00:55
<Hixie>
can i drop <textarea accept="">?
00:55
<Hixie>
or is anyone actually going to implement anything useful with it?
01:49
<erlehmann_>
Hixie: i'd use it as a complement for server side validating of (X)HTML content.
01:49
<erlehmann_>
Hixie: many blogs accept (X)HTML, but to not advertise it.
01:51
<Hixie>
hm?
01:52
<Hixie>
still to do: form submission, validity constraints api, encoding form data, data templates
01:52
<Hixie>
getting there!
01:52
<Hixie>
form submission attributes, rather, form submission itself is done
02:06
<erlehmann_>
Hixie: "hm?" what type of comment is that ?
02:09
<BenMillard>
erlehmann_, it probably meant "Hmm? How would that work?"
02:26
<Hixie>
erlehmann_: why would accept="" be relevant for that?
02:27
<erlehmann_>
can't i specify accept="application/xhtml+xml" and the browser would hint that to the user ?
02:27
<Hixie>
sure, but is any browser going to implement it?
02:29
<erlehmann_>
dunno. i'd probably rely on @pattern, with horrendous results
02:30
<Hixie>
<textarea pattern=""> is gone too actually :-)
02:30
<erlehmann_>
wat
02:31
<erlehmann_>
no wai
02:31
<erlehmann_>
http://www.whatwg.org/specs/web-forms/current-work/#the-pattern
02:31
<erlehmann_>
"and for the textarea element"
02:31
<Hixie>
yeah i dropped it in the transfer to html5
02:31
<Hixie>
couldn't find a good use case for it
02:32
<erlehmann_>
validating BBcode ?
02:33
<erlehmann_>
that's the most obvious thing
02:34
<erlehmann_>
willyoubutitbackplease ?
02:34
<Hixie>
can that even be done using regexps?
02:34
<erlehmann_>
lemme think.
02:34
<Hixie>
don't ask for things before you even know if they'll work for the reason you ask for them :-)
02:35
<erlehmann_>
well, i think it can work
02:35
<Hixie>
i can't see a pretty way to do it
02:35
<Hixie>
i'd imagine that custom JS with the setValidity() method would be much more effective
02:35
<erlehmann_>
[list of bbcodes] + [list of other characters] combined with wildcard = win ?
02:36
<erlehmann_>
like (let me build it)
02:36
<Hixie>
show me what the value of pattern="" would be. :-)
02:44
<erlehmann_>
wait. regexes are only capable of describing a chomsky type-3 language. what i want, is to correctly parse a type-2 language.
02:45
<erlehmann_>
therefore, i can absolutely not count opening and corresponding closing brackets with regular expressions, right ?
02:46
<BenMillard>
erlehmann_, that problem is something of a classic, although Philip` might have some ideas.
02:46
<erlehmann_>
:(
02:46
<Hixie>
like i said, not sure pattern="" is useful for <textarea>. :-)
02:46
<erlehmann_>
Hixie: you win. and again, i'm no match for the masters of the internet.
02:47
<erlehmann_>
BenMillard: nesting is the problem, i think. without nesting, no counting ...
02:48
<erlehmann_>
but what's with stuff like "enter a list of $stuff"
02:49
<erlehmann_>
though some of it becomes obsolete through the "add input" buttons, simply copying and pasting dozens of URLs into a textarea is easy at times
02:49
<erlehmann_>
gmail uses it
02:50
<erlehmann_>
and i recently encountered a guy who basically made trac tickets from CSV data
02:51
<BenMillard>
CSV is teh pwn
02:51
<BenMillard>
I much prefer it to XML
02:51
<erlehmann_>
(gmail uses it for mail adresses, of course)
02:51
<erlehmann_>
lol
02:51
<erlehmann_>
say, do you also prefer .INI over JSON ?
02:52
<erlehmann_>
:D
02:53
<Hixie>
well crap
02:53
<Hixie>
<form action=""> has nutty resolution semanics
02:57
<BenMillard>
erlehmann_, I prefer .ini over the Windows registry. Haven't used JSON.
