00:03
<jcranmer>
everything should use, er, DES
00:04
<olliej>
jcranmer: xor!
00:04
<olliej>
:D
00:04
<franksalim>
Philip`, I think that counts as making a comment about Debian
00:05
<jcranmer>
olliej: ROT-13... twice, for extra security!
00:05
<olliej>
jcranmer: couple with double xor!
00:05
<olliej>
jcranmer: it will be awesome!
00:06
<Philip`>
olliej: Stream ciphers use XOR anyway, so you're not proposing anything new :-p
00:06
<olliej>
Philip`: O_o
00:07
<Philip`>
(XOR with an approximation of a one-time pad, in particular)
01:24
<Hixie>
uh oh, adam's patience is cracking
01:35
<Hixie>
not only is this <keygen> thing introducing me to completely new technologies and standards
01:35
<Hixie>
it's introducing me to completely new (to me) standards ORGANISATIONS.
01:36
Hixie
really could have gone his whole week without learning about the DER part of ASN.1 and not been any the poorer
03:35
<Hixie>
i just wrote a spec i don't understand
03:35
<Hixie>
that generaly bodes badly
03:35
<Hixie>
i hope there are no errors in it
03:46
<olliej>
Hixie: keygen?
03:46
<Hixie>
yeah
05:01
<MikeSmith>
Hixie: about "key pair generator control", usual nit about possible ambiguity due to lack of hyphen for compound adjective (or whatever the correct grammar term is)
05:01
<MikeSmith>
that is, should probably best be "key-pair generator control"
05:02
<MikeSmith>
unless it means "key pair-generator control"
05:05
<MikeSmith>
Hixie: value of challenge attribute is just a string? not additional constraints on it?
05:48
<shepazu>
Hixie: ping
05:49
<Hixie>
hey
05:50
<Hixie>
(my connection is extremely flaky, i apologise in advance if i go dark unexpectedly)
05:51
<shepazu>
Hixie: I'm working on the final version of the WebApps charter
05:51
<shepazu>
and I wanted to give the real names and timetables for the deliverables
05:52
<shepazu>
so...
05:52
<Hixie>
let's see
05:52
<Hixie>
there are four i'm editing, right?
05:52
Hixie
goes to find out what they're called right now
05:52
<shepazu>
Server-Sent Events, Web Sockets API, Web Storage, and Web Workers
05:53
<Hixie>
sounds right
05:53
<Hixie>
my optimistic timetable is FPWD ASAP, LC in October, CR in 2012, and REC in 2022 at the latest
05:54
<Hixie>
REC will probably be far earlier for most of those
05:54
<shepazu>
there's also Web DOM (silly name, if you ask me) which could arguably be added as a real-name of DOM Level 4
05:54
<Hixie>
oh is XBL2 still on that list too?
05:54
<shepazu>
http://www.w3.org/2008/12/webapps/charter-2009-proposed
05:54
<shepazu>
XBL2, yes
05:55
<Hixie>
the reason for having "Web" in the DOM Core spec is that it's forking away from the DOM Core implemented by server-side libraries
05:55
<Hixie>
Does "DOM Level 4" include DOM Range and DOM Traversal?
05:56
shepazu
gets that, but is a little leary of not being explicit about that in the recent charter review
05:56
<shepazu>
jiminy, I just don't know... I guess it's subject to interpretation
05:56
<Hixie>
i might be roped into a DOM Range rewrite next year
05:56
<Hixie>
(there's growing pressure to fix that spec)
05:56
<shepazu>
we are talking about it during the domain call tomorrow, and we really should straighten it out
05:57
<Hixie>
(and add new APIs to it for things like rich text editing)
05:57
<shepazu>
Hixie: arguably, that could be part of DOM4
05:57
<shepazu>
but...
05:57
<shepazu>
I propose that we write all of this up in the wiki and get our ducks in a row
05:58
<shepazu>
I already feel like we are playing a little fast-and-loose on some of these
05:58
<shepazu>
Hixie: same timetable for all 4 of the "new" specs?
05:58
<shepazu>
ok, that's easy for me
05:58
<Hixie>
yeah
05:59
<Hixie>
xbl's timetable as in the spec now is probably ok, though if we miss the expected date for REC i won't be surprised
05:59
<Hixie>
hard for us to predict when REC will happen
05:59
<shepazu>
I imagine we will have a new charter (and AC review) before then
05:59
<Hixie>
yeah
06:00
<shepazu>
ok, if you could update the wiki with what you see as a sort of spec roundup and roadmap (for the specs you are involved in) that would help
06:00
<Hixie>
uri?
