it again?

[01:17:01.0000] <MikeSmith>
Hixie - yup

[01:18:00.0000] <Hixie>
MikeSmith: did you get a chance to the look at the server side of the annotation stuff by any chance?

[01:18:01.0000] <Hixie>
annevk: http://damowmow.com/playground/spec-annotation/001.html

[01:18:02.0000] <MikeSmith>
Hixie - nope, not much

[01:19:00.0000] <MikeSmith>
can make time this week unless you beat me to it

[01:19:01.0000] <MikeSmith>
I've had a few other things on my hands the last few weeks

[01:20:00.0000] <MikeSmith>
and weekends

[01:20:01.0000] <Hixie>
k

[01:20:02.0000] <Hixie>
no worries

[01:21:00.0000] <annevk>
it doesn't work in Opera

[01:21:01.0000] <annevk>
I get some errors in the console, such as 'backround' is not a CSS property

[01:21:02.0000] <Hixie>
reload

[01:21:03.0000] <annevk>
"Fatal network error: Not Found"

[01:21:04.0000] <Hixie>
yup

[01:21:05.0000] <Hixie>
no server side yet

[01:22:00.0000] <annevk>
when this tool works it will be awesome

[01:22:01.0000] <zcorpan>
Hixie: encodeAttributeValue and encodeTextValue need to use regexp with the "g" flag to replace all occurances

[01:22:02.0000] <Hixie>
(it also doesn't support the UI for creating an account, logging in, checking teh history of an entry, and viewing a user's history)

[01:22:03.0000] <Hixie>
zcorpan: good catch

[01:23:00.0000] <Hixie>
fixed

[02:15:00.0000] <othermaciej>
meh

[02:15:01.0000] <othermaciej>
/me can't sleep

[09:19:00.0000] <Catfish_Man>
annevk: around?

[09:20:00.0000] <anne-mac>
/me is

[09:20:01.0000] <Catfish_Man>
fair enough :)

[09:21:00.0000] <Catfish_Man>
had a question about getElementsByClassName case sensitivity; Specifically, I'm unconvinced that the case insensitive matching quirk shouldn't apply to it, for two reasons. The main one is that it seems really confusing to authors to have a situation where you attempt to match the exact same class(es) using two different methods, and get different results

[09:22:00.0000] <Catfish_Man>
the secondary not-actually-a-good-reason reason is that it's a lot harder for me to implement it inconsistently without a slight performance regression ;)

[09:22:01.0000] <Catfish_Man>
I've spoken with Hixie and sayrer (the moz implementer of it) and neither of them cares one way or the other, so I was curious what you thought

[09:23:00.0000] <anne-mac>
i guess i'm ok with it, if everyone is changing

[09:24:00.0000] <Catfish_Man>
is that a "don't care" or a "reluctant but willing"? If it's the latter, what's the reasoning? just avoidance of new quirks?

[09:25:00.0000] <anne-mac>
yeah, more if/else is annoying

[09:25:01.0000] <anne-mac>
although I know that in case of (at least) WebKit it's the reverse, so...

[09:25:02.0000] <Catfish_Man>
I'm actually surprised it's more if/else elsewhere as well

[09:25:03.0000] <Catfish_Man>
it seems like if the matching code is shared, in general the if would be as well

[09:26:00.0000] <anne-mac>
/me has only seen the WebKit patch

[09:26:01.0000] <anne-mac>
so maybe it's not such a big deal

[09:28:00.0000] <zcorpan>
keeping the case-insensitive quirk for all APIs that interact with classes makes sense to me. hmm, i wonder how svg's class="" works in quirks mode

[09:29:00.0000] <Catfish_Man>
zcorpan: hm, good question :)

[09:30:00.0000] <Catfish_Man>
brb, switching to wireless

[09:48:00.0000] <anne-mac>
quirks mode affects SVG too

[09:49:00.0000] <anne-mac>
in Safari 3 and Opera 9.5 at least

[09:56:00.0000] <zcorpan>
http://simon.html5.org/test/svg/quirks/001.htm

[09:56:01.0000] <zcorpan>
also firefox

[12:01:00.0000] <gsnedders>
/me smacks his head reading RFC2616

[12:08:00.0000] <Philip`_>
gsnedders: You should look where you're going instead of reading RFCs while walking, and then you'd be less likely to bash yourself on low-hanging obstacles

[12:08:01.0000] <gsnedders>
Philip`: or just try and make myself shorter

[12:12:00.0000] <Philip`>
That would work too, unless you become short enough that you keep hitting the undersides of tables

[12:15:00.0000] <gsnedders>
Philip`: well, just make me even shorter

[12:15:01.0000] <gsnedders>
:)

[12:15:02.0000] <gsnedders>
oh, wait, the undersides of doors…

[12:15:03.0000] <gsnedders>
hmm…

[13:25:00.0000] <gsnedders>
Hixie: ping

[13:27:00.0000] <Lachy>
can anyone recommend an easy to use svn client for OSX?

[13:27:01.0000] <gsnedders>
/me resists the temptation to say the CLI client :P

[13:28:00.0000] <gsnedders>
(honestly, having spent a year or so using various GUIs, I concluded the CLI was by far the best)

[13:28:01.0000] <Lachy>
CLI = command line interface?

[13:29:00.0000] <gsnedders>
yeah

[13:29:01.0000] <Lachy>
I'll have to learn how to use it properly

[13:30:00.0000] <Lfe>
If you're into TextMate - it has a svn bundle which is (imo) "ok"

[13:30:01.0000] <virtuelv>
/me is tempted to say "Don't use OS X"

[13:31:00.0000] <Dashiva>
/me is tempted to say "lol"

[13:31:01.0000] <virtuelv>
but yeah, I'd go with the command line client

[13:31:02.0000] <virtuelv>
'svn help' is a good place to start

[13:31:03.0000] <gsnedders>
(I'd go with the command line client on any POSIX OS)

[13:32:00.0000] <virtuelv>
gsnedders: well, for most of my Ubuntu use, I've found both esvn and rapidsvn to be good

[13:32:01.0000] <gsnedders>
virtuelv: I don't really like either :P

[13:34:00.0000] <Philip`>
There's KDESvn too

[13:34:01.0000] <Philip`>
which seems quite like TortoiseSVN, though less good

[13:37:00.0000] <Lachy>
TortoiseSVN is awesome. That's what I use on windows

[13:38:00.0000] <Lachy>
one option is that I just run TortoiseSVN on Windows using VMWare on my mac

[13:38:01.0000] <Lachy>
though it might not be such a bad idea to learn the command line client

[13:38:02.0000] <Philip`>
You could use Mercurial instead of SVN, and then you wouldn't be tempted by all these fancy graphical tools

[13:38:03.0000] <Hixie>
gsnedders: pong

[13:39:00.0000] <gsnedders>
Hixie: see what I said yesterday about Content-Type, or should I repeat it?

[13:39:01.0000] <gsnedders>
yeah, Mercurial > SVN

[13:39:02.0000] <Philip`>
Lachy: The command line is basically "svn up", "svn diff | less", "svn commit", so it's not that hard :-)

[13:40:00.0000] <Hixie>
gsnedders: hmm

[13:41:00.0000] <virtuelv>
Philip`: sometimes you don't have a choice of VCS

[13:41:01.0000] <gsnedders>
http://krijnhoetmer.nl/irc-logs/whatwg/20071202#l-202

[13:41:02.0000] <Hixie>
gsnedders: yeah i saw the question, was just thinking about it

[13:42:00.0000] <gsnedders>
Hixie: just for anyone else interested, more than anything, I thought that was what the hmm was :)

[13:42:01.0000] <Hixie>
:-)

[13:42:02.0000] <Hixie>
gsnedders: i would say that on the long run it is something we'd want to in the http spec, but that on the short run on a purely pragmatic basis we probably want it in the html5 spec for now, since it'll get more review that way

[13:45:00.0000] <gsnedders>
/me sighs at the irony: to reverse engineer error handling needed for HTTP an HTTP implementation is needed :\

[13:50:00.0000] <Dashiva>
It's hard to reverse engineer something without an implementation :)

[13:50:01.0000] <gsnedders>
That's true.

[13:50:02.0000] <Philip`>
When you don't have an implementation but need one, that's just engineering

[13:52:00.0000] <gsnedders>
Now, a dilemma: IIS is "quite strict" because of earlier security holes, but other servers aren't; clients aren't strict at all — do I use the same code-path for both requests and responses which isn't strict and risk IIS never implementing it?

[13:52:01.0000] <gsnedders>
Or do I add a second code-path for requests that is more strict?

[13:54:00.0000] <gsnedders>
Hixie: maybe I'll merge it in after -00 is published

[13:54:01.0000] <Philip`>
By "risk", I assume you mean "increase the already-significant risk of"?

[13:55:00.0000] <gsnedders>
Philip`: I got the impression that if it doesn't make it more lax (or, possibly slightly moreso than currently), they probably would

[13:55:01.0000] <Philip`>
gsnedders: Ah, okay, that sounds reasonable

[13:56:00.0000] <gsnedders>
(But that's purely my impression from private discussions, so it may be totally wrong :P)

[14:00:00.0000] <gsnedders>
Part of me wants to only have one code-path as that will reduce the attack area, but knowing that it won't ever be used by IIS is big issue with that.

[14:00:01.0000] <gsnedders>
It's also less work for me, which is always nice

[14:13:00.0000] <hsivonen>
so I gave a presentation on HTML5 to a handful of interested people

[14:14:00.0000] <hsivonen>
among the feedback I got:

[14:14:01.0000] <hsivonen>
HTML5 is already a couple of years late considering competition from Flash and Silverlight

[14:14:02.0000] <hsivonen>
and

[14:15:00.0000] <hsivonen>
concern about the W3C having dropped the development of browser-relevant stuff (except CSS) for quite a while

[14:15:01.0000] <Hixie>
yup, that's why we started

[14:16:00.0000] <Lachy>
I've tried the svn command line, but when I try "svn commit", this is the error I get:

[14:16:01.0000] <Lachy>
svn: Commit failed (details follow):

[14:16:02.0000] <Lachy>
svn: Could not use external editor to fetch log message; consider setting the $SVN_EDITOR environment variable or using the --message (-m) or --file (-F) options

[14:16:03.0000] <Lachy>
svn: None of the environment variables SVN_EDITOR, VISUAL or EDITOR is set, and no 'editor-cmd' run-time configuration option was found

[14:16:04.0000] <Hixie>
because the w3c wasn't doing anything and we saw that we were getting competition and we had to response

[14:17:00.0000] <Philip`>
hsivonen: Were they already concerned, or were they unaware of problems and only became concerned because of your presentation?

[14:17:01.0000] <hsivonen>
Philip`: unclear

[14:17:02.0000] <Philip`>
Lachy: It's lucky that the error message tells you several ways to solve the error yourself ;-)

[14:18:00.0000] <hsivonen>
Philip`: I got a feeling that the audience didn't have a precise idea about the events that have lead to HTML5

[14:18:01.0000] <Lachy>
how do I set the $SVN_EDITOR environment variable and what do I set it to?

[14:19:00.0000] <Hixie>
Lachy: just use the -m parameter

[14:19:01.0000] <hsivonen>
Philip`: but I also got the impression that people had observed spec stagnation on their own

[14:19:02.0000] <Hixie>
hsivonen: get them to speak to the "html4 is fine, why are we changing it" camp

[14:20:00.0000] <hober>
and the "html4 was big enough, we should dump the new parts of html5 (like parsing)" camp

[14:20:01.0000] <Lachy>
Hixie, thanks

[14:20:02.0000] <Philip`>
Lachy: Hmm, I can't remember how you set it on OS X, but it's probably via some command in a ~/.bashrc or something

[14:20:03.0000] <Philip`>
(I can't remember how you do it on Linux either)

[14:20:04.0000] <Lachy>
-m worked

[14:22:00.0000] <Philip`>
Lachy: I like having the external editor thing so it can show what files it's going to check in, and I can stop it if it's doing something stupid, but it's easy to revert mistakes so that's probably just me being unnecessarily fussy :-)

[14:22:01.0000] <Hixie>
that's what 'svn diff' is for

[14:22:02.0000] <Hixie>
i always diff before committing

[14:23:00.0000] <jgraham_>
Lachy: presumably export SVN_EDITOR=emacs or something would work, no? 

[14:23:01.0000] <Philip`>
Hixie: That introduces a race condition - my other self might change some files in the meantime and I wouldn't notice

[14:23:02.0000] <Philip`>
jgraham_: s/emacs/vim/

[14:23:03.0000] <hsivonen>
oh and most of the questions I got were about video codec patent licensing, MPEG-LA and codec pluggability in upcoming browsers

[14:24:00.0000] <gsnedders>
jgraham: yeah, and s/emacs/nano/ :)

[14:24:01.0000] <jgraham_>
nano?!

[14:24:02.0000] <gsnedders>
nano.

[14:24:03.0000] <Hixie>
nano is awesame

[14:25:00.0000] <Hixie>
for small edits

[14:25:01.0000] <hsivonen>
jgraham_: GNU pico

[14:25:02.0000] <gsnedders>
totally awesame.

[14:25:03.0000] <Philip`>
I like how Gentoo encourages nano by default

[14:25:04.0000] <Hixie>
but emacs is what one has to use for real work

[14:25:05.0000] <jgraham_>
I guess for this kind of thing nano is actually quite sensible

[14:25:06.0000] <hober>
for small edits, mg is pretty nice, since it's so similar to emacs

[14:25:07.0000] <Philip`>
since it's actually reasonably easy to use

[14:25:08.0000] <Hixie>
Philip`: i actually have that problem with the html5 spec, because the spec generation takes so long that i edit the spec before it's done (and before i see the diff to commit)

[14:25:09.0000] <gsnedders>
hsivonen: GNU pico? doesn't exist…

[14:26:00.0000] <gsnedders>
hsivonen: GNU nano is a pico clone

[14:26:01.0000] <hsivonen>
gsnedders: nano is the GNU approximation of pico

[14:26:02.0000] <Hixie>
that's what he meant

[14:26:03.0000] <gsnedders>
meh.

[14:26:04.0000] <gsnedders>
not clear.

[14:26:05.0000] <Philip`>
Not a very good approximation - they were off by three orders of magnitude

[14:26:06.0000] <gsnedders>
Philip`: which is what confused me so much.

[14:26:07.0000] <jgraham_>
/me has accidentally started an editor war 

[14:26:08.0000] <Hixie>
Philip`: i worked around that by working on a different file, which is copied over to my working copy when i generate the spec, so that even if i do edit the spec, the file that's diffed is the one i commit :_)

[14:26:09.0000] <hsivonen>
gsnedders: I thought jgraham_ was asking what nano was 

[14:27:00.0000] <hsivonen>
/me has trouble telling apart rhetorical and actual question marks at this hour

[14:27:01.0000] <hsivonen>
bed time

[14:27:02.0000] <hsivonen>
nn

[14:27:03.0000] <Hixie>
nn

[14:27:04.0000] <gsnedders>
Hixie: what does the generation actually do?

[14:27:05.0000] <gsnedders>
/me waves to hsivonen 

[14:27:06.0000] <Hixie>
gsnedders: cross-refs and section numbering, mostly

[14:28:00.0000] <Hixie>
gsnedders: also, sticking on the header

[14:28:01.0000] <gsnedders>
ah

[14:28:02.0000] <gsnedders>
anyone have any thoughts about using two code paths instead of one for HTTP request/response parsing?

[14:34:00.0000] <Philip`>
gsnedders: Would that be written as two separate versions of the parser, or would it work as one with "If (some condition), this is an error. Do (something crazy for legacy compatibility)" and a global "If you are a strict parser, stop at the first error"?

[14:35:00.0000] <gsnedders>
Philip`: for the tokeniser, the latter. For the actual parsing of header values, it would need to be two totally different algorithms.

[14:38:00.0000] <gsnedders>
(actually, for the parsing of header values, I could probably get away with, "There is an error if any header/value pair doesn't match its [RFC 2616] definition, with the following exceptions: …")

[14:39:00.0000] <gsnedders>
so it probably is doable without _too_ much extra text

[14:40:00.0000] <Philip`>
/me wonders why Thunderbird is complaining that cpop.corp.google.com's certificate expired two years ago

[14:41:00.0000] <gsnedders>
Also, seeming IIS gets away with using a strict parser, it would probably with worth saying that, "It is RECOMMENDED that requests are parsed with a strict parser, and that responses are parsed with a weak/lax(?) parser."

[14:41:01.0000] <Hixie>
/me wonders how Philip`'s computer can even get to a cert on cpop.corp.google.com

[14:42:00.0000] <gsnedders>
/me wonders whether weak or lax would be better

[14:43:00.0000] <Philip`>
gsnedders: "liberal"?

[14:43:01.0000] <gsnedders>
or follow RFC 3986's example of "non-strict" :)

[14:43:02.0000] <gsnedders>
(which means I only need to name one)

[14:44:00.0000] <Philip`>
Hixie: I have no idea, but I'm just downloading from pop.googlemail.com like normal, so it must be someone else's fault :-)

[14:44:01.0000] <Hixie>
weird

[14:44:02.0000] <gsnedders>
/me applies a SEP field around Gmail

[14:44:03.0000] <gsnedders>
(and paints it pink, otherwise it wouldn't work)

[14:46:00.0000] <Philip`>
http://groups.google.com/group/Gmail-Problem-solving/browse_thread/thread/137a15fe56fc1c2c?hl=en - ah, sounds like not just me

[14:47:00.0000] <gsnedders>
anyhow, enough talking to myself (more or less). If anyone has any comments on the HTTP stuff, feel free to drop me an email.

[14:47:01.0000] <gsnedders>
/me falls alseep

[14:48:00.0000] <Philip`>
The "continue" buttons on these certificate security warning boxes always look very tempting since they make the visible problem go away immediately


2007-12-04
[18:39:00.0000] <hober>
Lachy: congrats on the ALA publication!

[19:09:00.0000] <gavin_>
/me gets the same corp.google certificate error as Philip`

[19:09:01.0000] <gavin_>
/me has a feeling one of Google's POP servers is busted

[19:59:00.0000] <jwalden>
Lachy: question about http://www.alistapart.com/articles/previewofhtml5 -- why no link to or mention of http://www.double.co.nz/video_test/ and Mozilla's experimental builds?

[21:06:00.0000] <Hixie>
man, uris are hard to parse

[21:13:00.0000] <Hixie>
wow i'm still getting mail from thursday

[21:18:00.0000] <othermaciej>
I wonder how close the latest URI/IRI RFCs are to what it actually takes to parse real-world URLs in web content

[14:51:00.0000] <hendry>
i am a bit confused about <button>. should I be using it over input type=submit ?

[15:06:00.0000] <Hixie>
hendry: use whichever you want

[15:06:01.0000] <Hixie>
they're the same, basically

[15:07:00.0000] <gsnedders>
/me goes mad at his internet connection. can't push http-parsing tests to the web. stuck on my HD :\

[15:07:01.0000] <Dashiva>
input is simpler, button allows rich content (ooh, aah)

[15:15:00.0000] <hendry>
though i guess button isn't well supported by IE. not sure. distant memories

[15:20:00.0000] <Dashiva>
hendry: That's if you use the value

[15:26:00.0000] <hendry>
i wonder if i can have bg images in select's option forms

[15:28:00.0000] <hendry>
http://lab.artlung.com/form-image-background/ # hah, you can


2007-12-05
[17:06:00.0000] <Hixie>
lachy's post sure has gotten a lot of traction

[17:10:00.0000] <othermaciej>
on a list apart?

[17:10:01.0000] <Hixie>
yeah

[17:11:00.0000] <othermaciej>
it seemed really well geared towards content authors and yet with a lot of technical meat

[17:19:00.0000] <othermaciej>
people seem to flip out at the time estimate

[17:19:01.0000] <othermaciej>
I think for PR purposes it may be useful to have time estimates for earlier milestones than REC

[17:20:00.0000] <MikeSmith>
othermaciej - yeah, definitely. Many seem to assume that browser vendors will just be sitting around waiting for the complete spec to be published as REC

[17:20:01.0000] <Hixie>
i tried updating the faq recently

[17:20:02.0000] <Hixie>
and i'm actively working on the spec annotation script

[17:20:03.0000] <Hixie>
other suggestions welcome

[17:21:00.0000] <MikeSmith>
I'll try to see if I can add something on http://www.w3.org/html/wg/

[17:25:00.0000] <othermaciej>
what time estimates are plausibly possible to generate?

[17:25:01.0000] <othermaciej>
ideally we should be able to estimate FPWD with high precision in the not too distant future

[17:26:00.0000] <othermaciej>
"feature complete" would also be a good milestone, however rough

[17:26:01.0000] <othermaciej>
and "first LC" might be a good one too

[17:30:00.0000] <Dashiva>
Wasn't feature complete last month? :)

[17:32:00.0000] <Dashiva>
I thought alistapart was supposed to be part of the semantic faction, but all I see is people glorifying div soup in the comments...

[17:32:01.0000] <othermaciej>
negative commentary seems to be along the lines of either "omg, 15 years? I'll be dead by then"

[17:33:00.0000] <Philip`>
Do the stresses of web design have that much of an effect on life expectancy?

[17:33:01.0000] <othermaciej>
that or "wtf? why new elements instead of div soup?"

[17:33:02.0000] <Hixie>
my timetable is:

[17:33:03.0000] <Hixie>
 [X] First Working Draft in October 2007.

[17:33:04.0000] <Hixie>
 [ ] Last Call Working Draft in October 2009.

[17:33:05.0000] <Hixie>
 [ ] Call for contributions for the test suite in 2011.

[17:33:06.0000] <Hixie>
 [ ] Candidate Recommendation in 2012.

[17:33:07.0000] <Hixie>
 [ ] First draft of test suite in 2012.

[17:34:00.0000] <Hixie>
 [ ] Second draft of test suite in 2015.

[17:34:01.0000] <Hixie>
 [ ] Final version of test suite in 2019.

[17:34:02.0000] <Hixie>
 [ ] Reissued Last Call Working Draft in 2020.

[17:34:03.0000] <Hixie>
 [ ] Proposed Recommendation in 2022.

[17:34:04.0000] <Hixie>
(the [X] indicates the milestones that whatwg hit)

[17:34:05.0000] <Dashiva>
Well, then we should definitely be saying 2012

[17:35:00.0000] <Dashiva>
CSS2.1 seems to be doing fine at CR

[17:35:01.0000] <othermaciej>
I think "Call for contributions to the test suite" could be earlier

[17:35:02.0000] <othermaciej>
in fact I think it could be now

[17:36:00.0000] <othermaciej>
(with proviso that parts of the test suite might get invalidated by spec changes)

[17:36:01.0000] <Hixie>
only if you volunteer to rereview them when the spec changes from underneath them

[17:36:02.0000] <MikeSmith>
I wonder if what might help is to also show a timeline of implementation status in browsers

[17:36:03.0000] <Hixie>
(i don't mind getting tests now, i just don't think we should actually _call_ for tests yet)

[17:36:04.0000] <othermaciej>
the people who were around when we decided to start on the test suite that we thought would make good test suite reviewers agreed to do that

[17:36:05.0000] <MikeSmith>
e.g., show that canvas has already been implemented, and when

[17:36:06.0000] <Dashiva>
"The WebKit team has told the HTML WG that they want to implement the entirety of the HTML 5 spec in the very near future."

[17:36:07.0000] <Dashiva>
Did you actually say that, othermaciej?

[17:37:00.0000] <othermaciej>
(hsivonen, anne, Lachy, zcorpan, can't remember who else)

[17:37:01.0000] <Hixie>
(call for tests = going out and getting people on slashdot and reddit and co to volunteer to write tests)

[17:37:02.0000] <MikeSmith>
video, client-side DB storage in Webkit, etc.

[17:37:03.0000] <othermaciej>
Dashiva: that would be a bit of an overstatement

[17:37:04.0000] <othermaciej>
Dashiva: we are indeed implementing large parts, but I don't think I would say "entirety" or "very near future"

[17:37:05.0000] <Hixie>
MikeSmith: the data from the annotation script will be minable to get that kind of view dynamically

[17:37:06.0000] <MikeSmith>
Hixie - cool

[17:38:00.0000] <Dashiva>
othermaciej: http://www.alistapart.com/comments/previewofhtml5?page=4#33

[17:38:01.0000] <Dashiva>
Maybe you should comment before it becomes fact :)

[17:39:00.0000] <MikeSmith>
though having a static copy posted prominently might be good too

[17:39:01.0000] <Philip`>
/me wonders why anyone would volunteer to write tests

[17:39:02.0000] <Philip`>
It's usually kind of hard and takes lots of time and concentration and you don't get anything out of it except the ability to show your immense superiority over browser developers who keep failing your tests

[17:39:03.0000] <othermaciej>
I guess I should post, but I'm not sure how

[17:39:04.0000] <othermaciej>
do I need to register for an account first?

[17:39:05.0000] <Dashiva>
Philip`: That last part is it

[17:40:00.0000] <Dashiva>
Except it's kinda dual. For some people it's "Haha, I'm smarter than them" and for some it's "I've found a way to help improve the browse"

[17:40:01.0000] <Dashiva>
+r

[17:42:00.0000] <Hixie>
i updated the faq again

[17:46:00.0000] <MikeSmith>
othermaciej - yeah, you need to register first I think

[17:50:00.0000] <MikeSmith>
ability to do test-case development well seems very underappreciated and under-rewarded relative to its importance

[17:51:00.0000] <othermaciej>
in WebKit it is so valued that it's mandatory

[17:51:01.0000] <othermaciej>
(though admittedly regression test generation is not quite the same art as standards compliance test generation)

[17:51:02.0000] <Philip`>
I think the HTML WG should pay people $10 for every test case they write and have successfully reviewed

[17:52:00.0000] <Hixie>
and this money would come from...?

[17:52:01.0000] <Philip`>
That's somebody else's problem ;-)

[17:52:02.0000] <Hixie>
:-)

[17:52:03.0000] <Dashiva>
Is it also pink?

[17:54:00.0000] <MikeSmith>
maybe we should come up with some karma system. ++people who contribute test cases and good LC comments that actually result in improvements to a particular spec

[17:54:01.0000] <MikeSmith>
--people who [fill in the blank]

[17:54:02.0000] <Philip`>
We could use Mechanical Turk

[17:54:03.0000] <MikeSmith>
and we keep a Leader Board

[17:55:00.0000] <othermaciej>
we're thinking of doing an "HTML5 implementation status in WebKit, plus why we think it's kinda cool" blog post on Surfin' Safari

[17:55:01.0000] <Hixie>
MikeSmith: the html5 acknowledgments list is a list of people who have said at least one thing that has resulted in a change to the spec

[17:55:02.0000] <Hixie>
MikeSmith: (whether they asked for the change or not)

[18:02:00.0000] <MikeSmith>
Hixie - hopefully later drafts might could also include people who contribute test cases

[18:03:00.0000] <Hixie>
we'll see. (generally i think it's best to have them acknowledged on the test suite page)

[18:03:01.0000] <Dashiva>
So does that mean I'll get added retroactively if mouseenter/leave are added sometime?

[18:03:02.0000] <Hixie>
added retroactively?

[18:03:03.0000] <Dashiva>
Since I sent the mail like half a year ago, but it's going to take at least another six months before it's at the front of your queue :)

[18:03:04.0000] <Hixie>
yes, as i respond to mails i add people to the list

[18:04:00.0000] <Hixie>
but aren't you in the list already?

[18:04:01.0000] <MikeSmith>
/me realizes "might could" is ungrammatical, but lived in Texas where it's used all the time

[18:04:02.0000] <Hixie>
(as i respond to mails i add people to the list _if their mail results in a change_)

[18:04:03.0000] <Dashiva>
Don't think so, no

[18:04:04.0000] <Hixie>
huh

[18:05:00.0000] <Dashiva>
I've done spelling corrections and such, mostly

[18:05:01.0000] <Hixie>
oh i haven't gone through those yet

[18:06:00.0000] <Dashiva>
There are no acknowledgements of the vast browser-wing cabal either, just the regular cabal

[18:07:00.0000] <Hixie>
there are acknowledgements of the cabal?

[18:07:01.0000] <Dashiva>
The very last one

[18:10:00.0000] <Hixie>
oh that's not the cabal you think it is

[18:10:01.0000] <Hixie>
that's funny

[18:11:00.0000] <gavin>
is it the cabal that used to hang out on irc.m.o? :)

[18:11:01.0000] <Hixie>
yeah :-)

[18:12:00.0000] <Hixie>
well, one of them

[18:12:01.0000] <Hixie>
there were several over the years

[18:12:02.0000] <Dashiva>
Is there like a cabal of cabals, I wonder

[18:12:03.0000] <Hixie>
yes, but they don't know

[18:16:00.0000] <MikeSmith>
the Illuminati

[18:33:00.0000] <dglazkov>
mmm.... cabal

[18:34:00.0000] <Philip`>
Cabal broadband?

[18:35:00.0000] <dglazkov>
Isn't that the capital of Afghanistan?

[18:38:00.0000] <dglazkov>
Hixie, have you had a chance to look at the HTML5 player that I built?

[18:41:00.0000] <Hixie>
no

[18:41:01.0000] <Hixie>
uri?

[18:41:02.0000] <dglazkov>
http://attic.glazkov.com/player/

[18:43:00.0000] <Hixie>
nice

[18:43:01.0000] <dglazkov>
I posted on WHATWG about it a while back, so...

[18:44:00.0000] <dglazkov>
I am wondering if this is needed for anyone

[18:44:01.0000] <dglazkov>
'cause it is, there's still plenty of holes to patch

[18:44:02.0000] <Hixie>
yeah i think that went straight to my sql folder for future reading :-)

[18:44:03.0000] <dglazkov>
I meant "'cause _if_ it is.."

[18:51:00.0000] <dglazkov>
/me wonders if WebKit folks are breathing down his pioneering neck with their implementation

[19:43:00.0000] <othermaciej>
I deny any heavy breathing on the part of apple

[21:53:00.0000] <Hixie>
Gregory's comments frighten me further at the state of AT technology

[21:53:01.0000] <Hixie>
what's the point of adding semantic markup if AT users key of the _screen media CSS_ ?

[10:49:00.0000] <dglazkov>
wouldn't it be cool if I could just point my client to the IRC log server and have uninterrupted flow of conversation upon entry?

[10:53:00.0000] <rubys>
hsivonen: ping?

[13:14:00.0000] <othermaciej>
this article is so much more positive than the title: http://warpspire.com/features/html5-css3/

[14:45:00.0000] <jruderman>
"Recap: It’s 2008. We still don’t have full support for specifications released in 1998 and 2000"

[14:45:01.0000] <jruderman>
why do so many people think it's 2008?

[14:45:02.0000] <jruderman>
i saw an "As of 2008" in a wikipedia article the other day

[14:45:03.0000] <jruderman>
it's still 2007, right? i'm not crazy?

[14:46:00.0000] <Dashiva>
Well, cars and fashion and whatnot are in 2008 now

[14:46:01.0000] <mpt>
What's a "developer-developer"?

[14:49:00.0000] <hsivonen>
/me points out that the first release of HTML 4.0 was in 1997

[14:50:00.0000] <hsivonen>
so it's already a decade without waiting until 2008

[14:51:00.0000] <othermaciej>
jruderman: maybe they're using the Julian calendar?


2007-12-06
[16:49:00.0000] <Hixie>
heh, i just did a study to look at the average page rank of pages with different kinds of scripts

[16:49:01.0000] <Hixie>
and pages with rollovers on average have low pagerank :-P

[22:17:00.0000] <Hixie>
maybe instead of "high quality" and "low quality" as conformance levels, we should have "high priority" and "low piority" errors

[22:17:01.0000] <Hixie>
where high priority = things that really cause a lot of problems and should definitely be fixed

[22:18:00.0000] <Hixie>
and low priority = things that cause problems, but not disasters

[22:18:01.0000] <Hixie>
so for example a missing </style> is a high priority error since it hides the whole rest of the page

[22:18:02.0000] <othermaciej>
wouldn't most content model violation bugs then be low priority?

[22:19:00.0000] <Hixie>
but a style="" attribute in a production page is a low priority error since it just means that you can't easily add an alternative style sheet

[22:20:00.0000] <othermaciej>
now there's a marginally good argument against using style="" (though it applies equally to <style scoped> and even <style>, no?)

[22:20:01.0000] <Hixie>
<style> can have a title=""

[22:20:02.0000] <Hixie>
(without one it's a "persistent style sheet")

[22:21:00.0000] <Hixie>
(but it's easy to migrate from one persistent style sheet to a set of alternatives just by adding titles)

[22:22:00.0000] <othermaciej>
that seems about as good a reason for warning as it would be to warn on onclick because you can't have more than one handler set that way, and additional ones get slightly different behavior (scope chain, etc)

[22:23:00.0000] <othermaciej>
i.e. not completely bogus, but if it were a compiler I'd probably turn that warning off

[22:25:00.0000] <othermaciej>
and if we're talking mandatory vs. optional diagnostics, then it might be better to leave it up to individual validators to decide what warnings they emit in addition to errors, and compete on usefulness and quality of those

[22:25:01.0000] <othermaciej>
much as C compilers all give basically the same syntax errors, but compete on usefulness of optional warnings

[22:26:00.0000] <Hixie>
yeah

[22:26:01.0000] <Hixie>
maybe

[22:26:02.0000] <othermaciej>
(I suspect gcc has far more useful warnings than it would if they were all defined by the C standard)

[22:26:03.0000] <othermaciej>
it wouldn't hurt for the spec to editorially suggest some things as possible optional diagnostics

[22:27:00.0000] <othermaciej>
I should get going

[22:27:01.0000] <Hixie>
ttyl

[22:27:02.0000] <othermaciej>
(btw some people might find a warning on use of inline event handler attributes actually useful)

[22:27:03.0000] <othermaciej>
(since it's not uncommon to ban them as a house style for a particular site)

[22:44:00.0000] <Hixie>
i have no idea if sql does what i want it to do here

[22:44:01.0000] <Hixie>
but i guess i'll find out

[22:47:00.0000] <MikeSmith>
Hixie - sql for the status-annotations server-side stuff?

[22:47:01.0000] <Hixie>
yeah

[22:47:02.0000] <Hixie>
i'm doing:

[22:47:03.0000] <MikeSmith>
you got the database schema put together now?

[22:47:04.0000] <Hixie>
SELECT `annotation`, `section` FROM `section` GROUP BY `annotation` ORDER BY `change` DESC

[22:48:00.0000] <Hixie>
with a table that has annotation, section, change fields

[22:48:01.0000] <Hixie>
and i'm hoping that it'll give me the most recent section for each annotation

[22:49:00.0000] <MikeSmith>
seems like it will if the data is actually what it should be

[22:49:01.0000] <Hixie>
how about:

[22:49:02.0000] <Hixie>
SELECT `annotation`, `key`, `value`, CONCAT(`annotation`, ':', `key`) AS `group` FROM `implementations` GROUP BY `group` ORDER BY `change` DESC

[22:49:03.0000] <Hixie>
...giving me the most recent value for each annotation-key pair

[22:50:00.0000] <MikeSmith>
what is in "key"? just a generated id?

[22:50:01.0000] <Hixie>
key is a string

[22:50:02.0000] <Hixie>
one of 'firefox', 'ie', 'safari', 'opera

[22:50:03.0000] <Hixie>
'

[22:50:04.0000] <MikeSmith>
oh

[22:52:00.0000] <MikeSmith>
that don't look quite right to me

[22:52:01.0000] <Hixie>
oh?

[22:53:00.0000] <MikeSmith>
/me takes a look at mysql reference

[22:54:00.0000] <Hixie>
it's syntactically correct

[22:54:01.0000] <Hixie>
i just don't know if it's correct :-)

[22:54:02.0000] <MikeSmith>
ah, ok

[22:55:00.0000] <Dashiva>
hm

[22:57:00.0000] <Dashiva>
I suspect it would work because of how MySQL does things, even though it's not technically a reliable feature

[22:57:01.0000] <Hixie>
how would you do it reliably in sql?

[22:58:00.0000] <Dashiva>
I can only think of ugly ways at the moment (subqueries and whatnot)

[22:59:00.0000] <Hixie>
so yeah, like i said, i dunno if sql can do what i want of it :-)

[23:00:00.0000] <Dashiva>
Let's see how un-ugly I can imagine it

[23:04:00.0000] <MikeSmith>
Hixie - I think the high-priority vs low-priority distinction could be useful

[23:06:00.0000] <MikeSmith>
would be really useful to try to get some more implementors to work on conformance checkers, and for people to use those

[23:06:01.0000] <Hixie>
webkit is starting to have more and more of one

[23:06:02.0000] <MikeSmith>
in order to get some real feedback

[23:06:03.0000] <Hixie>
the iphone can even tell you about html parse errors

[23:06:04.0000] <MikeSmith>
nice

[23:07:00.0000] <MikeSmith>
so I think one real risk is that conformance-checker maintainers would start adding "ignore someClassofErrors" options

[23:08:00.0000] <MikeSmith>
due to pressure from users

[23:08:01.0000] <Hixie>
how much i care depends on what the error is

[23:08:02.0000] <Hixie>
or rather, what i think about it

[23:09:00.0000] <MikeSmith>
yeah, of course

[23:09:01.0000] <Hixie>
if it's because the spec has some requirement that really doesn't help anyone, then we should change the spec

[23:09:02.0000] <Hixie>
if the error is there because they're screwing users, then...

[23:13:00.0000] <Dashiva>
Hixie: If your MySQL is new enough to support subqueries, this should work

[23:13:01.0000] <Dashiva>
SELECT annotation, section FROM section WHERE (annotation, change) IN ( SELECT annotation, max(change) FROM section GROUP BY annotation );

[23:13:02.0000] <Dashiva>
I hope there's an index of annotation :)

[23:14:00.0000] <Hixie>
yeah

[23:14:01.0000] <Hixie>
there are indexes all over the place

[23:15:00.0000] <Hixie>
what on earth does that do? :-)

[23:15:01.0000] <Hixie>
fwiw i'm using mysql with VERSION() == 4.1.16-standard-log

[23:16:00.0000] <doublec>
Quick <video> question. If there are no <source> elements that are picked, and there is therefore no media resource, can a non-javascript enabled browser do anything meaningful to let the user know?

[23:16:01.0000] <Dashiva>
It gets the newest date for each annotation, and then selects the rows containing those annotation,date pairs

[23:17:00.0000] <Hixie>
wow

[23:17:01.0000] <Hixie>
that's some fancy syntax right there

[23:17:02.0000] <Hixie>
looks like this version of mysql doesn't do that

[23:17:03.0000] <Dashiva>
Yeah, MySQL only started doing subqueries recently. Might be 5.0

[23:18:00.0000] <Dashiva>
Could no doubt emulate it with a join on itself and so fancy tricks, though 

[23:18:01.0000] <Hixie>
you know, maybe i should have implemented the code that actually _adds_ annotations first

[23:18:02.0000] <Hixie>
leaving it til last is somewhat preventing me from actually _testing_ anything

[23:19:00.0000] <Hixie>
well we'll see if i just get what i need by luck first

[23:24:00.0000] <Hixie>
/me realises he's actually completely left off the ability to add sections, both in his UI, in his client-side logic layer, in his client-side network layer, in his server-side network layer, and in his server-side database logic

[23:24:01.0000] <Hixie>
"oops"

[23:27:00.0000] <maikmerten>
/me gives moral support to doublec's question

[23:27:01.0000] <MikeSmith>
Hixie - you get points for consistency at least

[23:28:00.0000] <maikmerten>
I was under the impression that if <video> fails it should display the fallback content to be fed into non-<video> browsers

[23:29:00.0000] <doublec>
does having no media resource count as fail though?

[23:29:01.0000] <doublec>
as they might add <source> dynamically later

[23:29:02.0000] <Hixie>
with no <source>, or no suitable media, you display a black rectangle

[23:29:03.0000] <Hixie>
according to the spec

[23:30:00.0000] <Hixie>
in a <video>-supporting UA, you never see the fallback

[23:30:01.0000] <Hixie>
again, according to the spec

[23:30:02.0000] <doublec>
good point

[23:30:03.0000] <Hixie>
(sorry, didn't see your question earlier)

[23:30:04.0000] <Hixie>
to answer your question, the UA is always allowed to be helpful

[23:31:00.0000] <Hixie>
it could, e.g., overlay a message on the <video>'s black screen saying "No media available..."

[23:31:01.0000] <Hixie>
there's no restriction on what the UI is or can do really

[23:31:02.0000] <Hixie>
and when JS is disabled, controls="" is implied, so the UI should be in full swing

[23:32:00.0000] <Hixie>
doublec: hth

[23:32:01.0000] <Hixie>
gonna go home now, will probably be online again in a bit

[23:32:02.0000] <doublec>
cool, thanks

[23:33:00.0000] <doublec>
where this came up is for user agents that don't support Ogg. Can a page detect that no meda resource was available and report a message themselves.

[00:11:00.0000] <Hixie>
doublec: yeah, there's an event

[00:11:01.0000] <Hixie>
iirc

[00:17:00.0000] <doublec>
ahh, the abort event, thanks

[04:07:00.0000] <Hixie>
http://www.w3.org/mid/bgdfl31rqfj95uhpkkjkmg0kc1bbg6612v⊙hbhd <- hahahaha, awesome, microsoft caught at it again

[04:07:01.0000] <Hixie>
hopefully that puts the final nail into the EOT coffin

[04:20:00.0000] <om_sleep>
harsh

[04:21:00.0000] <othermaciej>
now wait a second

[04:21:01.0000] <othermaciej>
if that's true, I don't understand why Microsoft was pushing EOT so hard

[04:21:02.0000] <othermaciej>
I assumed it was because they're a font vendor

[04:24:00.0000] <hsivonen>
othermaciej: perhaps the MS Core Fonts initiative has lead to some internal discussions

[04:25:00.0000] <hsivonen>
othermaciej: Paul Nelson's www-style future tense statement of what MS won't do is something they have done in the past

[04:25:01.0000] <othermaciej>
hsivonen: I don't understand how that would explain the seeming inconsistency in the message Hixie linked

[04:25:02.0000] <othermaciej>
(inconsistency demonstrated by the message, that is)

[04:25:03.0000] <hsivonen>
it doesn't

[04:26:00.0000] <hsivonen>
too bad that message is Member-only

[04:26:01.0000] <othermaciej>
you have Member access, right?

[04:26:02.0000] <hsivonen>
otherwise, it would be a good "recent successes" addition to the Björn Höhrmann Project :-)

[04:26:03.0000] <hsivonen>
othermaciej: I do

[04:27:00.0000] <othermaciej>
I'd guess Bjoern wouldn't mind posting it somewhere public too

[04:28:00.0000] <othermaciej>
hsivonen: what www-style statement are you referring to by the way?

[04:29:00.0000] <hsivonen>
othermaciej: http://www.w3.org/mid/49C257E2C13F584790B2E302E021B6F914F6B1D2⊙wswcmc

[04:30:00.0000] <hsivonen>
later amended, though: http://www.w3.org/mid/49C257E2C13F584790B2E302E021B6F914F6B1DD⊙wswcmc

[04:32:00.0000] <hsivonen>
the network effects, the development cost and the ways of making money on fonts do make fonts a hard problem for the Web and for any non-final form documents

[04:36:00.0000] <hsivonen>
I don't really know how to make fonts so that font designers get paid but users aren't bound by onerous licensing terms except developing free fonts under a patronage model

[04:36:01.0000] <hsivonen>
But that yields a practically necessary number of fonts

[04:37:00.0000] <hsivonen>
and the embedding issue is pretty much about high-quality fonts that go beyond what would be merely necessary for textual communication to work

[04:40:00.0000] <othermaciej>
whatever the complexities of the font business model, it seems clear that using a Microsoft-specific format with trivial DRM doesn't solve any of the problems and creates extra annoyance for web developers and browser developers, and even Microsoft seems to acknowledge that for technologies that they'd like to see succeed

[04:40:01.0000] <hsivonen>
I agree

[04:52:00.0000] <hsivonen>
clearly, making the font restriction too tight will drive a foundry out of business

[04:53:00.0000] <hsivonen>
a few years ago, I found a small private foundry that had great designs but wouldn't license them for embedding in PDF

[04:53:01.0000] <hsivonen>
it seems to me the foundry is out of business

[04:53:02.0000] <hsivonen>
the irony was, of course, that in order to communicate samples of their fonts, they had to embed them in PDFs

[04:54:00.0000] <othermaciej>
heh

[04:54:01.0000] <othermaciej>
fortunately they can't violate their own license, legally

[04:54:02.0000] <othermaciej>
but still amusing

[08:56:00.0000] <rubys>
is input/@size legal in html5?

[08:57:00.0000] <gsnedders>
rubys: no

[08:58:00.0000] <rubys>
what's the alternative?

[08:59:00.0000] <gsnedders>
CSS is just meant to be used instead

[09:44:00.0000] <csarven>
gsnedders how do non-presentational UAs allocate space for the input size?

[09:45:00.0000] <gsnedders>
csarven: non-presentational or non-CSS-supporting? non-presentational ones don't present anything :P

[09:46:00.0000] <csarven>
i was referring non-CSS-supporting UAs. e.g. Lynx

[09:47:00.0000] <gsnedders>
in some they always use the same width, others use @size

[09:48:00.0000] <csarven>
and if CSS is turned off CSS-supporting UAs? is it expected to default to browser stylesheet?

[09:49:00.0000] <gsnedders>
AFAIK they all use @size falling back to a default

[09:49:01.0000] <Dashiva>
I don't recall anyone switching width based on the font, though

[10:11:00.0000] <mpt>
/me has a vague memory of the meaning of size= being dependent on the width of an "e" in the relevant font

[10:14:00.0000] <gavin>
/me has a vague memory of a recent mozilla bug about mpt's vague memory

[10:23:00.0000] <Philip`>
/me has a memory of it being different in every browser

[10:42:00.0000] <rubys>
hsivonen: ping?

[10:54:00.0000] <annevk>
I think <input size> is probably something that should be allowed

[11:05:00.0000] <mpt>
<input size= has weak semantics in that it suggests approximately how long the entered data should be

[11:06:00.0000] <mpt>
(people will write less in a narrow field than a wide one)

[11:15:00.0000] <rubys>
any idea on how I can "recover" a nick?  nickserv doesn't seem to like the password I thought I used, and /info seems to indicate that this nick was registered by me over four years ago, and even used by me just over a week ago.

[11:19:00.0000] <gavin>
rubys: I'd just /msg someone in /stats p

[11:19:01.0000] <gavin>
this server's nickserv doesn't appear to have a "reset password and email it to me" option

[11:19:02.0000] <Dashiva>
Doesn't even ask for email on registration, it seems

[11:21:00.0000] <annevk>
gavin, I believe /msg does not work when you're not registered

[11:21:01.0000] <gavin>
for everyone? hrm

[11:21:02.0000] <annevk>
hsivonen, http://forums.whatwg.org/viewtopic.php?t=123 reports a bug in validator.nu (<form target> is marked non-conforming, but it is conforming)

[11:21:03.0000] <gavin>
I do seem to recall something like that

[11:21:04.0000] <gavin>
on other servers it's usually a off-by-default per-user switch

[12:07:00.0000] <annevk>
rubys, search forms should indeed be easier

[12:08:00.0000] <annevk>
that is, <form> <input type=text> <input type=submit> </form> should just be conforming

[12:10:00.0000] <rubys>
i think the problem is more general than that, the notion of "strictly inline content" probably should go away.

[12:11:00.0000] <annevk>
that might have been agreed on already

[12:11:01.0000] <annevk>
not sure if that solves the <form> thingie

[12:11:02.0000] <annevk>
btw, there's now http://html5.fr/ besides http://html5.jp/

[12:12:00.0000] <annevk>
(html5.fr made a validator button for validator.nu :) )

[12:15:00.0000] <Philip`>
Perhaps HTML5 could say that including a validator button on your web page is non-conforming

[12:15:01.0000] <Dashiva>
I made one too

[12:16:00.0000] <Dashiva>
Well, I stole one. But much the same :)

[12:16:01.0000] <Philip`>
It must be good for the validators' page-rank

[12:17:00.0000] <Dashiva>
http://dashiva.net/banner/valid-html5.png

[12:17:01.0000] <Dashiva>
I don't recall who made it

[12:24:00.0000] <annevk>
Simon I think

[12:24:01.0000] <rubys>
the downside of things that benefit the validator's page rank is that every crawler known to man will then proceed to pound your server to sawdust.

[13:37:00.0000] <annevk>
does anyone know how the document.all hiding works in Mozilla?

[13:38:00.0000] <gavin>
a bit

[13:39:00.0000] <annevk>
http://developer.mozilla.org/en/docs/DOM:document.all does not exist

[13:40:00.0000] <gavin>
do you have specific questions, or are you just looking for a comprehensive spec? :)

[13:41:00.0000] <annevk>
I wonder if document.all returns some quirky object that "booleanizes" to "false"

[13:41:01.0000] <annevk>
or if it's some other trickery

[13:42:00.0000] <Hixie>
hsivonen: yeah, they never did reply to http://lists.w3.org/Archives/Public/www-style/2007Nov/0274.html

[13:42:01.0000] <annevk>
and whether "booleanizes" is an actual JS concept or specifically added

[13:42:02.0000] <gavin>
other trickery

[13:42:03.0000] <Hixie>
/me sends a proposal to solve the inline/block problem

[13:42:04.0000] <Hixie>
but it basically requires us to stop allowing <ol> and co in paragraphs

[13:42:05.0000] <gavin>
I believe it hooks gets of document.all and returns "undefined" if it's in a "detecting" context

[13:44:00.0000] <annevk>
interesting

[13:44:01.0000] <gavin>
hmm

[13:44:02.0000] <gavin>
http://lxr.mozilla.org/seamonkey/source/dom/src/base/nsDOMClassInfo.cpp#7889

[13:45:00.0000] <annevk>
Hixie, you could make <ol> and <ul> imply <p> around them too

[13:46:00.0000] <gavin>
as I read that, it makes document.all exist if you try to get .length, .item(), or something that matches an id from it, but not exist otherwise

[13:46:01.0000] <annevk>
that file seems to beat HTML 5 in bytes

[13:47:00.0000] <gavin>
yeah, it's got a lot of icky DOM magic built into it :)

[13:48:00.0000] <Hixie>
annevk: but then you could never use <p> in <li>

[13:48:01.0000] <Hixie>
or <hx> in <li>

[13:48:02.0000] <Hixie>
etc

[13:49:00.0000] <annevk>
sorry, what I meant was that <aside> TEXT <ol> is like <aside> <p> TEST </p> <ol>

[13:49:01.0000] <Hixie>
right, that means not allowing <ol> in <p>

[13:50:00.0000] <annevk>
yes, but you marked that example as non-conforming

[13:50:01.0000] <annevk>
<aside> TEXT <ul> <li> <p> TEXT

[13:50:02.0000] <annevk>
to be more specific

[13:51:00.0000] <annevk>
oh, or that because of the <em> on the line before?

[13:51:01.0000] <Hixie>
an <ol> is either inline or block

[13:51:02.0000] <Hixie>
if it's inline, it can't contain blocks

[13:51:03.0000] <Hixie>
if it's inside a paragraph, it's inline

[13:52:00.0000] <Hixie>
if we allow <ol> inside paragraphs, then <aside> text <ol>... looks the same as <aside> <p> text <ol>...

[13:52:01.0000] <Hixie>
(assuming we do implied <p>s)

[13:53:00.0000] <Hixie>
(i think doing implied <p>s around only elements that can't be blocks would be even more confusing)

[13:53:01.0000] <annevk>
<aside> the way people use IMHO doesn't seem humble at all most of the time

[13:53:02.0000] <Hixie>
heh

[13:54:00.0000] <gavin>
that's completely wrong and you're stupid, IMHO

[13:54:01.0000] <annevk>
<aside> TEXT <ul> <li> TEXT <p> TEXT

[13:54:02.0000] <annevk>
why is the <p> non-conforming with implied paragraphs

[13:54:03.0000] <annevk>
?

[13:56:00.0000] <Hixie>
if we assume that <ul> can be inline, and if we assume that implied paragraphs wrap all potentially inline elements in a run, then the <ul> in that example is inline and inside a paragraph. Elements that are inline can't contain blocks.

[13:58:00.0000] <annevk>
ok, my proposal (on top of implied paragraphs) was to make <ul>, etc. block elements

[13:58:01.0000] <Hixie>
that's what i said. remove the ability for <ol> to be inside <p>.


2007-12-07
[17:37:00.0000] <Hixie>
SQL query:

[17:37:01.0000] <Hixie>
LOCK TABLES `section` , `status` , `tests` , `demos` , `implementations` , `changes` READ

[17:37:02.0000] <Hixie>
MySQL said: Documentation

[17:37:03.0000] <Hixie>
#1064 - You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near ' `status`, `tests`, `demos`, `implementations`, `changes` READ' at line 1 

[17:37:04.0000] <Hixie>
wtf

[17:37:05.0000] <Hixie>
what's wrong with that line

[17:37:06.0000] <Hixie>
all those tables exist

[17:37:07.0000] <Hixie>
OH

[17:37:08.0000] <Hixie>
i see

[17:38:00.0000] <Hixie>
4.x expects the lock type after each table name

[17:38:01.0000] <Hixie>
duh

[17:47:00.0000] <MikeSmith>
Hixie - I have a perhaps dumb question but I will ask anyway. Is post-serialized output of a source document idempotent if re-run through the conformant parser multiple times?

[17:49:00.0000] <Hixie>
in theory

[17:50:00.0000] <Hixie>
if you find any cases where it's not, let me know

[17:50:01.0000] <Hixie>
it is certainly the case that certain conforming documents can't round-trip safely

[17:50:02.0000] <Hixie>
whitespace tends to be dropped or moved near the start and end of documents, for historical reasons

[17:53:00.0000] <Dashiva>
But even if the whitespace is moved, they produce the same DOM, don't they?

[17:53:01.0000] <MikeSmith>
Hixie - at least running a doc through html5lib parse.py multiple times seems to result in linebreaks getting added cumulatively before the closing body tag in the serialized output

[17:53:02.0000] <MikeSmith>
Or maybe pilot error

[17:54:00.0000] <MikeSmith>
/me goes to check his test file again

[17:56:00.0000] <MikeSmith>
I was wrong. Sorry about the noise

[17:57:00.0000] <MikeSmith>
just gets added once

[18:01:00.0000] <Hixie>
Dashiva: yeah

[18:01:01.0000] <Philip`>
/me wonders what is the maximum filesize expansion if you pass something through an HTML5 parser then serialiser

[18:02:00.0000] <MikeSmith>
so with the following minimal source:

[18:03:00.0000] <MikeSmith>
<html><head><title>foo</title></head><body><p>foo</p>

[18:03:01.0000] <MikeSmith>
</body></html>

[18:03:02.0000] <MikeSmith>
I get a linebreak added before closing body tag in output

[18:04:00.0000] <MikeSmith>
/me shuts up and goes to read the spec for the parsing algoritm

[18:07:00.0000] <Philip`>
Hmm, I can get O(n^2) growth with a file like <b><b><b><b><u><u><u><u> </b> </b> </b> </b>

[18:07:01.0000] <Philip`>
/me wonders if that's the worst that's possible

[18:15:00.0000] <Dashiva>
What about lone <td> tags?

[18:15:01.0000] <Dashiva>
<td></table> repeated, maybe

[18:32:00.0000] <Hixie>
ok you can add annotations now, at least in the ui...

[18:32:01.0000] <Hixie>
but i suppose none of that will work until you can log in...

[20:20:00.0000] <Hixie>
you are in a twisty maze of callbacks, all alike

[23:45:00.0000] <Hixie>
i agree with your last e-mail, i think i tried to describe one other way of deciding where the line is in my own most recent email

[00:07:00.0000] <Hixie>
wtf, XMLHttpRequest.status == 0 ???

[00:09:00.0000] <hsivonen>
Hixie: was that my last email?

[00:10:00.0000] <Hixie>
was what your last e-mail?

[00:11:00.0000] <Hixie>
i meant the bit about unambiguous interop not being the line for what is invalid

[00:11:01.0000] <hsivonen>
Hixie: you said "i agree with your last e-mail"

[00:11:02.0000] <hsivonen>
Hixie: ok

[00:12:00.0000] <hsivonen>
I sense a huge bikeshed approaching

[00:15:00.0000] <Hixie>
well, we can try to make it more complicated :-)

[00:23:00.0000] <Hixie>
woot, we're down to just one XXX in this annotation script

[00:35:00.0000] <Hixie>
sql really is the wrong database technology for this

[01:53:00.0000] <annevk>
/me looks at anne-mac

[02:05:00.0000] <zcorpan>
hsivonen: in case you hadn't noticed... http://tinyurl.com/2wywpp s/error errors/errors/ and error 18 and 20 don't mark the right fragment of the source

[02:07:00.0000] <annevk>
yeah, I noticed that too when there were a larger number of errors

[02:08:00.0000] <hsivonen>
zcorpan: thanks, I hadn't noticed

[02:10:00.0000] <hsivonen>
hmm. the source highlight behavior in those results is *weird*

[02:11:00.0000] <hsivonen>
zcorpan: error errors fixed locally. will appear on site later today hopefully

[02:11:01.0000] <othermaciej>
annevk: finally decided to switch?

[02:12:00.0000] <zcorpan>
hsivonen: seems that at some point the line numbers get wrong

[02:13:00.0000] <hsivonen>
zcorpan: yes

[02:14:00.0000] <hsivonen>
zcorpan: debugging that won't be fun :-(

[02:14:01.0000] <annevk>
othermaciej, I bought a MacBook to play with

[02:14:02.0000] <annevk>
I'm still mostly using Ubuntu

[02:14:03.0000] <othermaciej>
annevk: cool, now you have no excuse for posting test results that don't include Safari :-)

[02:14:04.0000] <annevk>
heh

[02:14:05.0000] <annevk>
yeah, that's one advantage

[03:08:00.0000] <Hixie>
/me deploys his spec annotation script onto the actual spec without much testing

[03:15:00.0000] <Dashiva>
Hixie: 0 is nothing, wait until you encounter the ones around the 12029 range :)

[03:17:00.0000] <Hixie>
yeah but there was no reason for it

[03:17:01.0000] <Hixie>
the error disappeared shortly afterwards and i haven't been able to reproduce it

[03:18:00.0000] <Hixie>
i really need to write my own preprocessor for the html5 script

[03:18:01.0000] <Hixie>
bert's script is becoming painfully slow

[03:20:00.0000] <annevk>
please make it generic

[03:20:01.0000] <annevk>
having a public version of the CSS spec generator that is open source would be cool

[03:23:00.0000] <Hixie>
/me is still waiting for bert's script to respond

[03:23:01.0000] <Hixie>
weird

[03:23:02.0000] <Hixie>
doesn't normally take THIS long

[03:24:00.0000] <annevk>
it doesn't work for me

[03:28:00.0000] <MikeSmith>
Hixie, annevk - there's apparently a problem right now with a java process eating up all the cpu on cgi.w3.org

[03:29:00.0000] <Hixie>
ok

[03:29:01.0000] <MikeSmith>
btw, I think Bert's script has a dependencies on other tools he's writeen

[03:29:02.0000] <Hixie>
well then i'll deploy the annotation stuff while i'm asleep

[03:29:03.0000] <Hixie>
/me tells his script to wait 4 hours then regen the spec

[03:29:04.0000] <Hixie>
bed time

[03:29:05.0000] <Hixie>
nn

[03:30:00.0000] <MikeSmith>
night

[03:49:00.0000] <zcorpan>
hsivonen: <footer> allows inline-level content per spec

[03:53:00.0000] <hsivonen>
zcorpan: interesting. has it always been that way?

[03:55:00.0000] <zcorpan>
yees

[03:55:01.0000] <zcorpan>
s/e//

[03:57:00.0000] <hsivonen>
zcorpan: thanks. fixed locally

[03:57:01.0000] <zcorpan>
hsivonen: should i point that out on the list as well?

[03:59:00.0000] <hsivonen>
zcorpan: that would help, I think

[03:59:01.0000] <hsivonen>
I'm pushing out the change now

[04:00:00.0000] <zcorpan>
ok

[04:05:00.0000] <hsivonen>
zcorpan: fix deployed

[04:05:01.0000] <hsivonen>
MikeSmith: IRI checking changes deployed

[04:05:02.0000] <zcorpan>
hsivonen: nice

[04:06:00.0000] <MikeSmith>
hsivonen - sweet

[04:07:00.0000] <MikeSmith>
top-secret HTML5 version of the HTML WG home page now validates

[04:07:01.0000] <MikeSmith>
http://www.w3.org/html/wg/5.html

[04:09:00.0000] <hsivonen>
MikeSmith: <!DOCTYPE html>

[04:09:01.0000] <hsivonen>
<html lang=en xmlns=http://www.w3.org/1999/xhtml>

[04:09:02.0000] <hsivonen>
MikeSmith: that's weird

[04:09:03.0000] <MikeSmith>
hsivonen - it's generated from XHTML source

[04:10:00.0000] <hsivonen>
ah

[04:10:01.0000] <MikeSmith>
and conformant

[04:10:02.0000] <MikeSmith>
if weird

[04:10:03.0000] <MikeSmith>
I doubt Dan would go for the idea of doing it the other way around

[04:11:00.0000] <MikeSmith>
that is, maintaining source in HTML and generating the XHTML from that

[04:11:01.0000] <hsivonen>
MikeSmith: sure, it makes sense to maintain the source as XHTML

[04:12:00.0000] <hsivonen>
MikeSmith: I just didn't realize that you were doing that

[04:12:01.0000] <hsivonen>
MikeSmith: I'd expect an HTML5 serializer to drop xmlns

[04:12:02.0000] <MikeSmith>
html5lib parse.py util doesn't drop it a least

[04:12:03.0000] <MikeSmith>
I think it would make sense if it did, though

[04:13:00.0000] <hsivonen>
MikeSmith: in the Validator.nu parser, dropping it is configurable

[04:13:01.0000] <annevk>
you got to like what it does now :D

[04:13:02.0000] <MikeSmith>
heh

[04:14:00.0000] <hsivonen>
MikeSmith: which shows that allowing the xmlns isn't low-cost, as it had the cost of implementing that configurability

[04:14:01.0000] <hsivonen>
s/low-cost/no-cost/

[04:14:02.0000] <MikeSmith>
true that

[04:15:00.0000] <MikeSmith>
hsivonen - anyway, no longer any XSLT output method=html there. I just uses html5lib now

[04:16:00.0000] <hsivonen>
MikeSmith: great 

[04:16:01.0000] <MikeSmith>
plus some sed postprocessing to add linebreaks for prettification purposes

[04:25:00.0000] <hsivonen>
annevk: is step #3 at http://dev.w3.org/2006/webapi/XMLHttpRequest/#text-response-entity-body consistent with reality? shouldn't you override the XML normative reference to RFC 3023?

[04:26:00.0000] <zcorpan>
how about fixing rfc 3023? :)

[04:27:00.0000] <hsivonen>
zcorpan: feel free to try

[04:28:00.0000] <zcorpan>
at some point i might

[04:28:01.0000] <zcorpan>
try that is

[05:02:00.0000] <Philip`>
/me wonders what is a useful tool if you have thirty minutes to prepare a five minute presentation

[05:02:01.0000] <annevk>
/me is doing that now using HTML + some CSS

[05:04:00.0000] <zcorpan>
there is an opera show generator somewhere, though i don't know where and google seems to point at 404s :(

[05:04:01.0000] <Philip`>
Also I have no idea what computer I can use so it might be stuck with IE

[06:09:00.0000] <Dashiva>
Oh hey, cod@type got brought up again...

[06:11:00.0000] <hendry>
annevk: just looking at http://www.w3.org/TR/access-control So if the server responds with: Access-Control: allow <hello-world.invalid>, why would a request from evil-world.invalid obey that header?

[06:12:00.0000] <Dashiva>
As I understand it, it's not up to evil-world to decide. The browser does that before it serves responses

[06:12:01.0000] <Dashiva>
s/browser/UA/

[06:13:00.0000] <hendry>
you mean the UA that evil-world.invalid is using?

[06:15:00.0000] <Dashiva>
I'm not sure why it would

[06:16:00.0000] <Dashiva>
Any UA can send requests anywhere as it is. Some of them use restrictions (e.g. no cross-domain XHR). Access-control lets these restrictions have exceptions

[06:17:00.0000] <hendry>
I am having trouble visualising this. Is there some diagram I wonder.

[06:18:00.0000] <hendry>
So for UAs that are restricted (some policy?), Access-control is a way of defining some "safe sites"? Is that what you are saying Dashiva ?

[06:18:01.0000] <Dashiva>
Yes

[06:19:00.0000] <Dashiva>
Flash has a similar thing, if you're familiar with that

[06:19:01.0000] <hendry>
Dashiva: no i'm not

[06:20:00.0000] <hendry>
Dashiva: I wonder if Flash's security model is any good. and documented

[06:21:00.0000] <hsivonen>
hendry: you run a trusted UA inside your firewall

[06:21:01.0000] <hsivonen>
hendry: your trusted UA makes the decision whether the script from evil.invalid gets to talk with the servers on the inside of your network perimeter

[06:25:00.0000] <hendry>
ok, thanks for the explanation

[06:26:00.0000] <hendry>
i distantly recall that IE had some sort of privacy control features: http://www.microsoft.com/windows/ie/ie6/using/howto/privacy/config.mspx

[06:26:01.0000] <hendry>
i think it was some standard even WRT privacy policies

[06:27:00.0000] <hendry>
i can't see anything in FF similar. or am i looking in the wrong place

[06:28:00.0000] <hendry>
ah P3P

[06:29:00.0000] <hendry>
anyone know what's happening with http://www.w3.org/P3P/  ? is it dead? is it used? does it have a future?

[06:29:01.0000] <hsivonen>
is P3P actually used or was it just a distraction for shaking off the FTC regulators?

[06:30:00.0000] <hendry>
heh

[06:31:00.0000] <hsivonen>
Facebook shows that industry self-regulation isn't working

[06:31:01.0000] <hendry>
hsivonen: why's that?

[06:31:02.0000] <hendry>
/me logs into FB :)

[06:32:00.0000] <hsivonen>
hendry: Beacon

[06:34:00.0000] <hendry>
ok, ummm.. is there some blog entry explaining this all? What are you saying? They have not regulated people privacy to your requirements?

[06:34:01.0000] <Dashiva>
privacy doesn't exist on facebook

[06:35:00.0000] <hendry>
tbh i like their controls

[06:35:01.0000] <hendry>
i think a lot people (not just me) are fooled into a false sense of security :)

[06:36:00.0000] <hendry>
Dashiva: surely they have some limits. Privacy isn't boolean

[06:37:00.0000] <Dashiva>
Sure, you have privacy from other users.

[06:38:00.0000] <hendry>
But not from the sysops? Is that what you are saying?

[06:40:00.0000] <Dashiva>
Just read up on beacon :)

[06:41:00.0000] <hsivonen>
hendry: http://www.25hoursaday.com/weblog/2007/11/27/SomeThoughtsOnTheFacebookBeacon.aspx

[07:19:00.0000] <hendry>
just looking at FF3. it also doesn't seem to have a notion of a trusted zone. So how does the FF UA know it's in a trusted zone?

[07:25:00.0000] <hsivonen>
hendry: do you mean you'd like to give chrome access to certain http URIs?

[07:37:00.0000] <hsivonen>
Hixie: the HTML5 spec gives me an unresponsive script sheet in Firefox

[07:43:00.0000] <hendry>
hsivonen: well, I assume when you talk about a trusted UA in a firewall. Mustn't the UA be configured as to what to trust as a start point?

[07:43:01.0000] <hsivonen>
hendry: no. that's the point of the access-control spec

[07:44:00.0000] <hsivonen>
hendry: the servers tell what other servers they trust

[07:44:01.0000] <hendry>
ok, so the UA goes to a web site that returns those access-control headers

[07:44:02.0000] <hsivonen>
hendry: and presumably your inside firewall servers wouldn't say they their data is OK to leak to outside of firewall servers

[07:45:00.0000] <hsivonen>
hendry: the idea is that a site can control what other sites its data can be leaked to

[07:46:00.0000] <hendry>
Ok

[07:48:00.0000] <hendry>
and the UA enforces the policies right? so you need a "trusted UA" right? a UA that will obey these access control statements

[07:48:01.0000] <hsivonen>
hendry: yes

[07:49:00.0000] <hendry>
how can a web site be sure the UA accessing it will properly enforce it's access control rules?

[07:49:01.0000] <hsivonen>
hendry: they can't

[07:49:02.0000] <hsivonen>
hendry: by default XSS leaks are not allowed

[07:50:00.0000] <hendry>
you mean same origin stuff implemented on browsers? yeah, but there are ways of getting around that?

[07:51:00.0000] <hsivonen>
hendry: access-control is about letting sites relax same-origin

[07:55:00.0000] <hendry>
though aren't there well known ways around same-origin cross-site limitations? perhaps these holes will be closed mind.

[07:55:01.0000] <hsivonen>
hendry: in general, if a script from site A can read data from site B without cooperation from site B, that's a bug

[07:56:00.0000] <hsivonen>
hendry: there are already ways for a script at site A to issue GET and POST requests to site B without the cooperation of site B

[07:56:01.0000] <hsivonen>
hendry: but that's different from reading the responses

[07:59:00.0000] <hsivonen>
hendry: what kind of ways around same same origin you had in mind?

[08:00:00.0000] <hendry>
hsivonen: i did something the other day

[08:00:01.0000] <hendry>
which created a <script> tag and called in a JS from another site

[08:00:02.0000] <hendry>
i am not too sure if that is considered a same-origin hack

[08:01:00.0000] <hendry>
i think jquery does something similar with .load() or something like that

[08:01:01.0000] <hsivonen>
hendry: yeah, there's a hole if the remote script file embodies sensitive data

[08:02:00.0000] <hendry>
hsivonen: not too sure what is meant by your "reading the responses" line

[08:02:01.0000] <hsivonen>
hendry: which is a dumb thing for a script to do

[08:04:00.0000] <hsivonen>
hendry: I mean the script reading the contents of an HTTP request to another site in a way that it can send back to its own server

[08:06:00.0000] <hendry>
i need diagrams to explain some thoughts i have easily.

[08:07:00.0000] <hendry>
heh

[08:12:00.0000] <hendry>
/me just orders a X61... now i am thinking a tablet would have been useful after all :)

[08:13:00.0000] <Dashiva>
heh

[08:15:00.0000] <hendry>
i'll be in San Francisco next week. I think I'll buy a wacom bamboo tablet. I hope it'll work in Debian.

[08:17:00.0000] <Dashiva>
http://dashiva.net/test/access.jpg

[08:17:01.0000] <Dashiva>
Does that help?

[08:22:00.0000] <hendry>
Dashiva: You need  Access-Control: allow <good.example.org>

[08:22:01.0000] <hendry>
on resource.example.com, right?

[08:24:00.0000] <Dashiva>
yes

[08:25:00.0000] <hendry>
Dashiva: ok, it did help. wtf did you draw that in btw? :)

[08:25:01.0000] <Dashiva>
mspaint, what else? :)

[08:26:00.0000] <Philip`>
You should have used CanvasPaint

[08:26:01.0000] <hendry>
looks like dos paint

[08:26:02.0000] <Philip`>
(or SVG)

[08:27:00.0000] <hendry>
yes, canvas+svg need to get out there

[08:35:00.0000] <hendry>
still thinking here

[08:35:01.0000] <hendry>
this access-control: stuff is done by web sites

[08:35:02.0000] <hendry>
surely a web site user should be more in control

[08:36:00.0000] <hendry>
or is the web site supposed to provide an interface for the user to enter which sites data can be shared with

[08:36:01.0000] <hendry>
for the web site in turn to generate the right access-control headers

[08:40:00.0000] <Dashiva>
hendry: It can be done in headers, but also with PIs if the data is XML

[08:43:00.0000] <csarven>
is it appropriate for all documents on a site use the same h1? should h1 be reserved for heading that is specific to that document?

[08:44:00.0000] <csarven>
if the site is called 'Foo', should all documents on that site use 'Foo' in <h1> or would you reserve h1 for things like 'Welcome to Foo', 'About Foo'... and perhaps using 'Foo' in <address> only

[10:04:00.0000] <bradee-oh>
Hixie: multipage version of HTML5 currently MIA?

[11:19:00.0000] <Hixie>
SPEC NOW HAS ANNOTATION MECHANISM

[11:19:01.0000] <Hixie>
er

[11:19:02.0000] <Hixie>
mascaps

[11:19:03.0000] <Hixie>
miscaps

[11:19:04.0000] <Hixie>
bbiab

[11:20:00.0000] <Dashiva>
multipage is till MIA, it seems

[11:21:00.0000] <Dashiva>
*still

[11:51:00.0000] <Hixie>
huh, weird

[11:58:00.0000] <aroben>
Hixie: http://www.whatwg.org/specs/web-apps/current-work/multipage/ seems troubled

[11:59:00.0000] <aroben>
Hixie: annotation stuff is very cool though

[12:00:00.0000] <aroben>
Hixie: I guess it's impossible to mark an implementation as "pass" without a testsuite first

[12:06:00.0000] <Philip`>
Does non-native support count?

[12:07:00.0000] <Philip`>
(e.g. Excanvas)

[12:07:01.0000] <Philip`>
If so, who decides which non-native implementation should be counted?

[12:11:00.0000] <Philip`>
Hixie: If I middle-click on the show-history link, it doesn't warn me that it's broken

[12:13:00.0000] <Hixie>
no

[12:13:01.0000] <Hixie>
oh, interesting

[12:13:02.0000] <Hixie>
will fix once i have the multipage thing done

[12:13:03.0000] <Hixie>
still can't work out why that's broken

[12:14:00.0000] <Philip`>
Hixie: Double-click to edit is not intuitive

[12:15:00.0000] <Philip`>
(An "edit" link would be more obvious)

[12:16:00.0000] <Philip`>
It's not obvious what the "tests" or "demos" fields are meant to contain, at least when looking at an empty one

[12:18:00.0000] <Philip`>
"Section status: Widely implemented" seems an odd option, since some features are widely implemented at the same time as the spec for them is a first draft (particularly when it's specifying old unstandardised features)

[12:21:00.0000] <Hixie>
there actually is an edit link, it's display:none

[12:21:01.0000] <Hixie>
it looks ugly :-)

[12:21:02.0000] <Hixie>
but i can readd it i guess

[12:22:00.0000] <Hixie>
though if you know that double-click adds, it's not a far jump to double-click to edit

[12:23:00.0000] <Philip`>
/me turns stylesheets off, and sees the edit link now

[12:23:01.0000] <Philip`>
The editing UI is not so good without stylesheets, though

[12:23:02.0000] <Philip`>
Hixie: Oh, double-click adds?

[12:23:03.0000] <Hixie>
yeah

[12:24:00.0000] <Philip`>
The double-click to add behaviour is not intuitive ;-)

[12:24:01.0000] <Hixie>
yeah

[12:24:02.0000] <Hixie>
not sure how to do it better

[12:25:00.0000] <Philip`>
What is the scope of annotations? (Are they on <h3>s or something?)

[12:26:00.0000] <Hixie>
they look at whereever you double click, and find the first element above it to have an id

[12:26:01.0000] <Hixie>
which will typically be an <hx>

[12:26:02.0000] <Hixie>
first element child of body

[12:28:00.0000] <Philip`>
/me doesn't know if it'd be too distracting to have a little edit link beside each possible annotatable location

[12:29:00.0000] <Philip`>
s/edit/add/

[12:32:00.0000] <Hixie>
Philip`: i think having it non-intuitive is ok, the people who should be editing it can be told what the ui is

[12:32:01.0000] <Hixie>
it's not hard to explain

[12:32:02.0000] <Hixie>
multipage is back up

[12:32:03.0000] <Hixie>
but for some reason the annotations aren't showing up

[12:52:00.0000] <Hixie>
Philip`: as far as i can tell, the middle click of a link does give the warning

[12:52:01.0000] <Hixie>
well, error

[13:00:00.0000] <Philip`>
Hixie: Not in Opera 9.2

[13:00:01.0000] <Hixie>
probably a bug in opera

[13:00:02.0000] <Philip`>
(Also the Safari icon doesn't work in Opera)

[13:03:00.0000] <Hixie>
if you have any suggestions for changes to the section status options, let me know

[13:03:01.0000] <Hixie>
afk lunch

[13:16:00.0000] <hsivonen>
Hixie: it would help if the annotation UI highlighted the part of spec being annotated. I can't figure the scope of an annotation

[15:37:00.0000] <Hixie>
hsivonen: yeah... patches welcome :-) not sure how to do that


2007-12-08
[16:00:00.0000] <bradee-oh>
Hixie++  (for multipage being back up)

[16:00:01.0000] <Hixie>
heh

[16:00:02.0000] <Hixie>
i didn't realise the multipage page had a hardcoded reference to the script at the end of the spec

[16:00:03.0000] <Hixie>
so when i removed it...

[16:47:00.0000] <Hixie>
anyone heard of http://www.digistan.org/ ?

[16:47:01.0000] <hober>
nop

[16:48:00.0000] <hober>
*nope

[17:48:00.0000] <Hixie>
/me begins annotating the spec

[18:26:00.0000] <Hixie>
i wonder how to show the scope

[18:27:00.0000] <Hixie>
the scope of one of these annotations is from the header it's attached to, to either the next annotation or the next header of the same rank, whichever comes first

[18:40:00.0000] <Philip`>
Like some kind of vertical line in the margin extending to the end of the scope?

[21:10:00.0000] <Hixie>
Philip`: yeah, can't see how to do that without moving nodes around (which would mess up the code that works out what header to use)

[06:06:00.0000] <Philip`>
Bah, Links doesn't support <xmp>

[07:35:00.0000] <inimino>
Hixie: the link from the multi-page version to index-diff is a 404

[07:39:00.0000] <anne-mac>
/me wonders why the multipage version has a separate status.js

[07:51:00.0000] <Philip`>
Hixie: The opacity:0.05 on NONE status icons seems a bit low - on my screen I can't even see there's an icon there, unless I change my viewing angle so the screen colours get all distorted and then I can make out the shapes

[07:51:01.0000] <Philip`>
0.1 is a bit too faint too; 0.2 appears to work best for me

[07:52:00.0000] <Philip`>
Oh, then it conflicts with the "not fully supported" state

[07:55:00.0000] <Philip`>
I suppose the 0.05 is actually visible if I move my head low enough, so maybe it's just my screen's fault

[09:18:00.0000] <gsnedders>
/me wonders why there is a login box on the spec now

[09:18:01.0000] <Dashiva>
annotations

[09:19:00.0000] <gsnedders>
ah.

[09:34:00.0000] <Lachy>
I can't figure out how the annotations thing works. I registered an account, but there doesn't seem to be anything else in the document for adding annotations.

[09:35:00.0000] <krijnh>
I get status .

[09:35:01.0000] <Dashiva>
Check your error console? :)

[09:36:00.0000] <krijnh>
Ah, cool

[09:37:00.0000] <Philip`>
Lachy: Double click on the annotation boxes to edit, or double click somewhere else to add new annotations

[09:37:01.0000] <krijnh>
Lachy: double cli..

[09:37:02.0000] <krijnh>
/me slaps Philip` with a trout

[09:37:03.0000] <Dashiva>
Is it the same login as the issues list?

[09:37:04.0000] <Philip`>
/me wins

[09:37:05.0000] <krijnh>
:)

[09:41:00.0000] <Lachy>
Dashiva, I don't think so. The issues list login uses email address, this uses a username

[09:43:00.0000] <gsnedders>
Forums/blog login works

[09:45:00.0000] <gsnedders>
(not sure which)

[09:45:01.0000] <Dashiva>
My forum blog didn't. Probably blog then.

[09:45:02.0000] <krijnh>
Can we all put our own links to demos in there?

[09:45:03.0000] <gsnedders>
You don't seem to be able to annotate something which is just marked as a "section"

[09:48:00.0000] <Lachy>
we're probably going to get complaints about accessibility of the annotation script, since it seems impossible to add an annotation without double clicking

[09:48:01.0000] <krijnh>
Nah, can't think of anyone who would :)

[09:49:00.0000] <annevk>
if it's blog login then I've access to the blog now

[09:50:00.0000] <Lachy>
if it's the blog login, then I just registered a new account with the same username

[09:50:01.0000] <krijnh>
ERROR: Invalid username.

[09:50:02.0000] <krijnh>
Prolly not the same thing then

[09:50:03.0000] <gsnedders>
Very odd. I just used the login I've used elsewhere on the WHATWG site.

[09:51:00.0000] <Lachy>
I doubt Hixie would be messing around with the username table in the wordpress database

[09:51:01.0000] <Dashiva>
gsnedders: Maybe you were pre-added as a test

[09:51:02.0000] <gsnedders>
/me blames Hixie 

[09:51:03.0000] <krijnh>
Are there any guidelines on annotating somewhere?

[09:53:00.0000] <krijnh>
Or should everybody just randomly add stuff everywhere?

[09:53:01.0000] <Dashiva>
krijnh: let me fetch the quote

[09:53:02.0000] <Dashiva>
<Hixie> Philip`: i think having it non-intuitive is ok, the people who should be editing it can be told what the ui is

[09:54:00.0000] <krijnh>
Ah

[09:54:01.0000] <krijnh>
I probably shouldn't be editing then ;)

[09:55:00.0000] <gsnedders>
I just mucked around with various possible reasons how you might annotate something :)

[09:56:00.0000] <gsnedders>
/me sighs

[09:56:01.0000] <gsnedders>
I actually have a scroll bar in Mail.app for all my mailboxes :(

[09:58:00.0000] <Lachy>
the issues list says "Error loading the folder list: Internal Server Error. Let Hixie know." - Does he know already?

[10:01:00.0000] <annevk>
i think we could put some simple guidelines on the wiki

[10:01:01.0000] <annevk>
Lachy, probably not

[10:29:00.0000] <annevk>
http://www.kliper.nl/verschillen_tussen_html5_en_html.html

[10:30:00.0000] <annevk>
tekst loopt niet erg lekker

[10:30:01.0000] <annevk>
en Opera Software is vertaald :)

[10:30:02.0000] <Dashiva>
The text does not run smoothly, and Opera Software has been informed

[10:33:00.0000] <annevk>
vertaald = translated

[10:34:00.0000] <annevk>
ah, http://www.kliper.nl/blog_HTML5-verschillen_Verschillen_tussen_HTML4.htm says it's been machine translated

[10:37:00.0000] <Dashiva>
It's always partly hit-and-miss with dutch

[10:37:01.0000] <annevk>
/me has that with Norwegian

[10:51:00.0000] <krijnh>
Hehe

[10:52:00.0000] <krijnh>
<bed in> is het specifieke element voor plugininhoud

[10:52:01.0000] <krijnh>
:D

[14:27:00.0000] <Philip`>
Hixie: If I was going to bother adding stuff to the spec annotations, would I have to add five hundreds links to tests in the "Links to tests for this section" bit?

[14:28:00.0000] <Philip`>
Also, I expect "partial+buggy" would be a more accurate status option than the current choices in most cases

[14:28:01.0000] <annevk>
no

[14:29:00.0000] <Dashiva>
partial and/or buggy :)

[14:33:00.0000] <Philip`>
"The latest release version of %s passes all the available test cases for this feature." is vacuously true for many features since there are zero test cases - maybe there should be another option for "this feature is supported but there are not enough test cases to judge the quality of support"

[14:36:00.0000] <Philip`>
/me needs to remember to update his tests so that Opera 9.5 no longer gets a PASS for getImageData

[14:37:00.0000] <annevk>
how do we fail?

[14:38:00.0000] <Philip`>
Hixie: In Firefox 2, after I open the annotation thingy and click "Cancel", the space bar stops working for scrolling down the screen

[14:38:01.0000] <annevk>
/me sees no bug report

[14:39:00.0000] <Philip`>
annevk: .data is a CanvasPixelDataArray or something, and not an Array, so when somebody does getImageData(...).data.join(',') then it dies, like in https://bugzilla.mozilla.org/attachment.cgi?id=291126

[14:41:00.0000] <Dashiva>
Philip`: You should make that test use a single line for the problematic statement. The .join() doesn't even show in the console error :)

[14:43:00.0000] <Philip`>
That test isn't mine, so I can't do much to it :-)

[14:44:00.0000] <Philip`>
Actually, I could, since I could just modify it and upload a new version; but it's not meant to be testing getImageData anyway


2007-12-09
[04:53:00.0000] <gsnedders>
anyone able to be conned into writing some JavaScript for the HTTP parsing test suite?

[04:56:00.0000] <Philip`>
What is that likely to involve?

[06:07:00.0000] <gsnedders>
Philip`: Loading files from one folder, then checking that the headers/body/status match up with a JSON file in another folder (where the two files are related by filename)

[06:07:01.0000] <gsnedders>
http://hg.gsnedders.com/cgi-bin/hgwebdir.cgi/http-parsing/file/c550df97477b/tests/response/ shows the basic structure of the files

[07:35:00.0000] <Philip`>
gsnedders: How is it meant to determine the headers/body/status?

[07:35:01.0000] <Philip`>
(Using XHR, I guess)

[07:35:02.0000] <Philip`>
(since I can't think of any alternatives)

[07:36:00.0000] <Philip`>
gsnedders: How is it meant to know which files to load?

[08:52:00.0000] <gsnedders>
Philip`: Yeah, XHR. I guess the best would be to load /tests and use getElementsByTagName('a')

[08:53:00.0000] <gsnedders>
Philip`: (then assume the filenames in /expected have the extension .expected instead of .http)

[09:25:00.0000] <Philip`>
gsnedders: Ah, okay

[09:25:01.0000] <Philip`>
Why not write it yourself? :-)

[09:26:00.0000] <gsnedders>
Philip`: Because I suck and I've already wasted ages failing :)

[09:26:01.0000] <gsnedders>
(or rather, my knowledge of JS in general is minimal)

[09:41:00.0000] <Philip`>
gsnedders: Ah, fair enough :-)

[09:45:00.0000] <gsnedders>
Philip`: I'd say I'd add you to the list of contributors in the spec if you helped, but there again, you're already there :)

[09:58:00.0000] <hsivonen>
the current wikitruth about HTML5 does not acknowledge APIs

[10:18:00.0000] <Philip`>
gsnedders: It seems not entirely easy to test HTTP header parsing with XHR, because getAllResponseHeaders is largely undefined and inconsistently implemented

[10:18:01.0000] <gsnedders>
Philip`: surely getResponseHeader() would be easier?

[10:19:00.0000] <hsivonen>
I keep refinding issues that I've already reported...

[10:19:01.0000] <hsivonen>
ins/del datetime syntax this time

[10:19:02.0000] <Philip`>
gsnedders: Then the .expected format would have to be able to say that it expects a certain header to not exist

[10:20:00.0000] <gsnedders>
Philip`: some headers are created by the UA though, like the cache ones, in my experience

[10:20:01.0000] <Philip`>
(so you can test that things don't get misinterpreted as headers when they shouldn't)

[10:21:00.0000] <gsnedders>
Philip`: but surely some other header value or the body will be wrong?

[10:26:00.0000] <Philip`>
gsnedders: If you have "HTTP/1.0 200 OK\r\nSome-Garbage\r\nContent-Type: text/html\r\n\r\nBody" then I guess maybe the Some-Garbage would just get ignored, and wouldn't affect anything else, and the tests should check there isn't a parsed header named "Some-Garbage"

[10:26:01.0000] <gsnedders>
Philip`: ah. true.

[10:26:02.0000] <Philip`>
but maybe that's not too important

[10:27:00.0000] <Philip`>
but getAllResponseHeaders is useless anyway so there's not much choice

[10:32:00.0000] <gsnedders>
Philip`: does getResponseHeader() consistently return null when it doesn't exist?

[10:32:01.0000] <gsnedders>
(where it == the header)

[10:34:00.0000] <Philip`>
gsnedders: No - looks like FF returns null, IE/Opera return ""

[10:34:01.0000] <gsnedders>
ergh.

[10:34:02.0000] <gsnedders>
Could use that to check, but if that isn't consistent either.

[10:35:00.0000] <hsivonen>
/me tries to document the time element "vague" datetime format in one paragraph...

[10:35:01.0000] <Philip`>
/me wonders if XHR getResponseHeader uses the same HTTP parser as the rest of the browser engine does

[10:36:00.0000] <Philip`>
(given that IE's getAllResponseHeaders returns the originl raw HTTP header string, so it has at least some kind of lower-level hook)

[10:36:01.0000] <Philip`>
+a

[10:37:00.0000] <gsnedders>
Philip`: in IE and Opera, yes. In Fx it does for some headers, and not for others, IIRC.

[10:38:00.0000] <Philip`>
Hmm, is it impossible to do multiline createTextNode? IE requires "\r" and treats "\n" like a space, but FF requires "\n" and ignores a lone "\r"

[10:39:00.0000] <Philip`>
(...inside <pre>, in particular)

[10:39:01.0000] <hsivonen>
whoa. the vague time format sucks

[10:40:00.0000] <gsnedders>
(IIS and IE don't use the same HTTP parser, though, BTW)

[10:43:00.0000] <gsnedders>
Philip`: when does your term end, BTW?

[10:43:01.0000] <hsivonen>
gsnedders: did you send feedback about the time formats as part of your microsyntax review? I don't see your messages in the WHATWG issue folder

[10:43:02.0000] <gsnedders>
hsivonen: no, I never got around to doing them

[10:43:03.0000] <hsivonen>
gsnedders: ok

[10:48:00.0000] <Philip`>
gsnedders: Is something like http://canvex.lazyilluminati.com/misc/http-test-runner.patch vaguely useful?

[10:50:00.0000] <Philip`>
gsnedders: I don't really do terms now, since I've stopped being an undergraduate, so I can go home whenever I want as long as nobody minds :-)

[10:51:00.0000] <gsnedders>
/me is forgetting too much

[10:51:01.0000] <gsnedders>
Philip`: yeah

[10:51:02.0000] <Philip`>
(That is, as long as nobody minds who is important enough that I mind them minding)

[10:55:00.0000] <gsnedders>
/me thanks Philip` loads :P

[10:57:00.0000] <Philip`>
/me hopes his code isn't too horribly broken

[11:17:00.0000] <gsnedders>
/me hasn't tried running it yet

[11:17:01.0000] <gsnedders>
/me hopes Philip` did

[11:22:00.0000] <gsnedders>
/me wonders if the Connection header should be case sensitive

[11:22:01.0000] <hsivonen>
occasionally one has to wonder how this Web thing works at all

[11:35:00.0000] <mpt>
It's like an ogre

[13:13:00.0000] <gsnedders>
Onions are like ogres. They both have layers.

[13:15:00.0000] <mpt_>
That's what I was referring to, yes

[13:15:01.0000] <gsnedders>
I guessed :)


2007-12-10
[19:21:00.0000] <G0k>
hgey all

[19:23:00.0000] <G0k>
is there a way to see the history of changes to an html 5 spec section?

[19:56:00.0000] <MikeSmith>
G0k - no, not for a specific section, afaik

[19:57:00.0000] <G0k>
oki, thanks

[19:57:01.0000] <Dashiva>
You could blame your way through it with svn, but that's a hassle :)

[20:11:00.0000] <othermaciej>
if you have a specific question, people here might know the answer

[20:14:00.0000] <G0k>
well in http://www.whatwg.org/specs/web-apps/current-work/multipage/section-event0.html

[20:14:01.0000] <G0k>
the MessageEvent interface

[20:14:02.0000] <G0k>
i seem to recall the source being a document

[20:14:03.0000] <G0k>
now it's a window

[20:14:04.0000] <G0k>
is this a change or am i misremembering?

[20:14:05.0000] <othermaciej>
it was changed to work on Window

[20:14:06.0000] <othermaciej>
discussion is in the whatwg archive

[20:15:00.0000] <othermaciej>
basically so exceptions to same-origin policy can all be a the Window level

[20:16:00.0000] <G0k>
so how does that affect like..iframes?

[20:17:00.0000] <G0k>
if an iframe posts a message to its parent, it sees it as coming from the same window but a different domain?

[20:17:01.0000] <othermaciej>
iframes have a different Window than their parent

[20:17:02.0000] <G0k>
oh

[20:17:03.0000] <G0k>
heh sorry my dom skillz are a little weak

[20:19:00.0000] <othermaciej>
G0k: this part of the spec explains in detail how Window and Document fit together and work with frames: http://www.whatwg.org/specs/web-apps/current-work/multipage/section-windows.html#windows

[02:23:00.0000] <annevk>
/me still isn't really convinced by the postMessage change

[02:23:01.0000] <hsivonen>
annevk: do current impls do what you require?

[02:25:00.0000] <annevk>
Opera has the only implementation

[02:26:00.0000] <annevk>
Mozilla is implementing and wants changes because then all security exceptions supposedly are on Window

[02:26:01.0000] <hsivonen>
annevk: oh is this about cross-doc? I thought you were talking about XHR body on GET

[02:27:00.0000] <annevk>
a) that patch has gone nowhere and b) I'm not convinced there are no exceptions left on Document

[02:27:01.0000] <annevk>
oh, I'm not sure about that one

[02:27:02.0000] <annevk>
/me hopes to avoid editing the specification

[03:17:00.0000] <othermaciej>
annevk: in WebKit, it's definitely the nearl all security exceptions are on Window and none are on Document

[03:18:00.0000] <othermaciej>
(actually I guess it is fair to say that all exceptions are on Window but there are a few security checks in other places)

[03:29:00.0000] <annevk>
ok

[03:29:01.0000] <annevk>
I did get claims that .open() and .close() and such had similar things, but maybe that's something else

[03:29:02.0000] <annevk>
hmm

[03:36:00.0000] <othermaciej>
window.open() and window.close() are allowed from any other window

[03:36:01.0000] <othermaciej>
that's not true for most Window functions or properties

[03:37:00.0000] <othermaciej>
in the case of Document, in WebKit at least you can't even get the document object from a window that's not same-origin

[03:37:01.0000] <othermaciej>
I think in Gecko you can get it, but it throws on any property access

[03:37:02.0000] <annevk>
document.open() must have a similar check then

[03:37:03.0000] <annevk>
as it can be a synonym for window.open() depending on the number of arguments

[03:38:00.0000] <annevk>
ah, I see

[03:38:01.0000] <othermaciej>
it can be a synonym but I do not think it is required for web compatibility for document.open() to ever be callable from a window with a different origin

[03:38:02.0000] <othermaciej>
in WebKit that's not possible because you can't even get to the document object for another window

[03:38:03.0000] <annevk>
right

[03:38:04.0000] <annevk>
hmm

[03:44:00.0000] <hsivonen>
does html5lib have a SAX view to the parse tree?

[04:46:00.0000] <zcorpan>
hello whatwg

[04:46:01.0000] <othermaciej>
hello zcorpan 

[04:46:02.0000] <hsivonen>
zcorpan: hello

[04:46:03.0000] <hsivonen>
zcorpan: do you know what's happening with ARIA datatypes?

[04:47:00.0000] <hsivonen>
zcorpan: including the use of qnames there

[04:47:01.0000] <hsivonen>
zcorpan: and the general ugliness of XSD dependencies

[04:51:00.0000] <zcorpan>
hsivonen: not sure. xhtml2 wg wants curies, i think. pfwg and svgwg seem to want opaque strings, i think

[04:53:00.0000] <hsivonen>
zcorpan: even for stuff datatype='xsd:anyURI'?

[04:54:00.0000] <zcorpan>
hsivonen: example of what that would be?

[04:55:00.0000] <hsivonen>
zcorpan: dunno, but http://www.w3.org/TR/2007/WD-aria-state-20071019/#datatype

[04:55:01.0000] <hsivonen>
interestingly, the stated datatype is string--not qname

[04:56:00.0000] <hsivonen>
but the prose says qname

[04:56:01.0000] <hsivonen>
yay

[04:56:02.0000] <zcorpan>
"Clearly XSD base types are simpler for the user agent to understand."

[04:57:00.0000] <annevk>
being a ua rep: no

[04:59:00.0000] <hsivonen>
I wonder if it is too late to bring in the WF2 datatypes instead

[04:59:01.0000] <annevk>
where do they use datatypes?

[05:00:00.0000] <annevk>
ah, I see

[05:00:01.0000] <othermaciej>
zcorpan: they must have an interesting theoretical model of "the user agent"

[05:02:00.0000] <hsivonen>
zcorpan: btw, what's the case-(in)sensitivity story for ARIA in text/html? is in going to be lower-case only?

[05:03:00.0000] <zcorpan>
hsivonen: the values?

[05:54:00.0000] <hsivonen>
zcorpan: e.g. role='foo'

[05:54:01.0000] <hsivonen>
zcorpan: is foo case-sensitive?

[06:23:00.0000] <zcorpan>
hsivonen: i'd say yes

[06:23:01.0000] <zcorpan>
but i'm not sure if it's defined anywhere

[06:24:00.0000] <annevk>
i'd say no

[06:24:01.0000] <zcorpan>
annevk: why?

[06:24:02.0000] <annevk>
for consistency with other enumerated attributes

[06:25:00.0000] <zcorpan>
but role isn't really enumerated

[06:25:01.0000] <zcorpan>
it's more like class=""

[06:25:02.0000] <annevk>
it has a fixed amount of known values

[06:26:00.0000] <zcorpan>
/me finds a login button in the spec

[06:26:01.0000] <annevk>
i agree it's slightly different though given that it has fallback

[06:26:02.0000] <zcorpan>
yeah

[06:26:03.0000] <zcorpan>
case-sensitive is simpler

[06:28:00.0000] <annevk>
hmm, the annotating doesn't work in Opera

[06:30:00.0000] <annevk>
why is event-source considered for removal?

[06:31:00.0000] <hsivonen>
zcorpan: based on specs, I'd say case-sensitive as well

[06:32:00.0000] <hsivonen>
annevk: changing the case-senstivity of role from whatever Firefox does is probably counter-productive

[06:33:00.0000] <hsivonen>
annevk: I haven't tested what it does here

[06:34:00.0000] <zcorpan>
hmm, does anyone else see the safari icon? e.g. at <canvas>

[06:35:00.0000] <Philip`>
zcorpan: The login button is for editing annotations

[06:35:01.0000] <Philip`>
zcorpan: I see that icon in Firefox 2, but not in Opera 9.2

[06:35:02.0000] <hsivonen>
hmm. I've reached some kind of zen state. I feel no urge to reply to the <xmp> thread

[06:35:03.0000] <zcorpan>
Philip`: ok

[06:36:00.0000] <dglazkov>
hsivonen, aka "apathy"?

[06:36:01.0000] <annevk>
hsivonen, like selectors, media queries are not "CSS"

[06:36:02.0000] <hsivonen>
dglazkov: if you want to spin it negatively :-)

[06:37:00.0000] <Philip`>
/me hopes the <xmp> discussion can avoid collapsing into an XHTML discussion

[06:37:01.0000] <gsnedders>
Philip`: Fail.

[06:37:02.0000] <Philip`>
It hasn't collapsed that way yet, though it's on the edge :-)

[06:37:03.0000] <dglazkov>
/me doesn't like being negative

[06:37:04.0000] <dglazkov>
zen it is

[06:38:00.0000] <gsnedders>
Philip`: Oh, I can see into the future. You can't?

[06:43:00.0000] <hsivonen>
zcorpan: I see the safari icon

[06:44:00.0000] <hsivonen>
does the disclosure triangle in spec annotations work in any browser?

[06:45:00.0000] <Philip`>
hsivonen: Disclosure triangle? I thought the triangle was just pointing at the the attached section (except in Firefox where it gets drawn in the wrong place)

[06:45:01.0000] <hsivonen>
Philip`: oh.

[06:46:00.0000] <hsivonen>
Philip`: well, it's a usability bug then

[06:46:01.0000] <gsnedders>
/me still wants to know how he had an account without ever requesting one

[06:46:02.0000] <hsivonen>
Philip`: because I though it was a broken disclosure triangle

[06:46:03.0000] <Philip`>
hsivonen: I could be wrong, of course :-)

[06:47:00.0000] <hsivonen>
Philip`: your explanation makes sense considering the safari rendering

[06:47:01.0000] <gsnedders>
I always get, "Failed to save data: OK", when trying to save :\

[06:48:00.0000] <hsivonen>
gsnedders: the test data is broken in Firefox 2

[06:49:00.0000] <hsivonen>
gsnedders: you need a trunk build, it seems

[06:49:01.0000] <gsnedders>
Doesn't work in Saf3, either

[06:50:00.0000] <Philip`>
Somebody ought to do something about all this non-interoperability

[06:50:01.0000] <hsivonen>
gsnedders: in Safari 3 it works at least a bit more than in Firefox 2

[06:50:02.0000] <gsnedders>
Philip`: EMCAScript 5?

[06:50:03.0000] <gsnedders>
DOM5?

[06:50:04.0000] <annevk>
I would like to know what causes the breakage in Opera

[06:50:05.0000] <hsivonen>
I wonder if this is the Acid3 test

[06:50:06.0000] <gsnedders>
does it work in _anything_ except Fx3?

[06:51:00.0000] <OmegaJunior>
I thought ES4 was in the works.... ES5 doesn't exist yet.

[06:51:01.0000] <MikeSmith>
annevk - maybe Hallvord Steen could figure out pretty quickly

[06:52:00.0000] <gsnedders>
OmegaJunior: peh. we have XHTML5 while XHTML2 is in the works!

[06:58:00.0000] <MikeSmith>
Is something supposed to happen if I click on the triangle in an annotation

[07:00:00.0000] <Philip`>
MikeSmith: Probably not

[07:00:01.0000] <MikeSmith>
k

[07:00:02.0000] <Philip`>
MikeSmith: It seems it's just rendered differently in some browsers so that it no longer makes sense

[07:04:00.0000] <MikeSmith>
Philip` - ah, yeah, I see now. In Opera and Safari, it renders on the right edge of the annotation box

[07:08:00.0000] <annevk>
it's displayed on the wrong place in Firefox

[07:09:00.0000] <annevk>
oh

[08:02:00.0000] <MikeSmith>
annevk - are you running parallels or vmware on your Mac?

[08:05:00.0000] <annevk>
neither, I think

[08:08:00.0000] <MikeSmith>
annevk - I just mention because you said you were using Ubuntu also

[08:09:00.0000] <MikeSmith>
or some Linux

[08:10:00.0000] <MikeSmith>
you know if you use Parallels you can run OSX, Linux, Windows on the same machine

[08:10:01.0000] <MikeSmith>
without needing to reboot

[08:11:00.0000] <annevk>
you can also use two laptops :)

[08:12:00.0000] <MikeSmith>
sure, two is better than one of course

[08:12:01.0000] <MikeSmith>
and three would be even better than two

[08:12:02.0000] <MikeSmith>
you could hire a valet to help you carry them around

[08:13:00.0000] <annevk>
actually, I have three

[08:13:01.0000] <annevk>
although one is from zcorpan

[08:14:00.0000] <MikeSmith>
you and zcorpan could maybe share a valet

[08:14:01.0000] <MikeSmith>
alternating days or something

[08:18:00.0000] <Lachy>
VMWare is better than Parallels. That's what I use on  my Mac

[08:19:00.0000] <gsnedders>
yeah, that's what I concluded too.

[08:20:00.0000] <Philip`>
/me just uses Boot Camp and almost never boots OS X

[09:58:00.0000] <annevk>
hmm, people use CSS interop bugs as CSS hacks

[09:59:00.0000] <annevk>
guess there's nothing new there

[09:59:01.0000] <othermaciej>
isn't that the definition of a CSS hack?

[09:59:02.0000] <othermaciej>
(except that ideally you should use the failure case as the hack trigger)

[10:00:00.0000] <annevk>
it's not yet clear here what the hack is and what the failure

[10:01:00.0000] <annevk>
/me was just sending some results of interop testing to www-style

[10:01:01.0000] <annevk>
/me wonders when someone brings accessibility to the <xmp> discussion

[10:29:00.0000] <annevk>
hmm, I don't agree either with not changing the method names because of some Java API

[10:29:01.0000] <annevk>
grmbl

[10:30:00.0000] <othermaciej>
especially since the Java API was made by copying an Editor's Draft, with the agreement that it would not it being a work in progress and subject to change

[10:31:00.0000] <othermaciej>
now they ask not to change

[10:31:01.0000] <othermaciej>
that doesn't seem right

[10:31:02.0000] <annevk>
lengthComputable...

[10:57:00.0000] <annevk>
ah

[10:58:00.0000] <annevk>
people are still doing 2DGraphics in Java

[10:58:01.0000] <annevk>
http://www.cycloloco.com/shadowmaker/

[10:58:02.0000] <annevk>
if only shadows were implemented

[10:58:03.0000] <annevk>
we could <canvas>ify it

[11:04:00.0000] <Philip`>
<canvas> shadows wouldn't help much when you don't want the shadow to be exactly the same shape as the object

[11:05:00.0000] <Philip`>
You can do Gaussian blurring with canvas easily, but I can't think of an easy way to convert an image to greyscale

[11:06:00.0000] <Philip`>
You can't do shears in most implementations either

[11:10:00.0000] <Philip`>
Oh, actually, maybe greyscale is easy, by drawing a grey rectangle with destination-in

[11:12:00.0000] <Philip`>
Oh, actually, no current released implementation can do shearing

[11:13:00.0000] <Philip`>
(unless you fake it with rotations and scales)

[11:14:00.0000] <Philip`>
Anyway, you should implement canvas shadows anyway ;-)

[11:31:00.0000] <annevk>
heh

[13:18:00.0000] <anne-mac>
keeping globalStorage & co makes some sense given that SQL is asynchronous

[13:18:01.0000] <anne-mac>
I guess that will satisfy most requirements

[13:18:02.0000] <anne-mac>
btw, othermaciej, did you get around contacting the Firefox guys about changing the design of globalStorage?

[13:18:03.0000] <anne-mac>
especially its broken security model

[13:19:00.0000] <anne-mac>
maybe that can be changed in time

[13:28:00.0000] <othermaciej>
anne-mac: I wasn't planning to contact them, I think Hixie was

[13:29:00.0000] <anne-mac>
oh

[13:32:00.0000] <anne-mac>
Hixie, you around?

[13:35:00.0000] <anne-mac>
interesting, IE8 will have a new rendering engine

[13:36:00.0000] <anne-mac>
although on the other hand it's sort of problematic if that's going to be opt-in

[13:37:00.0000] <Lfe>
anne-mac: url? :-)

[13:37:01.0000] <anne-mac>
depending on what authors code againts, of course

[13:37:02.0000] <anne-mac>
http://www.molly.com/2007/12/05/conversation-with-bill-gates-about-ie8-and-microsoft-transparency/

[13:38:00.0000] <anne-mac>
"MOLLY HOLZSCHLAG: But they’re not letting people talk about it. I do realize that there is a new engine, "

[13:38:01.0000] <Lachy>
anne-mac, yeah, Chris mentioned that at web directions 2 months ago

[13:38:02.0000] <anne-mac>
apparently this has been announced earlier at some conference

[13:38:03.0000] <anne-mac>
right

[13:39:00.0000] <Lfe>
Well, by new - do they mean new from scratch or just a "version bump"?

[13:39:01.0000] <gsnedders>
What I think is really interesting is whether it'll replace the current standards mode or not.

[13:39:02.0000] <othermaciej>
Chris has announced that they are working on a new engine

[13:40:00.0000] <othermaciej>
where "engine" is pretty strictly restricted to layout, not a full rewrite of everything

[13:40:01.0000] <gsnedders>
othermaciej: how much of the layout?

[13:40:02.0000] <Lachy>
I wonder if they've tried to implement the HTML5 parsing algorithm in the new engine

[13:41:00.0000] <othermaciej>
I dunno

[13:41:01.0000] <othermaciej>
http://www.sitepoint.com/blogs/2007/10/01/wds07-bonus-feature-chris-wilson-microsoft/

[13:41:02.0000] <Lachy>
though given that we're yet to get any significant feedback from MS about the spec, it doesn't seem too likely

[13:42:00.0000] <anne-mac>
MS has been pretty much "inactive" everywhere apart from CSS

[13:42:01.0000] <gsnedders>
I guess even the layout itself is where a lot of the major bugs are.

[13:42:02.0000] <othermaciej>
layout is where one category of major bugs are

[13:42:03.0000] <anne-mac>
and even there...

[13:42:04.0000] <anne-mac>
one category :)

[13:43:00.0000] <gsnedders>
I mean, is the CSS parsing that bad? That probably depends on the border between parsing and layout.

[13:43:01.0000] <gsnedders>
/me doesn't even mention the DOM

[13:45:00.0000] <othermaciej>
there's still significant CSS parsing bugs, and certainly CSS processing bugs

[13:45:01.0000] <anne-mac>
yeah

[13:46:00.0000] <Hixie>
bugs in IE? say it isn't so!

[13:46:01.0000] <G0k>
i gotta say i fail to understand the "don't break the web" line

[13:46:02.0000] <othermaciej>
not that Gecko/Presto/WebKit don't have bugs

[13:46:03.0000] <Hixie>
/me is here now btw

[13:46:04.0000] <G0k>
they're going to have bug fixes which are going to break things, it's ridiculous to expect otherwise

[13:47:00.0000] <anne-mac>
Hixie, so 1) <event-source> considered for removal?! 2) annotation doesn't work anywhere... 3) ping Firefox about simplying client-side storage?

[13:47:01.0000] <G0k>
event-source wha?

[13:47:02.0000] <othermaciej>
G0k is working on implementing event-source for WebKit, so he'd probably be interested in hearing if it is considered for removal

[13:48:00.0000] <anne-mac>
G0k, don't worry, we'll just set up whatwg2

[13:48:01.0000] <Hixie>
1) it hasn't really had much demand either from authors or from implementors, and a number of issues have been raised suggested we should massively simplify it even if we keep it. it also seems redundant with the two-way network connection stuff and with multipart xhr. but if there really is demand, we could keep it. we just need to prove there is demand, same as any other feature.

[13:48:02.0000] <Hixie>
2) elaborate?

[13:48:03.0000] <Hixie>
3) yeah i should do that

[13:49:00.0000] <G0k>
heh well to be honest having both event-source and RemoveEventTarget seemed kinda redundant

[13:49:01.0000] <othermaciej>
is there a plan for multipart XHR?

[13:49:02.0000] <anne-mac>
firefox has an impl of that, but I don't really like it

[13:49:03.0000] <Hixie>
G0k: event-source and removeeventtarget are the same feature

[13:50:00.0000] <G0k>
heh awz

[13:50:01.0000] <Hixie>
anne-mac: i've heard people ask for it a lot, at least at google

[13:50:02.0000] <anne-mac>
Hixie, 2) it doesn't work in Opera, or Safari or stable versions of Firefox

[13:50:03.0000] <Hixie>
ah

[13:50:04.0000] <Hixie>
any idea why?

[13:51:00.0000] <anne-mac>
Hixie, same for event-source at Opera

[13:51:01.0000] <Hixie>
i don't think i used anything non-standard

[13:51:02.0000] <anne-mac>
not really

[13:51:03.0000] <Hixie>
well if there are any bugs in the code do let me know

[13:51:04.0000] <anne-mac>
debugging it seemed non-trivial

[13:51:05.0000] <Hixie>
afk bbiab

[13:52:00.0000] <othermaciej>
Annotation seems to work on in Safari 3 for me

[13:53:00.0000] <Hixie>
back

[13:53:01.0000] <othermaciej>
well actually I'm not sure if it is or not, since I don't know what it's supposed to look like

[13:53:02.0000] <Hixie>
click the login button at the top

[13:54:00.0000] <anne-mac>
Hixie, it doesn't make much sense to consider <event-source> for removal but not <datagrid> which is in comparison far less successful

[13:54:01.0000] <G0k>
how is event-source successful?

[13:54:02.0000] <Hixie>
oh datagrid will be removed too if nobody wants it

[13:55:00.0000] <Hixie>
but datagrid is much less mature and hasn't been looked at much by implementors yet

[13:55:01.0000] <Hixie>
and it has no alternative for authors

[13:55:02.0000] <anne-mac>
G0k, there's one implementation and one partial patch...

[13:55:03.0000] <G0k>
well two partial patches but i mean has anyone used it for anything?

[13:58:00.0000] <anne-mac>
editing the annotation in Safari seems to remove it

[13:59:00.0000] <anne-mac>
just happened to <canvas>

[13:59:01.0000] <Hixie>
check for js errors

[13:59:02.0000] <anne-mac>
G0k, there's a public chat client somewhere

[14:00:00.0000] <anne-mac>
Hixie, sorry, not familiar enough with Safari to get that running

[14:00:01.0000] <Hixie>
k

[14:00:02.0000] <G0k>
i guess i do feel like multipart XHR could very well eliminate 90% of the need for event source

[14:01:00.0000] <G0k>
even though i think the addEventSource API is about 1000x more elegant

[14:01:01.0000] <gsnedders>
Hixie: Saf shows no JS errors

[14:01:02.0000] <Hixie>
odd

[14:01:03.0000] <Hixie>
wonder what's going on

[14:01:04.0000] <anne-mac>
Opera doesn't do that either and yet it goes wrong with logging in

[14:01:05.0000] <anne-mac>
it's annoying

[14:02:00.0000] <Hixie>
what goes wrong?

[14:02:01.0000] <gsnedders>
anne-mac: you need the debug menu enabled (just google it) then its obvious enough in that menu

[14:02:02.0000] <anne-mac>
when you log in you get a single . instead of "Logged in as anne."

[14:02:03.0000] <Hixie>
odd

[14:03:00.0000] <gsnedders>
anne-mac: what should show that?

[14:03:01.0000] <Hixie>
that suggests that the login is failing without a message

[14:03:02.0000] <gsnedders>
anne-mac: login box?

[14:04:00.0000] <Hixie>
which means x.status != 200 and != 304

[14:04:01.0000] <gsnedders>
I get "OK." under the login button…

[14:04:02.0000] <Hixie>
anne-mac: reload and try again, i've changed the message that should appear

[14:04:03.0000] <gsnedders>
and trying to save anything results in, "Failed to save data: OK" (this is Saf 3/Mac)

[14:05:00.0000] <Hixie>
sounds like xhr bugs

[14:05:01.0000] <gsnedders>
now I get 403

[14:05:02.0000] <Hixie>
403 means you're not logged in

[14:05:03.0000] <gsnedders>
"Status: Error: OK (403)."

