07:40
<estellevw>
There were five global attributes related to microdata including itemid, itemprop, itemref, itemscope and itemtype that i remember seeing, but don't see in the spec right now. Am I missing something?
07:46
<myakura>
estellevw: they got split off from the main HTML5 spec about a month (or two) ago.
07:46
<estellevw>
thanks
07:47
<estellevw>
is it likely to come back in?
07:48
<myakura>
I guess the WHATWG version of HTML5 still incorporates it.
07:48
<myakura>
for W3C version, see http://www.w3.org/TR/microdata/ .
07:49
<estellevw>
yeah, looking at it
07:49
<estellevw>
i think it needs some editing. The attributes are in the body but not in the table of contents
07:51
<estellevw>
http://www.w3.org/TR/microdata/#attr-itemref and the like have lost their original stature.
07:52
<estellevw>
it states The following attributes are added as global attributes to HTML elements:
07:52
<estellevw>
* itemid
07:52
<estellevw>
* itemprop
07:52
<estellevw>
* itemref
07:52
<estellevw>
* itemscope
07:52
<estellevw>
* itemtype
07:52
<estellevw>
oops, sorry
07:52
<estellevw>
but if i recall correctly, those aren't listed in teh global attributes anymore
07:53
<estellevw>
the global attributes being here: http://www.w3.org/TR/html5/dom.html#global-attributes
07:56
<myakura>
because they are two separete specs; Microdata is build on top of HTML5 so those item* attributes cannot be defined in HTML5.
07:59
<estellevw>
ok, makes sense as to why role, and the aria-* attributes aren't listed as global even though they are too
07:59
<estellevw>
thanks
07:59
<myakura>
yeah..
08:00
<myakura>
just refer to the WHATWG version and you won't get confused :)
08:00
<myakura>
MikeSmith: aloha
08:01
<estellevw>
i can still get confused, but i'll have to blame myself ;)
11:28
<MikeSmith>
myakura: here now
11:29
<MikeSmith>
hsivonen: if/when you are around and have time to chat, please ping me
11:29
<MikeSmith>
in regard to http://dev.w3.org/html5/spec/text-level-semantics.html#guidance-for-conformance-checkers
11:42
<MikeSmith>
writing code for a conformance checker to check "The img element is part of the only paragraph directly in its section" or even "only non-whitespace content in the only paragraph directly in its section" is not easy
11:45
<annevk>
prolly also depends on how you implement things
11:46
<MikeSmith>
well, it's not practical at all to implement using a grammar-based schema, so we forget about that completely
11:48
<MikeSmith>
and it's quite complicated to implement in the Java code that validator.nu currently uses for things that can't be checked practically using grammar-based checking
11:50
<MikeSmith>
as far as I can see, it will require adding code that is very unlike existing code for any other checking that is being done by validator.nu
11:58
<othermaciej>
MikeSmith: I think that particular exception is not very well justified in the first place
11:58
<othermaciej>
MikeSmith: but yeah, to check an "only paragraph in its section" condition would require running the HTML5 outline algorithm I think
11:59
<MikeSmith>
It would be help to have the spec provide the intended rationale for that exception
12:00
<othermaciej>
MikeSmith: Hixie explained it a bit in the specific bug Laura filed about it
12:00
<MikeSmith>
ok
12:00
<othermaciej>
let me see if I can find it
12:01
<othermaciej>
MikeSmith: http://www.w3.org/Bugs/Public/show_bug.cgi?id=9217
12:02
<MikeSmith>
I see
12:02
<othermaciej>
MikeSmith: I would personally think it was fine to say "just use figure"
12:02
<MikeSmith>
well, I think we could all live without that particular exception
12:02
<MikeSmith>
yeah, what you said
12:03
<othermaciej>
because associating the relevant section header with the image is much more complicated
12:03
<othermaciej>
as much as you don't want to implement it in the validator, I *really* don't want to implement the outline algorithm solely for our accessibility code to handle this case
12:04
<MikeSmith>
yep
12:06
<othermaciej>
I'm adding a comment to the bug
12:06
<MikeSmith>
OK
12:10
<othermaciej>
done
12:11
<othermaciej>
I mentioned the validator issue but feel free to comment as well
12:12
<othermaciej>
the rate of incoming bugs is just ridiculous
12:12
<othermaciej>
can't believe we are up to 90 already, there were about 42 just a few days ago
12:13
<MikeSmith>
well, in some ways, incoming bugs is a sign of progress
12:13
<othermaciej>
it's a sign of people reviewing the spec
12:13
<MikeSmith>
e.g., with the media-accessibility bugs that Silvia raised
12:13
<othermaciej>
which is good
12:13
<MikeSmith>
true that
12:14
<othermaciej>
I don't feel bad about incoming bugs, as long as the rate of outgoing bugs is also high
12:47
<myakura>
MikeSmith: I talked with Shiraishi-san on Friday about what we're gonna talk in Fukuoka.