06:00
<shepazu>
even if it's beyond the scope of the current charter
06:00
<Hixie>
not sure which wiki this is
06:01
<shepazu>
anywhere in here: http://www.w3.org/2008/webapps/wiki/Main_Page
06:01
<shepazu>
I'm a little confused about where we are on XHR... I would have expected us to be long done with that
06:02
<Hixie>
i think there is a "soft" dependency on html5, in that anne is wanting to be in sync with things like the event queue mechanism
06:02
<shepazu>
happily, the Media Object/MAXIM spec looks to be obsoleted soon by the Media Annotations WG
06:02
<shepazu>
so, I think that will probably drop out in the next charter
06:03
<shepazu>
I'm also a little concerned about where the Origin header is headed
06:03
<shepazu>
last I saw, Adam Barth pretty much said he wasn't going forward with it in IETF, so I'm not sure where that leaves CORS
06:04
<shepazu>
and IETF collaboration in general...
06:04
<shepazu>
sigh.
06:04
<Hixie>
not sure what you want me to do on the wiki... should i just create a section called "hixie's drafts" and braindump into that? or a separate page?
06:05
<shepazu>
heh... that would be fine... it might give some context... I can massage it into the larger framework
06:05
<Hixie>
k
06:05
<shepazu>
yeah, a new page... "future development"
06:05
<shepazu>
thanks
06:05
<shepazu>
not urgent, but when you get time...
06:06
<shepazu>
the sooner the better, from the perspective of finalizing the current charter
06:06
<Hixie>
doing it now
06:07
<Hixie>
i created an account but it doesn't seem to let me edit?
06:07
<shepazu>
wait a mo
06:07
<Hixie>
ironically it seems fine with me _moving_ pages
06:08
<shepazu>
oh, mediawiki...
06:09
<shepazu>
ok, you should be able to edit now
06:09
<Hixie>
cool thanks
06:09
<shepazu>
yop
06:16
<Hixie>
http://www.w3.org/2008/webapps/wiki/Future_development
06:18
<Hixie>
shepazu: it's probably time to call the time of death on the Window Object spec, too
06:18
<Hixie>
realistically speaking i don't see anyone volunteering to do the significant work now required on that draft, and even if someone did come along, their time would probably be best spent doing more important things
06:19
<Hixie>
speaking of more important things, is WebIDL on your list?
06:19
<Hixie>
oh, there it is
06:20
<Hixie>
might be worth making that list alphabetical
06:23
<shepazu>
working on that now
06:24
<shepazu>
I'm not going to add or remove any deliverables until the WG formally decides to do so
08:16
<zcorpan>
Hixie: are you making <keygen> conforming?
08:20
<Hixie>
zcorpan: yes
08:22
<zcorpan>
cool
08:23
zcorpan
looks at the diff
08:41
<zcorpan>
Hixie: is <marquee> next? :)
08:43
<zcorpan>
wonder if marquee can be implemented using css transitions
10:08
MikeSmith
looks around for zcorpan
10:08
jgraham
has more success being in the same room
10:08
<MikeSmith>
hsivonen: http://bugzilla.validator.nu/attachment.cgi?id=75
10:09
<MikeSmith>
jgraham: is he there? poke him with a stick to wake
10:09
<zcorpan>
MikeSmith: pong
10:09
<MikeSmith>
hej
10:09
<jgraham>
They don't give us long enough sticks here
10:09
<MikeSmith>
zcorpan: http://bugzilla.validator.nu/show_bug.cgi?id=474
10:10
<MikeSmith>
jgraham: you get longer sticks with more seniority
10:10
<zcorpan>
MikeSmith: i think that's a case where hsivonen disagrees with what Hixie specced
10:10
<MikeSmith>
ah
10:11
<MikeSmith>
hmm, so I'm wondering to do with that test case as far as db.json/validator-tester.py goes
10:12
<zcorpan>
MikeSmith: i think i had filed a bug about that already
10:12
<MikeSmith>
oh
10:12
<zcorpan>
MikeSmith: since i have a test case :)
10:12
<MikeSmith>
I'll look and marked mine as a dupe if so
10:13
<MikeSmith>
http://bugzilla.validator.nu/attachment.cgi?id=75 is a proposed fix for your just-a-hash test
10:14
<MikeSmith>
ah, i find http://bugzilla.validator.nu/show_bug.cgi?id=323 also
10:14
<MikeSmith>
"<img usemap=#> validates"
10:14
<hsivonen>
MikeSmith: for the charset thing, what Hixie specced would be a lot of work for Simplified Chineses encodings for the sake of theoretical purity without helping authors
10:15
<hsivonen>
MikeSmith: http://bugzilla.validator.nu/attachment.cgi?id=75 looks good
10:15
<MikeSmith>
hsivonen: thanks
10:15
<MikeSmith>
hsivonen: about the encoding issue, I see
10:15
<MikeSmith>
so I'm wondering what to do about test cases for that kind of stuff
10:16
<hsivonen>
MikeSmith: I suppose the URLs could be listed in a text file in the same directory as the db.json
10:16
<MikeSmith>
hsivonen: just a static file, you mean?
10:17
<hsivonen>
MikeSmith: yes. to remember which tests didn't go into the db for any reason
10:17
<MikeSmith>
OK
10:17
<MikeSmith>
willful-violations-of-html5.txt ?
