2009-04-01
[18:03:00.0000] <Hixie>
philip: i like the subtle change at the end

[18:06:00.0000] <Philip`>
/me couldn't think of anything more interesting to do

[22:30:00.0000] <MikeSmith>
nico1: for more on whatwg, see  http://www.whatwg.org/

[22:31:00.0000] <MikeSmith>
note also that this channel and #html-wg and #webapps are logged, and you can read the logs at http://krijnhoetmer.nl/irc-logs/

[22:31:01.0000] <MikeSmith>
mostly we talk about HTML5 and browser stuff

[22:32:00.0000] <nico1>
thanks a lot MikeSmith !

[23:03:00.0000] <Niictar24>
Hrm

[23:04:00.0000] <Niictar24>
/me contemplates reading up on the content model

[23:25:00.0000] <MikeSmith>
nico1: http://validator.nu/

[23:26:00.0000] <MikeSmith>
http://about.validator.nu/

[23:27:00.0000] <MikeSmith>
http://hsivonen.iki.fi/thesis/html5-conformance-checker.xhtml

[23:30:00.0000] <MikeSmith>
http://www.whatwg.org/specs/web-apps/current-work/multipage/

[23:31:00.0000] <MikeSmith>
nico1: ↑

[23:31:01.0000] <MikeSmith>
http://www.whatwg.org/specs/web-apps/current-work/multipage/syntax.html#parsing

[23:34:00.0000] <MikeSmith>
nico1: http://blog.whatwg.org/validatornu-html-parser-120

[23:34:01.0000] <MikeSmith>
http://about.validator.nu/htmlparser/

[23:35:00.0000] <MikeSmith>
http://code.google.com/p/html5lib/

[23:36:00.0000] <MikeSmith>
http://code.google.com/p/html5lib/issues/list

[23:53:00.0000] <hsivonen>
sigh at the adactio post. some really should explain why we don't like landmarks, but I guess blogging on April 1 is a bad idea

[23:54:00.0000] <hsivonen>
(or, rather, why we don't like architectural forms as the means of expressing landmarks)

[00:32:00.0000] <MikeSmith>
hsivonen: so I'm discovering a bit about schematron 

[00:32:01.0000] <MikeSmith>
e.g., it has a <name/> element to report the name of the context node

[00:33:00.0000] <MikeSmith>
so one can do, e.g.:

[00:33:01.0000] <MikeSmith>
[[

[00:33:02.0000] <MikeSmith>
<rule context='h:h1|h:h2|h:h3|h:h4|h:h5|h:h6'>

[00:33:03.0000] <MikeSmith>
  <report test='ancestor::h:footer'>

[00:33:04.0000] <MikeSmith>
    The <name/> element must not

[00:33:05.0000] <MikeSmith>
    appear as a descendant of the

[00:33:06.0000] <MikeSmith>
    &#x201C;footer&#x201D; element.

[00:33:07.0000] <MikeSmith>
  </report>

[00:33:08.0000] <MikeSmith>
]]

[00:35:00.0000] <hsivonen>
MikeSmith: does that work in Jing?

[00:36:00.0000] <MikeSmith>
yeah, it does

[00:36:01.0000] <MikeSmith>
I just tested it

[00:36:02.0000] <MikeSmith>
it's in Schematron 1.5, and Jing supports that

[00:36:03.0000] <MikeSmith>
or does now at least

[00:36:04.0000] <hsivonen>
MikeSmith: interesting. I knew Rick Jelliffe's XSLT implementation had something like that but I always thought it was a later addition

[00:36:05.0000] <MikeSmith>
I'd guess that it might now have previously

[00:37:00.0000] <hsivonen>
that's possible

[00:37:01.0000] <MikeSmith>
anyway, if I use <name>, it could eliminate a whole lotta redunancy that's in assertions.sch now

[00:38:00.0000] <hsivonen>
seems like a good idea

[00:42:00.0000] <hsivonen>
MikeSmith: http://bugzilla.validator.nu/attachment.cgi?id=70 looks good, except I don't see you using the label locators at all for location reporting.

[00:42:01.0000] <MikeSmith>
/me looks back at patch

[00:42:02.0000] <hsivonen>
MikeSmith: what's the purpose of collecting forVals into a set first instead of doing the id matching as part of the stack walk?

[00:45:00.0000] <MikeSmith>
hsivonen: purpose is mainly that's just the simplest way it initially occurred to me to do it.

[00:45:01.0000] <MikeSmith>
hsivonen: I should just set a boolean instead?

[00:45:02.0000] <hsivonen>
MikeSmith: preferably, yes

[00:45:03.0000] <MikeSmith>
when it finds a match for the id value?

[00:45:04.0000] <MikeSmith>
hsivonen: OK

[00:46:00.0000] <MikeSmith>
hsivonen: about your other question, I'm still looking

[00:47:00.0000] <MikeSmith>
hsivonen: so you saying that it's better not to use the locator mechanism at all for this label case?

[00:48:00.0000] <hsivonen>
MikeSmith: or you could also emit a warning that gives the location of the label

[00:48:01.0000] <hsivonen>
MikeSmith: unless that turns out to suck as UI. I'm not sure if giving the location of the label would be helpful or annoying from a user POV

[00:49:00.0000] <danbri>
anyone got a nice simple demo page, showing the use of the HTML5 SQL API?

[00:50:00.0000] <danbri>
/me hoping i can pre-populate a database with sqllite elsewhere...

[00:50:01.0000] <billyjackass>
hsivonen: do you think it would be useful to emit those warnings? or overkill?

[01:07:00.0000] <Lachy>
I wish people would at least try to make april fools jokes sound plausible http://tech.slashdot.org/article.pl?sid=09/03/31/1950221&from=rss

[01:11:00.0000] <MikeSmith>
damn, the quality of postings from Slashdot writers is otherwise so exceptional

[01:11:01.0000] <MikeSmith>
Lachy: you've burst my bubble of deep appreciation for Slashdot

[01:12:00.0000] <Dashiva>
Lachy: That sounds a lot like... http://www.lunascape.tv/

[01:12:01.0000] <Hixie>
/me goes to replace his sarcasm detector

[01:13:00.0000] <Dashiva>
/me checks the whatwg blog

[01:13:01.0000] <Hixie>
blimey, the aria spec is over 80 pages long

[01:14:00.0000] <hsivonen>
MikeSmith: I put in warnings like that for some table integrity checking stuff. I don't have user feedback complaining about those.

[01:14:01.0000] <MikeSmith>
hsivonen: OK, I will add it then

[01:15:00.0000] <Dashiva>
http://labs.opera.com/news/2009/04/01/

[01:15:01.0000] <Dashiva>
This would be awesome

[01:22:00.0000] <Hixie>
hsivonen: so... aria. what should i say in the spec?

[01:23:00.0000] <Hixie>
hsivonen: should i just defer to the wai-aria spec and say that the attributes it defines can be used in html5 for accessibility purposes?

[01:24:00.0000] <Hixie>
or should i list something explicitly? or what?

[01:37:00.0000] <hsivonen>
Hixie: I had expected the ongoing task force to figure out what makes sense

[01:38:00.0000] <hsivonen>
Hixie: on the topic of UA conformance:

[01:38:01.0000] <hsivonen>
defer to the ARIA implementation guide and a (AFAIK) so far non-existent document on resolving native semantic and ARIA semantic conflicts in client implementation

[01:40:00.0000] <hsivonen>
Hixie: on the topic of authoring conformance: defer states and properties authoring to ARIA and role applicability to given HTML elements to so far non-existent document hopefully flowing out of the work of the task force

[01:41:00.0000] <hsivonen>
Hixie: if you want to address FUD along the lines of HTML5 hating ARIA, you could already put in notes that defer to the parts of the ARIA spec family that exist

[01:44:00.0000] <hsivonen>
Hixie: so to elaborate

[01:45:00.0000] <hsivonen>
Hixie: the task force documents accessibility API mappings for old HTML and ARIA (and perhaps new HTML5 stuff)

[01:45:01.0000] <hsivonen>
Hixie: from those mappings, it's possible to see what HTML bits and ARIA bits overlap

[01:46:00.0000] <hsivonen>
Hixie: then it's necessary to define what UAs should do when they get mixed signals (HTML semantics and ARIA role set inconsistently)

[01:47:00.0000] <hsivonen>
Hixie: or what should happen if an element has a native accessibiilty API mapping and has states and properties as aria-* attributes with no role

[01:47:01.0000] <hsivonen>
Hixie: hopefully in due course this will live in a W3C document somewhere

[01:48:00.0000] <hsivonen>
Hixie: so reference that for UA conformance

[01:48:01.0000] <hsivonen>
Hixie: then, *given* the behaviors documented in that document, some element/role combinations will be non-sensical. I think those should be flaggable as non-conforming by validators.

[01:49:00.0000] <hsivonen>
Hixie: but defining those cases needs to wait until the UA conformance reqs show what combinations are non-sensical to author

[01:50:00.0000] <hsivonen>
Hixie: if it were up to me, I'd put a statement of intent to this direction as a note in the HTML 5 spec

[01:51:00.0000] <hsivonen>
which reminds me that I should review the latest draft of ARIA

[02:19:00.0000] <jgraham>
/me agrees ith pretty much everything hsivoen said, fwiw

[02:33:00.0000] <hsivonen>
/me finally starts dogfooding HTML5 parsing builds (not an April fools joke)

[02:36:00.0000] <jgraham>
/me wonders if hsivonen will append (not an aprils fools joke to all statements made today)

[02:36:01.0000] <jgraham>
s/(// s//(/

[02:36:02.0000] <jgraham>
actually I didn't need any brackets

[02:36:03.0000] <jgraham>
sigh

[02:37:00.0000] <hsivonen>
jgraham: hopefully only the ones that seem potentially unbelievable

[02:38:00.0000] <hsivonen>
Actually, as far as a parser goes, I've been ready for dogfood for a while now. I was waiting for an image cache crasher to go away

[02:38:01.0000] <takkaria>
hm, that's like xkcd's (no pun intended)

[02:45:00.0000] <jgraham>
takkaria: Really? How would you use it? Like I hear they're making an HTML 5 (not an april fools joke)

[02:54:00.0000] <zcorpan>
Hixie: "<li>There can only be one character encoding declaration in the document.</p>"

[03:25:00.0000] <hsivonen>
hmm. identi.ca looks wrong with the HTML5 parser

[03:25:01.0000] <hsivonen>
all the content is "below the fold"

[03:26:00.0000] <jgraham>
hsivonen: You got new try builds avaliable?

[03:27:00.0000] <hsivonen>
jgraham: yes. https://build.mozilla.org/tryserver-builds/2009-03-30_05:34-hsivonen⊙if/

[03:27:01.0000] <hsivonen>
jgraham: it doesn't have all the latest namespace fixes

[03:27:02.0000] <hsivonen>
XPath in text/html is probably broken in that build

[03:33:00.0000] <jgraham>
hsivonen: That build reliably crashes when visiting the live dom viewer

[03:37:00.0000] <zcorpan>
hsivonen: identi.ca doesn't have any interesting parse errors

[03:39:00.0000] <hsivonen>
jgraham: on which platform? WFM on Mac.

[03:39:01.0000] <hsivonen>
oops. not WFM

[03:40:00.0000] <hsivonen>
jgraham: crashes here, too. thanks.

[03:40:01.0000] <zcorpan>
i still wonder why the parser should imply <colgroup>

[03:40:02.0000] <hsivonen>
aargh. the image cache crash is still there

[03:40:03.0000] <zcorpan>
and why the content model requires colgroup around col

[03:41:00.0000] <hsivonen>
jgraham: do you also see imgCacheEntry stuff on the top of the crash stack?

[03:41:01.0000] <hsivonen>
--> Lunch

[03:42:00.0000] <jgraham>
zcorpan: Does col without colgroup make any sense?

[03:45:00.0000] <zcorpan>
jgraham: just as much as tr without tbody. lone col is valid in html4 and xhtml1

[03:58:00.0000] <MikeSmith>
hsivonen: http://bugzilla.validator.nu/attachment.cgi?id=71

[04:00:00.0000] <MikeSmith>
hsivonen: just tweaked to set boolean hasFor instead of collecting the "for" values

[04:19:00.0000] <zcorpan>
http://canvex.lazyilluminati.com/misc/dom-viewer/?%3C!DOCTYPE%20html%3E%3Cp%3E%3Cobject%3E%3Cp%3E%3C%2Fobject%3E%3C%2Fp%3E%3Cscript%3Edocument.write(document.getElementsByTagName('object').length)%3C%2Fscript%3E

[04:19:01.0000] <zcorpan>
/me is really confused

[04:19:02.0000] <zcorpan>
try that in ie8

[04:21:00.0000] <Lachy>
Hixie, the Live DOM Viewer clipboard is broken

[04:21:01.0000] <Lachy>
clipboard.cgi returns 404

[04:23:00.0000] <Lachy>
oh, no, it's not. That's Philip`s DOM viewer.

[04:24:00.0000] <Lachy>
I didn't realise zcorpan had linked to it instead

[04:24:01.0000] <jgraham>
Lachy: The differencde is somewhat important :)

[04:25:00.0000] <hsivonen>
MikeSmith: looks good. You could also break immediately after hasFor = true;

[04:25:01.0000] <MikeSmith>
hsivonen: OK

[05:16:00.0000] <Dashiva>
Darn, we've been shut down

[05:21:00.0000] <Lachy>
Dashiva, do you mean by Mr Last Week?

[05:24:00.0000] <hsivonen>
Hixie: have you ran the HTML5 parser on Google's index to see if quirks-mode pages would break if <p><table> always closed the paragraph?

[05:25:00.0000] <Dashiva>
Lachy: Yes

[05:30:00.0000] <Lachy>
Dashiva, just ignore MLW. He's just an anonymous coward and a troll.

[05:30:01.0000] <Philip`>
Be careful - insults will only make him stronger

[05:31:00.0000] <hsivonen>
does anyone else have data on whether it's feasible to adopt the standards mode behavior for <p><table> in the quirks mode, too?

[05:31:01.0000] <Lachy>
hsivonen, what sort of data would be useful?

[05:32:00.0000] <gsnedders>
WTF? This is weird, yet seemingly per-spec.

[05:32:01.0000] <hsivonen>
Lachy: a Web crawl showing that the case doesn't occur on the Web or doesn't occur on the CSS-enabled Web

[05:32:02.0000] <hsivonen>
Lachy: or data about Opera having tried dropping the quirk and having gotten bug reports forcing a revert

[05:33:00.0000] <Lachy>
it could theoretically affect both rendering and scripting if scripts or styles depend on the table being within the p

[05:33:01.0000] <hsivonen>
true

[05:33:02.0000] <jgraham>
gsnedders: What?

[05:34:00.0000] <hsivonen>
it would be sad if the single parsing quirk in HTML5 ended up being Hixie's own making (through Acid2)

[05:34:01.0000] <Lachy>
I don't think we've tried dropping the quirk before

[05:34:02.0000] <Lachy>
what did acid2 require?

[05:34:03.0000] <gsnedders>
jgraham: http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cp%20style%3D%22display%3Arun-in%22%3E%3Cspan%3Efoo%3C%2Fspan%3E%2C%20%3Cspan%3Ebar%0A%3Cp%3Efoo

[05:34:04.0000] <zcorpan>
that <p><table> close the p

[05:34:05.0000] <hsivonen>
Lachy: Acid2 requires <p><table> to parse like <p></p><table>

[05:35:00.0000] <Lachy>
oh, good

[05:35:01.0000] <hsivonen>
good???

[05:36:00.0000] <Lachy>
yes, having the table close the p is always how it's worked in good browsers in standards mode, and was what was requried by html4

[05:36:01.0000] <zcorpan>
hsivonen: btw, have you found pages breaking because of <ul><li><ul></li>?

[05:36:02.0000] <hsivonen>
"what was required by html4" is a weak argument

[05:37:00.0000] <hsivonen>
zcorpan: I haven't

[05:37:01.0000] <zcorpan>
might be worth to research that markup pattern since ie7 and all other browsers except ie8 and html5-enabled gecko don't close the inner ul

[05:38:00.0000] <Lachy>
true. but it makes the most sense cause it allows the </p> to remain optional even when followed by a table

[06:24:00.0000] <Philip`>
hsivonen: If there was a version of the HTML parser library that reported occurrences of that quirk, I could run it on my collection of pages

[06:26:00.0000] <Philip`>
though that wouldn't be too useful if it's less common than ~0.01%

[06:27:00.0000] <hsivonen>
Philip`: what kind of reporting mechanism do you need?

[06:29:00.0000] <Philip`>
hsivonen: The easiest would probably be something that prints to stdout in a thread-safe way

[06:29:01.0000] <annevk42>
I doubt that's the only quirk you'll need though it would be nice. Reparsing is another issue :/

[06:29:02.0000] <Philip`>
Actually I suppose that wouldn't quite work since I'd need to print the URI too

[06:30:00.0000] <hsivonen>
annevk42: is reparsing so important that it trumps security?

[06:30:01.0000] <hsivonen>
Philip`: would a warning with an easily detectable string work?

[06:30:02.0000] <annevk42>
hsivonen, ideally it's not, but the security concerns are rather weak

[06:33:00.0000] <Philip`>
hsivonen: So I would just use setErrorHandler? That sounds easy enough

[06:34:00.0000] <hsivonen>
Philip`: ok

[06:34:01.0000] <Philip`>
(Does it matter if I use DOM or SAX?)

[06:34:02.0000] <hsivonen>
Philip`: no

[06:39:00.0000] <hsivonen>
Philip`: svn head now does warn("A \u201Ctable\u201D start tag caused a paragraph to close implicitly."); when a table closes a para

[06:42:00.0000] <Philip`>
hsivonen: Is there a chance you could convert it into a .jar file? :-)

[06:42:01.0000] <Philip`>
/me doesn't have a version checked out from SVN

[06:42:02.0000] <hsivonen>
Philip`: ok

[06:44:00.0000] <hsivonen>
Philip`: http://hsivonen.iki.fi/htmlparser-philip-2009-04-01.jar

[06:45:00.0000] <Philip`>
hsivonen: Thanks!

[06:47:00.0000] <Philip`>
hsivonen: Hmm, I get lots of org.w3c.dom.DOMException: INVALID_CHARACTER_ERR: An invalid or illegal XML character is specified.

[06:47:01.0000] <Philip`>
(which I didn't get using 1.0.7)

[06:47:02.0000] <Philip`>
when I'm using HtmlDocumentBuilder(XmlViolationPolicy.ALTER_INFOSET)

[06:48:00.0000] <Philip`>
By "lots" I mean "one", plus many NAMESPACE_ERR: An attempt is made to create or change an object in a way which is incorrect with regard to namespaces.

[06:50:00.0000] <hsivonen>
Philip`: hmm. I don't know what I've broken. Do you have a URL that triggers those?

[06:51:00.0000] <Philip`>
http://www.giftology.co.uk gives the NAMESPACE_ERR

[06:51:01.0000] <Philip`>
http://www.villatraining.ca/ gives the INVALID_CHARACTER_ERR

[06:53:00.0000] <hsivonen>
netquotevar:=""

[06:55:00.0000] <Philip`>
It also gives NAMESPACE_ERR on e.g. http://www.autobanga.lt/ which doesn't seem to use anything interesting except xml:lang

[06:56:00.0000] <Philip`>
(unless the code changed in the past few months)

[06:56:01.0000] <hsivonen>
seems like something is badly wrong in the XML sanity code

[06:57:00.0000] <Philip`>
I hope it's not my fault

[06:58:00.0000] <Philip`>
Whoops, there's more messages sent to ErrorHandler than I expected

[06:58:01.0000] <Philip`>
My output file is 3,483,189,354 bytes

[06:59:00.0000] <Philip`>
I see 8161 URLs (out of ~130K) that trigger the table-closed-p warning

[06:59:01.0000] <hsivonen>
Philip`: thanks. that seems bad on the face of it

[07:01:00.0000] <Philip`>
/me wonders if there's any value in uploading the list

[07:02:00.0000] <hsivonen>
Philip`: I think there would be. one could pick sites at random and view them with two parsers

[07:02:01.0000] <gsnedders>
Anyone from Opera aroud?

[07:02:02.0000] <gsnedders>
*around

[07:03:00.0000] <MikeSmith>
gsnedders: do I count?

[07:03:01.0000] <Philip`>
hsivonen: http://philip.html5.org/data/table-implicitly-closed-p.txt

[07:04:00.0000] <annevk42>
no

[07:04:01.0000] <gsnedders>
MikeSmith: Probably enough

[07:04:02.0000] <gsnedders>
Opera wants a transcript of grades, which implies something formal and official. I don't have any digital transcript.

[07:04:03.0000] <MikeSmith>
gsnedders: I'm still part of the conspiracy, according to some people at least

[07:05:00.0000] <gsnedders>
Or am I half asleep?

[07:05:01.0000] <hsivonen>
Philip`: thanks!

[07:07:00.0000] <zcorpan>
gsnedders: i just sent an email when i applied at opera

[07:08:00.0000] <Philip`>
http://www.reddit.com/r/programming/comments/893uo/html5_removes_tag_soup_support_goes_xml_only/ - apparently the commenters like the non-XML syntax

[07:08:01.0000] <gsnedders>
zcorpan: http://www.opera.com/company/jobs/opening/211/ does say to apply online, though

[07:08:02.0000] <zcorpan>
gsnedders: so?

[07:08:03.0000] <gsnedders>
zcorpan: That is not email :P

[07:09:00.0000] <zcorpan>
doesn't mean email is rejected

[07:09:01.0000] <zcorpan>
i'm not saying you should send an email, though

[07:09:02.0000] <gsnedders>
:P

[07:10:00.0000] <annevk42>
waha, http://philip.html5.org/tools/web-apps-tracker?from=2942&to=2943 is funny

[07:10:01.0000] <annevk42>
did someone announce that on blog.whatwg.org ?

[07:10:02.0000] <annevk42>
especially if you follow the diff all the way to the end, lol

[07:10:03.0000] <gsnedders>
dt? dimitri?

[07:10:04.0000] <gsnedders>
Yeah, right

[07:10:05.0000] <Philip`>
http://twitter.com/WHATWG/status/1427965755

[07:11:00.0000] <Philip`>
Clearly not enough people here read the WHATWG Twitter feed

[07:11:01.0000] <annevk42>
ah sweet

[07:11:02.0000] <gsnedders>
I guess the entire digg-like style of reddit is April st too?

[07:11:03.0000] <gsnedders>
*1st

[07:12:00.0000] <jgraham>
gsnedders: If you can't get a trasnscript at the moment you should apply first and worry later

[07:14:00.0000] <gsnedders>
http://digg.com/programming/Tag_Soup_Support_Removed_From_HTML5

[07:16:00.0000] <Philip`>
It seems much more pointless when it's not blending into the Twitter stream and hiding behind tinyurl

[07:19:00.0000] <robinduckett>
yo

[07:20:00.0000] <Philip`>
Hello

[07:21:00.0000] <robinduckett>
anyone know if the ie8 beta devs have a channel kicking about?

[07:21:01.0000] <jgraham>
I think kicking the ie8 devs, even over irc, is harsh

[07:21:02.0000] <robinduckett>
hmmm

[07:21:03.0000] <robinduckett>
Really? 8)

[07:21:04.0000] <jgraham>
Although maybe they deserve it for the way they parse <p><object><p>

[07:22:00.0000] <gsnedders>
But what Travis said was that they were considering fixing that within IE8

[07:22:01.0000] <MikeSmith>
robinduckett: the IE8 devs are kept in a hermetically sealed environment

[07:22:02.0000] <gsnedders>
(implied they were

[07:22:03.0000] <MikeSmith>
they're not allowed to be exposed directly to open standards

[07:22:04.0000] <robinduckett>
hahah

[07:22:05.0000] <Philip`>
robinduckett: I don't think they do IRC; the closest thing is the occasional chats like http://blogs.msdn.com/ie/archive/2009/03/24/ie-team-chat-wednesday-firestarter-event-thursday.aspx

[07:23:00.0000] <robinduckett>
lol msdn, haven't been there since I was 11/12 doing VB6

[07:23:01.0000] <gsnedders>
Well, there are normally a few around on irc.w3.org

[07:26:00.0000] <Philip`>
robinduckett: It's not MSDN, it's just a domain which happens to be hosted on msdn.com perhaps for the reason mentioned in http://blogs.msdn.com/oldnewthing/archive/2009/03/05/9459157.aspx#9460253

[07:26:01.0000] <robinduckett>
lol

[07:26:02.0000] <robinduckett>
okay

[07:27:00.0000] <MikeSmith>
robinduckett: did you have something in particular you wanted to ask them about?

[07:27:01.0000] <MikeSmith>
because they will tell you everything if you just ask them

[07:27:02.0000] <MikeSmith>
all their product plans, that kind of stuff

[07:27:03.0000] <Philip`>
/me sees http://blogs.msdn.com/ie/archive/2009/03/30/compatibility-view-list-and-ie8-rtw.aspx , with links to a spreadsheet of the list, and a claim that "The Compatibility View List is updated on a regular cadence (in a period mirroring IE security updates, approximately every 2 months)"

[07:27:04.0000] <robinduckett>
just a very strange bug i've been getting on IE8, but it's really intermittant, so i'm going to grab the latest release and see if it still happens

[07:28:00.0000] <gsnedders>
Even the internals to the brokenness of object in IE < 8

[07:28:01.0000] <robinduckett>
lol

[07:28:02.0000] <MikeSmith>
robinduckett: what's the bug?

[07:29:00.0000] <robinduckett>
sometimes, clicking a select element to get the drop down, causes the pop up blocker to show up and the drop down to dissapear, and once it does that once it wont work again until the browser is restarted, and then sometimes it will either work for the rest of the day or not at all

[07:29:01.0000] <robinduckett>
I thought it was a bug in my code, I was doing a car make/model drop down in a form

[07:29:02.0000] <robinduckett>
and it had a lot of elements, so i thought it was related, but it does it on a two option select box now

[07:30:00.0000] <robinduckett>
sometimes it works sometimes it doesn't

[07:30:01.0000] <robinduckett>
very confusing

[07:30:02.0000] <robinduckett>
so i'll try the latest just incase i'm using some kind of old version

[07:30:03.0000] <gsnedders>
/me has bitchin' cold

[07:30:04.0000] <gsnedders>
http://secret.gsnedders.com/cv.html — comments?

[07:31:00.0000] <gsnedders>
(Yes, I do have secret as a subdomain)

[07:34:00.0000] <MikeSmith>
gsnedders: I think "Author of a number of automatable test cases for parts of the specification." is worthy of more prominence

[07:34:01.0000] <MikeSmith>
but not sure how to suggest doing that

[07:35:00.0000] <gsnedders>
<em>!

[07:35:01.0000] <zcorpan>
gsnedders: although it probably doesn't matter, i think there's a subtle difference between "Public invited expert" and "Invited Expert"

[07:36:00.0000] <MikeSmith>
if you want to get attention from opera, maybe putting "Can write good test cases all day long." under Skills would be good

[07:36:01.0000] <gsnedders>
/me just went for what it formally is called in the W3C process

[07:36:02.0000] <Philip`>
MikeSmith: Suggest <font color=red>

[07:37:00.0000] <gsnedders>
Philip`: But I'm not using a WYSIWYG editor!

[07:37:01.0000] <MikeSmith>
CSS fire shadow thing

[07:38:00.0000] <MikeSmith>
however that's done

[07:39:00.0000] <zcorpan>
gsnedders: also list any test suites you have written

[07:39:01.0000] <MikeSmith>
gsnedders: have you written any test harness (as opposed to just test cases?)

[07:40:00.0000] <Philip`>
gsnedders: Have you found (and written tests for and reported and got fixed) any bugs in browsers?

[07:40:01.0000] <zcorpan>
gsnedders: omg it doesn't validate

[07:40:02.0000] <jgraham>
gsnedders: Also, go to a school that is not named after a curry

[07:41:00.0000] <zcorpan>
wait why does v.nu say h1 is interactive?

[07:41:01.0000] <zcorpan>
hsivonen: ^

[07:42:00.0000] <hsivonen>
zcorpan: must be copypaste somewhere

[07:42:01.0000] <zcorpan>
Content model: Flow content, but with no heading content descendants, no sectioning content descendants, no footer element descendants, and no address element descendants.

[07:43:00.0000] <gsnedders>
Why does it disallow header content?

[07:43:01.0000] <gsnedders>
And why is there some insect buzzing around in here?

[07:44:00.0000] <gsnedders>
Hixie: ^^

[07:44:01.0000] <MikeSmith>
zcorpan:  v.nu say h1 is interactive because of a dumb mistake I made

[07:44:02.0000] <MikeSmith>
sorry

[07:44:03.0000] <gsnedders>
/me slaps MikeSmith 

[07:44:04.0000] <MikeSmith>
it's fixed now

[07:44:05.0000] <MikeSmith>
in the source

[07:44:06.0000] <zcorpan>
gsnedders: maybe because it doesn't make sense to have sections in <address>?

[07:45:00.0000] <jgraham>
/me gets scared by this kind of content-model wrangling

[07:45:01.0000] <gsnedders>
zcorpan: But it isn't a section within address, per the outlining algorithm

[07:45:02.0000] <gsnedders>
http://gsnedders.html5.org/outliner/process.py?url=http%3A%2F%2Fsecret.gsnedders.com%2Fcv.html

[07:46:00.0000] <jgraham>
gsnedders: Anolis has tests, right? And SimplePie?

[07:46:01.0000] <gsnedders>
Yeah

[07:46:02.0000] <jgraham>
gsnedders: You might want to mention that

[07:47:00.0000] <jgraham>
(Also remove the borders from you tables)

[07:47:01.0000] <zcorpan>
yes; i think your chances would be better if you removed everything and just pointed to all tests you've written :)

[07:47:02.0000] <zcorpan>
though i'm not suggesting you remove anything

[07:48:00.0000] <jgraham>
So you're trying to decrease his chances?

[07:48:01.0000] <zcorpan>
no

[07:49:00.0000] <Philip`>
everything < tests < everything+tests

[07:49:01.0000] <Philip`>
hence it's better to not remove everything, but to still add tests

[07:49:02.0000] <jgraham>
gsnedders: In particular if there is anything notable about the testing strategy that you adopted for those projects it would be good to incluse

[07:50:00.0000] <gsnedders>
Basically, I learnt a lot about what not to do from SimplePie :)

[07:50:01.0000] <jgraham>
At least, I ould think it was good, although my opinion is irrelevant

[07:50:02.0000] <Philip`>
If there isn't, you've got a good half a day to invent and implement a testing strategy so you've got something to write about

[07:50:03.0000] <gsnedders>
zcorpan: Why would you say that?

[07:50:04.0000] <zcorpan>
gsnedders: say what?

[07:51:00.0000] <gsnedders>
zcorpan: Remove everything and link to tests

[07:51:01.0000] <jgraham>
gsnedders: He didn't

[07:51:02.0000] <gsnedders>
jgraham: He typed it, though, and then didn't suggest it

[07:51:03.0000] <zcorpan>
gsnedders: i just mean that experience with test cases is more important than the other stuff

[07:52:00.0000] <jgraham>
It certianly seems plausible that is is more importat then standard grade classical studies

[07:52:01.0000] <jgraham>
Although only getting a 2 in Latin is pretty inexcusable ;)

[07:53:00.0000] <gsnedders>
:P

[07:53:01.0000] <Dashiva>
Latin was fun

[07:53:02.0000] <zcorpan>
http://steve-yegge.blogspot.com/2007/09/ten-tips-for-slightly-less-awful-resume.html

[07:54:00.0000] <Lachy>
LOL, check out what YouTube have done to the videos today :-) http://www.youtube.com/watch?v=ExkUOOXkGtg&flip=1

[07:55:00.0000] <gsnedders>
Oh fun

[07:56:00.0000] <jgraham>
/me notes that US-style resumes seem to be a much more elaborate affair than european CVs

[08:02:00.0000] <Lachy>
jgraham, what's the difference between them?

[08:03:00.0000] <annevk42>
how did they flip the text?

[08:04:00.0000] <Lachy>
sıɥʇ ǝʞıן ʇsnɾ

[08:04:01.0000] <jgraham>
Lachy: Not sure, but all the sites to do with US resume writing talk about putting in your personal vision and all this stuff (although SY talks about taking it yout, which is good). 

[08:04:02.0000] <Philip`>
http://s.ytimg.com/yt/js/april_fools-vfl84826.js

[08:04:03.0000] <gsnedders>
annevk42: There are Unicode codepoints for them

[08:04:04.0000] <jgraham>
Whereas that doesn't seem to be expected here so much

[08:04:05.0000] <Philip`>
It's just a translation table into glyphs that look similar

[08:05:00.0000] <annevk42>
/me just found that file

[08:08:00.0000] <Philip`>
/me sees gsnedders engaging in shameless self-promotion by posting his CV to public-html

[08:13:00.0000] <gsnedders>
Philip`: I didn't mean that. I just wanted to show a real use-case!

[08:15:00.0000] <Philip`>
Oh, sure

[08:15:01.0000] <gsnedders>
jgraham: But tables without borders are ugly!

[08:15:02.0000] <jgraham>
gsnedders: Stop providing useful feedback and write your CV

[08:15:03.0000] <jgraham>
gsnedders: You are wwrong

[08:15:04.0000] <jgraham>
:)

[08:15:05.0000] <takkaria>
don't write your CV, watch BBC News 24 instead, it's far more entertaining

[08:16:00.0000] <gsnedders>
What do they have on at the moment?

[08:17:00.0000] <jgraham>
gsnedders: The News

[08:17:01.0000] <takkaria>
just reporters saying things like "the protesters are calling it financial fool's day, which is a pun or play on wards on 'april fool's day'" and "protesters have been using internet websites to organise"

[08:18:00.0000] <gsnedders>
Not organize?

[08:19:00.0000] <gsnedders>
:P

[08:20:00.0000] <takkaria>
they have some fun protests being covered too

[08:20:01.0000] <gsnedders>
/me hates writing this CV

[08:23:00.0000] <annevk42>
I'm not sure how this case-insensitive is supposed to work for CSS

[08:24:00.0000] <zcorpan>
should createElement use a list too and use the right namespace based on tag name?

[08:24:01.0000] <annevk42>
that would only work half the time

[08:25:00.0000] <annevk42>
e.g. "a", "audio", "video"

[08:26:00.0000] <jgraham>
what case-insensitive?

[08:26:01.0000] <jgraham>
Oh, hsivonen's email?

[08:26:02.0000] <annevk42>
I suppose some CSS bits become case-insensitive when the style sheet is referenced from an HTML document.

[08:27:00.0000] <hsivonen>
annevk42: the CSS parser knows whether an ident is in the role of an element selector or in the role of an attribute selector

[08:27:01.0000] <hsivonen>
right?

[08:27:02.0000] <Lachy>
this is interesting. http://www.addfullsize.com/ I'm not sure it's something that could or would be supported by browsers though.

[08:27:03.0000] <jgraham>
Lachy: I thought that was strange and crazy

[08:28:00.0000] <hsivonen>
annevk42: so when the CSS parser has lower-cased such an indent, it would check if the resulting ident is in a lower-to-camel mapping 

[08:28:01.0000] <jgraham>
A whole website dedicated to advocating an attribute?

[08:28:02.0000] <jgraham>
I double checked for 1st-April related effects

[08:28:03.0000] <hsivonen>
why hasn't the site admin emailed public-html?

[08:28:04.0000] <Philip`>
http://krijnhoetmer.nl/irc-logs/whatwg/20081027#l-302

[08:28:05.0000] <annevk42>
hsivonen, and the matcing would be on localName alone?

[08:28:06.0000] <Lachy>
yeah, the fact that he made a whole website for it instead of mailing the list is crazy

[08:29:00.0000] <Philip`>
It's months old

[08:29:01.0000] <hsivonen>
annevk42: right. isn't it already?

[08:29:02.0000] <annevk42>
hsivonen, e.g. foreignObject in some weird namespace would still be matched by foreignobject { ... } when used in CSS referenced from HTML?

[08:29:03.0000] <takkaria>
it has been argued about on the whatwg list already, I'm pretty sure

[08:29:04.0000] <hsivonen>
annevk42: right

[08:29:05.0000] <hsivonen>
annevk42: which is why textArea can't be on the list

[08:29:06.0000] <annevk42>
hsivonen, I'm not sure how Selectors case-insensitivity works to be honest :)

[08:30:00.0000] <hsivonen>
annevk42: I asked bz to be sure :-)

[08:30:01.0000] <zcorpan>
i thought selectors matched case-insensitively

[08:31:00.0000] <annevk42>
zcorpan, impl or theory?

[08:31:01.0000] <zcorpan>
both

[08:31:02.0000] <takkaria>
/me suggests making XML case-insensitive

[08:31:03.0000] <zcorpan>
takkaria: how about optionally case-insensitive?

[08:32:00.0000] <zcorpan>
oh wait

[08:32:01.0000] <annevk42>
zcorpan, e.g. if you createElementNS("x", "TEst") in HTML CSS matches that with "test" in Gecko, WebKit and Presto?

[08:32:02.0000] <zcorpan>
annevk42: yeah, that's what i had thought, but obviously i was wrong and i hadn't tested

[08:32:03.0000] <annevk42>
hsivonen, ok, it seems somewhat suboptimal but it is more performant than doing it while matching

[08:34:00.0000] <Philip`>
http://www.addfullsize.com/ - "Since there is no fullsize attribute for the <IMG> tag yet, I am using the longdesc attribute instead." - hmm

[08:34:01.0000] <zcorpan>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/50 

[08:35:00.0000] <jgraham>
Philip`: And that is why markup will always "suck"

[08:35:01.0000] <annevk42>
zcorpan, Opera is cool

[08:35:02.0000] <zcorpan>
so opera matches case-insensitively and webkit and gecko lowercase

[08:37:00.0000] <annevk42>
so how does this work for HTML attribute values that are case-insensitive?

[08:37:01.0000] <zcorpan>
attribute values will have to match case-insensitively rather than lowercasing

[08:37:02.0000] <zcorpan>
since they preserve the case in the dom

[08:37:03.0000] <hsivonen>
zcorpan: attribute values are very different from names in Gecko

[08:38:00.0000] <annevk42>
and do they match case-insensitively regardless of what element they are on?

[08:38:01.0000] <annevk42>
I'm guessing they do

[08:38:02.0000] <zcorpan>
annevk42: in opera, only html elements, iirc

[08:38:03.0000] <hsivonen>
annevk42: IIRC, only on HTML elements in Gecko, but I'd have to check

[08:39:00.0000] <hsivonen>
(though this whole area is bad for "IIRC")

[08:39:01.0000] <annevk42>
"HTML elements" being a set?

[08:39:02.0000] <zcorpan>
http://lists.w3.org/Archives/Public/public-html/2008Jan/0043.html

[08:39:03.0000] <hsivonen>
annevk42: HTML elements being elements that have their HTMLness bit set

[08:40:00.0000] <hsivonen>
annevk42: I *think* the bit is consistent with namespace

[08:40:01.0000] <hsivonen>
annevk42: also, I *think* only HTMLElement instances have the bit set

[08:40:02.0000] <hsivonen>
actually, it's not a bit that is "set" but a mask that classes respond to

[08:41:00.0000] <hsivonen>
anyway

[08:42:00.0000] <annevk42>
I guess it works well in practice and worse in the theoretical case of having two element scoped attributes with the same name but different value space rules :)

[08:42:01.0000] <zcorpan>
hsivonen: does gecko now restrict the attribute value case-insensitivity to html elements?

[08:44:00.0000] <annevk42>
no

[08:44:01.0000] <annevk42>
http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cstyle%3Etest[align%3DteST]{border%3A1em%20solid}%3C%2Fstyle%3E%3Cbody%3E%3Cscript%3Ex%3Ddocument.createElementNS%28%22x%22%2C%20%22test%22%29%3Bx.setAttribute%28%22align%22%2C%22TEST%22%29%3Bdocument.body.appendChild%28x%29%3C%2Fscript%3E

[08:44:02.0000] <annevk42>
it does it regardless of the element as long as the attribute is in some list

[08:44:03.0000] <hsivonen>
clearly, my "IIRC" was wrong

[08:44:04.0000] <annevk42>
though maybe the situation is different in XHTML?

[08:45:00.0000] <hsivonen>
indeed, the list is in the CSS parser, not in the selector matching code

[08:45:01.0000] <annevk42>
(Opera doesn't match for that case by the way...)

[08:45:02.0000] <hsivonen>
FWIW, the list is http://mxr.mozilla.org/mozilla-central/source/layout/style/nsCSSParser.cpp#2916

[08:45:03.0000] <zcorpan>
annevk42: last i checked gecko didn't do any case-insensitive matching of attribute values in xhtml

[08:46:00.0000] <annevk42>
making Selectors case-insensitive for attribute values was a stupid mistake

[08:46:01.0000] <hsivonen>
annevk42: I think that applies more broadly :-)

[08:47:00.0000] <hsivonen>
(that case-insensitivity was a mistake)

[08:47:01.0000] <annevk42>
I sort of like we ended up with lowercase being the norm for HTML :)

[08:48:00.0000] <annevk42>
but yeah, it doesn't buy us much at all

[08:48:01.0000] <hsivonen>
uppercase being the norm without case-insensitivity would have sucked

[08:48:02.0000] <zcorpan>
clearly my reverse-engineering skills were not enough to catch all attributes :(

[08:49:00.0000] <zcorpan>
hsivonen: ismap should be in that list

[08:50:00.0000] <hsivonen>
zcorpan: thanks

[08:53:00.0000] <annevk42>
are we going to extend the list?

[08:53:01.0000] <annevk42>
hmm

[08:53:02.0000] <hsivonen>
I thought HTML5 was going to

[08:53:03.0000] <robinduckett>
anyone know of a commercial in-page wysiwyg editor?

[08:54:00.0000] <jgraham>
Does anyone know of a library (preferably usable from python) that will allow me to send window manager events to a particular application (in particular that will allow me to make the application fullscreen)?

[08:54:01.0000] <jgraham>
For linux obviously

[08:54:02.0000] <annevk42>
hsivonen, I was hoping we'd not enlarge magic lists

[08:55:00.0000] <zcorpan>
i was missing accept-charset, checked, direction and lang in my list

[08:55:01.0000] <annevk42>
there's also contenteditable, spellcheck, required, etc.

[08:55:02.0000] <zcorpan>
yes...

[08:56:00.0000] <zcorpan>
/me mails the list

[08:56:01.0000] <annevk42>
I rather have the list fixed for eternity

[08:57:00.0000] <zcorpan>
why?

[08:57:01.0000] <hsivonen>
annevk42: like the list of elements that implies <p> close? like <footer>?

[08:57:02.0000] <zcorpan>
and <table>? :)

[08:58:00.0000] <annevk42>
hsivonen, that's actually a useful feature

[09:00:00.0000] <annevk42>
this case-insensitive attribute value checking only costs

[09:03:00.0000] <zcorpan>
annevk42: i think browserse are only doing it for compliance with css21 and html4

[09:03:01.0000] <zcorpan>
annevk42: ie gets away with doing case-sensitive matching

[09:03:02.0000] <zcorpan>
maybe we should change css21

[09:03:03.0000] <annevk42>
IE doesn't do it?

[09:03:04.0000] <hsivonen>
zcorpan: IE8 even?

[09:03:05.0000] <annevk42>
cool

[09:03:06.0000] <annevk42>
lets kill it

[09:03:07.0000] <zcorpan>
haven't tested ie8

[09:04:00.0000] <zcorpan>
snap, they changed it

[09:05:00.0000] <hsivonen>
sigh. like <p><table>

[09:05:01.0000] <annevk42>
should still be doable if everyone is willing

[09:05:02.0000] <annevk42>
it's pretty isolated piece of code in most impl I assume

[09:05:03.0000] <zcorpan>
/me leaves the convincing work to annevk42

[09:06:00.0000] <annevk42>
mwaha

[09:08:00.0000] <annevk42>
where is this even defined in CSS?

[09:08:01.0000] <annevk42>
http://www.w3.org/TR/CSS21/selector.html only talks about element names

[09:10:00.0000] <zcorpan>
"The case-sensitivity of attribute names and values in selectors depends on the document language." -- http://www.w3.org/TR/CSS21/selector.html#matching-attrs

[09:10:01.0000] <annevk42>
meh

[09:11:00.0000] <annevk42>
for elements it spells it out but here it doesn't

[09:11:01.0000] <annevk42>
implementations are arguably wrong too

[09:11:02.0000] <annevk42>
there's nothing in HTML4 that says align on <x> is case-insensitive

[09:12:00.0000] <annevk42>
(the value of align)

[09:12:01.0000] <zcorpan>
right

[09:12:02.0000] <zcorpan>
but that's invalid and html4 leaves handling of invalid documents undefined :)

[09:12:03.0000] <zcorpan>
so it's not wrong

[09:14:00.0000] <hsivonen>
I wonder how much lawyer time MS used to develop the TOS for their compat view spreadsheet

[09:14:01.0000] <annevk42>
so should I take on this fight with www-style?

[09:15:00.0000] <annevk42>
note also that a literal reading of CSS 2.1 does not allow createElementNS("x","TEst") not to be matched by "test"

[09:16:00.0000] <hsivonen>
Interesting. Y! has requested their domains be removed from the IE8 blacklist

[09:16:01.0000] <annevk42>
the other thing is that the only person who would care enough to reply would be Bert Bos prolly claiming that a) the DOM is irrelevant and b) that it's not valid HTML

[09:16:02.0000] <hsivonen>
annevk42: CSS delegates the decision to HTML :-)

[09:17:00.0000] <hsivonen>
(but yeah, www-style would be the right list, still)

[09:21:00.0000] <Philip`>
hsivonen: Also interesting that only about 9 other domains have requested removal, out of over 3000

[09:22:00.0000] <Philip`>
Also interesting: When you sort data in OO Calc, and sort on non-unique fields, the ordering of equivalent rows seems to be random, rather than doing something sane like a stable sort

[09:23:00.0000] <Dashiva>
Go quicksort

[09:29:00.0000] <annevk42>
e-mailed

[09:30:00.0000] <annevk42>
hsivonen, if you can prepare a fix for Gecko I and zcorpan can prolly sort it out for Opera...

[09:30:01.0000] <annevk42>
it would make HTML and XHTML more consistent as well

[09:31:00.0000] <annevk42>
i should've mentioned that in my e-mail

[09:34:00.0000] <zcorpan>
annevk42: i've advocated attribute values be case-insensitive in xhtml

[09:35:00.0000] <zcorpan>
since they are defined to be in xhtml5

[09:35:01.0000] <zcorpan>
annevk42: you need to get the css3.info selectors testsuite changed, too

[09:37:00.0000] <zcorpan>
yeah html5 could say that no attribute values are case-insensitive as far as selectors are concerned without changing css

[09:39:00.0000] <zcorpan>
annevk42: ie6 doesn't support attribute selectors :)

[09:41:00.0000] <annevk42>
zcorpan, the way this is implemented now would mean that all align attributes would have their values case-insensitively matched in XHTML...

[09:41:01.0000] <annevk42>
zcorpan, in XML even, that seems highly suboptimal

[09:42:00.0000] <annevk42>
zcorpan, yeah, that test suite would need a fix

[09:42:01.0000] <zcorpan>
annevk42: opera checks html namespace

[09:42:02.0000] <zcorpan>
annevk42: my proposal was to align with opera

[09:42:03.0000] <annevk42>
zcorpan, make this mess even more complicated? :)

[09:43:00.0000] <zcorpan>
basically, though we have already implemented the check

[09:43:01.0000] <zcorpan>
but i don't mind removing it altogether

[09:44:00.0000] <Philip`>
http://test.foaf-ssl.net/cert/make?Select=That%27s+ME&webid=http%3A%2F%2Fwww.advogato.org%2Fperson%2Fconnolly%2Ffoaf.rdf%23me

[09:45:00.0000] <Philip`>
They use <keygen>

[09:48:00.0000] <Philip`>
http://canvex.lazyilluminati.com/misc/cgi/issues.cgi/message/%3CC3CC87D2-2781-4F23-953B-FBD48639E556%40bblfish.net%3E - oh, someone pointed it out already

[09:49:00.0000] <Philip`>
(Well, not quite that exact implementation)

[09:57:00.0000] <mpt>
I'm surprised and disappointed that http://saveie6.com/ hasn't been mentioned here yet

[10:06:00.0000] <robinduckett>
mpt: ooh a new site to ddos

[10:39:00.0000] <gsnedders>
/me stretches

[10:50:00.0000] <gsnedders>
Anyone still around?

[10:50:01.0000] <Philip`>
No

[10:51:00.0000] <gsnedders>
Oh well

[10:51:01.0000] <gsnedders>
Tragic

[10:53:00.0000] <robinduckett>
fscking

[10:53:01.0000] <robinduckett>
IE8

[10:53:02.0000] <robinduckett>
sucks.

[10:54:00.0000] <jgraham>
gsnedders: Everyone's dead

[10:54:01.0000] <Philip`>
robinduckett: For any specific reasons?

[10:56:00.0000] <gsnedders>
"Mama, just killed a man, Put a gun against his head/Pulled my trigger, now he's dead/Mama, life had just begun/But now I've gone and thrown it all away"

[10:58:00.0000] <gsnedders>
Does <http://secret.gsnedders.com/cv.html> seem better yet?

[10:59:00.0000] <jgraham>
gsnedders: The table still has borders :)

[11:00:00.0000] <jgraham>
gsnedders: Have to go but generally looks OK

[11:00:01.0000] <gsnedders>
/me remembers to upload his changes

[11:00:02.0000] <gsnedders>
Now it doesn't!

[11:00:03.0000] <Philip`>
/me can quote random lyrics too

[11:00:04.0000] <Philip`>
"War, it's never been so much fun/War, it's never been so much fun/Go to your brother/Kill him with your gun/Leave him lying in his uniform/Dying in the sun!/War, it's never been so much fun/War, it's never been so much fun"

[11:06:00.0000] <gsnedders>
/me has his hair get in his way again, and remembers why he has for the past months always had it tied back

[11:32:00.0000] <tantek>
gsnedders, I took a brief look at your cv.html - well done with using hResume

[11:32:01.0000] <tantek>
perhaps consider adding it to http://microformats.org/wiki/hresume-examples-in-wild

[11:35:00.0000] <annevk42>
http://wrongtomorrow.com/ is great

[11:37:00.0000] <annevk42>
"Research has shown that experts make predictions at a rate worse than chance. This site exists in order to hold people and media outlets accountable for pretending to see into an unpredictable future."

[11:39:00.0000] <benh>
awesome

[11:45:00.0000] <Hixie>
hsivonen: long term, once aria is mature, what should html5 say? is it going to be the place that needs to define which cases are conflicts?

[11:50:00.0000] <annevk42>
Hixie, I guess it needs to define what is conforming

[11:55:00.0000] <gsnedders>
Hixie: You got any thoughts on address and headings, per my email?

[12:02:00.0000] <Hixie>
gsnedders: i haven't even read e-mail yet

[12:03:00.0000] <Hixie>
annevk42: any idea what it should say? i don't fancy going through the whole 80+ pages of the aria spec trying to work out what the right thing to say is...

[12:10:00.0000] <annevk42>
no sorry

[12:10:01.0000] <annevk42>
I think hsivonen has an idea, but we probably have to wait for the specifics of the implementation guidelines for HTML etc.

[12:18:00.0000] <benh>
Hixie: listing mutually incompatible roles and elements or attributes and elements might be an approach.

[12:18:01.0000] <benh>
e.g. input type="checkbox" role="button"

[12:19:00.0000] <annevk42>
type=checkbox role=checkbox also seems like an error to me

[12:20:00.0000] <benh>
annevk42: In that particular case, yes. In other similar cases, perhaps not.

[12:20:01.0000] <benh>
e.g. nav role="navigation" is probably not an author error.

[12:20:02.0000] <annevk42>
I'd say it is

[12:21:00.0000] <benh>
if ARIA roles and new elements are not always implemented side-by-side then authors are going to compensate by doubling them up

[12:21:01.0000] <benh>
so it's not helpful to authors to flag that as an error.

[12:21:02.0000] <gsnedders>
Hixie: I'd kinda like to know whether my invalid CV will become conforming before submitting it :)

[12:21:03.0000] <benh>
(maybe a warning?)

[12:22:00.0000] <annevk42>
it's not clear that ARIA roles and HTML5 elements mean identical things to screen readers so doubling them up should prolly be discouraged

[12:22:01.0000] <annevk42>
but it's all a bit sketchy

[12:23:00.0000] <benh>
I'm pretty sure I'd use role="navigation" on nav.

[12:23:01.0000] <benh>
I think a warning when the semantics partially overlap might be useful.

[12:23:02.0000] <benh>
but that's a human judgement call type thing

[12:23:03.0000] <benh>
not a valid/invalid boolean

[12:23:04.0000] <ojan>
sorry about the webkit failures

[12:24:00.0000] <annevk42>
ojan, wrong channel?

[12:24:01.0000] <ojan>
heh...oops

[12:24:02.0000] <ojan>
sorry 

[12:24:03.0000] <Philip`>
gsnedders: Just put a comment in your markup pointing to your email :-)

[12:25:00.0000] <Philip`>
Then it won't matter if it fails the validator because you've got an excuse, and also you've got a practical demonstration of your ability to contribute to standards-related activities

[12:28:00.0000] <annevk42>
http://www.w3.org/TR/wai-aria-implementation/ has a start on impl mapping btw

[12:31:00.0000] <annevk42>
so one of the comments I should make is that they cannot use the Nmtokens production to define whitespace separated because HTML and the DOM do not have whitespace normalization like XML

[12:31:01.0000] <benh>
/me wonders how current ARIA implementations handle input type="checkbox" role="button"

[12:31:02.0000] <annevk42>
(comments on ARIA)

[12:33:00.0000] <annevk42>
7.2 is redundant with the DOM spec

[12:35:00.0000] <annevk42>
7.3 has a confusing requirement about AT modifying the DOM and the Web application having to take those changes into account

[12:35:01.0000] <annevk42>
I thought it was supposed to work the other way around?

[12:35:02.0000] <benh>
7.3? doesn't http://www.w3.org/TR/wai-aria-implementation/ end with 7.2.3 ?

[12:36:00.0000] <annevk42>
I was asked to review http://www.w3.org/TR/wai-aria/

[12:36:01.0000] <benh>
oh sorry

[12:37:00.0000] <annevk42>
I can see how it was confusing :)

[12:38:00.0000] <annevk42>
and it all kinds of RDF stuff in there, who's going to use that?

[12:41:00.0000] <benh>
If they define a JSON representation of same, RDF enthusiasts I guess.

[12:41:01.0000] <annevk42>
JSON?

[12:42:00.0000] <benh>
annevk42: https://wiki.mozilla.org/Accessibility/JSON_ARIA

[12:42:01.0000] <annevk42>
and what's this crazy modelling about properties inheriting etc.

[12:42:02.0000] <annevk42>
geez, it's just a bunch of attributes to provide a low-level access api

[12:43:00.0000] <annevk42>
benh, oh yeah, I saw that before

[12:43:01.0000] <benh>
probably so you can specify complex structures to user agents in the same way as the spec specifies simpler ones.

[12:43:02.0000] <annevk42>
ARIA is already so complex and they're adding yet another layer of abstraction for the fun of it

[12:43:03.0000] <benh>
it's got a certain conceptual purity

[12:43:04.0000] <annevk42>
I'm sure that's going to work well

[12:44:00.0000] <annevk42>
IBM can prolly line up a few consultants a EUR 500 an hour that can implement it in your system harharhar

[12:45:00.0000] <benh>
an interesting question would be - can the implementation details of the specified complex structures be derived from the implementation details of the specified simple structures.

[12:45:01.0000] <benh>
if not, then the inheritance would seem to be window-dressing.

[12:45:02.0000] <benh>
well, for some value of "interesting"

[12:45:03.0000] <annevk42>
at that point you probably better have some turing-complete thing hooking in the AT but if a) ATs are happy with that and b) anyone is going to bother is another

[12:48:00.0000] <benh>
I can imagine using it to specify the name of a specialized widget in a page/app that people reuse.

[12:49:00.0000] <benh>
maybe

[12:49:01.0000] <benh>
the most immediate use seems to be building calendar roles.

[12:49:02.0000] <annevk42>
I'd rather we just fix the damn fomr control styling problem

[12:49:03.0000] <annevk42>
and get people to implement the new form controls

[12:50:00.0000] <annevk42>
having a somewhat low-level access api can still be useful, but what people are advocating it for are just hacks

[12:51:00.0000] <annevk42>
the main reason they did ARIA was to work around that problem and lack of updates from Microsoft

[12:52:00.0000] <annevk42>
(i.e. if that problem was solved it still wouldn't work in IE)

[12:53:00.0000] <benh>
well, you'd still want stuff like live regions

[12:55:00.0000] <annevk42>
live regions is a gap in HTML

[12:55:01.0000] <annevk42>
indeed

[12:56:00.0000] <annevk42>
a lot of ARIA is just gaps in HTML4. the problem is that the solution they designed to this problem was only solving it for the AT case and not for everyone (per the above constraints)

[12:57:00.0000] <annevk42>
HTML5 is part of the everyone approach, but we still haven't dealt with styling (pretty back track record)

[12:57:01.0000] <annevk42>
s/back/bad/

[12:59:00.0000] <benh>
/me doubts styling is going to be dealt with for years.

[13:00:00.0000] <benh>
enough to satisfy people using SVG/Canvas/images

[13:00:01.0000] <benh>
perhaps never.

[13:03:00.0000] <annevk42>
WebKit has a pretty decent appearance implementation

[13:04:00.0000] <annevk42>
you can turn of default form control rendering completely with that

[13:04:01.0000] <annevk42>
s/of/off/

[13:04:02.0000] <annevk42>
now if only it was defined a bit more clearly others could copy that

[13:05:00.0000] <annevk42>
we'd still need XBL for the more advanced controls of course

[13:05:01.0000] <benh>
indeed

[13:07:00.0000] <annevk42>
anyway, a solution to this problem doesn't seem that improbable

[14:15:00.0000] <annevk42>
Hixie "character (') U+0022 QUOTATION MARK character (")" misses an "or an" before U+0022

[14:16:00.0000] <annevk42>
Hixie, "critiera"

[14:35:00.0000] <jgraham>
Why does he aria implementation guide thing say that it is going to become a note. It seems like the most crucial part of the whole enterprise

[14:40:00.0000] <jgraham>
"

[14:40:01.0000] <jgraham>
For the core accessibility API properties of role, name, states, value, etc. there is typically only one of each of them. If there is a conflict, ARIA always wins, because ARIA is essentially an override. In other words, if the native markup says there a link, but the ARIA markup says it is a button, then it should be exposed as a button."

[14:40:02.0000] <Hixie>
annevk42: e-mail please :-)

[14:40:03.0000] <Hixie>
annevk42: or a bug

[14:58:00.0000] <gsnedders>
/me doesn't know what to write in a covering letter

[14:58:01.0000] <gsnedders>
/me sighs

[15:04:00.0000] <annevk42>
Hixie, haven't had a reply yet to the last one, but ok

[15:05:00.0000] <Hixie>
can you only send one e-mail at once? :-)

[15:05:01.0000] <Hixie>
i can't track IRC comments. if you comment on IRC, it'll be lost.

[15:06:00.0000] <gsnedders>
/me thinks we need an IRC -> email bot

[15:06:01.0000] <Hixie>
be my guest :-)

[15:06:02.0000] <gsnedders>
I'm trying to apply for an internship whose deadline is April 1st!

[15:07:00.0000] <Hixie>
i don't necessarily mean right now :-)

[15:07:01.0000] <gsnedders>
/me still doesn't know what to write in a covering letter

[15:07:02.0000] <gsnedders>
"Hi, I rule, hire me, kthxbai."

[15:07:03.0000] <olliej>
gsnedders: why wouldn't that work? :D

[15:08:00.0000] <gsnedders>
/me just shakes head

[15:08:01.0000] <gsnedders>
/me wonders how old olliej is anyway

[15:09:00.0000] <Dashiva>
Just write that you invented Forth

[15:11:00.0000] <gsnedders>
I live near the Forth, but I failed to invent it.

[15:11:01.0000] <annevk42>
just write you fancy the Web, browsers and Web standard games, have some experience with testing and would like to learn more

[15:12:00.0000] <gsnedders>
/me copies that

[15:13:00.0000] <gsnedders>
"Send us your CV and cover letter and indicate the preferred dates for your internship.

[15:13:01.0000] <gsnedders>
Please specify technical skills, relevant experience and what type of technical challenges you look for."

[15:15:00.0000] <annevk42>
sounds about right; admittedly I never applied for a proper job nor do I have a CV

[15:15:01.0000] <gsnedders>
heh

[15:15:02.0000] <gsnedders>
/me puts on some takkaria music

[15:18:00.0000] <gsnedders>
I guess I'll have to address the point of, "Enrollment in a 3-5 year university program, or completing such a program in 2009."