[14:05:04.0000] <Hixie>
in theory

[14:05:05.0000] <gsnedders>
that's trying to login.

[14:05:06.0000] <Hixie>
wait what?

[14:05:07.0000] <Hixie>
how the hell are you getting 403 OX

[14:05:08.0000] <Hixie>
OK

[14:06:00.0000] <Hixie>
that's not one of the things my script sends back

[14:06:01.0000] <gsnedders>
I try logging in.

[14:06:02.0000] <gsnedders>
Oh, I get it all right :)

[14:06:03.0000] <gsnedders>
:P

[14:06:04.0000] <Hixie>
do you have a login?

[14:06:05.0000] <gsnedders>
Hixie: it worked in Fx3b1

[14:06:06.0000] <gsnedders>
see, this is why I just get other people to write JavaScript for me

[14:07:00.0000] <Hixie>
what's your username?

[14:07:01.0000] <gsnedders>
/me stares at Philip` 

[14:07:02.0000] <gsnedders>
Hixie: gsnedders 

[14:07:03.0000] <anne-mac>
I get "        var node = response.documentElement;" null value around that line

[14:07:04.0000] <Hixie>
there's no "gsnedders" user registered

[14:08:00.0000] <gsnedders>
then how on earth did I login in Fx3b1!?

[14:08:01.0000] <Hixie>
anne-mac: when doing what in what browser?

[14:08:02.0000] <Hixie>
gsnedders: no idea

[14:09:00.0000] <Hixie>
there aree 14 users registered, including 3 annes, but you're not one of them :-)

[14:09:01.0000] <gsnedders>
OK, now I get, "Logged in as gsnedders."

[14:09:02.0000] <Hixie>
now you have one :-)

[14:09:03.0000] <anne-mac>
Hixie, when adding an "x" to the demo field in Safari

[14:09:04.0000] <anne-mac>
/me has anne, test, and test1000 registered, he thinks

[14:10:00.0000] <gsnedders>
Hixie: null value, line 580 of status.js trying to edit

[14:10:01.0000] <anne-mac>
/me got weird messages when registering

[14:10:02.0000] <Hixie>
the only way that can happen is if you got a 204... hmm...

[14:10:03.0000] <gsnedders>
and on that note, g'nite

[14:10:04.0000] <Hixie>
nn

[14:11:00.0000] <Hixie>
i can't see any way that get-one-annotation can return a 204...

[14:11:01.0000] <Hixie>
unless...

[14:11:02.0000] <Hixie>
oh i know what's going on

[14:11:03.0000] <Hixie>
i bet these browsers are all caching stuff

[14:11:04.0000] <Hixie>
/me adds cache-control: no-cache

[14:11:05.0000] <Hixie>
try now

[14:12:00.0000] <Hixie>
(clear your cache first)

[14:12:01.0000] <Hixie>
though frankly i don't understand what the browsers are caching if this helps it

[14:14:00.0000] <Hixie>
try now

[14:14:01.0000] <Hixie>
sorry

[14:14:02.0000] <Hixie>
had syntax error

[14:25:00.0000] <roc>
gah

[14:25:01.0000] <roc>
Colloquy is useless

[14:25:02.0000] <G0k>
yeah

[14:26:00.0000] <Hixie>
/me waves his irssi banner

[14:26:01.0000] <G0k>
it's kinda driving me crazy

[14:26:02.0000] <roc>
oops wrong channel

[14:26:03.0000] <Dashiva>
irssi isn't cool until it starts using python for scripting, 'cuz xkcd said so

[14:26:04.0000] <roc>
but thanks for the support

[14:31:00.0000] <othermaciej>
Colloquy rules!

[14:31:01.0000] <G0k>
it's pretty and nice except for the fact that it has some really inexplicable bugs

[14:32:00.0000] <G0k>
like sometimes you join channels and the text is invisible

[14:34:00.0000] <roc>
right, that's what's biting me

[14:38:00.0000] <Hixie>
did all the people who were having problems with the annotation script go to have lunch or something? :-)

[14:38:01.0000] <G0k>
they gave up

[14:38:02.0000] <anne-mac>
i had some food

[14:38:03.0000] <anne-mac>
i have issues with the safari inspector thingie not showing up now

[14:39:00.0000] <Philip`>
They were so enthralled they forgot to return to IRC

[14:40:00.0000] <Hixie>
anne-mac: we can get it to work in opera if you want instead

[14:40:01.0000] <Hixie>
let the safari guys worry about safari :-)

[14:40:02.0000] <Philip`>
/me wonders who will be the first to dare complain it doesn't work in IE

[14:41:00.0000] <anne-mac>
Hixie, when logging in I get "Error (0)"

[14:41:01.0000] <anne-mac>
in Opera

[14:41:02.0000] <Hixie>
in the login dialog?

[14:42:00.0000] <Hixie>
in an alert?

[14:42:01.0000] <anne-mac>
in the dialog

[14:42:02.0000] <Hixie>
and does it say exactly "Error (0)" or does it say "Error:  (0)"?

[14:42:03.0000] <anne-mac>
the latter

[14:42:04.0000] <anne-mac>
followed by a .

[14:42:05.0000] <Hixie>
well then that's an opera bug. that means xhr isn't returning a sensible status code.

[14:42:06.0000] <Hixie>
the 0 is x.status at readystate = 4

[14:43:00.0000] <anne-mac>
weird

[14:43:01.0000] <anne-mac>
thanks though

[14:45:00.0000] <Hixie>
(does the xhr spec ever allow 0 in rs4?

[14:45:01.0000] <Hixie>
)

[14:45:02.0000] <anne-mac>
if http does

[14:45:03.0000] <anne-mac>
actually, it does

[14:45:04.0000] <anne-mac>
or maybe not

[14:46:00.0000] <G0k>
hixie: so has anything been written w/r/t multipart XHR?

[14:46:01.0000] <Hixie>
not to my knowledge

[14:46:02.0000] <anne-mac>
no, not

[14:46:03.0000] <Hixie>
i'm trying to let anne deal with xhr issues

[14:46:04.0000] <Hixie>
as i have plenty on my plate already :-)

[14:46:05.0000] <G0k>
anne, let's talk xhr

[14:46:06.0000] <anne-mac>
sure

[14:47:00.0000] <Hixie>
anne-mac: if it turns out to be a bug in my script, please let me know. but i have no idea what it couldf be.

[14:48:00.0000] <anne-mac>
Hixie, does your script happen to use redirects on the server?

[14:49:00.0000] <G0k>
so it seems me like what we really want is more control over the the downloading process

[14:50:00.0000] <anne-mac>
http://dev.w3.org/2006/webapi/XMLHttpRequest-2/ has most of the new stuff, but I'm still awaiting conforming implementations of XHR1

[14:51:00.0000] <anne-mac>
G0k, more control, how?

[14:52:00.0000] <G0k>
like preventing connections from timing out, getting around the 2 connection limit, checking up on the connection download progress

[14:52:01.0000] <Hixie>
anne-mac: no, no redirects as far as i know

[14:52:02.0000] <Hixie>
(except maybe in the multipage version?)

[14:52:03.0000] <Hixie>
afk, brb

[14:57:00.0000] <anne-mac>
G0k, I'm not sure how to do one and two, but three is addressed

[14:57:01.0000] <G0k>
uhm. how about just allowing authors to set the connection timeout?

[14:58:00.0000] <anne-mac>
ah yeah, that's been suggested

[14:58:01.0000] <anne-mac>
send(data, timeout) or something like that

[14:59:00.0000] <G0k>
that sounds like progress yes.

[14:59:01.0000] <anne-mac>
 I've only seen one or two requests for that so far

[15:00:00.0000] <anne-mac>
and it's doable using setTimeout ... (although maybe slightly more annoying)

[15:01:00.0000] <G0k>
yeah but that sucks

[15:01:01.0000] <G0k>
you basically end up polling the thing

[15:06:00.0000] <G0k>
plus...i mean i guess it's an implementation issue but abort() isn't very reliable

[15:06:01.0000] <anne-mac>
in general we don't want to introduce new features just because others are badly implemented

[15:08:00.0000] <G0k>
yeah fair enough

[15:09:00.0000] <G0k>
i mean i feel like, in principle, we should be able to implement dom events with XHR

[15:09:01.0000] <G0k>
should be a design goal, that

[15:09:02.0000] <anne-mac>
dom events?

[15:09:03.0000] <G0k>
er...server-sent dom events

[15:09:04.0000] <G0k>
if they're really going bye bye

[15:10:00.0000] <anne-mac>
i also think that in general overloading xhr is a bad idea

[15:10:01.0000] <anne-mac>
it's already pretty bloated thanks to historic accidents

[15:10:02.0000] <G0k>
i also kinda feel like the name is sorta crappy

[15:10:03.0000] <G0k>
why not just make it HttpRequest?

[15:10:04.0000] <anne-mac>
with weird events being dispatched at random places

[15:11:00.0000] <anne-mac>
that would be similar to renaming <html> to <web>

[15:11:01.0000] <G0k>
but....no one uses it for xml

[15:12:00.0000] <anne-mac>
i'll try again; it already works, why break that?

[15:13:00.0000] <G0k>
just because it works doesn't mean that a new API wouldn't be clearer, less buggy, and more featureful

[15:13:01.0000] <anne-mac>
sounds like xhtml2 to me

[15:13:02.0000] <Hixie>
a new api means two apis

[15:13:03.0000] <Hixie>
two apis, one of which sucks, is worse than one api which sucks.

[15:14:00.0000] <G0k>
well there's already about 6 APIs for drawing a box

[15:14:01.0000] <Hixie>
it means more spec work, more implementation work, more testing, more tutorials, etc

[15:14:02.0000] <anne-mac>
G0k, that should not be an excuse to make it worse! :)

[15:14:03.0000] <Hixie>
bad things exist. they aren't a license to add more bad things.

[15:15:00.0000] <Hixie>
(the same arguments suggest removing event-source and co, btw)

[15:15:01.0000] <G0k>
so i presume <canvas> is going away too? :)

[15:15:02.0000] <anne-mac>
(i'm not convinced)

[15:17:00.0000] <Hixie>
anne-mac: that's why it's not already gone :-)

[15:17:01.0000] <G0k>
i mean we have to break backwards compatibility anyway to implement certain features

[15:18:00.0000] <Hixie>
G0k: how do you create, say, a fractal viewer, without <canvas>?

[15:18:01.0000] <G0k>
extreme SVG use?

[15:19:00.0000] <Hixie>
uh huh

[15:19:01.0000] <Philip`>
Lots and lots of tiny <div>s

[15:19:02.0000] <Hixie>
without taking down the browser.

[15:19:03.0000] <Hixie>
and without abusing semantics.

[15:19:04.0000] <Philip`>
That's just a quality-of-implementation issue

[15:19:05.0000] <Philip`>
<div> has no semantics :-)

[15:20:00.0000] <inimino>
Hixie: http://claudiusmaximus.goto10.org/GG/graphgrow.svg

[15:20:01.0000] <G0k>
and if we're going into semantics, then continuing to use the XMLHttpRequest name is especially ridiculous

[15:21:00.0000] <Hixie>
semantics are related to definitions, not names

[15:21:01.0000] <inimino>
(I think it would be more better with <canvas> though)

[15:22:00.0000] <anne-mac>
/me wonders why that doesn't show anything in Opera

[15:23:00.0000] <inimino>
/me looks for the link to the description page...

[15:23:01.0000] <G0k>
even then, the XHR definition of "scripted client functionality for transferring data between a client and a server" is not quite right

[15:25:00.0000] <G0k>
in any case, retaining backwards compatibility with oddities in the old ActiveX control's API seems equivalent to adding the marquee and blink tags back in

[15:27:00.0000] <G0k>
anyway i'm off for a bit

[15:29:00.0000] <inimino>
http://claudiusmaximus.blurty.com/#419

[15:30:00.0000] <inimino>
anne-mac: looks like it was tested only in Fx

[15:31:00.0000] <inimino>
when it works you can build fractals like this http://mjclement.com/images/screenshots/2007-12-09_4.png

[15:31:01.0000] <inimino>
and it uses SVG + JavaScript

[15:31:02.0000] <anne-mac>
interesting

[15:32:00.0000] <inimino>
I think an implementation using <canvas> could be faster

[15:32:01.0000] <anne-mac>
i wonder how the bugs are distributed

[15:32:02.0000] <inimino>
I didn't look at the code but I assume the SVG ends up rather large

[15:36:00.0000] <anne-mac>
there's some script error in Opera

[15:36:01.0000] <anne-mac>
looks like something is null/undefined where an object is expected

[15:36:02.0000] <anne-mac>
oh well, too late


2007-12-11
[18:11:00.0000] <Hixie>
i've changed the meanings of the implementation status settings

[18:13:00.0000] <Hixie>
"The World Wide Web Consortium, or W3C, a group devoted to publishing web standards, recently moved to approve the Ogg video and audio formats for inclusion into the forthcoming HTML5 standard."

[18:13:01.0000] <Hixie>
-- http://www.clipotech.com/2007/12/nokia-wants-w3c-to-remove-ogg-from.html

[18:14:00.0000] <Hixie>
we did? holy crap! i had no idea we were nearing FPWD publication

[18:14:01.0000] <Hixie>
that's awesome!

[18:16:00.0000] <othermaciej>
I don't understand what the brightness level of the logos means

[18:16:01.0000] <othermaciej>
in the status messages

[18:16:02.0000] <othermaciej>
is there a legend?

[18:16:03.0000] <Hixie>
there are tooltips

[18:20:00.0000] <othermaciej>
/me wonders where the line is between "no support" and "some support, with bugs"

[18:21:00.0000] <Hixie>
no support means strictly zero support

[18:21:01.0000] <Hixie>
as in, indistinguishable from what the UA would be doing if the spec didn't exist

[18:21:02.0000] <Hixie>
and if the implementors had never thought about it

[18:22:00.0000] <othermaciej>
what about when the spec defines something that was already in UAs?

[18:23:00.0000] <othermaciej>
(Selection is the case I'm looking at)

[18:23:01.0000] <Hixie>
then it's a common sense approach

[18:23:02.0000] <Hixie>
feel free to modify it if i made a mistake

[18:24:00.0000] <Hixie>
i wasn't sure who supported what

[18:24:01.0000] <Hixie>
off hand

[18:24:02.0000] <Hixie>
double-click to edit

[18:24:03.0000] <othermaciej>
I'm just asking so I know whether to suggest changes

[18:24:04.0000] <Hixie>
don't worry about suggesting changes, just double click the section and edit it :-)

[18:25:00.0000] <othermaciej>
ok, I will, if I can figure out how to change

[18:25:01.0000] <Hixie>
double clicking the section (or if the section already has an annotation, the annotation itself) will popup a dialog

[18:25:02.0000] <Hixie>
let me know if it doesn't work

[18:27:00.0000] <Hixie>
<Hixie> double clicking the section (or if the section already has an annotation, the annotation itself) will popup a dialog

[18:27:01.0000] <Hixie>
<Hixie> let me know if it doesn't work

[18:28:00.0000] <Hixie>
/me casually removes Ogg from the spec and sees what happens

[18:29:00.0000] <othermaciej_>
/me takes shelter

[18:58:00.0000] <kfish>
Hixie!

[19:04:00.0000] <kfish>
/me throws a tantrum on behalf of the free software community

[19:22:00.0000] <kfish>
/me reads the replacement text and revokes the tantrum

[19:23:00.0000] <kfish>
Hixie, actually you didn't casually remove Ogg, you made the case for Ogg stronger, so thankyou :-)

[19:28:00.0000] <Dashiva>
"Lift the cat who was amongst the pigeons up and put him back on his pedestal for now."

[19:28:01.0000] <Dashiva>
Poetic

[19:30:00.0000] <Hixie>
kfish: :-)

[19:47:00.0000] <Hixie>
so people keep asking for a way to do <code type="python">...</code> or equivalent

[19:51:00.0000] <Dashiva>
But the use cases are kinda lacking...

[19:52:00.0000] <Hixie>
the use cases are being able to automatically syntax-highlight the code

[19:52:01.0000] <Hixie>
e.g. using a second script

[19:52:02.0000] <Dashiva>
Yeah, and I don't see why that needs @type

[19:52:03.0000] <Hixie>
the question is, does class="" address that sufficiently

[19:52:04.0000] <Dashiva>
It's pure presentation, after all

[19:52:05.0000] <Hixie>
well, not really

[19:53:00.0000] <Hixie>
no more than "em" is "pure presentation"

[19:53:01.0000] <Dashiva>
Well, saying a code block is c++ is semantics, but coloring the block is presentation

[19:54:00.0000] <Hixie>
syntax highlighting can be much more than specific colours

[19:54:01.0000] <Hixie>
it could be fonts, voices, etc

[19:54:02.0000] <Hixie>
it's not _specific_ presentation

[19:55:00.0000] <Dashiva>
No, not specific. But some kind of.

[19:55:01.0000] <Hixie>
right. like <em>.

[19:55:02.0000] <Hixie>
or <p>

[19:55:03.0000] <Hixie>
or <h1>

[19:55:04.0000] <Hixie>
or...

[19:55:05.0000] <Dashiva>
<poem>

[19:55:06.0000] <Hixie>
right

[19:55:07.0000] <Dashiva>
If we do decide to add @type, who decides the values?

[19:56:00.0000] <Hixie>
yeah, that's one of the problems

[19:56:01.0000] <Dashiva>
Would microformats fit?

[19:56:02.0000] <Hixie>
another is that actually, as much as i thought i saw this often, i've only been able to find 2 e-mails in my piles

[19:57:00.0000] <Dashiva>
(Assuming they get rel)

[19:57:01.0000] <Hixie>
well, you could certainly define a set of classes as a microformat for <code class="">

[19:57:02.0000] <Hixie>
in fact the first e-mail i have on this was sent to mf-discuss

[19:58:00.0000] <Dashiva>
mediaformats?

[19:58:01.0000] <Hixie>
?

[19:58:02.0000] <Dashiva>
mf

[19:58:03.0000] <Hixie>
no :-P

[19:59:00.0000] <Dashiva>
Some mozilla then? :)

[19:59:01.0000] <Hixie>
microformat-discuss, silly :_P

[19:59:02.0000] <Dashiva>
oh

[19:59:03.0000] <Dashiva>
duh

[19:59:04.0000] <Hixie>
:-P

[20:00:00.0000] <Dashiva>
I imagine whoever ends up with it will have lots of fun deciding whether gcc-version-x.y.z is a separate code type because it has some special quirk

[20:00:01.0000] <Hixie>
hah

[20:03:00.0000] <Dashiva>
oh wow, 5 am. Time to poof

[20:03:01.0000] <Hixie>
later

[20:03:02.0000] <Hixie>
thanks for the help

[20:04:00.0000] <Dashiva>
I don't recall being much, but hey :P

[20:04:01.0000] <Hixie>
:-)

[20:04:02.0000] <Hixie>
always helpful to have someone to bounce things off

[20:43:00.0000] <MikeSmith>
Lachy - hei

[20:44:00.0000] <Lachy>
hi

[21:49:00.0000] <Hixie>
http://wiki.whatwg.org/wiki/FAQ#When_will_HTML_5_be_finished.3F updated again

[22:16:00.0000] <jruderman>
hmm "user agents will be required to support the marquee element, but authors must not use the marquee element in conforming documents."

[22:16:01.0000] <jruderman>
/me 's head explodes

[22:16:02.0000] <gavin>
why?

[22:18:00.0000] <jruderman>
requiring support for an entire disallowed feature... seems strange

[22:18:01.0000] <jruderman>
although i guess you could say the same about tag soup parsing or boring deprecated elements

[22:18:02.0000] <gavins>
right

[22:18:03.0000] <gavins>
web browsers need to support it, so it needs to be specified

[22:19:00.0000] <gavins>
seperate issue than whether or not it should be recommended to authors

[00:50:00.0000] <mitsuhiko>
flamewars ahead

[01:11:00.0000] <Hixie>
jruderman: where did i say they will be required to support it? i imagine the support will actually be limited to the dom aspects; the rendering will probably not be required

[01:12:00.0000] <jruderman>
Hixie: i pasted from http://wiki.whatwg.org/wiki/FAQ#Why_does_HTML5_legitimise_tag_soup.3F

[01:12:01.0000] <Hixie>
aah

[01:12:02.0000] <Hixie>
yeah

[01:12:03.0000] <Hixie>
the faq probably plays fast and loose with the truth in parts

[01:12:04.0000] <jruderman>
you might want to edit it to say that simply rendering it as block or inline-block or whatever is acceptable

[01:21:00.0000] <zcorpan>
hsivonen: yt?

[01:27:00.0000] <Hixie>
jruderman: well, we'll see waht the spec says

[01:38:00.0000] <Hixie>
i need a personal www-archive

[01:39:00.0000] <Hixie>
somewhere i can just cc that will make my e-mails public

[01:39:01.0000] <Hixie>
for things unrelated to the web

[01:42:00.0000] <hdh>
in the first example of #the-small, the footer has address and p as children, but one is block-level and the other inline

[01:43:00.0000] <hdh>
oops, address is block level too

[01:53:00.0000] <Hixie>
prepare for a lot of e-mails about random stuff that nobody cares about...

[02:28:00.0000] <Hixie>
hsivonen: you around?

[02:28:01.0000] <Hixie>
i was idly wondering if you had an opinion on the issue of block vs inline

[02:29:00.0000] <Hixie>
in particular, given the proposal of removing the distrinction, but considering that there are certain nestings we still want to disallow, whether you had any ideas on how to define that

[02:29:01.0000] <Hixie>
distinction, even

[02:30:00.0000] <Hixie>
e.g.  <section> bla bla <p> bla bla <ol><li> bla bla </li></ol></p></section>

[02:30:01.0000] <Hixie>
should be valid

[02:30:02.0000] <Hixie>
but

[02:30:03.0000] <Hixie>
e.g.  <section> bla bla <p> bla bla <ol><li> <p> bla bla </li></ol></p></section>

[02:30:04.0000] <Hixie>
should not be valid

[02:30:05.0000] <Hixie>
...because nesting paragraphs is meanignless

[02:30:06.0000] <Hixie>
meaningless even

[02:30:07.0000] <annevk>
the above being XML markup?

[02:30:08.0000] <Hixie>
yeah, all xml

[02:30:09.0000] <annevk>
i'm not convinced we should allow <ol> inside <p>

[02:31:00.0000] <Hixie>
html technicalities aside, do you agree that a single sentence cannot span more than one paragraph?

[02:33:00.0000] <annevk>
yeah, nested paragraphs don't make sense

[02:34:00.0000] <annevk>
I haven't quite up made my mind about nested links yet though

[02:34:01.0000] <Hixie>
no i mean can a sentence start in one paragraph and end in a sibling paragraph?

[02:34:02.0000] <Hixie>
ignore html for now

[02:34:03.0000] <Hixie>
i just mean with actual paragraphs

[02:34:04.0000] <aphid>
only fantastical paragraphs.  how else would they be connected?

[02:34:05.0000] <annevk>
I agree it doesn't make much sense, but that's how everyone is marking up their code

[02:35:00.0000] <annevk>
so maybe it should just work

[02:35:01.0000] <Hixie>
annevk: i'm confused as to what you are talking about

[02:35:02.0000] <aphid>
it shouldn't, thought isn't so hierarchical

[02:35:03.0000] <Hixie>
i'm just talking about normal human language

[02:35:04.0000] <Hixie>
like in a book

[02:36:00.0000] <othermaciej>
It's possible for a sentence to start in one paragraph and end in another

[02:36:01.0000] <othermaciej>
though usually it's a special effect with unusual punctuation

[02:36:02.0000] <othermaciej>
(like ... or an emdash)

[02:37:00.0000] <othermaciej>
you could argue these are two separate sentence fragments

[02:37:01.0000] <annevk>
yeah, &hellip;

[02:37:02.0000] <hdh>
or when people leave half a sentence in blog title, to be completed in the post

[02:37:03.0000] <Hixie>
i'm just thinking that a paragraph with a sentence like "i bought (1) a tree, (2) a chainsaw, and (3) an apple, and then went home."

[02:38:00.0000] <Hixie>
imho that belongs in one paragraph

[02:38:01.0000] <Hixie>
not five

[02:38:02.0000] <othermaciej>
that certainly makes sense as a paragraph, however, the presentation/semantics boundary becomes a bit fuzzy here

[02:38:03.0000] <othermaciej>
If instead you say:

[02:38:04.0000] <othermaciej>
I bought:

[02:38:05.0000] <Hixie>
(there are places in the html5 spec right now where i have sentences like that and html4 forces me to spread it over multiple paragraphs)

[02:38:06.0000] <othermaciej>
    (1) a tree

[02:39:00.0000] <othermaciej>
    (2) a chainsaw

[02:39:01.0000] <othermaciej>
    (3) an apple

[02:39:02.0000] <othermaciej>
and then went home

[02:39:03.0000] <othermaciej>
(end of example)

[02:39:04.0000] <othermaciej>
that's less clearly a paragraph

[02:39:05.0000] <Hixie>
possibly

[02:39:06.0000] <Hixie>
still

[02:39:07.0000] <Hixie>
we do want to be able to nest lists

[02:40:00.0000] <Hixie>
so <ol> <li> <ol> <li> needs to be valid

[02:40:01.0000] <Hixie>
hmm...

[02:40:02.0000] <othermaciej>
list nesting certainly makes sense

[02:40:03.0000] <othermaciej>
is that not valid now?

[02:40:04.0000] <Hixie>
sure

[02:40:05.0000] <Hixie>
i'm just thinking out loud

[02:41:00.0000] <Hixie>
i suppose making lists not be paragraphs does just basically mean that all we need to say is that inlines in block-level contexts get implied paragraphs wrapping them and then we're done

[02:41:01.0000] <Hixie>
<em> <ol>...</ol> </em> wouldn't be allowed

[02:41:02.0000] <Hixie>
presumably

[02:42:00.0000] <othermaciej>
usually I would not care much about issues like this but at least it's not about codecs

[02:42:01.0000] <Hixie>
heh

[02:42:02.0000] <Hixie>
i'm just idly pondering the issue

[02:42:03.0000] <Hixie>
it's one of the things we'll need to fix

[02:43:00.0000] <Hixie>
right now it's pretty much the #1 issue people run into when transitioning to html5

[02:43:01.0000] <Hixie>
(or html4 strict)

[02:49:00.0000] <othermaciej>
I think looser content rules could be good

[02:49:01.0000] <othermaciej>
And I agree that being able to use list markup for inline-level lists and nesting lists both seem like sensible use cases

[02:49:02.0000] <othermaciej>
also that nesting a paragraph inside a paragraph is not

[02:50:00.0000] <othermaciej>
also that nesting anything but its normal expected children in an element with a highly structured content model it generally not sensible

[02:50:01.0000] <othermaciej>
(things like lists, dables, etc)

[02:51:00.0000] <Hixie>
maybe i need to do a matrix if {block|inline elements} x {block|inline elements} and just figure out which is allowed where

[02:51:01.0000] <Hixie>
sure

[02:51:02.0000] <Hixie>
i'm just talking about inlines and blocks

[02:51:03.0000] <Hixie>
where they are allowed, and which elements claim to be them

[02:52:00.0000] <othermaciej>
I'm just stating the premises relating to this that seem fairly obvious

[02:52:01.0000] <Hixie>
yeah

[02:52:02.0000] <othermaciej>
hopefully some final rule results that is not completely ad-hoc

[02:53:00.0000] <Hixie>
yeah, that's what i'm hoping. maybe some new categorisation that is neither inline nor block

[02:53:01.0000] <Hixie>
but which denotes what's allowed where

[03:02:00.0000] <Philip`>
The idea of 'It costs us nothing to keep <some obscure element> because everyone has implemented it already' seems to conflict with that of 'We want to define HTML so that new competitors will be able to enter the browser space' (since they'll have to implement all these obscure historical elements from scratch)

[03:06:00.0000] <Hixie>
good god the ogg thing hit reddit.

[03:07:00.0000] <Hixie>
Philip`: <kbd> is extremely cheap to implement for browser vendors. the cost is mostly on the spec and tutorial side.

[03:08:00.0000] <othermaciej>
Hixie: multiple times

[03:27:00.0000] <zcorpan>
Hixie: personally, i think html5 should be more like html4 transitional wrt content model

[03:28:00.0000] <zcorpan>
drop structured inline concept, disallow <p><ol/></p>

[03:28:01.0000] <zcorpan>
allow <div>foo<p>bar</p></div>

[03:29:00.0000] <zcorpan>
doing so will save lots of time for authors understanding the rules

[03:29:01.0000] <zcorpan>
i think that the current rules will make more authors ignore conformance altogether, because they're too complicated and seemingly arbitrary

[03:30:00.0000] <zcorpan>
also, i think <div>foo<p>bar</p></div> is pretty harmless

[03:33:00.0000] <annevk>
html5.org on reddit

[03:34:00.0000] <annevk>
I better add some text links :evil:

[03:34:01.0000] <gsnedders>
<http://www.sitepoint.com/article/ie-standards-chris-wilson> for those who haven't seen

[03:34:02.0000] <gsnedders>
(says one or two things about IE8)

[03:34:03.0000] <annevk>
summary?

[03:38:00.0000] <gsnedders>
- Redoing layout engine (which was already known), first major revision to that code in years

[03:38:01.0000] <gsnedders>
- Not using Gecko/WebKit because of potential licensing issues

[03:39:00.0000] <gsnedders>
- If there was a bug in either, they'd be to blame and responsible for it

[03:39:01.0000] <gsnedders>
- Fixing security bugs in a codebase you don't know is hard

[03:40:00.0000] <gsnedders>
- If all standardise on one implementation you end up with how IE/Netscape are/were defining the de-facto rules for the web

[03:40:01.0000] <annevk>
I guess after six years you might forget a thing or two :)

[03:40:02.0000] <gsnedders>
Who actually did the security updates for IE while there was no IE team?

[03:40:03.0000] <annevk>
Hmm, yeah, that's certainly true. We already have to copy crazy-ass Firefox features :(

[03:41:00.0000] <gsnedders>
look at the latest Netscape, for example

[03:42:00.0000] <gsnedders>
There's various other interesting things in it, but I need to go

[03:48:00.0000] <gsnedders>
is it 10 or 20 years old that patents need to be to expire?

[03:48:01.0000] <gsnedders>
20?

[03:49:00.0000] <annevk>
they still seem to have the vendor-lock-in idea

[03:49:01.0000] <annevk>
or the infinite amount of quirks modes

[03:49:02.0000] <Hixie>
"You haven't read the HTML5 spec. You get to define your own tags and the CSS statement for your new tag may reuse the definitions of other tags like { display: p } for example."

[03:49:03.0000] <Hixie>
-- http://programming.reddit.com/info/62oek/comments/c02n4nm

[03:49:04.0000] <Hixie>
i can't tell if that's a troll or not...

[03:49:05.0000] <gsnedders>
/me blinks

[03:50:00.0000] <annevk>
I think he might have read Crockford HTML5

[03:50:01.0000] <gsnedders>
that makes no sense to me however many times I read it, yet alone make it wrong

[03:50:02.0000] <annevk>
which advocates that kind of stuff

[03:50:03.0000] <gsnedders>
anyhow, I really need to run off now

[03:50:04.0000] <gsnedders>
MPEG-1 is still too new, though, which is what I thought.

[03:50:05.0000] <annevk>
http://crockford.com/html/ search for "display: div"

[03:51:00.0000] <annevk>
Hixie, "Error loading the folder list: Internal Server Error. Let Hixie know." (on /issues/

[03:51:01.0000] <annevk>
)

[03:51:02.0000] <annevk>
(oh, already resolved)

[03:52:00.0000] <Hixie>
yeah the server is probably getting hit too hard

[03:53:00.0000] <Philip`>
Hixie: <tv-show> would be equally cheap as <kbd> for browser vendors to implement, and equally costly to spec (since the <kbd> spec is rewritten anyway, and they only need about one line of definition)

[03:54:00.0000] <Philip`>
so I'm not certain where the "much cheaper" comes from, in "since browser vendors aren't going to drop support for existing elements, existing elements end up being much cheaper to "add" than new elements"

[03:56:00.0000] <Hixie>
if we do nothing at all, e.g. if we simply don't have an html5 spec, then <kbd> continues to exist but <tv-show> doesn't exist

[03:56:01.0000] <Hixie>
to effect a change, we have to do something (a cost)

[03:56:02.0000] <Hixie>
adding <tv-show> thus has a cost

[03:57:00.0000] <Hixie>
i agree that at the end of the day <tv-show>, if it basically did nothing, would be cheap to add

[03:57:01.0000] <Hixie>
orders of magnitude cheaper than, say, <video>

[03:57:02.0000] <Philip`>
gsnedders: Normally 20, but browser developers have to care about the relevant laws and filing dates in every country in the world

[03:59:00.0000] <Philip`>
Hixie: Adding <kbd> to HTML5 has a similar cost, so that doesn't seem a convincing argument for adding <kbd> but not adding <tv-show>

[04:04:00.0000] <Hixie>
fair enough

[04:04:01.0000] <Hixie>
i don't know what would convince you, or, really, what we're actually trying to establish here

[04:08:00.0000] <Hixie>
right well enough of this firefighting, time to sleep

[04:08:01.0000] <Hixie>
nn

[04:08:02.0000] <alp>
Hixie: VP3 was from "On2", not "On3" btw

[04:09:00.0000] <alp>
On2 donated the VP3 codec before going on to create further editions which have had some success

[04:11:00.0000] <roc>
annevk: what "crazy-ass Firefox features"?

[04:12:00.0000] <roc>
because I'm trying to kill off getBoxObjectFor as fast as I can!

[04:12:01.0000] <Philip`>
Hixie: Maybe the cost of removing the feature is a more important point than the cost of adding it - it would make a few sites harder to migrate from HTML4 to HTML5, and annoy people who spent lots of time carefully semanticising their HTML4 since they'd have to strip it out

[04:14:00.0000] <annevk>
roc, extensions of Range, but some may be useful enough to consider extending the Range specification

[04:14:01.0000] <annevk>
some were not, but I believe you guys are removing these (or it was one) in Firefox 3 so it may get better

[04:14:02.0000] <Philip`>
Hixie: I don't want to be convinced about anything - I just wanted to note that your response to someone else didn't seem very compelling to me, so maybe a different response would be better in the future, if anyone cares enough to raise the point again :-)

[04:17:00.0000] <roc>
the extensions we currently have look pretty useful:

[04:17:01.0000] <roc>
http://mxr.mozilla.org/seamonkey/source/dom/public/idl/range/nsIDOMNSRange.idl

[04:17:02.0000] <mpt>
HTML5. Brought to you by the letter W, the number 5, and the word "Sadly".

[04:17:03.0000] <Dashiva>
Wow, that crockford html page was like reading about xhtml2 :)

[04:17:04.0000] <roc>
I actually want to add some more Range extensions

[04:18:00.0000] <roc>
like some form of getClientRects

[04:18:01.0000] <annevk>
I think createContextualFragment was causing us issues

[04:18:02.0000] <annevk>
I agree with getClientRects: http://dev.w3.org/csswg/cssom-view/#the-rangeview-interface

[04:19:00.0000] <roc>
ah excellent

[04:19:01.0000] <roc>
the wording sounds suspiciously similar to what I had in mind

[04:19:02.0000] <roc>
did I suggest it?

[04:19:03.0000] <roc>
:-)

[04:19:04.0000] <annevk>
i found it on a bug report I was cc'ed on :)

[04:20:00.0000] <roc>
another thing we need in Range is some way to get at the text nodes that comprise logical runs of text

[04:20:01.0000] <roc>
for things like Find

[04:21:00.0000] <roc>
or if a Web app wanted to implement its own spellchecking

[04:22:00.0000] <annevk>
I saw something about that elsewhere, about having an API to give a run of text some type of "class" which you can then use in ::selection("class") or something like that

[04:22:01.0000] <roc>
that might be different

[04:22:02.0000] <roc>
that sounds like an API for doing custom selections

[04:24:00.0000] <roc>
what I want is an API like "give me a list of text nodes that comprise the first 'text run' in the Range"

[04:24:01.0000] <roc>
taking into account various CSS features like block boundaries, generated content, etc

[04:25:00.0000] <annevk>
that might be tricky

[04:25:01.0000] <roc>
yeah

[04:25:02.0000] <roc>
it's tricky

[04:25:03.0000] <annevk>
given that generated content is some kind of pseudo-DOM

[04:25:04.0000] <roc>
yeah

[04:25:05.0000] <roc>
returning a list of text nodes doesn't quite work

[04:26:00.0000] <roc>
but some guy wanted to pay FIFTY THOUSAND DOLLARS for this approximately this feature for a Firefox extension

[04:27:00.0000] <annevk>
must be useful then :p

[04:27:01.0000] <roc>
/me must sleep

[04:36:00.0000] <hdh>
I have this tb when running html5lib runtests.sh http://pastebin.ca/812008

[05:12:00.0000] <Philip`>
hdh: Which version of html5lib is that?

[05:12:01.0000] <hdh>
I just pulled it from svn

[05:13:00.0000] <Philip`>
Hmm, I just get a load of "AttributeError: class simplejson has no attribute 'loads'"

[05:14:00.0000] <hdh>
I got that too, and installed python-simplejson (ubuntu) to exercise those tests

[05:14:01.0000] <annevk>
I guess you're both relying on the simplejson emulater which I believe is full of bugs

[05:15:00.0000] <annevk>
oh

[05:15:01.0000] <Philip`>
/me emerges simplejson

[05:15:02.0000] <Philip`>
"ValueError: Invalid \escape: "'": line 436 column 27 (char 13046)"

[05:17:00.0000] <Philip`>
/me tries to remember how to commit a fix

[05:18:00.0000] <hdh>
btw, the LICENSE file spells "Contributers"

[05:18:01.0000] <krijnh>
Silly Dutchies :)

[05:19:00.0000] <Philip`>
/me will fix that too :-)

[05:22:00.0000] <Philip`>
hdh: Should be fixed in SVN now

[05:22:01.0000] <annevk>
krijnh, I'm you're referring to me, it might be the British who've done that

[05:22:02.0000] <annevk>
(and fixed it subsequently, which is nice)

[05:23:00.0000] <krijnh>
annevk: Ah, sorry :)

[05:23:01.0000] <krijnh>
annevk: Are there more dutchies contributing?

[05:23:02.0000] <Philip`>
"svn blame" says "296 jgraham.cantab Contributers:"

[05:24:00.0000] <Teratogen>
what's going on with ogg?

[05:24:01.0000] <Teratogen>
why was it removed from the HTML5 spec?

[05:26:00.0000] <hdh>
/me needs to figure out how to revert his "contributors" commit, bzr-svn

[05:26:01.0000] <Teratogen>
I want ogg!

[05:27:00.0000] <hdh>
donate to who ever are working on theora

[05:29:00.0000] <annevk>
the main reason is potential submarine patents as I understand it

[05:29:01.0000] <annevk>
the other reasons seem mostly FUD

[05:35:00.0000] <Camaban>
I felt http://blog.kfish.org/2007/12/html5-for-free-media-today-on-whatwg.html gave me most of the relevant info needed to understand the OGG issue :)

[05:39:00.0000] <hdh>
Philip`: ok, no E now, but 3 Fs

[05:41:00.0000] <Philip`>
hdh: I get 4 Fs (test_absolute_uri_ref_with_space_in svg_attribute test_absolute_uri_refs_in_svg_attributes test_allow_html5_image_tag test_should_sanitize_tag_broken_up_by_null)

[05:42:00.0000] <hdh>
I don't have the by_null one

[05:42:01.0000] <Philip`>
I know very little about the html5lib code so I won't try fixing those :-)

[05:42:02.0000] <Philip`>
hdh: Oh, odd

[05:42:03.0000] <Philip`>
I get AssertionError: u'&lt;scr\xef\xbf\xbdipt&gt;alert("XSS")&lt;/scr\xef\xbf\xbdipt&gt;' != u'&lt;scr\ufffdipt&gt;alert("XSS")&lt;/scr\ufffdipt&gt;'

[05:44:00.0000] <Philip`>
...maybe because I don't have a ucs2 version of Python?

[05:44:01.0000] <hdh>
ubuntu's is ucs4

[05:45:00.0000] <Philip`>
"[-    ] ucs2 (dev-lang/python): Enable byte size 2 unicode (DON'T USE THIS UNLESS YOU KNOW WHAT YOU'RE DOING)"

[05:45:01.0000] <Philip`>
Not sure what the default on Gentoo is

[05:45:02.0000] <Philip`>
Oh, it's ucs4 if you don't explicitly specify ucs2

[05:46:00.0000] <Philip`>
so that sounds like it shouldn't differ from Ubuntu

[05:47:00.0000] <hdh>
weird, if I compile python without explicitly asking for ucs4, I get undefined PyFromUCS4 or somesuch when trying to use distro's c modules

[05:47:01.0000] <Philip`>
Gentoo says "enabling UCS2 support will break your existing python modules" which sounds reasonable

[05:48:00.0000] <Teratogen>
bring back ogg!

[05:48:01.0000] <hdh>
is preferred extension for theora ogv or ogm?

[05:49:00.0000] <Teratogen>
for ogg it's .ogg

[05:49:01.0000] <Philip`>
hdh: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-October/012861.html says ogv

[05:50:00.0000] <hdh>
thx

[05:51:00.0000] <Philip`>
/me wonders how many people will write <video ... type="video/ogg; codecs="theora, vorbis"">

[05:51:01.0000] <hdh>
can't they just use " ' ' "

[05:51:02.0000] <Philip`>
Can't we use single quotes in the examples somewhere? &quot; is just really ugly

[05:51:03.0000] <Philip`>
hdh: Good idea :-)

[05:52:00.0000] <hdh>
/me likes python's strings for that

[05:52:01.0000] <Philip`>
/me has no idea if the MIME type parameters allow "'"

[05:53:00.0000] <hdh>
Philip`: uhm, there's this one http://worsethanfailure.com/Articles/Notepad-Translation-Error.aspx

[05:54:00.0000] <hdh>
maybe google cache has the text

[05:55:00.0000] <Philip`>
(Looks like MIME types use RFC822's quoted-string which only allows double-quotes)

[06:01:00.0000] <hdh>
at least xsltproc keeps ' "" ' for me

[06:22:00.0000] <Teratogen>
bring back ogg!

[06:23:00.0000] <alp>
Teratogen: you're probably doing more to hurt the adoption of ogg than to help it by participating in the debate through the wrong channels

[06:24:00.0000] <Lfe>
alp: Something with trolls and feed? :-)

[06:24:01.0000] <alp>
Lfe: heh

[06:27:00.0000] <annevk>
<video ... type=...> is pretty silly in itself

[06:53:00.0000] <zcorpan>
annevk: <address> allows structured inline

[06:53:01.0000] <zcorpan>
so possibly, <address><ul><li>

[06:54:00.0000] <annevk>
hmm

[06:54:01.0000] <zcorpan>
but it's not clear to me why that would be better

[06:54:02.0000] <annevk>
it does indeed seem way better than a simple <br>

[06:56:00.0000] <Philip`>
<br> in <address> is useful to get an hCard linebreak

[07:07:00.0000] <krijnh>
Bah, damn 'tables are bad' people

[07:07:01.0000] <krijnh>
<ul class="article_order_details cart-overview"><li class="article_price"><span class="label">Price</span><span class="valuta"><span class="valuta-sign"> ...

[07:07:02.0000] <krijnh>
:/

[07:08:00.0000] <krijnh>
So much worse than a <br> in an <address>

[07:49:00.0000] <G0k>
hey all

[07:49:01.0000] <annevk>
and bye

[07:49:02.0000] <G0k>
arg

[07:49:03.0000] <G0k>
i think we should pick h.261 and be done with it. :)

[07:50:00.0000] <annevk>
does it do streaming?

[07:50:01.0000] <hsivonen>
zcorpan, Hixie: I'm here now

[07:50:02.0000] <annevk>
is it reasonably compact?

[07:50:03.0000] <G0k>
yes. yes.

[07:51:00.0000] <G0k>
and it was finalized in 1990, so patents should be gone by now

[07:51:01.0000] <annevk>
1990...

[07:51:02.0000] <annevk>
no

[07:51:03.0000] <annevk>
that takes at least 20 years depending on your jurisdiction I believe

[07:51:04.0000] <hsivonen>
Hixie: I have to think a bit more to formulate an opinion about block/inline

[07:52:00.0000] <hsivonen>
Hixie: what you have now makes semantic sense but is impractical for authors

[07:52:01.0000] <G0k>
yeah but...html5 wont be approved until 2010+, so....

[07:52:02.0000] <hsivonen>
Hixie: specifically, bimorhic seem more like a hard-to-grasp thing than block/inline

[07:52:03.0000] <hsivonen>
I'll get back to you on this

[07:53:00.0000] <G0k>
http://www.digitalpreservation.gov/formats/fdd/fdd000035.shtml

[07:53:01.0000] <G0k>
at least they seem to think there are no patents

[07:53:02.0000] <annevk>
hmm, html5.org is getting hit hard

[07:54:00.0000] <G0k>
the loc

[07:54:01.0000] <annevk>
and svn.whatwg.org probably too because html5.org doesn't do caching :)

[07:55:00.0000] <annevk>
heh, it went from about 1000 a week to 16000 in the week that is one day old

[07:56:00.0000] <G0k>
well duh, insulting an open source project like Ogg is possibly the best way to get hits

[07:56:01.0000] <G0k>
next we should say something nasty about vi and emacs

[07:57:00.0000] <annevk>
html5.org is not insulting anyone

[07:57:01.0000] <annevk>
it just hosts a service that provides interesting views on HTML5

[07:57:02.0000] <G0k>
yeah tell that to the mob outside the door

[07:58:00.0000] <madness>
normally lurk on the mailing list - I've noticed alot of talk on reddit about the ogg stuff :/

[07:58:01.0000] <annevk>
reddit is top referrer

[07:58:02.0000] <madness>
people who don't understand that issues are more complicated than they first appear spouting off if appears.

[07:58:03.0000] <G0k>
what ogg stuff? i don't seen any ogg stuff. that is unstuff.

[08:10:00.0000] <bakarat>
http://lists.whatwg.org/pipermail/help-whatwg.org/2007-December/000094.html <- says it all

[08:11:00.0000] <G0k>
"OGG" indeed

[08:11:01.0000] <G0k>
and how exactly do you "set" a de facto standard

[08:11:02.0000] <G0k>
de facto means you don't set it

[08:12:00.0000] <madness>
it's all a bit tin-foil-hat, to be honest

[08:12:01.0000] <G0k>
"at least know that ogg will be supported by all (standards-compatible) browsers"

[08:13:00.0000] <madness>
"people are questioning ogg, a darling of open source people, so it must be because of DRM/Vendor lock-in/insert other evil reason"

[08:13:01.0000] <G0k>
because you know...we have such great luck with that

[08:13:02.0000] <annevk>
oh, I missed help⊙wo was becoming active again

[08:13:03.0000] <Dashiva>
At least this means we're getting attention :)

[08:13:04.0000] <madness>
instead of taking the reason given, which is actually a sensible one

[08:13:05.0000] <madness>
(at least for theora)

[08:13:06.0000] <G0k>
i think someone needs to make a video explaining the whole situation

[08:13:07.0000] <madness>
not sure about vorbis..

[08:14:00.0000] <gsnedders>
G0k: encoded how? :)

[08:14:01.0000] <G0k>
MJPEG

[08:14:02.0000] <Teratogen>
bring back ogg!

[08:14:03.0000] <G0k>
no patents on that

[08:14:04.0000] <Teratogen>
ogg is free

[08:14:05.0000] <annevk>
I think Hixie is ok with that as long as you pay the Google bandwidth bill G0k

[08:14:06.0000] <gsnedders>
Thezilch: so? You still, as with anything invented within the last 20 years (including Ogg/Vorbis/Theora), run the risk of a submarine patents.

[08:15:00.0000] <Dashiva>
G0k: Isn't gif also up for grabs? :P

[08:15:01.0000] <G0k>
Dashiva: gif looks like crap

[08:15:02.0000] <G0k>
8-bit color for the suck

[08:16:00.0000] <gsnedders>
How does Opera currently deal with <video>? It's own decoders? The OS?

[08:16:01.0000] <madness>
Teratogen: have you something sensible to contribute to the debate, also ?

[08:16:02.0000] <G0k>
bundles libtheora

[08:16:03.0000] <gsnedders>
s/'//

[08:18:00.0000] <gsnedders>
G0k: which I assume means it doesn't work for anything apart from Theora?

[08:18:01.0000] <G0k>
gsnedders: nope

[08:18:02.0000] <gsnedders>
G0k: nope agreeing or disagreeing?

[08:18:03.0000] <gsnedders>
silly English…

[08:18:04.0000] <G0k>
gsnedders: it only works for theora

[08:18:05.0000] <gsnedders>
OK.

[08:19:00.0000] <G0k>
as far as I can tell, annevk might have better info

[08:19:01.0000] <gsnedders>
WebKit on OS X already supports H.261 which is a slight advantage

[08:19:02.0000] <G0k>
i kinda suggested H.261 as a joke

[08:19:03.0000] <gsnedders>
it and MJPEG are the two most likely specs IMO to be allowed

[08:20:00.0000] <G0k>
the fundamental fact that is being ignored here is that no one is going to make content with a non-competitive codec

[08:20:01.0000] <gsnedders>
G0k: YouTube?

[08:20:02.0000] <annevk>
our experimental builds support Ogg/Theora/Vorbis

[08:20:03.0000] <G0k>
gsnedders: i mean given the oppurtunity

[08:21:00.0000] <gsnedders>
YouTube is H.263 actually

[08:21:01.0000] <gsnedders>
But that's far too new

[08:22:00.0000] <G0k>
gsnedders: if safari/IE/firefox ship with h.264 decoders and theora/263/261 decoders, the reason to make content in the other format is approximately zero

[08:22:01.0000] <gsnedders>
G0k: Saf/IE won't ship Theora.

[08:22:02.0000] <G0k>
right but...ok, remove theora/263 from that last sentence

[08:22:03.0000] <gsnedders>
G0k: Win32 doesn't support H.264 yet IIRC

[08:23:00.0000] <G0k>
yeah but...xboxes do, zunes do...

[08:23:01.0000] <G0k>
it's only a matter of time before windows does

[08:23:02.0000] <gsnedders>
That's true.

[08:23:03.0000] <G0k>
and how many windows users have itunes/quicktime

[08:23:04.0000] <madness>
is there a case for mandating a container format but not a codec ?

[08:23:05.0000] <madness>
the submarine concern is about codecs, not containers, right ?

[08:23:06.0000] <gsnedders>
madness: both.

[08:24:00.0000] <G0k>
a mandated container is kinda silly

[08:24:01.0000] <gsnedders>
G0k: needed, though

[08:24:02.0000] <G0k>
and if we were doing that...Ogg is frankly kinda shitting

[08:24:03.0000] <G0k>
*shitty

[08:24:04.0000] <gsnedders>
allows anything as content, though, unlike the MPEG4 one

[08:24:05.0000] <G0k>
eh you can kinda force anything in MooV

[08:25:00.0000] <G0k>
plus it's really a standard

[08:25:01.0000] <gsnedders>
the spec disallows any non-MPEG4 codec within an MPEG4 container

[08:25:02.0000] <G0k>
yeah i mean the MP4 spec itself is pretty limiting

[08:25:03.0000] <G0k>
but the ISO..."metacontainer" i guess is really flexible

[08:26:00.0000] <gsnedders>
anyone know about the content of ISO/IEC 14496-12:2005?

[08:26:01.0000] <G0k>
somewhat

[08:27:00.0000] <G0k>
it was basically an ISO-ification of the quicktime atomic container format

[08:27:01.0000] <gsnedders>
I know the outline of the other MPEG4 container standard

[08:27:02.0000] <gsnedders>
G0k: yeah, I know tht

[08:27:03.0000] <gsnedders>
*that

[08:27:04.0000] <G0k>
i know they also use it for JPEG-2000

[08:27:05.0000] <gsnedders>
ISO/IEC 14496-14 is slightly different

[08:27:06.0000] <gsnedders>
Maybe -12 doesn't limit it to MPEG4

[08:27:07.0000] <G0k>
i believe -14 is the mp4-specific parts

[08:27:08.0000] <G0k>
and -12 is the base iso format

[08:28:00.0000] <G0k>
http://mp4ra.org/

[08:28:01.0000] <G0k>
apple maintains a registry of stuff based on it

[08:28:02.0000] <gsnedders>
is it MPEG4 specific?

[08:29:00.0000] <gsnedders>
(the extras in -14)

[08:29:01.0000] <gsnedders>
I know -14 limits content to MPEG4, but I don't know what else is different from 012

[08:29:02.0000] <gsnedders>
*-12

[08:29:03.0000] <G0k>
i believe so....one sec, lemme look through my ISO docs :)

[08:29:04.0000] <gsnedders>
/me doesn't have copies of many ISO docs

[08:29:05.0000] <gsnedders>
/me doesn't have the money

[08:30:00.0000] <gsnedders>
I think ISO8601 is the only one I have, plus a draft of C99 (whichever spec that is)

[08:30:01.0000] <G0k>
heh well it looks like i did the right thing after my last contract job and deleted it. :/

[08:31:00.0000] <Teratogen>
bring back ogg!

[08:31:01.0000] <G0k>
i believe -14 added stuff like how to pack the ESDS and junk in

[08:31:02.0000] <G0k>
and hinting

[08:32:00.0000] <G0k>
(which is patent encumbered sadly)

[08:32:01.0000] <madness>
/me /ignores Teratogen 

[08:32:02.0000] <gsnedders>
Teratogen: that isn't a very useful comment.

[08:32:03.0000] <gsnedders>
G0k: ESDS?

[08:32:04.0000] <G0k>
gsnedders: MPEG-4 codec specific information

[08:32:05.0000] <G0k>
then -15 added stuff for H.264

[08:34:00.0000] <G0k>
"Most of the specifications are related to the first MPEG-4 file format (MP4 version 1), which in turn was derived from the QuickTime file format defined by Apple Computer. More recently the MP4 file format was generalized into the ISO Base Media File Format, which defines a general structure for time-based media files. It in turn is used as the basis for other file formats in the family."

[08:34:01.0000] <gsnedders>
the ISO Base Media File Format still has the known patent issue, though

[08:34:02.0000] <zcorpan>
isn't the best way for people to push ogg to publish their content as ogg? content on the web is what can make big companies want to implement ogg, not a requirement in the spec

[08:35:00.0000] <G0k>
"but i can't publish my content as Ogg until Hixie says I'm allowed to!"

[08:35:01.0000] <gsnedders>
/me would look up the quote about us being Hixie's "followers"

[08:35:02.0000] <gsnedders>
I don't always agree with what Hixie has concluded. Just because I give people the reasons why the spec is as it is doesn't mean I agree with them.

[08:36:00.0000] <G0k>
i mean it's a shitty situation but I feel like the horrible truth is that the patent-free policy is going to need to get twisted a bit for the video part

[08:36:01.0000] <gsnedders>
/me ponders

[08:37:00.0000] <gsnedders>
MUST support H.261 and SHOULD support H.264

[08:37:01.0000] <gsnedders>
would anyone complain?

[08:37:02.0000] <G0k>
it would be the same as saying MUST support raw RGB

[08:37:03.0000] <G0k>
no one is going to use it

[08:37:04.0000] <G0k>
it's not going to reflect the current or future state of the web

[08:37:05.0000] <gsnedders>
It's widely supported though :P

[08:38:00.0000] <G0k>
but yeah i mean if we had to pick a baseline...MPEG-1/H.261 is a reasonable place to start

[08:38:01.0000] <gsnedders>
MPEG-1 is too new

[08:38:02.0000] <G0k>
well MPEG-1 is just H.261 with new gunk

[08:38:03.0000] <gsnedders>
MPEG-LA still collects money for it

[08:38:04.0000] <gsnedders>
It's still too new.

[08:39:00.0000] <G0k>
see i kinda wonder

[08:39:01.0000] <G0k>
this might be a good place for a bribe

[08:40:00.0000] <G0k>
wonder if MPEG-LA would go for some kind of web exception

[08:40:01.0000] <madness>
perhaps it needs to be recognised that whatever is specified is unlikely to be competitive if we want it to be free of submarine patents

[08:40:02.0000] <gsnedders>
G0k: nobody pays to encode MP3 anyway

[08:40:03.0000] <gsnedders>
(though legally you must)

[08:41:00.0000] <gsnedders>
I think what someone who strongly believes Ogg/Theora/Vorbis is patent free needs to do is say they'll pay for any patent infringements that come up.

[08:41:01.0000] <madness>
as a baseline, if I know I can use <insert standard here> I can provide that standard, knowing it will work everywhere, and then provide better to clients that can use it

[08:41:02.0000] <doublec>
gsnedders: even mpeg-la don't do that

[08:41:03.0000] <doublec>
for h.264

[08:41:04.0000] <gsnedders>
doublec: don't do what? patent infringement? they don't claim its free of other patents, though.

[08:41:05.0000] <G0k>
doublec: technically no, but since there are so many customers of MPEG LA, you get power in numbers

[08:42:00.0000] <madness>
even if it doesn't work as well, at least I know as a small web dev, that I can get my content to any client using <insert outdated codec here>

[08:42:01.0000] <doublec>
they don't offter to pay for patent infringements that come up

[08:42:02.0000] <doublec>
it's not possible to be free of submarine patents

[08:42:03.0000] <doublec>
for any format

[08:42:04.0000] <annevk>
hi doublec, what are the plans for <video> in Firefox?

[08:42:05.0000] <gsnedders>
doublec: it is, if the standards are old enough.

[08:42:06.0000] <doublec>
annevk: we are sticking with Ogg

[08:42:07.0000] <annevk>
I saw in an article that was republished along the globe it won't make Firefox 3, but when will it hit "the market"?

[08:43:00.0000] <Philip`>
gsnedders: That kind of 'guarantee' isn't very useful for a company that gets sued for $1.5B, because there's no way the guaranteer will be able to pay that

[08:43:01.0000] <gsnedders>
doublec: people are claiming that Ogg/Theora/Vorbis is patent free. MPEG-LA don't make such a claim. That's my point.

[08:43:02.0000] <gsnedders>
Philip`: I know, that's the very problem.

[08:43:03.0000] <doublec>
gsnedders: Ah, I see

[08:43:04.0000] <doublec>
I think they are saying (or should be saying) all known patents have been dealt with

[08:43:05.0000] <annevk>
people claim it's patent free? it's public knowledge Ogg/etc. aren't

[08:44:00.0000] <annevk>
right

[08:44:01.0000] <G0k>
if anyone has good patents on H.264/AAC, the probability that they wouldn't have sued already is disappearingly small

[08:44:02.0000] <doublec>
annevk: It's under discussion as to when exactly it will hit the market

[08:44:03.0000] <G0k>
and if they did, they'd have about a trillion dollar of corporate lawyers to fight

[08:45:00.0000] <annevk>
thx

[08:45:01.0000] <annevk>
that didn't help with MP3

[08:45:02.0000] <doublec>
The issue is that we don't want to ship a partial implementation of a spec, or something that may change and we end up shipping something broken.

[08:46:00.0000] <annevk>
I see, but you don't want to wait until 2022 either :)

[08:46:01.0000] <doublec>
Yep :)

[08:47:00.0000] <doublec>
I am hoping for a point release after 3

[08:47:01.0000] <annevk>
I personally don't think that it's an issue to implement a subset of the spec initially, as long as it's implemented per the spec

[08:47:02.0000] <doublec>
I personally agree

[08:47:03.0000] <gsnedders>
annevk: you got any links to any patents that cover Ogg/etc.?

[08:47:04.0000] <annevk>
everyone does that with SVG and CSS already, if you need arguments :)

[08:47:05.0000] <G0k>
gsnedders: much of On2's portfolio....

[08:48:00.0000] <gsnedders>
… that aren't covered by the RF grants?

[08:48:01.0000] <doublec>
All the people wanting it, and the other browsers implementing it are certainly helping the case too

[08:48:02.0000] <G0k>
gsnedders: it's important to note that there are some truely ridiculous MPEG-related video patents

[08:48:03.0000] <gsnedders>
G0k: I'm aware.

[08:50:00.0000] <G0k>
there's a patent on storing integers a little bit bigger than they need to be to avoid rounding errors

[08:50:01.0000] <G0k>
then there's a second patent on the same thing...except applied to a different codec

[08:55:00.0000] <G0k>
i wonder what would be involved in making a state of the art H.261 encoder

[08:56:00.0000] <G0k>
i suppose libavcodec's encoder would be close to that

[08:57:00.0000] <Philip`>
/me wonders how many people do their own video transcoding, compared to just uploading to YouTube

[08:58:00.0000] <G0k>
well yeah but now YouTube of the future will need to do that transcoding

[08:58:01.0000] <G0k>
re-encoding reall

[08:59:00.0000] <G0k>
*really

[08:59:01.0000] <gsnedders>
They already have H.263 and for some content H.264

[09:00:00.0000] <gsnedders>
http://pastebin.ca/812265 — anyone able to eyeball that?

[09:02:00.0000] <Philip`>
It seems a harder problem if authors are doing their own encoding, since they'll need easy-to-use suitably-priced tools, which means it's nice if e.g. Windows Movie Maker can generate it directly, whereas YouTube can pay for an encoder (I assume they already do for FLV) and can do all the fiddly setup

[09:03:00.0000] <G0k>
gsnedders: looks good to me

[09:05:00.0000] <madness>
gsnedders: looks sane.

[09:05:01.0000] <Philip`>
gsnedders: Maybe you should say explicitly that the companies are already taking the submarine patent risk on e.g. MPEG4 (and occasionally getting hit by it), i.e. it's not a unique problem to Theora, but they prefer not to take on the additional risk from a new codec

[09:05:02.0000] <gsnedders>
Philip`: ah, yeah, better add that.

[09:07:00.0000] <Philip`>
('MPEG4 has as much risk as Theora' may be true, but the relevant point is that MPEG4+Theora has more risk than MPEG4 alone)

[09:08:00.0000] <G0k>
yeah that's a really excellent point there

[09:09:00.0000] <Philip`>
gsnedders: Also, it's probably correcter to say Apple doesn't intend to implement Ogg just because the spec requires it (as opposed to "even if the spec requires it")

[09:09:01.0000] <G0k>
Apple/Microsoft/Google already have patent exposure on MPEG

[09:09:02.0000] <G0k>
not so on Ogg

[09:09:03.0000] <gsnedders>
Philip`: s/correcter/more correct/

[09:09:04.0000] <gsnedders>
:)

[09:09:05.0000] <gsnedders>
:P

[09:11:00.0000] <G0k>
the best part so who wants to start a petition to get MPEG-LA to grant an exception for MPEG patent use in open source web browsers?

[09:11:01.0000] <Philip`>
gsnedders: Also, "It won't be supported by all standards-compatible browsers" is untrue in the situation you were responding to, where support was required by the standards and hence anything without support would not be standards-compatible

[09:12:00.0000] <G0k>
isn't this what eventually happened with certain accessibility patents in HTML?

[09:12:01.0000] <gsnedders>
Philip`: I thought he/she meant the (Fx/Saf/Op)s of the world

[09:13:00.0000] <gsnedders>
Philip`: "It won't be supported by all (currently) standards-compatible browsers"?

[09:13:01.0000] <Philip`>
gsnedders: Hmm, I suppose that's a reasonable interpretation

[09:13:02.0000] <gsnedders>
That covers both.

[09:14:00.0000] <Philip`>
There's IE too, which is standards-compatible (though not perfectly standards-compatible, but no other web browser is either)

[09:14:01.0000] <gsnedders>
"IIt won't be supported by all (currently) "standards-compatible" browsers"? :P

[09:14:02.0000] <G0k>
by charter what wg members?

[09:15:00.0000] <gsnedders>
G0k: huh?

[09:15:01.0000] <G0k>
s/"all (currently) standards-compatible browsers"/"original WHAT WG member-made browsers

[09:16:00.0000] <gsnedders>
ergh.

[09:16:01.0000] <gsnedders>
That cuts out too many others.

[09:16:02.0000] <gsnedders>
Konqueror, iCab, etc.

[09:16:03.0000] <G0k>
sigh

[09:16:04.0000] <G0k>
the non-IE browsers

[09:16:05.0000] <gsnedders>
or recent non-Trident browsers :P

[09:17:00.0000] <G0k>
it would be strangely microsoftian if they announced IE8 would support Ogg now

[09:17:01.0000] <gsnedders>
we're overcomplicating this.

[09:17:02.0000] <gsnedders>
http://xkcd.com/309/

[09:18:00.0000] <G0k>
heh

[09:19:00.0000] <Dashiva>
There's so much FUD in the pro-OGG mails, it's scary :)

[09:19:01.0000] <G0k>
ok i have to state a complaint here

[09:19:02.0000] <G0k>
it's "Ogg"

[09:19:03.0000] <G0k>
not "OGG"

[09:19:04.0000] <gsnedders>
http://pastebin.ca/812288?

[09:20:00.0000] <gsnedders>
Philip`: that address MPEG4 submarine well enough?

[09:20:01.0000] <Dashiva>
G0k: It's easier to keep holding the shift key than to let go and hope it doesn't become OGg

[09:20:02.0000] <gsnedders>
Dashiva: peh. learn to type.

[09:20:03.0000] <gsnedders>
Dashiva: or do what I do in IRC and use very few capitals :)

[09:21:00.0000] <Dashiva>
But ogg is also wrong :P

[09:21:01.0000] <G0k>
ogg is easier on the eyes than OGG

[09:21:02.0000] <Philip`>
It's easier if someone on this channel called themselves Ogg, and then we could use tab completion

[09:21:03.0000] <Philip`>
s/It's/It would be/

[09:21:04.0000] <Dashiva>
You're a genius

[09:21:05.0000] <G0k>
it's like when people call apple's computing platform "MAC"

[09:21:06.0000] <gsnedders>
Nobody on freenode call Ogg. go ahead.

[09:22:00.0000] <gsnedders>
s/call/called/

[09:22:01.0000] <Philip`>
gsnedders: Sounds good to me now :-)

[09:22:02.0000] <gsnedders>
/me has found one more thing to add

[09:22:03.0000] <doublec>
"don't want to take the risk of being sued for submarine patents which are known to exist for Ogg/Vorbis/Theora"

[09:22:04.0000] <gsnedders>
/me wonders whether it is worth CC'ing to public-html to give all this reasoning there too, or whether that'll cause chaos

[09:22:05.0000] <doublec>
what are the known submarine patents?

[09:22:06.0000] <Dashiva>
So one of the mails suggests we should just let google decide... that's novel

[09:22:07.0000] <doublec>
if they were known, they wouldn't be submarine

[09:24:00.0000] <gsnedders>
/me nudges annevk 

[09:24:01.0000] <G0k>
heh you know

[09:24:02.0000] <G0k>
Google Video Codec 

[09:25:00.0000] <annevk>
uh?

[09:25:01.0000] <G0k>
kinda rolls off the tongue

[09:25:02.0000] <gsnedders>
"Also, if it a MUST everyone in the WG would be issuing a RF license covering any patents they hold covering Ogg/Vorbis/Theora to everyone else in the WG (as per <http://www.w3.org/Consortium/Patent-Policy-20040205/#def-essential>), which companies such as MS and Nokia have said they are unwilling to do"; does that sound all right as a penultimate paragraph?

[09:25:03.0000] <Dashiva>
ogg = open google graphics!

[09:25:04.0000] <gsnedders>
annevk: you among others have made mention of known patents covering Ogg/etc

[09:26:00.0000] <G0k>
FLV =  feeling lucky video

[09:26:01.0000] <annevk>
see http://en.wikipedia.org/wiki/Ogg_Theora

[09:26:02.0000] <annevk>
second paragraph

[09:26:03.0000] <gsnedders>
annevk: ah, you were just referring to those patents

[09:27:00.0000] <annevk>
everyone is, afaict

[09:27:01.0000] <Philip`>
gsnedders: Maybe conclude with some point encouraging people to try to establish Theora as a de-facto standard because that'll be the most effective way of getting browser support for it?

[09:28:00.0000] <gsnedders>
annevk: some people have said there are known patents outwith of those

[09:28:01.0000] <Philip`>
(to avoid sounding like you don't want Theora to succeed)

[09:28:02.0000] <G0k>
great, now we get a drive of people making the mediocre de facto. splendid.

[09:29:00.0000] <doublec>
gsnedders: yes, but a lot of people are saying things that aren't necessarily true. On both sides of course.

[09:29:01.0000] <Philip`>
I remember people saying their legal departments thought there was probably a significant risk of submarine patents, but don't remember anyone saying they knew of any in particular (or had even looked for any)

[09:29:02.0000] <gsnedders>
/me wouldn't have remembered it'd been said if it weren't someone he trusts

[09:29:03.0000] <doublec>
better I think would be to scratch the 'which are known to exist'

[09:30:00.0000] <doublec>
It's true they don't want to take on the risk of submarine patents

[09:30:01.0000] <gsnedders>
doublec: already done :)

[09:30:02.0000] <doublec>
cool :)

[09:30:03.0000] <doublec>
The risk is on their for whatever codec they choose. It's just they've already taken the risk for one particular codec.

[09:31:00.0000] <doublec>
and aren't keen to take it on for another

[09:31:01.0000] <doublec>
from what I can gather

[09:31:02.0000] <gsnedders>
doublec: well, more than one codec :P

[09:31:03.0000] <doublec>
and bonus if they actually get income from various patents related to those codecs too :)

[09:32:00.0000] <csarven>
how does a browser know (or the process that it goes through) that a deprecated block level element is to be displayed block in HTML 4.10 Strict?

[09:32:01.0000] <csarven>
(i only tested <menu> in Firefox2)

[09:33:00.0000] <gsnedders>
Philip`: "If you truly do want make no compromises yourself, you may be able to get the major browser manufacturers that are currently unwilling to implement Ogg/Vorbis/Theora by getting a critical mass of content already out there. Bare in mind, though, that MS still does not support MPEG-4 out of the box (except for Zune), despite the huge amount of MPEG-4 content already out there"?

[09:33:01.0000] <Philip`>
csarven: The same as how it knows an undeprecated block level element is to be displayed block

[09:34:00.0000] <Dashiva>
One point many seem to gloss over is that Apple and MS already had video codecs for their OS anyway, it's not like they paid for them for browser use

[09:34:01.0000] <csarven>
Philip` that means that the browser displays a set of elements in their default displays regardless of the doctype

[09:35:00.0000] <Philip`>
gsnedders: s/Bare/Bear/, and also the first sentence misparses as "... major browser manufacturers that are (currently unwilling to implement Ogg/...) by getting ..."

[09:35:01.0000] <gsnedders>
s/Theora/Theora to implement them/?

[09:36:00.0000] <G0k>
Dashiva: yeah but that's unfair to Opera + Mozilla

[09:36:01.0000] <Dashiva>
It's not about whether it's fair or not, it's just a fact

[09:37:00.0000] <csarven>
Philip` so then how does the browser differentiate between a set of elements thats to be displayed and the DTD?

[09:37:01.0000] <Philip`>
csarven: Yes - browsers support all elements regardless of the doctype, and the only effect of the doctype is to trigger certain buggy rendering behaviour

[09:37:02.0000] <gsnedders>
I also imply that MS support MPEG-4, which is untrue

[09:37:03.0000] <Philip`>
csarven: It just ignores the DTD

[09:38:00.0000] <csarven>
whats an example of a buggy rendering behaviour for the doctypes?

[09:38:01.0000] <Philip`>
gsnedders: That makes my brain parse it correctly :-)

[09:39:00.0000] <gsnedders>
Philip`: re-reading caused a fatal error in my brain too :)

[09:39:01.0000] <gsnedders>
csarven: IE box model

[09:39:02.0000] <doublec>
on the mailing list is mentioned a analysis by Dave Singer regarding codecs. Does anyone have a link to that?

[09:39:03.0000] <Lachy>
csarven, look up Quirks Mode

[09:40:00.0000] <gsnedders>
doublec: <http://www.w3.org/mid/p06240820c379fc1ecf69@%5B17.202.35.52%5D>

[09:40:01.0000] <doublec>
thanks gsnedders 

[09:41:00.0000] <Lachy>
csarven, http://developer.mozilla.org/en/docs/Mozilla_Quirks_Mode_Behavior

[09:41:01.0000] <csarven>
if i understand this correctly; the elements that the browser displays (how and what to do with it) is not borrowed from a specific DTD?

[09:42:00.0000] <csarven>
it belongs to SUPER_DTD of some sort?

[09:42:01.0000] <Lachy>
csarven, correct. The DTD is completely irrelevant to browsers.

[09:42:02.0000] <G0k>
aw man support vector machines are making my computer cry

[09:42:03.0000] <Philip`>
csarven: See what they said, or e.g. http://tinyurl.com/2act2m vs http://tinyurl.com/23pyfv

[09:43:00.0000] <Philip`>
(where "bgcolor" is supported in both cases, but invalid values in the bgcolor are only supported in one case, at least in a few of the browsers)

[09:43:01.0000] <csarven>
alright cool. thanks :)

[09:43:02.0000] <gsnedders>
I think <http://pastebin.ca/812313> addresses all the feedback

[09:43:03.0000] <gsnedders>
ergh. quotes aren't marked.

[09:44:00.0000] <gsnedders>
I think you guys know what the quotes are by now :)

[09:45:00.0000] <Philip`>
csarven: The set of elements supported by a browser is usually just spread through all its C++ code - they don't have a simple list saying what exists

[09:45:01.0000] <gsnedders>
<http://pastebin.ca/812315> marks the quotes

[09:45:02.0000] <Lachy>
gsnedders, why didn't you just reply in a proper email

[09:46:00.0000] <gsnedders>
Lachy: "proper" how?

[09:46:01.0000] <gsnedders>
Lachy: all this is is a draft :P

[09:48:00.0000] <Philip`>
gsnedders: Still s/Bare/Bear/ :-p

[09:48:01.0000] <gsnedders>
Philip`: see, that's why I said thought :)

[09:48:02.0000] <Lachy>
gsnedders, ok. didn't realise it was just a draft.

[09:49:00.0000] <gsnedders>
Lachy: the fact I've now been editing it for 49 minutes may make you think that, yeah :)

[09:49:01.0000] <gsnedders>
(and have now been dragged into an IM convo)

[09:52:00.0000] <Lachy>
gsnedders, I've only been watching the conversation for 5 minutes, so I didn't know

[09:53:00.0000] <gsnedders>
Is it worth copying it to public-html to explain a lot of it there, or will that cause more chaos than it is worth.

[09:53:01.0000] <Lachy>
when did MS and Nokia say they were unwilling to give up any patents they have covering ogg theora?

[09:54:00.0000] <gsnedders>
week or two ago

[09:54:01.0000] <Lachy>
it's highly unlikely that they actually have any patents covering it anyway

[09:54:02.0000] <gsnedders>
MS have a large number of video/audio coded patents, though

[09:55:00.0000] <Philip`>
gsnedders: Seems worthwhile to send/copy to whatwg⊙wo, but probably not worth dragging public-html into the unproductive discussion since there hasn't been any misinformation there yet

[09:56:00.0000] <Philip`>
Oh, was it in response to a whatwg⊙. post?

[09:56:01.0000] <Philip`>
Oh, no, it was the help@ one

[09:56:02.0000] <gsnedders>
Philip`: it was sent to both, IIRC

[09:56:03.0000] <Philip`>
/me was confused by the line wrapping and didn't recognise it

[09:57:00.0000] <gsnedders>
Philip`: yeah, both

[09:57:01.0000] <Lachy>
I'd rather not see the debate continue much longer. It will continue to be entirely unproductive

[09:58:00.0000] <Philip`>
gsnedders: Oh, so it was, but I don't seem to have received that message

[09:59:00.0000] <Philip`>
I have eight messages under "[whatwg] Removal of Ogg is *preposterous*" and nothing else :-/

[10:00:00.0000] <G0k>
i honestly can't read the word "preposterous" without thinking of http://www.youtube.com/watch?v=t-3qncy5Qfk

[10:02:00.0000] <Philip`>
I think we should find an area that isn't actually controversial, but make a controversial change, and then spread it on Digg and Reddit and everywhere and get people to sign up to the mailing list and argue against it, and then wait a few days and say "wow, you're right, let's change back" and then we'll have a few dozen more members than before

[10:02:01.0000] <gsnedders>
:D

[10:03:00.0000] <gsnedders>
So more irrational arguments?

[10:03:01.0000] <G0k>
yeah

[10:03:02.0000] <gsnedders>
Philip`: should we make XHTML illegal?

[10:04:00.0000] <G0k>
"Ron Paul and Denis Kucinich don't support HTML 5!" 

[10:05:00.0000] <G0k>
wow "an outrageous disaster"

[10:07:00.0000] <Dashiva>
gsnedders: On pain of death, even

[10:07:01.0000] <Philip`>
/me hasn't received any emails after http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-December/013168.html :-(

[10:07:02.0000] <Philip`>
Some of them weren't about Ogg, so I hope they turn up eventually

[10:09:00.0000] <G0k>
i'm suprised we haven't got any matroska trolls yet

[10:09:01.0000] <Dashiva>
matroska is the best codec in the world when used in a mov container

[10:09:02.0000] <Dashiva>
there, happy?

[10:09:03.0000] <G0k>
excellent

[10:09:04.0000] <G0k>
no wait

[10:10:00.0000] <G0k>
matroska is the best codec in the world when used in a theora container

[10:10:01.0000] <Philip`>
http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-December/013120.html ?

[10:10:02.0000] <G0k>
especially when combined with tarkin audio

[10:11:00.0000] <G0k>
heh ohz

[10:11:01.0000] <gsnedders>
http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-December/013186.html

[10:12:00.0000] <Philip`>
By the way, "Matroska trolls" is an excellent phrase

[10:12:01.0000] <G0k>
i guess i fail to understand why it's been deemed required to specify audio and video formats while image formats worked fine without specifications

[10:14:00.0000] <MikeSmith>
if Matroska trolls, then Oggres?

[10:15:00.0000] <annevk>
G0k, image formats will be in there in due course

[10:16:00.0000] <G0k>
MikeSmith: Imp-egs

[10:16:01.0000] <G0k>
(a stretch i know)

[10:17:00.0000] <annevk>
one of these slashdot posts points to http://www.w3.org/html/wg/html5/diff/ which seems wrong

[10:17:01.0000] <MikeSmith>
G0k - I raise the Goblet of Rock unto you for the effort

[10:20:00.0000] <G0k>
uhm

[10:21:00.0000] <G0k>
where the hell is rudd-o getting is information

[10:21:01.0000] <Dashiva>
G0k: Imp Eggs if so :)

[10:21:02.0000] <G0k>
" Where would we be today if the HTML 

[10:21:03.0000] <G0k>
spec didn't specify jpg, gif, and png as baseline standards for the image 

[10:21:04.0000] <G0k>
tag?"

[10:21:05.0000] <G0k>
you mean like...the way it doesn't?

[10:22:00.0000] <G0k>
all HTML 4.0.1 says is " Examples of widely recognized image formats include GIF, JPEG, and PNG."

[10:23:00.0000] <G0k>
says nothing about should, may, must....

[10:23:01.0000] <G0k>
In fact, you know what, I'm going to say that

[10:23:02.0000] <tndH>
the requirement for ogg should be in a <blink>, see if we get a new /. story every few seconds

[10:23:03.0000] <annevk>
HTML5 will require those, fwiw

[10:23:04.0000] <annevk>
probably also APNG

[10:24:00.0000] <G0k>
yeah but....the argument that JPEG/GIF/PNG only happened because the HTML standard mandated them is not valid

[10:24:01.0000] <annevk>
yeah, we've been lucky with those

[10:24:02.0000] <gsnedders>
annevk: but patents on them?

[10:25:00.0000] <gsnedders>
annevk: or are we just ignoring them?

[10:25:01.0000] <annevk>
/me is confused

[10:25:02.0000] <Lachy>
there were patents covering GIF compression, but not decomopression, so browsers could implement it freely

[10:26:00.0000] <Dashiva>
What does Apple gain (not save) from mpeg use growing?

[10:27:00.0000] <roc>
when browsers started supporting GIF they didn't know about the patents

[10:27:01.0000] <G0k>
Dashiva: takes market share away from Windows Media

[10:27:02.0000] <gsnedders>
Dashiva: patent charges

[10:27:03.0000] <G0k>
Dashiva: plus they have 200 million ipods which play MPEG

[10:28:00.0000] <gsnedders>
Lachy: and JPEG/PNG?

[10:28:01.0000] <G0k>
so they want their customers to be able to play them

[10:28:02.0000] <Lachy>
don't know about jpeg.

[10:29:00.0000] <Dashiva>
gsnedders: So they do get money? Okay, was wondering

[10:29:01.0000] <gsnedders>
Dashiva: I'm not sure what the cuts are between the patent holders. it's 0.75USD in total per implementation, Apple will likely get less than 0.10USD

[10:29:02.0000] <Lachy>
png was designed to avoid all patents somehow

[10:30:00.0000] <G0k>
there's also a dirty little secret about the MPEG patents

[10:30:01.0000] <G0k>
they have caps on how much money a company can pay

[10:30:02.0000] <G0k>
so Apple just pays the cap every year

[10:30:03.0000] <G0k>
and then they're covered

[10:31:00.0000] <Dashiva>
Also, re:jpeg/gif/png, IE6 had awesome support for PNG, amirite

[10:31:01.0000] <G0k>
yeah clearly mandating PNG as a W3C spec really helped there

[10:31:02.0000] <gsnedders>
G0k: $300k, IIRC

[10:32:00.0000] <G0k>
gsnedders: er i think it's more like $3.5 million now

[10:32:01.0000] <Dashiva>
/me realizes that if this conversation was about accessibility, there would be at least two formal objections raised by now

[10:32:02.0000] <gsnedders>
G0k: it was mentioned on public-html/whatwg a while back, I'm just trying to remember :P

[10:32:03.0000] <G0k>
yeah i mean the point is

[10:32:04.0000] <gsnedders>
you pay per shipped binary implementation of a standard up to that limit, though.

[10:32:05.0000] <G0k>
the way patent pools are set up

[10:32:06.0000] <G0k>
it's highly anti-competitive

[10:33:00.0000] <G0k>
because big companies just pay the cap

[10:33:01.0000] <G0k>
little companies are fucked

[10:33:02.0000] <gsnedders>
it's a huge number of copies you need to ship, though

[10:33:03.0000] <G0k>
yeah so...apple doesn't care about selling 100 million ipods, microsoft about 10 million xboxes...

[10:34:00.0000] <G0k>
me and my friends who want to sell 100,000 copies of software, we get screwed

[10:34:01.0000] <doublec>
4.25 million for 2007/2008 

[10:34:02.0000] <doublec>
5 million for 2009/2010

[10:35:00.0000] <doublec>
according to: http://www.mpegla.com/avc/AVC_TermsSummary.pdf

[10:35:01.0000] <G0k>
alright so...point is, apple net pays like 2 cents per ipod

[10:35:02.0000] <G0k>
me, i'd pay more like $2

[10:35:03.0000] <G0k>
that's a huge competitive advantage apple has

[10:36:00.0000] <G0k>
so yeah, that's another reason apple might have to make sure that mpeg spreads

[10:38:00.0000] <annevk>
so the storage spec changed, was Mozilla contacted?

[10:38:01.0000] <doublec>
Paying that money still doesn't prevent problems with patent holders who aren't part of the mpeg-la wanting to negotiate their own agreements

[10:39:00.0000] <doublec>
or those that try to work around the mpeg-la agreement: http://www.cdrinfo.com/Sections/News/Details.aspx?NewsId=21774

[10:39:01.0000] <gsnedders>
G0k: I truly doubt there is anything except submarines causing issues

[10:39:02.0000] <annevk>
there's https://bugzilla.mozilla.org/show_bug.cgi?id=337311#c18

[10:40:00.0000] <G0k>
doublec: they're going to have to fight those battles anyway, the question is whether they want to fight more battles too

[10:41:00.0000] <gsnedders>
G0k: to take the MP3 submarine case, it's the diff between 1.52 giga-dollars, and 3.04 giga-dollars, a big difference.

[10:41:01.0000] <gsnedders>
(yes, I did just use SI prefixes for money :))

[10:42:00.0000] <Dashiva>
It's a good idea if you ask me

[10:43:00.0000] <Dashiva>
Saves the whole "what is a billion" issue

[10:43:01.0000] <gsnedders>
yeah, that's half the reason I used it there

[10:43:02.0000] <gsnedders>
but there again, what is a dollar?

[10:43:03.0000] <G0k>
gsnedders: is that 1.52 * 2^30 dollars or 1.52 * 10^9?

[10:43:04.0000] <gsnedders>
G0k: see the SI prefixes.

[10:43:05.0000] <Dashiva>
gibi-dollars? :)

[10:43:06.0000] <gsnedders>
G0k: I don't go for the non-standard former option.

[10:43:07.0000] <G0k>
ah not gibi-bucks

[10:45:00.0000] <Dashiva>
Poor gsnedders, all alone against the email horde 

[10:45:01.0000] <gsnedders>
/me has stopped

[10:45:02.0000] <G0k>
i'm with you! kinda

[10:45:03.0000] <gsnedders>
well, I'm not writing anything that long again

[10:45:04.0000] <gsnedders>
anything I can answer in three sentences or less I'll do.

[10:47:00.0000] <Dashiva>
That makes me wonder, how would you summarize HTML5 in three sentences or less

[10:47:01.0000] <G0k>
Awesome. New. Standard.

[10:47:02.0000] <G0k>
or

[10:47:03.0000] <G0k>
Best. HTML. Ever.

[10:48:00.0000] <Dashiva>
Shouldn't it be "yet"? :)

[10:48:01.0000] <G0k>
well, summary needs to lose a little precision

[10:48:02.0000] <Dashiva>
true

[10:48:03.0000] <gsnedders>
Defined parsing for text/html. HTML/XHTML both defined in a single document relative to their DOM. <canvas>, <embed>, <video>, and <audio> defined and conformant.

[10:48:04.0000] <G0k>
technically, those are are sentence fragments

[10:49:00.0000] <G0k>
i think hixie should do one of his patent pending "see how often a feature is used" thing to see how many theora links actually exist

[11:05:00.0000] <Dashiva>
By the way, when do I get my share of the quick buck? :)

[11:27:00.0000] <bradee-goesHome>
I was just thinking the same thing

[11:27:01.0000] <bradee-goesHome>
Hixie: are you holding out on us???

[11:27:02.0000] <gsnedders>
I must admit I do like how everyone thinks that Hixie must've been paid for this.

[11:38:00.0000] <Philip`>
I find it strange that so many people get trapped inside King's College - whenever I'm there in an evening, there's always one or two people standing by the (locked) back gate, looking a bit lost and confused, occasionally shaking the gate a little, until a kind student with a key happens to come along and let them out

[11:41:00.0000] <Philip`>
/me assumes the average inter-arrival time is low enough that nobody is going to freeze to death while waiting, but isn't entirely confident about that

[11:43:00.0000] <gsnedders>
Philip`: the back gate, on the far side of the Cam?

[11:44:00.0000] <Philip`>
Yes

[11:44:01.0000] <gsnedders>
/me never knew that got locked

[11:45:00.0000] <gsnedders>
there again, I've never been through King's on my own

[11:45:01.0000] <gsnedders>
Always been with my mother (daughter of former teacher) or sister (who has an MA from Cambridge)

[11:47:00.0000] <Philip`>
It gets locked at about 18:30, at least around this time of year, but it seems most people don't know that until just when they're trying to get through it

[11:47:01.0000] <gsnedders>
Long way back, if you have to go back :\

[11:47:02.0000] <gsnedders>
Then a long way around.

[11:49:00.0000] <Philip`>
Observational evidence indicates that you can simply climb around the side of the gate where it's above a sometimes-river sometimes-ditch, so it's not exactly great security

[11:50:00.0000] <gsnedders>
Are the fences round Kings that small?

[11:53:00.0000] <Hixie>
holy crap. given how many people thought that the spec saying Ogg actually meant we had settled on Ogg, I'm extremely glad I made teh change.

[11:53:01.0000] <Philip`>
There aren't any fences at that point (except of the tiny "please keep off the grass" variety), perhaps since they assume (quite reasonably) the only people stupid enough to climb around that way are students who should be allowed in anyway

[11:54:00.0000] <gsnedders>
ah. true.

[11:54:01.0000] <gsnedders>
and it isn't and overly nice ditch.

[11:54:02.0000] <gsnedders>
*an

[11:55:00.0000] <Teratogen>
bring back ogg!

[11:56:00.0000] <annevk>
man, media query parsing is painful

[11:57:00.0000] <Philip`>
/me wonders if it's possible to make something equivalent to BNF that does all the error handling stuff that's needed in practice

[12:00:00.0000] <annevk>
Philip`, is it possible to do another HTML survey checking values of media="" ?

[12:01:00.0000] <annevk>
on <link> and <style> on hopefully a lot of sites?

[12:02:00.0000] <Philip`>
Looks like ~10% of pages have media, so it should be reasonable to collect some number of them

[12:02:01.0000] <hsivonen>
annevk: re: partial <video>: actually, it's important to get the <source> selection right enough not to poison it for the future

[12:03:00.0000] <annevk>
not having <source> would suffice for now

[12:04:00.0000] <roc>
hmm

[12:04:01.0000] <roc>
what's this about partial <video>?

[12:05:00.0000] <annevk>
i think it was about a release of Firefox that implemented a subset of HTML5

[12:05:01.0000] <annevk>
<video>

[12:05:02.0000] <roc>
was doublec involved?

[12:06:00.0000] <annevk>
yeah

[12:06:01.0000] <Philip`>
hsivonen: I think you suggested HttpClient in the past, so I was wondering if you happen to know if there's an easy way to cache responses? (Ideally it would work exactly like normal except it would never download the same file twice)

[12:06:02.0000] <roc>
regardless, can someone summarize to the list?

[12:06:03.0000] <roc>
/me has to go

[12:07:00.0000] <Philip`>
(I've not found anything better than manually saving the URI/headers/body into some kind of databasey thing)

[12:08:00.0000] <annevk>
Philip`, cool, that would be useful

[12:08:01.0000] <annevk>
/me needs to know how much web pages rely on crazy-ass stuff in media=""

[12:12:00.0000] <Hixie>
hey

[12:12:01.0000] <Hixie>
all this talk about codecs has done wonders for our membership

[12:12:02.0000] <annevk>
Hixie, you can do this too maybe

[12:12:03.0000] <Hixie>
we're up to 840+

[12:12:04.0000] <annevk>
Hixie, research values of media=""

[12:13:00.0000] <dglazkov>
Hixie, quick, zap something else from the spec. We can quadruple the membership by the end of the week!

[12:13:01.0000] <annevk>
whether or not we can drop the crazy-ass HTML4 rules for what's specified now in HTML5 is depending on such a thing

[12:13:02.0000] <Hixie>
probably won't happen for a while, but if you have a specific request, e.g. "top 200 values", send mail to ianh⊙gc

[12:13:03.0000] <Hixie>
annevk: ^

[12:13:04.0000] <Hixie>
dglazkov: hah

[12:14:00.0000] <annevk>
k, hopefully your future data will back up Philip`'s data :)

[12:14:01.0000] <hsivonen>
annevk: HTML5 probably shouldn't require JPEG but the subset of JFIF/JPEG supported by the IJG code (that is, no arithmetic coding).

[12:14:02.0000] <gsnedders>
For an I-D/RFC, should a reference that needs to be read to understand an informative section be normative or not?

[12:14:03.0000] <Hixie>
oh and anne, anytime you want to add caching to the diff script, please feel free to do so :-P

[12:15:00.0000] <annevk>
hehe

[12:15:01.0000] <gsnedders>
/me wonders why Hixie wants caching :P

[12:16:00.0000] <Hixie>
it certainly has nothing to do with the diff script being posted to reddit, no sirree

[12:17:00.0000] <annevk>
i've yet to get the first offer for advertizements on html5.org

[12:17:01.0000] <doublec>
does the opera build with <video> support <source>?

[12:17:02.0000] <annevk>
no

[12:18:00.0000] <annevk>
our implementation is pretty basic

[12:18:01.0000] <gsnedders>
Hixie: why would it? I mean, it wouldn't mean running a diff on every request, would it?

[12:18:02.0000] <hsivonen>
Philip`: I don't know about HttpClient caching. for Validator.nu, I don't want caching, because it is rare to want to validate a cache page (except if Slashdot links to a validation result)

[12:18:03.0000] <doublec>
I've got partial support for <source>. No media= yet

[12:18:04.0000] <annevk>
<video controls src=theora>

[12:18:05.0000] <Hixie>
gsnedders: right now every time you hit anne's page it causes my subversion server to use 50% CPU for a second or so

[12:18:06.0000] <annevk>
and some scripting

[12:18:07.0000] <annevk>
and a closing tag

[12:19:00.0000] <gsnedders>
Hixie: yeah, I know. :)

[12:19:01.0000] <doublec>
And we only use the <source> with a type of video/ogg

[12:19:02.0000] <gsnedders>
Hixie: I mean, we're never sarcastic, ever.

[12:19:03.0000] <Hixie>
hah

[12:19:04.0000] <gsnedders>
I mean, I don't need to do my homework due on 20071103.

[12:19:05.0000] <gsnedders>
sorry, 20071203.

[12:19:06.0000] <_Ivo>
Hixie: regarding one of your recent reddit posts, if everyone including Apple is looking out for a solution to the problem, it may be worth considering putting a MUST there to avoid the possible submarine issues.

[12:19:07.0000] <gsnedders>
not quite that overdue.

[12:20:00.0000] <gsnedders>
_Ivo: that would only cover patents held by WG members.

[12:20:01.0000] <gsnedders>
_Ivo: there are many more people who hold patents outwith of the WG.

[12:20:02.0000] <_Ivo>
gsnedders: I see.

[12:21:00.0000] <gsnedders>
/me bursts out laughing reading the RFC Editor's instructions2authors.txt

[12:22:00.0000] <gsnedders>
"Note that in past years the RFC Editor has sometimes published serious documents with April 1 dates.  Readers who cannot distinguish satire by reading the text may have a future in marketing".

[12:22:01.0000] <Dashiva>
So Hixie, how much did they pay you?

[12:22:02.0000] <dglazkov>
one

[12:22:03.0000] <dglazkov>
million

[12:22:04.0000] <dglazkov>
dollars!

[12:22:05.0000] <gsnedders>
earthlings

[12:22:06.0000] <hsivonen>
annevk: shipping with src='' support without <source> support seems like a distincly bad idea unless the spec is modified to allow both src='' and <source> as conforming such that the src='' is used as the last resort if every <source> fails

[12:22:07.0000] <csarven>
[15:19:37] <Hixie> all this talk about codecs has done wonders for our membership -- what membership are you referring to?

[12:23:00.0000] <hsivonen>
csarven: whatwg list most likely

[12:23:01.0000] <annevk>
hsivonen, no <video src> overrides <source>

[12:23:02.0000] <annevk>
it's a shorthand

[12:23:03.0000] <csarven>
/me checks to see if he is on that list-thingy

[12:23:04.0000] <hsivonen>
annevk: that sucks big time for forward compat

[12:23:05.0000] <csarven>
i think i pulled out last year :S

[12:23:06.0000] <annevk>
i don't follow

[12:24:00.0000] <Hixie>
_Ivo: how do you mean?

[12:24:01.0000] <Hixie>
Dashiva: seriously, if there was money in this, i'd be so much happier :-P

[12:24:02.0000] <Hixie>
csarven: whatwg

[12:24:03.0000] <hsivonen>
annevk: suppose Opera ships with src='' only and with Ogg/Vorbis/Theora only

[12:25:00.0000] <hsivonen>
annevk: then later a browser comes along that supports Ogg/Vorbis/Dirac

[12:25:01.0000] <Dashiva>
Hixie: Yeah, I suppose google took it all :P

[12:25:02.0000] <Hixie>
i'm a little disturbed that slashdot summaries are on a first-name basis with me

[12:25:03.0000] <hsivonen>
annevk: how would you author content such that it is conforming, doesn't use scripting to sniff Opera and would still cause Theora to play in legacy Opera and Dirac in new browsers?

[12:26:00.0000] <annevk>
you wouldn't

[12:26:01.0000] <hsivonen>
annevk: hence, sucks big time for forward compat

[12:26:02.0000] <annevk>
i'm not convinced that authors will offer multiple video streams

[12:26:03.0000] <roc>
I think we can implement <source> for FF3

[12:26:04.0000] <annevk>
but this is experimental

[12:26:05.0000] <roc>
I'm more worried about partial DOM APIs

[12:26:06.0000] <doublec>
A developer from another browser asked if I'd be willing to add support for <source> so they could demo a page on their browser at firefox at the w3c workshop

[12:27:00.0000] <doublec>
on the same page, with it falling back to the ogg source on firefox and mp4 on theirs

[12:27:01.0000] <doublec>
which is why I added the currently limited support

[12:28:00.0000] <annevk>
hmm, media= on <source>

[12:28:01.0000] <annevk>
yet another hole for interop errors

[12:28:02.0000] <Dashiva>
media="*"

[12:28:03.0000] <hsivonen>
annevk: I think the codecs hole is bigger

[12:29:00.0000] <annevk>
hsivonen, I would expect market pressure to get people to implement <source>

[12:29:01.0000] <doublec>
where is 'a valid media query' defined?

[12:29:02.0000] <doublec>
the link in the spec appears to go nowhere

[12:29:03.0000] <roc>
CSS?

[12:30:00.0000] <annevk>
http://www.w3.org/TR/css3-mediaqueries/

[12:30:01.0000] <doublec>
thanks

[12:30:02.0000] <hsivonen>
doublec: dbaron said on www-style that he has code

[12:30:03.0000] <doublec>
thanks hsivonen 

[12:30:04.0000] <annevk>
however, the parsing details for media="" are not entirely clear yet

[12:30:05.0000] <annevk>
but i guess that doesn't matter much initially

[12:31:00.0000] <hsivonen>
I guess I should write tests that/demos that mix src='' and <source>

[12:34:00.0000] <Hixie>
"Where would we be today if the HTML spec didn't specify jpg, gif, and png as baseline standards for the image tag?"

[12:34:01.0000] <Hixie>
hah

[12:34:02.0000] <Hixie>
i wonder if the person saying that knew what they were saying

[12:34:03.0000] <hsivonen>
annevk: it's not unreasonable to expect people to develop automated encoding solutions that take a video file and spit out multiple formats and HTML that uses <source> for them

[12:34:04.0000] <annevk>
i haven't seen a precedent for that

[12:35:00.0000] <hsivonen>
annevk: Google Video already gives you 2 formats automatically

[12:35:01.0000] <Philip`>
By "people", do you mean "normal people" or do you mean "YouTube/etc developers"?

[12:35:02.0000] <hsivonen>
annevk: it used to give you 4!

[12:35:03.0000] <roc>
Wikipedia serves video in multiple formats

[12:35:04.0000] <_Ivo>
[20:38] <roc> Wikipedia serves video in multiple formats << I think you meant Internet Archive; Wikipedia does only Theora

[12:35:05.0000] <hsivonen>
Philip`: I mean YouTube devs, blip.tv devs, and people who write scripts like the ones markp wrote for his video podcasts

[12:36:00.0000] <Philip`>
archive.org gives 13 different versions of one of the videos on their front page

[12:36:01.0000] <roc>
sorry, yes, you're right. I was thinking of something else

[12:36:02.0000] <Philip`>
http://www.archive.org/details/VisittoS1963 - mostly various MPEGs, plus Real and Cinepack

[12:37:00.0000] <Philip`>
and animated GIF, but I don't think that counts

[12:37:01.0000] <Philip`>
and embedded FLV too

[12:42:00.0000] <aphid>
right, if those 13 formats are uploaded by a user.  pretty sure they only transcode to mp4 and mpeg1 though

[12:42:01.0000] <Philip`>
hsivonen: About caching: Okay, thanks - I'll just try the obvious/dumb caching method

[12:42:02.0000] <hsivonen>
Philip`: HttpClient has a lot of config but it is tedious to find what is where

[12:43:00.0000] <hsivonen>
a couple of months ago I made an effort to find an online video hosting service that'd generate .ogg, .flv and .mp4 for me from any reasonable source file

[12:43:01.0000] <hsivonen>
I found none

[12:44:00.0000] <hsivonen>
blip.tv will host the files if the content provider encodes the files 

[12:44:01.0000] <hsivonen>
blip.tv only generates .flv

[12:44:02.0000] <hsivonen>
I was rather disappointed

[12:45:00.0000] <hsivonen>
(moreover, blip.tv is for episodic shows and doesn't work well as a hosting service for isolated videos)

[12:45:01.0000] <Philip`>
FFmpeg/MEncoder make it 'easy' to convert into pretty much anything, though with varying quality-of-implementation for different formats, and with lots of pain when trying to find the best encoder settings the first time you set it up

[12:47:00.0000] <hsivonen>
Philip`: that's why a driver script written by e.g. markp is needed

[12:47:01.0000] <doublec>
but using that for a commercial service would be risky?

[12:47:02.0000] <doublec>
according to ffmpeg's license faq

[12:47:03.0000] <hsivonen>
doublec: the word out there is that Google uses ffmpeg and x264

[12:48:00.0000] <Philip`>
hsivonen: Indeed - once you've got everything condensed into a script, then it's easy because you just feed it video files and get new video files popping out a (long) while later and everything is happy :-)

[12:48:01.0000] <hsivonen>
doublec: my conjecture is that the work around GPLv2 and patents by never distributing the software and by paying patent license fees

[12:49:00.0000] <Dashiva>
That rudd-O person is starting to annoy me. Especially when he CCs whatwg in a mail to... whatwg

[12:51:00.0000] <gsnedders>
goddamnit.

[12:51:01.0000] <gsnedders>
I just did that in a Reply All to him.

[12:51:02.0000] <gsnedders>
For an I-D/RFC, should a reference that needs to be read to understand an informative section be normative or not?

[12:52:00.0000] <annevk>
i'd suggest to reply less to obvious flame wars

[12:52:01.0000] <annevk>
give it a week or so

[12:53:00.0000] <Hixie>
a pearl of wisdom from anne there

[12:53:01.0000] <gsnedders>
Yeah, I probably should.

[12:54:00.0000] <Dashiva>
Well, it's good to have a few sensible replies to point to later when people start talking about being ignored :)

[12:54:01.0000] <gsnedders>
/me guesses the school spec only says you SHOULD (under the [RFC2119] meaning) do homework

[12:55:00.0000] <gsnedders>
i.e., if you have a good reason it's all right.

[12:57:00.0000] <Philip`>
Is doing homework a requirement for interoperability? If not, it shouldn't use a normative keyword at all

[12:57:01.0000] <Dashiva>
Depends. Some tasks are MUST requirements to pass the course

[12:57:02.0000] <Dashiva>
Philip`: Interoperability in grades

[12:58:00.0000] <gsnedders>
Philip`: for your knowledge to be interoperable with the final exam, yes.

[13:00:00.0000] <Philip`>
/me tries to remember how to write Java

[13:02:00.0000] <hsivonen>
Philip`: with Eclipse, it's all autocomplete--no remembering :-)

[13:02:01.0000] <gsnedders>
/me presses tab

[13:02:02.0000] <gsnedders>
oh, nice, a full HTML parser.

[13:04:00.0000] <hsivonen>
gsnedders: you need the Maven plug-in for that :-)

[13:04:01.0000] <Philip`>
The autocomplete in Visual Assist (for Visual C++) is the nicest I've ever used (though admittedly I haven't used many others, except Eclipse) - it occasionally autocompleted entire lines of code for me

[13:04:02.0000] <mpt>
annevk, or we could reply 2.5 years from now when none of the original posters are subscribed any more :-)

[13:04:03.0000] <hsivonen>
(I should remember to advertize the Mavenization of the Validator.nu parser)

[13:07:00.0000] <gsnedders>
/me sighs

[13:07:01.0000] <gsnedders>
Hixie: your website is blocked at my school

[13:08:00.0000] <gsnedders>
literally everything is.

[13:08:01.0000] <Hixie>
heh

[13:08:02.0000] <gsnedders>
Sometimes I wonder why they don't just use a whitelist.

[13:08:03.0000] <gsnedders>
It'd have the same affect more or less.

[13:08:04.0000] <gsnedders>
it is proof that Hixie is more famous than myself, annevk, or Eric Meyer.

[13:10:00.0000] <Dashiva>
I think it's because of his cats

[13:10:01.0000] <Dashiva>
They probably got him on a porn filter

[13:11:00.0000] <gsnedders>
That would explain why Molly is blocked too…

[13:14:00.0000] <MikeSmith>
/me wonders how come gsnedders don't post about HTML5 on his blog more

[13:14:01.0000] <gsnedders>
MikeSmith: because… my lust life is more interesting?

[13:15:00.0000] <gsnedders>
Or, more seriously, I don't have the time.

[13:15:01.0000] <gsnedders>
three posts in Oct and Nov, one in Sep…

[13:16:00.0000] <_Ivo>
An actual good comment from reddit: "So, let me get this: one member of the group wants a change, and it happens? I thought it was one member, one vote. The size or capitalization or your corporation shouldn't matter."

[13:16:01.0000] <gsnedders>
it's never been one member/one vote in the W3C.

[13:16:02.0000] <_Ivo>
It's supposedly two members, though, but the point remains.

[13:16:03.0000] <Dashiva>
one member, one vote is IEEE, isn't it?

[13:17:00.0000] <gsnedders>
Dashiva: IETF certainly, dunno about IEEE

[13:17:01.0000] <_Ivo>
So, how does the WHATWG achieves consensus?

[13:17:02.0000] <gsnedders>
_Ivo: the editor decides based on grounds of the strengths of the argument

[13:17:03.0000] <Hixie>
_Ivo: with more than 800 participants, we can never achieve true consensus. basically, i try to balance everyone's arguments and take the best approach each time.

[13:18:00.0000] <Hixie>
sometimes people disagree, and then we change the approach and try again.

[13:20:00.0000] <_Ivo>
Yes, but your entire argument seems to be based on the lack of consensus

[13:22:00.0000] <gsnedders>
_Ivo: remember the same spec is the W3C document, and that there you can formally object. we simply wouldn't have been able to publish the document in that state.

[13:23:00.0000] <_Ivo>
That is one good reason.

[13:24:00.0000] <gsnedders>
Vorbis has also been tested more in the real world (GTA:SA, UE2.5–3.0) than Theora (which has never had major companies shipping it)

[13:25:00.0000] <Dashiva>
ROSE Online used both ogg and mp3, sometimes it had both versions of the same tune even

[13:26:00.0000] <kingryan>
/me is amused by the amount of attention this issue has gotten on the web

[13:26:01.0000] <_Ivo>
gsnedders: That is certainly true.  Although, Linux companies like Red Hat and Cannonical do shipt Theora.

[13:26:02.0000] <gsnedders>
_Ivo: why was MS sued over MP3 not any other company? You wait till the biggest company you can get ships it.

[13:27:00.0000] <_Ivo>
gsnedders: of note is that while MS was sued, the case was settled and MS has paid nothing to Alacatel so far, and probably never will.

[13:28:00.0000] <gsnedders>
I'm well aware.

[13:28:01.0000] <_Ivo>
if Microsoft had paid Alcatel, they would go after Apple and their iPod next

[13:28:02.0000] <Dashiva>
The point is that they waited

[13:29:00.0000] <Hixie>
sweet!

[13:29:01.0000] <Hixie>
> Really if *anyone* should have any sway here (and I personally think

[13:29:02.0000] <Hixie>
> that no 1 or 2 companies should) it should be Google lets face it they

[13:29:03.0000] <Hixie>
> are the largest power on the Internet whether you love em/hate em/dont

[13:29:04.0000] <Hixie>
> know who they are..

[13:29:05.0000] <Hixie>
/me informs the writer of his employment status

[13:29:06.0000] <gsnedders>
/me bows down before Hixie and other googlities

[13:29:07.0000] <Dashiva>
uh-oh

[13:30:00.0000] <_Ivo>
Well, Google will likely attent the W3C video workshop.  Supposedly, an official statement will be made.

[13:30:01.0000] <Dashiva>
That could backfire, suddenly Google is running errands for the other companies

[13:30:02.0000] <_Ivo>
attend*

[13:32:00.0000] <_Ivo>
Is there any comment in Slashdot requesting people to post on the WHATWG list?  Because new comments just keep coming.

[13:32:01.0000] <Hixie>
yeah

[13:32:02.0000] <Hixie>
there is

[13:32:03.0000] <_Ivo>
oh great

[13:33:00.0000] <gsnedders>
I'll try and be in #video while the discussion regarding HTML 5 at the Video workshop is on tomorrow

[13:33:01.0000] <kingryan>
most of the comments are along the lines of "this is wrong. you're doing a bad thing. don't do it"

[13:33:02.0000] <kingryan>
which no suggestions of a better way

[13:33:03.0000] <_Ivo>
Yes, which is why they're not helping.

[13:34:00.0000] <kingryan>
and why I've stopped responding to them

[13:35:00.0000] <kingryan>
people seem to have the idea that if Ogg/* is in the spec, it will be implemented

[13:35:01.0000] <gsnedders>
kingryan: but they will! the spec says so! it must be true!

[13:35:02.0000] <Dashiva>
And that if it's not in the spec, nobody will ever touch it

[13:35:03.0000] <kingryan>
with no clear reason on how  you get from point A (spec) to point B (impl) to point C (deployment)

[13:36:00.0000] <kingryan>
Dashiva: right

[13:36:01.0000] <kingryan>
people don't realize that companies can still implement it

[13:36:02.0000] <kingryan>
and, in fact, the spec probably won't make much difference in terms of what containers/codec get implemented

[13:36:03.0000] <kingryan>
its like the tail wagging the dog

[13:37:00.0000] <_Ivo>
kingryan: Well, it will certainly have a bigger chance of being implemented if it's there than if it's not.  Not that big, but still notable.

[13:37:01.0000] <Dashiva>
I see Hixie brought out the "sadly" again :)

[13:38:00.0000] <Hixie>
hm?

[13:38:01.0000] <kingryan>
_Ivo: I don't think that's true

[13:38:02.0000] <Hixie>
_Ivo: not really

[13:38:03.0000] <Philip`>
I wonder if it's fortunate that I've got a three hour delay on whatwg emails, so there's no danger of me even thinking about replying to someone

[13:38:04.0000] <Hixie>
heh

[13:41:00.0000] <gsnedders>
do we define any specific behaviour for invalid byte sequences for a given character set?

[13:42:00.0000] <Dashiva>
Hixie: You used "sadly" 6 times in your latest mail.

[13:42:01.0000] <gsnedders>
how sad.

[13:42:02.0000] <Dashiva>
Reminded me of this: [13:24:38] <mpt> HTML5. Brought to you by the letter W, the number 5, and the word "Sadly".

[13:44:00.0000] <_Ivo>
Ah, it's that Amador guy who's asking people to e-mail WHATWG

[13:44:01.0000] <Hixie>
Dashiva: hah

[13:44:02.0000] <Hixie>
Dashiva: well it is a sad situation.

[13:44:03.0000] <Hixie>
:-)

[13:46:00.0000] <gsnedders>
/me cries

[13:47:00.0000] <kingryan>
/me laughs at http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-December/013210.html

[13:47:01.0000] <gsnedders>
kingryan: I mean, our feed parsers should totally send that!

[13:48:00.0000] <kingryan>
you mean "User-Agent: Please!" ?

[13:48:01.0000] <gsnedders>
no, the Accept header

[13:48:02.0000] <gsnedders>
IIS would send 400 Bad Request with an exclamation mark, I think

[13:49:00.0000] <gsnedders>
actually, ! is fine.

[13:50:00.0000] <Philip`>
annevk: Out of 32 sites, there's one <link media=screen> and one <style media=screen>

[13:50:01.0000] <Philip`>
/me tries to scale upwards a bit

[13:53:00.0000] <gsnedders>
/me laughs at "I really wish I had" in Hixie's latest email

[13:53:01.0000] <Philip`>
Hmm, I only just remembered that I've been sort of teaching people Java for the past two months, so I ought to actually know how to write it by now

[13:54:00.0000] <gsnedders>
/me concludes

[13:54:01.0000] <gsnedders>
I won't make any further comment on Ogg till after the workshop tomorrow is over.

[13:54:02.0000] <gsnedders>
(though I may spend a while writing an email addressing it)

[13:55:00.0000] <gsnedders>
/me runs off and falls alseep

[13:56:00.0000] <Lachy>
I really hate how the ogg discussion is spanning 3 separate mailing lists. I just wish it would stop!

[13:59:00.0000] <_Ivo>
I think igoring the lists for now would be a good idea.

[13:59:01.0000] <kingryan>
_Ivo: I agree

[13:59:02.0000] <_Ivo>
ignoring*

[13:59:03.0000] <kingryan>
I don't think there's anything any of us can say right now that will make a difference

[14:00:00.0000] <Hixie>
3?

[14:00:01.0000] <Dashiva>
What's the third?

[14:01:00.0000] <Philip`>
help⊙wo?

[14:01:01.0000] <Hixie>
cos help@ has received as much as www-archive@ on the subject, so far

[14:02:00.0000] <Dashiva>
whatwg@, help@, and...?

[14:03:00.0000] <Hixie>
oh actually help@ has had two, my bad

[14:03:01.0000] <Hixie>
public-html, presumably, is the second

[14:07:00.0000] <Dashiva>
That one mail about US-centricity? :)

[14:07:01.0000] <Hixie>
that was amusing

[14:07:02.0000] <Hixie>
given that more than half of the major browser vendors are based in the US...