10:17
<hsivonen>
yeah :-)
10:17
<MikeSmith>
heh
10:17
<MikeSmith>
I will make it so
10:19
<MikeSmith>
hsivonen: so I plan to do a patch for keygen, and want to also write some test cases for it
10:20
<hsivonen>
MikeSmith: great!
10:21
<MikeSmith>
should I add them into svn, or put them at my own site and use db.json?
10:22
<hsivonen>
I think having them is svn plus site somewhere & db.json would be ideal
10:22
<hsivonen>
I'm not sure if whattf.org autoupdates to serve the tests over http
10:23
<hsivonen>
MikeSmith: I think the tests will show up under http://syntax.whattf.org/relaxng/tests/ if you put them in svn
10:23
<hsivonen>
MikeSmith: I don't know how often fantasai's script refreshes directory
10:44
<MikeSmith>
hsivonen: OK, I'll mess around with it a bit
10:48
<gsnedders>
Woah. RSS only allows enclosure URIs to be in the HTTP scheme.
10:49
<gsnedders>
HTTPS is invalid.
10:51
<Philip`>
What about FTP?
10:52
<gsnedders>
Invalid.
10:53
<Philip`>
Shame
11:00
<annevk3>
hsivonen, you realize <keygen> is a major pain for you because Mozilla implements it like <isindex>?
11:05
<annevk3>
hmm, RFC2459 has been obsoleted several times...
11:07
<zcorpan>
Hixie: there is no parsing problem with <h1><p>Typeline</p><p>Mainline</p></h1>
11:14
<hsivonen>
annevk3: no, I haven't realized that yet
11:15
<hsivonen>
annevk3: what about <keygen> in XHTML?
11:17
<zcorpan>
hsivonen: html5 makes keygen a proper void element instead of an isindex-like parser macro
11:18
<Philip`>
<keygen> in XHTML in Firefox (2) doesn't do anything at all
11:19
<Philip`>
(http://canvex.lazyilluminati.com/misc/dom-viewer/x.html?%3C!DOCTYPE%20html%3E%0A%3Chtml%20xmlns%3D'http%3A%2F%2Fwww.w3.org%2F1999%2Fxhtml'%3E%3Cform%3E%3Ckeygen%20name%3D'x'%2F%3E%3Cinput%20type%3D'submit'%2F%3E%3C%2Fform%3E%3C%2Fhtml%3E )
11:19
<zcorpan>
hey how long have you had that xhtml viewer?
11:20
<Philip`>
Since at least 2007-08-25
11:21
<zcorpan>
you should make your tools easier to discover
11:21
<Philip`>
But then people might use them
11:22
<Philip`>
and it'll fill up my web server access logs because it sends a request for every keypress :-(
11:22
<zcorpan>
then increase the timeout
11:23
<jgraham>
You should make your tools easy to download so that we can run them locally and not fill your access logs :)
11:23
<annevk3>
Yeah, Gecko has no <keygen> support in XHTML.
11:23
<annevk3>
Opera does :)
11:24
<Philip`>
zcorpan: You should have been paying attention when I linked to it on http://krijnhoetmer.nl/irc-logs/whatwg/20070825#l-262
11:24
<Philip`>
I can't imagine why you'd fail to be aware of it
11:25
<annevk3>
WebKit expands it even in XHTML...
11:25
<zcorpan>
Philip`: maybe because the url was too similar to your html viewer that i didn't notice
11:26
<zcorpan>
annevk3: we expand isindex in xhtml, too, though
11:27
<Philip`>
jgraham: The server-side script is trivial (it just prints its query string) and the client-side stuff is just stolen from Hixie, so if anyone wants to set up something similar locally then it shouldn't take them any effort :-)
11:27
<Philip`>
zcorpan: Bah, excuses
11:28
Philip`
admittedly did have to look on his web server over SFTP to work out what the URL of the tool was
11:30
<annevk3>
zcorpan, yes, and <image>...
11:33
<hsivonen>
what does Dead DOM Viewer do?
11:33
<Philip`>
It does the opposite of what the Live one does, hence the name
11:34
<Philip`>
In particular it generates the page on the server rather than on the client
11:34
<hsivonen>
Philip`: parses XML on the server and reseriales?
11:34
<Philip`>
The purpose was to compare IE's processing of HTML in HTTP responses vs HTML in document.write
11:35
<Philip`>
and then I added the XHTML version too, for which there isn't really any point having the server-side version
11:35
<Philip`>
hsivonen: No, the server just echoes its query string
11:35
<hsivonen>
hmm. <image> is magic in WebKit but in body--not directly in html
11:36
<hsivonen>
why do I see HTML, HEAD, BODY in the dead viewer in WebKit?
11:36
<hsivonen>
oops. in Firefox
11:37
<Philip`>
(http://canvex.lazyilluminati.com/misc/dom-viewer/ is the non-XHTML version)
11:37
<hsivonen>
it's as though the Live and Dead were swapped in Gecko and WebKit. what's going on?