12:48
<MikeSmith>
OK
12:50
<myakura>
MikeSmith: and their plan is to: Takuya will do the keynote-like talk (.5hr), and You and I do about Web Standards in general (like what you did at DevFest)
12:50
<MikeSmith>
sounds good so far
12:51
<myakura>
MikeSmith: Oli for markup stuff, Shiraishi-san will talk about APIs, and Anne to do about "CSS latest status"
12:51
<myakura>
and I'm not sure you guys heard about that
12:51
<myakura>
and whether you guys are okay with that.
12:52
<boblet>
oh, good timing
12:52
<boblet>
hey all
12:54
<myakura>
boblet: heya.
12:54
<boblet>
myakura: by markup stuff do you mean sectioning elements etc?
12:54
<boblet>
or also CSS?
12:55
<myakura>
MikeSmith: btw we have an hour for each session but for yours and Anne's I need to interpret so you guys will have shorter time (40min?)
12:56
<MikeSmith>
myakura: that sounds fine
12:56
<myakura>
boblet: the table only shows "HTML5 Mark up" ...
13:00
<boblet>
myakura: will check it out. thanks for the info. Guess I better read Mike’s slides
16:34
<jgraham>
It seems like AT would want access to the document outline in any case
16:36
<jgraham>
And presumably webkit will have to implment it anyway once we get :section(n)
16:37
<annevk>
if we ever decide to do that
16:37
<jgraham>
Well sure
16:37
<jgraham>
If we don't decide to do that the whole outline algorithm is likely a waste of time
17:32
<micheil>
hmm.. would this be the place to ask a quick question about one of your spec proposals?
17:32
<jgraham>
Yes
17:33
<micheil>
okay, with the websocket protocol, would I be right in assuming that one server would be able to server multiple different sockets / data sets to clients based on the paths at which they connect?
17:34
<jgraham>
With the proviso I am not an expert on that spec, yes
17:34
<micheil>
okay
17:34
<jgraham>
The server can do whatever it wants based on anything the client sends including the path part
17:35
<jgraham>
(subject to it meeting the requirements for a successful connection of course)
17:36
<micheil>
yeah, I'm just trying to work out how to direct data about by using that, all the reference implementations I can find don't indicate anything on that
17:38
<jgraham>
Well when the server recieves the client handshake it parses out the path so you can arrange for it to be passed to the app
17:38
<jgraham>
Not sure what existing implementations do though
17:39
<micheil>
yeah, just trying to think if it's possible to send data to a specific connection or not
17:39
<micheil>
(ie, rather then just being a broadcast type service)
17:40
<micheil>
I suppose because the writes to the network stream would be done from within a connection, you'd handle it there maybe..