[14:08:00.0000] <Philip`>
HTTP has footers? How peculiar

[14:08:01.0000] <Hixie>
ok i have dealt with the majority of feedback on the ogg thing

[14:09:00.0000] <Hixie>
now i will shower and then possibly go to work

[14:09:01.0000] <Dashiva>
Philip`: They come after the content-body, presumable

[14:09:02.0000] <Dashiva>
*y

[14:10:00.0000] <hubick>
If I ask a question about the Theora mailing list discussion here, will people get mad? :)

[14:10:01.0000] <Dashiva>
Nah

[14:10:02.0000] <Philip`>
Depends on which sense of "mad" you mean :-)

[14:11:00.0000] <hubick>
I don't want to get in the way of any productive work happening here

[14:11:01.0000] <Dashiva>
Don't worry about that, Hixie's in the shower

[14:11:02.0000] <hubick>
the implication reading Hixie's latest email is that Theora is a possible submarine patent risk

[14:11:03.0000] <Dashiva>
Indeed

[14:12:00.0000] <hubick>
there are no other current options that I see... which raises the question, IF someone were to go create such an option

[14:12:01.0000] <hubick>
how would it not pose the same risk as Theora?

[14:12:02.0000] <Dashiva>
By being old enough

[14:12:03.0000] <hubick>
is there any actual hard evidance towards such a risk surrounding THeora, that some other format wouldn't have?

[14:13:00.0000] <hubick>
Is there an older format that fits the bill?

[14:13:01.0000] <doublec>
some have mentioned h.261 as being a possibility

[14:13:02.0000] <Dashiva>
There's nothing specific about Theora, any format not implemented by the giant cash cows is a risk

[14:13:03.0000] <doublec>
I haven't seen any actual confirmation that that is the case though

[14:14:00.0000] <kingryan>
doublec: confirmation of what? that h.261 is old enough?

[14:14:01.0000] <doublec>
that there are no currently valid patents covering it

[14:14:02.0000] <doublec>
Nokia's paper says something like 'might have expired'

[14:15:00.0000] <Philip`>
hubick: MPEG4 has similar submarine patent risks to Theora; but Microsoft/Apple/etc already support MPEG4, so they don't take on any extra risk by supporting it for <video>, whereas adding support for Theora (or any other format that isn't ancient and that they don't already support) would be an additional risk

[14:15:01.0000] <kingryan>
patents only last 20 years, so if h.261 is older than that, all patents must have expired or be invalid

[14:15:02.0000] <kingryan>
that doesn't mean that someone can't still try and sue you for it

[14:15:03.0000] <hubick>
Philip`: you can argue that you shouldn't implement ANY new tech because it could have submarine patents though... that is no way to operate

[14:16:00.0000] <Dashiva>
No, it's a matter of risk vs reward

[14:16:01.0000] <Philip`>
Video seems particularly bad for this kind of thing

[14:16:02.0000] <Dashiva>
Have you read Hixie's mail to the list, hubick?

[14:16:03.0000] <hubick>
Dashiva: yes

[14:16:04.0000] <doublec>
so h.261 is from 1990?

[14:17:00.0000] <jgraham_>
/me seems to have got his internet connection restored just in time for the great video debate

[14:18:00.0000] <roc>
does Microsoft actually ship MPEG4 in Windows?

[14:18:01.0000] <roc>
that's something I've never gotten a straight answer on

[14:19:00.0000] <Philip`>
jgraham_: If you've only just joined, you've missed a whole day of it already :-)

[14:19:01.0000] <hubick>
To be honest, I have never heard of h.261 until know... is it really a viable format?  My hunch is that it would be functionally "worse" than Theora, and thus tilting the "risk vs reward" factor in favor of Theora - mainly because, having believed the party line, I was led to believe Theora wasn't at significant risk.

[14:20:00.0000] <Philip`>
/me wonders where Windows keeps its list of installed codecs

[14:20:01.0000] <Dashiva>
It's hard to run an organization based on belief, unless it's a religion :)

[14:21:00.0000] <othermaciej>
wow, that's a lot of ogg email

[14:22:00.0000] <Dashiva>
And it's all your (and Nokia's) fault

[14:22:01.0000] <Dashiva>
/me hides

[14:22:02.0000] <roc>
Just because a standard is more than 20 years old doesn't eliminate your patent exposure.

[14:22:03.0000] <roc>
Your *implementation* also has to be more than 20 years old

[14:22:04.0000] <othermaciej>
Dashiva: I specifically said I don't care if the current statement is removed from the spec

[14:23:00.0000] <Dashiva>
othermaciej: But that's because a should can be ignored, isn't it?

[14:24:00.0000] <hubick>
/me would like to see a SHOULD for *some* (any) Free/Open format

[14:24:01.0000] <othermaciej>
roc: patents that are fundamental to a standard are a bigger risk than patents on specific implementation techniques

[14:24:02.0000] <othermaciej>
roc: since by definition you can't code around them

[14:24:03.0000] <roc>
sure,

[14:24:04.0000] <othermaciej>
roc: and since it is easier for patent trolls to tell who is infringing (anyone implementing the spec)

[14:25:00.0000] <Dashiva>
hubick: We don't want a should, that means it can be ignored

[14:25:01.0000] <roc>
patent trolls don't particularly care about the accuracy of their threats as far as I can tell

[14:25:02.0000] <hubick>
Dashiva: I mean, at the very least, something would be better than nothing

[14:26:00.0000] <Dashiva>
hubick: But the last resort comes many years from now, when the spec is being finished

[14:26:01.0000] <othermaciej>
There's a reasonable argument that H.264 has less submarine risk than Theora

[14:26:02.0000] <Dashiva>
Until then, we should strive for something better

[14:27:00.0000] <hubick>
Dashiva: then you put out an "acid5" test with some OGG in it (transformed via SVG, woo), and get popular grass roots push for vendors to implement that, like previous acid tests

[14:27:01.0000] <othermaciej>
specifically, it was developed as an open standard through a process with IP disclosure requirements, and including most of the key players with large codec patent portfolios

[14:27:02.0000] <othermaciej>
this means judges will likely invalidate any submarine patents held by those parties

[14:27:03.0000] <Dashiva>
hubick: And look how well IE7 passes that, yeah

[14:27:04.0000] <othermaciej>
but Theora has never gone through such a process

[14:28:00.0000] <othermaciej>
the same holds for other codecs developed as open standards under IP disclosure (H.261, H.263, MPEG-2, etc)

[14:28:01.0000] <hubick>
Dashiva: yeah, and it's a rally point for web devs to complain to MS about - they are under pressure for not supporting it

[14:28:02.0000] <Dashiva>
pressure doesn't matter if it doesn't lead to results

[14:28:03.0000] <roc>
sometimes pressure leads to results

[14:28:04.0000] <hubick>
results may yet come

[14:29:00.0000] <othermaciej>
I don't really know how Ogg Theora compares technically to H.261 or H.263 or other older open standard codecs

[14:30:00.0000] <roc>
I seem to recall reading that the Xiph people compared Theora to the MPEG patent pool

[14:30:01.0000] <othermaciej>
I think Vorbis is the codec where they did an extensive patent search

[14:32:00.0000] <othermaciej>
or at least, I know they did that for Vorbis, I have not heard similar claims from them about Theora

[14:35:00.0000] <hubick>
I just think that html5 has potential to be the chicken that solves the chicken vs egg problem with uptake of a free format, by pointing to *something*.  My worry is, will there be some free format pointed to?

[14:35:01.0000] <hubick>
Or maybe I'm wrong, and that will just make it into another XHTML

[14:36:00.0000] <othermaciej>
anyway I hope someone knows how to do a meaningful test to compare quality of Theora to other codecs where patents are expired (H.261) or effectively not enforced (H.263)

[14:36:01.0000] <Dashiva>
hubick: There's one thing that everyone involved must realize, specs don't force implementation

[14:36:02.0000] <othermaciej>
I personally have no idea how to do that kind of "shootout" test

[14:38:00.0000] <anne-mac>
Philip`, yeah, need a bit larger set than 32 :)

[14:39:00.0000] <hubick>
Dashiva: I would like to put video on the web on sites I run.  I will only do it in a universal and free format.  Having *some* kind of spec would focus/corrall others like me to authoring pages using that same common format, hopefully with the end result of spurring users to install the codec if they see it used in more and more places.

[14:40:00.0000] <Dashiva>
Then you should go start a movement to create theora content for the web

[14:41:00.0000] <Philip`>
anne-mac: Are there other attributes that are interesting to look at? (I might as well check as many things as possible at once)

[14:41:01.0000] <Dashiva>
Someone has to be the mutated proto-chicken that lays the egg

[14:41:02.0000] <hubick>
Dashiva: A movement like that is best spurred by some group that already has the ear of some number of developers: ie, whatwg :)

[14:42:00.0000] <Dashiva>
No, that would be circular reasoning

[14:43:00.0000] <hubick>
Specs don't force implementation, but there are a lot of web developers that will see a 'SHOULD' and do exactly that as a result.

[14:44:00.0000] <Hixie>
that's not been my experience

[14:44:01.0000] <Dashiva>
Developers are more famous for creating dirty hacks that make it work no matter what the consequences :)

[14:45:00.0000] <hubick>
Just because it's not the most common doesn't mean there aren't a lot of developers that try to do the right thing (ie, validate pages as XHTML)

[14:45:01.0000] <mpt>
http://diveintomark.org/archives/2004/08/16/specs

[14:45:02.0000] <anne-mac>
Philip`, sorry, I'm only interested in media=

[14:46:00.0000] <Dashiva>
hubick: Validate as xhtml and send the wrong content-type, yeah

[14:46:01.0000] <hubick>
Dashiva: yeah!

[14:46:02.0000] <Hixie>
hubick: i've studied literally billions of html pages as part of scans i've done here at google and my experience is that very, very few authors actually try to validate pages as XHTML

[14:47:00.0000] <Hixie>
despite everyone's claims, in fact, few authors actually try to use XHTML at all -- only 15% of the billions of pages I looked at even had the XHTML namespace

[14:47:01.0000] <hubick>
it seems I have outed myself as an attempted web purist

[14:47:02.0000] <Hixie>
and that (due to copy/paste) is many more pages than those actually attempting to use xhtml for real

[14:47:03.0000] <hubick>
15% of a billion isn't a lot?

[14:47:04.0000] <hubick>
you need to start somewhere

[14:47:05.0000] <Hixie>
xhtml has been around since 1999

[14:48:00.0000] <Hixie>
15% in 8 years is nothing

[14:48:01.0000] <Philip`>
(Argh, I want a particular method that was added in HttpClient 3.1, but Gentoo and Fedora only have 3.0)

[14:48:02.0000] <Philip`>
/me tries to work out how to use JARs again

[14:48:03.0000] <Dashiva>
You just use them

[14:48:04.0000] <Dashiva>
classpath

[14:49:00.0000] <Philip`>
(In Eclipse)

[14:50:00.0000] <Dashiva>
Right-click, "add to build path"

[14:50:01.0000] <Philip`>
(except I'll need to move this to another machine to run it faster)

[14:50:02.0000] <hubick>
Philip`: Project Menu->Properties Item->Java Build Path Item->Libraries tab

[14:51:00.0000] <Philip`>
Aha, seems to work now

[14:51:01.0000] <Philip`>
and it no longer complains about me downloading potentially infinitely large files

[14:53:00.0000] <Hixie>
it's hard to know how to react to people who say that apple and nokia are lying

[14:53:01.0000] <Hixie>
especially since google has the same concern and i know that google isn't lying...

[14:53:02.0000] <hubick>
Hixie: I think you should make more noise about free alternatives like h.261

[14:54:00.0000] <Hixie>
h.261 sucks

[14:54:01.0000] <jgraham_>
Hixie: Well of course you ould say Google aren't lying ;)

[14:54:02.0000] <jgraham_>
s/ould/would/

[14:54:03.0000] <Hixie>
jgraham_: well, right, that's why i haven't mentioned that

[14:54:04.0000] <Hixie>
jgraham_: people would just say that i was part of the conspiracy

[14:55:00.0000] <Hixie>
the problem is that even if you don't believe the submarine patent risk, you still have the problem that apple won't implement ogg

[14:55:01.0000] <Hixie>
so whether we are being played or not doesn't really matter, if what we want is interop

[14:55:02.0000] <Hixie>
sihg

[14:55:03.0000] <Hixie>
i'll reply when i get to work

[14:55:04.0000] <Hixie>
afk for now

[14:56:00.0000] <jgraham_>
Indeed. But I think it's not hard to understand the conspiracy theories. It's so common for companies, especially large companies, to lie

[14:57:00.0000] <Dashiva>
Even if they didn't lie, they would still be doing the same actions, though

[14:57:01.0000] <jgraham_>
so I think you'll just have to ignore the conspiracy theories and hope that people can find a technical solution

[14:57:02.0000] <hubick>
I have taken a survey of the room here at work, and Linux has more market share than Apple, and they will ship ogg, so you are set :)

[14:59:00.0000] <jgraham_>
Dashiva: I'm not sying they _are_ lying. I'm saying that they're suffering from "boy who cried wolf" syndrome (does that analogy translate?)

[14:59:01.0000] <hubick>
I'm not supposed to believe Apple is doing this just to protect their interest in pushing the whole world to using only Quick Time?

[15:01:00.0000] <othermaciej>
do you mean QuickTime the media container format, or QuickTime the media framework software?

[15:01:01.0000] <othermaciej>
the former is clearly false, Apple primarily advocates the MPEG-4 container and family of codecs

[15:01:02.0000] <hubick>
I mean the Whatever That Stuff To Do With Watching Video I Need To Get From Apple

[15:02:00.0000] <jgraham_>
hubick: Ultimately the question of why they are doing it can only be answered by seeing if they act in good faith to find a solution.

[15:02:01.0000] <othermaciej>
I don't think QuickTime downloads for Windows are a huge revenue driver for Apple, but I'm not privy to financial info on that

[15:02:02.0000] <hubick>
ask them to grant all patents on their format and donate it to the web then

[15:02:03.0000] <Dashiva>
jgraham: I know. I'm just saying, regardless of whether they lie or not, we have to consider their actions rather than their words

[15:03:00.0000] <hubick>
What about "Dirac" ?

[15:03:01.0000] <jgraham_>
hubick: AIUI Apple + Nokia don't actually hold all the necessary patents, just license them

[15:04:00.0000] <jgraham_>
hubick: I guess Dirac is a solution if it is agreed to be free of IPR issues

[15:05:00.0000] <hubick>
I'm guessing the same "possibility of submarine patents" argument will be made against Dirac.

[15:05:01.0000] <jgraham_>
Dashiva: I agree entirely. At this juncture the actions we need to consider are their efforts to find a acceptable codec

[15:05:02.0000] <othermaciej>
Apple's actions so far have involved negotiating with codec licensing groups, asking the w3c to do patent searches (it's risky for large corporations to do a patent search), and providing a long list of possible codecs for the w3c to consider (w/ a brief summary of the tradeoffs)

[15:06:00.0000] <hubick>
is this list public?

[15:06:01.0000] <othermaciej>
sure, it's included in a summary of codec issues that Dave Singer sent to public-html a few months ago

[15:07:00.0000] <anne-mac>
public-html: http://lists.w3.org/Archives/Public/public-html/

[15:08:00.0000] <hubick>
Is public-html why the www-html I am subscribed to seems so dead?

[15:09:00.0000] <mpt>
http://lists.w3.org/Archives/Public/public-html/2007Nov/0153.html

[15:09:01.0000] <othermaciej>
I don't think Dirac is materially different from Theora, from an IP risk point of view

[15:10:00.0000] <othermaciej>
the thing that makes a big difference for MPEG is broad-based open process with disclosure requirements

[15:10:01.0000] <othermaciej>
that mitigates a lot (but not all) of the risk

[15:10:02.0000] <othermaciej>
something that *would* help a lot is a codec developed through an RF-license open standards process

[15:11:00.0000] <othermaciej>
(that could include taking Dirac or Theora through such a process)

[15:11:01.0000] <anne-mac>
hubick, this could be true, yes

[15:12:00.0000] <anne-mac>
joining the HTML WG (and public-html): http://blog.whatwg.org/w3c-restarts-html-effort

[15:18:00.0000] <hubick>
I dunno if I should subscribe... all these issues just make me frustrated.  I don't know how you people involved with creating these standards keep up morale for so long while creating standards that may never see the light of implementation day.

[15:19:00.0000] <anne-mac>
the company I work for implements stuff, so that helps :)

[15:19:01.0000] <parcelbrat>
can someone point me to the logs for any discussion on the removal of the ogg codec's from the spec?

[15:19:02.0000] <hubick>
parcelbrat: http://lists.w3.org/Archives/Public/public-html/2007Nov/0153.html  (courtesy of mpt to me a moment ago)

[15:20:00.0000] <jgraham_>
/me finds the web standards stuff has more of a "people care abut this" vibe than Astrophysics

[15:20:01.0000] <parcelbrat>
hubick: thanks

[15:20:02.0000] <parcelbrat>
has it been a common topic?

[15:21:00.0000] <hubick>
parcelbrat: I been bugging them in here for almost an hour now :)

[15:21:01.0000] <anne-mac>
it has been discussed when <video> got introduced and during the technical planery a month ago

[15:21:02.0000] <anne-mac>
and now

[15:22:00.0000] <anne-mac>
same arguments each time iirc

[15:23:00.0000] <parcelbrat>
anne-mac: am i reading correctly that whatwg is still wanting support for the ogg formats? and it will be discussed more, or have they been tossed?

[15:24:00.0000] <Philip`>
jgraham_: There's as many HTML documents as there are stars in our galaxy (to within a couple of orders of magnitude), and the web is growing exponentially faster than the galaxy, so it's obviously the more interesting area :-)

[15:24:01.0000] <hubick>
parcelbrat: sounds like they are looking for alternatives, and a likely candidate may be h.261

[15:24:02.0000] <anne-mac>
parcelbrat, it will most certainly be discussed more

[15:25:00.0000] <anne-mac>
and W3C is looking into a patent search as I understand things

[15:25:01.0000] <anne-mac>
for Ogg stuff

[15:25:02.0000] <jgraham_>
parcelbrat: The WHATWG wants support for some freely-implementable formats

[15:28:00.0000] <parcelbrat>
so slashdot is over-reacting?

[15:28:01.0000] <parcelbrat>
http://yro.slashdot.org/yro/07/12/11/1339251.shtml

[15:28:02.0000] <jgraham_>
/me wonders how Philip` arrived at ~100 billion HTML documents

[15:28:03.0000] <gavin>
slashdot? overreacting? impossible!!

[15:28:04.0000] <jgraham_>
parcelbrat: If they're not then it would be the first time for this particular thing

[15:28:05.0000] <parcelbrat>
true

[15:29:00.0000] <hubick>
the bottom line is that that it *was* there and now it's not though.

[15:29:01.0000] <parcelbrat>
i knew there was a reason i avoid slashdot usually, ars is usually less dramatic

[15:29:02.0000] <parcelbrat>
hubick: was and not, but isn't banned, right?

[15:30:00.0000] <parcelbrat>
/me rhetorical

[15:30:01.0000] <hubick>
parcelbrat: I find Ars editors have just as biased takes on topics, if less obvious

[15:30:02.0000] <parcelbrat>
biased yes, just less dramatic

[15:30:03.0000] <Philip`>
jgraham_: Searching for e.g. "a" on Yahoo gives reportedly 20 billion matches, and Yahoo probably misses lots of pages so round it up to 10^11

[15:30:04.0000] <anne-mac>
 /. is written by contributors

[15:31:00.0000] <anne-mac>
this post happens to be written by a guy promoting his own post and providing all the incorrect links

[15:31:01.0000] <anne-mac>
who also posts to the WHATWG list a lot

[15:31:02.0000] <parcelbrat>
is he dramatic on the list too?

[15:31:03.0000] <hubick>
parcelbrat: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-December/013152.html

[15:35:00.0000] <parcelbrat>
<sarcasm>wow very weasely</sarcasm>

[15:35:01.0000] <jgraham_>
Philip`: I guess 10^(11±2) search-engine cached HTML documents might not be such a bad estimate, although that's not quite "total number of HTML documents"

[15:35:02.0000] <parcelbrat>
hubick: that definitely clears stuff up for me, thanks

[15:35:03.0000] <jgraham_>
s/cached/indexed/ I guess

[15:36:00.0000] <Philip`>
(Hooray, my cached downloader actually works)

[15:36:01.0000] <hubick>
Philip`: what are you building?

[15:37:00.0000] <Philip`>
hubick: Something to download and analyse lots of web pages (for a small value of "lots", like tens of thousands)

[15:38:00.0000] <anne-mac>
Philip`, didn't you already have sniffing for lots of pages?

[15:38:01.0000] <hubick>
Philip`: If you need graphs and haven't tried it yet, I highly recommend jFreeChart

[15:39:00.0000] <anne-mac>
/me though Philip` had hsivonen's stuff up and running

[15:39:01.0000] <Philip`>
anne-mac: Yes, but I've rewritten stuff so it caches the pages and doesn't have to download hundreds of megabytes of HTML every time I run it

[15:39:02.0000] <hubick>
hsivonen: which reminds me... I hope you received the patch I sent you for the htmlparser Maven metadata

[15:40:00.0000] <Philip`>
anne-mac: http://canvex.lazyilluminati.com/misc/media.xml is from 256 pages - is that about what you want?

[15:41:00.0000] <anne-mac>
/me liked 10s of 1000s

[15:41:01.0000] <Philip`>
I'm going to move it onto a better computer to run on more pages :-)

[15:41:02.0000] <Philip`>
assuming it's going to be doing the right thing

[15:41:03.0000] <anne-mac>
Philip`, I guess that page only shows something on Firefox or something?

[15:42:00.0000] <anne-mac>
/me gets a blank in Opera

[15:42:01.0000] <Philip`>
anne-mac: View source :-p

[15:42:02.0000] <anne-mac>
and Safari

[15:42:03.0000] <anne-mac>
hmm

[15:42:04.0000] <Philip`>
It's "XML" - I expect you've heard of that before

[15:42:05.0000] <anne-mac>
what's this thing you call view source? :p

[15:43:00.0000] <mpt>
View Source is for hackers

[15:43:01.0000] <parcelbrat>
what is this XML you speak of?

[15:43:02.0000] <Philip`>
/me is just generating a big XML dump of lots of headers and attributes and stuff, then using xml_grep to extract the @media values

[15:44:00.0000] <anne-mac>
so far people seem to comply to the arbitrary media= standards

[15:44:01.0000] <anne-mac>
i'm amazed

[15:45:00.0000] <Philip`>
parcelbrat: It's kind of like HTML, except the brackets are anglier

[15:45:01.0000] <parcelbrat>
/me classic

[15:46:00.0000] <Philip`>
/me has absolutely no idea how many concurrent downloading/processing threads to run

[15:47:00.0000] <jgraham_>
/me is sure the brackets are actually angrier; just look how upset they get when they don't get a partner...

[15:47:01.0000] <hubick>
At Linux world 2000 the Konqueror guys were telling me how awesome their browser was, so I loaded my home page which uses @media CSS tags in it, at which point it promptly disappeared *poof* from the screen, leaving them quite embarassed and quiet :)

[15:48:00.0000] <mpt>
<xml> ≪xml2≫

[15:48:01.0000] <parcelbrat>
oooh, xml2 is even anglier than xml!

[15:49:00.0000] <hubick>
mpt: so, xml will eventually evolve into Lisp then?

[15:49:01.0000] <mpt>
⋘xml3⋙

[15:49:02.0000] <parcelbrat>
(xmlisp)

[15:49:03.0000] <hubick>
isn't this JSON stuff basically that? :)

[15:50:00.0000] <parcelbrat>
pretty darn close

[15:50:01.0000] <Philip`>
http://canvex.lazyilluminati.com/misc/sexp.html

[15:50:02.0000] <hubick>
/me *lolz*

[15:50:03.0000] <parcelbrat>
/me um.....

[15:51:00.0000] <hubick>
what was the old sgml transform language?

[15:51:01.0000] <anne-mac>
dsssl?

[15:51:02.0000] <hubick>
yeah

[15:51:03.0000] <hubick>
wasn't it like that?

[15:52:00.0000] <anne-mac>
no idea

[15:52:01.0000] <hubick>
heh, I just started clicking links on http://www.jclark.com/dsssl/ and got like, four 404's in a row :(

[15:55:00.0000] <anne-mac>
Philip`, nice


2007-12-12
[16:03:00.0000] <Philip`>
/me wonders how long it'll take to run "sort -R" on 4.5M lines

[16:04:00.0000] <parcelbrat>
nice

[16:04:01.0000] <anne-mac>
we could have text/html+lisp

[16:05:00.0000] <parcelbrat>
yeah, but then you'd have to start catering to everyone

[16:05:01.0000] <parcelbrat>
text/html+ruby

[16:05:02.0000] <parcelbrat>
text/html+python

[16:05:03.0000] <parcelbrat>
text/html+vbs...

[16:05:04.0000] <parcelbrat>
/me keyboard breaks

[16:05:05.0000] <anne-mac>
neh, only those we like

[16:06:00.0000] <parcelbrat>
shouldn't those be application/html+<<lang>>

[16:06:01.0000] <Philip`>
We could support them all at first, then remove them all from the spec, and see which ones get complained about the most, and then just put those ones back in

[16:06:02.0000] <anne-mac>
application/* is overrated I think

[16:07:00.0000] <parcelbrat>
philip`: works for media, and got me here

[16:07:01.0000] <parcelbrat>
anne-mac: newbie question: why?

[16:08:00.0000] <anne-mac>
parcelbrat, take text/xml versus application/xml, the only reason to prefer the latter is theoretical concerns over character encoding

[16:08:01.0000] <Philip`>
Hmm, 6 minutes of CPU time to randomise the list

[16:08:02.0000] <anne-mac>
but all implementations treat them as being equivalent

[16:09:00.0000] <parcelbrat>
how would handle text/xml with character encoding? let http handle it?

[16:09:01.0000] <anne-mac>
most text/* formats have their own specific rules for determining the character encoding actually, all against the various RFCs from the stone age

[16:10:00.0000] <anne-mac>
text/xml defaults to US-ASCII unless it has a charset parameter that says otherwise

[16:10:01.0000] <anne-mac>
application/xml defaults to whatever the XML file says unless it has a charset parameter specified

[16:10:02.0000] <anne-mac>
in practice, text/xml is like application/xml

[16:12:00.0000] <parcelbrat>
makes sense

[16:12:01.0000] <parcelbrat>
and we know how well application/html+xml works ;)

[16:12:02.0000] <anne-mac>
xhtml+xml* ;)

[16:12:03.0000] <parcelbrat>
/me smacl

[16:13:00.0000] <parcelbrat>
s/smacl/smack

[16:14:00.0000] <hubick>
which makes me wonder if Firefox is ever gonna support */*+xml: https://bugzilla.mozilla.org/show_bug.cgi?id=155730

[16:17:00.0000] <Philip`>
Seems I can download/process 4K pages per minute

[16:18:00.0000] <parcelbrat>
hubick: yeah, lately, we've had a problem with Ruby on Rails error messages because <%= isn't a valid xml tag... the page isn't supposed to be coming across as xml... oh well

[16:18:01.0000] <Philip`>
I get loads of cookie spec violation warnings :-/

[16:22:00.0000] <Philip`>
Argh, and I've got ill-formed XML output too

[16:23:00.0000] <parcelbrat>
later ya'll

[16:23:01.0000] <Philip`>
<header uri="http://www.ganymede.cz/" name="Server" value="Apache/2.2.6 (Unix) mod_ssl/2.2.6 &#2; DAV/2 PHP/5.2.5"/>

[16:23:02.0000] <Philip`>
/me wishes XML was easy

[16:25:00.0000] <Philip`>
Hmm, there's five sites with &#2; in their Server

[16:26:00.0000] <Philip`>
<header uri="http://www.doxamus.ro/" name="Server" value="Apache/2.2.6 (Unix) mod_ssl/2.2.6 �N&#31;&#9;YNED�����N&#31;&#9;SAHP�����N&#31;&#9;TATS�����N&#31;&#9; mod_bwlimited/1.4 mod_auth_passthrough/2.1 FrontPage/5.0.2.2635 PHP/5.2.4"/>

[16:26:01.0000] <Philip`>
That's really not going to work

[16:35:00.0000] <Philip`>
anne-mac: http://www.cl.cam.ac.uk/~pjt47/misc/media.xml

[16:36:00.0000] <Philip`>
anne-mac: http://www.cl.cam.ac.uk/~pjt47/misc/media.txt too

[16:36:01.0000] <Philip`>
from 16 kilopages, minus about 500 with errors

[16:37:00.0000] <Philip`>
(This is only about 400MB of HTML, so I could do more fairly easily)

[16:37:01.0000] <Philip`>
(but probably not enough more to find really interesting things)

[16:42:00.0000] <Philip`>
s/about 500/940/

[16:42:01.0000] <Philip`>
(Also, binary kilo)

[16:52:00.0000] <Hixie>
"Re: [whatwg] HTML 5, OGG, competition, civil rights,  and persons with disabilities"

[16:52:01.0000] <Hixie>
/me fears looking at that e-mail

[16:52:02.0000] <bradee-oh>
lol

[16:53:00.0000] <Dashiva>
Hixie: Now you know how we feel about your mashup replies :P

[16:53:01.0000] <Hixie>
:-D

[16:56:00.0000] <Philip`>
hsivonen: I can process 16K pages in 20 seconds (wallclock time) - I think your parser is fast enough for me for now :-)

[16:57:00.0000] <_Ivo>
I can sadly say that that is a sad title.

[16:59:00.0000] <Philip`>
441% CPU usage? I think 'top' is lying to me...

[17:18:00.0000] <Hixie>
http://www.bluishcoder.co.nz/2007/12/video-element-and-ogg-theora.html is a good summary

[17:27:00.0000] <doublec_>
hotel network connections, sigh

[17:31:00.0000] <roc>
hey

[17:31:01.0000] <roc>
yeah, that was good Chris

[17:31:02.0000] <doublec_>
thanks :)

[17:31:03.0000] <doublec_>
I had to type it in using w3m over a ssh connection using bloggers interface.

[17:32:00.0000] <doublec_>
since the hotel network seems to kill any browser traffic over a certain size

[17:33:00.0000] <roc>
This is Avante?

[17:33:01.0000] <roc>
I don't remember that being a problem

[17:33:02.0000] <doublec_>
Yes, it's avante. And it is strange. I can receive fine. I can't even send via gmail.

[17:33:03.0000] <doublec_>
yet that's over ssl so I don' t understand it

[17:36:00.0000] <roc>
complain to the management, that's pretty important right now

[17:37:00.0000] <doublec_>
will do

[17:38:00.0000] <othermaciej>
mmmm, ogg flamage

[17:38:01.0000] <othermaciej>
toasty

[17:39:00.0000] <othermaciej>
Hixie: it argued that not recommending support for the Ogg Theora video codec will be harmful to the blind

[17:39:01.0000] <Hixie>
i've been trying to keep the flames to a minimum by asking for politeness off-list, i hope it helps

[17:39:02.0000] <Hixie>
wait, what?

[17:39:03.0000] <Hixie>
wow

[17:39:04.0000] <Hixie>
can't wait to read that

[17:41:00.0000] <Philip`>
http://www.cl.cam.ac.uk/~pjt47/misc/attributes.html - some numbers about values of an arbitrarily-chosen set of elements/attributes

[17:42:00.0000] <Hixie>
off hand those numbers match what i remember seeing

[17:42:01.0000] <Hixie>
except for a rev=PARENT

[17:42:02.0000] <Hixie>
i just saw rev=made and rev=stylesheet

[17:43:00.0000] <Philip`>
They all come from http://www.dcs.gla.ac.uk/~simon/quantum/

[17:43:01.0000] <Philip`>
(since I counted number of occurrences in total, not number of pages)

[17:44:00.0000] <Philip`>
Maybe number of pages would be more useful...

[17:44:01.0000] <Hixie>
ah yeah i found that number of occurances just never works

[17:44:02.0000] <Hixie>
there are too many gigantic pages that totally screw the count

[17:49:00.0000] <Philip`>
http://www.cl.cam.ac.uk/~pjt47/misc/attributes.html - now with number of pages too

[17:50:00.0000] <Philip`>
Not sure why I'm bothering to keep the number of occurrences too, but I guess it doesn't hurt

[17:50:01.0000] <Philip`>
Looks like cyan and magenta are the least favourite of the binary colours :-(

[17:51:00.0000] <othermaciej>
doublec_'s blog post is indeed a good summary

[17:51:01.0000] <Hixie>
/me pokes Philip` to look at his /msgs (and maybe to have him respond on w3.net, since he's not registered on freenode)

[17:56:00.0000] <othermaciej>
I posted on the codec thread

[17:57:00.0000] <othermaciej>
may knuth have mercy on my soul

[17:57:01.0000] <Philip`>
"The" codec thread? I thought there was about two dozen of them

[17:57:02.0000] <othermaciej>
one of them

[18:13:00.0000] <Philip`>
othermaciej: About "I've heard game vendors cited, not sure which ones": See http://wiki.xiph.org/index.php/Games_that_use_Vorbis

[18:14:00.0000] <othermaciej>
thanks, that doesn't list the vendors in an easy-to-find way

[18:14:01.0000] <Philip`>
http://www.unrealtechnology.com/features.php?ref=audio mentions Vorbis support quite prominently

[18:14:02.0000] <othermaciej>
main thing I wondered about was whether any Microsoft-published games use Ogg Vorbis

[18:15:00.0000] <othermaciej>
I think the audio issue is somewhat less important since (a) Vorbis has good quality and a somewhat more solid IP footing and (b) MP3 patents will expire in a few years, at which point MP3 is a suitable audio baseline

[18:15:01.0000] <Philip`>
I don't see any on the list that I recognise as Microsoft

[18:17:00.0000] <Philip`>
Oh

[18:17:01.0000] <Philip`>
Halo

[18:18:00.0000] <Philip`>
which was while they were owned by Microsoft

[18:19:00.0000] <Philip`>
s/they/Bungie/

[18:19:01.0000] <Hixie>
halo is certainly high profile

[18:20:00.0000] <Dashiva>
Fable too

[18:21:00.0000] <_Ivo>
and Gears of War

[18:22:00.0000] <Philip`>
_Ivo: Uh, I don't think that used Vorbis

[18:22:01.0000] <_Ivo>
as far as I know, it did

[18:22:02.0000] <_Ivo>
may be worth confirming

[18:23:00.0000] <Philip`>
Oh, looks like you're right

[18:24:00.0000] <Philip`>
e.g. http://utforums.epicgames.com/showthread.php?t=583980&page=10 says it has a vorbis.dll

[18:26:00.0000] <Hixie>
btw just so everyone is up to date, i'm thinking we should just drop the whole cross-references nonsense and replace it with some recommendations about using <a href="">

[18:26:01.0000] <Philip`>
Actually, maybe that's just left over from it being an Unreal Engine game - I have no idea if they really use the Ogg support

[18:26:02.0000] <othermaciej>
Hixie: the autolinking cross-references?

[18:26:03.0000] <Hixie>
yeah

[18:27:00.0000] <Hixie>
too much complexity for not much gain

[18:27:01.0000] <othermaciej>
Hixie: they did seem cute, but admittedly not that compelling over <a>

[18:28:00.0000] <Hixie>
good lord we got a lot of feedback on <cite>

[18:32:00.0000] <Hixie>
hm, i'm thinking, <cite> maybe should just be for a title of a work. any work, and even if it's not technically really cited.

[18:32:01.0000] <Hixie>
it seems that the typographic convention angle is more useful than the "this is a citation" angle

[18:32:02.0000] <Hixie>
and the line of what a citation is is a bit vague anyway

[18:35:00.0000] <roc>
Hixie: FWIW Apple and Nokia support software patents in general.

[18:36:00.0000] <othermaciej>
http://www.macobserver.com/article/2007/08/31.1.shtml

[18:37:00.0000] <othermaciej>
(that's not a counter-argument to supporting them in general, I don't know if Apple has an official position on that)

[18:43:00.0000] <roc>
http://www.macobserver.com/article/2007/08/02.12.shtml

[18:43:01.0000] <roc>
"However, Apple's chief patent counsel, Chip Lutton, contradicted Ms. Lee, and doesn't think the patent system is broken. In fact, "it's the best system in the world," he said. "

[18:48:00.0000] <othermaciej>
that certainly seems like support for the patent system in general

[18:49:00.0000] <roc>
it seems to indicate Apple is pretty happy with software patents in general

[18:49:01.0000] <roc>
because many other systems in the world don't have them

[18:50:00.0000] <othermaciej>
well, again, I'm not privy to Apple's official view on the matter, but I know that Apple has actively supported patent reform, and that this is likely to improve the situation at least somewhat

[18:51:00.0000] <othermaciej>
that probably makes a bigger difference than sound bites

[18:52:00.0000] <othermaciej>
I personally think software patents are broken and should either not exist or have much shorter terms than current terms, but that that is certainly not an official position

[18:53:00.0000] <roc>
I agree

[18:54:00.0000] <roc>
the "reforms" pushed by Apple and others are probably good things in themselves, but they're essentially self-interested attempts to tamp down the troll problem

[18:56:00.0000] <roc>
/me hopes that the US Supreme Court will just rule software patents invalid and all this will just blow away in the breeze

[18:57:00.0000] <othermaciej>
unfortunately that seems unlikely

[18:57:01.0000] <roc>
I would have thought so, but they've been most ornery about patents lately

[19:00:00.0000] <othermaciej>
honestly I'm not sure I get Apple's stance given which side of patent lawsuits we're usually on

[19:01:00.0000] <roc>
I know the feeling. I used to work for IBM

[19:06:00.0000] <MikeSmith>
Hixie - I won't shed any tears when the dfn cross-referencing thing gets dropped, but I think, gee wouldn't it be nice if we had a general xref mechanism?

[19:07:00.0000] <MikeSmith>
such that empty <xref href="#foo"> gets replaced with content of element at foo

[19:34:00.0000] <G0k>
uh

[19:34:01.0000] <G0k>
who the hell is this rudd-o clown?

[19:38:00.0000] <othermaciej>
G0k: he's clearly passionate about his beliefs

[19:38:01.0000] <G0k>
at this point i'm convinced he's an agent for MPEG LA

[19:38:02.0000] <G0k>
because he's doing more to discredit the Ogg crowd than anyone else I've seen

[19:43:00.0000] <aphid>
it's bulldada, by contrast he makes the rest of us look civilized and reasonable.

[19:43:01.0000] <aphid>
:D

[19:44:00.0000] <G0k>
uhg

[20:12:00.0000] <MikeSmith>
kfish - hei

[20:25:00.0000] <kfish>
yo MikeSmith 

[20:25:01.0000] <kfish>
cold in tokyo?

[20:44:00.0000] <MikeSmith>
kfish - yeah, too cold for me already

[20:44:01.0000] <MikeSmith>
both in Tokyo and out at Keio/SFC

[20:45:00.0000] <MikeSmith>
warmed up last night by eating shabu-shabu and drinking fugu hire-zake

[20:46:00.0000] <kfish>
nice :-)

[21:06:00.0000] <_Ivo>
This won't be nice of me asking, but is it possible to block Manuel Amador (Rudd-O) from the lists at least temporarily til he cools off?

[21:07:00.0000] <bradee-oh>
I see he promises an upcoming barrage of emails because of his handiwork at Digg.

[21:07:01.0000] <bradee-oh>
oh joy.

[21:08:00.0000] <bradee-oh>
sure has made it difficult to have actual discussions about actual standards issues today  *sigh*

[21:12:00.0000] <parcelbrat>
I'm one of the people who came to get clarification (earlier) due to his /. post. I'd actually like to stay involved for the real reason.

[21:12:01.0000] <parcelbrat>
That aside, I'm wondering if there has been a barage here from that too

[21:21:00.0000] <bradee-oh>
The quantity today has been *much* higher than usual, and much less productive.  yay!

[21:22:00.0000] <inimino>
/me wonders how many productive man-hours were lost

[21:24:00.0000] <Teratogen>
BRING BACK OGG!

[21:28:00.0000] <jruderman>
Teratogen: http://yro.slashdot.org/comments.pl?sid=385689&cid=21655557

[21:28:01.0000] <jruderman>
(sorry, i guess that was Oog)

[21:30:00.0000] <parcelbrat>
does Teratogen == rudd-o?

[21:34:00.0000] <Teratogen>
BRING BACK OGG NOW!

[21:36:00.0000] <_Ivo>
what the hell is a Oog?

[21:40:00.0000] <jruderman>
Oog, the open-source caveman, a legendary Slashdot troll

[21:44:00.0000] <aphid>
ogg is also the name of the stalinesque leader from some Nat'l Petroleum Institute produced cartoon that's on archive.org

[21:45:00.0000] <aphid>
http://www.archive.org/details/Destinat1956

[21:52:00.0000] <parcelbrat>
lol

[22:01:00.0000] <Hixie>
if anyone is on site5, feel free to point out on this thread that we didn't trade ogg for something proprietary: http://forums.site5.com/showthread.php?t=19941

[22:08:00.0000] <parcelbrat>
i noticed someone on site5 on #ror, but he already left

[22:43:00.0000] <Hixie>
I've added documentation to the annotation system

[23:37:00.0000] <hsivonen>
lots and lots of ogg email :-(

[23:43:00.0000] <doublec>
yes, Ogg is the favourite topic of the day

[23:44:00.0000] <Hixie>
more mail?

[23:45:00.0000] <doublec>
only a couple :)

[23:47:00.0000] <Hixie>
we've hit 855 members

[23:47:01.0000] <Hixie>
that's 50 more than this morning

[23:49:00.0000] <othermaciej>
Hixie: you need to make more controversial changes so we pass 1000

[23:51:00.0000] <Hixie>
seriously

[23:51:01.0000] <doublec>
and find a way to get money from each new member

[23:51:02.0000] <doublec>
since you aren't getting bribed to remove ogg :)

[23:52:00.0000] <doublec>
I think your reddit replies should get some sort of award

[23:53:00.0000] <othermaciej>
for "most reddit replies on a single thread"?

[23:53:01.0000] <roc_>
when Hixie does start taking bribes we're all going to look stupid

[00:48:00.0000] <Hixie>
hsivonen: i don't really see anything in http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2006-December/008849.html that i can respond to (other than the <q> issue), except maybe the suggestion in the parenthetical in point 1

[00:48:01.0000] <Hixie>
but i don't understand what that's suggesting

[00:48:02.0000] <Hixie>
can you advise?

[00:49:00.0000] <anne-mac>
that e-mail is not from hsivonen...

[00:49:01.0000] <hsivonen>
Hixie: looking at it now

[00:49:02.0000] <Hixie>
oh you're right, it's not

[00:49:03.0000] <Hixie>
oops

[00:50:00.0000] <Hixie>
well i don't know what to do with it then

[00:51:00.0000] <anne-mac>
i think it's more of a rant than a comment

[00:56:00.0000] <hsivonen>
Hixie: The main point I'd make is that non-heuristic machine consumption (i.e. naïvely trusting marked-up semantics and only using marked-up semantics) for dialogs, names of vessels, quotations, etc. does not have a plausible market-demand story. However, the elevator pitch for hCard and hCalendar is at least plausible: UI for adding event info or contact info to iCal, Address Book or similar app

[00:57:00.0000] <Hixie>
agreed

[00:57:01.0000] <Hixie>
doesn't really affect the spec though

[00:57:02.0000] <Hixie>
so...

[00:57:03.0000] <hsivonen>
the volume of the ogg thing has caused me to get out of sync with IRC and list email. :-(

[00:58:00.0000] <Hixie>
it wasn't that bad

[00:58:01.0000] <anne-mac>
oh, Hixie replies to a annevk⊙oc e-mail!

[00:58:02.0000] <Hixie>
:-)

[00:59:00.0000] <anne-mac>
reached the two year old e-mail mark? :)

[01:01:00.0000] <Hixie>
nah, just dealing with mail from buckets that have newer mail

[01:01:01.0000] <Hixie>
thought that particular mail was actually from the time my spam filter hated you

[01:01:02.0000] <Hixie>
though, even

[01:01:03.0000] <anne-mac>
could be

[01:02:00.0000] <Hixie>
no it definitely was

[01:02:01.0000] <Hixie>
i had to go fish it out of my gmail pile to reply to it

[01:02:02.0000] <anne-mac>
ok :)

[01:03:00.0000] <virtuelv>
Hixie: you're implying you never delete even spam?

[01:03:01.0000] <Hixie>
no, i saved a bunch of mail from anne back when my filter hated him

[01:03:02.0000] <Hixie>
but e-mails saved from gmail's spam folder don't get forwarded to my main imap server

[01:04:00.0000] <Hixie>
they just sit in my gmail inbox

[01:04:01.0000] <Hixie>
/me gets to an e-mail from Tina Holmboe

[01:04:02.0000] <Hixie>
/me decides to deal with that one later

[01:05:00.0000] <Hixie>
btw, what's the status with the role="" stuff? and what's the status with the forms stuff?

[01:05:01.0000] <anne-mac>
forms: nobody replied to my request for input

[01:06:00.0000] <anne-mac>
role: zcorpan knows it better than me at this point, being on the group and all

[01:06:01.0000] <anne-mac>
I think they like to call it role=, want to use namespaces in some places? and the rest is aria-xxx

[01:07:00.0000] <Hixie>
so i have a request here for an <attn> element, whose primary use case would be to do what waiaria does with update regions

[01:08:00.0000] <Hixie>
will i eventually be mentioning waiaria somewhere in html5?

[01:08:01.0000] <anne-mac>
i think so, yes

[01:08:02.0000] <zcorpan>
i think the idea is to define both role= and aria-xxx=, and the old namespaced attributes, even with the knowledge that browsers don't want to support the namespaced attributes

[01:08:03.0000] <zcorpan>
i really don't know why

[01:09:00.0000] <anne-mac>
Hixie, I guess HTML5 would defer to the aria draft for role= and attributes prefixed with aria-, would make it clear there are no DOM interfaces, and that's it

[01:10:00.0000] <anne-mac>
/me should probably go to work at some point

[01:10:01.0000] <Hixie>
no DOM interfaces? huh, that sucks. since it's primarily intended to be for scripts...

[01:10:02.0000] <Hixie>
zcorpan: can we fight for more sanity?

[01:11:00.0000] <Hixie>
zcorpan: i'm willing to help if needed. in particular, can we decouple from xhtml2's role stuff?

[01:11:01.0000] <zcorpan>
Hixie: not sure

[01:12:00.0000] <zcorpan>
but i think so

[01:14:00.0000] <zcorpan>
Hixie: the idea is that the attributes should be correct even in legacy browsers with no knowledge of aria, so that screen readers can pick it up from the dom

[01:14:01.0000] <zcorpan>
although i guess they could read js properties as well

[01:15:00.0000] <Hixie>
ah interesting

[01:18:00.0000] <zcorpan>
perhaps we can introduce a dom interface for aria 2, when we know what aria 2 will require

[01:26:00.0000] <zcorpan>
/me considers marking all ogg emails as read

[01:26:01.0000] <krijnh>
/me did

[01:27:00.0000] <krijnh>
Apart from the big Hixie replies

[01:31:00.0000] <hsivonen>
Philip`: re: 20 seconds: nice

[01:33:00.0000] <zcorpan>
hsivonen: btw, does your schema support role="fancy-checkbox checkbox"?

[01:34:00.0000] <zcorpan>
/me can see a disadvantage of allowing arbitrary roles; conformance checkers won't catch typos

[01:35:00.0000] <zcorpan>
role="chekcbox"

[01:35:01.0000] <jgraham_>
/me seems to be getting new ogg emails at roughly his time average rate of reading WHATWG emails so the unread number is almost constant

[01:35:02.0000] <hsivonen>
Hixie: I think <cite> should be defined to capture the Chicago Manual of Style title-of-work concept while at the same time saying that an author is not an evil person for using <i> instead and saying that you don't need to change legacy content that uses <cite> for names of people

[01:38:00.0000] <hsivonen>
hubick: I'm near email bankruptcy but I have received your patch. Sorry about the delay.

[01:45:00.0000] <hsivonen>
zcorpan: no, the proof-of-concept schema approach does not take address extensibility

[01:45:01.0000] <zcorpan>
hsivonen: ok

[01:45:02.0000] <hsivonen>
zcorpan: extensibility is not on the agenda for 1.0

[01:46:00.0000] <hsivonen>
zcorpan: moreover, extensibility and white-list-like validation are conflicting things

[01:46:01.0000] <zcorpan>
perhaps i should update the authoring conformance reqs accordingly

[01:56:00.0000] <Philip`>
Does the Ogg discussion indicate that HTML5 is actually relevant and people care about it, or would they complain as much about any other specification that did the same thing?

[01:56:01.0000] <Hixie>
hsivonen: well you never need to change legacy content

[01:56:02.0000] <Hixie>
hsivonen: it just might not be compliant to html5 :-)

[01:56:03.0000] <Hixie>
(i don't see any point explicit grandfathering in legacy content in that way)

[01:57:00.0000] <hsivonen>
Hixie: it appears that *some* authors don't think that way

[01:57:01.0000] <hsivonen>
Hixie: I'm inclined to think they are misguided, but still

[01:57:02.0000] <Hixie>
well if they want to be compliant, so much teh better

[01:57:03.0000] <Hixie>
could you send a mail elaborating on your idea for <cite>?

[01:57:04.0000] <Hixie>
i really want to address this inline vs block issue

[01:58:00.0000] <Hixie>
i don't really know hwo to do so

[01:58:01.0000] <Hixie>
maybe i really should do this matrix idea i mentioned

[01:58:02.0000] <hsivonen>
http://diveintomark.org/archives/2003/01/13/semantic_obsolescence

[01:59:00.0000] <hsivonen>
Hixie: OK. I'll send email on both issue

[01:59:01.0000] <hsivonen>
s

[01:59:02.0000] <Hixie>
well the inline thing i'm actually thinking of resolving right now

[01:59:03.0000] <Hixie>
so feel free to discuss that here

[01:59:04.0000] <othermaciej>
Hixie: I think the matrix would be a useful excercise to clarify what's important to capture in the content model reqirements

[01:59:05.0000] <othermaciej>
(there might be a bunch of "don't care" boxes)

[01:59:06.0000] <othermaciej>
Philip`: way to see a silver lining :-)

[02:00:00.0000] <zcorpan>
Hixie: did you see my thoughts about inline yesterday?

[02:01:00.0000] <zcorpan>
http://krijnhoetmer.nl/irc-logs/whatwg/20071211#l-361

[02:01:01.0000] <Hixie>
hsivonen: (and regarding mark's post, note that i work indirectly with him and he is up to date with the spec :-) )

[02:01:02.0000] <Hixie>
zcorpan: looking

[02:01:03.0000] <hsivonen>
Hixie: the following are random thoughts with no conclusion that I'd be comfortable with yet:

[02:02:00.0000] <hsivonen>
* so far it seems that authors don't like bimorphic

[02:02:01.0000] <Hixie>
zcorpan: i think there are pretty solid reasons for wanting to allow <p><ol/></p>, but i agree that it's conceptually a pain, especially given the html serialisation issue

[02:02:02.0000] <zcorpan>
indeed

[02:02:03.0000] <hsivonen>
* RELAX NG can validate bimorphic but the user experience sucks by default without schema-specific UI-level papering over

[02:02:04.0000] <Hixie>
hsivonen: (by which you mean they like mixing inline and block content?)

[02:03:00.0000] <hsivonen>
Hixie: yes

[02:03:01.0000] <hsivonen>
Hixie: cf. Sean Fraser, Sam Ruby and Dan Connolly

[02:03:02.0000] <Hixie>
yeah i agree that people want that

[02:03:03.0000] <Hixie>
i think we should allow that

[02:03:04.0000] <Hixie>
let's do this matrix thing

[02:03:05.0000] <Hixie>
/me finds a tool that can do easy editing of grids

[02:03:06.0000] <hsivonen>
* having general content models like "block" or "inline" makes schemas easier to write

[02:04:00.0000] <hsivonen>
* also makes conformance easier to teach

[02:04:01.0000] <zcorpan>
Hixie: (what's the matrix thing?)

[02:04:02.0000] <othermaciej>
google spreadsheets?

[02:04:03.0000] <othermaciej>
makes a grid, easy to share on the web

[02:04:04.0000] <hsivonen>
* having content model differences in XHTML5 is inconvenient

[02:04:05.0000] <hsivonen>
- Makes schema ugly

[02:04:06.0000] <Hixie>
othermaciej: way ahead of you :-)

[02:04:07.0000] <othermaciej>
zcorpan: which block/inline elements should be allowed in what others, assuming the rules were being designed from scratch

[02:05:00.0000] <hsivonen>
- Requires me to maintain a separate "HTML5-compatible subset of XHTML5" schema/mode

[02:05:01.0000] <hsivonen>
- Does not make sense with the "people should just use text/html" party line

[02:06:00.0000] <hsivonen>
- Does not make sense with the "apps should use XHTML internally but serialize to HTML5 for IE" party line

[02:06:01.0000] <Hixie>
yeah i agree that we should strive for no differences

[02:06:02.0000] <othermaciej>
Hixie: would you consider also multi-level nesting? (I guess that's most likely to affect <p>)

[02:06:03.0000] <Hixie>
which basically means the html serialisation wins

[02:06:04.0000] <Hixie>
or rather, can veto

[02:06:05.0000] <Hixie>
othermaciej: can you give an example where three-or-more-way nesting would have a different answer than two-way?

[02:07:00.0000] <hsivonen>
* I'm not convinced that achieving semantic purity with intra-paragraph lists is worth all the trouble

[02:07:01.0000] <Hixie>
othermaciej: i was just gonna do it on the basis of indirect nesting

[02:07:02.0000] <Hixie>
do we have a list of elements anywhere yet? i could use docs' magic list making feature but that seems unlikely to work perfectly here :-)

[02:08:00.0000] <othermaciej>
Hixie: I don't think there is such a case, but expressing the rules in terms of indirect nesting may be hard to understand

[02:08:01.0000] <othermaciej>
there are lists of elements, yes

[02:08:02.0000] <othermaciej>
I think zcorpan has one

[02:08:03.0000] <othermaciej>
http://simon.html5.org/html5-elements

[02:08:04.0000] <Hixie>
cool thanks

[02:08:05.0000] <Hixie>
othermaciej: well we'll worry about exactly how to express the rules later

[02:08:06.0000] <Hixie>
i just want an idea of what the ideal would be first

[02:09:00.0000] <hsivonen>
* OpenOffice.org Writer/Web makes an interesting case study of an editor with very strict block/inline boundaries

[02:09:01.0000] <othermaciej>
Hixie: if paragraphs could contain tables, then it might make sense to let a table in a paragraph contain a paragraph in a cell

[02:10:00.0000] <othermaciej>
I guess that is one plausible exception I can think of

[02:10:01.0000] <hsivonen>
* I don't know how to reconcile hand-authoring flexibility and importability in an app like OO.o Writer/Web except defining that editing apps MAY introduce a lot of layout-wrecking <p>s

[02:10:02.0000] <zcorpan>
some elements only allow inlines. others allow both block and inline. inline elements never allow blocks. something in that direction seems to be what authors believe the rules are, i think

[02:11:00.0000] <Hixie>
othermaciej: i'm not really convinced it would, but interesting

[02:11:01.0000] <hsivonen>
* Having different content models for <i> and <em> sucks big time

[02:12:00.0000] <hsivonen>
* In general, this whole strict inline thing probably sucks

[02:12:01.0000] <Hixie>
if you have google accounts post your e-mail addresses here (or msg me) so i can add you to this thing

[02:12:02.0000] <othermaciej>
maciej⊙gc

[02:12:03.0000] <zcorpan>
zcorpan⊙gc

[02:12:04.0000] <hsivonen>
hsivonen@

[02:13:00.0000] <krijnh>
krijnhoetmer⊙xn

[02:13:01.0000] <krijnh>
Is this about the mixing of inline/block content as well?

[02:14:00.0000] <hsivonen>
* I agree that the <div> content model in HTML 4 sucks semantically

[02:15:00.0000] <Hixie>
http://spreadsheets.google.com/ccc?key=pkNVM1HEQs-wsHB7s1M5Lbw

[02:15:01.0000] <hsivonen>
* But it seems the horse is out and the barn has burned

[02:15:02.0000] <Hixie>
feel free to fill in the cells you think look obvious

[02:15:03.0000] <Hixie>
if someone fills in something you disagree with, put a question mark after it

[02:16:00.0000] <hsivonen>
* We should probably allow Philip`'s Firefox <span> workaround

[02:17:00.0000] <zcorpan>
hmm, i'm a bit uncomfortable about that one. i breaks the rule "inlines never allows blocks"

[02:17:01.0000] <Philip`>
/me isn't sure that's worthwhile since it doesn't work in IE and is therefore a bit useless for most authors

[02:17:02.0000] <hsivonen>
* I don't like the direction of these points, since the direction is that <p> should allow inline and almost every other block container should allow %Flow

[02:18:00.0000] <zcorpan>
Philip`: indeed

[02:19:00.0000] <hsivonen>
zcorpan: isn't block-in-span an IEism that others have to implement for compat with content out there?

[02:19:01.0000] <Philip`>
<section><div class=section> lets you style .section in all browsers, and HTML5 UAs will get the sectioning correct

[02:20:00.0000] <zcorpan>
hsivonen: also h1-h6 and address

[02:20:01.0000] <Hixie>
feel free to fill in this spreadsheet too, btw, i don't plan on doing all 10000 cells myself :-P

[02:20:02.0000] <krijnh>
Ow, you're not? ;)

[02:20:03.0000] <zcorpan>
hsivonen: yes, but authors think it's disallowed

[02:20:04.0000] <Philip`>
Is there a Google Spreadsheet API to fill these things in automatically? :-)

[02:21:00.0000] <hsivonen>
as a schema writer and a closet markup purist, I like bimorhic and stuff

[02:21:01.0000] <hsivonen>
but as a validator front end developer and realist I don't 

[02:21:02.0000] <hsivonen>
I'm torn

[02:22:00.0000] <Hixie>
Philip`: it's not the api that's the hard part :-)

[02:22:01.0000] <Hixie>
i think it's clear that forcing a separation of inline and block isn't working in practice

[02:22:02.0000] <jgraham>
/me wishes he could stay for this discussion

[02:22:03.0000] <Hixie>
i think we can just say that paragraphs are implied

[02:23:00.0000] <Philip`>
/me wishes he could stay because it looks freezing outside

[02:23:01.0000] <Hixie>
done 50 of about 10000 so far

[02:23:02.0000] <jgraham>
Philip`: That too :)

[02:23:03.0000] <zcorpan>
(perhaps we should change address to allow %flow as well)

[02:23:04.0000] <hsivonen>
Hixie: yeah, but is there a credible story that'd allow the likes of OpenOffice.org Writer/Web to make all those implicit paragraphs explicit when it reserializes from its non-DOM internal datastructure?

[02:23:05.0000] <Hixie>
hsivonen: i don't think semantically we should disallow it

[02:24:00.0000] <Hixie>
hsivonen: of course as you say, it causes havoc with styling

[02:24:01.0000] <Hixie>
not sure what to do about that

[02:24:02.0000] <jgraham>
Hixie: Can you ad my jgraham.html@gmail account 

[02:24:03.0000] <Hixie>
done

[02:24:04.0000] <jgraham>
thx

[02:24:05.0000] <Hixie>
anyone can feel free to add other people btw

[02:25:00.0000] <Hixie>
in case y'all want to continue this when i go to bed :-)

[02:25:01.0000] <colione_>
olle.lundberg@gmail

[02:26:00.0000] <Hixie>
added

[02:26:01.0000] <colione_>
thnx

[02:28:00.0000] <Hixie>
woot, finished <abbr>

[02:29:00.0000] <Hixie>
hmm, should <caption> ... <address> </address> </caption> be legal

[02:31:00.0000] <zcorpan>
Hixie: no, caption should only allow inline (like h1-h6)

[02:31:01.0000] <Hixie>
makes sense

[02:32:00.0000] <Hixie>
hmm, should we allow a <dialog> to contain a <section> or <address> or <footer>...

[02:32:01.0000] <Hixie>
indirectly even

[02:33:00.0000] <krijnh>
What do you mean by indirectly contain?

[02:33:01.0000] <Hixie>
<dialog> <dt> krijnh <dd> <section> <p> What do you mean by indirectly contain? ...

[02:33:02.0000] <Hixie>
in fact, should we allow <ol>/<ul>/<dl> to contain any sectioning elements

[02:33:03.0000] <Hixie>
i'm thinking not really

[02:34:00.0000] <Hixie>
how about tables? should they be allowed to contain sectioning elements?

[02:34:01.0000] <hsivonen>
Indirectly does not sound good for my purposes...

[02:34:02.0000] <Hixie>
i'm sure we'll find a better way of phrasing this in due course

[02:34:03.0000] <krijnh>
Hixie: probably for table based layouts

[02:34:04.0000] <hsivonen>
Hixie: a <td> should have the some content model as <body> to allow real-world authoring patterns

[02:35:00.0000] <Hixie>
krijnh: well those aren't legal anyway

[02:35:01.0000] <hsivonen>
s/some/same/

[02:35:02.0000] <Hixie>
hsivonen: really? i'd have thought discouraging table-based layouts would be a plus here.

[02:35:03.0000] <Hixie>
they're already invalid

[02:35:04.0000] <hsivonen>
or more to the point, <td> should allow at least everything <body> allows

[02:35:05.0000] <Hixie>
we just can't catch them

[02:36:00.0000] <Hixie>
maybe <address> should only be allowed as a direct child of a sectioning element?

[02:36:01.0000] <Hixie>
hmm

[02:36:02.0000] <hsivonen>
Hixie: making table-based layout invalid is pointless until the CSS WG delivers a viable alternative *and* the top four browsers implement it

[02:36:03.0000] <Hixie>
hsivonen: table based layout has never been valid

[02:37:00.0000] <hsivonen>
Hixie: then perhaps we should make them valid

[02:38:00.0000] <hsivonen>
Hixie: at least with appropriate role='' pixie dust

[02:38:01.0000] <annevk>
annevankesteren⊙gc

[02:38:02.0000] <Hixie>
hsivonen: ew

[02:38:03.0000] <Hixie>
to the first part

[02:38:04.0000] <Hixie>
not so much the second

[02:38:05.0000] <Hixie>
annevk: done

[02:40:00.0000] <hsivonen>
Hixie: is the table intentionally missing <table> and table-internal stuff?

[02:40:01.0000] <Hixie>
yeah

[02:40:02.0000] <Hixie>
<td> and <th> cover those

[02:40:03.0000] <Hixie>
and <caption>

[02:40:04.0000] <Hixie>
they're the only "exit points" for tables

[02:40:05.0000] <Hixie>
i think i might drop <ol> <ul> <dl> <dialog> too for the same reason

[02:41:00.0000] <annevk>
Hixie, wasn't <p><strong>blah</strong></p> for <lede> or <lead>?

[02:41:01.0000] <Hixie>
annevk: it's not really marking importance, is it?

[02:41:02.0000] <Hixie>
actually <figure> should probably be taken out too

[02:42:00.0000] <Hixie>
it has a whole other set of issues

[02:42:01.0000] <krijnh>
Can an article contain an article?

[02:42:02.0000] <krijnh>
/me is very efficient - 2 no's already

[02:43:00.0000] <annevk>
yeah, for comments on an article

[02:43:01.0000] <hsivonen>
/me hopes the table will eventually generalize into a handful of content models to teach and to type into a schema

[02:43:02.0000] <Hixie>
krijnh: yes, blog comments are articles in articles

[02:43:03.0000] <Hixie>
hsivonen: i hope so too

[02:44:00.0000] <annevk>
Hixie, yeah, fair enough, just thought that was the idea earlier on

[02:45:00.0000] <othermaciej>
Hixie: does it make sense to have <td> on the horizontal axis?

[02:45:01.0000] <othermaciej>
lots of things can indirectly contain it but not directly

[02:45:02.0000] <othermaciej>
(similarly for other table structure)

[02:45:03.0000] <Hixie>
othermaciej: you mean the vertical axis? the question is "can elements in the left hand column contain elements on the top row"

[02:45:04.0000] <Hixie>
you mean on the top row?

[02:45:05.0000] <othermaciej>
yes

[02:46:00.0000] <Hixie>
i agree that we should remove all the inner table elements from the top row

[02:46:01.0000] <Hixie>
feel free to do so

[02:46:02.0000] <othermaciej>
ok

[02:48:00.0000] <othermaciej>
I will remove <html> <head> and <body> from the top row for similar reasons

[02:48:01.0000] <annevk>
bah, the CSS WG discussions should really be public

[02:48:02.0000] <Hixie>
i'm nuking option and optgroup from the first column since they're special

[02:49:00.0000] <Hixie>
select too, same reason

[02:49:01.0000] <krijnh>
Hixie: can't <details> contain additional contact information?

[02:49:02.0000] <othermaciej>
whoah, what's <nest>?

[02:49:03.0000] <Hixie>
krijnh: like <details> <legend> Contact information for this page </legend> <address> ... </address> </details> ?

[02:49:04.0000] <krijnh>
Yeah

[02:49:05.0000] <Hixie>
othermaciej: part of the data template feature. nuke it, it's seriously special.

[02:49:06.0000] <Hixie>
othermaciej: same wiht <datatemplate> and <rule>

[02:50:00.0000] <Hixie>
othermaciej: i'm probably gonna nuke that whole section anyway, in favour of some apis to make it easier to make your own templating language and graft it onto html

[02:50:01.0000] <Hixie>
othermaciej: there are too many different ways of doing templates, each with their own pros and cons, to really annoint any one model

[02:51:00.0000] <Hixie>
krijnh: sounds plausible...

[02:51:01.0000] <othermaciej>
removed

[02:51:02.0000] <krijnh>
So then an article can be put in a details as well

[02:51:03.0000] <othermaciej>
looking for other things too special to be worth gridding

[02:51:04.0000] <krijnh>
<details><legend>More articles on this subject</legend><article>foo</article><article>bar</article></details>

[02:52:00.0000] <Hixie>
seems reasonable

[02:52:01.0000] <krijnh>
Or would that be wrong use?

[02:52:02.0000] <Hixie>
it seems unexpected use, but i don't see that it would be wrong... dunno

[02:53:00.0000] <othermaciej>
base, link, meta, noscript, optgroup, option, param, source, style, script, title

[02:53:01.0000] <othermaciej>
from the first row

[02:53:02.0000] <othermaciej>
any objections?

[02:53:03.0000] <krijnh>
noscript ?

[02:54:00.0000] <othermaciej>
isn't noscript allowed anywhere and therefore not worth mentioning?

[02:54:01.0000] <othermaciej>
it does not seem affected by block/inline considerations

[02:54:02.0000] <othermaciej>
but I'll not remove it for now

[02:55:00.0000] <Hixie>
leave noscript for now, it's a weird case

[02:56:00.0000] <Hixie>
but the others can go for sure

[02:56:01.0000] <Hixie>
actually <style> might be worth leaving

[02:57:00.0000] <annevk>
<area> can be nuked from the top row probably

[02:58:00.0000] <Hixie>
why?

[02:58:01.0000] <Hixie>
<area>'s a tough one

[02:58:02.0000] <Hixie>
when it's allowed is unclear to me

[02:58:03.0000] <annevk>
only as descendent of <map>

[02:58:04.0000] <annevk>
or maybe only as child of <map>

[02:59:00.0000] <Hixie>
we're allowing things like <map><area/><area/></map> as well as things like <map><p>...<a/><area/>...</map>

[02:59:01.0000] <Hixie>
so far

[02:59:02.0000] <othermaciej>
I nuked <style> already, I can put it back if you really want it

[03:00:00.0000] <othermaciej>
(where <style scoped> is allowed is kind of interesting, but not really related to the core block/inline type issue)

[03:00:01.0000] <annevk>
Hixie, I'm not sure what the use of the latter is

[03:01:00.0000] <Hixie>
othermaciej: nuking is fine

[03:01:01.0000] <Hixie>
annevk: makes it easier to make sure you've got all your links and areas done together

[03:01:02.0000] <annevk>
isn't <area> a link?!

[03:01:03.0000] <annevk>
hmm

[03:02:00.0000] <annevk>
then again, I thought <map> was display:none, it isn't

[03:02:01.0000] <Hixie>
annevk: <area> is a link with a shape

[03:02:02.0000] <krijnh>
Hmm, could video contain section elements?

[03:02:03.0000] <annevk>
Hixie, per HTML4 so is <a>

[03:03:00.0000] <Hixie>
yeah but we dropped that long ago

[03:03:01.0000] <annevk>
and I'm not sure how that's relevant

[03:03:02.0000] <Hixie>
krijnh: i'd say yes, if the fallback is very detailed :-)

[03:03:03.0000] <Hixie>
annevk: ?

[03:03:04.0000] <krijnh>
Hixie: yeah, or has contact information?

[03:03:05.0000] <Hixie>
krijnh: right

[03:04:00.0000] <Hixie>
i could see one doing <video> <address> For a transcript, contact ...</address> </video>

[03:04:01.0000] <Hixie>
i guess

[03:04:02.0000] <krijnh>
Cause you put a ? at video->address :)

[03:07:00.0000] <Hixie>
fixed :-)

[03:07:01.0000] <annevk>
Hixie, I'm not sure how <area> being a shaped link is relevant

[03:07:02.0000] <Hixie>
annevk: when you're doing an image map, you want to provide both the shape link <area> and the fallback link <a>

[03:07:03.0000] <Hixie>
for each link

[03:07:04.0000] <Hixie>
easiest to do if you have them all together

[03:08:00.0000] <Hixie>
rather than as two blocks

[03:08:01.0000] <annevk>
isn't the algorithm for fallback to use <area>?

[03:08:02.0000] <annevk>
or is this in the case <map> is not supported?

[03:09:00.0000] <Hixie>
if <a>s aren't provided the UA uses <area>, but that kinda sucks compared to providing a custom fallback

[03:09:01.0000] <annevk>
that's not at all how the image map algorithm for fallback works

[03:10:00.0000] <Teratogen>
bring ogg back!

[03:18:00.0000] <hsivonen>
coming up with suggested definition of <cite> is hard

[03:18:01.0000] <hsivonen>
I'm bad at writing weasel words

[03:18:02.0000] <Dashiva>
hehe

[03:21:00.0000] <Hixie>
annevk: really?

[03:22:00.0000] <Hixie>
that was my intention...

[03:22:01.0000] <Hixie>
Teratogen: :-)

[03:24:00.0000] <Hixie>
hm, should <map> allow <section> in it then? or <article>?

[03:24:01.0000] <Hixie>
or <aside> or <address>? hmm

[03:26:00.0000] <annevk>
just <area> imo

[03:27:00.0000] <hsivonen>
is the diveintomark.org favicon one of the allegedly indecent IceWeasel icon suggestions?

[03:29:00.0000] <Hixie>
krijnh: ogg :-P

[03:29:01.0000] <krijnh>
:p

[03:30:00.0000] <krijnh>
Bring it back! ;)

[03:32:00.0000] <hsivonen>
Hixie: so I didn't send email about block/inline, because I dumped my points here instead

[03:33:00.0000] <Hixie>
k

[03:33:01.0000] <Hixie>
i think i mostly agreed with your points anyway

[03:33:02.0000] <Hixie>
it's turning out that there are some edge cases that i hadn't really thought of

[03:33:03.0000] <Hixie>
like should <nav> be able to contain <article>

[03:34:00.0000] <annevk>
no

[03:35:00.0000] <hdh>
/me imagines people use narration to guide users around the site

[03:36:00.0000] <krijnh>
<body><nav><section><article>An interesting article with lots and lots of interesting links</article></section></nav></body>

[03:37:00.0000] <krijnh>
Hixie: is that conditional formatting?

[03:37:01.0000] <krijnh>
Yes, cool

[03:38:00.0000] <Hixie>
yeah

[03:39:00.0000] <Hixie>
ok i'm gonna go sleep, i can't concentrate anymore

[03:39:01.0000] <Hixie>
feel free to continue editing :-)

[03:39:02.0000] <Hixie>
thanks for the help btw

[03:39:03.0000] <Hixie>
really helpful

[03:39:04.0000] <Hixie>
i'll try to continue this tomorrow

[03:40:00.0000] <hsivonen>
it is "interesting" how <article> and friends that would be "easy" to implement are less supported by browsers than "hard" stuff like <canvas> and <video>

[03:41:00.0000] <annevk>
it's not entirely clear what it would mean for rendering to support <section> and other sectioning elements

[03:42:00.0000] <krijnh>
It's also a dull feature to sell :)

[03:42:01.0000] <Camaban>
hsivonen: video and canvas are 'cool' and 'new', while article doesn't 'do' anything? :)

[03:43:00.0000] <annevk>
<article> together with <h1>-<h6> can in theory effect rendering, but it's unclear how

[03:43:01.0000] <hsivonen>
krijnh: yeah. it makes me wonder if using <article> instead of <div class='article'> will ever be compelling for authors

[03:44:00.0000] <annevk>
I think if CSS gave you something like :heading(2) to style all level two headers that might work

[03:46:00.0000] <krijnh>
hsivonen: do you think authors would use the new elements already, if IE/Fx didn't close unknown block level elements immediately?

[03:46:01.0000] <krijnh>
That's probably easy to fix behavior, but I don't think it would change anything

[03:47:00.0000] <othermaciej>
:heading(n) might be handy for implementing the default rendering of <h[1-6]> as well

[03:48:00.0000] <othermaciej>
hsivonen: we could easily support default rendering as a block for all the new semantic block elements

[03:48:01.0000] <othermaciej>
hsivonen: supporting headings styled in accordance with the outline algorithm would be hard and the spec doesn't say how to do that yet

[03:49:00.0000] <othermaciej>
(or whether)

[03:49:01.0000] <othermaciej>
I will tell you that we're interested in supporting any new elements and attributes that seem like low hanging fruit in WebKit in the fairly near future

[03:52:00.0000] <othermaciej>
(that would basically be irrelevant="", sectioning elements, dialog, m if someone decides it should have some special default style

[03:52:01.0000] <othermaciej>
)

[03:52:02.0000] <othermaciej>
figure would also be low-hanging fruit if not for the <legend> issue

[03:54:00.0000] <hsivonen>
krijnh: with the IE/Firefox situation, using the new elements is not worthwhile ATM from the author POV

[03:55:00.0000] <hsivonen>
othermaciej: Opera Mobile has a nice "scroll to content feature" it would be cool to have that in WebKit, too, and both taking <article> into account

[03:56:00.0000] <hsivonen>
actually, that's the only UA-side semantic treatment of <article> that I can come up with at the moment

[03:56:01.0000] <hsivonen>
skipping to content whether on mobile or in an aural browsing setup

[03:58:00.0000] <othermaciej>
yeah, supporting the new block-level elements would not have much value besides patriotism just yet

[04:12:00.0000] <annevk>
maybe [#heading=n]

[04:12:01.0000] <annevk>
at some point there was this idea of separating intrinsic attributes of pseudo-classes, but maybe that point is moot

[04:16:00.0000] <othermaciej>
"intrinsic attributes"?

[04:17:00.0000] <annevk>
td[#col=2] was another one

[04:18:00.0000] <hsivonen>
anyway, it seems that the whole <section> thing hinges upon a selector with reasonable perf and implementation characteristics

[04:19:00.0000] <krijnh>
Something for CSS5?

[04:21:00.0000] <othermaciej>
I'm not sure why that's better than pseudo-classes

[04:22:00.0000] <annevk>
me neither, I guess the idea might have been dropped already

[04:23:00.0000] <krijnh>
Why was it an idea in the first place then?

[04:24:00.0000] <hsivonen>
Is it reasonable to expect the CSS WG to have cycles to look into an outline-dependent selector any time soon?

[04:25:00.0000] <hsivonen>
they seem to have a lot on their plate even without new HTML5 needs

[04:27:00.0000] <othermaciej>
this is the WG that's bringing us ascii art layout

[04:27:01.0000] <othermaciej>
there doesn't have to be a reason

[04:32:00.0000] <annevk>
it should be pretty easy to draft a specific selector proposal for heading:

[04:32:01.0000] <hsivonen>
hmm. the TAG is going the way of SGML: the more common concepts have the longer names: resource vs. resource representation

[04:32:02.0000] <annevk>
especially as CSS only needs to define the syntax and say that it's up to languages to define when it actually matches

[04:33:00.0000] <hsivonen>
annevk: you still need a kind of rare person who groks CSS formatter internals well enough to assess the computational feasibility with dynamics DOM changes

[04:34:00.0000] <krijnh>
no

[04:34:01.0000] <krijnh>
no

[04:34:02.0000] <krijnh>
Oops :)

[04:35:00.0000] <annevk>
I can ask the guy who implemented selectors in Opera

[04:35:01.0000] <othermaciej>
yeah, I'm not sure the current html5 outline algorithm is computationally feasible for incremental rendering and dynamic DOM updates

[04:35:02.0000] <annevk>
I suppose other implementors have ways to find out themselves

[04:37:00.0000] <othermaciej>
it's written in terms of generating a hypothetical tree and walking it

[04:37:01.0000] <othermaciej>
but for selector matching you need something that's evaluated from the element point of view

[04:39:00.0000] <othermaciej>
so it's hard to tell if it can be efficient without doing a conversion to that type of algorithm first (and making sure it is actually equivalent)

[04:40:00.0000] <othermaciej>
it's not clear to me if changes outside an <hn> element can change its heading level

[04:41:00.0000] <othermaciej>
actually it's not that clear to me how the section tree affects heading levels

[04:48:00.0000] <annevk>
<section><section><h1> would be level 3 I think

[05:03:00.0000] <hasather>
The Ogg debate is like a hydra. You read one thread, and two others pop up during the time

[05:11:00.0000] <Dashiva>
And over half the mails are by the same two people

[05:13:00.0000] <Dashiva>
I'm strongly tempted to send a mail saying "So-called non-commercial entities have the option to pay for licences, they just choose not to."

[05:16:00.0000] <Philip`>
Open source entities don't have that option, since they couldn't distribute their code in a way that other people could modify and use

[05:18:00.0000] <Dashiva>
But that's just an arbitrary restriction they place on themselves. We're here to get interoperability, not to run errands for other organizations

[05:18:01.0000] <Philip`>
(unless they get a licence which allows everybody royalty-free usage of the patents for any purpose)

[05:18:02.0000] <Philip`>
(which is what Theora got, so it's not totally impossible)

[05:28:00.0000] <hsivonen>
I wonder what MPEG-LA estimates as the expected value of the overall H.264 licensing income over the lifetime of the patents discounted to present value

[05:33:00.0000] <Philip`>
/me wonders if it's worth trying some basic comparisons of non-state-of-the-art video codecs

[05:36:00.0000] <othermaciej>
if you know how to do such a thing then sure

[05:36:01.0000] <Philip`>
I don't know how to do it especially well, hence the "basic" :-)

[05:37:00.0000] <othermaciej>
might be interesting to try this shootout with H.261, MPEG-2, H.263, etc: http://osnews.com/story.php/19019/Theora-vs-h.264/

[05:41:00.0000] <hsivonen>
something that hasn't been explored: it is important to have RF decoding and *an* RF encoder, but it doesn't follow that there could not exist non-RF state-of-the-art encoders or hardware decoders

[05:41:01.0000] <hsivonen>
back when *compressed* GIF encoding was encumbered, there were RF decoders

[05:42:00.0000] <Philip`>
http://www.doom9.org/index.html?/codecs-quali-105-1.htm has Theora, but it's a few years old now and I don't know how much has changed

[05:42:01.0000] <hsivonen>
and RF encoders that sucked badly (i.e. produced a stream that decoded as lzw but was not compressed)

[05:43:00.0000] <Philip`>
(Also, DVDs are very different quality to what you'd publish on the web)

[05:44:00.0000] <othermaciej>
MPEG-LA could probably limit decoder revenue to hardware implementations, or mobile devices, or both, and not lose significant revenue

[05:45:00.0000] <othermaciej>
s/revenue/royalties/

[05:45:01.0000] <hsivonen>
mobile devices is still a field-of-use restriction that would not go well with Open Source

[05:46:00.0000] <othermaciej>
realistically nearly all handsets on which you could run interesting software have paid the license fee

[05:54:00.0000] <alp>
othermaciej: mobile distributors of webkit/gtk+ and webkit/qt at least would probably be happy with a royalty-free codec from what i gather but asking for licensing fees is pushing it. maybe the situation is different for other free browsers though

[05:54:01.0000] <othermaciej>
I don't think you need to pay royalties for the browser component if the phone already has a hardware or software decoder

[05:55:00.0000] <othermaciej>
which many do

[05:55:01.0000] <othermaciej>
could be wrong though, it's hard to understand the MPEG-LA's documents

[06:00:00.0000] <alp>
for various reasons it may be necessary to ship the codec with the browser (say if the hardware only allows video overlay and you need to support more complex rendering features, or if the hardware just doesn't support the codec at all in the first place)

[06:01:00.0000] <Philip`>
It's fun how FFmpeg is happy to be told to output an Ogg file, but actually it silently doesn't support it and outputs something totally different

[06:02:00.0000] <othermaciej>
fair enough, I don't really know how this works

[06:06:00.0000] <Philip`>
Hmm, H.261 doesn't support 320x240 :-(

[06:06:01.0000] <othermaciej>
is that too small or too big for it?

[06:06:02.0000] <othermaciej>
(I guess that is a problem in itself)

[06:08:00.0000] <Philip`>
"Valid sizes are 176x144, 352x288"

[06:08:01.0000] <Philip`>
Great flexibility!

[06:09:00.0000] <Philip`>
Also, it looks horrible quality

[06:10:00.0000] <Lachy_>
wow, that's terrible

[06:10:01.0000] <Lachy_>
what about motion jpeg or any of the other older alternatives?

[06:11:00.0000] <Lachy_>
motion jpeg just sounds like it would have huge file sizes

[06:13:00.0000] <annevk>
motion jpeg is not a serious option

[06:13:01.0000] <Philip`>
MJPEG doesn't look that awful compared to other codecs

[06:13:02.0000] <maikmerten>
H.261 didn't even expire

[06:14:00.0000] <maikmerten>
it's a 1990 standard

[06:15:00.0000] <Philip`>
(With my current utterly rubbish test setup, only looking at 5 seconds of video, it's quite better quality than H.261 but twice the filesize and I guess I need to fiddle with the encoder settings to get a fair comparison)

[06:15:01.0000] <maikmerten>
MJPEG doesn't stand a chance against even H.261

[06:15:02.0000] <maikmerten>
no motion compensation, no inter-frame coding...

[06:15:03.0000] <Philip`>
It can do more than two different frame sizes, though :-)

[06:15:04.0000] <maikmerten>
I'll give it that ;)

[06:16:00.0000] <maikmerten>
plus there is no MJPEG standard IIRC

[06:16:01.0000] <maikmerten>
there are a lot of codecs claiming "MJPEG"

[06:16:02.0000] <Philip`>
and it can do variable bitrates, which I assume H.261 can't since it was designed for streaming over ISDN

[06:17:00.0000] <maikmerten>
H.261 should be able to do VBR

[06:17:01.0000] <maikmerten>
thanks to the very nature of video compression codecs are VBR

[06:18:00.0000] <maikmerten>
and it's actually a lot of work to get them to do CBR

[06:18:01.0000] <maikmerten>
(bitrate reservoirs etc. etc.)

[06:18:02.0000] <MikeSmith>
/me finally gets to reading Hixie responses to messages he sent about <term> and <xref> stuff

[06:19:00.0000] <MikeSmith>
/me goes to re-read spec for <i>

[06:19:01.0000] <maikmerten>
(well, okay, granted, it's very possible to develop codecs with a fixed bitrate and CBR)

[06:19:02.0000] <Philip`>
maikmerten: If you want to do e.g. videoconferencing over an ISDN channel, that'd have to be CBR since you can't do buffering, and I thought that was roughly what H.261 was designed for

[06:20:00.0000] <maikmerten>
(but it's difficult to do e.g. with DCT based codecs)

[06:20:01.0000] <maikmerten>
why wouldn't you be allowed to send less data than the line allows you to send?

[06:20:02.0000] <Philip`>
(but you know more about this than I do so I'm probably wrong :-) )

[06:20:03.0000] <maikmerten>
in worst case just pad with zeros ;)

[06:20:04.0000] <Philip`>
You could send less but that'd be a waste of resources

[06:21:00.0000] <Philip`>
since you're not going to be able to reuse the spare bandwidth for anything else

[06:21:01.0000] <maikmerten>
(which, actually, is one way to get CBR if you manage to ceil the bitrate: Just fill up "unused" bits with crap)

[06:21:02.0000] <Philip`>
so you might as well send higher quality images instead of lowering the bitrate

[06:21:03.0000] <maikmerten>
well, albeit it's good to use up all the bandwidth you may not be able to actually deliver

[06:22:00.0000] <maikmerten>
think compressing a perfect black frame

[06:22:01.0000] <maikmerten>
DCT will give you a nice zero run

[06:22:02.0000] <maikmerten>
near-perfect compression

[06:22:03.0000] <maikmerten>
(only protocol overhead)

[06:22:04.0000] <Philip`>
You could send 64Kb/s of a really really precise shade of black

[06:22:05.0000] <maikmerten>
so it's actually *hard* to guarantee you're using up all bandwidth ;)

[06:22:06.0000] <Philip`>
Fair enough :-)

[06:22:07.0000] <Philip`>
/me has to go for some minutes

[06:23:00.0000] <maikmerten>
well, you mostly can't be any more precise than (0,0,0) with black ;)

[06:23:01.0000] <Philip`>
You can do (0.00000000, 0.00000000, 0.00000000) :-)

[06:23:02.0000] <maikmerten>
basically all video codecs are integer based

[06:23:03.0000] <Philip`>
/me is gone

[06:24:00.0000] <maikmerten>
plus 0.0000000 would just be another case of "fill up with crap" ;)

[06:38:00.0000] <Philip`>
maikmerten: Those extra significant figures are important in physics - something that I measure as 0.0kg is probably much heavier than something I measure as 0.000000kg :-)

[06:39:00.0000] <maikmerten>
well, only if you try to give an idea with what sort of precision you were working along with the value 

[06:39:01.0000] <maikmerten>
(which is often done)

[06:40:00.0000] <maikmerten>
(but for christ's (or any other religious figure's) sake - zero shall be zero ;-) )

[06:40:01.0000] <maikmerten>
actually my physics teacher always got semi-angry if we didn't compact numerical values ;)

[06:41:00.0000] <maikmerten>
(0.0340000 wouuld be 0.034)

[06:42:00.0000] <Philip`>
Eww - I've always been taught that significant figures are significant, and you can't just add them or drop them off whenever you fancy

[06:43:00.0000] <maikmerten>
well, you *have* to specify what precision is used

[06:43:01.0000] <Philip`>
and the worst thing you can possibly do is write 3½ instead of 3.5, because "½" implies some kind of mathematical precision that physics never has

[06:43:02.0000] <Philip`>
/me prefers Computer Science where everything is integers ;-)

[06:43:03.0000] <maikmerten>
aye

[06:52:00.0000] <Philip`>
/me wonders what typical internet video bitrate is

[06:53:00.0000] <maikmerten>
youtube used to serve 256 kbit/s h.263 with 64 kbit/s MP3

[06:53:01.0000] <maikmerten>
(the latter 22.05 kHz, mono - MP3 just is pretty far behind)

[06:54:00.0000] <Philip`>
Based on an extensive sample of two Youtube FLVs in my /tmp, they're 320kbit/s, so that sounds right

[06:55:00.0000] <maikmerten>
yup, two is basically a perfect statistical base ;)

[06:55:01.0000] <maikmerten>
but they're batch encoded with same settings anyway 

[06:57:00.0000] <hsivonen>
what press release is David Gerard referring to on whatwg@?

[06:58:00.0000] <annevk>
the one where Chris Double is quoted

[06:58:01.0000] <annevk>
about Opera and Mozilla pushing <video>

[06:58:02.0000] <hsivonen>
/me notes that dgerard talk about wikipedia and "we" in a way that assumes that everyone knows his wikipedia affiliation

[06:58:03.0000] <hsivonen>
annevk: ah the PC World article?

[06:59:00.0000] <annevk>
PC World just copied it

[06:59:01.0000] <annevk>
just like Washington Post and several others

[06:59:02.0000] <hsivonen>
hmm. something has gotten past my HTML5 radar

[06:59:03.0000] <Philip`>
Urgh, H.263 says "Valid sizes are 128x96, 176x144, 352x288, 704x576, and 1408x1152. Try H.263+."

[06:59:04.0000] <Philip`>
/me tries H.263+, which works

[07:01:00.0000] <hsivonen>
annevk: whose press release it was?

[07:01:01.0000] <annevk>
not sure what the original was

[07:01:02.0000] <maikmerten>
H.263+ is 1998

[07:01:03.0000] <maikmerten>
it's not close to expiring

[07:02:00.0000] <maikmerten>
H.263 itself is 1995/1996

[07:03:00.0000] <maikmerten>
it's more or less a direct predecessor to MPEG4 Part 2, IIRC

[07:03:01.0000] <Philip`>
What happened to H.262? :-)

[07:04:00.0000] <hsivonen>
annevk: I don't see a <video> press release from any of WHATWG, W3C, Mozilla or Opera

[07:04:01.0000] <maikmerten>
when MPEG ran out of puppies they began consuming future standards

[07:04:02.0000] <maikmerten>
look MPEG-3 ;)

[07:04:03.0000] <annevk>
hsivonen, there was no press release

[07:04:04.0000] <hsivonen>
annevk: ok

[07:04:05.0000] <annevk>
someone made an article that was reused all over the place (even localized)

[07:04:06.0000] <hsivonen>
ok

[07:05:00.0000] <maikmerten>
press releases are boring anyway.... "CEO of ..... says.... '...glad to be here and drive innovation..... customer satisfaction.... revenue... world domination'" - not sure I ever read a really interesting press release

[07:06:00.0000] <hdh>
the opera's dork release?

[07:06:01.0000] <hdh>
bork, maybe, the spelling escaped me

[07:11:00.0000] <hsivonen>
the silly thing about press releases is that all the substance has to be cast into quotations so that jounalists can print them as quotations and avoid stating anything controversial in text that isn't quoted

[07:11:01.0000] <hsivonen>
so to write a press release, one has to first come up with the points, then massage them into soundbytes and then figure out who agrees to be attributed with which soundbyte

[07:12:00.0000] <Camaban>
so journos can misinterpret them, mis-quote them, and be selective about what they quote to create a story ;)

[07:12:01.0000] <Philip`>
(Hmph, I tell these things to do 256Kb/s but they end up doing 430Kb/s instead :-( )

[07:13:00.0000] <Philip`>
(Maybe they're just not designed to scale so low?)

[07:16:00.0000] <maikmerten>
many codecs have limits on how low bitrate can be

[07:16:01.0000] <hsivonen>
comparing codecs properly is very hard, because the fixed parameter is what code you ship to the client and the tricks the encoder does plays such a big role

[07:16:02.0000] <maikmerten>
tried with lower resolution?

[07:16:03.0000] <maikmerten>
aye

[07:16:04.0000] <hsivonen>
so it is quite possible that you end up testing encoders instead of decoding specs

[07:17:00.0000] <hsivonen>
Philip`: are you testing ffmpeg encoders against each other?

[07:17:01.0000] <Philip`>
Testing just decoding specs isn't very useful in practice

[07:17:02.0000] <maikmerten>
well, at least for extremely old codecs "what we have now is as good as it'll ever get"

[07:17:03.0000] <Philip`>
since people will have to encode things, using what's available

[07:17:04.0000] <hsivonen>
Philip`: it isn't but testing a bad encoder or a good encoder with bad params isn't, either

[07:18:00.0000] <hsivonen>
I'd love to have a cheat sheet of QuickTime/H.264, x264, XiphQT and ffmpeg2theora tried and true magic params

[07:19:00.0000] <Philip`>
hsivonen: Yep, I'm just looking at FFmpeg for now, which is far from ideal and I won't claim this is an especially good comparison :-)

[07:19:01.0000] <hsivonen>
since the stuff other people encode tends to look better than what I get with naïve dabbling

[07:20:00.0000] <maikmerten>
most frontends e.g. don't expose all coding options

[07:20:01.0000] <maikmerten>
like in case of Theora I usually end up altering keyframe interval, the noise threshold or even the complete set of quantization tables

[07:20:02.0000] <hsivonen>
maikmerten: or worse, they do expose a zillion options that let you shoot yourself in the foot by overstepping your AVC profile bounds

[07:21:00.0000] <maikmerten>
well, that is a genuine opportunity for formats with profiles ;)

[07:21:01.0000] <hsivonen>
/me thinks AVC profiles and levels are an awfully bad idea from the interop POV

[07:21:02.0000] <maikmerten>
well, the argument was that at least it could be made sure restricted deviced could at least reliably support *something*

[07:22:00.0000] <maikmerten>
but I feel this has gone out of hand a bit

[07:23:00.0000] <maikmerten>
it's sometimes not quite easy to e.g. come up with a file that happens to play fine on both Playstation Portable and the iPod and some mobile phone etc. etc.

[07:24:00.0000] <hsivonen>
Google Video has the iPod/PSP magic figured out but afaik they aren't sharing it

[07:24:01.0000] <maikmerten>
especially for extremely sophisticated codecs it can make sense to have profiles that drop coding schemes that (wild example) increase CPU usage by 50% but only give 5% coding gain

[07:25:00.0000] <Philip`>
/me can't get below 1300Kb/s with MJPEG

[07:27:00.0000] <MikeSmith>
hsivonen, annevk - I believe the source for the "Mozilla, Opera Want to Make Video on the Web Easier" article was Jeremy Kirk of IDG

[07:27:01.0000] <MikeSmith>
pcworld article has the correct byline at least

[07:28:00.0000] <MikeSmith>
it was not a press release from Opera or Mozilla or whoever

[07:28:01.0000] <annevk>
yeah, I recall reading those names

[07:31:00.0000] <hsivonen>
MikeSmith: ok

[07:53:00.0000] <Philip`>
hsivonen: Would there be any chance of your HTML Parser including a brief summary of the changes between releases?

[07:56:00.0000] <hsivonen>
Philip`: perhaps next time.

[07:56:01.0000] <hsivonen>
Philip`: this time the main difference was Mavenization

[07:58:00.0000] <Philip`>
hsivonen: Okay - it's just useful to know if e.g. the main difference is something like Mavenization that I don't care about, or if it's important bug fixes and I should bother updating

[07:58:01.0000] <Philip`>
but since "updating" involves copying one file over another, it's not a significant issue at all :-)

[08:01:00.0000] <hsivonen>
Philip`: I can't remember if there was something else as well

[08:01:01.0000] <hsivonen>
/me looks at logs

[08:03:00.0000] <hsivonen>
Philip`: there was also a bug fix in case you run SAX Tree without a Locator

[08:05:00.0000] <Philip`>
hsivonen: Okay, thanks

[08:06:00.0000] <hsivonen>
Philip`: also, I eleminated a bogus import that referenced a Sun-internal class and caused badness 

[08:06:01.0000] <hsivonen>
Philip`: that's about it

[08:07:00.0000] <annevk>
hsivonen, see www-style for media queries

[08:07:01.0000] <annevk>
/me tries to fix stuff

[08:08:00.0000] <Philip`>
/me wonders what "pseudo-legal" really means

[08:10:00.0000] <hsivonen>
Philip`: my guess is that it means doing stuff that is not legal for a commercial entity to do in the United States but that is legal in e.g. Sweden or Hungary

[08:13:00.0000] <hsivonen>
annevk: aargh. I didn't realize there were comments, too

[08:15:00.0000] <annevk>
escapes, comments, error handling of syntax errors

[09:17:00.0000] <gsnedders>
/me probably shouldn't actually even look at the emails

[09:18:00.0000] <gsnedders>
woah.

[09:18:01.0000] <gsnedders>
102 in whatwg alone

[09:19:00.0000] <annevk>
November 2007: 110 e-mails

[09:19:01.0000] <annevk>
Decenber 2007: 208 e-mails so far

[09:30:00.0000] <gsnedders>
/me starts writing a reply

[09:30:01.0000] <gsnedders>
/me apologies and closes it

[09:31:00.0000] <Philip`>
gsnedders: If it's like your last two messages, it'll be sucked into my spam folder, so I won't even notice :-)

[09:31:01.0000] <gsnedders>
I just think there's no point.

[09:31:02.0000] <gsnedders>
There are formats with no valid patents that cover them (guaranteed, as they are too old) — the same can't be said for Theora

[09:32:00.0000] <gsnedders>
My preference is currently H.261/Vorbis (in some container, dunno what)

[09:36:00.0000] <annevk>
it's better to wait a week like I said, because  then reports from the video workshop will be in to better inform us what's going on

[09:37:00.0000] <doublec>
anyone here at the workshop right now (apart from me?)

[09:38:00.0000] <gsnedders>
annevk: see what I wrote to you last night?

[09:42:00.0000] <Philip`>
gsnedders: H.261 is limited to 176x144 and 352x288, which seems a bit rubbish

[09:42:01.0000] <annevk>
gsnedders, I think so, looked like a start

[09:42:02.0000] <gsnedders>
/me doesn't remember it being limited to specific resolutions

[09:42:03.0000] <gsnedders>
oh well.

[09:42:04.0000] <gsnedders>
MJPEG anyone? :P

[09:43:00.0000] <Philip`>
gsnedders: FFmpeg doesn't support other sizes

[09:43:01.0000] <gsnedders>
Philip`: wikipedia agrees with you

[09:43:02.0000] <Philip`>
H.263 seemingly supports the five sizes on http://en.wikipedia.org/wiki/Common_Intermediate_Format

[09:43:03.0000] <maikmerten>
MJPEG is even worse than GIF ;)

[09:43:04.0000] <Philip`>
(which all have a stupid naming convention)

[09:43:05.0000] <maikmerten>
GIF can at least say "keep this part unchanged from the last frame" ;)

[09:44:00.0000] <maikmerten>
MJPEG just codes everything, every frame

[09:44:01.0000] <Philip`>
maikmerten: But keyframes in GIF are gigantic :-)

[09:44:02.0000] <maikmerten>
H.263 is a mid-nineties standard

[09:44:03.0000] <maikmerten>
won't expire any time soon

[09:44:04.0000] <maikmerten>
Philip`, sure, GIF is horrid

[09:44:05.0000] <maikmerten>
I wasn't completely serious about GIF

[09:44:06.0000] <doublec>
maikmerten, so you suggest animated gif's as the baseline ;)

[09:45:00.0000] <maikmerten>
but GIF at least is able to exploit temporal redundancy ;)

[09:45:01.0000] <gsnedders>
I wasn't completely serious about MJPEG :P

[09:45:02.0000] <maikmerten>
doublec, well, would also save some implementation effort, right? ;)

[09:45:03.0000] <doublec>
absolutely :)

[09:45:04.0000] <maikmerten>
ah, good

[09:46:00.0000] <maikmerten>
having like 500 outdated and underperforming-till-it's-no-fun codec as baseline for sure isn't desirable ;)

[09:46:01.0000] <Philip`>
Couldn't you extend JPEG to proper 3D (2D+time, making use of redundancy in all directions) by just using a 3D DCT or something? :-)

[09:46:02.0000] <maikmerten>
Philip`, doing anything clever would again make you target to submarine patents

[09:47:00.0000] <maikmerten>
because you'd effectively develop a new codec

[09:49:00.0000] <maikmerten>
the only old old old codec I know that is really expired would be H.120 - 2 MBit/s video conferencing

[09:49:01.0000] <maikmerten>
oh joy

[09:49:02.0000] <maikmerten>
(it's from 1982)

[09:49:03.0000] <maikmerten>
(and no, I don't know a nowadays implementation)

[09:49:04.0000] <gsnedders>
H.261 at least is commonly shipped

[09:50:00.0000] <maikmerten>
1990

[09:50:01.0000] <maikmerten>
currently not old enough

[09:50:02.0000] <gsnedders>
H.261 is 1982

[09:50:03.0000] <gsnedders>
revised 1988

[09:50:04.0000] <maikmerten>
not to my knowledge

[09:50:05.0000] <gsnedders>
/me looks up

[09:50:06.0000] <Philip`>
I wonder about Bink

[09:51:00.0000] <gsnedders>
H.260 is that

[09:51:01.0000] <maikmerten>
H.261 is a 1990 ITU-T video coding standard originally designed for transmission over ISDN lines on which data rates are multiples of 64 kbit/s. It

[09:51:02.0000] <gsnedders>
H.261 is 1990

[09:51:03.0000] <maikmerten>
While H.261 was preceded in 1982 by H.120 [1][2] (which also underwent a revision in 1988 of some historic importance) as a digital video coding standard, H.261 was the first truly practical digital video coding standard (in terms of product support in significant quantities).

[09:51:04.0000] <maikmerten>
Wikipedia

[09:51:05.0000] <maikmerten>
^^ yeah, I know this is not the ultimate source

[09:51:06.0000] <maikmerten>
but H.261 in 1990 just makes sense from the history-point-of-view

[09:52:00.0000] <maikmerten>
it led to a direct line of successors to H.264

[09:52:01.0000] <gsnedders>
/me can't remember it

[09:52:02.0000] <gsnedders>
but there again, I wasn't yet alive in 1990 :P

[09:58:00.0000] <Teratogen>
bring back ogg!

[09:59:00.0000] <gsnedders>
Teratogen: _CAN YOU PLEAE MAKE A CONSTRUCTIVE COMMENT!?_

[09:59:01.0000] <Teratogen>
yes

[09:59:02.0000] <Teratogen>
BRING BACK OGG!

[09:59:03.0000] <gsnedders>
why?

[09:59:04.0000] <Teratogen>
because it's free!

[10:00:00.0000] <gsnedders>
So? What advantages does it have over, say, H.260 or Dirac?

[10:00:01.0000] <Teratogen>
freedom!

[10:00:02.0000] <maikmerten>
H.260 is oooooold beyond usefulness

[10:00:03.0000] <gsnedders>
All three are free (in terms of cost to license patents).

[10:00:04.0000] <Teratogen>
ogg totally rocks

[10:00:05.0000] <maikmerten>
Dirac is not finished yet and big players would still be scared about submarines

[10:01:00.0000] <Philip`>
If we just built web browsers on land instead of in the sea, submarines wouldn't be a problem at all

[10:01:01.0000] <gsnedders>
/me hugs Philip` 

[10:01:02.0000] <maikmerten>
I second that

[10:02:00.0000] <gsnedders>
Ogg is no help if it does not achieve interoperability between all browsers.

[10:02:01.0000] <maikmerten>
their choice.

[10:03:00.0000] <maikmerten>
there *is* no less-than-20-years old codec they'd accept

[10:03:01.0000] <gsnedders>
I don't particularly care whose choice it is. I want a video format I can use in every browser.

[10:03:02.0000] <Philip`>
gsnedders: You can use FLV

[10:04:00.0000] <gsnedders>
silly Philip`. that works in Flash, not any browser.

[10:04:01.0000] <gsnedders>
:P

[10:06:00.0000] <Philip`>
Flash works in any browser, and it works now, and in a few years it'll still work in more installed browsers than <video> even if IE8/FF3/Opera9.5/Safari4 add support

[10:06:01.0000] <gsnedders>
I know, that's true.

[10:06:02.0000] <Philip`>
and it'll support VP6, which is better than H.263

[10:06:03.0000] <gsnedders>
Philip`: it doesn't run on browsers on IA-64!

[10:08:00.0000] <doublec>
or any new hardware or devices that comes along

[10:08:01.0000] <doublec>
they'd have to rely on the flash vendor to port their software to it

[10:08:02.0000] <Philip`>
It doesn't run on Lynx either, and there's probably Lynx users than IA-64 users :-p

[10:08:03.0000] <Philip`>
s/probably/probably more/

[10:08:04.0000] <gsnedders>
Philip`: I dunno. probably more IA-64 users, but most won't run browsers on the it :P

[10:09:00.0000] <annevk>
/me wonders when the Ogg discussion stops

[10:10:00.0000] <gsnedders>
annevk: Christmas, because everyone is away :)

[10:10:01.0000] <Philip`>
I wonder how much it cost to get Flash on Opera Wii

[10:11:00.0000] <gsnedders>
/me is glad we don't require 100% consensus on everything for REC

[10:12:00.0000] <gsnedders>
http://xkcd.com/356/ — you know the worst part? I actually am now stuck thinking about that.

[10:13:00.0000] <Philip`>
gsnedders: Just do a numerical simulation :-p

[10:13:01.0000] <gsnedders>
I've moved on.

[10:13:02.0000] <gsnedders>
Better things to waste my time with.

[10:13:03.0000] <gsnedders>
(where <video> is worse)

[10:30:00.0000] <gsnedders>
hmm… vital maths test tomorrow. do I revise (i.e., learn stuff I missed when I was ill which will lead me to fail :P) or work on HTTP parsing, or write on my blog?

[11:10:00.0000] <Philip`>
gsnedders: By the way, would you be interested in information about HTTP response headers in the wild? There's some data I've got already, and some other stuff would be easy to collect, but I have no idea if it'd be useful for anything at all

[11:14:00.0000] <maikmerten>
gargh, somehow my messages to whatwg go to the moderation queue first because I subscribed as <blabla>@gmail.com but apparently Google is "correcting" the sender address to <blabla>@googlemail.com

[11:14:01.0000] <maikmerten>
is there a way to get the list accept that gmail.com "==" googlemail.com ?

[11:15:00.0000] <maikmerten>
(I really have to thank that brilliant german guy registering "GMail" as trademark and suing Google so all german users are "googlemail")

[11:16:00.0000] <gsnedders>
Philip`: yeah, sure. could you possibly drop it in an email to me?

[11:16:01.0000] <Philip`>
maikmerten: In Settings / Accounts / 'Send mail as', is that where it claims it's @gmail.com when actually it's not?

[11:17:00.0000] <maikmerten>
Philip`, I'll check

[11:17:01.0000] <maikmerten>
Philip`, wasn't aware there was a way to specify these things

[11:17:02.0000] <maikmerten>
anyway, thanks for the tip

[11:17:03.0000] <Philip`>
(The WHATWG moderation queue never gets moderated, so anything sent there will be lost)

[11:17:04.0000] <gsnedders>
wow. the email on whatwg won't stop.

[11:18:00.0000] <maikmerten>
Philip`, same policy as here at xiph.org ;)

[11:18:01.0000] <zcorpan>
oook. i've now cought up with whatwg email (mostly by marking large chunks as read)

[11:18:02.0000] <zcorpan>
not much that was interesting, actually

[11:19:00.0000] <maikmerten>
"You cannot send e-mail from maikmerten⊙gc"

[11:19:01.0000] <maikmerten>
I'll just unsubscribe and resubscribe with googlemail.com

[11:19:02.0000] <Philip`>
That sounds irritating

[11:22:00.0000] <Philip`>
I (in the UK) get a "Google Mail" logo rather than "Gmail", but it seems to be happy with my account staying as @gmail.com

[11:22:01.0000] <Philip`>
so I'm not quite sure how all this stuff works

[11:23:00.0000] <deltab>
which MLM is it?

[11:23:01.0000] <gsnedders>
My older email is @gmail.com

[11:23:02.0000] <gsnedders>
my newer is @googlemail.com

[11:24:00.0000] <maikmerten>
Philip`, may indeed be the "GMail" is a registered trademark in germany thingie

[11:24:01.0000] <gsnedders>
maikmerten: there was a dispute in the UK too

[11:24:02.0000] <maikmerten>
oh, wasn't aware of that

[11:24:03.0000] <deltab>
I think mailman supports alternate sending addresses

[11:25:00.0000] <maikmerten>
too bad I registered "AMail", "BMail" etc. but stopped at "FMail" ;)

[11:25:01.0000] <Philip`>
gsnedders: http://www.cl.cam.ac.uk/~pjt47/misc/headers.xml.bz2 is the headers from ~15K pages, as parsed by HttpClient

[11:26:00.0000] <Philip`>
(which is the only data I've got at the moment)

[11:27:00.0000] <gsnedders>
Philip`: it's served as application/xml!

[11:27:01.0000] <Philip`>
Uh, I think I don't care that it's served as ap...

[11:27:02.0000] <Philip`>
Yeah, that

[11:27:03.0000] <Philip`>
Not my web server :-p

[11:28:00.0000] <deltab>
maikmerten: instead of un/resubscribing you shoudl be able to use this: http://list.org/mailman-member/node22.html

[11:28:01.0000] <gsnedders>
/me remembers that SEE doesn't like large files

[11:28:02.0000] <maikmerten>
deltab, too late, I'm afraid.... :(

[11:28:03.0000] <maikmerten>
(and yeah, shame on me for not finding that myself)

[11:29:00.0000] <gsnedders>
Philip`: how's that done? just saving what it gives as headers?

[11:31:00.0000] <Philip`>
gsnedders: Yes, specifically via http://jakarta.apache.org/httpcomponents/httpclient-3.x/apidocs/org/apache/commons/httpclient/HttpMethod.html#getResponseHeaders()

[11:31:01.0000] <Philip`>
(i.e. using whatever kind of parsing code is provided by that)

[11:32:00.0000] <Philip`>
excluding anything that doesn't respond with 200

[11:33:00.0000] <Philip`>
and replacing control characters (except 9/A/D) with spaces

[11:34:00.0000] <Philip`>
(The output isn't necessarily grouped by uri, since it's processed multithreadedly)

[11:38:00.0000] <gsnedders>
Philip`: you got any issues with me publishing anything based on it?

[11:38:01.0000] <gsnedders>
/me assumes not seeming he's just linked to the data in a publicly logged IRC channel

[11:38:02.0000] <Philip`>
gsnedders: No - it's all from public web sites anyway, and I didn't ask them for permission ;-)

[11:43:00.0000] <gsnedders>
/me goes back to fumbling around with Python

[11:44:00.0000] <Philip`>
/me can't see an obvious way to get unparsed headers from HttpClient

[12:10:00.0000] <gsnedders>
Philip`: xml.parsers.expat.ExpatError: not well-formed (invalid token): line 14534, column 56 :(

[12:11:00.0000] <gsnedders>
Philip`: ISO-8859-1 file, with encoding undefined it seems

[12:11:01.0000] <Philip`>
gsnedders: Argh

[12:12:00.0000] <Philip`>
It was correct in my initial XML file, but it looks like xml_grep messes it up

[12:13:00.0000] <Philip`>
which is odd since it claims to be doing UTF-8

[12:13:01.0000] <hsivonen>
any recommendable JavaScript plug-in for Eclipse?

[12:14:00.0000] <roc>
The Aptana stuff is supposed to be good

[12:14:01.0000] <Philip`>
and if I set "--encoding utf-8" then it removes all linebreaks

[12:15:00.0000] <gsnedders>
/me is failing at Python

[12:17:00.0000] <gsnedders>
I'm getting a KeyError with self.headers[name] = [{"name": name, "uri": uri, "value": value}]

[12:19:00.0000] <Philip`>
gsnedders: Updated http://www.cl.cam.ac.uk/~pjt47/misc/headers.xml.bz2 so it should hopefully be utf-8 now

[12:19:01.0000] <Philip`>
Also I think I removed the <file> element

[12:20:00.0000] <gsnedders>
/me is just using getElementsByTagName() anyway

[12:20:01.0000] <Philip`>
gsnedders: I thought it only gave KeyError when trying to read a non-present key

[12:21:00.0000] <gsnedders>
Philip`: so did I.

[12:21:01.0000] <gsnedders>
"Raised when a mapping (dictionary) key is not found in the set of existing keys." — TFM

[12:21:02.0000] <Philip`>
About gEBTN: Ah, okay - I've been avoiding the DOM since my XML files are 200MB :-)

[12:22:00.0000] <gsnedders>
/me doesn't particularly care how long the scripts take to run

[12:22:01.0000] <gsnedders>
I'd be doing it in C if I did :P

[12:22:02.0000] <Philip`>
/me does care, because he's lazy and doesn't like waiting

[12:24:00.0000] <inimino>
gsnedders: self.headers['name'] perhaps?

[12:24:01.0000] <gsnedders>
inimino: no, name is a variable

[12:25:00.0000] <inimino>
oh, ok

[12:25:01.0000] <gsnedders>
(and set)

[12:25:02.0000] <gsnedders>
odd.

[12:25:03.0000] <gsnedders>
now with the new headers.xml it works

[12:25:04.0000] <Philip`>
That sounds impossible

[12:26:00.0000] <gsnedders>
totally real, though.

[12:26:01.0000] <gsnedders>
I love software :P

[12:27:00.0000] <Philip`>
I love only software that doesn't involve character encodings

[12:27:01.0000] <gsnedders>
Philip`: like what? :P

[12:27:02.0000] <Philip`>
Like anything that just uses integers :-)

[12:27:03.0000] <inimino>
Philip`: but not as character codes?

[12:28:00.0000] <gsnedders>
Philip`: but output? :P

[12:28:01.0000] <gsnedders>
Philip`: how do you encode those integers for display?

[12:28:02.0000] <Philip`>
inimino: Using integers for character codes is okay, as long as you never have to interpret those integers as characters :-)

[12:29:00.0000] <Philip`>
gsnedders: If you're only outputting and never inputting, then you don't have to care about encoding errors, because you just shove stuff through printf() and if the user gets garbage then it's their problem for not using ASCII

[12:29:01.0000] <gsnedders>
:D

[12:29:02.0000] <inimino>
heh

[12:32:00.0000] <Philip`>
/me 's current work's only user interface is via Telnet outputting to a serial console in a virtual machine connected to the host through UDP and then passed through Perl and Python scripts

[12:32:01.0000] <Philip`>
so there's pretty much no chance of me getting character encodings straight, so I'm just sticking with ASCII there to save myself the pain

[12:33:00.0000] <Philip`>
(Actually, it's a Telnet server and a netcat client, so all the magic Telnet commands get printed out to the screen)

[12:35:00.0000] <inimino>
/me guesses there is a story behind doing it that way

[12:37:00.0000] <hsivonen>
roc: thanks. installing now

[12:37:01.0000] <Philip`>
inimino: I'm testing some networking stuff, so it has to be done in VMs, and then that's the best way I've found to collect the output from them

[12:59:00.0000] <hsivonen>
annevk: where does HTML5 allow target='' on <form>?