11:38
<Philip`>
It's possible that it's a bit buggy, and particularly that it could get confused about updating the DOM views at the right times
11:40
<Philip`>
It's just a quick hack and probably shouldn't be relied on for anything important :-)
11:41
<Philip`>
(Even the name doesn't make any sense, but I've never cared enough to try fixing it)
11:44
<Lachy>
Hixie, now that webstorage has been split out, it's quite annoying and confusing that the where the openDatabase() method is listed in the IDL for Window is still in the HTML5 spec
11:48
<annevk3>
yeah, there should probably be [XXX] interface Window { } in Web Storag
11:48
<annevk3>
e
11:48
<annevk3>
same goes for localStorage/sessionStorage
12:08
<Lachy>
I'll go file a bug about it
12:16
<Lachy>
We need to get a WebStorage component added to Bugzilla for the WebApps WG. http://www.w3.org/Bugs/Public/enter_bug.cgi?product=WebAppsWG
12:17
<Lachy>
I just filed the bug as an HTML5 spec bug instead
14:19
<hsivonen>
http://stackoverflow.com/questions/725261/what-is-the-semantic-web
14:20
<Dashiva>
The semantic web is like a box of chocolates
14:32
<jgraham>
More people misunderstanding <header>
14:46
<hsivonen>
<header> as currently defined, is FAIL
14:46
<hsivonen>
s/,//
14:47
<annevk3>
maybe we should just stick to a bunch of <div>s
14:47
<hsivonen>
The W3C uses RDFa in text/html: http://www.w3.org/2001/sw/
14:51
<hsivonen>
<header> should probably be split into <banner> (the case people think <header> is for) and <hgroup> for grouping <h1> and <h2>
14:52
<jgraham>
<banner> sounds like <advert>
14:52
<hsivonen>
jgraham: banner is from ARIA landmarks
14:52
<hsivonen>
jgraham: I don't particulary like the connotations of 'banner'
14:52
<jgraham>
hsivonen: It's not my fault if aria does dumb things :)
14:53
<jgraham>
Or rather I don't think we should feel compelled to spread those dumb things to the rest of html
14:53
<hsivonen>
jgraham: well, HTML5 sucks here, too
14:53
<jgraham>
Sure
14:53
<jgraham>
(Doesn't aria also have role="advertising" or something?)
14:53
<hsivonen>
which reminds me that now that April 1st has passed, I should blog in reply to the adactio post.
14:54
<zcorpan>
so this page http://www.systemrequirementslab.com/bonus/t/analysis2.tmpl?registration_option_id=7840 fails to work given these rules http://www.whatwg.org/specs/web-apps/current-work/multipage/dom.html#dom-document-nameditem
14:54
<hsivonen>
jgraham: the semantics of banner were changed not to imply advertising
14:54
<zcorpan>
i wonder if the spec can be adjusted in some way
14:55
<zcorpan>
the page does <object id=ie><comment><object id=moz><table><tr><td>&nbsp;</table>
14:55
<annevk3>
zcorpan, why does it fail?
14:55
<zcorpan>
along with if (null != document.moz) ... else if (null != document.ie) ... else return // error
14:55
<jgraham>
hsivonen: Anyway HTML5 should probably have <header> which means role="banner" and <hgroup> which means "compound heading"
14:55
<annevk3>
ah, document.moz is not found?
14:55
<zcorpan>
both objects are ignored because they have "fallback"
14:56
<zcorpan>
right
14:56
<annevk3>
maybe the innermost <object> or <embed> should count and the rest be ignored?
14:56
<hsivonen>
jgraham: I think the words "header" and "heading" should be avoided in element naming
14:57
<jgraham>
hsivonen: Maybe. <banner> sucks at least as much though
14:57
<jgraham>
Or rather "Yes"
14:57
<zcorpan>
annevk3: i.e. an object is fallback-free if it doesn't have any object or embed descendants?
14:58
<annevk3>
zcorpan, ja
14:58
<Philip`>
jgraham: What's the use case for <header> in that case?
14:58
<zcorpan>
makes sense i guess. what about <applet>? any other elements?
14:59
<annevk3>
<applet name=x> <object name=x> </object> </applet> both are returned, no?