03:17
<Hixie>
still to do: validity constraints api, encoding form data, removing data templates
03:24
<Hixie>
still to do: validity constraints api, encoding form data
03:24
<Hixie>
nearly done!
03:37
<BenMillard>
Hixie, what do you think of the alt text for the two images in main content here: http://calthorpepark.hants.sch.uk/news/2008/points.htm
03:37
<BenMillard>
(and anyone else who's interested)
03:40
<Hixie>
the alt text is ok, but the paragraph before the first image isn't
03:40
<Hixie>
for non-visual users it reads:
03:40
<Hixie>
The first chart shows Marie Currie in the lead at the moment, but only just!
03:40
<Hixie>
Marie Curie lead with nearly 1,000 points. They are followed by Sommerville, Kilby and Da Vinci.
03:40
<Hixie>
At the start of the term all Houses started from zero points.
03:41
<BenMillard>
Hixie, hmm...the alt text was written by me, the surrounding text comes from staff
03:41
<BenMillard>
is it the repetition you don't like?
03:41
<Hixie>
well there's no graph
03:42
<BenMillard>
oh, I see
03:42
<Hixie>
or chart rather
03:42
<BenMillard>
what if I remove "The first chart shows"?
03:42
<BenMillard>
(and adjust grammar to suit)
03:43
<BenMillard>
Marie Currie are in the lead at the moment, but only just!
03:43
<Hixie>
sounds good to me :-)
03:43
<Hixie>
the alt text could also give the number for the others, e.g. They are followed by Sommerville (300), Kilby (200) and Da Vinci (100).
03:44
<BenMillard>
Hixie, I wasn't given the exact values yet but when I get them, I can add them.
03:44
<BenMillard>
I agree with putting parenthese around them, too :)
05:30
<Hixie>
still to do: encoding form data
05:43
<famicom>
hah
05:43
<famicom>
just add support for base64
05:47
<BenMillard>
Hixie, isn't it your bedtime? or are you offsetting your sleep cycle for TPAC?
06:26
<Hixie>
BenMillard: my bedtime? it's not even 11 yet. i normally go to bed around 4am.
06:32
<BenMillard>
Hixie, oh I forgot which timezone you're inj
10:21
<Philip`>
erlehmann_: You can't do /(x*)y\1/ with theoretical regular expressions either (for the same reason you can't test for matching brackets), but practical implementations don't care about that theory :-)
10:21
<erlehmann_>
Philip`: tell me more.
10:22
<Philip`>
(but most practical implementations still don't support testing for matching brackets)
10:22
<Philip`>
erlehmann_: What kind of more?
10:23
<erlehmann_>
practical implementations don't care ?
10:24
<Philip`>
Regular expression theory doesn't contain backreferences like \1, but they're (relatively) easy to implement and useful so everyone implements them anyway
11:15
<takkaria>
though implementations without them can run way way faster
11:26
Philip`
is reminded of http://blog.mozilla.com/dmandelin/2008/10/06/squirrelfishing-in-regexp-dnajs/
12:30
<hsivonen_>
http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2007Jun/0173.html
12:38
<hsivonen_>
Re: http://www.w3.org/2001/tag/2008/09/23-minutes : I wonder what "Losses of engineering quality" refers to
12:54
<Philip`>
"<timbl> I think it is important that the cases which don't meet the IRI spec are referred to as errors, even if the errors are ignored in HTML5 browser handling" - isn't that what http://www.whatwg.org/specs/web-apps/current-work/multipage/infrastructure.html#valid-url is about?
12:56
<hsivonen_>
I'm not sure what this clean/not-clean concern is about. I wonder if I should ask.
12:57
<raspberry-lemon>
html5: halal or haram?!?