[13:12:00.0000] <gsnedders>
Philip`: ValueError: too many values to unpack — your file is too big for what I want :P

[13:13:00.0000] <Philip`>
gsnedders: Uh, that sounds like an odd reason to get ValueError

[13:13:01.0000] <gsnedders>
Philip`: well, there are over ~15k items in the dictionary I'm trying to iterate over

[13:14:00.0000] <Philip`>
I don't see why that would be a problem

[13:15:00.0000] <gsnedders>
Philip`: and your saying of ~15k is wrong. ~115k would've been closer :) (116945 FYI)

[13:16:00.0000] <Philip`>
It's ~15K unique documents, mostly with >1 header each

[13:16:01.0000] <gsnedders>
ah. ~15k documents.

[13:16:02.0000] <gsnedders>
that is what you said, actually.

[13:19:00.0000] <gsnedders>
This is starting to get really annoying.

[13:19:01.0000] <gsnedders>
http://mail.python.org/pipermail/python-list/2006-June/387414.html

[13:19:02.0000] <gsnedders>
hmm

[13:20:00.0000] <gsnedders>
the value of the dictionary is a list

[13:20:01.0000] <Philip`>
What do you mean by "the value of"?

[13:21:00.0000] <gsnedders>
the value of every key is a list.

[13:22:00.0000] <Philip`>
The bit where you said [{"name": ...}] is a list because of the []

[13:22:01.0000] <Philip`>
or do you mean something else?

[13:23:00.0000] <gsnedders>
{"foo": "bar"} where "foo" is the key and "bar" the value

[13:24:00.0000] <Philip`>
/me doesn't understand the problem

[13:25:00.0000] <gsnedders>
/me now understanding the problem writes a tiny exemplar

[13:26:00.0000] <gsnedders>
Philip`: http://pastebin.ca/813896 — make that work.

[13:27:00.0000] <hsivonen>
Lachy__: http://html5.lachy.id.au/ could be better if the form was seeded with an HTML5 skeleton document

[13:27:01.0000] <Philip`>
gsnedders: foo.items()

[13:29:00.0000] <Philip`>
because "for k in foo" iterates over keys, whereas "for (k, v) in foo.items()" iterates over (key, value) pairs

[13:29:01.0000] <gsnedders>
and for k, v in foo?

[13:29:02.0000] <Philip`>
(http://docs.python.org/lib/typesmapping.html shows most of the useful functions)

[13:30:00.0000] <hsivonen>
Lachy__: the validate html5 button works only once for me on http://html5.lachy.id.au/ in Firefox 2

[13:30:01.0000] <hsivonen>
Lachy__: I have to reload to make the button work again

[13:30:02.0000] <Philip`>
gsnedders: That will iterates over keys, and try to unpack each key into a (k,v) tuple, which will raise ValueError because your keys are strings and can't be unpacked

[13:30:03.0000] <gsnedders>
ah

[13:31:00.0000] <Philip`>
(It's like "for x in foo: k, v = x")

[13:31:01.0000] <Philip`>
(in terms of iterating over keys rather than keys+values)

[13:31:02.0000] <gsnedders>
ah

[13:31:03.0000] <gsnedders>
/me doesn't pretend to be anything but a python n00b

[13:33:00.0000] <Philip`>
It mostly makes sense when you can see which rules apply - it's not like Perl where magical context-sensitive things happen and you'll never understand unless you read that particular detail in the documentation and newsgroups :-)

[13:36:00.0000] <Philip`>
Hmm, I want to watch some video on the web tomorrow, but it's streaming Windows Media and I'm not sure how to handle that

[13:36:01.0000] <Lachy__>
hsivonen, the validate button works for me all the time, without reloading

[13:38:00.0000] <Lachy__>
hsivonen, send me an email about adding a template document to it, and I'll see what I can do when I get back from Linkoping on Sunday

[13:39:00.0000] <hsivonen>
Lachy__: ok

[13:40:00.0000] <gsnedders>
/me has kinda given up hope at actually passing the maths test tomorrow, having missed a couple of weeks, and barely knowing one section

[13:41:00.0000] <Lachy>
gsnedders, maths isn't too hard

[13:41:01.0000] <Lachy>
which particular maths topics are you covering this year?

[13:43:00.0000] <gsnedders>
Lachy: http://www.sqa.org.uk/files/nq/C10012.pdf (which I think gives enough detail :P)

[13:43:01.0000] <Philip`>
Aha, mplayer works

[13:44:00.0000] <Philip`>
/me wonders if anyone happens to know how to record and watch a stream simultaneously

[13:44:01.0000] <gsnedders>
can anyone explain the reasoning behind there being both plus/minus and minus/plus signs?

[13:45:00.0000] <gsnedders>
Lachy: it's the stuff on page 17 of the PDF that I've missed

[13:45:01.0000] <Philip`>
gsnedders: They're useful for e.g. "+/- x = - (-/+ x)"

[13:46:00.0000] <Philip`>
i.e. representing two versions of the equation, where one has a mixture of + and -, and the other has the +s and -s flipped

[13:46:01.0000] <gsnedders>
/me doesn't see how that helps

[13:46:02.0000] <Lachy>
gsnedders, on which page can I find one of these minus-plus signs?

[13:46:03.0000] <Lachy>
I've never seen one before

[13:47:00.0000] <Philip`>
You can't say "+/- x = - (+/- x)" because that would be interpreted as "+x = -(+x) and -x = -(-x)" which is untrue

[13:47:01.0000] <gsnedders>
Lachy: the same page (p17 of the PDF, labelled within itself as p16)

[13:47:02.0000] <gsnedders>
Philip`: ah. so then you have to take both from the same side!

[13:47:03.0000] <Philip`>
In that cos example, it means "cos(A+B) = blah - blah; and cos(A-B) = blah + blah"

[13:48:00.0000] <gsnedders>
yeah, that makes sense now

[13:48:01.0000] <Philip`>
gsnedders: Yep

[13:48:02.0000] <gsnedders>
(I've just about done the basics of the top two formula)

[13:48:03.0000] <gsnedders>
and obviously the third is just rearranged

[13:48:04.0000] <Philip`>
(though sometimes +/- and -/+ are used in a context-sensitive way and don't actually work like that :-) )

[13:48:05.0000] <gsnedders>
(or rather, a rearranged copy of the above)

[13:49:00.0000] <Philip`>
The third/fourth are just taking A=B

[13:49:01.0000] <gsnedders>
Philip`: 15332 pages of HTTP headers, BTW

[13:50:00.0000] <gsnedders>
Philip`: I understand the first equality on the forth, but not the second/third equalities

[13:51:00.0000] <gsnedders>
Philip`: are all the headers from accessing the page once, or not?

[13:52:00.0000] <Philip`>
cos^2 x + sin^2 x = 1

[13:52:01.0000] <Philip`>
...is the relevant fact that should be known

[13:52:02.0000] <gsnedders>
that more or less makes sense from a graph, yeah.

[13:52:03.0000] <Philip`>
so that gives the cos^2 A - sin^2 A = 2cos^2 A - 1 and suchlike

[13:54:00.0000] <Lachy>
there was a time about 10 years ago where I would have been able to do those trig equations. Now I just stare at them blankly

[13:54:01.0000] <jgraham>
IRC clearly needs better support for maths notation

[13:54:02.0000] <gsnedders>
Philip`: LaTeX!

[13:54:03.0000] <roc>
IRC should be XML

[13:54:04.0000] <roc>
then we could post our HTML examples directly

[13:54:05.0000] <roc>
and use MathML

[13:55:00.0000] <gsnedders>
peh. start with the basics. get a universally accepted character encoding on IRC :)

[13:55:01.0000] <Philip`>
gsnedders: Each page was GETed once, but the <header uri> in the output is the result of redirections, so it's possible that some pages redirected to the same location

[13:55:02.0000] <gsnedders>
kk

[13:55:03.0000] <roc>
and play XSS pranks on each otehr

[13:55:04.0000] <Hixie>
lordy what a lot of main

[13:55:05.0000] <Lachy>
UTF-8 seems to be fairly widely accepted for IRC these days

[13:55:06.0000] <gsnedders>
Philip`: there's a page that claims to have 94 headers

[13:55:07.0000] <Philip`>
gsnedders: Maybe it'd be more helpful if I gave the original unique requested URI instead of the redirected result?

[13:55:08.0000] <Hixie>
half of this video mail has neither the word "video" nor the word "ogg" in it

[13:55:09.0000] <Hixie>
sheesh

[13:55:10.0000] <roc>
fear the wrath of Ogg!

[13:56:00.0000] <gsnedders>
Philip`: give the original URI for each request (i.e., a redirect has a different URI)

[13:56:01.0000] <jgraham>
Hixie: That's to make it hard to automatically redirect to /dev/null ;)

[13:56:02.0000] <Lachy>
I can't believe the whole ogg discussion is still going on, on far too many different lists

[13:57:00.0000] <Hixie>
what i'm amused by is that for every person sending 10 flames to one of the lists, i get a person e-mailing me privately telling me that they have my support and that they believe we're doing the right thing

[13:57:01.0000] <gsnedders>
I'm not totally sure whether it was the right thing to do.

[13:57:02.0000] <gsnedders>
Though I'm sure plenty of the people on the mailing list think I agree with you :)

[13:58:00.0000] <hsivonen>
/me finally replies to an ogg email

[13:58:01.0000] <gsnedders>
Philip`: least headers is 3

[13:59:00.0000] <hsivonen>
evidently, rudd-o hadn't read the spec before he started his slashdot campaign

[13:59:01.0000] <gsnedders>
Philip`: median is 7, mean is 8

[13:59:02.0000] <Hixie>
hah, the first comment on http://digg.com/tech_news/Nokia_and_Apple_seem_to_have_succeeded_in_suppressing_ogg is a complete non-sequitur

[13:59:03.0000] <Hixie>
hsivonen: shocking, that

[13:59:04.0000] <gsnedders>
hsivonen: why bother? it's only another damned technical document!

[13:59:05.0000] <gsnedders>
(on a totally unrelated note, I updated the to-do list on the tolerant http parsing spec)

[14:00:00.0000] <Lachy>
oh wow, I never expected accessibility to come up in the discussion: [whatwg] HTML 5, OGG, competition, civil rights,  and persons with disabilities

[14:01:00.0000] <hsivonen>
Lachy: I must have skipped that message.

[14:01:01.0000] <gsnedders>
(that's <http://stuff.gsnedders.com/draft-sneddon-http-parsing-00.html> or .txt)

[14:01:02.0000] <Philip`>
gsnedders: If I give the original URI, it's still going to return the final after-redirection request's headers, so if several URIs redirect to the same place then it'll repeat the redirection target's headers

[14:01:03.0000] <Philip`>
which isn't necessarily bad, but it's something to be aware of

[14:01:04.0000] <gsnedders>
Philip`: ergh.

[14:05:00.0000] <Philip`>
/me wonders how to select elements of type A or type B using the subset of XPath supported by xml_grep

[14:06:00.0000] <Hixie>
i think i have found an easy way to achieve my goals of replying to hundreds of e-mails by month's end

[14:06:01.0000] <Philip`>
Oh, looks like I can't do that

[14:10:00.0000] <Philip`>
gsnedders: http://www.cl.cam.ac.uk/~pjt47/misc/headers2.xml.bz2 has the original request URI, and some <redirect>s to point out what got redirected

[14:10:01.0000] <gsnedders>
/me just commited into hg the entire XML file

[14:11:00.0000] <Philip`>
The new XML file is indented differently, just to make fun diffs

[14:12:00.0000] <gsnedders>
what is the <redirect> element? just noting movement?

[14:12:01.0000] <Philip`>
Yes - it's added whenever the request URI and response URI differ

[14:12:02.0000] <gsnedders>
pay any attention to how many redirects it has?

[14:12:03.0000] <Philip`>
(i.e. when HttpClient did whatever magical redirection-handling it does)

[14:12:04.0000] <Philip`>
It has less than 100 redirects, but that's all I know

[14:13:00.0000] <Philip`>
(because otherwise it throws an exception and aborts)

[14:13:01.0000] <hsivonen>
Philip`: did you write your own spider based on HttpClient and the Validator.nu parser?

[14:13:02.0000] <gsnedders>
Philip`: different root element, too

[14:14:00.0000] <Philip`>
gsnedders: Yes, but you said you were using getElementsByWhatever so I assumed that wouldn't matter, and I used grep/echo/cat instead of xml_grep to extract the bits from my original XML file

[14:14:01.0000] <gsnedders>
Philip`: yeah, just an observation :)

[14:14:02.0000] <Philip`>
which is totally not the right way to do it :-)

[14:14:03.0000] <gsnedders>
http://hg.gsnedders.com/cgi-bin/hgwebdir.cgi/http-parsing/file/96df15d57efb/Philip%20Taylor%27s%20Header%20Data/README.txt — that all right?

[14:15:00.0000] <hsivonen>
Philip`: is the code that you are using to drive HttpClient in SVN somewhere?

[14:15:01.0000] <Philip`>
hsivonen: I don't think it's a spider since it doesn't follow links at all, but I did write my own thing to download/analyse a list of HTML files using HttpClient and the Validator.nu parser

[14:15:02.0000] <hsivonen>
Philip`: ok

[14:15:03.0000] <Philip`>
hsivonen: It isn't at the moment

[14:16:00.0000] <hsivonen>
Philip`: surely at that point you could make links from the parse tree feed back into the download list

[14:16:01.0000] <hsivonen>
although it probably isn't that simple

[14:17:00.0000] <Philip`>
hsivonen: I'm not sure exactly what you mean

[14:17:01.0000] <Philip`>
gsnedders: The last paragraph doesn't really make any sense :-)

[14:17:02.0000] <hsivonen>
If you analyse docs, presumably the contain links and those could be put on the list to download/analyse

[14:18:00.0000] <hsivonen>
but then there's robots.txt

[14:18:01.0000] <gsnedders>
Philip`: that's true, but I only did it quickly

[14:18:02.0000] <Philip`>
hsivonen: Ah, yes

[14:18:03.0000] <hsivonen>
and crawling in a reasonable breadth-first order etc, etc

[14:18:04.0000] <Philip`>
hsivonen: I'd prefer to use someone else's code rather than do all that work

[14:19:00.0000] <Philip`>
(I'm not even looking at robots.txt now, since that would double the number of requests I make)

[14:19:01.0000] <gsnedders>
Philip`: "It may not be grouped by URI fully, as it is not processed by a single thread"?

[14:19:02.0000] <Philip`>
gsnedders: Is that sentence needed at all?

[14:20:00.0000] <gsnedders>
Philip`: not really, but I may as well put it there in case anyone ever cares.

[14:20:01.0000] <Philip`>
gsnedders: It just means it might have <header uri=a/><header uri=b/><header uri=a/>, which isn't an extremely interesting observation

[14:20:02.0000] <hsivonen>
Philip`: the Internet Archive spider has the kitchen sink in it but seems to be picky about its execution environment according to docs

[14:21:00.0000] <hsivonen>
Philip`: also, the code base isn't particularly approachable due to the kitchen sink nature

[14:21:01.0000] <gsnedders>
Philip`: it means you can't do anything that assumes it's in order, which you might sometimes want to do

[14:23:00.0000] <Philip`>
hsivonen: Hmm, it does sound not entirely trivial

[14:23:01.0000] <Philip`>
I'm not sure how worthwhile it would be to do actual spidering, rather than sticking with the dmoz.org list

[14:24:00.0000] <hsivonen>
Philip`: isn't dmoz biased towards front pages?

[14:24:01.0000] <Philip`>
(particularly since I can't do especially extensive spidering - I'd prefer not to be making a hundred thousand requests, because it's kind of expensive in bandwidth)

[14:25:00.0000] <Philip`>
hsivonen: Yes, and to CNN

[14:25:01.0000] <hsivonen>
also, how alive is dmoz these days? does it represent current authoring?

[14:25:02.0000] <Philip`>
/me has no idea

[14:26:00.0000] <Philip`>
I can imagine getting much worse results from a spider that gets sucked into a single giant site, so I'm not sure how to make things definitely better

[14:27:00.0000] <Philip`>
(I'm not even sure what "better" means)

[14:29:00.0000] <hsivonen>
Philip`: knowledge about web site structures would probably be needed to make reasonable guesses

[14:30:00.0000] <hsivonen>
Philip`: without data I might guess that a sensible strategy would be taking a list of site roots, analyzing the front page, picking two site-internal links at random, analyzing those pages too and following one site-internal link from each of those

[14:30:01.0000] <hsivonen>
that would give front page plus 4 non-front pages for each site

[14:43:00.0000] <Philip`>
hsivonen: It would be good to have a way of evaluating the strategies, to see which ones actually work sensibly in practice, but I've got no idea how to do that either :-/

[14:51:00.0000] <tndH>
ooh, acronym/initialism debate again

[14:51:01.0000] <tndH>
feels nice to read that after all the ogg stuff

[14:53:00.0000] <tndH>
/me pronounces it "huttamerl 5", for what it's worth

[14:55:00.0000] <anne-mac>
evening

[14:57:00.0000] <roc>
you know what would be an interesting way out of the codec mess?

[14:57:01.0000] <anne-mac>
ooh, codecs

[14:57:02.0000] <anne-mac>
/me hides

[14:57:03.0000] <roc>
a GPLed H.264 implementation with a patent license grant for just that implementation (and derivatives)

[14:59:00.0000] <Philip`>
<video src="http://transcode.google.com/?fmt=h264;src=http://mycheaplyhostedsite.com/somevideo">

[14:59:01.0000] <anne-mac>
i don't want to get involved, but since this sound interesting, would that work cross-platform?

[15:00:00.0000] <doublec>
x264 sis a GPLed H.264 implementation

[15:00:01.0000] <Philip`>
roc: What if somebody took that implementation and created a derivative by deleting all the code and then pasting their own unrelated code into it, to get the patent licence?

[15:00:02.0000] <doublec>
but obviously would need the patent license grant

[15:00:03.0000] <Philip`>
s/to get/to get around/

[15:01:00.0000] <roc>
Philip`: that doesn't really matter. The main goal is to ensure that the patent license grant only applies to GPLed software

[15:01:01.0000] <roc>
so if they claim it's a derivative, then they have to keep the GPL too. They can't have it both ways

[15:01:02.0000] <Philip`>
roc: Ah, okay

[15:01:03.0000] <Philip`>
roc: I imagine Opera wouldn't like that so much :-)

[15:01:04.0000] <roc>
yeah well

[15:01:05.0000] <roc>
Mozilla wouldn't like it very much eitehr

[15:01:06.0000] <roc>
but it might be acceptable

[15:01:07.0000] <roc>
maybe

[15:02:00.0000] <roc>
the real main goal, of course, would be to exploit GPL virality to ensure that the patent grant would only apply to free software

[15:03:00.0000] <roc>
similar to the way Qt is GPLed so people writing closed source apps with it have to pay money

[15:03:01.0000] <anne-mac>
hmm, that doesn't sound very useful to us indeed

[15:04:00.0000] <anne-mac>
or content providers

[15:04:01.0000] <roc>
anne-mac: yeah, sorry, it's far from ideal

[15:05:00.0000] <Philip`>
What would the MPEG-LA get out of this?

[15:05:01.0000] <Philip`>
(I assume they don't want to just be nice to people)

[15:06:00.0000] <roc>
more players for H.264 vs VC-1

[15:06:01.0000] <Philip`>
(Possible answer: They would get money from content providers who are producing content in that format)

[15:06:02.0000] <roc>
which would appeal to some of the MPEG-LA members, I think

[15:07:00.0000] <Philip`>
(but that means content providers would have to be paying)

[15:07:01.0000] <roc>
yeah, I don't know about that situation

[15:07:02.0000] <Philip`>
((though small-scale content providers (e.g. bloggers) could just use MEncoder and not care about patents, like they do already with XviD and everything))

[15:07:03.0000] <roc>
but content providers may be a separate battle

[15:07:04.0000] <roc>
the patents might not even be the same

[15:09:00.0000] <roc>
doublec: aren't you at the workshop right now?

[15:22:00.0000] <doublec>
roc, yes

[15:22:01.0000] <roc>
ok

[15:22:02.0000] <doublec>
oddly someone representing license holders for a codec just spoke to me about something similar-ish

[15:22:03.0000] <roc>
interesting

[15:23:00.0000] <roc>
you should be doing that sort of thing instead of being on IRC :-)

[15:23:01.0000] <doublec>
so at least some people are interested in discussing options

[15:23:02.0000] <doublec>
:)

[15:23:03.0000] <Hixie>
/me wishes doublec the best of luck, really, and truly hopes doublec finds a solution

[15:23:04.0000] <roc>
give us a full report when you get back

[15:23:05.0000] <Hixie>
get together with dave singer

[15:23:06.0000] <doublec>
will do

[15:23:07.0000] <Hixie>
he's very interested in helping out too

[15:24:00.0000] <doublec>
i'll know more later, and will pass details on to Dave as well

[15:27:00.0000] <Hixie>
cool

[15:27:01.0000] <Hixie>
roc: re your mail, you are of course correct, i just meant given the current landscape

[15:27:02.0000] <Hixie>
roc: one way of forcing the issue would be to make there be a lot of theora content.

[15:27:03.0000] <roc>
yeah

[15:29:00.0000] <roc>
unfortunately my best idea there is "free pornography community"

[15:31:00.0000] <Hixie>
not necessarily a bad idea actually

[15:33:00.0000] <anne-mac>
hmm google blog search considers rudd-o's blog post to be the most relevant for "html5"

[15:34:00.0000] <anne-mac>
what is interesting is the amount of non-English content appearing on google blog search

[15:35:00.0000] <anne-mac>
it would be nice if we could somehow get in touch with all those people (or maybe we already are)

[15:46:00.0000] <Lachy>
I like this. Accusing Microsoft of using HTML5 as a distraction from their other problems with IE. http://inspireaction.mindandmedia.com/index.php/2007/12/11/why-is-microsoft-trying-to-distract-us-with-html-5/

[15:52:00.0000] <jruderman>
"force Microsoft to play by the rules"

[15:52:01.0000] <jruderman>
if he wants to attempt that, he's more than welcome to

[15:53:00.0000] <jruderman>
the rest of us are going to keep trying to improve the web

[15:55:00.0000] <anne-mac>
/me thinks it would make sense for <th>/<caption>/<legend>/<h1>-<h6> to share the content model

[15:59:00.0000] <anne-mac>
seems I can't edit though... neither Opera or Safari is supported by Google Docs :(


2007-12-13
[16:09:00.0000] <jruderman>
use firefox?

[16:09:01.0000] <anne-mac>
i guess

[16:10:00.0000] <Philip`>
Or IE?

[16:10:01.0000] <anne-mac>
i was trying to get away with using just Opera on my Mac but Google is against me

[16:11:00.0000] <anne-mac>
(on my other laptop I'm trying to get away with not having Flash, which works reasonably so far)

[16:11:01.0000] <anne-mac>
(but this one had it by default, so I guess I'm cheating)

[16:12:00.0000] <Philip`>
(You should get another one and set the firewall to block port 80, to complete your attempts at disabling significant portions of the web :-) )

[16:13:00.0000] <anne-mac>
I'm not convinced that's quite the same

[16:13:01.0000] <anne-mac>
:)

[16:14:00.0000] <Philip`>
/me doesn't do much unusual except disabling referer, and has only noticed that breaking about one site in the past several months

[16:16:00.0000] <anne-mac>
"Nokia and Apple have successfully pressured the WC3 board into dropping Ogg support from the HTML5 spec. Sheesh."

[16:16:01.0000] <Hixie>
i'm on the WC3 board? sweet!

[16:16:02.0000] <Hixie>
what's the WC3 board?

[16:16:03.0000] <Philip`>
A dartboard?

[16:17:00.0000] <Philip`>
I'm not sure what Warcraft 3 has to do with this, though

[16:17:01.0000] <anne-mac>
holly crap

[16:18:00.0000] <anne-mac>
Krzysztof e-mails way too much

[16:18:01.0000] <Hixie>
i've seen worse

[16:18:02.0000] <Hixie>
but yeah

[16:19:00.0000] <Philip`>
Fortunately I'm protected from the onslaught by a 2.5 hour email lag again

[16:19:01.0000] <Hixie>
heh

[16:21:00.0000] <anne-mac>
hmm, WCAG 2.0 in LC

[16:37:00.0000] <MikeSmith>
Is there any kind of whatwg timeline/history at the whatwg wiki?

[16:42:00.0000] <Hixie>
yeah

[16:42:01.0000] <Hixie>
look on the faq

[16:42:02.0000] <Hixie>
it's one of the questions

[16:43:00.0000] <Hixie>
also, http://lists.w3.org/Archives/Public/www-archive/2006Nov/0045.html

[16:46:00.0000] <MikeSmith>
Hixie - thanks

[16:48:00.0000] <MikeSmith>
about timeline, what I was looking for was something that showed when whawg first published Web Forms 2.0, when first Web Apps was published

[16:48:01.0000] <MikeSmith>
that kind of stuf

[16:49:00.0000] <Hixie>
oh

[16:49:01.0000] <MikeSmith>
I been thinking about adding it to the timeline on the HTML WG home page

[16:49:02.0000] <Hixie>
first draft of webforms 2, then known as xforms basic, came out in october/november 2003, just after the xforms PR went to vote

[16:49:03.0000] <MikeSmith>
or at least linking to something

[16:49:04.0000] <MikeSmith>
/me nods

[16:50:00.0000] <MikeSmith>
how about first Web Apps?

[16:50:01.0000] <Hixie>
whatwg itself (along with first draft of wa1, now known as html5) came out the week after the adobe meeting

[16:50:02.0000] <Hixie>
march 2004?

[16:50:03.0000] <Hixie>
about then

[16:50:04.0000] <MikeSmith>
OK

[16:50:05.0000] <Hixie>
my blog commented on it all, if you want exact dates

[16:50:06.0000] <Lachy>
MikeSmith, some of that info is listed here http://www.whatwg.org/specs/

[16:50:07.0000] <MikeSmith>
Lachy - thanks

[16:51:00.0000] <MikeSmith>
Hixie - yeah, I want to go back and read through your blog archive, dbaron's, some others

[16:51:01.0000] <anne-mac>
Web Forms 2.0 has all the dated links

[16:51:02.0000] <anne-mac>
in the draft

[16:51:03.0000] <MikeSmith>
anne-mac - thanks

[16:52:00.0000] <anne-mac>
first (W3C Member-only): http://lists.w3.org/Archives/Member/w3c-archive/2003Sep/att-0014/hfp.html

[16:52:01.0000] <MikeSmith>
I remember dbaron once posted a summary about his perspective on the adobe meeting and the results of it

[16:53:00.0000] <Hixie>
wow, september

[16:53:01.0000] <Hixie>
that's even earlier than i remember

[16:53:02.0000] <dbaron>
This one?  http://dbaron.org/log/2004-06#e20040609a

[16:54:00.0000] <dbaron>
I revised my opinions a bit later on... http://dbaron.org/log/2006-08#e20060818a

[16:54:01.0000] <MikeSmith>
dbaron - yep, thanks much

[16:54:02.0000] <Hixie>
my early blog entries on the matter are quite amusing in retrospect

[16:54:03.0000] <Hixie>
since they mention the whatwg before it was announced

[16:54:04.0000] <Hixie>
but without saying so :_)

[16:54:05.0000] <MikeSmith>
dbaron - thanks, I remember reading that one too

[16:55:00.0000] <dbaron>
Hixie, but you did that intentionally...

[16:55:01.0000] <Hixie>
yes

[16:56:00.0000] <Hixie>
i am easily amused :-P

[16:58:00.0000] <MikeSmith>
I'm hoping to have some time later today to write up draft timeline of the events from the adobe workshop up until the announcement of the current HTML WG

[16:58:01.0000] <Lachy>
what was the adobe workshop?

[16:59:00.0000] <Hixie>
the web apps workship

[16:59:01.0000] <Hixie>
shop

[16:59:02.0000] <MikeSmith>
Lachy - http://www.w3.org/2004/04/webapps-cdf-ws/ 

[17:00:00.0000] <Lachy>
ah, is that the one after which the whatwg formed?

[17:00:01.0000] <Hixie>
it's the one after which the whatwg was announced

[17:00:02.0000] <Hixie>
we actually had the whatwg ready to go at least a month earlier but we wanted to give the w3c a chance

[17:01:00.0000] <MikeSmith>
Lachy - dbaron blogged just after it at http://dbaron.org/log/2004-06#e20040607a 

[17:01:01.0000] <MikeSmith>
Brendan Eich at http://weblogs.mozillazine.org/roadmap/archives/005632.html

[17:01:02.0000] <Hixie>
i also blogged many detailed blogs on it both during and after

[17:02:00.0000] <dbaron>
about a month and a half before, looks like

[17:02:01.0000] <Hixie>
based on registration date?

[17:02:02.0000] <Hixie>
of the domain?

[17:02:03.0000] <Hixie>
sounds about right

[17:03:00.0000] <MikeSmith>
I guess the following is the minutes of the workshop

[17:03:01.0000] <MikeSmith>
http://lists.w3.org/Archives/Public/public-webapps-cdf-discuss/2004Jun/att-0000/2004jun01.html

[17:03:02.0000] <dbaron>
Hixie, no, based on email threads I have

[17:04:00.0000] <MikeSmith>
or maybe above is just position statements, I dunno

[17:04:01.0000] <MikeSmith>
hmm, no, I guess that's the minutes

[17:05:00.0000] <Hixie>
dbaron: ah

[17:05:01.0000] <MikeSmith>
anyway, I will add whatever I come up with to the http://esw.w3.org/topic/HTML/ wiki

[17:06:00.0000] <MikeSmith>
but if somebody else has time and interest in doing it before I get to it, feel free

[17:06:01.0000] <MikeSmith>
either there or at the whatwg wiki

[17:07:00.0000] <MikeSmith>
I guess if we really wanted to be ambitious, it could go back to events since publication of HTML 4.0 or 4.01 rec

[17:08:00.0000] <Hixie>
there aren't many between that and wf2 :-P

[17:08:01.0000] <Hixie>
that was the problem :-P

[17:11:00.0000] <MikeSmith>
I guess showing that there weren't HTML milestones during those years could be another part of the usefulness of it

[17:11:01.0000] <MikeSmith>
or showing what else was going on instead

[17:17:00.0000] <Hixie>
heh

[17:18:00.0000] <Hixie>
wow, globalStorage is already on more sites than <t:video>

[17:19:00.0000] <Hixie>
including on http://maps.live.com/localsearch/

[17:19:01.0000] <Hixie>
!!

[17:19:02.0000] <Hixie>
microsoft use html5!

[17:23:00.0000] <gavin>
"<i>What working group</i> is going to work on extending HTML..." :)

[17:25:00.0000] <gavin>
(from http://ln.hixie.ch/?start=1086387609&count=1 )

[17:26:00.0000] <Hixie>
hey sweet, there's a book that mentions globalStorage: http://safari.adobepress.com/9780132242066

[17:37:00.0000] <roc>
wow!!!

[17:37:01.0000] <othermaciej_>
Hixie: that's funny

[17:38:00.0000] <Hixie>
so it seems that we will break in the region of 200 sites by simplifying globalStorage

[17:39:00.0000] <Hixie>
we can work around one of the problems by just making globalStorage[domain] === globalStorage

[17:39:01.0000] <Hixie>
(just in firefox)

[17:39:02.0000] <Hixie>
but it seems that there is a bigger problem

[17:39:03.0000] <Hixie>
which is that people are using the .value accessor

[17:39:04.0000] <Hixie>
which i wasn't expecting

[18:15:00.0000] <Hixie>
<cite><audio>...</audio></cite> could be valid i guess... right...?

[00:05:00.0000] <hsivonen>
Hixie: what was "the Adobe meeting"?

[00:06:00.0000] <hsivonen>
canceling the question. noticed Lachy asked it too

[00:15:00.0000] <Hixie>
heh

[00:17:00.0000] <Hixie>
can anyone think of a use case for <address> <blockquote> ... ?

[00:18:00.0000] <hsivonen>
Hixie: indented contact info

[00:19:00.0000] <Hixie>
-_-

[00:20:00.0000] <Hixie>
<dialog> <dt> Fred <dd> <audio> makes sense

[00:20:01.0000] <Hixie>
but what about <dialog> <dt> <audio> ?

[00:22:00.0000] <Hixie>
i guess we should allow   <sup><audio/></sup> though i'm hard pressed to come up with any valid use cases for that either

[00:23:00.0000] <hsivonen>
Hixie: what's the worse case scenario if we do allow those?

[00:23:01.0000] <Hixie>
people make mistakes that aren't caught by the validators, i think

[00:24:00.0000] <Hixie>
i really need a third option, "yes", "no", and "dumb"

[00:26:00.0000] <Hixie>
/me removes ins and del from the table since really they're orthogonal to the block/inline thing

[00:27:00.0000] <hsivonen>
what's the right way to abstract away createElement vs. createElementNS these days?

[00:28:00.0000] <hsivonen>
/me is reviewing and integrating zcorpan's message collapsing script

[00:32:00.0000] <Hixie>
abstract?

[00:38:00.0000] <hsivonen>
Hixie: zcorpan wrote a script for the validator.nu UI that collapses duplicate instances of the same error message

[00:38:01.0000] <hsivonen>
I want UI scripts to work with &out=xhtml as well

[00:38:02.0000] <hsivonen>
or if that's too hard, I want to turn them off properly instead of letting them fail in an ugly way

[00:43:00.0000] <Hixie>
ah

[00:43:01.0000] <Hixie>
the html5 way is to just use createElement or to use createElementNS with the xhtml ns

[00:43:02.0000] <Hixie>
the way that actually works...

[00:43:03.0000] <Hixie>
you need to use createElement in text/html, for IE

[00:43:04.0000] <Hixie>
but createElementNS() with the XHTML ns will work everywhere, i believe

[00:44:00.0000] <Hixie>
everywhere else, i mean

[00:44:01.0000] <Hixie>
even text/html (though you'll be creating the "xhtml" version of the nodes)

[00:44:02.0000] <hsivonen>
Hixie: and presumably, IE doesn't have createElementNS?

[00:44:03.0000] <Hixie>
(or some such weirdness)

[00:44:04.0000] <Hixie>
correct

[00:44:05.0000] <hsivonen>
so I could do if (!document.createElementNS) {

[00:45:00.0000] <hsivonen>
document.createElementNS = function() {...

[00:45:01.0000] <Hixie>
that might work

[00:46:00.0000] <hsivonen>
of course, I don't have IE to test with, so backporting to IE after Firefox/Opera/Safari work is at the bottom of the list...

[00:47:00.0000] <hsivonen>
perhaps I live in a reality distortion field, but having easy access only to Mac/Ubuntu is not unique to me

[00:48:00.0000] <Hixie>
it's becoming much more common

[00:48:01.0000] <Hixie>
i wonder what <datagrid>'s content model should be

[00:48:02.0000] <hsivonen>
when I gave the talk about HTML5 a week and a half ago, the audience seemed to be an all Mac/Ubuntu crowd

[00:49:00.0000] <hsivonen>
and one person commented that he doesn't test in IE because it doesn't run on his systems

[00:49:01.0000] <Hixie>
hehe

[00:49:02.0000] <Hixie>
it's a good trend. i just hope that it stops at 20%

[00:49:03.0000] <Hixie>
i wouldn't want IE's monopoly to be replaced by another monopoly

[00:49:04.0000] <Hixie>
we should have four browsers at 20% each and a mirad of other filling the remaining 20%

[00:49:05.0000] <Hixie>
four browser engines, rather

[00:50:00.0000] <Hixie>
myriad, even

[00:51:00.0000] <Hixie>
you know right now <canvas> is inline

[00:51:01.0000] <Hixie>
but it really should be either inline or inline or block

[00:51:02.0000] <Hixie>
in legacy terms

[00:51:03.0000] <Hixie>
er, either only block, or inline/block

[00:52:00.0000] <Hixie>
same with video/audio

[00:53:00.0000] <hsivonen>
Hixie: while you are at it, please consider the <figure> content model, too

[00:54:00.0000] <hsivonen>
that one has a clear bug now

[00:54:01.0000] <hsivonen>
and being able to showcase a block of text would make sense, too

[00:55:00.0000] <hsivonen>
the bug being the telescoping thing that you have to terminate <video> with <img> or <embed>

[00:56:00.0000] <hsivonen>
hmm. is iterating over childNodes in a for loop faster than iterating over the .nextSibling linked list?

[00:56:01.0000] <Hixie>
well i expect we'll just allow figure to contain anything

[00:56:02.0000] <Hixie>
as in, <figure> <legend> Listing 1 </legend> <pre> <code> ... </code> </pre> </figure>

[00:56:03.0000] <hsivonen>
Hixie: right

[00:57:00.0000] <Hixie>
so it'll solve itself

[00:57:01.0000] <Hixie>
i'm seriously considering allowing <figure> to basically contain anything

[00:57:02.0000] <Hixie>
even like <article>

[00:57:03.0000] <Hixie>
anything media or what we now think of as block-level

[00:57:04.0000] <hsivonen>
perhaps I shouldn't care about DOM perf right now

[00:58:00.0000] <Hixie>
don't pre-optimise :-)

[00:58:01.0000] <hsivonen>
the DOM is weird, because the internal impl can make different access patterns have counterintuitive perf characteristics if you consider the API

[00:59:00.0000] <hsivonen>
IIRC, Gecko uses a child node array internally instead of neighbor links like Xerces

[00:59:01.0000] <jgraham_>
/me isn't sure what <figure><article> would do but is generally in favour of less restrictive content models

[00:59:02.0000] <jgraham_>
s/do/mean/ maybe

[01:00:00.0000] <hsivonen>
/me likes neighbor link tree access

[01:00:01.0000] <Hixie>
jgraham_: well it'd be a figure of an entire article, e.g. maybe  <figure> <legend> Sample blog comment </legend> <article> <p>v1agr4!!!1</p> </article> </figure>  or something

[01:00:02.0000] <Hixie>
(not a quoted blog comment)

[01:01:00.0000] <Hixie>
the more i look at this the more i find that we have a whole bunch of subtly different element types

[01:02:00.0000] <jgraham_>
Yeah, that seems quite tenuous to be. But I don't see that as a reason to not allow it :)

[01:03:00.0000] <Hixie>
no disagreement there

[01:07:00.0000] <anne-mac>
Hixie, in order for all those simple elements to get implemented the rendering section really needs to be written

[01:07:01.0000] <hsivonen>
zcorpan: is the className redundancy with irrelevant in the collapsing script for IE?

[01:07:02.0000] <anne-mac>
especially for section, etc.

[01:08:00.0000] <hsivonen>
Was there a way to trigger attribute selectors dynamically in IE7?

[01:08:01.0000] <hsivonen>
should I support IE6?

[01:08:02.0000] <Hixie>
anne-mac: well most of them are just display:block

[01:08:03.0000] <Hixie>
anne-mac: but actually, i was hoping for implementation feedback :-)

[01:08:04.0000] <Hixie>
IE6 is around as big as IE7 iirc

[01:09:00.0000] <hsivonen>
Hixie: it's qualitatively different, though :-)

[01:10:00.0000] <anne-mac>
Hixie, I mean that we probably need something like :heading(n)

[01:10:01.0000] <anne-mac>
Hixie, implementation feedback for rendering has failed for the last 10 years or so except for trivial issues

[01:10:02.0000] <anne-mac>
see forms

[01:11:00.0000] <Philip`>
/me wishes people didn't make sites that only work with IPv6

[01:11:01.0000] <Hixie>
anne-mac: yeah well

[01:11:02.0000] <Hixie>
anne-mac: it's worked as well as for other things

[01:11:03.0000] <hsivonen>
Philip`: wow. what site?

[01:12:00.0000] <Hixie>
anne-mac: i agree that pseudo-classes would be helpful; not sure it's critical to getting html implemented though

[01:13:00.0000] <hsivonen>
Hixie: I think the HTML5 outline algorithm is mostly pointless without making sure that authors *see* the effects of the outline from day 1

[01:13:01.0000] <hsivonen>
Hixie: that is, h1-h6 should have traditional style rules in the UA style sheet

[01:13:02.0000] <anne-mac>
I side with hsivonen although I'm willing to be convinced we should simply add them as display:block

[01:13:03.0000] <hsivonen>
when there's no section parent

[01:13:04.0000] <hsivonen>
ancestor, rather

[01:13:05.0000] <Philip`>
hsivonen: streaming.ist-ring.eu, though maybe that's just because they haven't quite got around to adding an IPv4 address yet

[01:14:00.0000] <Hixie>
hsivonen: well, h1 should auto-style, i agree. (not h2-h6, for compat reasons)

[01:14:01.0000] <hsivonen>
Hixie: what's the compat problem with 

[01:14:02.0000] <hsivonen>
section h2:depth(n)

[01:14:03.0000] <hsivonen>
?

[01:15:00.0000] <hsivonen>
Hixie: that is, limiting the autostyling to cases where there is at least one section ancestor?

[01:16:00.0000] <hsivonen>
Hixie: does the outline depth of a heading ever depend on stuff the follows the heading in document order?

[01:16:01.0000] <anne-mac>
<h1>xxx</h1>heading 1<h3>xxx</h3>heading 2 iirc

[01:17:00.0000] <Hixie>
hsivonen: the idea is to allow authors to get the right rendering in legacy UAs

[01:17:01.0000] <hsivonen>
anne-mac: in that case, as far a I can tell, the outline depth of a given node depends on stuff before it in the document order

[01:18:00.0000] <Hixie>
hsivonen: but then we want compatible rendering across browsers

[01:18:01.0000] <hsivonen>
Hixie: well, section doesn't play nice in IE/Firefox, so...

[01:18:02.0000] <Hixie>
hsivonen: having it autostyle in newer browsers and not in the others would just lead to very confused authors and users with content that doens't do what the author intended

[01:19:00.0000] <hsivonen>
Hixie: leaving the styling to authors would lead to mostly academic applicability of the outline algorithm

[01:19:01.0000] <jgraham_>
/me agrees with hsivonen 

[01:19:02.0000] <Hixie>
hsivonen: the intention of hte algorithm was mostly just to define what it was, so that if people _do_ try to implement one, they at least do the same one

[01:20:00.0000] <hsivonen>
Hixie: btw, I think it would help a great deal if the outline algorithm were restated from the point of view of a node whose outline depth is being computed

[01:20:01.0000] <zcorpan>
hsivonen: ie6, yes (re className)

[01:20:02.0000] <Hixie>
hsivonen: you mean http://www.whatwg.org/specs/web-apps/current-work/#associatedSection ?

[01:20:03.0000] <hsivonen>
Hixie: such that it would be straight-forward to evaluate on a per-node basis

[01:20:04.0000] <zcorpan>
hsivonen: className += '' (or something similar) is needed to trigger a reflow in ie7

[01:21:00.0000] <hsivonen>
zcorpan: thanks. can reflow be forced in IE6 similarly? or does IE6 even support attribute selectors? I forget

[01:22:00.0000] <anne-mac>
Hixie, you want a clear algorithm to compute the heading level so implementors can say something about the impact of supporting :heading(n)

[01:22:01.0000] <zcorpan>
hsivonen: ie6 doens't support attr selectors

[01:22:02.0000] <hsivonen>
Hixie: not only walking back to find heading but walking back to find outline depth

[01:22:03.0000] <hsivonen>
zcorpan: ok. thanks

[01:23:00.0000] <hsivonen>
Hixie: I think the depth of a node in the outline should be computable on a per-node basis by walking the document order backwards

[01:23:01.0000] <Hixie>
hm

[01:23:02.0000] <jgraham_>
Yeah, the current algorithm is a) black magic and b) quite far away from what many types of UA are likely to want to implement

[01:24:00.0000] <hsivonen>
Hixie: and, conversely, on a batch basis by walking from the root forward without backtracking while keeping some variables

[01:24:01.0000] <Hixie>
anne-mac: i never actually intended for <h1> styling to be based on implied sections and so forth

[01:24:02.0000] <hsivonen>
Hixie: i.e. the isolated selector re-evaluation case and the SAX/batch selector case

[01:24:03.0000] <Hixie>
anne-mac: my original intent was for <h1> styling to be based exclusively on the number of <section> ancestors

[01:24:04.0000] <Hixie>
hsivonen: i guess that must be possible

[01:25:00.0000] <Hixie>
hsivonen: i haven't really thought about it

[01:25:01.0000] <Hixie>
hsivonen: the algorithms in the spec are intended to be optimised when implemented

[01:25:02.0000] <Hixie>
hsivonen: but i'm happy to change the spec's algorithm at some point if we find a better way of phrasing it

[01:25:03.0000] <anne-mac>
it's not just <section>, it's also <article>, etc.

[01:25:04.0000] <zcorpan>
hsivonen: function createHtmlElement(tagName) { return document.createElementNS ? document.createElementNS("http://www.w3.org/1999/xhtml", tagName) : document.createElement(tagName); }

[01:25:05.0000] <Hixie>
anne-mac: my original intent was for <h1> styling to be based exclusively on the number of <section> ancestors

[01:25:06.0000] <anne-mac>
but yeah, implied headings too

[01:26:00.0000] <anne-mac>
ok

[01:26:01.0000] <hsivonen>
zcorpan: thanks

[01:26:02.0000] <Hixie>
anne-mac: of course we later added other elements...

[01:26:03.0000] <Hixie>
anne-mac: which may make that no longer really the best solution

[01:26:04.0000] <othermaciej>
the OggStorm seems to have passed

[01:28:00.0000] <hsivonen>
Hixie: it seems to me that stating the spec in terms of algorithms is helpful when the general point of view of the algorithm matches the way implementors will need to look at the issue

[01:29:00.0000] <hsivonen>
Hixie: however, requiring each implementor to work out the invariats / declarative statement of what the algorithm does and recasting it to another algorithm from a different POV doesn't seem useful

[01:29:01.0000] <hsivonen>
Hixie: if the implementor is expected to find a very different algorithm, a declarative statement of the identities/constraints would be more useful

[01:30:00.0000] <othermaciej>
are y'all talking about the sectioning algorithm?

[01:31:00.0000] <Hixie>
if we can find a way to phrase it that way that works too

[01:31:01.0000] <hsivonen>
othermaciej: yes

[01:31:02.0000] <othermaciej>
I must admit as written it's not directly useful to a UA that wants to efficiently style headings by level

[01:31:03.0000] <othermaciej>
but I'm not actually sure it even defines an effective heading level

[01:32:00.0000] <othermaciej>
(should both the rank and level of section nesting affect the effective heading level? only level of section nesting?)

[01:32:01.0000] <jgraham_>
othermaciej: Both, in different ways

[01:33:00.0000] <anne-mac>
i'm fine with only doing h1 auto-styling btw, although it might encourage people to create pages that are not backwards compatible

[01:33:01.0000] <anne-mac>
you'd do h1:heading(n) in the browser basically

[01:34:00.0000] <hsivonen>
zcorpan: is the purpose of addValueAttrs() to prevent the UA from renumbering the list items when some are hidden?

[01:34:01.0000] <othermaciej>
jgraham_: can you be more specific?

[01:35:00.0000] <othermaciej>
the sectioning algorithm does specifically allow having lower rank headers than the first inside a section

[01:35:01.0000] <hsivonen>
Hixie: thinking about it more, I think the we need two things:

[01:35:02.0000] <othermaciej>
without creating a section break

[01:35:03.0000] <hsivonen>
1) a forward-only sweep that keeps variables as it goes and assigns depth to nodes

[01:36:00.0000] <hsivonen>
2) A way to walk back from a tainted node to an untainted node and then do a partial sweep from there onwards

[01:37:00.0000] <hsivonen>
it would be a bonus if the partial sweep could end at the last sibling of the tainted node and wouldn't require resweeping the whole doc until the end

[01:38:00.0000] <Hixie>
send mail

[01:38:01.0000] <Hixie>
i'm far from working on that right now

[01:38:02.0000] <hsivonen>
Disclaimer: I have no idea if it is even footprint-wise acceptable to keep a level variable and taint flag on nodes

[01:38:03.0000] <Hixie>
it's content models all the way til the new year for me, i think

[01:38:04.0000] <hsivonen>
Hixie: ok

[01:38:05.0000] <othermaciej>
if the styling is solely based on rank and number of section ancestors that might not be too hard

[01:39:00.0000] <anne-mac>
i think it should also be based on previous siblings

[01:39:01.0000] <othermaciej>
keeping flags on HTMLHeaderElement nodes only is likely acceptable

[01:39:02.0000] <Hixie>
hsivonen: but in general i agree with your suggestions and requests

[01:39:03.0000] <anne-mac>
<h1>1</h1>xx<h3>2</h3>

[01:40:00.0000] <anne-mac>
<h1>1</h1>xx<div><h3>2</h3></div> makes this more complicated I guess

[01:40:01.0000] <Hixie>
(also <h3>1</h3>xx<h2>1</h2> iirc)

[01:40:02.0000] <krijnh>
I don't get 'backwards compatibility' with h1-h6 combined with section/article, when section/article isn't even supported now..

[01:41:00.0000] <krijnh>
When people start using <section> <hx> surely they don't care about backwards compatibility, right?

[01:42:00.0000] <krijnh>
(Sorry, didn't mean to ruin the conversation)

[01:43:00.0000] <hsivonen>
krijnh: as far as I can tell, <section> is backwards-compat poison, yes

[01:43:01.0000] <krijnh>
hsivonen: I know

[01:43:02.0000] <zcorpan>
Hixie: it seems wf2 doesn't allow <form target>. is that on your radar or should i send email?

[01:43:03.0000] <Hixie>
mail

[01:43:04.0000] <zcorpan>
ok

[01:43:05.0000] <Hixie>
but wf2 is blocked until the forms tf comes back

[01:44:00.0000] <hsivonen>
zcorpan: I speculatively treated the non-allowance as a spec bug last night

[01:45:00.0000] <othermaciej>
oh great, that means the forms tf needs to do something

[01:45:01.0000] <hsivonen>
othermaciej: no, they don't. target isn't architectural consistency. (or so I'd hope)

[01:46:00.0000] <hsivonen>
Hixie: whatever happened with WF2 call for impls from the WHATWG POV?

[01:47:00.0000] <Hixie>
still waiting :-)

[01:47:01.0000] <Hixie>
afk

[01:48:00.0000] <othermaciej>
hsivonen: the Forms TF needs to do something at some point to remove the blocker to forms progress

[01:51:00.0000] <anne-mac>
there's a deadline

[01:55:00.0000] <Philip`>
What happens if the deadline is exceeded?

[01:56:00.0000] <Hixie>
anne-mac: yeah because timelines have worked really well so far in this working group

[01:57:00.0000] <Hixie>
also, as far as i can tell there isn't a deadline per so, just a desired timeline :-)

[01:58:00.0000] <anne-mac>
well, if there isn't any progress by then I think there's enough evidence that it doesn't work

[02:00:00.0000] <anne-mac>
zcorpan, where does WF2 state it has removed a feature from HTML4?

[02:00:01.0000] <hsivonen>
Hixie: what's the deal with HTML5 zapping the HTMLHeaderElement interface?

[02:02:00.0000] <Hixie>
there wasn't anything left in it, since i haven't put the align="" attributes, etc in the idl

[02:05:00.0000] <hsivonen>
Hixie: but browsers need to support HTMLHeaderElement until the end of the world anyway

[02:06:00.0000] <Hixie>
yeah, there's a bunch of stuff i need to re-add, probably when i do the rendering section

[02:07:00.0000] <anne-mac>
let me know when you change your mind about IDL interfaces being for authors

[02:07:01.0000] <anne-mac>
so I can complain about document.write() again

[02:07:02.0000] <anne-mac>
:)

[02:08:00.0000] <hsivonen>
Hixie: email sent about the outline algorithm. I hope the email captures the main points.

[02:09:00.0000] <othermaciej>
I don't think it makes sense to leave "bad" things out of the IDL interfaces

[02:10:00.0000] <othermaciej>
it makes life harder for implementors (since we have to reconstruct the real IDL interfaces) and conformance checking script use of DOM APIs seems pointless and infeasible

[02:10:01.0000] <Hixie>
anne-mac: almost all idl changes are on hold until bindings for dom is done

[02:10:02.0000] <Hixie>
hsivonen: thanks

[02:10:03.0000] <Hixie>
othermaciej: agreed

[02:11:00.0000] <Hixie>
othermaciej: i just haven't done any of the presentational attributes yet, dom or not

[02:12:00.0000] <othermaciej>
sounds reasonable

[02:13:00.0000] <anne-mac>
/me suggested to heycam to name it Web IDL instead

[02:14:00.0000] <hsivonen>
do the Samsung products support Theora, too, or only Ogg and Vorbis?

[02:14:01.0000] <Lachy>
woah! Microsoft have shipped ogg vorbis in Halo for PC

[02:15:00.0000] <othermaciej>
hsivonen: I'd be surprised if any products support Theora already given that the bitstream format wasn't frozen until a few months ago

[02:15:01.0000] <othermaciej>
(not impossible though)

[02:15:02.0000] <hsivonen>
Lachy: yeah, *Vorbis* has shipped in a lot of high-$$$ games

[02:16:00.0000] <aphid>
metavid is theora based, we have a few thousand hours of congressional video

[02:17:00.0000] <othermaciej>
it's unfortunate that the Xiph press release conflates Ogg, Vorbis and Theora

[02:17:01.0000] <hsivonen>
othermaciej: yeah.

[02:21:00.0000] <zcorpan>
anne-mac: <form target> is not allowed in html4 strict

[02:22:00.0000] <anne-mac>
sure

[02:23:00.0000] <maikmerten>
the Ogg family is considered to be a set of codecs which is simply called "Ogg". Granted, the release could have been more specific.

[02:23:01.0000] <zcorpan>
anne-mac: i had presumed that wf2 built on top of html4 strict, but it actually just says html4, so i don't know

[02:24:00.0000] <anne-mac>
it is true that the XHTML Module definition at the end doesn't mention <form target>, <input target> or <button target>

[02:24:01.0000] <maikmerten>
anyway, the press release at least totally applies to Ogg Vorbis. Albeit most discussion is centered on video the audio part shouldn't be neglected either

[02:24:02.0000] <Hixie>
maybe i should just allow any nesting, except that each element adds to a set of elements that are banned as descendants

[02:26:00.0000] <Hixie>
e.g. <header> can't contain <footer> but can contain anything else

[02:26:01.0000] <Hixie>
oh also <section>, <article>, <aside> wouldn't be allowed

[02:26:02.0000] <othermaciej>
maik|afk: it seems to say things that mostly apply to Ogg Vorbis, but then talks about the <video> element and video in general

[02:26:03.0000] <Hixie>
but <nav> maybe would? maybe not.

[02:26:04.0000] <Hixie>
gah

[02:27:00.0000] <anne-mac>
where is the <a> element on the v axis?

[02:27:01.0000] <Hixie>
oops, maybe i dropped it

[02:27:02.0000] <Hixie>
added it

[02:28:00.0000] <othermaciej>
oops, I was wrong about when the Theora bitstream format was frozen

[02:28:01.0000] <anne-mac>
oh, <dt> should probably be similar to <h1>-<h6> too

[02:28:02.0000] <othermaciej>
mea culpa

[02:29:00.0000] <maikmerten>
othermaciej, while this is true the itent is also to show xiph.org is knowing what it is doing

[02:30:00.0000] <maikmerten>
and of course the millions of deployments *do* apply to Theora

[02:30:01.0000] <Hixie>
are there any deployments of theora that are done by companies with big targets on their backs? because that's the only reason we're not requiring it at the moment (the submarine patent issue)

[02:31:00.0000] <hsivonen>
maikmerten: it seems to me that Ogg proponents arguing about Theora risks is like a Green party making statements about pollution: if it isn't 100% correct, the incorrect part gets all the attention

[02:31:01.0000] <maikmerten>
I consider Novell rather big

[02:31:02.0000] <maikmerten>
hsivonen, same for Nokia position papers, unfortunately ;)

[02:32:00.0000] <maikmerten>
(but that is another cup of tea)

[02:32:01.0000] <Hixie>
novell is pretty small these days, heck they're below $3bn market cap and living mostly on chicken feed from microsoft

[02:33:00.0000] <hsivonen>
is IBM shipping Theora in any of their Linux things?

[02:33:01.0000] <maikmerten>
but at least it's clear that at least some of the Ogg codecs are rather widely deployed, well tested and adopted by big players, still e.g. word has been that "Ogg" is not an option

[02:33:02.0000] <maikmerten>
what about Ogg Vorbis? What about Ogg Speex? (Xbox live)

[02:33:03.0000] <Hixie>
as far as i know the only argument is about ogg theora

[02:34:00.0000] <Hixie>
i don't know that anyone has made any particular claims about vorbis

[02:34:01.0000] <hsivonen>
maikmerten: almost all of the naysaying in the HTML5 context has been about Theora. Xiph using Vorbis references is not helpful there

[02:34:02.0000] <roc>
IBM doesn't ship Linux

[02:34:03.0000] <maikmerten>
hsivonen, if IBM is shipping a full blown linux desktop: I'd think so

[02:35:00.0000] <Philip`>
Should the <audio> baseline be the same as the audio component of <video>?

[02:35:01.0000] <roc>
they aren't

[02:35:02.0000] <anne-mac>
hsivonen, Vorbis was in the spec before though

[02:35:03.0000] <Philip`>
(*...of the <video> baseline)

[02:35:04.0000] <Hixie>
audio isn't really anywhere near as important as video

[02:36:00.0000] <othermaciej>
Philip`: possibly, although simple lossless formats might also be worth supporting for things like sound effects

[02:36:01.0000] <hdh>
... outside myspace

[02:37:00.0000] <hsivonen>
although it is well-known in Maemo circles that Nokia specifically avoids shipping Vorbis, too

[02:37:01.0000] <maikmerten>
hsivonen, albeit this is correct most of the press release states that albeit major proponents of MPEG give the public image that MPEG is somehow safe (they know otherwise and won't deny when being asked in detail) in fact it's not quite that easy

[02:37:02.0000] <stijntje>
hdh: doesn't myspace use flash for audio anyway, presumably to make copying it harder (which would be dead easy with <audio src="...">?

[02:39:00.0000] <othermaciej>
it's not that hard to grab the audio from myspace

[02:39:01.0000] <othermaciej>
(at least with Safari's activity window)

[02:39:02.0000] <Philip`>
stijntje: That would require people to 'view source' and work their way through masses of horrid HTML to find the URI, which sounds harder than just copying Flash's cache of the file from its temp directory

[02:40:00.0000] <Philip`>
(unless their browser adds a 'save this file' button to the <audio> UI)

[02:40:01.0000] <stijntje>
considering myspace's use of HTML, that would make sense ;)

[02:40:02.0000] <roc>
I bet there's a Firefox extension for this

[02:40:03.0000] <hdh>
then there's commons.wikimedia, librivox, gutenberg(?), streaming sites like jamendo

[02:41:00.0000] <stijntje>
othermaciej: it is not hard indeed, in the way it is not hard to copy a "protected" (transparent .gif on top) photo on flickr too, but it does stop the "masses" from doing it

[02:41:01.0000] <stijntje>
of course I'm not working at MySpace so I wouldn't know whether that's really the reason to use flash, I can imagine interoperability and usability would play a significant role too :P

[02:41:02.0000] <othermaciej>
hah, I didn't know flickr did silly things like that

[02:42:00.0000] <hdh>
myspace example was an uneducated guess from xkcd #134 anyway

[02:42:01.0000] <Hixie>
maikmerten: at least in the html5 context, i don't think anyone has claimed that mpeg is safe

[02:43:00.0000] <roc>
http://lifehacker.com/software/downloads/download-of-the-day--video-downloader-firefox-extension-170659.php

[02:43:01.0000] <Hixie>
maikmerten: i can't comment on the "public image" aspect, but i assure you that has no bearing on the technical decisions of the spec

[02:43:02.0000] <hdh>
doesn't a royalty-free requirement exclude all things under MPEG-LA?

[02:44:00.0000] <Hixie>
unless MPEG-LA change their minds, yes

[02:44:01.0000] <Hixie>
well, there are some obsolete MPEG codecs we could use

[02:44:02.0000] <Hixie>
but they're obsolete.

[02:45:00.0000] <stijntje>
are there any alternatives to theora (patent issues-wise) that are not?

[02:46:00.0000] <Philip`>
Dirac is the only similar project I've heard of

[02:46:01.0000] <Philip`>
and that seems somewhat too experimental at the moment

[02:46:02.0000] <stijntje>
yeah, me too, but afaik it's still in its early stages

[02:46:03.0000] <roc>
Dirac is not done and has a lot less patent analysis than Theora

[02:46:04.0000] <roc>
AFAIK

[02:47:00.0000] <maikmerten>
Hixie, right

[02:47:01.0000] <maikmerten>
Hixie, this release is bringing nothing new to the table

[02:47:02.0000] <hsivonen>
Dirac uses arithmetic coding. When are the IBM patents on that going to hit the 20 year mark?

[02:47:03.0000] <maikmerten>
Hixie, but it was necessary to e.g. contain the damage Nokia was doing to us in the public image

[02:48:00.0000] <hsivonen>
The Dirac FAQ is rather vague on that pointt

[02:48:01.0000] <maikmerten>
Hixie, like to fight the impression xiph.org would be new/geeky/not trustworthy and generally not used at all

[02:48:02.0000] <hdh>
there's a #dirac here

[02:49:00.0000] <maikmerten>
Hixie, the background is that xiph.org is being flooded right now by press people - so we can deal with the "FAQ" in a release

[02:49:01.0000] <Hixie>
maikmerten: sure, i wasn't commenting on the press release, i understand the reasons for it

[02:50:00.0000] <Hixie>
cool, being flooded with press people is a good problem to have :-)

[02:50:01.0000] <maikmerten>
didn't I yesterday state that press releases are boring ;)

[02:50:02.0000] <maikmerten>
s/release/press release, obviously

[02:50:03.0000] <othermaciej>
as some famous person once said, "there's no such thing as bad publicity"

[02:52:00.0000] <Philip`>
othermaciej: Is there any evidence to back up that statement? :-)

[02:52:01.0000] <Hixie>
bed time

[02:52:02.0000] <Hixie>
nn

[02:53:00.0000] <stijntje>
gn

[02:58:00.0000] <heycam>
anne-mac, yeah sorry about the delay with that.  over the christmas break, once the next batik release is out, i'll have some time for it.

[03:01:00.0000] <Lachy>
this is interesting. Not sure what I think of it yet (although I'm not really allowed to say much about it). http://www.opera.com/pressreleases/en/2007/12/13/

[03:03:00.0000] <hsivonen>
wow

[03:03:01.0000] <stijntje>
"Opera requests the Commission to implement two remedies to Microsoft�s abusive actions. First, it requests the Commission to obligate Microsoft to unbundle Internet Explorer from Windows" <- wouldn't that just lead to another Windows N no one will buy?

[03:10:00.0000] <hsivonen>
hmm. Opera Mobile is not linked from the press release like desktop and mini

[03:10:01.0000] <hsivonen>
and the Opera 9 coming soon for S60 banner is gone

[03:10:02.0000] <othermaciej>
stijntje: I think Microsoft can achieve that all by itself :-)

[03:15:00.0000] <hasather>
hsivonen: Opera 9 Mobile was announced for Windows Mobile and UIQ some days ago, but no word on the S60 version yet

[03:15:01.0000] <othermaciej>
now seems like an odd time to file an antitrust complaint about Microsoft bundling the browser

[03:15:02.0000] <hsivonen>
hasather: oh. interesting

[03:17:00.0000] <hasather>
hsivonen: see here, I asked Daniel about the S60 version in the comments: http://operawatch.com/news/2007/12/here-comes-opera-mobile-9.html

[03:19:00.0000] <hsivonen>
hasather: thanks

[04:07:00.0000] <stijntje>
mp3: Extince - Le Fanclub

[04:07:01.0000] <stijntje>
woops

[04:07:02.0000] <stijntje>
apologies, wrong channel :(

[04:09:00.0000] <gsnedders>
/me notices the whole zero conversations that stijntje disturbed :)

[04:10:00.0000] <stijntje>
:D

[04:14:00.0000] <othermaciej>
if I want to vertically center-align text in two different font sizes using CSS, can I do that?

[04:14:01.0000] <othermaciej>
(sorry for slightly off-topic question, figured people here might know)

[04:16:00.0000] <hsivonen>
now that I'm working on front end UI JS, are there other Validator.nu front-end JS requests besides duplicate message collapsing and swapping in a <textfield> or a file upload control?

[04:17:00.0000] <othermaciej>
/me is terrified to discover that you can relative-position a float

[04:18:00.0000] <Philip`>
othermaciej: You should use <table valign=middle>

[04:19:00.0000] <othermaciej>
Philip`: isn't it a sin against all that is righteous to use layout tables?

[04:19:01.0000] <Philip`>
othermaciej: It's not really layout, it's just presentation

[04:19:02.0000] <othermaciej>
(I got what I wanted to working but the result is a little ugly)

[04:20:00.0000] <stijntje>
would display: table-cell + vertical-align: middle work? notwithstanding that it doesn't work in IE7

[04:20:01.0000] <othermaciej>
I'm not sure if that would properly vertically align two pieces of text in two different fonts

[04:21:00.0000] <othermaciej>
I think I would need them to be in two different table cells

[04:21:01.0000] <Philip`>
<div style="display:table"><div style="display:table-row"><div style="display:table-cell;vertical-align:middle">...

[04:21:02.0000] <Philip`>
That's not got evil <table> therefore it's good

[04:21:03.0000] <stijntje>
not saying so, just wondering whether it'd work 

[04:22:00.0000] <gsnedders>
stijntje: well, yeah, it'd work

[04:22:01.0000] <Philip`>
Does CSS define that the outer table and table-row get inferred somehow?

[04:22:02.0000] <othermaciej>
what I have is:

[04:22:03.0000] <othermaciej>
<h2><span style="float: left; position: relative; bottom: 0.25em; font-size: 2em;">&#x2600;&nbsp;</span> My Heading Stuff</h2>

[04:23:00.0000] <othermaciej>
the need for the nbsp is unfortunate

[04:23:01.0000] <othermaciej>
(I will move the style rules into the stylesheet once I decide how this should work)

[04:26:00.0000] <Philip`>
<div style="display:inline-block;vertical-align:middle"> works in Opera at least

[04:26:01.0000] <gsnedders>
Philip`: how do you do bar charts in HTML?

[04:26:02.0000] <Philip`>
gsnedders: Do you mean the ones like http://www.cl.cam.ac.uk/~pjt47/misc/attributes.html ?

[04:26:03.0000] <stijntje>
firefox doesn't support inline-block afaik

[04:26:04.0000] <gsnedders>
Philip`: yeah

[04:27:00.0000] <othermaciej>
Philip`: I ruled out inline-block due to lack of Firefox/IE support

[04:27:01.0000] <Philip`>
gsnedders: View source :-)

[04:27:02.0000] <Philip`>
-moz-inline-block sometimes works in Firefox

[04:27:03.0000] <gsnedders>
Philip`: so the JS and CSS?

[04:27:04.0000] <Philip`>
gsnedders: Yes, and the HTML

[04:28:00.0000] <Philip`>
Ignore the incorrectness of the word "histogram" as used in the code

[04:33:00.0000] <gsnedders>
wow. Excel 11 Mac's HTML output isn't that bad.

[04:46:00.0000] <gsnedders>
/me wonders why the code is failing on him

[04:50:00.0000] <gsnedders>
Philip`: the entire first column gets changed to NaN here :\

[04:52:00.0000] <gsnedders>
ah. whitespace between <tr> and <td>

[04:59:00.0000] <hsivonen>
zcorpan: if the script is expected to prevent li renumbering in Firefox, it isn't working

[05:02:00.0000] <hsivonen>
zcorpan: script deployed

[05:02:01.0000] <hsivonen>
zcorpan: thank you

[05:05:00.0000] <hsivonen>
hmm. now expanding a grouped message is counter-intuitive when expanding

[05:05:01.0000] <hsivonen>
because you don't notice where it expands

[05:08:00.0000] <stijntje>
http://img520.imageshack.us/img520/9563/negerted2.png

[05:08:01.0000] <stijntje>
nvm

[05:17:00.0000] <annevk>
othermaciej, Firefox 3 does inline-block...

[05:18:00.0000] <hsivonen>
hmm. I guess the message grouping UI could be better, if the collapsed messages were moved into a sublist in <details>

[05:18:01.0000] <hsivonen>
but <details> lacks impls.

[05:19:00.0000] <hsivonen>
and if people start scripting <details> impls now, it'll poison the possibility of introducing native impls without breaking the Web.

[05:19:01.0000] <hsivonen>
hmm.

[05:20:00.0000] <hsivonen>
Quick poll: if I change the "collapse all" feature to "group by message", do I need to provide an undo button?

[05:21:00.0000] <hsivonen>
for ungrouping?

[05:25:00.0000] <zcorpan>
hsivonen: http://simon.html5.org/temp/validator-nu-collapse.html has ungrouping

[05:25:01.0000] <zcorpan>
and some bugs fixed, in case you were working from an older copy

[05:25:02.0000] <hsivonen>
zcorpan: yeah. looks like I accidentally broke ungrouping

[05:26:00.0000] <hsivonen>
zcorpan: I downloaded yesterday evening version

[05:27:00.0000] <zcorpan>
hsivonen: ok, then it should be the latest version

[05:28:00.0000] <zcorpan>
hsivonen: it doesn't renumber in opera

[05:28:01.0000] <hsivonen>
zcorpan: fixed

[05:28:02.0000] <zcorpan>
er, at least not on my sample

[05:29:00.0000] <hsivonen>
zcorpan: what's the purpose of addValueAttrs then?

[05:30:00.0000] <zcorpan>
hsivonen: to prevent renumbering. it works on my copy but not on yours

[05:30:01.0000] <hsivonen>
weird

[05:31:00.0000] <hsivonen>
I'm tempted to reimplement it as O(n) and with grouping the various locations in a sublist

[05:31:01.0000] <zcorpan>
ok, that might result in a better ui

[05:37:00.0000] <zcorpan>
/me is very confused as to why collapsing renumbers in v.nu but not in my copy

[05:50:00.0000] <Philip`>
gsnedders: It doesn't do anything clever about handling whitespace - it just assumes the HTML is written precisely like how I wrote it, without doing anything useful like documenting that fact :-)

[05:50:01.0000] <gsnedders>
Philip`: :)

[06:09:00.0000] <hsivonen>
what's the difference of textContent and innerText?

[06:11:00.0000] <hsivonen>
is it just a gratuitous naming difference?

[06:32:00.0000] <gsnedders>
http://http-parsing.gsnedders.com/

[06:34:00.0000] <hsivonen>
gsnedders: hmm. the x-pingback frequency suggests that dmoz doesn't link to blogs much

[06:34:01.0000] <gsnedders>
/me doesn't know how Philip` chose the sample

[06:36:00.0000] <gsnedders>
Philip`: what's interesting is we have one header name which is blank, and one that is 0. Looks like the parsing isn't perfect :)

[06:40:00.0000] <hsivonen>
set-cookie2 looks like a miserable failure

[06:46:00.0000] <Philip`>
gsnedders: It's a random selection from the ~4.5M URIs in a dmoz.org dump from a few months ago (after removing duplicates)

[06:46:01.0000] <Philip`>
(where "random" means "sort -R | head")

[06:50:00.0000] <Philip`>
gsnedders: The one with an apparently blank header is a pretty broken server

[06:50:01.0000] <Philip`>
(http://www.nightwing.com.au/index.htm)

[06:52:00.0000] <Philip`>
It sends Content-Type twice, and "Client-Junk: :", etc, though I've got no idea why it's parsed like it is

[06:53:00.0000] <Philip`>
Oh, wait

[06:53:01.0000] <Philip`>
It does different things if I use curl instead of GET

[06:53:02.0000] <Philip`>
It has a header line ":"

[06:59:00.0000] <Philip`>
http://www.4silverhillspares.co.uk/cgi-bin/home.pl says "Pragma: no-cache" "0: no-cache"

[07:06:00.0000] <Philip`>
http://wm01.allmusic.com/cg/amg.dll says "Content-Type: text/html" "Content-Length: 31155" "Content:" - I'm not sure if they were intentionally adding that header last to identify the message content

[07:07:00.0000] <Philip`>
Where does Page-Completion-Status come from?

[07:08:00.0000] <Philip`>
Oh, ColdFusion

[07:11:00.0000] <zcorpan>
is there a reference somewhere for how much of the web is text/html?

[07:12:00.0000] <zcorpan>
i've seen the number 98% but i don't know where it came from

[07:12:01.0000] <Philip`>
98% of what?

[07:12:02.0000] <Philip`>
(e.g. does that include image files and scripts and stylesheets and everything else?)

[07:13:00.0000] <zcorpan>
(no)

[07:13:01.0000] <zcorpan>
what <a href> points to

[07:13:02.0000] <zcorpan>
i guess

[07:14:00.0000] <Philip`>
How would you count sites with a million pages or sites with an infinite number of pages?

[07:16:00.0000] <zcorpan>
will you do a crawl and then tell me how many text/html results you got if i suggest how to count sites? :)

[07:17:00.0000] <zcorpan>
/me does s/98%/The majority/

[07:17:01.0000] <zcorpan>
s/majority/vast majority/

[07:18:00.0000] <Philip`>
I can find that 99.7% of dmoz.org is text/html with the particular User-Agent and Accept headers that I have

[07:18:01.0000] <Philip`>
If you can suggest what "a crawl" is, I could probably do something like that :-)

[07:22:00.0000] <Philip`>
gsnedders: Maybe you should indicate whether e.g. a site that sends two Set-Cookies is counted as one or two

[07:30:00.0000] <gsnedders>
wow. almost all the Date values are valid.

[07:32:00.0000] <Philip`>
I assume people writing CGI/PHP/etc don't bother outputting Date at all, so the Dates are generated by Apache/IIS/etc and are therefore more competently designed and tested

[07:32:01.0000] <Philip`>
*so the Dates which exist are ...

[07:35:00.0000] <gsnedders>
/me realises he actually has a bug in his code so even more are valid :P

[07:35:01.0000] <Philip`>
Try checking cookies for validity

[07:36:00.0000] <gsnedders>
OK, one more date is valid. :P

[07:36:01.0000] <Philip`>
(At least HttpClient complained a lot about cookies, so I guess there's interesting problems there)

[07:40:00.0000] <gsnedders>
http://http-parsing.gsnedders.com/ now has Date values

[07:42:00.0000] <Philip`>
"Everything done. Thank you for downloading a media file containing proprietary and patented technology. Core dumped ;)" - mplayer is a bit peculiar when it reaches the end of a stream

[07:42:01.0000] <Philip`>
gsnedders: That makes it sound like people actually say "Date: RFC822"

[07:43:00.0000] <gsnedders>
I couldn't think of any better way to phrase it though, quickly, Philip` 

[07:43:01.0000] <gsnedders>
(and I'm not having each and every RFC 822 date separately :))

[07:43:02.0000] <Philip`>
"Most Common Date Formats"?

[07:45:00.0000] <Philip`>
and probably show the asctime-formatted date, rather than saying "asctime()"

[07:45:01.0000] <gsnedders>
Philip`: it's valid, though

[07:46:00.0000] <Philip`>
gsnedders: It's confusing since it makes me wonder whether someone literally sent "Date: asctime()"

[07:46:01.0000] <hsivonen>
IIRC The httpclient docs say you need netscape emulation instea of rfc stuff

[07:47:00.0000] <Philip`>
hsivonen: It says that for cookies (http://jakarta.apache.org/httpcomponents/httpclient-3.x/cookies.html) - do you mean it says the same for Date headers?

[07:51:00.0000] <zcorpan>
/me finds that http://dev.w3.org/csswg/cssom-view/ has a different "the body element" than html5

[07:51:01.0000] <zcorpan>
/me thinks cssom-view's definition is better

[07:52:00.0000] <Philip`>
There's quite a bit of variation in cookie expiry dates - "01-Jan-2038", "02 Jan 1970", "04-Dec-37", lots of "Friday" vs "Fri", a "12:40:25 AM GMT", some missing spaces/commas, etc

[07:52:01.0000] <hsivonen>
Philip`: no, just cookies

[07:53:00.0000] <gsnedders>
Philip`: is it clearer now?

[07:54:00.0000] <Philip`>
hsivonen: Okay

[07:54:01.0000] <Philip`>
gsnedders: It is

[07:54:02.0000] <gsnedders>
Philip`: now, should I include dates from other headers there?

[07:55:00.0000] <Philip`>
Server: Microsoft-IIS/6.0

[07:55:01.0000] <Philip`>
File 'c: \mysql\share\charsets\?.conf' not found (Errcode: 2)

[07:55:02.0000] <Philip`>
Character set '#33' is not a compiled character set and is not specified in the 'c: \mysql\share\charsets\Index' file

[07:55:03.0000] <Philip`>
X-Powered-By: PHP/4.4.7

[07:56:00.0000] <Philip`>
(from http://www.ordinepsicologi.piemonte.it/) doesn't look quite right

[07:57:00.0000] <Philip`>
gsnedders: Depends on what your data is meant to be used for and whether that'll be useful, I guess

[07:57:01.0000] <gsnedders>
how can you merge multiple lists? :P

[07:57:02.0000] <Philip`>
In Python?

[07:57:03.0000] <gsnedders>
just use .append() multiple times?

[07:57:04.0000] <gsnedders>
yeah

[07:57:05.0000] <Philip`>
[1,2] + [3,4]

[07:58:00.0000] <Philip`>
or I guess there's some concatenate method

[07:58:01.0000] <Philip`>
http://docs.python.org/lib/typesseq-mutable.html - list1.extend(list2)

[07:59:00.0000] <gsnedders>
/me doesn't really understand their equivalents

[08:00:00.0000] <gsnedders>
Expires is the only other header allowed in responses

[08:00:01.0000] <gsnedders>
and Last-Modified

[08:00:02.0000] <Philip`>
gsnedders: Do you mean the s[len(s):len(s)] = x equivalent?

[08:00:03.0000] <gsnedders>
yeah

[08:02:00.0000] <Philip`>
s[a:b] refers to the part of the list from index a up to index b-1, so s[len(s):len(s)] refers to the zero-length part just after the final item in the list, and if you assign something there then it gets spliced into s

[08:07:00.0000] <gsnedders>
Odd. The second most common date format (after valid RFC 822) is "-1".

[08:08:00.0000] <gsnedders>
Philip`: ah. slightly odd way to do it logically.

[08:08:01.0000] <Philip`>
gsnedders: That's why they have .append() and .extend() instead :-)

[08:09:00.0000] <gsnedders>
most of the invalid dates are relatively consistent in their format, which is nice.

[08:11:00.0000] <gsnedders>
defining date parsing isn't nice, though

[08:11:01.0000] <gsnedders>
/me points at HTML 5

[08:22:00.0000] <gsnedders>
Defining date formats (including invalid ones) as ABNF seems simpler

[08:23:00.0000] <gsnedders>
Do any elements close open formatting elements?

[08:23:01.0000] <Philip`>
HTML5 isn't necessarily a good example of the neatest way to write algorithms :-)

[08:27:00.0000] <gsnedders>
There are a finite number of ways to do so that leave nothing that's unclear, though

[08:27:01.0000] <Philip`>
Experimentation suggests that nothing closes open formatting elements

[08:28:00.0000] <Philip`>
at least when you do <b> ... <x></x> for all tag names x ... </b>

[08:37:00.0000] <Philip`>
Hixie: status-documentation.html says "s quite stable" and probably needs more vowels

[08:53:00.0000] <zcorpan>
/me looks at the matrix

[08:54:00.0000] <zcorpan>
<footer><address> seems like it should be allowed

[08:54:01.0000] <gsnedders>
wow. 1547 Content-Location headers.

[08:57:00.0000] <Philip`>
/me wonders what Content-Location is meant to be used for

[08:59:00.0000] <gsnedders>
Philip`: Base URI to resolve relative URIs to

[09:00:00.0000] <Philip`>
gsnedders: Ah

[09:00:01.0000] <gsnedders>
IIS in some cases sends an bad URI by default

[09:01:00.0000] <gsnedders>
Philip`: http://support.microsoft.com/kb/218180

[09:02:00.0000] <Philip`>
I guess http://page.freett.com/cahfm/ is a bad one since its Content-Location points at :8080 which isn't listening to connections

[09:02:01.0000] <gsnedders>
the norm is an internal network IP

[09:03:00.0000] <Philip`>
Are you allowed relative URIs, like "Content-Location: rfc1596.html"?

[09:04:00.0000] <Philip`>
/me assumes so, since ietf.org is doing that

[09:05:00.0000] <gsnedders>
off the top of my head, no

[09:05:01.0000] <gsnedders>
you are, actually

[09:05:02.0000] <gsnedders>
Location must be absolute, though

[09:06:00.0000] <gsnedders>
caused all kinds of issues in older versions of Saf that enforced that

[09:11:00.0000] <gsnedders>
/me wonders what this page-completion-status header is

[09:12:00.0000] <gsnedders>
ColdFusion apparently

[09:24:00.0000] <Philip`>
gsnedders: http://krijnhoetmer.nl/irc-logs/whatwg/20071213#l-771

[09:24:01.0000] <gsnedders>
heh.

[09:24:02.0000] <gsnedders>
I missed that.

[09:25:00.0000] <gsnedders>
/me wonders how he should deal with Server values

[09:25:01.0000] <Philip`>
Use a pie chart!

[09:25:02.0000] <gsnedders>
mmm… π…

[09:26:00.0000] <gsnedders>
But do I keep the whole string, the server/version, or just the server?

[09:27:00.0000] <Philip`>
Do a pie chart with one slice per server, then cut the slices by version, then cut the slice cuts by string

[09:27:01.0000] <Philip`>
Or do three separate bar charts

[09:28:00.0000] <gsnedders>
And what about invalid values? :P

[09:29:00.0000] <Philip`>
Create an "unknown" category, since they're not really invalid since there's no syntax constraints

[09:30:00.0000] <gsnedders>
No, it must be a token :P

[09:32:00.0000] <gsnedders>
/me finds bug in RFC2616

[09:32:01.0000] <Philip`>
That says it must be >= 1 token, I think

[09:32:02.0000] <Philip`>
but its definition of the meaning of the tokens is incompatible with reality

[09:33:00.0000] <gsnedders>
actually, it's unclear only

[09:33:01.0000] <gsnedders>
with 1*(product-token | comment) are you allowed implied *LWS?

[09:33:02.0000] <gsnedders>
I guess they count as two adjacent words

[09:33:03.0000] <Philip`>
I thought implied LWS was everywhere where it's not explicitly disallowed

[09:34:00.0000] <gsnedders>
yeah

[09:34:01.0000] <gsnedders>
it's anywhere between two adjacent words.

[09:34:02.0000] <gsnedders>
but is a repetition two words?

[09:34:03.0000] <gsnedders>
/me shrugs

[09:34:04.0000] <Philip`>
Parsing it as tokens is silly since "(compatible;" isn't a useful token

[09:35:00.0000] <gsnedders>
that's a comment!

[09:35:01.0000] <Philip`>
Oh, it is?

[09:35:02.0000] <gsnedders>
(comment)

[09:35:03.0000] <Philip`>
Oh, okay

[09:36:00.0000] <gsnedders>
that's perfectly valid

[09:36:01.0000] <gsnedders>
and it's why ( and ) aren't allowed in token

[09:37:00.0000] <Philip`>
It makes more sense when I actually read what it says

[09:39:00.0000] <gsnedders>
That's normal.

[09:48:00.0000] <gsnedders>
Philip`: http://http-parsing.gsnedders.com/#first-server-product-tokens

[09:48:01.0000] <gsnedders>
<http://http-parsing.gsnedders.com/#first-server-product-token> even

[09:50:00.0000] <Philip`>
gsnedders: The scroll position gets messed up after it creates the graphics :-(

[09:50:01.0000] <gsnedders>
in Saf3/Mac it's fine :P

[09:51:00.0000] <Philip`>
Are people intentionally sticking with Apache/1.3.37 for as long as possible?

[09:52:00.0000] <gsnedders>
we have two responses from Apache/1.2.6. quite complaining.

[09:52:01.0000] <gsnedders>
*quit]

[09:52:02.0000] <gsnedders>
**quit

[09:55:00.0000] <gsnedders>
why on earth would anyone send ;charset=none

[11:23:00.0000] <zcorpan>
http://tinyurl.com/2fa6ko - html5 content model sketch

[11:23:01.0000] <zcorpan>
Hixie: ^

[11:38:00.0000] <gsnedders>
zcorpan: could you possibly use data URIs for anything more crazy? :)

[11:38:01.0000] <zcorpan>
gsnedders: hmm, how about video? :)

[11:39:00.0000] <gsnedders>
zcorpan: I'll take that as a yes :)

[11:40:00.0000] <hsivonen>
hmm. DOM replaceChild has weird argument order

[11:41:00.0000] <gsnedders>
empty header values are oddly popular.

[11:41:01.0000] <hsivonen>
(and it's in general weird compared to child.replace(replacement))

[11:42:00.0000] <hsivonen>
the tiny url doesn't work in Firefox 2

[11:42:01.0000] <hsivonen>
security?

[11:42:02.0000] <zcorpan>
hsivonen: yeah, i never remember what the right order is of the arguments (re replaceChild)

[11:42:03.0000] <zcorpan>
hsivonen: wfm in firefox 3

[11:43:00.0000] <gsnedders>
Saf3/Mac too

[11:44:00.0000] <kingryan>
doesn't work in camino

[11:44:01.0000] <hsivonen>
doh. I'm missing collapsing-relevant styles on validator.n

[11:44:02.0000] <hsivonen>
u

[11:45:00.0000] <gsnedders>
Philip`: <http://www.nightwing.com.au/> — that has a header to outdo all other headers.

[11:46:00.0000] <kingryan>
gsnedders: the empty one?

[11:46:01.0000] <gsnedders>
yeah

[11:46:02.0000] <gsnedders>
kingryan: we were talking about headers without values earlier

[11:46:03.0000] <zcorpan>
hsivonen: now renumbering works correctly

[11:47:00.0000] <zcorpan>
hsivonen: aha, was that why renumbering didn't work before?

[11:48:00.0000] <hsivonen>
zcorpan: I don't know

[11:48:01.0000] <zcorpan>
hsivonen: that would make sense

[11:48:02.0000] <hsivonen>
zcorpan: I didn't fix anything else yet in deployment

[11:48:03.0000] <Philip`>
gsnedders: http://krijnhoetmer.nl/irc-logs/whatwg/20071213#l-764 - old news :-p

[11:48:04.0000] <gsnedders>
Philip`: you found that one too. damn.

[11:49:00.0000] <hsivonen>
my sandbox copy has not O(n) rewritten message grouping but no collapsing

[11:49:01.0000] <hsivonen>
s/not/now/

[11:49:02.0000] <hsivonen>
there should probably be collapsing of the locations, too

[11:50:00.0000] <zcorpan>
hsivonen: it would be cool (and easy to impl :) ) if the sub lists had the same numbers as they would have when not collapsed, if you're going to do sub lists

[11:51:00.0000] <hsivonen>
zcorpan: hmm. I hadn't thought of that but OK

[11:51:01.0000] <hsivonen>
If I collapse automatically after grouping, I need to provide an "expand" all feature as well as "ungroup"...

[11:52:00.0000] <zcorpan>
hsivonen: it would be a side effect of just moving the <li>s into the sub list. the number also gives a hint about where the problem is in the source

[11:52:01.0000] <hsivonen>
zcorpan: ok. I let the outer list be renumbered

[11:52:02.0000] <hsivonen>
zcorpan: so the outer list counts the number of different messages

[11:52:03.0000] <hsivonen>
zcorpan: and I'll make the inner list items have the ungrouped numbers

[11:53:00.0000] <zcorpan>
hsivonen: ok

[11:53:01.0000] <hsivonen>
zcorpan: btw, I use innerHTML as the comparison key to avoid innerText/textContent differences

[11:54:00.0000] <zcorpan>
hsivonen: i don't know, perhaps having 1,2,3 on the inner lists are better anyway

[11:54:01.0000] <zcorpan>
hsivonen: ok

[11:55:00.0000] <hsivonen>
hmm. is the innerHTML getter interoperable in application/xhtml+xml?

[11:55:01.0000] <hsivonen>
wrong question

[11:56:00.0000] <hsivonen>
does innerHTML in application/xhtml+xml return something stable in each of Firefox/Safari/Opera?

[11:56:01.0000] <zcorpan>
stable as in won't change in the future?

[11:58:00.0000] <hsivonen>
stable as in in the same browser instance return the same string for the same-shaped tree fragment and different strings for different fragments

[11:58:01.0000] <zcorpan>
yes

[11:59:00.0000] <zcorpan>
(unless there's a bug i don't know about)

[11:59:01.0000] <hsivonen>
zcorpan: I implemented list number keeping. it looks a bit weird. I'm not sure if it is useful or just weird

[11:59:02.0000] <hsivonen>
zcorpan: ok. thanks

[11:59:03.0000] <zcorpan>
/me looks

[11:59:04.0000] <zcorpan>
deployed?

[12:00:00.0000] <hsivonen>
zcorpan: not yet

[12:00:01.0000] <zcorpan>
ok

[12:00:02.0000] <zcorpan>
i can imagine that it would be weird. feel free to remove it :)

[12:01:00.0000] <zcorpan>
i had in mind that a given error would have the same number regardless of view, but perhaps it's not that useful

[12:01:01.0000] <zcorpan>
so that authors could refer to them. "I don't understand error 16! help!"

[12:02:00.0000] <hsivonen>
zcorpan: I think I'm leaving it in for a while and see what comments I get on IRC

[12:03:00.0000] <zcorpan>
hsivonen: having the outer list not keeping the numbers probably makes the idea collapse, because then you couldn't refer to a specific error that was in the outer list

[12:08:00.0000] <hsivonen>
zcorpan: no problem :-) the location of the first error instance is moved to the inner list as well

[12:09:00.0000] <zcorpan>
hsivonen: ok. but what about errors that are "alone" and then renumbered (in the outer list)?

[12:10:00.0000] <zcorpan>
or do they also get an inner list?

[12:13:00.0000] <hsivonen>
zcorpan: they get a single-item inner list

[12:13:01.0000] <zcorpan>
hsivonen: aha. then it's all good :)

[12:14:00.0000] <hsivonen>
were the WebKit animatable css properties documented somewhere? in particular, will trasitions to and from display: none; be animatable

[12:14:01.0000] <hsivonen>
?

[12:15:00.0000] <roc_>
there's a thread in www-style with the proposal

[12:16:00.0000] <zcorpan>
display:none animation would be cool

[12:16:01.0000] <hsivonen>
roc_: Hyatt's first email doesn't seem to have a list of animatable properties 

[12:16:02.0000] <roc_>
oh

[12:17:00.0000] <roc_>
I hope 'display' is not animatable

[12:17:01.0000] <roc_>
that sounds both useless and hard to implement

[12:17:02.0000] <zcorpan>
some js libraries support display:none animation, sort of

[12:17:03.0000] <hsivonen>
oh, there's a value-based definition of what's animatable

[12:17:04.0000] <zcorpan>
roc_: collapsing things is often animated

[12:18:00.0000] <roc_>
yeah but you can't exactly have a smooth transition from display:block to display:inline

[12:18:01.0000] <hsivonen>
roc_: when something changes from display: none; to display: block;, it would be cool to have the newly displayed block slide from under the block above

[12:18:02.0000] <zcorpan>
perhaps collapsing annimation can be faked with 'height' animation (with overflow:hidden)

[12:19:00.0000] <hsivonen>
but it kinda sucks to have to use overflow: hidden; height: 0; as a surrogate for display:none

[12:19:01.0000] <zcorpan>
i guess

[12:19:02.0000] <roc_>
not really

[12:19:03.0000] <zcorpan>
a lot of things in css suck :)

[12:19:04.0000] <roc_>
animating to display:none is under-constrained

[12:19:05.0000] <roc_>
which way should the element slide?

[12:20:00.0000] <hsivonen>
I guess whether animating height sucks depends on how well legacy browsers handle overflow: hidden;

[12:20:01.0000] <roc_>
or should it perhaps zoom out from zero size?

[12:20:02.0000] <roc_>
etc

[12:20:03.0000] <hsivonen>
roc_: I see the problem. but wanting to animate collapsing will be a common case

[12:21:00.0000] <hsivonen>
roc_: and currently, display:none is the way things are collapsed

[12:23:00.0000] <Philip`>
If I remember correctly, Apple's proposal was for 'animation of CSS properties', which is distinct from 'animation of HTML documents using CSS'

[12:24:00.0000] <hsivonen>
oh great. more ogg email

[12:25:00.0000] <Philip`>
and so it can't animate nicely between display:block and display:none because there's no intermediate state to animate the property through

[12:25:01.0000] <Philip`>
as a fundamental part of the feature's design

[12:26:00.0000] <Philip`>
(and so you'd need something totally different to get the desired effect)

[12:36:00.0000] <hsivonen>
interesting. in Firefox HTML DOM mode, elements created with createElementNS don't get the DOM uppercasing tagName treatment

[13:53:00.0000] <Hixie>
well last night's ogg flame wars sure were more polite than the previous day's

[13:53:01.0000] <Dashiva>
Is that RoggO guy still around?

[13:55:00.0000] <Hixie>
yeah, a little

[14:04:00.0000] <hsivonen>
now another commentator has inadvertently revealed not havin read the video part of the spec

[14:05:00.0000] <hsivonen>
I wonder what would be the right UI placement for a "group messages" button

[14:05:01.0000] <hsivonen>
clearly, the "Collapse All" button is now in the wrong place

[15:12:00.0000] <Hixie>
zcorpan: i think your proposal is pretty good

[15:12:01.0000] <Hixie>
i think i'm going to take this opportunity to discard the "block" and "inline" terms though

[15:12:02.0000] <Hixie>
in favour of something like "paragraph-level" and "phrasing-level" or something

[15:14:00.0000] <Hixie>
i don't know that we want video to be strictly speaking inline/phrasing-level... after all it has to contain blocks (i agree it's transparent)... consider this:

[15:15:00.0000] <Hixie>
 <section> The screen should now look like <object ...> <pre> <samp> ... </samp> </pre> </object> which...

[15:15:01.0000] <Hixie>
i suppose it's ok for paragraphing to be loosely defined

[15:16:00.0000] <Hixie>
in fact i guess we can throw all of that to the winds and just say that you apply css rules to determine "blocks" for the purposes of rendering, even if you don't support css itself

[15:18:00.0000] <Hixie>
*ponder* *ponder*

[15:31:00.0000] <jruderman>
Hixie: neat, so "block in inline" will unambiguously refer to the CSS layout issue rather than the HTML parsing issue

[15:58:00.0000] <Hixie>
jruderman: yeah


2007-12-14
[01:08:00.0000] <krijnh>
Ping

[01:09:00.0000] <krijnh>
Shitty connection today :/

[01:17:00.0000] <krijnh>
Pong

[03:00:00.0000] <krijnh>
Grmbl

[03:01:00.0000] <hsivonen>
the source line numbering is right

[03:02:00.0000] <krijnh>
Hey, I'm back

[03:02:01.0000] <hsivonen>
but the parser line numbering goes into the weed but there's nothing particularly strange in the source at that point...

[03:10:00.0000] <madness>
Camaban: re: opera, yes, that too

[03:44:00.0000] <Philip`>
"[whatwg] [html5] Unsubcribe me!" -- that's not fair - artificially inflated subscriber numbers are the most useful outcome of the Ogg issue, but it will be lost if people just unsubscribe :-(

[03:44:01.0000] <stijntje>
hahaha wtf

[03:44:02.0000] <stijntje>
:SD

[03:45:00.0000] <stijntje>
"$$ka-ching$$"

[03:47:00.0000] <Lachy>
looks like mo0n_sniper just subscribed to send that pointless message and now wants to unsubscribe

[03:49:00.0000] <othermaciej>
that was pretty sad, yeah

[03:59:00.0000] <hsivonen>
yay. I managed to minimize Sam's dogfood case to 5 lines of XML

[04:00:00.0000] <maikmerten>
stijntje, yeah, I totally love this professional unsubscripe request, too

[04:00:01.0000] <maikmerten>
"hit and run"

[04:00:02.0000] <maikmerten>
*unsubscribe

[04:00:03.0000] <hsivonen>
unsophisticated advocacy doesn't reflect well on any community :-(

[04:01:00.0000] <maikmerten>
yeah, and I always get a slight stomache when seing things like this

[04:02:00.0000] <stijntje>
did I miss something? I don't seem to have the mail Shannon is replying to in his latest

[04:02:01.0000] <maikmerten>
err... stomachache

[04:04:00.0000] <hsivonen>
stijntje: I don't see what (s)he is replying to, either

[04:04:01.0000] <stijntje>
might be a private mail from Dave

[04:05:00.0000] <hsivonen>
in-reply-to points to the same domain as message-id

[04:05:01.0000] <maikmerten>
which reminds me how I often just hit "reply" on a mailing list post and thus accidentially answer off-list

[04:05:02.0000] <stijntje>
that happens to me too

[04:06:00.0000] <maikmerten>
back in ye olde days "reply" would always get back to the list

[04:06:01.0000] <maikmerten>
I'm not sure why all lists seem to have changed this

[04:07:00.0000] <maikmerten>
(I now hit "reply all" and then kill all non-list addresses so people won't get the same thing twice)

[04:08:00.0000] <stijntje>
is Shanon a male or female name, by the way?

[04:09:00.0000] <hsivonen>
stijntje: it is at least female, but I don't know it works as a male name, too

[04:09:01.0000] <stijntje>
hsivonen: thank you

[04:10:00.0000] <maikmerten>
on the stomachache thing: It's nice if people show support for free formats and all... but why can some guys not stick to minimum social standards? (even if impulse and passion drives them away)

[04:10:01.0000] <stijntje>
or read up on the thing about which they're upset :)

[04:12:00.0000] <maikmerten>
(having said that I must confess I also did write some messages to the list I wished I wouldn't have sent)

[07:28:00.0000] <hsivonen>
What foreign namespaces are reasonable in XHTML5 other than SVG, MathML and RDF (plus whatever happens in the RDF subtree)?

[07:30:00.0000] <hsivonen>
is there any good-for-the-Web reason to use XLink in XHTML5?

[07:33:00.0000] <hsivonen>
the XBL2 spec has an example that embeds XBL2 in SVG. is it reasonable to embed XBL2 in XHTML5?

[07:33:01.0000] <Philip`>
There's XLink in SVG subtrees

[07:34:00.0000] <hsivonen>
Philip`: yeah, embedding SVG covers that case

[07:35:00.0000] <annevk>
it's reasonable

[07:36:00.0000] <annevk>
XLink should be avoided when possible

[07:36:01.0000] <hsivonen>
annevk: ok. what's our XBL2 in text/html story?

[07:36:02.0000] <annevk>
CSS

[07:36:03.0000] <annevk>
for now anyway

[07:36:04.0000] <hsivonen>
ok. XHTML5 > HTML5 then :-)

[07:37:00.0000] <annevk>
euh, run

[07:37:01.0000] <hsivonen>
does RDF have some kind of notion of correctness that doesn't involve knowledge about a particular vocabulary/ontology/whatever

[07:37:02.0000] <hsivonen>
?

[07:38:00.0000] <hsivonen>
I'm wondering if I should make RDF a black hole or plug in Jena

[07:38:01.0000] <hsivonen>
(though I'm only going to plug in Jena if it's useful and low-hanging fruit)

[07:39:00.0000] <hsivonen>
is it legitimate to use XHTML5/SVG/MathML elements inside an RDF subtree?

[07:39:01.0000] <annevk>
given how widely RDF is deployed I would prioritize other things personally

[07:39:02.0000] <hsivonen>
annevk: yeah.

[07:39:03.0000] <hsivonen>
black hole it is

[07:52:00.0000] <Dashiva>
RDF is like prolog, it's all backwards

[07:53:00.0000] <Dashiva>
Talking about correctness is tricky

[07:54:00.0000] <gsnedders>
I. am. correct.

[07:54:01.0000] <gsnedders>
That wasn't hard.

[07:54:02.0000] <gsnedders>
/me looks up logs to see what he just said

[08:38:00.0000] <gsnedders>
god. another 51 emails on whatwg.

[08:38:01.0000] <G0k>
i have a solution for you

[08:38:02.0000] <stijntje>
mostly ogg drama

[08:38:03.0000] <G0k>
if contains 'ogg' -> junk

[08:42:00.0000] <gsnedders>
:P

[08:51:00.0000] <oxygenws>
in globalStorage, what happen to IP browsing sites?

[08:59:00.0000] <G0k>
so has anyone even considered implementing the network connections stuff (section 6.3) ?

[09:40:00.0000] <Hixie>
wow, lots of mail

[09:41:00.0000] <inimino>
I have replied to some off-list

[09:43:00.0000] <Hixie>
the internet gods aren't wanting me to reply to this mail

[09:43:01.0000] <Hixie>
every time i try, my connection dies

[09:44:00.0000] <inimino>
it seems there are widely divergent opinions on the coercive value of a SHOULD in the spec

[09:44:01.0000] <Hixie>
it seems pretty academic, since all the browser vendors who would be coerced have already implemented Theora, and the others have said they won't do it

[09:44:02.0000] <inimino>
and perhaps of the value of <video> without a real agreed-upon baseline

[09:45:00.0000] <Hixie>
i think it's pretty clear that <video> without a common codec would be a failure

[09:45:01.0000] <inimino>
right

[09:45:02.0000] <inimino>
I think a token SHOULD mentioning Ogg could be as great of a failure

[09:46:00.0000] <Hixie>
sure, since it wouldn't be a common codec

[09:47:00.0000] <inimino>
specifically if some other format is actually interoperable across any sizeable group of UAs

[09:48:00.0000] <Hixie>
if some format is interoperable, and it doesn't suck too much, it'll be successful

[09:48:01.0000] <Hixie>
regardless of what the spec says

[09:48:02.0000] <inimino>
right

[09:48:03.0000] <inimino>
and regardless of whether it is interoperable by free software

[09:48:04.0000] <inimino>
s/interoperable/implementable/

[10:36:00.0000] <gsnedders>
Flow 5.0: <http://www.engadget.com/2007/12/14/art-installation-made-up-of-hundreds-of-case-fans-is-full-of-air/>

[13:29:00.0000] <Philip`>
Is it bad to write web content using __defineGetter__?

[13:30:00.0000] <Philip`>
(particularly for something around canvas-3d, so I can make some assumptions about what browsers users will have)

[13:31:00.0000] <Dashiva>
It's part of es4, so sure

[13:31:01.0000] <gavin_>
is it part of es3?

[13:32:00.0000] <gavin_>
does it work in safari/opera?

[13:32:01.0000] <Philip`>
I guess I only really care about Firefox 3 and Opera 10 and Safari 4, since my code won't work in anything else

[13:32:02.0000] <gavin_>
/me is just curious

[13:33:00.0000] <Philip`>
I vaguely remembering hearing that Opera 9.5 added it

[13:33:01.0000] <Philip`>
s/ering/er/

[13:37:00.0000] <Philip`>
Woah, my code works :-o

[13:38:00.0000] <Philip`>
/me can now automatically create a shader object to say e.g. shader.lightPosition = [1,2,3] to set the correspondingly-named GLSL variable, instead of lots of ugly getUniformLocation/uniformf calls

[13:47:00.0000] <Dashiva>
gavin: Not in es3, mozilla started it in their js 1.7 or so. Opera added support in 9.5. Not sure about Safari

[13:51:00.0000] <gavin_>
er, __defineGetter__?

[13:51:01.0000] <gavin_>
pretty sure that existed long before js1.7

[13:52:00.0000] <gavin_>
it was landed in Mozilla CVS on 2000-03-02 15:19 with the comment "Added ECMA3 compliant getter/setter syntax."

[13:53:00.0000] <gavin_>
now, that commit message may not mean __defineGetter__ itself

[13:53:01.0000] <gavin_>
but that does suggest that it was pre-1.7 :)

[13:58:00.0000] <Dashiva>
I expect "ECMA3 compliant" means it doesn't extend the syntax, like { get a : function(){} } and function getter a() {}

[14:17:00.0000] <Philip`>
"most users don't install third-party products" ... except for Flash?

[14:17:01.0000] <Hixie>
most users don't install flash, it comes pre-installed

[14:18:00.0000] <Philip`>
Oh, does Windows do that?

[14:18:01.0000] <Hixie>
not sure what the status is these days. it used to, i believe.

[14:18:02.0000] <Hixie>
most OEMs do too

[14:21:00.0000] <Philip`>
Hmm, apparently at least XP does

[14:21:01.0000] <Philip`>
That sounds boringly non-anti-competitive

[14:21:02.0000] <Hixie>
XP is from before SilverLight

[14:21:03.0000] <Hixie>
:-)

[14:21:04.0000] <parcelbrat>
i didn't think MS was allowed to do that

[14:21:05.0000] <Hixie>
anyway, gotta go get lunch

[14:21:06.0000] <parcelbrat>
(be non-anti-ompetitive)

[14:26:00.0000] <Philip`>
parcelbrat: Indeed - I think we should sue them for it

[14:26:01.0000] <parcelbrat>
You and me, sholdn't be a problem ;)

[14:27:00.0000] <Philip`>
/me starts writing the press release and preparing interviews

[14:27:01.0000] <parcelbrat>
/me hides his money

[15:07:00.0000] <Philip`>
/me hates rotateX/rotateY/rotateZ

[15:12:00.0000] <stijntje>
my XP didn't pre-install Flash last time I installed it (2 weeks ago(

[15:12:01.0000] <stijntje>
)

[15:12:02.0000] <Philip`>
/me tries to find someone's affine decomposition code to steal

[15:12:03.0000] <dbaron>
Was Apple's CSS animations proposal discussed on the WHATWG list in addition to www-style?  I sent some comments in response to the www-style post; not sure if I should send them elsewhere.

[15:13:00.0000] <Hixie>
i don't think so

[15:13:01.0000] <Philip`>
/me doesn't think so either

[15:14:00.0000] <Philip`>
/me really doesn't want to port http://www.soe.ucsc.edu/~pang/160/f98/Gems/GemsIV/polar_decomp/Decompose.c to JS

[15:15:00.0000] <othermaciej>
dbaron: no, it wasn't discussed on the WHATWG list

[15:20:00.0000] <parcelbrat>
C to JS OUCH!

[15:22:00.0000] <gsnedders>
hmm…

[15:22:01.0000] <gsnedders>
nobody sending any unacceptable Content-Length values.

[15:23:00.0000] <Philip`>
gsnedders: Are there any >256KB? (I set that as the maximum to accept, so hopefully HttpClient still downloaded all the headers and first 256KB properly and just stopped before downloading the rest, but I'm not positive)

[15:24:00.0000] <Philip`>
(At least, I *think* it was 256KB...)

[15:25:00.0000] <gsnedders>
/me needs to decide what to do with invalid Etag values

[15:26:00.0000] <gsnedders>
"59584b8bdef9e2e442f03cf8945a7229" does not match the invalid value 59584b8bdef9e2e442f03cf8945a7229

[15:26:01.0000] <gsnedders>
Which means the standard quoted-string error handling

[15:26:02.0000] <gsnedders>
…doesn't work

[15:28:00.0000] <Philip`>
/me wonders whether rotateX goes clockwise or anticlockwise

[15:28:01.0000] <Philip`>
/me also wonders which direction is clockwise

[15:29:00.0000] <Philip`>
/me uses his standard 3d graphics approach of guessing blindly, then fiddling with minus signs until it renders correctly

[15:29:01.0000] <gsnedders>
/me points in a clockwise direction

[15:29:02.0000] <gsnedders>
silly IRC limitations

[15:30:00.0000] <gsnedders>
Philip`: max Content-Length value is 9992

[15:31:00.0000] <gsnedders>
(which is amazing small, methinks)

[15:31:01.0000] <Philip`>
If you turn a clock around, clockwise goes backwards, and if you don't have an actual clock face to start with then you can't really tell which way is backwards :-(

[15:31:02.0000] <Philip`>
gsnedders: Try sorting numerically, not alphabetically :-p

[15:31:03.0000] <gsnedders>
Philip`: that was :P

[15:32:00.0000] <gsnedders>
if header["value"] > max:\n\tmax = header["value"]

[15:32:01.0000] <Philip`>
Put some int(...) in there to make sure it's numbers

[15:32:02.0000] <gsnedders>
Philip`: they all are

[15:33:00.0000] <gsnedders>
/me opens the file and finds a larger one straight away

[15:33:01.0000] <gsnedders>
/me wonders how Python does > on a string

[15:34:00.0000] <Philip`>
238242 is the maximum

[15:34:01.0000] <Philip`>
(9992 is the maximum if you do string comparisons)

[15:35:00.0000] <gsnedders>
/me is nowadays used to >, + and the like casting to int/float implicitly

[15:35:01.0000] <Philip`>
Hmm, it looks suspiciously like I dropped anything >256KB

[15:36:00.0000] <gsnedders>
now I get 3487 as the max value with int() :\

[15:36:01.0000] <gsnedders>
oh. wait.

[15:36:02.0000] <gsnedders>
I printed the wrong value.

[15:36:03.0000] <gsnedders>
I printed the last content-length value :P

[15:36:04.0000] <Philip`>
http://www.untraveledroad.com/USA/Nevada/Elko.htm - "Content-Length: 00006330"

[15:37:00.0000] <gsnedders>
well, as a normal integer, all leading zeros are pointless.

[15:37:01.0000] <Philip`>
Are they allowed in theory/practice?

[15:38:00.0000] <gsnedders>
in theory allowed, in practice as far as I know they are too


2007-12-15
[16:08:00.0000] <Philip`>
/me wishes Opera had uneval

[16:18:00.0000] <Philip`>
Argh, rotateX uses degrees :-(

[16:19:00.0000] <Philip`>
Hooray, now I have a monkey and a box in Opera which looks about the same as in Firefox, except it's totally unshaded

[16:49:00.0000] <Philip`>
Now my monkey is shaded approximately the same in both, except rubbisher and slower in Opera

[16:49:01.0000] <Philip`>
but at least it works, which is handy

[17:19:00.0000] <Philip`>
http://philip.html5.org/demos/canvas/3d/x3d/test.html - should work in Firefox 3 beta with Canvas 3D extension and OpenGL 2.0 hardware; and in Opera 9.5 builds with opera-3d

[17:26:00.0000] <Philip`>
http://philip.html5.org/demos/canvas/3d/x3d/output-ff.png vs http://philip.html5.org/demos/canvas/3d/x3d/output-opera.png

[17:27:00.0000] <Philip`>
(The Opera version is several times slower, and has much more hardcoded lighting stuff)

[17:27:01.0000] <Philip`>
(and only does per-vertex lighting, not per-pixel)

[17:27:02.0000] <hdh>
wow, blender's monkey head

[17:27:03.0000] <Philip`>
Indeed :-)

[22:48:00.0000] <jacobolus>
Hixie: around?

[22:48:01.0000] <jacobolus>
I notice that the server-sent events things now have a big "might remove this" notice.  does that mean altered, or scrapped altogether?

[22:49:00.0000] <Hixie>
either

[22:49:01.0000] <Hixie>
they're kinda redundant with xhr with multipart, and with the bidirectional networking stuff

[22:49:02.0000] <jacobolus>
Hixie: you see http://cometdaily.com/2007/12/11/the-future-of-comet-part-1-comet-today/ ?

[22:50:00.0000] <jacobolus>
Hixie: I have most of an article about server-sent events as well, talking about opera's implementation

[22:50:01.0000] <jacobolus>
also, apparently there are issues w/ multipart xhr 

[22:50:02.0000] <jacobolus>
namely, there's no notice when connections are severed

[22:51:00.0000] <Hixie>
yeah we'd have to make sure we covered all the bases

[22:51:01.0000] <Hixie>
if we don't remove event-source, we'll probably simplify it a lot

[22:51:02.0000] <Hixie>
right now it's way overengineered

[22:53:00.0000] <jacobolus>
Hixie: yeah, that's a big part of my article :)

[22:53:01.0000] <jacobolus>
which is mostly actually about getting more ppl to send you feedback :)

[22:54:00.0000] <Hixie>
:-)

[22:54:01.0000] <Hixie>
to be honest even if we really want event-source and even if it was perfect, we might still have to remove them

[22:54:02.0000] <Hixie>
depends on whether browsers implement it or not

[22:54:03.0000] <jacobolus>
Hixie: if you have time to read this, let me know what you think :) (assuming the thing ever pastes)

[22:54:04.0000] <Hixie>
that goes for every feature of course

[22:54:05.0000] <Hixie>
the article?

[22:54:06.0000] <Hixie>
i read it

[22:55:00.0000] <jacobolus>
yeah, trying to paste to a pastebin

[22:55:01.0000] <jacobolus>
it's stalling

[22:55:02.0000] <jacobolus>
/me pasted http://pastie.textmate.org/private/xsvmjyopvftavgbcbjpexa

[22:55:03.0000] <jacobolus>
no no, the draft of the next one

[22:56:00.0000] <jacobolus>
Hixie: ↑

[22:56:01.0000] <Hixie>
/me finishes reading it

[23:03:00.0000] <jacobolus>
hsivonen: I did a web search for "java applet suck", and a comment you made in here came up in the top page of google results :)

[23:06:00.0000] <jacobolus>
bb5

[23:20:00.0000] <jacobolus>
Hixie: thoughts?

[23:22:00.0000] <jacobolus1>
Hixie: hrm, i was apparently disconnected.  did i miss anything?

[23:36:00.0000] <jacobolus>
well.  off to sleep.  bb later sometime :)

[01:17:00.0000] <othermaciej>
Hixie: people have started on event-source for WebKit, so it would be good to have revision/simplification sooner rather than later

[01:17:01.0000] <othermaciej>
Hixie: I do tend to agree that sending arbitrary server-sent events to arbitrary nodes seems like overkill

[01:51:00.0000] <MikeSmith>
hsivonen - you around?