14:59
<zcorpan>
i was thinking about applet in object
14:59
<jgraham>
Philip`: Authors like being able to structure their source in a readable way. It probably also helps with CSS selectors and maybe with a11y
15:00
<hsivonen>
it would also be nice if Opera Mobile took it into account when guessing where the main content starts
15:00
<annevk3>
zcorpan, good point
15:01
<Philip`>
jgraham: <div id=header> lets them structure their source in the same way, with the same readability
15:02
<annevk3>
zcorpan, <img> too
15:02
<zcorpan>
annevk3: not sure about <img>
15:02
<Philip`>
jgraham: Oh, if it maybe helps with a11y then we are morally required to include it
15:03
<zcorpan>
annevk3: consider that page having a spacer gif instead of &nbsp;
15:04
<annevk3>
zcorpan, it seems that for Mozilla <applet>, <img> etc. do not matter
15:04
<annevk3>
but neither does <embed>
15:04
<annevk3>
or <object>
15:05
<annevk3>
sigh
15:05
<jgraham>
Philip`: <div id=header><div id=something><div></div>[...]</div> is not as easy to read as <header><div id=something><div></div>[...]</header>
15:06
<jgraham>
(similarly div#header is slightly harder to read in CSS than header)
15:07
<zcorpan>
annevk3: moz does something like if (document.getElementById(x) && blah) else blah for document.x
15:07
<beowulf>
i like <banner>
15:07
<webben>
beowulf: banner is very ambiguous
15:07
<Philip`>
jgraham: Try using indentation and quotes and then it'll be fine :-)
15:07
<webben>
e.g. sometimes used for ads
15:08
<Philip`>
But people won't use <banner> for ads because it would be trivial to filter
15:08
<annevk3>
zcorpan, what is && blah?
15:08
<Philip`>
therefore it'll be used only by people writing page headers
15:09
<beowulf>
webben: probably, though not in my case, i've been reading about newspaper layout, terms, whatnot
15:10
<jgraham>
Philip`: Indentation helps a bit. But it's still easier to read if the end tag matches the start tag
15:10
<jgraham>
(otherwise everyone wwould love lisp :) )
15:10
<beowulf>
and <banner> is probably not used for ads or anything right now :)
15:10
<Philip`>
http://zepheira.com/news/releases/20070711.html mentions PURLs being critical to the Semantic Web - it seems strange to me that a Web should depend so much on a centralised repository; I guess they don't want to accept the true nature of the web, with its instability and ephemerality and general rubbishness
15:10
<Philip`>
jgraham: s/lisp/Python/
15:10
<zcorpan>
annevk3: && document.getElementById(x) is one of img, embed, etc
15:11
<Philip`>
jgraham: and everyone does love Python
15:11
<Philip`>
jgraham: therefore disproving your argument
15:11
Philip`
wins
15:11
<jgraham>
Philip`: No because python doesn't have end tags
15:11
<annevk3>
zcorpan, I can't really reproduce that
15:11
<jgraham>
or whatever
15:11
<Philip`>
jgraham: It does have end tags - they're just invisible DEDENT tokens
15:11
<jgraham>
:p
15:12
<jgraham>
Philip`: Right but they're not in the source
15:12
<jgraham>
Er, that made no sense
15:12
<Philip`>
jgraham: How can you tell the invisible tokens aren't in the source?
15:12
<Philip`>
jgraham: They're invisible!
15:12
<zcorpan>
annevk3: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3Cembed%20id%3Dx%3E%3Cembed%20id%3Dx%3E%3Cscript%3Ew%28document.x%29%3C%2Fscript%3E
15:13
<Philip`>
jgraham: It's possible that I'm completely misunderstanding your argument
15:13
webben
is getting a vision of a lolcat version of this "invisible end tag" argument.
15:13
<jgraham>
I think what I mean is that the python philosophy of uniform indentation as a marker doesn't transfer well onto the web
15:14
<jgraham>
Because the web doesn't enforce uniform indentation
15:14
<annevk3>
zcorpan: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3Cimg%20name%3Dx%3E%3Cembed%20id%3Dx%3E%3Cscript%3Ew(document.x)%3C%2Fscript%3E
15:15
<Philip`>
jgraham: but people do seem to prefer programming languages where end tags aren't like start tags, e.g. C (and JavaScript and Java etc) and Python, vs Basic (if ... endif) and Bash (if ... fi)
15:15
<Philip`>
s/people/I/
15:16
<Philip`>
Anyway, I've forgotten how this discussion started or what I was trying to say
15:16
<Philip`>
s/or/and/
15:16
<annevk3>
zcorpan, maybe reverse engineering WebKit is more sane or does it also fail on that page?
15:16
<zcorpan>
annevk3: webkit also fails
15:16
<zcorpan>
annevk3: webkit, opera and html5 do basically the same thing
15:18
<hsivonen>
whee! "The XPath evaluator must never call this with a null or empty argument, because the result of doing this is undefined."
15:18
<jgraham>
hsivonen: Is that a normative condition on user code?
15:18
<jgraham>
s/user/web developer/
15:18
<annevk3>
zcorpan, yeah. I guess doing what Firefox does breaks some other set of pages?
15:19
<hsivonen>
jgraham: it's a normative statement on XPath implentations, so no
15:19
<zcorpan>
annevk3: likely. it's also less sane and less compatible with ie
15:19
<zcorpan>
last i looked
15:19
<jgraham>
hsivonen: Ah, OK
15:19
<hsivonen>
jgraham: though it seems to me calling into user code can be undefined :-)
15:19
<hsivonen>
in any case :-)
15:20
<Philip`>
But is it defined that the user code must follow certain rules?
15:20
<annevk3>
zcorpan, ignoring <object> if it has a descendant named element maybe?