17:41
<jgraham>
If I follow you, that sounds right
17:42
<jgraham>
From the point of view of the server each client is a seperate connection so you can read and write to each client independently
17:43
<micheil>
hopefully
17:43
<jgraham>
I don't really see how it could work otherwise
17:44
<micheil>
I suppose I should just work out the raw tcp send switching first, then add on the websocket protocol
17:46
<MikeSmith>
jgraham: conformance checkers should not require access to the document outline
17:47
<jgraham>
MikeSmith: I didn't mean to have any opinion on conformance checkers
17:47
<jgraham>
Just other types of UA
17:47
<MikeSmith>
I see
17:47
<MikeSmith>
I can see that AT having access to the document outline would be good
17:48
<MikeSmith>
but then there are many things that AT should be doing that they are not currently
17:49
<jgraham>
Right, but I disagree with othermaciej's assertion that implementing the outline algorithm in the accessibility code would soley be for this case
17:53
<MikeSmith>
yeah, it'd certainly seem it could end j useful for
17:54
<MikeSmith>
.. end up being useful for lth
17:54
<MikeSmith>
oops
17:54
<MikeSmith>
...useful for other things
17:54
<MikeSmith>
(my fingers are cold)
17:55
<jgraham>
ah, I was imagining that you had insane keyboard macros that could expand a few characters into whole sentences
17:55
<jgraham>
and they had broken
17:55
<MikeSmith>
heh
17:59
<MikeSmith>
I just came back from sento, in 水風呂 water that's 17.1 degrees
18:04
jgraham
assumes that is the kanji for "fucking cold"
18:05
<micheil>
it'd have to be less then 17 degrees outside here
18:06
<jgraham>
Well sure it is less than 17 degrees C outside here too
18:06
<jgraham>
But I wouldn't go in water that temperature
18:06
<jgraham>
The thermal conduction is a killer
18:07
<micheil>
actually. 20ºC here is cold.
18:08
<AryehGregor>
It's barely above 20°C here now, and it's starting to be summer already.
18:09
<micheil>
jgraham: I think the key to my websocket problem is in the http parsers (I'm working with node.js btw)
18:10
<micheil>
it's not even winter yet and it's 7.6ºC outside.
18:10
<micheil>
then again, I'm usually getting 30-45º in summer.
18:17
<MikeSmith>
micheil: I think maybe websocket isn't meant to be used with existing http parsers
18:17
<micheil>
MikeSmith: it isn't. however, the key to building a robust node.js websocket server implementation lies in how http servers work
18:18
<MikeSmith>
I'm not familiar with node.js or what its main use cases are
18:19
<micheil>
async / evented javascript with network and file i/o
18:19
<micheil>
built on the same javascript engine that powers chrome / chromium
18:38
<jgraham>
micheil: Do you get raw socket access with node.js?
18:38
<jgraham>
If you do I would start from there rather than with the existing http parser
18:38
<micheil>
yeah, raw tcp / sockets
18:39
<micheil>
although, node's http parser is built on it's net sockets, so I can use that as reference
18:45
<jgraham>
micheil: You might be just as well working from the simple echo server example in the documentation
18:46
<micheil>
not realy
18:46
<micheil>
not everything is documented, so there are some different ways to do things under the hood
18:46
<jgraham>
Or you sould take the existin websocket-in-node.js implementation and check that it supports the new handshake
18:46
<jgraham>
*could
18:46
<micheil>
(I wish the current net implementation was around when I started working on node-smtp-client)
18:47
<micheil>
jgraham: nawh, that takes away half the fun :P
18:47
jgraham
should finish his websocket-server-in-python-diesel implementation
18:48
<micheil>
heh heh, what I'm wanting to get from this is pretty specific
18:55
<AryehGregor>
Hmm, so Google is funding efficient ARM decoding of Theora. I still have hope that they'll switch YouTube from H.264 someday.
18:59
<jgraham>
AryehGregor: Well I don't doubt they will. They question is will it be to theora and soon or to H.264.next and not-so-soon
18:59
<AryehGregor>
Heh.