[01:52:00.0000] <MikeSmith>
I thought I remembered hsivonen saying that he had added some support to html5.validator.nu for checking for allowed values for the rel attribute based on scraping http://wiki.whatwg.org/wiki/RelExtensions

[01:58:00.0000] <hsivonen>
MikeSmith: I didn't add it yet. just discussed options with Hixie

[01:58:01.0000] <MikeSmith>
OK

[01:59:00.0000] <MikeSmith>
hsivonen - so html5.validator.nu isn't doing an conformance checking on rel values now, right?

[01:59:01.0000] <hsivonen>
MikeSmith: right

[01:59:02.0000] <MikeSmith>
OK, thanks

[01:59:03.0000] <hsivonen>
afk. switchin locations. will be back in 45 minutes

[02:03:00.0000] <hsivonen>
back here. bus gone already

[07:32:00.0000] <zcorpan_lap>
hsivonen: the group messages feature is great :)

[07:33:00.0000] <hsivonen>
zcorpan_lap: thanks

[07:34:00.0000] <hsivonen>
zcorpan_lap: re: <semantics>: annotation/annotation-xml are so under-defined that I seriously doubt that there are useful implementations 

[07:35:00.0000] <zcorpan_lap>
hsivonen: ok

[07:36:00.0000] <hsivonen>
I think I'm going to leave annotation-xml as an anything-goes black hole for now

[07:36:01.0000] <hsivonen>
like RDF...

[07:38:00.0000] <zcorpan_lap>
hsivonen: http://validator.nu/?doc=http%3A%2F%2Fwww.w3.org%2F&schema=http%3A%2F%2Fsyntax.whattf.org%2Frelaxng%2Fhtml5full.rnc+http%3A%2F%2Fsyntax.whattf.org%2Frelaxng%2Fassertions.sch+http%3A%2F%2Fhsivonen.iki.fi%2Fcheckers%2Fall%2F&showsource=yes -- the location link for error 2 doesn't point to the right place

[07:38:01.0000] <zcorpan_lap>
s/error/message/

[07:39:00.0000] <hsivonen>
Hixie: IIRC you said that RDF is allowed where metadata elements are allowed. are there other foreign metadata elements that I should allow in XHTML5?

[07:40:00.0000] <hsivonen>
zcorpan_lap: hmm. strange. I have a parser fix in my tree that I couldn't deploy yesterday. I try deploying now

[07:41:00.0000] <zcorpan_lap>
hsivonen: the link points to cl2c109 but the id is l2c109

[07:41:01.0000] <hsivonen>
ooh. a front-end bug

[07:46:00.0000] <hsivonen>
zcorpan_lap: fixed. thanks

[07:46:01.0000] <hsivonen>
(deployed too)

[07:51:00.0000] <hsivonen>
can rdf:RDF legitimately have non-xmlns attributes?

[07:52:00.0000] <hsivonen>
Hixie: should only rdf:RDF be allowed as an RDF subtree root or should rdf:Description be allowed as well?

[07:53:00.0000] <zcorpan_lap>
hsivonen: ie doesn't know about HTMLOListElement

[07:54:00.0000] <hsivonen>
zcorpan_lap: ok. what test would you suggest instead?

[07:54:01.0000] <hsivonen>
I have a hard time interpreting http://www.w3.org/TR/rdf-syntax-grammar/#start

[07:55:00.0000] <zcorpan_lap>
hsivonen: .tagName.toLowerCase() == "ol"

[07:55:01.0000] <hsivonen>
zcorpan_lap: not theoretically pure, but OK

[07:56:00.0000] <zcorpan_lap>
hsivonen: indeed. not sure what else would work cross-browser

[07:57:00.0000] <hsivonen>
zcorpan_lap: does IE support nodeName? it would be more robust in case I ever put whitespace there

[07:58:00.0000] <zcorpan_lap>
hsivonen: it does

[07:58:01.0000] <zcorpan_lap>
hsivonen: you could check .start

[07:58:02.0000] <zcorpan_lap>
returns 1 if <ol> otherwise undefined

[07:59:00.0000] <zcorpan_lap>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Col%3E.%3Cscript%3Ew(document.body.firstChild.start)%3C%2Fscript%3E

[07:59:01.0000] <hsivonen>
hmm. I'm not sure which test is theoretically less pure...

[08:00:00.0000] <zcorpan_lap>
firefox returns -1 though

[08:02:00.0000] <hsivonen>
zcorpan_lap: yeah, I just noticed it broke firefox

[08:02:01.0000] <zcorpan_lap>
but -1 is truthy so you could just do if (n.start)

[08:02:02.0000] <hsivonen>
I did if (n.start) and it broke

[08:04:00.0000] <zcorpan_lap>
weird

[08:04:01.0000] <hsivonen>
hmm. file copying was what broke

[08:04:02.0000] <zcorpan_lap>
ok

[08:04:03.0000] <hsivonen>
now the script is on the server for real

[08:04:04.0000] <hsivonen>
zcorpan_lap: thanks

[08:05:00.0000] <zcorpan_lap>
np

[08:05:01.0000] <hsivonen>
XHTML in Safari still broken

[08:06:00.0000] <zcorpan_lap>
i think that's document.body

[08:06:01.0000] <zcorpan_lap>
hmm, now ie7 crashes

[08:06:02.0000] <zcorpan_lap>
:)

[08:13:00.0000] <hsivonen>
zcorpan_lap: any advice on what I need to do to stop it from crashing?

[08:13:01.0000] <zcorpan_lap>
i don't know what it is that makes it crash

[08:14:00.0000] <zcorpan_lap>
it's not the .start check, i think

[08:15:00.0000] <hsivonen>
I think I'm doing relatively tame DOM operations

[08:15:01.0000] <zcorpan_lap>
it's not createHtmlElement either

[08:18:00.0000] <zcorpan_lap>
hsivonen: you may need to set .type = 'button' before you append the input to the p

[08:19:00.0000] <zcorpan_lap>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cscript%3E%20function%20createHtmlElement(tagName)%7B%0D%0A%20%20%20%20return%20document.createElementNS%20%3F%20document.createElementNS(%22http%3A%2F%2Fwww.w3.org%2F1999%2Fxhtml%22%2C%20tagName)%20%3A%20document.createElement(tagName)%0D%0A%7D%0D%0Avar%20currentOl%20%3D%20document.body.firstChild%3B%0D%0Afunction%20installGrouping

[08:19:01.0000] <zcorpan_lap>
a%20%3D%20createHtmlElement('p')%0D%0A%09var%20button%20%3D%20createHtmlElement('input')%0D%0A%09para.appendChild(button)%0D%0A%09button.type%20%3D%20'button'%0D%0A%7D%0D%0AinstallGroupingToggle()%0D%0A%3C%2Fscript%3E

[08:20:00.0000] <hsivonen>
zcorpan_lap: better now (after reload)?

[08:21:00.0000] <zcorpan_lap>
still crashes

[08:22:00.0000] <hsivonen>
zcorpan_lap: you were right about document.body and Safari

[08:28:00.0000] <zcorpan_lap>
/me has no idea what makes ie crash :(

[08:29:00.0000] <hsivonen>
zcorpan_lap: does it crash onload?

[08:29:01.0000] <hsivonen>
that is, before pushing the button?

[08:29:02.0000] <zcorpan_lap>
yes

[08:29:03.0000] <zcorpan_lap>
but not before you've validated something

[08:30:00.0000] <zcorpan_lap>
i.e. just loading http://validator.nu/ works fine

[08:30:01.0000] <hsivonen>
well, crashing isn't user-friendly

[08:30:02.0000] <zcorpan_lap>
:)

[08:30:03.0000] <hsivonen>
I guess it's better if I just disable the script for IE

[08:31:00.0000] <hsivonen>
I wonder what's the best way to *really* detect IE without depriving other browsers of the script

[08:31:01.0000] <hsivonen>
should I just put the instanceof thing back?

[08:32:00.0000] <zcorpan_lap>
that would work

[08:32:01.0000] <zcorpan_lap>
the safest way otherwise, i think, is conditional comments (either in html or js)

[08:35:00.0000] <hsivonen>
zcorpan_lap: I put instanceof back

[08:35:01.0000] <Philip`>
Can you report the bug to the IE developers?

[08:35:02.0000] <hsivonen>
Philip`: me? I haven't even witnessed it firsthand

[08:37:00.0000] <Philip`>
hsivonen: I guess "you" means "someone who can reproduce the bug"

[08:37:01.0000] <Philip`>
/me doesn't know if there's still any public bug reporting mechanism

[08:38:00.0000] <zcorpan_lap>
"Now that the final version of IE7 has been released, the IE Feedback site is temporarily closed. It will re-open in the future."

[08:38:01.0000] <zcorpan_lap>
it has been "temporarily closed" the past year or so

[08:39:00.0000] <hsivonen>
did anne write an XBL2 schema in RELAX NG?

[08:40:00.0000] <hsivonen>
/me works on a massive überschema reorganization

[08:51:00.0000] <hsivonen>
/me has trouble getting trang to cooperate...

[09:23:00.0000] <hasather>
hsivonen: the schemas are here: http://dev.w3.org/cvsweb/2006/waf/xbl-schemas/

[09:23:01.0000] <hsivonen>
hasather: thanks

[09:27:00.0000] <hsivonen>
hmm. anne is not here. that xbl2 schema lacks license info

[09:30:00.0000] <hsivonen>
has Hixie really specced XBL id to take the value constraints of xsd:ID or has Anne taken liberties with the schema?

[09:34:00.0000] <hasather>
hsivonen: as far as I know, that was never specced

[09:35:00.0000] <hsivonen>
hasather: OK

[09:37:00.0000] <hasather>
hsivonen: I may be wrong though, but I remember there were some discussion around ID in XBL

[09:38:00.0000] <hsivonen>
hasather: there were at least the discussions where nwalsh said it should use xml:id

[09:38:01.0000] <hasather>
yea

[09:49:00.0000] <gsnedders>
Old glow sticks are so boring.

[10:47:00.0000] <jruderman>
gsnedders: break it open and pour it in homestar's mountain dew

[11:26:00.0000] <hsivonen>
hmm. the message grouping script has a silly bug when source is shown and there are no messages...

[11:34:00.0000] <zcorpan_lap>
hsivonen: i guess you could just set an id on the messages list

[11:34:01.0000] <zcorpan_lap>
although if you replace instanceof with getElementById then you'll probably get the ie crasher again :)

[11:35:00.0000] <hsivonen>
zcorpan_lap: I wanted to avoid issues like that :-)

[11:40:00.0000] <hsivonen>
let me say again that RELAX NG DTD Compatibility sucks

[11:47:00.0000] <hsivonen>
w00t. my local copy validates intertwingly.net/blog

[11:47:01.0000] <hdh>
what can I use instead of XFN?

[11:48:00.0000] <hsivonen>
hdh: for what purpose?

[11:48:01.0000] <hdh>
in html5 the rel values is a limited set → no XFN

[11:48:02.0000] <hsivonen>
hdh: please edit the wiki :-)

[11:52:00.0000] <hdh>
"Unless otherwise specified, a keyword must not be specified more than once per rel attribute." ← "keyword" here means link type?

[11:53:00.0000] <hsivonen>
yes

[12:08:00.0000] <hdh>
hm, unknown rel types aren't conformance error, and the spec does say that "if none of the values in the attribute are recognised by the UA", i.e. allowing types outside it

[12:08:01.0000] <hdh>
s/that//

[12:13:00.0000] <Philip`>
hdh: 4.12.3.21 says "Other than the types defined above, only types defined as extensions in the WHATWG Wiki RelExtensions page may be used with the rel attribute on link, a, and area elements."

[12:14:00.0000] <Philip`>
so it sounds like it's intending that unknown types (i.e. not in spec nor wiki) are an error

[12:14:01.0000] <Philip`>
though it should probably say "must" somewhere, and I can't see that it does

[12:16:00.0000] <hsivonen>
any opinions on what I should do with inkscape/sodipodi product-specific cruft?

[12:17:00.0000] <hdh>
search for prefix:, no hit → delete xmlns: ; remove all style= and test if things look ugly afterwards; run throught xmllint to drop whitespace

[12:17:01.0000] <hdh>
I haven't tried the second one yet

[12:18:00.0000] <hsivonen>
hdh: I meant: what should the validator do

[12:18:01.0000] <Philip`>
hsivonen: It would be non-helpful if it was an error, since there's no point in authors cleaning it up

[12:18:02.0000] <Philip`>
(as far as I'm aware)

[12:18:03.0000] <hdh>
warn as unknown namespace?

[12:18:04.0000] <hsivonen>
Philip`: will I be product-biased if I allow sodipodi/inkscape cruft specifically?

[12:19:00.0000] <Philip`>
though maybe it could have one warning saying "you're wasting some space with this stuff - try the 'export as plain SVG' option"

[12:19:01.0000] <hsivonen>
or of course I would, but would I be bad?

[12:20:00.0000] <Philip`>
hsivonen: I don't see why it would be bad, as long as you gave the same allowance for any other product that people use and which is harmless

[12:20:01.0000] <Philip`>
(where "that people use" could mean "that people complain to you about")

[12:21:00.0000] <hsivonen>
Philip`: ok

[12:21:01.0000] <Philip`>
I could be wrong, of course :-)

[12:21:02.0000] <hsivonen>
/me wonders if that's a slippery slope to allowing RDFa in XHTML5 :-)

[12:22:00.0000] <Philip`>
Hmm, can you validate RDF that's inside HTML comments?

[12:22:01.0000] <hsivonen>
Philip`: I don't

[12:23:00.0000] <hsivonen>
Philip`: however, I'm right now punching holes for real RDF in the XHTML5 and SVG 1.1 schemas

[12:26:00.0000] <Philip`>
/me wonders if anyone would want to embed X3D inline in XHTML

[12:26:01.0000] <Philip`>
(That would be easier if it had a namespace, I suppose)

[12:27:00.0000] <Philip`>
(but it's easy to make one up)

[12:28:00.0000] <zcorpan_lap>
hsivonen: it seems weird to allow sodipodi/inkscape cruft if it's not allowed per the svg spec

[12:30:00.0000] <Philip`>
"SVG allows inclusion of elements from foreign namespaces anywhere with the SVG content. In general, the SVG user agent will include the unknown elements in the DOM but will otherwise ignore unknown elements."

[12:30:01.0000] <Philip`>
"Additionally, SVG allows inclusion of attributes from foreign namespaces on any SVG element. The SVG user agent will include unknown attributes in the DOM but with otherwise ignore unknown attributes."

[12:30:02.0000] <Philip`>
That sounds like it's allowed

[12:30:03.0000] <Philip`>
(via http://www.w3.org/TR/SVG11/extend.html )

[12:31:00.0000] <zcorpan_lap>
ok. then you wouldn't have to have knowledge about sodipodi/inkscape cruft specifically

[12:32:00.0000] <Philip`>
You would have to if you want to whitelist known extensions, since otherwise you couldn't detect people incorrectly using xmlns="http://www.w3.org/1999/html" or whatever

[12:32:01.0000] <hsivonen>
I'm not a fan of anything-goes holes

[12:32:02.0000] <hsivonen>
Philip`: indeed

[12:33:00.0000] <hsivonen>
btw, the above "ignore" text doesn't say if ignoring an element means ignoring the subtree or behaving as if the content of the element appeared in place of the element

[12:34:00.0000] <hsivonen>
I'm inclined to restrict foreignObject to XHTML and MathML only

[12:35:00.0000] <hdh>
is xfn not allowed on <link>?

[12:35:01.0000] <hsivonen>
hdh: it is if it's on the wiki. if not, it will be once you've edited the wiki

[12:35:02.0000] <hdh>
I'm editing it, but not sure what to put in the Effect on link column

[12:36:00.0000] <hsivonen>
"(Other XML-compatible metadata languages, including ones not based on RDF, can be used also.)" Any examples of "other"?

[12:36:01.0000] <zcorpan_lap>
is there a spec for xfn?

[12:36:02.0000] <hdh>
http://gmpg.org/xfn/11 

[12:36:03.0000] <hsivonen>
http://gmpg.org/xfn/11

[12:37:00.0000] <hsivonen>
I'm slow

[12:37:01.0000] <zcorpan_lap>
thanks

[12:37:02.0000] <hdh>
actually a profile, which is gone in html5

[12:38:00.0000] <hsivonen>
doesn't matter

[12:38:01.0000] <hsivonen>
no one checks the profile attribute upon processing

[12:39:00.0000] <Philip`>
I thought GRDDL does

[12:40:00.0000] <Philip`>
(since the profile points to some XSLT to convert the XHTML to RDF, according to this diagram)

[12:40:01.0000] <hsivonen>
ok, someone checks

[12:40:02.0000] <hsivonen>
except the XFN profile isn't XSLT. oops.

[12:41:00.0000] <hdh>
and xslt is a stylesheet

[12:41:01.0000] <hsivonen>
a "transformation" would be more politically correct on this channel :-)

[12:42:00.0000] <Philip`>
http://www.w3.org/2003/12/rdf-in-xhtml-xslts/grokXFN.xsl

[12:42:01.0000] <hdh>
aye, I'm a newcomer

[12:43:00.0000] <hsivonen>
Philip`: is there some mapping magic from a gmpg.org URI to that URI?

[12:45:00.0000] <Philip`>
hsivonen: Don't know - I'm not sure if the aim is to put the XSLT version at the gmpg.org URI, or to have an external mapping that doesn't need cooperation, or something else

[12:46:00.0000] <Philip`>
I vaguely remembering hearing something about those possibilities, except I've forgotten it all now

[12:46:01.0000] <Philip`>
s/remembering/remember/

[12:46:02.0000] <hsivonen>
Does illustrator put any private-namespace cruft in SVG?

[12:47:00.0000] <gsnedders>
hsivonen: no

[12:47:01.0000] <hsivonen>
gsnedders: thanks

[12:47:02.0000] <gsnedders>
(at least, I've never seen so in CS3)

[12:50:00.0000] <Philip`>
The examples at http://www.adobe.com/svg/demos/samples.html don't have any public namespaces (and aren't well-formed) either :-/

[12:50:01.0000] <Philip`>
and have Content-Type: text/plain

[12:51:00.0000] <hsivonen>
XML is hard.

[12:52:00.0000] <Philip`>
(Doesn't look like those samples come from Illustrator anyway)

[12:52:01.0000] <hsivonen>
I'd like to see real-world cases of RDF inside SVG inside XHTML

[12:54:00.0000] <Philip`>
http://www.adobe.com/svg/svgfiles/interactive_filters.svgz

[12:54:01.0000] <gsnedders>
Philip`: how come on http://tinyurl.com/2dsa9f line 105 it is returned before it is set on line 98? Am I getting something wrong with how variables bounce around from function to function?

[12:54:02.0000] <Philip`>
is apparently from Illustrator

[12:54:03.0000] <Philip`>
<svg  xmlns:i="&ns_ai;" xmlns:xlink="&ns_xlink;" xmlns:a="http://ns.adobe.com/AdobeSVGViewerExtensions/3.0/" width="400" height="304.9" viewBox="0 0 400 304.9" style="overflow:visible;enable-background:new 0 0 400 304.9" xml:space="preserve">

[12:54:04.0000] <Philip`>
and <!ENTITY ns_ai "http://ns.adobe.com/AdobeIllustrator/10.0/"> etc

[12:54:05.0000] <hdh>
er, the second newest blog post from Shelley

[12:55:00.0000] <hsivonen>
hdh: thanks

[12:56:00.0000] <Philip`>
The <svg> is still in no namespace, but at least Firefox displays the XML tree instead of a parse error

[12:56:01.0000] <hdh>
/me 's firehose has been planet intertwingly lately

[12:56:02.0000] <Philip`>
/me wonders if Adobe cares about anything other than the Adobe SVG Viewer

[12:56:03.0000] <hsivonen>
Philip`: eww. private elements

[12:56:04.0000] <hsivonen>
no fun

[12:57:00.0000] <gsnedders>
/me opens Ai CS3

[12:58:00.0000] <Philip`>
gsnedders: XMLHttpRequest is asynchronous, so .send() returns before it's sent the request and called onreadystatechange and set ok

[12:58:01.0000] <zcorpan_lap>
hmm

[12:59:00.0000] <gsnedders>
hsivonen: CS3 has SVG and Xlink namespaces only, no DOCTYPE

[12:59:01.0000] <hsivonen>
hmm. Shelley relies on entities in XML on the Web...

[12:59:02.0000] <zcorpan_lap>
if illustrator doesn't emit xmlns="...svg", then it wouldn't be helpful to require xmlns="...svg" for text/html

[13:00:00.0000] <gsnedders>
Philip`: will setting it to false break anything?

[13:00:01.0000] <Philip`>
gsnedders: Setting what to false?

[13:00:02.0000] <gsnedders>
async

[13:02:00.0000] <Philip`>
gsnedders: Hmm, I guess it'd probably work if you change both requests to be synchronous, though you might end up with complaints about scripts taking too long to run

[13:03:00.0000] <hsivonen>
SVG element metadata not allowed as child of SVG element svg in this context. 

[13:03:01.0000] <hsivonen>
I wonder why that happens...

[13:03:02.0000] <hsivonen>
Bad value 1.0 for attribute version on SVG element svg.

[13:03:03.0000] <hsivonen>
that's useful... Not

[13:04:00.0000] <hsivonen>
"The first occurrance of the ID value defs3 was here." hazards of mere inclusion...

[13:05:00.0000] <Philip`>
s/occurrance/occurrence/

[13:08:00.0000] <Philip`>
From http://burningbird.net/technology/standards-take-work/ : "sodipodi:docbase="/Users/richardthompson/Desktop"" - that looks like unintentional information leakage

[13:08:01.0000] <Dashiva>
Say hsivonen, would the validator be able to validate DOM trees too (in a suitable serialization format, like the one zcorpan was working on)?

[13:08:02.0000] <hsivonen>
Philip`: thanks. now fixed in my local copy (not deployed)

[13:09:00.0000] <hsivonen>
Dashiva: if you contribute a parser for zcorpan's format to SAX

[13:10:00.0000] <Philip`>
Dashiva: (Why not a suitable serialisation format like XML?)

[13:10:01.0000] <Dashiva>
Because not all DOM trees can be represented as XML

[13:10:02.0000] <Philip`>
Ah

[13:10:03.0000] <hsivonen>
Dashiva: it would be nice to have a Firefox extension that'd take a snapshot of the DOM and would send in to Validator.nu

[13:11:00.0000] <Dashiva>
Yeah, I got the idea from opera's validation option, and thought to extend it to a page after e.g. onload scripts had done their magic

[13:11:01.0000] <Philip`>
It would be nicer to have one that does local validation, so it doesn't send e.g. your filled-in password and credit card number fields to Validator.nu

[13:12:00.0000] <Dashiva>
But how big is the validator, with all dependencies? :)

[13:12:01.0000] <Philip`>
Do you count the JVM as a dependency? :-)

[13:12:02.0000] <hsivonen>
Dashiva: it's big

[13:13:00.0000] <hsivonen>
and there's the JVM startup delay

[13:13:01.0000] <Philip`>
Hopefully you would only have the startup delay when launching Firefox the first time, and ideally that would happen asynchronously

[13:14:00.0000] <Philip`>
s/the first time//

[13:16:00.0000] <hdh>
/me just noticed the RDF tags for Shelley's tree in akregator

[13:16:01.0000] <Dashiva>
Might be better to simply run a pre-hook that removes value attributes from password fields :)

[13:17:00.0000] <hsivonen>
Dashiva: if you are OK with non-XMLizable DOM proclaimed invalid, the current Web service API goes a long way with XHR

[13:17:01.0000] <Philip`>
After you've sent your hundred-kilobyte document the validator a few dozen times, you'd be better off in terms of bandwidth if you downloaded a local copy plus all the dependencies

[13:17:02.0000] <hsivonen>
btw, what does XHR do if you try to POST a tree that doesn't serialize to XML?

[13:18:00.0000] <Philip`>
(particularly since developers will run this from their own computers, so it's not transmitting from the web server, so it'll be a slower network connection)

[13:19:00.0000] <hsivonen>
If someone wants to write a .zip packaging action for build/build.py, I'd be happy to integrate it

[13:23:00.0000] <zcorpan_lap>
setting opera:config#UserPrefs|ValidationURL to http://validator.nu/ seems to work

[13:23:01.0000] <hsivonen>
zcorpan_lap: cool

[13:23:02.0000] <hdh>
very

[13:23:03.0000] <hsivonen>
unintentional compatibility

[13:24:00.0000] <zcorpan_lap>
although encoding information is lost

[13:24:01.0000] <zcorpan_lap>
hsivonen: it would be useful to be able to suggest and/or override encoding

[13:25:00.0000] <hsivonen>
zcorpan_lap: on file upload?

[13:25:01.0000] <hsivonen>
zcorpan_lap: or in general_

[13:25:02.0000] <hsivonen>
?

[13:25:03.0000] <zcorpan_lap>
at least file upload

[13:26:00.0000] <hsivonen>
zcorpan_lap: feature request recorded. thanks

[13:26:01.0000] <hdh>
web developer "validate local html" doesn't have uri setting, but the other one do

[13:28:00.0000] <zcorpan_lap>
/me tried to enter http://validator.nu/?showsource=yes in opera:config but it didn't have the desired effect :(

[13:30:00.0000] <hsivonen>
zcorpan_lap: the query string is ignored in the file upload and textarea modes

[13:31:00.0000] <zcorpan_lap>
hsivonen: why?

[13:34:00.0000] <hsivonen>
zcorpan_lap: there's different code that reads params from multipart/form-data

[13:34:01.0000] <hsivonen>
zcorpan_lap: recorded as a feature request

[13:34:02.0000] <zcorpan_lap>
hsivonen: ok. thanks

[13:46:00.0000] <gsnedders>
Philip`: you're original only ever could be waiting on one thing at a time anyway

[13:47:00.0000] <gsnedders>
Philip`: making it synchronous would have no difference from your original

[13:47:01.0000] <Philip`>
From X3D / H-Anim: "float[4] rotation 0 0 1 0" ... "The rotation field specifies a rotation of the coordinate system of the Joint object as a quaternion." - it would be so much better if they actually defined quaternions and how they're represented, rather than making you guess whether it's xyzw or wxyz

[13:48:00.0000] <Philip`>
gsnedders: The original asynchronous one lets the browser get on with other things in the meantime, whereas the synchronous one prevents the browser from doing anything much to the page while the script is running and waiting

[13:48:01.0000] <gsnedders>
Philip`: yeah, I know. But how the original version of the script is written, it never tries to load more than one file at once

[13:48:02.0000] <othermaciej>
Philip`: quaternion! crazy

[13:49:00.0000] <Philip`>
gsnedders: The server is single-threaded so it can only load one file at a time anyway :-)

[13:49:01.0000] <gsnedders>
Philip`: the browser get on with what other things? surely other things outwith JS are unaffected by it?

[13:50:00.0000] <gsnedders>
(forgive my naïvety about this)

[13:50:01.0000] <Philip`>
/me actually likes quaternions

[13:51:00.0000] <Philip`>
gsnedders: Maybe things like noticing DOM changes (e.g. when log messages are added) and re-rendering, except I don't know which browsers get hung up on that in practice

[13:52:00.0000] <hsivonen>
I was excited about the RNC "external" keyword when I learned about it but now I'm disappointed because I don't know how to make circular references

[13:52:01.0000] <gsnedders>
Philip`: but if the script isn't running, we don't add log mesages

[13:53:00.0000] <gsnedders>
/me wonders whether to make the log an ordered list

[13:53:01.0000] <Philip`>
gsnedders: The script is 'running' in the sense that it's in the middle of executing a JS function, while waiting for the synchronous XHR, and it never stops running until every test has been processed

[13:53:02.0000] <gsnedders>
Philip`: I mean, the script isn't moving forward.

[13:53:03.0000] <gsnedders>
Philip`: so therefore we don't add log messages, we're just waiting at one line

[13:54:00.0000] <gsnedders>
Philip`: but you think that browsers don't actually check the DOM till JS has totally finished executing? :\

[13:55:00.0000] <hsivonen>
is SVG 1.1 Full a superset of SVG 1.0 and SVG 1.2 Full a superset of SVG 1.1 Full?

[13:55:01.0000] <Philip`>
gsnedders: It does (I assume) something effectively like "function run_tests() { log('Test 1'); xhr_request_stuff(); log('Test 2'); xhr_request_stuff(); ... }" and it doesn't return until everything's finished, so the browser doesn't get a chance to dispatch events (since it has to pretend the script is running single-threadedly)

[13:55:02.0000] <Philip`>
...I think

[13:56:00.0000] <Philip`>
gsnedders: I think Opera and Firefox vary in whether/when they redraw the screen in response to DOM changes if a script is still running, but I can't remember the details now

[13:56:01.0000] <gsnedders>
Silly browsers.

[13:56:02.0000] <Philip`>
They'd be sillier if they allowed arbitrary concurrency between scripts :-)

[13:57:00.0000] <hsivonen>
any opinions on what I should do with <svg version='1.0'>? complaining about "1.0" seem useless

[13:58:00.0000] <hsivonen>
but then silently accepting version='1.0' with 1.1 features seems bogus, too

[13:58:01.0000] <gsnedders>
Philip`: events in the readystatechange event sense, or?

[13:58:02.0000] <Philip`>
hsivonen: Use a SVG 1.0 schema?

[13:59:00.0000] <Philip`>
gsnedders: Yes, but I guess readystatechange in synchronous XHR is special

[14:00:00.0000] <hsivonen>
Philip`: is there one that a) I don't need to write, b) has reasonable quality and c) doesn't conflict with the production names from the 1.1 schema?

[14:00:01.0000] <Philip`>
hsivonen: No idea. Those do sound like possible difficulties

[14:00:02.0000] <hsivonen>
point c) is particularly noteworthy if I'm expected to support 7 flavors of SVG

[14:01:00.0000] <Philip`>
Can't you have namespaced schemas? :-)

[14:02:00.0000] <hsivonen>
Philip`: I could, but it's a non-trivial find-and-replace job

[14:02:01.0000] <hsivonen>
or I could do that step in RNG and only then convert to RNC

[14:03:00.0000] <hsivonen>
what I really want is an SVG5 that gives me a license to treat version='' as a talisman

[14:08:00.0000] <othermaciej>
does SVG actually require particular treatment of version?

[14:08:01.0000] <Philip`>
X3D says "a script url prefixed with ecmascript: (or the deprecated javascript:) shall contain ECMAScript source" - is ecmascript: used by anything else?

[14:08:02.0000] <othermaciej>
no

[14:08:03.0000] <hsivonen>
othermaciej: I don't know

[14:08:04.0000] <Philip`>
I wish they gave examples in the XML syntax rather than the VRML syntax

[14:09:00.0000] <hsivonen>
so far, it seems to me that version='' only causes purity issues for validators

[14:10:00.0000] <Philip`>
/me wonders if he should bother trying to implement a subset of X3D timing/events/scripting, or just make up something easier to do animation

[14:20:00.0000] <gsnedders>
How bizarre. 'Got "connection" header "", expected "close"' — Op9.2

[14:25:00.0000] <Philip`>
http://www.web3d.org/specifications/x3d-3.0.xsd - hmm, text/plain

[14:25:01.0000] <Philip`>
Actually, 'text/plain; charset=iso-8859-1' and then <?xml version="1.0" encoding="UTF-8"?>

[15:19:00.0000] <Philip`>
Hmm, an X3D spec example uses a TimeSensor field named "fraction", which actually doesn't exist in the definition of TimeSensor (whereas there is a "fraction_changed" field), but it works anyway in at least one particular X3D viewer

[15:20:00.0000] <Philip`>
/me won't support that bug for now


2007-12-16
[16:01:00.0000] <Philip`>
Is html5.org down?

[16:09:00.0000] <hasather>
Philip`: this might be it: http://www.dreamhoststatus.com/2007/12/15/annoying-network-issues/

[16:09:01.0000] <jgraham_>
Philip`: It seems to be down for me as weel

[16:09:02.0000] <jgraham_>
s/weel/well/

[16:14:00.0000] <Philip`>
"An OrientationInterpolator interpolates between two orientations by computing the shortest path on the unit sphere between the two orientations. The interpolation is linear in arc length along this path." - hmm, it'd be much more obvious if they simply said to use quaternion slerp, since I wouldn't have to work out whether that's what they actually mean

[16:14:01.0000] <jacobolus>
Hixie: sorry to keep bugging you.  but any thoughts about that draft?

[16:15:00.0000] <jacobolus>
(http://pastie.textmate.org/private/xsvmjyopvftavgbcbjpexa)

[16:39:00.0000] <Philip`>
Hooray, a shiny monkey head on a declaratively animated rotating cube

[22:38:00.0000] <Hixie>
http://groups.google.com/group/show-in-a-box/browse_thread/thread/2e53cb4c3df04578/b196aaf9676688fe?hl=en&q=html5#b196aaf9676688fe needs some subscribe-and-comment love

[22:38:01.0000] <Hixie>
if anyone's up for it

[22:40:00.0000] <Hixie>
and if anyone wants to explain to this guy that DRM really isn't part of the problem, be my guest. He event cites me but apparently believes my summary isn't an accurate representation of the situation. http://www.sourfizz.com/2007/12/15/html5-working-group-rejects-open-media-formats/

[23:41:00.0000] <doublec>
I've subscribed to the google groups one and will reply

[00:22:00.0000] <hsivonen>
whoa. MathML 3.0

[00:23:00.0000] <hsivonen>
aaargh. they still use entities an *redefine* a bunch of them

[00:25:00.0000] <hsivonen>
hmm. the MathML 3.0 draft uses some mighty weird bolding in code. I wonder if the weirdness comes from TeX4ht or similar

[00:33:00.0000] <Hixie>
doublec: cool, thanks

[02:37:00.0000] <Philip`>
Is someone able to delete wiki pages, like http://wiki.whatwg.org/wiki/Help:Contents and http://wiki.whatwg.org/wiki/Brooklyn_indie_market ?

[02:39:00.0000] <anne-mac>
in theory

[02:41:00.0000] <Philip`>
/me assumes it's limited to sysops

[02:42:00.0000] <Philip`>
Aha, thanks

[02:59:00.0000] <anne-mac>
you should ask lachy or so to get you upgraded to sysops

[04:15:00.0000] <Philip`>
index.php?pid=2&&opcje=a:1:{i:0;s:6:"wyszuk";}

[04:16:00.0000] <Philip`>
What kind of syntax is that? It looks somewhat reminiscent of JSON, but it isn't

[04:59:00.0000] <inimino>
Philip`: that's from PHP

[04:59:01.0000] <tndH>
yeah, it's output from serialize()

[04:59:02.0000] <inimino>
right

[05:01:00.0000] <Philip`>
Ah, thanks

[05:56:00.0000] <anne-mac>
we need to update http://www.whatwg.org/specs/ ...

[05:58:00.0000] <gsnedders>
what a bizarre bug in Opera.

[05:59:00.0000] <gsnedders>
/me looks at XMLHttpRequest

[05:59:01.0000] <gsnedders>
/me notices other stuff

[06:00:00.0000] <gsnedders>
anne-mac: AFAIK nothing actually checks getreponseheader() matches field-name

[06:00:01.0000] <Dashiva>
gsnedders: Bizarre bugs are the best kind

[06:00:02.0000] <Philip`>
Does anybody use XPointer?

[06:00:03.0000] <Dashiva>
Like when setting a className on a table cell causes it to quadruple in width, even though the class doesn't exist

[06:01:00.0000] <hdh>
me, with this fx extension http://www.codedread.com/fxpointer/

[06:02:00.0000] <gsnedders>
Dashiva: ah. the bug I found is considerably less bizarre.

[06:02:01.0000] <gsnedders>
Dashiva: XMLHttpRequest.getResponseHeader("connection") always returns "" in Opera

[06:03:00.0000] <Dashiva>
There was also that one where innerHTML of responseXML had no closing tags

[06:04:00.0000] <Dashiva>
hm

[06:04:01.0000] <Philip`>
hdh: Ah, that looks potentially useful

[06:05:00.0000] <Philip`>
/me is planning to use XPointer for an extension of X3D so you can include bits of uncooperative external X3D files

[06:05:01.0000] <Dashiva>
gsnedders: Must be some special condition you've found, it works in the general case. Got a testcase?

[06:05:02.0000] <Philip`>
(since the normal X3D inclusion mechanism seems to require cooperation and explicit exports, which is a pain)

[06:06:00.0000] <gsnedders>
Dashiva: <http://hg.gsnedders.com/cgi-bin/hgwebdir.cgi/http-parsing/file/29a9ec814634/tests/response/> — I need to make an actual test-case though

[06:08:00.0000] <hdh>
http://blog.codedread.com/archives/2007/10/30/fxpointer-link-exactly/ has a link to an userjs version for opera

[06:10:00.0000] <Philip`>
Hmm, looks like that only supports the element(confusing and fragile child node numbering) system

[06:10:01.0000] <Philip`>
whereas I want XPath because that's more fun

[06:11:00.0000] <Philip`>
(and XPath is supported by all the browsers I care about)

[06:11:01.0000] <gsnedders>
Dashiva: it may be only when the only header is a Connection: Close

[06:16:00.0000] <hdh>
the userjs demo has an xpath, #xmlns(e=http://www.w3.org/1999/xhtml)xpointer(//e:li[4])

[06:18:00.0000] <Philip`>
hdh: Oh, oops, I was looking at the wrong .js file on that page

[06:20:00.0000] <Philip`>
/me will try copying some of that code

[06:28:00.0000] <gsnedders>
http://http-parsing.gsnedders.com/#date-formats makes sense now I see http://support.microsoft.com/kb/234067

[06:45:00.0000] <gsnedders>
hmmm.

[06:51:00.0000] <gsnedders>
even odder things happen when you have the developer console open in Opera

[06:52:00.0000] <gsnedders>
and if there are other headers too

[06:59:00.0000] <Philip`>
Odd things happening in a web browser? That's unfortunately unsurprising :-(

[07:09:00.0000] <gsnedders>
/me has reported the bug to Opera Software, but has been unable to write a minimalistic test case, so is just linking to the revision of the test suite

[07:11:00.0000] <gsnedders>
/me writes in an email "specs be damned"

[07:14:00.0000] <Philip`>
/me wonders if that email is about braces not being allowed in URIs

[07:14:01.0000] <gsnedders>
yeah

[07:16:00.0000] <Philip`>
/me wonders if he should look for braces in other URI attributes

[07:17:00.0000] <Philip`>
/me wonders if anyone has a list of what attributes take URIs

[07:17:01.0000] <gsnedders>
another way to make sure a spec never gets implemented by UAs: require HTML is to be parsed by SGML parsers.

[07:17:02.0000] <gsnedders>
Philip`: I tend to just search through the HTML 4.01 DTD for %uri

[07:17:03.0000] <gsnedders>
(which misses out non-conforming tags/attributes, but hey)

[07:20:00.0000] <Philip`>
http://www.cl.cam.ac.uk/~pjt47/misc/uris-with-braces.txt

[07:21:00.0000] <Philip`>
Oops

[07:21:01.0000] <Philip`>
/me fixes minor bug

[07:23:00.0000] <Philip`>
"If RFC 3986 defined what to do with non-conformant URIs" ... and if browsers implemented those requirements perfectly

[07:26:00.0000] <gsnedders>
:P

[07:27:00.0000] <Dashiva>
I don't think it would work even then

[07:28:00.0000] <Dashiva>
The issue is on the markup level, so either the UAs let it work (current situation), or they refuse to accept those URLs (not going to happen)

[07:30:00.0000] <gsnedders>
Dashiva: well, it depends what the defined handling was anyway

[07:31:00.0000] <hsivonen>
putting GUIDs in URIs strikes me as awful URI design

[07:32:00.0000] <Philip`>
I think clean URI design comes quite low on the prioritised list of problems that many authors have

[07:33:00.0000] <gsnedders>
"No, they rely on user agents not checking for syntactically correct URIs. They aren't using URIs." — reply to me

[07:35:00.0000] <Philip`>
Are they using syntactically incorrect URIs, or are such things defined to not exist?

[07:35:01.0000] <gsnedders>
I guess they don't exist.

[07:35:02.0000] <gsnedders>
(technically, if you take RFC 3986 literally, a string is either a conforming URI or it isn't — nothing else)

[07:35:03.0000] <gsnedders>
*it isn't a URI whatsoever

[07:36:00.0000] <Philip`>
It does sound sensible for other specs (like HTML) to define how non-conforming URIs are handled, since different cases will prefer it to be handled in different ways

[07:36:01.0000] <gsnedders>
WebKit for one handles it all in one place, regardless of where it has come from

[07:37:00.0000] <gsnedders>
Under the same argument HTML error handling should be defined by other specs (like HTTP)

[07:37:01.0000] <Philip`>
e.g. the URI RFC can't say that clicking on an HTML link with an invalid URI should log an error message and perform no action, because it's not its job to define anything about HTML

[07:38:00.0000] <Dashiva>
It seems like two different perspectives. One says "If it doesn't follow the rules, it's not an URI, who cares". The other says "This was used as an URI, but doesn't follow the rules, what do we do?"

[07:38:01.0000] <gsnedders>
Philip`: It can say to strip leading/trailing whitespace, and escape any other invalid characters as pct-encoded

[07:38:02.0000] <gsnedders>
Philip`: which is what is used everywhere, AFAIK

[07:39:00.0000] <Philip`>
"everywhere" does not include HttpClient, it seems

[07:39:01.0000] <Philip`>
Invalid uri 'http://www.springer.com/east/home?SGWID=5-102-70-1033334-detailsPage=journal|description|description': Invalid query

[07:39:02.0000] <Philip`>
(is what it complains about)

[07:40:00.0000] <gsnedders>
s/everywhere/WebKit\/Gecko\/Trident\/Presto/ then

[07:40:01.0000] <Philip`>
Elsewhere: "GET failed: HTTP/1.1 508 unused" - huh?

[07:41:00.0000] <gsnedders>
/me headdesks

[07:43:00.0000] <Philip`>
java.lang.IllegalArgumentException: Invalid uri 'http://www.metmuseum.org/special/se_event.asp?OccurrenceId={A238C1BA-B848-11D3-936D-00902786BF44}': Invalid query

[07:43:01.0000] <Philip`>
Ah, HttpClient does complain about braces too

[07:45:00.0000] <Philip`>
Quite a few invalid redirects too

[07:45:01.0000] <Philip`>
Error on http://gutenberg.spiegel.de/autoren/dauthend.htm : Invalid redirect location: http://gutenberg.spiegel.de/?id=19&autor=Dauthendey,  Max&a

[07:45:02.0000] <Philip`>
utor_vorname= Max&autor_nachname=Dauthendey

[07:45:03.0000] <Philip`>
(...minus the linebreak)

[07:51:00.0000] <gsnedders>
more silly arguments on digg.

[07:51:01.0000] <gsnedders>
equal crime rates in a place where guns are illegal and one where they aren't, therefore you're safer if you have a gun.

[07:52:00.0000] <hdh>
just ban bullets and let people use their metal butts

[07:53:00.0000] <gsnedders>
That's the Swiss solution, IIRC

[07:53:01.0000] <Philip`>
If guns are banned, more bank robbers will have to use bananas instead, which are a good source of potassium, so everyone wins

[07:53:02.0000] <gsnedders>
(everybody who has done national service there has a gun, but the ammunition is sealed — the prison sentences for breaking that seal are c. 20 years IIRC)

[08:26:00.0000] <hdh>
POST is non-idempotent, so why don't input[type=button][method=POST] have a pushed state like a:visited?

[08:29:00.0000] <Philip`>
hdh: Shouldn't it be the other way around? Since POST is non-idempotent, it's different every time you click the button, whereas :visited is meant for things that are the same as what you've seen before

[08:32:00.0000] <hdh>
oh yes

[09:27:00.0000] <hdh>
what's the progress' "denominator punctuation character"?

[09:28:00.0000] <gsnedders>
hdh: follow the reference to "steps for finding one or two numbers of a ratio in a string"

[09:28:01.0000] <G0k>
hey all

[09:28:02.0000] <G0k>
args

[09:28:03.0000] <hdh>
gsnedders: thanks

[09:28:04.0000] <gsnedders>
G0k: see you like to stay around

[09:29:00.0000] <G0k>
fucking colloquy

[09:29:01.0000] <G0k>
it lets me join a channel

[09:29:02.0000] <G0k>
then forgets to echo anything in it

[09:29:03.0000] <G0k>
so i have to rejoin

[09:29:04.0000] <gsnedders>
no, it does output stuff into it

[09:29:05.0000] <gsnedders>
the bug is in WebKit :P

[09:29:06.0000] <G0k>
how does it work sometimes but not other times?

[09:29:07.0000] <G0k>
and is there some other workaround?

[09:29:08.0000] <gsnedders>
G0k: Clear the window

[09:30:00.0000] <gsnedders>
G0k: shortcut is cmd + k, no idea what menu

[09:30:01.0000] <G0k>
cmd+k?

[09:30:02.0000] <G0k>
ah

[09:30:03.0000] <G0k>
yeah

[09:30:04.0000] <G0k>
so i have an idea wrt server sent events

[09:31:00.0000] <gsnedders>
http://bugs.webkit.org/show_bug.cgi?id=14171

[09:31:01.0000] <G0k>
ah

[09:31:02.0000] <G0k>
well. poop.

[09:31:03.0000] <G0k>
anyway

[09:32:00.0000] <G0k>
so 90% of the proposed uses of server sent events has been for chat-style stuff

[09:32:01.0000] <G0k>
so rather than invent that new dom event stream protocol

[09:32:02.0000] <G0k>
how about we just use XMPP?

[09:33:00.0000] <gsnedders>
G0k: and add DOM hooks to it?

[09:34:00.0000] <G0k>
well i mean you could send dom events as the payload

[09:34:01.0000] <G0k>
or just use (for example) the messaging and presence functionality already in it

[09:38:00.0000] <Philip`>
Hmm, using X3D inline in XHTML doesn't actually seem like a stupid idea any more, when I can import stuff like meshes from external files

[09:42:00.0000] <hdh>
and some nice CPU; I can't try x3d out, nobody builds for 64bit gnu/linux

[09:43:00.0000] <Philip`>
hdh: Do you mean normal X3D plugins, or Mozilla's canvas-3d extension?

[09:44:00.0000] <hdh>
I mean the two builds on http://philip.html5.org/demos/canvas/3d/x3d/test.html

[09:45:00.0000] <Philip`>
Ah, okay

[09:45:01.0000] <Philip`>
(I've just been running on Windows, since I don't have useful graphics hardware on my Linux computer)

[09:46:00.0000] <G0k>
we can combine 3D canvases with server sent evetns

[09:46:01.0000] <hdh>
quake clone?

[09:46:02.0000] <Philip`>
I would assume the Mozilla extension could be built, but Mozilla is not the quickest thing to compile

[09:47:00.0000] <Philip`>
I've already done an online multiplayer FPS using Canvex :-)

[09:47:01.0000] <zcorpan_lap>
Philip`: pointer?

[09:47:02.0000] <hdh>
I got the video-enabled fx3 here; I'll see about the plugin

[09:47:03.0000] <G0k>
yeah 3d games ftw

[09:48:00.0000] <Philip`>
(using pipelined XHR (in Opera) and dynamically-inserted <script>s (in Firefox) to get improved latency)

[09:48:01.0000] <Philip`>
zcorpan_lap: It's at file:/// somewhere, unfortunately

[09:56:00.0000] <zcorpan_lap>
Philip`: ok

[10:02:00.0000] <Philip`>
/me updates his code

[10:03:00.0000] <Philip`>
Now you can export some static meshes from e.g. Blender, and then write something like http://philip.html5.org/demos/canvas/3d/x3d/scene.x3d to combine them and animate them, which seems useful

[10:05:00.0000] <Philip`>
I think that mostly addresses my concerns over how anyone could make complex scenes, which have too much mesh data to be editable in a text editor and too much complex XML linking to edit in a 3D editor

[10:06:00.0000] <Philip`>
though now I need to find some way to import an external mesh and then inject some extra nodes into it, like replacing its <Material>

[10:07:00.0000] <Philip`>
/me guesses XSLT might be too complex

[10:51:00.0000] <Philip`>
Oh, maybe XSLT isn't that bad after all

[11:57:00.0000] <Dashiva>
So the spec says that whitespace "after body" should be treated as "in body", and that whitespace in "trailing end phase" should be treated as in "main phase", which becomes "in body".

[11:57:01.0000] <Dashiva>
"in body" says "Append the character to the current node"

[11:58:00.0000] <Dashiva>
So if you have whitespace between </body> and </html>, and after </html>, is that one or two whitespace text nodes at the end of body?

[12:28:00.0000] <webben>
Is there no way to hide a machine-friendly version of a human-friendly duration in HTML5 as currently drafted?

[12:29:00.0000] <webben>
e.g. in XHTML 2 as currently drafted one could do: <span content="PT2M43S">almost three minutes</span>

[12:30:00.0000] <Dashiva>
Isn't that what <time> does?

[12:30:01.0000] <Philip`>
<time> seems to only do absolute times

[12:30:02.0000] <webben>
A duration is not a datetime.

[12:31:00.0000] <Philip`>
Is http://www.whatwg.org/specs/web-apps/current-work/multipage/section-common1.html#time-offsets relevant?

[12:32:00.0000] <webben>
Philip`: If there were a DURATION with a VALUE attribute, it might be. Otherwise, it's just a way of parsing such values.

[12:35:00.0000] <webben>
The context for this is I'm wondering how hAudio durations could be correctly represented in HTML5. This problem of hiding machine-friendly versions of human-friendly data is just going to recur again and again.

[12:36:00.0000] <Philip`>
Does hAudio need machine-friendly vs human-friendly versions? I would assume something like "5:02" or "1:58:21" would be easy for both machines and humans

[12:38:00.0000] <Philip`>
/me doesn't know if there are i18n concerns with writing times in that format

[12:38:01.0000] <webben>
Philip`: Funnily enough, that's the argument I've just put to the list (so I'll wait and see counterexamples or complaints from parser implementers). There is certainly one case where authors might want to not use a format like that: when you're describing a piece of audio in a paragraph of extended descriptions.

[12:38:02.0000] <webben>
*description.

[12:39:00.0000] <webben>
e.g. "and the tedious concerto lasts <duration value="1:04:24">a whole hour</duration>!

[12:40:00.0000] <webben>
My argument is that the i18n problems in this particular case probably don't justify hacking around in HTML 4.01, but it would be nice to have a clean way to represent such things in HTML5.

[12:41:00.0000] <webben>
e.g. if you look at http://www.top100.cn/Product/Product.aspx?productid=S0093820000 Chinese users are apparently happy with arabic numerals in minutes and seconds for durations.

[12:44:00.0000] <Philip`>
Looks like there's not as much variation as with dates/times, which is nice

[12:44:01.0000] <Philip`>
/me sees that page has "2007-3-5"

[15:10:00.0000] <Lachy>
is anyone else having trouble connecting to irc.w3.org?

[15:10:01.0000] <gsnedders>
I'm on it fine

[15:10:02.0000] <Lachy>
weird. I'm getting unknown host errors

[15:11:00.0000] <gsnedders>
just disconnected and reconnected without issue

[15:13:00.0000] <Lachy>
hmm. now it's working

[15:14:00.0000] <csarven>
6665 is okay for me


2007-12-17
[17:01:00.0000] <othermaciej>
I wonder how Andy Clarke came to his conclusions about how to improve CSS

[17:14:00.0000] <spuffle>
got a link? :)

[17:17:00.0000] <othermaciej>
http://www.stuffandnonsense.co.uk/malarkey/more/css_unworking_group/

[17:17:01.0000] <othermaciej>
http://www.stuffandnonsense.co.uk/malarkey/more/csswg_proposals/

[17:17:02.0000] <othermaciej>
short version, he would like to get rid of implementors and have the W3C pay web designers to be on the working group full time

[17:17:03.0000] <othermaciej>
or something

[17:17:04.0000] <othermaciej>
because Opera sued Microsoft

[17:19:00.0000] <spuffle>
ahha

[17:19:01.0000] <spuffle>
always gotta make a bold statement

[17:25:00.0000] <spuffle>
defintely +1 for leadership

[17:26:00.0000] <spuffle>
It's just an opinion though, everyone is entitled to theirs

[17:40:00.0000] <othermaciej>
leadership is a good thing, given the right leader

[17:40:01.0000] <othermaciej>
mainly what puzzles me is why he thinks less implementor involvement would be good

[17:40:02.0000] <othermaciej>
I think a key problem is not enough implementor involvement

[17:40:03.0000] <othermaciej>
(and the fact that few other parties have the needed skills and the time to spend)

[17:43:00.0000] <spuffle>
he just wants to see it run like a business

[17:43:01.0000] <spuffle>
with a more corporate-style hierachy

[17:44:00.0000] <roc>
I think he's in "X isn't working, panic, let's try Y" mode

[17:48:00.0000] <othermaciej>
running it like a business certainly wouldn't involve removing the participants that are actual businesses

[17:48:01.0000] <othermaciej>
"let's treat it like a software project - get rid of the software vendors and let the users drive it from their wishlist"

[17:48:02.0000] <othermaciej>
that ain't how software gets shipped

[17:50:00.0000] <othermaciej>
damn, I posted some wordy blog comments

[17:53:00.0000] <jruderman>
i just read http://www.stuffandnonsense.co.uk/malarkey/more/css_unworking_group/ and i don't understand why he thinks Opera's antitrust complaint has anything to do with the development of CSS3

[17:54:00.0000] <othermaciej>
me neither

[17:56:00.0000] <jruderman>
"I was surprised to learn that one of the reasons why CSS2.1 is only now nearing its candidate recommendation status is because its features are now widely supported by browsers."

[17:57:00.0000] <jruderman>
in reality, is it just because there are multiple interoperable implementations?

[17:57:01.0000] <jruderman>
in which case microsoft dragging its feet doesn't really slow things down

[17:58:00.0000] <othermaciej>
multiple interoperable implementations, and the fact that implementations showed problems

[17:58:01.0000] <othermaciej>
I think CSS2.1 may have actually already hit CR once and went back to WD

[17:59:00.0000] <othermaciej>
/me hopes it wasn't rude to answer the question about how to learn more about CSS internals with a suggestion to get involved w/ an open source browser engine

[17:59:01.0000] <othermaciej>
obviously, I have some selfish interests there

[18:01:00.0000] <hdh>
well, that's still a 1/3 chance :)

[18:02:00.0000] <othermaciej>
I think there's only 2 significant open source browser engines

[18:02:01.0000] <othermaciej>
but maybe I'm counting wrong

[18:03:00.0000] <hdh>
khtml isn't dropped yet

[18:05:00.0000] <hdh>
btw, who's the other maciej?

[18:09:00.0000] <othermaciej>
someone else

[18:34:00.0000] <roc>
othermaciej: I believe Dillo, links, and Amaya are all under active development

[18:35:00.0000] <roc>
oh, significant :-)

[18:36:00.0000] <othermaciej>
I don't think Dillo is

[18:36:01.0000] <othermaciej>
and trying to do QA on Amaya would just be depressing

[23:00:00.0000] <hsivonen>
Hixie: is there a design reason why video src='' overrides <source>s instead of being the last resort if no <source> matches?

[23:01:00.0000] <Hixie>
hsivonen: yeah, i thought it would be confusing for the test order to be different than the source order

[23:01:01.0000] <Hixie>
and i further thought it would look stupid for the two-codec case to have the two codecs at different indent levels and syntaxes

[23:02:00.0000] <hsivonen>
Hixie: it seems to me that having src='' as the last resort (if changed before any impl ships) would make the selection more resilient against partial implementations and would give a server-side sniffing point as the last resort

[23:03:00.0000] <hsivonen>
In particular, Anne's comments about Opera's implementation made me concerned that with the current spec a vendor shipping without <source> support would effectively poison the selection mechanism

[23:04:00.0000] <Hixie>
i agree with both points, but i don't think they justify the author confusion

[23:04:01.0000] <othermaciej>
I think it would be simplest for src to be the first resort

[23:04:02.0000] <hsivonen>
but if src='' were the last choice, that wouldn't be a problem

[23:04:03.0000] <Hixie>
now if a vendor shipped with just src="" and did poison the case, that would be a reason to change, but hopefully that won't happen

[23:05:00.0000] <Hixie>
<video src="3"><source src="1"><source src="2"></video> is extremely confusion imho

[23:05:01.0000] <hsivonen>
ok.

[23:05:02.0000] <Hixie>
er, my grammar sucks

[23:05:03.0000] <Hixie>
but you get the point

[23:05:04.0000] <othermaciej>
what about making the first src be the first to try? (I guess that would require video to have a type attribute if it doesn't already)

[23:06:00.0000] <Hixie>
that's more compelling, but i don't see why authors would want that. <video> doesn't have the attributes <source> does, and it looks weird:

[23:06:01.0000] <hsivonen>
part of the problem is that it's too easy to ship with semi-broken type support

[23:06:02.0000] <othermaciej>
It's also unfortunate that video has no good way to handle the "<video> is supported, but none of the codecs I want are available" case

[23:06:03.0000] <Hixie>
   <video src="video.mpg">

[23:06:04.0000] <Hixie>
     <source src="video.ogg">

[23:06:05.0000] <Hixie>
   </video>

[23:07:00.0000] <Hixie>
othermaciej: i don't believe, by and large, that authors will care about that case

[23:07:01.0000] <othermaciej>
like say I wanted to use <video> with an Ogg source only, and otherwise fall back to a Java applet

[23:07:02.0000] <Hixie>
<object> use in the wild suggests that's far rarer than we would hope

[23:07:03.0000] <hsivonen>
I'm very pessimistic about the abilit of implementations to propertly introspec for dynamically loaded codecs

[23:07:04.0000] <othermaciej>
or use <video> with an H.264 source only, and otherwise fall back to a Flash or Quicktime plugin

[23:08:00.0000] <Hixie>
we do have events that fire if authors do want to handle that case

[23:08:01.0000] <Hixie>
and if it turns out to be common (which i seriously doubt) we can always support that later via some hack

[23:08:02.0000] <othermaciej>
I guess that's better than nothing

[23:08:03.0000] <hsivonen>
speaking applets, has anyone started an effort to implement <video> in IE using JS and Cortado?

[23:08:04.0000] <Hixie>
but i'm skeptical

[23:08:05.0000] <othermaciej>
what's Cortado?

[23:08:06.0000] <hsivonen>
othermaciej: pure Java Ogg/Theora/Vorbis player applet

[23:10:00.0000] <othermaciej>
thinking about video formats makes my brain hurt

[23:11:00.0000] <othermaciej>
but I do have the concern that if multiple browsers support <video> with different sets of codecs, per the current spec that makes life harder for authors than if only one browser supports <video>

[23:11:01.0000] <othermaciej>
(well, I guess it depends on how much they care about single encoding of the video vs. consistent API)

[23:11:02.0000] <othermaciej>
(but QuickTime offers something pretty close to the <video> API in the latest version, by design)

[23:12:00.0000] <othermaciej>
(and I suspect the same is doable with Flash)

[23:13:00.0000] <othermaciej>
anyway we'll see what happens

[23:13:01.0000] <othermaciej>
it's not hard to think of hacks to explicitly allow for such use

[23:15:00.0000] <Hixie>
if we can't get a common codec, changing the fallback rules will be the least of our problems

[23:16:00.0000] <hsivonen>
today embedding YouTube Flash involves pasting in magic boilerplate

[23:17:00.0000] <hsivonen>
it wouldn't be impossible to develop pasteable magic boilerplate that used native Ogg support in Opera and Gecko, XiphQT in Safari and Cortado in IE

[23:17:01.0000] <hsivonen>
far from ideal, though

[23:21:00.0000] <othermaciej>
not sure most web authors will do that much work to get non-browser-native support for a worse codec than Flash's non-browser-native support provides

[23:21:01.0000] <othermaciej>
yet another reason the codec situation sucks

[23:22:00.0000] <othermaciej>
(obviously some ideologically committed content providers like wikipedia would do so)

[23:22:01.0000] <hsivonen>
othermaciej: that depends on whether the authors are doing something that shows up on MPEG-LAs radar, I guess

[23:24:00.0000] <othermaciej>
fortunately still bandwidth and image formats have evolved to the point that patent-encumbered improvements are not worth it

[23:24:01.0000] <othermaciej>
audio could be close to that point

[23:24:02.0000] <othermaciej>
maybe video too someday

[23:24:03.0000] <othermaciej>
(withness the lack of excitement over JPEG2000)

[23:25:00.0000] <othermaciej>
*witness

[23:25:01.0000] <Hixie>
microsoft are really starting to royally piss me off with this EOT bullshit

[23:25:02.0000] <othermaciej>
are they still riding that horse?

[23:25:03.0000] <Hixie>
yet another reason to start the whatcsswg

[23:26:00.0000] <hsivonen>
othermaciej: not only is JPEG2000 legally undesirable, it isn't even technically competitive with the old JPEG even in the huffman form of the old JPEG when the level of compression you want is no visible artifacts

[23:26:01.0000] <othermaciej>
oh, I gotta read the latest bit of that thread

[23:26:02.0000] <hsivonen>
JPEG2000 is a win technically only if you want to compress so much that there are visible artifacts

[23:26:03.0000] <othermaciej>
that's sad

[23:26:04.0000] <Hixie>
othermaciej: i moved the thread to the public list

[23:27:00.0000] <hsivonen>
in that case, the artifacts of the old JPEG get more ugly faster than the JPEG2000 artifacts

[23:27:01.0000] <Hixie>
which i'm sure will make them so happy

[23:30:00.0000] <othermaciej>
their latest mail on the subject in the secret list is just embarassing

[23:31:00.0000] <othermaciej>
I can't believe they got pwned that hard and are still trying to argue their point

[23:35:00.0000] <hsivonen>
Did Björn send his message to any public archive later?

[23:36:00.0000] <MikeSmith>
hsivonen - which message from Björn do you mean?

[23:37:00.0000] <hsivonen>
MikeSmith: http://lists.w3.org/Archives/Member/w3c-css-wg/2007OctDec/0329.html

[23:37:01.0000] <othermaciej>
someone else could send the same information to a public list

[23:38:00.0000] <othermaciej>
although I'm not sure if the Microsoft request it was responsive to has been made public

[23:40:00.0000] <Hixie>
my new macbook pro has this weird thing where every now and then the network stalls

[23:40:01.0000] <MikeSmith>
so were the MS reps to the CSS WG aware of this already at the time of the face to face in Boston?

[23:40:02.0000] <Hixie>
just opening the wireless menu makes it work again

[23:40:03.0000] <Hixie>
i wonder what's up with that

[23:41:00.0000] <MikeSmith>
Hixie - is that when you're using a wireless connection or when you're using Ethernet?

[23:41:01.0000] <Hixie>
wireless

[23:42:00.0000] <MikeSmith>
I've not had a problem like that when using Airport, but I have when using my wireless HSDPA modem

[23:42:01.0000] <MikeSmith>
specifically when using Skype

[23:42:02.0000] <Hixie>
my mac mini hasn't ever had that as far as i know

[23:42:03.0000] <Hixie>
nor my powerbook

[23:42:04.0000] <Hixie>
but my mac book pro, and the macbook my girlfriend has, both have it

[23:45:00.0000] <hsivonen>
does anyone remember off-hand if Gecko/Opera/WebKit trim SVG enumerated-value attribute values before comparing against an internal token literal?

[23:45:01.0000] <MikeSmith>
Hixie - are you running Tiger on your other machines?

[23:45:02.0000] <hsivonen>
that is, does clip-rule='  nonzero  ' work?

[23:46:00.0000] <Hixie>
the mac mini ran tiger then leopard, as did the macbook. the powerbook and the mac book pro are both tiger.

[23:49:00.0000] <othermaciej>
hsivonen: trim what? whitespace?

[23:50:00.0000] <hsivonen>
othermaciej: yes

[23:51:00.0000] <hsivonen>
this is interesting: http://lists.w3.org/Archives/Public/www-math/2007Nov/0007.html

[23:51:01.0000] <othermaciej>
not easy to tell from code

[23:51:02.0000] <hsivonen>
othermaciej: I guess I just have to write a test case or two

[23:51:03.0000] <hsivonen>
thanks

[02:51:00.0000] <Hemebond>
Where can I find a list of "problems" the WHATWG has with XHTML2?

[02:53:00.0000] <annevk>
If they are not in our FAQ I don't think we've written them up

[02:53:01.0000] <annevk>
Then again, the HTML5 draft does mention the relationship of HTML5 with XHTML2

[02:53:02.0000] <Hemebond>
Hi Anne.

[02:54:00.0000] <Hemebond>
I'll go have another look.

[02:58:00.0000] <Hemebond>
Can't see anything in the FAQ.

[02:59:00.0000] <Lachy>
Hemebond, I don't believe any such list exists, but is there anything specific you would like to know about?

[03:00:00.0000] <Hemebond>
I wanted to read about the problems developers had implementing XHTML2.

[03:00:01.0000] <Hemebond>
I remember when I first questioned the direction of HTML5, the respondent said XHTML2 was a mess to implement.

[03:00:02.0000] <Hemebond>
Too many gaps.

[03:00:03.0000] <Hemebond>
Too much room for interpretation.

[03:01:00.0000] <Lachy>
yes, things are too underspecified to be implemented

[03:01:01.0000] <Lachy>
for instance, there is no error handling defined anywhere

[03:01:02.0000] <Hemebond>
XHTML2 is still under development, no?

[03:02:00.0000] <Lachy>
yes

[03:03:00.0000] <Hemebond>
"All the browser vendors have already said they will support HTML 5 (yes, that includes MS) and all but MS have said they won't support XHTML 2"

[03:03:01.0000] <Hemebond>
Is that true?

[03:03:02.0000] <Lachy>
where is that quoted from?

[03:03:03.0000] <Hemebond>
Slashdot comment.

[03:03:04.0000] <Lachy>
link?

[03:04:00.0000] <Hemebond>
http://developers.slashdot.org/article.pl?sid=07/12/16/1656245&from=rss

[03:04:01.0000] <Hemebond>
http://developers.slashdot.org/comments.pl?sid=390646&cid=21718302 < actual comment

[03:04:02.0000] <Hemebond>
Wait... no it's not

[03:04:03.0000] <Hemebond>
http://developers.slashdot.org/comments.pl?sid=390646&cid=21718278 < that is

[03:04:04.0000] <akaroa>
It's not as simple as that Hemebond. There's a lot of politics involved.

[03:04:05.0000] <om_sleep>
I don't know if there are any formal promises, but all the browser vendors are involved in the HTML working group but not the XHTML2 working group

[03:04:06.0000] <Hemebond>
Ah politics....

[03:05:00.0000] <Hemebond>
How I loath thee.

[03:05:01.0000] <om_sleep>
and all but MS are already implementing some HTML5 features

[03:05:02.0000] <othermaciej>
and most have said vaguely negative things about XHTML2 at times

[03:05:03.0000] <Hemebond>
And yet most haven't finished implementing the specs that are "finished".

[03:06:00.0000] <annevk>
so 1) XHTML2 doesn't address web applications and 2) it goes against the design principles for HTML

[03:06:01.0000] <Hemebond>
Design principals?

[03:06:02.0000] <annevk>
http://www.w3.org/TR/html-design-principles/

[03:06:03.0000] <Hemebond>
Cheers.

[03:06:04.0000] <othermaciej>
depends on what you mean by "finished implementing" and "finished spec"

[03:07:00.0000] <annevk>
they have been mostly "implicit" so far, but now they have a pointer :)

[03:07:01.0000] <othermaciej>
the latest specs that have REC status (the most mature status for a W3C spec) are CSS1, HTML4.01 and DOM2

[03:08:00.0000] <othermaciej>
(wait, I'm wrong, DOM 3 Core is also REC)

[03:08:01.0000] <hsivonen>
and HTML 4.01 shouldn't be a REC by today's requirements

[03:08:02.0000] <othermaciej>
non-IE browsers actually have very good support for DOM1 and DOM2

[03:08:03.0000] <othermaciej>
most browsers are pretty good at CSS1

[03:08:04.0000] <othermaciej>
HTML4.01 leaves a lot of things unspecified, so it's hard to tell if you really implemented it or not

[03:08:05.0000] <hsivonen>
(when some CSS1 definitions are updated per CSS2.1)

[03:09:00.0000] <annevk>
(by today's requirements we wouldn't have RECs, which may indicate we need something else)

[03:09:01.0000] <hsivonen>
indeed

[03:09:02.0000] <othermaciej>
I like the IETF's standards track maturity levels

[03:09:03.0000] <annevk>
maybe that can work

[03:10:00.0000] <Lachy>
I never understood the IETF maturity levels, they seemed to be quite meaningless

[03:11:00.0000] <hsivonen>
there's a lot of stuff that one needs to implement on the Web but that hasn't been elevated on the IETF ladder

[03:12:00.0000] <othermaciej>
the point is that you don't need to reach the top of the IETF ladder to be a successful spec

[03:13:00.0000] <hsivonen>
also, the IETF hasn't addressed Web realities e.g. the text/* encoding mess

[03:13:01.0000] <Lachy>
where does the IETF list the maturity levels of their specs?

[03:13:02.0000] <othermaciej>
ftp://ftp.rfc-editor.org/in-notes/bcp/bcp9.txt

[03:14:00.0000] <Hemebond>
http://www.w3.org/TR/html-design-principles/ < was this made up by the HTML5 crew or has it always been there?

[03:14:01.0000] <othermaciej>
Hemebond: it was created by the HTML Working Group

[03:15:00.0000] <othermaciej>
(edited by me and Anne)

[03:15:01.0000] <Hemebond>
uhuh

[03:16:00.0000] <othermaciej>
by IETF process, "Draft Standard" has the interoperable implementation requirement, but still leaves the possibility of changes from further implementation experience if needed

[03:16:01.0000] <othermaciej>
most things don't reach "Internet Standard"

[03:16:02.0000] <webben>
othermaciej: Not /all/ the browser vendors are involved in HTML WG.

[03:16:03.0000] <webben>
(Just the most famous ones.)

[03:16:04.0000] <othermaciej>
webben: ok, the four highest market share and best-known browser vendors, plus some others

[03:17:00.0000] <hsivonen>
Hemebond: if you want an older reference, there's the Opera/Mozilla Position Paper from The Workshop

[03:17:01.0000] <hsivonen>
/me looks up the URI

[03:17:02.0000] <hsivonen>
Hemebond: http://www.w3.org/2004/04/webapps-cdf-ws/papers/opera.html

[03:17:03.0000] <Hemebond>
cheers

[03:18:00.0000] <othermaciej>
(browsers not involved in the WG add up to less than 0.16% market share)

[03:18:01.0000] <hsivonen>
Hemebond: you probably won't find any official statements about XHTML2 from browser vendors

[03:19:00.0000] <hsivonen>
but then, you probably won't find any official announcements to implement HTML5, either

[03:19:01.0000] <Hemebond>
Not really after official statements or anything. Just trying to understand the reason for avoiding XHTML2, that's all.

[03:23:00.0000] <hsivonen>
Hemebond: some things to consider: 1) XHTML2 involves a discontinuity: there isn't a smooth upgrade path for authors. 2) XHTML2 isn't designed to leverage the existing HTML/XHTML implementations in browsers to the extent XHTML5 is and 3) the payoff of being the first-mover to implement XHTML2 does not appear to be favorable when compared to the cost of the implementation

[03:23:01.0000] <webben>
Hemebond: Basically, the HTML5 draft provides a lot of additional features that developers can go right out and use today (e.g. Canvas, with JS/VML fakery in IE). XHTML 2 doesn't. So there's a lot more pressure from developers on browser vendors to implement HTML5 features.

[03:23:02.0000] <webben>
Hemebond: XHTML2 does have useful features that HTML5 doesn't have, but it also breaks a lot of backwards compatibility and isn't designed to degrade gracefully in the one browser that pages absolutely must work in.

[03:24:00.0000] <Hemebond>
You should paste all this into the FAQ or something...

[03:24:01.0000] <Hemebond>
FAQ-for-hemebond

[03:24:02.0000] <Hemebond>
.html

[03:24:03.0000] <Hemebond>
Thanks for that.

[03:24:04.0000] <webben>
So moving to XHTML 2 (from a developer perspective) would have costs that generally outweigh the gains.

[03:25:00.0000] <webben>
Hemebond: There are cases where XHTML 2 as currently drafted might be a better choice than HTML5. e.g. for server-side storage of documents.

[03:26:00.0000] <akaroa>
webben: how is that?

[03:26:01.0000] <Hemebond>
akaroa: transform to HTML using server-side scripting

[03:26:02.0000] <Hemebond>
or XSL

[03:26:03.0000] <Hemebond>
I've thought of using XHTML as a storage markup. Or even opendoc

[03:27:00.0000] <webben>
akoroa: Well, just for one example, XHTML 2 allows you to store a machine-friendly version of human-friendly data (as commonly needed for microformats) with the content attribute.

[03:27:01.0000] <webben>
<span content="PT3M23S">about three minutes</span>

[03:29:00.0000] <webben>
XHTML 2 also has more extension mechanisms like arbitrary roles and mixing in markup from other (perhaps your own) namespaces.

[03:30:00.0000] <webben>
Although mixing in namespaces can be done in XHTML 5 too.

[03:33:00.0000] <webben>
Hemebond: You might also want to look at DocBook and TEI.

[03:33:01.0000] <Hemebond>
Ah yeah, I meant DocBook above.

[03:33:02.0000] <Hemebond>
TEI is new to me though.

[03:34:00.0000] <webben>
http://www.tei-c.org/index.xml

[03:34:01.0000] <Hemebond>
Just found it. Cheers.

[03:34:02.0000] <akaroa>
I am still not convinced that there is a use case for XHTML2. HTML5 and XHTML5 should take care of everything right?

[03:34:03.0000] <webben>
akaroa: "Should" and "will" aren't the same thing. :)

[03:35:00.0000] <akaroa>
If not, we can still add more features to the spec

[03:36:00.0000] <webben>
akaroa: Well yes, both drafts could change.

[03:36:01.0000] <hsivonen>
having more features doesn't necessarily mean better when it comes to deciding the storage format of a back end system

[03:36:02.0000] <Hemebond>
or front end system

[03:36:03.0000] <othermaciej>
XHTML2 has decided to focus on being an authoring format

[03:37:00.0000] <Camaban>
The IBM article that the slashdot link referred to talked about the different philosophies of HTML5 and XHTML2, XHTML2 striving for a purer seperation, and HTML5 striving to be more readily 'practical', is that not a big issue that should be mentioned when talking about the differences between the 2?

[03:37:01.0000] <Hemebond>
Thanks for the info folks, will have to read up more tomorrow. Night.

[03:38:00.0000] <othermaciej>
I think HTML5 does a reasonably good job with separation of concerns

[03:38:01.0000] <othermaciej>
I'd state the main differences as:

[03:38:02.0000] <othermaciej>
- XHTML2 doesn't really care about compatibility with HTML4, or even XHTML1 very much

[03:38:03.0000] <Hemebond>
I may as well lurk here and record the discussion....

[03:38:04.0000] <Hemebond>
night

[03:38:05.0000] <webben>
Camaban: That's sort of true, although the WHATWG line has tended to be that things like <I> are just fine when it comes to separating content and presentation.

[03:39:00.0000] <othermaciej>
- XHTML2 is more concerned with making things really generic by moving semantics and behavior to attributes

[03:39:01.0000] <webben>
Hemebond: good night :)

[03:39:02.0000] <Camaban>
some argue the seperation point when HTML5 is wanting to not only define HTML, but the audio/video codecs used, and I believe bits to do with Bluetooth among others?

[03:39:03.0000] <othermaciej>
- HTML5 is interested in supporting web applications as well as web documents, XHTML2, not so much

[03:40:00.0000] <Camaban>
not sure where I sit on these issues, it's an area that interests me, and has me stuck on what i think of the XHTML2 Vs HTML5 thing overall

[03:40:01.0000] <akaroa>
It's not just about whether XHTML2 is good or not, it's about whether it can fit on to the web alongside HTML5 or not. I don't believe it can, and believe that XHTML5 should replace XHTML2

[03:41:00.0000] <Camaban>
othermaciej: that's a pretty fundamental difference to HTML as we know it, and I've seen suggestions that if that's the case, should it not be called soemthing else altogether?

[03:42:00.0000] <othermaciej>
Camaban: what's a pretty fundamental differences?

[03:42:01.0000] <Camaban>
/me shrugs

[03:42:02.0000] <othermaciej>
er

[03:42:03.0000] <othermaciej>
"fundamental difference"

[03:42:04.0000] <webben>
Camaban: Not necessarily. Part of the issue here is that electronic document formats tend to be application formats too in all but name.

[03:42:05.0000] <Camaban>
othermaciej: supporting web apss as well as web docs

[03:42:06.0000] <Camaban>
*apps

[03:42:07.0000] <othermaciej>
Camaban: people write web apps in HTML today

[03:43:00.0000] <othermaciej>
HTML5, instead of trying to stop them, tries to make this work better

[03:43:01.0000] <webben>
Camaban: If you look at formats like ODF, the Office formats, PDF they all have support for dynamic forms, scripting, media formats etc

[03:43:02.0000] <Camaban>
but HTML4/XHTML aren't designed to be 'web app' languages, it's jsut people have worked out how to combine them with javascript and stuff to do it

[03:44:00.0000] <webben>
Camaban: That's not obvious.

[03:44:01.0000] <hsivonen>
webben: although in PDF, the dynamism is an awfully bad idea

[03:44:02.0000] <othermaciej>
that's true, they haven't been designed that way historically (although forms and client-side image maps are clearly app-like features)

[03:44:03.0000] <webben>
hsivonen: Why is it a worse idea in PDF than any of the others?

[03:44:04.0000] <othermaciej>
(as is support for scripting and events)

[03:45:00.0000] <Camaban>
If HTML5, from my limited understanding (hence I generally just lurk round here) is designed to deal with web apps specifically (as well as web pages), that's quite a philisophical difference

[03:45:01.0000] <othermaciej>
HTML5 tries to pave over the rough edges

[03:45:02.0000] <webben>
Camaban: HTML has seen shifts in philosophy before.

[03:45:03.0000] <othermaciej>
well, web apps like gmail, flickr, upcoming, google maps, and so forth are not going away

[03:45:04.0000] <webben>
Camaban: e.g. the realization that authors needed more control over suggested presentation, without compromising media independence of content.

[03:46:00.0000] <Camaban>
as I said, I'm unconvinced either way at the moment, I'm trying to resolve some of the for/agaisnt arguements I've seen

[03:46:01.0000] <othermaciej>
it seems the web is going more in that direction

[03:46:02.0000] <webben>
(leading to dropping loads of presentational elements and attributes in favour of CSS)

[03:46:03.0000] <othermaciej>
so your options are either to improve HTML, so that web apps like that fit into the web better

[03:46:04.0000] <othermaciej>
or you can ignore the issue

[03:46:05.0000] <othermaciej>
or you can invent a new, wholly different approach for web apps that is not based on HTML at all

[03:46:06.0000] <hsivonen>
webben: PDF (1.4) is solves the main problem it was designed to solve phenomenally good. the dynamism risks breaking what PDF is phenomenally good at

[03:47:00.0000] <webben>
hsivonen: Well, that's certainly been the case with HTML too. e.g. scripting and frames etc regularly breaking the usability of the web.

[03:47:01.0000] <hsivonen>
webben: fortunately, Adobe now has less need to use PDF as a presentation layer for dynamic apps because they have Flex

[03:48:00.0000] <Camaban>
othermaciej: which is what I've seen some people suggesting, maybe not wholly new, but something not called 'hypertext markup language'

[03:48:01.0000] <webben>
Like XUL.

[03:48:02.0000] <hsivonen>
s/good/well/

[03:48:03.0000] <webben>
or XAML.

[03:49:00.0000] <webben>
(or XForms to some extent.)

[03:49:01.0000] <othermaciej>
or Silverlight

[03:49:02.0000] <othermaciej>
or Java (for non-markup ways of doing it)

[03:49:03.0000] <hsivonen>
I wonder if there's a spec for Inkscape's private markup

[03:49:04.0000] <othermaciej>
or SVG

[03:49:05.0000] <othermaciej>
or Flex

[03:50:00.0000] <hsivonen>
webben: the difference is that people want to use HTML for apps. telling them otherwise is not productive

[03:51:00.0000] <Camaban>
maybe part of the problem is, the people who are pushing the semantic view point, using things for their intended use, rather than just using 'what works', and turning HTML into a web app language doesn't quite fit with some of that

[03:51:01.0000] <webben>
hsivonen: I don't think people want to use HTML for apps per se. I think people want to write apps that work in popular browsers, and HTML is the cheapest way to do that.

[03:51:02.0000] <hsivonen>
webben: but twisting PDF from an excellent final-form digital paper into an enterprise app UI layer was a vendor-initiated move to leverage the installed base of a product to gain foothold in a different market

[03:51:03.0000] <hsivonen>
webben: granted

[03:51:04.0000] <annevk>
is there any realistic alternative to HTML for doing web apps?

[03:51:05.0000] <webben>
Flash and (now) Silverlight.

[03:51:06.0000] <annevk>
if not then I'm not sure if this discussion is productive

[03:52:00.0000] <othermaciej>
Flash is the only currently realistic alternative

[03:52:01.0000] <annevk>
I should have said non-proprietary, of course

[03:52:02.0000] <othermaciej>
Silverlight may get there but does not have the deployment

[03:52:03.0000] <annevk>
can't have a Web that depends on a single vendor

[03:52:04.0000] <othermaciej>
Java is widely available but sucks really hard

[03:52:05.0000] <hsivonen>
webben: but while HTML and Flash make sense as app platforms from the developer POV, what developer would want to choose Adobe Intelligent Document Platform is the UI layer of an app?

[03:52:06.0000] <othermaciej>
I'm measuring "realistic" solely by feasibility of wide deployment for a web developer

[03:52:07.0000] <Camaban>
annevk: is HTML the way it *should* be done though? or is this just somehting that's happening because it seems to work?

[03:52:08.0000] <webben>
hsivonen: I don't know. :)

[03:53:00.0000] <annevk>
Camaban, I don't really see much issues with using HTML for it

[03:53:01.0000] <annevk>
along with CSS improvements, API improvements, etc., of course

[03:53:02.0000] <webben>
annevk: Well, the success of HTML5 web apps is rather dependent on MS actually implementing things. Otherwise you might as well use something like XUL.

[03:53:03.0000] <othermaciej>
Camaban: I dunno, can you think of a reason it shouldn't be done?

[03:53:04.0000] <annevk>
webben, no, XUL would be a horrible option

[03:54:00.0000] <annevk>
everyone, including Moz, is in agreement on that afaict

[03:54:01.0000] <annevk>
stuff from XUL is useful though and is making its way to HTML5 and CSS

[03:54:02.0000] <Camaban>
I'm slightly nervous of the idea of an HTML spec including some of the things it's looks like it's going to include, but I don't know if that's because I genuinely htink it's a bad idea, or jsut because it alters my preconceptions of what HTML is and what I 'know', it negates some of my expertise :)

[03:55:00.0000] <webben>
annevk: And yet there's a community of XUL developers (hence XULRunner etc..)

[03:55:01.0000] <webben>
/me isn't an XUL developer, he should add.

[03:55:02.0000] <othermaciej>
a lot of the stuff in html5 is all about web documents, not web apps at all, particularly

[03:55:03.0000] <annevk>
webben, sure, there's a community of people doing Flash too

[03:56:00.0000] <othermaciej>
some stuff is hybird (<details> could be thought of as an app feature or for an interactive document, <video> certainly makes sense in the context of an interactive multimedia hypertext document)

[03:56:01.0000] <othermaciej>
some stuff is pretty web-app-oriented (AJAX history support, <canvas>)

[03:57:00.0000] <akaroa>
webben: I can't think of anything that can done be in Flash or silverlight that can't be done using Web Standards in the future.

[03:57:01.0000] <othermaciej>
I can understand being nervous about all the new stuff, but change is the one constant in the tech industry, learning to embrace it and adapt is a strength

[03:57:02.0000] <Camaban>
to be honest, I think some people might be pacified if it was called something like 'multimedia markup lanuage', instead :)

[03:57:03.0000] <webben>
SQL

[03:57:04.0000] <othermaciej>
akaroa: "in the future" is a pretty big out

[03:58:00.0000] <webben>
akaroa: Yes. But it's coping with current browsers that leads people to hack up HTML solutions.

[03:58:01.0000] <annevk>
Camaban, it was called Web Applications 1.0 at some point

[03:58:02.0000] <othermaciej>
Camaban: well, for all sorts of compatibility reasons, it has to keep text/html and application/xhtml+xml as the MIME types

[03:58:03.0000] <annevk>
people preferred HTML5

[03:58:04.0000] <othermaciej>
Camaban: and <!DOCTYPE html> as the doctype

[03:58:05.0000] <webben>
And they'll continue to do so until (unless) a killer app galvanizes end users to change their systems to support something else.

[03:58:06.0000] <othermaciej>
and use most of the tags historically defined for HTML

[03:59:00.0000] <akaroa>
othermaciej: I meant with the tools that we have now (CSS, SVG), but improved

[03:59:01.0000] <othermaciej>
so renaming it would probably be more confusing than helpful

[03:59:02.0000] <Camaban>
ah, that stuff I wasn't aware of

[03:59:03.0000] <othermaciej>
akaroa: does "improved" include adding features?

[04:00:00.0000] <othermaciej>
and in some cases radical performance improvements?

[04:01:00.0000] <Camaban>
I can see the reasons and desirability for some of the things that HTML5 is doing, I'm just struggling to be convinced that after what was shoved down our throats about XHTML --> XML being the future and all the well defined sepeartion of things etc.... that taking something of a different approach is suddenly meant to be better

[04:02:00.0000] <akaroa>
othermaciej: yes, adding features I guess. I'm not sure what you mean. I thought you would agree with me  :) 

[04:02:01.0000] <webben>
othermaciej: Well, there are a lot of radical performance improvements in stuff like SVG scripting (e.g. the forthcoming tamarin-based improved JS implementation in Moz.)

[04:03:00.0000] <othermaciej>
akaroa: I'm just saying, it's pretty open-ended to say "you'll be able to do everything in the future"

[04:03:01.0000] <othermaciej>
I do agree that extending web standards is good

[04:03:02.0000] <othermaciej>
and improving their implementations

[04:03:03.0000] <akaroa>
I guess I said it all wrong sorry.

[04:03:04.0000] <othermaciej>
webben: core JS execution is almost never the bottleneck for scripted SVG animations, in my experience

[04:04:00.0000] <webben>
othermaciej: What is?

[04:04:01.0000] <othermaciej>
DOM access, layout, painting

[04:04:02.0000] <annevk>
yeah, it's not so much JavaScript

[04:04:03.0000] <webben>
Why should DOM access be intrinsically slower than (say) manipulating Flash objects via actionscript.

[04:04:04.0000] <othermaciej>
(WebKit is pretty fast on scripted SVG animations, and not because our JS is way better than everyone else's)

[04:04:05.0000] <othermaciej>
(yet)

[04:04:06.0000] <webben>
Is that due to something intrinsic or just down to sluggish implementations?

[04:05:00.0000] <othermaciej>
webben: it doesn't need to be - I'm just saying that Tamarin won't address any of those

[04:05:01.0000] <webben>
oh okay

[04:05:02.0000] <othermaciej>
WebKit has a much faster core DOM than Gecko for instance

[04:05:03.0000] <othermaciej>
as does Opera

[04:06:00.0000] <Philip`>
JS performance hurts my per-frame lighting calculations for 3D canvas stuff in Opera, so I wouldn't mind that being faster :-)

[04:07:00.0000] <othermaciej>
this is a fairly good core DOM benchmark

[04:07:01.0000] <othermaciej>
Philip`: it's certainly worth improving!

[04:07:02.0000] <othermaciej>
JS performance is one of the top things I personally code on actually

[04:07:03.0000] <othermaciej>
but it takes a lot of different things to make an excellent web engine

[04:08:00.0000] <akaroa>
othermaciej: I meant we have a pretty good "web-toolbox" with what is in the (x)html5 spec now + CSS3 SVG etc.

[04:08:01.0000] <Philip`>
(Firefox does that particular code insanely faster, because it's running as highly optimised code on highly parallel specialised hardware, rather than as JavaScript, which is a better way to solve this problem)

[04:09:00.0000] <othermaciej>
yeah

[04:09:01.0000] <othermaciej>
in some cases the best way around JS perf bottlenecks is not to use JS at all

[04:09:02.0000] <othermaciej>
thus Apple's CSS animation proposal

[04:09:03.0000] <othermaciej>
btw this is a pretty good core DOM benchmark if anyone would like to test: http://www.hixie.ch/tests/adhoc/perf/dom/artificial/core/001.html

[04:46:00.0000] <hdh>
 /part

[05:48:00.0000] <Philip`>
/me hopes the URI template topic can stay on topic :-)

[05:50:00.0000] <zcorpan>
/me probably should have let that email stay as a draft :|

[06:13:00.0000] <hsivonen>
RDF and XMLNS show up in the URI template thread...

[08:05:00.0000] <gsnedders>
/me should take a screenshot of all my mailboxes in Mail.app

[08:06:00.0000] <gsnedders>
but that's hard seeming there's a scrollbar

[08:06:01.0000] <hsivonen>
666 messages and 1337 messages?

[08:06:02.0000] <gsnedders>
no

[08:07:00.0000] <gsnedders>
(though, according to Last.fm, I stopped listening to Queen after 1337 plays, so I kinda force myself to never listen to any of their stuff)

[09:58:00.0000] <zcorpan>
Hixie: do you have any idea why the annotation script doesn't work in opera?

[09:59:00.0000] <zcorpan>
Hixie: it says "Error: (0)."

[10:00:00.0000] <zcorpan>
which is

[10:00:01.0000] <zcorpan>
        } else if ((x.status != 200) && (x.status != 304)) {

[10:00:02.0000] <zcorpan>
          failure("Error: " + x.statusText + " (" + x.status + ")");

[10:00:03.0000] <zcorpan>
...in doXMLHttpRequest()

[12:06:00.0000] <krijn>
Damn peer

[12:12:00.0000] <Philip`>
/me sees http://www.bbc.co.uk/iplayer/ is now cross-platform since it uses Flash

[12:13:00.0000] <webben>
where cross-platform means "more than one platform" presumably

[12:14:00.0000] <Philip`>
It means "not Windows Media Player"

[12:15:00.0000] <Philip`>
I can't tell what codec it's using, but it's going over RTMP (unlike e.g. YouTube which is HTTP (I think))

[12:15:01.0000] <webben>
Windows Media is probably usable on more platforms than Flash. You can run mplayer on 64-bit *nix I presume.

[12:16:00.0000] <webben>
but I guess they're probably using a lot of DRM or something

[12:16:01.0000] <Philip`>
Windows Media Player + Windows Media DRM

[12:16:02.0000] <Philip`>
which I guess mplayer doesn't handle so well

[12:16:03.0000] <webben>
probably not

[12:18:00.0000] <Philip`>
Looks like they're still using Windows Media for the download option (with DRM), and Flash only for streaming (which is hard to save to disk, so it has the same effect as DRM)

[12:20:00.0000] <Philip`>
I guess they're unlikely to use Ogg :-(

[12:21:00.0000] <webben>
Could one stream Ogg in a way that would make it "hard" to save?

[12:24:00.0000] <Philip`>
I guess RTMP mostly works on the basis of being complex and undocumented, so most people won't bother writing tools to download it and to play it back

[12:24:01.0000] <webben>
Sounds optimistic.

[12:24:02.0000] <Philip`>
and the 'undocumented' bit wouldn't go down very well in the Ogg community

[12:24:03.0000] <webben>
Given the existence of projects like Gnash.

[12:24:04.0000] <webben>
(which depend on ignoring documentation of complex software for their very existence)

[12:25:00.0000] <Philip`>
http://osflash.org/documentation/rtmp

[12:25:01.0000] <webben>
heh

[12:25:02.0000] <Philip`>
with lots of "Fix Me!" labels

[12:27:00.0000] <Philip`>
It's just a pain to decode these things, and it looks like it's enough of a pain that nobody has added support for "mplayer -dumpstream rtmp://..." yet

[13:14:00.0000] <Hixie>
i hate how i have to hit _three_ keys to page up and down in a terminal window on mac

[13:14:01.0000] <Hixie>
(specifically a mac laptop, though even a desktop still requires two keys)

[13:16:00.0000] <gsnedders>
peh. it depends on the keyboard!

[13:17:00.0000] <gavin_>
my mac laptop requires two keys (fn+up)

[13:17:01.0000] <gavin_>
but I had to tweak terminal.app keymappings

[13:18:00.0000] <Hixie>
yeah i'm tweaking settings too

[13:18:01.0000] <gsnedders>
I think it's just fn+up here too

[13:18:02.0000] <gsnedders>
But there again I've little idea what the defaults are

[13:20:00.0000] <Hixie>
ok i can do it with one hand now

[13:20:01.0000] <Hixie>
shift-up and shift-down

[13:24:00.0000] <Hixie>
i don't suppose there's a way to control what double-click selection behaviour is, either?

[13:24:01.0000] <Hixie>
s/either/too/

[13:24:02.0000] <Hixie>
on putty i have it set to just select uri characters, which makes it easy to select uris

[13:24:03.0000] <gavin_>
/me doesn't know of any

[13:25:00.0000] <gsnedders>
Hixie: AFAIK no

[13:25:01.0000] <gavin_>
pre-leopard Cmd+DblCLick was "load link", not it's Cmd+Shift+DblClick

[13:25:02.0000] <gavin_>
*now

[13:25:03.0000] <gavin_>
kind of a pain

[13:25:04.0000] <hsivonen>
indeed

[13:26:00.0000] <Hixie>
yeah but that selects | characters

[13:26:01.0000] <Hixie>
so it doesn't work well for me either

[13:34:00.0000] <hsivonen>
Philip`: did you decrypt the overlap in interleave error with no interleave in sight?

[13:41:00.0000] <hsivonen>
interleave problem found

[13:45:00.0000] <Hixie>
hm?

[13:51:00.0000] <hsivonen>
Hixie: I tried to interleave rdf:RDF with *

[14:10:00.0000] <roc>
othermaciej: "although I'm not sure if the Microsoft request it was responsive to has been made public" ... those earlier discussions never were made public (as far as I can tell), that's the problem

[14:11:00.0000] <roc>
This is a powerful reason why the private w3c-css-wg list needs to die.

[14:15:00.0000] <Hixie>
hsivonen: ah, did it not work?

[14:16:00.0000] <hsivonen>
Hixie: it didn't since duplicate names in interleave are prohibited and * is a duplicate of everything

[14:16:01.0000] <hsivonen>
(this is a really annoying feature of RELAX NG)

[14:16:02.0000] <Hixie>
ah

[14:17:00.0000] <Hixie>
i don't think we want to allow RDF _anywhere_, btw

[14:17:01.0000] <Hixie>
<select> <rdf...> would probably not work well

[14:17:02.0000] <Hixie>
same with various other cases

[14:17:03.0000] <hsivonen>
Hixie: I thought you said earlies that RDF was OK where metadata elements are OK, i.e. in <head>

[14:17:04.0000] <Hixie>
i'd recommend just allowing rdf at metadata and block/inline (soon to become "prose") entry points

[14:18:00.0000] <Hixie>
i may have misunderstood what you meant

[14:18:01.0000] <hsivonen>
I got rid of the * that was a bug

[14:18:02.0000] <hsivonen>
I didn't mean * quantifier. I meant the * name class

[14:19:00.0000] <hsivonen>
currently, I don't allow RDF on the prose level at all

[14:19:01.0000] <Hixie>
that works for me too

[14:19:02.0000] <hsivonen>
only in XHTML <head> and SVG <metadata>

[14:20:00.0000] <Philip`>
hsivonen: I fixed the "Error: Both operands of interleave contain text" by giving up on the RNG conversion and just using XSD

[14:20:01.0000] <hsivonen>
Philip`: heh

[14:21:00.0000] <Philip`>
(Now I've imported the XSLT schema and so the XSD->RNG converter just gives up and dies before getting that far)

[14:22:00.0000] <Philip`>
(Incidentally, I'm probably doing stuff stupidly since I really have no idea about any of this, but at least it seems to be working at the moment)

[14:29:00.0000] <hsivonen>
Hixie: FYI, http://golem.ph.utexas.edu/~distler/blog/archives/001475.html#c013846

[14:32:00.0000] <Hixie>
yeah, we need to figure something out

[14:32:01.0000] <Hixie>
the parsing issue is the biggest problem, frankly

[15:57:00.0000] <Philip`>
/me sees that at least one of the iPlayer shows has a sign language interpreter overlayed (seemingly with no way to disable it), which is a bit distracting


2007-12-18
[16:09:00.0000] <Hixie>
many shows in the uk get broadcast twice, once with and once without the sign interpreter

[16:09:01.0000] <Hixie>
including, e.g., cartoons

[16:22:00.0000] <Hixie>
hsivonen: how do i distinguish RDF from MathML and SVG in a generic sense?

[16:23:00.0000] <Hixie>
as in, "elements that are from non-XHTML namespaces and whose semantics define metadata" might describe RDF

[16:23:01.0000] <Hixie>
but i can't see how to describe MathML/SVG

[16:28:00.0000] <othermaciej>
what is the purpose of the distinction?

[16:34:00.0000] <Hixie>
i want to allow rdf anywhere metadata is allowed

[16:34:01.0000] <Hixie>
and want to allow svg anywhere prose content is allowed

[16:35:00.0000] <Hixie>
(without calling them out by name, since i don't want to close teh set of namespaces allowed)

[16:37:00.0000] <Hixie>
25 points to anyone who finds a reason why we can't allow <ol>, <div>, <ul>, <p>, etc, in <address>

[16:39:00.0000] <othermaciej>
"elements ... whose semantics define metadata" and "elements ... whose semantics define content" maybe?

[16:39:01.0000] <othermaciej>
that's pretty vague

[16:39:02.0000] <Hixie>
that's what i have but i argued with myself that metadata is a kind of content

[16:40:00.0000] <othermaciej>
what isn't a kind of content?

[16:41:00.0000] <othermaciej>
(you could also ask, what isn't a kind of metadata?)

[16:41:01.0000] <Hixie>
yeah

[16:41:02.0000] <Hixie>
that's my problem

[16:42:00.0000] <webben>
Philip`: Do you have a link for that? The impression I got was that you'd get a choice of non-interpreted vs interpreted.

[16:43:00.0000] <Hixie>
while i'm boiling the oceans over here getting rid of hte block vs inline distinction, maybe i should make <body> optional too... http://www.glazman.org/weblog/dotclear/index.php?post/2005/05/27/1055-adam-2

[16:43:01.0000] <Hixie>
(see the paragraph starting "Steven Pemberton")

[16:44:00.0000] <Philip`>
webben: http://www.bbc.co.uk/iplayer/page/item/b008d36h.shtml is one (linked from the front page, with no link to an alternative version that I can see)

[16:44:01.0000] <webben>
ta

[16:44:02.0000] <othermaciej>
Hixie: isn't <body> already optional?

[16:44:03.0000] <Philip`>
(It's quite probable that the Windows downloadable version has more options than the Flash streaming one)

[16:45:00.0000] <Hixie>
othermaciej: not in the dom, only in the html serialisation

[16:45:01.0000] <othermaciej>
ah

[17:03:00.0000] <Hixie>
as part of the liberation of html elements, i'm allowing <style scoped> virtually anywhere

[17:03:01.0000] <Hixie>
let me know if you think <style scoped> should be a little more oppressed

[17:07:00.0000] <othermaciej>
where was it previously not allowed?

[17:09:00.0000] <Hixie>
only in section, nav, div, and article

[17:09:01.0000] <Hixie>
(i'm still requiring that it be at the top of its parent)

[18:29:00.0000] <deltab>
om_afk: in SGML terms at least, the body element is mandatory but its tags are optional

[18:30:00.0000] <deltab>
likewise html, head, and tbody

[18:34:00.0000] <Hixie>
hsivonen is going to have a field day when this is done

[18:34:01.0000] <Hixie>
there are so many subtle changes

[18:35:00.0000] <Philip`>
Only one field day? That doesn't sound too bad

[18:35:01.0000] <Hixie>
hsivonen is good

[18:35:02.0000] <Hixie>
:-P

[01:34:00.0000] <Hixie>
my edits are making the heading association algorithm even harder to understand

[01:35:00.0000] <Hixie>
so i'll definitely have to rewrite it at some point

[02:58:00.0000] <Hixie>
good lord <ins> is going to make my head explode

[02:59:00.0000] <Hixie>
it just totally screwed up my entire carefully designed content model

[02:59:01.0000] <Hemebond>
Why?

[02:59:02.0000] <Hemebond>
You working on HTML5?

[02:59:03.0000] <Hixie>
yeah

[02:59:04.0000] <Philip`>
Replace it by an attribute, so you can have <p ins> and <b ins> and <blockquote ins> and whatever, which won't affect the content model

[03:00:00.0000] <othermaciej>
the XHTML2 solution!

[03:00:01.0000] <Hixie>
<article> <ins> <em> hello </em> </ins> <ins> world </ins> of pain <ins> <p> how are you </p> </ins> </article>

[03:00:02.0000] <Hixie>
that has to become 2 paragraphs

[03:00:03.0000] <Hixie>
but it is legal

[03:01:00.0000] <Hixie>
and <article> <em> hello <ins> world </ins> </em> </article> is legal too

[03:01:01.0000] <Hemebond>
ins is block?

[03:01:02.0000] <Hixie>
it's... something

[03:01:03.0000] <Philip`>
It's not like people do much processing of <ins> today, so it wouldn't hurt much to change it

[03:01:04.0000] <Hemebond>
Old INS was only inline, right?

[03:02:00.0000] <Hixie>
<ins> was neither inline nor block in html4

[03:02:01.0000] <Hixie>
it was an sgml inclusion

[03:03:00.0000] <Hemebond>
*boggle*

[03:03:01.0000] <Lachy>
Hixie, I thought this issue was solved using transparent content models, and that your first example wouldn't be legal

[03:16:00.0000] <Hixie>
what's wrong with the first example?

[03:17:00.0000] <Hixie>
the transparent content models don't really help determine where <ins> is allowed in the first place, only what's allowed inside it

[03:17:01.0000] <Lachy>
Hixie, because <article> <em></em> <p></p> </article> isn't allowed

[03:18:00.0000] <Hixie>
it is in the new model

[03:19:00.0000] <Lachy>
why?

[03:19:01.0000] <Hixie>
we're removing the requirement that <p> be required around paragraphs

[03:19:02.0000] <Hixie>
apparently it was causing too much confusion

[03:19:03.0000] <Lachy>
who was confused by it?

[03:19:04.0000] <Hixie>
sam ruby, e.g.

[03:20:00.0000] <Lachy>
oh, he's confused by a lot of things, and I generally disagree with him

[03:20:01.0000] <Hixie>
heh

[03:20:02.0000] <Hixie>
he wasn't the only one

[03:21:00.0000] <Hixie>
basically paragraphs are implied much like sections get implied

[03:21:01.0000] <Hixie>
at least, that's the model i'm writing up

[03:21:02.0000] <Hixie>
who knows if we'll keep it

[03:22:00.0000] <Lachy>
but that makes it possible to do things like this: <div><p>foo</p> bar <p>baz</p></div> and it represents 3 paragraphs.

[03:23:00.0000] <Lachy>
I really don't like that model

[03:23:01.0000] <Hemebond>
Ugh

[03:23:02.0000] <Hixie>
me either, but it's hard to find a good argument for why that's not ok

[03:23:03.0000] <Hixie>
(and note that html4 strict allows it)

[03:23:04.0000] <Lachy>
where are the arguments that it is ok?

[03:23:05.0000] <Hixie>
(so a lot of people will hit that when transitioning to html5)

[03:24:00.0000] <Hemebond>
Allows what? Dropping of tags?

[03:24:01.0000] <Hixie>
Lachy: it doesn't cause anyone any harm, and it's unambiguous

[03:24:02.0000] <Hixie>
Hemebond: allows mixing inline and block content in <div> elements

[03:25:00.0000] <Hemebond>
Ah, right.

[03:26:00.0000] <Hemebond>
Hixie: Are you the BDFL for WHATWG?

[03:27:00.0000] <Hixie>
pretty much

[03:27:01.0000] <Hemebond>
The intawebs just told me so.

[03:27:02.0000] <Hixie>
(except maybe for the "FL" part)

[03:27:03.0000] <Hixie>
Lachy: see, kinda hard to come up with a good counter argument

[03:28:00.0000] <Lachy>
Hixie, not hard, just busy with other stuff right now.

[03:29:00.0000] <Hixie>
well, if you can find an argument, do let me know, because i couldn't, and that backed me into the corner of allowing it.

[03:30:00.0000] <Hixie>
and i don't like it

[03:33:00.0000] <Lachy>
can you point me to Sam's arguments?

[03:37:00.0000] <Hixie>
he didn't really argue for anything as i recall

[03:37:01.0000] <Hixie>
but see the dogfood thread in public-html, e.g.

[03:38:00.0000] <Hixie>
wow i'm really confused about how to do <ins> and <del> with the implied paragraph stuff

[03:38:01.0000] <Hixie>
i'd better sleep and work on this tomorro

[03:38:02.0000] <Hixie>
nn

[03:42:00.0000] <Camaban>
what the... so hang on, if <p> isn't required, does that mean whitespace in HTML is going to become something to think about? (aside from a couple of browser bugs)

[03:44:00.0000] <Lachy>
Camaban, no

[03:44:01.0000] <Camaban>
presume I'm missing something, I'll go look it up

[03:45:00.0000] <Hemebond>
What makes it an implied <P>?

[03:45:01.0000] <Lachy>
Camaban, white space can't become significant because it would seriously break compatibility

[03:45:02.0000] <OmegaJunior>
It already does... browser bug in MSIE

[03:46:00.0000] <Hemebond>
And isn't whitespace kept in Mozilla too?

[03:46:01.0000] <Camaban>
Lachy: that's what I'm thinking, but if <div><p>foo</p> bar <p>baz</p></div> represents 3 paragraphs, then it looks like whitespace would be significant

[03:46:02.0000] <Hemebond>
Between parent and child elements?

[03:46:03.0000] <OmegaJunior>
It shouldn't...

[03:47:00.0000] <Lachy>
Hemebond, <p> will never be implied directly. But rather, the concept of an implied paragraph is a little more abstract

[03:47:01.0000] <Hemebond>
I believe (maybe they fixed it) if you iterated through the children of an <OL>, you'd pick up the whitespace/line-returns too.

[03:47:02.0000] <Lachy>
Camaban, even if you took out the whitepace in that example, it would still represent 3 paragraphs in the new model

[03:47:03.0000] <OmegaJunior>
Yes, Hemebond, that is normal

[03:47:04.0000] <Hemebond>
Lachy: So it isn't marked up with anything except the DIV?

[03:47:05.0000] <OmegaJunior>
But it shouldn't be displayed by the browser

[03:48:00.0000] <Hemebond>
OmegaJunior: But.... why is there whitespace between elements where there shouldn't be any other child elements?

[03:48:01.0000] <Hemebond>
<OL> can only have <LI> as a child element, no?

[03:48:02.0000] <Camaban>
ok, so how would soemthing like http://pastebin.com/d367d5a2c be interpreted? as 1 paragraph, or 2?

[03:48:03.0000] <OmegaJunior>
Hemebond: because the source contains new-line or carriage-return?

[03:48:04.0000] <OmegaJunior>
That counts as whitespace, too.

[03:49:00.0000] <Hemebond>
OmegaJunior: I'm talking about the DOM though.

[03:49:01.0000] <OmegaJunior>
Yes

[03:49:02.0000] <Lachy>
the whitespace will still be present in the DOM, but as far as semantics are concerned, it's very much irrelevant

[03:49:03.0000] <Lachy>
I gotta go

[03:49:04.0000] <OmegaJunior>
The DOM sees spaces, carriage returns, new-lines, tabs, etc. as white-space text nodes

[03:49:05.0000] <OmegaJunior>
unless they're already inside a text

[03:50:00.0000] <Hemebond>
OmegaJunior: I know.

[03:50:01.0000] <Philip`>
"Inter-element whitespace, comment nodes, and processing instruction nodes must be ignored when establishing whether an element matches its content model or not, and must be ignored when following algorithms that define document and element semantics."

[03:50:02.0000] <Philip`>
(says HTML5)

[03:50:03.0000] <OmegaJunior>
Excellent

[03:50:04.0000] <OmegaJunior>
Let's hope the browser vendors will implement it

[03:51:00.0000] <othermaciej>
that's not a requirement for browsers

[03:51:01.0000] <OmegaJunior>
How so?

[03:51:02.0000] <Philip`>
so they're in the DOM, and included in DOM childNodes and stuff, but are not relevant for conformance or for HTML5-defined algorithms (e.g. to find the list items in a list)

[03:51:03.0000] <othermaciej>
that rule applies to documents and validators

[03:52:00.0000] <Philip`>
othermaciej: The "must be ignored when following algorithms ..." sounds like browser conformance

[03:52:01.0000] <othermaciej>
"when following algorithms that define document and element semantics"

[03:52:02.0000] <OmegaJunior>
othermaciej: it applies to documents and validators but not to browsers? That seems contradictory.

[03:52:03.0000] <Philip`>
at least to the extent that browsers run those algorithms

[03:52:04.0000] <othermaciej>
in particular it wouldn't apply to the parsing algorithm, as written

[03:52:05.0000] <Philip`>
(which should be a non-zero extent, I guess)

[03:52:06.0000] <othermaciej>
since that does not define document or element semantics

[03:53:00.0000] <OmegaJunior>
No... but it interprets it...

[03:53:01.0000] <Hemebond>
The DOM isn't the document?

[03:53:02.0000] <OmegaJunior>
No, just a model of it

[03:53:03.0000] <Hemebond>
Which the browser uses to build the document.

[03:54:00.0000] <othermaciej>
I don't know what algorithms are considered to define document and element semantics and which browsers must follow

[03:54:01.0000] <othermaciej>
off hand

[03:54:02.0000] <othermaciej>
but I don't think the quote from Philip` is primarily aimed at browser algorithms that would be relevant here, thus, "let's hope browser vendors will implement it" is not really a proper response

[03:55:00.0000] <OmegaJunior>
I disagree

[03:55:01.0000] <Philip`>
/me agrees with othermaciej :-)

[03:55:02.0000] <othermaciej>
it's true that <div><p>foo</p> <p>bar</p></div> should generate a text node in the DOM for the space, but not render it

[03:56:00.0000] <OmegaJunior>
What good does it do, to describe how whitespace should be handled, if that description doesn't apply to how a browser should handle whitespace while parsing and displaying a document?

[03:56:01.0000] <othermaciej>
but the reasons are the HTML5 parsing algorithm and CSS whitespace collapsing rules

[03:56:02.0000] <othermaciej>
there are descriptions of how whitespace should be handled while parsing and displaying a document

[03:56:03.0000] <othermaciej>
for parsing, that is the parsing algorithm

[03:57:00.0000] <othermaciej>
for display, it is CSS (mainly whitespace collapsing, possibly some other stuff)

[03:57:01.0000] <Hemebond>
the parsing alg is different to the alg for generating a DOM?

[03:57:02.0000] <Philip`>
The parsing algorithm is what generates the DOM

[03:57:03.0000] <OmegaJunior>
othermaciej, And as everyone will understand, the parsing algorythem is developed without any contact with the rest of the HTML5 spec.

[03:57:04.0000] <OmegaJunior>
Difficult word, algorythm

[03:57:05.0000] <othermaciej>
the HTML5 parsing algorithm describes how to take a text/html document and generate a DOM (with various possible side effects along the way)

[03:58:00.0000] <Hemebond>
This is confusing. And messy.

[03:58:01.0000] <OmegaJunior>
So how whitespace should be handled in a document has no bearing on how a parser should handle whitespace in a document?

[03:58:02.0000] <OmegaJunior>
Hemebond: I agree

[03:59:00.0000] <Hemebond>
I think I'll go learn XHTML2.

[03:59:01.0000] <Philip`>
If you have <div><p>foo</p> <p>bar</p></div> in XHTML, you get a text node between the two 'p' nodes, exactly the same as in HTML

[04:00:00.0000] <Philip`>
(in terms of the DOM, which the document gets parsed into)

[04:00:01.0000] <OmegaJunior>
Yes, I'd expect as much.

[04:00:02.0000] <Philip`>
(and if you write scripts to manipulate the DOM, you'll hit those text nodes)

[04:00:03.0000] <othermaciej>
the web is messy

[04:00:04.0000] <Hemebond>
Very.

[04:00:05.0000] <Philip`>
so XHTML2 won't make this any less confusing :-)

[04:00:06.0000] <Hemebond>
So why is HTML5 trying to validate that mess instead of cleaning it up?

[04:00:07.0000] <OmegaJunior>
The HTML5-project should try to avoid adding to the mess.

[04:01:00.0000] <Philip`>
What should HTML5 do instead?

[04:01:01.0000] <Philip`>
(in this particular case)

[04:01:02.0000] <Hemebond>
Well, is inline content allowed in a DIV?

[04:02:00.0000] <Philip`>
Today HTML5 says "Content model:  Zero or more style elements, followed by either zero or more block-level elements, or inline-level content (but not both)."

[04:02:01.0000] <Philip`>
(but that is being rewritten by Hixie)

[04:02:02.0000] <Philip`>
so it depends on what "is" is :-)

[04:03:00.0000] <OmegaJunior>
Yes, inline-content is allowed in a DIV, but not together with block-content.

[04:03:01.0000] <OmegaJunior>
(current spec)

[04:03:02.0000] <Hemebond>
Then document is invalid and should fail.

[04:04:00.0000] <Philip`>
But the current spec says inter-element whitespace is not significant for whether a document is valid; otherwise it'd be horrible because you couldn't e.g. write a list with one <li>item...</li> per line

[04:05:00.0000] <Hemebond>
I was thinking of the original example that had text between the nodes.

[04:05:01.0000] <Hemebond>
Whitespace between elements should be ignored.