[15:39:00.0000] <annevk42>
gsnedders, why does your name need to be part of the address?

[15:39:01.0000] <gsnedders>
annevk42: Because that's what hCard says so, because that's what vCard says :P

[15:40:00.0000] <gsnedders>
s/so//

[15:40:01.0000] <annevk42>
it looks broken

[15:43:00.0000] <gsnedders>
How do you finish a covering letter?

[16:04:00.0000] <gsnedders>
So, I click the button and then it gets stuck "Contacting 'www.opera.com'"

[16:04:01.0000] <gsnedders>
Service Temporarily Unavailable

[16:04:02.0000] <gsnedders>
The server is temporarily unable to service your request due to maintenance downtime or capacity problems. Please try again later.

[16:04:03.0000] <gsnedders>
Additionally, a 503 Service Temporarily Unavailable error was encountered while trying to use an ErrorDocument to handle the request.

[16:04:04.0000] <gsnedders>
Oh fun

[16:05:00.0000] <Hixie>
if we did want to do a review process for the whatwg, i guess we could use a google calendar to suggest areas to review each week

[16:07:00.0000] <gsnedders>
Hmmm…

[16:07:01.0000] <gsnedders>
This rules. I'm going to miss the application deadline because the form doesn't work :P

[16:09:00.0000] <Hixie>
is there no e-mail alternative?

[16:09:01.0000] <gsnedders>
Nope

[16:19:00.0000] <annevk42>
oh that sucks

[16:23:00.0000] <gsnedders>
/me has it come back to life

[16:23:01.0000] <gsnedders>
/me reads RFC 5514

[16:28:00.0000] <jcranmer>
gsnedders: seesh, you're so slow

[16:28:01.0000] <jcranmer>
you're only reading it now?

[16:40:00.0000] <Hixie>
annevk42, hsivonen: looks like what you said about aria conflicts with aria

[16:40:01.0000] <Hixie>
specifically, the wai-aria doc says that a host language must support all roles on all elements

[16:43:00.0000] <Hixie>
abarth rocks http://www.w3.org/mid/7789133a0904010118g437592e0vb61b67e2dd18feb2⊙mgc

[16:59:00.0000] <gsnedders>
jcranmer: Yeah :(


2009-04-02
[17:46:00.0000] <Hixie>
ok I'm going to create a Google Calendar, and it'll list sections for review each week

[17:46:01.0000] <Hixie>
anyone interested in helping set this up?

[18:04:00.0000] <Hixie>
hm, this doesn't seem like a good way of doing this

[18:05:00.0000] <Hixie>
/me ponders

[18:08:00.0000] <MikeSmith>
Hixie: I might otherwise offer to help but I'm the middle of (re)learning Java so that I can try to write useful patches for v.nu conformance-checking backend without making my lack of coding skills obvious

[18:22:00.0000] <Hixie>
MikeSmith: heh

[18:24:00.0000] <Hixie>
maybe the better solution is to just get people to see if the spec satisfies their pet peeve rather than trying to get people to review the spec section by section

[18:24:01.0000] <Hixie>
since the latter really isn't something most people have the skillset to do sanely

[18:28:00.0000] <Philip`>
There are still some people who would be willing and able to review many sections, and those people's interests would overlap to some extent, and so it would be good to prioritise them to maximise coverage

[18:28:01.0000] <Philip`>
s/prioritise/organise/

[18:30:00.0000] <othermaciej>
I can probably do a lot of useful review but I have no idea how to prioritize it in a secion-based way

[18:31:00.0000] <MikeSmith>
hmm, Glazman resigning from CSS WG

[18:31:01.0000] <MikeSmith>
http://www.glazman.org/weblog/dotclear/index.php?post/2009/04/01/Moving-on

[18:32:00.0000] <MikeSmith>
oh

[18:32:01.0000] <MikeSmith>
April Fools

[18:32:02.0000] <MikeSmith>
?

[18:32:03.0000] <MikeSmith>
maybe I should actually read the posting

[18:33:00.0000] <MikeSmith>
ah, got me

[18:37:00.0000] <MikeSmith>
Hixie: maybe Sierk Bornemann's comments in "Disregard of RFC 4329 and IANA MIME Media Types" are an April Fools joke

[18:44:00.0000] <Hixie>
MikeSmith: it's an old bug, i doubt it

[18:44:01.0000] <Hixie>
Philip`: any suggestions on how to do that?

[18:46:00.0000] <MikeSmith>
Hixie: I think it planned it well in advance. He seems like a really clever guy. I like especially the part where he says, "the browser teams of Mozilla, KDE Konqueror, Safari and Opera... did not have such problems like you in accepting and implementing these ... They simply did it. Without questioning the decision of IANA and IETF."

[18:47:00.0000] <MikeSmith>
s/think it planned/think he planned/

[18:47:01.0000] <Hixie>
i wish i could think you were right

[19:22:00.0000] <Philip`>
Hixie: No

[19:31:00.0000] <Philip`>
Hixie: "the MIME type used to refer to JavaScript in this specification is text/javascript, since that is the most commonly used type." - most commonly used where? http://canvex.lazyilluminati.com/misc/stats/scripttypes2.html says the HTTP Content-Type 1.5 years ago was almost always "application/x-javascript" instead

[19:31:01.0000] <Philip`>
Hixie: Also: s/JavsScript/JavaScript/

[19:37:00.0000] <Hixie>
hm, most commonly used isn't what i meant

[19:37:01.0000] <Hixie>
most recognisable, maybe?

[19:38:00.0000] <Hixie>
it's most commonly used in <script type="">, i think

[19:38:01.0000] <Hixie>
but i haven't checked recently

[19:39:00.0000] <Philip`>
http://canvex.lazyilluminati.com/misc/stats/scripts2.html - it was the most common <script type> by far

[19:43:00.0000] <Hixie>
ok

[19:43:01.0000] <Hixie>
then i'll claim that's what i meant :-P

[19:43:02.0000] <Hixie>
after all, i don't think i talk about the HTTP type anywhere

[20:22:00.0000] <Hixie>
Philip`: yt?

[20:31:00.0000] <Hixie>
if anyone has any comments on http://wiki.whatwg.org/wiki/Reviewing_HTML5 let me know, otherwise i'll post it to the list and the blog tomorrow

[21:21:00.0000] <Hixie>
/me sends aria last call comments

[21:36:00.0000] <MikeSmith>
Hixie: so is it not your plan for the HTML5 spec itself to define which ARIA attributes are conformant on particular elements?

[22:33:00.0000] <Hixie>
MikeSmith: as far as i can tell, the ARIA spec disallows that. I sent them e-mail asking if they could change that.

[22:34:00.0000] <MikeSmith>
Hixie: I see. I wasn't aware that the ARIA spec explicitly disallowed it.

[22:35:00.0000] <MikeSmith>
Hixie: I'm wondering how useful it will be in the long run to put ARIA-attribute checking into conformance checkers

[22:35:01.0000] <MikeSmith>
I mean, isn't the real use-case for ARIA that the attributes get added to the DOM dynamically?

[22:36:00.0000] <MikeSmith>
instead of in markup

[22:44:00.0000] <Hixie>
yeah

[22:45:00.0000] <Hixie>
i'm hoping we will get dynamic conformance checkers on the long run

[22:45:01.0000] <Hixie>
not holding my breath though

[23:33:00.0000] <sayrer>
hmm, these what blog posts have become increasingly snide

[23:34:00.0000] <sayrer>
whatwg blog posts

[23:34:01.0000] <sayrer>
I would like to publish a whatwg blog post about that

[23:36:00.0000] <Hixie>
go ahead :-)

[23:36:01.0000] <Hixie>
the blog is open to anyone

[23:36:02.0000] <sayrer>
that's what I thought

[23:36:03.0000] <sayrer>
how do I do it?

[23:37:00.0000] <Hixie>
if you register for an account and log in, you can write the post

[23:37:01.0000] <Hixie>
then ask lachy to submit it

[23:37:02.0000] <sayrer>
alright, let's see

[23:38:00.0000] <Hixie>
keep things nice and happy though :-)

[23:38:01.0000] <Hixie>
/me is off to play his game console

[23:38:02.0000] <Hixie>
bbl

[23:38:03.0000] <sayrer>
I think markp is failing at nice and happy, fwiw

[00:17:00.0000] <hsivonen>
/me notes that markp's copy and paste from IRC logs lower cases content

[00:20:00.0000] <hsivonen>
Hixie: as benh says, we'll probably be stuck with <nav role=navigation> indefinitely. so that probably needs to be made conforming even for the long term accompanied by IRC, blog and twitter whining how it sucks that the W3C has a separate WG patching HTML only from an accessibility point of view

[00:21:00.0000] <hsivonen>
Hixie: (and whining how implementors opted to implement role=navigation first instead of exposing <nav> right away)

[00:24:00.0000] <hsivonen>
annevk42: I think in general, ARIA should just stay away from XSD. (nmtoken or otherwise)

[00:26:00.0000] <hsivonen>
annevk42: regarding ARIA working around MS but being supported in IE8 nonetheless: do you believe that all top four would have some support for ARIA if ARIA weren't designed in such a way that no vendor can block it?

[01:12:00.0000] <Hixie>
sayrer: how so?

[01:12:01.0000] <sayrer>
Hixie, just snide stuff like

[01:13:00.0000] <sayrer>
regarding the meta charset thing

[01:13:01.0000] <sayrer>
"This is mainly to address the desire of a few overly vocal authors to be able to serve the same markup in both text/html and application/xhtml+xml modes."

[01:13:02.0000] <sayrer>
or

[01:13:03.0000] <Hixie>
is that snide? isn't it precisely accurate?

[01:14:00.0000] <sayrer>
do you add things because people are overly vocal?

[01:14:01.0000] <sayrer>
I thought not

[01:14:02.0000] <Hixie>
in that particular case, yes, it was added because sam complained loudly

[01:14:03.0000] <Hixie>
i probably wouldn't have added it if he hadn't asked for it

[01:15:00.0000] <Hixie>
because i think serving polyglot documents is silly

[01:15:01.0000] <sayrer>
did he make a bad argument?

[01:15:02.0000] <Hixie>
but he had some use cases for it

[01:15:03.0000] <Hixie>
so in it goes

[01:15:04.0000] <Hixie>
no, what makes you ask that?

[01:15:05.0000] <sayrer>
so, it seems snide to require people to make their case, and then call them "overly vocal" for doing so

[01:16:00.0000] <sayrer>
and then there's

[01:16:01.0000] <sayrer>
"Those who complained that HTML 5 was 'too bloated' will have a little less to complain about now that several parts of it have been published as separate documents. On the other hand, critics who complained about these things as a cover for other agendas will have to continue complaining a little while longer."

[01:16:02.0000] <Hixie>
again, that just seems accurate

[01:16:03.0000] <Hixie>
how is that snide?

[01:17:00.0000] <Hixie>
i think you're reading far more into mark's humorous comments than is intended

[01:17:01.0000] <sayrer>
who has a hidden agenda?

[01:17:02.0000] <Hixie>
the only way that last comment could be offensive is if someone had a hidden agenda, which no-one does as far as i know

[01:17:03.0000] <sayrer>
but the comment seems to claim they exist

[01:18:00.0000] <sayrer>
not very productive

[01:18:01.0000] <sayrer>
I wouldn't call it "precisely accurate" at all

[01:18:02.0000] <Hixie>
given how productive mark is and how useful his posts are, i think it's reasonable to let him inject some of his humour in there

[01:18:03.0000] <Hixie>
in what way is it not accurate?

[01:19:00.0000] <sayrer>
I think it's reasonable to let him write whatever he wants

[01:20:00.0000] <sayrer>
you just said no one has a hidden agenda, but Mark wrote

[01:20:01.0000] <sayrer>
"critics who complained about these things as a cover for other agendas"

[01:21:00.0000] <sayrer>
I would call that "haha only serious"

[01:21:01.0000] <Hixie>
why, are there people with hidden agendas?!

[01:21:02.0000] <sayrer>
oh probably

[01:21:03.0000] <Hixie>
o_O

[01:21:04.0000] <Hixie>
who?

[01:22:00.0000] <sayrer>
I dunno, the problem with the comment is that it places some people above reproach

[01:22:01.0000] <Hixie>
how ever so?

[01:23:00.0000] <Hixie>
are you seriously suggesting that mark was suggesting that there really are people with hidden agendas??

[01:23:01.0000] <sayrer>
yes

[01:23:02.0000] <Hixie>
ok in that case allow me to disillusion you

[01:24:00.0000] <Hixie>
mark may have a sense of humour that some would call "british"

[01:24:01.0000] <sayrer>
oh, I understand that kind of humour

[01:24:02.0000] <Hixie>
but i assure you that he's not actually suggesting there are people acting in bad faith

[01:24:03.0000] <sayrer>
Agree to disagree

[01:24:04.0000] <Hixie>
he likes needling people (just see his blog)

[01:24:05.0000] <Hixie>
apparently he needled you this time :-)

[01:26:00.0000] <sayrer>
no, I just think he's smug, and I don't mind saying so

[01:26:01.0000] <sayrer>
it's not happy and nice

[01:26:02.0000] <Hixie>
mark is one of the most self-effacing guys out there

[01:27:00.0000] <sayrer>
ok

[01:33:00.0000] <MikeSmith>
+1 to more jackassery on the whatwg blog

[01:33:01.0000] <sayrer>
http://twitter.com/diveintomark/status/1344958151

[01:34:00.0000] <MikeSmith>
the whatwg blog should ideally reflect the diversity of opinion on #whatwg, and the color/character of #whatwg discussions

[01:34:01.0000] <sayrer>
what a self-effacing guy

[01:35:00.0000] <Hixie>
are you sure you know what "smug" and "self-effacing" mean? that twitter had nothing at all to do with mark, how could it possibly be either?

[01:35:01.0000] <Hixie>
i'm very confused about your attitude here

[01:35:02.0000] <sayrer>
yes, I know what both of those mean. do you?

[01:36:00.0000] <Hixie>
yes, and both require one to say something about oneself

[01:36:01.0000] <Hixie>
that twitter said nothing about mark

[01:36:02.0000] <sayrer>
that isn't true

[01:37:00.0000] <Hixie>
smug: Exhibiting or feeling great or offensive satisfaction with *oneself*

[01:37:01.0000] <Hixie>
self-effacing: Not drawing attention to *oneself*

[01:37:02.0000] <sayrer>
what do you need to do to pass judgement on others, in public?

[01:38:00.0000] <Hixie>
who is passing judgment on something?

[01:38:01.0000] <sayrer>
especially regarding their moral fiber or motivations

[01:38:02.0000] <Hixie>
sam himself has acknowledged that his history document missed numerous key points of the history of html, how is mark's statement again anything but factual?

[01:39:00.0000] <sayrer>
would you say it is "unencumbered as it was by facts, context, or perspective of any kind" ?

[01:39:01.0000] <Hixie>
yes, the whole document was woefully inaccurate and wildly devoid of context

[01:39:02.0000] <sayrer>
so, no facts, context, or perspective of any kind?

[01:39:03.0000] <Hixie>
(literally starting with the first sentence)

[01:40:00.0000] <Hixie>
it certainly had an interesting perspective

[01:40:01.0000] <sayrer>
no facts?

[01:40:02.0000] <Hixie>
so possibly the statement was inaccurate in that regard

[01:40:03.0000] <Hixie>
i'll certainly grant you that mark, like myself, is prone to exaggeration

[01:40:04.0000] <sayrer>
no context?

[01:40:05.0000] <sayrer>
of any kind?

[01:41:00.0000] <sayrer>
which part of the statement was true?

[01:42:00.0000] <Hixie>
mark's twitter was an exaggerated indication of inaccuracy in sam's document, and sam's document was indeed wildly inaccurate

[01:43:00.0000] <Hixie>
sure, it was exaggerated for humorous value, just like mark's blog posts have humorous comments which you apparently take seriously

[01:43:01.0000] <sayrer>
by exaggerated, you mean wrong?

[01:43:02.0000] <Hixie>
no, i mean exaggerated.

[01:43:03.0000] <Hixie>
why are you offended at an obvious joke (people with hidden agendas? really? you thought that was serious?), and why do you now try to insult someone who has spent so much of his time helping us bring html to a greater audience?

[01:44:00.0000] <sayrer>
he gets paid to that, afaik

[01:44:01.0000] <Hixie>
so...?

[01:45:00.0000] <sayrer>
anyway, you think "HTML is being developed outside of the W3C by a number of browser implementers, excluding Microsoft. "

[01:45:01.0000] <sayrer>
is wrong?

[01:45:02.0000] <Hixie>
yes

[01:45:03.0000] <sayrer>
why?

[01:46:00.0000] <Hixie>
HTML is being developed inside the W3C by 350 people and outside the W3C by over 950 people, and Microsoft contributes roughly equal amounts to html both inside and outside the w3c

[01:46:01.0000] <Hixie>
the statement is wildly misleading from start to finish

[01:46:02.0000] <sayrer>
you mean exaggerated?

[01:46:03.0000] <Hixie>
no, i mean misleading

[01:46:04.0000] <Hixie>
there's no exaggeration there

[01:47:00.0000] <Hixie>
if anything it's the opposite

[01:47:01.0000] <Hixie>
misleading through omission

[01:47:02.0000] <sayrer>
I don't know. you mean the ommission of development inside the W3C?

[01:48:00.0000] <sayrer>
too many mms. anyway, there is some W3C mailing list traffic

[01:48:01.0000] <Hixie>
and the omission of people other than browser vendors, who form the bulk of the contributors, and the omission of the way microsoft's involvement is not any substantially more productive in the w3c than out

[01:48:02.0000] <Hixie>
but i'm really not interested in discussing sam's document, i don't think it is particularly important

[01:48:03.0000] <sayrer>
OK

[01:48:04.0000] <Hixie>
(i and others have already mentioned the numerous mistakes with it)

[01:48:05.0000] <Hixie>
(to him)

[01:49:00.0000] <sayrer>
it's mostly quotes

[01:49:01.0000] <sayrer>
I guess there's a bit of commentary alongside

[01:50:00.0000] <Hixie>
if you're interested in a significantly more balanced and accurate history of html, btw, MikeSmith wrote one on the esw wiki

[01:50:01.0000] <MikeSmith>
heh

[01:50:02.0000] <MikeSmith>
well, there are some that think that is not more balanced or more accurate at all

[01:50:03.0000] <MikeSmith>
which is why my name is at the top of it now

[01:51:00.0000] <MikeSmith>
I was asked to put it there to make it clear that, well, it was written by an actual person

[01:51:01.0000] <Hixie>
i said "more balanced", not that it was perfect :-P

[01:51:02.0000] <MikeSmith>
a person with opinions

[01:51:03.0000] <sayrer>
what is the link?

[01:51:04.0000] <MikeSmith>
http://esw.w3.org/topic/HTML/history

[01:51:05.0000] <sayrer>
I wonder if it spends a lot of time discussing XHTML

[01:52:00.0000] <MikeSmith>
there are subpages like http://esw.w3.org/topic/HTML/DraconianErrorHandling

[01:52:01.0000] <MikeSmith>
sayrer: yeah, both flavors

[01:52:02.0000] <MikeSmith>
e.g., http://esw.w3.org/topic/HTML/XHTML2/objections

[01:53:00.0000] <sayrer>
frankly, this one just looks uncontroversial

[01:53:01.0000] <sayrer>
so I guess there is less to argue about

[01:53:02.0000] <jgraham>
/me wasn't aware that controversy was a goal

[01:54:00.0000] <sayrer>
it is orthogonal to accuracy

[01:54:01.0000] <MikeSmith>
sayrer: anyway, you will notice that Mark Pilgrim's name shows up in that history a lot

[01:54:02.0000] <sayrer>
yours?

[01:54:03.0000] <MikeSmith>
yeah

[01:54:04.0000] <sayrer>
where?

[01:54:05.0000] <MikeSmith>
everywhere

[01:54:06.0000] <MikeSmith>
he has been like a kind of Greek chorus throughout this whole tragic comedy

[01:55:00.0000] <sayrer>
in the linked documents?

[01:55:01.0000] <MikeSmith>
I probably could have dropped a whole more more Pilgrim in there

[01:55:02.0000] <sayrer>
boy oh boy

[01:55:03.0000] <MikeSmith>
but I tried to use it judiciously

[01:55:04.0000] <sayrer>
like... SVG Tiny 1.2 Candidate Recommendation published. 

[01:55:05.0000] <MikeSmith>
Mark's like a fine spice

[01:55:06.0000] <sayrer>
what is that about?

[01:55:07.0000] <MikeSmith>
filler

[01:56:00.0000] <sayrer>
haha

[01:56:01.0000] <MikeSmith>
to make the peanut gallery happy

[01:56:02.0000] <sayrer>
oh, XML On The Web Has Failed

[01:58:00.0000] <sayrer>
it has an entire section titled "Enter RFC 3023"

[01:58:01.0000] <MikeSmith>
sayrer: feel free to add anything to that history page, if you find gaps

[01:58:02.0000] <sayrer>
no thanks

[01:58:03.0000] <MikeSmith>
well, you can even say mean stuff there that you know will really piss people off

[01:59:00.0000] <MikeSmith>
but just put a smiley after it

[01:59:01.0000] <MikeSmith>
and then everything will be OK

[01:59:02.0000] <sayrer>
oh right

[01:59:03.0000] <MikeSmith>
that's the magic of the smiley

[01:59:04.0000] <sayrer>
british humour

[01:59:05.0000] <MikeSmith>
I want to make a smiley that's a little icon of me smiling

[01:59:06.0000] <MikeSmith>
but flipping the bird at the same time

[01:59:07.0000] <MikeSmith>
I would use that one a lot if I had it

[02:00:00.0000] <sayrer>
I'm surprised this hasn't gotten old yet

[02:00:01.0000] <sayrer>
"you're a cunt :)"

[02:00:02.0000] <sayrer>
so many forms of that on the internet

[02:01:00.0000] <jgraham>
MikeSmith: Just make a really popular typeface and thn use some PUA glyphs for extra smilie characters

[02:02:00.0000] <MikeSmith>
sayrer: you know what playing the dozens is?

[02:02:01.0000] <sayrer>
yes

[02:02:02.0000] <MikeSmith>
well, there's only one way to lose that game

[02:03:00.0000] <sayrer>
I thought there were a couple

[02:03:01.0000] <sayrer>
but which did you have in mind?

[02:04:00.0000] <MikeSmith>
the one where you get genuinely mad or hurt by what the other dude says

[02:06:00.0000] <sayrer>
you can also lose by being an actual asshole

[02:06:01.0000] <MikeSmith>
nope

[02:06:02.0000] <MikeSmith>
not in that game

[02:07:00.0000] <MikeSmith>
it's irrelevant

[02:07:01.0000] <MikeSmith>
I'm not defending it as a great game, by the way

[02:07:02.0000] <sayrer>
there are lots of seemingly rude things that are allowed, sure

[02:07:03.0000] <zcorpan__>
"The fact that different user agents deal with protocol violations in different ways is a good thing." -- http://lists.w3.org/Archives/Public/ietf-http-wg/2009AprJun/0030.html

[02:08:00.0000] <takkaria>
uh-oh

[02:10:00.0000] <MikeSmith>
zcorpan__: that is a gem

[02:10:01.0000] <MikeSmith>
[[

[02:10:02.0000] <MikeSmith>
It cannot be standardized in a way that would be safe for safety- 

[02:10:03.0000] <MikeSmith>
critical

[02:10:04.0000] <MikeSmith>
environments such as health care, where failure to display the errors

[02:10:05.0000] <MikeSmith>
could very well result in serious injury or death.

[02:10:06.0000] <MikeSmith>
]]

[02:10:07.0000] <MikeSmith>
solid gold, that

[02:10:08.0000] <MikeSmith>
the "serious injury or death" card

[02:10:09.0000] <sayrer>
yeah, close to godwin's law

[02:11:00.0000] <sayrer>
on the other hand, they aren't protocol violations if everyone handles them the same way

[02:11:01.0000] <Hixie>
aah, mikesmith is ready roy's latest offering

[02:11:02.0000] <jgraham>
"The correct way to interoperate with broken Web

[02:11:03.0000] <jgraham>
content is to display a very large error message that explains why

[02:11:04.0000] <jgraham>
it is broken"

[02:12:00.0000] <jgraham>
I see Roy is pitching for that UI led job he's always coveted

[02:12:01.0000] <MikeSmith>
sayrer: they are because the Bible and/or the Founding Fathers said they should be violations

[02:12:02.0000] <jgraham>
*lead

[02:13:00.0000] <takkaria>
I do wonder what to do with a problem like Roy

[02:13:01.0000] <MikeSmith>
jgraham: very well aligned with market realities

[02:13:02.0000] <hsivonen>
It would be "fun" to generate a version of the HTML5 parser for Gecko that halts on the first parse error

[02:13:03.0000] <Hixie>
sayrer: i think there are things that are clearly errors even if all UAs interoperate on the behaviour

[02:13:04.0000] <Hixie>
sayrer: e.g. <span></em> is clearly an error

[02:13:05.0000] <jgraham>
takkaria: I guess sending him off to work as a children's nanny won't work?

[02:14:00.0000] <sayrer>
<b><i>hmm</b></i> ?

[02:14:01.0000] <Hixie>
that's another example

[02:14:02.0000] <Hixie>
the markup doesn't match the resulting DOM at all

[02:14:03.0000] <sayrer>
where is the error?

[02:15:00.0000] <sayrer>
we should have a good reason to call it an error

[02:15:01.0000] <sayrer>
it's certainly ugly

[02:15:02.0000] <takkaria>
jgraham: I have a feeling his approach to children when they do soething wrong is display a very large erorr message that explains what they've done wrong, probably not too lenient on the error-correcting side of things...

[02:15:03.0000] <annevk42>
if you put a character after </b> you cannot tell what DOM you get

[02:15:04.0000] <annevk42>
unless you know how an HTML parser works

[02:15:05.0000] <annevk42>
therefore it makes sense to flag the simple case as well

[02:15:06.0000] <sayrer>
we're assuming a conformant HTML parser

[02:16:00.0000] <sayrer>
what else would we talk about?

[02:16:01.0000] <annevk42>
I am too

[02:16:02.0000] <MikeSmith>
takkaria: we don't want to do anything. he's actually a secret double agent of the whatwg cabal

[02:16:03.0000] <annevk42>
I'm just saying that getting nodes duplicated is not what you expect to happen

[02:16:04.0000] <sayrer>
but you can check it in your browser and see if it is what you want

[02:16:05.0000] <MikeSmith>
he appears in times of trouble to act as a sort of common foil to unite the cabal

[02:16:06.0000] <takkaria>
how about we call things parse errors where the DOM and the text/html don't intuitively match up?

[02:18:00.0000] <annevk42>
sayrer, that doesn't really lead to understanding how it works or why it works in unexpected ways

[02:18:01.0000] <sayrer>
so parse errors have a pedagogical goal?

[02:18:02.0000] <jgraham>
sayrer: What if it isn't what you want? How do you tell where you made a mistake?

[02:19:00.0000] <sayrer>
a fair point!

[02:19:01.0000] <sayrer>
the difference between a C compiler and a lint tool, right?

[02:20:00.0000] <jgraham>
Hmm.  Not sure that's a good analogy. biab

[02:20:01.0000] <takkaria>
I don't know what other goal they would have

[02:21:00.0000] <takkaria>
HTML prior to HTML5 never had parse errors

[02:21:01.0000] <roc>
it's important to have a common enemy to unite against

[02:22:00.0000] <roc>
hsivonen: you need to implement that strict parser, ship it to Roy and force him to use it exclusively

[02:23:00.0000] <sayrer>
"The error handling for parse errors is well-defined"

[02:23:01.0000] <sayrer>
pretty much says it all

[02:24:00.0000] <annevk42>
I doubt it says the same to you and me

[02:24:01.0000] <sayrer>
well, what does it say to you?

[02:25:00.0000] <zcorpan__>
it's just a statement of fact, might not actually be true

[02:26:00.0000] <zcorpan__>
just like this one http://html5.org/tools/web-apps-tracker?from=2951&to=2952

[02:27:00.0000] <Hixie>
zcorpan__: btw, please send comments on typos and things by e-mail, i can't track them when you just mention them on IRC

[02:27:01.0000] <annevk42>
sayrer, that requirements for authors are different from implementors

[02:27:02.0000] <sayrer>
when I see <b><i>hmm</b></i>

[02:28:00.0000] <sayrer>
I definitely agree there is an aspect "this may not mean what you think it means"

[02:28:01.0000] <MikeSmith>
/me likes roc idea of inviting people to try browsing the Web with a strict-HTML-parser browser for one day or whatever. or even 1 hour. or ten minutes

[02:28:02.0000] <annevk42>
actually, the simple case does what you think it means

[02:28:03.0000] <sayrer>
beside the point

[02:28:04.0000] <annevk42>
problems starts arising when you enter a few more characters between the tags

[02:28:05.0000] <sayrer>
but also invalid?

[02:28:06.0000] <sayrer>
or a parse error?

[02:28:07.0000] <annevk42>
sure

[02:28:08.0000] <sayrer>
why, it does what I think it means

[02:29:00.0000] <Hixie>
<b>a<i>b</b>c</i> is different from <x>a<y>b</x>c</y>, too

[02:29:01.0000] <annevk42>
it's more likely you made a typo

[02:29:02.0000] <zcorpan__>
Hixie: k

[02:29:03.0000] <sayrer>
I agree that is possible

[02:29:04.0000] <sayrer>
but it's not an interoperability problem

[02:30:00.0000] <Hixie>
sayrer: do you think we should allow <blink>, too?

[02:30:01.0000] <sayrer>
yes

[02:30:02.0000] <Hixie>
ok

[02:30:03.0000] <takkaria>
if you mentally do s/parse error/warning/, does that solve your problems?

[02:30:04.0000] <Hixie>
i am confident that your opinion is far from the consensus

[02:30:05.0000] <sayrer>
not really

[02:30:06.0000] <sayrer>
no big deal

[02:30:07.0000] <sayrer>
it's not like I'm going to propose blink

[02:31:00.0000] <sayrer>
since, well, who cares

[02:31:01.0000] <annevk42>
if you don't believe in having differences for authors and impl at all I can see why we differ from opinion :)

[02:31:02.0000] <sayrer>
I think there is value in helping authors understand what they have written

[02:31:03.0000] <zcorpan__>
sayrer: it's easy to make everything allowed - just don't read the spec and don't use a validator

[02:32:00.0000] <sayrer>
that is true for authors

[02:32:01.0000] <zcorpan__>
yes i meant as an author

[02:32:02.0000] <zcorpan__>
if you're not an author, why do you care about conformance requirements on authors? :)

[02:32:03.0000] <sayrer>
I am obviously an author

[02:33:00.0000] <sayrer>
but I don't care about them at all, true

[02:33:01.0000] <sayrer>
I have noticed that most of the controversy in the spec centers around author requirements

[02:34:00.0000] <Hixie>
removing teh author requirements won't remove the controversies

[02:34:01.0000] <zcorpan__>
removing the author requirements doesn't mean that the controversy is removed, too

[02:34:02.0000] <Hixie>
it would add a lot more :-)

[02:34:03.0000] <sayrer>
how so?

[02:34:04.0000] <Hixie>
well for a start i would start bitching :-)

[02:34:05.0000] <sayrer>
but why?

[02:34:06.0000] <sayrer>
is more the question

[02:34:07.0000] <zcorpan__>
people would go "omg html5 allows <font> again! and tag soup! what a mess!"

[02:35:00.0000] <Hixie>
because i think have conformance rules for authors is an extremely valuable tool for quality assurance and for learning the language

[02:35:01.0000] <sayrer>
sounds like what the IETF would call a BCP document

[02:35:02.0000] <sayrer>
and that is something that would be valuable, I agree

[02:36:00.0000] <sayrer>
I am not sure I would call them "conformance rules"

[02:36:01.0000] <sayrer>
still thinking of a better way to put it

[02:36:02.0000] <annevk42>
if you do not want authoring conformance at all it might be easier to start with that next time rather than why <i><b>test</i></i> is non-conforming

[02:37:00.0000] <annevk42>
would save me some time

[02:37:01.0000] <Hixie>
call them what you like, so long as not following them is considered an error

[02:37:02.0000] <Hixie>
and is not allowed

[02:37:03.0000] <sayrer>
it is obviously allowed

[02:37:04.0000] <sayrer>
as HTML5 is written currently

[02:38:00.0000] <Hixie>
given that html5 explicitly disallows the use of incorrect syntax and incorrect semantics, "obvious" is not the word i would use

[02:38:01.0000] <Hixie>
"not" maybe would be the word

[02:39:00.0000] <sayrer>
hmm, there don't seem to be a lot of teeth there

[02:40:00.0000] <sayrer>
like, Apple's HTML5 homepage had incorrect semantics

[02:40:01.0000] <Hixie>
teeth?

[02:40:02.0000] <sayrer>
well, if you use English incorrectly, people might think of less of you, or maybe not even understand you

[02:40:03.0000] <Hixie>
right, and it's incorrect

[02:41:00.0000] <Hixie>
just like using incorrect html

[02:41:01.0000] <Hixie>
it's not allowed by english syntax and grammar

[02:41:02.0000] <Hixie>
it's not allowed by html syntax and grammar

[02:41:03.0000] <sayrer>
but with HTML5, you can do that and people will read your web page just fine

[02:41:04.0000] <Hixie>
same concept

[02:41:05.0000] <sayrer>
and not notice at all

[02:41:06.0000] <Hixie>
so?

[02:41:07.0000] <Hixie>
why would we want them to notice

[02:41:08.0000] <sayrer>
we don't

[02:42:00.0000] <zcorpan__>
you can make them notice by having a draconian html parser which is allowed in the spec

[02:42:01.0000] <sayrer>
so why wag a finger over grammar

[02:42:02.0000] <zcorpan__>
the v.nu html parser is draconian for certain cases in streaming mode

[02:42:03.0000] <roc>
the only real solution is for Google to dish out search points based on HTML5 conformance

[02:43:00.0000] <sayrer>
adsense rates are down, might not work ;)

[02:43:01.0000] <MikeSmith>
sayrer: authors love to be told exactly what to do, even they don't understand the reasons

[02:43:02.0000] <sayrer>
yep

[02:43:03.0000] <annevk42>
Hixie, proposed blog post seems fine

[02:44:00.0000] <annevk42>
Hixie, though I'd indicate they can e-mail public-html-comments as well

[02:44:01.0000] <annevk42>
Hixie, that's easier to post to than whatwg⊙wo

[02:44:02.0000] <Hixie>
annevk42: good idea

[02:44:03.0000] <sayrer>
so the main reason for these conformance rules is so that people won't say bad things?

[02:45:00.0000] <sayrer>
like "ooooh, tag soup! <blink>!"

[02:46:00.0000] <annevk42>
it's recommending how  to author HTML

[02:46:01.0000] <annevk42>
it doesn't prevent anyone from doing something else

[02:46:02.0000] <sayrer>
actually, it's requiring

[02:46:03.0000] <Hixie>
there are dozens of reasons for having authoring requirements, many of which have been discussed tonight

[02:46:04.0000] <Hixie>
- helping people learn how to write effective markup

[02:46:05.0000] <annevk42>
sayrer, only if you follow the spec :)

[02:46:06.0000] <Hixie>
- helping people track down typos and other errors

[02:46:07.0000] <MikeSmith>
sayrer: authors also love to have conformance rules so that they can beat other authors over the head with them, when those other authors make mistakes

[02:46:08.0000] <Hixie>
- helping people write markup that reflects what they intend

[02:47:00.0000] <Hixie>
- encouraging best practices

[02:47:01.0000] <sayrer>
that last one seems bogus

[02:47:02.0000] <MikeSmith>
we need to support that for-beating-others-over-the-head-with use case

[02:47:03.0000] <Hixie>
- satisfying the needs of authors who feel the need for rules to guide them

[02:47:04.0000] <sayrer>
or at least redundant

[02:47:05.0000] <Hixie>
you may agree with some or none of these

[02:48:00.0000] <sayrer>
I actually find the list quite valuable

[02:48:01.0000] <sayrer>
are there more?

[02:48:02.0000] <zcorpan__>
- make people aware of accessiblity issues? at least that seems to be a point pf wants to make about alt

[02:48:03.0000] <Hixie>
there's lots, yes

[02:48:04.0000] <sayrer>
do they all start with "help people..."

[02:48:05.0000] <Hixie>
zcorpan__: yeah, that's similar to the best practices thing

[02:48:06.0000] <sayrer>
?

[02:48:07.0000] <Hixie>
sayrer: no idea

[02:49:00.0000] <sayrer>
well, I guess your current list is enough for me to work with

[02:50:00.0000] <sayrer>
but I would like to know whether any of the author conformance requirements actually have an effect on conformant HTML parsers

[02:50:01.0000] <sayrer>
it doesn't seem that they do

[02:50:02.0000] <sayrer>
which I think is a fine design, I should add

[02:51:00.0000] <Hixie>
author conformance requirements have nothing to do with parsers

[02:51:01.0000] <annevk42>
http://www.w3.org/mid/e9dffd640904011708l3dd36d28xc6e797791777f4df⊙mgc maybe we need WTP, like HTTP, but for Web browsers :p

[02:52:00.0000] <zcorpan__>
Web HTTP

[02:53:00.0000] <sayrer>
Hixie, ok, it's almost like they could be in a different document

[02:55:00.0000] <Hixie>
i think that would be a maintenance nightmare

[02:55:01.0000] <Hixie>
but as it stands the html5 spec can be filtered to show only the author stuff

[02:56:00.0000] <Hixie>
which is a middle ground that seems to address the same needs

[02:56:01.0000] <sayrer>
can it do the opposite?

[02:56:02.0000] <Hixie>
only show implementation stuff? no, but it can highlight implementation stuff, which is about as good

[02:56:03.0000] <Hixie>
(in practice implementations need to know what author rules are because some implementations are conformance checkers)

[02:57:00.0000] <Hixie>
(and annotation the document in more detail would be far too much work)

[02:57:01.0000] <sayrer>
is the author stuff marked up with a specific class?

[02:57:02.0000] <Hixie>
(for me, at least)

[02:57:03.0000] <sayrer>
I haven't looked since you did your great big edit

[02:57:04.0000] <Hixie>
the implementation requirements have class="impl"

[02:58:00.0000] <sayrer>
I can't believe that HTTP WG thread

[02:58:01.0000] <annevk42>
(but in a way that you cannot extract all class="impl" and have it make sense)

[02:58:02.0000] <Hixie>
right, imeplementation requirements share definitions with author stuff

[02:58:03.0000] <Hixie>
anyway

[02:58:04.0000] <Hixie>
i should sleep

[02:58:05.0000] <Hixie>
nn

[02:59:00.0000] <MikeSmith>
cool to see the http mailing list providing some real entertainment value again

[02:59:01.0000] <MikeSmith>
imaginary numbers or no less important than real numbers

[02:59:02.0000] <takkaria>
annevk42: I wonder how large the non-html/js/jpeg portion of the web is

[03:00:00.0000] <annevk42>
large enough for the HTTP WG to exist and not care about it

[03:00:01.0000] <annevk42>
not care about the html/js/jpeg portion, that is

[03:01:00.0000] <zcorpan>
so they care about the ms word and pdf portion?

[03:02:00.0000] <annevk42>
harhar

[03:02:01.0000] <MikeSmith>
why make handling of existing Web content such a huge priority when we have these whole undiscovered galaxies where clients and servers exchange stuff other than things like HTML, Javascript, and JPGs?

[03:02:02.0000] <MikeSmith>
the logic is perfectly clear

[03:02:03.0000] <sayrer>
I suggested publishing Barth's document as informational

[03:02:04.0000] <sayrer>
everyone shook their head at me

[03:03:00.0000] <MikeSmith>
sayrer: I think they shook their beards

[03:03:01.0000] <annevk42>
really? I'd think that's what the HTTP WG guys want

[03:03:02.0000] <sayrer>
they didn't want to publish it without a specific section in the HTTP standard saying "something might happen here"

[03:03:03.0000] <MikeSmith>
the council of wise men who don't suffer fools lightly

[03:04:00.0000] <sayrer>
I don't really see why

[03:04:01.0000] <annevk42>
it might kill people apparently

[03:04:02.0000] <sayrer>
since the value of the information will be the same, no matter what the HTTP spec says

[03:04:03.0000] <MikeSmith>
I hope Adam has learned his lesson feels sufficiently chastised

[03:04:04.0000] <MikeSmith>
all that nonsense about doing real research and collecting data. who cares about that?

[03:04:05.0000] <annevk42>
I think it should be more than informational personally

[03:04:06.0000] <sayrer>
why?

[03:04:07.0000] <MikeSmith>
it will be

[03:05:00.0000] <sayrer>
what does it matter?

[03:05:01.0000] <MikeSmith>
sayrer: because it will get implemented

[03:05:02.0000] <MikeSmith>
or something like it

[03:05:03.0000] <sayrer>
W3C standards are only recommendations

[03:05:04.0000] <annevk42>
sayrer, by that metric we might as well not bother at all and just make it publicly available

[03:05:05.0000] <annevk42>
e.g. by sending a copy to www-archive

[03:05:06.0000] <sayrer>
JSON is only an informational RFC

[03:05:07.0000] <MikeSmith>
sayrer: the ones that aren't just recommendations are called "implementations"

[03:06:00.0000] <MikeSmith>
or "implemented"

[03:06:01.0000] <annevk42>
sayrer, I don't quite get that either, though I understand JSON has some interop issues

[03:08:00.0000] <sayrer>
DNS is an informational RFC too

[03:08:01.0000] <sayrer>
the status doesn't matter

[03:08:02.0000] <MikeSmith>
sayrer: it's moot point anyway

[03:08:03.0000] <annevk42>
sayrer, mkay

[03:08:04.0000] <sayrer>
what do you mean?

[03:09:00.0000] <MikeSmith>
they clearly aren't going to publish it regardless

[03:09:01.0000] <sayrer>
you think?

[03:09:02.0000] <sayrer>
why?

[03:09:03.0000] <zcorpan>
hmm. here's an idea -- have the same tree builder as in html but use a different namespace and do namespace fixups on attributes when in foreign

[03:09:04.0000] <zcorpan>
then support /> on "unknown" elements everywhere and do element case fixup everywhere

[03:10:00.0000] <MikeSmith>
sayrer: because it's too clearly aligned with the real world

[03:10:01.0000] <sayrer>
I think it is too conservative actually

[03:10:02.0000] <sayrer>
maybe they would go for it!

[03:10:03.0000] <annevk42>
zcorpan, idea for what?

[03:11:00.0000] <zcorpan>
annevk42: handling svg and mathml in html in an html-y way

[03:11:01.0000] <zcorpan>
allowing proper speculative tokenization

[03:12:00.0000] <zcorpan>
even past <svg> ... <select><style> ...

[03:13:00.0000] <zcorpan>
hsivonen: ^

[03:13:01.0000] <zcorpan>
sicking: ^

[03:14:00.0000] <annevk42>
I'm not really following, but ok

[03:14:01.0000] <zcorpan>
instead of having insertion modes, have a flag for foreign content and use the insertion modes you would use for html (except you insert elements with a different namespace)

[03:16:00.0000] <Philip`>
Hixie: No

[03:16:01.0000] <Philip`>
but I am now

[03:16:02.0000] <Philip`>
/me is having fun upgrading his kernel on Gentoo and switching to a new wireless driver architecture and wondering why kdm no longer displays anything on the screen

[03:17:00.0000] <annevk42>
zcorpan, and then have special handling for some elements?

[03:18:00.0000] <annevk42>
zcorpan, e.g. <image>

[03:18:01.0000] <annevk42>
zcorpan, but why can't you speculate past <svg> currently?

[03:19:00.0000] <hsivonen>
annevk42: currently, you can do speculative tree building past <svg> but you can't do cheap speculative tokenization without tree building

[03:20:00.0000] <zcorpan>
annevk42: forgot about <image>

[03:20:01.0000] <hsivonen>
I want to instrument the HTML5 parser in Gecko and get some data about the feasibility of doing speculative tree building only before I express more opinions on this topic

[03:20:02.0000] <zcorpan>
but yeah, you could check the flag for <image>

[03:21:00.0000] <hsivonen>
zcorpan: case fixup everywhere might be a good idea regardless of speculative parsing if we end up with a magic mapping for selectors and getElementsByTagName

[03:22:00.0000] <annevk42>
bz said that making selectors and gEBTN work like HTML5 currently requirescould work

[03:22:01.0000] <annevk42>
/me much rather has that

[03:24:00.0000] <hsivonen>
annevk42: it would bloat all interned node ns/local/prefix interned struct by another pointer and would require changes to code is unexpected places

[03:24:01.0000] <hsivonen>
annevk42: but clearly, storing pre-lowercased local names for each node would work

[03:25:00.0000] <annevk42>
you could also do 2 pointer checks at matching time

[03:25:01.0000] <hsivonen>
annevk42: what's the other pointer check?

[03:25:02.0000] <annevk42>
isHTMLElement && equallowercasetoken || equaltoken

[03:26:00.0000] <annevk42>
might need parenthesis there

[03:26:01.0000] <hsivonen>
if there were pre-lowercased atoms for each node, it would make sense to have those on all nodes anyway to avoid an extra check

[03:27:00.0000] <zcorpan>
if the parser camelcases everywhere, should createElement similarly camelcase?

[03:27:01.0000] <hsivonen>
anyway, doing it "right" adds some RAM footprint and requires a review of a lot of code only to support non-conforming cases that currently Gecko, WebKit and IE don't support and only Opera currently supports

[03:28:00.0000] <hsivonen>
zcorpan: maybe, but you'd still get the wrong namespace

[03:28:01.0000] <annevk42>
hsivonen, that wouldn't allow for gEBTN("textarea") to not match textArea but gEBTN("textArea") to match textArea

[03:28:02.0000] <zcorpan>
hsivonen: yep

[03:28:03.0000] <hsivonen>
annevk42: that can be solved by not supporting textArea

[03:29:00.0000] <annevk42>
same example for any other camelCase name

[03:29:01.0000] <zcorpan>
there are no other conflicting camelcase names

[03:29:02.0000] <zcorpan>
or are there?

[03:29:03.0000] <hsivonen>
AFAIK, textArea is the only conflict

[03:29:04.0000] <annevk42>
I don't think so, but you might still not want them to match case-insensitively

[03:30:00.0000] <hsivonen>
and textArea is a bad idea in the first place since it duplicates CSS formatter functionality

[03:30:01.0000] <annevk42>
and it also doesn't help with createElementNS() nodes

[03:30:02.0000] <annevk42>
and I think it would be nice if it did work with createElementNS nodes

[03:30:03.0000] <hsivonen>
annevk42: it helps with conforming createElementNS() nodes

[03:30:04.0000] <annevk42>
only if you change createElementNS

[03:31:00.0000] <zcorpan>
if you want to select elements with a certain case, use getEBTNNS

[03:31:01.0000] <hsivonen>
annevk42: the only issue is if a WG somewhere creates a camelCaseML for XML and then realizes they need to ship an Ajax library to hack support into text/html

[03:31:02.0000] <annevk42>
Is (isHTMLElement && equallowercasetoken) || equaltoken really that bad?

[03:32:00.0000] <hsivonen>
annevk42: it's worse than simply (caseFoldAtom == queryAtom)

[03:34:00.0000] <hsivonen>
I'd be surprised if Opera implemented what it currently does in a way other than doing what Maciej said on the list

[03:35:00.0000] <hsivonen>
as far as I can tell, adding a pointer to all Gecko node infos and WebKit qualified names would be the Right Way to implmement real ASCII-case-insensitivity

[03:40:00.0000] <annevk42>
hsivonen, we match createElementNS("x","X") with gEBTN("x")

[03:40:01.0000] <annevk42>
(not sure that's what I want either)

[03:41:00.0000] <hsivonen>
annevk42: that neither proves nor disproves my hypothesis of Opera internals :-)

[03:42:00.0000] <hsivonen>
my hypothesis isn't blackbox testable except by observing memory footprint and timing execution

[03:43:00.0000] <annevk42>
fair enough

[03:55:00.0000] <annevk42>
sweet

[03:55:01.0000] <annevk42>
the Forms WG published a whole bunch of Member confidential information

[03:55:02.0000] <annevk42>
http://lists.w3.org/Archives/Public/public-forms/2009Apr/att-0000/2009-04-01.html

[03:55:03.0000] <annevk42>
regarding the AC meeting

[04:03:00.0000] <annevk42>
not too interesting though

[04:04:00.0000] <MikeSmith>
annevk42: I should remove that

[04:04:01.0000] <MikeSmith>
but I don't think I actually have the ability to remove attachments from teh archive

[04:04:02.0000] <annevk42>
I hope someone makes a copy

[04:14:00.0000] <hsivonen>
already gone :-/

[04:15:00.0000] <MikeSmith>
temporarily gone

[04:15:01.0000] <annevk42>
I do have a copy

[04:18:00.0000] <Philip`>
/me wonders what genius thought the new wireless drivers should make the LED blink at about 4Hz every time a packet is sent or received

[04:24:00.0000] <jgraham>
Philip`: Presumably one that doesn't suffer from photosensitive epilepsy

[04:26:00.0000] <annevk42>
Hixie, your ARIA comments went to the wrong list

[04:26:01.0000] <annevk42>
Hixie, they need to go to public-pfwg-comments⊙wo

[04:35:00.0000] <hsivonen>
Hixie: why is there "XXX quirks" on 'hr' start tag? as far as I can tell, none of Gecko, WebKit, Opera and IE8 have that quirk

[04:39:00.0000] <hsivonen>
hmm. Safari and Opera split the text node on <body>foo</h1>bar

[04:41:00.0000] <annevk42>
I guess I should stop reviewing non-normative materal in WAI-ARIA

[04:45:00.0000] <hsivonen>
MikeSmith: I redeployed V.nu

[04:46:00.0000] <MikeSmith>
hsivonen: cool. hoping I didn't introduce any regresssions

[04:49:00.0000] <annevk42>
you should probably add tests for code you touch :)

[04:50:00.0000] <zcorpan>
hsivonen: ie nests the hr in p if there was no text in the p

[04:52:00.0000] <hsivonen>
zcorpan: ah. I tested with text

[04:52:01.0000] <hsivonen>
zcorpan: any indication of Web actually needing that?

[04:52:02.0000] <zcorpan>
seems not since other browsers don't do it

[04:53:00.0000] <zcorpan>
actually opera nests hr in p if there's another element in between

[04:53:01.0000] <zcorpan>
but i haven't tested whether styling agrees

[04:54:00.0000] <zcorpan>
nothing surprising with the styling

[04:55:00.0000] <hsivonen>
so it seemss the only quirk left in the spec XXX that everyone implements is <p><table>

[04:55:01.0000] <zcorpan>
gecko nests hr for <p><span><hr>, too

[04:56:00.0000] <hsivonen>
but it's the <span>, not <hr> that is special, right?

[05:07:00.0000] <zcorpan>
in gecko yes

[05:08:00.0000] <zcorpan>
in opera span is just as special as em and i, but unknown elements don't cause nesting

[05:09:00.0000] <annevk42>
doesn't IE nest <hr> inside <p>?

[05:09:01.0000] <zcorpan>
annevk42: sometimes

[05:10:00.0000] <zcorpan>
hsivonen: actually the styling doesn't agree with the dom in opera for <p><span><hr>x

[05:10:01.0000] <zcorpan>
we style it more like as if the p was closed

[05:12:00.0000] <hsivonen>
/me wonders what the Opera CSS frame constructor is like

[05:12:01.0000] <zcorpan>
i could tell you but then we'd have to hire you ;)

[05:17:00.0000] <zcorpan>
hsivonen: ie inserts an empty element for most blockish end tags

[05:17:01.0000] <zcorpan>
hsivonen: like e.g. </plaintext>

[05:18:00.0000] <zcorpan>
hsivonen: only </p> is needed for compat :)

[05:19:00.0000] <Philip`>
/me supposes </plaintext> is pretty rare in practice

[05:19:01.0000] <Philip`>
seeing as even people legitimately using the obsolete element won't use it

[05:20:00.0000] <Philip`>
but it wouldn't surprise me to see a non-zero number of occurrences of it anyway :-(

[05:20:01.0000] <hsivonen>
zcorpan: Interesting. that suggests that the </h1> stuff isn't important

[05:20:02.0000] <zcorpan>
hsivonen: indeed

[05:21:00.0000] <annevk42>
OMG: http://www.w3.org/TR/wai-aria/#textequivalentcomputation

[05:21:01.0000] <annevk42>
make that OMG: http://www.w3.org/TR/2009/WD-wai-aria-20090224/#textequivalentcomputation

[05:21:02.0000] <annevk42>
'cause hopefully they fix this mess

[05:23:00.0000] <annevk42>
a lot of this ARIA stuff is just written down way more fricking complex than necessary

[05:23:01.0000] <annevk42>
you don't need RDF or OWL

[05:23:02.0000] <annevk42>
not for conformance checkers anyway, which is what the initial claim in the draft is

[05:24:00.0000] <annevk42>
it's a bit sad that all this is so over engineered, I imagine they spent a lot of time on these things for zero benefit

[05:25:00.0000] <annevk42>
I'm now actually making comments, though not on most of the editorial mayhem: http://lists.w3.org/Archives/Public/public-pfwg-comments/2009AprJun/

[05:26:00.0000] <annevk42>
OMG2: http://www.w3.org/TR/2009/WD-wai-aria-20090224/rdf_model.png

[05:32:00.0000] <Philip`>
Oh, my LEDs stopped blinking

[05:32:01.0000] <Philip`>
and I'm not sure why, since I tried to run some script I found on the internet to disable the blinking but it had no effect for at least an hour

[05:33:00.0000] <hsivonen>
/me wonders if there's a generic UML<->RDF mapping

[05:33:01.0000] <MikeSmith>
annevk42: point taken (about tests for v.nu patches)

[05:46:00.0000] <MikeSmith>
speaking of playing the dozens, I really miss that mookid conneg dude

[05:50:00.0000] <hsivonen>
Argh. I broke XSLT.

[05:50:01.0000] <hsivonen>
hmm. maybe not

[05:54:00.0000] <hsivonen>
does anyone happen to know if Web compat requires http://mxr.mozilla.org/mozilla-central/source/content/html/content/test/347174transform.xsl to 'work' without neither the XHTML namespace nor the html output method?

[06:17:00.0000] <hsivonen>
Is ARIA Best Practices in LC or just ARIA itself?

[06:21:00.0000] <annevk42>
just ARIA

[06:28:00.0000] <annevk42>
someone argued to me about live-region yesterday

[06:28:01.0000] <annevk42>
live-regions are apparently dropped

[06:28:02.0000] <jgraham>
?!

[06:28:03.0000] <jgraham>
I thought live regions were supposed to be the most useful part

[06:28:04.0000] <annevk42>
or is that just a comment?

[06:29:00.0000] <annevk42>
ah, never mind

[06:29:01.0000] <annevk42>
it was simplified or something: http://lists.w3.org/Archives/Public/public-pfwg-comments/2009JanMar/0035.html

[06:29:02.0000] <annevk42>
/me read it too quickly initially

[06:29:03.0000] <annevk42>
I'm getting sick of all these auto-responders

[06:30:00.0000] <MikeSmith>
"make tests, not war"

[06:32:00.0000] <annevk42>
is that a DanC quote?

[06:32:01.0000] <annevk42>
:)

[06:40:00.0000] <annevk42>
http://www.w3.org/mid/201536C5-96EB-4029-B4A7-D3FB0EAF25FA⊙ao "I strongly believe that specialized mime-types will be useful in the future!" (MathML 3.0 will have 3 apparently.)

[06:42:00.0000] <hsivonen>
sigh. does it still have xml:id, too?

[06:43:00.0000] <annevk42>
I guess some people missed the news about media types, XML, etc.

[06:48:00.0000] <hsivonen>
has anyone filed a formal objection about xml:id in MathML 3, yet?

[06:48:01.0000] <annevk42>
i filed a comment

[06:49:00.0000] <hsivonen>
annevk42: has it been handled yet?

[06:49:01.0000] <annevk42>
no

[06:49:02.0000] <hsivonen>
ok

[06:50:00.0000] <annevk42>
but I filed my comment in April and they have published something since that date

[06:52:00.0000] <hsivonen>
annevk42: happens all the time with HTML5 :-)

[06:52:01.0000] <annevk42>
true

[06:53:00.0000] <hsivonen>
considering all the MUST around @alt, ARIA sure says a lot of SHOULD for authoring reqs

[06:54:00.0000] <annevk42>
hehe

[06:57:00.0000] <MikeSmith>
I wish we could all avoid using the words "formal objection"

[06:57:01.0000] <MikeSmith>
it is a blunt instrument

[06:57:02.0000] <jgraham>
MikeSmith: I would like to make a formal objection to that wish

[06:57:03.0000] <MikeSmith>
jgraham: good smartass-ery

[06:58:00.0000] <MikeSmith>
we need more of that

[06:58:01.0000] <MikeSmith>
need to pad out whatwg blog with more

[06:59:00.0000] <jgraham>
MikeSmith: Actually I thought it was pretty lame. I almost didn't bother. Then I realised that consstency is the hobgoblin of little minds, so I ought to be consistent in my lameness.

[06:59:01.0000] <MikeSmith>
heh

[06:59:02.0000] <MikeSmith>
even better

[06:59:03.0000] <annevk42>
hobgoblin is the name of my dreamhost server

[06:59:04.0000] <Philip`>
The W3C should require all formal objections to actually be formal, and hand-written on paper with an official letterhead and wax seal

[07:00:00.0000] <MikeSmith>
yeah, and delivered in person

[07:00:01.0000] <MikeSmith>
and you only get one per lifetime

[07:00:02.0000] <jgraham>
And signed in blood

[07:00:03.0000] <jgraham>
Your own. 

[07:00:04.0000] <jgraham>
Lots.

[07:00:05.0000] <MikeSmith>
and if you ever try to file a 2nd one, we have somebody who comes and visits you and kills you

[07:00:06.0000] <jgraham>
Hence the one per lifetime restriction

[07:01:00.0000] <MikeSmith>
draconian error handling

[07:01:01.0000] <hsivonen>
like suggesting new C++ features?

[07:06:00.0000] <hsivonen>
what kind of sharp instruments are available?

[07:07:00.0000] <MikeSmith>
hsivonen: assertions of technical merit

[07:16:00.0000] <jgraham>
MikeSmith: Arguments from technical merit don't seem to be sharp enough to cut through philosophical divides

[07:18:00.0000] <MikeSmith>
blunt objections don't work any better

[07:24:00.0000] <Philip`>
Blunt objects work better than blunt objections

[07:24:01.0000] <jgraham>
MikeSmith: Sure

[07:25:00.0000] <jgraham>
Philip`: Do sharp objects work better than blunt objects?

[07:26:00.0000] <MikeSmith>
sharp object slide in between the ribs better

[07:28:00.0000] <zcorpan>
Hixie: why did you change the forum antispam message from "FIVE"? it's annoying to have to copy-paste every time

[07:28:01.0000] <Philip`>
MikeSmith: If you're trying to slide a blunt object between people's ribs, you're doing it wrong

[07:28:02.0000] <Philip`>
I suggest smashing them on the back of the head

[07:29:00.0000] <MikeSmith>
kneecaps

[07:31:00.0000] <Philip`>
That's less fatal

[07:32:00.0000] <Philip`>
(That's just an observation of fact, not a comment on whether I think that's a benefit or a drawback)

[08:02:00.0000] <MikeSmith>
blockquote is not allowed in footer, right?

[08:02:01.0000] <MikeSmith>
neither in header nor footer?

[08:03:00.0000] <MikeSmith>
hmm, or it is allowed in footer?

[08:03:01.0000] <MikeSmith>
if so, I wonder why

[08:04:00.0000] <Philip`>
If online purchasing systems redirecting me to https://www.securesuite.co.uk (who I've never heard of and have no reason to trust) and telling me to enter my online banking password weren't bad enough, I've now found one that opens that site in a snazzy drop-shadowed iframe, which means I can't tell if the page inside the iframe even has a valid certificate and isn't https://a-really-dodgy-domain.info/

[08:06:00.0000] <zcorpan>
MikeSmith: should be allowed if i'm reading the spec correctly

[08:06:01.0000] <zcorpan>
MikeSmith: blockquote is sectioning root but sectioning root is not sectioning content

[08:06:02.0000] <MikeSmith>
hmm

[08:07:00.0000] <MikeSmith>
so why should it not be allowed in header as well?

