08:21
<zcorpan_>
hmm, wonder if "the title element" is right
08:24
<hsivonen>
zcorpan_: why?
08:25
<zcorpan_>
it seems too restrictive relative to what browsers implement
08:25
<zcorpan_>
and only opera seems to reflect document changes
08:26
<zcorpan_>
e.g., if you move the title element from head to body, does it stop being the title element?
08:27
<zcorpan_>
if you insert a new title before the title element, which is "the title element"?
08:34
<hsivonen>
zcorpan_: seems like an opportunity for test-based detailed review. :-)
08:36
<zcorpan_>
hsivonen: indeed :)
08:36
zcorpan_
has 10 tests already
08:39
<othermaciej>
10 tests for what?
08:39
<zcorpan_>
document.title
08:39
<zcorpan_>
getting
08:45
<hsivonen>
I have no idea how to make a sane streaming approximatin of the AAA
08:45
<hsivonen>
will leave it undone for now
08:46
<zcorpan_>
that isn't drocanian?
08:46
<hsivonen>
zcorpan_: no, non-conforming and non-draconian
08:47
<hsivonen>
zcorpan_: halting is draconian
08:47
<hsivonen>
zcorpan_: sorry. I misread you
08:47
<hsivonen>
zcorpan_: yes, an approximation that isn't draconian
08:47
<hsivonen>
(and is thereby non-conforming)
08:48
<zcorpan_>
ok
09:04
<zcorpan_>
wow
09:05
<zcorpan_>
safari for windows treats .xht files as text/html
09:40
<zcorpan_>
where is SVGDocument.title defined?
09:46
<hsivonen>
zcorpan_: Section 5.17 in SVG 1.1
09:51
<zcorpan_>
thanks
10:35
zcorpan_
notes that http://www.whatwg.org/specs/web-forms/current-work/#form-submission mixes block and inline content
12:02
<Xiven->
hi, hixie just wanted me to pass on the message to anyone at Opera, that he's currently in the Oslo office canteen, so come along and say hi
13:47
<zcorpan_>
Document.renameNode isn't implemented anywhere, is it
13:51
<hsivonen>
hmm. I wonder what renameNode does to implemented interfaces.
13:51
hsivonen
looks up DOM3
13:52
<hsivonen>
oh. it returns a new object
14:05
<zcorpan_>
what are the use-cases of renameNode?
14:05
<zcorpan_>
seems like something that could be dropped in DOM5 Core... ;)
14:08
<hsivonen>
zcorpan_: suppose you have a script that walks an XHTML2 document tree and converts it to XHTML5. you'd renameNode separator to hr
14:08
<zcorpan_>
wouldn't you use XSLT for that?
14:09
<hsivonen>
zcorpan_: not according to http://hixie.ch/advocacy/xslt
14:10
<zcorpan_>
ok
14:10
<zcorpan_>
the bigger problem though is serving xhtml2 over the wire in the first place
14:10
<hsivonen>
sure
14:11
<hsivonen>
I don't know but I guess that renameNode exists to have full API-provided mutability
14:11
<Philip`>
You could be using DOM methods on the server side to transform something into XHTML5
14:11
<hsivonen>
I'd expect to see more use case for renameNode on server-side Java apps than in scripts running in browsers
14:11
<hsivonen>
cases
14:12
<zcorpan_>
yeah
14:12
<zcorpan_>
for client side i can only see the use-case of writing obscure test cases
14:13
<hsivonen>
on the face of it, it looks like a "nice to have", "for completeness" feature
14:13
<zcorpan_>
so it could have been useful to me today if it was implemented anywhere
14:13
<zcorpan_>
yeah
14:23
<billyjack>
fwiw, I expect we are going to see a lot more developers making use of client-side XSLT now that browser support has improved significantly
14:23
<MikeSmith>
for example, Webkit adding support for the node-set() function
14:24
<MikeSmith>
it is painful to try to get much done on XSLT 1.0 without node-set()
14:24
<MikeSmith>
s/done on/done itn/
14:24
<MikeSmith>
done in
14:24
<MikeSmith>
in in in
14:41
<Philip`>
When counting tag frequency, it's a good idea to not parse PDF files
15:14
<gsnedders>
Philip`: what makes you think that? :P
15:16
<gsnedders>
now, back to writing a 1:1 implementation of the common microsyntaxes
15:17
<Philip`>
The number of tag names which are ~150 characters long and only ever appear once and are full of non-ASCII characters does suggest that some filtering out of non-HTML files would help give slightly cleaner results :-)
15:18
<hsivonen>
the internationalization bomb is about to go off regarding ratios...