[04:06:00.0000] <OmegaJunior>
Agreed

[04:06:01.0000] <OmegaJunior>
That whitespace exists for the benefit of the source programmers and readers.

[04:06:02.0000] <Philip`>
Okay - that one is currently invalid, but apparently there has been no convincing reason why it should be invalid, and it causes problems for people migrating from HTML and provides no benefits

[04:07:00.0000] <Hemebond>
Boohoo. Their code is shit.

[04:07:01.0000] <Philip`>
Their code is valid HTML4

[04:07:02.0000] <Hemebond>
Yeah.

[04:07:03.0000] <Hemebond>
So they can stay with HTML4.

[04:07:04.0000] <Hemebond>
Or clean it up.

[04:07:05.0000] <Hemebond>
HTML4 allows inline and block in the same div?

[04:07:06.0000] <OmegaJunior>
Yeah

[04:08:00.0000] <Hemebond>
Awesome.

[04:08:01.0000] <Hemebond>
Why not allow it here in HTML5 then?

[04:08:02.0000] <OmegaJunior>
Hemebond: because the rendering is undefined

[04:08:03.0000] <OmegaJunior>
Which leads to differences between browser

[04:08:04.0000] <Hemebond>
I see.

[04:09:00.0000] <Hemebond>
Hmmm

[04:09:01.0000] <Hemebond>
Render it inline, at the top of the DIV.

[04:09:02.0000] <Hemebond>
With block elements rendering below it.

[04:09:03.0000] <Hemebond>
Haha

[04:09:04.0000] <Philip`>
http://validator.w3.org/check?fragment=%3C!DOCTYPE%20HTML%20PUBLIC%20%22-//W3C//DTD%20HTML%204.0//EN%22%3E%20%3Ctitle%3E%3C/title%3E%20%3Cdiv%3E%3Cp%3Efoo%3C/p%3E%20bar%20%3Cp%3Ebaz%3C/p%3E%3C/div%3E

[04:09:05.0000] <Hemebond>
That'd mess with people.

[04:09:06.0000] <Philip`>
This Page Is Valid HTML 4.0 Strict!

[04:09:07.0000] <Hemebond>
w3.org validator means nothing.

[04:10:00.0000] <Hemebond>
I mentally slap anyone who claims spec conformance by using that validator.

[04:10:01.0000] <OmegaJunior>
o.O.

[04:10:02.0000] <Hemebond>
In this case of course...

[04:10:03.0000] <Hemebond>
it does match.

[04:10:04.0000] <Camaban>
and the one you use is.....

[04:11:00.0000] <Hemebond>
Got a link somewhere....

[04:11:01.0000] <Philip`>
http://htmlhelp.com/cgi-bin/validate.cgi?url=http%3A%2F%2Fphilip.html5.org%2Fmisc%2Fdivblockinline.html&warnings=yes

[04:11:02.0000] <OmegaJunior>
To my knowledge the w3 validator is authoritive?

[04:11:03.0000] <Hemebond>
It's no longer developed as far as I know.

[04:11:04.0000] <Hemebond>
And not complete.

[04:12:00.0000] <Hemebond>
According to ex-w3 members.

[04:12:01.0000] <Hemebond>
Philip`: Yes, I think that's the one.

[04:12:02.0000] <Philip`>
"Congratulations, no errors!"

[04:12:03.0000] <Hemebond>
Yeah... I know.

[04:12:04.0000] <Hemebond>
Because it is valid HTML4.

[04:12:05.0000] <Hemebond>
My comment was just slightly OT.

[04:13:00.0000] <Hemebond>
Just an aside.

[04:14:00.0000] <Philip`>
Hmm, Validome doesn't work at all - it says the doctype is missing

[04:15:00.0000] <Philip`>
Oh, right, it's just broken in Opera

[04:16:00.0000] <Philip`>
Hemebond: The W3C validator is still actively developed

[04:16:01.0000] <Camaban>
http://validator.w3.org/whatsnew.html#t2007-10-11 no longer developed? 

[04:16:02.0000] <Hemebond>
*shrug*

[04:16:03.0000] <Hemebond>
Just repeating what I read.

[04:16:04.0000] <Camaban>
and that actually the validator at w3.org and htmlhelp are identical

[04:18:00.0000] <Hemebond>
1:30am

[04:18:01.0000] <Hemebond>
Night all.

[04:20:00.0000] <Philip`>
Good night

[04:20:01.0000] <OmegaJunior>
Night!

[05:22:00.0000] <hdh>
/me thought of tagging white-space node with irrelevant

[05:40:00.0000] <Philip`>
hdh: Adding the 'irrelevant' attribute to white-space text nodes in the DOM? I thought only Element nodes could have attributes

[05:42:00.0000] <hdh>
hm, okay

[12:04:00.0000] <nikola_tesla>
/me changes topic to ''

[12:16:00.0000] <gsnedders>
/me wonders what the point of \paragraph and \subparagraph is in LaTeX

[12:21:00.0000] <hsivonen>
isn't \paragraph a bit like \theorem or somesuch--not a generic paragraph?

[12:27:00.0000] <zcorpan>
is it possible to throw proper DOMExceptions with javascript? e.g. INVALID_STATE_ERR ?

[12:28:00.0000] <zcorpan>
throw new Error('INVALID_STATE_ERR') isn't it

[12:28:01.0000] <gsnedders>
hsivonen: \theorem? :P

[12:32:00.0000] <Philip`>
He stole our topic :-(

[12:33:00.0000] <gsnedders>
/me blames him

[12:33:01.0000] <Philip`>
/me wonders if anyone has a backup copy

[12:34:00.0000] <hdh>
WHATWG (HTML5) -- http://www.whatwg.org/ -- Logs: http://krijnhoetmer.nl/irc-logs/ -- Please leave your sense of logic at the door, thanks!

[12:34:01.0000] <gsnedders>
/me changes topic to 'WHATWG (HTML5) -- http://www.whatwg.org/ -- Logs: http://krijnhoetmer.nl/irc-logs/ -- Please leave your sense of logic at the door, thanks!'

[12:34:02.0000] <Philip`>
Ah, thanks :-)

[12:34:03.0000] <hdh>
:) pidgin powered

[12:41:00.0000] <hsivonen>
gsnedders: isn't there a \theorem in LaTeX?

[12:41:01.0000] <gsnedders>
hsivonen: no idea. I've only started properly learning LaTeX tonight :P

[13:20:00.0000] <jgraham_>
Er, well LaTeX can have \anything-you-like with the right style file :)

[13:21:00.0000] <jgraham_>
But I think \theorem is pretty standard

[13:23:00.0000] <jgraham_>
see http://www.math.uiuc.edu/~hildebr/tex/theorems.html - it seems to be included in amslatex

[13:25:00.0000] <jgraham_>
actually I think I'm wrong about that \begin{theorem} seems to be in ordinary LaTeX

[14:15:00.0000] <gsnedders>
/me wonders whether the following will confuse his computing teacher:

[14:15:01.0000] <gsnedders>
"Reliability is the consistency that the program creates the correct output 

[14:15:02.0000] <gsnedders>
for any given input (to the extent that output can be found given 

[14:15:03.0000] <gsnedders>
the halting problem)."

[14:17:00.0000] <Philip`>
I don't see why the halting problem is at all relevant

[14:17:01.0000] <gsnedders>
actually, it shouldn't create any output, should it, therefore meaning it is irrelevant

[14:17:02.0000] <gsnedders>
/me headdesks

[14:17:03.0000] <Philip`>
Uh, I'm not sure what you mean

[14:18:00.0000] <gsnedders>
Philip`: don't worry, just how I ended up with the halting problem being relevant in my head (which is stupid)

[14:19:00.0000] <Philip`>
It might be relevant that something which doesn't give an incorrect answer because it doesn't give any answer (i.e. it doesn't terminate) is incorrect, but that's not the halting problem

[14:19:01.0000] <gsnedders>
Philip`: mainly down to whether it should be creating any output for a certain input that makes the algorithm fall into the halting problem

[14:20:00.0000] <gsnedders>
/me has already forgotten half his reasoning for it anyway

[14:20:01.0000] <gsnedders>
and trying to re-create it is failing

[14:21:00.0000] <Philip`>
It's only a problem if you want to prove whether the program does or does not produce output, and I'm not sure why you'd want to prove that :-)

[14:22:00.0000] <gsnedders>
Philip`: I did, indirectly, without explaining myself anywhere near enough :)

[14:22:01.0000] <gsnedders>
And it's so vaguely relevant and the halting problem isn't even taught so you couldn't really rely on my teacher having a clue what it is :\

[14:22:02.0000] <Philip`>
The halting problem is largely irrelevant in reality anyway, because a tool which decides "this program does terminate" or "I'm not sure this program terminates" is possible and useful (and e.g. implemented by Microsoft)

[14:24:00.0000] <Philip`>
Also, computers are finite

[14:24:01.0000] <jgraham_>
/me can suddenly access whatwg.org again :)

[14:24:02.0000] <Philip`>
So I don't understand why people use the halting problem as an excuse so often :-)

[14:26:00.0000] <gsnedders>
Philip`: "… we're only immortal for a limited time."

[14:29:00.0000] <Philip`>
(I also don't understand http://www.ietf.org/internet-drafts/draft-gregorio-uritemplate-02.txt talking about Turing completeness, because XML DTDs are not Turing complete and you can do things like the billion laughs attack in non-Turing-complete languages easily since a billion is finite)

[14:32:00.0000] <gsnedders>
/me doesn't really under turning completeness

[14:33:00.0000] <jgraham_>
Philip`: Do you have any data on how many pages use <form method=foo> with foo not in the set (get, post, put, delete)?

[14:34:00.0000] <Philip`>
If a language is Turing complete, it kind of means you can emulate a Turing machine in the language and you can emulate that language on a Turing machine, and hence the language is equivalent in expressiveness to every other Turing complete language, except the details are a bit more complex and that's probably inaccurate :-)

[14:35:00.0000] <jgraham_>
(Turing machines are kinda hard to come by. Specifically the infinite rolls of tape aren't stocked by Tesco)

[14:36:00.0000] <Philip`>
(That's not a problem - you can predict how fast the tape will be used up, and just buy new rolls and sellotape them onto the end of the previous one when it's about to be used up)

[14:36:01.0000] <gsnedders>
(What's a turning machine?)

[14:36:02.0000] <hsivonen>
jgraham_: the tape only needs to be long enough for whatever you are computing if the program terminates :-)

[14:36:03.0000] <hsivonen>
of course, figuring out if it terminates so you can go buy tape is hard

[14:37:00.0000] <Philip`>
http://c2.com/cgi/wiki?BloopFloopAndGloop - you can do vaguely useful things without being Turing complete

[14:37:01.0000] <Philip`>
gsnedders: (Do you mean "Turing" rather than "turning"?)

[14:37:02.0000] <gsnedders>
Philip`: (yes)

[14:37:03.0000] <gsnedders>
(I always do that)

[14:37:04.0000] <Philip`>
(http://en.wikipedia.org/wiki/Turing_machine is the obvious place to look :-) )

[14:37:05.0000] <Philip`>
(but, er, it looks quite a long page)

[14:39:00.0000] <gsnedders>
yeah

[14:39:01.0000] <gsnedders>
it is, and confusing.

[14:39:02.0000] <gsnedders>
/me wonders what \relax is in LaTeX

[14:40:00.0000] <Philip`>
http://www.abelard.org/turpap2/tp2-ie.asp is more confusing

[14:40:01.0000] <jgraham_>
gsnedders: if you ever have quite a lot of time to spare you might enjoy something like The Emperor's New Mind by Roger Penrose. It touches on Turing machines, Godel's theorem and lots of other interesting things in a mostly-comprehensible way

[14:41:00.0000] <gsnedders>
/me wishes he had more spare time

[14:41:01.0000] <jgraham_>
/me wishes he made better use of his time

[14:41:02.0000] <gsnedders>
/me finishes computing homework for today

[14:42:00.0000] <gsnedders>
I better do the homework for the Friday before last tomorrow, then.

[14:43:00.0000] <hdh>
last tomorrow? end of the world ?-)

[14:43:01.0000] <Philip`>
jgraham_: http://www.cl.cam.ac.uk/~pjt47/misc/method.txt is total number of occurrences

[14:43:02.0000] <Philip`>
(out of ~15K pages)

[14:44:00.0000] <Philip`>
(with XML escaping that shouldn't be there)

[14:44:01.0000] <Philip`>
/me fixes the escaping

[14:45:00.0000] <jgraham_>
Philip`: Thanks

[14:45:01.0000] <Philip`>
http://www.showroomnavi.com/ - <form name="area01" method=post" action="cgi-bin/awsearch.cgi" accept-charset="x-euc-jp"> - that is all the post" ones

[14:50:00.0000] <hdh>
410 fits better, no?

[14:51:00.0000] <zcorpan>
/me made a new http://simon.html5.org/tools/js/innerhtml-viewer/

[14:52:00.0000] <zcorpan>
with features that are not in the spec (yet)! (see the js file)

[15:20:00.0000] <zcorpan>
Hixie: try this in firefox: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3Caddress%3E%3Ctable%3E%3C%2Ftable%3E%3C%2Faddress%3E%3Caddress%3E%3Col%3E%3C%2Fol%3E%3C%2Faddress%3E%3Caddress%3E%3Cul%3E%3C%2Ful%3E%3C%2Faddress%3E%3Caddress%3E%3Caddress%3E%3C%2Faddress%3E%3C%2Faddress%3E%3Caddress%3E%3Ch1%3E%3C%2Fh1%3E%3C%2Faddress%3E (re why <address><ol> can't be allowed, although i guess we could allow it despite

[15:22:00.0000] <Hixie>
aw man

[15:22:01.0000] <Hixie>
firefox screws us again

[15:27:00.0000] <Hixie>
hey anyone got any publicly visible java implementations of the tree construction phase?

[15:31:00.0000] <jgraham_>
Hixie: Doesn't hsivonen's imp. do what you want (I feel I'm missing something)

[15:31:01.0000] <Hixie>
is it online?

[15:31:02.0000] <Hixie>
i guess it must be

[15:31:03.0000] <jgraham_>
Do you want it to print the tree online or just to download it?

[15:32:00.0000] <Hixie>
oh i just meant the source

[15:32:01.0000] <jgraham_>
http://about.validator.nu/htmlparser/

[15:32:02.0000] <Hixie>
sweet

[15:33:00.0000] <jgraham_>
Hmm. I wonder why gecko adds random attributes to heading elements in misnested formatting elements

[15:33:01.0000] <Hixie>
which attributes?

[15:34:00.0000] <hdh>
_moz_something

[15:34:01.0000] <Hixie>
yeah it's the "something" part i was wondering about

[15:35:00.0000] <jgraham_>
actually to the formatting elements themselves 

[15:35:01.0000] <jgraham_>
_moz-rs-heading=""

[15:35:02.0000] <Hixie>
odd

[15:35:03.0000] <Hixie>
dunno

[15:36:00.0000] <jgraham_>
It seems to be so that they can set font-size:inherit

[15:59:00.0000] <Hixie>
wohayhay, an actual deadline for publication of html5 fpwd


2007-12-19
[16:00:00.0000] <hober>
yay!

[16:01:00.0000] <hober>
/me wonders if that'll speed up the release of the msft review

[16:31:00.0000] <webben>
Vaguely OT question: Does anyone have an opinion on the best way of declaring a XML syntax for XHTML5-like languages? Is Relax NG capable of declaring a syntax of that sort of complexity, or would one need schema?

[16:32:00.0000] <Hixie>
what do you mean by "declaring a XML syntax"?

[16:33:00.0000] <webben>
Hixie: What elements can be contained by what. What content is allowed in attributes.

[16:33:01.0000] <webben>
yeah basically nesting rules + data types.

[16:33:02.0000] <Hixie>
i prefer to use english prose to define that

[16:33:03.0000] <webben>
Hixie: I'm talking about for validation purposes.

[16:33:04.0000] <hdh>
validator.nu is using RN, the files are in http://svn.versiondude.net/whattf/syntax/

[16:34:00.0000] <Hixie>
oh, for validatior

[16:34:01.0000] <Hixie>
validation, even

[16:34:02.0000] <webben>
hdh: Ah, interesting. Thanks! :)

[16:34:03.0000] <hdh>
I added them to nxml-mode, but got stuck with w: datatype lib

[16:34:04.0000] <Hixie>
which schema language is best probably depends on the specific rules of the language

[16:34:05.0000] <Hixie>
personally i'd just implement the whole thing in code, but that's just me :-)

[16:37:00.0000] <webben>
/me isn't sure what the code for that would look like... but that might be an option further down the road.

[16:40:00.0000] <Philip`>
You could invent your own schema language to cope with HTML5's requirements

[16:40:01.0000] <Philip`>
(Warning: may not be trivial)

[16:40:02.0000] <Hixie>
it could be really trivial

[16:41:00.0000] <Hixie>
the schema language could be designed so that the empty string represents the html5 grammar, and then the schema would just be "".

[16:41:01.0000] <Philip`>
Specifying that schema language is hard, though

[16:41:02.0000] <Hixie>
Specifying that schema language is my day job.

[16:42:00.0000] <inimino>
Philip`: not if you can use HTML5 as a normative reference

[16:44:00.0000] <Philip`>
Specifying that schema language well (i.e. built up in a formal way from primitive components) is still hard :-)

[16:44:01.0000] <Hixie>
implementing the schema language is hard, yes

[16:45:00.0000] <Hixie>
that's hsivonen's day job :-)

[16:45:01.0000] <inimino>
hehe true

[16:57:00.0000] <Hixie>
sweet kittens. if it's not my site that's down it's the w3c's.

[16:57:01.0000] <Hixie>
how am i supposed to get any work done.

[16:57:02.0000] <hdh>
CSS to emulate irrelevant="" would be {width: 0; height: 0; visibility: hidden;}, right?

[16:58:00.0000] <Hixie>
[irrelevant] { display: none; }

[16:58:01.0000] <hdh>
ok, thanks

[16:58:02.0000] <kingryan>
Hixie: w3.org appears to work for me

[16:59:00.0000] <Hixie>
lists.w3.org was down briefly

[16:59:01.0000] <Hixie>
it's back up now

[16:59:02.0000] <kingryan>
ah

[17:05:00.0000] <gsnedders>
/me needs some soul searching help

[17:06:00.0000] <Hixie>
Lachy: someone's talking about you here http://etfb.livejournal.com/83307.html

[17:07:00.0000] <gsnedders>
LiveJournal! yay! :P

[17:15:00.0000] <Hixie>
holy. war. on. kittens.

[17:15:01.0000] <Hixie>
now my pubrules checking web service host is down.

[17:19:00.0000] <Philip`>
In the future, we will be able to download applications and run them on our own computers, which will solve all these downtime problems

[17:20:00.0000] <gsnedders>
/me points at what Eric Smit (forgive my spelling) said a few days ago: 90% of everything can be done online

[17:21:00.0000] <Philip`>
/me finds it's more useful to do 100% of something, rather than getting stuck just before the end because someone else broke something outside your control

[17:21:01.0000] <Hixie>
Philip`: there are so many reasons that wouldn't work...

[17:43:00.0000] <csarven>
/me thinks <aside> is a bad name

[17:44:00.0000] <gsnedders>
/me moves csarven aside

[17:44:01.0000] <csarven>
good one :)

[17:44:02.0000] <Hixie>
csarven: what do you propose instead?

[17:45:00.0000] <csarven>
Hixie <section> would have been sufficient imo

[17:45:01.0000] <Hixie>
<section> implies you should read it

[17:46:00.0000] <Hixie>
i wanted something for, e.g., the examples and notes in the spec

[17:46:01.0000] <csarven>
what to read is subjective

[17:46:02.0000] <Hixie>
<section> implies its contents are an intergal, inline, part of its container

[17:47:00.0000] <csarven>
id="sidebar" -> <aside> -- im not sure if this is really better. id="sidebar" -> id="supplemental_information" would have been okay and perhaps <supplemental>

[17:47:01.0000] <csarven>
<aside> indicates some sort of a physical direction

[17:48:00.0000] <csarven>
(imo)

[17:48:01.0000] <Hixie>
supplemental is harder to type

[17:48:02.0000] <othermaciej>
<aside> sounds less presentational than <sidebar>

[17:48:03.0000] <hdh>
something for the footnotes

[17:48:04.0000] <othermaciej>
but it also sounds like you'd use it for inline parenthetical remarks

[17:48:05.0000] <gsnedders>
I wanted to work on HTML 5. You probably don't care, but I was away working for money instead, getting paid to drop Ogg.

[17:48:06.0000] <Philip`>
<blockaside> would stop it sounding like it's inline

[17:48:07.0000] <gsnedders>
(where the latter sentence is an aside)

[17:48:08.0000] <hdh>
and longer to say

[17:48:09.0000] <gsnedders>
My former maths teachers was always talking about asides

[17:48:10.0000] <gsnedders>
how about… <sidenote>?

[17:48:11.0000] <Hixie>
<aside> is fine

[17:48:12.0000] <gsnedders>
or isn't that what you're meaning?

[17:49:00.0000] <csarven>
Hixe <supp> or how about <wazzup> :P

[17:49:01.0000] <hdh>
inline aside can use small?

[17:49:02.0000] <csarven>
Hixie

[17:49:03.0000] <hdh>
oops, remember it wrong

[17:49:04.0000] <gsnedders>
Hixie: still livin' high the life after you pay-off from MS/Apple/Nokia? :P

[17:49:05.0000] <Philip`>
<div lessrelevant>

[17:50:00.0000] <gsnedders>
<div gsneddersisawesome>

[17:50:01.0000] <othermaciej>
wait, I thought Google was bribing Apple

[17:50:02.0000] <othermaciej>
to stop XHTML2 so that AdSense can keep working

[17:50:03.0000] <othermaciej>
did we renegotiate the bribe agreement?

[17:50:04.0000] <gsnedders>
othermaciej: I thought they needed to stop XHTML totally, which they already failed at.

[17:51:00.0000] <csarven>
Hixie to me <supp> (short for supplemental) or <aux> (short for auxiliary)

[17:51:01.0000] <othermaciej>
I think <sidebar> would not be so bad

[17:51:02.0000] <othermaciej>
that's what they are called in print

[17:51:03.0000] <othermaciej>
even when not literally on the side

[17:52:00.0000] <othermaciej>
but I don't care about the name that much

[17:52:01.0000] <csarven>
<aside> really comes across as something that exists in a physical space. to me HTML documents don't (or perhaps shouldn't) convey that information

[17:52:02.0000] <csarven>
<sidebar> is worse imo because again that suggests a location

[17:52:03.0000] <hdh>
should footnotes be inside <aside> too

[17:53:00.0000] <Hixie>
/me mumbles something about giving you all cans of paint so you can paint the bikeshed yourselves :-P

[17:53:01.0000] <csarven>
supplemental information (that we usually see on the side as a sidebar) can be located anywhere.. it could be before or below the document even

[17:53:02.0000] <csarven>
Hixie oh c'mon. just reasoning out. i'm willing to hear the problem with my suggestion

[17:53:03.0000] <othermaciej>
http://en.wikipedia.org/wiki/Aside

[17:53:04.0000] <othermaciej>
it's not about physical space

[17:53:05.0000] <Hixie>
csarven: i don't see the problem with <aside>

[17:54:00.0000] <gsnedders>
Hixie: can I have pink and blue, so I can paint my bikeshed pink with blue spots?

[17:55:00.0000] <gsnedders>
/me waves g'nite before he can get his paint (just throw it at him after he leaves)

[17:56:00.0000] <csarven>
Hixie lets start here then: do you think <aside> is more accurate then <supplemental> ?

[17:56:01.0000] <Hixie>
mildly so, but not especially, no

[17:56:02.0000] <Hixie>
there are dozens of words that would be equally good

[17:57:00.0000] <Hixie>
<sidebar>, <aside>, <comment>, <note>, <supplemental>, etc

[17:57:01.0000] <Hixie>
some have problems (e.g. <comment> is display:none in some old versions of IE)

[17:57:02.0000] <Hixie>
(and supplemental is too long)

[17:58:00.0000] <Hixie>
but at the end of the day, they're not especially better or worse than each other

[17:58:01.0000] <othermaciej>
the name is not terribly important

[17:58:02.0000] <hdh>
aux is a nice fit for Dvorak, supp too :)

[18:00:00.0000] <csarven>
/me wonders if <supplemental> would be the longest tag out there

[18:00:01.0000] <othermaciej>
on the one hand, the time to fine-tune tag names is drawing short (as browsers start implementing); but on the other hand, it's not clear that it would be the best use of time to do so

[18:00:02.0000] <othermaciej>
unless there's an actual practical problem with specific element names

[18:00:03.0000] <kingryan>
/me is laughing at "Mr "Burn All Oggs" himself"

[18:01:00.0000] <Hixie>
the names were fine tuned months ago

[18:01:01.0000] <Hixie>
(<aside> was originally <sidebar>, <time> originally <t>, etc)

[18:01:02.0000] <csarven>
im glad <sidebar> changed to <aside>

[23:22:00.0000] <Hixie>
http://www.petitiononline.com/lortow3/petition.html <-- one of the most pointless petitions ever

[23:25:00.0000] <Hixie>
the text was removed because people said they'd ignore it... but sure... let's put it back... i'm sure they won't ignore it then!

[23:40:00.0000] <hsivonen>
<aside> WFM

[23:57:00.0000] <Hemebond>
?

[23:57:01.0000] <Hemebond>
Why would WHATWG define a codec requirement?

[23:58:00.0000] <Hixie>
Hemebond: to encourage convergence on a common codec

[23:58:01.0000] <Hixie>
so that people don't have to encode everything two or more times

[23:58:02.0000] <hober>
Hemebond: it's required in order for <video> to actually be interoperable

[00:35:00.0000] <Hemebond>
But...

[00:35:01.0000] <Hemebond>
hmm

[00:35:02.0000] <Hemebond>
Isn't the web suppose to be dumb?

[00:38:00.0000] <hsivonen>
the IP layer is dumb. the Web is lazy.

[01:09:00.0000] <othermaciej>
any Opera folks around?

[01:42:00.0000] <hsivonen>
on case anyone is wondering, I am aware that Validator.nu is in a semi-broken state. 

[01:43:00.0000] <hsivonen>
(I should have had the common sense not to install kernel updates right now...)

[02:06:00.0000] <Hemebond>
Hey.... question.

[02:07:00.0000] <Hemebond>
XHTML is modular... can be extended. Could browser makers not create their own modules to add in stuff they wanted? Kind of like extensions in OpenGL?

[02:08:00.0000] <Philip`>
HTML can be extended without being modular or namespaced - see e.g. Apple adding elements like <canvas>

[02:09:00.0000] <hsivonen>
Hemebond: the modularity is in the dtd and has nothing to do with browsers

[02:09:01.0000] <hsivonen>
canvas did not need modularization to happen

[02:09:02.0000] <Hemebond>
Except that it muddies the markup with elements/tags that don't exist in the DTD. No?

[02:09:03.0000] <Philip`>
Hemebond: That's easy to solve by just ignoring the DTD :-)

[02:12:00.0000] <Philip`>
If you add an extension to (X)HTML by any method, it's not going to be in the standard DTD, and there is not much value in creating a new non-standard DTD with the extension because that doesn't solve any useful problems except for letting DTD-based validators claim your non-standard document is valid

[02:15:00.0000] <Hemebond>
I thought you could... like namespaces and such. Like mixing in <svg:blah> and such.

[02:16:00.0000] <Hemebond>
I guess that's not a module/extension.

[02:16:01.0000] <Hemebond>
If browser makers created a seperate namespace for their extensions.

[02:16:02.0000] <Hemebond>
A bit like CSS does.

[02:17:00.0000] <Philip`>
If we had <apple:canvas>, and people started using it, what would happen when Mozilla started supporting the same feature?

[02:18:00.0000] <Hemebond>
What happens in CSS when more than one browser supports an extension?

[02:19:00.0000] <webben>
Philip`: What would happen? They could implement <apple:canvas> (indeed in XHTML5 one could write apple:canvas, since that's a prefix not a namespace)

[02:20:00.0000] <Hemebond>
And if it ever became a part of the proper "standard", it would move to the XHTML namespace.

[02:20:01.0000] <Philip`>
Hemebond: People write "-moz-border-radius: 5px; -webkit-border-radius: 5px; border-radius: 5px;" which isn't very pretty and favours certain browsers and will still break if the official CSS definition of border-radius is changed in the future

[02:20:02.0000] <jgraham_>
/me wonders how likely MS would be to implement something in the http://apple.com/ns/ namespace

[02:21:00.0000] <jgraham_>
(that NS URI is made u of course)

[02:21:01.0000] <krijn>
/me just wrote -moz-border-radius: 10px 0 0 10px; -webkit-border-top-left-radius: 10px; -webkit-border-bottom-left-radius: 10px;

[02:21:02.0000] <webben>
Philip`: If they only added border-radius when finalized, it wouldn't break

[02:22:00.0000] <Philip`>
webben: It seems that people don't want to wait for that, because they want their current pages to work in future CSS3 UAs, e.g. in http://virtuelvis.com/archives/2004/11/imageless-rounded-corners

[02:22:01.0000] <webben>
jgraham: As likely as they are to implement an Apple invention like canvas in the first place. Because of exceptional market demand. (So far not much sign of that, and they've got Silverlight to push...)

[02:22:02.0000] <webben>
Philip`: Yes... but that's a problem with extensions that aren't "namespaced" not extensions that are.

[02:23:00.0000] <webben>
Or a problem with people using experimental drafts.

[02:23:01.0000] <webben>
it's a different problem.

[02:23:02.0000] <jgraham_>
webben: You think? I would have thought that they would be much less likely to implement something that required their competitors name to be associated with a useful feature

[02:24:00.0000] <Hemebond>
They would put it in their own namespace.

[02:24:01.0000] <webben>
jgraham_: Apple's name is associated with canvas. Namespaced or not.

[02:24:02.0000] <Hemebond>
Each browser would put it in their own namespace.

[02:26:00.0000] <Philip`>
webben: It's a problem with vendor-namespaced extensions that are later standardised into differently-namespaced extensions, because people will want to use the differently-namespaced extension to get forward-compatibility, even though they can only test pre-standard implementations, and so it limits one of the benefits of using vendor-namespaced extensions

[02:26:01.0000] <webben>
Hemebond: have nested <apple:canvas><microsoft:canvas><firefox:canvas><camino:canvas> could make CSS selection and DOM manipulation triky.

[02:26:02.0000] <Philip`>
webben: That would make Opera users unhappy too :-)

[02:26:03.0000] <jgraham_>
webben: Really? By what fraction of the web authoring population? How often is it drummed in? Do you think the association between netscape and <script> is stronger or weaker than it would have been if it had been called <netscape:script>. What about Microsoft and CSS?

[02:27:00.0000] <webben>
Philip`: Yes... but removing vendor-namespaced extensions makes the problem worse, because /everyone/ is forced to use differently-namespaced extensions that break.

[02:27:01.0000] <Hemebond>
webben: That's what happens when you implement non-standard extensions.

[02:29:00.0000] <Hemebond>
http://www.opengl.org/resources/features/OGLextensions/

[02:30:00.0000] <Philip`>
I guess vendor-namespaced extensions work well when nobody uses that extension, so the trick is to make sure it doesn't become popular until after it's been standardised

[02:30:01.0000] <webben>
jgraham_: Unknown about current knowledge. I agree the association would be stronger. Not sure about the CSS reference. I still think that if MS is rabidly avoiding other companies' tech it will only implement out of market demand: and that includes canvas. Having Silverlight makes an implementation much less likely.

[02:31:00.0000] <webben>
jgraham_: In other words, yes it would make a difference but probably not enough of a difference to be a deciding factor.

[02:31:01.0000] <Philip`>
I have a vague memory that Vista's GL implementation supports some APPLE_* extensions, but I might be totally wrong about that and I'm too lazy to check right now

[02:32:00.0000] <Philip`>
(NVIDIA definitely has some e.g. ATI_* extensions, though)

[02:32:01.0000] <Hemebond>
Windows has no OpenGL implementation. Implementations are provided by ICD drivers.

[02:33:00.0000] <Philip`>
Windowses up to XP had a software renderer for OpenGL 1.1, and Vista has GL 1.4 implemented on top of DirectX, if I remember correctly

[02:34:00.0000] <webben>
http://msdn2.microsoft.com/en-us/library/ms537767.aspx : "The Microsoft implementation of OpenGL for the Windows operating system is industry-standard graphics software with which programmers can create high-quality still and animated three-dimensional color images."

[02:34:01.0000] <Hemebond>
Ah, you're right.

[02:36:00.0000] <Philip`>
http://www.delphi3d.net/hardware/viewreport.php?report=1566 - hmm, maybe that's the Vista one, and there's no Apple :-(

[02:37:00.0000] <Philip`>
/me blames his memory

[02:44:00.0000] <webben>
It's notable that Microsoft's aversion to other brands didn't stop them calling IE "Mozilla".

[02:44:01.0000] <webben>
(in the UA string)

[02:45:00.0000] <Hemebond>
Haha yeah.

[02:46:00.0000] <Hemebond>
But most users would never see that so they're fairly safe.

[02:47:00.0000] <webben>
I doubt the most users view source.

[02:47:01.0000] <webben>
*that

[02:53:00.0000] <Philip`>
Most people who saw the IE UA string wouldn't know what "Mozilla" meant

[02:53:01.0000] <Philip`>
(back when it was adopted, when it was just an internal codename used by Netscape)

[02:54:00.0000] <Philip`>
or at least I had no idea what it meant for years after I first saw it :-)

[02:54:01.0000] <webben>
they do now

[02:54:02.0000] <Philip`>
IE can't change it now because it'd break compatibility

[02:54:03.0000] <webben>
yes, so it's still a demonstration of the power of the market

[02:54:04.0000] <Philip`>
regardless of how much they like/dislike advertising their competitors hundreds of billions of times a day

[02:55:00.0000] <webben>
Oh I'm sure they like nothing better ;)

[02:56:00.0000] <othermaciej>
Hixie likes to tell people they should make their user agent way simpler

[02:56:01.0000] <Hemebond>
I thought they had removed it.

[02:57:00.0000] <othermaciej>
but it would break a lot of content

[02:57:01.0000] <othermaciej>
this is Safari's latest UA string:

[02:57:02.0000] <othermaciej>
Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10_5; en-us) AppleWebKit/525.1+ (KHTML, like Gecko) Version/3.0.4 Safari/525.1

[02:58:00.0000] <othermaciej>
it claims to be three different browsers that it actually isn't

[02:58:01.0000] <Philip`>
HTTP_USER_AGENT = Opera/9.24 (X11; Linux i686; U; en)

[02:58:02.0000] <othermaciej>
some other WebKit-based browsers include both Safari and their own name, so they get up to claiming 4 browsers that they are not

[02:59:00.0000] <othermaciej>
at the time we developed it, a lot of the werid details really mattered for compatibility

[02:59:01.0000] <othermaciej>
Opera does indeed have a nice clean one

[03:00:00.0000] <Philip`>
You should solve all your compatibility concerns by having a DOCTYPE switch, and only sending the old ugly UA string to pages written before the switch

[03:00:01.0000] <othermaciej>
has it always had that, or is that a recent change?

[03:00:02.0000] <othermaciej>
but the UA string goes to the web server

[03:00:03.0000] <othermaciej>
doctype switch is too late to know which to send in the http request

[03:00:04.0000] <Hemebond>
Are there any web browsers that actually do everything properly?

[03:00:05.0000] <Hemebond>
No hacks whatsoever,

[03:01:00.0000] <Hemebond>
Supports the specs correctly.

[03:01:01.0000] <Philip`>
I think it was Opera 9 that stopped sending the Mozilla string to everyone

[03:01:02.0000] <Hemebond>
No "error correction" and such?

[03:01:03.0000] <Philip`>
Hemebond: And no bugs?

[03:01:04.0000] <krijn>
http://www.user-agents.org/index.shtml?n_s

[03:01:05.0000] <Hemebond>
One that has a proper SGML parser.

[03:02:00.0000] <othermaciej>
Hemebond: I'm pretty sure the answer is no

[03:02:01.0000] <Hemebond>
lame

[03:02:02.0000] <Philip`>
I imagine a project to make a 'correct' browser would die out quite quickly due to a lack of users

[03:02:03.0000] <webben>
Hemebond: I don't think any maintained browser interprets <br /> in text/html as <br>&gt;

[03:02:04.0000] <othermaciej>
there's a lot of specs, supporting them 100% correctly and with no error handling or extensions would lead to a browser that can't handle many important web sites

[03:03:00.0000] <webben>
Hemebond: (which is what SGML would require) .... so no.

[03:03:01.0000] <webben>
Hemebond: Lynx's strict SGML mode might be close though. But Lynx doesn't support all of HTML 4.01 (no one does)

[03:03:02.0000] <Philip`>
/me wonders what would happen if browsers downloaded DTDs, and a billion users started accessing www.w3.org every day

[03:03:03.0000] <othermaciej>
it's not clear what it would even mean to support all of HTML 4.01

[03:04:00.0000] <webben>
Philip`: Why would they download a DTD rather than using a local copy.

[03:04:01.0000] <Philip`>
webben: Are DTDs allowed to change?

[03:04:02.0000] <krijn>
/me wonders if he should use <!doctype html> on the to-be-launched-today Fronteers website :/

[03:04:03.0000] <webben>
Philip`: They probably shouldn't be.

[03:04:04.0000] <krijn>
People will hate me if I do, and others will hate me if I don't

[03:05:00.0000] <Philip`>
krijn: Some people will love you regardless of what you do, so it doesn't matter that much :-)

[03:05:01.0000] <webben>
Philip`: Although they do occasionally have bugs, which might that problematic.

[03:05:02.0000] <webben>
Philip`: It hardly makes sense to check on a daily basis, anyhow.

[03:06:00.0000] <krijn>
Philip`: they don't care about which sites I build though ;p

[03:06:01.0000] <webben>
Philip`: I guess they could also set some long-term expiration headers.

[03:07:00.0000] <webben>
firefox addons doesn't seem to go down from Fx constantly checking for updated extensions.

[03:08:00.0000] <Hemebond>
Don't extensions check their own sites for updates?

[03:08:01.0000] <webben>
Hemebond: A lot of extensions are hosted on addons.

[03:08:02.0000] <webben>
AFAIK

[03:08:03.0000] <Philip`>
Firefox addon update messages are smaller than DTDs, and the addon system was designed for lots of users, e.g. by having its own domain name rather than sharing with the rest of www.mozilla.org

[03:09:00.0000] <webben>
it was designed a lot later than 1999 ;)

[03:09:01.0000] <webben>
Philip`: You should only have to do a HEAD request to check for DTD updates.

[03:10:00.0000] <Philip`>
I remember Netscape complaining recently about the cost of sending RSS DTDs from their domain to feed readers

[03:10:01.0000] <webben>
Well, they actually pulled the page for a while.

[03:10:02.0000] <Philip`>
Do web browsers use HEAD much? (I've seen a small number of servers which handle it very buggily)

[03:10:03.0000] <webben>
Philip`: Dunno. They /should/ do, but there we go.

[03:11:00.0000] <webben>
Seems to work pretty nicely with curl whenever I try it.

[03:11:01.0000] <Philip`>
(One server sent four random bytes (uninitialised memory?) in the Content-Type for HEADs)

[03:12:00.0000] <Philip`>
http://www.louvre.fr/llv/commun/home.jsphas

[03:12:01.0000] <webben>
Awesome. That's because people who sit down to code web servers don't start by thinking "I should really read the spec for this HTTP nonsense."

[03:12:02.0000] <Philip`>
Content-Type: p#�/html;charset=UTF-8

[03:13:00.0000] <webben>
(that's always the scary thing about this stuff: not that authors don't read specs, but that so many implementers don't either)

[03:13:01.0000] <Philip`>
Oops

[03:13:02.0000] <Philip`>
s/has//

[03:13:03.0000] <Philip`>
(in the URL)

[03:14:00.0000] <annevk>
and that people writing specs make mistakes too, etc.

[03:15:00.0000] <webben>
annevk: If implementers read the specs, they could point out the mistakes and thump the people writing the specs to fix em ;)

[03:16:00.0000] <webben>
All I know is that when someone starts talking about coding a web server, I shouldn't be having to point them in the direction of the HTTP spec. :)

[03:16:01.0000] <annevk>
this happens, but given that so far no complicated specs that made REC required two complete interoperable impl people don't review enough

[03:17:00.0000] <annevk>
HTTP is a good example of not being good enough

[03:17:01.0000] <webben>
HTTP wasn't on the same standards pipeline.

[03:20:00.0000] <takkaria>
where's the schedule for fpwd?

[03:54:00.0000] <Hemebond>
1am. Night all.

[04:21:00.0000] <hsivonen>
Philip`: the *nix readme for Netscape said, IIRC, "And remember, it is spelled N-e-t-s-c-a-p-e but pronounced Mozilla." but the Mac and Windows masses didn't see that

[04:26:00.0000] <hsivonen>
Re: DTD discussion: http://hsivonen.iki.fi/no-dtd/

[04:36:00.0000] <webben>
hsivonen: Your FAQ similarly seems to assume that it would need a GET request not a HEAD request. HEAD could still cause DOS, but it might be less likely to.

[04:36:01.0000] <webben>
"actually retrieving the referenced DTDs"

[04:37:00.0000] <webben>
rather than "actually checking if the referenced DTDs have been modified"

[04:38:00.0000] <webben>
And also that the HTML spec does seem to recommend not connecting to the web to retrieve DTDs (at least, not every time): http://www.w3.org/TR/html4/sgml/intro.html#h-19.2

[04:38:01.0000] <webben>
(admittedly, your discussion is mainly about XML ... maybe the XML spec says something else)

[04:39:00.0000] <kig>
argh, implementing svg on the 2d canvas is rather challenging. no gradient/pattern transforms, no dashed strokes, no markers, lack of filters and blendmodes, no masks, no text. and no getTransform either

[04:39:01.0000] <webben>
kig: Why are you trying to implement SVG on canvas?

[04:39:02.0000] <hsivonen>
webben: if you implement your idea in the default Xerces resolver and manage to get it deployed so that the deployers don't get autobanned from w3.org, I would gladly revise my statements in face of new evidence

[04:40:00.0000] <webben>
hsivonen: Xerces doesn't use a local override?

[04:41:00.0000] <hsivonen>
webben: xerces is widely used and a piece of software where one might try out stuff like this more realistically than in browsers

[04:42:00.0000] <hsivonen>
webben: If I'm not mistaken, the defaults in Java XML parsers are *very* naive

[04:44:00.0000] <kig>
webben: to use svg images on a canvas

[04:44:01.0000] <webben>
kig: Oh I see.

[04:49:00.0000] <hsivonen>
would it good or bad if Validator.nu had a preset for XHTML 1.0 *Transitional* + Ruby + RDF + SVG 1.1 + MathML 2.0?

[04:49:01.0000] <hsivonen>
it seems to me that adhering to Strict XHTML content models and using embedded SVG or orthogonal issues in the real world

[04:50:00.0000] <webben>
hsivonen: So a bit like modular XHTML with the legacy module?

[04:51:00.0000] <hsivonen>
webben: yeah

[04:51:01.0000] <webben>
hsivonen: I should think they are more related issues in the real world than the ideal world, actually ;)

[04:53:00.0000] <webben>
hsivonen: I'm not really sure what the presets are for.

[04:54:00.0000] <webben>
What if you wanted to validate (say) XHTML 1.0 Transitional + Ruby + MathML but not RDF and SVG.

[04:54:01.0000] <webben>
maybe it would be better to have some tickboxes

[04:56:00.0000] <hsivonen>
webben: if you want to do that, the current approach is that you download the schema driver and put an edited copy on your on HTTP server

[04:56:01.0000] <hsivonen>
webben: probably too hard for most users

[04:57:00.0000] <hsivonen>
webben: but it seems to me that people want either a single language or full compound doc support

[04:58:00.0000] <hsivonen>
webben: although some people might actually want more restricted compound document combos, putting it all in the UI is problematic

[04:58:01.0000] <webben>
hsivonen: That's why I think tickboxes would be better.

[04:58:02.0000] <webben>
So I want Strict (tick), and Ruby (tick), or maybe just go ahead and tick All.

[04:59:00.0000] <hsivonen>
webben: positive checkboxes are problematic from the point of view of validator URI design :-(

[05:00:00.0000] <webben>
hsivonen: &components=xhtmlstrict+ruby+rdf ?

[05:00:01.0000] <hsivonen>
webben: in order to make the URIs simple in the common case, the checkboxes should exclude stuff

[05:00:02.0000] <hsivonen>
and that would be bad UI

[05:01:00.0000] <webben>
hmm

[05:02:00.0000] <webben>
hsivonen: what does the URI bit for presets look like atm?

[05:05:00.0000] <kig>
webben: this is how far i've gotten: http://glimr.rubyforge.org/cake/canvas.html#SVGParser

[05:05:01.0000] <hsivonen>
webben: horrible if you actually choose a preset but nice if you don't choose a preset and let it pick one from the root namespace

[05:06:00.0000] <webben>
kig: Can't you just overlay svg objects using the browser's svg renderer rather than trying to parse SVG?

[05:06:01.0000] <kig>
not if i want to draw them on the canvas.. 

[05:07:00.0000] <hsivonen>
webben: the basic idea of the generic facet is that the interface is uniform regardless of what schemas you use. some schemas just are faster than others because they are precompiled in RAM and some are faster than completely remote schemas because the source bytes are stored locally

[05:08:00.0000] <hsivonen>
webben: the basic idea of the HTML5 facet is that there are no tweakable settings

[05:08:01.0000] <webben>
hsivonen: You could have tickboxes and still not have any in the uri by default.

[05:08:02.0000] <webben>
a missing components key could just mean "no preset"

[05:10:00.0000] <hsivonen>
webben: yeah, I think I have to sit down and design a checkbox UI at some point

[05:10:01.0000] <hsivonen>
webben: I don't know how to do it such that

[05:10:02.0000] <hsivonen>
1) It's good UI

[05:10:03.0000] <hsivonen>
2) The URIs are nice

[05:10:04.0000] <hsivonen>
3) Users can bring their own schemas

[05:11:00.0000] <hsivonen>
4) I don't need to build an insane number of schemas

[05:11:01.0000] <webben>
hsivonen: Schemas can be declared in the submitted documents.

[05:11:02.0000] <hsivonen>
5) I don't to kill perf by generating schemas on the fly

[05:11:03.0000] <hsivonen>
6) The default do the right thing for most people

[05:11:04.0000] <webben>
hsivonen: Well, if you modularize it, you can create all the schemas from those combinations /once/.

[05:12:00.0000] <hsivonen>
webben: putting schema-related syntax in the document itself is wrong and putting validation service-specific syntax there is even worse

[05:13:00.0000] <webben>
hsivonen: I'm just talking about schema declarations (if declarations is the right term). Nothing specific to using your validator.

[05:13:01.0000] <hsivonen>
s/I don't to kill perf/I don't hove to kill perf/

[05:13:02.0000] <hsivonen>
s/hove/have/

[05:13:03.0000] <hsivonen>
typo++

[05:14:00.0000] <hsivonen>
webben: doctype and xsi:schemaLocation are bugs

[05:14:01.0000] <webben>
xsi:schemaLocation="http://www.w3.org/MarkUp/SCHEMA/xhtml11.xsd" (is that called a "declaration")

[05:14:02.0000] <hsivonen>
webben: that's Considered Harmful

[05:14:03.0000] <webben>
hsivonen: Well, yes, but here you're hitting a side-effect of that principle.

[05:15:00.0000] <hsivonen>
yes

[05:15:01.0000] <webben>
the alternative is for people to throw a schema uri into their validator query: but that's just as prone to single-point failure.

[05:15:02.0000] <hsivonen>
I'm assuming that most people are OK with validating against the compound Web language 

[05:16:00.0000] <hsivonen>
that is, the schema allowing SVG is not a real problem for people who want to check if they mistyped their XHTML or MathML

[05:17:00.0000] <webben>
No. The people who must want to do subsetting would go rather beyond what you could easily represent in a UI anyway.

[05:17:01.0000] <webben>
e.g. subsetting for comments on blogs

[05:18:00.0000] <webben>
(or indeed blog posts for that matter)

[05:18:01.0000] <webben>
where you might want to allow SVG but disallow SCRIPT for instance

[05:20:00.0000] <hsivonen>
webben: indeed. which is why the XHTML Modularization failed. and which is why a UI that doesn't involve user-written driver schemas can become too complex and still lack the options the user wants

[05:21:00.0000] <webben>
hsivonen: I think XHTML modularization's problems have more to do with it being utterly unclear what exactly it's for, the spec never having been finished, and schemas being a bit of a nightmare as a tool.

[05:22:00.0000] <webben>
(compared to the HTML DTD, navigating XHTML schemas is horrible)

[05:23:00.0000] <hsivonen>
webben: I think XHTML Modularization failed when their flagship use case, XHTML MP, chose a subset that doesn't follow the prescribed module boundaries

[05:23:01.0000] <webben>
Doesn't that suggest they got the module boundaries very wrong?

[05:24:00.0000] <hsivonen>
webben: yes. hence, fail

[05:24:01.0000] <webben>
ideally a module should be a set of related things that you would never not want together. if you have dependencies that you might or might not want, those should be in different modules.

[05:25:00.0000] <webben>
(i.e. modules should be about relatedness, the relationships between modules should be about dependencies)

[05:37:00.0000] <hsivonen>
Philip`: does X3D have a namespace URI?

[05:37:01.0000] <annevk>
seems it was not mentioned here yet: http://lists.w3.org/Archives/Public/www-archive/2007Dec/0072.html

[05:37:02.0000] <annevk>
kig, Opera 9.5 adds some SVG integration for <canvas>

[05:37:03.0000] <annevk>
kig, so hopefully in due course you won't need your own library anymore and you can just paint SVG graphics directly

[05:42:00.0000] <webben>
annevk: Does it preserve text alternatives for SVG graphics when it does so?

[05:43:00.0000] <webben>
or does it effectively unsemanticize the SVG much like using it as a CSS background image might?

[05:51:00.0000] <Philip`>
hsivonen: No, it isn't in a namespace

[05:51:01.0000] <Philip`>
(That has been complained about years ago, but the Web3D group works in secret, and I've got no idea if they're doing anything about it)

[05:52:00.0000] <hsivonen>
Philip`: ok. 

[05:52:01.0000] <hsivonen>
Philip`: Google shows that different people are using different bogus namespaces, so I was wondering if there's a correct one...

[05:53:00.0000] <Philip`>
hsivonen: http://www.xj3d.org/javadoc2/overview-summary.html under "XML Integration" is the one I've seen, so I copied that into my implementation

[05:53:01.0000] <Philip`>
(with a compatibility transform so you can load 'legacy' non-namespaced X3D documents)

[05:55:00.0000] <Philip`>
http://www.web3d.org/specifications/x3d-schema-changelog.txt has some comments mentioning namespaces and suggesting xmlns:x3d="http://www.web3d.org/specifications/x3d-3.0.xsd", which seems a really bad idea because it'll be different for every version

[05:56:00.0000] <hsivonen>
Philip`: should I add a UI string for http://www.web3d.org/specifications/x3d-namespace ?

[05:57:00.0000] <Philip`>
hsivonen: Doesn't seem worthwhile at the moment, since probably nobody uses it

[05:57:01.0000] <hsivonen>
Philip`: OK

[05:58:00.0000] <hsivonen>
(I have UI strings for various stuff that might occur on purpose or by mistake in document types roughly advertised in the presets)

[05:59:00.0000] <hsivonen>
(but I don't have UI strings for Atom extensions or RDF ontologies)

[06:00:00.0000] <Philip`>
I can only see three places on the web which mention http://www.web3d.org/specifications/x3d-namespace and one of those places is the #whatwg logs

[06:01:00.0000] <hsivonen>
ok

[06:03:00.0000] <Philip`>
/me will post something to the X3D mailing list, and will complain about the namespace and see if anyone has helpful suggestions

[06:03:01.0000] <Philip`>
(because I'm assuming the lack of a namespace is a bad thing)

[06:08:00.0000] <hsivonen>
well, if the namespace is lacking, it should be "" and people shouldn't use bogonamespaces

[06:09:00.0000] <hsivonen>
bogonamespaces are definitely a bad thing for interop

[06:09:01.0000] <Philip`>
Bogonamespaces?

[06:09:02.0000] <hsivonen>
Philip`: someone pulls a URI from their sleeve in order to have *some* namespace

[06:09:03.0000] <Philip`>
Ah

[06:10:00.0000] <Philip`>
But if everyone picks the same bogonamespace, that's good for interoperability, so that's alright :-)

[06:10:01.0000] <hsivonen>
yeah

[06:12:00.0000] <Philip`>
/me wonders if there are standard arguments against non-namespaced XML languages

[06:12:01.0000] <webben>
Philip`: Wouldn't that boil down to people using the same names for different things?

[06:13:00.0000] <webben>
Philip`: Also note the argument that in order to change functionality for the same names, you must use a new namespace (e.g. WHATWG critique of XHTML 2 WG's decision to reuse XHTML1's namespace for different functionality)

[06:13:01.0000] <webben>
in other words, used right, namespaces preserve interoperability and compatibility

[06:14:00.0000] <Dashiva>
Isn't the latter rather circular?

[06:14:01.0000] <Dashiva>
Namespaces must be separate to avoid collisions; collisions are being made, therefore we must split namespace

[06:14:02.0000] <webben>
Dashiva: No. Because it's ultimately grounded in giving clients a way to distinguish two names with similar but different functionality.

[06:14:03.0000] <hsivonen>
case studies RSS 2.0 (which you probably don't want to use in any argument about anything) and DocBook

[06:14:04.0000] <webben>
Dashiva: Namespaces must be separate to avoid collisions between names.

[06:15:00.0000] <webben>
obviously if you have collisions /within/ namespaces, that's bad

[06:15:01.0000] <Dashiva>
But that's what you're trying to prove, not your premise

[06:15:02.0000] <Philip`>
I suppose a problem is <html><X3D><script></X3D></html>, since the XHTML layer has to either ignore the entire subtree rooted at the unknown-to-XHTML <X3D> element, which is bad for extensibility of XHTML; or it has to ignore that element and process the child elements, in which case it'll incorrectly execute the script 

[06:15:03.0000] <webben>
Dashiva: Well, in a world without namespacing of any sort, functionality for names couldn't change without breaking backwards compat.

[06:15:04.0000] <Philip`>
(assuming that's meant to be an X3D script element, not an XHTML script element)

[06:16:00.0000] <Dashiva>
webben: That is true

[06:16:01.0000] <webben>
And a name, once chosen, would be universal throughout documents even if it's wildly inappropriate.

[06:16:02.0000] <Philip`>
(and assuming the non-XHTML languages always have a root element which is not in XHTML)

[06:16:03.0000] <webben>
over time, you'd have ever more complicated and non-intuitive names to embrace new functionality

[06:16:04.0000] <Dashiva>
So instead you make all names complicated and non-intuitive

[06:16:05.0000] <webben>
and you'd need a central registry which could actually manage all those names (even names used in private companies for private purposes)

[06:17:00.0000] <webben>
to ensure that when a public standard uses a new name it doesn't conflict with internal XML.

[06:18:00.0000] <webben>
I think namespacing is sound and used in most languages (e.g. CPAN). Whether the particular way XML does namespacing (the whole weird URIs, and prefixes, and null namespaces + schemas + doctypes craziness) makes any sense is another matter

[06:18:01.0000] <Dashiva>
Also true, but it's quite a significant matter

[06:18:02.0000] <webben>
Very significant. Yes.

[06:18:03.0000] <webben>
possibly grounds for revision to XML at some point

[06:19:00.0000] <Dashiva>
If we could do namespaces in an easy, transparent way I don't see many arguments against it at all

[06:23:00.0000] <hsivonen>
if anyone knows of legitimate annotation-xml content that is not MathML, SVG, XHTML or OpenMath, please let me know

[06:24:00.0000] <kig>
(re: svg on canvas, another bonus of it is that as long as you have javascript and implement the canvas api, you can use svg in your drawing (say, on the iPhone and on the OpenGL canvas, etc.))

[06:25:00.0000] <webben>
hsivonen: What is "legitimate annotation-xml"?

[06:27:00.0000] <hsivonen>
webben: stuff that is deployed in annotation-xml and is appropriate for sending over the wire in the Web context

[06:27:01.0000] <Philip`>
kig: About getTransform: why not do all the transform matrix calculations in JS, then just use setTransform (or decompose into scale/rotate/translate) once before rendering?

[06:28:00.0000] <kig>
i am doing that now, and i'd rather not

[06:28:01.0000] <Philip`>
(That's what I've done for X3D in canvas-3d, and it seems to work alright)

[06:28:02.0000] <kig>
(both)

[06:29:00.0000] <kig>
i mean, the underlying canvas implementation must have knowledge of the ctm to be able to draw anything, so duplicating the math in JS is a waste and a bother

[06:32:00.0000] <Philip`>
The underlying implementation might be an external library which the browser uses, and the library doesn't necessarily have to expose the CTM to the browser

[06:32:01.0000] <Philip`>
but at least Cairo has a cairo_get_matrix, and it'd be a silly library to not expose that kind of information, so that's probably not a practical problem

[06:34:00.0000] <Philip`>
/me guesses the next version of canvas will have dashed lines, since that seems the most demanded feature after text rendering

[06:34:01.0000] <kig>
are skew matrices invertible :?

[06:34:02.0000] <hsivonen>
but that CTM is the CTM of the back end relative to the world. it isn't the "CTM" relative of the canvas relative to page 

[06:35:00.0000] <Philip`>
kig: Yes, since they're not singular

[06:35:01.0000] <Philip`>
(Well, assuming it's not skewed infinitely into a straight line, because that'd be silly)

[06:37:00.0000] <kig>
i have a "decompose matrix into translate+rotate+scale" to do setTransform, transform and isPointInPath in JS, but i don't know what to do with skews

[06:38:00.0000] <Philip`>
Oh, do you mean decomposing skews into translate+rotate+scale?

[06:39:00.0000] <kig>
and my crappy matrix inverter uses the decomposition thing too :<

[06:41:00.0000] <kig>
Philip`: yeah

[06:41:01.0000] <Philip`>
kig: Seems easier to use something like http://www.j3d.org/matrix_faq/matrfaq_latest.html#Q23 for inverting

[06:42:00.0000] <kig>
nice, thanks!

[06:43:00.0000] <Philip`>
I'm not aware of a totally obvious way to decompose skews - it looks like they should be solvable as rotate+scale+rotate, but I don't know any details

[06:43:01.0000] <kig>
http://pastie.caboo.se/130487 <- what i have now

[06:49:00.0000] <Philip`>
Ah, okay, that seems quicker than the general 3x3 case

[06:49:01.0000] <Philip`>
Is it actually correct, though?

[06:50:00.0000] <Philip`>
(I get a quite different answer if I ask Maxima how to invert a 3x2 matrix)

[06:51:00.0000] <kig>
it works here, at least. first inverts scale, then inverts rotation, then inverts translation

[06:51:01.0000] <kig>
but might also be that my whole matrix math lib is broken, but in the right way

[06:52:00.0000] <Philip`>
It says invert(invert([1,2,3,4,5,6])) = 3.6764705882352944,2.1551724137931028,2.2058823529411766,0.8620689655172411,15.941176470588237,8.672413793103447

[06:54:00.0000] <Philip`>
function invert(m) { var d = 1/(m[0]*m[3]-m[1]*m[2]); return [ m[3]*d, -m[1]*d, -m[2]*d, m[0]*d, m[2]*m[5]-m[3]*m[4], m[1]*m[4]-m[0]*m[5] ];

[06:54:01.0000] <Philip`>
}

[06:54:02.0000] <Philip`>
seems to work better

[06:55:00.0000] <Philip`>
(Maybe that's just a problem with inverting skewed matrices)

[06:55:01.0000] <kig>
1,2,3,4,5,6 is skewed, yeah

[06:55:02.0000] <Philip`>
Argh

[06:56:00.0000] <Philip`>
return [ m[3]*d, -m[1]*d, -m[2]*d, m[0]*d, d*(m[2]*m[5]-m[3]*m[4]), d*(m[1]*m[4]-m[0]*m[5]) ];

[06:56:01.0000] <Philip`>
works betterer

[06:56:02.0000] <Philip`>
and is actually correct

[06:59:00.0000] <kig>
yeah, that works for skews too 8)

[06:59:01.0000] <Philip`>
Doesn't help with decomposing skews for canvases without setTransform, though...

[07:03:00.0000] <hsivonen>
error 503 mod_jk sucks

[07:04:00.0000] <kig>
Philip`: it needs to be var d = (m[0]*m[3]-m[1]*m[2]); if (Math.abs(d) < 0.005 /* or such */) return [1,0,0,1,0,0]; var id = 1 / d; ...

[07:05:00.0000] <kig>
otherwise zero division on e.g. [1,0,0,0,0,0]

[07:05:01.0000] <Philip`>
Zero division is non-fatal, so returning an incorrect result ([1,0,0,1,0,0]) instead of another incorrect result ([Inf,Inf,...]) doesn't seem that much of an improvement :-)

[07:20:00.0000] <Philip`>
Aha - apparently A = Q R S R^T for any A and for suitable orthonormal Q and R and diagonal S

[07:21:00.0000] <Philip`>
(so you can decompose anything into rotations and scales)

[07:24:00.0000] <Philip`>
(The trick is in finding Q and R, I suppose)

[07:29:00.0000] <Philip`>
http://www.me.rochester.edu/courses/ME444/MatrixPolarDecomp.pdf - that looks non-trivial

[07:31:00.0000] <kig>
i found this for doing a skew with rotation and scale: http://newsgroups.cryer.info/comp/graphics.algorithms/200602/07/0602072134.html

[07:33:00.0000] <Philip`>
Ah, that looks like a slightly easier approach

[07:34:00.0000] <kig>
and apparently graphics gems vol 2 ch 7 describes how to extract the skew from a matrix http://tog.acm.org/GraphicsGems/AllGems.TOC

[07:34:01.0000] <Philip`>
though you still need to do a bit of work if you want to decompose an arbitrary matrix, rather than just a [[1,a],[0,1]] one

[07:34:02.0000] <kig>
http://tog.acm.org/GraphicsGems/gemsii/unmatrix.c

[07:35:00.0000] <Philip`>
I assume you don't care about 4x4 matrices, though

[07:36:00.0000] <kig>
might be useful for unbreaking the opera-3d api..

[07:36:01.0000] <Philip`>
Ah, true

[07:38:00.0000] <Philip`>
Looks like much nicer code than http://www.soe.ucsc.edu/~pang/160/f98/Gems/GemsIV/polar_decomp/Decompose.c though I don't know if it's differently featured in a significant way

[07:41:00.0000] <Philip`>
Oh, that polar decomposition says it's designed to produce matrices which can be animated nicely, whereas the Graphics Gems II one tries "to synthesize an arbitrary matrix from a limited set of primitives, without regard for meaningfulness of the decomposition"

[07:41:01.0000] <Philip`>
(says http://www.cs.wisc.edu/graphics/Courses/838-s2002/Papers/polar-decomp.pdf)

[08:18:00.0000] <webben>
Is there an easy way to check that a relax ng schema is a valid schema?

[08:30:00.0000] <hsivonen>
webben: yes. put the schema URI in the Validator.nu schema field and put any document URI in the document field

[08:30:01.0000] <webben>
hsivonen: Cool, thanks.

[09:05:00.0000] <webben>
hsivonen: I donwloaded the validator locally and got it running. But when I point it at documents and schema at http://localhost:1024 it says "IO Error: Attempted to connect to localhost." ... yet if I take the same URIs and paste them into a firefox tab address bar, they resolve and work fine

[09:05:01.0000] <webben>
That doesn't happen to ring any bells about what might not be working, does it?

[09:05:02.0000] <hsivonen>
webben: It's a deliberate tin-foil hat. I guess it is a bad one

[09:06:00.0000] <webben>
tin-foil hat against what?

[09:06:01.0000] <Philip`>
It's a tin-foil hat with a large hole in the top of it

[09:07:00.0000] <Philip`>
(since you can e.g. use DNS to make the validator connect to 127.0.0.1)

[09:07:01.0000] <webben>
hsivonen: Is there an easy way to take off the tin-foil hat?

[09:08:00.0000] <hsivonen>
webben: if you run python build/build.py checkout now, it the hat will come off

[09:08:01.0000] <Philip`>
webben: util/src/nu/validator/xml/PrudentHttpEntityResolver.java

[09:08:02.0000] <Philip`>
and comment out the bit with "127.0.0.1".equals(host) etc

[09:08:03.0000] <hsivonen>
Philip`: now commented out in svn

[09:09:00.0000] <Philip`>
validator.w3.org seems a bit cleverer since it resolves the IP address and fails if it's not a public range

[09:10:00.0000] <webben>
hmm the hat seems not to have come off

[09:10:01.0000] <webben>
I'll try checkout a second time.

[09:12:00.0000] <hsivonen>
XMLFilterImpl is messing up my careful ErrorHandler setup :-(

[09:13:00.0000] <webben>
yeah, behavior is unchanged

[09:15:00.0000] <Philip`>
webben: I assume you built and ran the updated version too? :-)

[09:16:00.0000] <webben>
Philip`: checkout then run

[09:16:01.0000] <webben>
oh hang on

[09:16:02.0000] <webben>
did i need a build in between there?

[09:17:00.0000] <hsivonen>
webben: python build/build.py checkout build run

[09:18:00.0000] <webben>
ah :) sorry. Thanks.

[09:18:01.0000] <webben>
that seems to work

[09:42:00.0000] <hsivonen>
Do I read Dean Edrigdes message correctly that he is basically talking about using the spec to force IE to implement application/xhtml+xml?

[09:45:00.0000] <gsnedders>
hsivonen: that's how I read it too

[09:48:00.0000] <Philip`>
Don't forget Konqueror!

[09:58:00.0000] <webben>
Philip`: Konqueror has a sort-of XHTML mode already, I think.

[09:59:00.0000] <Philip`>
It uses its HTML parser for application/xhtml+xml in the currently released versions, and when I last looked they weren't actively working on fixing that

[10:02:00.0000] <webben>
Philip`: The impression I got from talking to the devs about a year ago was that it's XHTML-handling is a bit more complicated than just pretending its broken HTML.

[10:02:01.0000] <webben>
*its

[10:27:00.0000] <gsnedders>
ergh.

[10:27:01.0000] <gsnedders>
/me has to answer a question on why the people writing the software are not the best people to test it

[10:38:00.0000] <kig>
ouch, all browsers handle canvas image pattern transforms differently

[10:39:00.0000] <kig>
firefox's being the most useful permutation of the bunch

[10:42:00.0000] <kig>
opera conforms to the spec's "transformations screw everything up", firefox does "transformations affect patterns and path points created after the transformation", safari does what firefox does, but if there's a translation and rotation, it gives up and draws nothing.

[11:00:00.0000] <kig>
http://dark.fhtr.org/imageTest.html

[11:07:00.0000] <kig>
ah, safari only fails on no-repeat. go figure.

[11:25:00.0000] <Philip`>
webben: Ah, you're probably right - I think it did weird things when I last tried

[11:25:01.0000] <Philip`>
but with e.g. http://canvex.lazyilluminati.com/83/play.xhtml it definitely looks like it's using an HTML parser instead of an XML one, and messing up the doctype

[11:26:00.0000] <webben>
Philip`: Serving it a custom DTD is probably not playing fair ;)

[11:27:00.0000] <Philip`>
kig: http://philip.html5.org/tests/canvas/suite/tests/index.2d.pattern.html covers the areas where I noticed differences between browsers

[11:27:01.0000] <Philip`>
webben: Any XML parser should be able to handle that with no problem :-)

[11:27:02.0000] <webben>
Philip`: Should, no doubt. :)

[11:28:00.0000] <Philip`>
If it doesn't, it's not an XML parser, it's just a tag-soup-which-looks-a-bit-like-XML parser :-)

[11:28:01.0000] <webben>
Philip`: I did call it a "sorta-XHTML mode"

[11:29:00.0000] <Philip`>
I don't remember ever seeing XML parse errors from Konqueror either

[11:30:00.0000] <Philip`>
I would have assumed an XML parser would be the first thing someone would add to a sorta-XHTML browser

[11:30:01.0000] <Philip`>
(but I would have quite possibly assumed wrong)

[11:31:00.0000] <webben>
XML parsers? Who needs em? ;)

[11:34:00.0000] <Philip`>
It's lucky that Konqueror isn't quite as popular as IE, else it would be destroying XHTML

[11:34:01.0000] <Philip`>
(or at least destroying application/xhtml+xml)

[11:35:00.0000] <kig>
Philip`: btw, did you try this patch? https://bugzilla.mozilla.org/show_bug.cgi?id=401790

[11:35:01.0000] <Philip`>
kig: I haven't tried it

[11:37:00.0000] <Philip`>
/me should get around to working on the Mozilla canvas regression tests again, to add all the currently-failing tests (though keep them disabled) and remind the developers how many they still fail

[11:38:00.0000] <gsnedders>
Philip`: 42 plx kthxbai

[11:40:00.0000] <gsnedders>
on the subject of tests, I need to answer, "Explain why the people responsible for 'writing' the software are not the best people to test it."

[11:40:01.0000] <kig>
gsnedders: what do you think?

[11:41:00.0000] <gsnedders>
kig: I think anyone who can't test their own code because they think it's perfect needs to get out of idealism.

[11:43:00.0000] <kig>
and this applies to other fields as well?

[11:45:00.0000] <kig>
e.g. teachers are the best judges of the quality of their classes?

[11:46:00.0000] <gsnedders>
if they can escape from an idealistic view, they can be very good judges.

[11:54:00.0000] <Philip`>
gsnedders: The people writing the software have a negative incentive to find bugs, because a stupid manager will think it reflects badly on the programmer's abilities, and they will have to fix the bugs now instead of letting them accumulate for a couple of releases and then quitting the job and leaving the mess to someone else

[11:54:01.0000] <gsnedders>
/me wonders whether to copy that verbatim :P

[11:55:00.0000] <Philip`>
Also if your testing is trying to determine whether the code follows some specification, the programmer might misinterpret the specification, so you want a different person to be checking it

[11:57:00.0000] <Philip`>
Also, maybe a programmer's time is more expensive than QA monkey's time

[11:57:01.0000] <Philip`>
s/than/than a/

[11:57:02.0000] <gsnedders>
Philip`: oh, this is all waterfall model, so there has to been a specification :P

[11:57:03.0000] <Philip`>
There's plenty of reasons, of varying levels of dubiousness :-)

[11:58:00.0000] <gsnedders>
Philip`: I'm trying to think of any that aren't dubious. Not easy :P

[11:59:00.0000] <Philip`>
/me sees video discussion on the X3D mailing list, and a link to http://lists.xiph.org/pipermail/theora-dev/2003-February/000435.html which seems to be saying 'Theora is relatively safe because nobody cares enough to sue us'

[12:00:00.0000] <jwalden>
I think the common reason is "I can get more done and provide more benefit to users by spending time fixing bugs instead of writing tests"

[12:00:01.0000] <gsnedders>
I think all the booklet says is bias and pride in their own work.

[12:01:00.0000] <gsnedders>
Anyone who knows me knows that I'm always slagging off my code :P

[12:01:01.0000] <Philip`>
jwalden: Is that the actual reason, or just the stated reason which covers up the fact that testing is boring and programmers want to do fun things instead? :-)

[12:02:00.0000] <jwalden>
Philip`: I think it's the real reason, and I definitely know a person or two where it's probably accurate (at least one because when he writes the tests he spends so much time making them rigorous)

[12:03:00.0000] <jwalden>
and he usually writes them anyway, but I can see utility in having someone else doing at least some of the in-depth testing for him

[12:03:01.0000] <gsnedders>
How widely used is the waterfall model anyway (i.e., how irrelevant is what we're being taught)?

[12:06:00.0000] <krijn>
Is it okay to do Location: http://example.com/#fragment in HTTP?

[12:07:00.0000] <gsnedders>
krijn: Location must be an absolute URI, so yeah

[12:07:01.0000] <krijn>
Including the fragment identifier, I mean

[12:08:00.0000] <gsnedders>
AFAIK yeah

[12:08:01.0000] <krijn>
I thought that had issues, some years ago

[12:08:02.0000] <krijn>
But that could've been PHP screwing up

[12:09:00.0000] <hsivonen>
gsnedders: the person writing the code might not have come up with all the possible ways of abusing the code. It helps to have someone else figure out ways to put the code into unexpected states.

[12:09:01.0000] <gsnedders>
krijn: yeah, it is definitely allowed

[12:09:02.0000] <krijn>
gsnedders: Okay, thanks :)

[12:09:03.0000] <gsnedders>
Philip`: Location headers in the data set wold be nice :P

[12:09:04.0000] <gsnedders>
/me nudges Philip` 

[12:11:00.0000] <krijn>
/me slaps Mathias ;p

[12:14:00.0000] <Mathias>
Well I'm guessing it was an old version of PHP messing up with the # character.

[12:14:01.0000] <Mathias>
/me slaps Krijn right back

[12:15:00.0000] <gsnedders>
PHP--

[12:15:01.0000] <krijn>
Yeah, who uses PHP anyway? ;P

[12:16:00.0000] <gsnedders>
Me, as little as possible :P

[12:16:01.0000] <Mathias>
<?php echo 'My PHP is all I have.'; ?>

[12:19:00.0000] <csarven->
anyone know a good frontend dev person looking for full-time work?

[12:21:00.0000] <csarven->
(from Montreal)

[12:22:00.0000] <gsnedders>
I meet the good, person, looking, work parts, but not the rest.

[12:23:00.0000] <csarven->
trying to find someone for http://faq.css-standards.org/Integrator :)

[12:31:00.0000] <gsnedders>
/me wonders if it is possible to call pride "naïve"

[13:13:00.0000] <Lachy>
apparently IE8 will pass acid 2 http://blogs.msdn.com/ie/archive/2007/12/19/internet-explorer-8-and-acid2-a-milestone.aspx

[13:20:00.0000] <jruderman>
does that mean it has support for data: URLs?

[13:20:01.0000] <Lachy>
yes, if data URLs are tested in acid2. I can't recall.

[13:20:02.0000] <gsnedders>
jruderman: almost certainly yes

[13:20:03.0000] <gsnedders>
Lachy: yeah, they are

[13:20:04.0000] <jruderman>
any other features we can say the same about?

[13:20:05.0000] <gsnedders>
jruderman: hopefully we'll get a list of new features/bug fixes in IE8

[13:21:00.0000] <jruderman>
based on the acid 2 news, i mean

[13:21:01.0000] <gsnedders>
jruderman: display:table*

[13:21:02.0000] <gsnedders>
jruderman: position: fixed

[13:21:03.0000] <gsnedders>
jruderman: whatever bugs were blocking it

[13:21:04.0000] <Hixie>
we made a no-data-url version of the test for them, so not necessarily

[13:21:05.0000] <gsnedders>
Hixie: I'm aware, that's why I'm slightly hesitant about it :)

[13:21:06.0000] <gsnedders>
jruderman: (the list of things for acid2 that is)

[13:22:00.0000] <gsnedders>
Hixie: but there again, the filename list implies data URI support

[13:23:00.0000] <jruderman>
"Change Description: Reverse integration from green branch. Includes full implementation of ACID2" hmm

[13:23:01.0000] <jruderman>
oh, so that's what "dataprot" means

[13:27:00.0000] <jruderman>
it will be interesting to see what triggers "IE8 standards mode"

[13:27:01.0000] <gsnedders>
jruderman: see html-wg logs (linked from /topic)

[13:27:02.0000] <gsnedders>
jruderman: or join yourself

[13:28:00.0000] <jruderman>
?

[13:28:01.0000] <jwalden>
Lachy: re <http://www.alistapart.com/articles/previewofhtml5>, could you update it to add a link for Mozilla's experimental video builds?  a decent link target would be <http://www.double.co.nz/video_test/>

[13:28:02.0000] <gsnedders>
jruderman: http://krijnhoetmer.nl/irc-logs/html-wg/20071219

[13:28:03.0000] <Lachy>
jwalden, sorry, no can do. It can't be updated

[13:28:04.0000] <Lachy>
I wish someone had told me about them before

[13:29:00.0000] <jwalden>
Lachy: why not?  I wasn't aware ALA was anything beyond the web presentation where fixedness would matter

[13:30:00.0000] <jruderman>
an opt-in meta tag? eww

[13:34:00.0000] <hsivonen>
zcorpan: bad stuff. the source location bug on Musings is not reproducible when I copy the doc to another server

[13:34:01.0000] <hsivonen>
might be something bad around buffer boundaries

[13:34:02.0000] <jwalden>
now hope for sufficient evangelism that it can be made clear that the tag grants permission for IE to break your layout in the name of standards compat

[13:34:03.0000] <hsivonen>
hard to debug

[13:34:04.0000] <jwalden>
i.e. what doctype was supposed to be

[13:39:00.0000] <gsnedders>
Hixie: watching the video, it's the real acid2 test with data URI et all

[13:40:00.0000] <Hixie>
cool

[13:40:01.0000] <Hixie>
man, i agree with shawn's rant in http://www.w3.org/mid/994fc8d00712190734y3dea47derd9ba296d5232d541⊙mgc

[13:42:00.0000] <gsnedders>
they have worn printed copies of the CSS 2.1 :P

[14:59:00.0000] <Philip`>
/me wonders if the IE8 Acid2 thing makes Opera's antitrust complaint significantly less valid

[15:12:00.0000] <roc>
not really

[15:12:01.0000] <roc>
Acid2 is just a slice of a very large standards pie

[15:24:00.0000] <othermaciej>
Microsoft themselves used to downplay the importance of Acid2 quite a bit

[15:24:01.0000] <Hixie>
tehy still do

[15:25:00.0000] <Hixie>
even in that post they (correctly) point out it's just one test of many

[15:25:01.0000] <Hixie>
and not an especially good test from a standadrs-compliance perspective

[15:25:02.0000] <Hixie>
acid3 is mildly better in that regard, as constructed at the moment


2007-12-20
[18:45:00.0000] <Hixie>
hsivonen: yt?

[18:47:00.0000] <Hixie>
what should the conformance criteria for <iframe>'s contents be?

[18:47:01.0000] <Hixie>
especially given validation with live <script> elements, and the dynamic validity of <iframe> elements added to the dom in a browser on the fly

[18:55:00.0000] <webben>
It's interesting that IEBlog commentators often call for MS to implement HTML5, including stuff like Web Forms 2.0 and VIDEO. Just wondering whether people think that would be a good thing (because there would be more implementor feedback) or a bad thing (since effectively it would prevent the spec being changed).

[18:55:01.0000] <Hixie>
microsoft implementing html5 would be a good thing, especially if they actually gave us feedback

[19:23:00.0000] <Hixie>
/me prepares a 6000 line patch for the spec

[19:24:00.0000] <Hixie>
-1037/+851

[19:29:00.0000] <Hixie>
"block" and "inline" are a thing of the past

[19:41:00.0000] <takkaria>
fun

[19:43:00.0000] <Hixie>
i hope someone checks that patch for mistakes

[19:43:01.0000] <Hixie>
my eyes glaze over every time i look at it

[19:43:02.0000] <Hixie>
maybe i should go eat

[22:32:00.0000] <Hixie>
so, anyone read the changes? :-)

[22:32:01.0000] <Hemebond>
Nope.

[22:33:00.0000] <Hemebond>
What did you change?

[22:34:00.0000] <takkaria>
Hixie: I did

[22:34:01.0000] <takkaria>
seemed reasonable

[22:34:02.0000] <takkaria>
but I'm an armchair spectator wrt HTML5, I don't do much web work anymore...

[22:35:00.0000] <Hixie>
takkaria: hehe

[22:35:01.0000] <Hixie>
Hemebond: the content models

[22:35:02.0000] <hdh>
Hemebond: diff http://html5.org/tools/web-apps-tracker?from=1151&to=1152

[22:35:03.0000] <Hemebond>
""block" and "inline" are a thing of the past"

[22:35:04.0000] <Hemebond>
WTF

[22:36:00.0000] <takkaria>
nothing glaring popped out at me, but I think that perhaps some kind of diagram might be useful for us visual types

[22:39:00.0000] <Hixie>
yeah

[22:39:01.0000] <Hixie>
diagrams are hard, though, don't wanna do that until i'm sure it's ok :-)

[22:39:02.0000] <Hixie>
also i'm not exactly sure what to diagram

[22:40:00.0000] <takkaria>
/me nods

[22:42:00.0000] <takkaria>
on not knowing what to diagram, I have no suggestions there, either.  there's certainly no rush to create them, the prose is quite readable

[22:47:00.0000] <Hixie>
cool

[22:47:01.0000] <Hixie>
what i really want is hsivonen's opinion

[22:47:02.0000] <Hixie>
since he's implementing this stuff

[22:53:00.0000] <Hixie>
so looking at the acid2 video from microsoft we see some interesting things

[22:53:01.0000] <Hixie>
they implemented data: really early on in the process

[22:54:00.0000] <Hixie>
also for a long time they had the eyes rendering twice

[23:09:00.0000] <Hemebond>
IE8 passes the Acid2 test?

[23:10:00.0000] <Hemebond>
That's.... good right?

[23:14:00.0000] <Hixie>
yup

[23:15:00.0000] <Hemebond>
Does the Acid2 test use all the spec? Or just some bits.

[23:21:00.0000] <Hixie>
a tiny part

[23:22:00.0000] <Hemebond>
Thought so.

[00:02:00.0000] <hsivonen>
Hixie: I'm not familiar enough with <iframe> content usage patterns to say what should be conforming.

[00:05:00.0000] <hsivonen>
/me sees the comeback of significant inline in a light uncheckable form

[00:16:00.0000] <inimino>
"When a transparent or semi-transparent element has no parent" <-- when would that happen?

[00:18:00.0000] <hsivonen>
inimino: in DOM fragments 

[00:19:00.0000] <inimino>
ah, I see

[00:20:00.0000] <Hemebond>
But... would they not still have a parent when being displayed?

[00:20:01.0000] <hsivonen>
Hemebond: no, they get a parent when inserted into the document

[00:21:00.0000] <Hemebond>
Yeah...

[00:21:01.0000] <hsivonen>
Hixie: did you ping glazou about the content model changes?

[00:21:02.0000] <Hemebond>
How do you display a DOM fragment?

[00:21:03.0000] <hsivonen>
Hemebond: by inserting it into a document.

[00:22:00.0000] <Hemebond>
Exactly... so it will have a parent when being displayed.

[00:22:01.0000] <hsivonen>
the sentence inimino quoted has littly practical value and mostly makes readers start pondering what it is about

[00:23:00.0000] <Hemebond>
Curse you Hixie!

[00:24:00.0000] <Lachy>
the acid 2 test is broken. apparently a file as gone missing http://www.webstandards.org/files/acid2/test.html

[00:24:01.0000] <Lachy>
this copy still works though http://www.hixie.ch/tests/evil/acid/002/

[00:25:00.0000] <inimino>
I'm trying to understand the value of defining conformance for document fragments such that an element pulled from a conforming document could be a non-conforming fragment

[00:26:00.0000] <hsivonen>
Lachy: have you pinged Molly about it?

[00:29:00.0000] <hsivonen>
Hixie: regarding making body optional: We can't make it optional in DOMs parsed from text/html. If we do make it optional in DOMs parsed from application/xhtml+xml and in synthetic DOMs, there will be yet another special case to consider where a conforming DOM is unserializable as text/html

[00:30:00.0000] <hsivonen>
Hixie: moreover, given that the main problem with <body> is CSS behavior in the XML mode, perhaps it would be better for CSS to take mitigating actions in the spec like it did for HTML

[00:31:00.0000] <hsivonen>
whoa! Hixie made <title> optional!

[00:33:00.0000] <hsivonen>
I'm glad I took a local copy of the spec file before Hixie landed the content model changes...

[00:40:00.0000] <Lachy>
hsivonen, no

[00:41:00.0000] <hsivonen>
Hixie: it might be possible to write the permitted contexts for <style scoped> in a more obvious way

[00:42:00.0000] <hsivonen>
Hixie: "Where prose content is expected before any non-<style> sibling elements and before text nodes that are not inter-element whitespace

[00:42:01.0000] <hsivonen>
(the last part of the requirement is going to suck big time for RELAX NG)

[00:43:00.0000] <hsivonen>
(or, rather, RELAX NG sucks big time for mixed content where the positions of the text nodes are restricted)

[00:46:00.0000] <hsivonen>
Hixie: out of curiosity, do you have implementor feedback on the acceptability of <style scoped> except from Hyatt?

[00:49:00.0000] <hsivonen>
http://www.w3.org/mid/c0be00e90712181835j5f8cff31mf897fba71f307998⊙mgc

[00:56:00.0000] <Hixie>
hsivonen: ok. if you find out anything about iframes, do let me know, i basically have no idea how to define it right now.

[00:57:00.0000] <Hixie>
hsivonen: the comeback of "significant inline" actualy happened at the same time as it was removed (assuming you mean that you "should" not have empty elements)

[00:57:01.0000] <Hixie>
hsivonen: haven't spoken to glazou yet, wanted to get your take on it first

[00:57:02.0000] <Hixie>
hsivonen: agreed on <body>. I'll note that.

[00:58:00.0000] <Hixie>
hsivonen: I made <title> optional? oops

[00:58:01.0000] <Hixie>
hsivonen: re <style scoped>, no, not yet.

[01:00:00.0000] <Hixie>
hsivonen: ok, i have a fix for <title> pending

[01:00:01.0000] <hsivonen>
Hixie: ok

[01:01:00.0000] <hsivonen>
/me continues reviewing the changes

[01:01:01.0000] <Hixie>
and i've removed the note i have about body (originally a proposal from glazou)

[01:01:02.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/index-diff should be a diff of the big change, btw

[01:02:00.0000] <hsivonen>
Hixie: that diff has green-on-green styling for inserted text

[01:03:00.0000] <Hixie>
fixed

[01:03:01.0000] <hsivonen>
Hixie: thanks

[01:04:00.0000] <Hixie>
i recommend taking a copy of that page, it'll get nuked when i next regen the spec

[01:05:00.0000] <hsivonen>
Hixie: copy taken

[01:06:00.0000] <Hixie>
i also have a fix for <style scoped> pending

[01:07:00.0000] <hsivonen>
thanks

[01:07:01.0000] <Hixie>
i think that's everything you mentioned

[01:07:02.0000] <hsivonen>
yes

[01:08:00.0000] <hsivonen>
/me needs to learn to think that prose means %Flow

[01:08:01.0000] <Hixie>
yeah the new terms take some getting used to.

[01:08:02.0000] <Hixie>
i think the win from not clashing with css is good, though

[01:12:00.0000] <hsivonen>
hmm. now that bimorphic is gone, <del> becomes less complex unless there are new hidden complexities

[01:14:00.0000] <hsivonen>
does Heading content participate in any parent-child content models?

[01:14:01.0000] <hsivonen>
or only exclusions?

[01:16:00.0000] <hsivonen>
<address> has probably one of the worst usefulness/complexity ratios

[01:20:00.0000] <hdh>
/me recommends tracking whatwg with git-svn or bzr-svn or something like such

[01:21:00.0000] <hsivonen>
hdh: what do git-svn and bzr-svn do and what's the benefit over using svn directly?

[01:21:01.0000] <hdh>
can't get an overview like that index-diff though

[01:21:02.0000] <Hixie>
hsivonen: i think <del> is simple. but i'm not 100% sure.

[01:22:00.0000] <Hixie>
hsivonen: Heading is involved in the (increasingly badly explained and to be rewritten) outline algorithms

[01:22:01.0000] <Hixie>
hsivonen: <address> could be simplified by making it phrasing-only as in html4

[01:22:02.0000] <hdh>
it copies the whole svn history to my machine, so I can get the diffs offline

[01:23:00.0000] <hdh>
I checked only the recent revs, my internet access is rather costy

[01:23:01.0000] <hdh>
costly even

[01:24:00.0000] <hsivonen>
Hixie: The must not about <p><br></p> is going to be violated all the time

[01:25:00.0000] <Philip`>
svnsync can copy the whole SVN history too, without involving any other source-control system, assuming the server is new enough

[01:25:01.0000] <Philip`>
(It's a standard part of SVN 1.4, I think)

[01:26:00.0000] <Hixie>
hsivonen: sure. so is the requirement not to use <table> for layout, etc. those violations are real violations, though.

[01:26:01.0000] <Hixie>
(and always have been)

[01:26:02.0000] <hsivonen>
Hixie: evidently, these MUSTs are not held to the same standard as the Ogg SHOULD :-)

[01:27:00.0000] <Hixie>
hsivonen: sure, one's an author requirement and the other a UA requirement

[01:27:01.0000] <Hixie>
the bar for UA requirements is much higher since there are so many fewer of them

[01:27:02.0000] <Hixie>
uas, that is

[01:27:03.0000] <Hixie>
as opposed to authors

[01:28:00.0000] <hdh>
Philip`: didn't know that, thanks

[01:29:00.0000] <hsivonen>
Is there a spec-blessed word for the elements that are prose but not phrasing?

[01:30:00.0000] <Hixie>
no

[01:30:01.0000] <Hixie>
i ended up not really needing one

[01:30:02.0000] <Hixie>
if you need one i can add one

[01:30:03.0000] <Hixie>
you get the honour of deciding what it should be

[01:30:04.0000] <hsivonen>
I'm going to need a word, but then bimorphic wasn't in the spec, either

[01:31:00.0000] <Hixie>
(and exactly what it should cover)

[01:31:01.0000] <hsivonen>
Hixie: ok. I'll think about it when I start implementing this

[01:31:02.0000] <Hixie>
k

[01:31:03.0000] <hsivonen>
(not going to happen before Christmas)

[01:31:04.0000] <Hixie>
does the general idea look sound?

[01:31:05.0000] <Hixie>
(and is it an improvement?)

[01:32:00.0000] <hsivonen>
Hixie: it isn't an improvement considering a markup-aesthetic gut feeling, but it is a big improvement considering actual experience with applying the previous model to actual pages

[01:33:00.0000] <hsivonen>
so I guess the practical needs should override the aesthetic gut feeling

[01:33:01.0000] <Hixie>
cool, that's what i was aiming for

[01:33:02.0000] <Hixie>
i tried to address the aesthetic aspect with the new definition of paragraph

[01:34:00.0000] <hsivonen>
I congratulate myself for postponing improving bimorphic violation error messages

[01:35:00.0000] <Hixie>
hehe

[01:35:01.0000] <Hixie>
now i want to know what lachy's gonna think :-)

[01:35:02.0000] <Hixie>
i'm sure he'll hate it

[01:35:03.0000] <Hixie>
i guess i should post to a list

[01:36:00.0000] <hsivonen>
there he is

[01:36:01.0000] <Hixie>
wow, good timing

[01:36:02.0000] <Hixie>
Lachy: so the spec changed

[01:36:03.0000] <Hixie>
Lachy: i'm sure you're going to hate it

[01:37:00.0000] <Hixie>
Lachy: but i want to know your opinion anyway :-)

[01:37:01.0000] <hsivonen>
Hixie: did you consider allowing <span> to contain more stuff the the extent interoperably parsed by current browsers?

[01:37:02.0000] <hsivonen>
s/the the/to the/

[01:37:03.0000] <Hixie>
hsivonen: not while i was doing this change, no

[01:37:04.0000] <Hixie>
hsivonen: it's been proposed before though, and i'm sure i'll have to respond to that feedback at some point

[01:37:05.0000] <hsivonen>
Hixie: ok

[01:37:06.0000] <Hixie>
hsivonen: i'm not sure it's worth it, though, if it's just a migration thing

[01:38:00.0000] <hsivonen>
Migration is very important considering out design principles :-)

[01:38:01.0000] <hsivonen>
in fact, migration considerations are a big thing going for HTML5 compared to competing specs :-)

[01:39:00.0000] <hsivonen>
s/out/our/

[01:39:01.0000] <hsivonen>
typo++

[01:39:02.0000] <Hixie>
i guess

[01:39:03.0000] <Hixie>
i dunno

[01:39:04.0000] <Lachy>
/me looks

[01:39:05.0000] <Hixie>
it seems like a big hack to me

[01:39:06.0000] <Hixie>
why not use <div>?

[01:40:00.0000] <hsivonen>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cp%3Efoo%3Cspan%3E%3Col%3E%3Cli%3Eitem%3C%2Fli%3E%3Cli%3Eitem%3C%2Fli%3E%3C%2Fol%3E%3C%2Fspan%3Ebar%3C%2Fp%3E

[01:41:00.0000] <hsivonen>
that works in Firefox

[01:41:01.0000] <hsivonen>
this doesn't: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cp%3Efoo%3Cdiv%3E%3Col%3E%3Cli%3Eitem%3C%2Fli%3E%3Cli%3Eitem%3C%2Fli%3E%3C%2Fol%3E%3C%2Fdiv%3Ebar%3C%2Fp%3E

[01:42:00.0000] <Hixie>
oh, for putting lists inside paragraphs

[01:42:01.0000] <Hixie>
i thought you meant for <section>s and company

[01:42:02.0000] <Hixie>
i basically concluded we should give up with the whole <ol>s-in-paragraphs thing

[01:42:03.0000] <Hixie>
the benefits aren't compelling

[01:44:00.0000] <hsivonen>
Hixie: there's demand and allowing the span hack doesn't seem to have a serious downside

[01:44:01.0000] <Hixie>
is there really a demand?

[01:44:02.0000] <Hixie>
the pain is pretty big

[01:44:03.0000] <hsivonen>
Hixie: the pain?

[01:46:00.0000] <hsivonen>
Hixie: well, even if J. Random Author doesn't care, the issue has come up again and again among markup enthusiasts and experts

[01:46:01.0000] <Hixie>
e.g. worrying about ending up with nested <p>s

[01:46:02.0000] <Hixie>
so has <acronym>, our solution to that was to just chop it off altogether

[01:47:00.0000] <hsivonen>
Hixie: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cp%3Efoo%3Cspan%3E%3Col%3E%3Cli%3E%3Cp%3Eitem%3C%2Fp%3E%3C%2Fli%3E%3Cli%3E%3Cp%3Eitem%3C%2Fp%3E%3C%2Fli%3E%3C%2Fol%3E%3C%2Fspan%3Ebar%3C%2Fp%3E works for me in Firefox

[01:49:00.0000] <Hixie>
yeah

[01:49:01.0000] <Hixie>
that's a bad thing

[01:49:02.0000] <Hixie>
what does it mean to have a paragraph containing a paragraph?

[01:49:03.0000] <Hixie>
that's not typographically sane

[01:49:04.0000] <hsivonen>
well, I already have the schema bits around for allowing lists whose items are phrasing only

[01:52:00.0000] <Hixie>
yeah but then we're back to people being confused and... it just doesn't seem worth the effort

[01:52:01.0000] <Lachy>
Hixie, you're right about one thing: I hate it!

[01:52:02.0000] <Hixie>
hsivonen: i mean, what do we gain?

[01:52:03.0000] <Hixie>
Lachy: :-)

[01:53:00.0000] <Hixie>
Lachy: i tried to mitigate the damage with the new implied paragraph concept

[01:53:01.0000] <hsivonen>
Hixie: we fix a long-standing spec bug for people doing conversions from LaTeX and DocBook

[01:53:02.0000] <Philip`>
http://janus.state.me.us/legis/statutes/17-a/title17-Asec1105-A.html has subparagraphs in lists in paragraphs in lists in paragraphs

[01:54:00.0000] <Hixie>
hsivonen: we fix a longstanding bug for roundtripping from and back to LaTeX and DocBook, but... do we care?

[01:55:00.0000] <hsivonen>
I kinda care, yes

[01:55:01.0000] <Hixie>
Philip`: actually, i only see one paragraph there

[01:55:02.0000] <Hixie>
hsivonen: so what are the elements you would allow there?

[01:55:03.0000] <Hixie>
hsivonen: just <ol> and <ul>?

[01:55:04.0000] <hsivonen>
Hixie: yes

[01:55:05.0000] <hdh>
dl?

[01:56:00.0000] <Hixie>
hsivonen: and limit all decendants to phrasing-only?

[01:56:01.0000] <hsivonen>
Hixie: I'd be OK with phrasing-only <li>s in that case

[01:56:02.0000] <Hixie>
(and <li>)

[01:57:00.0000] <Lachy>
Hixie, for h1-h6, shouldn't the expected context say something like "Where prose content is expected, except where heading content is forbidden"

[01:58:00.0000] <hdh>
Any content model that expects prose content also expects phrasing content ← should the second expects be accepts?

[01:59:00.0000] <hsivonen>
all my elaborate <del> magic went to waste :-)

[02:01:00.0000] <Lachy>
it's not all bad I guess. Prose content seems to be rougly equivalent to %flow; in the HTML4 DTD

[02:01:01.0000] <hsivonen>
(I bet that some people who went Strict to be cool will feel betrayed.)

[02:02:00.0000] <Hixie>
Lachy: yeah basically prose = flow and phrasing = inline

[02:02:01.0000] <Hixie>
hsivonen: i know i do...

[02:02:02.0000] <Hixie>
Lachy: well, everything is allowed where it's allowed except where it's forbidden

[02:04:00.0000] <Hixie>
hsivonen: so... this is going to require parser changes too, i take it

[02:05:00.0000] <hsivonen>
Hixie: what's "this"?

[02:05:01.0000] <Hixie>
the whole span-ol thing

[02:05:02.0000] <hsivonen>
oh :-( the whole point was not requiring parsing changes

[02:06:00.0000] <Hixie>
seems the spec isn't what safari and mozilla do for <p><span><ol>

[02:07:00.0000] <Hixie>
in fact it's not what they do for any formatting element

[02:07:01.0000] <Hixie>
even other blocks handle being put inside an inline.

[02:07:02.0000] <Hixie>
huh.

[02:08:00.0000] <Hixie>
i wonder what i need to do to make the spec work

[02:08:01.0000] <Hixie>
the parser work rather

[02:09:00.0000] <hsivonen>
Does http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cp%3Efoo%3Cspan%3E%3Col%3E%3Cli%3E%3Cp%3Eitem%3C%2Fp%3E%3C%2Fli%3E%3Cli%3E%3Cp%3Eitem%3C%2Fp%3E%3C%2Fli%3E%3C%2Fol%3E%3C%2Fspan%3Ebar%3C%2Fp%3E work in IE like it works in Firefox and Safari?

[02:09:01.0000] <Hixie>
i don't have IE right here

[02:09:02.0000] <hsivonen>
in Opera, "bar" is misplaced compared to Firefox and Safari

[02:10:00.0000] <hsivonen>
IIRC, Gecko parsing was changed to allow blocks in <span>, because IE did and there were sites depending on it

[02:10:01.0000] <hsivonen>
If Firefox, Safari and IE agree here, I think the parsing spec should be changed to match, although I would have hoped that parsing was done and a closed issue

[02:11:00.0000] <jgraham_>
hsivonen: That was a bit optimistic, no?

[02:11:01.0000] <hsivonen>
jgraham_: yeah

[02:13:00.0000] <Hixie>
man, the parsing spec doesn't match browsers at all when it comes to handling of block-level start tags inside inlines

[02:13:01.0000] <Philip`>
hsivonen: In IE6, the OL is a child of both the SPAN and then of BODY; and "bar" comes after that (as a child of BODY)

[02:14:00.0000] <hsivonen>
Philip`: I don't understand what you mean

[02:14:01.0000] <hdh>
body > span > ol?

[02:16:00.0000] <Philip`>
data:text/plain,%23comment%3A%20CTYPE%20ht%0D%0AHTML%0D%0A%20%20%20%20HEAD%0D%0A%20%20%20%20%20%20%20%20TITLE%0D%0A%20%20%20%20BODY%0D%0A%20%20%20%20%20%20%20%20P%0D%0A%20%20%20%20%20%20%20%20%20%20%20%20%23text%3A%20foo%0D%0A%20%20%20%20%20%20%20%20%20%20%20%20SPAN%0D%0A%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20OL%0D%0A%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20LI%0D%0A%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20%20

[02:16:01.0000] <Philip`>
Uh, that's too long

[02:16:02.0000] <Hixie>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%3Cp%3E%3Ci%3E%3Cp%3E%3C%2Fp%3E%3C%2Fi%3E%3C%2Fp%3E is crazy

[02:17:00.0000] <Hixie>
luckily, safari agrees with the spec on that one

[02:17:01.0000] <Philip`>
http://tinyurl.com/2bslye

[02:17:02.0000] <Hixie>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%3Cp%3E%3Ci%3E%3Cdiv%3E%3Cp%3E%3C%2Fp%3E%3C%2Fdiv%3E%3C%2Fi%3E%3C%2Fp%3E is crazy even in safari

[02:18:00.0000] <Hixie>
i'm not sure exactly what to do here

[02:27:00.0000] <hsivonen>
Hixie: why is div's content model transparent instead of prose?

[02:28:00.0000] <hsivonen>
yay, <font> is phrasing

[02:29:00.0000] <hsivonen>
"XXX publish a "Valid HTML5!" button with a kitten on it. Made by an artist. (Doodle?)"

[02:29:01.0000] <hsivonen>
didn't zcorpan already take care of that action item?

[02:30:00.0000] <hsivonen>
Hixie: the big content model landing looks good

[02:30:01.0000] <Philip`>
I'm not sure he got the "Made by an artist" bit :-)

[02:30:02.0000] <Hixie>
<div> being transparent _should_ make no difference to being prose, but transparent better conveys the intent. i might change it back to prose though.

[02:30:03.0000] <Hixie>
<font> is phrasing but note that it is transparent too

[02:31:00.0000] <hsivonen>
Hixie: the main practical problem for me will involve banning text siblings before <style scoped>, <param> and <source>

[02:31:01.0000] <Hixie>
so it allows blocks inside it

[02:31:02.0000] <hsivonen>
Hixie: not a big deal in Java, though

[02:31:03.0000] <hsivonen>
Hixie: making <font> transparent sucks

[02:31:04.0000] <Hixie>
yeah well <font> in general sucks and is an open issue

[02:31:05.0000] <hsivonen>
Hixie: I'd prefer changing div's content model to prose

[02:32:00.0000] <Hixie>
the idea is to jsut allow <font> <div> <p> ... bla

[02:32:01.0000] <Hixie>
since everyone who's gonna use <font> is gonna do that anyway

[02:32:02.0000] <Hixie>
(and it seems no more harmful than allowing font in the firtst place)

[02:32:03.0000] <hsivonen>
Hixie: every time you say something out of the ordinary, everyone has to stop and think what the spec *really* means

[02:32:04.0000] <Hixie>
hsivonen: yeah, that was somewhat the intent. but i'll change it to prose.

[02:47:00.0000] <hsivonen>
:-( :-( Creative Commons deploys now qNames in content: http://wiki.creativecommons.org/CCPlus

[02:47:01.0000] <hsivonen>
s/now/new/

[02:56:00.0000] <Hemebond>
Hixie: That's actually something that kind of puts me off HTML5... legitimising crap.

[02:57:00.0000] <Hixie>
well, telling people they're wrong without really a good reason hasn't worked so far

[02:57:01.0000] <Hemebond>
What are you talking about? The web is finally starting to clean itself up thanks to the standards movement.

[02:58:00.0000] <Hemebond>
There are already specs that allow FONT and that rubbish.

[02:58:01.0000] <Hixie>
oh i agree that we shouldn't allow <font>

[02:58:02.0000] <Hemebond>
Let the (insert derogatory term) use that spec.

[02:59:00.0000] <Hixie>
(though if we don't allow it, it's not clear what to do with wysiwyg software, since the state of the art doesn't yet have a solution for semantic editor ui)

[02:59:01.0000] <Hemebond>
yet

[03:00:00.0000] <Hemebond>
I think there are WYSIWYG editors out there, web-based, that create fairly clean markup.

[03:00:01.0000] <Hemebond>
I guess I'm always going to prefer the direction of XHTML2.

[03:00:02.0000] <Hemebond>
Also...

[03:01:00.0000] <Hemebond>
I remember someone saying a day or so ago about how crap XUL was...

[03:01:01.0000] <Hemebond>
Was that you?

[03:01:02.0000] <Hemebond>
Or how it was a bad idea to use it for web apps.

[03:01:03.0000] <Hixie>
i don't recall talking about xul recently

[03:01:04.0000] <Hixie>
you can join the xhtml2 working group, btw, they're as open as the html5 working group

[03:02:00.0000] <Hixie>
and while there are wysiwyg editors that generate clean markup, i haven't seen any that generate correct non-presentational semantic markup that are usable by your average user

[03:02:01.0000] <akaroa>
Hixie: that's not actually the case (xhtml2)

[03:02:02.0000] <Hixie>
what isn't the case?

[03:03:00.0000] <akaroa>
I tried to join and got the : " no, members only" deal 

[03:04:00.0000] <Hemebond>
Hixie: Even if I could join XHTML2 WG, which is pointless cause I don't know shit, I'm really just trying to understand the reason for the revolt.

[03:04:01.0000] <Philip`>
akaroa: Do you mean joining just the mailing list, or joining the group as an invited expert?

[03:05:00.0000] <Philip`>
*public invited expert, or whatever they're called

[03:05:01.0000] <Hixie>
akaroa: recently?

[03:05:02.0000] <akaroa>
Then SP emailed back and said that I'd have to submit a paragraph saying how I'd be of use to the group and that this would be sent to the managment for consideration

[03:05:03.0000] <Hixie>
weird

[03:05:04.0000] <Hixie>
they have the same charter as the html working group as far as i can tell

[03:07:00.0000] <Hixie>
did you follow the instructions on this page?: http://www.w3.org/2004/01/pp-impl/32107/instructions

[03:07:01.0000] <Philip`>
As far as I can see, the charter doesn't specify the policy for accepting invited expert requests

[03:07:02.0000] <akaroa>
Hixie: Yes

[03:08:00.0000] <Hixie>
akaroa: did you submit a paragraph?

[03:08:01.0000] <akaroa>
no

[03:08:02.0000] <hsivonen>
does the XHTML2 WG have (public) Invited Experts at all or only W3C Invited Experts?

[03:08:03.0000] <Philip`>
and http://www.w3.org/2007/04/html-ie-faq#sameproc says it is the same process as normal, which requires approval from various people and isn't just automatic

[03:08:04.0000] <akaroa>
I gave up because I believe the future of xhtml is with xhtml5

[03:08:05.0000] <Hixie>
akaroa: at which step in those steps did they ask you for a paragraph?

[03:09:00.0000] <akaroa>
At first I got the no we don't have invited experts, then SP emailed and said that I'd need to write and ask etc...

[03:10:00.0000] <Hixie>
well that's weird

[03:10:01.0000] <hsivonen>
Hemebond: if a user selects text in a WYSIWYG editor and chooses a color for the text, what markup should be generated?

[03:10:02.0000] <Hixie>
i didn't realise they were screening members

[03:10:03.0000] <Hixie>
oh well

[03:10:04.0000] <Hixie>
yet another reason 5>2

[03:11:00.0000] <akaroa>
Hixie: yes

[03:11:01.0000] <Philip`>
I expect they look at the HTML WG and don't exactly want to copy that type of community

[03:20:00.0000] <Dashiva>
Did anyone else get double content in the latest commit-watchers mail?

[03:20:01.0000] <Hemebond>
hsivonen: Well, it wouldn't work the way other WYSIWYG editors.

[03:20:02.0000] <Hemebond>
You would need to change the way they construct documents.

[03:20:03.0000] <Hixie>
indeed. the problem is that nobody has yet shown how to do that.

[03:21:00.0000] <Philip`>
Dashiva: You mean other than the way it normally sends both 'index' and 'source' diffs?

[03:21:01.0000] <hsivonen>
Hemebond: do you have a plan for marketing the paradigm shift?

[03:21:02.0000] <Dashiva>
This time I got index+source, massive whitespace, index+source again, massive whitespace

[03:21:03.0000] <hsivonen>
back in 2002, I thought that problems could be solved if only editors had a different UI paradigm

[03:21:04.0000] <Hemebond>
Yes, I have a 5-point plan I prepared earlier.

[03:21:05.0000] <Hemebond>
/me pulls the proposal out of the oven.

[03:22:00.0000] <Philip`>
/me didn't read that 300KB commit message at all

[03:22:01.0000] <Hemebond>
*sniff* ahh, absolute rubbish.

[03:23:00.0000] <Philip`>
Dashiva: Hmm, I seem to have just one set of content in that message

[03:23:01.0000] <akaroa>
Hixie: 5>2, yes. You can't develop open standards for the open web in closed working groups. Open standards need to be developed in open groups

[03:23:02.0000] <Dashiva>
Philip`: Okay, probably a bug in my client then

[03:23:03.0000] <Hixie>
anyone know of any bugs in browsers in the DOM relative to what the DOM2 specs say?

[03:24:00.0000] <hsivonen>
Hixie: getAttribute return value when attribute not set. (but one could argue it is a spec bug now)

[03:24:01.0000] <hsivonen>
Hixie: namespace stuff in XHR-loaded DOMs

[03:24:02.0000] <Dashiva>
hrefs being resolved vs not is also a bit vague

[03:25:00.0000] <Hixie>
XHR isn't (yet) a spec, sadly

[03:26:00.0000] <Hixie>
these are things for acid3, i'm trying to limit myself to things i can prove are bugs using just specs published in 2004 or earlier.

[03:26:01.0000] <Hixie>
dom bugs, specifically

[03:26:02.0000] <hsivonen>
Hixie: actually, I think Acid3 should test that getAttribute violates the spec in the interoperable way

[03:27:00.0000] <Hixie>
acid3 can't test anything that the 2004-or-earlier specs don't explicitly require

[03:27:01.0000] <Philip`>
Opera supports getElementById('id\0garbage'), but that's not really a DOM problem because it affects the entire browser API but I like complaining about it anyway

[03:27:02.0000] <Hixie>
(though i can avoid testing things i know are wrong)

[03:27:03.0000] <Hixie>
Philip`: that counts

[03:28:00.0000] <Hixie>
you have test 38

[03:29:00.0000] <Hixie>
load http://www.hixie.ch/tests/evil/acid/003/NOT_READY_PLEASE_DO_NOT_USE.html in opera, wait for the animation to stop, and click the "Acid3" text

[03:29:01.0000] <Hixie>
see if the alert mentioned test 38 fails

[03:29:02.0000] <Philip`>
/me will cherish it

[03:30:00.0000] <Philip`>
Fails in 9.2 and 9.5 and works in FF2, which sounds right

[03:31:00.0000] <Philip`>
Maybe you should keep search engines away from http://www.hixie.ch/tests/evil/acid/003/reference.html so it isn't the top result for "acid3" and makes people think their browser scored 100%

[03:32:00.0000] <Hixie>
it'll fix itself when acid3 actually exists

[03:33:00.0000] <hsivonen>
Hixie: judging the smoothness of the animation is problematic

[03:34:00.0000] <Hixie>
hsivonen: really?

[03:34:01.0000] <Hemebond>
You could run it without a tic and time how long it takes to run.

[03:34:02.0000] <Hemebond>
Liek a timedemo.

[03:34:03.0000] <hsivonen>
Hixie: in Firefox 2 on Mac, I'm not sure if the animation counts as smooth

[03:35:00.0000] <Hixie>
hsivonen: then it probably doesn't

[03:39:00.0000] <hsivonen>
Hixie: standards-minded people might feel less betrayed if you allowed Content-Style-Type text/css as a conforming talisman

[03:39:01.0000] <Hixie>
with what UA requirements if it's not text/css?

[03:40:00.0000] <hsivonen>
Hixie: assuming that UAs today ignore it, requiring it to be ignored

[03:40:01.0000] <Hixie>
ugh

[03:42:00.0000] <hsivonen>
I'm not a fan of talismans, but Validome telling people to put it in and then Validator.nu telling people to take it out seems unproductive

[03:43:00.0000] <Hixie>
i see one possible fix...

[03:43:01.0000] <Hixie>
:-)

[03:45:00.0000] <hsivonen>
perhaps it is good to point out the bogosity of some talismans. hard call

[03:46:00.0000] <Hixie>
/me is watching microsoft's acid2 video

[03:46:01.0000] <Hixie>
man, microsoft sure are nice about this test now that they pass it

[03:46:02.0000] <Hixie>
they even said it was well-constructed :-D

[04:05:00.0000] <othermaciej>
heh

[04:06:00.0000] <othermaciej>
it's funny given how much they complained in the past that the test was silly, that competitions to pass tests like these are meaningless, and that browsers shouldn't compete on standards compliance

[04:19:00.0000] <Camaban>
well, they shouldn't compete on standards, but they will while some (mostly) support them and some make a real hash of them

[04:25:00.0000] <Hixie>
othermaciej: if you have any ideas for acid3, do let me know

[04:25:01.0000] <othermaciej>
Hixie: I need to look over what it currently tests at some point

[04:26:00.0000] <othermaciej>
maybe now is a good time, if I can't manage to get to sleep soon

[04:26:01.0000] <Hixie>
http://www.hixie.ch/tests/evil/acid/003/NOT_READY_PLEASE_DO_NOT_USE.html

[04:26:02.0000] <Hixie>
it's much more modular than acid2

[04:26:03.0000] <othermaciej>
my least favorite IE quirk is the direct JS property <--> DOM attribute mapping

[04:27:00.0000] <Hixie>
and i'm only testing things that i can pretty strongly argue are required in specs that were in CR or better in 2004

[04:27:01.0000] <othermaciej>
for elements

[04:27:02.0000] <othermaciej>
but I have to figure out what aspects of that are non-compliant

[04:27:03.0000] <othermaciej>
I'm pretty sure getAttribute() returning a non-string is non-compliant

[04:27:04.0000] <othermaciej>
getElementById finding things by name also seems like a clear violation

[04:31:00.0000] <Hixie>
what attribute returns a non-string?

[04:32:00.0000] <othermaciej>
any event listener attribute would return a function object if the attribute is set

[04:32:01.0000] <Hixie>
ooo ok

[04:33:00.0000] <hsivonen>
doesn't any getAttribute in IE return whatever the corresponding DOM property would return?