[08:07:01.0000] <zcorpan>
it is allowed in header afaict

[08:07:02.0000] <zcorpan>
header just bans sectioning content

[08:09:00.0000] <zcorpan>
MikeSmith: i think blockquote was sectioning content before, which is probably why v.nu disallows it

[08:09:01.0000] <MikeSmith>
zcorpan: 

[08:09:02.0000] <MikeSmith>
zcorpan: OK

[08:10:00.0000] <zcorpan>
hmm.. <header><blockquote><h1></h1></blockquote></header> seems to be allowed in the content model

[08:11:00.0000] <MikeSmith>
yeah, 'cause I can't see any constraint assertions in the current spec to indicate that blockquote is not allowed in header or footer  

[08:11:01.0000] <zcorpan>
but the outline ignores teh blockquote

[08:11:02.0000] <zcorpan>
i think

[08:11:03.0000] <MikeSmith>
maybe that's the issue

[08:11:04.0000] <jgraham>
zcorpan: Yes

[08:11:05.0000] <jgraham>
Because the <h1> is part of the quote so it is not the heading of the current page

[08:12:00.0000] <zcorpan>
where's gsnedders' outline tool?

[08:12:01.0000] <MikeSmith>
hsivonen: blockquote in header and footer?

[08:14:00.0000] <zcorpan>
gsnedders: please add a textarea to your outliner

[08:14:01.0000] <MikeSmith>
v.nu assertions.sch still says, "The sectioning element blockquote..."

[08:14:02.0000] <MikeSmith>
which it's not a sectioning element

[08:14:03.0000] <MikeSmith>
or not now at least

[08:15:00.0000] <MikeSmith>
makes me thinks maybe the rationale for that constraint is outdated/not in sync with current spec

[08:15:01.0000] <zcorpan>
so clearly the content model for header needs to be smarter about h1 and blockquote

[08:21:00.0000] <zcorpan>
jgraham: gsnedders outliner doesn't ignore the blockquote when it's a descendant of header

[08:22:00.0000] <jgraham>
zcorpan: I guess that's a bug

[08:22:01.0000] <zcorpan>
jgraham: in the spec or in the impl?

[08:23:00.0000] <jgraham>
In the impl.

[08:23:01.0000] <jgraham>
At least, if my understanding of the rationale behind the current spec is right

[08:25:00.0000] <jgraham>
zcorpan: http://james.html5.org/outliner.html has the same issue so I guess maybe it is a spec bug?

[09:01:00.0000] <zcorpan>
the pf people seem to think authors are angels and therefore we can expose them to aria-labelledby and aria-describedby etc and disabled users will benefit by both short and long descriptions

[09:01:01.0000] <zcorpan>
meanwhile authors generally can't use alt correctly (if at all)

[09:02:00.0000] <jgraham>
The pf people seem to have the alarming notion that authors generally know what they are doing and therefore should be given as much control as possible

[09:02:01.0000] <zcorpan>
yeah, that's what i meant with "angels" :)

[09:03:00.0000] <jgraham>
<cynic>Or that authors have no idea what they are doing and so having complex technology results in lots of consulting jobs</cynic> :)

[09:03:01.0000] <jgraham>
Note that I am not seriously suggesting the second option

[09:04:00.0000] <jgraham>
Because that would be ascribing to malice what can be adequately explained by naivity

[09:04:01.0000] <beowulf>
is that malice?

[09:05:00.0000] <jgraham>
beowulf: Making accessibility intentionally hard to line one's own pockets would be malice

[09:06:00.0000] <jgraham>
(also: most a11y advocates I have met do not strike me as malicious)

[09:08:00.0000] <zcorpan>
MikeSmith: i get b.v.nu email even though i'm not explicitly cc-ed

[09:08:01.0000] <gsnedders>
zcorpan, jgraham: I pretty much exactly implement the spec, so almost certainly spec

[09:08:02.0000] <jgraham>
gsnedders: Me too, but neither of us have great test coverage so…

[09:09:00.0000] <jgraham>
(the fact that we have the same bug does suggest a spec bug)

[09:09:01.0000] <gsnedders>
I just think <http://hg.gsnedders.com/anolis/file/b6d93515d41e/anolislib/processes/outliner.py> is so close to the spec that it's unlikely to be an impl bug :)

[09:09:02.0000] <jgraham>
gsnedders: iirc the spec is so hard to understand that anything is posssible

[09:09:03.0000] <gsnedders>
jgraham: When did you last read the spec?

[09:10:00.0000] <gsnedders>
jgraham: Hixie made quite a few edits just before he wrote URLs

[09:10:01.0000] <gsnedders>
(Which is when I was writing that impl.)

[09:10:02.0000] <zcorpan>
i tried to read the spec and i couldn't tell whether it should ignore sectioning roots in header or not

[09:11:00.0000] <jgraham>
gsnedders: Maybe before that. After the algorithm changed from the treewalker-based one though

[09:11:01.0000] <gsnedders>
jgraham: A lot of my comments were about spec clarity, as there were things where I wasn't sure what to do because the spec was ambiguous

[09:12:00.0000] <Philip`>
zcorpan: Maybe they think that a non-zero number of authors are angels and therefore if we expose them to aria-labelledby and aria-describedby etc then a non-zero number of disabled users will benefit, and so the effort will have been worthwhile

[09:13:00.0000] <gsnedders>
zcorpan: Everything should be ignored in a header, me thinks

[09:13:01.0000] <zcorpan>
Philip`: not if a greater number of authors use aria-labelledby and aria-describedby incorrectly so that users will have a better experience overall if the attributes are ignored

[09:13:02.0000] <Philip`>
(It's clearly impossible to make everything better for everyone, so the scope is limited to making some things better for some people)

[09:14:00.0000] <jgraham>
Philip`: That doesn't sound like a good way of making an optimum solution

[09:14:01.0000] <jgraham>
For example, some disabled people would be better off if no one had spent any time working on aria but had fixed a few existing websites

[09:14:02.0000] <zcorpan>
gsnedders: what do you use as the heading text in the outline for <header>?

[09:14:03.0000] <jgraham>
to be more accessible

[09:15:00.0000] <gsnedders>
zcorpan: See paragraph two of #the-header-element

[09:15:01.0000] <gsnedders>
"For the purposes of document summaries, outlines, and the like, the text of header elements is defined to be the text of the highest ranked h1–h6 element descendant of the header element, if there are any such elements, and the first such element if there are multiple elements with that rank. If there are no such elements, then the text of the header element is the empty string."

[09:15:02.0000] <Philip`>
zcorpan: Depends on whether the attributes have a negative effect on user experience when used incorrectly, and it seems easy to imagine that users just won't press the 'access description of this element' key when they know the site isn't going to be using it usefully

[09:16:00.0000] <zcorpan>
gsnedders: well that's the thing. that paragraph should ignore h1-h6 in descendants that are sectioning roots

[09:16:01.0000] <gsnedders>
zcorpan: Well, spec bug

[09:17:00.0000] <gsnedders>
/me is just what Hixie told him to :P

[09:18:00.0000] <zcorpan>
Philip`: presumably the AT would tell the user that there is a description available, though?

[09:18:01.0000] <jgraham>
gsnedders: That won't stand up in a court of law when your outliner is used in a life-critical system and someone DIES

[09:21:00.0000] <jgraham>
Sorry dunno what came over me there, guess I should take a rest

[09:22:00.0000] <jgraham>
Er, anyway, moving on

[09:23:00.0000] <jgraham>
Philip`: Your argument still seems like a logical fallacy

[09:23:01.0000] <jgraham>
Since you are asserting that doing anything that helps a non-zero number of people is good, even if it isn't the optimum solution

[09:28:00.0000] <Philip`>
jgraham: Now you're just engaging in the fallacy of "argument by logical fallacy"

[11:22:00.0000] <gsnedders>
/me has managed to hang tesseract

[11:34:00.0000] <gsnedders>
Wow. Fun. It hangs ocrad too

[11:34:01.0000] <gsnedders>
This is one awesome image.

[11:35:00.0000] <gsnedders>
Ah, ocrad comes out of hang, but output is totally wrong

[11:38:00.0000] <gsnedders>
/me crops image into segments and throws it at tesseract again, and it works straight away

[11:38:01.0000] <gsnedders>
Oh well

[12:37:00.0000] <Hixie>
hsivonen: re XXX quirks things -- the XXX comments in the spec are what i put in when i don't want to think about something at all

[12:38:00.0000] <Hixie>
hsivonen: e.g. if i happen to notice a mistake, don't have the inclination to fix it straight away, and want to move on to something more important (at the time)

[12:38:01.0000] <Hixie>
hsivonen: so "XXX quirks" might just mean i happen to notice something fishy while doing some other testing, and i should look more closely

[12:38:02.0000] <Hixie>
hsivonen: it doesn't mean "there is definitely a quirk to add here"

[13:15:00.0000] <mpilgrim>
there is a pending (non-spam!) draft on the WHATWG blog

[13:15:01.0000] <mpilgrim>
by sayrer

[13:15:02.0000] <mpilgrim>
any clue if he's done with it and wants it published?

[13:18:00.0000] <mpilgrim>
it seems done but still says "draft", not "pending review"

[13:19:00.0000] <jgraham>
I can deal with that

[13:19:01.0000] <gavin_>
he is online on moznet

[13:19:02.0000] <jgraham>
If he wants it published

[13:20:00.0000] <mpilgrim>
i'm in there anyway, publishing another "this week"

[13:21:00.0000] <Philip`>
/me thinks it's going by a fairly loose definition of "this"

[13:22:00.0000] <mpilgrim>
/me slinks away in shame

[13:22:01.0000] <mpilgrim>
it's been a hectic quarter

[13:23:00.0000] <mpilgrim>
i'm catching up now so i can resume regular weekly posts in Q2

[13:23:01.0000] <Philip`>
I'm not complaining, just commenting :-)

[13:25:00.0000] <jgraham>
Maybe we should rename it "Some past week in HTML5, untill Google finally invents time travel, at which point we'll need several new tenses to describe just what is going on anyway"

[13:26:00.0000] <Philip`>
"A week in HTML5"

[13:26:01.0000] <mpilgrim>
pretty sure the past tense is taken ;)

[13:48:00.0000] <Hixie>
/me grumbles

[13:48:01.0000] <Hixie>
openDatabase() should have been a constructor

[13:48:02.0000] <Hixie>
i hate watching myself screw up the web

[13:48:03.0000] <gsnedders>
Well it's hardly as if _I_ wrote the spec that's now interoperably implemented!

[13:49:00.0000] <Lachy>
Hixie, re http://www.w3.org/Bugs/Public/show_bug.cgi?id=6477 - what made you decide to add support for input.height and .width if browsers don't support it?

[13:49:01.0000] <Hixie>
Lachy: i don't remember. they were added long before i added that comment.

[13:49:02.0000] <Hixie>
i probably thought browsers did have them

[13:50:00.0000] <Lachy>
ok

[13:50:01.0000] <Lachy>
I suppose they shouldn't be too hard to implement. Not sure if it's really worth the effort though.

[13:51:00.0000] <gsnedders>
Is it bad that more than half the money (in terms of value!) in my wallet is coppers?

[13:51:01.0000] <Lachy>
gsnedders, that depends if by "coppers" you mean really low value coins, like 1 and 2 pence peices (or whatever currency you use)

[13:52:00.0000] <gsnedders>
Yeah

[13:52:01.0000] <jcranmer>
gsnedders: you have copper pound notes in the UK?

[13:52:02.0000] <gsnedders>
jcranmer: No

[13:52:03.0000] <gsnedders>
Coppers as in coins worth 0.01 GBP and 0.02 GBP

[13:54:00.0000] <Lachy>
jcranmer, originally, 1 GBP was actually measured as 1 pound of silver. Imagine carrying several of those around in your wallet :-)

[13:54:01.0000] <Hixie>
i thought he meant police men

[13:54:02.0000] <gsnedders>
heh

[13:54:03.0000] <jcranmer>
"

[14:05:00.0000] <Philip`>
/me was unaware until recently that Isaac Newton was heavily involved with currency reform

[14:06:00.0000] <Hixie>
it's a big plot point in the baroque cycle

[14:06:01.0000] <Philip`>
That's how I became aware of it :-)

[14:07:00.0000] <Philip`>
(It's probably unwise to take everything in those books as strictly true, but Wikipedia seems to agree that he was involved)

[14:07:01.0000] <Hixie>
indeed

[14:08:00.0000] <gsnedders>
http://stuff.gsnedders.com/atom-iri/ — start of IRI testing!

[14:08:01.0000] <gsnedders>
Not that I got very far, because it took me a while to work out how to produce those files

[14:35:00.0000] <Philip`>
I love how selecting "run as administrator" on a 1.5GB .exe in Vista results in it freezing for ten minutes while silently copying the entire file into c:\windows\temp, and then it runs out of disk space and doesn't clean up after itself and leaves me with 0 bytes of free space

[14:39:00.0000] <Lachy>
Philip`, which program has a 1.5GB exe file?

[14:39:01.0000] <Philip`>
Lachy: An installer for some game

[15:52:00.0000] <Hixie>
are there any exception objects on the Web other than DOMException? And is it ok to make up new ones?

[16:04:00.0000] <heycam>
Hixie, there's an SVGException. don't see any reason not to make a new one.

[16:06:00.0000] <Hixie>
ok

[16:57:00.0000] <Hixie>
ok, the invitation for people to review the spec is Out There


2009-04-03
[17:03:00.0000] <Lachy>
Hixie, "Anyone who helps [...] will get their name in the acknowledgements section." - So what's in it for those of us who's name is already in there? I didn't see any other incentives, why should I bother ;-).

[17:03:01.0000] <Lachy>
(aside from the fact that I already get paid to do this as part of my job)

[17:04:00.0000] <Hixie>
so many possible replies

[17:04:01.0000] <Hixie>
i'm torn between "if your name is already in there, you already know the rush that comes from sending feedback, and so you don't need any other incentive" and "if your name is already in there, i don't want more of your damn e-mails". :-P

[17:07:00.0000] <Niictar24>
Wow, that thing is long

[17:08:00.0000] <Niictar24>
839 pages (!)

[17:08:01.0000] <Hixie>
the html5 spec? yeah

[17:10:00.0000] <Lachy>
/me wonders how many people are going to try printing that

[17:10:01.0000] <Hixie>
everyone at microsoft, i expect

[17:10:02.0000] <Hixie>
other than that, i hope nobody!

[17:10:03.0000] <Niictar24>
/me was planning on it before he finished downloaded it

[17:10:04.0000] <Niictar24>
downloading*

[17:10:05.0000] <Niictar24>
Not so much now

[17:10:06.0000] <Lachy>
wow, the acknowledgement section takes up 2 pages on its own.

[17:11:00.0000] <Hixie>
and that's after extracting the websockets, eventsource, and webstorage sections, too

[17:11:01.0000] <Niictar24>
After this invite, that will probably get much bigger

[17:11:02.0000] <Hixie>
i'm going for the award for "most community involvement in a web specification's development"

[17:11:03.0000] <Hixie>
:-P

[17:12:00.0000] <Lachy>
I'm not so sure it will get too much bigger.  The invite has so far only reached the limited number of people who are already following the spec development in some capacity.

[17:13:00.0000] <Niictar24>
I read somewhere there are over 800 subscribers

[17:13:01.0000] <Niictar24>
If each one of them finds and reports a typo?

[17:13:02.0000] <Niictar24>
Subcsribers to the WHATWG mailing list, I mean

[17:13:03.0000] <MikeSmith>
Hixie: people who submit more feedback should get their names in a bigger point size

[17:13:04.0000] <MikeSmith>
like tag clouds

[17:14:00.0000] <Hixie>
MikeSmith: hah

[17:14:01.0000] <Hixie>
Niictar24: 950 last i checked

[17:14:02.0000] <Niictar24>
And coloured, too!

[17:14:03.0000] <Dashiva>
What if you gave points per submission, and people could use those points to pimp their name in a custom way

[17:15:00.0000] <Lachy>
Hixie, it's really nice when you win one of those self-created and self-awarded awards. It'll make you feel really proud.

[17:15:01.0000] <Hixie>
help vote this up so we get more input :-) http://www.reddit.com/r/technology/comments/89lgh/get_your_name_in_the_html5_acknowledgements_help/

[17:15:02.0000] <Hixie>
Dashiva: that'd be hilarious

[17:16:00.0000] <Niictar24>
Hixie: When the PDF references page numbers, were those generated dynamically?

[17:16:01.0000] <Hixie>
yes

[17:16:02.0000] <Niictar24>
Ok

[17:22:00.0000] <Lachy>
I hope other ISPs in the US don't follow Time Warner's lead and start introducing ridiculously low bandwidth caps of 40GB/month

[17:22:01.0000] <MikeSmith>
/me wants to read rsayre whatwg blog posting

[17:22:02.0000] <Lachy>
I thought I left that nonsense behind in Australia. Not looking forward to being stuck with that again when I move to the US

[17:22:03.0000] <MikeSmith>
Lachy: you moving to the US?

[17:23:00.0000] <Lachy>
yes

[17:23:01.0000] <MikeSmith>
where to?

[17:23:02.0000] <Lachy>
I'm transferring to the Mountain View office

[17:23:03.0000] <Lachy>
I thought I'd told everyone that already.

[17:24:00.0000] <MikeSmith>
Lachy: you didn't tell me. I feel left out

[17:24:01.0000] <MikeSmith>
hmm, maybe annevk42 told me earlier

[17:24:02.0000] <MikeSmith>
Lachy: congrats on that

[17:24:03.0000] <MikeSmith>
great place in the world to work

[17:24:04.0000] <MikeSmith>
Tatsuki is great too

[17:25:00.0000] <Lachy>
where's Tatsuki?

[17:25:01.0000] <MikeSmith>
the head of the office there

[17:25:02.0000] <Hixie>
lachy will presumably become opera's rep to all the clandestine whatwg cabal meetings now

[17:25:03.0000] <jcranmer>
/me has to keep track of too many cabals these days

[17:26:00.0000] <MikeSmith>
Lachy: or head of the consumer business at least

[17:26:01.0000] <MikeSmith>
he used to run the Japan office

[17:26:02.0000] <MikeSmith>
great guy, from Kyushu. crazy about outdoors, hiking and stuff

[17:26:03.0000] <Lachy>
MikeSmith, oh, I thought you meant Tatsuki was a place, not a person. That's a weird name.

[17:26:04.0000] <MikeSmith>
normal name in Japan

[17:26:05.0000] <MikeSmith>
I hear there's a few more people there now too

[17:27:00.0000] <Lachy>
sure, but Japan isn't a normal place. :-)

[17:27:01.0000] <MikeSmith>
Opera in Mountain View, I mean

[17:27:02.0000] <Dashiva>
uh-oh, incoming rathole

[17:27:03.0000] <MikeSmith>
Lachy: you should come to Japan for a working visit when you can

[17:27:04.0000] <Lachy>
I first need to make up some bogus reason for me to be there

[17:28:00.0000] <Lachy>
or a real reason, if there is one

[17:28:01.0000] <Dashiva>
If you find one, let me know

[17:29:00.0000] <MikeSmith>
a real reason is that you can learn about the part of the business that involves product dev for getting browsers preinstalled on millions of devices

[17:29:01.0000] <MikeSmith>
dev, testing, deployment for that

[17:30:00.0000] <MikeSmith>
customer requirements and functional-spec development for that

[17:32:00.0000] <blooberry>
for as close as I am, I hardly ever visit the Opera Mt.View office

[17:33:00.0000] <MikeSmith>
blooberry: did you and kaz ever meet up much?

[17:34:00.0000] <MikeSmith>
I guess he's moved elsewhere now

[17:34:01.0000] <MikeSmith>
blooberry: you are in Seattle area, right?

[17:34:02.0000] <Hixie>
man, abarth has the patience of a saint

[17:34:03.0000] <Lachy>
Hixie, is abarth still debating with Roy?

[17:34:04.0000] <MikeSmith>
Hixie: indeed

[17:35:00.0000] <blooberry>
mikesmith: we meet up as often as we can, but he's moved out of the area. But I live in the SF bay area

[17:35:01.0000] <MikeSmith>
blooberry: ah, OK

[17:35:02.0000] <Hixie>
blooberry should move up to the peninsula

[17:35:03.0000] <MikeSmith>
Lachy: lots of fun recently on http list again

[17:35:04.0000] <blooberry>
MikeSmith: kaz is in Japan right now even

[17:35:05.0000] <MikeSmith>
blooberry: yeah, I ran into him last week

[17:36:00.0000] <blooberry>
way? cool. 8-}

[17:37:00.0000] <blooberry>
hope you don't have too many bruises. ;-}

[17:37:01.0000] <MikeSmith>
heh

[17:37:02.0000] <Niictar24>
Hixie: can IRC be used to point out possible corrections? Or should it be sent to your email address or to the WHATWG mailing list?

[17:38:00.0000] <Hixie>
Niictar24: any of the places mentioned on the blog post would be fine, IRC will lead to the comments being lost though

[17:38:01.0000] <Hixie>
Niictar24: unless you can get someone (you maybe!) to volunteer to track IRC comments and send them as e-mail each day!

[17:38:02.0000] <Hixie>
Niictar24: which would be awesome btw!

[17:39:00.0000] <Hixie>
Niictar24: especially since anne and zcorpan keep insisting on sending IRC comments and grumble when i ask them to send mail :-)

[17:39:01.0000] <Niictar24>
lol

[17:39:02.0000] <MikeSmith>
need bot

[17:39:03.0000] <Niictar24>
I thought you had all the time in the world to read the IRC chat logs

[17:40:00.0000] <Hixie>
i read them, but i don't do anything with them and usually i'm in the middle of edits when i see the comments

[17:40:01.0000] <Hixie>
so i can't deal with them straight away

[17:41:00.0000] <Lachy>
Hixie, it would be nice if there was an IRC bot that could help track IRC comments for you

[17:42:00.0000] <MikeSmith>
Niictar24: one way that some people use is to first discuss on IRC, then just copy-paste the IRC discussion into an e-mail message or new bugzilla issue

[17:42:01.0000] <Niictar24>
Fair enough

[17:43:00.0000] <Niictar24>
So anyway, the example on page 170 of the letter format PDF for the <nav> element could be cleaned up I think

[17:44:00.0000] <Hixie>
Lachy: that would be cool too

[17:44:01.0000] <Hixie>
Lachy: find a volunteer to run one :-)

[17:44:02.0000] <jcranmer>
well, I have some python code for a dynamically-reconfigurable bot

[17:45:00.0000] <Niictar24>
But its just clean up so I'll just email it away

[17:46:00.0000] <Philip`>
I'm guessing PDF page numbers aren't very useful, since they'll change after every edit

[17:47:00.0000] <Philip`>
Lachy: Write a script to scrape krijnh's logs once a day and look for specially-marked comments

[17:47:01.0000] <Niictar24>
Actually, yea

[17:47:02.0000] <Niictar24>
What's the best way to reference something?

[17:48:00.0000] <jcranmer>
see!

[17:48:01.0000] <jcranmer>
oh, right

[17:48:02.0000] <Philip`>
Niictar24: Maybe the section title and a quote

[17:48:03.0000] <jcranmer>
the quit message is broken

[17:48:04.0000] <Hixie>
Niictar24: copy and paste text from that area, along with the section title

[17:48:05.0000] <Hixie>
what Philip` said

[17:49:00.0000] <Philip`>
What Hixie said

[17:49:01.0000] <Niictar24>
Ha

[17:49:02.0000] <Niictar24>
Kay

[17:50:00.0000] <Niictar24>
On another sort of half-way-related note, the image/figure/bubble kind of thing under "3.4.1 Kinds of content" is just a black blob in Acrobat Reader 7

[17:51:00.0000] <jcranmer>
/me wonders how acceptable python's XML/DOM stuff is with whatwg specification

[17:51:01.0000] <Hixie>
Niictar24: prince might not support SVG properly

[17:51:02.0000] <Niictar24>
Somehow I could imagine that one being likely

[17:51:03.0000] <Niictar24>
/me wishes SVG was fully supported everywhere

[17:52:00.0000] <Niictar24>
Which is silly, I know

[17:54:00.0000] <Philip`>
Even in Lynx?

[17:54:01.0000] <jcranmer>
duh

[17:54:02.0000] <jcranmer>
who doesn't want SVG in Lynx?

[17:55:00.0000] <jcranmer>
"Lynx... supporting SVG in a terminal before IE supported it!"

[17:56:00.0000] <heycam>
hmm does cairo have an aalib backend?

[17:58:00.0000] <Niictar24>
Sigh, I keep forgetting about that one

[17:58:01.0000] <jcranmer>
how hard could it be?

[17:59:00.0000] <Niictar24>
Someone should be creative and make it happen. Convert the SVG image to ASCII art. Simple!

[18:00:00.0000] <Niictar24>
Completely practical, too

[18:00:01.0000] <jcranmer>
apparently

[18:01:00.0000] <jcranmer>
someone was actually working on it

[18:01:01.0000] <Niictar24>
What? Really?

[18:02:00.0000] <jcranmer>
judging from the mailing lists, yes

[18:02:01.0000] <MikeSmith>
lynx is a pager, not a browser

[18:03:00.0000] <Niictar24>
"Pager"?

[18:04:00.0000] <MikeSmith>
Niictar24: like more(1) or less(1)

[18:09:00.0000] <heycam>
but you can press enter to follow a link, is that not browsing?

[18:09:01.0000] <heycam>
i like that w3m lets me click on links in my xterm

[18:18:00.0000] <olliej>
Hixie: i've been thinking about the cookies, etc issue chrome folk brought up

[18:18:01.0000] <olliej>
Hixie: i'm not sure why the spec needs to change

[18:18:02.0000] <olliej>
Hixie: the correct behaviour is for the browser to do the right thing

[18:19:00.0000] <olliej>
Hixie: and make sites on the same domain not run javascript concurrently

[18:19:01.0000] <olliej>
Hixie: that avoids spec changes, it doesn't add artificial complexity to all webpages to deal with one browser

[18:20:00.0000] <MikeSmith>
heycam: it's a kind of browsing, but not lacking

[18:21:00.0000] <MikeSmith>
so maybe we need another word

[18:21:01.0000] <MikeSmith>
subbrowser

[18:21:02.0000] <MikeSmith>
like subgenius

[18:21:03.0000] <MikeSmith>
s/but not lacking/but lacking/

[18:22:00.0000] <MikeSmith>
webpager

[18:23:00.0000] <MikeSmith>
heycam: vim has a way to follow links too

[18:23:01.0000] <heycam>
with a plugin?

[18:23:02.0000] <MikeSmith>
heycam: nah, I mean in :help stuff

[18:23:03.0000] <MikeSmith>
^] or whatever it is

[18:23:04.0000] <heycam>
yeah

[18:24:00.0000] <heycam>
seems to be a help file syntax specific thing tho

[18:24:01.0000] <heycam>
not sure i would class vim as a webpager from that :)

[18:24:02.0000] <MikeSmith>
true

[18:24:03.0000] <MikeSmith>
lynx doesn't even implement the DOM, right?

[18:25:00.0000] <heycam>
:r! lynx -dump http://www.google.com/

[18:25:01.0000] <heycam>
no i don't think it does script

[18:25:02.0000] <MikeSmith>
I don't see how it could rightly do CSS to any degree eiterh

[18:25:03.0000] <heycam>
w3m does reasonable table layout

[18:26:00.0000] <heycam>
don't know if it looks at css tho

[18:26:01.0000] <MikeSmith>
I don't mean just the rendering, but internally how it could deal with managing it

[18:26:02.0000] <MikeSmith>
I use w3m (and w3mmee) and elinks and lynx a lot

[18:27:00.0000] <MikeSmith>
but calling them browsers is like calling Amaya a browser

[18:27:01.0000] <heycam>
:)

[18:29:00.0000] <dave_levin>
olliej: ping

[18:29:01.0000] <MikeSmith>
[[

[18:29:02.0000] <MikeSmith>
First Murderer: We are men, my liege.

[18:29:03.0000] <MikeSmith>
Macbeth: Ay, in the catalogue ye go for men,

[18:29:04.0000] <MikeSmith>
As hounds and greyhounds, mongrels, spaniels, curs,

[18:29:05.0000] <MikeSmith>
Shoughs, water-rugs, and demi-wolves are clipt

[18:29:06.0000] <MikeSmith>
All by the name of dogs. 

[18:29:07.0000] <olliej>
dave_levin: pong

[18:29:08.0000] <MikeSmith>
]]

[18:30:00.0000] <dave_levin>
olliej: I was just reading the cookie stuff.  Is the problem there for any mutlithreaded UA?

[18:30:01.0000] <olliej>
dave_levin: only if the multithread UA doesn't actually put any effort into ensuring correct semantics

[18:30:02.0000] <olliej>
dave_levin: multiple sites on the same domain cannot run js concurrently

[18:31:00.0000] <dave_levin>
olliej: But a worker can run concurrently with the page and I thought that was the problem.

[18:31:01.0000] <olliej>
dave_levin: workers are a special class of problem

[18:31:02.0000] <olliej>
dave_levin: and cookie access can be controlled

[18:32:00.0000] <olliej>
dave_levin: standard js running in the main page context doesn't have any such luck

[18:33:00.0000] <olliej>
dave_levin: i think it would be easy to get the correct semantics

[18:33:01.0000] <olliej>
without having pages on the same domain running synchronously all the time

[18:34:00.0000] <olliej>
dave_levin: basically when js references the cookie list (or local storage, etc) that js takes a lock that is not released until js has finished executing

[18:34:01.0000] <dave_levin>
olliej: Ok, I haven't been following that very closely.  I had thought the issue came up due to trying to get workers to have access to the cookies.

[18:34:02.0000] <dave_levin>
olliej: And the workers may execute for a long time (like a ray tracer).

[18:35:00.0000] <olliej>
dave_levin: nope, workers are a different kettle of fish

[18:35:01.0000] <dave_levin>
olliej: thx for your time.  I was simply curious.

[18:35:02.0000] <olliej>
dave_levin: while there are isues, you can actually control access (as they don't just have access to the cookies from get go)

[18:41:00.0000] <dave_levin>
olliej: Oh I see. It came out of the worker discussion but then became this other thing. And you simply want the spec to be less demanding about the specifics of how this should be done.  Instead the spec should simply say what the guarantee is and leave it up to the UA to figure out how to satisfy that.

[18:41:01.0000] <olliej>
dave_levin: yes

[18:41:02.0000] <dave_levin>
olliej: fwiw, that sounds reasonable to me :)

[18:41:03.0000] <dave_levin>
olliej: But I not involved in that.

[18:42:00.0000] <olliej>
dave_levin: yeah, the proposal i saw earlier in the list before i just started ignoring it felt technically unsound

[19:45:00.0000] <Hixie>
olliej: all pages in a browsing context have to be in one serialisable scripting unit, and all pages in one origin have to be in one serialisable scripting unit.

[19:45:01.0000] <Hixie>
olliej: so it's more than just per-origin

[19:49:00.0000] <roc>
yeah

[19:49:01.0000] <roc>
but olliej is still right -)

[19:55:00.0000] <Hixie>
about what?

[19:59:00.0000] <roc>
about the big issue

[20:06:00.0000] <Hixie>
not sure i understand. are you agreeing with a particular proposal?

[20:06:01.0000] <Hixie>
i agree that concurrency is a big issue, that's why the spec has all this prose in it about making sure things like cookies and localstorage work :-)

[20:19:00.0000] <MikeSmith>
Hixie: does the current spec disallow blockquote as a descendant of header?

[20:19:01.0000] <MikeSmith>
and as a descendant of address?

[20:36:00.0000] <roc>
Hixie: I think what you have is great

[20:37:00.0000] <roc>
it seems some Chrome people aren't happy about it

[20:53:00.0000] <Hixie>
MikeSmith: not currently

[21:36:00.0000] <MikeSmith>
Hixie: did the spec previously disallow blockquote as a descendant of header?

[22:25:00.0000] <olliej>
roc_: i think S3.1 had local storage 

[22:25:01.0000] <roc_>
yeah?

[22:25:02.0000] <olliej>
but conceivably there could have been api changes since that original version

[22:25:03.0000] <olliej>
definitely had the sql db apis

[22:25:04.0000] <roc>
yeah

[22:25:05.0000] <olliej>
i may have 3.1 around

[22:25:06.0000] <olliej>
one mo

[22:26:00.0000] <roc>
the DB APIs are all async (aren't they?) and transactional so they may not be a problem

[22:27:00.0000] <roc>
apart from the orthogonal "what SQL dialect is it anyway" problem

[22:27:01.0000] <Hixie>
MikeSmith: i don't think so

[22:28:00.0000] <MikeSmith>
Hixie: OK

[22:28:01.0000] <olliej>
roc: hmmm, doesn't look like it, just sqldb

[22:28:02.0000] <Hixie>
roc: there is a sync DB API (as of today), for Workers, but that one is still transactional

[22:28:03.0000] <MikeSmith>
I'm wondering if hsivonen maybe added the blockquote constraints because of the outlining algorithm

[23:34:00.0000] <Hixie>
abarth keeps outclassing me

[23:56:00.0000] <MikeSmith>
Hixie: I like the part where other dude says "Your rants are getting irritating."

[01:27:00.0000] <hsivonen>
hmm. why don't simple element selectors in text/html match non-http://www.w3.org/1999/xhtml element in Opera?

[01:28:00.0000] <hsivonen>
do style sheets in text/html in Opera have an implicit http://www.w3.org/1999/xhtml @namespace rule?

[02:20:00.0000] <annevk42>
hsivonen, interesting find

[02:58:00.0000] <annevk42>
"But your rants are getting irritating." some guy addressing abarth

[02:58:01.0000] <annevk42>
wtf

[02:59:00.0000] <Hixie>
holy crap, the call for reviews is bringing in a lot of reviews!

[02:59:01.0000] <Hixie>
excellent!

[02:59:02.0000] <Hixie>
and now to bed!

[03:29:00.0000] <zcorpan>
hsivonen: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/59 - matches for me

[03:41:00.0000] <hsivonen>
zcorpan: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/60

[03:41:01.0000] <hsivonen>
I smell a magic list :-)

[03:42:00.0000] <annevk42>
whoa, that's bad

[03:44:00.0000] <annevk42>
hsivonen, if nobody does XLink for MathML maybe they can just zap it?

[03:45:00.0000] <annevk42>
hsivonen, I have the feeling they're considering that

[03:45:01.0000] <annevk42>
/me would prefer that tbh

[03:45:02.0000] <zcorpan>
i'm wondering whether it's feasible to use <a href> in the html namespace mixed with mathml

[03:46:00.0000] <annevk42>
that sounds like layout trouble

[03:46:01.0000] <annevk42>
though where MathML allows variables I think HTML content can be inserted

[03:46:02.0000] <jgraham>
href everywhere for mathML sounds appealing. But it sounds appealing in HTML too…

[03:47:00.0000] <zcorpan>
<a href> could either wrap the whole <math> element or be inside <mi> etc

[03:47:01.0000] <jgraham>
Although s/everywhere/a well defined set of places where it makes sense/

[03:47:02.0000] <zcorpan>
<a href> has the advantage that it works with legacy UAs (and search engines)

[03:48:00.0000] <zcorpan>
disadvantage is that it's less flexible -- you can't link an mrow etc

[03:48:01.0000] <zcorpan>
though you can't link <tr> in html, either

[03:48:02.0000] <jgraham>
Being able to link parts of expressions seems useful

[03:48:03.0000] <jgraham>
like larger than single tokens but smaller than the whole thing

[03:49:00.0000] <annevk42>
MathML doesn't really have the compat issue as it's not widely deployed yet

[03:50:00.0000] <jgraham>
Does it have the implementation issue? It seems like href everywhere is no worse than xlink

[03:50:01.0000] <jgraham>
Also, isn't there some layout funkiness in MathML that could break if you insert random extra elements in the DOM?

[03:53:00.0000] <annevk42>
XLink is mostly bad because it is not well defined and we do not really need the additional namespace

[03:57:00.0000] <annevk42>
I don't really see why MathML could not have an href= attribute on several elements

[03:58:00.0000] <annevk42>
other than maybe the inconsistency it gives with HTML and SVG

[04:04:00.0000] <annevk42>
oh, MathML did change back to id=""

[04:05:00.0000] <annevk42>
http://www.w3.org/TR/MathML3/appendixh.html still has some pointers to usage of xml:id="" though

[04:27:00.0000] <annevk42>
hsivonen, twitter creates the tinyurls

[04:27:01.0000] <Philip`>
<a href> presumably doesn't help much when MathML is being used outside of HTML

[04:30:00.0000] <hsivonen>
annevk42: the twitterific tweet character limit seems to look at whatever URL I paste there--not a tinyurl version of it

[04:35:00.0000] <annevk42>
hsivonen, that wasn't the question :)

[04:37:00.0000] <hsivonen>
annevk42: ok. still wondering about non-tinyurl.com short urls on twitter

[04:38:00.0000] <hsivonen>
seems like the SMS gateway is only useful as an excuse to keep the microblogging micro

[04:38:01.0000] <hsivonen>
I imagine it would be extremely annoying to actually receive tweets as SMS

[04:44:00.0000] <Philip`>
More useful for sending, probably

[04:45:00.0000] <Lachy>
hsivonen, it can also be expensive if you're on a plan that makes you pay to receive SMSs

[04:56:00.0000] <jgraham>
There are plans that may you pay to recieve SMS? That seems… open to abuse

[04:58:00.0000] <Lachy>
jgraham, AIUI, that's how the system commonly works in the USA

[04:58:01.0000] <Philip`>
I thought that was how it always worked in the US (and nowhere else)

[04:59:00.0000] <jgraham>
So, what happens if I send someone 10,000 SMS messages?

[04:59:01.0000] <Lachy>
I don't know of any other contries that have such a backwards system

[04:59:02.0000] <Philip`>
Then the phone company is laughing all the way to the bank

[05:00:00.0000] <Philip`>
(particularly since you have to pay to send too)

[05:01:00.0000] <jgraham>
Horray for free market capitalism

[05:01:01.0000] <Lachy>
the USA also makes you pay to receive calls on a mobile too

[05:02:00.0000] <hsivonen>
EU operators also laugh all the way to the bank when the customer is roaming

[05:03:00.0000] <hsivonen>
the Commission really should do something about data roaming, too

[05:03:01.0000] <Lachy>
hsivonen, excessive roaming costs seem to be charged worldwide, not just in the EU

[05:04:00.0000] <jgraham>
Paying to recieve calls is marginally less silly since you can lways choose not to answer the call

[05:04:01.0000] <jgraham>
(it is still pretty silly though)

[05:05:00.0000] <Lachy>
the only reasonable exception is for premium rate subscription services, cause if you're stupid enough to subscribe, then you deserve to pay

[05:06:00.0000] <Lachy>
but then, at least in Australia, they charge $0.55 AUD to receive, which is nearly triple the cost of sending a normal SMS

[05:07:00.0000] <Lachy>
(actually, they can charge more cause it's up to the service to set the fee)

[05:23:00.0000] <Philip`>
/me wonders how the cost per megabyte of SMSs compares to e.g. a floppy disk carried by a limousine

[05:32:00.0000] <hsivonen>
does the spec lack document.all definition or am I bad at reading?

[05:36:00.0000] <annevk42>
it lacks it

[05:36:01.0000] <hsivonen>
I filed a bug

[05:36:02.0000] <hsivonen>
as it happens, I broke document.all locally

[05:37:00.0000] <annevk42>
document.all is such a pain

[05:37:01.0000] <hsivonen>
it would be nice to know how it is supposed to work in order to unbreak it the right way

[05:37:02.0000] <hsivonen>
fixing Gecko to use the XHTML namespace in text/html breaks all sorts of interesting things

[05:39:00.0000] <hsivonen>
sigh. document.all is quirks-mode-only in Gecko but works in the standards mode in Opera and WebKit

[05:40:00.0000] <jgraham>
Philip`: It seems like a floppy disk full of SMS messages would be about 1000 pounds which is roughly 8 hours of limo hire. Unless I am wrong by a factor of 1000 somewhere

[05:42:00.0000] <hsivonen>
sigh. document.all isn't interoperable to begin with

[05:43:00.0000] <annevk42>
did you see what I just wrote on document.all? :)

[05:43:01.0000] <hsivonen>
Opera puts name=x in document.all["x"] but Gecko and WebKit don't

[05:44:00.0000] <hsivonen>
aaargh. IE8 XSS protection on live dom viewer strikes again

[05:45:00.0000] <hsivonen>
Hixie: could you turn the XSS protection off, please?

[05:47:00.0000] <hsivonen>
looks like IE8 agrees with Gecko and WebKit

[05:47:01.0000] <hsivonen>
/me decides Opera behavior is a bug here

[05:50:00.0000] <annevk42>
IE8 in quirks mode?

[05:50:01.0000] <hsivonen>
standards mode

[05:51:00.0000] <hsivonen>
quirks, too

[05:52:00.0000] <annevk42>
also for elements like <input name=x> or <img name=x>?

[05:58:00.0000] <hsivonen>
aargh. are those magic?

[05:58:01.0000] <hsivonen>
sigh.

[06:00:00.0000] <annevk42>
https://bugs.opera.com/browse/CORE-8133

[06:00:01.0000] <annevk42>
oops

[06:00:02.0000] <annevk42>
meant to copy http://www.whatwg.org/specs/web-apps/current-work/multipage/dom.html#dom-document-getelementsbyname

[06:00:03.0000] <annevk42>
which lists magic elements

[06:01:00.0000] <hsivonen>
thanks

[06:03:00.0000] <hsivonen>
ah. Gecko changed here earlier this month, so I was wrong to use the latest release as my baseline of sane behavior

[06:50:00.0000] <Philip`>
"Woohoo, I'm gonna be famous!" - who's going to break the news to them that that typo was already reported on Bugzilla? :-)

[06:51:00.0000] <Lachy>
Philip`, where is that quote from?

[06:52:00.0000] <Philip`>
Lachy: WHATWG list

[06:52:01.0000] <jgraham>
Philip`: I think you are focusing on the wrong misapprehension

[06:56:00.0000] <Dashiva>
/me wonders how many of those typos could be fixed with a spell checker

[07:01:00.0000] <Lachy>
Dashiva, http://www.w3.org/TR/html5/single-page/,spell

[07:03:00.0000] <Fyrd>
Philip`: That would have been me. Bah, knew that had to be too easy. :)

[07:03:01.0000] <Fyrd>
Guess I'll just have to work harder for immortality.

[07:04:00.0000] <jgraham>
Lachy: or, alternatively M-x flyspell-mode

[07:04:01.0000] <Lachy>
jgraham, what?

[07:05:00.0000] <jgraham>
Lachy: Inline spellchecking in Emacs. It seems more useful as a solution since it allows the errors to be corrected

[07:06:00.0000] <Lachy>
there's something weird about that W3C spell checker, since it's freezing Minefield

[07:11:00.0000] <Philip`>
Fyrd: You could also hope that Hixie gets confused by timezones and thinks your email was sent before the Bugzilla bug was filed

[07:13:00.0000] <annevk42>
that spell checker should ignore <code>

[07:14:00.0000] <Philip`>
That wouldn't help with typos in code comments

[07:14:01.0000] <annevk42>
grmbl

[07:14:02.0000] <annevk42>
I wrote getElementsByClassName

[07:14:03.0000] <annevk42>
hopefully people can read over that

[07:15:00.0000] <Fyrd>
Philip`: Ooh, yeah, maybe! :)

[07:15:01.0000] <Fyrd>
Actually I have some other errors up my sleeve.

[07:20:00.0000] <annevk42>
Fyrd, you'll now be known as the guy that tried to get his name in HTML5 by reporting a spelling error... and failed! :D

[07:21:00.0000] <Fyrd>
annevk42: :(

[07:21:01.0000] <Dashiva>
That's probably better fame-wise than being one among hundreds of acks in the spec

[07:21:02.0000] <Fyrd>
annevk42: Oh well, I'll take it.

[07:22:00.0000] <annevk42>
Fyrd, I can see how this might not be funny to you. Fortunately you were planning on doing more review :)

[07:24:00.0000] <Fyrd>
It is funny, I just like making sad faces.

[07:25:00.0000] <Fyrd>
Besides, I'm sure the XHTML2 group would be more than happy to receive my spellcheck powers! Ha!

[07:27:00.0000] <Philip`>
But they won't make you famous unless you join their WG

[07:28:00.0000] <jgraham>
You can get famous in the XHTML2 WG? I thought you languished in academic obscurity

[07:28:01.0000] <Fyrd>
Ah well. Never mind then.

[07:28:02.0000] <Philip`>
HTML5 is happy to give equal fame to cats as to browser developers

[07:28:03.0000] <Fyrd>
Maybe I should have my cats report typos then.

[07:29:00.0000] <Fyrd>
My time-travelling cats, that is.

[07:29:01.0000] <jgraham>
Yeah it true that the acknowledgements list is too long. Maybe I should train some cats to kill the other people on it

[07:29:02.0000] <Fyrd>
Anyway, back to work now. I guess I'll report future stuff in Bugzilla.

[07:31:00.0000] <Philip`>
I don't think there's much value in using Bugzilla rather than the list

[07:32:00.0000] <Philip`>
I suppose there might be a bit of value in checking both Bugzilla and the list for duplicates before posting, but that's not too important since we'll let Hixie deal with it :-)

[07:32:01.0000] <Philip`>
so it's probably best to just do whatever's easier

[07:32:02.0000] <Fyrd>
Well, at least I can try to be less redundant by seeing what's already been reported there.

[07:33:00.0000] <Fyrd>
Okay, thanks!

[07:35:00.0000] <annevk42>
hsivonen, where you put the dual tokens leaks in one way or another

[07:35:01.0000] <Lachy>
using bugzilla for reporting spec bugs is useful because it keeps the volume of traffic on the mailing list lower

[07:36:00.0000] <annevk42>
hsivonen, so the spec will have to dicate something

[07:36:01.0000] <Philip`>
Lachy: Why is lowering volume on the mailing list a goal?

[07:37:00.0000] <jgraham>
For typos it generates the same amount of mail

[07:37:01.0000] <Lachy>
because it means discussions like that on bug 6684 don't generate too much noise

[07:37:02.0000] <hsivonen>
annevk42: It doesn't leak if you make createElementNS lowercase and throw upon importing non-lowercase node from an XML doc

[07:38:00.0000] <hsivonen>
annevk42: but I'm not convinced that plugging the leak is worthwhile

[07:38:01.0000] <Lachy>
jgraham, that's because bugzilla has been misconfigured to send new bugs to public-html instead of just public-html-bugzilla

[07:39:00.0000] <Philip`>
Lachy: Almost the entire purpose of the mailing list is to have discussions, so it seems discussions should go there

[07:39:01.0000] <annevk42>
hsivonen, what exactly is your strategy?

[07:39:02.0000] <hsivonen>
annevk42: do you mean implementation strategy?

[07:39:03.0000] <Philip`>
(Bugzilla might be more useful for things there's no point in anyone except the editor reading and no need for discussion, like typos)

[07:40:00.0000] <Lachy>
Philip`, in principle, yet. But experience with public-html suggests that it's useful to try and keep as much unproductive discussion off the list as possible, so that the little bit of productive discussion that goes on there doesn't get drowned out in the noise

[07:41:00.0000] <annevk42>
hsivonen, what is the alternative to what I posted for getElementsByTagName?

[07:43:00.0000] <hsivonen>
annevk42: the alternative is that each element node has a pointer to an interned compound name struct (aka. node info in Gecko and qualified name in WebKit) and this struct has pointers to four interned items:

[07:43:01.0000] <hsivonen>
namespace, local name, prefix and ASCII-lowercased local name

[07:44:00.0000] <hsivonen>
then in XML, selectors preserve case and match against local name

[07:44:01.0000] <hsivonen>
and in HTML selectors are lower-case and match against ASCII-lowercased local name

[07:45:00.0000] <hsivonen>
making SVG-in-text/html nodes give the appearance of case-insensitive selector matching

[07:46:00.0000] <annevk42>
that's not just a different strategy, that gives different results

[07:46:01.0000] <hsivonen>
and the third alternative is using a single local name atoms on both sides and magic fixup lists that can't support textArea

[07:46:02.0000] <hsivonen>
annevk42: different results

[07:47:00.0000] <annevk42>
i think what bz said makes the most sense

[07:47:01.0000] <annevk42>
that's also what i said on IRC yesterday... :)

[07:47:02.0000] <hsivonen>
so, yeah I guess there's no way around the spec forcing which side you put the dual atoms on

[07:48:00.0000] <hsivonen>
annevk42: yesterday I didn't know that bz would be OK with HTMLness compare plus pointer compare instead of just pointer compare :-)

[07:48:01.0000] <Fyrd>
Hey, shouldn't this page: http://www.whatwg.org/specs/ include the newly separated specs? Or are they listed somewhere else?

[07:49:00.0000] <annevk42>
they moved to the W3C

[07:49:01.0000] <annevk42>
http://dev.w3.org/html5/

[07:51:00.0000] <Fyrd>
Ah, okay, thanks.

[09:24:00.0000] <gsnedders>
/me wonders about a !feedback bot

[10:52:00.0000] <gsnedders>
<h1>foo</h1><section><h1>bar</h1></section><p>where — what section is "where" associated with?

[11:40:00.0000] <jgraham>
gsnedders: Foo

[11:40:01.0000] <jgraham>
asasuming the spec makes sense

[11:42:00.0000] <gsnedders>
jgraham: That's what I'd logically expect. I'm not entirely convinced the spec makes sense though.

[11:46:00.0000] <Philip`>
/me is concerned about overuse of the terms "logic" and "sense"

[12:07:00.0000] <mpilgrim>
man, that HTTPbis thread is epic

[12:08:00.0000] <mpilgrim>
it's like arguing with a toddler

[12:08:01.0000] <mpilgrim>
"Would you like an apple or an orange?"

[12:08:02.0000] <mpilgrim>
"I want a pony!"

[12:08:03.0000] <mpilgrim>
"No. Would you like an apple, or would you like an orange?"

[12:08:04.0000] <mpilgrim>
"I want a pony!"

[12:08:05.0000] <mpilgrim>
"All right, I'll make the choice for you."

[12:08:06.0000] <mpilgrim>
"NOOOOOOO!"

[12:10:00.0000] <mpilgrim>
"Would you like browsers to behave according to a documented algorithm or an undocumented algorithm?"

[12:10:01.0000] <mpilgrim>
"I want a pony!"

[12:11:00.0000] <mpilgrim>
"No. Would you like browsers to behave according to a documented or undocumented algorithm?"

[12:11:01.0000] <mpilgrim>
"I want a pony!"

[12:11:02.0000] <mpilgrim>
"Would you like me to make the choice for you?"

[12:11:03.0000] <mpilgrim>
"NOOOOOOOO!"

[12:11:04.0000] <Dashiva>
It's a sneaky way to say "I want neither of those", maybe

[12:12:00.0000] <gavin_>
it seems to me that the most productive way forward is to push for Adam's option a) (remove content-type restrictions) rather than b) (add the sniffing algorithm to the HTTP spec)

[12:12:01.0000] <gavin_>
the argument that "http forbids us from doing this" isn't a very strong one in my mind

[12:12:02.0000] <gavin_>
it should be fixed, but it's hardly the critical problem some people are making it out to be

[12:13:00.0000] <mpilgrim>
I predict the actual solution will be c) HTTPbis continues to ignore reality, and HTML5 or some sub-spec will contain the phrase "The following algorithm is a willful violation of blah blah blah"

[12:14:00.0000] <gavin_>
I think an a) solution is possible with a small phrasing change... but maybe even that's being optimistic

[12:14:01.0000] <gavin_>
I don't think your option c) is a huge problem

[12:15:00.0000] <mpilgrim>
nor do i

[12:15:01.0000] <mpilgrim>
note that i am not actually a fan of content-sniffing

[12:15:02.0000] <mpilgrim>
at all

[12:15:03.0000] <mpilgrim>
but documented crap is better than undocumented crap

[12:15:04.0000] <gavin_>
indeed

[12:17:00.0000] <annevk2>
having it documented will hopefully discourage additional sniffing

[12:17:01.0000] <mpilgrim>
hahaha

[12:17:02.0000] <mpilgrim>
no, that won't happen either

[12:18:00.0000] <annevk2>
I'm still hoping :)

[12:18:01.0000] <mpilgrim>
there are whole new avenues of sniffing yet to come

[12:18:02.0000] <mpilgrim>
audio formats

[12:18:03.0000] <mpilgrim>
video formats

[12:19:00.0000] <annevk2>
fonts

[12:19:01.0000] <mpilgrim>
fonts, indeed

[12:19:02.0000] <annevk2>
admittedly media types suck

[12:20:00.0000] <jcranmer>
x-cabal/x-html

[12:21:00.0000] <annevk2>
doing the right thing for fonts (getting font/otf) is apparently very hard so we really went ahead with the draft

[12:21:01.0000] <annevk2>
so we never really*

[12:33:00.0000] <gsnedders>
/me predicts mpilgrim is right and solution c prevails… again

[12:34:00.0000] <mpilgrim>
somebody should add my prediction to http://wrongtomorrow.com/

[12:39:00.0000] <Philip`>
It's kind of hard to *not* implement content-sniffing for fonts when nobody's bothered registering a type, even if everyone has the best of intentions and doesn't want sniffing

[12:40:00.0000] <Philip`>
Someone should set up wrongtoday.com and put the HTTP spec on it

[12:40:01.0000] <annevk2>
yeah, we should have just forced people to use font/ttf or font/otf and then bother about registration later

[12:40:02.0000] <annevk2>
it's too late now though, I think

[12:41:00.0000] <Philip`>
The TTF/OTF distinction is a bit meaningless, which makes things confusing

[12:41:01.0000] <Philip`>
(since OTF is a backward-compatible extension to TTF, and can still use .ttf files)

[12:41:02.0000] <Philip`>
(unless you use the non-backward-compatible parts of OTF, in which case you can use .ttf or .otf)

[12:41:03.0000] <Philip`>
(I think)

[12:42:00.0000] <Philip`>
Hmm, that reminds me, http://fonts.philip.html5.org/ is still breaking the law

[12:43:00.0000] <Philip`>
since it sets the font names to "Subset of [original name]", and the Open Font License says you mustn't use any part of the original name for your new font, and I forgot to implement that

[12:43:01.0000] <Philip`>
but nobody has complained yet, so that's okay

[12:47:00.0000] <annevk2>
Philip`, the idea would be that they would both be handed to the same processor. I wasn't quite convinced by this part of the story but the other people involved in the discussion thought that having a distinct mapping for both .ttf and .otf was worth it even though in practice it would be meaningless.

[12:48:00.0000] <Philip`>
As long as you can use either extension and either content-type and it will work without needing to understand the distinction, that sounds fair enough

[14:05:00.0000] <Hixie>
http://www.w3.org/mid/ee0b30dd0904030511w559c76bey3a6ddb4a95b5d90c⊙mgc would be grounds for banning in the whatwg, i wonder what the w3c did about it

[14:06:00.0000] <Hixie>
(of course the e-mail it is replying to would be grounds for permanent banning on the grounds that it's a linkspam bot)

[14:07:00.0000] <Hixie>
(but that's another story)

[14:11:00.0000] <gavin_>
why would that be grounds for banning?

[14:12:00.0000] <Hixie>
oh i guess i misread it and didn't realise he was using the word "spammer" literally as opposed to as an insult

[14:12:01.0000] <Hixie>
never mind

[14:29:00.0000] <Philip`>
Is Ojan Vafai just trying to make clickjacking attacks far easier?

[14:30:00.0000] <ojan>
Philip`: how so?

[14:30:01.0000] <ojan>
Philip`: if you meant that as a serious question, the answer is no :)

[14:37:00.0000] <Philip`>
ojan: Because one of the difficulties with clickjacking is positioning and scrolling the target page so the link or button or whatever is in the right place, and your proposal was to provide a feature that lets you much more easily scroll the target page to wherever you want

[14:37:01.0000] <Philip`>
ojan: Also, it wasn't an entirely serious question ;-)

[14:46:00.0000] <ojan>
Philip`: the use case i gave is actually the one that inspired the idea. things like image search (e.g. MS image search or google image search) would work a lot better if they could scroll the image into view

[14:50:00.0000] <Hixie>
philip is right that it would make clickjacking easier

[14:50:01.0000] <Hixie>
but i think in practice it doesn't make it significantly easier

[14:50:02.0000] <Hixie>
in that it's already possible to "scroll" an iframe by just making it high enough

[15:31:00.0000] <jwalden>
Philip`: that's not exactly hard, is it?  iframe in an iframe with the outer of the two abspos-ing the inner one to produce exactly the pixels desired

[15:39:00.0000] <Philip`>
jwalden: It's harder to get decent cross-browser cross-platform support that way, since the absolute positioning will depend on fonts and layout and stuff

[15:39:01.0000] <jwalden>
eh...hitting the 2/3 case is good enough if that's what you're trying to do

[15:40:00.0000] <Philip`>
Not if you're using it to target a specific individual, and don't know what browser they're going to be using

[15:40:01.0000] <jwalden>
the money in attacks is in affecting a large audience

[15:40:02.0000] <jwalden>
anyway, it's like buffer overflows, once you have one you know you're pretty much hosed, even if it might take a lot of effort and potentially be machine-specific

[15:41:00.0000] <Philip`>
Not if you're tricking one person into clicking the "send one million dollars to this bank account" button on their online banking ystem

[15:41:01.0000] <jwalden>
or heap smashes

[15:41:02.0000] <Philip`>
s//s/

[15:42:00.0000] <Philip`>
Anyway, yes, if you're vulnerable to normal clickjacking then that should be fixed, and it doesn't matter if the attack is made easy, so this doesn't seem like a real problem :-)

[15:42:01.0000] <Philip`>
s/easy/easier/


2009-04-04
[17:33:00.0000] <sayrer>
"Mozilla screws another loyal customer"

[17:33:01.0000] <sayrer>
that's rich

[17:33:02.0000] <gavin_>
context?

[17:34:00.0000] <sayrer>
http://delicious.com/wearehugh

[17:36:00.0000] <sayrer>
I like mkaply's comment too

[17:36:01.0000] <sayrer>
"Can someone point me to one of these supposed malware Firefox Community Editions?"

[17:36:02.0000] <sayrer>
yes, supposed.

[17:39:00.0000] <sayrer>
I like that neither of them care how dishonest their arguments are

[17:39:01.0000] <sayrer>
or how hypocritical

[17:39:02.0000] <sayrer>
they just want to win

[04:58:00.0000] <Philip`>
http://ocoins.info/ - <span class="Z3988" title="ctx_ver=Z39.88-2004&amp;rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&amp;rft.issn=1045-4438"></span> - that doesn't look overwhelmingly elegant

[05:04:00.0000] <annevk2>
I like their 'Why "Z3988"?'

[05:04:01.0000] <annevk2>
It basically explains they do not read specs and just figure out what works from browsers.

[05:05:00.0000] <annevk2>
(CSS can work with punctation in class names, you just need to escape it properly.)

[05:14:00.0000] <Dashiva>
Why did they drop the -2004 though?

[05:35:00.0000] <gsnedders>
Dashiva: Because versioning in the root of evil

[05:42:00.0000] <Dashiva>
They didn't drop the versioning, though

[05:42:01.0000] <Dashiva>
They just moved it from the class name to elsewhere

[05:46:00.0000] <gsnedders>
Well it's hardly as if I read that document.

[08:30:00.0000] <gsnedders>
Hixie: What's the point in the 30% margin at the bottom of body?

[08:41:00.0000] <Philip`>
gsnedders: So that links to #foo will display foo at the top of the screen, even if that content is near the bottom of the page

[08:42:00.0000] <gsnedders>
But the final anchors are quite a way up!

[12:30:00.0000] <Hixie>
gsnedders: multipage version

[12:30:01.0000] <gsnedders>
Ah, that makes a bit more sense.

[14:00:00.0000] <annevk2>
http://twitter.com/aral/status/1452067278

[14:37:00.0000] <Philip`>
Calling it "HTML5" seemed more arrogant

[14:37:01.0000] <gsnedders>
Should it have been 2.0

[14:37:02.0000] <gsnedders>
*?

[14:41:00.0000] <Niictar24>
Maybe he thought the 1.0 suggested that this was the first attempt at "web applications" and anything before it didn't count?

[14:42:00.0000] <webben>
Flash 5 ;)


2009-04-05
[03:22:00.0000] <Hixie>
crap, misentered the url in the recent twitter

[03:22:01.0000] <Hixie>
hm.. how do i set up a redirect on wordpress...

[03:23:00.0000] <Lachy>
Hixie, .htaccess

[03:24:00.0000] <Hixie>
can you set one up? i think the blog is under your account

[03:24:01.0000] <Hixie>
oh hey

[03:24:02.0000] <Hixie>
it works already

[03:24:03.0000] <Hixie>
that's awesome

[03:24:04.0000] <Lachy>
do you mean this link http://blog.whatwg.org/?p=659 ?