15:20
hsivonen
wonders what happens if XPathNSResolver returns the empty string
15:20
<annevk3>
zcorpan, it seems a bit risky though
15:23
<annevk3>
zcorpan, and it's hard to get good data on this though I suppose we could try to get anedotical data
15:23
<annevk3>
anecdotal, even
15:23
<zcorpan>
i'm all for doing research :)
15:27
<annevk3>
maybe Philip` can find pages that have <object id=... or name=...> + child nodes
15:28
<zcorpan>
ignoring <param> and whitespace
15:28
<zcorpan>
Philip`: ^
15:31
<hsivonen>
do markers ever come off the list of active formatting elements in ways other than via "clear up to last marker"?
15:41
<Philip`>
annevk3: I can find <object id> and <object name> easily
15:50
<hsivonen>
aaargh. Gecko allows an XPathNSResolver to return "" to bind a prefix to no namespace
15:51
<hsivonen>
doesn't work in WebKit
15:52
<hsivonen>
nor in Opera
15:52
<hsivonen>
maybe there's hope for zapping that behavior from Gecko, too.
15:53
<zcorpan>
Philip`: can you easily find <object id/name> that have child elements other than <param>?
15:54
<Philip`>
annevk3: Actually, I could find them easily if I had a decent XML grep tool that actually worked
15:54
<Philip`>
but I don't
15:55
<Philip`>
so you can have http://philip.html5.org/data/object-with-id-or-name.xml.bz2 which is all <object>s
15:55
<Philip`>
and feel free to write code to extract what you want from it
15:55
<Philip`>
(That's directed to zcorpan too :-) )
15:58
<zcorpan>
Philip`: thanks
15:59
Philip`
hopes the structure of that file is sufficiently obvious
15:59
<Philip`>
(Hooray for human-readable file formats)
16:05
<zcorpan>
hmm bad idea to open that xml file in a browser
16:06
<Philip`>
It's only 20MB
16:06
<Philip`>
Oh, but I suppose it has a few <object>s
16:06
<zcorpan>
almost 27000 object elements
16:06
<Philip`>
which browsers sillily try to interpret
16:06
<Philip`>
despite it not even being in an <html>
16:07
zcorpan
removes namespace declarations
16:07
Philip`
didn't add namespace declarations on purpose, they were just what his XML serialiser did
16:07
<zcorpan>
oops, my script didn't escape <
16:08
<zcorpan>
ok so 2600 don't have an embed descendant
16:09
<zcorpan>
<object classid="clsid:d27cdb6e-ae6d-11cf-96b8-444553540000" height="475" id="mymovie" width="830">
16:09
<zcorpan>
<param name="movie" value="../images/home.swf"/>
16:09
<zcorpan>
<object data="../images/home.swf" height="475" type="application/x-shockwave-flash" width="830">
16:09
<zcorpan>
<p>
16:09
<zcorpan>
<img alt="Our applications fields : Veterinary Lab, Haematology and Pathology" height="475" src="../images/homepics.jpg" width="830"/>
16:11
<zcorpan>
http://www.orphee-medical.com/ ...if the inner object had an id then i think we'd want document.id to match it despite the <img>
16:13
<zcorpan>
298 have nested object elements
16:26
<zcorpan>
copy-paste has resulted in lots of "FAIL (the browser should render some flash content, not this)."
16:27
<zcorpan>
Hixie: take note - don't write copy-paste examples as test cases
16:28
zcorpan
will continue tomorrow
16:32
<Philip`>
Where does that FAIL come from originally?
16:32
<Philip`>
All I see is http://www.adobe.com/jp/devnet/flash/articles/swfobject_05.html
16:49
<hsivonen>
I'm amused to find that Opera already violates XPath invariants in a way that cannot be explained by a custom DOM-to-XDM mapping
17:13
jgraham
notes that hsivonen is probably the first person in the history of all creation to be aused by XPath
17:13
<jgraham>
*amused
17:19
Hixie
gets off the phone with a yahoo searchmonkey engineer
17:19
<Hixie>
who had very interesting insights into rdf in html
17:20
<gsnedders>
Wait, what?
17:20
<gsnedders>
"Hixie gets off the phone" — Does not compute.
17:20
<Hixie>
i had to come to work and go to my cube to do this
17:20
<jgraham>
Hixie: Are you just teasing us with confidential information?
17:20
<Hixie>
first time i've been here in months
17:20
<Hixie>
jgraham: no
17:21
<gsnedders>
Hixie: I guessed you were at work from the lack of a phone anywhere else :P
17:21
<Hixie>
paraphrasing, and apologies if i misrepresent his opinions, he pointed out that microformats were prettier but there was the perception of lack of extensibility
17:22
<jgraham>
So far, so few surprises
17:22
<Hixie>
and that the lack of a clear mapping to RDF actually did hurt their use of microformats
17:22
<gsnedders>
GRDDL!
17:22
<Hixie>
but that the rdfa syntax, use of URIs as identifiers, and the need to use explicit URIs that aren't the Web page's own URI as subjects were all things that caused confusion
17:23
<jgraham>
gsnedders: Doesn't GRDDL have some weird thing that the author needs to provide the GRDDL?