12:58
<Philip`>
hsivonen_: I assumed it was the usual distinction of valid HTML which authors should be told about / error handling of invalid HTML which browsers should be told about and should be hidden from authors
12:59
<Dashiva>
Apparently they want to split HTML5 into XHTML5 and HTML5-Tidy
13:00
<Philip`>
Dashiva: That sounds quite like how HTML5 works already, if you read "XHTML5" as "DOM" and "HTML5-Tidy" as "the HTML5 parsing algorithm"
13:01
<Philip`>
at least in the sense of converting into a form that is guaranteed to be free from syntax errors
13:01
<Dashiva>
Philip`: Except the parsing algorithm also parses the clean markup
13:01
<Philip`>
(because it has no syntax)
13:01
<Philip`>
Dashiva: HTML5-Tidy processes clean markup too
13:02
<Dashiva>
But it's not needed for it
13:02
<Dashiva>
DOM alone is pretty useless :)
13:02
<Philip`>
You can generate the DOM from scripts, without involving the parser much at all :-)
13:03
<Philip`>
Anyway, I'm wrong and I see your point :-)
13:03
<Dashiva>
But, but... then we can't continue
13:03
<Dashiva>
If I wanted to see a conspiracy, I'd say the tag wants to split the clean and non-clean so they can later deprecate the non-clean, or say people should only implement the clean :)
13:06
<Philip`>
I'd guess it's just because uncleanliness offends their sensibilities, and they want the W3C to be producing nice technologies, not HTML5
13:07
<Philip`>
Mr Last Week: Could you change your visited-link and date-line colour? It's very hard to read against the background, at least on my monitor with the sun shining on it
13:08
<Dashiva>
jar: Balance of power is different. In C the language had no power relative to the "browsers" (CPUs), so new CPUs could dictate language changes
13:09
<Dashiva>
I don't see how it's so very different now
13:09
Philip`
can't think of any examples of how C changed because of CPUs
13:11
<hsivonen_>
Philip`: CPU development has caused non-standard vectorization extensions to C, AFAIK
13:12
<Philip`>
I suppose that's true, but C specifically allows compilers to do that kind of thing
13:12
<Philip`>
It's just distributed extensibility, using the __ prefix to avoid collisions with future versions of the standard language
13:13
<Philip`>
And those extensions are just adding datatypes and things that look like library functions, and aren't really changing the language
13:14
<Philip`>
Oh, there's OpenMP which I guess is a more significant change to the language
13:15
<Philip`>
but that's got Open in its name so it must be a nice non-proprietary standard, and it isn't CPU-specific at all
13:16
<Philip`>
(and it's even backward-compatible)
13:22
<hsivonen_>
I guess Objective-C is just distributed extensibility, too, then?
13:23
<Philip`>
That's just a new language
13:26
<hsivonen_>
but every valid C program is a valid Objective-C program, right?
13:32
<Philip`>
Quite possibly
13:33
<Philip`>
but the difference is that it's called Objective-C, not C
13:33
<Philip`>
and nobody release software they claim is a C compiler, when actually it's an Objective-C compiler
13:33
<Philip`>
*releases
13:33
MikeSmith
prefers "Subjective C"
13:34
<MikeSmith>
"Objective C" would be a good name for a rapper
13:34
<jcranmer>
Surjective?
13:34
<Philip`>
so it's very different to releasing a C compiler, which happens to have been extended with some new features
13:36
<Philip`>
Objective C is the bunker on the left which we'll clear out once enemy resistance at the first two objectives has been suppressed
13:44
<MikeSmith>
Philip`: heh
13:48
<hsivonen>
well. I sent email asking what was meant with separation of clean and not-clean
15:36
<annevk2>
from DreamHost: "A suggestion you voted against has been marked completed!" yay, I'm all excited
15:37
<annevk2>
lol, the e-mail even suggests to re-submit the idea in case it was not completed to my satisfaction
15:41
<zcorpan>
what was the suggestion?