[03:25:00.0000] <Hixie>
no i meant "http://blog.whatwg.org/help-us-review-htm", but it automatically forwards it already

[03:25:01.0000] <Hixie>
so no need

[03:26:00.0000] <Lachy>
ah, I looked at the Hixie twitter acct, not the whatwg one

[03:27:00.0000] <Hixie>
following discussions on twitter is night on impossible

[03:27:01.0000] <Hixie>
i've no idea wtf gsnedders and cwilso are talking about

[03:27:02.0000] <Hixie>
and i follow them both

[03:27:03.0000] <Lachy>
the way WP has been set up is to treat blog.whatwg.org/(.*) as the title, and then it just searches the db for the closest match

[03:27:04.0000] <Hixie>
so you'd think i'd be able to work it out

[03:27:05.0000] <Hixie>
Lachy: indeed

[03:28:00.0000] <Lachy>
I generally don't follow twitter conversations much

[03:28:01.0000] <Lachy>
I would follow more if I could get tircd working, but I can't get my IRC client to connect to it properly

[03:28:02.0000] <Hixie>
oh?

[03:29:00.0000] <Hixie>
what does it do?

[03:29:01.0000] <Hixie>
i don't see why x-chat would fail

[03:29:02.0000] <Hixie>
you're joining the #twitter channel?

[03:29:03.0000] <Hixie>
and sending it the password?

[03:30:00.0000] <Lachy>
yes. X-Chat just says "Connected. Now logging in...", but never proceeds any further.

[03:30:01.0000] <Hixie>
what does tircd's console say?

[03:31:00.0000] <Lachy>
it gives an error.  I will try now and see

[03:32:00.0000] <Lachy>
$ ./tircd.pl tircd.cfg

[03:32:01.0000] <Lachy>
[Sun Apr  5 11:36:23 2009] tircd 0.7 started, using config from: tircd.cfg.

[03:32:02.0000] <Lachy>
[Sun Apr  5 11:36:23 2009] Listening on: 127.0.0.1:6667.

[03:32:03.0000] <Lachy>
[Sun Apr  5 11:36:33 2009] 127.0.0.1 connected.

[03:32:04.0000] <Lachy>
[Sun Apr  5 11:36:51 2009] [Lachy] 127.0.0.1 disconnected.

[03:32:05.0000] <Lachy>
[Sun Apr  5 11:36:51 2009] [Lachy] storage_path is not set or is not writable, not saving configuration.

[03:32:06.0000] <Hixie>
i wonder why xchat is disconnecting

[03:33:00.0000] <Hixie>
you sure you got the right password?

[03:33:01.0000] <Lachy>
I disconnected it myself, cause otherwise it just sits there doing nothing forever

[03:33:02.0000] <Hixie>
oh

[03:33:03.0000] <Lachy>
I've left if go for several minutes on previous attempts with no luck

[03:33:04.0000] <Hixie>
after it says logging in, if you type /motd, or /join #twitter, what happens?

[03:34:00.0000] <gsnedders>
Hixie: fail.

[03:34:01.0000] <Hixie>
gsnedders: ?

[03:34:02.0000] <gsnedders>
Hixie: We're talking about what @karlpro said

[03:34:03.0000] <gsnedders>
But he says a lot so that might not help :)

[03:34:04.0000] <Hixie>
i figured that out, but couldn't work out what karl had said

[03:36:00.0000] <Lachy>
[11:39]  * - tircd Message of the Day -

[03:36:01.0000] <Lachy>
[11:39]  * - This code is uber alpha, if you got this far, consider yourself lucky

[03:36:02.0000] <Lachy>
[11:39]  * - 

[03:36:03.0000] <Lachy>
[11:39]  * - /join #twitter to get started!  

[03:36:04.0000] <Lachy>
[11:39]  * - 

[03:36:05.0000] <Lachy>
[11:39]  * - /stats to set configurable options.

[03:36:06.0000] <Lachy>
[11:39]  * - 

[03:36:07.0000] <Lachy>
[11:39]  * - Please submit bug reports here: http://code.google.com/p/tircd/issues/list  

[03:36:08.0000] <Lachy>
[11:39]  * - 

[03:36:09.0000] <Lachy>
[11:39]  * - Thanks for using tircd,

[03:36:10.0000] <Lachy>
[11:39]  * - 

[03:36:11.0000] <Lachy>
[11:39]  * - @tircd on twitter. <tircd⊙co>

[03:36:12.0000] <Lachy>
[11:39]  * End of /MOTD command.

[03:36:13.0000] <Hixie>
and did you try /join #twitter?

[03:36:14.0000] <Lachy>
[11:39]  * NICKSERV :Unable to send direct message.  Perhaps NICKSERV isn't following you?

[03:36:15.0000] <Lachy>
[11:39]  * Disconnected (Remote host closed socket).

[03:36:16.0000] <Lachy>
yes

[03:36:17.0000] <Hixie>
why are you sending a msg to NICKSERV?

[03:37:00.0000] <Hixie>
if you get the motd, as you do, that means you are connecting properly

[03:37:01.0000] <Lachy>
[Sun Apr  5 11:41:30 2009] 127.0.0.1 connected.

[03:37:02.0000] <Lachy>
Can't call method "http_code" on an undefined value at ./tircd.pl line 910.

[03:37:03.0000] <Lachy>
$

[03:37:04.0000] <Hixie>
o_O

[03:37:05.0000] <Hixie>
interesting

[03:37:06.0000] <Hixie>
/me looks at tircd.pl line 910

[03:38:00.0000] <Lachy>
I have x-chat setup to automatically send the password, and it must be trying /msg nickserv with it by default

[03:38:01.0000] <Hixie>
that's not the kind of password you should be sending

[03:38:02.0000] <Lachy>
that's how it works with freednode at least

[03:38:03.0000] <Hixie>
you should be sending an IRC password, not a nick services password

[03:39:00.0000] <Lachy>
oh, ok.

[03:39:01.0000] <Lachy>
now it works

[03:39:02.0000] <Hixie>
:-)

[03:40:00.0000] <jgraham>
/me is thoroughly confused about InCoreignContent and isn't sure if he is silly or if the spec is broken

[03:40:01.0000] <jgraham>
*InForeignContent

[03:40:02.0000] <Hixie>
spec could definitely be broken

[03:40:03.0000] <Lachy>
it still complains about storage_path not being writable. Maybe I need to set that to something different in the cfg file

[03:40:04.0000] <jgraham>
I think waht hsivonen says about making foreigness a flag makes sense

[03:40:05.0000] <Hixie>
Lachy: i just ignored that message

[03:41:00.0000] <Hixie>
jgraham: it would be helpful to see an example showing that the two ways of speccing it aren't black-box identical

[03:41:01.0000] <jgraham>
The case that html5lib chokes on at the mometn is like <svg><foreignContent><table></table></foreignContent><g></g>

[03:41:02.0000] <Lachy>
ok

[03:42:00.0000] <Hixie>
jgraham: send me a mail, i'll have a look at some point

[03:42:01.0000] <jgraham>
Because the <table> changes the insertion mode to InTable so the <g> is processed InBody, notInForeignContent

[03:42:02.0000] <Lachy>
now I just need to figure out how to make tircd.pl run automatically on startup, instead of requiring me to start it manually in terminal

[03:43:00.0000] <Hixie>
use a launchd script

[03:43:01.0000] <Lachy>
how?

[03:43:02.0000] <jgraham>
Hixie: You a mail or WHATWG or public-html or…

[03:43:03.0000] <jgraham>
Maybe I should start twittering bugs. It seems to be what all the cool kids are doing

[03:44:00.0000] <Hixie>
jgraham: whichever (direct mail if you want me to just look at it and say if it's a spec bug or not)

[03:44:01.0000] <Hixie>
Lachy: hold o

[03:44:02.0000] <Hixie>
n

[03:45:00.0000] <Hixie>
Lachy: put a file in ~/Library/LaunchAgents

[03:45:01.0000] <Hixie>
google for the syntax

[03:45:02.0000] <Hixie>
i don't seem to have ever done an on-boot one

[03:46:00.0000] <Hixie>
only cronjob-like things

[03:46:01.0000] <Lachy>
ok. I'll look into that later

[03:47:00.0000] <Lachy>
right now, I have to organise my trip to London and Paris next week

[03:47:01.0000] <Hixie>
nn

[03:48:00.0000] <jgraham>
Hixie: Mail sent. Goodnight

[03:53:00.0000] <jgraham>
Lachy: Londaon and Paris? Now you're just showing off

[03:56:00.0000] <Lachy>
jgraham, the coach tour to Paris departs from London. But I wanted to go to London anyway to see my friends there

[03:56:01.0000] <jgraham>
Oh, OK if you're going on a coach tour it just makes you sound old :)

[03:59:00.0000] <Lachy>
the tour should be too full of old people. It's going to Euro Disney, so I expect families will be on it

[06:27:00.0000] <gsnedders>
I want to go out somewhere and read a book. Only issue: I can't decide where to go.

[06:34:00.0000] <Dashiva>
The library

[06:36:00.0000] <gsnedders>
That's indoors.

[06:36:01.0000] <gsnedders>
And I think shut.

[06:37:00.0000] <Dashiva>
The lawn in front of the library then

[06:37:01.0000] <gsnedders>
There isn't one.

[06:37:02.0000] <gsnedders>
n00b

[06:37:03.0000] <Dashiva>
The lawn L satisfying min(d(L,library))

[06:38:00.0000] <gsnedders>
http://maps.google.co.uk/maps?f=q&source=s_q&hl=en&geocode=&q=st+andrews,+fife&sll=53.800651,-4.064941&sspn=25.146196,39.550781&ie=UTF8&ll=56.3397,-2.795996&spn=0.001432,0.002414&t=h&z=19

[06:38:01.0000] <gsnedders>
The library is the building in the middle of that picture

[06:38:02.0000] <gavin_>
I see lawns on that map

[06:38:03.0000] <Dashiva>
Looks like a great lawn by the bus stop

[06:39:00.0000] <gsnedders>
Me too. But several are in church yard and aren't open.

[06:39:01.0000] <gsnedders>
That's a private garden.

[06:39:02.0000] <gavin_>
doesn't look very private to me

[06:39:03.0000] <gsnedders>
And the nearest lawn I could sit on <http://maps.google.co.uk/maps?f=q&source=s_q&hl=en&geocode=&q=st+andrews,+fife&sll=53.800651,-4.064941&sspn=25.146196,39.550781&ie=UTF8&ll=56.338778,-2.798131&spn=0.001432,0.002414&t=h&z=19> is technically school ground and thus depressing

[06:46:00.0000] <gsnedders>
/me heads off, intending to go to the beach, but maybe will just sit around on the Old Course isntead

[06:46:01.0000] <gsnedders>
*instead

[06:46:02.0000] <gsnedders>
(It is common ground on a Sunday)

[06:48:00.0000] <gsnedders>
(Legally it is common ground all the time, it's just golf is forbidden (by law) on a Sunday)

[12:21:00.0000] <Hixie>
jgraham: i assume you meant <foreignObject> not <foreignContent>

[12:24:00.0000] <Hixie>
hsivonen: i don't really understand how a flag for foreign content would work. Wouldn't it just be like having a second "state" which was either "foreign" or "whatever the secondary mode is", with the secondary mode being what the insertion mode is now?

[12:24:01.0000] <Hixie>
what's the benefit of doing it that way?

[12:25:00.0000] <Hixie>
it seems like you end up with an extra "if" statement per token

[12:27:00.0000] <gsnedders>
/me is failing to work out what to write as a final sentence of his Englsih dissertation

[12:28:00.0000] <Hixie>
i recommend not writing it like that sentence

[12:28:01.0000] <gsnedders>
That, on the whole, is a good idea.

[12:42:00.0000] <jgraham>
Hixie: Yes, I meant foreignObject 

[12:42:01.0000] <Hixie>
i replied to the e-mail

[12:43:00.0000] <jgraham>
Hixie: Yeah, I saw, thanks

[12:47:00.0000] <jgraham>
Tip for browser developers: if search can't find a string in a document, it probably also won't finda longer string for which that is a substring

[12:51:00.0000] <jgraham>
Woo. html5lib passes the tiny number of foreign content tests

[12:51:01.0000] <jgraham>
(it turned out that the least bug was a missing break statement)

[12:52:00.0000] <jgraham>
*last

[13:01:00.0000] <gsnedders>
This is horrible.

[13:02:00.0000] <gsnedders>
How can I end 4,500 words?

[13:03:00.0000] <svl>
"Rocks fall, everbody dies"?

[13:09:00.0000] <Philip`>
"And then I woke up."

[13:12:00.0000] <Dashiva>
Exeunt omnes

[13:12:01.0000] <jgraham>
QED

[13:13:00.0000] <jgraham>
gsnedders: What is the second-to-last sentence?

[13:14:00.0000] <gsnedders>
well, provided the absolute final paragraph doesn't become more than one sentence:

[13:14:01.0000] <gsnedders>
"The outcome of the other loves and lusts of Van are never given, though may well be equally morbid. Lucette's suicide does not stop Van and Ada's love: it brings them back together after their father banishing them from each other, for their love to continue unabated, with all the passion it had prior to their separation."

[13:15:00.0000] <Philip`>
You could replace the ':' with a '.', and then your second-to-last sentence would become two sentences and therefore would provide the last sentence.

[13:15:01.0000] <gsnedders>
That would only make a statement about one of the three texts though

[13:16:00.0000] <jgraham>
Alternatively go with "Which would never happen in reality, so proving how silly the whole thing is."

[13:17:00.0000] <jgraham>
(Actully it sounds like you could just stop it there. Unless you were tring to make some oher point in which case you should consider mentioning that)

[13:18:00.0000] <gsnedders>
This is why I dislike several hundred word conclusions: I never know how to conclude them.

[13:21:00.0000] <gsnedders>
I dislike what I just wrote as it results in a final page with one word.

[13:21:01.0000] <gsnedders>
"The destructiveness of the relationships within all three novels is ultimately caused by the idealistically perfect love that the protagonists thrive upon: in an age-old romantic novel, love would prevail; in these novels, all does not end well."

[13:22:00.0000] <gsnedders>
Right, now at 5,339 words.

[13:23:00.0000] <gsnedders>
(Limit is 4,500)

[13:39:00.0000] <noocx>
hi

[13:39:01.0000] <Philip`>
Hello

[13:39:02.0000] <noocx>
can you tell me where should i find ui experts who knows faceted classification?

[13:40:00.0000] <Philip`>
/me can't, since he has no idea what it is

[13:41:00.0000] <noocx>
http://flamenco.berkeley.edu

[13:41:01.0000] <Philip`>
This probably isn't the right channel for that kind of thing

[13:42:00.0000] <noocx>
yes, that's why i asked

[13:48:00.0000] <jgraham>
gsnedders: If you have to cut 1000 words, it seems silly to worry about the aesthetics of the final page layout at this stage :)

[13:48:01.0000] <gsnedders>
jgraham: That's why I'm not caring at this stage :)

[13:48:02.0000] <gsnedders>
jgraham: I also need to reset the entire thing in Times New Roman, 12pt, ragged right, for some reason best known to the SQA.

[13:48:03.0000] <jgraham>
noocx: why do you think this channel is the right place to find the people who do know abou that kind of thing?

[13:48:04.0000] <noocx>
jgraham: found on google:D

[13:48:05.0000] <jgraham>
gsnedders: Also: start cutting words and then reread the ending. Once you have condensed your essay and reduced the repetition, you you will have a much clearer idea of what your essay is about and how to summarise it

[13:48:06.0000] <jgraham>
noocx:Ironically if Google were perfect people presumably wouldn't be working on novel search interfaces

[13:48:07.0000] <jgraham>
/me waits for Philip` to point out that that isn't really true

[13:50:00.0000] <Philip`>
Are you assuming people never reinvent wheels?

[13:50:01.0000] <gsnedders>
But I thought all shapes already had a name!

[14:08:00.0000] <gsnedders>
/me sighs

[14:08:01.0000] <gsnedders>
getting my mother to use a computer is highly stressful


2009-04-06
[00:29:00.0000] <hsivonen>
Hixie: the foreign flag is a two-state separate variable that the tree builder checks for each start tag token before checking the instertion mode aka. 'secondary' insertion mode

[00:37:00.0000] <Hixie>
hsivonen: that seems like just an extra "if" statement per token

[00:37:01.0000] <Hixie>
hsivonen: what's the benefit?

[00:38:00.0000] <hsivonen>
Hixie: the benefit depends on how insertion modes have been implemented

[00:39:00.0000] <hsivonen>
Hixie: it seems to me that this eliminates a second method / function pointer-based method call per foreign start tag token

[00:41:00.0000] <hsivonen>
Hixie: I suppose on could eliminate the extra if by rolling the foreign behavior into the main mode switch and by changing the value of the variable to switch on and re-switching when falling back onto the secondary mode

[00:42:00.0000] <Hixie>
well i don't have an opinion on how you should implement it

[00:43:00.0000] <Hixie>
but i don't see why a separate switch is better for the way the spec describes it

[00:43:01.0000] <hsivonen>
was jgraham's infinite loop his bug or a spec bug?

[00:44:00.0000] <Hixie>
his bug

[00:44:01.0000] <Hixie>
(and unrelated to the insertion mode)

[00:46:00.0000] <hsivonen>
ok

[01:30:00.0000] <hsivonen>
which WG should spec the behavior of XPath being matched against a DOM that is an 'HTML document' for the purposes of the flag defined in HTML 5?

[01:30:01.0000] <Hixie>
my solution would be "don't use XPath"

[01:30:02.0000] <hsivonen>
which WG should spec the behavior of the XSLT 'html' output mode in the case of in-browser DOM-to-DOM transform

[01:30:03.0000] <Hixie>
why is "use the infoset as defined for xml" not satisfactory?

[01:30:04.0000] <hsivonen>
Hixie: we still need a spec

[01:31:00.0000] <hsivonen>
Hixie: no

[01:31:01.0000] <Hixie>
"why" not "is" :-)

[01:32:00.0000] <hsivonen>
Hixie: for compatibility with existing content, when an XPath expression is used via document.evaluate, a name expression where the namespace is no namespace and local name is /local/,

[01:32:01.0000] <Hixie>
sounds like xpath would have to define this, if it needs changes to xpath semantics

[01:32:02.0000] <hsivonen>
must evaluate to true when compared against a DOM element node name whose namespace is http://www.w3.org/1999/xhtml and local name is /local/

[01:33:00.0000] <hsivonen>
Hixie: it would be nice to have an informative note about this in HTML 5

[01:34:00.0000] <Hixie>
tell me what it should say in a bug

[01:34:01.0000] <hsivonen>
Hixie: I'll file bugs on XSLT, XPath and HTML 5.

[01:34:02.0000] <Hixie>
i know only enough about xpath and xslt to know to avoid them

[01:34:03.0000] <hsivonen>
Hixie: well, that doesn't help me when just following HTML 5 would break mochitests and existing content

[01:35:00.0000] <Hixie>
just remove the relevant features :-)

[01:38:00.0000] <hsivonen>
ooh. there's an XSLT 2.1 in the works

[01:38:01.0000] <hsivonen>
/me goes looking for public WDs

[01:39:00.0000] <hsivonen>
hrm. I can't find a public WD for XSLT 2.1.

[01:45:00.0000] <MikeSmith>
/me looks for XSLT 2.1 also

[01:49:00.0000] <hsivonen>
where does the XSLT 2.0 spec even define the serialization algorithms?

[01:51:00.0000] <MikeSmith>
hsivonen: seems that XSLT 2.1 draft is member-only at this point

[01:51:01.0000] <MikeSmith>
as is pretty much all their draft content

[01:52:00.0000] <hsivonen>
MikeSmith: thanks. I'll file bugs against the released specs, then.

[01:52:01.0000] <MikeSmith>
hsivonen: I find http://www.w3.org/TR/xslt-xquery-serialization/ 

[01:53:00.0000] <hsivonen>
If the WG wants to make it know to the general public whether their drafts have the same bugs, they can change the bugzilla fields

[01:53:01.0000] <hsivonen>
s/know/known/

[01:53:02.0000] <MikeSmith>
yeah

[01:53:03.0000] <hsivonen>
thanks. I didn't realize the serialization spec was separate.

[01:54:00.0000] <hsivonen>
I was performing random-access reads to XSLT 2.0 and missed normative references if there were any

[02:06:00.0000] <zcorpan>
hsivonen: maybe you should treat new html5 elements when validating html4 the same as legacy elements when validating as html5

[02:07:00.0000] <zcorpan>
hsivonen: i.e. whine about <article> being a new element but don't skip the subtree

[02:07:01.0000] <jgraham>
Hixie: Actually I need to check that the infinite loop isn't a spec bug because I think I have an if statement left that was supposed to prevent it but, without looking at the code, it seems like it should be unnecessary now

[02:07:02.0000] <hsivonen>
zcorpan: good idea!

[02:07:03.0000] <Hixie>
jgraham: is that the same bug you mailed me about?

[02:08:00.0000] <jgraham>
/me thinks some more and realises he's not sure it is unnecessary

[02:08:01.0000] <jgraham>
Hixie: No

[02:08:02.0000] <Hixie>
ah ok

[02:08:03.0000] <Hixie>
so what is the if statement in question?

[02:09:00.0000] <jgraham>
Hixie: It was about getting a <svg> or <math> element processed as if In Body when In Foreign Content

[02:09:01.0000] <jgraham>
Which I think sets the secondard insertion mode to In Foreign Content

[02:09:02.0000] <Hixie>
is there a sample markup snippet demonstrating what you mean?

[02:10:00.0000] <jgraham>
Hixie: Probably something like <svg><foreignObject><math> 

[02:10:01.0000] <jgraham>
iirc

[02:10:02.0000] <jgraham>
(maybe you need something else)

[02:10:03.0000] <jgraham>
/me goes to look at the code

[02:10:04.0000] <zcorpan>
/me filed http://bugzilla.validator.nu/show_bug.cgi?id=472

[02:11:00.0000] <hsivonen>
thanks!

[02:11:01.0000] <Hixie>
jgraham: i don't see why that would be a problem

[02:13:00.0000] <jgraham>
Hixie: The problem I was having, which may or may not have just been a bug, was that both the insertion mode and the secondary insertion mode were In Foreign Content, so hen you hit something that tried to switch the insertion mode to the secondary insertion mode and reprocess the token an infinite loop occurred

[02:14:00.0000] <jgraham>
/me realises he doesn't have the right tree on this computer, gives up

[02:15:00.0000] <Hixie>
oh yeah

[02:15:01.0000] <MikeSmith>
hsivonen: about bug 472, the obsolete-element checking for HTML5 is just in the Assertions.java code, right?

[02:15:02.0000] <hsivonen>
MikeSmith: yes. and legacy.rnc makes them non-errors on the RNG layer

[02:15:03.0000] <Hixie>
jgraham: i guess we shouldn't change the insertion mode if it's already in foreign, for "math" and "svg" start tag tokens

[02:16:00.0000] <jgraham>
Hixie: I think that was my if statement :)

[02:18:00.0000] <Hixie>
jgraham: fixed

[02:18:01.0000] <MikeSmith>
hsivonen: so implementing a similar mechanism for warning about HTML5 elements in HTML4 checking would require adding a corresponding legacy.rnc with the HTML5 elements, and adding an Assertions.java for HTML4 checking?

[02:19:00.0000] <hsivonen>
MikeSmith: yes. although now the HTML 4-related assertions are in real .sch

[02:19:01.0000] <MikeSmith>
OK

[02:19:02.0000] <hsivonen>
MikeSmith: only because I haven't gotten around to reimplementing that .sch in Java

[02:20:00.0000] <MikeSmith>
hsivonen: I see. well, if you want, I can take a shot at implementing zcorpan's suggestion and send you a patch to review

[02:22:00.0000] <jgraham>
Hixie: Thanks

[02:27:00.0000] <hsivonen>
MikeSmith: that would be nice

[02:29:00.0000] <MikeSmith>
hsivonen: OK. so this is the schema code in validator/schema/xhtml10/ , right?

[02:30:00.0000] <hsivonen>
MikeSmith: yes

[02:30:01.0000] <MikeSmith>
OK

[02:30:02.0000] <hsivonen>
I've now filed 4 bugs: one on XSLT and XPath each and two on HTML 5 about making notes pointing out the same things.

[02:32:00.0000] <MikeSmith>
zcorpan: btw, about http://bugzilla.validator.nu/show_bug.cgi?id=467

[02:32:01.0000] <zcorpan>
MikeSmith: yes?

[02:35:00.0000] <MikeSmith>
zcorpan: do you have tests that are not already included in the automated set of tests that validator.nu runs when it builds the "test" target?

[02:36:00.0000] <MikeSmith>
or is it that the tests need to be incorporated in the JSON mechanism that's there?

[02:36:01.0000] <zcorpan>
MikeSmith: i don't know which tests are run

[02:36:02.0000] <MikeSmith>
OK

[02:36:03.0000] <hsivonen>
MikeSmith: the JSON db is not included in the 'test' run yet

[02:36:04.0000] <MikeSmith>
ah

[02:36:05.0000] <hsivonen>
MikeSmith: 'test' runs fantasai's tests

[02:36:06.0000] <zcorpan>
MikeSmith: i'm not aware of any validator-specific tests i have around other than those at http://simon.html5.org/test/validator/

[02:36:07.0000] <MikeSmith>
hsivonen: I see

[02:38:00.0000] <Hixie>
sicking: document.write() when not in a parser-triggered <script> always implied a document.open()

[02:38:01.0000] <hsivonen>
(also, when I was developing Web Forms 2.0 validation, 'test' ran a modified version of annevk's Web Forms 2.0 suite, but I can't distribute those files.)

[02:38:02.0000] <Hixie>
sicking: there is never a race condition that might insert data in a random place in the input stream

[02:38:03.0000] <Hixie>
sicking: (this matches implementations, and is, i thought, clear in the spec)

[02:39:00.0000] <MikeSmith>
zcorpan: I see now. I was thinking you had more there, under the content-model subdir

[02:39:01.0000] <hsivonen>
Hixie: does it match Gecko/WebKit or only IE/Opera?

[02:39:02.0000] <Hixie>
everyone as far as i know

[02:39:03.0000] <hsivonen>
Hixie: ok. my recollection is different, but I agree with your point.

[02:39:04.0000] <Hixie>
there are occasional cases where you can trick some browsers into inserting text in arbitrary places in the input stream, but those seem to be rare bugs

[02:41:00.0000] <MikeSmith>
hsivonen, zcorpan: anyway, I see now that the purpose of the json thing is to allow testing of remote documents, right?

[02:41:01.0000] <Hixie>
bed time

[02:41:02.0000] <Hixie>
nn

[02:42:00.0000] <zcorpan>
MikeSmith: yep

[02:43:00.0000] <MikeSmith>
zcorpan: OK. and it requires that v.nu report only a single error message?

[02:43:01.0000] <hsivonen>
MikeSmith: no, the harness only compares the first message from remote against the first message from db

[02:43:02.0000] <MikeSmith>
OK

[02:44:00.0000] <hsivonen>
MikeSmith: http://wiki.whatwg.org/wiki/Validator.nu_Full-Stack_Tests

[02:44:01.0000] <MikeSmith>
thanks

[02:46:00.0000] <zcorpan>
MikeSmith: after having fixed a bug, you can do python validator-tester.py adduri url-of-test

[02:47:00.0000] <MikeSmith>
oh, cool

[03:10:00.0000] <MikeSmith>
hsivonen: minor problem with this test:

[03:11:00.0000] <MikeSmith>
http://hsivonen.iki.fi/test/moz/unescaped-ampersand.html

[03:13:00.0000] <MikeSmith>
failing because db say to expect the error to start at column 7, but v.nu reports it at column 4

[03:13:01.0000] <MikeSmith>
should I go ahead and update it before adding other tests?

[03:18:00.0000] <MikeSmith>
hsivonen: hmm, message output for that case no longer matches either

[03:18:01.0000] <MikeSmith>
it was:

[03:18:02.0000] <MikeSmith>
Text after \u201c&\u201d did not match an entity name. Probable cause: \u201c&\u201d should have been escaped as \u201c&amp;\u201d.

[03:19:00.0000] <MikeSmith>
but v.nu now says:

[03:19:01.0000] <MikeSmith>
u201c&\u201d did not start a character reference. (\u201c&\u201d probably should have been escaped as \u201c&amp;\u201d.

[03:21:00.0000] <MikeSmith>
hsivonen: current result seems correct, so should I just update the db.json to that?

[03:22:00.0000] <hsivonen>
MikeSmith: please do

[03:22:01.0000] <hsivonen>
MikeSmith: the test harness doesn't care about the text of the error message

[03:36:00.0000] <MikeSmith>
hsivonen: OK

[03:38:00.0000] <MikeSmith>
I'll check that in and other the other of zcorpan tests that aren't in there yet

[03:39:00.0000] <hsivonen>
Can an XPath expression compiler tell from context if a given name expression will match against element names or against attribute names?

[04:12:00.0000] <Lachy>
looks like there's still some confusion about elements vs. tags. http://blog.whatwg.org/help-us-review-html5#comment-40489 I wonder what can be done to clarify the spec, even though it's already technically correct as is

[04:20:00.0000] <beowulf>
Lachy: i think i undertand elements vrs tags but i don't understand those two parts of the spec as quoted in that comment...

[04:21:00.0000] <Lachy>
beowulf, why is it hard to understand that the root element is still present in the document, even though its tags may be omitted from the serialisation?

[04:22:00.0000] <beowulf>
because i'm an html author, not a ua author?

[04:23:00.0000] <beowulf>
but i get it now

[04:23:01.0000] <Philip`>
Depends on whether you view "the document" as a DOM or as a set of tags, I guess

[04:25:00.0000] <Lachy>
yeah, that can be a little confusing for authors since the exact meaning of "the document" depends on the context in which its used

[04:52:00.0000] <hsivonen>
did Web Apps inherit all the old DOM specs for maintenance purposes?

[05:06:00.0000] <annevk2>
hsivonen, sort of, yes

[05:07:00.0000] <hsivonen>
annevk2: what's the preferred means of reporting spec change requests?

[05:09:00.0000] <annevk2>
emailing public-webapps or www-dom

[05:09:01.0000] <hsivonen>
annevk2: thanks.

[05:09:02.0000] <hsivonen>
I guess I need to read up on the reuse of the www-dom list

[05:10:00.0000] <annevk2>
it's used for event stuff but has traditionally been the place for issues with dom core as well

[05:10:01.0000] <annevk2>
since nothing much is happening with dom core until simon's draft is ready it hasn't been used for that recently I suppose

[05:11:00.0000] <hsivonen>
annevk2: is anyone managing updates to DOM Level 3 XPath? 

[05:11:01.0000] <annevk2>
not really

[05:11:02.0000] <hsivonen>
annevk2: should I expect Web Apps to update that spec, or should I push for a delta-spec in the "APIs in HTML Documents" section of HTML5?

[05:12:00.0000] <annevk2>
there was an idea to standardize on an XSLT API as well but that hasn't happened either

[05:12:01.0000] <hsivonen>
does existing content pass HTML DOMs into such an API?

[05:12:02.0000] <annevk2>
DOM3XPath is nothing more than a note so I suspect webapps would be the best place

[05:12:03.0000] <hsivonen>
ok

[05:12:04.0000] <annevk2>
I'm not really familiar with the XSLT API

[05:13:00.0000] <hsivonen>
ok

[05:13:01.0000] <hsivonen>
thanks

[05:13:02.0000] <annevk2>
though I believe Gecko/Presto/WebKit have one in common and IE has another

[05:13:03.0000] <hsivonen>
having it in 3 browsers makes me suspect someone out there is already using it on HTML DOMs

[05:14:00.0000] <hsivonen>
so the XPath hack needs to apply to XPath in the browser in general--not just to document.evaluate

[05:14:01.0000] <zcorpan>
didn't google docs or something use xslt in the dom?

[05:16:00.0000] <zcorpan>
does someone have an opinion about Element.parentElement?

[05:16:01.0000] <zcorpan>
s/some/any/

[05:16:02.0000] <annevk2>
doesn't seem harmful

[05:17:00.0000] <annevk2>
and we have firstElementChild and such too...

[05:17:01.0000] <zcorpan>
yeah

[05:18:00.0000] <jgraham>
/me is against it

[05:19:00.0000] <jgraham>
Unless it becomes part of ElementTraversal

[05:19:01.0000] <jgraham>
Because it is totally redundant with parentNode

[05:19:02.0000] <annevk2>
why does it matter what it is part of?

[05:19:03.0000] <annevk2>
I think ElementTraversel should just be integrated in DOM Core fwiw

[05:19:04.0000] <jgraham>
annevk2: At the moment it is not part of anything at all, right?

[05:20:00.0000] <annevk2>
it's part of IE, WebKit and Opera

[05:20:01.0000] <jgraham>
What I really mean is that if it were in some spec I would accept it as inevitable

[05:20:02.0000] <jgraham>
annevk2: The fact that Firefox doesn't support it suggests that it is not needed for the web

[05:21:00.0000] <zcorpan>
nextElementSibling wasn't needed for the web, either

[05:21:01.0000] <jgraham>
AFAICT the only difference between parentNode and parentElement is that in one case you do while(node != document){} and in the other you do while(node != null){}

[05:22:00.0000] <jgraham>
nextElementSibling actually has a technical advantage for the common case where you want to iterate over elements but not other node types

[05:22:01.0000] <jgraham>
Although arguably it is redundant with jQuery

[05:22:02.0000] <zcorpan>
you can have elements in a DocumentFragment

[05:23:00.0000] <annevk2>
both array index and linked list also is redundant

[05:23:01.0000] <jgraham>
annevk2: ?

[05:24:00.0000] <annevk2>
I'm saying that redundancy is a "feature" when it comes to the DOM

[05:24:01.0000] <annevk2>
it's all over the place

[05:24:02.0000] <zcorpan>
let's add more!

[05:25:00.0000] <jgraham>
So your argument is that the DOM is so unweildy that we may as well add any extra feature because it only adds epsilon extra ugliness/confusion?

[05:26:00.0000] <zcorpan>
always doing whlie (n = n.parentElement) seems simpler than keeping track of what to check for in different cases

[05:27:00.0000] <jgraham>
I am toatlly unconvinced that the extra simplicity is worthwhile

[05:27:01.0000] <annevk2>
jgraham, I don't think parentElement is ugly or confusing

[05:27:02.0000] <jgraham>
annevk2: Do you think it is ugly or confusing to have two different methods that do almost exactly the same thing?

[05:28:00.0000] <jgraham>
(I agree that, in isolation, parentElement is not particularly bad)

[05:28:01.0000] <annevk2>
not necessarily, apparently :)

[05:28:02.0000] <zcorpan>
parentNode and parentElement have descriptive enough names to not be so confusing

[05:28:03.0000] <hsivonen>
parentElement makes sense for languages that don't do duck typing

[05:29:00.0000] <hsivonen>
it would make more sense in the Java DOM than in the JavaScript DOM

[05:29:01.0000] <jgraham>
hsivonen: Presumably the Java DOM has to support .parentNode anyway

[05:30:00.0000] <hsivonen>
jgraham: yes, but you need to cast it into Element explicitly to get at stuff that isn't on Node

[05:30:01.0000] <jgraham>
/me wishes we could stop designing APIs that are supposed to be programming-language neutral and just design good APIs for the web

[05:30:02.0000] <annevk2>
I think we're doing that now

[05:30:03.0000] <jgraham>
annevk2: Really?

[05:31:00.0000] <annevk2>
when we design new APIs, yes

[05:31:01.0000] <jgraham>
annevk2: I guess localStorage is a pretty js-friendly api

[05:31:02.0000] <annevk2>
see e.g. overloading of postMessage 

[05:32:00.0000] <hsivonen>
does WebKit implement XSLT as DOM-to-DOM like Gecko or as DOM-to-characters-to-DOM like IE?

[05:33:00.0000] <hsivonen>
Opera is like IE here, right?

[05:33:01.0000] <hsivonen>
(I wonder why, though. If Gecko can get away with DOM-to-DOM, why not Opera?)

[05:33:02.0000] <Philip`>
Loads of pages use parentElement

[05:34:00.0000] <jgraham>
Really? I wonder how Firefox gets away with not supporting it…

[05:34:01.0000] <hsivonen>
jgraham: yes, Firefox does DOM-to-DOM

[05:35:00.0000] <hsivonen>
jgraham: no disabling output escaping in the transform or document.write in the result

[05:35:01.0000] <annevk2>
hsivonen, jgraham meant parentElement

[05:35:02.0000] <hsivonen>
oh

[05:35:03.0000] <jgraham>
/me was rather confused :)

[05:35:04.0000] <annevk2>
hsivonen, I'm not sure why we do what we do with respect to XSLT

[05:35:05.0000] <Philip`>
http://www.runningstrong.com/ -

[05:35:06.0000] <Philip`>
  if (elem.parentNode) ... // IE5, IE6 and Netscape 6

[05:35:07.0000] <Philip`>
  else if (elem.parentElement && elem.parentElement.setAttribute) ... // IE4.

[05:35:08.0000] <Philip`>
  else // Netscape 4.6x or 4.7x

[05:35:09.0000] <Philip`>
    //alert("Must be Netscape! do nothing");

[05:36:00.0000] <Philip`>
(Lots of pages have that code)

[05:36:01.0000] <Philip`>
http://www.evo-solutions.com/ - firedobj=ns6? firedobj.parentNode : firedobj.parentElement

[05:37:00.0000] <Philip`>
http://villacapri.8m.com/ - while (EventTag.parentElement&&(EventTag.tagName!="A")){ EventTag=EventTag.parentElement; }

[05:39:00.0000] <Philip`>
http://www.ozteknik.com/ - if (TempJudge && SrcTecMenuIE ){ if (thisobj.parentElement.className=="SrcTecMenuClass"+thisobj.parentElement.id+"On") ... }

[05:39:01.0000] <Philip`>
and lots of others

[05:40:00.0000] <Philip`>
(In that last case, SrcTecMenuIE = (document.all))

[05:40:01.0000] <zcorpan>
hsivonen: i think opera does reparse because dom-to-dom broke various htmlness issues that a reparse fixed given our architecture

[05:41:00.0000] <hsivonen>
zcorpan: ah. DOM-to-DOM is magic in Gecko in its handling of local names

[05:42:00.0000] <hsivonen>
clearly, there would have been a need for a spec for this stuff, since browser differ radically in their implementation approach

[05:42:01.0000] <zcorpan>
Philip`: thanks. i guess that's reason enough to add it to web dom core

[05:42:02.0000] <Philip`>
zcorpan: Depends on whether you want to get the IE browser-sniffing path or the Netscape browser-sniffing path

[05:43:00.0000] <zcorpan>
Philip`: hmm

[05:45:00.0000] <annevk2>
the ones that check seem to check for parentNode first

[05:46:00.0000] <zcorpan>
var ns6=document.getElementById&&!document.all

[05:47:00.0000] <zcorpan>
rather they seem to check for other things than parentElement in their checks

[05:48:00.0000] <zcorpan>
so if you're not "netscape" you might get the parentElement code path

[05:49:00.0000] <annevk2>
why does http://www.whatwg.org/specs/web-apps/current-work/?slow-browser#named-access-on-the-window-object list the named elements twice?

[05:49:01.0000] <zcorpan>
http://www.ozteknik.com/ just uses it without checking

[05:50:00.0000] <hsivonen>
looks like we don't have interop here: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/64

[05:51:00.0000] <hsivonen>
but here we have interop: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/65

[05:51:01.0000] <zcorpan>
hsivonen: i think gecko is alone in disabling document.all in standards mode

[05:52:00.0000] <hsivonen>
oh. right. good point

[05:53:00.0000] <hsivonen>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/67

[05:53:01.0000] <hsivonen>
Gecko shows a different class compared to WebKit and Opera

[05:55:00.0000] <zcorpan>
ie8, opera and webkit all say [object HTMLCollection]

[05:57:00.0000] <zcorpan>
i guess parentElement should be on Node?

[05:57:01.0000] <annevk2>
is it on Node now?

[05:57:02.0000] <gavin_>
that "HTML document.all class" dates back all the way to implementation of "undetectable document.all" support

[05:58:00.0000] <gavin_>
/me can't find any justification for it

[05:58:01.0000] <gavin_>
bet we could change it!

[05:59:00.0000] <zcorpan>
hmm the live dom viewer save link doesn't work in ie8

[05:59:01.0000] <zcorpan>
it's on Node in webkit

[06:00:00.0000] <zcorpan>
HTMLElement in opera

[06:00:01.0000] <zcorpan>
Element in ie8

[06:01:00.0000] <zcorpan>
the ElementTraveral attributes are going to be on Node, right?

[06:02:00.0000] <annevk2>
maybe

[06:03:00.0000] <zcorpan>
i seem to remember feedback saying they should be on Node but that would be deferred to 2.0 because of willingness to get the spec to REC quickly

[06:08:00.0000] <jgraham>
/me accidentially reads the alt requirements again, realises, again, that they are works of pure fiction compared to what will actually happen

[06:16:00.0000] <Philip`>
jgraham: The requirements in the spec?

[06:16:01.0000] <jgraham>
Philip`: Yes.

[06:17:00.0000] <jgraham>
/me probably shouldn't say anything related to alt at all ever

[06:20:00.0000] <zcorpan>
http://simon.html5.org/specs/web-dom-core#dom-node-parentelement

[06:27:00.0000] <Philip`>
jgraham: I've generally given up trying to read the spec to work out what my alt text should be, because it seems to go into so much detail that I can't find any relevant advice

[06:29:00.0000] <zcorpan>
Philip`: would the spec be easier to understand if it said "Conformance requirements regarding the usage of the alt attribute are given in WCAG 2.0."?

[06:29:01.0000] <Lachy>
Philip`, is that because it doesn't cover your specific case, or because you can't figure out which cases apply to your images?

[06:29:02.0000] <Lachy>
zcorpan, no.

[06:29:03.0000] <Philip`>
zcorpan: No

[06:30:00.0000] <jgraham>
Philip`: I wasn't even trying to write alt text. A search just landed me there and I thought "if I had a dollar for every instance of a screenshot with a precise description of the screen contents on a real website, I would have exactly zero dollars"

[06:30:01.0000] <zcorpan>
please say so to pf people

[06:30:02.0000] <Philip`>
Lachy: I can't figure out whether it does cover my specific case

[06:31:00.0000] <Lachy>
Philip`, can you point to specific case of yours as an example?

[06:33:00.0000] <Philip`>
Lachy: There was http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2008-April/014446.html a while ago

[06:34:00.0000] <Philip`>
The spec has been updated, so maybe it now deals with those specific cases, but the alt description is eight pages long so I'm not going to bother reading it to check

[06:35:00.0000] <Philip`>
(I could cope much better with more general advice, like saying that pages should remain understandable if all the images are replaced by their alt text)

[06:35:01.0000] <zcorpan>
Philip`: you can check the toc to see which cases the spec lists

[06:37:00.0000] <zcorpan>
Philip`: it also has a section called General guidelines

[06:38:00.0000] <jgraham>
Yeah there could be a conformance requirement like "Pages must be understandable when images are not avaliable" and the alt attribute sould just say "the attribute provides alternative text for an image which is often required comply with the requirements in section x.y.z"

[06:38:01.0000] <jgraham>
That would annoy the pf people no end

[06:39:00.0000] <Philip`>
zcorpan: Having the general guidelines hidden underneath a dozen tedious case descriptions is not a good way to make me discover them :-)

[06:40:00.0000] <zcorpan>
Philip`: bug hixie to move them up

[06:41:00.0000] <zcorpan>
hmm http://msdn.microsoft.com/en-us/library/ms536654(VS.85).aspx

[06:42:00.0000] <Philip`>
Reading the TOC doesn't really help either - if I e.g. want to write a page describing a game that's decorated with some screenshots that link to full-size images when you click on them, the first one I noticed when randomly scanning the TOC was "4.8.2.1.6 A purely decorative image that doesn't add any information", which says to use alt="", which is wrong

[06:42:01.0000] <Lachy>
Philip`, specifically which case in that email?  The first one mentioned about the google logo now seems to be fairly well addressed in the spec

[06:43:00.0000] <Philip`>
Lachy: I don't know, I haven't bothered checking the latest version of the spec because I don't care enough :-p

[06:44:00.0000] <Lachy>
Philip`, ok. You're making it difficult to undrestand how to address your problems

[06:44:01.0000] <Lachy>
*understand

[06:45:00.0000] <jgraham>
Lachy: It seems pretty clear. Make that section of the spec shorter

[06:48:00.0000] <Philip`>
Lachy: Adding text to the spec to clarify the existing cases or cover new cases to address my problems would just make my higher-level problem (that I'd like to write acceptable alt text but the spec is too verbose and hard to read) worse

[06:50:00.0000] <Lachy>
making the spec detailed enough to cover all cases is in direct conflict with the desire to make it more concise

[06:51:00.0000] <Philip`>
It can cover all cases by being more general, rather than by being more detailed

[06:51:01.0000] <Lachy>
would it be better to instead make the spec easier to navigate and somehow helping people to more easily identify applicable cases

[06:52:00.0000] <Philip`>
That suggests that authors will write an <img>, think "ooh, I need some alt text", then open the HTML5 spec, navigate to the applicable case, read the instructions, and then write their alt text

[06:52:01.0000] <jgraham>
We could write a text-based adventure frontend to the spec

[06:52:02.0000] <Philip`>
which is horribly inefficient

[06:52:03.0000] <Lachy>
since I'm going to have to eventually explain this in the authoring guide, it would be good to figure out how to make that more approachable than the spec.

[06:52:04.0000] <Philip`>
and therefore unlikely to happen

[06:52:05.0000] <jgraham>
"You are in a maze of twisy reccomendations, all alike"

[06:53:00.0000] <jgraham>
"It is dark, you are likely to be eaten by a lawyer"

[06:53:01.0000] <beowulf>
can't you just decide to dump @alt? 

[06:55:00.0000] <Philip`>
It's more realistic if the author writes an <img>, thinks "ooh, I need some alt text", then remembers an approximation of the nice concise description of alt they read in the HTML5 spec six months ago, and then writes their alt text, and nobody points at them and laughs and says "you're a moron, section 4.8.2.1.11.6.9 specifically addresses this image and says you should write this specific text"

[06:55:01.0000] <zcorpan>
Lachy: copy the general guidelines section, then say "for more detail or suggestions for specific cases, refer to the spec"

[07:57:00.0000] <annevk2>
hsivonen has awoken the XML gods: http://twitter.com/elharo/statuses/1462873257

[07:59:00.0000] <hsivonen>
annevk2: too bad the removal of the Selectors special case introduced an XPath special case

[08:00:00.0000] <annevk2>
better XPath special case than special cases in the more commonly used features

[08:01:00.0000] <MikeSmith>
"The only thing they hold sacred are 10-year old browser bugs."

[08:01:01.0000] <MikeSmith>
beautiful

[08:01:02.0000] <MikeSmith>
we should start a quotations page

[08:01:03.0000] <hsivonen>
it's a DOM Level 2 / XHTML 1.0 issue, actually...

[08:02:00.0000] <hsivonen>
or Namespaces issue...

[08:02:01.0000] <annevk2>
yeah, it's clear from the accusation that he hasn't really followed what this is about

[08:02:02.0000] <MikeSmith>
we should do like Elijah, line up the priests of Baal and put them all to a test

[08:03:00.0000] <hsivonen>
of course, this wouldn't be an issue if XPath had had a namespace wildcard by default like Selectors

[08:03:01.0000] <hsivonen>
a lot of issues can be traced down to Namespaces sucking in the first place

[08:04:00.0000] <hsivonen>
or more precisely, the changing naming from a single string into a pair

[08:04:01.0000] <MikeSmith>
and everybody knew it was f*ucked up from teh beginning

[08:04:02.0000] <gsnedders>
MikeSmith: :)

[08:05:00.0000] <annevk2>
which can be traced back to RDF which is why XML has namespaces

[08:05:01.0000] <hsivonen>
indeed

[08:05:02.0000] <annevk2>
a Semantic Web seed carefully planted to make the current Web more complicated...

[08:09:00.0000] <annevk2>
http://twitter.com/distobj/status/1462908234 more rambling without reading

[08:13:00.0000] <zcorpan>
<title role="html:h1">?

[08:14:00.0000] <hsivonen>
wouldn't architectural forms be more like <div html="h1">?

[08:15:00.0000] <MikeSmith>
ah, more gems

[08:16:00.0000] <MikeSmith>
"the IETF tries to master entropy"

[08:20:00.0000] <takkaria>
"The correct solution is simple: require namespace well-formedness for HTML 5 documents. Until the spec takes that simple step, ..."

[08:22:00.0000] <jcranmer>
where is this?

[08:22:01.0000] <takkaria>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=6777

[08:22:02.0000] <Philip`>
"The difference between the IETF and the WHATWG is that the WHATWG respects the laws of thermodynamics"

[08:23:00.0000] <MikeSmith>
 like Christian Democrats asserting that existing social institutions work just fine, but the problem is just that people don't follow the (arbitrary) rules, therefor the solution is to fine ways to compel/force more people to just bend to their will and follow the rules

[08:23:01.0000] <gsnedders>
http://www.youtube.com/user/ie8videos

[08:24:00.0000] <MikeSmith>
faith-based specification development

[08:25:00.0000] <MikeSmith>
maybe they can get some US government funding for their work

[08:25:01.0000] <jcranmer>
/me wants to smack comment 8

[08:27:00.0000] <MikeSmith>
the problem with the whatwg is that it's soft on crime

[08:28:00.0000] <jcranmer>
does she realize she's asking the impossible?

[08:28:01.0000] <annevk2>
she?

[08:30:00.0000] <gsnedders>
A female name, like Anne :P

[08:30:01.0000] <Philip`>
jcranmer: Things that are impossible just take longer

[08:31:00.0000] <Lachy>
jcranmer, comment 8 where?

[08:31:01.0000] <jcranmer>
Elliotte -- names that end in -te tend to be feminine

[08:31:02.0000] <jcranmer>
Lachy: 10:27 < takkaria> http://www.w3.org/Bugs/Public/show_bug.cgi?id=6777

[08:32:00.0000] <gsnedders>
MikeSmith: Also, to make you happy, I've now almost finished dissertation on Nabokov

[08:32:01.0000] <Philip`>
jcranmer: Like "Pete"?

[08:32:02.0000] <jcranmer>
I said "tend to be"

[08:32:03.0000] <jcranmer>
besides, Pete is usually short for Peter

[08:32:04.0000] <MikeSmith>
gsnedders: you would make me really happy if you told me that you burned it after you finally have it done

[08:32:05.0000] <Lachy>
jcranmer, Elliotte sounds more like a male name. But that depends how its pronounced

[08:33:00.0000] <MikeSmith>
gsnedders: that's what all great creative types do

[08:33:01.0000] <MikeSmith>
Sibelius, etc.

[08:35:00.0000] <Philip`>
And there's that guy on Youtube who built an image out of matchsticks and then burnt them all

[08:37:00.0000] <Philip`>
/me wonders if Elliotte is an unusually rare name, or if elharo just has a lot more page-rank than all the others

[09:05:00.0000] <zcorpan>
hsivonen: you should try to get http://v.nu/

[09:08:00.0000] <Philip`>
If that's taken, try http://ν.nu

[09:08:01.0000] <hsivonen>
zcorpan: earlier, when it was up for auction, I decided that it wasn't worth my money (or Mozilla's money) at its price.

[09:08:02.0000] <hsivonen>
zcorpan: I forgot what the price was.

[09:09:00.0000] <zcorpan>
ok

[09:09:01.0000] <Philip`>
/me wonders why .nu

[09:09:02.0000] <hsivonen>
zcorpan: also, I didn't like the idea of paying a squatter

[09:09:03.0000] <MikeSmith>
hsivonen: I think it was that they auction the single-letter domains, and it starts at 500 euro or so

[09:10:00.0000] <MikeSmith>
hsivonen: I don't think it's necessarily a squatter

[09:10:01.0000] <MikeSmith>
or maybe it was in this case

[09:10:02.0000] <hsivonen>
Philip`: it was the least ugly (the only? I forget) validator.tld domain available 

[09:11:00.0000] <hsivonen>
Philip`: also, you are supposed to pronounce "nu" like English "new"--not like French "nu" :-)

[09:11:01.0000] <Philip`>
Ah, okay

[09:11:02.0000] <Philip`>
/me pronounces it like English ν :-)

[09:11:03.0000] <Philip`>
(which is like "new")

[09:12:00.0000] <zcorpan>
hsivonen: or like swedish "nu" (which means "now") :)

[09:12:01.0000] <gsnedders>
/me pronounces like French "nu" :)

[09:12:02.0000] <Philip`>
(at least according to the mathsy people who I've heard use that term)

[09:13:00.0000] <hsivonen>
MikeSmith: I think it was more expensive than 500 EUR the last time

[09:13:01.0000] <hsivonen>
but like I said, I forgot

[09:14:00.0000] <gsnedders>
/me thinks he should start editing his English dissertation

[09:14:01.0000] <gsnedders>
/me sighs

[09:14:02.0000] <gsnedders>
When will I finish it!?

[09:14:03.0000] <gsnedders>
(Likely answer: hours before it has to be sent off.)

[09:15:00.0000] <hsivonen>
whoa. Hixie is not on IRC. screen died perhaps?

[09:17:00.0000] <Philip`>
gsnedders: Why waste hours of valuable editing time before it has to be sent off?

[09:17:01.0000] <hsivonen>
does Chrome support XSLT?

[09:18:00.0000] <hsivonen>
according to w3schools, yes

[09:18:01.0000] <hsivonen>
/me ducks having used w3schools as a reference

[09:18:02.0000] <zcorpan>
hsivonen: w3c can't be wrong

[09:34:00.0000] <gsnedders>
Wikipedia is so much easier than trying to find information from an entire novel

[09:43:00.0000] <Lachy>
http://www.quirksmode.org/blog/archives/2009/04/making_time_saf.html

[09:45:00.0000] <hsivonen>
perhaps HTML 5 should say that <time> is not safe for historians

[09:45:01.0000] <hsivonen>
I thought it already did

[09:45:02.0000] <hsivonen>
so more clearly than now

[09:48:00.0000] <Lachy>
I thought it did too, but it seems that it isn't clear enough

[09:48:01.0000] <Lachy>
I don't understand what PPK means by allowing "arbitrary year-naming systems to be specified"

[09:52:00.0000] <zcorpan>
seems pointless to have a cutoff date at 1870 or 1918

[09:54:00.0000] <Philip`>
Seems pointless to have a cutoff date at 1

[09:54:01.0000] <zcorpan>
the point with having it at 1 is that it keeps the syntax simple

[09:56:00.0000] <Lachy>
Philip`, the cut off date was set to 1 because anything else would be competely arbitrary and it made things simple without requiring syntactic changes

[09:57:00.0000] <annevk3>
times/dates suck

[09:58:00.0000] <Philip`>
Lachy: Negative infinity would be less arbitrary than 1

[09:59:00.0000] <Philip`>
(I guess 0 would be less arbitrary than 1 too)

[09:59:01.0000] <Lachy>
negative dates would require additional syntax and give teh false impression that time is designed for historical dates

[09:59:02.0000] <Lachy>
there was no year 0, so we had to pick 1

[10:00:00.0000] <Philip`>
Syntax is a pretty trivial issue

[10:01:00.0000] <Lachy>
sure, the changes to the algorithm and implementation needed to support dates beginning with a '-' is trivial. But completely unnecessary

[10:01:01.0000] <Philip`>
There's a year 0 in the ISO 8601 proleptic Gregorian calendar

[10:01:02.0000] <Philip`>
and it's just as meaningful as year 1

[10:01:03.0000] <Lachy>
yes, but that's confusing because year 0 in 8601 is 1BCE

[10:02:00.0000] <Philip`>
/me doesn't find that confusing :-)

[10:02:01.0000] <Lachy>
Philip`, ask some random people whether -0004-01-01 is 3BCE or 4BCE.

[10:03:00.0000] <Lachy>
or 5BCE

[10:03:01.0000] <Philip`>
Random people probably would never recognise it as a date

[10:03:02.0000] <Lachy>
well, HTML authors who are familiar with the ISO-8601 syntax

[10:04:00.0000] <Lachy>
i.e. people who use hCalendar

[10:13:00.0000] <Philip`>
Lachy: Ask the same people whether 2 September 1666 is 1666-09-02 or 1666-09-12 and they'll probably get it wrong too, but we're not trying to protect them from that

[10:16:00.0000] <Lachy>
Philip`, I'm sure they would get that wrong too. But choosing to explicitly add support for negative years when we know it's problematic is different from supporting pre 16th century dates simply as a result of the syntax

[10:16:01.0000] <Lachy>
actually, 17th century, but whatever.

[10:18:00.0000] <krijnh>
Are people here okay with me moving the logs over to my new website, http://www.krijn-engineering.nl/ ?

[10:20:00.0000] <Lachy>
krijnh, if you set up a redirect, yes

[10:21:00.0000] <Philip`>
krijnh: I like your use of a 404 in a <bgsound>

[10:21:01.0000] <krijnh>
Philip`: what? Ow, that's a bug then :/ Damnit

[10:22:00.0000] <Philip`>
krijnh: You need to be more careful when you're uploading your web site from Word

[10:23:00.0000] <annevk3>
you'd think krijnh would know a bit about HTML after logging this channel for two years

[10:23:01.0000] <Lachy>
krijnh, wtf? Why have you published a page created with Microsoft Office?

[10:24:00.0000] <krijnh>
How do you know it's made with Office?

[10:24:01.0000] <Lachy>
/me collects krijnh's whatwg cabal membership card.

[10:24:02.0000] <krijnh>
I thought I stripped out that code :/

[10:24:03.0000] <Lachy>
look at the source.

[10:24:04.0000] <annevk3>
punk'd

[10:24:05.0000] <annevk3>
:D

[10:25:00.0000] <Philip`>
krijnh: Surely you should write that page with SVG rather than VML

[10:26:00.0000] <krijnh>
*sigh*

[10:26:01.0000] <krijnh>
You're way too elitist you!

[10:26:02.0000] <krijnh>
Not everybody can handcode HTML :(

[10:26:03.0000] <Philip`>
You don't need to handcode it - just use Inkscape

[10:27:00.0000] <Philip`>
and then use Amaya to merge it with the HTML

[10:27:01.0000] <krijnh>
In fact I did, and after that I copied it to Word, cause the margins were easier to manage in Word

[10:27:02.0000] <annevk3>
but Word also kills babies

[10:27:03.0000] <krijnh>
The rounded corners were pretty hard to create in Word, so that's why I used Inkscape

[10:27:04.0000] <krijnh>
But it's not finished yet

[10:28:00.0000] <krijnh>
I think the design should be centered

[10:29:00.0000] <Philip`>
Needs more clipart

[10:29:01.0000] <hsivonen>
is the use of OOo allowed?

[10:30:00.0000] <krijnh>
I'll tell it to my webdesinger

[10:31:00.0000] <hsivonen>
I blame all the uppercase tags on my site on OOo.

[10:33:00.0000] <Philip`>
You should use Tidy

[10:34:00.0000] <hsivonen>
why?

[10:34:01.0000] <Philip`>
To make things tidy

[10:34:02.0000] <Philip`>
like lowercasing tags

[10:35:00.0000] <hsivonen>
the ROI is unclear to me

[10:36:00.0000] <annevk3>
ISO 8601 is not Y100K proof

[10:37:00.0000] <Philip`>
Human civilisation is not Y100K proof

[10:37:01.0000] <tantek>
Lachy, Philip' - supporting day-precise negative years or even years 18th century and earlier is asking for a heap of trouble/complexity

[10:37:02.0000] <tantek>
no matter the format

[10:37:03.0000] <annevk3>
Philip`, that's unclear

[10:38:00.0000] <Philip`>
hsivonen: If you "blame" something for uppercase tags, presumably you dislike uppercase tags, and hence there would be some aesthetic value in lowercasing them

[10:38:01.0000] <Philip`>
annevk3: History is probably a good indicator, in the absence of any other evidence

[10:38:02.0000] <annevk3>
Philip`, it might be that the calendar format we have is not Y100K proof though

[10:39:00.0000] <Philip`>
Not even the Romans lasted a hundred thousand years

[10:39:01.0000] <annevk3>
Philip`, Romans are a subset of human civilisation...

[10:40:00.0000] <gsnedders>
tantek: I don't think anyone apart from those who are saying we must support such things or else time is pointless thinks that isn't the case

[10:41:00.0000] <tantek>
let them see this chart then, and write the *location*-sensitive code themselves: http://en.wikipedia.org/wiki/Gregorian_calendar#Timeline

[10:41:01.0000] <Philip`>
tantek: We've already asked for the heap of trouble by trying to support dates at all - if we don't support ancient dates then we have some arbitrary cutoff date instead, which will just cause different kinds of trouble

[10:41:02.0000] <gsnedders>
Oh yeah, they've brought up the complexity caused by that, and said we should just provide a means to provide dates in other calendars to circumvent that problem

[10:42:00.0000] <tantek>
Philip' no need to have an "arbitrary" cutoff date, simply use 1926 as the cut-off and cite http://en.wikipedia.org/wiki/Gregorian_calendar#Timeline as the justification

[10:42:01.0000] <tantek>
and invite counter-proposals with *better* justifications/citations

[10:43:00.0000] <Philip`>
tantek: Sounds pretty arbitrary to base it on when Turkey switched calendar

[10:43:01.0000] <tantek>
Philip' - it's simply the last country to switch over in that list

[10:43:02.0000] <Philip`>
and introduces unnecessary complexity if e.g. a UK-based site wants to mark up people's dates of birth

[10:43:03.0000] <gsnedders>
tantek: They have given counter-proposals of allowing an attribute to specify the calendar in use for the date format

[10:44:00.0000] <tantek>
thus from 1926 on, you know that Gregorian dates are consistent worldwide

[10:44:01.0000] <hsivonen>
Philip`: s/Turkey/Greece/ ?