15:18
<gsnedders>
hsivonen: I saw. We seem to be managing fine with how UAs currently work, though…
15:19
<Philip`>
I still get a relatively large number of tags named scr"+"ipt - I'm not sure if that suggests it would be worth using a whole HTML parser when trying to do surveys of tag frequency, just to find and exclude the non-PCDATA bits
15:20
<hsivonen>
gsnedders: what do you mean? are there shipped implementations of the ratio algorithm except mine inside the conformance checker?
15:20
<gsnedders>
hsivonen: I mean for the other places where numbers are used (@height and @width will be the most widespread)
15:21
<hsivonen>
gsnedders: those are attributes--not element content
15:22
<gsnedders>
hsivonen: that's true, but how much of a difference should it really make?
15:22
<hsivonen>
gsnedders: it makes *all* the difference as far as internationazation politics go
15:22
<hsivonen>
gotta go
15:23
<Philip`>
Do 'international' people actually use number characters other than 0-9?
15:23
<Philip`>
I don't know how to find out, but at least http://www.aljazeera.net/ and http://divyabhaskar.co.in/ appear to just have normal digits everywhere
15:26
<Philip`>
...and I believe I remember a Hebrew-speaking person say they never use the Hebrew number system
15:44
<Dashiva>
Philip`: Which context?
15:48
<Philip`>
Dashiva: Which context for what?
15:48
<Dashiva>
Never using hebrew number system
15:51
<Philip`>
Ah - I can't really remember now, since it was a long time ago and I wasn't paying that much attention at the time :-p
15:52
<Philip`>
I think it was originally in the context of i18ning a game's UI, but I can't remember what context their comment was in
15:53
<Philip`>
(I wanted to support duodecimal Quenya numbers too, but I eventually realised that was just stupid)
16:05
<MikeSmith>
Philip` - Kanji for numbers are used quite commonly in Japanese and Chinese text
16:05
<MikeSmith>
particuarly when it's written vertically
16:05
<MikeSmith>
but also in e-mail
16:05
<MikeSmith>
and other contexts I'm probably forgetting about
16:06
<MikeSmith>
addresses
16:06
<MikeSmith>
postal addresses
16:41
<Dashiva>
MikeSmith: Calculations and time seems to be pretty dominated by roman numerals, though
20:05
<zcorpan_>
hsivonen: opera's fosterparenting doesn't seem generic handling in the css... (using divs with display:table etc doesn't result in the same thing)
20:06
<zcorpan_>
s/seem/seem to be/
20:08
<zcorpan_>
hsivonen: opera seems to use an anonymous caption element around the misplaced content
20:09
<zcorpan_>
(which is similar to what ie does -- only difference is ie actually adds the anonymous element to the dom)
20:09
<hsivonen>
zcorpan_: ok. interesting
20:10
<zcorpan_>
e.g. try caption-side:bottom
20:11
<zcorpan_>
or * { display:block }
20:14
<hsivonen>
I wonder why hyatt and Hixie didn't put a caption there. not worth the trouble?
20:15
<zcorpan_>
instead of moving things around in the dom?
20:20
<hsivonen>
no, in addition to moving to a different place
20:22
<zcorpan_>
what's the benefit?
20:23
<hsivonen>
zcorpan_: matching IE
20:24
<hsivonen>
zcorpan_: probably not worth it
20:24
<zcorpan_>
it does affect rendering in some cases, but firefox and safari seem to not need it
20:25
<zcorpan_>
and if you were to add a real "caption" element to the dom, you might break other things
20:25
<zcorpan_>
ie adds an element with the empty string as tag name
20:26
<zcorpan_>
how do you select that with css? :)
20:27
<hsivonen>
oh. crazy.
20:28
<zcorpan_>
in any case, if we were to use the caption approach, then we wouldn't need to move things around in the dom
20:29
<othermaciej>
empty string as tag name? for what?
20:30
<zcorpan_>
othermaciej: parent element of text node "foo" in <table><tr><td>x</tr>foo<tr><td>y</table>
20:31
<othermaciej>
awesome
20:33
<zcorpan_>
come to think of it, opera probably uses the same anonymous element approach with misnested block in inline when the inline styling gets applied to text nodes even though the inline element isn't an ancestor to the text node
23:57
<MikeSmith>
Dashiva - about usage in Japan of kanji numerals vs roman numerals: Yeah, true that calculations are dominated by roman numerals
23:57
<MikeSmith>
but not true about time
23:58
<MikeSmith>
at least in writing about time in the body of a message or document or whatever
23:58
<MikeSmith>
when it's quite common to write stuff like 三時半
23:59
<MikeSmith>
=3:30
23:59
<MikeSmith>
or 一時間半
23:59
<MikeSmith>
to talk about something that takes 1.5 hours