00:17
<Philip`>
Hixie: You have a bogus </colgroup> that is breaking my attempt to run the spec-splitter :-(
00:17
<Hixie>
oops
00:18
<Hixie>
um
00:18
<Hixie>
why is it broken
00:18
<Hixie>
source seems ok
00:18
<Philip`>
and the spec-splitter happily dies and then uploads it non-existent output and thus deletes the multipage spec
00:18
<Hixie>
what's broken about it?
00:19
<Philip`>
<table class=apple-table-examples><col></colgroup><col><col><col><thead><tr><th> <th>2008 <th>2007 <th>2006
00:19
<Hixie>
wtf
00:19
<Hixie>
that must be a libhtml bug
00:19
<Hixie>
or libxml
00:19
<Hixie>
or whatever we're using
00:19
<Philip`>
That's quite possible
00:19
<hober>
I've written a blog post about how to use HTML5's new sectioning elements.
00:19
<hober>
Would anyone like to review it before I publish? I want to be sure I'm making good recommendations. I'd rather the article reduce, not add, to the confusion around these elements.
00:20
<Philip`>
(I'm using libxml for parsing, because html5lib is unbearably slow)
00:20
<Hixie>
hober: can't review right now, but i'd be happy to do so after my meeting which starts in 10 minutes
00:20
<Philip`>
(Also, this is the index document, not the source document)
00:20
<Philip`>
lxml.etree.XMLSyntaxError: Unexpected end tag : colgroup, line 29957, column 53
00:21
<hober>
Hixie: ok, I'll /msg you the link; please review at your leisure
00:21
<Philip`>
(using etree.HTMLParser)
00:22
<Philip`>
Is this something that is easily fixable without me doing anything?
00:26
<TabAtkins>
hober, I'll take a look.
00:28
<Philip`>
annevk2: http://www.whatwg.org/specs/web-apps/current-work/multipage/introduction.html - now with added TOC
00:29
<Philip`>
http://www.whatwg.org/specs/web-apps/current-work/multipage/commands.html - it's a little peculiar on certain pages, but I think it's probably about correct
00:29
<TabAtkins>
Philip`, I love the new TOCs.
00:29
<Philip`>
Hmm, the toc.js script messes things up a bit :-(
00:29
<Philip`>
Hixie: Is there some way toc.js could be disabled on the mini-TOCs in the multipage view?
00:30
Philip`
temporarily switched to the html5lib parser, but it's too resource-intensive for him to be particularly happy running it
04:59
Hixie
is proud of his table examples
07:47
<othermaciej>
hi all
09:19
<annevk3>
Philip`, sweet
09:20
<annevk3>
Philip`, for the introduction page the "full table of contents" should probably not be there
09:20
<annevk3>
Philip`, and the link above it, it's not present on other pages either
09:20
<Philip`>
annevk3: I blame Hixie for that
09:21
<Philip`>
since toc.js adds the full table of contents automatically
09:21
<annevk3>
ah
09:21
<annevk3>
i see, it's also present on full table of contents
09:21
<annevk3>
maybe toc.js should recognize a class name or something
09:21
<Philip`>
Hixie: It's all your fault :-(
09:22
<Philip`>
annevk3: I think I have to use class="toc" to get the right styling, but could add an extra class to make toc.js ignore it
09:22
<annevk3>
that could work, or hixie adds a class to the main toc; either way
09:41
<annevk3>
Hixie, pretty sure that if the fieldset is disabled the contents of the legend element are too
09:42
<annevk3>
in Opera they are
09:42
<annevk3>
which is per spec
09:42
<annevk3>
Hixie, so the fieldset example will never work
09:43
<annevk3>
Hixie, I'd move the checkbox into a <p> or so before the fieldset
10:17
<Hixie>
annevk3: should we fix that?
10:17
<Hixie>
the spec i mean?
10:17
<Hixie>
seems like a reasonable use case
10:17
<Hixie>
anyway i gotta sleep now
10:28
<annevk3>
I guess we could
10:29
<Lachy>
would it work to just say that controls within the <legend> don't get disabled along with the fieldset?
10:30
<annevk3>
spec-wise it's no problem
10:30
<annevk3>
just wondering about implementation complexity with dynamic updates and all
10:30
<annevk3>
not more crazy I guess than other stuff, but the question is whether it's worth it
10:31
<Lachy>
I'd be more convinced if there were real sites built with equivalent functionality
10:31
<annevk3>
sites prolly will if the styling issue gets solved
12:53
Philip`
tries to work out whether he should be using zh or zh-CN or zh-Hans
12:53
<Philip`>
Languages are too confusing :-(
12:57
<annevk2>
don't bother tagging it?
12:57
annevk2
tries to work out which alias table Safari is using for ICU
12:57
<Dashiva>
zh-x-Philip`
12:58
<Philip`>
annevk2: I need to at least have some language code for the filename
12:58
<Philip`>
so I might as well use the correct one
12:59
<annevk2>
why is that?
13:00
<Philip`>
Because of a pointless desire for correctness
13:00
<annevk2>
carry on then :)
13:04
<Philip`>
So which one is correcter? :-)
13:10
<and>
Philip`: zh-CN looks wronger/older; zh-Hans is preciser (so likelier to be incorrect).
13:11
<and>
(Apart from that, CN would be associated with mainland China, and Hans with simplified ideograms, which originated there.)
13:12
<annevk2>
was to be expected: http://meiert.com/en/blog/20090911/html-5-or-html5/
13:13
<annevk2>
Philip`, zh seems safest if you're unsure
13:14
<Philip`>
http://blog.whatwg.org/spelling-html5 - "Isn’t 5 a *version number*? As in HTML 4.01. Shouldn’t it actually be HTML 5?"