[10:44:02.0000] <hsivonen>
oops.

[10:44:03.0000] <hsivonen>
/me read the wrong sentence on wikipedia

[10:44:04.0000] <tantek>
gsnedders, based on what evidence that people specify alternative calendars? 

[10:44:05.0000] <tantek>
RFC2445 similarly has such a field, but in practice it is always GREGORIAN

[10:45:00.0000] <hsivonen>
(someone should mention Turkey in the wikipedia prose in addition to the image that doesn't respond to Find on Page)

[10:47:00.0000] <Philip`>
tantek: If Saudi Arabia decided tomorrow to switch from the Islamic calendar to the Gregorian, should we restrict <time> to use 2009 as the cut-off?

[10:48:00.0000] <annevk3>
reading http://tools.ietf.org/html/rfc3339#appendix-A I wonder why some people argue we should simply reference ISO 8601; seems like a disaster

[10:48:01.0000] <tantek>
Philip - I call theoretical. E.g.: if aliens showed up tomorrow with a better datetime format, should we switch to use it?

[10:49:00.0000] <annevk3>
i think we should form a comittee and compromise between all the various formats

[10:50:00.0000] <gsnedders>
tantek: Evidence? They're trolls, they don't need evidence.

[10:50:01.0000] <annevk3>
(that RFC syntax also does not allow Y10K)

[10:50:02.0000] <tantek>
lacking evidence, trolls can be ignored.

[10:52:00.0000] <tantek>
annevk3 - in that case, here is my proposed new calendar: http://tr.im/newcal

[10:52:01.0000] <Philip`>
tantek: So the cut-off date should be defined as the date that the latest country officially switched to the Gregorian calendar according to Wikipedia (ignoring China which is complex) before 2009?

[10:53:00.0000] <tantek>
Philip', see above, you are free/encouraged to provide "counter-proposals with *better* justifications/citations"

[10:53:01.0000] <tantek>
(wow PBWiki 2 forced upgrade (really a downgrade) really screwed up some markup)

[10:55:00.0000] <Philip`>
tantek: I counter-propose setting the cut-off date at 1, which is justified by being useful for a wider range of use cases and by being easier to remember and by not depending on quite so many pieces of irrelevant trivia :-)

[10:55:01.0000] <annevk3>
of course ECMAScript is even worse as it does not define Date parsing at all :/

[10:57:00.0000] <tantek>
Philip - that's a reasonable counterproposal, you may want to add some wording about sticking to ISO8601 and for any dates in locations before they supported the Gregorian calendar, that they MUST use Proleptic Gregorian dates per http://en.wikipedia.org/wiki/Gregorian_calendar#Proleptic_Gregorian_calendar

[10:59:00.0000] <tantek>
/me has begun using ordinal ISO8601 dates with a hyphen http://en.wikipedia.org/wiki/ISO_8601#Ordinal_dates for filing/dating/signing purposes.

[11:01:00.0000] <gsnedders>
tantek: Why ordinal?

[11:01:01.0000] <Philip`>
/me can only use year/month/day, because he can only remember year and his watch only tells him month/day

[11:03:00.0000] <tantek>
gsnedders - easier to do day math (how many days between date x and date y), and convert between Gregorian and NewCalendar

[12:43:00.0000] <Hixie>
hsivonen: having one cake and eating another kinda misses the point of the expression :-P

[12:50:00.0000] <tantek>
Hixie, clearly the expression must have been underspecified.

[12:51:00.0000] <hsivonen>
Hixie: do you have an actual solution to the problem (other than "don't implement XPath")?

[12:51:01.0000] <Hixie>
i don't understand the problem in enough detail to have an educated suggestion

[12:52:00.0000] <hsivonen>
The problem is this:

[12:52:01.0000] <hsivonen>
XPath has been exposed to content before the namespace unification

[12:52:02.0000] <hsivonen>
therefore, existing expressions expect HTML elements to be in no namespace

[12:53:00.0000] <hsivonen>
and XHTML elements to be in http://www.w3.org/1999/xhtml namespace

[12:53:01.0000] <annevk3>
arguably XPath for HTML is undefined so we can do whatever we want

[12:53:02.0000] <hsivonen>
We'd want to change things so that you can write one kind of new expressions for both *and* keep the old expressions out there matching, too

[12:54:00.0000] <hsivonen>
By default, the syntax people actually use doesn't wildcard the namespace as is the case with Selectors

[12:54:01.0000] <hsivonen>
instead, the namespace is fixed to be the 'no namespace'

[12:54:02.0000] <hsivonen>
in the XPath data model, a node can be in only one namespace

[12:55:00.0000] <hsivonen>
so to support both future expressions that use the http://www.w3.org/1999/xhtml namespace and past expressions that use no namespace, WebKit violates the XPath data model

[12:56:00.0000] <hsivonen>
I have tentatively implemented the exact same violation in a patch for Gecko

[12:56:01.0000] <hsivonen>
annevk3: we can do whatever we want only if we are OK with breaking existing expressions

[12:56:02.0000] <jgraham>
hsivonen: It seems like an entirely reasonable and necessary violation to me

[12:57:00.0000] <jgraham>
Since I agree that there is a constraint that existing expressions should not be broken

[12:57:01.0000] <annevk3>
hsivonen, I mean we can "violate" XPath because technically it is not a violation as XPath for text/html is not defined

[12:57:02.0000] <hsivonen>
arguably, the violation could be adjusted to be a bit more sane if a no-namespace expression no longer matched no-namespace elements

[12:57:03.0000] <hsivonen>
when now in WebKit, it matches both no-namespace nodes and HTML nodes

[12:58:00.0000] <hsivonen>
annevk3: It's a violation of XPath if you cannot explain the behavior in terms of the XPath data model

[12:58:01.0000] <jgraham>
hsivonen: Would there still be a way to match no-nemaspace elements?

[12:58:02.0000] <hsivonen>
jgraham: no

[12:58:03.0000] <hsivonen>
jgraham: but one might argue no one should have no-namespace nodes in HTML trees anyway

[12:59:00.0000] <annevk3>
hsivonen, saying that no namespace and the XHTML namespace are identical for HTML trees is not good enough?

[12:59:01.0000] <jgraham>
One could argue lots of things :)

[13:00:00.0000] <hsivonen>
annevk3: one might claim that the mapping to XDM not be bijective and that both no-namespace nodes and XHTML nodes flatten into one in XDM

[13:00:01.0000] <jgraham>
I think making XPath match either no-namespace or HTML namespace makes a lot of sense.

[13:00:02.0000] <annevk3>
XPath is indeed an exception btw, the DOM, Selectors, usually default to any namespace

[13:00:03.0000] <hsivonen>
annevk3: but that's still not enough to explain the WebKit behavior in terms of XDM

[13:00:04.0000] <gsnedders>
hsivonen: Can it not just be specified in terms of DOM-like interface in HTML 5?

[13:00:05.0000] <hsivonen>
annevk3: because a node can have only one namespace in XDM but we want both no-namespace and http://www.w3.org/1999/xhtml-namespace expressions to match against the same nodes

[13:01:00.0000] <gsnedders>
i.e., this should follow DOM Level 3 XPath, but HTML elements should behave as if they are in no namespace

[13:01:01.0000] <hsivonen>
gsnedders: the interface requires an additonal datum that isn't in XDM:  the HTMLness flag on owner doc

[13:01:02.0000] <annevk3>
"This is a willful violation of the XDM."

[13:01:03.0000] <hsivonen>
gsnedders: unless you violate XML to XDM mapping in XHTML, too...

[13:01:04.0000] <gsnedders>
IMO XPath is the wrong place to spec this

[13:01:05.0000] <annevk3>
hsivonen, we can argue that in our XDM no namespace is the XHTML namespace

[13:02:00.0000] <annevk3>
hsivonen, that there's no distinction in our XDM

[13:02:01.0000] <annevk3>
hsivonen, for the purposes of the XDM, anyway

[13:02:02.0000] <annevk3>
hsivonen, but I'm not that interested in spec-lawyering, lets just deal with the fallout in 5 years rather than now :)

[13:03:00.0000] <hsivonen>
annevk3: the thing is, in terms of XDM, you only get to flatten things in the tree

[13:03:01.0000] <hsivonen>
annevk3: you don't get to flatten things in the expressions themselves

[13:05:00.0000] <jgraham>
hsivonen: annevk3 has a point. It is pretty clear what implementations have do do. We should ship the software then spec the required behaviour

[13:05:01.0000] <jgraham>
Since it's hard to argue with interoperable running code

[13:07:00.0000] <jgraham>
/me wonders if that is somehoe not playing by the rules

[13:07:01.0000] <hsivonen>
jgraham: well, as you can see, what I did with code was to do exactly what WebKit does

[13:07:02.0000] <Hixie>
doing what webkit does seems like the simplest solution

[13:07:03.0000] <gsnedders>
What does WebKit do, exactly?

[13:07:04.0000] <gsnedders>
Does it only apply for HTML documents, or for all documents?

[13:07:05.0000] <hsivonen>
Hixie: the simpler solution is not to unify HTML and XHTML XDMs

[13:08:00.0000] <hsivonen>
Hixie: so that even future XPath would always have to be different for the two

[13:08:01.0000] <hsivonen>
Hixie: even if JS and Selectors Just Worked

[13:08:02.0000] <hsivonen>
it seems like an undesirable outcome not to be able to write unified XPath for HTML and XHTML, though

[13:10:00.0000] <Hixie>
hsivonen: s/simplest/best/ then

[13:16:00.0000] <tantek>
/me is a bit amazed.

[13:17:00.0000] <tantek>
which spec does Webkit violate? (URL?)

[13:17:01.0000] <tantek>
and is there a description for how it violates it, and *why*? (URL aside from #whatwg IRC logs)

[13:18:00.0000] <Hixie>
two bonus points to anyone who can guess the [[Class]] name of the element created by document.createElement('keygen')

[13:20:00.0000] <hsivonen>
tantek: XPath. and http://www.w3.org/Bugs/Public/show_bug.cgi?id=6777 

[13:20:01.0000] <hsivonen>
(sorry, gotta run now)

[13:20:02.0000] <gsnedders>
/me guesses something totally illogical, thus doesn't guess

[13:22:00.0000] <gsnedders>
(Actually, I know it is something totally illogical: if that weren't the case, Hixie wouldn't be asking.)

[13:22:01.0000] <Hixie>
you're wrong

[13:22:02.0000] <gsnedders>
Sod.

[13:22:03.0000] <Hixie>
it is not one thing totally illogical

[13:22:04.0000] <Hixie>
it is THREE things totally illogical

[13:22:05.0000] <Hixie>
depending on the browser

[13:22:06.0000] <Hixie>
HTMLSelectElement, HTMLSpanElement, or HTMLElement

[13:22:07.0000] <annevk3>
HTMLElement, HTMLSpanElement, and HTMLSelectElement are not that illogical

[13:23:00.0000] <Hixie>
given that keygen should be exposing .disabled, they are

[13:23:01.0000] <annevk3>
you win :)

[13:24:00.0000] <annevk3>
the tree it gives in the DOM in Firefox/WebKit is also funny

[13:25:00.0000] <Hixie>
firefox just treats keygen as a magic macro in the parser, like <isindex>

[13:26:00.0000] <gsnedders>
/me realizes he is procrastinating, again

[13:26:01.0000] <Hixie>
but webkit actually creates a <select> element even for document.createElement('keygen')

[13:26:02.0000] <annevk3>
this will bite hsivonen

[13:27:00.0000] <annevk3>
(Opera doesn't actually expose .disabled)

[13:27:01.0000] <gsnedders>
(The version of U2's "The Fly" on "U2 Live from Boston" is awesome.)

[13:27:02.0000] <Hixie>
oh this is rich! guess what the return value of  document.createElement('keygen').childNodes.length  is

[13:28:00.0000] <gsnedders>
2?

[13:28:01.0000] <Hixie>
per DOM Core, it cannot be anything but 0

[13:28:02.0000] <Hixie>
but webkit makes it 3

[13:28:03.0000] <gsnedders>
I was only one out!

[13:41:00.0000] <tantek>
hsivonen, thanks for another laugh. your bug report elicited some quite (untintended I'm sure) humorous responses/requests.

[13:42:00.0000] <tantek>
e.g. "The specific step I would like to see happen is for HTML 5 to mandate namespace

[13:42:01.0000] <tantek>
well-formedness and require draconian error handling."  LOL!

[13:42:02.0000] <Hixie>
it's not clear to me what elharo meant by that

[13:42:03.0000] <Hixie>
the literal interpretation is so obviously not going to work that i can only assume i misunderstood it

[13:43:00.0000] <tantek>
also, referring to "text/html" as "one special case" (same comment in bug thread)

[13:43:01.0000] <tantek>
it's always nice when people make it quite clear how out of touch they are with real world web development/deployment/maintenance.

[13:45:00.0000] <annevk3>
hmm, Opera <keygen> has more event handler attributes so it does have some differences from e.g. <span>

[13:46:00.0000] <Hixie>
if you think that's fun you should see roy's comments on ietf-http-wg

[13:46:01.0000] <Hixie>
about how web browsers are but a minor part of the http ecosystem and that their needs can basically be dismissed since they're not that important on the global scale of things

[13:47:00.0000] <tantek>
Hixie, URL? that sounds worth the read

[13:49:00.0000] <tantek>
/me is collecting URLs to expand the documentation of why/how namespaces have failed on the web.

[13:49:01.0000] <annevk3>
FYI: http://wiki.whatwg.org/wiki/Namespace_confusion

[13:50:00.0000] <Hixie>
threat starting here: http://lists.w3.org/Archives/Public/ietf-http-wg/2009JanMar/thread.html#msg290 continuing here http://lists.w3.org/Archives/Public/ietf-http-wg/2009AprJun/thread.html#msg0

[13:50:01.0000] <Hixie>
thread, even

[13:50:02.0000] <annevk3>
amusing typo

[14:00:00.0000] <tantek>
annevk3 - indeed - have seen it and am documenting here (cross-linked) http://microformats.org/wiki/namespaces-considered-harmful

[14:00:01.0000] <tantek>
recently found and added this: http://www.xml.com/pub/a/2004/07/21/dive.html

[14:01:00.0000] <tantek>
worth reading. note how old it is.

[14:04:00.0000] <Philip`>
http://google.com/codesearch/p?hl=en#J7osgojbryc/src/java/org/web3d/x3d/jaxp/X3DSAVAdapter.java&q=X3DSAVAdapter&l=847 is fun

[14:05:00.0000] <Philip`>
if(!space.equals("xsd")) etc

[14:06:00.0000] <Philip`>
(Xj3D is the closest thing to an 'official' implementation of X3D)

[14:32:00.0000] <tantek>
Philip', nice example. I've added it to http://microformats.org/wiki/namespaces-considered-harmful#namespaced_content_has_failed .  I wonder how many such examples could find by using google code search to look for similar string prefix literal instances/compares with common namespace prefix usages.

[14:43:00.0000] <annevk3>
we should somehow deal with the JavaScript versioning issue

[14:43:01.0000] <annevk3>
https://bugzilla.mozilla.org/show_bug.cgi?id=487070

[14:43:02.0000] <annevk3>
otherwise it starts leaking all over the place

[14:44:00.0000] <annevk3>
(I don't have a proposal :/)

[14:47:00.0000] <gsnedders>
To what extent is form@method followed?

[14:47:01.0000] <gsnedders>
Are only certain methods allowed?

[14:48:00.0000] <jgraham>
annevk3: Yeah, we totlly need a good solution for that

[14:51:00.0000] <jgraham>
It seems like in the short term Mozilla could do something like mozImportScripts but even ES5 vs ES3 needs a solution

[14:53:00.0000] <annevk3>
gsnedders, yes, with GET as fallback

[14:53:01.0000] <gsnedders>
That's what I thought.

[14:53:02.0000] <annevk3>
gsnedders, RTFS

[14:53:03.0000] <gsnedders>
annevk3: That doesn't mean we have interop on it now.

[14:54:00.0000] <gsnedders>
(I don't care about TFS, I care about what is shipping)

[14:54:01.0000] <jgraham>
/me was reading S as "Source"

[14:54:02.0000] <annevk3>
hehe, my answer works either way :p

[14:55:00.0000] <jgraham>
annevk3: You work for a proprietary browser company :)

[14:55:01.0000] <annevk3>
gsnedders, I suspect they follow the spec, but you'd have to test to be sure

[14:56:00.0000] <annevk3>
jgraham, I didn't say it was easy :p

[14:57:00.0000] <jgraham>
step 1: accept a job as a developer

[14:58:00.0000] <jgraham>
step 2: leave

[14:58:01.0000] <jgraham>
step 3: accept a MS job that gives you access to the IE source

[14:58:02.0000] <jgraham>
Seems easier to just ask really 

[14:58:03.0000] <sayrer>
tantek, you might want to look at <o:p> for your namespace document

[15:06:00.0000] <sicking>
Hixie, re doc.write()

[15:06:01.0000] <sicking>
Hixie, that is not compatible with what at least firefox does. In a fairly catastrophicly incompatible way

[15:07:00.0000] <sicking>
Hixie, is that what IE does?

[15:09:00.0000] <Hixie>
sicking: as far as i can tell, the spec matches what Firefox does

[15:09:01.0000] <Hixie>
sicking: and what IE does

[15:09:02.0000] <Hixie>
sicking: (with some minor edge case exceptions that are unlikely to be hit in practice)

[15:09:03.0000] <sicking>
Hixie, in firefox if you use document.write while the document is still loading, from for example a setTimeout, we do not wipe the current doc

[15:10:00.0000] <sicking>
Hixie, we insert into the stream, whereever that happens to be

[15:11:00.0000] <Hixie>
ah you are correct, i just looked at my notes

[15:11:01.0000] <Hixie>
firefox does something ridiculous (rewind the tokeniser if half-way through a token, etc)

[15:12:00.0000] <Hixie>
so i dismissed firefox's behavior as something nobody would rely on

[15:12:01.0000] <sicking>
Hixie, indeed. Firefox is a guaranteed race condition

[15:12:02.0000] <sicking>
Hixie, does IE wipe the doc?

[15:12:03.0000] <Hixie>
yes

[15:12:04.0000] <Hixie>
http://www.hixie.ch/tests/adhoc/dom/level0/write/007.html

[15:12:05.0000] <Hixie>
http://www.hixie.ch/tests/adhoc/dom/level0/write/008.html

[15:13:00.0000] <sicking>
Hixie, so i'm a little scared to implement that, given how catastrophically things fail for gecko-specific content

[15:13:01.0000] <sicking>
Hixie, would prefer to make it a no-op instead

[15:14:00.0000] <Hixie>
i'll make this a no-op if you're willing to change the inner-window outer-window thing to what IE does. :-)

[15:14:01.0000] <sicking>
what does IE do?

[15:14:02.0000] <Hixie>
throw an exception if you run script in a window that's not active

[15:14:03.0000] <sicking>
not always

[15:15:00.0000] <Hixie>
my comment wasn't really a serious proposal

[15:15:01.0000] <sicking>
i know

[15:15:02.0000] <Hixie>
my point is that we can't just never do what IE does

[15:15:03.0000] <sicking>
agreed, but don't we have a lot of IE cruft in there?

[15:15:04.0000] <Hixie>
not really

[15:15:05.0000] <sicking>
like innerHTML, offsetLeft, XMLHttpRequest

[15:15:06.0000] <Hixie>
only innerHTML of those is in HTML5

[15:16:00.0000] <Hixie>
and it's specced to match firefox

[15:16:01.0000] <Hixie>
not IE

[15:16:02.0000] <Hixie>
i don't really see why doing it the way IE does here is a bad thing

[15:16:03.0000] <Hixie>
it seems to be what everyone else does

[15:16:04.0000] <Hixie>
even firefox does it in most cases, just not timeouts

[15:17:00.0000] <Hixie>
e.g. from an onload="" it'll blow away the doc, no?

[15:17:01.0000] <sicking>
anything after DOMContentLoaded

[15:17:02.0000] <sicking>
not at all based on where it's called from, only when

[15:17:03.0000] <sicking>
so onmouseout, xhr.readystatechange, setInterval, etc, are all the same

[15:18:00.0000] <sicking>
i would be fine with trying to do what the spec says though

[15:18:01.0000] <Hixie>
if there is actual content that this breaks, then i'd be glad to consider it, but realistically speaking i imagine other browsers would be just as scared of making it a no-op

[15:18:02.0000] <sicking>
it seems very unlikely that someone relies on current behavior, other than through bugs

[15:18:03.0000] <sicking>
true

[15:18:04.0000] <sicking>
sure, lets do this and if we find pages that break we can reeval

[15:19:00.0000] <Hixie>
cool

[15:49:00.0000] <Hixie>
hsivonen is going to love <keygen>. </sarcasm>

[15:51:00.0000] <gsnedders>
/me takes a deep breath

[15:52:00.0000] <gsnedders>
(and then proceeds as in the "Any other end tag" section below)

[16:27:00.0000] <Hixie>
anyone know what opera supports when it comes to <keygen>? dsa? pqg? keyparams?

[16:27:01.0000] <Hixie>
annevk3?

[16:29:00.0000] <annevk3>
I think type="pkcs10" is an extension we have

[16:30:00.0000] <annevk3>
Yngve e-mailed you with details

[16:30:01.0000] <annevk3>
Not archived unfortunately but I guess that doesn't really matter

[16:32:00.0000] <Hixie>
his e-mail didn't seem to cover pqg/keyparams

[16:32:01.0000] <Hixie>
does that mean opera doesn't support them?

[16:33:00.0000] <annevk3>
It means I don't know

[16:33:01.0000] <annevk3>
sorry

[16:34:00.0000] <Hixie>
any chance you can find out? :-)

[16:34:01.0000] <Hixie>
/me considers dropping keygen after all

[16:34:02.0000] <Philip`>
What's wrong with testing and reverse-engineering? :-)

[16:35:00.0000] <Hixie>
i have no idea what to test or how

[16:35:01.0000] <Philip`>
Look at the specification to see what it's meant to do, and then just see if it does that or not

[16:36:00.0000] <Hixie>
what is this specification you speak of

[16:37:00.0000] <Philip`>
Well, maybe it's not so much a specification as an opportunity for a specification

[16:37:01.0000] <Philip`>
which is close enough

[16:38:00.0000] <annevk3>
Hixie, any chance you can reply and ask yourself? I understand less of this than you

[16:39:00.0000] <Hixie>
Philip`: until i looked at the gecko source code, i had no idea that there was a keyparams attribute at all, it's not documented anywhere

[16:39:01.0000] <Hixie>
Philip`: and until i saw an obscure e-mail on the topic in some archives, i didn't know that "ec" was a valid keytype

[16:39:02.0000] <Hixie>
Philip`: i have no idea how to reverse engineer what opera does short of asking them

[16:39:03.0000] <Hixie>
annevk3: yeah, probably will

[16:40:00.0000] <Hixie>
annevk3: basically all i'm looking for is a list of all the attributes and values that do something on keygen

[16:40:01.0000] <Hixie>
annevk3: ideally with some documentation as to what they do and what their format is

[16:40:02.0000] <annevk3>
there is http://web.archive.org/web/20070401073244/wp.netscape.com/eng/security/comm4-keygen.html

[16:40:03.0000] <annevk3>
but it's not that useful

[16:42:00.0000] <Philip`>
/me supposes it can be tricky to reverse-engineer a feature whose specific purpose is to produce results that cannot be reverse-engineered even with all the computing power in the world

[16:43:00.0000] <olliej>
Philip`: heheh

[16:43:01.0000] <annevk3>
https://developer.mozilla.org/En/HTML/HTML_Extensions/KEYGEN_Tag has no documentation on the Gecko extension either

[16:48:00.0000] <Hixie>
lol

[16:48:01.0000] <Hixie>
after tracing this code very carefully all the way through to figure out how gecko uses the pqg attribute for dsa keys

[16:49:00.0000] <Hixie>
and going through multiple quite obscure and poorly documented codepaths

[16:49:01.0000] <Hixie>
i find this:

[16:49:02.0000] <Hixie>
605         case CKM_DSA_KEY_PAIR_GEN:

[16:49:03.0000] <Hixie>
606             // XXX Fix this! XXX //

[16:49:04.0000] <Hixie>
607             goto loser;

[16:49:05.0000] <Hixie>
they don't support dsa at all!

[16:52:00.0000] <Hixie>
and i have no way to tell if changing an attribute value changes the result

[16:52:01.0000] <Hixie>
since the result is PSEUDORANDOM BY DESIGN

[16:52:02.0000] <Hixie>
gah

[16:52:03.0000] <olliej>
Hixie: keygen should generate a random 16bit value that gets used character by character using efficient xor encryption of utf16 data :D

[16:52:04.0000] <Hixie>
i hate <keygen>

[16:52:05.0000] <olliej>
Hixie: nice and easy to spec out

[16:53:00.0000] <olliej>
Hixie: and efficient

[16:53:01.0000] <olliej>
;D

[16:53:02.0000] <Hixie>
olliej: i don't think we'd get many interoperable implementations

[16:53:03.0000] <olliej>
Hixie: of xor encryption? it's one line, even browser developers cna't screw up compatibility for that

[16:53:04.0000] <olliej>
surely :D

[16:53:05.0000] <Hixie>
oh i don't think complexity is why it'd be ignored...

[16:55:00.0000] <olliej>
Hixie: oh you mean there just wouldn't be many implementations at all

[16:55:01.0000] <olliej>
vs. multiple incompatible ones

[16:55:02.0000] <Philip`>
You could try patching the source code to remove the randomness, to make it easier to test

[16:55:03.0000] <Philip`>
/me refrains from making any comments about Debian

[16:55:04.0000] <olliej>
personally i'm for speccing out keygen based on xor encryption

[16:55:05.0000] <olliej>
but mostly because i want to see the followup whatwg mail from security people :D

[16:56:00.0000] <olliej>
i suspect it would be entertaining


2009-04-07
[17:03:00.0000] <jcranmer>
everything should use, er, DES

[17:04:00.0000] <olliej>
jcranmer: xor!

[17:04:01.0000] <olliej>
:D

[17:04:02.0000] <franksalim>
Philip`, I think that counts as making a comment about Debian

[17:05:00.0000] <jcranmer>
olliej: ROT-13... twice, for extra security!

[17:05:01.0000] <olliej>
jcranmer: couple with double xor!

[17:05:02.0000] <olliej>
jcranmer: it will be awesome!

[17:06:00.0000] <Philip`>
olliej: Stream ciphers use XOR anyway, so you're not proposing anything new :-p

[17:06:01.0000] <olliej>
Philip`: O_o

[17:07:00.0000] <Philip`>
(XOR with an approximation of a one-time pad, in particular)

[18:24:00.0000] <Hixie>
uh oh, adam's patience is cracking

[18:35:00.0000] <Hixie>
not only is this <keygen> thing introducing me to completely new technologies and standards

[18:35:01.0000] <Hixie>
it's introducing me to completely new (to me) standards ORGANISATIONS.

[18:36:00.0000] <Hixie>
/me really could have gone his whole week without learning about the DER part of ASN.1 and not been any the poorer

[20:35:00.0000] <Hixie>
i just wrote a spec i don't understand

[20:35:01.0000] <Hixie>
that generaly bodes badly

[20:35:02.0000] <Hixie>
i hope there are no errors in it

[20:46:00.0000] <olliej>
Hixie: keygen?

[20:46:01.0000] <Hixie>
yeah

[22:01:00.0000] <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)

[22:01:01.0000] <MikeSmith>
that is, should probably best be "key-pair generator control"

[22:02:00.0000] <MikeSmith>
unless it means "key pair-generator control"

[22:05:00.0000] <MikeSmith>
Hixie: value of challenge attribute is just a string? not additional constraints on it?

[22:48:00.0000] <shepazu>
Hixie: ping

[22:49:00.0000] <Hixie>
hey

[22:50:00.0000] <Hixie>
(my connection is extremely flaky, i apologise in advance if i go dark unexpectedly)

[22:51:00.0000] <shepazu>
Hixie: I'm working on the final version of the WebApps charter

[22:51:01.0000] <shepazu>
and I wanted to give the real names and timetables for the deliverables

[22:52:00.0000] <shepazu>
so...

[22:52:01.0000] <Hixie>
let's see

[22:52:02.0000] <Hixie>
there are four i'm editing, right?

[22:52:03.0000] <Hixie>
/me goes to find out what they're called right now

[22:52:04.0000] <shepazu>
Server-Sent Events, Web Sockets API, Web Storage, and Web Workers

[22:53:00.0000] <Hixie>
sounds right

[22:53:01.0000] <Hixie>
my optimistic timetable is FPWD ASAP, LC in October, CR in 2012, and REC in 2022 at the latest

[22:54:00.0000] <Hixie>
REC will probably be far earlier for most of those

[22:54:01.0000] <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

[22:54:02.0000] <Hixie>
oh is XBL2 still on that list too?

[22:54:03.0000] <shepazu>
http://www.w3.org/2008/12/webapps/charter-2009-proposed

[22:54:04.0000] <shepazu>
XBL2, yes

[22:55:00.0000] <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

[22:55:01.0000] <Hixie>
Does "DOM Level 4" include DOM Range and DOM Traversal?

[22:56:00.0000] <shepazu>
/me gets that, but is a little leary of not being explicit about that in the recent charter review

[22:56:01.0000] <shepazu>
jiminy, I just don't know... I guess it's subject to interpretation

[22:56:02.0000] <Hixie>
i might be roped into a DOM Range rewrite next year

[22:56:03.0000] <Hixie>
(there's growing pressure to fix that spec)

[22:56:04.0000] <shepazu>
we are talking about it during the domain call tomorrow, and we really should straighten it out

[22:57:00.0000] <Hixie>
(and add new APIs to it for things like rich text editing)

[22:57:01.0000] <shepazu>
Hixie: arguably, that could be part of DOM4

[22:57:02.0000] <shepazu>
but...

[22:57:03.0000] <shepazu>
I propose that we write all of this up in the wiki and get our ducks in a row

[22:58:00.0000] <shepazu>
I already feel like we are playing a little fast-and-loose on some of these

[22:58:01.0000] <shepazu>
Hixie: same timetable for all 4 of the "new" specs?

[22:58:02.0000] <shepazu>
ok, that's easy for me

[22:58:03.0000] <Hixie>
yeah

[22:59:00.0000] <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

[22:59:01.0000] <Hixie>
hard for us to predict when REC will happen

[22:59:02.0000] <shepazu>
I imagine we will have a new charter (and AC review) before then

[22:59:03.0000] <Hixie>
yeah

[23:00:00.0000] <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

[23:00:01.0000] <Hixie>
uri?

[23:00:02.0000] <shepazu>
even if it's beyond the scope of the current charter

[23:00:03.0000] <Hixie>
not sure which wiki this is

[23:01:00.0000] <shepazu>
anywhere in here: http://www.w3.org/2008/webapps/wiki/Main_Page

[23:01:01.0000] <shepazu>
I'm a little confused about where we are on XHR... I would have expected us to be long done with that

[23:02:00.0000] <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

[23:02:01.0000] <shepazu>
happily, the Media Object/MAXIM spec looks to be obsoleted soon by the Media Annotations WG

[23:02:02.0000] <shepazu>
so, I think that will probably drop out in the next charter

[23:03:00.0000] <shepazu>
I'm also a little concerned about where the Origin header is headed

[23:03:01.0000] <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

[23:04:00.0000] <shepazu>
and IETF collaboration in general...

[23:04:01.0000] <shepazu>
sigh.

[23:04:02.0000] <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?

[23:05:00.0000] <shepazu>
heh... that would be fine... it might give some context... I can massage it into the larger framework

[23:05:01.0000] <Hixie>
k

[23:05:02.0000] <shepazu>
yeah, a new page... "future development"

[23:05:03.0000] <shepazu>
thanks

[23:05:04.0000] <shepazu>
not urgent, but when you get time...

[23:06:00.0000] <shepazu>
the sooner the better, from the perspective of finalizing the current charter

[23:06:01.0000] <Hixie>
doing it now

[23:07:00.0000] <Hixie>
i created an account but it doesn't seem to let me edit?

[23:07:01.0000] <shepazu>
wait a mo

[23:07:02.0000] <Hixie>
ironically it seems fine with me _moving_ pages

[23:08:00.0000] <shepazu>
oh, mediawiki...

[23:09:00.0000] <shepazu>
ok, you should be able to edit now

[23:09:01.0000] <Hixie>
cool thanks

[23:09:02.0000] <shepazu>
yop

[23:16:00.0000] <Hixie>
http://www.w3.org/2008/webapps/wiki/Future_development

[23:18:00.0000] <Hixie>
shepazu: it's probably time to call the time of death on the Window Object spec, too

[23:18:01.0000] <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

[23:19:00.0000] <Hixie>
speaking of more important things, is WebIDL on your list?

[23:19:01.0000] <Hixie>
oh, there it is

[23:20:00.0000] <Hixie>
might be worth making that list alphabetical

[23:23:00.0000] <shepazu>
working on that now

[23:24:00.0000] <shepazu>
I'm not going to add or remove any deliverables until the WG formally decides to do so

[01:16:00.0000] <zcorpan>
Hixie: are you making <keygen> conforming?

[01:20:00.0000] <Hixie>
zcorpan: yes

[01:22:00.0000] <zcorpan>
cool

[01:23:00.0000] <zcorpan>
/me looks at the diff

[01:41:00.0000] <zcorpan>
Hixie: is <marquee> next? :)

[01:43:00.0000] <zcorpan>
wonder if marquee can be implemented using css transitions

[03:08:00.0000] <MikeSmith>
/me looks around for zcorpan

[03:08:01.0000] <jgraham>
/me has more success being in the same room

[03:08:02.0000] <MikeSmith>
hsivonen: http://bugzilla.validator.nu/attachment.cgi?id=75

[03:09:00.0000] <MikeSmith>
jgraham: is he there? poke him with a stick to wake 

[03:09:01.0000] <zcorpan>
MikeSmith: pong

[03:09:02.0000] <MikeSmith>
hej

[03:09:03.0000] <jgraham>
They don't give us long enough sticks here

[03:09:04.0000] <MikeSmith>
zcorpan: http://bugzilla.validator.nu/show_bug.cgi?id=474

[03:10:00.0000] <MikeSmith>
jgraham: you get longer sticks with more seniority

[03:10:01.0000] <zcorpan>
MikeSmith: i think that's a case where hsivonen disagrees with what Hixie specced

[03:10:02.0000] <MikeSmith>
ah

[03:11:00.0000] <MikeSmith>
hmm, so I'm wondering to do with that test case as far as db.json/validator-tester.py goes

[03:12:00.0000] <zcorpan>
MikeSmith: i think i had filed a bug about that already

[03:12:01.0000] <MikeSmith>
oh

[03:12:02.0000] <zcorpan>
MikeSmith: since i have a test case :)

[03:12:03.0000] <MikeSmith>
I'll look and marked mine as a dupe if so

[03:13:00.0000] <MikeSmith>
http://bugzilla.validator.nu/attachment.cgi?id=75 is a proposed fix for your just-a-hash test

[03:14:00.0000] <MikeSmith>
ah, i find http://bugzilla.validator.nu/show_bug.cgi?id=323 also

[03:14:01.0000] <MikeSmith>
"<img usemap=#> validates"

[03:14:02.0000] <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

[03:15:00.0000] <hsivonen>
MikeSmith: http://bugzilla.validator.nu/attachment.cgi?id=75 looks good

[03:15:01.0000] <MikeSmith>
hsivonen: thanks

[03:15:02.0000] <MikeSmith>
hsivonen: about the encoding issue, I see

[03:15:03.0000] <MikeSmith>
so I'm wondering what to do about test cases for that kind of stuff

[03:16:00.0000] <hsivonen>
MikeSmith: I suppose the URLs could be listed in a text file in the same directory as the db.json

[03:16:01.0000] <MikeSmith>
hsivonen: just a static file, you mean?

[03:17:00.0000] <hsivonen>
MikeSmith: yes. to remember which tests didn't go into the db for any reason

[03:17:01.0000] <MikeSmith>
OK

[03:17:02.0000] <MikeSmith>
willful-violations-of-html5.txt ?

[03:17:03.0000] <hsivonen>
yeah :-)

[03:17:04.0000] <MikeSmith>
heh

[03:17:05.0000] <MikeSmith>
I will make it so

[03:19:00.0000] <MikeSmith>
hsivonen: so I plan to do a patch for keygen, and want to also write some test cases for it

[03:20:00.0000] <hsivonen>
MikeSmith: great! 

[03:21:00.0000] <MikeSmith>
should I add them into svn, or put them at my own site and use db.json?

[03:22:00.0000] <hsivonen>
I think having them is svn plus site somewhere & db.json would be ideal

[03:22:01.0000] <hsivonen>
I'm not sure if whattf.org autoupdates to serve the tests over http

[03:23:00.0000] <hsivonen>
MikeSmith: I think the tests will show up under http://syntax.whattf.org/relaxng/tests/ if you put them in svn

[03:23:01.0000] <hsivonen>
MikeSmith: I don't know how often fantasai's script refreshes directory

[03:44:00.0000] <MikeSmith>
hsivonen: OK, I'll mess around with it a bit

[03:48:00.0000] <gsnedders>
Woah. RSS only allows enclosure URIs to be in the HTTP scheme.

[03:49:00.0000] <gsnedders>
HTTPS is invalid.

[03:51:00.0000] <Philip`>
What about FTP?

[03:52:00.0000] <gsnedders>
Invalid.

[03:53:00.0000] <Philip`>
Shame

[04:00:00.0000] <annevk3>
hsivonen, you realize <keygen> is a major pain for you because Mozilla implements it like <isindex>?

[04:05:00.0000] <annevk3>
hmm, RFC2459 has been obsoleted several times...

[04:07:00.0000] <zcorpan>
Hixie: there is no parsing problem with <h1><p>Typeline</p><p>Mainline</p></h1>

[04:14:00.0000] <hsivonen>
annevk3: no, I haven't realized that yet

[04:15:00.0000] <hsivonen>
annevk3: what about <keygen> in XHTML?

[04:17:00.0000] <zcorpan>
hsivonen: html5 makes keygen a proper void element instead of an isindex-like parser macro

[04:18:00.0000] <Philip`>
<keygen> in XHTML in Firefox (2) doesn't do anything at all

[04:19:00.0000] <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 )

[04:19:01.0000] <zcorpan>
hey how long have you had that xhtml viewer?

[04:20:00.0000] <Philip`>
Since at least 2007-08-25

[04:21:00.0000] <zcorpan>
you should make your tools easier to discover

[04:21:01.0000] <Philip`>
But then people might use them

[04:22:00.0000] <Philip`>
and it'll fill up my web server access logs because it sends a request for every keypress :-(

[04:22:01.0000] <zcorpan>
then increase the timeout

[04:23:00.0000] <jgraham>
You should make your tools easy to download so that we can run them locally and not fill your access logs :)

[04:23:01.0000] <annevk3>
Yeah, Gecko has no <keygen> support in XHTML.

[04:23:02.0000] <annevk3>
Opera does :)

[04:24:00.0000] <Philip`>
zcorpan: You should have been paying attention when I linked to it on http://krijnhoetmer.nl/irc-logs/whatwg/20070825#l-262

[04:24:01.0000] <Philip`>
I can't imagine why you'd fail to be aware of it

[04:25:00.0000] <annevk3>
WebKit expands it even in XHTML...

[04:25:01.0000] <zcorpan>
Philip`: maybe because the url was too similar to your html viewer that i didn't notice

[04:26:00.0000] <zcorpan>
annevk3: we expand isindex in xhtml, too, though

[04:27:00.0000] <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 :-)

[04:27:01.0000] <Philip`>
zcorpan: Bah, excuses

[04:28:00.0000] <Philip`>
/me admittedly did have to look on his web server over SFTP to work out what the URL of the tool was

[04:30:00.0000] <annevk3>
zcorpan, yes, and <image>...

[04:33:00.0000] <hsivonen>
what does Dead DOM Viewer do?

[04:33:01.0000] <Philip`>
It does the opposite of what the Live one does, hence the name

[04:34:00.0000] <Philip`>
In particular it generates the page on the server rather than on the client

[04:34:01.0000] <hsivonen>
Philip`: parses XML on the server and reseriales?

[04:34:02.0000] <Philip`>
The purpose was to compare IE's processing of HTML in HTTP responses vs HTML in document.write

[04:35:00.0000] <Philip`>
and then I added the XHTML version too, for which there isn't really any point having the server-side version

[04:35:01.0000] <Philip`>
hsivonen: No, the server just echoes its query string

[04:35:02.0000] <hsivonen>
hmm. <image> is magic in WebKit but in body--not directly in html

[04:36:00.0000] <hsivonen>
why do I see HTML, HEAD, BODY in the dead viewer in WebKit?

[04:36:01.0000] <hsivonen>
oops. in Firefox

[04:37:00.0000] <Philip`>
(http://canvex.lazyilluminati.com/misc/dom-viewer/ is the non-XHTML version)

[04:37:01.0000] <hsivonen>
it's as though the Live and Dead were swapped in Gecko and WebKit. what's going on?

[04:38:00.0000] <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

[04:40:00.0000] <Philip`>
It's just a quick hack and probably shouldn't be relied on for anything important :-)

[04:41:00.0000] <Philip`>
(Even the name doesn't make any sense, but I've never cared enough to try fixing it)

[04:44:00.0000] <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

[04:48:00.0000] <annevk3>
yeah, there should probably be [XXX] interface Window { } in Web Storag

[04:48:01.0000] <annevk3>
e

[04:48:02.0000] <annevk3>
same goes for localStorage/sessionStorage

[05:08:00.0000] <Lachy>
I'll go file a bug about it

[05:16:00.0000] <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

[05:17:00.0000] <Lachy>
I just filed the bug as an HTML5 spec bug instead

[07:19:00.0000] <hsivonen>
http://stackoverflow.com/questions/725261/what-is-the-semantic-web

[07:20:00.0000] <Dashiva>
The semantic web is like a box of chocolates

[07:32:00.0000] <jgraham>
More people misunderstanding <header>

[07:46:00.0000] <hsivonen>
<header> as currently defined, is FAIL

[07:46:01.0000] <hsivonen>
s/,//

[07:47:00.0000] <annevk3>
maybe we should just stick to a bunch of <div>s

[07:47:01.0000] <hsivonen>
The W3C uses RDFa in text/html: http://www.w3.org/2001/sw/

[07:51:00.0000] <hsivonen>
<header> should probably be split into <banner> (the case people think <header> is for) and <hgroup> for grouping <h1> and <h2>

[07:52:00.0000] <jgraham>
<banner> sounds like <advert>

[07:52:01.0000] <hsivonen>
jgraham: banner is from ARIA landmarks

[07:52:02.0000] <hsivonen>
jgraham: I don't particulary like the connotations of 'banner'

[07:52:03.0000] <jgraham>
hsivonen: It's not my fault if aria does dumb things :)

[07:53:00.0000] <jgraham>
Or rather I don't think we should feel compelled to spread those dumb things to the rest of html

[07:53:01.0000] <hsivonen>
jgraham: well, HTML5 sucks here, too

[07:53:02.0000] <jgraham>
Sure

[07:53:03.0000] <jgraham>
(Doesn't aria also have role="advertising" or something?)

[07:53:04.0000] <hsivonen>
which reminds me that now that April 1st has passed, I should blog in reply to the adactio post.

[07:54:00.0000] <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

[07:54:01.0000] <hsivonen>
jgraham: the semantics of banner were changed not to imply advertising

[07:54:02.0000] <zcorpan>
i wonder if the spec can be adjusted in some way

[07:55:00.0000] <zcorpan>
the page does <object id=ie><comment><object id=moz><table><tr><td>&nbsp;</table>

[07:55:01.0000] <annevk3>
zcorpan, why does it fail?

[07:55:02.0000] <zcorpan>
along with if (null != document.moz) ... else if (null != document.ie) ... else return // error

[07:55:03.0000] <jgraham>
hsivonen: Anyway HTML5 should probably have <header> which means role="banner" and <hgroup> which means "compound heading"

[07:55:04.0000] <annevk3>
ah, document.moz is not found?

[07:55:05.0000] <zcorpan>
both objects are ignored because they have "fallback"

[07:56:00.0000] <zcorpan>
right

[07:56:01.0000] <annevk3>
maybe the innermost <object> or <embed> should count and the rest be ignored?

[07:56:02.0000] <hsivonen>
jgraham: I think the words "header" and "heading" should be avoided in element naming

[07:57:00.0000] <jgraham>
hsivonen: Maybe. <banner> sucks at least as much though

[07:57:01.0000] <jgraham>
Or rather "Yes" 

[07:57:02.0000] <zcorpan>
annevk3: i.e. an object is fallback-free if it doesn't have any object or embed descendants?

[07:58:00.0000] <annevk3>
zcorpan, ja

[07:58:01.0000] <Philip`>
jgraham: What's the use case for <header> in that case?

[07:58:02.0000] <zcorpan>
makes sense i guess. what about <applet>? any other elements?

[07:59:00.0000] <annevk3>
<applet name=x> <object name=x> </object> </applet> both are returned, no?

[07:59:01.0000] <zcorpan>
i was thinking about applet in object

[07:59:02.0000] <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

[08:00:00.0000] <hsivonen>
it would also be nice if Opera Mobile took it into account when guessing where the main content starts

[08:00:01.0000] <annevk3>
zcorpan, good point

[08:01:00.0000] <Philip`>
jgraham: <div id=header> lets them structure their source in the same way, with the same readability

[08:02:00.0000] <annevk3>
zcorpan, <img> too

[08:02:01.0000] <zcorpan>
annevk3: not sure about <img>

[08:02:02.0000] <Philip`>
jgraham: Oh, if it maybe helps with a11y then we are morally required to include it

[08:03:00.0000] <zcorpan>
annevk3: consider that page having a spacer gif instead of &nbsp;

[08:04:00.0000] <annevk3>
zcorpan, it seems that for Mozilla <applet>, <img> etc. do not matter

[08:04:01.0000] <annevk3>
but neither does <embed>

[08:04:02.0000] <annevk3>
or <object>

[08:05:00.0000] <annevk3>
sigh

[08:05:01.0000] <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>

[08:06:00.0000] <jgraham>
(similarly div#header is slightly harder to read in CSS than header)

[08:07:00.0000] <zcorpan>
annevk3: moz does something like if (document.getElementById(x) && blah) else blah for document.x

[08:07:01.0000] <beowulf>
i like <banner>

[08:07:02.0000] <webben>
beowulf: banner is very ambiguous

[08:07:03.0000] <Philip`>
jgraham: Try using indentation and quotes and then it'll be fine :-)

[08:07:04.0000] <webben>
e.g. sometimes used for ads

[08:08:00.0000] <Philip`>
But people won't use <banner> for ads because it would be trivial to filter

[08:08:01.0000] <annevk3>
zcorpan, what is && blah?

[08:08:02.0000] <Philip`>
therefore it'll be used only by people writing page headers

[08:09:00.0000] <beowulf>
webben: probably, though not in my case, i've been reading about newspaper layout, terms, whatnot

[08:10:00.0000] <jgraham>
Philip`: Indentation helps a bit. But it's still easier to read if the end tag matches the start tag

[08:10:01.0000] <jgraham>
(otherwise everyone wwould love lisp :) )

[08:10:02.0000] <beowulf>
and <banner> is probably not used for ads or anything right now :)

[08:10:03.0000] <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

[08:10:04.0000] <Philip`>
jgraham: s/lisp/Python/

[08:10:05.0000] <zcorpan>
annevk3: && document.getElementById(x) is one of img, embed, etc

[08:11:00.0000] <Philip`>
jgraham: and everyone does love Python

[08:11:01.0000] <Philip`>
jgraham: therefore disproving your argument

[08:11:02.0000] <Philip`>
/me wins

[08:11:03.0000] <jgraham>
Philip`: No because python doesn't have end tags

[08:11:04.0000] <annevk3>
zcorpan, I can't really reproduce that

[08:11:05.0000] <jgraham>
or whatever

[08:11:06.0000] <Philip`>
jgraham: It does have end tags - they're just invisible DEDENT tokens

[08:11:07.0000] <jgraham>
:p

[08:12:00.0000] <jgraham>
Philip`: Right but they're not in the source

[08:12:01.0000] <jgraham>
Er, that made no sense

[08:12:02.0000] <Philip`>
jgraham: How can you tell the invisible tokens aren't in the source?

[08:12:03.0000] <Philip`>
jgraham: They're invisible!

[08:12:04.0000] <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

[08:13:00.0000] <Philip`>
jgraham: It's possible that I'm completely misunderstanding your argument

[08:13:01.0000] <webben>
/me is getting a vision of a lolcat version of this "invisible end tag" argument.

[08:13:02.0000] <jgraham>
I think what I mean is that the python philosophy of uniform indentation as a marker doesn't transfer well onto the web

[08:14:00.0000] <jgraham>
Because the web doesn't enforce uniform indentation

[08:14:01.0000] <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

[08:15:00.0000] <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)

[08:15:01.0000] <Philip`>
s/people/I/

[08:16:00.0000] <Philip`>
Anyway, I've forgotten how this discussion started or what I was trying to say

[08:16:01.0000] <Philip`>
s/or/and/

[08:16:02.0000] <annevk3>
zcorpan, maybe reverse engineering WebKit is more sane or does it also fail on that page?

[08:16:03.0000] <zcorpan>
annevk3: webkit also fails

[08:16:04.0000] <zcorpan>
annevk3: webkit, opera and html5 do basically the same thing

[08:18:00.0000] <hsivonen>
whee! "The XPath evaluator must never call this with a null or empty argument, because the result of doing this is undefined."

[08:18:01.0000] <jgraham>
hsivonen: Is that a normative condition on user code?

[08:18:02.0000] <jgraham>
s/user/web developer/

[08:18:03.0000] <annevk3>
zcorpan, yeah. I guess doing what Firefox does breaks some other set of pages?

[08:19:00.0000] <hsivonen>
jgraham: it's a normative statement on XPath implentations, so no

[08:19:01.0000] <zcorpan>
annevk3: likely. it's also less sane and less compatible with ie

[08:19:02.0000] <zcorpan>
last i looked

[08:19:03.0000] <jgraham>
hsivonen: Ah, OK

[08:19:04.0000] <hsivonen>
jgraham: though it seems to me calling into user code can be undefined :-)

[08:19:05.0000] <hsivonen>
in any case :-)

[08:20:00.0000] <Philip`>
But is it defined that the user code must follow certain rules?

[08:20:01.0000] <annevk3>
zcorpan, ignoring <object> if it has a descendant named element maybe?

[08:20:02.0000] <hsivonen>
/me wonders what happens if XPathNSResolver returns the empty string

[08:20:03.0000] <annevk3>
zcorpan, it seems a bit risky though

[08:23:00.0000] <annevk3>
zcorpan, and it's hard to get good data on this though I suppose we could try to get anedotical data

[08:23:01.0000] <annevk3>
anecdotal, even

[08:23:02.0000] <zcorpan>
i'm all for doing research :)

[08:27:00.0000] <annevk3>
maybe Philip` can find pages that have <object id=... or name=...> + child nodes

[08:28:00.0000] <zcorpan>
ignoring <param> and whitespace

[08:28:01.0000] <zcorpan>
Philip`: ^

[08:31:00.0000] <hsivonen>
do markers ever come off the list of active formatting elements in ways other than via "clear up to last marker"?

[08:41:00.0000] <Philip`>
annevk3: I can find <object id> and <object name> easily

[08:50:00.0000] <hsivonen>
aaargh. Gecko allows an XPathNSResolver to return "" to bind a prefix to no namespace

[08:51:00.0000] <hsivonen>
doesn't work in WebKit

[08:52:00.0000] <hsivonen>
nor in Opera

[08:52:01.0000] <hsivonen>
maybe there's hope for zapping that behavior from Gecko, too.

[08:53:00.0000] <zcorpan>
Philip`: can you easily find <object id/name> that have child elements other than <param>?

[08:54:00.0000] <Philip`>
annevk3: Actually, I could find them easily if I had a decent XML grep tool that actually worked

[08:54:01.0000] <Philip`>
but I don't

[08:55:00.0000] <Philip`>
so you can have http://philip.html5.org/data/object-with-id-or-name.xml.bz2 which is all <object>s

[08:55:01.0000] <Philip`>
and feel free to write code to extract what you want from it

[08:55:02.0000] <Philip`>
(That's directed to zcorpan too :-) )

[08:58:00.0000] <zcorpan>
Philip`: thanks

[08:59:00.0000] <Philip`>
/me hopes the structure of that file is sufficiently obvious

[08:59:01.0000] <Philip`>
(Hooray for human-readable file formats)

[09:05:00.0000] <zcorpan>
hmm bad idea to open that xml file in a browser

[09:06:00.0000] <Philip`>
It's only 20MB

[09:06:01.0000] <Philip`>
Oh, but I suppose it has a few <object>s

[09:06:02.0000] <zcorpan>
almost 27000 object elements

[09:06:03.0000] <Philip`>
which browsers sillily try to interpret

[09:06:04.0000] <Philip`>
despite it not even being in an <html>

[09:07:00.0000] <zcorpan>
/me removes namespace declarations

[09:07:01.0000] <Philip`>
/me didn't add namespace declarations on purpose, they were just what his XML serialiser did

[09:07:02.0000] <zcorpan>
oops, my script didn't escape <

[09:08:00.0000] <zcorpan>
ok so 2600 don't have an embed descendant

[09:09:00.0000] <zcorpan>
<object classid="clsid:d27cdb6e-ae6d-11cf-96b8-444553540000" height="475" id="mymovie" width="830">

[09:09:01.0000] <zcorpan>
        <param name="movie" value="../images/home.swf"/>

[09:09:02.0000] <zcorpan>
        <object data="../images/home.swf" height="475" type="application/x-shockwave-flash" width="830">

[09:09:03.0000] <zcorpan>
          <p>

[09:09:04.0000] <zcorpan>
<img alt="Our applications fields : Veterinary Lab, Haematology and Pathology" height="475" src="../images/homepics.jpg" width="830"/>

[09:11:00.0000] <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>

[09:13:00.0000] <zcorpan>
298 have nested object elements

[09:26:00.0000] <zcorpan>
copy-paste has resulted in lots of "FAIL (the browser should render some flash content, not this)."

[09:27:00.0000] <zcorpan>
Hixie: take note - don't write copy-paste examples as test cases

[09:28:00.0000] <zcorpan>
/me will continue tomorrow

[09:32:00.0000] <Philip`>
Where does that FAIL come from originally?

[09:32:01.0000] <Philip`>
All I see is http://www.adobe.com/jp/devnet/flash/articles/swfobject_05.html

[09:49:00.0000] <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

[10:13:00.0000] <jgraham>
/me notes that hsivonen is probably the first person in the history of all creation to be aused by XPath

[10:13:01.0000] <jgraham>
*amused

[10:19:00.0000] <Hixie>
/me gets off the phone with a yahoo searchmonkey engineer

[10:19:01.0000] <Hixie>
who had very interesting insights into rdf in html

[10:20:00.0000] <gsnedders>
Wait, what?

[10:20:01.0000] <gsnedders>
"Hixie gets off the phone" — Does not compute.

[10:20:02.0000] <Hixie>
i had to come to work and go to my cube to do this

[10:20:03.0000] <jgraham>
Hixie: Are you just teasing us with confidential information?

[10:20:04.0000] <Hixie>
first time i've been here in months

[10:20:05.0000] <Hixie>
jgraham: no

[10:21:00.0000] <gsnedders>
Hixie: I guessed you were at work from the lack of a phone anywhere else :P

[10:21:01.0000] <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

[10:22:00.0000] <jgraham>
So far, so few surprises

[10:22:01.0000] <Hixie>
and that the lack of a clear mapping to RDF actually did hurt their use of microformats

[10:22:02.0000] <gsnedders>
GRDDL!

[10:22:03.0000] <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

[10:23:00.0000] <jgraham>
gsnedders: Doesn't GRDDL have some weird thing that the author needs to provide the GRDDL?

[10:23:01.0000] <gsnedders>
I dunno.

[10:23:02.0000] <gsnedders>
I thought it was something along the lines of @profile being magic

[10:23:03.0000] <jgraham>
/me thought that GRDDL was basically a way of using an XSLT stylesheet to transform some HTML into some RDF

[10:23:04.0000] <gsnedders>
Yeah, basically

[10:24:00.0000] <jgraham>
And that the author had to put the link to the sty;esheet in @profile

[10:24:01.0000] <Hixie>
no

[10:24:02.0000] <Hixie>
they have to put a link to a page that has a link to the xslt

[10:25:00.0000] <jgraham>
Ah, yes, I guess it it important to be correct on that point. @profile is an option for the link :)

[10:26:00.0000] <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

[10:26:01.0000] <jgraham>
?

[10:27:00.0000] <gsnedders>
Are free video downloads from iTunes encumbered?

[10:36:00.0000] <Philip`>
jgraham: And by something with "a clear mapping to RDF"

[10:38:00.0000] <jgraham>
Philip`: Good point

[10:49:00.0000] <Hixie>
jgraham: yeah

[10:51:00.0000] <jgraham>
Good luck convincing W3C that URIs are part of the problem, not part of the solution :)

[10:52:00.0000] <Hixie>
no need, so long as we define a way to map strings to URIs without having to declare prefixes explicitly

[10:53:00.0000] <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

[10:53:01.0000] <jgraham>
It also seems like the only plausible way forward

[11:06:00.0000] <Hixie>
i don't understand what http://www.w3.org/mid/b6bb4d890904070441p248bedccsaf3f2971f282c18⊙mgc would mean, concretely

[11:11:00.0000] <Philip`>
Something about taking fancy new web technologies and making them usable by normal web developers, I guess

[11:11:01.0000] <Philip`>
though that's still not exactly concrete

[11:14:00.0000] <Philip`>
Also it might be talking about making HTML more attractive to teenage girls

[11:16:00.0000] <webben>
Looks like it just means deliver the key features needed for webapps fast.

[11:16:01.0000] <mpilgrim>
i think it's more along the lines of "we need a site like chromeexperiments.com to showcase html 5"

[11:16:02.0000] <mpilgrim>
html5experiments.com?

[11:16:03.0000] <mpilgrim>
i think that's just called annevankesteren.nl

[11:20:00.0000] <Philip`>
That's the catwalk

[11:22:00.0000] <gsnedders>
mpilgrim: But he does SVG too! That's an SVG experiment!

[11:24:00.0000] <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

[11:29:00.0000] <jgraham>
Compsring html5 to topshop? This must be some sort of new low

[11:30:00.0000] <gsnedders>
I thought we had already got as low as we can go.

[11:32:00.0000] <jgraham>
Do we have any examples of really crappy but popular html 5? We could call them the html 5 "new look"

[11:33:00.0000] <jgraham>
Straw poll: "data is" or "data are"?

[11:33:01.0000] <gsnedders>
is

[11:34:00.0000] <gsnedders>
(Even though I'm well-aware that's technically incorrect.)

[11:34:01.0000] <Philip`>
Avoid phrases that force you to make the distinction, because you'll annoy somebody no matter which you choose

[11:35:00.0000] <Philip`>
('is' will annoy people because it's wrong, 'are' will annoy people because it's stupid and ugly)

[11:35:01.0000] <inimino>
depends if you are treating data as a plural or as a mass noun

[11:35:02.0000] <jgraham>
Philip`: The hidden context is teaching english to non-native speakers

[11:35:03.0000] <gsnedders>
data is the plural of datum, so it should never be singular

[11:36:00.0000] <Philip`>
Have you ever heard anybody use the word "datum"?

[11:36:01.0000] <jcranmer>
I've used it once or twice

[11:36:02.0000] <jgraham>
gsnedders: I don't think that is a good argument becaue it doesn't preclude data also being used as a mass noun

[11:36:03.0000] <gsnedders>
Philip`: Yes

[11:37:00.0000] <Philip`>
particularly somebody who isn't sounding unbelievably pretentious when saying it? ;-)

[11:37:01.0000] <gsnedders>
Philip`: Um, no.

[11:37:02.0000] <gsnedders>
Philip`: But I rarely hear people who don't sound unbelievably pretentious

[11:37:03.0000] <jcranmer>
I believe ArcGIS uses datum

[11:37:04.0000] <jgraham>
gsnedders: Just by speaking?