16:11
<gfan>
Hi, all. I am new to WHATWG and interested in Mobile Device API. Can someone tell me where to find out current WHATWG's efforts on Mobile Device API? Are there any cooperation between WHATWG and OpenAjax on Deviec API? Thanks.
16:12
<annevk2>
we're currently not working on such a thing and there are no plans afaik
16:13
<gfan>
annevk2: thanks. does WHATWG have some work on geolocation?
16:14
<Lachy_>
gfan, I'm fairly sure that the geolocation work is within the W3C, not the WHATWG
16:15
<gfan>
lachy_: thanks.
16:15
<hasather>
gfan: http://www.w3.org/2008/geolocation/
17:17
gsnedders
has no idea whether he'll be online from tomorrow till Sunday next week
17:18
<gsnedders>
So if anyone wants to tell me anything before TPAC, either tell me now or phone me
17:24
<ehird>
gsnedders: donkeys!
17:25
<erlehmann__>
HURR DURR I'M A WEB BROWSER
17:25
<ehird>
erlehmann__: :DDDDDDDD
17:27
<takkaria>
http://www.appleinsider.com/articles/08/09/25/apple_proposes_improvements_to_safari_browsing_experience.html is interesting
17:27
<takkaria>
they seem to be patenting a bit of the workings of the video tag :)
17:31
<gsnedders>
brb
18:48
<gsnedders>
hahahaha at my appearance on lastweekinhtml5
18:53
Philip`
hates it when adding a single line to his LaTeX document makes it suddenly become ten lines larger than the page limit
18:55
<Philip`>
(But yesterday it was 1.5 pages over the limit, so it's not doing too badly)
19:02
<Philip`>
Is it typical for French books to have the text on their spine the opposite way around to English books?
19:45
<john_fallows>
i see in the Server-sent Events section that servers are expected to send a "keep-alive" comment to prevent intermediate proxy servers from closing an otherwise dormant connection
19:47
<john_fallows>
how do we ensure that event latency is not adversely affected by intermediate proxy buffering?
19:48
<Hixie>
john_fallows: i am not aware of any way to avoid intermediate proxies from screwing everything up
19:50
<john_fallows>
would it help to specify that the browser should setup a tunnel with HTTP CONNECT first?
19:50
<Hixie>
so... who wants to be on the TPAC panel talking about the webarch document
19:50
<Hixie>
john_fallows: you can do that with WebSockets
19:51
<Hixie>
john_fallows: having them be different allows you to pick the one that works the best for a particular situation
19:51
<Hixie>
john_fallows: (dynamically)
19:51
<Hixie>
though frankly i've been tempted to just drop server-sent events altogether
19:53
<john_fallows>
i think Server-sent Events has a number of compelling features, and a clean integration with the browser event model
19:54
<Hixie>
so does WebSocket
19:55
<john_fallows>
yeah, but the events in WebSocket are not driven by the content, and can never be because it doesn't standardize the protocol (beyond framing), nor would it make sense to do so
19:57
<john_fallows>
also, Server-sent Events can recover from a broken TCP connection, making it very useful for broadcast failover
19:59
<john_fallows>
btw, we just released our WebSocket server on kaazing.org in case you are interested
20:01
<john_fallows>
there are clearly benefits to having Server-sent Events with a standard protocol syntax targeted at broadcast scenarios
20:01
<john_fallows>
and WebSockets with no standard payload syntax, targeted at bidirectional communication
20:01
<john_fallows>
almost like having UDP and TCP :-)
20:05
<Hixie>
fair enough :-)
20:05
<franksalim>
Hixie, john_fallows: i think the most compelling reason to keep SSE in addition to WS is that SSE is still HTTP and can be driven by current web servers
20:06
<john_fallows>
true - whereas WebSocket just uses HTTP-compatible bytes for the handshake
20:09
<john_fallows>
specifying HTTP CONNECT for Server-sent Events would seem to address the potential buffering issue though
20:11
<Hixie>
please send e-mail to the list to remind me to look into it
20:11
<john_fallows>
ok, sounds good - will do.