13:14
<Philip`>
That seems a good argument to call it "HTML5", so people don't think it's a version number
13:16
<annevk2>
and, new grouping is nice btw
13:16
<Philip`>
annevk2: I'm sure it's simplified Chinese, written by someone from mainland China, so the issue is just what code to use (rather than what language it is)
13:17
<Philip`>
The person who wrote it called it zh_cn, but e.g. http://www.w3.org/International/articles/language-tags/ seems to indicate zh-Hans is newer and more consistent and accurate
13:20
<annevk2>
are you sure it's not zh-Hant?
13:20
<annevk2>
anyway, http://tools.ietf.org/html/rfc5646 seems to suggest a couple of those are at least out of date (though still valid for backcompat reasons)
13:22
<and>
Hant/Hans means traditional/simplified Han characters (ideograms).
13:23
<and>
The use of CN to mean Hans is obsolete, but probably still fine to indicate mainland China.
13:23
<and>
(as I understand it, anyway)
13:24
<annevk2>
language tags certainly became a whole lot more complicated since I last looked at it
13:29
<and>
Hierarchically constructed tags like zh-Hans-CN are nice in theory, but I am not sure they are handled reasonably by matching algorithms.
13:32
<and>
annevk2: Glad you like it. I tried to choose clear and uncontroversial categories, but the labels are not completely self-explanatory (e.g., Mac CJK encodings belong to the CJK group, not to the Mac group). I hope I did not change anything in the process.
13:32
<annevk2>
oh, I haven't checked that
13:33
<annevk2>
it makes it a lot easier to compare browsers anyway
13:34
<annevk2>
once the data is more complete I plan to write a scraper that allows you to run comparisons on it
13:34
<and>
Well, I tried to be careful.
13:34
<and>
That is a good idea. We might want to add the IANA registry as a "browser" as well.
13:35
<annevk2>
yeah
13:35
<annevk2>
unfortunately it seems i'm booked for most of the weekend at least until sunday evening so it'll have to wait a little
14:14
gsnedders
hears loud roar of small jet engines, and looks up at window, and sees some of the red arrows flying fairly low and fast
16:24
<Philip`>
http://philip.html5.org/demos/rdfa/google-rich-snippets.html - it's a bit pathetic really
16:25
<Dashiva>
wau
16:26
<Dashiva>
Looks like something you'd make with regular expressions :)
16:29
gsnedders
sighs at British train prices
16:58
<gsnedders>
So a ticket from Reading to Leuchars is the same price as one from Birmingham New Street to Leuchars. That makes sense.
16:59
<gsnedders>
That's untrue. There is a slight difference.
17:06
<Philip`>
http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2009Sep/0124.html
17:06
Philip`
hopes he didn't make too many stupid mistakes
17:11
<Dashiva>
Philip`: I expect one response will be "But nobody is going to use v:g/#Person so it's just silly to support it"
17:14
<Philip`>
Then I might respond that some people seriously considered xmlns:http="http:" as a valid trick to transition to non-prefix-based URLs, which would be broken by this kind of thing
17:15
<Dashiva>
They were silly too
17:16
<Philip`>
Recent experience indicates RDFa people are happy with requiring specific behaviour for invalid input anyway, even when it's weird edge cases
17:17
<Philip`>
(and any disagreement is just about how explicitly and clearly it needs to be specified, vs being determined by the spec authors' intentions)
17:47
<Philip`>
"we will [...] deviate from the standard in some cases to be forgiving of webmaster errors. For example, we expect that some webmasters will forget the xmlns attribute entirely." (says a person from Google)
17:47
<Philip`>
Hooray for standards and interoperability!
17:48
<hober>
I think the Google parsing behavior sounds reasonable, at least as a forgiving, "real-world" processor of RDFa bits and pieces that survive copy&paste
17:49
<hober>
It seems like the extent to which they're deviating from RDFa-as-specified is the extent to which RDFa-as-specified isn't a web-ready tech...
17:53
<Steve^>
Should an element with display: none be included in the outline?
17:53
<Philip`>
If they want something that's forgiving and web-ready and works in the real world, and don't think RDFa is suitable for that, then they shouldn't pretend to support RDFa
17:55
<Philip`>
Then they could simplify it further (e.g. not use xmlns at all), and would no longer be flagrantly violating standards
17:55
<Philip`>
It'd be good if they documented whatever they implement, too
17:55
<hober>
indeed
17:56
<hober>
I'm just saying I think the RDFa folks should take a different lesson than the one I think they'll take from this
17:56
<Philip`>
Steve^: Yes
17:57
<Philip`>
since CSS is just an optional rendering thing
17:57
<Philip`>
Maybe <whatever hidden> shouldn't be included, though
17:57
<Philip`>
("shouldn't" in terms of what's sensible, not necessarily what the spec says)
18:02
<Steve^>
It is good to avoid the "untitled section" in an outline
18:14
<Philip`>
(For future reference: That quote is from http://www.w3.org/mid/6934bc6b0909120945l7e721b90pda3ec06ca4d4c917⊙mgc which doesn't exist since it hasn't hit the list yet)
18:14
<Philip`>
(The Google RDFa quote, that is)
18:20
<jgraham>
Steve^: Yes. Indeed if you use one of the techniques that shows things to AT but not to ordinary browsers (e.g. the left -10000 thing) it is even better
21:15
<hober>
OK, I've published my blog post about how to use sectioning elements correctly http://edward.oconnor.cx/2009/09/using-the-html5-sectioning-elements
21:37
<Lachy>
hober, did you make any changes since I reviewed it earlier?
21:37
<hober>
only a handful of litle edits; nothing major
23:56
gsnedders
wonders if trying to change at Clapham Junction is a good idea