[04:33:01.0000] <othermaciej>
yes

[04:33:02.0000] <othermaciej>
a bunch of those happen to be strings

[04:33:03.0000] <hsivonen>
also, there's the getAttribute('className') thing

[04:33:04.0000] <othermaciej>
event handler attributes are a case where it won't even convert to string to look right

[04:34:00.0000] <othermaciej>
there was some blog post about IE DOM bugs recently that covered this

[04:34:01.0000] <othermaciej>
but I can't remember where I saw it

[04:35:00.0000] <hsivonen>
might have been sitepoint

[04:35:01.0000] <othermaciej>
ah: http://www.sitepoint.com/blogs/2007/11/30/internet-explorer-doesnt-just-suck-it-also-blows/

[04:36:00.0000] <Hixie>
ok cool, i'll look through those tomorrow

[04:36:01.0000] <Hixie>
thanks for the help

[04:36:02.0000] <Hixie>
send me further ideas at ian⊙hc

[04:37:00.0000] <Hixie>
going to bed now

[04:37:01.0000] <othermaciej>
roger that

[04:37:02.0000] <othermaciej>
good night

[05:45:00.0000] <Lachy>
wow, this video is a huge download. http://channel9.msdn.com/showpost.aspx?postid=367207

[05:46:00.0000] <Lachy>
for 584MB, it had better be a good quality video 

[05:49:00.0000] <Camaban>
for a 32min vid? it had better be good quality...

[05:57:00.0000] <Lachy>
oh no. <p> can no longer contain <ol> or <ul>, even for XHTML. :-(

[05:58:00.0000] <webben>
I wonder if we could have some sort of p-continuation element that would provide the same semantic in HTML and XHTML

[05:58:01.0000] <webben>
because that's what it actually is ... it's when a paragraph continues after a block

[06:06:00.0000] <webben>
it would also provide a nice styling hook : continuation {text-indent:0;margin-top:1ex;}

[06:14:00.0000] <othermaciej>
if a looser content model is really desirable it would make more sense to add a new element for paragraphs in addition to <p>

[06:14:01.0000] <othermaciej>
but it's probably not worth it

[06:15:00.0000] <Camaban>
webben: something like that would be cool for lists too :) quite how to do it in a 'correct' way I'm not sure on though

[07:02:00.0000] <webben>
othermaciej: A looser content model isn't the aim. Being able to express the media independent semantic of and simply style them is the aim.

[07:03:00.0000] <webben>
*of paragraph continuations

[07:03:01.0000] <webben>
one could do it with a <paragraph> or <para> element, but then you have two elements that mean the same thing

[07:03:02.0000] <webben>
if one were designing a lang from scratch, doubtless a different content model would be the way to go

[07:05:00.0000] <webben>
i suppose another possibility would be <p type="continuation">

[07:05:01.0000] <othermaciej>
<para>some text <ol>...</ol> more text</para> reads better to me than <p>some text</p><ol>...</ol><continuation>more text</continuation>

[07:05:02.0000] <othermaciej>
and certainly expresses the semantics better

[07:06:00.0000] <webben>
othermaciej: The semantic is the same in either case. It's a nicer way to do it, but having 2 elements do the same thing would probably be unpopular.

[07:06:01.0000] <webben>
(I wouldn't mind writing para ... not sure others would feel the same way)

[07:08:00.0000] <hdh>
the type=cont way would need reparenting elements somehow, so I'd go for allowing <ol> inside <p>

[07:09:00.0000] <Camaban>
I'd kind of prefer 'para', it maybe gives more a mental clue as to the purpose of the tag.  I'm not sure how many people at the moment really associate <p> with 'a paragraph of text', though I guess that could be a remnant of using whatever code makes it look right

[07:14:00.0000] <webben>
hdh: You /can't/ allow OL inside P in text/html since it breaks backwards compat.

[07:14:01.0000] <webben>
(browsers close P when they hit an OL)

[07:15:00.0000] <webben>
the main problem with type="continuation" is that it would turn into <p type="continuation" class="continuation"> for styling in IE6 (which doesn't support attribute selectors)

[07:16:00.0000] <webben>
which is a lot more typing than <continuation> 

[07:16:01.0000] <hdh>
webben: ah sorry, I remembered it to be the p > span > ol example

[07:28:00.0000] <gsnedders>
http://tom-frank.blogspot.com/ — what can I say?

[08:07:00.0000] <krijn>
Are the IRC logs working for anybody?

[08:12:00.0000] <gsnedders>
krijn: yeah

[08:12:01.0000] <krijn>
:/

[08:38:00.0000] <webben>
Hixie: Just a thought that it might be good if someone familiar with the issues wrote a WHATWG FAQ item about why VIDEO is a good idea as opposed to defining a video API for object.

[09:08:00.0000] <krijn>
Is IE3 still used today? :|

[09:10:00.0000] <Camaban>
webben: heh

[09:12:00.0000] <Camaban>
webben: I agree though, documenting the why behind some of this stuff would be good

[09:22:00.0000] <hdh>
what does 1H08 mean in the IE8 post?

[09:22:01.0000] <hdh>
oh, first half

[09:24:00.0000] <hdh>
that page use <font>...

[09:25:00.0000] <hdh>
generated by MS Office

[09:25:01.0000] <Philip`>
http://blogs.msdn.com/ie/archive/2007/12/19/internet-explorer-8-and-acid2-a-milestone.aspx ? I don't see any "<font" on there

[09:26:00.0000] <hdh>
<p class="MsoNormal" style="margin: 0in 0in 0pt;"><font face="Calibri">

[09:26:01.0000] <hdh>
/html/body/form/div[2]/div/div[3]/div[2]/div/div/div/p[3]/font

[09:26:02.0000] <hdh>
wtf is that form doing there?

[09:27:00.0000] <hdh>
argh, I mean http://myitforum.com/cs2/blogs/rtrent/archive/2007/12/19/ie8-news-on-standards-and-acid2-test-announcement.aspx

[09:28:00.0000] <hdh>
sorry MS

[09:30:00.0000] <Philip`>
"I watched the Channel 9 video and the UI of IE8 looks awfully similar to IE7" - hmm, I only noticed the UI looking totally different to IE7 though quite like IE3 - it just had a single toolbar with a few buttons, presumably since it was for debugging the engine and not for browsing the web

[09:30:01.0000] <hdh>
qt's demo browser even has a search box

[09:30:02.0000] <Philip`>
/me wonders what sites Chris Wilson was measuring when he said that half use 'standards' mode

[09:31:00.0000] <hdh>
defaulting to Qt doc searching to boot

[09:35:00.0000] <hdh>
1997.webhistory.org ← first time I see such name

[09:40:00.0000] <hdh>
why didn't they link to http://channel9.msdn.com/showpost.aspx?postid=367207 but channel9's / ?

[09:50:00.0000] <webben_>
hdh: webhistory is awesome ... they've got archives (searchable of yahoo-groups) from the earliest days of HTML.

[09:50:01.0000] <webben_>
(They're not complete, because there was a failure on the original mailserver.)

[09:50:02.0000] <hdh>
kinda wayback machines for news:// ?

[09:50:03.0000] <webben_>
hdh: Sorry ... no just for the mailing list for the original HTML specs

[09:51:00.0000] <hdh>
sorry, I missed the last phrase xD

[09:51:01.0000] <webben_>
hdh: http://1997.webhistory.org/lists/lists.html

[09:53:00.0000] <hdh>
If the material can be digitally uploaded, you will be able to submit it to the archives via the Internet once this capability is in place. Until then, please contact the Webmaster at webmaster⊙wo ← that IE8 video qualifies here?

[09:54:00.0000] <webben_>
hdh: Arguably... however, I'm not sure that the project is active.

[09:54:01.0000] <webben_>
It doesn't seem to have changed since 2003 or something.

[09:54:02.0000] <webben_>
You could always try mailing em and see what happens though.

[09:54:03.0000] <hdh>
well, at least the hosting is still up

[10:01:00.0000] <hdh>
/me googled 'make dl look like table' and found nothing, has anyone here did that?

[10:05:00.0000] <hdh>
okay, adding css to the query helps; ftr, someone did http://www.macrankin.co.uk/web_projects/tables/table_forge_3.html via http://www.killersites.com/mvnforum/mvnforum/viewthread?thread=4416

[10:33:00.0000] <zcorpan>
Hixie: shouldn't the content model for <del> be just "transparent"?

[10:35:00.0000] <zcorpan>
Hixie: <figure> still talks about embedded content

[10:39:00.0000] <zcorpan>
Hixie: it seems to me that <noscript> has similar content model requirements as <iframe> would have

[10:42:00.0000] <zcorpan>
Hixie: furthermore, iirc, opera has a setting to disable frames, where <noframes> and <iframe> is parsed as pcdata (like <noscript> when scripting is disabled)

[10:42:01.0000] <zcorpan>
Hixie: same with plugins and <noembed>

[10:43:00.0000] <zcorpan>
(though, having three toggles in the html parser sucks)

[10:55:00.0000] <zcorpan>
Hixie: why does one need to log in to view test cases and demos that the spec annotation system links to?

[11:00:00.0000] <zcorpan>
Hixie: while we're on content models, i think <table> should allow <col> as children

[11:01:00.0000] <zcorpan>
in fact, i'm not sure the parser shouldn't imply <colgroup> around <col>s (only ie does that)

[11:06:00.0000] <zcorpan>
Hixie: why allow prose content in <th>?

[11:06:01.0000] <zcorpan>
(and not in <h1>, <dt>, etc)

[12:37:00.0000] <zcorpan>
Hixie: nested <menu>s are now banned

[12:38:00.0000] <Hixie>
i have a fix pending for <del> transparent; figure's mention of embedding content will be gone

[12:39:00.0000] <Hixie>
noscript and iframe are similar and i nearly used the exact same text but it doesn't work because iframe could have <script> elements inside it, which could detect the validator testing the page for correctness

[12:39:01.0000] <Hixie>
the log in problem is an oversight

[12:39:02.0000] <Hixie>
i don't see why <table> should allow <col>

[12:39:03.0000] <zcorpan>
<table><col> is allowed in html4 and xhtml1

[12:39:04.0000] <Hixie>
i wasn't going to allow prose in <th> but i'm assuming we need to for back compat, i'm willing to try a change if we want

[12:40:00.0000] <Hixie>
and i'll look into <menu> nesting

[12:41:00.0000] <zcorpan>
Hixie: i think the common case for prose in <th> is <th><p> or <th><div>, which on the one hand seem pretty harmless but otoh pretty bogus

[12:41:01.0000] <Hixie>
yeah

[12:41:02.0000] <Hixie>
not sure what to do about that

[12:41:03.0000] <zcorpan>
i've also seen <th><h3> though, from authors who want to be overly explicit

[12:41:04.0000] <zcorpan>
which will probably just screw up things

[12:42:00.0000] <Hixie>
i can allow heading content too

[12:43:00.0000] <Hixie>
i could say phrasing content or heading content but not both

[12:44:00.0000] <Hixie>
i dunno

[12:44:01.0000] <Hixie>
hsivonen: any opinion?

[12:44:02.0000] <zcorpan>
i'm not sure why heading in <th> would be useful

[12:44:03.0000] <Hixie>
yeah

[12:44:04.0000] <zcorpan>
wouldn't it screw up the outline?

[12:44:05.0000] <Hixie>
i've changed it to phrasing content

[12:44:06.0000] <Hixie>
we'll see if we get complaints

[12:44:07.0000] <zcorpan>
ok

[12:45:00.0000] <Hixie>
<table><col> i recall had some issue with parsing which is why i didn't allow it

[12:45:01.0000] <zcorpan>
likely :)

[12:45:02.0000] <Hixie>
i think

[12:45:03.0000] <zcorpan>
ie implies <colgroup>

[12:45:04.0000] <zcorpan>
but it also implies <tbody>, which is not required :)

[12:47:00.0000] <Hixie>
fixed requiring login for showing tests and demos in a separate tab

[12:48:00.0000] <Hixie>
yeah

[12:51:00.0000] <gsnedders>
does http://pastebin.ca/825838 seem overly harsh?

[12:52:00.0000] <Philip`>
I hope that's not intended for posting to public-html :-p

[12:52:01.0000] <gsnedders>
No

[12:53:00.0000] <gsnedders>
:P

[12:53:01.0000] <hober>
gsnedders: looks pretty good to me. -public-html +www-archive that baby

[12:53:02.0000] <gsnedders>
I'm not that dumb

[12:53:03.0000] <gsnedders>
hober: yeah, and cc'd to chairs, I was planning

[12:53:04.0000] <gsnedders>
s/dumb/dumb normally/

[12:53:05.0000] <gsnedders>
:)

[12:54:00.0000] <Philip`>
Why CC to chairs?

[12:55:00.0000] <gsnedders>
Philip`: issue with conduct on mailing lists

[12:57:00.0000] <gsnedders>
http://lists.w3.org/Archives/Public/www-archive/2007Dec/0087.html

[12:59:00.0000] <kingryan>
gsnedders: thanks for that email. I had given up on that thread being productive.

[13:00:00.0000] <gsnedders>
kingryan: Oh, I scrapped various drafts of emails like that to various people over the months.

[13:08:00.0000] <Philip`>
/me doesn't like it when the spec says "you must only do x where y", because he doesn't find it immediately clear whether that's "you must do x, and must not do it without y" (i.e. "you must do x and y") or "you may do x, and must not do it without y" (i.e. "you must not do x unless y")

[13:09:00.0000] <Philip`>
(e.g. "Authors must only use elements in the HTML namespace in the contexts where they are allowed, as defined for each element")

[13:12:00.0000] <inimino>
that sentence doesn't seem to match the "only x where y" form

[13:17:00.0000] <Philip`>
It's not a syntactic match, but there's the same ambiguity of "authors must not use elements in the HTML namespace except in the contexts where they are allowed" vs "authors must not use elements which aren't in the HTML namespace, in contexts where HTML elements are allowed"

[13:17:01.0000] <Philip`>
Actually, maybe that's still not the same

[13:18:00.0000] <Philip`>
But it's still the "z must only w" form, so it's not too far off :-)

[13:20:00.0000] <Hixie>
Philip`: how would you phrase it?

[13:21:00.0000] <Philip`>
Hixie: As "Authors must not use elements in the HTML namespace except in the contexts where they are allowed" or "Authors must not use elements which aren't in the HTML namespace, in contexts where HTML elements are allowed"

[13:21:01.0000] <Hixie>
that disallows wrong use, but doesn't allow correct use

[13:24:00.0000] <Philip`>
The spec doesn't explicitly allow me to wear socks, but that doesn't make me a non-conforming HTML5 UA if I do - the default state for unspecified things is that they are allowed (I think)

[13:25:00.0000] <Hixie>
i guess

[13:25:01.0000] <Hixie>
can you send mail with those suggestions to the whatwg list?

[13:25:02.0000] <Hixie>
i'll fix it at some point

[13:25:03.0000] <Hixie>
(i want to think about exactly how best to phrase it)

[13:26:00.0000] <Philip`>
"must only" seems an abuse of the RFC2119 definition of "must", since it relies on parsing English to understand what are the requirements and allowances in the sentence

[13:27:00.0000] <Philip`>
/me will mail

[13:28:00.0000] <Philip`>
Incidentally, s/preceeded/preceded/ in the concept of "preceeded or followed"

[13:28:01.0000] <Hixie>
yeah i agree that the must only sentence is poor

[13:29:00.0000] <Hixie>
i guess it's time for me to regen the spec with all the fixes

[13:35:00.0000] <Hixie>
right, spec regenned and checked in

[13:38:00.0000] <Lachy>
woot! hasLayout is gone from IE8! :-) http://lists.w3.org/Archives/Public/www-style/2007Dec/0151.html

[13:39:00.0000] <webben>
lots of good news coming out of the IE team ... finally

[13:40:00.0000] <gsnedders>
/me blames middle-management of MS

[13:47:00.0000] <Philip`>
Is http://lists.whatwg.org dead?

[14:34:00.0000] <Lachy>
Philip`, yes, it appears to be

[14:46:00.0000] <Hixie>
Philip`: back up

[15:29:00.0000] <jwalden>
Hixie: outlook for a fixed version of acid2 (regarding fonts and their effects on how the chin displays), unofficially most likely, is very low, correct?

[15:31:00.0000] <Hixie>
there is a zero chance of the acid2 test changing to handle fonts with metrics much different from Arial and Times New Roman

[15:31:01.0000] <Hixie>
but i'll try to make sure to avoid that mistake in acid3

[15:33:00.0000] <jwalden>
Hixie: even unofficially, to be clear?  I'm creating a reftest for it and dbaron's last nagging question is about the fonts issue

[15:33:01.0000] <Hixie>
it would take me days to work out what was going on and fix the test

[15:33:02.0000] <jwalden>
about what I expected

[15:34:00.0000] <jwalden>
good to know, thanks :-)

[15:34:01.0000] <Hixie>
i'm happy to host an unofficial fixed version if someone makes one

[15:34:02.0000] <jwalden>
heh

[15:34:03.0000] <Hixie>
but frankly it's easier just to add Times Now Roman and Arial to the list of requirements for running the test

[15:34:04.0000] <Hixie>
in practice those fonts are available royalty free

[15:34:05.0000] <Hixie>
so it's not a huge blocker

[15:35:00.0000] <Hixie>
not ideal, though, i agree


2007-12-21
[16:14:00.0000] <jruderman>
Hixie: is the requirement to *have* one of those fonts, or to have one of those fonts set as the default?

[16:18:00.0000] <Lachy>
Yay! Spec to by published Jan 22

[16:18:01.0000] <Lachy>
(see mail from MikeSmith on public-html)

[16:18:02.0000] <othermaciej>
that's cool

[16:22:00.0000] <Lachy>
Selectors API should be published as LC tomorrow too

[16:53:00.0000] <Hixie>
jruderman: have them set as your default serif and sans-serif, i think

[16:54:00.0000] <jruderman>
Hixie: what do you think of making the test ask for those fonts, so if you have something else set as your default, you still see a pass?

[16:55:00.0000] <Hixie>
sounds fine to me

[16:55:01.0000] <jruderman>
jwalden: ^^

[16:56:00.0000] <jwalden>
too late, just committed -- followup commit, I guess!

[16:56:01.0000] <jwalden>
although...

[16:57:00.0000] <jwalden>
/me wonders whether tinderboxen will all have the specific fonts by name

[16:57:01.0000] <jwalden>
I'm going to Pretend This Works until we find out it doesn't

[16:58:00.0000] <jruderman>
ok

[17:49:00.0000] <Hixie>
Lachy: there's still gonna be a vote...

[17:50:00.0000] <Lachy>
yes,  I realise that

[17:50:01.0000] <Lachy>
I just forgot to mention it in my blog entry

[18:01:00.0000] <Hixie>
/me is doing a study of Server: headers and wonders how to parse "Server: <a href=http://forpsi.com>FORPSI</a> Apache"

[18:09:00.0000] <hdh>
/me read Opera's first request the wrong way; it means IE on other platforms, not another Windows N

[18:10:00.0000] <hdh>
argh, wrong again, gotta catch some sleep, sorry

[18:11:00.0000] <jwalden>
pretend it doesn't exist -- that doesn't match the RFC 2616 productions ;-)

[18:11:01.0000] <jwalden>
(that's tongue-in-cheek if it wasn't clear)

[18:18:00.0000] <Philip`>
/me saw http://www.vsemi-deseti.cz/ with the <a href=http://forpsi.com> Server header

[18:20:00.0000] <Philip`>
(http://http-parsing.gsnedders.com/ parses it into first token "<a")

[18:21:00.0000] <Philip`>
gsnedders: You should probably sort by string when the frequencies are identical, to make the long list of 1s easier to scan through

[18:22:00.0000] <Hixie>
(yeah i parsed it as a <a token)

[19:33:00.0000] <Hixie>
i love the accusation in http://www.w3.org/mid/476B3222.3060309⊙5cn that we "hijacked the group"

[19:33:01.0000] <Hixie>
given the history of html5

[19:34:00.0000] <takkaria>
/me gets a 404 for that url

[19:34:01.0000] <Hixie>
it'll work in a few minutes

[19:35:00.0000] <Hixie>
"Microsoft-IIS/5.0,Apache-Coyote/1.1" seems like an odd Server header to send

[19:37:00.0000] <takkaria>
I wonder what possesses people to write strings like that

[19:37:01.0000] <takkaria>
surely it's not for the same reason that UA strings are so screwy?

[19:37:02.0000] <roc>
it probably is

[19:37:03.0000] <Hixie>
there are a lot of "Microsoft-IIS/5.0,Apache-Coyote/1.1" headers in my data

[19:37:04.0000] <roc>
to work around baroque sniffing techniques

[19:38:00.0000] <Hixie>
ah that seems to be ebay.

[19:39:00.0000] <takkaria>
roc: server-sniffing seems like an odd kind of thing to do, though...

[19:40:00.0000] <roc>
maybe

[19:40:01.0000] <takkaria>
mind, I do seem to remember reading about some weird secret pact IE and IIS use to speed up connections

[19:40:02.0000] <Lachy>
Hixie, that link doesn't work yet http://lists.w3.org/Archives/Public/www-archive/2007Dec/0090.html

[19:40:03.0000] <roc>
proxies do all kinds of stupid things

[19:40:04.0000] <roc>
but who knows

[19:40:05.0000] <Hixie>
that particular string seems to just be what ebay sends

[19:43:00.0000] <takkaria>
Hixie: out of interest, why the survey of Server strings?

[19:43:01.0000] <Hixie>
takkaria: http://lists.w3.org/Archives/Public/public-appformats/2007Dec/0060.html

[19:45:00.0000] <takkaria>
oh, ta

[19:45:01.0000] <takkaria>
/me should subscribe to more of these w3 lists so he stops missing things like that

[19:45:02.0000] <roc>
but then you'd have to read them

[19:45:03.0000] <Hixie>
yeah i'm not sure you're missing out

[19:46:00.0000] <takkaria>
I've got pretty good at skimreading, it probably won't add much to my daily skim

[19:47:00.0000] <Hixie>
http://xn--74h.damowmow.com/ lists the mailing lists i read

[19:47:01.0000] <Hixie>
(second column)

[19:47:02.0000] <Hixie>
those that are faded out are the ones i am subscribed to

[19:48:00.0000] <takkaria>
impressive

[19:49:00.0000] <Hixie>
(there's probably many more that i don't care about that i'm subscribed to too)

[19:49:01.0000] <Hixie>
(and some i've just forgotten from the list, like appformats)

[19:49:02.0000] <Hixie>
Lachy: yeah, that's the one. i think someone should explain to him exactly what the history of html5 is :-)

[19:52:00.0000] <takkaria>
I don't mind reading through technical bits and reasonable discussion, it keeps the Web bit of my head active, and it's easy to skip through trolling and flamebait

[20:59:00.0000] <Hixie>
hopefully appformats will appreciate my desire to use actual data

[23:48:00.0000] <hsivonen>
Hixie: great post on browser and server adoption rates

[23:56:00.0000] <hsivonen>
"RDF is meant for babies and grandmothers alike; and its text formats read like a romance novel, only less racey."

[23:56:01.0000] <hsivonen>
(from www-archive)

[00:07:00.0000] <hsivonen>
Philip`: I believe the Chris Wilson's standards mode stats on the video were the same that he posted to public-html about IIRC 200 top sites 

[00:11:00.0000] <hsivonen>
Hixie: phrasing content or heading kinda makes sense and is expressible in RELAX NG. The RELAX NG error messages would suck, though, in the case where there is a heading followed by something else. And that case is probably just rare enough not to justify special-casing.

[00:11:01.0000] <hsivonen>
(special-casing in the validator UI layer that is)

[00:12:00.0000] <hsivonen>
Hixie: semantically, for non-layout tables, phrasing-only <th>s should be fine

[00:16:00.0000] <hsivonen>
takkaria: you need to do server sniffing if you support HTTP pipelining

[02:01:00.0000] <zcorpan>
Hixie: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cstyle%3E%20p%20%7B%20border%3Asolid%20%7D%20%3C%2Fstyle%3E%0D%0A%3Cp%3EFoo%20%3Cspan%3E%3Col%3E%3Cli%3Ebar%3C%2Fli%3E%3C%2Fol%3E%3C%2Fspan%3E%20baz%3C%2Fp%3E seems to not work in ie and opera...

[02:04:00.0000] <zcorpan>
Hixie: while http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cstyle%3E%20a%20%7B%20display%3Ablock%3B%20border%3Asolid%20%7D%20%3C%2Fstyle%3E%0D%0A%3Ca%20href%3D%3E%3Ch1%3Efoo%3C%2Fh1%3E%3Cp%3Ebar%3C%2Fp%3E%3C%2Fa%3E works in ie, opera, safari and (most of the time) firefox... blocks in <a> has, afaict, higher demand from authors than lists in <p>

[02:39:00.0000] <webben>
hsivonen: What if you want buttons for sorting tables by table headers?

[02:39:01.0000] <webben>
(or hiding/showing columns)

[02:39:02.0000] <hsivonen>
webben: buttons are phrasing, aren't they?

[02:40:00.0000] <webben>
hsivonen: What if you want a whole form (e.g. an input for filtering plus a button to submit)?

[02:42:00.0000] <webben>
actually what would be really nice is to able to declare part of th the heading proper

[02:42:01.0000] <webben>
and be able to reserve part of it for UI additions

[02:42:02.0000] <hsivonen>
webben: wouldn't it be more likely that the author would want every column heading button to be part of one form (either wrapping the table or using the form='' attribute)?

[02:43:00.0000] <webben>
hsivonen: Not necessarily, since you might have other UI on the table that does other things (and might say use POST rather than GET)

[02:44:00.0000] <webben>
the hard part is working out how to get it so that only the phrasing part of th is treated as the heading

[02:44:01.0000] <webben>
but I think that's  a problem that will need to be solved regardless of what it's formal content model is

[02:45:00.0000] <webben>
I guess abbr actually does that

[02:45:01.0000] <webben>
<th abbr="Address">Address<form><input><button></button></form></th>

[02:51:00.0000] <hsivonen>
webben: is abbr='' supported by UAs? abbr='' has the same bidi issues as title=''

[02:55:00.0000] <webben>
hsivonen: Not sure: doesn't seem to have been tested: http://esw.w3.org/topic/HTML/TableAccessibility. I agree it has the same issues as title. They both of the advantage of already existing.

[02:55:01.0000] <webben>
hsivonen: maybe a nested <short> element would be nicer.

[02:55:02.0000] <hsivonen>
webben: pre-existence is an advantage if it exists in impls and not just in specs

[02:56:00.0000] <webben>
it would also be good to know exactly how such UAs/AT deal with forms and form controls in th

[02:57:00.0000] <webben>
Another thing this would need solving for is footnotes. I had to create tables for a funds site where a lot of the TH's had footnotes. We don't have any footnote markup yet (awaiting on CSS WG as I understand it) but when we do it will need to support multiple sub-blocks).

[03:02:00.0000] <webben>
hsivonen: JAWS has supported it since at least version 5 apparently: http://www.unt.edu/ACSGAL/jaws50/files/tables_challenge.html

[03:06:00.0000] <webben>
I have a feeling WE doesn't support it, but I'd need to confirm that.

[03:11:00.0000] <hsivonen>
webben: ok

[03:34:00.0000] <webben>
hsivonen: Okay, a quick test Window-Eyes 6.1 supports it too. (Don't know about previous versions.)

[03:34:01.0000] <webben>
*quick test shows

[03:39:00.0000] <hsivonen>
webben: OK. have you checked whether this is on Hixie's plate to ponder?

[03:40:00.0000] <webben>
hsivonen: Nah, I was just thinking incoherently in response to your content model suggestion above.

[03:41:00.0000] <Ben`>
Is the Acid2 test broken? For some reason when I view it in Opera (9.24) it looks messed up.

[03:43:00.0000] <zcorpan>
/me looks at http://simon.html5.org/css21-properties and finds that he wrapped the whole table in a <form>

[03:44:00.0000] <zcorpan>
Ben`: yeah, it's a resource that should be 404 but is now 200

[03:45:00.0000] <hsivonen>
who's in charge of the webstandards.org server? I left a comment on Molly's blog 

[03:45:01.0000] <Ben`>
zcorpan: Ok, thanks.

[03:49:00.0000] <webben>
zcorpan: Those are all GET type queries, not a mixture of querying and modification.

[04:01:00.0000] <hsivonen>
gotta love it when you hit something like this in a library: // FIXME missing two end-of-buffer cases

[04:10:00.0000] <Lachy>
krijn, the logs seem to be down. server isn't responding.

[04:12:00.0000] <krijn>
Lachy: mysql is playing stupid :/

[04:12:01.0000] <krijn>
They work here btw

[04:15:00.0000] <hsivonen>
so Ælfred2 is broken and Xerces is complex

[04:18:00.0000] <hsivonen>
Ælfred2 has just 2 classes (but they contain partly bogus code)

[04:18:01.0000] <hsivonen>
Xerces has layers and layers of interfaces and abstract classes

[04:18:02.0000] <hsivonen>
/me hasn't even found the code the actually implements the Xerces tokenizer

[04:24:00.0000] <hsivonen>
if it weren't for DTDs, I could grab the HTML5 tokenizer and hack until I had an XML tokenizer...

[04:24:01.0000] <hsivonen>
bad, bad DTDs

[04:27:00.0000] <hsivonen>
Xerces in unapproachable. Writing my own DTD handling doesn't seem like smart use of effort.

[04:28:00.0000] <hsivonen>
so I'm stuck with Ælfred2

[04:29:00.0000] <webben>
hsivonen: Can't you use ngsgml to handle XML DTDs for you?

[04:29:01.0000] <hsivonen>
webben: it's not Java

[04:29:02.0000] <hsivonen>
webben: so no

[04:31:00.0000] <webben>
hsivonen: I'd be tempted to hook in a shell process and use its output.

[04:31:01.0000] <webben>
e.g. http://search.cpan.org/dist/XML-DTD/lib/XML/DTD.pm

[04:32:00.0000] <webben>
maybe that would be too slow, I dunno.

[04:33:00.0000] <hsivonen>
webben: more to the point, it doesn't make sense to use an SGML tool written in C when there are XML tools written in Java

[04:33:01.0000] <webben>
hsivonen: I thought you were saying those are broken or over-complex?

[04:34:00.0000] <webben>
are there other Java options?

[04:34:01.0000] <hsivonen>
webben: yes, but I don't see pulling in an SGML tool and slapping it on with duct tape an improvement

[04:34:02.0000] <hsivonen>
webben: I'm looking at Piccolo and Crimson now

[04:34:03.0000] <hsivonen>
webben: though I suspect Piccolo isn't suitable modifiable, either

[04:35:00.0000] <hsivonen>
there's already quite a bit of sunk cost in my fork of Ælfred2

[04:35:01.0000] <webben>
hmmm ... you can /generate/ DTDs... but that's probably not helpful ;) http://saxon.sourceforge.net/dtdgen.html

[04:36:00.0000] <hsivonen>
OK. Piccolo is seriously non-hackable for my purposes.

[04:36:01.0000] <hsivonen>
a big part of the problem is the need to tweak the source location reporting carefully

[04:37:00.0000] <webben>
http://matra.sourceforge.net/ ... don't suppose that's any use

[04:37:01.0000] <hsivonen>
out-of-the box XML parsers don't report the source location in a way that fits the highlight feature of Validator.nu

[04:37:02.0000] <webben>
api here: http://matra.sourceforge.net/javadoc/0.8b/index.html

[04:40:00.0000] <hsivonen>
webben: thanks. that one looks like an overkill, too

[04:40:01.0000] <hsivonen>
if Crimson doesn't turn out to be exceptionally good, I'll just bite the bullet and hack Ælfred2 until it does what I want

[04:49:00.0000] <hsivonen>
Crimson internals are remarkably similar to Ælfred2

[04:50:00.0000] <hsivonen>
these parsers would benefit greatly from an HTML5-style loop around a "data state"

[04:51:00.0000] <webben>
hsivonen: What's a  "loop around 'a data state'"?

[04:51:01.0000] <hsivonen>
webben: I mean the loop you'd use to implement the HTML5 parsing algorithm data state

[04:58:00.0000] <gsnedders>
 Can we get a new version of Christmas? Christmas 5?

[04:59:00.0000] <hdh>
the Bible 5 hasn't get to that part yet, chance to tweak it your way!

[05:01:00.0000] <hsivonen>
I wonder what would happen if the XML mode of Validator.nu had a non-conforming parser that just gave a fatal error upon a doctype

[05:04:00.0000] <hsivonen>
whew. I finally found the bug

[05:32:00.0000] <hsivonen>
zcorpan: no works: http://validator.nu/?doc=http%3A%2F%2Fgolem.ph.utexas.edu%2F~distler%2Fblog%2F&schema=http%3A%2F%2Fc.validator.nu%2Fdebug%2F&showsource=yes

[05:32:01.0000] <hsivonen>
s/no/now/

[06:20:00.0000] <Philip`>
hsivonen: If that fatal-doctype thing happened, I'd have to fix http://canvex.lazyilluminati.com/83/play.xhtml so that it still pointlessly validated :-(

[06:21:00.0000] <Philip`>
which might not be a bad thing

[06:22:00.0000] <hsivonen>
Philip`: yeah, that's a use case

[07:06:00.0000] <hsivonen>
wow, I didn't even remember how much I had to patch the XML parser to make it do the right thing with character encodings

[07:07:00.0000] <hsivonen>
/me is diffing the forked parser against the current upstream HEAD

[09:30:00.0000] <hsivonen>
hmm. http://validator.nu/?doc=http%3A%2F%2Fwww-math.mit.edu%2F18.013A%2FMathML%2Fchapter07%2Fsection01.xhtml&parser=xmldtd&laxtype=yes

[09:36:00.0000] <hsivonen>
looks like a parser bug...

[09:38:00.0000] <hsivonen>
aagh. not a parser bug

[09:38:01.0000] <hsivonen>
Namespaces in XML spec silliness

[09:38:02.0000] <hsivonen>
:-(

[09:46:00.0000] <gsnedders>
Philip`: ping

[10:51:00.0000] <Philip`>
gsnedders: Pung

[10:51:01.0000] <gsnedders>
Philip`: can you see how mange pages have [if IE] on them (i.e., have no version check?)?

[10:51:02.0000] <Philip`>
s/mange/many/ ?

[10:51:03.0000] <gsnedders>
yeah

[10:51:04.0000] <gsnedders>
/me is a guy

[10:51:05.0000] <gsnedders>
/me can't multitask

[10:52:00.0000] <Philip`>
I think I can

[10:52:01.0000] <gsnedders>
I can do somethings at once

[10:52:02.0000] <Philip`>
Is there an easy regexp to find all the various conditional comments?

[10:52:03.0000] <Philip`>
(since that's more interesting than simply looking for [if IE])

[10:52:04.0000] <gsnedders>
how about…

[10:53:00.0000] <gsnedders>
no, that won't work.

[10:53:01.0000] <gsnedders>
matching things like [if (gt IE 5)&(lt IE 7)] is hard

[10:54:00.0000] <Philip`>
<!(--\[[^]]*\]>|\[[^]]*\]--)> should get the relevant bits, I think

[10:54:01.0000] <Philip`>
/me goes off to eat instead of doing this

[10:56:00.0000] <gsnedders>
Philip`: somethings use <![if lt IE 5]> with no --

[11:21:00.0000] <Philip`>
gsnedders: Okay, I'll try looking for <!-*\[[^]]*\]-*>

[11:21:01.0000] <Philip`>
(assuming there's no nested square brackets allowed)

[11:22:00.0000] <Philip`>
By the way, did you mention looking for Location headers a while ago?

[11:23:00.0000] <Philip`>
(If so, I can't find them trivially, since I've got HttpClient doing its normal automatic redirection-following and only saved the post-redirection headers)

[11:25:00.0000] <Philip`>
/me wonders how to do regular expressions in Java

[11:26:00.0000] <hsivonen>
Philip`: java.util.regex?

[11:46:00.0000] <alp>
i'm tracking down a bug in WebKit's cairo canvas implementation. our IDL says the last parameter for arc() is "in boolean clockwise" while the html5 spec says "in boolean anticlockwise"

[11:49:00.0000] <Philip`>
alp: When I last looked, all browsers implemented it as anticlockwise

[11:50:00.0000] <Philip`>
(e.g. 2d.path.arc.angle.1 on http://philip.html5.org/tests/canvas/suite/tests/results.html )

[11:51:00.0000] <Philip`>
gsnedders: http://philip.html5.org/data/ccs.html

[11:54:00.0000] <gsnedders>
Philip`: thx

[11:54:01.0000] <gsnedders>
Philip`: and ergh.

[11:54:02.0000] <gsnedders>
Philip`: exactly why they need a new switch: everything just tests for IE with no version :\

[11:57:00.0000] <Philip`>
Has anyone documented IE's quirks mode logic?

[11:57:01.0000] <gsnedders>
Philip`: for entering quirks mode?

[11:57:02.0000] <Philip`>
Yes

[11:58:00.0000] <gsnedders>
Philip`: MS publishes it somewhere

[11:58:01.0000] <Philip`>
/me should try to ask questions that actually say what the question is, instead of just the general topic

[11:58:02.0000] <Philip`>
I remember it having weird stuff like matching against the string "NETSCAP", but never saw that in any documentation

[11:58:03.0000] <gsnedders>
http://msdn2.microsoft.com/en-us/library/bb250395.aspx#cssenhancements_topic2

[11:58:04.0000] <gsnedders>
ah

[11:59:00.0000] <gsnedders>
Philip`: try asking Chris Wilson or Tantek

[11:59:01.0000] <gsnedders>
/me expects you'd get an answer from one

[12:00:00.0000] <Philip`>
Ah, it was http://krijnhoetmer.nl/irc-logs/html-wg/20070714#l-67

[12:02:00.0000] <gsnedders>
http://lists.w3.org/Archives/Public/www-style/2007Dec/0186.html

[12:07:00.0000] <alp>
Philip`: thanks. (filed as http://bugs.webkit.org/show_bug.cgi?id=16558 for reference)

[12:09:00.0000] <Philip`>
Maybe Microsoft could set up something like their phishing filter, so IE asks MS's server whether a given URI should be quirks or IE7-bugs or IE8-bugs or real-standards mode, and MS can quickly update their database to include sites which need buggy rendering and which are important enough and complex enough that their owner complains to MS instead of updating their site

[12:10:00.0000] <jwalden>
*cough* privacy *cough* </wonk>

[12:10:01.0000] <jwalden>
surely somebody'd bring that up

[12:11:00.0000] <Philip`>
Privacy is so 20th century

[12:12:00.0000] <Philip`>
Anyway, they could do it like Firefox instead where it downloads a database of broken sites and just does checks against that local list

[12:13:00.0000] <hsivonen>
An admin-configurable mode setting for an "intranet zone" would probably address many intranet concerns: http://weblogs.mozillazine.org/roc/archives/2007/11/intranet_explor.html

[12:14:00.0000] <gsnedders>
Philip`: Fx works like that? oh :P

[12:14:01.0000] <hsivonen>
doesn't Opera's JS fixup system download patching scripts?

[12:14:02.0000] <roc>
gsnedders: Fx does not work like that

[12:14:03.0000] <roc>
it works like that for phishing and malware detection, sure

[12:14:04.0000] <gsnedders>
/me blames Philip` 

[12:14:05.0000] <roc>
but we don't classify quirks-mode pages as malware :-)

[12:16:00.0000] <Philip`>
That's why I said "like Firefox" instead of "precisely the same as Firefox", continuing the analogy between phishing filters and quirks-mode filters :-)

[12:16:01.0000] <jruderman>
you could have said "like Firefox does for phishing" :)

[12:17:00.0000] <othermaciej>
if they always put unknown doctypes in full standards mode, that somewhat mitigates the damage of multiple quirks modes

[12:17:01.0000] <othermaciej>
but now I see why they want the HTML doctype to be ever-changing

[12:21:00.0000] <jwalden>
they should just say <!DOCTYPE html> is super-hyper-mega-standards mode where they can and will break non-conformant pages without prejudice in the name of supporting standards

[12:21:01.0000] <jwalden>
and evangelize the heck out of that message

[12:21:02.0000] <roc>
/me grumbles that every time he hears about intranet compatibility issues Boeing is always the example

[12:22:00.0000] <jwalden>
planely their intranet needs improvement

[12:22:01.0000] <hsivonen>
that kind of policy would work better if the Web devs were kept aware of constant change by churning out a new release every 6 months or more frequently

[12:22:02.0000] <roc>
Boeing is the reason MS can't roll out IE security updates fast, Boeing is the reason there need to be a bazillion compat modes

[12:23:00.0000] <roc>
They should use some other examples

[12:24:00.0000] <roc>
I agree, we probably can't (and maybe shouldn't) prevent IE offering one mode per IE engine, but we should have an "all standards, all the time, always the latest IE engine" mode

[12:26:00.0000] <roc>
they're making an interesting bed for themselves though, when IE10 is shipping 4 entirely separate layout engines, especially for mobile and QA

[12:26:01.0000] <gsnedders>
Philip`: ccs.html says 2007-17-17 :P

[12:27:00.0000] <Philip`>
What happens when a super-hyper-mega-standards compliant thing like Dreamweaver or Wordpress uses <!doctype html> because their users expect them to be super-hyper-mega standards compliant, and then millions of separate people use those things, and then actually they didn't quite follow the standards (or the standards change) and they break?

[12:27:01.0000] <jruderman>
IE8 actually contains older layout engines? as opposed to having backwards-compatibility switches within its engine?

[12:27:02.0000] <gsnedders>
jruderman: yeah

[12:27:03.0000] <jruderman>
yikes

[12:27:04.0000] <gsnedders>
jruderman: IE7 has Trident IV too

[12:27:05.0000] <Philip`>
gsnedders: Uh

[12:27:06.0000] <Philip`>
gsnedders: You pass the observation test!

[12:27:07.0000] <gsnedders>
Philip`: no, I don't

[12:27:08.0000] <gsnedders>
Philip`: someone else on www-style does :)

[12:28:00.0000] <jwalden>
Philip`: that's where the evangelism comes in, and Wordpress at least requires updates for security anyway

[12:28:01.0000] <gsnedders>
jwalden: loads of old copies of WP out there

[12:28:02.0000] <gsnedders>
(WP, ergh)

[12:28:03.0000] <jwalden>
not ones anyone cares about :-)

[12:28:04.0000] <jruderman>
wordpress needs auto-update

[12:28:05.0000] <Philip`>
gsnedders: Fixed

[12:28:06.0000] <jwalden>
yes, it does

[12:29:00.0000] <roc>
jruderman: they might merge the engines somehow, but that sounds like a maintenance nightmare

[12:29:01.0000] <hsivonen>
WP's standards update track record isn't that great. case study: Atom

[12:30:00.0000] <gsnedders>
/me mentions <http://comox.textdrive.com/pipermail/wp-hackers/2007-May/012987.html> and makes no further comment on WP

[12:30:01.0000] <gsnedders>
roc: IE6 is still maintained for older Windows releases, so they update the engine that way anyway

[12:30:02.0000] <gsnedders>
roc: and the maintenance has nothing to do with the IE team at all :P

[12:31:00.0000] <roc>
I mean ongoing development maintainance, not maintenance of old versions

[12:34:00.0000] <jruderman>
roc: i don't understand what you're saying. if a different team takes care of maintaining the old versions, why does keeping them around cause "ongoing development maintainance" issues?

[12:36:00.0000] <gsnedders>
IE8 Standards Mode sometimes/always (I wasn't totally clear on what Chris meant) needs a re-parse of the HTML, FWIW

[12:41:00.0000] <roc>
I'm talkiing about maintenance activities like cleaning up the code to lay the groundwork for future improvements

[12:42:00.0000] <roc>
basically a single code base that implements N "modes" of which at least N-1 must never change behaviour in any way sounds really hard to work with

[12:43:00.0000] <roc>
but the obvious alternative, shipping N different versions of the code, sounds bad too

[12:43:01.0000] <roc>
maybe some automatic tool takes the N versions and weaves them together, sharing identical code where possible...

[12:43:02.0000] <hober>
roc: the worst part is the burden that could very well place on non-IE browsers, who might also be required to implement all N modes (they already implement 3? 4?)

[12:44:00.0000] <roc>
hober: no

[12:44:01.0000] <hsivonen>
I wouldn't be too surprised if the CSS formatter of IE8 new standards mode were separate code and the quirks mode and IE7 standards mode CSS formatters were one thing with conditional branches

[12:44:02.0000] <roc>
because at least in the case of Gecko, we have set expectations appropriately

[12:44:03.0000] <gsnedders>
hsivonen: Chris has said publicly that IE7 has both Trident IV and Trident V in it

[12:44:04.0000] <roc>
hsivonen: yeah

[12:45:00.0000] <gsnedders>
which implies two things

[12:45:01.0000] <roc>
oh OK, never mind then

[12:45:02.0000] <gsnedders>
(where Chris == Chris Wilson)

[12:45:03.0000] <hsivonen>
gsnedders: ok

[12:45:04.0000] <roc>
the problem is not so much IE8 but that the pain must increase with every release

[12:45:05.0000] <hober>
roc: right. where does that leave the web in 10, 15, 20 years?

[12:45:06.0000] <roc>
hmm, the tool approach would be interesting, but would have its own limitations and surprises

[12:46:00.0000] <gsnedders>
I think part of their issue is they don't want to reimplement anything because people rely on bugs they might not kow about

[12:46:01.0000] <roc>
hober: it leaves IE in a pickle and Gecko and Webkit looking better and better, I reckon

[12:46:02.0000] <hober>
roc: that's the optimistic outcome, sure

[12:46:03.0000] <roc>
the frozen intranets will either be stuck with IE swiss-army-knife or they'll figure out how to modernize themselves

[12:47:00.0000] <roc>
the frozen intranets are really a different thing from the public Web

[12:47:01.0000] <hober>
right, I'm completely uninterested in such intranets

[12:47:02.0000] <roc>
which is why I suggested in my blog Microsoft fork IE

[12:47:03.0000] <hsivonen>
It'll be interesting to see what the IE8 new standards more DOM will be like

[12:48:00.0000] <hsivonen>
for example, will it be the old one but the non-tree links turned off?

[12:48:01.0000] <hsivonen>
I'd expect a new CSS formatter to want a tree DOM

[12:48:02.0000] <gsnedders>
I'm wondering if they're doing one major thing a release: CSS parsing in IE7, layout engine in IE8…

[13:04:00.0000] <krijn>
gsnedders: you really seem to like WP ;)

[13:04:01.0000] <gsnedders>
krijn: :P

[13:05:00.0000] <krijn>
/me doesn't like it either

[13:05:01.0000] <krijn>
Too bloated

[13:06:00.0000] <krijn>
So in stead I've built my own bloated thing :\

[13:06:01.0000] <gsnedders>
I used Habari 0.4-alpha :P

[13:07:00.0000] <Philip`>
/me wonders what the least bloated blog management system could be

[13:07:01.0000] <krijn>
QMS!

[13:07:02.0000] <krijn>
;>

[13:07:03.0000] <krijn>
Not open source though

[13:08:00.0000] <Philip`>
Maybe a blog could just an Atom file, which you edit in your favourite text editor and then upload with FTP or whatever, and then the system checks it in the feed validator and publishes it if everything's okay

[13:08:01.0000] <Philip`>
s/just an/just be an/

[13:09:00.0000] <hober>
Philip`: that's almost exactly what my blog software is

[13:09:01.0000] <gsnedders>
Philip`: PyBlosxom?

[13:09:02.0000] <Philip`>
hober: Why only "almost"? :-)

[13:09:03.0000] <hober>
basically, I have a directory tree which contains a bunch of files, some of which are atom entry documents.

[13:09:04.0000] <gsnedders>
Philip`: that uses just files, not Atom

[13:09:05.0000] <hober>
the publish script is basically an atom-aware cp -r

[13:10:00.0000] <Philip`>
*Multiple* files? That's just bloat :-(

[13:10:01.0000] <hober>
heh.

[13:12:00.0000] <gsnedders>
Philip`: less to send to the feed validator, though :P

[13:15:00.0000] <hober>
yeah, were it a single file, it would be about 100,000 lines

[13:36:00.0000] <gsnedders>
Bebo is so awesome. I just got a java.lang.ArrayIndexOutOfBoundsException.

[13:36:01.0000] <jwalden>
that got mentioned in the web2.0 video, dinnit?

[13:58:00.0000] <hsivonen>
jwalden: the Richter Scales video? yes.

[14:19:00.0000] <Hixie>
man the idea that we can get two complete and correct implementations of html5 in 3 years is so funny

[14:19:01.0000] <Philip`>
/me sees http://www.snellspace.com/wp/?p=832 for a way to handle degradation of http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-December/013493.html

[14:20:00.0000] <hsivonen>
annevk: https://bugzilla.mozilla.org/show_bug.cgi?id=409404 Any comment on Opera behavior?

[14:21:00.0000] <Philip`>
I assume it needs two implementations which each do the entire spec correctly, rather than n implementations where each feature is correct in >= 2 of them?

[14:21:01.0000] <annevk>
svg:focusable is wrong

[14:21:02.0000] <annevk>
it's focusable in no namespace

[14:21:03.0000] <othermaciej>
IE ships multiple engines? really?

[14:21:04.0000] <othermaciej>
(was reading scrollback)

[14:22:00.0000] <Hixie>
IE8 will have three, as i understand it

[14:22:01.0000] <Hixie>
IE8, IE7, and IE5.5

[14:22:02.0000] <annevk>
hopefully Opera works nothing like what's suggested in that bug

[14:22:03.0000] <annevk>
if it does it would be a bug

[14:23:00.0000] <annevk>
(and I don't think Opera does, as far as I know we we "ignore" version= and baseProfile=)

[14:23:01.0000] <hsivonen>
annevk: ok

[14:23:02.0000] <Hixie>
Philip`: yeah, i'm really tempted to just tell people to use a quick script for that, given how easy it is to do

[14:23:03.0000] <Hixie>
Philip`: it really doesn't seem groundbreaking enough to warrant browser-side support

[14:25:00.0000] <annevk>
hsivonen, I don't think I need to comment on it, right?

[14:26:00.0000] <hsivonen>
annevk: no, I'll mention your opinion

[14:30:00.0000] <hsivonen>
I guess I should allow focusable in Validator.nu in due course even though the base schema is 1.1

[14:30:01.0000] <hsivonen>
ARIA, too

[14:32:00.0000] <annevk>
are people sure IE actually ships with _different_ layout engines?

[14:33:00.0000] <annevk>
/me thought the APIs had some type of case/switch stuff for different versions

[14:33:01.0000] <gsnedders>
annevk: Chris Wilson ha said multiple times.

[14:33:02.0000] <gsnedders>
*has

[14:34:00.0000] <gsnedders>
cmd+w on wrong tab

[14:34:01.0000] <gsnedders>
ergh.

[14:37:00.0000] <annevk>
hmm ok

[14:37:01.0000] <annevk>
that doesn't seem like a winning strategy to me

[14:38:00.0000] <annevk>
unless they can lock a large set of pages into each engine...

[14:39:00.0000] <gsnedders>
/me waves g'nite

[14:39:01.0000] <Philip`>
hsivonen: The "DOM Consistency" principle seems to be about consistency of the DOM from the perspective of scripts, and about minimising syntactic differences in the HTML and XHTML representations of the same DOM; so I'm not sure how it's relevant to an author requirement to use some XHTML features unless you have a good reason otherwise (where e.g. needing to convert the document to HTML is a good reason)

[14:41:00.0000] <hsivonen>
Philip`: I'm not saying XHTML5 authors should put stuff in script. I'm saying we shouldn't say that they shouldn't in normative language

[14:41:01.0000] <othermaciej>
I'm curious how the multiple engine thing will work out for their promised future "desktop grade" browser for Windows Mobile

[14:41:02.0000] <othermaciej>
maybe it will include only one of the engines

[14:42:00.0000] <hsivonen>
when was such a browser promised?

[14:43:00.0000] <Philip`>
hsivonen: That sounds reasonable to me; I don't understand the reference to DOM Consistency, though

[14:45:00.0000] <hsivonen>
Philip`: it permits doing stuff that you can do in HTML5 also in XHTML5

[14:45:01.0000] <hsivonen>
in DOM terms

[14:50:00.0000] <Philip`>
hsivonen: DOM Consistency sounds like it's only talking about UA conformance, and not about document conformance at all - it's only relevant to the design of the HTML parsing algorithm (HTML and 'equivalent' XHTML should be parsed the same) and the design of the HTML language (since HTML-serialised DOMs should look as similar as possible to XHTML-serialised DOMs)

[14:52:00.0000] <hsivonen>
Philip`: it's also relevant to not having gratuitously different conformance requirements so that you can take XHTML5 without foreign namespaces on throw it at a serializer that gives you HTMl5

[14:53:00.0000] <Philip`>
...but it seems the problem may be with what 'equivalent' means - if HTML5 said you shouldn't use <script> in XHTML, then <script type=image/svg+xml><svg .../></script> and <svg xmlns="http://whatever">...</svg> would be the equivalent representations of the same high-level concept

[14:54:00.0000] <Philip`>
(which is a significant syntactic difference between the 'same' document in two serialisations)

[14:54:01.0000] <hsivonen>
Philip`: I don't consider those equivalent in DOM terms

[14:54:02.0000] <hsivonen>
Philip`: not even on a higher level as one requires a script to aid it

[14:55:00.0000] <Philip`>
(Replace "svg" with "some-custom-xml-dialect", so in both cases it's a lump of unknown XML which requires a script to aid it)

[14:58:00.0000] <Philip`>
<html lang=en> and <html xml:lang="en"/> are not equivalent in DOM terms, but they are conceptually equivalent at a slightly higher level, and that kind of difference is what DOM Consistency says should be minimised

[14:59:00.0000] <Philip`>
so similarly <script type=custom-stuff> and <custom-stuff xmlns="..."> are not equivalent in DOM terms, but are equivalent at a higher level, so that difference should be minimised too

[15:00:00.0000] <Philip`>
so I think I agree with your email :-)

[15:01:00.0000] <hsivonen>
good. :-)

[15:03:00.0000] <annevk>
so I watched parts of the IE8 Acid2 video

[15:04:00.0000] <annevk>
there's a scrollbar in their rendering...

[15:55:00.0000] <annevk>
https://bugzilla.mozilla.org/show_bug.cgi?id=409436 - whitespace in gecko - someone should track that

[15:55:01.0000] <annevk>
maybe i should point to html5...

[15:57:00.0000] <annevk>
/me -> bed


2007-12-22
[16:03:00.0000] <Philip`>
Hmm, silly Gentoo

[16:03:01.0000] <Philip`>
I upgraded Opera to 9.25, and now when I click on the Opera icon it launches Opera 9.50 Alpha inside Wine

[16:28:00.0000] <kig>
http://glimr.rubyforge.org/cake/canvasmage.html hihih

[16:30:00.0000] <Philip`>
kig: Doesn't work in Opera 9.2 :-(

[16:30:01.0000] <Philip`>
"Syntax error while loading: line 55 of linked script at http://glimr.rubyforge.org/cake/canvasmage.js :   }"

[16:31:00.0000] <Philip`>
(maybe due to a trailing comma in a { ... } list?)

[16:32:00.0000] <kig>
thanks, took it out

[16:32:01.0000] <Philip`>
kig: Works better now - thanks :-)

[16:32:02.0000] <Philip`>
The creature looks pretty broken in Opera, but that's probably just arc bugs

[16:33:00.0000] <kig>
circle strokes with gaps and visible polygon segments?

[16:34:00.0000] <Philip`>
kig: http://tinyurl.com/3y9kp3

[16:35:00.0000] <kig>
dfsdg tinyurl data urls don't quite work in firefox

[16:35:01.0000] <Philip`>
Use Opera ;-)

[16:36:00.0000] <inimino>
with its buggy <canvas>? ;-)

[16:36:01.0000] <Philip`>
/me ought to find a better way of uploading small files

[16:36:02.0000] <kig>
wow, that is messed up

[16:36:03.0000] <Philip`>
inimino: It's not really more buggy than anyone else's <canvas> :-)

[16:37:00.0000] <inimino>
Philip`: Perhaps, but when I tried to use <canvas> last the only bugs I hit were in Opera

[16:38:00.0000] <jwalden>
kig: dunno if you have time, but is tinyurl trying to 302 redirect to them or something?

[16:38:01.0000] <kig>
jwalden: it gives 301 with the data: url

[16:38:02.0000] <jwalden>
I seem to recall that intentionally not being supported, sec

[16:38:03.0000] <kig>
which looks funny when trying to wget

[16:38:04.0000] <inimino>
(specifically just arc() stuff, that was about as far as I got)

[16:40:00.0000] <kig>
Philip`: does it look any better now?

[16:40:01.0000] <Philip`>
kig: It looks one pixel high

[16:41:00.0000] <kig>
eh?

[16:42:00.0000] <Philip`>
The example images are about 640x1 and 320x1 and 1200x1, like lines with some yellow/black/white gradient

[16:42:01.0000] <kig>
ghhh

[16:42:02.0000] <Philip`>
(By the way, your .json files aren't really JSON because you'd need to quote the key strings)

[16:43:00.0000] <jwalden>
kig: https://bugzilla.mozilla.org/show_bug.cgi?id=211999#c19 seems to say it's fixed in 3.0, doesn't sound like it'll be fixed for 2.0 unless somebody complains

[16:44:00.0000] <kig>
jwalden: alright

[16:44:01.0000] <Philip`>
jwalden: Can I complain? I like having a tinyurlfs - but then again I like telling people to use Opera so maybe it's good that FF2 doesn't support it ;-)

[16:45:00.0000] <jwalden>
Philip`: complain in the bug if you want :-)

[16:45:01.0000] <jwalden>
redirecting to data URLs isn't especially mainstream, so I'm not super-worried if it doesn't work, to be honest

[16:46:00.0000] <Philip`>
kig: By the way, you could perhaps use ctx.__defineSetter__('fillStyle', ...) in Firefox and Opera 9.5 (and maybe others?) to catch people saying 'ctx.fillStyle = ...', instead of requiring 'ctx.setFillStyle(...)'

[16:46:01.0000] <jwalden>
think Safari added getter/setter support recently, dunno if it's in 3 or not

[16:48:00.0000] <kig>
Philip`: yeah, that'd be nice

[16:50:00.0000] <Philip`>
kig: It's no longer one pixel tall - it looks about the same as it did before

[16:50:01.0000] <Philip`>
s/about/exactly/

[16:50:02.0000] <kig>
one pixel tall was using "draw, insert canvas, remove image"

[16:50:03.0000] <Philip`>
s/exactly/exactly except for a little bit around the beak/

[16:51:00.0000] <kig>
now it's using "draw, img.src = canvas.toDataURL()"

[16:51:01.0000] <Hixie>
acid2 is apparently fixed

[16:51:02.0000] <Philip`>
kig: Hmm, odd

[16:51:03.0000] <kig>
but i'm doing hackery with copying the image style to the canvas style, so that might screw things up

[16:54:00.0000] <kig>
Philip`: try now

[16:54:01.0000] <Philip`>
kig: By the way, the CanvasMage thing is quite neat :-)

[16:56:00.0000] <Philip`>
It looks slightly more broken in Opera, since the red rectangle is now a wide triangle

[16:57:00.0000] <kig>
alright, that's screwed up

[16:57:01.0000] <kig>
i'll have to try and debug on opera tomorrow

[16:57:02.0000] <Philip`>
Oh

[16:57:03.0000] <Philip`>
It's random

[16:57:04.0000] <Philip`>
(It tends to change each time I reload the page)

[16:58:00.0000] <Philip`>
Probably same bug as http://canvex.lazyilluminati.com/misc/operanondet.html which gets rendered as http://canvex.lazyilluminati.com/misc/operanondet.png (but differently each time you reload)

[16:59:00.0000] <kig>
okay, opera gets the wtf award of the day

[17:01:00.0000] <Philip`>
That's fixed in 9.50

[17:01:01.0000] <Philip`>
The whole image looks correct in 9.50, except for some gaps where curves don't join up neatly

[17:02:00.0000] <Philip`>
/me would probably suggest giving up on Opera 9.2 because it's far too broken

[17:03:00.0000] <kig>
yeah :/

[17:03:01.0000] <Philip`>
(or emulate the curve-drawing functions with JS)

[17:04:00.0000] <kig>
i thought of breaking circles into four beziers but then i found another shiny bauble

[17:04:01.0000] <Philip`>
Hmm, I thought it was impossible to represent circles as Beziers, but maybe I'm misremembering

[17:05:00.0000] <Philip`>
http://pages.sbcglobal.net/tymmott-smith/Math/Bezier.html - "we cannot get a circle with a Bezier curve" - ah, I wasn't wrong

[17:05:01.0000] <Philip`>
though maybe you can approximate something good enough

[17:06:00.0000] <kig>
flash (iirc) uses quadratic curves for everything

[17:06:01.0000] <Philip`>
Flash doesn't have to worry about interoperability or about passing annoying test cases that try to ensure curves are exactly the right shape ;-)

[17:08:00.0000] <kig>
did some tests with canvasmage animations, but they tend to be a bit big. 300k for 50 frames uncompressed, 30k gzipped

[17:09:00.0000] <kig>
1k procedural

[17:10:00.0000] <Philip`>
kig: What is the advantage of saving images/animations rather than just re-executing the original canvas script every time someone views it?

[17:11:00.0000] <kig>
faster to execute, more evil

[17:13:00.0000] <Philip`>
Maybe you could make the output much smaller by limiting the number of decimal places in all the floats, since those won't compress well

[17:13:01.0000] <kig>
e.g. image/animation based on heavy physics simulation, might take hours to generate, but the command array runs at 60fps

[17:15:00.0000] <kig>
yeah, let's see

[17:19:00.0000] <Philip`>
/me doesn't like how he keeps writing <script src=...> and forgetting the </script>

[17:21:00.0000] <kig>
stripping decimals might save 10-20%, but circles get gaps 

[17:22:00.0000] <Philip`>
Can you strip fewer decimals?

[17:31:00.0000] <kig>
no visible gaps with four decimals, 10% smaller gzipped

[17:32:00.0000] <kig>
(599b vs 664b, i need a bigger test file)

[17:32:01.0000] <Philip`>
Try recording the output of Canvex ;-)

[17:33:00.0000] <kig>
no image support yet :<

[17:43:00.0000] <kig>
but hmm, maybe it'll work. let's see

[17:45:00.0000] <kig>
would have to replace render_ctx with a RecordingContext and add frame markers to the command stream

[17:45:01.0000] <jruderman>
Philip`: i forget the </script> even for inline scripts

[17:45:02.0000] <jruderman>
Philip`: i wish firefox would warn me instead of just showing a blank page

[17:46:00.0000] <jruderman>
Philip`: and i'm the one who insisted on that behavior :P

[01:21:00.0000] <annevk>
hmm, I didn't even get that bug about version= and baseProfile=

[04:01:00.0000] <salty-horse>
hi. seen this? http://stephenfry.com/blog/?p=32

[04:02:00.0000] <annevk>
"due out soon"

[04:02:01.0000] <annevk>
heh

[04:03:00.0000] <annevk>
seems also that WHATWG is now a synonym for W3C...

[04:04:00.0000] <salty-horse>
you're welcome to comment with corrections

[04:06:00.0000] <Philip`>
"all subscribe to the Working Hypertext Application Technology Working Group mailing list (whatwg.org/mailing-list#specs) and exert pressure" - hmm, that probably won't have the desired effect

[04:08:00.0000] <hsivonen>
ouch. I'm a bit disappointed. I expected more accurate posts on stephenfry.com

[04:09:00.0000] <Lachy>
omg, that's the most innacurate crap I've read in a while

[04:09:01.0000] <Lachy>
since when did Nokia claim any ownership of ogg?

[04:10:00.0000] <Lachy>
and Apple have explicitly stated that they're not pushing their own formats

[04:11:00.0000] <Philip`>
Lachy: What someone explicitly states is not necessarily what they are really doing

[04:11:01.0000] <Philip`>
so saying "Apple doesn't like ..." is not clearly wrong (whereas saying "Apple says it doesn't like ..." would be, since they don't say that)

[04:13:00.0000] <Philip`>
(and saying "Nokia claims ownership of some elements of Ogg" is wrong since they haven't claimed that (as far as I'm aware))

[04:45:00.0000] <kig>
heh, svgtext.svg is 64k, 27k gzipped. svgtext.json is 79k, 22k gzipped

[04:55:00.0000] <hdh>
heh, that stephenfry post also conflates open source and open standard, not a good move

[05:03:00.0000] <Dashiva>
If it's not open source, it's not open at all!

[05:05:00.0000] <hdh>
some business people twitch at open source (see "cancer"); but none can dispute the words "open standard"

[05:28:00.0000] <tndH>
okay, so now guardian readers think html 5 is evil?

[05:29:00.0000] <gsnedders>
tndH: what? in today's? where?

[05:29:01.0000] <tndH>
http://www.guardian.co.uk/technology/2007/dec/22/internet.digitalmedia

[05:30:00.0000] <tndH>
(same as the stephenfry.com link before)

[05:31:00.0000] <hsivonen>
well, the silver lining is that now the WHATWG list is noted on the Guardian

[05:31:01.0000] <gsnedders>
We're in print!

[05:32:00.0000] <hsivonen>
(if you believe that all publicity is eventually good)

[05:35:00.0000] <gsnedders>
Hey, more uninformed emails!

[05:44:00.0000] <tndH>
all i can imagine while reading through the stephen fry article is the QI siren going off and deducting him 10 points

[06:01:00.0000] <salty-horse>
are you going to email the guardian/post a comment?

[06:05:00.0000] <maikmerten>
err....

[06:05:01.0000] <maikmerten>
"The problem is that Nokia claims ownership of some elements of Ogg, which is essentially the file format for the streaming and delivery part of the codec"

[06:05:02.0000] <maikmerten>
that is... not correct

[06:06:00.0000] <maikmerten>
Nokia to my knowledge never made such a claim

[06:13:00.0000] <gsnedders>
salty-horse: the issue is most of the copies of that article will be in print, and already sold.

[06:14:00.0000] <maikmerten>
actually here in Germany HTML5 is in print, too

[06:14:01.0000] <maikmerten>
in Magazin für Computertechnik (c't)

[06:14:02.0000] <maikmerten>
again mostly about <video> and codecs

[06:16:00.0000] <Philip`>
There will eventually be a point at which we have enough bandwidth and processing power to store video losslessly at the maximum resolution the human eye can perceive, so this codec thing is only a temporary problem

[06:17:00.0000] <gsnedders>
Philip`: the total bandwidth needed once we get more and more online would be huge, though

[06:18:00.0000] <maikmerten>
and even then there's the question why anyone would want to be so wasteful with resources

[06:18:01.0000] <salty-horse>
gsnedders, he could issue a clarification in the next column -- it's not the first time a print article was wrong

[06:18:02.0000] <gsnedders>
salty-horse: that's true

[06:18:03.0000] <maikmerten>
video can be compressed by a factor > 10 without anyone noticing anything

[06:18:04.0000] <maikmerten>
(usual factor 100 and bigger)

[06:18:05.0000] <gsnedders>
/me is half asleep doing too much

[06:19:00.0000] <gsnedders>
/me sigha

[06:19:01.0000] <gsnedders>
*sighs

[06:19:02.0000] <gsnedders>
IE6 is still the most common UA to my site :\

[06:19:03.0000] <gsnedders>
/me blames search referrers such as "why does life get complicated when you think you gay"

[06:20:00.0000] <takkaria>
hah

[06:20:01.0000] <gsnedders>
(OK, that only got one, but you get the ideA)

[06:20:02.0000] <gsnedders>
*idea

[06:24:00.0000] <jwalden>
hsivonen: too bad Hixie actually didn't think to make it a task force instead of a working group ;-)

[06:25:00.0000] <gsnedders>
WTF?

[06:25:01.0000] <gsnedders>
(as in WHATTF?)

[06:27:00.0000] <jwalden>
yes, exactly

[06:27:01.0000] <jwalden>
I remember him saying one time that if he'd thought of it he'd have done it :-)

[06:30:00.0000] <hsivonen>
jwalden: well, fantasai has whattf.org

[06:30:01.0000] <takkaria>
I dislike Christmas; people stop posting so often to mailing lists and I find myself reading slashdot instead

[06:30:02.0000] <jwalden>
well-played!

[06:44:00.0000] <hsivonen>
so I'm thinking that I should provide a setting for filtering out foreign namespaces

[06:44:01.0000] <hsivonen>
should it filter out tags or the entire subtree rooted at a dropped element?

[06:45:00.0000] <hsivonen>
I'm leaning towards dropping the entire subtree

[06:45:01.0000] <hsivonen>
(consider foreign metadata elements in <head> or SVG <metadata> and what would happen if only text content were left)

[07:54:00.0000] <gsnedders>
/me is amazed that Lachy is actually twittering about snow in Norway

[08:35:00.0000] <annevk>
html5.org got over a 100.000 page requests on 11 December

[08:36:00.0000] <annevk>
normal amount is about 800-1000 a day

[08:36:01.0000] <gsnedders>
/me blinks, then realises annevk is doing the European norm of . being a thousand separator

[08:39:00.0000] <Dashiva>
gsnedders: Lachy is new in town :)

[08:40:00.0000] <gsnedders>
Dashiva: I know, it still seems odd.

[08:40:01.0000] <gsnedders>
/me wants snow here (where here = east coast of Scotland)

[08:40:02.0000] <takkaria>
/me wants snow in the middle of Manchester, UK, but it's not going to happen. :)

[08:41:00.0000] <annevk>
/me would like 20 degrees in NL

[08:41:01.0000] <gsnedders>
haven't had proper snow here since 1997

[08:41:02.0000] <annevk>
(Celcius)

[08:42:00.0000] <gsnedders>
we have 275K here

[08:43:00.0000] <gsnedders>
(i.e., 2°C)

[14:56:00.0000] <gsnedders>
oh I do love PHP. randomly using a 1-based array for unpack() :\

[14:58:00.0000] <hdh>
on PHP, I just came across this, http://zestyping.livejournal.com/233348.html

[14:58:01.0000] <gsnedders>
hdh: not taught? it shouldn't exist!

[14:59:00.0000] <hdh>
http://zestyping.livejournal.com/124503.html?thread=690519#t690519 ← is that you ?-P

[14:59:01.0000] <gsnedders>
bo

[14:59:02.0000] <gsnedders>
*no

[14:59:03.0000] <hdh>
:)

[15:01:00.0000] <gsnedders>
/me wonders whether to rely on PHP have this bizarre behaviour forever or not

[15:03:00.0000] <Philip`>
Regardless of that particular bizarre behaviour, it's probably quite safe to rely on PHP having some bizarre behaviour forever, and thus use Perl or Python or Ruby or anything else :-)

[15:03:01.0000] <gsnedders>
no, Philip`, the PHP devs claim to not break b/c then break it in some bizarre way that breaks everything :)

[15:03:02.0000] <gsnedders>
(and I don't have choice about programming language, sadly)

[15:04:00.0000] <hdh>
/me is still temped towards touching some PHP, the HTML embedness feels nice, like XSLT

[15:06:00.0000] <gsnedders>
hdh: don't.

[15:06:01.0000] <gsnedders>
hdh: As someone who's used PHP for years, I can wholeheartedly say you don't want to use PHP.

[15:06:02.0000] <Philip`>
hdh: There are about three million different HTML templating systems in pretty much every language, so that's no reason to choose PHP :-)

[15:07:00.0000] <hdh>
uhm, ye, I haven't checked any out yet

[15:08:00.0000] <hdh>
/me spent some moments looking for a nice [Tt] and [Ff], and found "WƮƑ?", to avoid spiders making his page as non-safe

[15:09:00.0000] <hdh>
s/making/marking/

[15:10:00.0000] <gsnedders>
/me wonders how his UTF-32 decoder is more expensive than his UTF-16 one

[15:11:00.0000] <gsnedders>
re-running it makes UTF-32 < UTF-8 < UTF-16 which is still odd

[15:11:01.0000] <gsnedders>
third time I get UTF-32 < UTF-16 < UTF-8 which is the expectation

[15:13:00.0000] <Philip`>
/me is currently using Template Toolkit in Perl, but doesn't really like it because it has seemingly-arbitrary syntactic restrictions so e.g. you can't use expressions inside [...]-style list literals

[15:15:00.0000] <hdh>
/me got a flash like "hook cheetah to pyblosxom"; a name to hunt down

[15:40:00.0000] <Philip`>
/me must remember not to set up web pages that do stuff like <a href="edit/[%message_id]">[% message_title %]</a> when message_title might be an empty string

[15:40:01.0000] <Philip`>
Uh, s/[%message_id]/[% message_id %]/


2007-12-23
[16:53:00.0000] <hdh>
/me votes for <m> staying, it makes more sense than strong i b em for highlighting blockquotes

[17:14:00.0000] <Lachy_>
hdh, where was there any discussion of removing <m>?

[17:14:01.0000] <hdh>
#the-m says so

[17:15:00.0000] <hdh>
"This section has a large number of outstanding comments    and will likely be rewritten or removed from the spec."

[17:16:00.0000] <Lachy_>
nooo! It absolutely must stay

[17:16:01.0000] <hdh>
ai'ight, two votes :)

[05:56:00.0000] <annevk>
http://lists.w3.org/Archives/Public/ietf-http-wg/2007OctDec/0409.html

[06:01:00.0000] <MikeSmith>
annevk - I commented over on #html-wg about your link pointing to XForms WG aiming at Turing completeness

[06:01:01.0000] <MikeSmith>
If all working groups were expected/required to working under the expectation that all major features of their specs should be explicitly mentioned in their charters, I wonder what the verdict would be on the XForms WG resolution to add a capability for user-defined functions.

[06:01:02.0000] <MikeSmith>
looking at the XForms charter, almost none of its actual features are mentioned in the scope section or anywhere else in that charter at all.

[06:02:00.0000] <annevk>
yeah, dunno

[06:03:00.0000] <annevk>
I thought part of the idea of XForms was that it was not turing complete so that you can "reason" about it or something

[06:13:00.0000] <MikeSmith>
annevk - I think it's great for it to evolve along to meet whatever is needed by the people who are actually using it or want to use it. But I think it would be hard for anybody to claim that the group's charter gives an accurate description of the scope of what XForms has evolved into or what they seem to be further evolving it into.

[06:14:00.0000] <MikeSmith>
If that charter were held to the same standard that some seem to expect the HTML WG charter to be held to, it would need to be a lot more explicit about particular features than it is now. 

[08:27:00.0000] <gsnedders>
nice emails on http-wg

[08:27:01.0000] <gsnedders>
"Specifying behaviour for handling every conceivable error isn't possible"

[08:28:00.0000] <takkaria>
surely if you can conceive of an error then you can say how you should handle it?

[08:28:01.0000] <gsnedders>
takkaria: oh, you just define what happens for any possible next byte.

[08:28:02.0000] <gsnedders>
like HTML5 :P

[08:29:00.0000] <gsnedders>
But I must remember seeming I'm doing it for HTTP outwith the WG that it isn't possible.

[08:29:01.0000] <Philip`>
takkaria: I can conceive of every positive integer, but I can't necessarily say how every positive integer should be handled since they might each need different handling

[08:31:00.0000] <gsnedders>
I guess if it is infinite then you can't

[08:31:01.0000] <gsnedders>
(if you don't use a state machine)

[08:31:02.0000] <gsnedders>
(or something else that can run forever)

[08:35:00.0000] <Philip`>
If you're willing/able to define your error handling in a finite set of rules, then that's alright, but that's a special case and doesn't apply in general

[08:39:00.0000] <gsnedders>
for HTTP you can definitely define error handling, though not always for what is transmitted over it

[09:55:00.0000] <gsnedders>
Hixie: http://ln.hixie.ch/?start=1047614078&count=1 — a perfectly linear one which she climbs quicker and quicker? or maybe she gains mass?

[09:57:00.0000] <Philip`>
Or a mountain which is less steep near the summit, so you go faster and have larger momentum as you climb?

[09:59:00.0000] <hdh>
/me lol @ the twitter status

[10:01:00.0000] <gsnedders>
Philip`: but if it's steeper getting up to that less steep part, your average velocity would surely be lower?

[10:02:00.0000] <gsnedders>
because you'd waste so much time on the steep part

[10:02:01.0000] <Philip`>
gsnedders: Why is average velocity relevant?

[10:03:00.0000] <Philip`>
(Also, can I be pedantic and point out that average speed is more useful than average velocity, particularly if you're going up then back down the same side of the mountain?)

[10:04:00.0000] <gsnedders>
Philip`: momentum = mass * average velocity

[10:06:00.0000] <Philip`>
momentum = instantaneous mass * instantaneous velocity :-p

[10:06:01.0000] <gsnedders>
/me headdesks

[10:06:02.0000] <gsnedders>
that's true.

[10:07:00.0000] <gsnedders>
which means the fact it is velocity is irrelevant.

[10:08:00.0000] <Philip`>
That's relevant since it means momentum is a vector quantity

[10:08:01.0000] <gsnedders>
and of course we're comparing it

[10:10:00.0000] <Philip`>
We're implicitly considering the magnitude of it before comparing car vs mountain-climber, I guess

[10:11:00.0000] <gsnedders>
Philip`: but if it is less steep near the summit, surely she'd need to move really quickly to keep the momentum growing

[10:12:00.0000] <Philip`>
She could be going really slowly on the steep bits, and then just start moving slightly quicker as it becomes continually less steep - there's no need to ever go really quickly

[10:12:01.0000] <gsnedders>
but comparatively

[10:14:00.0000] <Philip`>
She should start off at 30 kg m s^-1 where it's really steep near the start, and then end up at 30.01 kg m s^-1 towards the end where it's much less steep, which would still be gaining momentum but wouldn't be comparatively really quick

[10:15:00.0000] <gsnedders>
but to compensate for the difference in direction?

[10:22:00.0000] <gsnedders>
I'm sure enough of us could come up with mad theories about how it works out :)

[14:23:00.0000] <Lachy__>
http://blog.fawny.org/2007/12/23/janefonda/

[14:23:01.0000] <Lachy__>
That's more complaints about HTML5 accessibility issues

[14:28:00.0000] <takkaria>
that article is high on rhetoric and appeals to emotion using a bad analogy

[14:29:00.0000] <gsnedders>
oh, I've seen worse analogy.

[14:30:00.0000] <takkaria>
I don't believe I was disputing that. :)

[14:30:01.0000] <takkaria>
the last paragraph is amusing too

[14:30:02.0000] <takkaria>
all the bits of HTML4 and XHTML1 that don't work very well now will continue not working very well into the future. :)

[14:31:00.0000] <gsnedders>
For example: a potential divider is somehow related to a ladder coming out of water.

[14:35:00.0000] <csarven>
"benevolent dictator, Hixie, with Maciej as court jester" woah

[14:35:01.0000] <gsnedders>
well, Hixie is a benevolent dictator :P

[14:36:00.0000] <csarven>
/me wonders if he ever gives it a break with the complaints and bashing

[14:36:01.0000] <gsnedders>
/me takes a sausepan, and bashes csarven about over the head

[14:36:02.0000] <gsnedders>
sorry, I had to

[14:37:00.0000] <takkaria>
he says benevolent dictator as if it's a bad thing

[14:37:01.0000] <Philip`>
gsnedders: That doesn't make no sense - the (gravitational) potential difference between the top of the ladder and the water, and the water and the bottom of the ladder, is analogous to the electrical potential difference in the divider, e.g. the two differences add up to a constant but you can move things up and down within a certain range

[14:37:02.0000] <Philip`>
It's not an especially great analogy, since it doesn't give any insight into additional behaviours, but it's not like it's totally wrong :-)

[14:37:03.0000] <gsnedders>
Philip`: yeah, I did get it.

[14:38:00.0000] <gsnedders>
Philip`: But how my teacher explained it left me more confused than before he started.

[14:41:00.0000] <Philip`>
I liked the idea of coulomb beetles (charge) walking through pipes (circuits), with each beetle carrying a backpack with some number of Mars bars (voltage) of which some get removed when passing through components in the circuit

[14:44:00.0000] <ray>
better than a malevolent dictator

[14:44:01.0000] <gsnedders>
Philip`: that's nice

[15:02:00.0000] <othermaciej_>
csarven: who declared me court jester? that's kinda cool :-)

[15:03:00.0000] <gsnedders>
othermaciej_: http://blog.fawny.org/2007/12/23/janefonda/ 

[15:03:01.0000] <gsnedders>
othermaciej_: not a good context :P

[15:04:00.0000] <othermaciej_>
dictators don't have jesters

[15:07:00.0000] <othermaciej>
wow, I'm part of the politburo

[15:09:00.0000] <takkaria>
oh, a bad metphor to complement the bad analogy :P

[15:09:01.0000] <othermaciej>
that was a pretty childish post

[15:10:00.0000] <othermaciej>
"add longdesc... for the children! the poor disabled children!"

[15:10:01.0000] <takkaria>
or they'll have to hack the system

[15:12:00.0000] <othermaciej>
Joe seems to think accessibility is too important to apply normal standards of evidence for what features are truly effective

[15:12:01.0000] <othermaciej>
I think it's too important *not* to

[15:12:02.0000] <Lachy_>
I didn't get the analogy Joe was trying to make using the story about the airline passenger in a wheelchair

[15:13:00.0000] <othermaciej>
he's trying to say that it's the working group's moral duty to do all required research to justify accessibility features

[15:13:01.0000] <othermaciej>
(perhaps unaware of how much detailed research has been done)

[15:13:02.0000] <Lachy_>
of course we do

[15:14:00.0000] <othermaciej>
(and perhaps assuming any research that doesn't justify every single one of HTML4.01's features that are "for accessibility" is prima facie invalid because the experts don't agree)

[15:15:00.0000] <Lachy_>
we need research to find out what accessibility features actually work and are benefitial to add or retain, not just automatically accept features because they are for accessibility, especially if those features don't actually work in practice

[15:19:00.0000] <gsnedders>
JAWS implements @longdesc, does it not?

[15:20:00.0000] <webben>
gsnedders: Both JAWS and Window-Eyes implement longdesc.

[15:20:01.0000] <webben>
gsnedders: As do add-ons for Firefox and IE.

[15:20:02.0000] <gsnedders>
webben: the latter is less important for accessibility alone, though, really

[15:21:00.0000] <webben>
gsnedders: I strongly disagree. If you're colorblind, you might well not use a screen reader but still need a longdesc.

[15:21:01.0000] <webben>
ditto with screen magnifier users

[15:21:02.0000] <gsnedders>
that's true

[15:22:00.0000] <webben>
Now I know that HAL doesn't support longdesc and neither does VoiceOver (there's no interface in WebKit; I lost my code which added one and haven't had time to recode it for submission, unfortunately).

[15:22:01.0000] <webben>
I don't know about the current situation with NVDA and Orca.

[15:23:00.0000] <othermaciej>
the main problems people have documented with longdesc are that the vast majority of longdesc attributes out there are wrong in various ways

[15:24:00.0000] <othermaciej>
so following longdesc whenever possible would be actively harmful to the user

[15:24:01.0000] <othermaciej>
some evidence was also presented that many blind users are unaware longdesc even exists

[15:25:00.0000] <webben>
Actually, I think a lot of that documentation shows longdesc being misinterpreted as a string value rather than a URI.

[15:25:01.0000] <othermaciej>
taken together that makes the feature seem somewhat unsuccessful in achieving its intended purpose

[15:25:02.0000] <webben>
Which means it can be used for error correction of alternative text where necessary.

[15:25:03.0000] <othermaciej>
I do not think that's the case

[15:26:00.0000] <othermaciej>
what Hixie found is that many longdescs are links to the image itself, links to the site's top-level page, 404s, irrelevant spam links, stuff like that

[15:26:01.0000] <webben>
The "blind users" being "unaware" is entirely down to the UI and documentation of assistive tech. (And in fact it's quite clearly documented.) Never mind we're talking about tiny sample sizes of people and a tiny sample of sites.

[15:27:00.0000] <webben>
othermaciej: A UA that presented a 404 as a longdesc would be a bit silly.

[15:27:01.0000] <gavin_>
presumably hixie search through his large web archive

[15:27:02.0000] <webben>
I don't think there's any element or attribute not misused by spammers.

[15:27:03.0000] <gavin_>
I wonder whether the people arguing for its inclusion care about "the web"

[15:27:04.0000] <gavin_>
rather than "the web they frequently use"

[15:28:00.0000] <webben>
I also pointed Hixie to various examples of longdesc used correctly.

[15:28:01.0000] <webben>
(real examples from government websites for example)

[15:29:00.0000] <webben>
I think arguing from what AT currently does is highly problematic, given how UAs are trying to hack around broken web content, itself highly determined by flaws in normal UAs.

[15:29:01.0000] <webben>
*given how ATs

[15:29:02.0000] <webben>
it's not a technical argument really.

[15:30:00.0000] <webben>
and in any case Window-Eyes and JAWS do in fact support it.

[15:31:00.0000] <othermaciej>
webben: well, you don't know it's gonna be a 404 until the user requests to follow it

[15:31:01.0000] <webben>
othermaciej: It's a longdesc not an anchor or form control. It should be safe for a HEAD request at least.

[15:32:00.0000] <othermaciej>
webben: I suppose you could pre-emptively load all longdesc contents to decide whether to even show the UI to the user, however, that's not what we do with normal hyperlinks

[15:32:01.0000] <othermaciej>
HEAD requests are not reliable

[15:32:02.0000] <othermaciej>
many servers out there give mismatched HEAD results

[15:33:00.0000] <takkaria>
from the research I did, I found one longdesc being used right in Philip's directory

[15:33:01.0000] <webben>
othermaciej: Well there's an argument for doing that for normal hyperlinks (preloading) ... particularly ones where alt is missing (image links). The main problem is the misuse of GET requests for destructive actions, although the RoR fiasco with web accelerator may have reduced that abuse. But I very much doubt the same problem exists at any significant scale with longdesc URIs.

[15:34:00.0000] <takkaria>
out of 62 pages that used longdesc

[15:34:01.0000] <webben>
takkaria: Yeah, I found more examples than "1".

[15:34:02.0000] <webben>
(not in Philip's directory, though)

[15:34:03.0000] <takkaria>
fair enough.  if you cite government websites, though, then that's hardly regular usage

[15:36:00.0000] <othermaciej>
webben: most longdesc attributes do not point to longdesc-specific content, so I don't know if it would be materially different (other than the fact that longdesc in general is so rare)

[15:37:00.0000] <Lachy_>
Most of the good examples of long descriptions include a regular link, making longdesc redundant

[15:38:00.0000] <Lachy_>
and the information often includes information that would be useful for more than just those with assistive technology

[15:39:00.0000] <webben>
Lachy_: There's no reason why longdesc shouldn't be in primary UI.

[15:39:01.0000] <Lachy_>
the best solution I've seen for the longdesc issue is rel="longdesc"

[15:39:02.0000] <webben>
Lachy_: That's not very helpful since it can't be programmatically associated with the image.

[15:39:03.0000] <Lachy_>
it can be when used within a <figure>

[15:40:00.0000] <webben>
what about when it's not in a figure?

[15:40:01.0000] <webben>
and what about when your photoshop-mock maker complains about the visible longdesc link?

[15:41:00.0000] <webben>
visible associations are ideal, but they are problematic in the real world.

[15:41:01.0000] <webben>
(witness the microformat's movements endless trouble with "visible" yet "hidden" data)

[15:41:02.0000] <Lachy_>
hidden metadata is more problematic in the real world, and londesc="" simply doesn't work in practice

[15:41:03.0000] <othermaciej>
the link anchor can be around the image

[15:41:04.0000] <webben>
It /works/ fine. Saying it doesn't work is like saying alt doesn't work because some people write garbage alt text.

[15:42:00.0000] <webben>
othermaciej: Not if the image is a link already.

[15:42:01.0000] <othermaciej>
if it is a link already, then likely a separate longdesc link is unnecessary

[15:43:00.0000] <webben>
othermaciej: For example. An image of painting of Lady of Shalott linking to the poem. A long description for the painting.

[15:43:01.0000] <Lachy_>
since authors can always include a <figure> around the image to explicitly associate the <a rel="longdesc" ...>, the case where there is no figure could be considered an error. But that doesn't prevent any heurisitics from making a reasonable guess about which image it's associated with

[15:43:02.0000] <othermaciej>
I dunno, I don't really have a horse in the longdesc race

[15:44:00.0000] <webben>
Lachy_: I think you're wildly underestimating how difficult authors make it to write effective heurestics for that sort of case.

[15:44:01.0000] <othermaciej>
however, it seems to be the case that very few (less than 1 in 100,000) pages use it, and of those, at best 1 in 1000 use it correctly (that's just eliminate longdescs that simple testing can show you are wrong)

[15:44:02.0000] <othermaciej>
those are pretty poor odds

[15:44:03.0000] <webben>
I know AT have had enormous trouble associating fields with labels; I know from experience it's difficult even to conclusively associate permalinks with blog articles.

[15:44:04.0000] <Lachy_>
webben, there's still the issue that longdesc doesn't work in practice. Authors rarely use properly, and when they do, they most often include a redundant link anyway

[15:45:00.0000] <webben>
Lachy_: "work in practice" and "is widely used" are completely different things.

[15:45:01.0000] <webben>
Especially when even accessibility resources have long continued under an apprehension that longdesc isn't supported.

[15:45:02.0000] <othermaciej>
it's not just rarely used - the rare times it is used, it is usually used wrong

[15:46:00.0000] <othermaciej>
(at least that's what the stats I have seen seem to show)

[15:46:01.0000] <webben>
Even that isn't the same thing as "not work in practice".

[15:46:02.0000] <Lachy_>
given that it's supported with extremely bad UI, treating it as if it's not supported makes little difference

[15:46:03.0000] <othermaciej>
I guess it depends on how you define work

[15:46:04.0000] <webben>
e.g. if responsible authors assume it doesn't work, then you'd /expect/ a higher proportion of misuse by spammers

[15:46:05.0000] <othermaciej>
I would assume the goal of the feature is to convey useful information to the blind and visually impaired

[15:47:00.0000] <othermaciej>
so far it does not seem to have become an effective way to do that

[15:47:01.0000] <webben>
othermaciej: That's not exactly the stated rationale in the HTML4 spec.

[15:48:00.0000] <webben>
although it's certainly a goal.

[15:48:01.0000] <othermaciej>
this would lead me to at least wonder if other ways of associating a long description might be better

[15:48:02.0000] <webben>
othermaciej: It's perfectly effective if you let people know that you're using longdesc and they have supporting AT or an add-on.

[15:49:00.0000] <webben>
Although again I'd tend to say that OBJECT's nesting model is a cleaner solution.

[15:49:01.0000] <othermaciej>
it's true, HTML 4.01 says "This attribute specifies a link to a long description of the image", but I think it has to be assumed it's not intended for general audiences because then a normal link would be better

[15:50:00.0000] <webben>
othermaciej: Yeah but it's emphatically not just targeted at disabled users.

[15:50:01.0000] <webben>
note the rationale given for ALT: http://www.w3.org/TR/html4/struct/objects.html#adef-alt

[15:53:00.0000] <webben>
In an ideal world, it would always be good to give authors the opportunity to express accessibility information/metadata visibly. In the real world, publishers are never going to always choose to do so. Therefore there will always be a set of publishers willing to express accessibility information/metadata "invisibly" (to some degree) and not visibly.

[15:55:00.0000] <takkaria>
that set seems to be very small today, though

[15:55:01.0000] <webben>
So while I think the "hidden metadata is evil" argument is a powerful argument for allowing ways to explicitly markup visible content as being a genuine equivalent for some other bit of content, I don't think it's a real-world argument for preventing authors providing "invisible" equivalents/metadata.

[15:56:00.0000] <webben>
takkaria: On the contrary, it's vast. ALT is one of the more successful accessibility features. Microformats have been surprisingly successful, given that while they preach visible metadata they practice a lot of invisible metadata.

[15:56:01.0000] <webben>
(a lot of key data is buried in TITLE attributes)

[15:57:00.0000] <webben>
I can't imagine a solution where to provide text for a image button you would have needed to make it visible would have been half as successful.

[15:57:01.0000] <webben>
It certainly doesn't parallel the model adopted for desktop UIs.

[15:57:02.0000] <takkaria>
true, but given that many people are willing to use alt text, the fact that most of those same people aren't willing to use longdesc, means that the number of people who actually want to provide longdesc must be fairly small

[15:58:00.0000] <webben>
takkaria: Again clearly false, given alt is /heavily/ evangelized, and longdesc, where mentioned at all by evangelists, has generally being accompanied by caveats about it not being supported.

[15:58:01.0000] <Dashiva>
alt is just text. longdesc requires a whole webpage.

[15:59:00.0000] <webben>
Dashiva: Yes. Which is the advantage of the OBJECT model for equivalent content.


2007-12-24
[16:00:00.0000] <Dashiva>
Doesn't that just reduce to using <object> for images?

[16:00:01.0000] <webben>
Dashiva: Although OTOH it provides reuse potential in practice.

[16:00:02.0000] <webben>
Dashiva: It's generally said that OBJECT is too broken now.

[16:00:03.0000] <webben>
and we can't change the content model for IMG, and browsers implement all similar words as though they were images

[16:00:04.0000] <Dashiva>
I seem to recall IE8 including image-object fixes for their acid2 run

[16:00:05.0000] <webben>
which is where we get proposals for things like PICTURE

[16:01:00.0000] <webben>
but PICTURE is supported by nothing, whereas AT already supports longdesc

[16:01:01.0000] <webben>
*some AT

[16:01:02.0000] <webben>
*as though they were the IMG element, I should say.

[16:02:00.0000] <webben>
Dashiva: Yeah, part of the problem here is AT needs to have (if anything) more backwards compat than anything else, since the upgrade time for AT is much longer than the upgrade time for ordinary browsers.

[16:03:00.0000] <Dashiva>
Well, that's not a given 

[16:03:01.0000] <webben>
Dashiva: It's a reasonable assumption. To assume anything else is to gamble on Orca and NVDA being huge successes.

[16:03:02.0000] <webben>
(basically)

[16:04:00.0000] <Dashiva>
Or gamble on AT upgrading being non-broken

[16:04:01.0000] <webben>
It's like assuming from the start that Firebird was going to win the browser wars ... actually worse, since AT is a much more difficult industry to crack

[16:05:00.0000] <webben>
Dashiva: A lot of "AT" brokenness is a result of: poor accessibility frameworks (MSAA was pretty useless by most accounts), broken browsers (forcing AT to do the browser's job), and having to cope with real-world broken web content.

[16:05:01.0000] <Dashiva>
That's not part of the horrible upgrade cycle, though

[16:05:02.0000] <webben>
and equally important real-world avant-garde web content.

[16:06:00.0000] <Dashiva>
Every time AT comes up in a discussion, we're reminded that <large number> of users still use <version current-4>

[16:06:01.0000] <webben>
Dashiva: given the target is moving all the time (new browser bugs cascading to AT, new forms of broken web content, etc.), it is in a way.

[16:06:02.0000] <webben>
Dashiva: Yes.

[16:06:03.0000] <webben>
Dashiva: That's more to do with expense though.

[16:07:00.0000] <Dashiva>
And that's a problem

[16:07:01.0000] <webben>
And way outside the ability of the HTML WG/web community generally to do much about.

[16:07:02.0000] <Dashiva>
It's within the ability of AT companies, though

[16:07:03.0000] <webben>
(Well, other than inventing new accessible web-based apps which kill MS Office etc.)

[16:07:04.0000] <webben>
Dashiva: It's not necessarily within the ability of FS and GW-Micro to make their software cheaper without going out of business.

[16:08:00.0000] <webben>
(Not that I have any special insight into their finances.)

[16:08:01.0000] <Dashiva>
As a Opera guy I'm all for accessibility, but I do get somewhat frustrated when we're asked to both run accord for AT software being buggy -and- for refusing to allow their users to upgrade

[16:09:00.0000] <webben>
Dashiva: Well, they do tend to make the intra-version upgrades free.

[16:09:01.0000] <webben>
Dashiva: And they are (I think) making some efforts to make the upgrade path less expensive.

[16:10:00.0000] <webben>
GW-Micro introduced some sort of monthly payment scheme which included upgrades IIRC.

[16:11:00.0000] <webben>
Dashiva: No disagreement that it makes it difficult for the browser and webdev community ... but we are really only a part of what they have to deal with.

[16:12:00.0000] <Dashiva>
Well, we're all here to cooperate and compromise, what are they offering? :)

[16:12:01.0000] <webben>
Offering for what?

[16:12:02.0000] <webben>
(Or what for what?)

[16:13:00.0000] <Dashiva>
In contributions

[16:13:01.0000] <webben>
what sort of contributions?

[16:14:00.0000] <Dashiva>
Activity. In the working group

[16:16:00.0000] <webben>
Dashiva: Well, I'd certainly like to see more involvement (although HTML WG already has three AT vendors). But (to some degree), it's the browser makers' job to make sure they get a language that can be represented in the OS-level accessibility framework (just like with ODF etc.). And in the case of MS and Apple, both on the WG, it's very much their job to improve the framework where necessary.

[16:17:00.0000] <webben>
And involvement with HTML WG would probably mean less bugfixes.

[16:18:00.0000] <webben>
These companies aren't the size of Google, MS, Microsoft, Apple, after all. They don't have proficient developers to spare.

[16:18:01.0000] <Dashiva>
Strictly feaking, bugfixes don't really matter because we still have to take the old versions into considerations :)

[16:19:00.0000] <webben>
Dashiva: Well, okay, raising the cost of bugfixes and hence the cost of upgrading.

[16:19:01.0000] <othermaciej>
I have some influence over what Apple does with VoiceOver and its integration with the browser

[16:19:02.0000] <othermaciej>
but whatever I do, I can't make things work better in JAWS

[16:20:00.0000] <othermaciej>
apparently very few blind people use macs, although I'm told the experience is pretty good

[16:20:01.0000] <webben>
othermaciej: It is pretty good! (although WebKit has some areas to improve as we've discussed before).

[16:20:02.0000] <webben>
/me subscribes to the macvisionaries mailing list, so he hears a lot of user feedback.

[16:21:00.0000] <webben>
Hopefully Office 2008 will use the accessibility framework, that would be a big boost. Although I'm not that hopeful that it will.

[16:23:00.0000] <webben>
othermaciej: I agree you can't do much about JAWS yet (although that will change when WebKit gets an MSAA/IAccessible2 implementation).

[16:23:01.0000] <webben>
But there are at least two people with experience of JAWS as normal end-users ... and there's MS as a sort of MSAA representative.

[16:24:00.0000] <webben>
Don't get me wrong; would like to see an FS rep too!

[16:24:01.0000] <webben>
Isn't Aaron Leventhal on the WG now?

[16:25:00.0000] <webben>
Yeah he is (http://www.w3.org/2000/09/dbwg/details?group=40318&public=1) ... he's regularly bug-fixing Moz. for JAWS.

[16:31:00.0000] <weinig>
does any one know if there is a test suite for the Selectors API yet?

[16:32:00.0000] <weinig>
/me realizes this is not a whatwg spec and feels silly

[16:44:00.0000] <dglazkov>
they shoot misanthropes, don't they?

[16:47:00.0000] <Dashiva>
They do whatever they is defined to be does

[16:48:00.0000] <webben>
misanthrope shooters definitely do ;)

[16:49:00.0000] <dglazkov>
Dashiva, don't forget the definition of "is" :)

[17:33:00.0000] <othermaciej>
webben: we definitely take requests for accessibility fixes in WebKit, of course

[17:33:01.0000] <othermaciej>
weinig: you could ask on #webapi on this server, or ask Lachy or anne, but I don't think there is an official test suite

[17:34:00.0000] <weinig>
/me nods

[17:34:01.0000] <othermaciej>
weinig: however, that spec is going into Last Call, I would suggest you should review and get others (Catfish_Man, maybe dethbakin and mitz) to try to review too

[17:34:02.0000] <weinig>
othermaciej: I was planning on it

[17:34:03.0000] <webben>
othermaciej: Yeah I know (I've logged some). I need to find time to do some hacking myself for it too. :)

[17:35:00.0000] <othermaciej>
webben: that is welcome as well

[17:35:01.0000] <othermaciej>
I need to schedule a meeting with the VoiceOver people to talk about ARIA

[17:36:00.0000] <othermaciej>
lemme know if you have other things that might be worth discussing with or mentioning to them

[17:36:01.0000] <othermaciej>
I think a lot of HTML5 elements will help with accessibility

[17:36:02.0000] <othermaciej>
proper use of <section> / <article> could allow for better navigation

[17:36:03.0000] <othermaciej>
I dunno how focused they are on accessibility in the browser though, relative to native apps

[17:38:00.0000] <webben>
othermaciej: From a webkit perspective, I'd say making navigation from bit to bit around the page more predictable is the most important (but unfortunately probably the trickiest) thing.

[17:38:01.0000] <webben>
othermaciej: Then there are some showstopper bugs like multiple selects being broken to the extent that they can't be read.

[17:38:02.0000] <webben>
Table support of some sort would be good.

[17:39:00.0000] <webben>
Another thing would be the ability to skip links.

[17:39:01.0000] <webben>
VO users sometimes complain about not being able to get to the meat of pages quick enough.

[17:39:02.0000] <othermaciej>
I dunno what you mean about the multiple selects

[17:39:03.0000] <webben>
I think being able to skip groups of links (as you can with JAWS and WE)

[17:39:04.0000] <webben>
would help with that

[17:40:00.0000] <webben>
(Obviously HTML5 nav could help in the future too.)

[17:40:01.0000] <webben>
othermaciej: http://bugs.webkit.org/show_bug.cgi?id=15817

[17:41:00.0000] <webben>
(for the multiple select bug)

[17:42:00.0000] <webben>
othermaciej: Skipping links would also be good for non-VO users (keyboard/switch users)

[17:43:00.0000] <othermaciej>
webben: what defines a "group of links"?

[17:44:00.0000] <webben>
othermaciej: Good question :) In the case of WE and JAWS, you can set a threshold number of non-link characters between links for something to still count as a group of links (I assume that's to cope with pipe characters used in navigation)

[17:45:00.0000] <othermaciej>
"you can set a threshold" would probably not be an acceptable answer for Mac OS X human interface

[17:45:01.0000] <webben>
so say you set 3 as a threshhold ... <a href="...">foo</a> | <a href="...">baz</a> is part of a group but <a href="...">foo</a> and also <a href="...">baz</a> is not.

[17:45:02.0000] <othermaciej>
but series of links w/ no more than some number of characters in between is sorta reasonable

[17:45:03.0000] <webben>
othermaciej: Well you could devise some sort of algorithm to do something similar I guess.

[17:45:04.0000] <othermaciej>
I guess we could have a shortcut for that when tab to links is on

[17:46:00.0000] <webben>
othermaciej: "tab to links" just switches the keybindings round

[17:46:01.0000] <webben>
it doesn't actually change functionality

[17:46:02.0000] <webben>
so I'm not sure why one would restrict the functionality to one keybinding scenario or another

[17:47:00.0000] <webben>
(especially as that would complicate further a feature I've already had to attempt to explain more than once on macvisionaries)

[17:47:01.0000] <othermaciej>
I don't think it makes sense to have a "tab skipping group of links" shortcut when not in "tab to links" mode

[17:48:00.0000] <othermaciej>
(can't remember if full keyboard access still enables tabbing to links by default)

[17:48:01.0000] <webben>
othermaciej: Why not have an option tab skipping group of links

[17:48:02.0000] <webben>
when in option tab to links mode

[17:48:03.0000] <webben>
othermaciej: I don't /think/ it does. I'm pretty sure I have to set that manually.

[17:48:04.0000] <othermaciej>
well, it would have to be an obscure shortcut

[17:49:00.0000] <othermaciej>
if you have one to suggest we can consider it

[17:49:01.0000] <webben>
othermaciej: With JAWS you just press N

[17:50:00.0000] <othermaciej>
tab, option-tab, shift-tab and shift-option-tab are all taken

[17:50:01.0000] <webben>
granted you can't use N in the main UI though!

[17:50:02.0000] <othermaciej>
I don't think we'd add a keyboard shortcut on an unmodified key in the default UI

[17:50:03.0000] <webben>
yep ... my point is more that obscure is not necessarily better

[17:51:00.0000] <webben>
othermaciej: I suppose one thing to discuss is whether you could use a VO-prefixed shortcut (ctrl option) in the main UI space

[17:51:01.0000] <othermaciej>
the real problems are first, having a good shortcut, second figuring out how to make it reasonable discoverable, and third, figuring out what to call the feature

[17:51:02.0000] <othermaciej>
"skip group of links" is not a very clear name

[17:51:03.0000] <othermaciej>
since the default setting in the normal UI is that focus navigation skips all links

[17:51:04.0000] <webben>
othermaciej: It's a lot clearer than what the WE manual calls it!

[17:51:05.0000] <webben>
"Skip to text" IIRC

[17:52:00.0000] <webben>
othermaciej: Skip navigation might be better.

[17:52:01.0000] <webben>
(since that's the basic idea)

[17:52:02.0000] <othermaciej>
so this is supposed to be a heuristic to skip what are intended to be groups of navigation links?

[17:52:03.0000] <othermaciej>
to get to the main content faster?

[17:53:00.0000] <othermaciej>
I think a heuristic for voiceover to skip to main content might be the best solution for this use case

[17:53:01.0000] <webben>
othermaciej: That's the main purpose (although you can also have subsidiary groups of links scattered through the page)

[17:53:02.0000] <webben>
e.g. tag links following a blog entry

[17:54:00.0000] <webben>
(or social media sharing links etc.)

[17:55:00.0000] <webben>
I'd say getting to the main content is roughly 70% of the need.

[17:56:00.0000] <webben>
(very roughly)

[17:56:01.0000] <webben>
Getting to the main content may be even harder to create a reliable algorithm for, however.

[18:04:00.0000] <webben>
othermaciej: Another thing would be fixing the context menu so it can be operated with VO as in other apps.

[18:04:01.0000] <webben>
one of the reasons for that is so that users can add bookmarklets easily

[18:05:00.0000] <webben>
(which requires activating the context menu and selecting Add Link

[18:05:01.0000] <webben>
... that's another thing that came up on macvisionaries

[18:05:02.0000] <webben>
(sorry this list is a bit random)

[18:05:03.0000] <webben>
there's a bug filed for context menu and it has a radar

[18:07:00.0000] <webben>
othermaciej: oh another big thing: fix read all, so people can sit back, relax and have a webpage read to them :) http://bugs.webkit.org/show_bug.cgi?id=15487

[18:14:00.0000] <webben>
othermaciej: This was another bug that came out of discussions on macvisionaries: http://bugs.webkit.org/show_bug.cgi?id=15310

[18:14:01.0000] <webben>
(allow folks to select ranges of text with the keyboard)

[18:15:00.0000] <othermaciej>
does VoiceOver have a standard mechanism for keyboard selection of non-editable text?

[18:16:00.0000] <webben>
othermaciej: Not sure. Where else do you find non-editable documents that you'd want to copy?

[18:16:01.0000] <othermaciej>
Mail

[18:17:00.0000] <othermaciej>
iChat

[18:17:01.0000] <webben>
don't those both use WebKit for display too?

[18:17:02.0000] <othermaciej>
IRC clients

[18:17:03.0000] <othermaciej>
ironically all these things tend to use WebKit

[18:17:04.0000] <othermaciej>
but that's an implementation detail

[18:18:00.0000] <othermaciej>
many alerts also have the error text selectable

[18:18:01.0000] <othermaciej>
iChat didn't use WebKit before Leopard

[18:20:00.0000] <webben>
othermaciej: As far as I can tell, in Tiger, text is selectable in a plain text msg in Mail but not an HTML msg.

[18:21:00.0000] <othermaciej>
webben: really?

[18:21:01.0000] <webben>
but I may be missing something

[18:21:02.0000] <othermaciej>
how do you start a selection with the keyboard?

[18:21:03.0000] <othermaciej>
(or do you just mean that selection modification shortcuts work once you have a mouse selection)

[18:21:04.0000] <webben>
othermaciej: you get a normal character cursor when you focus on the plain text msg area

[18:21:05.0000] <webben>
so you can use shift

[18:21:06.0000] <webben>
shift + arrow keys

[18:23:00.0000] <webben>
othermaciej: I think you need VO on for that to work though

[18:24:00.0000] <othermaciej>
ah

[18:24:01.0000] <othermaciej>
probably, yes

[18:31:00.0000] <webben>
othermaciej: Thinking about it, groups mode provides some of that skipping functionality already (or it would if VO announced "navigation group" rather than just "group") ... it just doesn't handle the jumping to main content bit.

[22:18:00.0000] <MikeSmith>
webben - you still around?

[22:19:00.0000] <MikeSmith>
/me sees that webben mentioned "although HTML WG already has three AT vendors" and wonders which three AT vendors that might be

[22:20:00.0000] <othermaciej>
I assume Apple is one

[22:23:00.0000] <MikeSmith>
othermaciej - If that's what he meant, I guess MS would be one also.

[22:30:00.0000] <othermaciej>
I guess Vista includes a built-in screenreader

[22:31:00.0000] <othermaciej>
not sure how it compares to JAWS

[22:31:01.0000] <othermaciej>
Apple ships VoiceOver, not just accessibility APIs

[22:46:00.0000] <MikeSmith>
I wonder how much direct participation WAI has had from AT implementors

[22:48:00.0000] <MikeSmith>
As far as I can see, no one from Freedom Scientific has ever participated in WAI

[22:49:00.0000] <MikeSmith>
would be great to have them in the HTML WG but if they've never bothered to get involved in WAI, I guess we shouldn't hold our breath

[06:37:00.0000] <gsnedders>
Do any UAs already support @rel= noreferrer?

[06:42:00.0000] <krijn>
gsnedders: re http://www.russellbeattie.com/blog/i-just-added-twitter-into-my-feed#comment-45594 - wouldn't that just become ]]&gt; in a feed?

[06:43:00.0000] <gsnedders>
krijn: sure, but any unescaped ampersand or less-than sign after?

[06:43:01.0000] <krijn>
Unescaped?

[06:44:00.0000] <gsnedders>
<title type="html"><![CDATA[ my broken tweet ]]> & test</title>

[06:44:01.0000] <gsnedders>
sorry…

[06:44:02.0000] <gsnedders>
<title type="html"><![CDATA[ my broken tweet ]]> & test]]></title>

[06:44:03.0000] <gsnedders>
yeah, the actual second ]]> is fine. But the ampersand is a well-formness error.

[06:44:04.0000] <krijn>
Same question; wouldn't that just become &amp; test .. ? :)

[06:45:00.0000] <gsnedders>
no, it's a fatal error in XML

[06:46:00.0000] <krijn>
<title type="html"><![CDATA[ my broken tweet ]]&gt; &amp; test]]></title>

[06:46:01.0000] <krijn>
That would just work, right?

[06:46:02.0000] <gsnedders>
no

[06:46:03.0000] <krijn>
Ow

[06:46:04.0000] <gsnedders>
yeah, that'd work

[06:46:05.0000] <gsnedders>
but it leaves a raw & in the HTML :P

[06:47:00.0000] <krijn>
Wasn't it there in the first place?

[06:47:01.0000] <krijn>
/me is too stupid to get this ;)

[06:47:02.0000] <gsnedders>
everything that SP outputs is fine for embedded in HTML: Atom uses escaped HTML.

[06:47:03.0000] <gsnedders>
s/for/to be/

[06:48:00.0000] <gsnedders>
using htmlspecialchars() would've been better than using CDATA sections

[06:48:01.0000] <krijn>
I never use CDATA sections, so perhaps that's why I don't get it :)

[06:48:02.0000] <gsnedders>
Don't use them.

[06:48:03.0000] <gsnedders>
:)

[06:49:00.0000] <gsnedders>
http://hsivonen.iki.fi/producing-xml/#cdata

[06:49:01.0000] <krijn>
Yeah, that's why I never use them ;)

[06:51:00.0000] <takkaria>
argh, this house is so christmass

[06:57:00.0000] <Philip`>
I saw some shops convert their Halloween displays immediately into Christmas displays - I think we need to invent a new holiday in mid-December to take some of the attention away, since otherwise there's too much build-up to Christmas and one gets fed up by the time it's actually here

[06:58:00.0000] <takkaria>
heh

[06:58:01.0000] <takkaria>
I'm up for that

[06:58:02.0000] <takkaria>
mind you, next year I think there's agreement amongst my siblings to have a non-commercial christmas

[07:01:00.0000] <gsnedders>
Philip`: Oh, Tesco was selling stuff for Halloween in August

[07:50:00.0000] <Philip`>
"P.S. I might turn this into a blog message because hyperlinks and plain text don't mix well." - surely that's what HTML email is for? :-)

[07:50:01.0000] <Philip`>
(I don't actually know how common or useful HTML email is, since I get Thunderbird to convert everything to plain text...)

[07:56:00.0000] <takkaria>
hyperlinks mix fine

[07:56:01.0000] <takkaria>
just use [1] and [2]

[08:01:00.0000] <gsnedders>
wow. it seems even spam bots are on holiday!

[08:04:00.0000] <jgraham>
/me finds Ben's citation style in email rather hard to read

[08:06:00.0000] <jgraham>
Specifically things like "Joe's [previous] table collection for

[08:06:01.0000] <jgraham>
PDF/UA was my starting point" make me think that [previous] is some sort of editorial addition, which doesn't make any sense because it's not a quote...

[08:12:00.0000] <takkaria>
gsnedders: oh, I thought I'd finally sorted out my spam filters...

[10:38:00.0000] <takkaria>
nutter alert on public-html

[11:13:00.0000] <kig>
Philip`: ever implemented svg's elliptical arc path segments?

[11:19:00.0000] <Philip`>
kig: Nope

[11:20:00.0000] <kig>
hm, maybe inkscape has an implementation

[11:20:01.0000] <Philip`>
kig: Looks like it just needs a bit of geometry calculation :-)

[11:21:00.0000] <kig>
i like how the elliptical arc implementation notes are as long as all the other implementation notes combined

[11:21:01.0000] <Philip`>
Do you want to emulate it as a series of straight lines, or as a rotated/scaled arc, or something else?