20:11
<Hixie>
but i'm not sure if we'll be able to do that -- if that worked, why wouldn't browsers just do that for all connections?
20:15
<john_fallows>
i suppose not all connections have a requirement to minimize latency of incremental parts of the response
20:15
<john_fallows>
i'll organize my thoughts and send an email to the list
20:22
<john_fallows>
btw, i'll also send a separate email regarding server-initiated flush to force a client acknowledgment and immediate reconnect, without affecting recovery timeout in network or server failure scenario
22:11
<Philip`>
annevk: Is it possible to set up wildcard domains on html5.org, like to make *.philip.html5.org equivalent to philip.html5.org?
22:14
<Philip`>
(or possibly to make them all point to some other server, if Dreamhost doesn't let you make the web server handle all those domains)
22:15
<gsnedders>
Hixie: ping
22:15
<gsnedders>
BenMillard: ping
22:16
<Hixie>
gsnedders: in meeting but here
22:17
<gsnedders>
Hixie: Need I bring any PS2 stuff? You have two controllers?
22:17
<Hixie>
i was gonna ask you the same question
22:17
<Hixie>
i don't plan to bring anything
22:17
<Hixie>
unless you want me to
22:18
<Hixie>
(i have to go through the US customs so the less i bring the better)
22:18
<gsnedders>
Hixie: Do you have GT4?
22:18
<Hixie>
um
22:18
<Hixie>
maybe
22:18
<Hixie>
not sure
22:18
<Hixie>
if i do i haven't unlocked anything
22:19
<gsnedders>
Heh.
22:19
<gsnedders>
There's only really any point in you bringing PS2 if we have two copies of GT4 so we can have it running over LAN and not split-screen
22:19
<gsnedders>
As for not unlocking anything, I can bring memory card (my disk of it is screwed, and I can't bring my PS2 with me, mainly because it's old-style bulk)
22:20
<Hixie>
i'd really rather not bring my ps2
22:20
<Hixie>
but we'll see
22:20
<gsnedders>
On the whole, I'd say don't bother
22:21
<Hixie>
k
22:41
<BenMillard>
gsnedders, still here?
22:42
<gsnedders>
BenMillard: yea
22:42
<BenMillard>
gsnedders, just saw you pinged me in the logs
22:42
<gsnedders>
BenMillard: and dropped email
22:42
<BenMillard>
I only have 1 bonafide PS controller
22:42
<BenMillard>
and no LAN cable
22:42
<gsnedders>
BenMillard: So should I bring another?
22:42
<gsnedders>
BenMillard: LAN cable is no issue, we will only have 1 PS2
22:42
<BenMillard>
ah, ok
22:42
<BenMillard>
gsnedders, yes, do please bring 1 controller
22:43
<gsnedders>
BenMillard: OK. Maybe I'll bring a game or two too :)
22:43
<BenMillard>
gsnedders, sure...and your memory card if it's got anything useful on it
22:43
<gsnedders>
BenMillard: It might, might not
22:43
<gsnedders>
BenMillard: Maybe I'll bring GT3…
22:43
<BenMillard>
gsnedders, I can bring that
22:43
<gsnedders>
BenMillard: So can I. I'm in the middle of a race on it atm :P
22:44
<BenMillard>
the way GT3 models oversteer and drifting is my favourite of any game
22:44
<gsnedders>
I started GT3 from scratch on Monday to practice with slower cars :P
22:45
<BenMillard>
Hixie, you don't need to bring anything PS2-related unless it's an extra game (like GTA: foo?)
22:48
<BenMillard>
gsnedders, easier if I bring both GT games, since my discs are probably in better condition than yours
22:48
<gsnedders>
BenMillard: My GT3 works fine, that's why I'm playing that and not GT4 now :)
22:48
<BenMillard>
ok, if you really want to bring your copy then we'll use that
22:49
<gsnedders>
I have plenty to bring already :)
22:49
<gsnedders>
And I have to lug it around for a week first
22:49
<BenMillard>
GRR! so are you bringing it or am I? make your mind up!