[11:37:05.0000] <gsnedders>
jgraham: yeah

[11:38:00.0000] <jgraham>
gsnedders: You have really chavvy standards :)

[11:38:01.0000] <gsnedders>
:D

[11:38:02.0000] <gsnedders>
jgraham: And dude, chav is _so_ last year.

[11:38:03.0000] <jgraham>
gsnedders: Probably depends where you are from a bit

[11:39:00.0000] <Philip`>
gsnedders: "_so_" is so last year, too

[11:39:01.0000] <gsnedders>
(Well, really it's several years ago here.)

[11:39:02.0000] <gsnedders>
It's all about the neds nowadays

[11:40:00.0000] <jcranmer>
... ばか

[11:40:01.0000] <gsnedders>
Huh?

[11:40:02.0000] <gsnedders>
UnknownLanguageException

[11:41:00.0000] <jcranmer>
can't recognize hiragana, can you?

[11:41:01.0000] <gsnedders>
/me catches the exception, and carries on

[11:41:02.0000] <Philip`>
noun  1. bonehead [sl.]  2. clod [sl.]  3. stupidity

[11:42:00.0000] <Philip`>
Being unable to translate between languages is _so_ last year

[11:43:00.0000] <jcranmer>
...

[11:43:01.0000] <jcranmer>
あほ

[11:44:00.0000] <Philip`>
"Idiot"?

[11:44:01.0000] <Philip`>
Google Translate isn't so good on that

[11:45:00.0000] <jcranmer>
closer to asshole, IIRC

[11:46:00.0000] <jcranmer>
much stronger form of ばか

[11:47:00.0000] <Philip`>
/me 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

[11:47:01.0000] <Philip`>
which includes "Idiot, right (State of Arizona)" - I'm not sure what that's meant to mean

[11:48:00.0000] <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)."

[11:49:00.0000] <gsnedders>
What's .ta?

[11:49:01.0000] <jcranmer>
`Ajo' sounds like it's from spanish origins

[11:49:02.0000] <jcranmer>
so it's pronunciation is like `Aho'

[11:49:03.0000] <jcranmer>
or, should I say, ah-ho

[11:50:00.0000] <jcranmer>
transliterate that into Japanese to get, in the rōmaji form aho

[11:50:01.0000] <jcranmer>
あほ == aho

[11:50:02.0000] <Philip`>
I guess it's http://en.wikipedia.org/wiki/Ajo,_Arizona but that page doesn't talk about idiots at all

[11:50:03.0000] <jcranmer>
google's not picking up the fact that it's a transliterations

[11:50:04.0000] <gsnedders>
Meh. It's meant to be raining in Dundee for the next couple of days.

[11:51:00.0000] <jcranmer>
yep

[11:52:00.0000] <jcranmer>
the wikipedia page has it in katakana--which is generally how you write transliterated proper names in Japanese

[11:52:01.0000] <jcranmer>
so あほ and アホ are both aho, but the former would be the word aho and the latter the place Aho

[11:53:00.0000] <jcranmer>
or the name, if someone were cruel enough to name someone that :-)

[11:53:01.0000] <jgraham>
gsnedders: Move to Fort William. Then it would always be raining so you wouldn't get so disappointed

[11:53:02.0000] <gsnedders>
/me went to Fort William once…

[11:53:03.0000] <gsnedders>
It wasn't raining.

[11:53:04.0000] <gsnedders>
Also: I doubt Fort William has anywhere to buy a suit from.

[11:55:00.0000] <jgraham>
gsnedders: I think it does. But I only went there once too (and I think it was raining)

[11:56:00.0000] <gsnedders>
(If I didn't need to get a suit, I would have no reason to go to Dundee.)

[11:56:01.0000] <gsnedders>
It's no better in Edinburgh, so I don't gain anything by going their instead.

[11:56:02.0000] <jgraham>
gsnedders: There are places that will sell suits where it is unlikely to be raining

[11:57:00.0000] <jgraham>
Saudia Arabia for example

[11:57:01.0000] <gsnedders>
jgraham: They are not however practical places to go to get a suit

[11:57:02.0000] <erlehmann>
can anyone with working ff 3.5 or opera 10 test this ? http://blog.dieweltistgarnichtso.net/interview-moot-of-4chan-part-1

[11:57:03.0000] <takkaria>
I daresay Saudia Arabia is a more practical place than say, Devon

[11:58:00.0000] <gsnedders>
Wednesday afternoon is meant to be all right in Dundee.

[11:58:01.0000] <takkaria>
or Blackpool

[11:58:02.0000] <gsnedders>
takkaria: Why?

[11:58:03.0000] <takkaria>
well, Blackpool's a bit naff really, isn't it?

[11:58:04.0000] <gsnedders>
Dunno. Never been there.

[11:59:00.0000] <erlehmann>
got it — the server must serve the file using the correct MIME type

[11:59:01.0000] <jgraham>
gsnedders: On he other hand if the only options are dundee or Edinburgh, at least Edinburgh is nice

[12:00:00.0000] <gsnedders>
That is true.

[12:01:00.0000] <gsnedders>
Also, Dundee wastes less time

[12:59:00.0000] <gsnedders>
erlehmann: "spending […] every" should be "spending… I was up every"

[12:59:01.0000] <erlehmann>
thanks

[12:59:02.0000] <gsnedders>
His speech is too quick

[13:01:00.0000] <gsnedders>
erlehmann: "site […] seven servers," "site. I rearranged seven servers,"

[13:03:00.0000] <erlehmann>
gsnedders, actually, i did have the impression of understanding him when sitting on the opposing side of the desk

[13:04:00.0000] <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 :)

[13:04:01.0000] <erlehmann>
i hate AMR now

[13:04:02.0000] <erlehmann>
another reason to root my new googlephone

[13:11:00.0000] <Philip`>
http://hg.mozilla.org/mozilla-central/rev/cfd5af38abc6

[14:13:00.0000] <gsnedders>
/me guesses he ought to let the outliner just take a textarea

[14:13:01.0000] <gsnedders>
Seemingly multiple people ahve asked for it

[14:26:00.0000] <Lachy>
MikeSmith, yt?

[16:27:00.0000] <annevk3>
yay email


2009-04-08
[17:01:00.0000] <kimus>
hi, can I strip unwanted tags with html5lib?

[17:01:01.0000] <kimus>
i'm testing  it and the max I could get is a escaped tag

[17:03:00.0000] <annevk5>
you could try filtering the DOM

[17:03:01.0000] <kimus>
hi annevk5, humm, ... that sucks :-)

[17:04:00.0000] <annevk5>
I don't know much about the current feature set of html5lib though

[17:05:00.0000] <kimus>
I just wanted a python module to 'normalize' user entered HTML 

[17:05:01.0000] <gsnedders>
I don't think there is any easy way, kimus 

[17:05:02.0000] <gsnedders>
I could be totally wrong

[17:06:00.0000] <kimus>
gsnedders: http://htmlpurifier.org/

[17:06:01.0000] <kimus>
works fine for php :-)

[17:06:02.0000] <gsnedders>
I mean in python html5lib

[17:07:00.0000] <annevk5>
writing a function that filters the DOM cannot be that hard

[17:08:00.0000] <annevk5>
though it depends on how complex you want to make it of course :)

[17:09:00.0000] <kimus>
annevk5: it works fine like it's know... I just dont want to escape not allowed tags

[17:10:00.0000] <kimus>
not allowed attributes are removed

[17:11:00.0000] <kimus>
html5lib removes not allowed attributes... but there's no option to remove not allowed tags :-S

[17:16:00.0000] <kimus>
i'm going to change the sanitize_token

[17:17:00.0000] <kimus>
in my class implementation

[17:18:00.0000] <kimus>
annevk5 and gsnedders: done :-D

[17:19:00.0000] <gsnedders>
:P

[17:24:00.0000] <kimus>
annevk5 and gsnedders: http://paste.ubuntu.com/146546/

[17:24:01.0000] <kimus>
was not hard... but...

[17:51:00.0000] <gsnedders>
Hixie: When are you planning on doing ref now?

[00:03:00.0000] <MikeSmith>
Hixie: <keygen> needs to be added to the list of void elements in 8.1.2

[00:05:00.0000] <Hixie>
fixed

[00:08:00.0000] <MikeSmith>
k

[00:08:01.0000] <MikeSmith>
Hixie: there are not additional constraints on the value of the challenge attribute? any string?

[00:10:00.0000] <MikeSmith>
s/not additional constraints/no additional constraints/

[00:16:00.0000] <Hixie>
MikeSmith: no, why would there?

[00:17:00.0000] <MikeSmith>
Hixie: dunno. because I have no idea what it really is or what normal values people/apps would have for it

[00:18:00.0000] <Hixie>
it's just a string used to salt the SPKAC structure

[00:18:01.0000] <Hixie>
presumably to prevent replay attacks, though i don't really understand what the replay attack would be

[00:18:02.0000] <Hixie>
since you wouldn't have the private key...

[00:19:00.0000] <MikeSmith>
I see

[00:20:00.0000] <MikeSmith>
/me reads blooberry's http://www.blooberry.com/indexdot/html/tagpages/k/keygen.htm now

[00:20:01.0000] <Hixie>
that isn't particularly detailed

[00:20:02.0000] <MikeSmith>
I'm looking at the links

[00:20:03.0000] <Hixie>
good luck!

[00:20:04.0000] <MikeSmith>
it really sucks that AOL shut down developer.netscape.com without redirecting anything

[00:21:00.0000] <MikeSmith>
idiots

[00:21:01.0000] <MikeSmith>
ah great

[00:21:02.0000] <MikeSmith>
all of wp.netscape.com too

[00:21:03.0000] <MikeSmith>
they get a gold star for consistency at least

[00:25:00.0000] <Hixie>
bed time

[00:25:01.0000] <Hixie>
nn

[00:26:00.0000] <zcorpan>
morning!

[00:27:00.0000] <zcorpan>
<p>Alternative content</p> - http://nord-trondelag.sv.no/

[00:27:01.0000] <zcorpan>
yay for copy-paste

[00:29:00.0000] <MikeSmith>
hsivonen: (or anybody) what ARIA attribute set would be appropriate for keygen?

[00:29:01.0000] <zcorpan>
MikeSmith: tabindex?

[00:31:00.0000] <MikeSmith>
zcorpan: ARIA has a tabindex attribute?

[00:31:01.0000] <MikeSmith>
tabindex is a global attribute anyway, right?

[00:33:00.0000] <zcorpan>
don't remember if tabindex is global in aria

[00:33:01.0000] <zcorpan>
but yeah

[00:34:00.0000] <zcorpan>
hmm maybe they removed tabindex from aria

[00:34:01.0000] <zcorpan>
"An implementing host language MUST provide support for the author to make arbitrary elements focusable, such as the tabindex attribute in HTML."

[00:34:02.0000] <MikeSmith>
that makes sense

[00:35:00.0000] <MikeSmith>
zcorpan: I meant from among the sets that hsivonen already has in the schema

[00:35:01.0000] <MikeSmith>
http://www.w3.org/html/wg/markup-spec/#implicit-aria

[00:35:02.0000] <MikeSmith>
maybe http://www.w3.org/html/wg/markup-spec/#common.attrs.aria.implicit.input

[00:35:03.0000] <zcorpan>
earlier tabindex was specified in aria because they wanted to use html4 and html5 was so long away from rec so not relevant

[00:35:04.0000] <MikeSmith>
ah

[00:36:00.0000] <MikeSmith>
so as far as keygen, isn't it essentially a specialized input control for submitting a public key?

[00:37:00.0000] <zcorpan>
wasn't "rude" removed from aria?

[00:37:01.0000] <MikeSmith>
zcorpan: rude?

[00:37:02.0000] <zcorpan>
aria-live = "off" | "polite" | "assertive" | "rude"

[00:38:00.0000] <MikeSmith>
wow

[00:53:00.0000] <MikeSmith>
zcorpan: per spec, <keygen keytype=""> should be valid, right? (because it defaults to "RSA")

[00:54:00.0000] <MikeSmith>
http://www.whatwg.org/specs/web-apps/current-work/multipage/forms.html#attr-keygen-keytype

[00:55:00.0000] <MikeSmith>
hmm

[00:55:01.0000] <MikeSmith>
"If an enumerated attribute is specified, the attribute's value must be an ASCII case-insensitive match for one of the given keywords that are not said to be non-conforming, with no leading or trailing whitespace."

[00:56:00.0000] <MikeSmith>
so keytype="" is not conforming

[01:01:00.0000] <hsivonen>
MikeSmith: not really knowing: would the closest analogue to keygen ARIA-wise input type=file?

[01:01:01.0000] <MikeSmith>
hsivonen: yeah, I suppose

[01:01:02.0000] <hsivonen>
MikeSmith: in any case, keygen is so special that overriding the role makes no sense

[01:02:00.0000] <MikeSmith>
hsivonen: OK, so no need to put any ARIA attributes on it at all?

[01:05:00.0000] <MikeSmith>
Philip`: about https://bugzilla.mozilla.org/attachment.cgi?id=371594

[01:05:01.0000] <MikeSmith>
oops

[01:05:02.0000] <MikeSmith>
https://bugzilla.mozilla.org/show_bug.cgi?id=477564

[01:07:00.0000] <MikeSmith>
might be useful to have some data on what percentage of forms have hundreds of checkboxes

[01:07:01.0000] <MikeSmith>
https://bugzilla.mozilla.org/show_bug.cgi?id=477564#c14

[01:08:00.0000] <MikeSmith>
"Thankfully, pages with 1000 checkboxes are few and far between."

[01:08:01.0000] <MikeSmith>
suspect that may not be so true

[01:08:02.0000] <MikeSmith>
at least I'd think there are a significant number with several hundreds of checkboxes

[01:16:00.0000] <Philip`>
MikeSmith: The most I see is 230

[01:17:00.0000] <Philip`>
MikeSmith: and there's 5 with >100

[01:17:01.0000] <MikeSmith>
Philip`: OK

[01:17:02.0000] <MikeSmith>
thanks

[01:17:03.0000] <Philip`>
/me wonders if it's worth commenting on that in that bug

[01:18:00.0000] <MikeSmith>
Philip`: worth it if you care about the bug at all, not worth it if you don't

[01:18:01.0000] <MikeSmith>
I guess

[01:18:02.0000] <MikeSmith>
not sure how much I care about

[01:19:00.0000] <MikeSmith>
seems like anybody designing forms with 500 or more checkboxes shouldn't be doing that anyway

[01:19:01.0000] <MikeSmith>
if they care about usability at all

[01:21:00.0000] <Philip`>
MikeSmith: A UI like http://www.sciencedirect.com/science/journals doesn't seem too unusable

[01:22:00.0000] <MikeSmith>
Philip`: hmm, and loads much faster than the HTML5 spec, at least

[01:25:00.0000] <hsivonen>
MikeSmith: either no ARIA or only states and properties that make sense if some of them make sense

[01:25:01.0000] <hsivonen>
MikeSmith: the real answer depends on what the Task Force is doing

[01:25:02.0000] <zcorpan>
i could imagine a single-page survey with a table and lots of checkboxes

[01:26:00.0000] <hsivonen>
I wonder if the TF has keygen on their tables

[01:26:01.0000] <zcorpan>
Statement | Fully agree | Partially agree | Neither agree nor disagree | Partially disagree | Fully disagree | Don't know

[01:27:00.0000] <zcorpan>
6 times the number of statements, and surveys often have hundreds of statements

[01:27:01.0000] <MikeSmith>
hsivonen: how about I omit any ARIA attributes from keygen schema for now, but put a "#REVISIT: ARIA attributes" comment in?

[01:27:02.0000] <MikeSmith>
zcorpan: yeah, true

[01:27:03.0000] <hsivonen>
MikeSmith: seems ok

[01:28:00.0000] <MikeSmith>
zcorpan: but better design is to put them into multiple pages

[01:28:01.0000] <MikeSmith>
I would think

[01:28:02.0000] <MikeSmith>
hsivonen: thanks

[01:28:03.0000] <zcorpan>
MikeSmith: why?

[01:29:00.0000] <zcorpan>
MikeSmith: it might be on the same page but on different slides implemented with javascript show/hide

[01:29:01.0000] <MikeSmith>
yeah, I suppose

[01:30:00.0000] <MikeSmith>
but I would hate to be in the middle of a hundreds-question survey and have my browser crash

[01:30:01.0000] <MikeSmith>
for one thing

[01:30:02.0000] <MikeSmith>
so that I have to go back and re-answer them all

[01:30:03.0000] <zcorpan>
MikeSmith: that's why firefox implemented the feature to restore the answers, no?

[01:30:04.0000] <MikeSmith>
or have some problem during form submission

[01:30:05.0000] <MikeSmith>
yeah, but I mean in general

[01:31:00.0000] <MikeSmith>
as far as designing across browsers

[01:31:01.0000] <zcorpan>
the page could do ajaxy autosave

[01:31:02.0000] <Philip`>
zcorpan: Those would be radio buttons, not checkboxes

[01:32:00.0000] <zcorpan>
Philip`: right

[01:34:00.0000] <MikeSmith>
hsivonen: after I added assertions-checking for img@usemap as a descendant of a|bb|button, the following test case is no longer valid:

[01:34:01.0000] <MikeSmith>
http://syntax.whattf.org/relaxng/tests/html5core/valid/043.html

[01:36:00.0000] <MikeSmith>
hsivonen: seems like it's not necessary for the img instances in that particular test to have a usemap attribute

[01:37:00.0000] <MikeSmith>
so may I should just remove the usemap attributes there?

[01:37:01.0000] <zcorpan>
or add tests that use usemap but not <a> ancestor

[01:38:00.0000] <MikeSmith>
zcorpan: yah, was now just thinking the same

[01:38:01.0000] <MikeSmith>
but not add, just change that existing test to not use <a> ancestors 

[01:38:02.0000] <zcorpan>
not sure what it tests, the comments seem to be using pre-content model overhaul terminology

[01:39:00.0000] <zcorpan>
MikeSmith: it also has ismap which requires <a> ancestor

[01:39:01.0000] <MikeSmith>
ah

[01:41:00.0000] <zcorpan>
<a href><img ismap><img ismap></a> seems to be allowed per spec

[01:43:00.0000] <hsivonen>
MikeSmith: ok to remove the usemap

[01:44:00.0000] <MikeSmith>
hsivonen: k

[01:48:00.0000] <MikeSmith>
hmm, should not object@usemap also be classified as interactive content?

[01:51:00.0000] <hsivonen>
MikeSmith: topic notwithstanding, yes if img@usemap is classified as interactive

[01:52:00.0000] <MikeSmith>
OK, I'll raise in bugzilla

[01:53:00.0000] <MikeSmith>
hsivonen: for the img@usemap test, I'll delete from the existing test and just add another new test that has img with usemap but not with ismap

[01:54:00.0000] <hsivonen>
MikeSmith: OK. the existing tests often test too many things in one file

[01:54:01.0000] <MikeSmith>
yeah

[01:55:00.0000] <MikeSmith>
I guess I could just reduce the new test to a simple test that has a map and an img with usemap

[02:28:00.0000] <jgraham>
/me has a page with several thousand checkboxes

[02:28:01.0000] <jgraham>
Although it is not finished yet so the idea might not work in practice

[02:46:00.0000] <zcorpan>
/me is having difficulties understanding what html5 says should happen for <frameset rows=",86">

[02:46:01.0000] <MikeSmith>
hsivonen: any clues as to why the same test case is in both tests/html5full-xhtml/valid/043.xhtml and in tests/html5core/valid/043.xhtml ?

[04:30:00.0000] <MikeSmith>
hsivonen: patch for htmlparser handling for <keygen> -

[04:30:01.0000] <MikeSmith>
http://bugzilla.validator.nu/attachment.cgi?id=76

[04:31:00.0000] <MikeSmith>
tested and works as expected (my local v.nu recognizes <keygen> as a void element)

[04:32:00.0000] <MikeSmith>
but I realize maybe some handling also needs to be added to cases other than IN_BODY

[04:33:00.0000] <MikeSmith>
(though that case is the only one for which the spec makes specific mention of <keygen>)

[05:14:00.0000] <annevk42>
Philip`, is it easy to find more pages with broken Location headers?

[05:16:00.0000] <annevk42>
Philip`, it would be cool to have a "out of X amount of links Y amount of links had a Location header and of those Z amount of links had a broken Location header (relative, invalid character). List links that with broken headers: ..."

[05:29:00.0000] <jgraham>
http://intertwingly.net/blog/2009/04/08/HTML-Reunification

[05:30:00.0000] <annevk42>
Browser vendors no longer like HTML5?

[05:33:00.0000] <hsivonen>
I'm very annoyed by the proposal to break DOM Consistency for "language extensions".

[05:33:01.0000] <hsivonen>
why do "language extensions" have to go poke the "here be dragons" area? can't extensions simply not dabble with colons?

[05:34:00.0000] <annevk42>
And how does ubiquity-xforms qualify as language extension? Surely non-browsers would have to implement it natively short of implementing an AI-complete algorithm to figure out how it works from the attached ECMAScript library?

[05:35:00.0000] <jgraham>
Heh, I thought hsivonen would like that :)

[05:36:00.0000] <hsivonen>
annevk42: as far as I can tell, ubiquity-xforms needs to use colons only to keep up the appearances of the XForms syntax going strong

[05:36:01.0000] <jgraham>
annevk42: The presence of js libraries seems to blur the distinction between langauge extension and platform extension considerably

[05:36:02.0000] <jgraham>
e.g. <canvas> can be implemented in js for IE

[05:37:00.0000] <annevk42>
The inevitable result of this is everyone sending proprietary markup over the wire.

[05:38:00.0000] <hsivonen>
it seems to me that it's unsafe to assume that successful Firefox extension functionality wouldn't migrate to the core feature sets of browsers

[05:38:01.0000] <annevk42>
That "works" because of a bunch of ECMAScript libraries.

[05:38:02.0000] <jgraham>
I am kindof saddened that we have resorted to "consenus" politics being more important than the technical arguments that have been made

[05:38:03.0000] <hsivonen>
jgraham: oh, indeed

[05:39:00.0000] <annevk42>
hsivonen, true, but we're far way from something interoperable I think

[05:39:01.0000] <hsivonen>
annevk42: I mean the argument is that RDFa is a "language extension" that Firefox extensions could be sensitive to

[05:40:00.0000] <annevk42>
ah yeah, you mean it would make it a platform extension down the road?

[05:40:01.0000] <hsivonen>
annevk42: if it that setup becomes successful, why wouldn't the client-side feature migrate into Firefox itself or into the native feature set of other browsers that don't support Firefox extensions but need to support successful features?

[05:40:02.0000] <jgraham>
hsivonen: Indeed

[05:41:00.0000] <jgraham>
Although I think that is in some sense the goal

[05:41:01.0000] <jgraham>
Albeit not explicitly stated

[05:41:02.0000] <hsivonen>
if it's the goal, then the extension should play by the rules that are needed to avoid sucky browser core code down the road

[05:42:00.0000] <annevk42>
I hadn't reached the part around "It needs to be noted that such nodes are placed into the DOM today differently by HTML and XML parsers." yet. That's seems like a terrible solution.

[05:42:01.0000] <jgraham>
hsivonen: Sure. But W3C people are being told that their baby is ugly and they are trying to find ways to route around the objections rather than consider their merits

[05:43:00.0000] <jgraham>
It seems to me that one possible outcome of this is to break XML on the web entirely

[05:45:00.0000] <zcorpan>
"ARIA introduces elements and attributes ..." - aria introduces elements?

[05:45:01.0000] <hsivonen>
no

[05:46:00.0000] <hsivonen>
unless <access> is making its way into ARIA while people aren't looking.

[05:48:00.0000] <annevk42>
"lesser-distributed extensibility"

[05:48:01.0000] <annevk42>
You got to love the terms all these architecture astronauts come up with. I wonder if any normal author can still follow what this is about.

[05:51:00.0000] <takkaria>
I thought we had implementor consensus on <q> these days

[05:53:00.0000] <jgraham>
Yeah, I thought we had converged on the insane HTML4 requirements

[05:54:00.0000] <annevk42>
We have.

[05:56:00.0000] <zcorpan>
chris has still not replied about their ua style sheet

[05:56:01.0000] <zcorpan>
although i guess Philip` reverse engineered it anyway

[06:03:00.0000] <rubys1>
re <q>, that was a result of my discussions with Steven where he claimed that the platform portions of XHTML2 had consensus.  He said that what is reflected in XHTML2 was based on input from the CSS working group.

[06:08:00.0000] <rubys>
@jgraham, re: "I am kindof saddened that we have resorted to "consenus" politics being more important than the technical arguments that have been made"

[06:08:01.0000] <rubys>
author conformance requirements are codified personal opinion.

[06:09:00.0000] <annevk42>
platform portions of XHTML2 had consensus with whom?

[06:09:01.0000] <jgraham>
rubys: Sure if you're talking about things like "should I quote my attributes"

[06:10:00.0000] <annevk42>
<q> has quotation marks in IE8, Safari, Firefox, Opera, Chrome, etc.

[06:10:01.0000] <rubys>
Apparently the change in the way the Q element was defined was done jointly between the (then) HTML working group and the CSS Working group.

[06:10:02.0000] <jgraham>
Not if you're talking about "can I invent new names with a colon in them"

[06:10:03.0000] <annevk42>
rubys, for the HTML <q> element or the XHTML2 <q> element?

[06:10:04.0000] <annevk42>
anyway, all user agents render quotation marks

[06:11:00.0000] <annevk42>
I'm fine with changing it, but I don't really see it happen

[06:11:01.0000] <rubys>
IE renders quotation marks?

[06:11:02.0000] <annevk42>
I just said so

[06:11:03.0000] <rubys>
wow, I thought it was otherwise

[06:12:00.0000] <annevk42>
IE8 supports "all" of CSS 2.1 and could therefore implement support for <q>

[06:12:01.0000] <annevk42>
and they did

[06:12:02.0000] <annevk42>
(against what HTML5 suggested at the time, but we changed HTML5 after that happened)

[06:13:00.0000] <rubys>
I just ran a test with IE8 and don't see the quotes.

[06:14:00.0000] <rubys>
ah, it was in IE7 standards mode

[06:16:00.0000] <rubys>
http://www.w3.org/MarkUp/2009/ED-xhtml2-20090205/mod-text.html#sec_9.8.

[06:17:00.0000] <rubys>
http://dev.w3.org/html5/spec/Overview.html#the-q-element

[06:17:01.0000] <annevk42>
I know XHTML2 contradicts HTML4

[06:18:00.0000] <annevk42>
XHTML2 originally had a <quote> element

[06:18:01.0000] <annevk42>
But since their original bold design choices they're tweaking it more and more to be just like HTML4 with weird incompatible changes. See e.g. the <img> element.

[06:45:00.0000] <Philip`>
annevk42: Something like http://philip.html5.org/data/invalid-location.txt ? (Feel free to munge it into a nicer format)

[06:46:00.0000] <Philip`>
zcorpan: Like http://philip.html5.org/data/ie8-quotes.txt ? It's not really reverse engineering (since I couldn't figure out where the quoting rules were stored), it's just testing a lot of possible values :-)

[06:48:00.0000] <annevk42>
Philip`, cool, though that doesn't tell how many pages attempted to do a redirect in total

[06:49:00.0000] <Philip`>
annevk42: I don't have that information, since I only got the headers that HttpClient finally returned (or the exceptions it threw instead of returning a response)

[06:49:01.0000] <annevk42>
ok

[06:51:00.0000] <Philip`>
(http://www.dotnetdotcom.org/ says about 16% of some completely different set of URLs returned a redirect status code)

[06:54:00.0000] <Philip`>
(Numbers a bit meaningless anyway, because e.g. half of the errors were fixed by Yahoo Maps being fixed)

[06:54:01.0000] <hsivonen>
perhaps <q> isn't the most important part of the blog post and Steven isn't the best source on implmentor consensus on <q>

[06:55:00.0000] <annevk42>
that's a bit too logical for my taste

[06:56:00.0000] <gsnedders>
rubys2: Tinyurls always do that

[06:59:00.0000] <gsnedders>
/me needs to buy a suit… today

[07:01:00.0000] <Philip`>
Does it need to fit?

[07:07:00.0000] <gsnedders>
Philip`: I think a suit that doesn't fit isn't formal enough

[07:08:00.0000] <mpilgrim>
hsivonen: indeed. let's not all waste time discussing the exact shade of blue painted on the side of the nuclear bomb

[07:09:00.0000] <gsnedders>
/me still wants pink with blue spots

[07:10:00.0000] <Philip`>
"I want to explore the idea of dropping the assumption that the current HTML working group has the sole responsibility for, and absolute dominion over, authoring guidelines." - it seems more like validator developers have absolute dominion in practice

[07:11:00.0000] <Philip`>
(to the same extent that browser developers have absolute dominion over e.g. parsing requirements)

[07:12:00.0000] <Dashiva>
Frontpage and dreamweaver might have even more sway than validators

[07:12:01.0000] <gsnedders>
/me goes crazy

[07:14:00.0000] <beowulf>
re: tinyurl, what about a @alt for <a> ...

[07:15:00.0000] <hsivonen>
Philip`: shh! don't tell people about the absolute dominion!

[07:17:00.0000] <hsivonen>
https://twitter.com/jreschke/statuses/1476412112

[07:18:00.0000] <Philip`>
/me agrees with jgraham about not seeing a clear distinction between platform and language features

[07:18:01.0000] <Philip`>
because the platform is already extensive enough that you can do pretty much anything on it

[07:19:00.0000] <Philip`>
e.g. you could implement <canvas> using lots of coloured <div>s, with XHR and a PNG decoder written in JS so you can implement drawImage

[07:20:00.0000] <Philip`>
but it's better (for performance) to move that into the platform

[07:20:01.0000] <mpilgrim>
don't spend too much time trying to dissect it

[07:20:02.0000] <mpilgrim>
it's a false dichotomy

[07:21:00.0000] <mpilgrim>
just a way of trying to change the subject

[07:21:01.0000] <Philip`>
and similarly XForms could be implemented with scripts, but that's not going to help users of ELinks or Opera Mini or NoScript or WWW::Mechanize::FormFiller

[07:24:00.0000] <rubys2>
*shrug*  TV Raman suggested it, and a room full of people seemed to grok it

[07:25:00.0000] <Philip`>
mpilgrim: I'm not sure what it's trying to change the subject away from

[07:32:00.0000] <takkaria>
lots of intuitively simple distinctions turn out to be somewhat false, though

[07:32:01.0000] <takkaria>
or at least, not as simple as they at first seem

[07:33:00.0000] <jgraham>
It seems to be trying to be an attemp to create a false divison such that each group of interested parties can be told that their desires are being met by one part of the divison

[07:35:00.0000] <hsivonen>
jgraham++

[07:38:00.0000] <Philip`>
rubys2: Maybe those people weren't aware that 2D graphics has been implemented as a scripting library, or that ubiquity-xforms violates fundamental assumptions of XML (e.g. if I change the prefix from "xf" to "xf2" then it stops work entirely)

[07:40:00.0000] <mpilgrim>
still not the most important part

[07:40:01.0000] <Philip`>
(and if I create new elements through scripts then it doesn't work)

[07:41:00.0000] <mpilgrim>
(hint: the only sentence that matters is the one with the word "king" in it. everything else is smokescreen.)

[07:41:01.0000] <Philip`>
(so it's kind of an interesting hack but not an adequate substitute for a real feature)

[07:42:00.0000] <Philip`>
s/work/working/ a few lines ago

[07:50:00.0000] <hsivonen>
mpilgrim: what's your take on specs that say "foo is not allowed"?

[07:52:00.0000] <jgraham>
/me wonders why it is made out to be a big deal whether XHTML2 and HTML5 converge or not

[07:54:00.0000] <Dashiva>
I could guess that XHTML2 failing would be bad PR for the XML world

[07:54:01.0000] <Philip`>
Is "Reunification" meant to refer to unification of HTML5 and XHTML2?

[07:55:00.0000] <annevk42>
Philip`, I think so, yes

[07:55:01.0000] <hsivonen>
Dashiva: does Sam care about bad PR for the XML world?

[07:55:02.0000] <annevk42>
Philip`, I'm not sure when that became a goal

[07:55:03.0000] <MikeSmith>
where is the word "reunification" used?

[07:56:00.0000] <Philip`>
MikeSmith: In the title

[07:56:01.0000] <Philip`>
(but nowhere else)

[07:56:02.0000] <MikeSmith>
which title ?

[07:56:03.0000] <Philip`>
MikeSmith: http://intertwingly.net/blog/2009/04/08/HTML-Reunification

[07:57:00.0000] <Philip`>
MikeSmith: That's literally hours old; surely you must have seen it already, via RSS and Planets and Twitter and IRC

[07:58:00.0000] <jgraham>
It is interesting that no one expects the same dynamics to apply in other fields. Like no one tries to reunify different programming languages

[07:58:01.0000] <mpilgrim>
Philip`: rubys2 talks about "unification" in the same way the romulans talk about it (c.f. http://memory-alpha.org/en/wiki/Unification_II_(episode) )

[07:58:02.0000] <Philip`>
jgraham: Sure they do

[07:58:03.0000] <MikeSmith>
Philip`: sorry, i was too busy having dinner and drinking with a beautiful woman

[07:58:04.0000] <Philip`>
jgraham: C99 and C++0x have been moving towards more commonality

[07:58:05.0000] <MikeSmith>
I suppose it would be judged wrong of me to say that I think this whole thing was a lost cause

[07:59:00.0000] <MikeSmith>
(if I were to actualy say that)

[07:59:01.0000] <jgraham>
MikeSmith: What whole thing?

[07:59:02.0000] <MikeSmith>
jgraham: the marriage of heaven and hell

[07:59:03.0000] <Philip`>
MikeSmith: Even if it's a lost cause we'll still continue searching for it, because it might turn up in the last place we look

[07:59:04.0000] <MikeSmith>
proposed marriage

[07:59:05.0000] <jgraham>
Philip`: In some sense C++ is supposed to be a superset of C

[08:00:00.0000] <jgraham>
So that makes more sense

[08:00:01.0000] <Philip`>
Also we might find some other interesting things that fell down the back of a cupboard while we're searching

[08:00:02.0000] <MikeSmith>
the heaven and hell parts being different depending on your political beliefs

[08:00:03.0000] <jgraham>
This is more like C and lisp unifying :)

[08:01:00.0000] <hsivonen>
jgraham: comparing XHTML2 to lisp seems insulting to lisp

[08:01:01.0000] <MikeSmith>
I'm finding that sam's blog doesn't read quite the same after consuming a full bottle of wine

[08:01:02.0000] <Philip`>
jgraham: C++ started that way, but changed incompatibly, and C changed too (though much less), and now both languages are evolving to remove some of the differences; similarly XHTML2 and HTML5 started from the existing (X)HTML language, and evolved differently

[08:02:00.0000] <mpilgrim>
i find it's much more understandable if you read it out loud in the voice of the comic book guy from the simpsons

[08:02:01.0000] <MikeSmith>
but I do notice the word "XForms" in there

[08:02:02.0000] <hsivonen>
Philip`: the difference is that people actually use C and C++. A lot.

[08:03:00.0000] <Philip`>
hsivonen: Perhaps jgraham should have said "Like no one tries to reunify different programming languages when one is used by a billion people and the other is used by two people"

[08:03:01.0000] <MikeSmith>
"Any additional features required of browsers in support for XForms would have to meet a very high bar indeed"

[08:03:02.0000] <MikeSmith>
emphasis on the "indeed"

[08:04:00.0000] <MikeSmith>
damn, hsivonen, that's quite a comment

[08:05:00.0000] <Philip`>
MikeSmith: Indeed, since ubiquity-xforms requires browsers to be infinitely fast so you don't get an ugly flash of not-yet-munged content appearing while the scripts load

[08:05:01.0000] <mpilgrim>
/me feels a "translation from ruby-speak to english" coming on

[08:05:02.0000] <MikeSmith>
speak of the devil

[08:05:03.0000] <Philip`>
(Or, I suppose, browsers could implement something like XBL2)

[08:06:00.0000] <Philip`>
(but for text/html)

[08:06:01.0000] <mpilgrim>
nah, not even worth my time

[08:06:02.0000] <Philip`>
(though I know nothing about XBL2 except that it's the magic solution to all problems that is sometimes promoted by WHATWG people)

[08:06:03.0000] <MikeSmith>
rubys: we are performing hermeneutics on your blog entry

[08:06:04.0000] <hsivonen>
MikeSmith: what's quite a comment?

[08:06:05.0000] <MikeSmith>
long

[08:06:06.0000] <MikeSmith>
lots of words

[08:07:00.0000] <MikeSmith>
i think you quoted the most of the original text

[08:08:00.0000] <MikeSmith>
/me arrives at Yokohama and seat frees up so that I don't have to keep typing with one hand

[08:08:01.0000] <rubys>
/me looks up "hermeneutics"

[08:09:00.0000] <Philip`>
Sounds like a dirty word

[08:09:01.0000] <MikeSmith>
I don't know that it means either. I overheard it in Vacation Bible School

[08:10:00.0000] <hsivonen>
MikeSmith: it seems that it's unnecessary to repeat the code for the KEYGEN switch-case. I'll look into it carefully tomorrow. Gotta go now.

[08:10:01.0000] <Philip`>
MikeSmith: Definitely a dirty word, then

[08:10:02.0000] <MikeSmith>
hsivonen: wait 2 seconds

[08:10:03.0000] <MikeSmith>
please

[08:10:04.0000] <hsivonen>
MikeSmith: ok

[08:10:05.0000] <MikeSmith>
http://bugzilla.validator.nu/attachment.cgi?id=77

[08:11:00.0000] <hsivonen>
MikeSmith: on the face of it, it seems that case INPUT: case KEYGEN: /* old code */ would work

[08:11:01.0000] <MikeSmith>
patch for the syntax part

[08:11:02.0000] <MikeSmith>
when you got time

[08:11:03.0000] <MikeSmith>
hsivonen: OK

[08:11:04.0000] <MikeSmith>
hsivonen: anyway, no rush

[08:11:05.0000] <MikeSmith>
just wanted to send you links before I forget

[08:11:06.0000] <hsivonen>
MikeSmith: looks good.

[08:11:07.0000] <MikeSmith>
k

[08:12:00.0000] <MikeSmith>
I'll write some tests as well

[08:13:00.0000] <LeifHS>
zcorpan: regarding <object> etc, what exactly does "fallback-free" mean? Where is the effect of considering object fallback-free explained?

[08:14:00.0000] <Philip`>
LeifHS: http://www.whatwg.org/specs/web-apps/current-work/multipage/dom.html#fallback-free

[08:15:00.0000] <MikeSmith>
wow, LeifHS on IRC

[08:15:01.0000] <LeifHS>
Mike, everyithing can happen ...

[08:15:02.0000] <MikeSmith>
LeifHS: yeah, that's what makes life worth living

[08:16:00.0000] <MikeSmith>
damn, 9 minutes of battery left, and 25 minutes to my destination

[08:16:01.0000] <Philip`>
LeifHS: More usefully: http://www.whatwg.org/specs/web-apps/current-work/#fallback-free and then wait a minute for everything to load and burn a load of CPU power, and then click "fallback-free" to see a list of references, which should be all its effects

[08:17:00.0000] <Philip`>
MikeSmith: Type slower to conserve battery power

[08:17:01.0000] <MikeSmith>
heh

[08:17:02.0000] <MikeSmith>
well, I got keygen patches for v.nu done today

[08:18:00.0000] <MikeSmith>
so I guess I can claim I've accomplisehd something

[08:18:01.0000] <MikeSmith>
and it's only 23:23

[08:18:02.0000] <MikeSmith>
usually by this time of the day, I've accomplished nothing at all

[08:23:00.0000] <LeifHS>
Phillip: All I see is that it is said that it is considered "fallback-free". The issue I am after is what it means if an OBJECT with EMBED is considered fallback-free. That, in my mind, means that the definition of how OBJECT works changes. But to be sure, I wanted something that told me what effeect it has.

[08:23:01.0000] <LeifHS>
Sorry, meant "Philip".

[08:26:00.0000] <annevk42>
It only means something when it is said to mean something

[08:29:00.0000] <Philip`>
LeifHS: Clicking the definition to see references should show the effects - it appears to just be involved in document.nameofobjectelement

[08:30:00.0000] <Philip`>
(so you can't access non-fallback-free object elements via that mechanism)

[08:54:00.0000] <jgraham>
/me wonders why rubys's blog says "This comment has been blocked" if I type anything into the comment box

[08:54:01.0000] <jgraham>
Not that I had anything in particular to say but it is still quite unfriendly

[09:03:00.0000] <LeifHS>
Philip, I've read the definition, but if you have external references in min, htne have not seen them.

[09:08:00.0000] <Philip`>
LeifHS: It's not referred to externally from HTML5 (if that's what you mean), and the only internal references are the previous few paragraphs

[09:09:00.0000] <Philip`>
(and those paragraphs are describing how document.nameofobjectelement will (or will not) find the object with that name, depending on whether it's fallback-free)

[09:10:00.0000] <zcorpan>
/me wonders what is being unclear

[09:13:00.0000] <LeifHS>
Scripting is not my strong sidde, but how can we decide what should be available for scripting without a look at how UAs are treating OBJECT fallback as such? And the  trouble is that <embed> is treated as a <source> element by Webkit (and partly by IE). In all the OBJECT-with-EMBED examples that zcorpan looked at, any markup beside EMBED will be revelealed if EMBED is revealed. Except in...

[09:13:01.0000] <LeifHS>
...Webkit. See http://www.malform.no/html5/object+youtube. To make an OBJECT-with-EMBED be seen as fallback free, would - it seems to me - be to sanction the WebKit behaviour. (See the thread about interopable object in HTMLwg) 

[09:14:00.0000] <rubys>
jgraham: weird

[09:14:01.0000] <rubys>
it shouldn't do that.... and obviously doesn't for others.

[09:14:02.0000] <rubys>
is it reproducible?

[09:17:00.0000] <Philip`>
/me wonders if rubys's blog still puts a "submit" button directly where the "preview" button one was about to click on was located at the moment one started to click one's mouse button

[09:17:01.0000] <zcorpan>
LeifHS: this issue actually has nothing to do with fallback. "fallback-free" is just a term that could be chosen better

[09:18:00.0000] <Philip`>
rubys: I get "This comment has been blocked" as soon as I type a letter into the comment box

[09:18:01.0000] <zcorpan>
LeifHS: it's more like "is-intended-for-internet-explorer-and-should-not-be-visible-for-other-browsers"

[09:18:02.0000] <zcorpan>
LeifHS: for the purposes of document.foo lookup

[09:19:00.0000] <jgraham>
rubys: Yes, it is reproducable. Based on IP address?

[09:20:00.0000] <jgraham>
/me has no way to verify the IP address thing

[09:23:00.0000] <jgraham>
Oh, I didn't notice that Philip` had the same problem

[09:28:00.0000] <annevk5>
LeifHS, <embed> _is_ exactly like <img>, <br>, <source>, etc. in all browsers

[09:28:01.0000] <LeifHS>
sorry - bye for now. L

[09:38:00.0000] <zcorpan>
it seems the xml core wg are unaware of google existing

[10:16:00.0000] <Philip`>
zcorpan: I thought Google was the cause of XML's failure on the web, because AdSense relies on document.write and their business model would be destroyed by XHTML - surely the WG can't be unaware of such facts

[10:19:00.0000] <zcorpan>
Philip`: if they are aware of google, then they don't seem to know that people use google rather than going to w3.org to find out what <?xml-model?> does upon finding it in view source

[10:20:00.0000] <zcorpan>
actually "xml-model" is probably not the best name because it gives unrelated results in google

[10:21:00.0000] <Philip`>
Maybe they think Google will point users at the relevant page on w3.org

[10:22:00.0000] <Philip`>
in which case somebody could tell them about w3schools.com's pagerank

[10:36:00.0000] <rubys>
jgraham, Philip`: I believe I've isolated the problem, and disabled that check for now until I can correct it.

[11:24:00.0000] <Hixie>
zcorpan: rows=",86" is treated like rows="1,86"

[11:29:00.0000] <zcorpan>
Hixie: not "1*,86"?

[11:29:01.0000] <rubys>
jgraham, Philip: I believe I've fixed the problem.  My code was comparing GMT time to local times and falsely concluded that there was a DOS attack in progress.

[11:29:02.0000] <Hixie>
zcorpan: er yes, sorry

[11:29:03.0000] <zcorpan>
Hixie: ok, then that's good

[11:30:00.0000] <zcorpan>
Hixie: thanks for speccing out framesets

[11:31:00.0000] <zcorpan>
Hixie: it has already saved us the work of reverse engineering ie :)

[11:32:00.0000] <zcorpan>
s/the/some/

[11:36:00.0000] <Hixie>
:-)

[11:44:00.0000] <gsnedders>
Hixie: When are you planning on doing ref now?

[11:48:00.0000] <gsnedders>
hsivonen: As for RelaxNG/XSD, AFAIK lxml has built-in support for XSD but not RelaxNG

[11:59:00.0000] <Philip`>
gsnedders: Like http://www.xml.com/cs/user/view/cs_msg/2576 ?

[11:59:01.0000] <Hixie>
gsnedders: ref?

[11:59:02.0000] <gsnedders>
/me is n00b'd

[11:59:03.0000] <gsnedders>
Hixie: references

[11:59:04.0000] <Hixie>
gsnedders: august

[12:00:00.0000] <Philip`>
/me makes an evidence-free assumption that that feature wasn't removed from lxml

[12:00:01.0000] <Philip`>
(libxml2 definitely supports RelaxNG, though)

[12:06:00.0000] <gsnedders>
I know that libxml2 does :P

[12:07:00.0000] <Philip`>
/me decides that offering to proofread someone's PhD dissertation may not have been the most effort-saving thing to do

[12:14:00.0000] <Philip`>
/me is now up to page 94

[12:19:00.0000] <Philip`>
http://philip.html5.org/demos/xforms/example-xf.html vs http://philip.html5.org/demos/xforms/example-xf2.html

[13:26:00.0000] <hsivonen>
http://webbackplane.com/node/47

[13:27:00.0000] <gmiernicki>
shouldnt the video on that page be in <video> as a theora .ogg?

[13:27:01.0000] <Philip`>
No, it should be a transcript

[13:27:02.0000] <gmiernicki>
some of us don't install flash ;D

[13:29:00.0000] <hsivonen>
gmiernicki: the publisher isn't an HTML5 proponent

[13:29:01.0000] <Philip`>
Some of us install Flash but only so that it works in Firefox 2 and not a browser which said person regularly uses, because it tends to crash in Opera

[13:30:00.0000] <smedero>
the publisher apparently isn't interested in anyone hearing that video either.

[13:30:01.0000] <Philip`>
(and said person hasn't figured out how to make the plugin work in his copy of Firefox 3)

[13:30:02.0000] <smedero>
i've got the volume jacked up to eleven... and I can barely make out portions of it

[13:32:00.0000] <hsivonen>
available in another encumbered format at http://blip.tv/file/897562/

[13:34:00.0000] <Philip`>
Does anyone have a pointer to a script library that implements RDFa in a browser?

[13:35:00.0000] <Philip`>
http://www.w3.org/2006/07/SWD/RDFa/impl/js/ - oh, that

[13:35:01.0000] <smedero>
perhaps http://code.google.com/p/rdfquery/

[13:36:00.0000] <smedero>
http://www.jenitennison.com/blog/node/94 seems to describe the project better than the google project page does

[13:40:00.0000] <hsivonen>
what's http://lists.w3.org/Archives/Public/public-xml-core-wg/2009Apr/att-0001/1193.htm about?

[13:41:00.0000] <hsivonen>
did the ISO committee decide to do shemaLocation as PI?

[13:41:01.0000] <hsivonen>
*schemaLocation

[13:41:02.0000] <hsivonen>
despite RNG not having it by design

[14:00:00.0000] <inimino>
so is this HTML Reunification post suggesting that libraries will implement namespace support in JavaScript?

[14:00:01.0000] <inimino>
including looking for and handling xmlns attributes?

[14:00:02.0000] <hsivonen>
inimino: so it seems

[14:01:00.0000] <inimino>
wow

[14:01:01.0000] <Hixie>
wait what? that's the most ridiculous thing i've heard all day. Who suggested that?

[14:02:00.0000] <inimino>
I'm trying to make sense of the paragraph that contains " Since HTML parsers are unaware of xmlns attributes, it is entirely too dangerous to encourage the use of default namespaces in conformant HTML."

[14:02:01.0000] <inimino>
that's the only interpretation I've considered that makes that make any sense, but I'm probably misunderstanding the whole post

[14:03:00.0000] <Hixie>
what does that even mean

[14:03:01.0000] <Hixie>
who wrote this? is this a lastweek crazy post or something?

[14:03:02.0000] <hsivonen>
Hixie: http://intertwingly.net/blog/2009/04/08/HTML-Reunification#c1239205696

[14:04:00.0000] <inimino>
"libraries that operate on the resulting DOM will consistently evaluate namespaces after the reparenting operation."

[14:04:01.0000] <hsivonen>
Hixie: Mr. Last Week homed in on the crux of Sam's post: the gatekeeper and king maker sentence

[14:04:02.0000] <inimino>
as far as I can tell, the idea as that we can shoe-horn in distributed extensibility by just treating names with colons transparently

[14:05:00.0000] <Hixie>
oh, it's a blog post

[14:05:01.0000] <Hixie>
comment i mean

[14:05:02.0000] <inimino>
and then I guess the idea is that JS libraries will implement something like Namespaces in XML at a layer above that

[14:05:03.0000] <Hixie>
sam gets lots of people commenting crazy stuff on his blog, i wouldn't pay much attention to them

[14:06:00.0000] <Hixie>
(i didn't look to see who wrote the comment)

[14:06:01.0000] <rubys1>
that would be Sam.

[14:06:02.0000] <inimino>
it's not a comment, it's it Sam's post

[14:06:03.0000] <inimino>
s/it/in/

[14:06:04.0000] <Hixie>
oh

[14:08:00.0000] <Hixie>
hsivonen: gatekeeper and king maker?

[14:08:01.0000] <rubys1>
Mr. Last Week homed into what MarkP views as the crux of the post.

[14:09:00.0000] <Hixie>
did mark blog also?

[14:09:01.0000] <Hixie>
man i really need to start reading blogs again

[14:09:02.0000] <rubys1>
mark showed up in IRC.

[14:09:03.0000] <Hixie>
oh

[14:09:04.0000] <Hixie>
i had assumed he was talking about lastweek too

[14:09:05.0000] <Hixie>
d'oh

[14:10:00.0000] <smedero>
see: http://krijnhoetmer.nl/irc-logs/whatwg/20090408#l-502

[14:10:01.0000] <hsivonen>
rubys1: fwiw, I agree with markp on what the crux was but opted to focus on the DOM consistency thing in my comments

[14:10:02.0000] <rubys1>
/me happens to think that author conformance requirements is the crux, as I stated up front, but go figure.

[14:10:03.0000] <Hixie>
/me sighs and goes to read sam's blog

[14:13:00.0000] <Hixie>
it's probably a bad sign that i don't understand what you're saying

[14:14:00.0000] <Hixie>
though it's news to me that the HTML WG has any power of "which features are permissible and which are not in the open web"

[14:14:01.0000] <Hixie>
i mean, the html5 and the whatwg is a pretty good example of how that's not true

[14:14:02.0000] <gsnedders>
But the evil cabal must live on!

[14:14:03.0000] <hsivonen>
rubys1: don't authoring conformance reqs count as gatekeeping?

[14:15:00.0000] <Hixie>
rdfa is another great example -- there are people, including sites like myspace, using it quite happily in html

[14:15:01.0000] <hsivonen>
Hixie: including w3.org!

[14:15:02.0000] <Hixie>
exactly

[14:15:03.0000] <Hixie>
anyway

[14:16:00.0000] <Hixie>
the rdfa discussion will be resolved in a matter of weeks

[14:16:01.0000] <Hixie>
it's been on the calendar for april for months

[14:16:02.0000] <Hixie>
and i don't understand the rest of the post!

[14:16:03.0000] <Hixie>
so...

[14:16:04.0000] <Philip`>
Depends on your definition of "resolved"

[14:16:05.0000] <tantek>
Hixie, html5/whatwg and microformats was just the start. now there's a whole Open Web Foundation that is "is an independent non-profit dedicated to the development and protection of open, non-proprietary specifications for web technologies." http://openwebfoundation.org/

[14:16:06.0000] <Hixie>
/me goes back to trying to get <datagrid> fixed

[14:16:07.0000] <Hixie>
tantek: indeed

[14:17:00.0000] <gsnedders>
As opposed to the Web Foundation?

[14:17:01.0000] <Hixie>
Philip`: "To make a firm decision about"

[14:17:02.0000] <Hixie>
Philip`: "To find a solution to"

[14:20:00.0000] <hsivonen>
Hixie: still trying to make <datagrid> a huge replaced element instead of shipping a MIT-licensed JS lib implemented on top of <table> & ARIA?

[14:20:01.0000] <Hixie>
yes

[14:20:02.0000] <Hixie>
except not "huge"

[14:20:03.0000] <Hixie>
:-)

[14:20:04.0000] <Philip`>
Hixie: Depends on whether your definition of "resolved" means just from your perspective, or from the perspective of the HTML WG and the people who will disagree with your decision and won't let the matter be closed

[14:21:00.0000] <Hixie>
Philip`: i mean from the perspective of the WHATWG, since this is #whatwg. It's still not clear to me how things are resolved in the HTMLWG.

[14:23:00.0000] <Philip`>
Most IRC discussion about the HTML WG occurs in here, so it's good to clarify which is meant :-)

[14:24:00.0000] <Hixie>
Philip`: granted :-)

[14:24:01.0000] <hsivonen>
fwiw, the comment thread branches into http://edward.oconnor.cx/2009/04/DOM-Consistency

[14:25:00.0000] <smedero>
rubys1: "I’d like to retain intact the design principles (link to http://www.w3.org/TR/html-design-principles/) for platform design as worked out by the browser vendors." Perhaps I'm being nitpicky but FWIW, I thought these were worked out by the HTML WG as a whole in April 2007 and WG consensus to publish them was polled in Nov 2007.

[14:26:00.0000] <rubys1>
smedero: that's a fair point

[14:28:00.0000] <rubys1>
refresh

[14:29:00.0000] <smedero>
ahh, good

[14:50:00.0000] <gsnedders>
/me stabs Google Books

[14:50:01.0000] <gsnedders>
/me wants to see page 215 of urn:isbn:9780520220805

[14:55:00.0000] <Hixie>
buy it

[14:55:01.0000] <Hixie>
:-)

[14:55:02.0000] <jcranmer>
meh, torrents are easier

[14:56:00.0000] <gsnedders>
Hixie: I made the mistake of going to not one, but three book stores today. I think I have too much already. :)

[14:56:01.0000] <gsnedders>
Hixie: Also, I need it just to check reference from secondary source, so I'm just going to reference the secondary source instead :P

[14:56:02.0000] <Hixie>
heh

[14:57:00.0000] <gsnedders>
(Basically it contains Nabokov saying in a letter he's working on something called, "The Kingdom by the Sea" — it would later become "Lolita".)

[14:57:01.0000] <gsnedders>
(The original title is a reference to Poe's "Annabel Lee".)

[14:58:00.0000] <gsnedders>
/me goes back to working on his English dissertation

[14:58:01.0000] <gsnedders>
/me sighs

[15:05:00.0000] <gsnedders>
/me tries to make sense of his English teachers scribbling on his draft

[15:05:01.0000] <gsnedders>
This would be simpler if it weren't for the fact that he often ended up scribbling down things about what we had ended up tangentially discussed and thus aren't entirely obviously related.

[15:12:00.0000] <Philip`>
gsnedders: Can't you just cite Wikipedia? It saves a lot of this bother

[15:19:00.0000] <Hixie>
blog admins, i changed from wp-cache to wp-super-cache and added wp-spamfree

[15:19:01.0000] <Hixie>
let me know if it breaks anything

[15:19:02.0000] <Dashiva>
wp-ultra-cache coming next fall

[15:59:00.0000] <annevk5>
it's funny that when ubiquity-xforms is put to test it fails

[15:59:01.0000] <annevk5>
the disturbing thing is of course that some day they get all the details right (if they at all care) and claim we should support it because it's simple

[16:02:00.0000] <Hixie>
the day they get all the details right will be the day it looks like wf2 :-)

[16:02:01.0000] <Hixie>
/me ducks

[16:03:00.0000] <Hixie>
(wf2 is what resulted from our attempt to adapt xforms for html, so i'm not really joking)

[16:07:00.0000] <gsnedders>
Hixie: What you need to fight spam is escape @name and @id on the comment form using entities. Spam bots are dumb.

[16:07:01.0000] <gsnedders>
(That's basically all the spam protection on my blog. Works awesomely.)

[16:08:00.0000] <Hixie>
i don't particularly care about fighting spam, i was just enabling the plugins dreamhost asked me to enable

[16:09:00.0000] <gsnedders>
Hixie: What? Is spam like CSRF and not a problem for the web?

[16:56:00.0000] <takkaria>
neither of Philip's demos worked for me

[16:57:00.0000] <takkaria>
(the xforms ones)


2009-04-09
[17:00:00.0000] <takkaria>
oh, the first one seems to do something now

[17:18:00.0000] <Philip`>
takkaria: You have to wait for it to load a zillion script files

[17:18:01.0000] <Philip`>
(Presumably it'd be possible to pack them all into a single .js file if you cared about performance)

[17:19:00.0000] <Philip`>
/me hopes he is actually testing it correctly, and didn't forget to change some of the prefixes or something

[17:40:00.0000] <LeifHS>
annevk5:  Yes, <embed> is like <source> - kind of. However, <source> is part of the <video> element - it is an extension of <video>. What I meant was that WebKit is treating <embed> as if is part of <object> (the same way that <source> is part of <video>). 

[17:40:01.0000] <karlcow>
/me always wonders why talented people becomes suddenly blind when they are passionate about their tech

[17:50:00.0000] <LeifHS>
zcorpan: Is it possible for you to create an arch typical exsample of what you mean?

[01:31:00.0000] <Hixie>
/me now has a basical proof of concept of his new <datagrid> API design

[01:32:00.0000] <Hixie>
now i just have to check how plausible it is

[01:32:01.0000] <Hixie>
from the authoring side...

[01:37:00.0000] <Hixie>
this is one of the first times that i've designed something for which i think a structure a bit like a B-tree would actually be a pretty good fit

[01:45:00.0000] <MikeSmith>
wow

[01:45:01.0000] <MikeSmith>
B-tree

[01:45:02.0000] <MikeSmith>
that's a blast from the past

[01:46:00.0000] <MikeSmith>
Hixie: seems like datagrid is ultimately going to sink or swim based on how much implementor commitment there is

[01:47:00.0000] <MikeSmith>
and given that there's not be much implementor commitment forthcoming so far, I wonder where that's going to leave it

[01:48:00.0000] <MikeSmith>
if we really want to get to LC by this Fall

[01:48:01.0000] <Hixie>
my impression is that browser vendors seem in agreement that it would be a useful feature, but that it isn't a priority

[01:51:00.0000] <MikeSmith>
Hixie: I think it's something more than just a useful feature

[01:52:00.0000] <Hixie>
oh?

[01:52:01.0000] <MikeSmith>
to me, it's something that would appeal quite a bit to Web developers

[01:52:02.0000] <Hixie>
that's what i mean by "useful feature" :-)

[01:53:00.0000] <Hixie>
note that by LC we don't have to have commitements to implement, only agreement that the features should be in the language

[01:53:01.0000] <MikeSmith>
Hixie: it's the "isn't a priority" to browser developers that's the big stumbling block

[01:53:02.0000] <Hixie>
in fact it's only when entering CR that we have to list features that might be at risk

[01:53:03.0000] <Hixie>
and even then we have until REC to see them implemented

[01:54:00.0000] <Hixie>
so lack of implementation commitements is not a big deal so long as implementors don't disagree that it would be good to implement eventually

[01:54:01.0000] <Hixie>
commitments

[01:54:02.0000] <MikeSmith>
I guess

[01:55:00.0000] <MikeSmith>
I would personally rather not see us take a particular feature into CR without a clear commitment from multiple browser vendors to implement it

[01:55:01.0000] <Hixie>
sure

[01:56:00.0000] <MikeSmith>
maybe we need to light a fire under some asses as far as datagrid

[01:56:01.0000] <hsivonen>
myvidoop.com no longer shows an EV cert. Am I being MITMed?

[01:56:02.0000] <MikeSmith>
e.g., threat "this is going to be removed from HTML5 unless we get a clear indication of vendor support"

[01:56:03.0000] <MikeSmith>
hsivonen: MITMed

[01:57:00.0000] <MikeSmith>
?