18:59
<JonathanNeal>
updated ie print protector; added summary element, wrapped entire script in ie conditional (instead of individual fns), and compressed window/document variables (all based on remy sharp suggestions)
19:10
<micheil>
hmm.. due to the way the websocket protocol is, I'm not likely to need much of a really heavy parser for the handshake (initial GET request), am I?
19:11
<jgraham>
micheil: You need to parse out the headers to get the websocket-sec-key (or whatever they are called) values
19:12
<jgraham>
And you need the random bytes
19:12
<micheil>
yeah, but I want need the same level of parser as what a standard http server would
19:12
<jgraham>
and there are some requirements about when you must drop the connection
19:12
<micheil>
hmm.. random bytes..
19:12
micheil
checks it
19:12
<jgraham>
But no, a full HTTP stack isn't necessary
19:13
<micheil>
hmm..
19:13
<micheil>
random bytes as in the ^n:ds[4u from the spec?
19:14
<jgraham>
The 8 bytes after the end of the part that looks like HTTP headers
19:14
<micheil>
which, according to http spec should probably be: ...headers..\r\n\r\nDATA\r\n\r\n
19:14
<micheil>
I think
19:15
<jgraham>
The HTTP spec isn't really relevant
19:15
<jgraham>
From the point of view of the HTTP spec they are the first 8 bytes of the body or so, I think
19:19
<micheil>
yeah, although, I'm just working out what I'll need to do to parser it
19:36
<Hixie>
simplest answer to that is to follow the rules in the parser section that tell you how to write the parser :-)
19:36
<Hixie>
section 5.1
19:36
<micheil>
Hixie: heh, I suppose you'd be the man to ask :P
19:37
<micheil>
Hixie: is there any sane way that you read those documents?
19:40
<Hixie>
how do you mean?
19:47
<micheil>
well, there seems to be a lot of extra formatting on the IEFT / RFC type documents, is there a cleaner way to view them?
20:40
<jgraham>
micheil: If you can use the complete.html spec on the WHATWG site. It is a bitch to load but once it is loaded it is rather good to read
21:23
<annevk>
http://adactio.com/journal/1654/ -- I still think we should drop <article> and get <content>
21:24
<annevk>
though maybe just dropping <article> for now and see what patterns emerge
21:40
<Hixie>
annevk: i don't understand why people are so confused by them... they're completely differnet
21:40
<Hixie>
one is for chapters and the other is for syndicatable content
21:40
<Hixie>
they're different use cases with almost no overlap
21:42
<Dashiva>
Hixie: The confusion seems to be with the descriptions, maybe it's just a matter of condensing it down to a better explanation
21:47
<Hixie>
the descriptions are completely different
21:47
<Hixie>
i think the problem is that the element names aren't intuitive based on the descriptions
21:48
<Hixie>
but that's not a big problem, people will learn the difference in due course
21:49
<Hixie>
it's like <dl>, <ul>, and <ol>. Given just those element names and then descriptions of what they're for, how would you guess which was which?
21:50
<annevk>
hmm, would've been good to test that out
21:51
<othermaciej>
<article> and <section> as names sound very distinct
21:51
<othermaciej>
maybe the problem is with the descriptions
21:54
<othermaciej>
"The section element represents a generic document or application section." --> ""The section element represents a generic section of a document or application"
21:54
<othermaciej>
I bet that by itself would reduce confusion
21:54
<othermaciej>
if you skim fast, the current text looks like "The ____ element represents a document"
21:55
<othermaciej>
"The article element represents a component of a page that consists of a self-contained composition in a document, page, application, or site and that is intended to be independently distributable or reusable, e.g. in syndication."
21:55
<othermaciej>
and that one reads as "The ____ element represents a component" if you skim
21:56
<othermaciej>
could be something like "The article element represents a self-contained composition in a document, page, application, or site that is intended to be independently distributable or reusable, e.g. in syndication."
21:56
<othermaciej>
Hixie: ^
21:56
<othermaciej>
(annevk also)