17:23
<gsnedders>
I dunno.
17:23
<gsnedders>
I thought it was something along the lines of @profile being magic
17:23
jgraham
thought that GRDDL was basically a way of using an XSLT stylesheet to transform some HTML into some RDF
17:23
<gsnedders>
Yeah, basically
17:24
<jgraham>
And that the author had to put the link to the sty;esheet in @profile
17:24
<Hixie>
no
17:24
<Hixie>
they have to put a link to a page that has a link to the xslt
17:25
<jgraham>
Ah, yes, I guess it it important to be correct on that point. @profile is an option for the link :)
17:26
<jgraham>
Hixie: So the summary is that Yahoo would be better served by something with simpler syntax than RDFa but an easier extensibility story than microformats
17:26
<jgraham>
?
17:27
<gsnedders>
Are free video downloads from iTunes encumbered?
17:36
<Philip`>
jgraham: And by something with "a clear mapping to RDF"
17:38
<jgraham>
Philip`: Good point
17:49
<Hixie>
jgraham: yeah
17:51
<jgraham>
Good luck convincing W3C that URIs are part of the problem, not part of the solution :)
17:52
<Hixie>
no need, so long as we define a way to map strings to URIs without having to declare prefixes explicitly
17:53
<jgraham>
I guess I can't complain since I have previously advocated that approach. But it does seem like adding another layer of indirection to paper over the problem
17:53
<jgraham>
It also seems like the only plausible way forward
18:06
<Hixie>
i don't understand what http://www.w3.org/mid/b6bb4d890904070441p248bedccsaf3f2971f282c18⊙mgc would mean, concretely
18:11
<Philip`>
Something about taking fancy new web technologies and making them usable by normal web developers, I guess
18:11
<Philip`>
though that's still not exactly concrete
18:14
<Philip`>
Also it might be talking about making HTML more attractive to teenage girls
18:16
<webben>
Looks like it just means deliver the key features needed for webapps fast.
18:16
<mpilgrim>
i think it's more along the lines of "we need a site like chromeexperiments.com to showcase html 5"
18:16
<mpilgrim>
html5experiments.com?
18:16
<mpilgrim>
i think that's just called annevankesteren.nl
18:20
<Philip`>
That's the catwalk
18:22
<gsnedders>
mpilgrim: But he does SVG too! That's an SVG experiment!
18:24
<Philip`>
Topshop would be making it so ordinary people can create sites like that, or like Gmail, without it being insanely hard and insufficiently general-purpose
18:29
<jgraham>
Compsring html5 to topshop? This must be some sort of new low
18:30
<gsnedders>
I thought we had already got as low as we can go.
18:32
<jgraham>
Do we have any examples of really crappy but popular html 5? We could call them the html 5 "new look"
18:33
<jgraham>
Straw poll: "data is" or "data are"?
18:33
<gsnedders>
is
18:34
<gsnedders>
(Even though I'm well-aware that's technically incorrect.)
18:34
<Philip`>
Avoid phrases that force you to make the distinction, because you'll annoy somebody no matter which you choose
18:35
<Philip`>
('is' will annoy people because it's wrong, 'are' will annoy people because it's stupid and ugly)
18:35
<inimino>
depends if you are treating data as a plural or as a mass noun
18:35
<jgraham>
Philip`: The hidden context is teaching english to non-native speakers
18:35
<gsnedders>
data is the plural of datum, so it should never be singular
18:36
<Philip`>
Have you ever heard anybody use the word "datum"?
18:36
<jcranmer>
I've used it once or twice
18:36
<jgraham>
gsnedders: I don't think that is a good argument becaue it doesn't preclude data also being used as a mass noun
18:36
<gsnedders>
Philip`: Yes
18:37
<Philip`>
particularly somebody who isn't sounding unbelievably pretentious when saying it? ;-)
18:37
<gsnedders>
Philip`: Um, no.
18:37
<gsnedders>
Philip`: But I rarely hear people who don't sound unbelievably pretentious
18:37
<jcranmer>
I believe ArcGIS uses datum
18:37
<jgraham>
gsnedders: Just by speaking?
18:37
<gsnedders>
jgraham: yeah
18:38
<jgraham>
gsnedders: You have really chavvy standards :)
18:38
<gsnedders>
:D
18:38
<gsnedders>
jgraham: And dude, chav is _so_ last year.
18:38
<jgraham>
gsnedders: Probably depends where you are from a bit
18:39
<Philip`>
gsnedders: "_so_" is so last year, too
18:39
<gsnedders>
(Well, really it's several years ago here.)
18:39
<gsnedders>
It's all about the neds nowadays
18:40
<jcranmer>
... ばか
18:40
<gsnedders>
Huh?
18:40
<gsnedders>
UnknownLanguageException
18:41
<jcranmer>
can't recognize hiragana, can you?