[01:57:01.0000] <hsivonen>
MikeSmith: Man In The Middle

[01:57:02.0000] <Hixie>
/me watches the EV cert security model fall apart

[01:58:00.0000] <zcorpan>
hsivonen: do what normal people do and discard any dialogs and don't notice the lack of dialogs

[01:58:01.0000] <MikeSmith>
EV cert, despite whatever faults it has, is approximately one gazillion times better than non-EV certs

[01:59:00.0000] <Hixie>
for cert vendors, sure

[01:59:01.0000] <hsivonen>
MikeSmith: how so. It's causing me distress now. Probably for no good reason.

[01:59:02.0000] <hsivonen>
If I'm being MITMed, how do I email vidoop, when I should assume that by now the MX record in my DNS cache is poisoned as well?

[02:00:00.0000] <Hixie>
certs in general can answer the question "is this who i think it is", which is never a question users ask. They ask (implicitly) the question "am I being attacked", which is not possible to answer using any kind of SSL cert that I know of, EV or otherwise.

[02:01:00.0000] <Hixie>
(and EV certs can answer the first question with $450 more confidence than non-EV certs, that's about it.)

[02:01:01.0000] <MikeSmith>
Hixie: EV certs have documented rules for identity

[02:01:02.0000] <Hixie>
right. "$450 more confidence".

[02:01:03.0000] <MikeSmith>
when you pay (whatever dollar amount) for a non-EV cert, you are a sucker

[02:02:00.0000] <MikeSmith>
because you are paying for nothing

[02:02:01.0000] <Hixie>
you are paying for someone to sign it

[02:02:02.0000] <MikeSmith>
CAs are bound to do zero vetting otherwise

[02:02:03.0000] <Philip`>
You're paying for something that prevents passive network attackers from reading all the data send to/from your web site, which seems quite useful

[02:03:00.0000] <Philip`>
*sent

[02:05:00.0000] <Hixie>
Philip`: you can get that with a self-signed cert

[02:05:01.0000] <Philip`>
Hixie: You can't since it won't work in Firefox 3

[02:05:02.0000] <MikeSmith>
the EV-cert work is similar to a lot of things in that it's really easy to make jackass comments about it from the outside after the fact, throwing rocks, without any understanding of the difficulties involved in ever having tried to do something like it all

[02:05:03.0000] <Hixie>
Philip`: that's a bug (it should encrypt but not show any ui about it)

[02:05:04.0000] <Hixie>
imho, anyway

[02:06:00.0000] <Hixie>
MikeSmith: as someone who is involved in the security work of three browsers, i feel pretty qualified in the subject. I'm fully aware that there has not been anything better proposed. That doesn't mean it's good.

[02:06:01.0000] <hsivonen>
I emailed vidoop. I wonder if they'll answer and with what answer.

[02:07:00.0000] <hsivonen>
Hixie: hey, that's the argument against the Public Suffic List

[02:07:01.0000] <Hixie>
i'm not saying we shouldn't have EVs

[02:07:02.0000] <Hixie>
and the public suffix list sucks too, yes

[02:07:03.0000] <Hixie>
:-)

[02:07:04.0000] <zcorpan>
/me thinks they'll answer "What is MITM?"

[02:09:00.0000] <Hixie>
(and before someone thinks i'm only picking on technologies i'm not working on, html sucks in many ways too. It, like the other two, just happens to be the best we have in the real world.)

[02:09:01.0000] <hsivonen>
zcorpan: their core business is making people feel more secure, so one would hope they'd know what MITM is

[02:10:00.0000] <MikeSmith>
Hixie: yeah, the point is that there was not anything better proposed, and that coming up with anything at all involved burning up a lot of time negotiating things with the CAs, and not all browser projects cared to actually show up to bother to take the time to actually involve themselves actively in the discussions

[02:10:01.0000] <Hixie>
oh?

[02:10:02.0000] <Hixie>
who wasn't involved?

[02:11:00.0000] <MikeSmith>
Opera was involved, Microsoft was involved, George Staikos was involved

[02:14:00.0000] <Hixie>
interesting

[02:14:01.0000] <Hixie>
anyway, bed time

[02:15:00.0000] <Hixie>
nn

[02:23:00.0000] <MikeSmith>
hsivonen: about http://bugzilla.validator.nu/show_bug.cgi?id=437

[02:23:01.0000] <MikeSmith>
I'm trying to figure out which part of the code it is that adds the hyperlinking back to the spec

[02:24:00.0000] <hsivonen>
MikeSmith: nu.validator.spec.html5.Html5SpecBuilder

[02:24:01.0000] <MikeSmith>
/me looks now

[02:28:00.0000] <MikeSmith>
hsivonen: also, seems to be some brokenness in currently linking

[02:29:00.0000] <MikeSmith>
test, with, e.g., <!DOCTYPE html><title>foo</title><img>

[02:29:01.0000] <MikeSmith>
links to <img> doc end up as:

[02:29:02.0000] <MikeSmith>
http://www.whatwg.org/specs/web-apps/current-work/#null

[02:29:03.0000] <hsivonen>
MikeSmith: that code is *very* brittle :-(

[02:30:00.0000] <MikeSmith>
hsivonen: i see

[02:31:00.0000] <hsivonen>
I really should get to fixing outstanding V.nu bugs, but I'm still blocked on eliminating a key memory management issue from the C++ version of the parser.

[02:32:00.0000] <MikeSmith>
hsivonen: yeah, understood

[02:33:00.0000] <MikeSmith>
I'm happy to help with, at the very least, dealing with some of the low-hanging fruit

[02:33:01.0000] <MikeSmith>
as far a v.nu bugs

[02:33:02.0000] <zcorpan>
/me wonders how to ensure that text doesn't use non-XML 1.0 Char characters in a text-based CMS

[02:33:03.0000] <hsivonen>
MikeSmith: I very much appreciate your help. The recent fixes have been great.

[02:34:00.0000] <hsivonen>
zcorpan: run everything through a magic regexp?

[02:35:00.0000] <zcorpan>
i guess

[02:37:00.0000] <Philip`>
zcorpan: Remove all non-ASCII characters, and all below 0x20

[02:38:00.0000] <Philip`>
Alternatively: Don't try to stop people inputting non-XML 1.0 Char characters

[02:38:01.0000] <jgraham>
zcorpan: html5lib has a big regexp somewhere

[02:38:02.0000] <Philip`>
and just make sure you deal with the issue when outputting to XML

[02:38:03.0000] <zcorpan>
Philip`: i need non-ascii characters

[02:38:04.0000] <jgraham>
zcorpan: Why are you trying to produce AML with a text-based CMS?

[02:38:05.0000] <jgraham>
*XML

[02:38:06.0000] <Philip`>
(because otherwise you'll forget to validate some of the input, and get invalid text in your database, and then it'll be a pain to get rid of)

[02:39:00.0000] <zcorpan>
jgraham: i want to have an Atom feed and the CMS i have is text-based

[02:39:01.0000] <jgraham>
s/somewhere/in ihatexml.py/

[02:39:02.0000] <zcorpan>
lol

[02:39:03.0000] <Philip`>
Escape on output, it's the only way to be sure :-)

[02:40:00.0000] <jgraham>
zcorpan: I feel there must be a "now I have two problems" type comment in here somewhere

[02:41:00.0000] <MikeSmith>
heh

[03:19:00.0000] <MikeSmith>
http://www.whatwg.org/specs/web-apps/current-work/multipage/editing.html#spelling-and-grammar-checking

[03:20:00.0000] <MikeSmith>
"The spellcheck attribute is an enumerated attribute whose keywords are the empty string, true and false."

[03:21:00.0000] <MikeSmith>
so <p spellcheck>foo</p> should be invalid, right?

[03:21:01.0000] <Philip`>
That looks like an empty string to me

[03:21:02.0000] <MikeSmith>
but <p spellcheck="">foo</p> is valid

[03:21:03.0000] <MikeSmith>
Philip`: <p spellcheck> is the same as empty string?

[03:21:04.0000] <Philip`>
The tokeniser doesn't make any distinction between those two cases, unless I'm horribly mistaken

[03:21:05.0000] <MikeSmith>
OK

[03:21:06.0000] <Philip`>
and I'm pretty sure I'm not horribly mistaken

[03:21:07.0000] <Philip`>
though I could be wrong about that

[03:21:08.0000] <jfkthame>
Philip`: are you the one behind http://fonts.philip.html5.org/ by any chance?

[03:22:00.0000] <Philip`>
jfkthame: I am

[03:22:01.0000] <MikeSmith>
hsivonen: <p spellcheck> vs. <p spellcheck=""> ?

[03:22:02.0000] <jfkthame>
Philip`: cool page! however, there's a problem with at least one of the fonts

[03:22:03.0000] <jfkthame>
please see https://bugzilla.mozilla.org/show_bug.cgi?id=487549

[03:23:00.0000] <Philip`>
jfkthame: "Not authorized"

[03:23:01.0000] <jfkthame>
aaarrrgghhh, sorry....

[03:23:02.0000] <Philip`>
jfkthame: You could CC me on the bug if you want :-)

[03:24:00.0000] <jfkthame>
bugzilla id?

[03:24:01.0000] <Philip`>
philip.taylor⊙ccau

[03:24:02.0000] <jgraham>
Has Philip` inadvertently found yet another browser security bug

[03:24:03.0000] <Philip`>
If I had to guess wildly, it would be that it crashes on OS X

[03:24:04.0000] <jfkthame>
right - due to an invalid kern table in one of the subsetted fonts

[03:25:00.0000] <jfkthame>
you're cc'd

[03:27:00.0000] <jfkthame>
we could potentially do some more validation, but it's not feasible to exhaustively test fonts before handing them over to the OS or font rendering library, so if that isn't completely robust, bad fonts will always be a risk

[03:29:00.0000] <Philip`>
/me wonders why his copy of the font doesn't look anything like a font, and then realises after several minutes that it's been gzipped

[03:31:00.0000] <jfkthame>
looking at the original PixAntiqua.ttf, i see that it has the apple-format kern table, and offhand it looks valid

[03:31:01.0000] <jfkthame>
i guess the subsetting code is at fault, then, when it strips out all the irrelevant kern pairs from the subset font

[03:36:00.0000] <jfkthame>
ah, looks like the problem is Kern.pm (from Font::TTF).... it always packs the kern table header in the old format, without checking the version number to see whether that's correct

[03:37:00.0000] <Philip`>
/me was coming to the same conclusion :-)

[03:37:01.0000] <Philip`>
Is kern table version 1 documented somewhere?

[03:38:00.0000] <jfkthame>
http://developer.apple.com/textfonts/TTRefMan/RM06/Chap6kern.html

[03:38:01.0000] <jgraham>
/me notes there seems very little point in the bug being marked security sensitive now since most of the information needed to reproduce it seems to be documented here :)

[03:38:02.0000] <Philip`>
I suppose I'd really have to support it properly, to fix up all the glyph IDs, and can't just pass it through :-(

[03:39:00.0000] <Philip`>
but I'm too lazy to do that

[03:39:01.0000] <Philip`>
but I could make it just drop the entire kern table

[03:40:00.0000] <Philip`>
jgraham: It's not a clearly exploitable bug, and there's probably hundreds of other ways you could make font renderers crash because they weren't designed for untrusted input, so I suppose the relative risk is quite low :-)

[03:40:01.0000] <jfkthame>
jgraham: yup - 'fraid so - it's basically just another case of "feed random data to the font system and you'll probably crash something"

[03:41:00.0000] <Philip`>
Doesn't fill me with great confidence in the security of browsers if they're dependent on such things :-/

[03:42:00.0000] <takkaria>
font renderers obviously need to stop trusting input

[03:42:01.0000] <jfkthame>
you mean "the security of operating systems", don't you? the browser is just a way to get a font onto the user's system

[03:43:00.0000] <Philip`>
I mean the security of the process of using my web browser to view a (untrusted) page

[03:43:01.0000] <jfkthame>
takkaria: right - it's no different from jpeg or png decoders or whatever - just that we've been attacking those for longer and so they've had a lot more hardening

[03:44:00.0000] <Philip`>
Were JPEG and PNG decoders as crash-prone as TTF decoders are, when they were first used in browsers?

[03:44:01.0000] <Philip`>
/me supposes it's quite possible they were

[03:44:02.0000] <Philip`>
(Has anyone tried making Theora files that crash Firefox?)

[03:44:03.0000] <jfkthame>
ISTR plenty of security alerts in that area over the years

[03:47:00.0000] <jfkthame>
/me needs to go for now - Philip`, have fun with the Theora idea!

[03:48:00.0000] <Philip`>
jfkthame: Thanks for pointing me at the bug :-)

[03:50:00.0000] <gsnedders>
Oh, yeah, it's Thursday today.

[03:50:01.0000] <gsnedders>
/me forgot

[03:55:00.0000] <Philip`>
/me fixes his code to drop the kern table

[03:55:01.0000] <Philip`>
I wonder how many OS X users' browsers my page crashed before that fix

[03:58:00.0000] <jgraham>
gsnedders: You never could get the hang of Thursdays?

[03:59:00.0000] <gsnedders>
jgraham: Nah, it's just it's school holidays so I have no concept of time anymore.

[03:59:01.0000] <Philip`>
/me is not on holiday but has no concept of time anyway

[04:44:00.0000] <eighty4>
gsnedders: so no concept of time == holiday?

[05:02:00.0000] <gsnedders>
eighty4: No, if holiday then no concept of time. The inverse does not hold true.

[05:03:00.0000] <eighty4>
:)

[05:03:01.0000] <eighty4>
I'm having 2 days completly alone tomorrow... during that time I'll probably have no concept of time

[05:15:00.0000] <remysharp>
Is this a sensible place to ask html5 questions - in particular about using it for new sites/pages?

[05:16:00.0000] <gsnedders>
remysharp: Yes

[05:16:01.0000] <gsnedders>
remysharp: But I would say that, wouldn't I? :)

[05:16:02.0000] <jgraham>
gsnedders: The right answer was no. We fail on the first clause (Is this a sensible place)

[05:16:03.0000] <gsnedders>
jgraham: True.

[05:17:00.0000] <remysharp>
Hmm, okay - how about senseless questions?

[05:17:01.0000] <jgraham>
remysharp: However it is still a good place to ask HTML 5 questions

[05:17:02.0000] <gsnedders>
jgraham: But giving the right answer would be sensible.

[05:17:03.0000] <jgraham>
Perhaps the best

[05:17:04.0000] <jgraham>
Just note the /topic

[05:18:00.0000] <remysharp>
So, in the <footer> element, if I have a list of elements -

[05:18:01.0000] <remysharp>
As per this example: http://dev.w3.org/html5/spec/Overview.html#the-nav-element

[05:18:02.0000] <remysharp>
shouldn't the list be in a nav element?

[05:18:03.0000] <remysharp>
/me hmm - was that the right link...

[05:19:00.0000] <gsnedders>
remysharp: "Not all groups of links on a page need to be in a nav element — only sections that consist of primary navigation blocks are appropriate for the nav element."

[05:19:01.0000] <gsnedders>
remysharp: They aren't primary navigation links

[05:19:02.0000] <remysharp>
gsnedders: ah, right, so only "primary" - great - thanks.

[05:20:00.0000] <remysharp>
So I've been using a few live examples around the web to help me code my html5 page -

[05:20:01.0000] <remysharp>
and I've been looking at the uxlondon.com site - 

[05:20:02.0000] <remysharp>
which uses the <div class="section"> method to get around having to use JS to trigger IE to see html

[05:20:03.0000] <remysharp>
so - my question is - 

[05:20:04.0000] <jgraham>
(On the opther hand it does make some sense that the primary navigation might be partially in the footer so the content model restriction maybe doesn't make sense)

[05:21:00.0000] <remysharp>
is there a real reason why all their 'section's contain 'article's?

[05:21:01.0000] <remysharp>
I wouldn't have always nested an article in a section

[05:21:02.0000] <remysharp>
and the spec description of an article is user generate content or article blog entry, etc.

[05:22:00.0000] <remysharp>
example: http://uxlondon.com/speakers/

[05:22:01.0000] <jgraham>
remysharp: At first glance, no

[05:22:02.0000] <remysharp>
any thoughts on that at all?  or perhaps just a design choice they went with

[05:23:00.0000] <gsnedders>
At a further glance, no.

[05:23:01.0000] <jgraham>
remysharp: I think their markup is unnecessarily redundant

[05:23:02.0000] <remysharp>
okay, good - that's what I was thinking too - so at least I'm partially following it.

[05:23:03.0000] <remysharp>
Using the same page as an example (the output) 

[05:24:00.0000] <remysharp>
I am marking up a speakers page for my own project:

[05:24:01.0000] <remysharp>
and I've got a list of speakers - which normally I'd put in a <ul>

[05:24:02.0000] <remysharp>
however - I kind of want to put each one in a <section> element with the whole thing nested inside an <article> element

[05:25:00.0000] <remysharp>
would that make sense?  (or would you like a quick example of what I mean?)

[05:25:01.0000] <gsnedders>
You're asking about sense again…

[05:25:02.0000] <remysharp>
!! :-D

[05:25:03.0000] <remysharp>
yeah, sorry!

[05:25:04.0000] <gsnedders>
/me reads the actual question

[05:26:00.0000] <jgraham>
remysharp: Something like <article><h2>Speakers</h2><section><h3>J. Smith</h3>[...]

[05:26:01.0000] <remysharp>
I suspect a quick mock of the markup I'm suggesting might help

[05:26:02.0000] <remysharp>
yeah - jgraham that looks like what I was thinking

[05:26:03.0000] <remysharp>
so no use of <ul> at all

[05:26:04.0000] <jgraham>
It's not really clear why the outside bit would be <article> rather than <section> but I guess it barely makes any difference anyway

[05:26:05.0000] <remysharp>
but my normal html4 approach would be to use a list element, but with html5, I kinda don't want to

[05:26:06.0000] <gsnedders>
remysharp: I'd go for <article><h2>Speakers</h2><dl><dt>Mr John Smith<dd><p>Mr John Smith is awesome.<p>He even wrote <cite>My Magical Wonderland</cite></dl></article>

[05:27:00.0000] <remysharp>
but isn't Mr John Smith a header within a section?

[05:27:01.0000] <jgraham>
gsnedders: Doesn't work so well if you ant to generate a toc that has the speakers listed

[05:27:02.0000] <jgraham>
*want

[05:27:03.0000] <gsnedders>
/me shrugs

[05:27:04.0000] <beowulf>
i'd go with gsnedders

[05:28:00.0000] <beowulf>
in terms of that question :)

[05:28:01.0000] <jgraham>
remysharp: I think, assuming each spaeker will have a little description, that using section+headers is fine

[05:28:02.0000] <beowulf>
though i'd probably s/article/section

[05:28:03.0000] <gsnedders>
jgraham: I do have a vague clue about what the outlining algorithm says :P

[05:29:00.0000] <remysharp>
damn - is there a prefered pastebin?

[05:30:00.0000] <gsnedders>
The one that Google takes you to when you click on "I'm feeling lucky!" because it takes the least amount of effort to find.

[05:30:01.0000] <jgraham>
remysharp: Try your markup in http://gsnedders.html5.org/outliner/

[05:30:02.0000] <remysharp>
jgraham: ta

[05:30:03.0000] <gsnedders>
(OMG! I WROTE THAT1)

[05:30:04.0000] <jgraham>
remysharp: (That is not a pastebin)

[05:30:05.0000] <remysharp>
yeah, sure - oh

[05:30:06.0000] <remysharp>
there's not going to be url is there?

[05:30:07.0000] <remysharp>
hixie has one I believe - that saves the url...trying that

[05:30:08.0000] <jgraham>
remysharp: Ask gsnedders :)

[05:31:00.0000] <jgraham>
Oh, yeah you could use the LDV

[05:31:01.0000] <gsnedders>
remysharp: Just allowing a textarea?

[05:31:02.0000] <gsnedders>
Yeah, that's on my to-do list.

[05:32:00.0000] <remysharp>
right - there you: 

[05:32:01.0000] <remysharp>
http://tr.im/iv56

[05:34:00.0000] <beowulf>
i'd say the first header is redundant unless it wraps the <p> and that the outer article is a section

[05:34:01.0000] <jgraham>
remysharp: No need for the <header> element unless you plan to make a subheading

[05:34:02.0000] <gsnedders>
The header elements are both needless

[05:34:03.0000] <beowulf>
but i'm easily the dumbest person in the room, just to clarify

[05:35:00.0000] <remysharp>
There would obviously be a header element at the top of the page

[05:35:01.0000] <remysharp>
but you're saying the name shouldn't be in a header

[05:35:02.0000] <jgraham>
remysharp: No, no no :)

[05:36:00.0000] <gsnedders>
I'm saying you gain nothing by having it in a header element

[05:36:01.0000] <jgraham>
<header><h2>Foo</h2></header> === <h2>Foo</h2>

[05:36:02.0000] <gsnedders>
A header element only is of use when you having a heading and subheading and you only want the heading to appear in the TOC

[05:36:03.0000] <remysharp>
sorry, yeah, doesn't gain anything

[05:36:04.0000] <jgraham>
The extra <header> is redundant

[05:36:05.0000] <gsnedders>
<header><h2>Foo</h2><h3>Bar</h3></header> === <h2>Foo</h2> in terms of TOC too

[05:37:00.0000] <remysharp>
gsnedders: why does the h3 text get lost then?

[05:37:01.0000] <remysharp>
(in the TOC)

[05:37:02.0000] <remysharp>
or does it read the highest level heading and use that?

[05:37:03.0000] <gsnedders>
jgraham: Explain!

[05:37:04.0000] <gsnedders>
:P

[05:37:05.0000] <gsnedders>
remysharp: The highest level heading is all that's used

[05:37:06.0000] <remysharp>
cool - that makes sense.

[05:38:00.0000] <remysharp>
so if a TOC was generated from that page - if I omitted the <header> on the names of speakers, their names wouldn't appear in the TOC - is that right?

[05:38:01.0000] <beowulf>
gsnedders: what does <header><h1>Foo</h1><h1>Bar</h1></header> come out as in the TOC?

[05:38:02.0000] <gsnedders>
beowulf: Foo

[05:39:00.0000] <gsnedders>
remysharp: They would. The h3 element would make them.

[05:39:01.0000] <remysharp>
right - got you.

[05:39:02.0000] <remysharp>
so header is only if there's mixed content and you want a specific (or the highest) to be used.

[05:39:03.0000] <remysharp>
that makes sense as to why it's utterly redundant in my example.

[05:39:04.0000] <gsnedders>
beowulf: (It's the first highest order header)

[05:40:00.0000] <gsnedders>
(There are, however, bugs in what the spec currently says.)

[05:40:01.0000] <remysharp>
So, on that same topic, would it be fair to say that in this example, the <header> is redundant:

[05:40:02.0000] <remysharp>
<header><h1>My site</h1><p>Tag line</p></header>

[05:41:00.0000] <gsnedders>
remysharp: No

[05:41:01.0000] <jgraham>
remysharp: Technically, no

[05:41:02.0000] <gsnedders>
(the p element will not be associated with any section, IIRC)

[05:42:00.0000] <jgraham>
It won't have any observable effect on the outline, but semantically it is right

[05:42:01.0000] <gsnedders>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=6750

[05:42:02.0000] <jgraham>
Oh, maybe it does have some observable effect

[05:42:03.0000] <jgraham>
then

[05:42:04.0000] <gsnedders>
jgraham: Not really

[05:42:05.0000] <gsnedders>
jgraham: Unless you have a UI which shows what element is linked to what section, it doesn't.

[05:42:06.0000] <jgraham>
gsnedders: That is an observable effect

[05:43:00.0000] <gsnedders>
Well, not in any current implementation of the algorithm seeming both of our implementations just build a TOC.

[05:43:01.0000] <jgraham>
gsnedders: In principle though

[05:43:02.0000] <remysharp>
just backpeddling to the question about using sections within an article instead of a <ul><li> collection - does that look right in essence? 

[05:44:00.0000] <jgraham>
remysharp: Yes

[05:44:01.0000] <jgraham>
It is better than <ul> for sure

[05:44:02.0000] <remysharp>
awesome. total head f**k based on getting used to using lists for everything, but feel right.

[05:44:03.0000] <remysharp>
*feels right

[05:45:00.0000] <beowulf>
jgraham: why is it better than a list?

[05:46:00.0000] <jgraham>
beowulf: Even extant AT will allow you to nagivate easilly by header elements, for example

[05:58:00.0000] <Philip`>
If you want to generate a TOC of all the speakers, use <span class="author"> and then write a script that extracts all the author names and sticks them into a TOC list

[05:58:01.0000] <Philip`>
which is, like, two lines of code

[05:59:00.0000] <hsivonen>
Hmm. Is keygen really meant to differ from input 'in select'?

[06:16:00.0000] <MikeSmith>
hsivonen: I wondered about that too, when looking at your treebuilder code

[06:16:01.0000] <MikeSmith>
maybe worth a bugzilla to clarify

[06:20:00.0000] <hsivonen>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/73

[06:21:00.0000] <hsivonen>
like <input> in Gecko. Not like <input> in Opera and Safari.

[06:24:00.0000] <hsivonen>
MikeSmith: bug filed just in case

[06:30:00.0000] <gsnedders>
MikeSmith: Which Irish author is it you want me to read, again?

[06:43:00.0000] <hsivonen>
could someone please point out to me why the following XSD regexp: "\s*(none|xMinYMin|xMidYMin|xMaxYMin|xMinYMid|xMidYMid|xMaxYMid|xMinYMax|xMidYMax|xMaxYMax)\s+(meet|slice)?\s*"

[06:43:01.0000] <hsivonen>
does not match the word 'none'?

[06:44:00.0000] <hsivonen>
ooh

[06:44:01.0000] <hsivonen>
now I see it

[06:44:02.0000] <hsivonen>
\s+

[06:45:00.0000] <hsivonen>
that's a classing mis-use of XSD \s, BTW

[06:45:01.0000] <hsivonen>
XSD \s does *not* equal XML whitespace

[06:46:00.0000] <gsnedders>
WHAT!?

[06:46:01.0000] <hsivonen>
gsnedders: welcome to the world of i18n political correctness. Zs FTW!

[06:46:02.0000] <gsnedders>
Not White_Space and not Zs?

[06:46:03.0000] <hsivonen>
gsnedders: IIRC, Zs

[06:47:00.0000] <gsnedders>
That doesn't even include U+000A IIRC!

[06:47:01.0000] <hsivonen>
s/classing/classical/

[06:48:00.0000] <hsivonen>
s/al// I suppose

[06:48:01.0000] <gsnedders>
Yeah

[06:48:02.0000] <gsnedders>
It's now right :)

[06:48:03.0000] <Philip`>
/me would be unable to resist the temptation to write that regexp with (none|x(Min|Mid|Max)Y(Min|Mid|Max))

[06:49:00.0000] <Philip`>
(Fortunately I would be able to resist (none|xM(in|id|ax)YM(in|id|ax)) because that's just crazy)

[06:51:00.0000] <Philip`>
(Shouldn't the regexp start with "(defer\s+)?"?)

[06:53:00.0000] <hsivonen>
defer?

[06:54:00.0000] <Philip`>
That's what http://www.w3.org/TR/SVG/coords.html#PreserveAspectRatioAttribute says

[06:55:00.0000] <hsivonen>
Philip`: good point. 

[06:56:00.0000] <hsivonen>
is defer only allowed on <image>?

[06:56:01.0000] <Philip`>
Sounds like it's allowed everywhere, but ignored except on <image>

[06:57:00.0000] <hsivonen>
heycam: ^

[06:57:01.0000] <hsivonen>
heycam: the delta of the V.nu copy of the SVG 1.1 schema and the W3C copy is growing

[06:59:00.0000] <hsivonen>
MikeSmith: I deployed your recent checkins.

[07:05:00.0000] <MikeSmith>
hsivonen: thanks

[07:06:00.0000] <MikeSmith>
gsnedders: Flann O'Brien

[07:10:00.0000] <heycam>
hsivonen, spec problem?

[07:11:00.0000] <heycam>
or just a problem with the SVG 1.1 DTD not being as restrictive as it could be?

[07:37:00.0000] <hsivonen>
heycam: schema bug

[07:37:01.0000] <hsivonen>
heycam: schema bug in the RELAX NG schema

[07:55:00.0000] <heycam>
hsivonen, ok

[07:55:01.0000] <heycam>
so that rng isn't really official

[07:56:00.0000] <heycam>
we're going to be making a new one soon, for 1.1, but starting from the 1.2T rng

[07:58:00.0000] <heycam>
ah i see that regex you quote is from that unofficial rng

[07:58:01.0000] <gsnedders>
MikeSmith: Well, it's my birthday in just over a week ;P

[07:59:00.0000] <jgraham>
gsnedders: How old are ypu going to be 5? 6? I lose track

[07:59:01.0000] <gsnedders>
jgraham: 7

[07:59:02.0000] <gsnedders>
Sorry, I lie. 10, in an as of yet undecided base.

[08:00:00.0000] <heycam>
hsivonen, erk, seems like the regex in the 1.2T relaxng is horribly wrong!

[08:00:01.0000] <heycam>
\s*(none|xMidYMid)\s*(meet)?\s*

[08:00:02.0000] <heycam>
/me raises an issue

[08:02:00.0000] <krikey>
I was sat like billy no mates in #html5

[08:02:01.0000] <krikey>
someone could have coma and got me :)

[08:02:02.0000] <heycam>
http://www.w3.org/Graphics/SVG/WG/track/issues/2257

[08:02:03.0000] <jgraham>
Hard to get someone if you're in a coma

[08:03:00.0000] <krikey>
ye

[08:03:01.0000] <heycam>
/me goes to watch some more daily show

[09:26:00.0000] <Philip`>
heycam: That issue fails to mention that it accepts strings like "nonemeet" too

[09:26:01.0000] <jgraham>
http://www.ecma-international.org/news/PressReleases/PR_Ecma_finalises_major_revision_of_ECMAScript.htm

[09:27:00.0000] <cryzed>
Hey :)

[09:27:01.0000] <cryzed>
Is someone from the html5lib for Python here?

[09:27:02.0000] <jgraham>
cryzed: Yes

[09:28:00.0000] <cryzed>
great :)

[09:28:01.0000] <jgraham>
"TC39 members will create and test implementations of the candidate specification to verify its correctness and the feasibility of creating interoperable implementations". I wonder if they mean interoperable implemetations that can ship on the web

[09:28:02.0000] <cryzed>
So basically I want to know

[09:29:00.0000] <cryzed>
Do I get the lxml.html parser with this : html5lib.HTMLParser(tree=treebuilders.getTreeBuilder("lxml"))?

[09:29:01.0000] <Philip`>
jgraham: Try reading the next sentence

[09:29:02.0000] <Philip`>
"The test implementations will also be used for web compatibility testing to ensure that the revised specification remains compatible with existing web applications."

[09:29:03.0000] <jgraham>
Philip`: Oh.

[09:30:00.0000] <Philip`>
cryzed: No - that uses html5lib's HTML5 parser (and constructs an lxml document from it), not lxml's non-standard HTML parser

[09:30:01.0000] <jgraham>
cryzed: Yes, unless you use the latest svn in which case the best option is to use html5lib.parse(input, tree="lxml")

[09:31:00.0000] <cryzed>
Well

[09:31:01.0000] <cryzed>
what now?

[09:31:02.0000] <jgraham>
Oh, sorry, I misunderstood the question

[09:31:03.0000] <jgraham>
cryzed: What do you actually want to do

[09:31:04.0000] <cryzed>
wait

[09:31:05.0000] <cryzed>
http://paste.pocoo.org/show/ad9DwVRGDKIhXNgqSk8j/

[09:31:06.0000] <jgraham>
/me is not doing very well at reading at the moment

[09:31:07.0000] <cryzed>
I want to parse my blog with the html5lib

[09:32:00.0000] <cryzed>
and then scrape it with the resulting elementtree

[09:32:01.0000] <cryzed>
Unfortunately I don't really find any documentation for the ElementTree except http://effbot.org/zone/pythondoc-elementtree-ElementTree.htm#elementtree.ElementTree.ElementTree-class

[09:32:02.0000] <jgraham>
cryzed: That looks vaugely sensible. What is the problem?

[09:33:00.0000] <cryzed>
http://paste.pocoo.org/show/KHOsRJHJNnHNvS9ec0Bg/ that should work

[09:33:01.0000] <cryzed>
doesn't though

[09:33:02.0000] <jgraham>
http://codespeak.net/lxml/tutorial.html

[09:33:03.0000] <cryzed>
http://paste.pocoo.org/show/ld7f4BUjcUZbnyFG4vGd/

[09:33:04.0000] <cryzed>
ah...

[09:33:05.0000] <cryzed>
the case..

[09:34:00.0000] <jgraham>
findall()?

[09:34:01.0000] <cryzed>
yes

[09:34:02.0000] <cryzed>
and I DO need to supply

[09:34:03.0000] <cryzed>
an argument

[09:34:04.0000] <cryzed>
which argument do I need to supply to find all tags?

[09:34:05.0000] <cryzed>
.

[09:34:06.0000] <cryzed>
?

[09:35:00.0000] <jgraham>
if you just want all the child nodes you can just do "for item in element:"

[09:35:01.0000] <cryzed>
etree.findall(".//*"): that works aswell

[09:35:02.0000] <cryzed>
thanks though jgraham  :)

[09:35:03.0000] <cryzed>
Is the argument which I pass to findall called a "xpath"?

[09:35:04.0000] <cryzed>
ElementPath

[09:36:00.0000] <cryzed>
found it..

[09:36:01.0000] <cryzed>
sorry I seem to ask only stupid questions

[09:36:02.0000] <jgraham>
cryzed: If you know xpath and are using lxml you can do element.xpath(xpath_expression)

[09:37:00.0000] <cryzed>
Or I can just use the ElementPath?

[09:37:01.0000] <cryzed>
http://effbot.org/zone/element-xpath.htm

[09:37:02.0000] <jgraham>
e.g. element.xpath(".//a") finds all a descendants

[09:37:03.0000] <jgraham>
cryzed: ElementPaths are like a subset of XPath 1.0

[09:38:00.0000] <cryzed>
Hrmm, I don't think I do need full xpath support, thanks for the tip though

[09:38:01.0000] <cryzed>
jgraham, I read on the lxml.html documentation

[09:38:02.0000] <cryzed>
about the following function:

[09:38:03.0000] <cryzed>
*method

[09:38:04.0000] <cryzed>
.text_content():

[09:38:05.0000] <jgraham>
Also if you use a really up to date lxml you can probably get CSS Selectors

[09:39:00.0000] <cryzed>
this isn't available in the lxml etree, right?

[09:39:01.0000] <jgraham>
cryzed: html5lib just generates an lxml tree. It has all the features of whichever lxml you have installed

[09:39:02.0000] <cryzed>
well, yes

[09:39:03.0000] <cryzed>
the problem is

[09:40:00.0000] <cryzed>
lxml.html

[09:40:01.0000] <cryzed>
the html Etree is a special tree

[09:40:02.0000] <cryzed>
How do I tell html5lib

[09:40:03.0000] <cryzed>
to use the lxml.html tree?

[09:40:04.0000] <jgraham>
Oh, yeah

[09:40:05.0000] <smedero>
jgraham: I don't think the .text_content() method exist in lxml.etree 

[09:40:06.0000] <cryzed>
Is there any way to use the lxml.html tree?

[09:40:07.0000] <jgraham>
So, I don't think you can at the moment because of some weirdness in the way that lxml is set up

[09:40:08.0000] <cryzed>
I think this would be really comfortable for webscraping

[09:41:00.0000] <jgraham>
At least that is my recollection from when I implemented this stuff a while ago

[09:41:01.0000] <Philip`>
etree.tostring(node, method='text')

[09:41:02.0000] <Philip`>
might be similar to node.text_content()

[09:41:03.0000] <smedero>
yeah, that should be in the ballpark

[09:41:04.0000] <jgraham>
It's something like you can't create comments in lxml.html or...

[09:41:05.0000] <jgraham>
.xpath(".//text()) works

[09:42:00.0000] <jgraham>
.xpath(".//text()")

[09:42:01.0000] <cryzed>
thanks

[09:42:02.0000] <cryzed>
I found in the lxml.html implementation the following

[09:42:03.0000] <cryzed>
_collect_string_content(self)

[09:42:04.0000] <cryzed>
should work if there is no other way

[09:59:00.0000] <cryzed>
the .text attribute

[09:59:01.0000] <cryzed>
works beautifully

[10:01:00.0000] <cryzed>
..not

[10:02:00.0000] <jgraham>
cryzed: for <a><b>foo</b>bar</a> a.text == None

[10:02:01.0000] <jgraham>
b.text == "foo"

[10:02:02.0000] <jgraham>
b.tail =="bar"

[10:03:00.0000] <cryzed>
is there any way to get the whole text and protect the formatting?

[10:03:01.0000] <cryzed>
for example replace <br>

[10:03:02.0000] <cryzed>
with \n

[10:03:03.0000] <jgraham>
cryzed: Not an easy way that I know of

[10:03:04.0000] <cryzed>
hrm

[10:03:05.0000] <jgraham>
You would need to walk the tree, normalize whitespace and make whatever replacements you want

[10:05:00.0000] <cryzed>
http://paste.pocoo.org/show/5gbDet52tiXSMRWkLZ4y/ ?

[10:06:00.0000] <jgraham>
That will only do children of the td

[10:07:00.0000] <cryzed>
yes

[10:07:01.0000] <cryzed>
that's what I want actually

[10:08:00.0000] <cryzed>
sorry if I start to get annoying

[10:08:01.0000] <cryzed>
but why doesn't that work: http://paste.pocoo.org/show/cfmGTUkDaWSCr1ehDH3X/ ?

[10:13:00.0000] <jgraham>
Trying to get the "blockquote" attribute of a "blockquote" element?

[10:14:00.0000] <cryzed>
well.. kinda

[10:14:01.0000] <cryzed>
^^

[10:14:02.0000] <cryzed>
It works in BeautifulSoup :D...

[10:14:03.0000] <jgraham>
Er, what does it do?

[10:14:04.0000] <jgraham>
I mean if you really have <blockquote blockquote=something> I guess it should work

[10:15:00.0000] <cryzed>
it should get me the text WITHOUT markup WITH formatting out of the blockquote tags

[10:15:01.0000] <cryzed>
>IN BETWEEN HERE<

[10:16:00.0000] <gsnedders>
with formatting without markup? how?

[10:16:01.0000] <cryzed>
well

[10:16:02.0000] <cryzed>
for example

[10:16:03.0000] <jgraham>
cryzed: If you just want the test you can do element.xpath(".//text()")

[10:16:04.0000] <gsnedders>
jgraham: You don't need XPath for that! Peh!

[10:17:00.0000] <cryzed>
<pre>That's some fancy text <br>comment'</pre>

[10:17:01.0000] <cryzed>
Should result to

[10:17:02.0000] <jgraham>
If you want to do some formatting on the text you need to decide what formatting you want

[10:17:03.0000] <cryzed>
That's some fancy text

[10:17:04.0000] <cryzed>
comment

[10:17:05.0000] <cryzed>
gsnedders, what should I use?

[10:17:06.0000] <gsnedders>
jgraham: return etree.tostring(element, encoding=unicode, method='text', with_tail=False) is better than that

[10:18:00.0000] <cryzed>
I can't acces this function

[10:18:01.0000] <jgraham>
And implement that by e.g. walking the tree replacing <br> with "\n" and adding the .tail of the br to the right place

[10:18:02.0000] <jgraham>
gsnedders: Define "better"

[10:18:03.0000] <gsnedders>
jgraham: Quicker

[10:18:04.0000] <gsnedders>
jgraham: The result is identical :P

[10:19:00.0000] <jgraham>
gsnedders: Seems unlikely to be a problem in this case

[10:19:01.0000] <jgraham>
It is much longer to type and easier to get wrong (maybe)

[10:19:02.0000] <gsnedders>
/me just wraps it in a function :P

[10:19:03.0000] <cryzed>
etree doesn't got the attribute .tostring

[10:19:04.0000] <gsnedders>
cryzed: from lxml import etree

[10:20:00.0000] <cryzed>
I originally only wanted to import html5lib :|..

[11:01:00.0000] <cryzed>
Is it a good idea to use the latest svn build?

[11:01:01.0000] <gsnedders>
cryzed: Absolutely.

[11:02:00.0000] <gsnedders>
cryzed: It's a better idea than using the latest release

[11:02:01.0000] <cryzed>
Okay

[11:02:02.0000] <cryzed>
btw

[11:02:03.0000] <cryzed>
gsnedders, if I had an custom BeautifulSoup.py

[11:02:04.0000] <cryzed>
*a

[11:02:05.0000] <cryzed>
Could I somehow tell the treebuilder

[11:02:06.0000] <cryzed>
to use this this BeautifulSoup?

[11:03:00.0000] <cryzed>
http://furyu-tei.sakura.ne.jp/archives/BSXPath.zip I found this

[11:03:01.0000] <cryzed>
it has xpath support as it seems

[11:03:02.0000] <Philip`>
html5lib's BeautifulSoup support is not particularly reliable

[11:03:03.0000] <cryzed>
Yeah, but it works

[11:03:04.0000] <cryzed>
I did some things with it

[11:03:05.0000] <cryzed>
And I somehow think that using BeautifulSoup is easier than lxml

[11:04:00.0000] <Philip`>
It works as long as you don't do one of the things that doesn't work :-)

[11:04:01.0000] <cryzed>
e.g?

[11:05:00.0000] <Philip`>
e.g. http://code.google.com/p/html5lib/issues/detail?id=80

[11:05:01.0000] <cryzed>
oh..

[11:06:00.0000] <cryzed>
lxml

[11:06:01.0000] <cryzed>
doesn't work any better

[11:06:02.0000] <cryzed>
lol

[11:06:03.0000] <cryzed>
html5lib.HTMLParser(tree=html5lib.treebuilders.getTreeBuilder("lxml")).parse("<a><div><div><a>")

[11:06:04.0000] <cryzed>
>>> etree.tostring(e)

[11:06:05.0000] <cryzed>
'<html><head/><body><a/><div><a></a><div><a></a><a/></div></div></body></html>'

[11:06:06.0000] <Philip`>
That works much better, since it gives the right output and doesn't throw exceptions :-p

[11:06:07.0000] <cryzed>
right?

[11:06:08.0000] <cryzed>
oh.. you are right

[11:06:09.0000] <gsnedders>
/me guesses you want the html5lib serializer

[11:07:00.0000] <Philip`>
cryzed: "Right" according to the HTML5 spec

[11:07:01.0000] <cryzed>
Oke

[11:07:02.0000] <cryzed>
Guys let me tell what I want to do

[11:07:03.0000] <Philip`>
Bit peculiar how it mixes "<a></a>" and "<a/>"...

[11:07:04.0000] <cryzed>
Oke

[11:07:05.0000] <cryzed>
I'm sure some of you know 4chan

[11:08:00.0000] <cryzed>
I want to write a Python "API" for it

[11:08:01.0000] <cryzed>
take a look at this page for example:

[11:08:02.0000] <cryzed>
http://zip.4chan.org/x/res/1648607.html

[11:08:03.0000] <cryzed>
So

[11:08:04.0000] <cryzed>
each reply has got an id attribute and a class

[11:08:05.0000] <cryzed>
Well, basically I want to use lxml

[11:08:06.0000] <cryzed>
to scrape this site

[11:08:07.0000] <cryzed>
get the text between the blockquotes

[11:09:00.0000] <cryzed>
remove the tags

[11:09:01.0000] <cryzed>
and replace the <br> tags with \n

[11:09:02.0000] <cryzed>
and save it into a variable

[11:09:03.0000] <cryzed>
at the this is all getting wrapped in a class called Reply

[11:09:04.0000] <cryzed>
*at the end

[11:10:00.0000] <cryzed>
http://iohosaf.pastebin.com/d5b10f992

[11:10:01.0000] <cryzed>
This is what I've got so far

[11:10:02.0000] <cryzed>
in the implementation with BeautifulSoup

[11:10:03.0000] <cryzed>
works, but is ugly imho

[11:13:00.0000] <Philip`>
(On your original question about using a different BeautifulSoup.py: It may be sufficient to just put it in Python's search path, like by using PYTHON_PATH=some-directory-which-contains-that-file)

[11:13:01.0000] <cryzed>
Philip`, okay

[11:13:02.0000] <cryzed>
I could just rename it to BeautifulSoup.py

[11:13:03.0000] <cryzed>
and place it locally

[11:13:04.0000] <cryzed>
next to my script

[11:13:05.0000] <cryzed>
probably

[11:13:06.0000] <Philip`>
I'm not certain but I think that ought to get picked up when html5lib tries loading it

[11:14:00.0000] <Philip`>
/me has to go away

[11:15:00.0000] <cryzed>
Well oke

[11:16:00.0000] <cryzed>
thanks Philip` 

[11:16:01.0000] <cryzed>
I'll get myself the newest html5lib

[11:16:02.0000] <cryzed>
and just tune my old lib

[11:16:03.0000] <cryzed>
with BeautifulSoup a bit

[11:16:04.0000] <cryzed>
and hope that it works

[12:42:00.0000] <hsivonen>
"I don't think anyone wants to break plugins" - iPhone, anyone?

[12:42:01.0000] <hsivonen>
(quote from HTML WG telecon minutes)

[12:43:00.0000] <smedero>
it was a surreal a couple of moments... for sure.

[12:43:01.0000] <gsnedders>
hsivonen: Well, arguably it never broke

[13:45:00.0000] <tantek>
hsivonen - neither iPhone nor BlackBerry browsers support Flash / plugins

[14:07:00.0000] <gsnedders>
/me feels bad…

[14:07:01.0000] <gsnedders>
Appealing to authority :(

[14:08:00.0000] <takkaria>
some authority is good

[14:08:01.0000] <gsnedders>
annevk?

[14:16:00.0000] <gsnedders>
ARGH!

[14:16:01.0000] <gsnedders>
/me gets annoyed at Google again

[14:16:02.0000] <gsnedders>
I can't simply look for anything relating to Lolita without getting a ton of results of porn

[14:26:00.0000] <Philip`>
gsnedders: Maybe you shouldn't be using the image search with SafeSearch off

[14:27:00.0000] <gsnedders>
Philip`: I'm not using image search. That doesn't help me write an English dissertation.

[14:40:00.0000] <gsnedders>
Hmm… Nabokov almost always uses to sob and rarely to cry…

[15:42:00.0000] <Hixie>
heycam: yt?

[15:43:00.0000] <Hixie>
i have a method take takes as its argument an array of values

[15:43:01.0000] <Hixie>
the values are typed

[15:44:00.0000] <Hixie>
er, i mean, an array of arrays of values, which are typed

[15:44:01.0000] <Hixie>
let's say, it's an array of arrays of DOMString, long pairs

[15:44:02.0000] <Hixie>
e.g.

[15:44:03.0000] <Hixie>
foo([['a', 1], ['b', 2], ['c', 3]]);

[15:44:04.0000] <Hixie>
is there a sane way to describe that in WebIDL?


2009-04-10
[18:20:00.0000] <heycam>
Hixie, nope, no way to specify that

[18:21:00.0000] <Hixie>
how should i do it?

[18:22:00.0000] <heycam>
i do have to fix up arrays/sequences quite a bit, but maybe you could do sequence<sequence<Object>>?

[18:23:00.0000] <Hixie>
k

[20:40:00.0000] <karlcow>
[13:08]  <cryzed> And I somehow think that using BeautifulSoup is easier than lxml

[20:40:01.0000] <karlcow>
Did you see http://blog.ianbicking.org/2008/12/10/lxml-an-underappreciated-web-scraping-library/

[22:48:00.0000] <john_fallows>
Hixie: fyi, i just subscribed to hybi

[00:09:00.0000] <Hixie>
john_fallows: cool

[00:09:01.0000] <john_fallows>
just read through all the hybi archives *whew*

[00:10:00.0000] <Hixie>
heh

[00:14:00.0000] <john_fallows>
tell me if i understand rHTTP correctly - someone wanted to relax the client-initiated constraints of HTTP, so they used 101 Switching Protocols with Upgrade header, and then chose the same half-duplex protocol only in the opposite direction?

[00:21:00.0000] <john_fallows>
rHTTP seems to be missing the point, once you have completed the Upgrade, you could choose to leverage the full power of raw TCP at that point, letting you implement rHTTP or any other protocol

[00:23:00.0000] <john_fallows>
of course, JavaScript has no binary type (yet) so you would need minimal framing to determine raw binary (just like TCP) versus text - giving us WebSocket as currently specified

[00:24:00.0000] <Hixie>
you've pretty much summed up what appears to be the current state of affairs

[00:25:00.0000] <Hixie>
there also seems to be a lack of understanding of the web browser security model

[00:26:00.0000] <Hixie>
which doesn't help

[00:26:01.0000] <Hixie>
but anyway

[00:26:02.0000] <john_fallows>
that's a good point, and critically important in understanding where WebSocket came from

[00:27:00.0000] <john_fallows>
there was also some discussion about formalizing how proxies should behave in the presence of an Upgrade header

[00:28:00.0000] <john_fallows>
i thought that the Upgrade header was hop-by-hop by design, so it wouldn't be appropriate to give it any meaning beyond the current hop - necessitating CONNECT instead, just like https, right?

[00:31:00.0000] <john_fallows>
or is the concern more about transparent HTTP proxies that would not receive a CONNECT ?

[00:31:01.0000] <Hixie>
i'm not an expert when it comes to proxies

[00:32:00.0000] <john_fallows>
in any case, it's probably a good idea for the IETF to formalize how proxies (transparent or otherwise) should behave in the presence of a 101 Switching Protocols response

[00:36:00.0000] <Hixie>
mcarter and i had dinner the other day and discussed some of the feedback, btw; in particular around framing, and other ways to do the handshake

[00:37:00.0000] <Hixie>
i think we concluded that there were some options but none especially compelling

[00:37:01.0000] <Hixie>
in particular mcarter had some ideas around faking the SSL handshake instead of the HTTP one

[00:37:02.0000] <Hixie>
(faking a zero-encryption handshake)

[00:39:00.0000] <Hixie>
but that doesn't really work well when you think about it, sadly

[00:40:00.0000] <john_fallows>
i think i've also sent similar feedback to the public mailing list and you pointed out that it might introduce security concerns - which i agree with

[00:41:00.0000] <john_fallows>
btw, there are at least 2 ways to fake a zero encryption handshake

[00:41:01.0000] <Hixie>
oh?

[00:41:02.0000] <john_fallows>
1) negotiate null encryption via SSL

[00:42:00.0000] <john_fallows>
2) negotiate full encryption via SSL, and then use the subsequent "Upgraded" protocol to negotiate removal of the encryption

[00:42:01.0000] <Hixie>
true

[00:43:00.0000] <Hixie>
anyway. long and short of it is, i didn't make any changes

[00:43:01.0000] <Hixie>
not really sure what to do on hybi though

[00:43:02.0000] <john_fallows>
i think my previous feedback to the mailing list was #1, but #2 has more potential :)

[00:43:03.0000] <Hixie>
right now it's just me pushing a websocket-like approach

[00:44:00.0000] <Hixie>
and a bunch of others pushing elaborate multi-level protocols

[00:44:01.0000] <john_fallows>
well, i'll be around there to help others to understand your brilliance :)

[00:44:02.0000] <Hixie>
websocket is more mcarter's brilliance than mine at this point :-)

[00:45:00.0000] <Hixie>
ok time to go to bed

[00:45:01.0000] <Hixie>
nn

[00:45:02.0000] <john_fallows>
don't know if you realize this - but mcarter and i both designed a lot of his proposals together :-)

[00:46:00.0000] <john_fallows>
see you on hybi

[00:46:01.0000] <john_fallows>
nn

[01:02:00.0000] <sayrer>
/me reads @hybi and feels sorry for Hixie

[01:33:00.0000] <Philip`>
http://www.crummy.com/2009/04/09/0 - "I'll be writing an html5lib tree builder and packaging it and the lxml builder in Beautiful Soup for a while, but I think long-term the TreeBuilders should live with their parent projects."

[01:39:00.0000] <hsivonen>
I think the reception of Web Socket on hybi is sad but not surprising.

[01:57:00.0000] <hsivonen>
(it seems like the usual case of various people wanting to do very different things but getting bundled as one group because the may seem to be some commonality)

[02:03:00.0000] <hsivonen>
vipood suggests that I used several DNS servers to check the IP address of their site to restore my confidence until they restore the EV cert

[02:09:00.0000] <MikeSmith>
hsivonen: so what was the problem?

[02:36:00.0000] <hsivonen>
MikeSmith: a "hiccup"

[02:37:00.0000] <MikeSmith>
aka "fuckup" (I'd guess)

[02:42:00.0000] <hsivonen>
Hixie: did you happen to consider data-* elements, too when you added data-* attributes to the spec?

[03:04:00.0000] <MikeSmith>
hsivonen: I just tested and checked in the schema/assertions changes for <keygen>

[03:04:01.0000] <MikeSmith>
and 12+12 new tests (html+xhtml)

[03:05:00.0000] <MikeSmith>
writing the tests actually helped me catch a couple things I'd forgottent to add to the assertions.sch file

[03:07:00.0000] <MikeSmith>
so I guess I should complain less about making tests being a PITA

[03:08:00.0000] <hsivonen>
MikeSmith: nice!

[03:10:00.0000] <hsivonen>
Hixie: context for data-* element question: http://intertwingly.net/blog/2009/04/08/HTML-Reunification#c1239352657

[03:29:00.0000] <hsivonen>
http://lists.w3.org/Archives/Public/public-xhtml2/2009Apr/0015.html

[03:35:00.0000] <MikeSmith>
hsivonen: lemme please know when you got latest v.nu deployed

[03:40:00.0000] <hsivonen>
MikeSmith: ok. I'm currently at an IP address that isn't allowed to trigger deployment

[03:40:01.0000] <hsivonen>
http://twitter.com/bradneuberg/status/1478859862

[03:43:00.0000] <hsivonen>
http://steve-yegge.blogspot.com/2009/04/have-you-ever-legalized-marijuana.html

[03:43:01.0000] <hsivonen>
the spec item list in there is illustrative of many drive-by Web spec enhancement requests

[03:44:00.0000] <hsivonen>
particularly the point starting "WHAT THE EVER-LOVING *FUCK* ARE YOU PEOPLE SMOKING?"

[03:44:01.0000] <takkaria>
oh, I just ordered that book

[03:45:00.0000] <MikeSmith>
which "canvas shim" is Brad referring to, I wonder

[03:45:01.0000] <MikeSmith>
hsivonen: I like the "I personally do this all the time, except that I'm usually the ugly one." part

[03:48:00.0000] <MikeSmith>
/me scrolls to find the parts about smoking marijuana

[03:49:00.0000] <MikeSmith>
"I have smoked marijuana (and inhaled it, deeply) on more occasions than I can count. And yet I'm almost undoubtedly smarter than your kid that you're so goddamned worried about."

[03:49:01.0000] <MikeSmith>
well put

[03:54:00.0000] <MikeSmith>
hsivonen: have you read Fred Brooks's "The Mythical Man-Month"?

[03:57:00.0000] <hsivonen>
MikeSmith: yes 

[03:58:00.0000] <hsivonen>
My boss at a previous job gave it to me when I left saying that I was a "productive individual"

[03:58:01.0000] <MikeSmith>
heh

[03:58:02.0000] <MikeSmith>
it's surprising to me how many people I've known in software product development who've not read that book

[03:58:03.0000] <MikeSmith>
this Steve Yegge post reminds me of it

[03:59:00.0000] <hsivonen>
s/previous/past/ I should say

[04:00:00.0000] <MikeSmith>
the decision-makers unfortunately don't read things like Steve's blog, nor Brooks

[04:00:01.0000] <MikeSmith>
they read instead stuff like "Who Moved My Cheese"

[04:00:02.0000] <MikeSmith>
or something like that

[04:01:00.0000] <MikeSmith>
books with like 10 words on each page, plus cartoons

[04:01:01.0000] <MikeSmith>
and then they distribute them to everybody in the company, because they think that stuff is brilliant

[04:02:00.0000] <MikeSmith>
I don't know how people get into those positions

[04:02:01.0000] <MikeSmith>
what the criteria are

[04:02:02.0000] <MikeSmith>
success certainly isn't a criterion in many cases

[04:03:00.0000] <MikeSmith>
I read today that the guy who's running Motorola's mobile business is the highest-paid software executive in the US

[04:03:01.0000] <MikeSmith>
despite the fact that part of their business is a massive failure

[04:07:00.0000] <MikeSmith>
the mobile part of their business being a failure, I mean

[04:18:00.0000] <takkaria>
good article, that

[04:18:01.0000] <MikeSmith>
aye

[04:19:00.0000] <takkaria>
I do work with a UK national drug law reform NGO and they're putting together proposals which try and answer most of his questions

[04:20:00.0000] <takkaria>
and yes, it is actually a pain in the arse to work out even how you go about decriminalising just possession well

[04:20:01.0000] <MikeSmith>
I posted a comment there under the alias "Winston"

[04:21:00.0000] <MikeSmith>
not really

[04:21:01.0000] <MikeSmith>
anyway, his "Because laws are pretty much like programs." premise is wrong

[04:21:02.0000] <MikeSmith>
as some of the comments point out

[04:22:00.0000] <MikeSmith>
but that doesn't mean most of his points aren't still valid

[04:26:00.0000] <takkaria>
the biggest obstacle to countries doing anything positive towards drug legalisation are the UN treaties on the matter, until they're gone there's a limit to what states can do in terms of a regulatory framework

[04:26:01.0000] <hsivonen>
MikeSmith: "Winston" overlooks the issue that stuff is way more complex than "my body is mine" in the case of distributed health care costs and the moral imperative not to just let sick people die

[04:26:02.0000] <MikeSmith>
hsivonen: yeah, true

[04:27:00.0000] <MikeSmith>
"my body is mine" is not the point

[04:27:01.0000] <MikeSmith>
plus, I think he made up the word "ayahuasca"

[04:28:00.0000] <MikeSmith>
ah, geez

[04:28:01.0000] <MikeSmith>
Wikipedia article on ayahuasca is great

[04:28:02.0000] <MikeSmith>
"any of various psychoactive infusions or decoctions prepared from the Banisteriopsis spp. vine, usually mixed with the leaves of the Psychotria bush"

[04:28:03.0000] <MikeSmith>
seems like even more made-up words

[04:30:00.0000] <MikeSmith>
hsivonen: obesity is a bigger problem that marijuana, as far as distributed health-care costs

[04:30:01.0000] <takkaria>
I think "And this is why Britain will continue to have a monarchy for centuries. The benefits of being a republic will never outweigh the effort of arging about what to put on postage stamps." is the best comment

[04:31:00.0000] <MikeSmith>
/me looks forward to the day when a fat marijuana bud appears on a postage stamp

[04:33:00.0000] <hsivonen>
MikeSmith: very likely. my point is that health care (and the investment in a person's education) make policy issues more complex than argumenting from pure individualism

[04:33:01.0000] <MikeSmith>
hsivonen: yeah, absolutely

[04:34:00.0000] <MikeSmith>
this is the key flaw in the so-called "libertarian" political philosophy

[04:35:00.0000] <MikeSmith>
I say "so-called" because that's what it's known by the last 50 years or so in the US at least

[04:35:01.0000] <MikeSmith>
but it's something quite different from what political libertarianism originally meant

[04:36:00.0000] <hsivonen>
the political vocabulary in the U.S. around liber* is quite transformed

[04:36:01.0000] <MikeSmith>
I don't mean the health-care issue specifically, but the general issue policy issues being more complex than pure individualism

[04:37:00.0000] <MikeSmith>
hsivonen: the political vocabulary of the US is like a pidgin language

[04:37:01.0000] <MikeSmith>
one of the few places that's worse is Japan, where there's no real debate at all

[04:37:02.0000] <MikeSmith>
nor no real political parties

[04:54:00.0000] <hsivonen>
http://m.alistapart.com/ what happened to CSS and media queries?

[05:00:00.0000] <hsivonen>
heh. jd worrying about the barrier for entry for new browsers: http://blogs.adobe.com/jd/2009/03/pervin_the_standards.html

[05:02:00.0000] <MikeSmith>
I bet jd is the kind of guy who spits a lot when he talks

[05:03:00.0000] <MikeSmith>
and I guarantee he does not smoke marijuana

[05:06:00.0000] <MikeSmith>
I wonder, if Adobe fired jd, what he would do

[05:12:00.0000] <gsnedders>
Is the barrier to entry not always going to get higher as more and more features are added?

[05:13:00.0000] <gsnedders>
So should we have never added the img element because it raised the barrier of entry?