[11:22:00.0000] <kig>
can't use (easily) transformed arcs in opera, so bezier curves

[11:22:01.0000] <Philip`>
/me sees that SVG Tiny 1.2 has removed elliptical arcs

[11:23:00.0000] <annevk>
kig, if something in Opera is causing problems please tell, maybe it's fixable :)

[11:24:00.0000] <Philip`>
annevk: I think the issue is with Opera following the spec and applying transformations when drawing paths, rather than applying them while constructing paths

[11:26:00.0000] <annevk>
hmm

[11:26:01.0000] <annevk>
I think everyone agreed that following the specification for that case would be good, but I guess if Firefox still doesn't do it we're in trouble

[11:27:00.0000] <kig>
safari doesn't do it either afaik

[11:28:00.0000] <annevk>
at some point in the past we checked with Safari and Firefox guys and both agreed to change their impl

[11:30:00.0000] <kig>
for what it's worth, applying transforms to paths at drawing time is just about useless

[11:32:00.0000] <kig>
the one use-case is "draw this path several times with different transforms", but path construction cost pales in comparison with stroking, patterns and gradients, so it's not a very efficient optimization either

[11:32:01.0000] <annevk>
http://bugs.webkit.org/show_bug.cgi?id=13669

[11:34:00.0000] <kig>
(not to mention that you'd need to sort your scene by path at drawing time, which is a pain)

[11:35:00.0000] <annevk>
https://bugzilla.mozilla.org/show_bug.cgi?id=380338

[11:35:01.0000] <annevk>
vlad is somehow convinced we agreed that Firefox / Safari behavior was the way to go

[11:35:02.0000] <annevk>
afaict no such thing happened

[11:37:00.0000] <kig>
alright

[11:38:00.0000] <annevk>
i'm willing to believe the non-spec behavior is more useful, but I'm not sure about getting this fixed in time for Opera 9.5

[11:44:00.0000] <kig>
oh, btw, there's a small bug in Opera 9.5 canvas, full circle arc strokes have a gap with large line widths. i'll make a testcase, sec

[11:46:00.0000] <annevk>
it would be good if <canvas> was updated once again...

[11:46:01.0000] <annevk>
(the spec)

[11:46:02.0000] <annevk>
addressing all of Philip`'s outstanding comments and such

[11:49:00.0000] <kig>
http://glimr.rubyforge.org/cake/tests/arcStroke.html

[11:53:00.0000] <annevk>
thanks

[11:54:00.0000] <annevk>
/me files a bug

[11:54:01.0000] <kig>
thanks

[12:00:00.0000] <annevk>
Philip`, are your tests for transforms aligned with the spec or Firefox and Safari?

[12:09:00.0000] <Philip`>
annevk: We need an HTML55, to fix HTML5 so that it reflects current practice and fixes all the errors related to things like <canvas> :-)

[12:11:00.0000] <Philip`>
annevk: They should match the spec - see http://philip.html5.org/tests/canvas/suite/tests/results.html / 2d.path.transformation.*

[12:39:00.0000] <kig>
hooray, inkscape's arcs-to-beziers code works

[12:45:00.0000] <webben>
MikeSmith: Microsoft: MSAA (backbone of Windows accessibility) + Windows accessibility features and applications (Narrator, a primitive screen reader). Apple: Apple Accessibility API + VoiceOver. T. V. Raman: Emacspeak.

[12:46:00.0000] <annevk>
hmm ok, so if we fix our implementation a good reference would be breaking all Philip`'s tests

[12:47:00.0000] <webben>
MikeSmith: also IBM: various products from the shutdown HPR screen reading software to defunct LSR project and the (currently still vapourware) accessible media player they said they were going to release.

[12:48:00.0000] <webben>
IBM also make ViaVoice; I'm not sure that's under any active development any more either. MS make a speech recognition product which is AT software.

[12:50:00.0000] <Philip`>
annevk: I'd be happy to update my tests if people agree the spec should be changed

[12:58:00.0000] <Philip`>
(I think there's at least several other tests that rely on the spec's transformation behaviour, particularly in relation to patterns, so those would break as well if Opera changes)

[13:45:00.0000] <Philip`>
Is there a way to use prompt() in IE7 without it popping up a warning in the information bar instead?

[13:49:00.0000] <Philip`>
Hmm, it seems the answer is no

[13:50:00.0000] <Philip`>
Is there any reason they effectively removed prompt(), rather than changing it to use a non-standard appearance so it won't get confused with an 'official' browser window and used for UI spoofing?

[14:04:00.0000] <webben>
Well, it's a modal dialog I guess.

[14:09:00.0000] <annevk>
Philip`, ok, cool, I guess I'll know something in January, just ping me around then

[14:12:00.0000] <annevk>
from previous iterations going through this issue it seems it's slightly painful to do it differently

[14:12:01.0000] <annevk>
SVG does it like the <canvas> spec does it now it seems

[14:15:00.0000] <gsnedders>
PHP is so annoying. There's the function, php_charmask(), in the C sourcecode that I need access to in the userland. Ergh. :\

[14:31:00.0000] <gsnedders>
The behaviour is bizarre too.

[14:32:00.0000] <gsnedders>
It takes a character list (like "a..z" representing everything between the two, in terms of ASCII codes), and the error handling…

[14:32:01.0000] <gsnedders>
"a..b..c" parses to array("a", "b", ".", "c")

[15:17:00.0000] <Lachy>
OMG, Dean Edridge is attacking me, Anne and Hixie by private email to me


2007-12-25
[10:23:00.0000] <nickshanks>
is anyone likely to be doing a survey of current element usage on the web?

[10:39:00.0000] <MikeSmith>
nickshanks - you've seen http://code.google.com/webstats/ ?

[10:40:00.0000] <nickshanks>
mike: yes, i am asking if ian or someone like him is going to do an update

[10:42:00.0000] <nickshanks>
i want to add <image></image> to HTML, but people have said that you can't do that, as sites already use that when they mean <img>. But no-one's ever provided any stats to back this up

[10:42:01.0000] <nickshanks>
and ian said he was too busy last time i asked (about mid-2006)

[10:43:00.0000] <MikeSmith>
nickshanks - I see. I think he's since run some one-off reports for specific things. maybe if you remind him

[10:43:01.0000] <MikeSmith>
Philip` runs some now and then

[10:43:02.0000] <MikeSmith>
but on a smaller set of pages

[10:44:00.0000] <nickshanks>
how big a sample? where does he get the sample from?

[10:44:01.0000] <MikeSmith>
At least some of them from dmoz.org I think

[10:44:02.0000] <MikeSmith>
not sure how big a sample

[10:45:00.0000] <MikeSmith>
Shawn Medero (smedero) from the HTML WG has access to a very big index and has talked about setting up some reporting mechanism for that

[10:45:01.0000] <nickshanks>
anyway, i think adding <image></image> and deprecating <img alt=""> would be a step forward, even if only in standards mode

[10:46:00.0000] <nickshanks>
hmm, interesting. i'm not familiar with him

[10:46:01.0000] <MikeSmith>
smedero over on #html-wg on irc.freenode.net

[10:46:02.0000] <MikeSmith>
err

[10:46:03.0000] <MikeSmith>
I mean

[10:46:04.0000] <MikeSmith>
irc.w3.org

[10:46:05.0000] <nickshanks>
yep :) i guessed

[10:47:00.0000] <MikeSmith>
but definitely worth asking Philip` about it too if you can catch him 

[10:48:00.0000] <nickshanks>
do you think this would be a good idea? i draw parallels with the resently introduced <video> and <audio> elelments. previously the idea was shot down by people saying "why not just use <object> ?" but now i have a counterpoint to that

[11:05:00.0000] <Lachy>
nickshanks, <image> is already handled by the browsers exactly the same as <img>. It can't be changed without breaking back compat

[11:05:01.0000] <nickshanks>
yeah, i know. i still want it changed in standards mode. it can stay the same in tag soup mode

[11:10:00.0000] <nickshanks>
anyway, will have to come back later

[11:12:00.0000] <annevk>
yeah, we're not going to do that

[11:16:00.0000] <annevk>
2 promille of the page out there are using <image> and introducing more standards mode / quirks mode differences is silly

[11:17:00.0000] <annevk>
s/page/pages/ ...

[12:06:00.0000] <Philip`>
nickshanks (if you see this message): See http://www.imdb.com/ for an example of <image> on a site which browsers probably don't want to break

[12:09:00.0000] <Philip`>
http://canvex.lazyilluminati.com/survey/2007-07-17/analyse.cgi/pages/tag/image lists where I found <image> in ~8K pages

[12:10:00.0000] <annevk>
interesting, that matches the 2 promille figure

[12:10:01.0000] <annevk>
even though this sample is much smaller, hmm

[12:20:00.0000] <Lachy>
/me wonders if "promille" is supposed to mean "permille"

[12:26:00.0000] <annevk>
/me wonders why promille is Dutch-specific

[12:27:00.0000] <annevk>
anyway: ‰

[12:27:01.0000] <annevk>
oh, seems Finnish and German have it too

[12:28:00.0000] <webben>
I thought "standards mode" and "tag sop mode" where the same thing in HTML5?

[12:28:01.0000] <webben>
*soup

[12:28:02.0000] <webben>
(in terms of HTML parsing anyhow)

[12:28:03.0000] <annevk>
there's quirks mode, limited quirks mode and no quirks mode in HTML5

[12:29:00.0000] <annevk>
parsing currently only deals with no quirks mode, although it does do DOCTYPE processing already

[12:29:01.0000] <webben>
oh okay

[12:31:00.0000] <Philip`>
annevk: Why should the smaller sample size make it not match? :-)

[12:32:00.0000] <webben>
<picture> is probably a more promising line of inquiry, given it has no legacy

[12:33:00.0000] <annevk>
doing away with <img> would be silly imo

[12:33:01.0000] <annevk>
and there's always <object>

[12:33:02.0000] <annevk>
otherwise

[12:33:03.0000] <webben>
Does object for images actually work well enough to be used?

[12:33:04.0000] <kig>
promille is from latin, permille is the english bastardization (i guess)

[12:33:05.0000] <webben>
or does it require conditional comments to make it work?

[12:34:00.0000] <Philip`>
(I think 15 out of 7739 gives standard deviation around 4, so it's too small to get an especially accurate figure out)

[12:34:01.0000] <Philip`>
(*too small a sample)

[12:35:00.0000] <webben>
also: could one specify an API for images included via OBJECT?

[12:35:01.0000] <Philip`>
webben: See http://canvex.lazyilluminati.com/survey/2007-07-17/analyse.cgi/pages/tag/picture for <picture>

[12:35:02.0000] <webben>
Philip`: I'm more worried about what UAs do or don't do with it.

[12:35:03.0000] <Philip`>
<!--webbot bot="PhotoAlbum" U-Include="photogallery/photo18322/real.htm" clientside TAG="BODY" startspan --> <picture file-href="photogallery/photo18322/real_p.htm" />

[12:36:00.0000] <webben>
Philip`: but thanks for the link :)

[12:36:01.0000] <Philip`>
Argh, second link has Java and crashed Opera :-(

[12:36:02.0000] <annevk>
apart from getImageData() I'm not sure what API you'd want

[12:37:00.0000] <annevk>
and even getImageData() is not really needed given that there's drawImage()

[12:37:01.0000] <Philip`>
<!--webbot bot="PhotoAlbum" U-Include="photogallery/photo25215/real.htm" clientside TAG="BODY" startspan --> <picture file-href="photogallery/photo25215/real_p.htm" />

[12:38:00.0000] <Philip`>
Looks like both pages using <picture> get it via the same PhotoAlbum software

[12:39:00.0000] <webben>
annevk: Can getImageData() be used with OBJECT?

[12:39:01.0000] <webben>
(how large the API is less of a question than whether having an image-specific API is even possible for OBJECT)

[12:39:02.0000] <webben>
IIRC one of the rationale's for forking OBJECT to create VIDEO was that it made providing a dedicated API easier.

[12:40:00.0000] <Philip`>
http://support.microsoft.com/kb/813736 - ah, those <picture>s are from Microsoft FrontPage

[12:40:01.0000] <annevk>
for images, a dedicated API on <object> is less of an image, but this is all highly theoretical anyway

[12:40:02.0000] <annevk>
less of a problem, even :)

[12:40:03.0000] <webben>
annevk: It's not theoretical.

[12:40:04.0000] <annevk>
it is, there's no proposed API for images other than <canvas>

[12:41:00.0000] <annevk>
and <img> already works pretty damn well

[12:41:01.0000] <annevk>
compared to using <object> for video that is

[12:54:00.0000] <kig>
annevk, Philip`: re: canvas spec, i'll mail some comments on the ml, based on writing an svg renderer on canvas

[12:55:00.0000] <annevk>
cool

[12:55:01.0000] <annevk>
hopefully hixie will take another look at again as well

[13:33:00.0000] <G0k>
hey all

[13:34:00.0000] <G0k>
how's the party?

[13:35:00.0000] <takkaria>
"not on IRC"

[13:36:00.0000] <G0k>
ouch, cut my heart out why don't you

[13:39:00.0000] <takkaria>
:P

[13:39:01.0000] <takkaria>
hey, I'm on IRC too


2007-12-26
[18:09:00.0000] <hdh>
I'm confused as to with <dfn title="A">B</dfn>, is A or B "the term being defined", or must they be identical?

[18:09:01.0000] <hdh>
I would think that the title attr gives a definition

[04:55:00.0000] <gsnedders>
Can anyone think of an automated method to test list of integer parsing?

[05:02:00.0000] <annevk>
input output newline ?


2007-12-27
[02:02:00.0000] <Lachy_>
good morning everyone :-)

[02:04:00.0000] <annevk>
morning

[02:06:00.0000] <Lachy_>
any idea when Hixie will be returning?

[02:07:00.0000] <annevk>
returning from what?

[02:07:01.0000] <annevk>
(not really)

[02:28:00.0000] <Lachy>
well, I assume he's taking some sort of break, since I haven't seen him here for a few days.

[06:58:00.0000] <annevk>
https://bugzilla.mozilla.org/show_bug.cgi?id=392975 :)

[07:39:00.0000] <gsnedders>
/me posts a really short comment

[12:19:00.0000] <kig>
hahah, found my js port of the C impl in real-time fluid dynamics for games

[12:19:01.0000] <kig>
it's slooow

[12:20:00.0000] <Philip`>
/me is not surprised

[12:20:01.0000] <kig>
http://glimr.rubyforge.org/cake/tests/fluid.html and fluid_satan.html 

[12:21:00.0000] <kig>
http://www.dgp.toronto.edu/people/stam/reality/Research/pdf/GDC03.pdf

[12:22:00.0000] <Philip`>
I ported a simple 3D polygon renderer from a toy language that compiled into JVM bytecode, to JS; it was orders of magnitude slower :-(

[12:30:00.0000] <gsnedders>
JS,

[12:30:01.0000] <gsnedders>
slow?

[12:32:00.0000] <Philip`>
Can you do fluid dynamics with GPU shaders?

[12:33:00.0000] <gsnedders>
In JS? No. But that's not what you're meaning.

[12:34:00.0000] <Philip`>
Running JS on a GPU would be slightly insane, I think

[12:39:00.0000] <kig>
blackout :<

[12:42:00.0000] <Philip`>
/me 's home server indicates there has not been a power outage for 267 days and 14 hours, which is quite surprisingly reliable

[12:43:00.0000] <gsnedders>
I think when I feared there had been a powercut elsewhere in town I think there was just thick fog, so I couldn't see lights anywhere else

[12:47:00.0000] <Philip`>
Hmm, it's kind of annoying that it costs at least $50 to join the more interesting X3D mailing lists

[12:52:00.0000] <Philip`>
It also seems a bit annoying when someone responds to questions with 'for more detailed answers, buy my book' (~£30 on Amazon Marketplace), though they gave fairly detailed answers on the (free) X3D list too so I won't complain :-)

[13:39:00.0000] <kig>
webkit's CSS -webkit-transform is awesome

[14:34:00.0000] <Philip`>
How many web-related specifications try to define graceful error recovery to a reasonable extent, rather than being fatal or leaving it undefined? (HTML5 and CSS are the ones that seem most obvious to me)

[14:37:00.0000] <Lachy_>
Philip`, XBL, XHR, Selectors API and presumably any other spec worked on by Hixie, Anne, myself and others who think like us

[14:40:00.0000] <Philip`>
Hmm, I can't see anything in Selectors API which authors must not do and that UAs must handle in a certain way

[14:47:00.0000] <Hixie>
Philip`: CSS was the first spec to really do it

[14:47:01.0000] <Hixie>
though that was just for forward compat, not for bug compat and interop

[14:47:02.0000] <Hixie>
so it didn't define all cases

[14:47:03.0000] <Hixie>
we tried to fix that in 2.1

[14:55:00.0000] <gsnedders>
Hixie: 2.1 doesn't seem to have complete error handling when I last looked, though

[14:55:01.0000] <Hixie>
what got overlooked?

[14:55:02.0000] <gsnedders>
/me can't remember

[14:55:03.0000] <gsnedders>
I don't have the time to look right now

[14:55:04.0000] <gsnedders>
/me is half asleep

[15:37:00.0000] <Hixie>
it's very sad that despite not actually targetting IE at all, IE7 fails every test in acid3 so far

[15:38:00.0000] <Philip`>
Does it run the test framework correctly?

[15:38:01.0000] <Philip`>
(That was a significant problem for me with canvas tests in Safari and Firefox...)

[15:38:02.0000] <Hixie>
oh it doesn't fail _every_ test

[15:39:00.0000] <Hixie>
it passes 80 and 83

[15:39:01.0000] <Hixie>
wait 83 isn't done yet

[15:39:02.0000] <Hixie>
so it passes exactly one test

[15:39:03.0000] <Hixie>
i wonder who fails it

[15:40:00.0000] <Hixie>
seems nobody fails 80

[15:40:01.0000] <Hixie>
maybe i should change it

[15:40:02.0000] <gsnedders>
:D

[15:41:00.0000] <gsnedders>
the fun of acid tests… "wait, nobody fails this. better change this then…"

[15:42:00.0000] <Hixie>
oh hold on

[15:42:01.0000] <Hixie>
IE7 _does_ fail 80

[15:42:02.0000] <Hixie>
it just fails it in a way that does something odd to the results

[15:45:00.0000] <takkaria>
Hixie: have the ie8 team talked to you at all about more acid tests, or are you just doing acid3 because its time has come?

[15:47:00.0000] <Hixie>
neither

[15:48:00.0000] <Hixie>
i have asked the IE team to suggest tests, but they have refused to do so

[15:48:01.0000] <Hixie>
and i've been working on acid3 for at least 6 months now, on and off

[15:49:00.0000] <G0k>
what kind of stuff does acid3 test?

[15:50:00.0000] <Hixie>
dom

[15:50:01.0000] <G0k>
does anyone come close to passing it?

[15:50:02.0000] <Hixie>
any suggestions are welcome, and should ideally come in the form of a 1-20 line JS function that returns true or false


2007-12-28
[16:18:00.0000] <Lachy_>
Hixie, do you have any estimate for when you expect to complete acid 3?

[16:18:01.0000] <Hixie>
nope

[16:19:00.0000] <Hixie>
still many tests to come up with

[16:19:01.0000] <Lachy_>
yeah, I've noticed. I had a look at it last week to see how well Opera does

[16:19:02.0000] <Lachy_>
it's not too bad, but it certainly fails a lot

[18:04:00.0000] <Philip`>
    String attrib = attribs.getQName(i);

[18:04:01.0000] <Philip`>
    if(attrib.startsWith("xmlns:")) {

[18:04:02.0000] <Philip`>
        String space = attrib.substring(6);

[18:04:03.0000] <Philip`>
        if(!space.equals("xsd")) {

[18:04:04.0000] <Philip`>
            ...

[18:04:05.0000] <Philip`>
Could I be forgiven for thinking that's not the optimal way to process XML namespaces?

[18:13:00.0000] <othermaciej>
Philip`: that's some sad looking code

[18:18:00.0000] <Philip`>
As far as I can see, the code looks for <X3D xmlns:foo="[the X3D namespace URI]"> (where <X3D> has exactly qname "X3D", assuming it's not nested inside another <X3D>, and where foo != "xsd") and after that point it converts <foo:bar> into <bar> before its normal element processing

[18:19:00.0000] <Philip`>
which seems quite horrendously wrong

[18:35:00.0000] <othermaciej>
is xsd conventionally the prefix for something specific?

[18:35:01.0000] <othermaciej>
wondering why code would even check for that

[18:36:00.0000] <Philip`>
XML Schema

[18:37:00.0000] <Philip`>
The code does 'if(attrib.equals("xsd:noNamespaceSchemaLocation"))' to enable some extra processing or something

[18:39:00.0000] <Philip`>
The X3D world seems unpleasantly confused about namespaces and DTDs and schemas and whatnot

[23:01:00.0000] <Hixie>
/me makes extra sure to remove font metric dependencies in Ahem3

[23:01:01.0000] <Hixie>
er

[23:01:02.0000] <Hixie>
Acid3

[23:10:00.0000] <Hixie>
sigh, i still have so many tests to write and no idea what to do with them

[23:12:00.0000] <jruderman>
acid-style tests are overrated, imo. it's more meaningful to say "Browser X passes the entire test suite for Y" than "Browser X passes Acid N"

[23:16:00.0000] <Hixie>
i agree

[23:17:00.0000] <Hixie>
but try getting media attention for a test suite

[23:52:00.0000] <jruderman>
hehe

[00:06:00.0000] <Hixie>
i can't believe that with basically no effort on my part, IE fails every test

[00:07:00.0000] <Hixie>
i only targetted IE specifically for a couple of tests

[00:21:00.0000] <MacDome>
/me wonders which tests

[00:21:01.0000] <MacDome>
which test suite rather

[00:24:00.0000] <Hixie>
acid3

[00:25:00.0000] <Hixie>
http://www.hixie.ch/tests/evil/acid/003/NOT_READY_PLEASE_DO_NOT_USE.html

[00:25:01.0000] <Hixie>
if you have any suggestions of what to test, btw, do let me know

[00:25:02.0000] <MacDome>
yeah, Safari seems to fail pretty hard

[00:26:00.0000] <Hixie>
i'm trying to make safari and moz fail

[00:26:01.0000] <Hixie>
so i'm especially interested in bugs that affect those browsers

[00:26:02.0000] <MacDome>
Hixie: well you already found the strange JS parsing forgiveness bug in FF

[00:27:00.0000] <Hixie>
actually that's in the spec as far as i can tell

[00:27:01.0000] <MacDome>
the fact that somehow it didn't barf at "table." on a line by itself

[00:27:02.0000] <Hixie>
oh that

[00:27:03.0000] <Hixie>
yeah i need to add that back

[00:27:04.0000] <MacDome>
add that back!?

[00:27:05.0000] <MacDome>
it's a bug in FF AFAICT

[00:27:06.0000] <Hixie>
right

[00:27:07.0000] <Hixie>
need to catch it

[00:27:08.0000] <MacDome>
I guess.

[00:28:00.0000] <G0k>
but are bugs in the js parser really relevant?

[00:28:01.0000] <MacDome>
you'd have to load a separate JS file which you expected to fail to parse

[00:28:02.0000] <Hixie>
eval baby

[00:28:03.0000] <MacDome>
G0k: Acid2 had all sorts of irrelevant stuff :)

[00:28:04.0000] <Hixie>
G0k: yes, that's what i'm trying to test

[00:28:05.0000] <Hixie>
G0k: JS and DOM

[00:29:00.0000] <G0k>
mightn't you accidentally make the test incompatible with future ECMAScript version?

[00:29:01.0000] <G0k>
or extensions?

[00:29:02.0000] <Hixie>
maybe

[00:30:00.0000] <Hixie>
but i try to avoid things i know will happen

[00:31:00.0000] <G0k>
say so has anyone implemented the cross-document messaging thing?

[00:31:01.0000] <MacDome>
you're using all sorts of CSS3 which we don't have yet :)

[00:31:02.0000] <MacDome>
by "all sorts" I mean CSS3 selectors

[00:32:00.0000] <Hixie>
i'm only testing things that were in CR or better as of 2004

[00:32:01.0000] <annevk>
G0k, Opera has

[00:32:02.0000] <Hixie>
so at least 3 year old technologies

[00:32:03.0000] <annevk>
but Hixie changed the spec to make our impl incompatible with what's in the spec...

[00:32:04.0000] <Hixie>
everyone should have had time to implement them

[00:33:00.0000] <G0k>
annevk: say i had an idea i wanted to discuss with you

[00:33:01.0000] <annevk>
sure

[00:33:02.0000] <G0k>
so the dom event stream thing

[00:33:03.0000] <annevk>
/me just woke up, has time :)

[00:33:04.0000] <MacDome>
/me wonders if Acid 3 will include SVG

[00:33:05.0000] <Hixie>
no, only DOM and JS stuff

[00:34:00.0000] <G0k>
so since there's been talk about throwing it out and most of the real suggested uses of the event stream stuff has been chat-style apps...

[00:34:01.0000] <G0k>
what about just specifying a DOM interface to XMPP or some other already existing messaging protocol?

[00:35:00.0000] <G0k>
xmpp would be particularly cool since it can be tunneled over HTTP and has a defined way of including XMLy stuff

[00:35:01.0000] <Hixie>
good lord that would be obscenely complicated

[00:35:02.0000] <Hixie>
XMPP over HTTP

[00:36:00.0000] <G0k>
it already exists....

[00:36:01.0000] <Hixie>
one of hte requirements is that you can implement this from scratch in perl or python with 10-20 lines of code

[00:36:02.0000] <Hixie>
fully conforming

[00:36:03.0000] <G0k>
you can implement an SQL database in 10-20 lines of perl or python, fully conforming?

[00:37:00.0000] <Hixie>
by authors, i mean

[00:37:01.0000] <Hixie>
for author technologies

[00:37:02.0000] <Hixie>
author-side, that is

[00:37:03.0000] <MacDome>
yay for Acid3 crashing Safari :)

[00:38:00.0000] <Hixie>
really?

[00:38:01.0000] <Hixie>
i've been avoiding crash bugs

[00:38:02.0000] <krijnh>
Doesn't crash Safari 3 on Windows

[00:38:03.0000] <MacDome>
http://bugs.webkit.org/show_bug.cgi?id=16632

[00:39:00.0000] <MacDome>
the inspector crashed

[00:39:01.0000] <G0k>
i mean there's already lots of perl/python XMPP libs

[00:39:02.0000] <MacDome>
by somehow causing us to go down an unsupported code path in our Image code

[00:39:03.0000] <G0k>
sending a message is easily a one liner

[00:40:00.0000] <annevk>
you are making the web dependent upon the XMPP protocol though, which does sound like overkill

[00:42:00.0000] <G0k>
i guess i just really really don't understand the point of inventing a new protocol here

[00:42:01.0000] <G0k>
there are so many network event style protocols....and instead we're taking the one that isn't (HTTP) and trying to beat it into submission

[00:43:00.0000] <annevk>
this one is over HTTP

[00:44:00.0000] <annevk>
and a lot simpler than XMPP

[00:45:00.0000] <G0k>
yeah but it has an installed base of very approximately zero, and basically no existing server-side APIs

[00:46:00.0000] <G0k>
i mean the beauty of it not "needing" an API because it's simple is cool

[00:46:01.0000] <G0k>
but how well is that going to scale?

[00:47:00.0000] <G0k>
other protocols are at least somewhat widely implemented

[00:48:00.0000] <annevk>
the main thing this needs for adoption is more client-side impl

[00:49:00.0000] <G0k>
but does it really do anything that, say, slow download XHR doesn't do already?

[00:51:00.0000] <annevk>
it requires less connections and is more convenient

[00:52:00.0000] <annevk>
you can probably implement this on top of an <iframe> too, but it's not pretty

[00:55:00.0000] <annevk>
hehe https://bugzilla.mozilla.org/show_bug.cgi?id=349255

[00:56:00.0000] <Lachy>
Hixie, is acid3 affected at all if the user doesn't have Ahem installed?

[00:58:00.0000] <Lachy>
Hixie, you could use Ahem as a downloadable font, using @fontface stuff from CSS2

[00:59:00.0000] <annevk>
that wasn't really a CR three years ago

[01:05:00.0000] <G0k>
ok well my final argument is that XMPP can be implemented as a javascript library for non-supported UAs: http://zeank.in-berlin.de/jsjac/

[01:05:01.0000] <G0k>
which I think would be much harder for the event stream spec, at least as its written right now

[01:06:00.0000] <MacDome>
Hixie: it would be nice if Acid3 would do a bit of logging :)

[01:06:01.0000] <MacDome>
as in, if it displayed some sort of message for each failed test

[01:06:02.0000] <krijnh>
MacDome: Click the <h1>

[01:06:03.0000] <MacDome>
interesting

[01:07:00.0000] <G0k>
*it's

[01:08:00.0000] <MacDome>
/me is trying to figure out why buckets/result/intructions renders incorrect in WebKit

[01:13:00.0000] <MacDome>
/me thinks that somehow #result is supposed to be placing itself below <h1> and it's failign to do so correctly

[01:16:00.0000] <annevk>
hmm, Hixie changed Acid3!

[01:16:01.0000] <annevk>
Opera renders it worse and fails more

[01:20:00.0000] <MacDome>
Hixie, krijnh: yeah, but assertions with actual logs are much easier to debug. :)  I could point you towards our decent set of assertion functions for our JS tests if you like

[01:20:01.0000] <G0k>
is test 1 really fair? i mean isn't i possible that a DOM exists for PNG images that we just don't know about yet? :)

[01:21:00.0000] <Lachy>
at least outputting the results to something other tan an alert would be more useful so the results could be copied and pasted

[01:22:00.0000] <annevk>
hmm, you can't copy alert text?

[01:24:00.0000] <annevk>
btw, I dropped the advanced CSS value interfaces from the CSSOM

[01:24:01.0000] <annevk>
CSS animations seem better suited than using ECMAScript, even with convenient APIs

[01:25:00.0000] <Lachy>
annevk, not in all browsers. Only in Opera.

[01:25:01.0000] <annevk>
bah

[01:25:02.0000] <annevk>
/me notes that the test URI is pretty clear about the state of the test, though

[01:26:00.0000] <krijnh>
You think so? :)

[01:27:00.0000] <MacDome>
annevk: yeah, just hoping to provide some early feedback :)

[01:27:01.0000] <annevk>
Hixie, may I suggest that instead of a text/html MIME type you give that test a "bogus" MIME type?

[01:27:02.0000] <annevk>
Hixie, such as Content-Type:bogus

[01:27:03.0000] <MacDome>
eeew

[01:28:00.0000] <annevk>
Hixie, because when we previously implemented proper support for MIME types and CSS loading it turned out Gecko didn't

[01:28:01.0000] <annevk>
and we broke a site

[01:29:00.0000] <G0k>
yeah i mean.....DOM doesn't define how to not parse other documents

[01:31:00.0000] <G0k>
whereas recognizing an invalid mime type is indeed a good UA behavior

[01:31:01.0000] <annevk>
I mean this test: <link rel="stylesheet" href="empty.css"><!-- text/html file (should be ignored, <h1> will go red if it isn't) -->

[01:32:00.0000] <G0k>
ohz

[01:33:00.0000] <G0k>
that is a tough one, since it's really bad behavior but i'm sure fixing that will break sites

[01:34:00.0000] <annevk>
well, Gecko does it in standards mode, but it only does it if the MIME type actually parses correctly

[01:34:01.0000] <annevk>
for some value of "correctly"

[01:34:02.0000] <annevk>
I believe it checks if "/" or "\" occurs in the MIME type string...

[01:34:03.0000] <G0k>
so if it were text/bogus, what would happen?

[01:34:04.0000] <annevk>
then it would not apply in Gecko

[01:35:00.0000] <annevk>
but if it was "bogus" it would

[01:35:01.0000] <G0k>
hm. and what if it were omited?

[01:35:02.0000] <G0k>
if that...even makes snese

[01:35:03.0000] <annevk>
if Content-Type is not there it would also apply

[01:35:04.0000] <annevk>
in Gecko

[01:36:00.0000] <annevk>
we decided not to implement that and just ignore 404 URIs

[01:36:01.0000] <annevk>
which was the main problem

[01:47:00.0000] <MacDome>
Hixie: I'm not even sure if you want me filing these... :) but in case you wanted to know: http://bugs.webkit.org/show_bug.cgi?id=16636 http://bugs.webkit.org/show_bug.cgi?id=16637

[01:51:00.0000] <annevk>
MacDome, IE doesn't do proper DOM exceptions, Firefox does...

[01:52:00.0000] <MacDome>
annevk: if so, then it seems that that test is in disagreement with the DOM Core 2 spec

[01:52:01.0000] <annevk>
I think we emulated the e.SYNTAX_ERR stuff as well, but I'm not sure if it's supported by any spec

[01:55:00.0000] <MacDome>
annevk: and since we auto-gen our dom interfaces from the IDLs, we just match the official IDL in this case

[01:59:00.0000] <annevk>
I guess DOM Bindings / Web IDL should address this case

[01:59:01.0000] <annevk>
if it doesn't already

[02:54:00.0000] <annevk>
grmbl, getComputedStyle and various other CSSOM APIs are really quite crap in terms of documentation, implementation, etc.

[02:54:01.0000] <annevk>
DOM Level 2 Style is sort of like HTML4

[02:55:00.0000] <Hixie>
Lachy: in theory acid3 doesn't depend on fonts at all

[03:00:00.0000] <Hixie>
annevk: i can't find a spec that defines how to handle Content-Type: bogus

[03:01:00.0000] <annevk>
Hixie, I can't find a spec that defines that when <link rel=stylesheet href=test> returns Content-Type:text/html it can't be parsed as CSS

[03:02:00.0000] <annevk>
also, I don't think it's acceptable for other browsers to fix that bug in a way that Firefox handles it

[03:03:00.0000] <Hixie>
HTTP section 7.2.1 paragraph 3 sentence 2 is what i'm basing the current test on

[03:04:00.0000] <Hixie>
Content-Type: text/html is a given type, so sniffing is not allowed

[03:04:01.0000] <Hixie>
but I can't see anything telling me whether Content-Type: bogus is a given type or not

[03:04:02.0000] <Hixie>
so I could argue it both ways

[03:04:03.0000] <Hixie>
so it can't go in an acid test

[03:05:00.0000] <annevk>
hmm, given that text Firefox behavior (apart form the quirky parsing) does make some sense, but still...

[03:05:01.0000] <annevk>
I rather ignore Content-Type algother and just honer 404 and such...

[03:07:00.0000] <Hixie>
oh i agree

[03:07:01.0000] <Hixie>
well

[03:08:00.0000] <Hixie>
i don't think we should ever treat a text/html file as a stylesheet

[03:08:01.0000] <Hixie>
(except in quirks)

[03:08:02.0000] <Hixie>
but i agree that bogus should be treated like text/html

[03:08:03.0000] <Hixie>
i just can't test it in the acid3 test

[03:09:00.0000] <annevk>
why would we want a quirks there?

[03:09:01.0000] <annevk>
it's not like we have any trouble doing this for <script> or <img>

[03:09:02.0000] <Hixie>
there are lots of css files sent as not-text/css last i checked

[03:10:00.0000] <Hixie>
i have a lot of trouble, i just can't see any way to save it

[03:10:01.0000] <Hixie>
with those two

[03:10:02.0000] <Hixie>
the problem with css is that text/html files often end up containing CSS inline, and if you apply that to the other doc, it usually isn't what you meant to do

[03:10:03.0000] <annevk>
the only cases of that in wild are 404 docs

[03:11:00.0000] <annevk>
which you'd ignore anyway

[03:12:00.0000] <annevk>
Content-Type:bogus was actually used a on somewhat important website which is why Opera doesn't adhere to MIME types for CSS files

[03:12:01.0000] <Hixie>
well, 404 pages aren't always 404 Not Found

[03:12:02.0000] <Hixie>
witness the webstandards.org site recently

[03:12:03.0000] <Hixie>
but yeah

[03:12:04.0000] <Hixie>
i'd love to change the rules here

[03:13:00.0000] <Hixie>
i just don't see that we can convince the httpwg to do so

[03:13:01.0000] <Hixie>
and under that regime, the test is correct

[03:13:02.0000] <annevk>
i believe the HTTP WG doesn't care about "error handling"

[03:14:00.0000] <Hixie>
that does seem to be the case

[03:14:01.0000] <annevk>
so if we decide that the case of <link rel=stylesheet> pointing to a non text/css file is non-conforming but nevertheless must be treated as text/css they are probably happy

[03:17:00.0000] <Hixie>
i doubt it very much

[03:17:01.0000] <Hixie>
since that's not necessarily an error

[03:17:02.0000] <Hixie>
xslt, e.g.

[03:18:00.0000] <annevk>
surely what's conforming for <link rel=stylesheet> is not up to them?

[03:18:01.0000] <annevk>
(this is probably stretching it)

[03:21:00.0000] <Hixie>
it's not a matter of <link rel>

[03:21:01.0000] <Hixie>
<link rel> is just one of several ways of including a stylesheet

[03:21:02.0000] <Hixie>
including @import, Link:, etc

[03:21:03.0000] <Hixie>
as well as just direct access e.g. by an editor

[03:59:00.0000] <Hixie>
anyone got IE?

[03:59:01.0000] <Hixie>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%3A%3Cscript%3Eif%20(true)%20function%20a()%20{%20document.write(%27FAIL%27)%20}%20a()%3B%3C%2Fscript%3E

[03:59:02.0000] <Hixie>
does it say FAIL?

[04:01:00.0000] <annevk>
yes

[04:02:00.0000] <annevk>
well, ":FAIL"

[04:04:00.0000] <Lachy>
Hixie, the test looks wrong to me. Why shouldn't it print ":FAIL"?

[04:04:01.0000] <Hixie>
ES3 section 12.4

[04:04:02.0000] <Hixie>
and 12.5

[04:04:03.0000] <annevk>
does that say you can't define functions within if() statements or something?

[04:05:00.0000] <Hixie>
somewhat

[04:05:01.0000] <Lachy>
is it because function a() is out of scope when it's invoked?

[04:06:00.0000] <annevk>
I don't think if/else blocks create a new scope

[04:07:00.0000] <othermaciej>
Hixie: I think printing FAIL is correct behavior there

[04:08:00.0000] <Hixie>
othermaciej: it's not, per spec. i think the spec should change though. (webkit is the only compliant browser at the moment)

[04:09:00.0000] <othermaciej>
actually maybe not, due to lack of semicolon

[04:09:01.0000] <Lachy>
WebKit does something weird. It doesn't print "FAIL" for that test, but If you include the braces around the if block, then it does.

[04:10:00.0000] <othermaciej>
this should certainly print FAIL:

[04:10:01.0000] <othermaciej>
<!DOCTYPE html>:<script>if (true) { function a() { document.write('FAIL') }}; a();</script>

[04:10:02.0000] <othermaciej>
function declarations are processed before execution

[04:10:03.0000] <Hixie>
Lachy: that's correct per spec

[04:10:04.0000] <othermaciej>
and have either function or global scope

[04:10:05.0000] <othermaciej>
is the first version a syntax error?

[04:10:06.0000] <Hixie>
yes, if you add braces it should parse

[04:10:07.0000] <Hixie>
without braces it's a syntax error per spec

[04:11:00.0000] <othermaciej>
ah, I see

[04:11:01.0000] <webben>
because of "an ExpressionStatement cannot start with the function keyword because that might make it ambiguous with a FunctionDeclaration." ?

[04:11:02.0000] <othermaciej>
an if statement's condition is a statement, not a source element

[04:11:03.0000] <othermaciej>
a block is a valid statement

[04:12:00.0000] <othermaciej>
but a function declaration is not

[04:12:01.0000] <othermaciej>
it's a source element, but not a statement

[04:12:02.0000] <Lachy>
oh, ok. now I understand

[04:12:03.0000] <Lachy>
so only webkit passes

[04:12:04.0000] <othermaciej>
(and it's not an expression statement either, because that is disallowed)

[04:12:05.0000] <othermaciej>
good night all

[04:27:00.0000] <Dashiva>
Hixie: How does adding braces change it? The contents of a BlockStatement are Statements, just like the tail of an if

[04:29:00.0000] <Hixie>
hm, you are correct

[04:31:00.0000] <webben>
But without the brackets it's an expressionstatement not a block isn't it?

[04:31:01.0000] <Dashiva>
No, it's invalid syntax in both cases

[04:32:00.0000] <Dashiva>
ExpressionStatement can't start with "function", and nothing else matches

[04:33:00.0000] <webben>
sorry ... yes, I mean with brackets it would be a block.

[04:33:01.0000] <Dashiva>
yeah

[04:34:00.0000] <webben>
and http://bclary.com/2004/11/07/#a-12.1 doesn't say blocks can't start with function declarations

[04:34:01.0000] <Dashiva>
It says a Block contains Statement, and that's the same place we were in with the if

[04:34:02.0000] <Dashiva>
(well, StatementList)

[04:35:00.0000] <webben>
hmm

[04:36:00.0000] <Dashiva>
It's not attractive code, though. As macie mentioned, function declarations are processed first, so it's either a) always add the function, regardless of if test failing or b) break with es3 behavior (this is more like let x = function(){})

[04:37:00.0000] <Hixie>
yeah what's sad is that the spec doesn't match realitty

[04:37:01.0000] <webben>
so when http://developer.mozilla.org/en/docs/Core_JavaScript_1.5_Reference:Statements:function refers to function as a statement, is that a misnomer?

[04:37:02.0000] <Hixie>
oh well

[04:37:03.0000] <Dashiva>
webben: Yes and no

[04:37:04.0000] <Hixie>
if any of you come up with anything for me to test, do let me know

[04:38:00.0000] <Hixie>
nn

[04:38:01.0000] <Dashiva>
nn

[04:38:02.0000] <Dashiva>
webben: In spec terms, it's a SourceElement, which can be either a Statement or a function decl. But outside grammar parsing, calling it a statement makes for better text

[04:39:00.0000] <Dashiva>
I'm curious what WebKit does, though. If they crash on the function statement, or on the function invocation

[04:51:00.0000] <webben>
Dashiva: If it /can/ be a Statement, then can't it be the first statement in a block's statement list?

[04:52:00.0000] <Dashiva>
But it isn't a Statement in the grammar sense of the word

[04:53:00.0000] <webben>
ah I see

[05:15:00.0000] <othermaciej>
Dashiva: if (true) function a() {} is a parse error in webkit, so it fails at the declaration, not the invocation

[05:17:00.0000] <othermaciej>
webkit accepts SourceElements in many places that technically require a StatementList, probably for compatibility reasons

[05:19:00.0000] <annevk>
hmm, is that fixed in ECMA4?

[05:24:00.0000] <othermaciej>
I don't know

[05:25:00.0000] <othermaciej>
I can't find enough of a spec draft for ES4 to tell

[05:25:01.0000] <othermaciej>
I guess there is a grammar published so you can check that

[05:26:00.0000] <Dashiva>
The spec is still fragmented and wiki-dead and all that, I think

[05:26:01.0000] <othermaciej>
it doesn't really exist in a readable form; makes me wonder how they plan to finalize it next year

[05:27:00.0000] <annevk>
hmm

[05:27:01.0000] <annevk>
crap, WebKit implemented getPropertyCSSValue

[05:27:02.0000] <annevk>
I hope that's an internal API only

[05:27:03.0000] <othermaciej>
someone sent me a link to an semi-officiall grammar though

[05:28:00.0000] <othermaciej>
./css/CSSStyleDeclaration.idl:        CSSValue           getPropertyCSSValue(in DOMString propertyName);

[05:28:01.0000] <othermaciej>
nope

[05:28:02.0000] <othermaciej>
(is getPropertyCSSValue a bad thing?)

[05:28:03.0000] <annevk>
http://www.w3.org/mid/Pine.LNX.4.05.10310302134070.1173-100000⊙lif

[05:29:00.0000] <othermaciej>
grammar here: http://www.ecmascript.org/es4/spec/grammar.pdf

[05:31:00.0000] <othermaciej>
annevk: is WebKit the only engine to implement the old CSSOM?

[05:31:01.0000] <othermaciej>
(that seems unlikely)

[05:32:00.0000] <annevk>
maybe Gecko has some bits of it

[05:32:01.0000] <othermaciej>
Mozilla appears to have getPropertyCSSValue

[05:32:02.0000] <othermaciej>
and the CSSValue interface

[05:33:00.0000] <othermaciej>
(well, from searching their source code it seems that way)

[05:34:00.0000] <annevk>
In Opera the method throws a NOT_IMPLEMENTED_ERR and in Gecko it always returns null

[05:34:01.0000] <annevk>
it won't be part of the next CSSOM if it's up to me

[05:35:00.0000] <othermaciej>
div id="foo" style="color: red">x</div>

[05:35:01.0000] <othermaciej>
<script>

[05:35:02.0000] <othermaciej>
var foo = document.getElementById("foo");

[05:35:03.0000] <othermaciej>
alert(document.defaultView.getComputedStyle(foo, "").getPropertyCSSValue("color"));

[05:35:04.0000] <othermaciej>
</script>

[05:35:05.0000] <othermaciej>
does what I expect in Firefox 2

[05:36:00.0000] <othermaciej>
(and Firefox 3 Beta 1)

[05:37:00.0000] <othermaciej>
it's pretty likely some content depends on it, even more so that some content depends on the CSSValue interface and related interfaces

[05:38:00.0000] <othermaciej>
(it does return null in Firefox on foo.style though, which is certainly weird)

[05:39:00.0000] <annevk>
we haven't had a single bug report on not supporting it

[05:39:01.0000] <annevk>
and IE doesn't support it

[05:39:02.0000] <othermaciej>
does IE support getComputedStyle()?

[05:39:03.0000] <annevk>
fair point

[05:40:00.0000] <othermaciej>
we have had bug reports relating to getComputedStyle before (due to the fact that document.defaultView was originally not the same document as window)

[05:40:01.0000] <annevk>
oh, yeah, pages rely on getComputedStyle

[05:41:00.0000] <annevk>
i guess I can file a bug on WebKit and Gecko and see what happens

[05:41:01.0000] <othermaciej>
to remove getPropertyCSSValue?

[05:41:02.0000] <annevk>
yeah

[05:42:00.0000] <annevk>
is CSSStyleDeclaration implemented on various objects btw that behave slightly different?

[05:42:01.0000] <othermaciej>
what's the reason to remove it?

[05:42:02.0000] <annevk>
that it has been obsoleted by the CSS WG

[05:42:03.0000] <othermaciej>
that's not a good reason

[05:42:04.0000] <annevk>
and that the API is awkward for the Web

[05:42:05.0000] <othermaciej>
that one, sadly, isn't either

[05:43:00.0000] <annevk>
if anything we want something like .style.width.px

[05:43:01.0000] <othermaciej>
neither of those is worth breaking content, even if it's Apple-only content like Dashboard widgets

[05:43:02.0000] <annevk>
bloat would be another reason

[05:43:03.0000] <othermaciej>
I'd certainly be open to having a nicer CSS API in addition

[05:44:00.0000] <othermaciej>
the CSSValue part of the API specifically is not a lot of code

[05:44:01.0000] <annevk>
i doubt anything relies on it, but you could be right

[05:44:02.0000] <othermaciej>
getComputedStyle is the hard part, if anything

[05:44:03.0000] <othermaciej>
(and CSS parsing)

[05:45:00.0000] <annevk>
CSS parsing is madness

[05:45:01.0000] <othermaciej>
it's better than HTML parsing :-)

[05:45:02.0000] <annevk>
/me always thought CSS was a simple language

[05:45:03.0000] <othermaciej>
or JavaScript parsing

[05:46:00.0000] <annevk>
b\000061ckground

[05:46:01.0000] <annevk>
or b\061\063kground

[05:48:00.0000] <othermaciej>
escape sequences aren't particularly complicated to parse

[05:48:01.0000] <othermaciej>
(all handled by the lexer)

[05:49:00.0000] <annevk>
true

[05:50:00.0000] <annevk>
i probably don't get enough of the CSS grammar yet

[05:51:00.0000] <annevk>
it's not entirely clear to me what the intersection is of the appendix and the core grammar etc.

[05:54:00.0000] <annevk>
Mozilla is not working anymore on getPropertyCSSValue per that CSS WG message

[05:54:01.0000] <annevk>
so that's encouraging

[05:54:02.0000] <annevk>
(info from bugs in bugzilla)

[05:55:00.0000] <othermaciej>
it looks like ES4 does allow function definitions in blocks, but not in places that expect a single statement (as far as I can tell from the grammar)

[05:56:00.0000] <othermaciej>
(side note: we appear to have the UnknownRule interface too, but I dunno if that gets used by anything)

[05:58:00.0000] <annevk>
UknownRule is incompatible with the CSS parser so that sounds interesting :)

[05:59:00.0000] <othermaciej>
I think the interface exists but nothing actually ever makes one

[06:02:00.0000] <annevk>
to get back to my earlier question, to properly define CSSStyleDeclaration would I have to define two separate implementations of it?

[06:02:01.0000] <annevk>
one that is readonly and one that is read-write

[06:02:02.0000] <annevk>
and maybe there's also a computed / non-computed distinction?

[06:02:03.0000] <othermaciej>
probably, yes

[06:03:00.0000] <othermaciej>
we internally have CSSSMutableStyleDeclaration and CSSComputedStyleDeclaration classes

[06:04:00.0000] <othermaciej>
they both implement the CSSStyleDeclaration interface but the Computed version either ignores or throws on all attempts at mutation (can't remember which)

[06:04:01.0000] <annevk>
Mozilla actually exposes the latter as interface name

[06:04:02.0000] <annevk>
it throws

[06:04:03.0000] <annevk>
/me read bits of the source

[06:05:00.0000] <annevk>
one of the larger problems with CSSValue and any kind of dedicated CSS API is that it's hard to make them scale

[06:06:00.0000] <othermaciej>
but yeah, it would be nicer if there was a read-write interface inheriting from a read-only interface, not sure what the consequences of changing now would be

[06:06:01.0000] <annevk>
at some point background-image could have returned CSSUrlValue

[06:06:02.0000] <othermaciej>
there's probably not much risk of someone depending on existence of a method that always throws

[06:06:03.0000] <annevk>
now it needs to return a list

[06:07:00.0000] <annevk>
can I actually do that in IDL?

[06:07:01.0000] <annevk>
inherit and modify members

[06:08:00.0000] <annevk>
another option would to have separate interfaces where you simply don't have modifying stuff on one of them...

[06:08:01.0000] <othermaciej>
you can inherit and add members

[06:08:02.0000] <gsnedders>
annevk: peh! anyone on the HTTPbis mailing list is a member of the WG! Some people in the WG care about error handling :P

[06:08:03.0000] <othermaciej>
but I dunno if it's praactically changeable at this point

[06:09:00.0000] <othermaciej>
probably easier to just spec computed styles as implementing the same interface but throwing on mutation

[06:10:00.0000] <gsnedders>
annevk: SimplePie (another feed parser) has similar behaviour to UFP for text/xml as of the latest dev copy

[06:11:00.0000] <gsnedders>
it also uses content-type sniffing based on HTML 5.

[06:11:01.0000] <gsnedders>
Only real breakage is text/plain feeds

[06:12:00.0000] <annevk>
othermaciej, yeah ok, it doesn't really matter either way I suppose

[06:12:01.0000] <gsnedders>
The other breakages have mainly been impl bugs

[06:14:00.0000] <gsnedders>
/me wonders how you're meant to deal with application/octet-stream

[06:30:00.0000] <annevk>
the other annoying thing is that getComputedStyle does not return the computed style

[06:30:01.0000] <annevk>
<body style=width:20em> the computed style of 'width' is not 320px afaict

[06:32:00.0000] <annevk>
or without 'width' it should be auto

[06:32:01.0000] <annevk>
however, I get back 1452px instead

[06:33:00.0000] <annevk>
currentStyle in IE is much closer to giving back the computed style it seems

[08:21:00.0000] <annevk>
bah, updates to <style>.media.mediaText and presumably similar stuff has no effect

[08:22:00.0000] <annevk>
(for a normal persistent style sheet)

[08:23:00.0000] <annevk>
(haven't tested Safari)

[08:24:00.0000] <annevk>
actually, IE7 does do updates but it does not have MediaList object...

[08:24:01.0000] <annevk>
in IE7 <style>.sheet.media is a simple string

[08:25:00.0000] <annevk>
(<style>.media above should've been <style>.sheet.media as well)

[08:45:00.0000] <annevk>
actually, Opera seems to do dynamic updates too, but not always

[08:46:00.0000] <annevk>
and <style>.media is not updated after changes from <style>.sheet.media


2007-12-29
[16:50:00.0000] <heycam>
hi people.  are there ecmascript objects that stringify in non-default ways, apart from window.location and window.getSelection()?

[16:51:00.0000] <MacDome>
one can always override toString...

[16:52:00.0000] <heycam>
yeah.  i'm wondering what other (host) objects do that.  (or perhaps stringify in a different way, e.g. by having a different [[DefaultValue]] method.)

[17:06:00.0000] <othermaciej>
heycam: in WebKit, Range and HTMLAnchorElement also have custom toString methods (though not sure if that is necessary or correct)

[17:07:00.0000] <othermaciej>
javascript:alert(document.links[0]) on most pages will show the result

[17:07:01.0000] <heycam>
ok, thanks

[17:10:00.0000] <othermaciej>
heycam: looks like Mozilla does at least the HTMLAnchorElement one as well (serializing as its own href)

[17:10:01.0000] <heycam>
yeah i just noticed.  maybe the spec should say something?

[17:10:02.0000] <othermaciej>
heycam: it's also possible that there are more classes in WebKit which serialize in a custom way, but do it too obscurely for my simple search to find

[17:11:00.0000] <othermaciej>
(I found Selection, Location, Range and HTMLAnchorElement)

[17:11:01.0000] <heycam>
(ie7 also does the HTMLAnchorElement thing)

[17:12:00.0000] <othermaciej>
I suspect serializing anchors that way is needed for compatibility, so yes, HTML5 should probably say something

[17:18:00.0000] <gavin__>
Gecko appears to have custom toStrings for HTMLAnchorElements and HTMLAreaElements, Range, Selection

[17:19:00.0000] <gavin__>
and some CSSOM related objects that I don't really know about

[17:19:01.0000] <gavin__>
othermaciej's caveat applies

[17:20:00.0000] <heycam>
thanks gavin__

[17:20:01.0000] <gavin__>
I didn't find Location because its magical, for example

[17:20:02.0000] <heycam>
how's the magic implemented there?

[17:20:03.0000] <gavin__>
well it's magic for reasons other than having a custom tostring

[17:21:00.0000] <heycam>
ah

[17:21:01.0000] <gavin__>
oh, actually I'm wrong

[17:21:02.0000] <gavin__>
I just missed it

[17:21:03.0000] <gavin__>
http://lxr.mozilla.org/seamonkey/source/dom/src/base/nsLocation.cpp#979

[17:22:00.0000] <heycam>
/me wonders if [[DefaultValue]] is ever overridden in imlementations

[19:22:00.0000] <Hixie>
ok now IE is just TRYING to fail these tests

[19:22:01.0000] <Hixie>
i finally found something that EVERY BROWSER EXCEPT IE failed

[19:23:00.0000] <Hixie>
and i added it to acid3

[19:23:01.0000] <Hixie>
and now it fails in IE, because IE handles code in eval() differently from code in a js block

[19:23:02.0000] <Hixie>
so IE still fails every test

[19:23:03.0000] <Hixie>
EVEN THE ONES I ADDED SPECIFICALLY TO GIVE IT A NON-ZERO STARTING SCORE

[19:24:00.0000] <hober>
heh

[20:26:00.0000] <othermaciej>
Hixie: code in eval() is supposed to be treated differently from code in a JS block in some ways

[20:26:01.0000] <MacDome>
sadly Acid3 seems to have borked again... oh well

[20:26:02.0000] <MacDome>
and we're now 92%!

[20:26:03.0000] <MacDome>
huaazh

[20:45:00.0000] <Hixie>
othermaciej: different how?

[21:53:00.0000] <MacDome>
Hixie: sadly the HMTL5 spec does not support your form.elements.length == form.elements testcase.  http://bugs.webkit.org/show_bug.cgi?id=16656

[21:53:01.0000] <MacDome>
unless I missed something in my reading thereof

[21:54:00.0000] <Hixie>
FIXED

[21:54:01.0000] <Hixie>
er

[21:54:02.0000] <Hixie>
fixed

[21:54:03.0000] <Hixie>
that was a typo

[21:54:04.0000] <Hixie>
i meant to test form.elements.length against form.length

[22:29:00.0000] <MacDome>
Hixie: oh good.  well, I just sent an email about it :)

[22:29:01.0000] <MacDome>
Hixie: said email contained a list of all WebKit bugs re: Acid3

[22:29:02.0000] <MacDome>
which I imagine you will find interesting :)

[22:29:03.0000] <Hixie>
cool

[22:30:00.0000] <MacDome>
/me goes to look at why the latest Acid3 file causes webkit to fail completely

[22:30:01.0000] <MacDome>
ah, it's fixed again, good

[22:30:02.0000] <MacDome>
except now we're 88% instead of 91% :(

[22:32:00.0000] <Hixie>
teehee

[22:32:01.0000] <Hixie>
do you know of any bugs in safari or mozilla that i should test?

[22:33:00.0000] <MacDome>
well, XHTML support is riddled w/ bugs :)

[22:33:01.0000] <MacDome>
embedding bugs are a big deal for me

[22:33:02.0000] <MacDome>
Hixie: I really think the test could benefit from using something like these: http://trac.webkit.org/projects/webkit/browser/trunk/LayoutTests/fast/js/resources/js-test-pre.js

[22:33:03.0000] <Hixie>
i mean DOM or JS things

[22:33:04.0000] <MacDome>
for making it easier to debug

[22:34:00.0000] <MacDome>
Hixie: I'll think about it next time I'm sorting through old bugs

[22:34:01.0000] <Hixie>
you mean reporting something more than the test number?

[22:34:02.0000] <Hixie>
i'm not sure exactly what one would need to report

[22:34:03.0000] <MacDome>
Hixie: breaking the ifs down into single tests

[22:35:00.0000] <MacDome>
instead of if (foo || bar || baz) fail!

[22:35:01.0000] <MacDome>
it could be

[22:35:02.0000] <MacDome>
assert(!foo);

[22:35:03.0000] <MacDome>
assert(!bar)

[22:35:04.0000] <MacDome>
etc

[22:35:05.0000] <Hixie>
and have assert thrown an exception or something?

[22:35:06.0000] <Hixie>
i could i guess

[22:35:07.0000] <MacDome>
sure

[22:35:08.0000] <MacDome>
and then the wrapper could catch it

[22:35:09.0000] <Hixie>
seems tdious

[22:35:10.0000] <MacDome>
it would certainly make debugging easier. to have things be single-line.  but maybe it's not worth it

[22:36:00.0000] <MacDome>
it could throw an exception containing the assertion text.

[22:36:01.0000] <MacDome>
which would allow you to report things better

[22:36:02.0000] <Hixie>
then only the first bit that fails would be run

[22:36:03.0000] <MacDome>
but again, maybe not worth it

[22:36:04.0000] <Hixie>
but i guess that's ok

[22:37:00.0000] <Hixie>
(of each test)

[22:37:01.0000] <MacDome>
well, you obviously have lots of time to play with differnet methods for differnet tests

[22:38:00.0000] <MacDome>
perhaps you'll find one you like better than the current

[22:38:01.0000] <MacDome>
from a debugging standpoint, watching for when "ok" turns false, and then checking each part of an if () clause can be a bit tedius

[22:38:02.0000] <Hixie>
true

[22:38:03.0000] <MacDome>
/me goes back to debugging the latest build

[22:41:00.0000] <Hixie>
how do i distinguish a DOMException object from a string thrown by 'throw' in a spec-compliant way that works in IE?

[22:42:00.0000] <MacDome>
Hixie: ha!  it looks like test 56 is impossible to pass.  You never set ok to true! :)

[22:43:00.0000] <MacDome>
/me assumes he meant to initialize ok to true

[22:43:01.0000] <Hixie>
fixed

[22:43:02.0000] <MacDome>
Hixie++ # for testing reserved words!

[22:43:03.0000] <MacDome>
Hixie: that was one thing I *really* wanted to see moz fix

[22:44:00.0000] <Hixie>
any others?

[22:44:01.0000] <MacDome>
or rather.. to see all browsers agree on

[22:44:02.0000] <Hixie>
i have no idea how to handle this exception thing

[22:44:03.0000] <MacDome>
Hixie: can't you grab the prototype off of an object in IE?

[22:44:04.0000] <MacDome>
Hixie: or maybe .toNumber() the exception?

[22:45:00.0000] <Hixie>
is there a spec that guarantees either of those do anything sane for DOMException? (specifically, a spec that was in CR or better in 2004)

[22:46:00.0000] <Hixie>
i'll just throw an object with a 'message' property

[22:46:01.0000] <Hixie>
that works everywhere

[22:47:00.0000] <Hixie>
i'm amused that test 85 is the only test ie can pass

[22:47:01.0000] <Hixie>
i'm not especially targetting IE either

[22:47:02.0000] <Hixie>
i'm really only worrying about mozilla and safari bugs, by and large

[22:48:00.0000] <Hixie>
(except for things other people have told me to test)

[22:48:01.0000] <Hixie>
(like ie's attribute mess)

[22:48:02.0000] <MacDome>
Hixie: it woudl appear that the 30s are misnumbered in comment

[22:48:03.0000] <MacDome>
it says we fail "39" but according to the comments there is no 39 :)

[22:49:00.0000] <MacDome>
/me goes to look at the DOMException spec

[22:50:00.0000] <MacDome>
what the hell is an "exception" instead of an "interface" in idl

[22:51:00.0000] <Hixie>
how is there no 39

[22:52:00.0000] <MacDome>
/me reads http://www.w3.org/TR/DOM-Bindings/#idl-exceptions

[22:53:00.0000] <MacDome>
Hixie: so is e.code accessible within IE?

[22:54:00.0000] <othermaciej>
Hixie: maybe not the difference you noticed, but there are some intended behavior differences for eval code, for example "var" declarations create bindings that are *not* DontDelete under eval

[22:54:01.0000] <Hixie>
MacDome: that's one of the things i test i believe

[22:54:02.0000] <Hixie>
othermaciej: ah

[22:54:03.0000] <Hixie>
othermaciej: is that in the spec?

[22:55:00.0000] <MacDome>
Hixie: you were just asking for ways to identify DOMExceptions, I had assumed you were looking to replace the e.HIERARCHY_.... check

[22:55:01.0000] <MacDome>
which fails in Safari and seems to disagree w/ the spec

[22:55:02.0000] <othermaciej>
Hixie: some differences are spec'd here: http://bclary.com/2004/11/07/#a-10.2.2

[22:56:00.0000] <MacDome>
interesting.  IE always makes .constructor DontEnum

[22:57:00.0000] <MacDome>
even if it might be

[22:57:01.0000] <othermaciej>
Hixie: ah, in fact that section includes the difference I mentioned

[22:57:02.0000] <Hixie>
MacDome: no, i was looking for ways to change the framework to handle custom exceptions. i cheated instead.

[22:57:03.0000] <Hixie>
othermaciej: cool

[22:57:04.0000] <MacDome>
Hixie: I wonder if you can grab at e.constructor and check to see if it's a DOMException that way... I'm not actually sure what e.constructor will get you, othermaciej would probably know

[22:58:00.0000] <Hixie>
i can only rely on things that either work 100% reliably from DOM Level 0, or things that were in specs at CR or later in 2004

[22:58:01.0000] <Hixie>
DOM Bindings wasn't even close to either

[22:59:00.0000] <MacDome>
november 2000: http://www.w3.org/TR/DOM-Level-2-Core/core.html#ID-17189187 ?

[22:59:01.0000] <Hixie>
MacDome: where's the ".constructor" part of that?

[22:59:02.0000] <MacDome>
I was thinking .constructor was on Object

[23:00:00.0000] <Hixie>
exceptions aren't necessarily Objects

[23:00:01.0000] <Hixie>
they're host objects

[23:00:02.0000] <Hixie>
which basically (as of 2004) had no defined behaviour

[23:00:03.0000] <Hixie>
insofar as .constructor goes

[23:00:04.0000] <Hixie>
as far as i can tell

[23:00:05.0000] <MacDome>
/me wonders what an exception can be if not an "Object" in JS.  I guess it could be another primitive type...

[23:01:00.0000] <Hixie>
host objects don't have to be any primitive type as i understand it

[23:09:00.0000] <MacDome>
Er r or Obj e c t s 

[23:09:01.0000] <MacDome>
I nst ances of Er r or obj ect s ar e t hr own as except i ons when r unt i me er r or s occur . The Er r or obj ect s may al so 

[23:09:02.0000] <MacDome>
ser ve as base obj ect s f or user - def i ned except i on cl asses. 

[23:10:00.0000] <MacDome>
Hixie: according to ECMA e.constructor.prototype.name == "Error" for all runtime exceptions

[23:11:00.0000] <MacDome>
Hixie: assuming I'm reading 15.11 correctly

[23:11:01.0000] <Hixie>
sure, but user-thrown exceptions and DOM-thrown exceptions aren't runtime exceptions

[23:11:02.0000] <MacDome>
correct, they are not required to be based from Error

[23:11:03.0000] <MacDome>
Hixie: user exceptions at least

[23:11:04.0000] <MacDome>
Hixie: I'm not sure about DOM exceptions

[23:11:05.0000] <Hixie>
DOM exceptions are effectively user-defined

[23:11:06.0000] <Hixie>
or rather, host-defined

[23:11:07.0000] <Hixie>
(which is even less useful to us)

[23:11:08.0000] <Hixie>
dom bindings addresses all this, luckily

[23:13:00.0000] <MacDome>
well, throw can throw any arbitrary value, so that's no hel

[23:13:01.0000] <MacDome>
p

[23:14:00.0000] <Hixie>
yeah, i just settled on throwing a { message: "" } object

[23:14:01.0000] <MacDome>
nice new error messages!

[23:14:02.0000] <MacDome>
I just reloaded and saw them

[23:14:03.0000] <Hixie>
i'm up to test 36

[23:15:00.0000] <MacDome>
bah.  once again you check e.NAMESPACE_ERR  which isn't supported by any spec I've seen

[23:16:00.0000] <MacDome>
/me bitches instead of fixing the "bug"

[23:16:01.0000] <Hixie>
DOM2 Core, appendix E

[23:18:00.0000] <MacDome>
nice new tests, btw.

[23:19:00.0000] <MacDome>
Hixie: sure, and javascript:alert(DOMException.HIERARCHY_REQUEST_ERR) is valid in Safari

[23:19:01.0000] <MacDome>
Hixie: but just because it's a property on the constructor funtion doesn't mean it's a property on the prototype

[23:20:00.0000] <Hixie>
do you agree that instances of Node have the node type constants on their objects?

[23:21:00.0000] <MacDome>
sure.  it's defined for any object conforming to the Node interface: http://www.w3.org/TR/DOM-Level-2-Core/core.html#ID-1950641247

[23:21:01.0000] <MacDome>
http://www.w3.org/TR/DOM-Level-2-Core/core.html#ID-17189187 DOMException seems to have no such requirement

[23:21:02.0000] <Hixie>
the text defining that  for JSis exactly the same as the text defining the constants should be on all exception objects

[23:22:00.0000] <MacDome>
I don't follow

[23:22:01.0000] <MacDome>
Foo.bar != (new Foo).bar

[23:22:02.0000] <Hixie>
the only reason Node objects have those constants is the text in appendix E that says that Node.ELEMENT_NODE on the "Prototype Object Node" is present

[23:23:00.0000] <MacDome>
(the easy solution is obviously for us to implement these constants as part of the DOMException interface), I'm just not sure I understand how that's implied form the spec yet

[23:23:01.0000] <Hixie>
which is the same text as the text that says that DOMException.INDEX_SIZE_ERR on the "Prototype Object DOMException" is present

[23:23:02.0000] <MacDome>
ah...

[23:23:03.0000] <Hixie>
so either the constants should be present in both cases, or in neither cae

[23:24:00.0000] <Hixie>
case

[23:24:01.0000] <MacDome>
Prototype Object DOMException

[23:24:02.0000] <Hixie>
now i agree that the text in the spec sucks, which is why we need the DOM Bindigns spec

[23:24:03.0000] <MacDome>
I can see how that could be used to mean that the prototype for the object DOMException shoudl have those constants

[23:24:04.0000] <MacDome>
ok, I'll concede Prototype Object DOMException saves your case

[23:24:05.0000] <Hixie>
but i can't see any way to interpret that spec which leads to exceptions being different from nodes in this regard

[23:25:00.0000] <MacDome>
Hixie: well, ignoring appendix E, I think my point is valid

[23:25:01.0000] <Hixie>
appendix E is the only reason we have anything in JS at all

[23:25:02.0000] <MacDome>
Hixie: the rest of the spec clearly demands that (new Node).ELEMENT_NODE be valid.

[23:25:03.0000] <MacDome>
and demands that HIERARCHY_REQUEST_ERR be defined

[23:25:04.0000] <MacDome>
but makes no such demand on (new DOMException).HIERARCHY_REQUEST_ERR

[23:25:05.0000] <Hixie>
i disagree; i don't see anything that says how to interpret that idl other than appendix E

[23:26:00.0000] <MacDome>
/me shrugs

[23:26:01.0000] <MacDome>
/me goes to prepare the fix

[23:27:00.0000] <Hixie>
hehe

[23:27:01.0000] <Hixie>
i've done half the tests with error messages

[23:27:02.0000] <Hixie>
i'll do the other half in a bit

[23:28:00.0000] <MacDome>
/me enjoys having such a formidable spec opponent. :)

[23:28:01.0000] <Hixie>
:-)

[23:31:00.0000] <MacDome>
/me sighs.  DOMException.idl appears non-autogenerated :(

[00:03:00.0000] <MacDome>
sigh.  Hixie's got webkit down to 85% :(

[00:05:00.0000] <MacDome>
actually 85% for 3.04, 88% for TOT

[00:06:00.0000] <MacDome>
Hixie: soooo much easier to debug, btw.  thanks.

[00:46:00.0000] <othermaciej>
I think MacDome doesn't realize that the test will not seem legit if WebKit passes out of the box

[00:49:00.0000] <Hixie>
i'm far more interested in seeing compliant browsers than in having my life be easy in making the test seem legit

[00:50:00.0000] <othermaciej>
yeah, but if he keeps fixing bugs, you'll have to keep adding tests

[00:50:01.0000] <Hixie>
that is a problem i welcome with open arms

[00:50:02.0000] <othermaciej>
I mean, not that I'm against people fixing more bugs

[01:06:00.0000] <hsivonen>
Hixie: you could test that XHTML docs implement HTMLDocument and check that document.body and stuff works

[01:06:01.0000] <Hixie>
is there a way to do that purely from DOM?

[01:07:00.0000] <Hixie>
(i'm trying to avoid external files)

[01:09:00.0000] <hsivonen>
Hixie: I'm not sure if it is possible with 2004 specs. with HTML 5 it should be. :-)

[01:09:01.0000] <hsivonen>
Hixie: unless there's a spec for DOMParser

[01:10:00.0000] <Hixie>
testing html5 in an acid test will be for acid4, maybe :-)

[01:10:01.0000] <hsivonen>
which I doubt

[01:10:02.0000] <Hixie>
there's dom3 load and save, but i'm pretending that doesn't exist

[01:10:03.0000] <hsivonen>
(I doubt DOMParser has a spec that is)

[01:11:00.0000] <hsivonen>
Hixie: anyway, document.body would be a test that'd help acid3 look legitimate by WebKit not passing :-)

[01:11:01.0000] <hsivonen>
in XHTML that is

[01:11:02.0000] <Hixie>
heh

[01:12:00.0000] <Hixie>
i could createDocument() an XHTML doc

[01:18:00.0000] <othermaciej>
hsivonen: XHTML documents implementing HTMLDocument is not required by any spec but HTML5 afaik

[01:18:01.0000] <Hixie>
and dom2 html

[01:18:02.0000] <othermaciej>
I thought it was allowed but not required by dom2 html

[01:18:03.0000] <othermaciej>
/me checks his references

[01:19:00.0000] <Hixie>
it's required as much as HTMLDocument is required for text/html, as far as i can tell

[01:19:01.0000] <Hixie>
of course back in those days specs were so vague...

[01:20:00.0000] <othermaciej>
yeah, it seems to be required for XHTML 1.0 as much as for HTML 4.01, which is to say, not very clearly required at all

[01:20:01.0000] <othermaciej>
wow, so many statements in DOM 2 HTML are just outright factually false

[01:20:02.0000] <Hixie>
do paste them here

[01:21:00.0000] <othermaciej>
"In many cases, these enhancements are not applicable to a general DOM because they rely on the presence of a predefined DTD."

[01:21:01.0000] <krijnh>
http://www.digg.com/software/Ian_Hickson_s_having_a_bit_of_trouble_writing_ACID3 :P

[01:22:00.0000] <othermaciej>
Hixie: I do know of one DOM Level 1 Core thing that I think every browser gets wrong per spec (but which probably can't be fixed)

[01:22:01.0000] <Hixie>
yeah i'm avoiding those

[01:23:00.0000] <Hixie>
and yeah, the DTD stuff is funny

[01:23:01.0000] <othermaciej>
elt.getAttribute("nonexistentAttribute") is supposed to return empty string instead of null

[01:23:02.0000] <Hixie>
back in those days people believed it

[01:23:03.0000] <hsivonen>
DTDs have the questionable honor of being the second most misunderstood part of XML right after namespaces

[01:27:00.0000] <othermaciej>
out of the official w3c DOM tests, WebKit fails 6 of the DOM1 Core ones, 1 of the DOM2 Core, 1 of the DOM2 Events, and 5 of the DOM2 HTML

[01:27:01.0000] <othermaciej>
(in some cases a few more than that in xhtml instead of html, not sure why)

[01:29:00.0000] <Hixie>
othermaciej: is there a summary of those failures somewhere?

[01:29:01.0000] <Hixie>
e.g. bugzilla query?

[01:30:00.0000] <othermaciej>
not in bugzilla

[01:31:00.0000] <othermaciej>
if you have a WebKit tree checked out, WebKitTools/Scripts/check-dom-results gives counts of tests that appear to succeed, and you can search for "fail" in .txt files in LayoutTests/dom/* to find which specific tests fail

[01:31:01.0000] <othermaciej>
a lot of the DOM3 Core ones fail due to unimplemented useless features

[01:31:02.0000] <Hixie>
i do not have a checkout right now

[01:31:03.0000] <othermaciej>
most of the others are deliberate failures for web compatibility I think

[01:31:04.0000] <Hixie>
k

[01:31:05.0000] <othermaciej>
but not sure

[01:31:06.0000] <othermaciej>
I looked at the DOM 1 Core failures

[01:32:00.0000] <othermaciej>
3 are because we allow implicit adoption when inserting a node from another document

[01:32:01.0000] <othermaciej>
which we added because Gecko allowed it and some sites started to depend on it

[01:32:02.0000] <othermaciej>
(or maybe it was intranet "enterprise" products w/ a web interface)

[01:33:00.0000] <othermaciej>
one is due to returning null for getAttribute of a nonexistent attribute

[01:33:01.0000] <Hixie>
yeah we should just allow that

[01:33:02.0000] <Hixie>
i don't understand why implicit adoption wasn't allowed

[01:33:03.0000] <Hixie>
and getAttribute's null thing needs fixing

[01:33:04.0000] <Hixie>
in the spec

[01:33:05.0000] <othermaciej>
yes, yes it does

[01:34:00.0000] <Hixie>
Web DOM Core 4

[01:34:01.0000] <othermaciej>
asking for that to be fixed was one of my first unpleasant experiences with the w3c

[01:34:02.0000] <Hixie>
or 5 i guess :-)

[01:34:03.0000] <othermaciej>
many years ago

[01:34:04.0000] <Hixie>
oh?

[01:34:05.0000] <othermaciej>
let me see if I can find the thread in the archives

[01:35:00.0000] <othermaciej>
ah, the two other failures are also due to not throwing WRONG_DOCUMENT_ERR

[01:35:01.0000] <othermaciej>
so 5 failures for allowing implicit adoption, one for returning null from getAttribute

[01:35:02.0000] <othermaciej>
I can tell you that last time I tested, every other browser failed a lot more of these

[01:35:03.0000] <othermaciej>
even the DOM1 Core tests

[01:36:00.0000] <othermaciej>
I think the DOM2EV test we fail is due to dispatching events in both capture and bubble phases to the target node

[01:36:01.0000] <othermaciej>
(technically they should only bubble, but that causes web compat issues due to Gecko's longstanding behavior in this regard)

[01:37:00.0000] <Hixie>
yeah, i'm not testing that one either

[01:37:01.0000] <Hixie>
i'm not testing anything that we want to change

[01:38:00.0000] <othermaciej>
http://lists.w3.org/Archives/Public/www-dom/2005OctDec/0024.html

[01:38:01.0000] <othermaciej>
http://lists.w3.org/Archives/Public/www-dom/2005OctDec/0025.html

[01:38:02.0000] <othermaciej>
both had floods of angry responses

[01:39:00.0000] <othermaciej>
here's another interesting message from around the same time: http://lists.w3.org/Archives/Public/www-dom/2005OctDec/0076.html

[01:39:01.0000] <othermaciej>
(incompatibility between DOM 2 Core and DOM 3 Core)

[01:40:00.0000] <othermaciej>
(on a minor extremely boring edge case issue though)

[01:46:00.0000] <othermaciej>
hey, proof that I'm stubbornly consistent in my beliefs: http://lists.w3.org/Archives/Public/www-dom/2005OctDec/0047.html

[01:52:00.0000] <Hixie>
proof that i am: http://lists.w3.org/Archives/Public/www-html/1998Dec/0003.html

[01:57:00.0000] <Hixie>
heh, i found that html4 had no real conformance criteria back in 1999: http://lists.w3.org/Archives/Public/www-html/1999Feb/0010.html

[01:59:00.0000] <othermaciej>
heh

[01:59:01.0000] <othermaciej>
well, we have higher standards now

[02:00:00.0000] <Hixie>
it's sad that from 1998 to 2004 we basically made no progress

[02:00:01.0000] <Hixie>
i'm glad we have finally shaken ourselves free of the old guard

[02:00:02.0000] <othermaciej>
well, some of that was the state of affairs in the browser market during that period

[02:23:00.0000] <Hixie>
i need more tests

[02:23:01.0000] <Hixie>
let me know if you come up with any

[02:23:02.0000] <Hixie>
bed time now

[05:42:00.0000] <Philip`>
http://www.natural-innovations.com/wa/xhtml.html - hmm, XHTML was sometimes referred to as HTML 5.0?

[05:46:00.0000] <webben>
Philip`: http://groups.google.com/group/mail.www-html/browse_thread/thread/140306dc90e21ce3/a0bef1d6fde4d332?lnk=st&q=voyager+w3c+%22html+5.0%22#a0bef1d6fde4d332 perhaps

[05:57:00.0000] <Philip`>
webben: Ah, thanks

[05:58:00.0000] <Philip`>
http://www.w3.org/TR/1998/WD-html-in-xml-19981205/ talks about Voyager, but I can't find any 'official' references to an HTML 5.0 so I guess that term just fell out of favour before catching on

[14:21:00.0000] <Philip`>
http://intertwingly.net/blog/?q=%00 - yay for XML

[14:25:00.0000] <Philip`>
(Also http://golem.ph.utexas.edu/instiki/search?query=%00 etc)

[14:26:00.0000] <Philip`>
(Also http://golem.ph.utexas.edu/instiki/save/HomePage etc)

[14:26:01.0000] <Philip`>
Are there any non-trivial dynamically-generated XHTML sites that successfully ensure well-formedness?

[14:32:00.0000] <hsivonen>
Philip`: Validator.nu used to be a dynamically-generated app (even if not site) that ensured well-formedness, but now it may have holes like that--I have not tested properly for forbidden characters after I did some major rework in the front end

[14:33:00.0000] <hsivonen>
bullet-proofing it against \0, etc. is one of the many items I should take care of

[14:34:00.0000] <hsivonen>
actually, I could find out right now

[14:35:00.0000] <hsivonen>
yep, Validator.nu is broken after the major rework :-(

[14:35:01.0000] <hsivonen>
my options are:

[14:35:02.0000] <hsivonen>
1) Hacking the Xalan serializer (hard)

[14:35:03.0000] <hsivonen>
2) Writing a SAX filter (easy but potentially inefficient)

[14:35:04.0000] <hsivonen>
and 

[14:36:00.0000] <hsivonen>
3) Writing my own XML serializer (tedious with namespaces)

[14:38:00.0000] <hsivonen>
hmm. and 4) filtering in a Writer after the serializer

[14:39:00.0000] <Philip`>
Those all sound like great fun

[15:49:00.0000] <G0k>
oy all

[15:49:01.0000] <G0k>
are there any test cases for cross-document messaging?


2007-12-30
[16:04:00.0000] <jgraham>
G0k: I'm not aware of any off the top of my head

[16:05:00.0000] <jgraham>
But I have rather less than encyclopedic knowledge on the subject

[16:06:00.0000] <G0k>
k

[17:37:00.0000] <Hixie>
/me learns about setAttributeNodeNS

[17:37:01.0000] <Hixie>
wtf

[19:06:00.0000] <weinig>
Hixie: for real, even the DOM 3 spec doesn't seem to really understand what it is for

[20:15:00.0000] <Hixie>
41 tests so far

[20:17:00.0000] <MacDome>
sah

[20:17:01.0000] <MacDome>
sigh

[20:17:02.0000] <MacDome>
rather

[20:17:03.0000] <Hixie>
hm?

[20:19:00.0000] <MacDome>
Hixie: was just looking at our current failure status

[20:19:01.0000] <Hixie>
did it get worse?

[20:19:02.0000] <MacDome>
yeah, 87 instead of 88

[20:19:03.0000] <Hixie>
hehe

[20:19:04.0000] <MacDome>
Hixie: some of these are intentionally "broken" in Safari: http://bugs.webkit.org/show_bug.cgi?id=6179

[20:19:05.0000] <Hixie>
hmm, 59 tests still for me to write

[20:22:00.0000] <Hixie>
MacDome: i'm willing to take keywords out if you think some should be taken out

[20:23:00.0000] <Hixie>
MacDome: what do you suggest?

[20:23:01.0000] <MacDome>
Hixie: well, that bug links to various test cases which FF and WK prepared to test reserved words, and we tested against IE to match IE

[20:23:02.0000] <Hixie>
i removd the future reserved words

[20:24:00.0000] <MacDome>
I believe we match IE: https://bugzilla.mozilla.org/attachment.cgi?id=206492&action=view

[20:25:00.0000] <MacDome>
interesting: http://developer.mozilla.org/en/docs/ECMAScript_DontEnum_attribute

[20:29:00.0000] <Hixie>
interesting indeed

[20:36:00.0000] <MacDome>
Hixie: did you find some list of what .constructor properties should be DontDelete or ReadOnly?

[20:36:01.0000] <MacDome>
they're all DontEnum

[20:37:00.0000] <Hixie>
i only tested Function's

[20:37:01.0000] <MacDome>
Hixie: I'll make a more comprehensive test and fix them all if I can find a list

[20:40:00.0000] <MacDome>
/me is confused by what (new Function).prototype would mean

[20:40:01.0000] <MacDome>
I guess that's the function prototype, which is differnet from the function constructor

[20:41:00.0000] <MacDome>
but is (new Function).prototype == Function.prototype? I would think not.

[20:41:01.0000] <MacDome>
/me always gets turned around by JS

[20:43:00.0000] <kig>
new Function().prototype, the prototype property of the newly-created function object (i.e. the prototype to be used when the function is used as a constructor)

[20:43:01.0000] <kig>
...i think

[20:44:00.0000] <Hixie>
(new Function).prototype is null, no?

[20:44:01.0000] <Hixie>
or am i thinking of (new function() { }).prototype

[20:45:00.0000] <takkaria>
/me chuckles

[20:45:01.0000] <kig>
it's an empty object

[20:46:00.0000] <kig>
(at least in firefox)

[20:46:01.0000] <kig>
new Function().__proto__ == Function.prototype => true, new Function().prototype == Function.prototype => false

[20:47:00.0000] <MacDome>
yeah

[20:47:01.0000] <MacDome>
Hixie: so you're grabbing at (function() {}).prototype

[20:48:00.0000] <MacDome>
ah

[20:54:00.0000] <MacDome>
/me doesn't know what (function() {}).prototype means

[20:55:00.0000] <Hixie>
i've no idea what specifically you are talking about, btw

[20:58:00.0000] <MacDome>
Hixie: I'm just tryign to relearn prototypes in JS for the 18th time.  /me goes back to reading http://mckoss.com/jscript/object.htm

[20:58:01.0000] <MacDome>
/me gets confused by .constructor vs. .prototype vs. .__proto__

[21:00:00.0000] <MacDome>
/me thinks .__proto__ == .contructor.prototype

[21:15:00.0000] <Hixie>
  __proto__ is [[Prototype]] in the spec

[22:37:00.0000] <Hixie>
46 done, 54 to go

[23:37:00.0000] <heycam>
Hixie, what spec is __proto__ in?  i couldn't see it in ecma-262.

[23:37:01.0000] <heycam>
is it a JavaScript-specific thing?

[23:37:02.0000] <Hixie>
it's proprietary, yes

[23:38:00.0000] <heycam>
which spec?

[23:42:00.0000] <heycam>
/me returns to beer

[23:55:00.0000] <Hixie>
it's proprietary, as in, no spec

[01:00:00.0000] <hsivonen>
Philip`: Validator.nu is now non-XML-character-proof: http://validator.nu/?out=xhtml&doc=http%3A%2F%2Fgolem.ph.utexas.edu%2Finstiki%2Fsearch%3Fquery%3D%2500&showsource=yes

[05:52:00.0000] <Philip`>
hsivonen: Looks like that site is now non-XML-character-proof (at least to some extent) so it doesn't work as a demo any more :'-(

[05:57:00.0000] <Philip`>
hsivonen: By the way, when a page like http://golem.ph.utexas.edu/instiki/search?query=%00 returns XHTML in the error response, it would be very helpful if the validator could validate that XHTML (instead of simply complaining about the HTTP status)

[05:59:00.0000] <hsivonen>
Philip`: yeah, validating error pages is on the feature request list

[07:16:00.0000] <hsivonen>
what's the point of SVG Print?

[07:21:00.0000] <Philip`>
hsivonen: Replacing PostScript in printers? (That's just a totally uninformed guess based on the name)

[07:29:00.0000] <hsivonen>
Philip`: is there a good technical reason to seek to replace PostScript or PDF as the means of driving printers?

[07:31:00.0000] <hsivonen>
It seems to me that a RIP that eats PDF 1.4 is the peak of development as far as driving printers goes

[07:34:00.0000] <Philip`>
Looks like Canon is quite involved in editing SVG Print - is it something that's been driven by them, or do other companies care too?

[07:47:00.0000] <mpt>
So Adobe pushed SVG as an alternative to Flash/Flex, until they bought Macromedia

[07:48:00.0000] <mpt>
Now Canon is pushing SVG Print as an alternative to PDF?

[07:50:00.0000] <hsivonen>
and HP has been involved in XHTML-Print...

[07:50:01.0000] <mpt>
My irony meter is warbling

[07:55:00.0000] <hsivonen>
seeking to replace Flash makes sense because Flash is a single-vendor technology

[07:55:01.0000] <hsivonen>
seeking to replace PDF makes no sense

[07:56:00.0000] <hsivonen>
since PDF is a remarkably interoperably implemented multivendor technology in practice 

[07:56:01.0000] <hsivonen>
even if the spec comes from a single vendor

[08:07:00.0000] <mpt>
So the next illogical step would be for Quark to jump on board

[14:44:00.0000] <gsnedders>
/me wonders why he has no email from whatwg@whatwg since the 20th

[14:51:00.0000] <Lachy__>
gsnedders, check that you're still subscribed and then check that there isn't some spam filter blocking it somewhere

[14:52:00.0000] <gsnedders>
broken gmail filters trying to apply non-existant labels, and archiving it.

[14:52:01.0000] <Lachy__>
ok

[14:52:02.0000] <gsnedders>
fix the filters and everything appears

[15:24:00.0000] <jgraham>
So... slightly further off topic than the keyboard discussion in public-html; does anyone have any good/bad experience with the in ear sound isolating styl earbud headphones

[15:24:01.0000] <jgraham>
?

[15:25:00.0000] <jgraham>
(add extra letters where required)

[15:26:00.0000] <anne-mac>
i bought a set and they didn't really stay in my ear so I gave them away...

[15:26:01.0000] <anne-mac>
(if you're referring the ones sold for the iPod among other things)

[15:26:02.0000] <jgraham>
do you remember which ones?

[15:27:00.0000] <jgraham>
(things like http://www.westone.com/content/215.html http://www.shure.com/PersonalAudio/Products/Earphones/ESeries/us_pa_E2c_content )

[15:27:01.0000] <anne-mac>
http://www.amazon.com/Apple-M9394G-B-iPod-Headphones/dp/B0001AP8CE

[15:29:00.0000] <anne-mac>
my experience was similar to this one: http://www.amazon.com/review/R12VD3WYE1JU5/ref=cm_cr_rdp_perm

[15:29:01.0000] <anne-mac>
although I did not bring them back

[15:30:00.0000] <takkaria>
jgraham: I have some shure e2cs

[15:31:00.0000] <jgraham>
Yeah, the general issue with these things seems to be finding ones that fit comfortably. 

[15:31:01.0000] <takkaria>
they're pretty good

[15:31:02.0000] <takkaria>
they come with three different sizes of three different materials of earbud

[15:31:03.0000] <jgraham>
takkaria: Do they provide noticeable reduction in background noise?

[15:32:00.0000] <takkaria>
yup

[15:32:01.0000] <jgraham>
/me cares more about reducing background noise than about perfect audio quality

[15:32:02.0000] <takkaria>
they allow me to walk around town happily without my ipod ast full volume

[15:32:03.0000] <jgraham>
Excellent

[15:33:00.0000] <takkaria>
around the middle of Manchester, nothing less

[15:33:01.0000] <takkaria>
I know other people who've had good luck with the e2cs, no-one who's used the apple ones

[15:34:00.0000] <jgraham>
/me tries hard to avoid making some remark about the general danger of walking around manchester, let alone with one sense compromised 

[15:34:01.0000] <inimino>
I tried some Ultimate Ears ones but they did not fit my ear canals

[15:34:02.0000] <jgraham>
My interest in the Westone ones stems from a site that said they were the most comfortable fit for most people

[15:35:00.0000] <takkaria>
nothing dangerous about walking round manchester

[15:35:01.0000] <takkaria>
cycling is a different matter, but I don't wear earphones for that. :)

[15:35:02.0000] <jgraham>
Unfortunately, they seem to be relatively hard to find in the UK 

[15:35:03.0000] <jgraham>
takkaria: :)

[15:38:00.0000] <jgraham>
Thanks everyone :)

[15:48:00.0000] <takkaria>
np

[15:59:00.0000] <virtuelv>
jgraham: have you looked into the Shure headphones?


2007-12-31
[16:01:00.0000] <virtuelv>
nm, /me reads backlog

[16:01:01.0000] <virtuelv>
I have some Koss Spark Plugs lying around

[16:01:02.0000] <virtuelv>
they're pretty horrible, and I've just always ended up using my Grados instead

[16:01:03.0000] <virtuelv>
which are everything but in-ear

[16:01:04.0000] <jgraham>
virtuelv: If you have any insight I'm still very much interested

[16:02:00.0000] <virtuelv>
jgraham: sound-quality-wise, they are supposed to be among the better in-ear solutions you can get

[16:02:01.0000] <jgraham>
Which ones?

[16:02:02.0000] <virtuelv>
problem is, they have an associated price tag

[16:02:03.0000] <virtuelv>
jgraham: shure in general

[16:02:04.0000] <virtuelv>
I only know shure from microphones

[16:03:00.0000] <virtuelv>
(the SM58 is a fantastic piece of kit, but that's beside the point)

[16:04:00.0000] <jgraham>
My audiophile credentials are severely limited so I wonder how much sound quality issues will affect me

[16:05:00.0000] <virtuelv>
jgraham: while there is a lot of bollock science in audiophile circles, they at least have understood the quality of not getting tired listening to headphones

[16:06:00.0000] <virtuelv>
my golden rule is something like this: Don't buy audio equipment by mail order, unless you have pre-listened to them extensively

[16:06:01.0000] <virtuelv>
example: I recently (a week ago), got rid of my old DVD/CD player, and bought one with a darker sound

[16:07:00.0000] <virtuelv>
less impressive at first, but it allows me to listen for more extensive periods of time

[16:07:01.0000] <virtuelv>
(in this regard, ipods suck, they cause listening fatigue)

[16:07:02.0000] <virtuelv>
so it's not really about being "audiophile", it's about "which pair of headphones would I want to use for eight hours on end"

[16:08:00.0000] <gavin>
I don't think I would ever want to use headphones for 8 hours on end regardless of how comfortable they were :)

[16:09:00.0000] <jgraham>
virtuelv: whilst that sounds eminently sensible, I have difficulty imaganing where I could go to try on in ear headphones

[16:09:01.0000] <virtuelv>
jgraham: your nearest hi-fi-pusher

[16:09:02.0000] <virtuelv>
and by that, I mean people who are serious about it

[16:09:03.0000] <virtuelv>
fwiw, http://www.stereophile.com/headphones/504shure/

[16:11:00.0000] <virtuelv>
you just have to get over the possible trauma of having someone else's earwax shared

[16:13:00.0000] <jgraham>
I guess it's worth a try.

[16:13:01.0000] <virtuelv>
jgraham: is there any specific reason for going with in-ear?

[16:15:00.0000] <jgraham>
virtuelv: They seem to be effective at cutting down background noise. At least that's the impression I get. 

[16:16:00.0000] <virtuelv>
jgraham: yes, but so are active solutions, like Sennheiser PXC 450

[16:16:01.0000] <virtuelv>
looks ridiculous walking around town, but on flights and similar, they are guaranteed to be much more comfortable

[16:16:02.0000] <virtuelv>
http://www.sennheiser.com/nl/icm_nl.nsf/root/500643

[16:18:00.0000] <virtuelv>
350's are slightly cheaper, but I haven't read any reviews yet

[16:18:01.0000] <jgraham>
virtuelv: There seem to be to problems with the active noise cancelling things - they are comparatively expensive for the same level of background reduction, and the active cancelling /seems/ to be more low-frequency-specific

[16:18:02.0000] <jgraham>
(I was originally looking at the active noise canceling ones)

[16:21:00.0000] <jgraham>
/me -> bed

[16:21:01.0000] <virtuelv>
as long as you are aware of the disadvantages, though

[16:21:02.0000] <virtuelv>
listening fatigue being one

[16:22:00.0000] <virtuelv>
it's about optimizing the solution for the use-case

[20:29:00.0000] <doublec>
<erg> git pull http://littledan.onigirihouse.com/factor.git

[20:29:01.0000] <doublec>
you can also: git pull git://onigirihouse.com/git/littledan.git

[20:29:02.0000] <doublec>
it's faster

[20:37:00.0000] <doublec>
oops, wrong channel

[20:52:00.0000] <Hixie>
53 more tests to write

[20:53:00.0000] <Hixie>
anyone got any tests they know that safari, mozilla, or opera fail?

[20:57:00.0000] <G0k>
http://www.css3.info/selectors-test/test.html fails some in safari

[20:57:01.0000] <Hixie>
wow, sweet test suite

[20:57:02.0000] <G0k>
and firefox

[20:58:00.0000] <Hixie>
i really want primarily dom and js tests, but these will definitely be useful

[21:01:00.0000] <G0k>
by js testing, do you mean like testing compliance with the ECMAScript spec?

[21:04:00.0000] <G0k>
oh and of course http://bugs.webkit.org/attachment.cgi?id=9554

[21:04:01.0000] <G0k>
which i think only opera handles nicely right now

[21:04:02.0000] <Hixie>
yes

[21:05:00.0000] <Hixie>
JS spec and DOM specs

[21:07:00.0000] <G0k>
i actually had a question for you about the same-origin policy

[21:07:01.0000] <G0k>
should you be able enumerate variables across domains?

[21:07:02.0000] <Hixie>
webkit gets http://bugs.webkit.org/attachment.cgi?id=9554 correct

[21:08:00.0000] <Hixie>
(though i initially was confused sicne red and yellow are typically fail colours :-) )

[21:08:01.0000] <Hixie>
no

[21:08:02.0000] <Hixie>
or wait

[21:08:03.0000] <Hixie>
what do you mean?

[21:08:04.0000] <G0k>
hixie: they re-broke it in the latest nightly

[21:08:05.0000] <G0k>
ok, so

[21:08:06.0000] <G0k>
well actually, test case: http://bugs.webkit.org/attachment.cgi?id=9554

[21:08:07.0000] <G0k>
erps

[21:08:08.0000] <G0k>
http://mapseekret.com/staticmedia/document_a.html

[21:08:09.0000] <G0k>
that

[21:09:00.0000] <G0k>
basically, should a document from one domain be able to use a for..in loop to find the names of the variables in the window from another domain?

[21:09:01.0000] <G0k>
opera and firefox don't let you

[21:09:02.0000] <G0k>
safari currently does

[21:09:03.0000] <G0k>
safari does stop you from getting the actually values, but not the variable names

[21:09:04.0000] <G0k>
well actually, i put it better in this bug: http://bugs.webkit.org/show_bug.cgi?id=16387

[21:10:00.0000] <G0k>
but is that actually a same-origin violation?

[21:10:01.0000] <Hixie>
i thought i was using the latest nightly

[21:10:02.0000] <Hixie>
the second test case does indeed fail

[21:10:03.0000] <Hixie>
/me takes note of it

[21:11:00.0000] <G0k>
well alright the latest one i just compiled is a little wonkey in general

[21:11:01.0000] <Hixie>
waiiit

[21:11:02.0000] <Hixie>
i thought 9554 was a different attachment

[21:11:03.0000] <Hixie>
but it's the same one

[21:11:04.0000] <Hixie>
and it failed once and succeeded the other time

[21:11:05.0000] <Hixie>
wtf

[21:11:06.0000] <G0k>
are you sure you're not mixing up builds?

[21:11:07.0000] <Hixie>
re the same-origin thing, yes that's a security bug

[21:11:08.0000] <Hixie>
no, only one build

[21:12:00.0000] <G0k>
kdoke

[21:18:00.0000] <MacDome>
/me is surprsied we'd break an existing test

[21:18:01.0000] <MacDome>
would suggest that the test never got landed

[21:21:00.0000] <Hixie>
it seems intermittent

[21:21:01.0000] <MacDome>
http://build.webkit.org/waterfall is generally quite good about catching regressions

[22:57:00.0000] <weinig>
Hixie: http://bugs.webkit.org/show_bug.cgi?id=16387 :)

[22:57:01.0000] <weinig>
Hixie: that's what you get for commenting in a bug that I should have fixed already

[22:57:02.0000] <G0k>
damnit, my test suggestion is foiled

[22:58:00.0000] <G0k>
i should get hixie to CC all my bugs. :)

[00:11:00.0000] <Hixie>
can anyone think of a way of detecting if an element is matching a selector without using some 2004-or-earlier standards-based mechanism other than getComputedStyle()?

[00:14:00.0000] <MacDomeSleep>
Hixie: well, in IE there is _hasLayout :)

[00:14:01.0000] <MacDomeSleep>
certain styles would certainly change the DOM behavior

[00:14:02.0000] <Hixie>
maybe you don't fully understand "2004-or-earlier standards-based mechanism" :-P

[00:14:03.0000] <Hixie>
er wait

[00:14:04.0000] <Hixie>
i misspoke

[00:14:05.0000] <Hixie>
i meant to ask "can anyone think of a way of detecting if an element is matching a selector using some 2004-or-earlier standards-based mechanism other than getComputedStyle()?"

[00:15:00.0000] <Hixie>
the without was a misedit

[00:15:01.0000] <Hixie>
appologies

[00:15:02.0000] <MacDomeSleep>
yeah,I understood the firstime

[00:15:03.0000] <MacDomeSleep>
Imust have read betweent he lines

[00:16:00.0000] <MacDomeSleep>
Hixie: I would use a style which I knew changed dom behavior

[00:16:01.0000] <MacDomeSleep>
things like height

[00:16:02.0000] <MacDomeSleep>
display

[00:16:03.0000] <MacDomeSleep>
etc.

[00:17:00.0000] <Hixie>
right but how can i tell the style has changed, without using getComputedStyle?

[00:17:01.0000] <Hixie>
and still following the specs?

[00:17:02.0000] <Hixie>
(from script)

[00:18:00.0000] <MacDomeSleep>
if you set a height with the selector, wouldn't element.height change?

[00:18:01.0000] <Hixie>
element.innerHeight isn't standards-based

[00:19:00.0000] <MacDomeSleep>
http://trac.webkit.org/projects/webkit/browser/trunk/WebCore/dom/Element.idl

[00:20:00.0000] <MacDomeSleep>
I can't think of a standards based method

[00:21:00.0000] <MacDomeSleep>
Hixie: since all the useful properties on element in that file, are non-standards based

[00:21:01.0000] <Hixie>
bummer

[00:21:02.0000] <Hixie>
i don't like breaking IE on a whole series of tests just for not having getComputedStyle

[00:23:00.0000] <MacDomeSleep>
wasn't getComputedStyle back in like 2000?

[00:23:01.0000] <Hixie>
yeah

[00:23:02.0000] <Hixie>
i'm totally using it at least once

[00:23:03.0000] <MacDomeSleep>
http://www.w3.org/TR/DOM-Level-2-Style/css.html#CSS-ViewCSS

[00:23:04.0000] <Hixie>
just feels a bit over the top to totally screw IE on 16 tests in a row just for one bug

[00:24:00.0000] <Hixie>
oh well

[00:24:01.0000] <Hixie>
i'm not really targetting IE anyway

[00:24:02.0000] <MacDomeSleep>
it sorta sucks that IE screws developers over based on one bug :)

[00:24:03.0000] <MacDomeSleep>
or rather... one thousand bugs

[00:24:04.0000] <Hixie>
me giggles as he discovers his selectorTest() function works

[00:24:05.0000] <Hixie>
^/

[00:27:00.0000] <Hixie>
ok this is awesome

[00:27:01.0000] <Hixie>
i have a little framework for selectors tests

[00:27:02.0000] <Hixie>
now i can go crazy!

[00:27:03.0000] <Hixie>
mwuhahahaha

[00:27:04.0000] <G0k>
he's a mad man

[00:27:05.0000] <G0k>
someone sedate him

[00:28:00.0000] <MacDome>
/me flails wildly in Hixie's direction, utterly failing.

[00:28:01.0000] <MacDome>
/me ends up sedating himself

[00:28:02.0000] <Hixie>
heh

[03:28:00.0000] <Philip`>
Hixie: If you set an img width with CSS, then get the width DOM attribute and test if it equals the width set with CSS (to see if the selector matched), is that standards-based or is it just an undocumented feature?

[03:29:00.0000] <anne-mac>
standards-based per HTML5

[03:39:00.0000] <hsivonen>
IIRC, what the getComputedStyle spec says about returned units is utterly bogus

[03:39:01.0000] <anne-mac>
that's different from the width DOM attribute of <img> though

[03:39:02.0000] <hsivonen>
so if you want to build a test that is both Good for the Web and adheres to specs as of 2004, it might be hard

[03:40:00.0000] <anne-mac>
oh, like that

[03:40:01.0000] <anne-mac>
it's also pretty vague, so...

[03:40:02.0000] <anne-mac>
/me hopes to fix getComputedStyle in due course

[03:41:00.0000] <hsivonen>
"The CSSStyleDeclaration is read-only and contains only absolute values."

[03:41:01.0000] <hsivonen>
my recollection was that it said absolute lengths

[03:41:02.0000] <hsivonen>
if absolute values means something else, please disregard my comment above

[03:42:00.0000] <hsivonen>
IIRC, the sensible approach is to normalize lengths to px which CSS calls "relative" not "absolute"

[03:44:00.0000] <anne-mac>
per CSS the computed value for border-top-width is the absolute length...

[03:44:01.0000] <anne-mac>
so I guess that's what getComputedStyle means too

[03:44:02.0000] <anne-mac>
it's not what getComputedStyle does though

[05:54:00.0000] <webben>
hsivonen: btw validator (local build) is giving me a "Caused by: org.xml.sax.SAXException: Malformed spec: Expected dt to be context dt but it was not." error whenever I run it, in case you know what that's about

[06:31:00.0000] <hsivonen>
webben: try starting with --html5load=http://about.validator.nu/spec.html

[06:31:01.0000] <hsivonen>
webben: evidently, I need to bundle a known-good copy of the spec with the software

[06:32:00.0000] <hsivonen>
webben: sorry about the inconvenience

[06:32:01.0000] <webben_>
python build/build.py run --html5load=http://about.validator.nu/spec.html gives me the same error.

[06:33:00.0000] <hsivonen>
Philip`: I pushed out a vastly updated http://about.validator.nu/

[06:33:01.0000] <hsivonen>
webben: you need python build/build.py --html5load=http://about.validator.nu/spec.html run

[06:33:02.0000] <anne-mac>
hsivonen, you missed the other error Philip` pointed out

[06:33:03.0000] <anne-mac>
"but only it if the DTD"

[06:34:00.0000] <hsivonen>
anne-mac: thanks

[06:34:01.0000] <webben_>
hsivonen: Ah great, that works. Thank you. :)

[06:34:02.0000] <hsivonen>
anne-mac: fixed

[07:35:00.0000] <zcorpan>
div[align=left] {

[07:35:01.0000] <zcorpan>
}

[07:35:02.0000] <zcorpan>
<div align='LEFT'></div>

[07:35:03.0000] <zcorpan>
The CSS selector should match the HTML fragment because the value of the align attribute should be case insensitive in a HTML document

[07:35:04.0000] <zcorpan>
-- http://www.css3.info/selectors-test/test-attribute-equal.html#attribute-equal

[07:35:05.0000] <zcorpan>
hmm

[07:36:00.0000] <zcorpan>
html5 needs to define this

[07:36:01.0000] <zcorpan>
Hixie: ^

[07:38:00.0000] <zcorpan>
/me would like to run the selectors tests in quirks mode

[07:41:00.0000] <zcorpan>
bah, it doesn't test case sensitivity of .class

[07:48:00.0000] <dbaron>
zcorpan, There's a list of which attributes count for that in CSS2.1, I think.

[07:48:01.0000] <dbaron>
zcorpan, although actually it's defined in HTML4 as well...

[07:49:00.0000] <zcorpan>
i think html5 needs to define it

[07:49:01.0000] <zcorpan>
also for xhtml5

[07:50:00.0000] <dbaron>
hrm, actually not -- we just implement using the [CI] / [CS] in HTML4

[07:51:00.0000] <zcorpan>
does that mean ascii case insensitive or unicode case insensitive?

[07:52:00.0000] <zcorpan>
usemap="" is unicode case insensitive, iirc. some others are ascii case insensitive. although i haven't tested their case sensitivity wrt selectors

[07:53:00.0000] <hsivonen>
zcorpan: why is usemap case-insensitive in any way?

[07:54:00.0000] <zcorpan>
hsivonen: dunno

[07:54:01.0000] <gsnedders>
hsivonen: aren't @id normally case-insensitive?

[07:54:02.0000] <zcorpan>
usemap points to name=""

[07:54:03.0000] <hsivonen>
zcorpan: have you tested selector case-insensitivity of with enumerated-value attributes in XHTML?

[07:54:04.0000] <zcorpan>
(per html4 anyway)

[07:54:05.0000] <zcorpan>
hsivonen: no

[07:55:00.0000] <hsivonen>
gsnedders: id is supposed to be case-sensitive

[07:55:01.0000] <gsnedders>
hsivonen: and HTML is supposed to be SGML. What things are supposed to mean is irrelevant :)

[07:55:02.0000] <hsivonen>
zcorpan: is usemap case-insensitive only when comparing againts name or also with id?

[07:56:00.0000] <zcorpan>
i guess i'll make a thorough test suite for this in due course

[07:56:01.0000] <zcorpan>
hsivonen: in ie and opera it's insensitive for both

[07:56:02.0000] <zcorpan>
hsivonen: in saf and moz it's insensitive for name (html only) and sensitive for id (xhtml only)

[07:57:00.0000] <hsivonen>
I currently implement usemap checking using code-point-for-code-point equality 

[07:58:00.0000] <hsivonen>
zcorpan: fun

[07:58:01.0000] <hsivonen>
zcorpan: but WebKit supports name in XHTML, too, doesn't it?

[07:59:00.0000] <gsnedders>
hsivonen: yeah

[09:25:00.0000] <zcorpan>
hsivonen: not last time i tested, iirc

[09:30:00.0000] <zcorpan>
hsivonen: "If you the automatic choice of parser ..." s/you // in http://about.validator.nu/

[09:32:00.0000] <zcorpan>
hsivonen: btw, i'm not sure about the utility of allowing text/xsl as a lax content type, considering that iirc moz and ie don't actually accept it as a content type

[09:32:01.0000] <zcorpan>
(but ie requires "text/xsl" in the type pseudo-attribute)

[09:33:00.0000] <webben_>
/me can understand why one might want to treat name as ID in HTML for IE compat, but can't see the point in corrupting NAME in XHTML in the same way.

[09:34:00.0000] <zcorpan>
webben_: for HTML compat :)

[09:34:01.0000] <webben_>
It's a silly habit in HTML to begin with, given that NAME is not unique.

[09:35:00.0000] <zcorpan>
<a name> or <map name> must be unique per html4

[09:35:01.0000] <zcorpan>
<div id="foo"><map name="foo"> is not conforming html4

[09:35:02.0000] <webben_>
zcorpan: Do you mean getElementById never returns <input type="radio" name="foo"> ?

[09:36:00.0000] <zcorpan>
that's not what i meant, and i don't see how you make that connection, but it might well do in ie

[09:37:00.0000] <zcorpan>
(it shouldn't, though)

[09:37:01.0000] <webben_>
zcorpan: If it's not the case that getElementById is special-cased for A and MAP, then I'm not sure how that uniqueness restriction helps.

[09:38:00.0000] <Dashiva>
webben_: How is getElementById on a duplicate id any different from on a duplicate name?

[09:38:01.0000] <webben_>
Dashiva: You're not supposed to have duplicate IDs.

[09:38:02.0000] <Dashiva>
That's not enforced, though

[09:38:03.0000] <zcorpan>
webben_: there are other interfaces than gEBI

[09:38:04.0000] <webben_>
Not much is enforced.

[09:38:05.0000] <webben_>
You are supposed to have duplicate NAMEs for radio buttons however.

[09:38:06.0000] <Dashiva>
Besides, many uses of name are also supposed to be unique

[09:38:07.0000] <zcorpan>
webben_: such as <img usemap>

[09:40:00.0000] <Dashiva>
webben_: There's not much of a use case for using gEBI on a radio button name, though

[09:41:00.0000] <webben_>
I don't see any usecase for using gEBI on NAMEs full-stop, especially in XML.

[09:41:01.0000] <Dashiva>
Easy access to form elements is one

[09:41:02.0000] <webben_>
I can see the backwards compat argument in HTML, although it's not massively strong given that relying on such behavior would break webapps in a lot of browsers.

[09:41:03.0000] <zcorpan>
Dashiva: dom0 has simpler access to form elements

[09:42:00.0000] <webben_>
I think ID makes form access pretty easy actually.

[09:42:01.0000] <Dashiva>
Sure, if you have a reference to the form

[09:42:02.0000] <webben_>
you need ID for form labels anyhow

[09:42:03.0000] <webben_>
(well, if you want them to actually work)

[09:42:04.0000] <webben_>
*field labels

[09:43:00.0000] <Dashiva>
If you use labels, and don't use containing labels, yes

[09:43:01.0000] <webben_>
Either way, I think that's a stronger argument for a getElementByAttributeValue('name','foo') or something

[09:43:02.0000] <Dashiva>
getElementsByName exists, yes

[09:43:03.0000] <webben_>
at least that would do what it says on the tin, and be usable for many more situations

[09:44:00.0000] <webben_>
e.g. getting all password inputs in a form.

[09:44:01.0000] <webben_>
s/password/text (probably a better example)

[09:45:00.0000] <zcorpan>
text inputs might not have a type attribute :)

[09:46:00.0000] <webben_>
They might not. That's an argument for something like getElementsBy (http://developer.yahoo.com/yui/docs/YAHOO.util.Dom.html)

[15:31:00.0000] <gsnedders>
http://geekninja.blogspot.com/2007/12/html5s-canvas-tag-are-we-using-it.html

[15:42:00.0000] <kig>
as webcore is lgpl and canvas is a part of webcore and lgpl has a patent release clause...

[15:49:00.0000] <Hixie>
Philip`: .width doesn't just reflect the content attribute?

[15:51:00.0000] <Philip`>
Hixie: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cimg%20src%3Dimage%20style%3Dwidth%3A100%25%20onload%3Dw(this.width)%3E

[15:51:01.0000] <Philip`>
At least Opera and Firefox return the rendered width, and I think I remember IE doing that too

[15:51:02.0000] <Hixie>
well crap

[15:51:03.0000] <Hixie>
does html5 require that?

[15:51:04.0000] <Philip`>
I think it does

[15:52:00.0000] <Philip`>
HTML5 says "The DOM attributes height and width must return the rendered height and width of the image, in CSS pixels, if the image is being rendered, and is being rendered to a visual medium, or 0 otherwise. [CSS21]"

[15:53:00.0000] <Philip`>
(As far as I could see when I last wanted to do this, it's impossible to find the actual bitmap size of the image, which is a bit irritating)

[15:54:00.0000] <Philip`>
((except for maybe adding it to the document and having it rendered and then seeing how big it is, but that's nasty))