18:41
gsnedders
catches the exception, and carries on
18:41
<Philip`>
noun 1. bonehead [sl.] 2. clod [sl.] 3. stupidity
18:42
<Philip`>
Being unable to translate between languages is _so_ last year
18:43
<jcranmer>
...
18:43
<jcranmer>
あほ
18:44
<Philip`>
"Idiot"?
18:44
<Philip`>
Google Translate isn't so good on that
18:45
<jcranmer>
closer to asshole, IIRC
18:46
<jcranmer>
much stronger form of ばか
18:47
Philip`
was looking at http://translate.google.com/translate?prev=_t&hl=en&ie=UTF-8&u=http%3A%2F%2Fja.wikipedia.org%2Fwiki%2FWikipedia%3A%E7%B4%A2%E5%BC%95_%E3%81%82%E3%81%BB&sl=ja&tl=en&history_state0=ja%7Cen%7C%E3%81%82%E3%81%BB
18:47
<Philip`>
which includes "Idiot, right (State of Arizona)" - I'm not sure what that's meant to mean
18:48
<Philip`>
"Idiot, right (Ajo), southern Arizona, United States, the city is located in PIMA County. Towns and cities in the official land belongs not to be determined, CDP. The population is 3,705 people (2000 census)."
18:49
<gsnedders>
What's .ta?
18:49
<jcranmer>
`Ajo' sounds like it's from spanish origins
18:49
<jcranmer>
so it's pronunciation is like `Aho'
18:49
<jcranmer>
or, should I say, ah-ho
18:50
<jcranmer>
transliterate that into Japanese to get, in the rōmaji form aho
18:50
<jcranmer>
あほ == aho
18:50
<Philip`>
I guess it's http://en.wikipedia.org/wiki/Ajo,_Arizona but that page doesn't talk about idiots at all
18:50
<jcranmer>
google's not picking up the fact that it's a transliterations
18:50
<gsnedders>
Meh. It's meant to be raining in Dundee for the next couple of days.
18:51
<jcranmer>
yep
18:52
<jcranmer>
the wikipedia page has it in katakana--which is generally how you write transliterated proper names in Japanese
18:52
<jcranmer>
so あほ and アホ are both aho, but the former would be the word aho and the latter the place Aho
18:53
<jcranmer>
or the name, if someone were cruel enough to name someone that :-)
18:53
<jgraham>
gsnedders: Move to Fort William. Then it would always be raining so you wouldn't get so disappointed
18:53
gsnedders
went to Fort William once…
18:53
<gsnedders>
It wasn't raining.
18:53
<gsnedders>
Also: I doubt Fort William has anywhere to buy a suit from.
18:55
<jgraham>
gsnedders: I think it does. But I only went there once too (and I think it was raining)
18:56
<gsnedders>
(If I didn't need to get a suit, I would have no reason to go to Dundee.)
18:56
<gsnedders>
It's no better in Edinburgh, so I don't gain anything by going their instead.
18:56
<jgraham>
gsnedders: There are places that will sell suits where it is unlikely to be raining
18:57
<jgraham>
Saudia Arabia for example
18:57
<gsnedders>
jgraham: They are not however practical places to go to get a suit
18:57
<erlehmann>
can anyone with working ff 3.5 or opera 10 test this ? http://blog.dieweltistgarnichtso.net/interview-moot-of-4chan-part-1
18:57
<takkaria>
I daresay Saudia Arabia is a more practical place than say, Devon
18:58
<gsnedders>
Wednesday afternoon is meant to be all right in Dundee.
18:58
<takkaria>
or Blackpool
18:58
<gsnedders>
takkaria: Why?
18:58
<takkaria>
well, Blackpool's a bit naff really, isn't it?
18:58
<gsnedders>
Dunno. Never been there.
18:59
<erlehmann>
got it — the server must serve the file using the correct MIME type
18:59
<jgraham>
gsnedders: On he other hand if the only options are dundee or Edinburgh, at least Edinburgh is nice
19:00
<gsnedders>
That is true.
19:01
<gsnedders>
Also, Dundee wastes less time
19:59
<gsnedders>
erlehmann: "spending […] every" should be "spending… I was up every"
19:59
<erlehmann>
thanks
19:59
<gsnedders>
His speech is too quick
20:01
<gsnedders>
erlehmann: "site […] seven servers," "site. I rearranged seven servers,"
20:03
<erlehmann>
gsnedders, actually, i did have the impression of understanding him when sitting on the opposing side of the desk
20:04
<gsnedders>
erlehmann: It's just about comprehsible, speech like that, when you're close to them, they're not projecting their voice, and it's not recorded :)
20:04
<erlehmann>
i hate AMR now
20:04
<erlehmann>
another reason to root my new googlephone
20:11
<Philip`>
http://hg.mozilla.org/mozilla-central/rev/cfd5af38abc6
21:13
gsnedders
guesses he ought to let the outliner just take a textarea
21:13
<gsnedders>
Seemingly multiple people ahve asked for it
21:26
<Lachy>
MikeSmith, yt?
23:27
<annevk3>
yay email