22:49
<gsnedders>
You are
22:49
<BenMillard>
ok
22:49
<gsnedders>
That's the implication of that comment :P
22:50
<gsnedders>
Ooo… I could bring NFS:HP2 (the best NFS game IMO), F1:05, and Tokyo Xtreme Racer
22:50
<BenMillard>
I remmber the first Hot Pursuit on PC, was great in 2P
22:50
<BenMillard>
F1 games can be hilariously unrealistic, so that might be intertaining
22:50
<gsnedders>
Oh, the AI is lovely. You can easily lap them several times on hard.
22:51
<gsnedders>
(NFS as always has a rubber-band AI, so you never really thrash it)
22:51
<gsnedders>
(unless you turn the rubber-band off, and do an eight lap race)
23:32
<annevk2>
Hixie, can't you be on the panel?
23:32
<Hixie>
i'll be editing html5.
23:33
<annevk2>
hmm, what if I edit it an hour for you (preferably a different hour, so I can watch the show :) )
23:33
<Hixie>
i don't see the point in participating. what difference does it make if the webarch document changes, even if we do convince people to change it?
23:34
<annevk2>
i was mostly concerned with the entertainment value
23:34
<Hixie>
if you want entertainment... the meeting is in a casino.
23:34
<Hixie>
and gsnedders and BenMillard will have a PS2
23:38
<annevk2>
i think i'd have more of a laugh during that panel, but we'll see :)
23:41
<annevk2>
Philip`, dunno
23:41
<annevk2>
(re wildcard domains)
23:42
Philip`
is trying to write some quick tests, but forgot how buggy browsers are
23:43
<Philip`>
(My current problem is that I can add a certain key into Firefox's globalStorage, which is impossible to remove again)
23:44
<Philip`>
annevk2: Is there a thing where you can type in subdomain names and IPs, so you could put something like "*.philip" in the subdomain thing and point it at my server?
23:44
<annevk2>
I can make a custom DNS record
23:44
<annevk2>
I'm not sure if it does wildcards
23:44
<annevk2>
but give me an IP and I'll try
23:45
<Philip`>
92.243.11.39 would be nice
23:46
<annevk2>
"We have now added the A record for *.philip.html5.org with value 92.243.11.39. Our DNS servers will start serving this record within a couple of minutes."
23:47
<Hixie>
if that works i'll be impressed
23:47
<Philip`>
$ hostx foo.philip.html5.org ns1.dreamhost.com
23:47
<Philip`>
foo.philip.html5.org A 92.243.11.39
23:48
<Philip`>
Looks like that does work - thanks!
23:48
<Hixie>
wow
23:48
<Hixie>
i am shocked
23:48
<Hixie>
that's awesome
23:48
<Hixie>
time to set up wildcard domains everywhere!
23:48
Philip`
wonders what to do now that he has control of an infinite number of origins
23:48
<Hixie>
not infinite
23:48
<Hixie>
just very very large
23:49
<Hixie>
(dns names are length-limited and character-set limited)
23:49
<Philip`>
Close enough to infinite for all practical purposes :-)
23:49
<Philip`>
and for many impractical purposes too
23:49
<annevk2>
btw, DreamHost still claims www.philip and ftp.philip, fwiw
23:50
Hixie
turned off ftp
23:50
<annevk2>
ok, both btw and fwiw is a bit too much
23:50
<Hixie>
and you can turn off www, iirc
23:50
<Hixie>
dunno if that affects dns though
23:50
<annevk2>
I thought you could only make it redirect
23:50
<annevk2>
(which I do for all sites)
23:52
<Hixie>
hm maybe
23:52
<Hixie>
i thought you could disable it
23:52
<Hixie>
i forget