[05:17:00.0000] <hsivonen>
/me mumbles about having to create a special Flash host process as a barrier to entry for Chrome

[05:55:00.0000] <roc>
I happen to agree with jd that CSS animations is a slightly dubious application of CSS

[05:55:01.0000] <roc>
I think that's about the only thing I agree with him about, though

[05:56:00.0000] <gsnedders>
/me posts a short comment, and will almost certainly not look at jd's blog for another year or so, and completely forget he ever published this comment

[07:13:00.0000] <hsivonen>
at the office now. Will redeploy v.nu

[11:05:00.0000] <gsnedders>
Philip`: Do you not have a working XML DOM Viewer?

[11:15:00.0000] <Philip`>
gsnedders: You mean like http://canvex.lazyilluminati.com/misc/dom-viewer/x.html ?

[11:54:00.0000] <gsnedders>
Philip`: Yes

[12:15:00.0000] <Philip`>
gsnedders: In that case, I don't have such a thing

[12:31:00.0000] <sayrer>
boy, webkit does not like hsivonen's downloadable font pages

[12:32:00.0000] <sayrer>
I thought Firefox was acting wonky

[12:34:00.0000] <hsivonen>
sayrer: just a flash of unfontified content or something worse?

[12:34:01.0000] <sayrer>
hsivonen, Firefox has a flash

[12:34:02.0000] <sayrer>
hsivonen, WebKit draws with no face at all

[12:34:03.0000] <sayrer>
Safari4 beta, I should say

[12:35:00.0000] <hsivonen>
sayrer: oh. that's bad.

[12:35:01.0000] <hsivonen>
sayrer: thanks

[12:35:02.0000] <sayrer>
hsivonen, once all the fonts are downloaded, it picks one

[12:35:03.0000] <hsivonen>
I thought the worst rendering issue was that the sans-serif font doesn't hint nicely in ClearType

[12:35:04.0000] <hsivonen>
it has been designed for real anti-aliasing

[12:39:00.0000] <sayrer>
http://people.mozilla.com/~sayrer/2009/04/hsivonen.png

[15:37:00.0000] <jwalden>
hsivonen: "sprikled" in the producing-xml page

[16:32:00.0000] <weinig_>
sayrer: out of curiosity, what would you expect hsivonen's producing-xml page to look like before the fonts have loaded?

[16:33:00.0000] <sayrer>
weinig_, no idea. I was using Firefox and saw a flash of unfontedness, and checked out Safari 4 to see how you guys handled it

[16:34:00.0000] <weinig_>
sayrer: ah


2009-04-11
[19:54:00.0000] <gsnedders>
/me wonders where the j came from in olliej 

[19:55:00.0000] <olliej>
gsnedders: hehe

[19:55:01.0000] <gsnedders>
olliej: Middle name?

[19:55:02.0000] <olliej>
gsnedders: yup

[19:55:03.0000] <olliej>
gsnedders: many years ago ollie wasn't available

[19:55:04.0000] <gsnedders>
olliej: Damnit! I wanted you to say no to that so I could guess maiden name!

[19:55:05.0000] <olliej>
gsnedders: and realistically it was either olliej or n|tr0 ;)

[19:57:00.0000] <gsnedders>
olliej: Well, I rule by having a far less common name. :P

[19:57:01.0000] <gsnedders>
And you fail by going offline. :P

[02:12:00.0000] <Philip`>
http://search.live.com/results.aspx?q=pdf - first search result is "a PDF Viewer for RISC OS", which doesn't sound terribly helpful - I guess I'll have to stick with Google for a while longer :-(

[02:45:00.0000] <gsnedders>
Oh, the tyranny!

[03:12:00.0000] <hsivonen>
/me wonders how Java AppEngine deal with threads and cross-thread static objects

[03:14:00.0000] <hsivonen>
"A Java application cannot create a new java.lang.ThreadGroup nor a new java.lang.Thread."

[03:14:01.0000] <hsivonen>
as expected...

[03:14:02.0000] <Philip`>
hsivonen: http://code.google.com/appengine/docs/java/runtime.html - "A Java application cannot create a new java.lang.ThreadGroup nor a new java.lang.Thread."

[03:14:03.0000] <Philip`>
Oops, too late

[03:17:00.0000] <hsivonen>
boohoo. they used JDK logging instead of a log4j mockup :-(

[03:17:01.0000] <hsivonen>
I don't see how static members work

[03:17:02.0000] <Philip`>
The Java WORA idea seems to be pretty much dead

[03:18:00.0000] <hsivonen>
what are the latest news on the projected death of Sun itself?

[03:19:00.0000] <hsivonen>
(IBM buying Sun, that is)

[03:19:01.0000] <Philip`>
I think that's expected in about 5 billion years

[03:19:02.0000] <hsivonen>
not $5 billion?

[03:20:00.0000] <Philip`>
(but I've watched Sunshine so I know it might happen within about 50 years unless some people successfully detonate a giant nuclear bomb inside the Sun)

[03:42:00.0000] <hsivonen>
if I ported v.nu to app engine, the quotas would prevent validating the html5 spec

[04:32:00.0000] <takkaria>
Philip`: that's the third result on google, too

[04:35:00.0000] <gsnedders>
takkaria: Challenge: what song did I have stuck in my head when I woke up this morning?

[04:37:00.0000] <takkaria>
It's A Sin by the Pet Shop Boys

[04:38:00.0000] <gsnedders>
/me is listening to Death Row by System Fault from Perfect Gentlemen EP

[04:38:01.0000] <gsnedders>
takkaria: ^^

[04:38:02.0000] <takkaria>
oh, I'm sorry to hear that. :)

[04:39:00.0000] <gsnedders>
takkaria: Damned you linking to such things on Last.fm!

[04:39:01.0000] <gsnedders>
/me deletes four hundred words from his English dissertation

[04:41:00.0000] <gsnedders>
(It's still over 1100 words too long.)

[04:46:00.0000] <hsivonen>
https://build.mozilla.org/tryserver-builds/2009-04-10_10:28-hsivonen⊙if/

[04:47:00.0000] <hsivonen>
the latest batch of html5 parsing tryserver build sucks less than html5 parsing builds ever!

[04:48:00.0000] <takkaria>
win

[04:51:00.0000] <hsivonen>
(they no longer seem to crash on cnn.com)

[12:48:00.0000] <hsivonen>
jwalden: fixed the typo in http://hsivonen.iki.fi/producing-xml/ thanks!

[15:17:00.0000] <Lachy>
Bonjour

[15:17:01.0000] <Lachy>
/me is in Paris

[15:24:00.0000] <jcranmer>
Comment ça va?

[15:26:00.0000] <Philip`>
Mon aéroglisseur est plein d'anguilles?

[15:30:00.0000] <Lachy>
Philip`, LOL :-D

[16:03:00.0000] <Philip`>
Wow, I think I woke my MBP out of sleep mode by sending a static shock through the metal shelf it's resting on

[16:03:01.0000] <takkaria>
hm, tasty


2009-04-12
[09:38:00.0000] <Philip`>
http://intertwingly.net/blog/2004/10/20/Attractive-Nuisance#c1239547517 - the line between markup and text seems to get ignored quite frequently

[10:25:00.0000] <eighty4>
Any of the authors to html4.01 in here?

[10:27:00.0000] <eighty4>
we got a discussions about the semantic meaning of <hr> and <center> and why center is deprecated but hr isn't anyone happen to know the actual reason?

[11:08:00.0000] <Dashiva>
*sigh*

[11:09:00.0000] <Dashiva>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=6684#c29

[12:57:00.0000] <MathiasBynens>
I remember reading this in the HTML5 specs, but I can't seem to find the exact link again... Is it okay to use vertical pipe characters to separate navigation list items, f.e. <p id="navigation"><a href="#link-1">Bla</a> | <a href="#link-2">Foo</a> | <a href="#link-3">Bar</a></p>

[12:57:01.0000] <Hixie>
sure

[12:57:02.0000] <jcranmer>
what about &middot; ?

[12:57:03.0000] <MathiasBynens>
Any hints on where this is defined in the spec? ;)

[12:58:00.0000] <jcranmer>
i.e., •

[13:04:00.0000] <MathiasBynens>
jcranmer, the character you just typed isn't &middot;

[13:05:00.0000] <MathiasBynens>
This is: �

[13:06:00.0000] <jcranmer>
huh, must be because the text is bold at the same time

[13:18:00.0000] <Hixie>
MathiasBynens: paragraphs can contain anything they want, text-wise

[13:18:01.0000] <Hixie>
MathiasBynens: i don't think the spec says what text is ok, it's just assumed that, well, anything the spec doesn't disallow is allowed


2009-04-13
[21:14:00.0000] <MikeSmith>
eighty4: the HTML5 draft redefines <hr> as representing a "thematic break"

[21:14:01.0000] <MikeSmith>
basically the same purpose that the XHTML2 spec uses the <seperator> element for

[21:14:02.0000] <MikeSmith>
http://www.w3.org/TR/xhtml2/mod-structural.html#edef_structural_separator

[05:24:00.0000] <MikeSmith>
hsivonen: btw, NicolasRaoul is the same who posted a v.nu-related patch to the whatwg implementors list today

[05:24:01.0000] <MikeSmith>
http://lists.whatwg.org/pipermail/implementors-whatwg.org/2009-April/000253.html

[05:25:00.0000] <MikeSmith>
he's at W3C/Keio for while part-time as an intern, and been looking from something Web-related and new to help work on

[05:26:00.0000] <MikeSmith>
and he's got some Java chops

[05:27:00.0000] <MikeSmith>
so I suggested he might take a look at v.nu -- the FIXMEs and such -- and see if there's anything that looked interesting

[11:04:00.0000] <Philip`>
http://www.cssquirrel.com/comic/?comic=13

[13:16:00.0000] <sgarrity>
If none of the source videos provided in a <video> element are playable by the client, is it the intended behaviour that you get the video player, but with no viewable video, and not the fall-back content?

[13:30:00.0000] <Hixie>
heycam`: empty string, but wow, i somehow didn't put that in the spec! will fix

[14:14:00.0000] <sayrer>
hmm, Josh Soref works for Nokia I think. Not Mozilla in any case

[14:16:00.0000] <sayrer>
Nokia's Maemo browser ("MicroB") and some other things are based on Mozilla

[16:15:00.0000] <annevk2>
annevk2? hmm I thought I tunred my laptop at home off...

[16:15:01.0000] <annevk2>
oh, annevk5 then

[16:16:00.0000] <annevk5>
hmm ok, so I was connected to both freenode.net and .org at the same time from the same laptop


2009-04-14
[17:51:00.0000] <heycam`>
Hixie, thanks

[18:01:00.0000] <sayrer>
"just" use BEEP

[18:01:01.0000] <sayrer>
lol wut

[18:02:00.0000] <Hixie>
if the beep advocates can write a DOM API spec for BEEP, adapt BEEP to work in the browser same-origin security model, and convince browser vendors to implement it, I have no objections to it

[18:02:01.0000] <Hixie>
but i think it would still be good to have a simple protocol like websocket as well :-)

[18:03:00.0000] <Hixie>
(not necessarily websocket itself)

[18:03:01.0000] <sayrer>
yeah, that's basically my issue with the entire discussion

[18:04:00.0000] <sayrer>
I almost wrote a message asking maciej whether WebSocket prevented the design he was adovocating

[18:05:00.0000] <Hixie>
i figure we can add an optional second argument to the constructor, and have that be part of the handshake -- and if the server doesn't send back the same string, abort with a suitable error code.

[18:05:01.0000] <Hixie>
WebSocket-Subprotocol: or something

[18:05:02.0000] <Hixie>
maybe we can even get the IANA to give us a registry for subprotocol types

[18:05:03.0000] <sayrer>
I think rfc3117 is my favorite

[18:06:00.0000] <sayrer>
until it goes off into BEEP land

[18:06:01.0000] <sayrer>
it's like the finale of battlestar galactica

[18:06:02.0000] <doublec>
no spoilers please...

[18:06:03.0000] <doublec>
:)

[18:07:00.0000] <doublec>
for those of us in way behind tv land

[18:07:01.0000] <doublec>
i'll forever be anticipating rfc3117 references in msg now

[18:07:02.0000] <doublec>
s/msg/bsg

[18:08:00.0000] <sayrer>
fortunately there is nothing overt!

[18:08:01.0000] <doublec>
hehe

[18:09:00.0000] <sayrer>
mmm, this hybi list is intensely unproductive

[18:09:01.0000] <sayrer>
amazing

[18:09:02.0000] <sayrer>
I just don't get it

[18:10:00.0000] <sayrer>
1) solve the smallest problem you possibly can

[18:10:01.0000] <sayrer>
2) declare victory

[18:10:02.0000] <sayrer>
3) ???

[18:10:03.0000] <sayrer>
4) Profit!

[18:10:04.0000] <sayrer>
why do anything else?

[18:13:00.0000] <Hixie>
sayrer: a lot of the people on that list are already on step 4 and are wondering why we are talking about changing their steps 1 and 2

[18:14:00.0000] <sayrer>
Hixie, are you being sympathetic to me or them?

[18:14:01.0000] <Hixie>
neither :-)

[18:15:00.0000] <Hixie>
i believe that a lot of the pushback we're getting is because there is already a thriving business around these involved protocols

[18:15:01.0000] <sayrer>
good point

[18:15:02.0000] <sayrer>
maybe we should ask for a new WG

[18:17:00.0000] <sayrer>
"The good thing about reinventing the wheel is that you can get a round one."

[18:17:01.0000] <sayrer>
I rather like that quote

[18:18:00.0000] <sayrer>
and it was about XML vs JSON

[18:18:01.0000] <Hixie>
i'm reading rfc3117 again -- it feels like when you hit section 3.4 it suddenly goes from being a useful overview of the state of affairs to being an exercise in second system syndrome.

[18:20:00.0000] <sayrer>
yeah

[18:26:00.0000] <sayrer>
but it's great because it is 8 years old and still a useful overview

[02:01:00.0000] <annevk5>
/me joins the hybi discussion but isn't sure he should've done so

[02:02:00.0000] <annevk5>
/me wonders what http://twitter.com/kevinwhinnery/statuses/1511868625 is about

[02:12:00.0000] <MikeSmith>
zcorpan: yeah, I noticed the sub/sup case also

[02:13:00.0000] <MikeSmith>
zcorpan: dealing with those actually turns out to be a major PITA

[02:14:00.0000] <MikeSmith>
at least for someone like me with limited coding skills

[02:14:01.0000] <MikeSmith>
but I messed around with it again today and have a fix

[02:14:02.0000] <zcorpan>
MikeSmith: the spec should be less human readable and be easier to scrape

[02:14:03.0000] <MikeSmith>
he

[02:14:04.0000] <MikeSmith>
heh

[02:14:05.0000] <MikeSmith>
yeah

[02:14:06.0000] <MikeSmith>
Hixie: please optimize the script for scraping!

[02:14:07.0000] <jgraham>
/me sees reams of hiby discussion, hopes that it is not too interesting

[02:15:00.0000] <jgraham>
s/hiby/hybi/

[02:15:01.0000] <zcorpan>
MikeSmith: you can look for <dl class="element"> and its previous element sibling

[02:15:02.0000] <zcorpan>
s/element/heading element/

[02:16:00.0000] <MikeSmith>
I could, but it'd mean more rewriting and testing to make sure I didn't break stuff

[02:16:01.0000] <MikeSmith>
zcorpan: it would be slightly easier if Hixie consistently used id=the-foo-element for all element h4s

[02:16:02.0000] <annevk5>
jgraham, it seems mostly about bidirectional HTTP / BEEP vs something close but not quite raw TCP sockets

[02:16:03.0000] <MikeSmith>
zcorpan: though that wouldn't work for h1-h6 or sub/sup case

[02:17:00.0000] <annevk5>
jgraham, obviously these two are pretty far apart so the discussion isn't really going anywhere

[02:19:00.0000] <jgraham>
annevk5: Oh. Well I guess it is safe to ignore for a while at least then

[02:33:00.0000] <annevk5>
http://twitpic.com/3aqou

[02:42:00.0000] <annevk5>
http://decafbad.com/blog/2009/04/13/i-like-revcanonical so besides that we're trying to remove rev, isn't canonical meant to be same-origin?

[03:08:00.0000] <hsivonen_>
annevk5: is this rev=canonical different from Google/Y!/MS rel=canonical?

[03:09:00.0000] <hsivonen_>
magic 8 ball says Web authors won't be able to use the distinction of rel and rev right here

[03:09:01.0000] <annevk5>
it is

[03:10:00.0000] <annevk5>
something like rel=shorturl would be better I think

[03:11:00.0000] <hsivonen_>
I can imagine all sorts of blog posts about evil HTML5 raining on the rev=canonical backpattery parade

[03:11:01.0000] <svl>
Mostly (from what I've seen) it's been "let's all use this en-masse, so html5 will be forced to include this".

[03:12:00.0000] <annevk5>
there's that too, but mostly using it just doesn't really add up

[03:12:01.0000] <annevk5>
canonical is for resource equivalence, not URL-after-redirect equivalence; canonical also has a same-origin restriction

[03:13:00.0000] <hsivonen_>
the tinyurl centralization problem would be mostly solved if twitter just showed the real URLs in tweets on every medium except the ephemeral SMSs

[03:24:00.0000] <Philip`>
I've seen quite a few people starting to use TinyURL in blog comments and blog posts and other places where it's completely unnecessary

[03:27:00.0000] <annevk5>
i blogged about rev=canonical; lets see if someone picks it up

[03:27:01.0000] <annevk5>
maybe in a few years all standards will be made through blogging

[03:28:00.0000] <hsivonen_>
/me mumbles about feed autodiscovery

[03:28:01.0000] <annevk5>
true that :)

[03:29:00.0000] <annevk5>
would be nice if rel=feed got some more impl love

[03:34:00.0000] <annevk5>
rel=canonical too btw

[03:35:00.0000] <annevk5>
there's also http://revcanonical.appspot.com/

[03:36:00.0000] <annevk5>
and apparently someone added "short" to http://wiki.whatwg.org/wiki/RelExtensions

[03:41:00.0000] <hsivonen_>
why am I not seeing the standards suck guy in Minefield on Anne's blog?

[03:44:00.0000] <annevk5>
does Minefield support SVG referenced from 'content'?

[03:58:00.0000] <annevk5>
heycam, FYI, empty string is fine for ID because an empty ID cannot exist

[03:59:00.0000] <heycam>
annevk5, fine with me, was just looking for it being defined one way or the other

[04:04:00.0000] <annevk5>
heycam, yeah, weird omission (it's defined for all other DOMString reflecting thingies)

[04:14:00.0000] <Philip`>
/me sees http://o-micron.blogspot.com/ (with some storage-related stuff)

[04:15:00.0000] <annevk5>
that's from the Oracle guy who's in favor of something like Atom for storage, no?

[04:18:00.0000] <Philip`>
Looks like it probably is

[04:18:01.0000] <hsivonen_>
annevk5: I don't know if Minefield is supposed to support SVG in CSS 'content'

[04:19:00.0000] <hsivonen_>
looks like the XHTML2 WG rep used the wrong name for their WG: http://lists.w3.org/Archives/Public/public-pfwg-comments/2009AprJun/0037.html

[04:19:01.0000] <annevk5>
they call themselves the XHTML WG quite often

[04:19:02.0000] <annevk5>
and call us the HTML5 WG

[04:20:00.0000] <annevk5>
"us" being the HTML WG, not the WHATWG

[04:20:01.0000] <annevk5>
whenever that happens I wonder if it's on purpose

[04:21:00.0000] <jgraham>
I always assume that it is on purpose

[04:22:00.0000] <annevk5>
outside the W3C it's often referred to as the HTML5 WG as well

[04:23:00.0000] <Philip`>
Maybe it's to disambiguate from the old HTML WG

[04:25:00.0000] <jgraham>
Somebody should teach the Oracle guy the golden rule of academic qualifications, which is that you should never, ever put them after your name

[04:25:01.0000] <Philip`>
What about putting them before your name?

[04:25:02.0000] <jgraham>
/me was waiting for that

[04:26:00.0000] <Philip`>
Or in the middle of your name?

[04:26:01.0000] <Philip`>
John "Dr" Smith, etc

[04:27:00.0000] <jgraham>
Philip`: At least nobody can be expected to take those things seriously

[04:27:01.0000] <Philip`>
What if you skip the name entirely and just call yourself Doctor?

[04:27:02.0000] <jgraham>
Although they have roughly the same effect of making you look like a right twat

[04:28:00.0000] <jgraham>
Philip`: Only works if you are a time lord

[04:30:00.0000] <hsivonen>
/me assumes that "some browsers" in ARIA is an euphemism for IE

[04:31:00.0000] <annevk5>
yeah...

[04:33:00.0000] <annevk5>
hsivonen, thanks for replying to that XHTML2 WG comment btw

[04:33:01.0000] <annevk5>
I was wondering whether I should do it

[04:34:00.0000] <hsivonen>
annevk5: I almost wrote feedback about the same spec sentence before seeing the XHTML2 WG comment.

[04:35:00.0000] <hsivonen>
the name "ARIA Best Practices" irks me. ARIA is so new, I don't believe Best Practices have emerged yet

[04:35:01.0000] <hsivonen>
should be Authoring Guide or something like that

[04:36:00.0000] <hsivonen>
Best Practices and Guidelines are weasel words in general

[04:36:01.0000] <annevk5>
if it ever gets widely adopted :/

[05:14:00.0000] <zcorpan>
what's focal locus?

[05:16:00.0000] <hsivonen>
zcorpan: the point of focus?

[05:21:00.0000] <virtuelv>
"In mathematics, a locus (Latin for "place", plural loci) is a collection of points which share a property. "

[05:22:00.0000] <virtuelv>
http://en.wikipedia.org/wiki/Locus_(mathematics)

[05:47:00.0000] <hsivonen>
/me has 63 public-pfwg-comments autoreplies

[06:07:00.0000] <annevk5>
I complained about that and was told it was a special list :/

[06:08:00.0000] <jgraham>
The pfwg list is insane. I can only imagine that the point of it is to deter comments on the spec

[06:09:00.0000] <annevk5>
/me has to run

[06:47:00.0000] <remysharp>
A question about <article> - I've seen in examples that <article> is nested within a <section> - but the page I put together goes as: <article><section> - is there any specific reason I might be wrong, or is it actually down to content layout?

[06:48:00.0000] <jgraham>
remysharp: That is good and more likely sounding than the other way around

[06:49:00.0000] <remysharp>
sorry - just to be clear are you saying that it *should* be <article><section> then?

[06:50:00.0000] <jgraham>
e.g. <article><h1>A Celebrity Did Something</h1><section><h2>Summary</h2></section><section><h2>So civilization will end</h2></section></article>

[06:50:01.0000] <remysharp>
perfect - that's what I understood the spec to mean.

[06:50:02.0000] <remysharp>
cheers for that.

[06:51:00.0000] <jgraham>
remysharp: Neither way is "right" it's just more obvious why an article would have multiple <section>s than a <section> would have multiple articles

[06:52:00.0000] <jgraham>
(an example of the latter could be <body><h1>The News</h1><section><h2>Celebrity News</h2><article><h3>A Celebrity Did Something</h3>[...]

[06:53:00.0000] <remysharp>
Aye, sure.  that makes sense, and that's what I understood.

[06:53:01.0000] <remysharp>
I've just come across quite a few examples that used the article for smaller chunks, and I wasn't completely clear why (when they may have been related to later chunks of content).

[06:55:00.0000] <jgraham>
remysharp: The wwonderful thing about sematics is that there are so many of them to choose from

[07:01:00.0000] <remysharp>
jgraham: agreed - thanks again for the advice.

[13:19:00.0000] <annevk5>
http://www.mnot.net/blog/2009/04/14/rev_canonical_bad is a slightly better explanation of the rev=canonical problem

[13:21:00.0000] <annevk5>
I think I agree with rubys that it would be good to indicate in HTML5 which attributes and elements are obsolete (they're not deprecated). But I also thought that it was planned already. Anybody who knows?

[13:43:00.0000] <Hixie>
annevk5: anything that isn't in html5 is obsolete

[13:58:00.0000] <Philip`>
Hixie: That seems to unhelpfully ignore the distinction between things that used to have some meaning in HTML and are not in HTML5, and things that have never had any meaning at all in HTML

[13:58:01.0000] <Philip`>
(e.g. <link rev> vs <link wlkthglkrh>)

[13:58:02.0000] <Hixie>
not sure why that distinction is needed

[13:59:00.0000] <Philip`>
People will see <link rev> being used somewhere, then look in the latest HTML spec to see what it is, and encounter a complete absence of information, which is less helpful to them than a note that it was defined in some other spec and has been omitted for whatever reasons

[14:03:00.0000] <Hixie>
*shrug*

[14:03:01.0000] <Hixie>
if someone wants to make a list of features that should have such notes, file a bug and i'll add it somewhere

[14:03:02.0000] <Philip`>
(Even people who fully understand the HTML5 process won't be able to deduce that <link rev> was omitted from HTML5 when they fail to find any references to it, since it might just be that they're searching in the wrong section)

[14:04:00.0000] <Philip`>
(or they might be using Opera which seems to have bugs in its search feature that make it miss things in the spec)

[14:10:00.0000] <tantek>
Hixie, HTML 4.0 has the notion/label of "Obsolete" for features that have been outright dropped from the previous version 3.2 (in contrast to "Deprecated", meaning, still supported but discouraged).

[14:10:01.0000] <tantek>
http://www.w3.org/TR/html4/appendix/changes.html#h-A.3

[14:11:00.0000] <Hixie>
like I said, I don't mind listing some obsolete features if people want. There's already some level of that in the "downplayed errors" list

[14:12:00.0000] <tantek>
It may be reasonable for HTML4 to have a similar "changes from HTML4.01" section with a list of such "Obsolete" elements and attributes, optionally with a short description or link to why it was dropped etc.

[14:12:01.0000] <tantek>
er, reasonable for HTML5 to have a similar ... section

[14:13:00.0000] <tantek>
Hixie, as we're tangentially on the rev subject, do you have a canonical post somewhere with the summary of the poor usability in practice of the "rev" attribute?

[14:13:01.0000] <tantek>
The best I've written up so far is: http://microformats.org/wiki/rev#Should_rev_even_be_used

[14:14:00.0000] <tantek>
which cites your http://code.google.com/webstats/2005-12/linkrels.html

[14:14:01.0000] <tantek>
anything newer/better?

[14:15:00.0000] <Philip`>
http://philip.html5.org/data/link-rel-rev.txt has some newer data

[14:15:01.0000] <Hixie>
i don't think i have written anything up other than mails to the whatwg list

[14:17:00.0000] <tantek>
I particularly like <link rev="1.02" ...>

[14:17:01.0000] <tantek>
(from Philip`s URL)

[14:18:00.0000] <tantek>
at least vote-links show up in Philip`s data ("vote-for", "vote-against", "vote-abstain")

[14:19:00.0000] <sayrer>
      2 <link rev="argonet" ...>

[14:19:01.0000] <sayrer>
weird

[14:19:02.0000] <sayrer>
      1 <link rev="D. Bailey Management" ...>

[14:21:00.0000] <Hixie>
the long tail of markup oddities is very long on the web

[14:21:01.0000] <Hixie>
very

[14:21:02.0000] <Hixie>
very

[14:21:03.0000] <Hixie>
long

[14:25:00.0000] <tantek>
Thanks Philip`. I've added that URL to http://microformats.org/wiki/rel-faq#Should_rev_even_be_used

[14:39:00.0000] <takkaria>
argonet?  heh, I wonder how that ende up there

[14:42:00.0000] <Philip`>
http://www.comune.barzago.lc.it/ - <link rev="argonet" href="http://www.argonet.it/" title="Sviluppato da Argonet: comunicazione, web design, consulenza e sviluppo.">

[14:43:00.0000] <Philip`>
http://www.comune.tremenico.lc.it/ too

[15:58:00.0000] <Hixie>
/me finds himself writing long winding sentences and wonders if maybe half way through he should plant a signpost with a map and "you are here"

[16:37:00.0000] <MikeSmith>
about http://krijnhoetmer.nl/irc-logs/whatwg/20090414#l-407 - validator.nu has some special handling/reporting of a few obsolete elements

[16:37:01.0000] <MikeSmith>
e.g., http://validator.nu/?=&doc=http://dev.w3.org/html5/tests/validation/full/invalid/obsolete/center.html

[16:39:00.0000] <MikeSmith>
it reports them as "Error: The center element is obsolete."

[16:40:00.0000] <MikeSmith>
whereas it would otherwise report them as "Error: Element center not allowed as child of element body in this context."

[16:47:00.0000] <sayrer>
the center element is obsolete?

[16:47:01.0000] <MikeSmith>
it is in the HTML5 draft, yeah

[16:47:02.0000] <sayrer>
obsolete, works great, news at 11

[16:47:03.0000] <Hixie>
it doesn't work great

[16:47:04.0000] <Hixie>
it's like the <font> element

[16:48:00.0000] <Hixie>
leads to terribly inaccessible sites

[16:48:01.0000] <Hixie>
and is a maintenance nightmare

[16:48:02.0000] <Hixie>
even HTML4 acknowledges that

[16:48:03.0000] <sayrer>
 maintenance nightmare doesn't seem like a good argument

[16:48:04.0000] <Hixie>
to you maybe :-)

[16:48:05.0000] <sayrer>
since the alternatives are still available

[16:48:06.0000] <MikeSmith>
sayrer: it's not valid in XHTML strict either, right?

[16:48:07.0000] <sayrer>
XHTML is not relevant to me

[16:49:00.0000] <sayrer>
(or anyone else on the Web, but I digress)

[16:49:01.0000] <sayrer>
Hixie, how does it lead to inaccessible sites?

[16:50:00.0000] <MikeSmith>
<center> is not in HTML 4.01 strict either

[16:51:00.0000] <Hixie>
sayrer: it scares me that you are editing an html specification and are asking that question

[16:51:01.0000] <sayrer>
well, that's not really an answer

[16:52:00.0000] <Hixie>
i really don't have the time to teach you elementary web design, sorry

[16:52:01.0000] <Hixie>
i'm sure you can find a plethora of tutorials on the subject on the web though

[16:53:00.0000] <sayrer>
I have seen many examples of the font element being used in an inaccessible fashion

[16:53:01.0000] <MikeSmith>
anyway, fwiw, the way the obsolete-element handling is implemented in v.nu is that the obsolete elements are part of the v.nu/whattf HTML5 RelaxNG schema -- treated as valid (that's the only way to get the schema to recognize them -- but an additional assertions-checking step further along in the processing pipeline then catches and reports them

[16:53:02.0000] <sayrer>
but why is it different than an equivalent style attribute?

[16:54:00.0000] <Hixie>
it's not. style="" attributes are just as bad and should be limited to things like prototyping and temporary workarounds.

[16:55:00.0000] <sayrer>
Hixie, ok, is style="" obsolete in HTML5 as well?

[16:55:01.0000] <Hixie>
it was, for a long time. there are some edge cases for which it is useful (primarily prototyping), which is why it is currently allowed.

[16:56:00.0000] <sayrer>
Hixie, <font> works better for fragments

[16:56:01.0000] <sayrer>
sad but true, perhaps

[16:57:00.0000] <Hixie>
i really am not interested in arguing this case right now; i need to finish the <datagrid> redesign.

[16:57:01.0000] <Hixie>
there are ample essays on the subject on the web that can enlighten you on the subject though.

[16:57:02.0000] <sayrer>
well, no one invited you to argue it. carry on!


2009-04-15
[17:11:00.0000] <MikeSmith>
hsivonen: I wonder if it would make sense to switch the order of processing in v.nu such that the assertions checking is done prior to the schema checking in the processing pipeline

[17:12:00.0000] <webben>
sayrer: "how does it lead to inaccessible sites?" it doesn't _necessarily_.

[17:12:01.0000] <webben>
sayrer: Presentational markup can produce inaccessible sites indirectly if style rather than semantic markup is depended upon for communication.

[17:12:02.0000] <sayrer>
webben, I definitely agree there are bad things

[17:13:00.0000] <sayrer>
webben, I don't think "presentational markup" is a meaningful term however

[17:13:01.0000] <webben>
sayrer: You can (of course) produce exactly the same accessibility problem using generic elements (hello div and span) plus CSS.

[17:14:00.0000] <webben>
sayrer: do you think "markup conveying formatting information rather than explicitly implying the type of content and its relationship to other content" is meaningful?

[17:14:01.0000] <sayrer>
webben, no

[17:14:02.0000] <sayrer>
sometimes, yes. but not as a general principle

[17:14:03.0000] <webben>
when yes and when no?

[17:15:00.0000] <webben>
or, why yes when yes, and why no when no

[17:15:01.0000] <sayrer>
webben, presentational information can lend context to words in the same way that intonation can when words are spoken

[17:16:00.0000] <webben>
there's a difference between humans inferring information from context and machines inferring informationm, that they can then present to humans in a regularized fashion, from explicit instruction.

[17:16:01.0000] <sayrer>
accessibility is about humans, I though

[17:16:02.0000] <sayrer>
thought

[17:17:00.0000] <webben>
yes. and being about to regularize information into a tailored format is about making information accessible to humans.

[17:17:01.0000] <webben>
*able to

[17:17:02.0000] <sayrer>
I think the goal is a good one

[17:18:00.0000] <sayrer>
I'm not sure banning the <font> element is going to help much

[17:18:01.0000] <sayrer>
I could see the value of doing that if one could say "rest assured, you are creating an accessible site" in its absence

[17:19:00.0000] <webben>
Why would you expect it to be a silver bullet?

[17:19:01.0000] <sayrer>
that is not what I said

[17:19:02.0000] <sayrer>
or what I meant, rather

[17:19:03.0000] <webben>
Seems implied in what you said. Can you rephrase perhaps?

[17:20:00.0000] <sayrer>
banning <font> does not prevent the exact problems we are discussing

[17:20:01.0000] <sayrer>
since they are all still there with different markup

[17:21:00.0000] <webben>
the substitution of ambiguous formatting information for (more) unambiguous information about types and relationships?

[17:21:01.0000] <sayrer>
webben, my view is that these elements are syntactic sugar for CSS properties

[17:21:02.0000] <sayrer>
they are no more evil than the CSS, and it takes some judgement to figure out whether they are ok

[17:22:00.0000] <webben>
I think it's confused to say they are syntactic sugar for CSS properties.

[17:22:01.0000] <sayrer>
why?

[17:23:00.0000] <webben>
Because what's "objectionable" about font and style is that they bake formatting information into the structured content layer.

[17:23:01.0000] <sayrer>
that is not an accessibility objection

[17:24:00.0000] <sayrer>
I definitely agree that there are maintainability concerns, and accessibility concerns

[17:26:00.0000] <webben>
sayrer: If your point is it's possible to bake formatting information into the structured content layer, without making it impossible for a machine to separate the two, that's strictly true.

[17:26:01.0000] <sayrer>
webben, concrete examples are better

[17:26:02.0000] <webben>
Concrete examples of what, sorry?

[17:27:00.0000] <sayrer>
webben, can you create something that is strictly worse, from an accessibility perspective, that is only enabled by <font> or <center>?

[17:32:00.0000] <webben>
sayrer: "enabled", probably not. "encouraged", probably yes. encouragement isn't enablement though.

[17:32:01.0000] <sayrer>
webben, ok, I believe that might be true. what is an example?

[17:33:00.0000] <webben>
an example of encouragement?

[17:33:01.0000] <sayrer>
yeah

[17:33:02.0000] <sayrer>
I am particularly interested in cases that aren't of the form "you aren't smart enough to understand CSS"

[17:33:03.0000] <webben>
Not entirely sure what that means.

[17:34:00.0000] <webben>
I think the context is WYSIWYG and the idea that by changing the font you've communicated that something is e.g. a heading, or a field label, or some code.

[17:34:01.0000] <sayrer>
webben, it seems to me that lots of things we consider to be "good markup" are getting mixed up. Good maintainability, good accessibility, etc.

[17:35:00.0000] <webben>
Mmm. "mixed up" or ... happen to be produced by the same practices.

[17:35:01.0000] <sayrer>
and a lot of it is motivated by a sophomoric division of "presentation" and "data"

[17:35:02.0000] <webben>
"sophomoric" seems like a way to criticise something without having to actually state your criticism.

[17:36:00.0000] <sayrer>
that is not the definition! :)

[17:36:01.0000] <sayrer>
but anyway, "the idea that by changing the font you've communicated that something is e.g. a heading, or a field label, or some code"

[17:36:02.0000] <sayrer>
sure

[17:36:03.0000] <sayrer>
but do we get an element for each visual subtlety out there?

[17:37:00.0000] <webben>
Not in HTML.

[17:37:01.0000] <webben>
CSS came along before that happened! ;)

[17:37:02.0000] <sayrer>
ok, we're in the same place I think

[17:37:03.0000] <sayrer>
different ideas of what's right

[18:00:00.0000] <Hixie>
/me ponders whether to make it possible to insert rows into a datagrid dynamically

[18:00:01.0000] <Hixie>
and how this should affect the IDs of previously-inserted elements

[18:08:00.0000] <MikeSmith>
Hixie: it would be great if some smart student developer could take on implementing datagrid in Mozilla or Webkit as a GSOC project

[18:09:00.0000] <MikeSmith>
Hixie: how long do you reckon it will take you to finish the datgrid redesign?

[18:11:00.0000] <roc_>
/me wonders if datagrid is a good idea

[18:12:00.0000] <roc>
I'll wait and see :-)

[18:12:01.0000] <Hixie>
MikeSmith: a few days.

[18:12:02.0000] <Hixie>
MikeSmith: i'm mostly done, but writing it up takes a while.

[18:13:00.0000] <Hixie>
ironically, my proof of concept implementation was done in a few hours, speccing it is taking much longer.

[18:14:00.0000] <MikeSmith>
Hixie: that seems like the way such things usually go

[18:14:01.0000] <MikeSmith>
hmm, Webkit not participating in GSOC this year

[18:15:00.0000] <MikeSmith>
http://socghop.appspot.com/program/accepted_orgs/google/gsoc2009

[18:15:01.0000] <MikeSmith>
unfortunately

[18:21:00.0000] <olliej>
MikeSmith: it was a lot of work for relatively little gain last year

[18:24:00.0000] <MikeSmith>
olliej: I see. I had been hoping the XBL2 stuff would have gotten further along

[18:25:00.0000] <olliej>
MikeSmith: the xbl guy was basically doomed -- there's no way a GSoC student has a real hope of getting xbl done -- you need to have an astonishingly good knowledge of the way everything needs to work

[18:26:00.0000] <MikeSmith>
olliej: yeah, I can imagine

[18:26:01.0000] <MikeSmith>
Keishi Hattori's Web Inspector stuff mostly got done, didn't it?

[18:26:02.0000] <olliej>
MikeSmith: yeah, and thingy

[18:27:00.0000] <Niictar24>
Hrm, does a nightly build like minefield automatically update to the latest build?

[18:27:01.0000] <olliej>
michaelangelo got some good work done

[18:27:02.0000] <olliej>
but the majority of devs did little to no work

[18:27:03.0000] <Niictar24>
Or do I need to download a new one if I want to be super up to date?

[18:27:04.0000] <olliej>
and consumed many hours from actual developers

[18:27:05.0000] <olliej>
Niictar24: it updates automagically

[18:27:06.0000] <olliej>
Niictar24: on mac

[18:28:00.0000] <olliej>
Niictar24: but #webkit would be a more appropriate place to ask

[18:28:01.0000] <MikeSmith>
olliej: Michaelangelo did WF2 stuff?

[18:28:02.0000] <olliej>
MikeSmith: i think he ended up doing a mix of inspector and other stuff

[18:28:03.0000] <MikeSmith>
OK

[18:28:04.0000] <olliej>
can't remember the specifics, just remember liking his patches :D

[18:29:00.0000] <MikeSmith>
WEb Inspector is quite a piece of work - definitely has raised the bar as far as expectations for developer tools, and UI for them

[18:29:01.0000] <MikeSmith>
olliej: what thingy?

[18:29:02.0000] <Niictar24>
Webkit? Know I know about that channel, but are they going to tell me to go somewhere else when I ask about a mozilla product? :P

[18:30:00.0000] <olliej>
Niictar24: oh i completely miread your quetion sorry

[18:30:01.0000] <Rik`>
MikeSmith: I remember he added the autofocus attribute

[18:30:02.0000] <olliej>
Niictar24: i read "nightly build" as "webkit nightly"

[18:30:03.0000] <MikeSmith>
Rik`: cool

[18:32:00.0000] <MikeSmith>
Niictar24: there's also a Help > Check for updates

[18:32:01.0000] <MikeSmith>
in Minefield

[18:33:00.0000] <MikeSmith>
and in the Minefield preferences, you can set what the update behavior should be

[18:34:00.0000] <MikeSmith>
you can have it automatically download and install updates in the background

[18:34:01.0000] <Niictar24>
Thank ye

[18:34:02.0000] <MikeSmith>
or can have it prompt you each time if finds a new update

[18:34:03.0000] <MikeSmith>
I forget what the default is

[19:20:00.0000] <annevk5>
hmm ok, I suppose i'll file a bug at some point on obsolete feature

[19:20:01.0000] <annevk5>
s

[19:20:02.0000] <annevk5>
simply taking HTML4 as reference I think, not want to make it too complicated

[00:21:00.0000] <hsivonen>
MikeSmith: it's doable to change the RNG/assertion order. Is there a particular pair of messages that would be better that way?

[00:24:00.0000] <MikeSmith>
hsivonen: obsoleted elements

[00:24:01.0000] <MikeSmith>
see also my comment at http://bugzilla.validator.nu/show_bug.cgi?id=480

[00:25:00.0000] <MikeSmith>
hsivonen: was the design choice to include the obsoleted element in the schema because of the order of the error messages, or for other reasons?

[00:25:01.0000] <hsivonen>
MikeSmith: so switch order AND remove legacy.rnc?

[00:26:00.0000] <MikeSmith>
hsivonen: if you remove legacy.rnc, I think it might not be important to change the order at all

[00:26:01.0000] <hsivonen>
MikeSmith: IIRC, the design choice was to have a single but more sensible error message

[00:26:02.0000] <MikeSmith>
OK

[00:26:03.0000] <MikeSmith>
that I can understand

[00:26:04.0000] <MikeSmith>
having a single message

[00:27:00.0000] <MikeSmith>
and having it not be the near-useless message "Error: Element foo not allowed as child of element bar in this context" that jing gives

[00:27:01.0000] <hsivonen>
no, now I remember the real reason

[00:27:02.0000] <hsivonen>
I wanted the subtree not be suppressed

[00:27:03.0000] <MikeSmith>
ah

[00:27:04.0000] <MikeSmith>
yeah

[00:27:05.0000] <MikeSmith>
hmm

[00:27:06.0000] <MikeSmith>
that's sort of critical

[00:27:07.0000] <hsivonen>
but in the general case, not making Jing suppress the subtree was badness

[00:27:08.0000] <MikeSmith>
really?

[00:27:09.0000] <MikeSmith>
why?

[00:28:00.0000] <hsivonen>
I think I had a good reason to change that, yes

[00:28:01.0000] <hsivonen>
but I've forgotten the details

[00:28:02.0000] <hsivonen>
that is, the vanilla Jing didn't suppress errors from bogus subtrees

[00:28:03.0000] <MikeSmith>
oh

[00:28:04.0000] <hsivonen>
hmm. maybe it was just that Jing's pattern recovery sucked really badly

[00:29:00.0000] <hsivonen>
so pretty much any element became an error

[00:29:01.0000] <hsivonen>
or something like that

[00:30:00.0000] <MikeSmith>
hsivonen: so maybe it would be useful to experiment with tentatively removing legacy.rnc and see?

[00:33:00.0000] <hsivonen>
MikeSmith: I think it would be worthwhile to see if there have been good patches to Jing trunk before experimenting

[00:33:01.0000] <MikeSmith>
OK

[00:37:00.0000] <MikeSmith>
hsivonen: btw, as I mentioned before, NicolasRaoul is here at W3C Japan for a while, half days (he studies Japanese at a language school in mornings, is here at Keio U. in the afternoons -- and he has some time and interest in helping with patches and potential enhancement to v.nu

[00:37:01.0000] <MikeSmith>
I was talking with him about Jing this morning

[00:38:00.0000] <MikeSmith>
and I guess you saw the patch he Aelfred2 patch he posted to the whatwg implementors list

[00:38:01.0000] <hsivonen>
MikeSmith: I'll look at his patch right soon, once I've unblocked another patch over at b.m.o

[00:38:02.0000] <hsivonen>
s/right /

[00:39:00.0000] <MikeSmith>
hsivonen: OK. btw, you still trying to solve that mystery crash?

[00:41:00.0000] <hsivonen>
MikeSmith: no, I resolved it on Friday

[00:41:01.0000] <hsivonen>
(yay!)

[00:41:02.0000] <MikeSmith>
ah great

[00:41:03.0000] <MikeSmith>
congrats on that

[00:41:04.0000] <MikeSmith>
that must have been pretty frustrating

[00:41:05.0000] <MikeSmith>
hsivonen: speaking of patches, please put this one in your queue:

[00:42:00.0000] <MikeSmith>
http://bugzilla.validator.nu/attachment.cgi?id=81

[00:42:01.0000] <hsivonen>
yes. OS X gave me totally bogus stacks, but I failed to inspect the stack for bogosity until vlad pointed out to me that it was bogus

[00:42:02.0000] <hsivonen>
with the right stack, it was less mysterious

[00:42:03.0000] <MikeSmith>
hsivonen: ah, I saw your tweet about that

[00:43:00.0000] <hsivonen>
interestingly, the OS X crash reporter and Mac Valgrind gave consistent bogus stacks. I guess they use the same stack grabbing mechanism.

[00:43:01.0000] <MikeSmith>
yeah, would see so

[00:44:00.0000] <MikeSmith>
surprised that Valgrind wouldn't use its own thing and not trust the built-in

[00:44:01.0000] <MikeSmith>
instead

[00:44:02.0000] <MikeSmith>
that sounds like a major OS bug

[00:44:03.0000] <MikeSmith>
but not a totally uncommon kind of one

[00:45:00.0000] <MikeSmith>
I remember seeing problems kind of like that when I was involved with customer work for customers running AIX, HP-UX, Tru64, etc.

[00:47:00.0000] <MikeSmith>
I learned a lot about OS bugs and about how untrustworthy some so-called enterprise/carrier-grade OS/platforms can be

[01:06:00.0000] <zcorpan>
hsivonen: your site is now harder to read :(

[01:07:00.0000] <hsivonen>
zcorpan: which browser/OS?

[01:07:01.0000] <hsivonen>
zcorpan: due to colors or fonts or something else?

[01:07:02.0000] <zcorpan>
hsivonen: opera/windows

[01:07:03.0000] <zcorpan>
hsivonen: font for the body text

[01:08:00.0000] <hsivonen>
zcorpan: it looks great on Mac and on Linux with no hinting or little hinting in FreeType

[01:08:01.0000] <hsivonen>
zcorpan: ClearType sucks :-(

[01:08:02.0000] <hsivonen>
zcorpan: I don't know what I should do. 

[01:08:03.0000] <hsivonen>
zcorpan: Opera and Firefox need to ship font renderer that don't suck on Windows the way Safari does, I guess

[01:08:04.0000] <zcorpan>
/me tries disabling cleartype

[01:10:00.0000] <zcorpan>
well that made it even worse :/

[01:10:01.0000] <hsivonen>
the heading fonts is the one that *really* sucks with any hinting

[01:11:00.0000] <hsivonen>
(either Windows hinting or FreeType hinting)

[01:11:01.0000] <hsivonen>
zcorpan: do I have any other course of actions than removing the fonts from Mac and Linux users, too?

[01:12:00.0000] <hsivonen>
which reminds me that Safari makes it really hard to test the issue sayrer reported if the affected font happens to reside on one's hard drive

[01:13:00.0000] <hsivonen>
secondary hard drive even!

[01:13:01.0000] <hsivonen>
it as though Safari had total information awareness of all fonts around all my hard drives

[01:14:00.0000] <zcorpan>
hsivonen: dunno. maybe try some other font that is easier to read with cleartype? maybe one that isn't italics

[01:15:00.0000] <hsivonen>
zcorpan: italics???

[01:15:01.0000] <hsivonen>
http://api.browsershots.org/png/original/b6/b6e53e1d3771b5ebc8bd9a738257b2aa.png looks interesting

[01:15:02.0000] <zcorpan>
hsivonen: everything is in italics for me. but might be a bug in opera

[01:17:00.0000] <MikeSmith>
hsivonen: it might be useful to have an "Error messages" component for v.nu bugzilla

[01:18:00.0000] <MikeSmith>
or "Error reporting"

[01:19:00.0000] <zcorpan>
hsivonen: i see that it's not italics in safari and firefox

[01:20:00.0000] <zcorpan>
hsivonen: it looks good there

[01:26:00.0000] <zcorpan>
hsivonen: http://img19.imageshack.us/img19/8633/hsivonenfontopera.png

[01:26:01.0000] <hsivonen>
zcorpan: should I file an Opera bug? I'm seeing it on my virtual machine.

[01:27:00.0000] <Philip`>
hsivonen: You should add an EOT file for IE

[01:27:01.0000] <hsivonen>
MikeSmith: why not file the bugs under the components that are responible for errors

[01:27:02.0000] <hsivonen>
Philip`: nope :-)

[01:27:03.0000] <zcorpan>
hsivonen: yes please

[01:28:00.0000] <hsivonen>
zcorpan: ok

[01:28:01.0000] <hsivonen>
Philip`: I wanted to have an IE-only feature for a draft CSS feature, but it sucked too much in practice to use

[01:31:00.0000] <Philip`>
hsivonen: If you don't have an EOT file, then the IE developers will be discouraged from adding TTF support because it will make your site start using weird ugly fonts :-)

[01:34:00.0000] <MikeSmith>
hsivonen: for some cases, I'm not sure users will have any idea what component is responsible

[01:34:01.0000] <MikeSmith>
for some cases I'm not sure I know myself

[01:34:02.0000] <MikeSmith>
hsivonen: I meant mostly of reporting of bugs from general public

[01:34:03.0000] <Philip`>
/me supposes he can just disable styles entirely when reading hsivonen's site

[01:35:00.0000] <MikeSmith>
e.g., if somebody sees a typo in an error message and wants to report it

[01:35:01.0000] <MikeSmith>
or a formatting problem (e.g., in the spe-scraped output)

[01:35:02.0000] <MikeSmith>
hsivonen: btw, v.nu bugzilla just gave me a message I've not seen before:

[01:36:00.0000] <MikeSmith>
undef error - Insecure dependency in exec while running with -T switch at /usr/share/perl5/Mail/Mailer/sendmail.pm line 22.

[01:37:00.0000] <zcorpan>
MikeSmith: i've got that a few times iirc

[01:40:00.0000] <Hixie>
set your path to ''

[01:40:01.0000] <Hixie>
local $ENV{PATH} = '';

[01:41:00.0000] <Hixie>
before calling exec(), system(), ``, or qx()

[01:41:01.0000] <Hixie>
or open('|')

[01:42:00.0000] <Hixie>
(and use absolute paths)

[01:53:00.0000] <hsivonen>
zcorpan: DSK-251114

[01:53:01.0000] <hsivonen>
MikeSmith: the perl thing is a know bug that I don't know a fix to

[01:54:00.0000] <hsivonen>
Hixie: I'll try that. thanks!

[01:55:00.0000] <Hixie>
thank me if it works :-)

[01:55:01.0000] <hsivonen>
Philip`: does the serif font suck or do you feel like disabling the sans serif font?

[01:56:00.0000] <hsivonen>
/me wishes FreeType and ClearType stopped this hinting business. it's so last century

[01:59:00.0000] <jgraham>
/me doesn't like hsivonen's choice of background colour :)

[02:00:00.0000] <hsivonen>
maybe I should try 42 shades and get some data :-)

[02:01:00.0000] <zcorpan>
hsivonen: thanks

[02:04:00.0000] <Philip`>
hsivonen: The serif font is the only one that doesn't get rendered really badly (in Firefox on Linux)

[02:04:01.0000] <hsivonen>
Philip`: what's your FreeType hinting setting?

[02:04:02.0000] <Philip`>
(It doesn't seem like a hinting issue - the shapes are just wrong)

[02:04:03.0000] <Philip`>
hsivonen: Don't know

[02:04:04.0000] <hsivonen>
Philip`: do you have a suggestion for a Free as in Freedom sans-serif font that isn't trite and doesn't suck

[02:05:00.0000] <hsivonen>
Philip`: MgOpen Cosmetica looks good with full AA and no hinting and looks terrible with hinting

[02:06:00.0000] <hsivonen>
I expect many more Mac users to use fonts that utterly suck on Windows and with FreeType hinting

[02:06:01.0000] <Philip`>
hsivonen: You could use the one called "sans-serif" :-)

[02:07:00.0000] <hsivonen>
I wanted to use Linux Biolinum, but I didn't find any public test version of it

[02:08:00.0000] <hsivonen>
Cosmetica even ships with Ubuntu for Greek use cases

[02:08:01.0000] <hsivonen>
Philip`: I claim the wrongness of shapes is hinting

[02:09:00.0000] <hsivonen>
Philip`: because the issue goes away if I tweak FreeType hinting settings

[02:09:01.0000] <Philip`>
hsivonen: How do you tweak the hinting settings?

[02:09:02.0000] <Philip`>
/me has to go away and will look later

[02:10:00.0000] <hsivonen>
Philip`: under Ubuntu, System->Preferences->Appearance->Fonts

[02:14:00.0000] <hsivonen>
w00t. the latest libertine package comes with biolinum!!!

[02:18:00.0000] <hsivonen>
perhaps I should experiment with the OTF variants, too. in order to get a different glyph rasterizer

[02:26:00.0000] <hsivonen>
I love this comment in an XML of all places: "many versions of nwalsh docbook stylesheets have bogus URLs; so this can't be an error..."

[02:26:01.0000] <hsivonen>
I guess Draconian software needs to work with existing content :-)

[02:44:00.0000] <MikeSmith>
hsivonen: btw, I noticed that the spec snapshot that v.nu is using is gone pretty stale -- it's from November

[02:44:01.0000] <MikeSmith>
though i guess there have not been so many content-model/attribute changes since then

[02:44:02.0000] <MikeSmith>
there is addition of keygen, though, for one

[02:44:03.0000] <MikeSmith>
so message links won't be generated for that until spec snapshot is updated

[02:46:00.0000] <MikeSmith>
and spec snapshot update will break the build unless Html5SpecBuilder.java is patched

[02:46:01.0000] <hsivonen>
MikeSmith: yeah, that's what tends to happen :-(

[02:46:02.0000] <MikeSmith>
yeah, I learned a bit from looking at that code

[02:47:00.0000] <MikeSmith>
I'm wondering why you just didn't use your HTML parser instead

[02:48:00.0000] <MikeSmith>
to populate urisByElement etc.

[02:48:01.0000] <MikeSmith>
hmm, you did I gues

[02:48:02.0000] <zcorpan>
MikeSmith: you filed the same bug thrice

[02:48:03.0000] <hsivonen>
MikeSmith: http://bugzilla.validator.nu/attachment.cgi?id=81 looks good, but I didn't test it

[02:48:04.0000] <MikeSmith>
zcorpan: oops, sorry

[02:49:00.0000] <hsivonen>
MikeSmith:  "Malformed spec: no element currentName.", locator);

[02:49:01.0000] <hsivonen>
MikeSmith: that looks like my bug

[02:49:02.0000] <hsivonen>
currentName should be a concatenated variable, not part of the literal

[02:50:00.0000] <MikeSmith>
hsivonen: I have tested it as much as I could figure to

[02:51:00.0000] <MikeSmith>
I have some simple per-element tests at http://dev.w3.org/html5/tests/validation/full/invalid/unknown-attribute/

[02:52:00.0000] <MikeSmith>
putting an unknown attribute on each element was the simplest single-invalidity test case I could think of that would generated the error messages with those links

[02:53:00.0000] <zcorpan>
MikeSmith: what about <embed>?

[02:53:01.0000] <MikeSmith>
zcorpan: embed allows any unknown attribute

[02:54:00.0000] <MikeSmith>
I have another test for it

[02:54:01.0000] <MikeSmith>
some other tests

[02:54:02.0000] <MikeSmith>
http://dev.w3.org/html5/tests/validation/full/invalid/bad-value/embed-height.html

[02:55:00.0000] <zcorpan>
MikeSmith: i think 0 is allowed now

[02:55:01.0000] <MikeSmith>
zcorpan: only on <img>

[02:55:02.0000] <MikeSmith>
I think

[02:55:03.0000] <zcorpan>
oh

[02:55:04.0000] <zcorpan>
ok

[02:56:00.0000] <zcorpan>
consistency++

[03:10:00.0000] <MikeSmith>
hsivonen: OK, fixed the problem with the variable name literal in the message string, and re-tested

[03:10:01.0000] <MikeSmith>
http://bugzilla.validator.nu/attachment.cgi?id=83

[03:12:00.0000] <MikeSmith>
though I'm not sure how to malform the spec in such a way as to cause that code to be exercise

[03:15:00.0000] <hsivonen>
MikeSmith: looks good for check-in

[03:24:00.0000] <MikeSmith>
hsivonen: cool, thanks

[03:39:00.0000] <MikeSmith>
hsivonen: http://svn8.cvsdude.com/vvc/whattf/validator?view=revision&revision=311

[03:40:00.0000] <MikeSmith>
checked in snapshot of latest spec along with the SpecBuilder change

[04:29:00.0000] <annevk5>
meh, getting IE8 to run is a pain

[04:29:01.0000] <annevk5>
just told me two restarts are required

[04:46:00.0000] <annevk5>
hmm, IE8 dispatches the storage event on the document object?!

[04:52:00.0000] <annevk5>
in Safari I can't seem to get it dispatch at all

[04:52:01.0000] <annevk5>
(and curiously onstorage exists on both document and window in Safari)

[04:56:00.0000] <jgraham>
Note for the curious: getting one's thesis bound remotely is both stressful and insanely expensive

[04:57:00.0000] <annevk5>
I thought the IE team had fixed the issues :/

[04:57:01.0000] <hsivonen>
/me wonders how much "insanely" is

[04:57:02.0000] <hsivonen>
jgraham: would 22.50 EUR per copy incl. VAT be "insanely"?

[04:57:03.0000] <jgraham>
hsivonen: It is going to end up cosing roughly 60GBP/copy

[04:57:04.0000] <hsivonen>
wow

[04:58:00.0000] <hsivonen>
jgraham: did you get leather covers or something?

[04:58:01.0000] <Dashiva>
Gold plating

[04:58:02.0000] <jgraham>
hsivonen: No cloth covers are mandatory though

[04:58:03.0000] <jgraham>
And cost about 30GBP/copy

[04:59:00.0000] <jgraham>
The remaining cost is printing which I can't easilly do and get the documents to the binders

[04:59:01.0000] <jgraham>
and delivery of the documents to the depertment

[04:59:02.0000] <Dashiva>
There's no binding service at the uni?

[04:59:03.0000] <jgraham>
Dashiva: I think they recommend the commercial service

[05:00:00.0000] <jgraham>
It is likely that I have done this in a more expensive way than is strictly necessary

[05:01:00.0000] <annevk5>
IE also throws for trying to set storage keys to invalid XML characters now

[05:02:00.0000] <annevk5>
which notably violates the spec

[05:03:00.0000] <jgraham>
annevk5: Is there a testsuite?

[05:04:00.0000] <jgraham>
Dashiva: It turns out that there is a Uni binding service that is actually just a frontend for the commercial service

[05:04:01.0000] <jgraham>
With the same prices and so on

[05:05:00.0000] <annevk5>
I'm just toying with the Live DOM Viewer

[05:06:00.0000] <Dashiva>
That's... weak

[05:06:01.0000] <annevk5>
writing an update on IE8 and standards

[05:06:02.0000] <annevk5>
I actually thought they had fixed most except for a few things but it's a bit worse than that :/

[05:07:00.0000] <jgraham>
annevk5: Right but I'm wondering if there is a testsuite or not becuase having testsuites for features that are known to be in development is good

[05:07:01.0000] <jgraham>
and we have known that storage is in development for some time now

[05:07:02.0000] <jgraham>
(and obviously it has been released)

[05:07:03.0000] <hsivonen>
/me is again amazed at the epic design failure that .h files are

[05:08:00.0000] <jgraham>
So if there were a WG testsuite we could have pointed them at it would have been good

[05:08:01.0000] <Dashiva>
hsivonen: I'm more amazed by the people who defend them

[05:09:00.0000] <jgraham>
In general it would be good if the HTMLWG created a community around testing features that are known to be in development for various browsers

[05:09:01.0000] <jgraham>
So that the whole web isn't affected by lame QA at one company

[05:10:00.0000] <jgraham>
/me wonders what happened to takkaria's test czar plan

[05:10:01.0000] <annevk5>
sure, everyone has bugs

[05:11:00.0000] <annevk5>
but dispatching an event on a completely different object? come on

[05:11:01.0000] <annevk5>
I mean, these are not edge case bugs

[05:11:02.0000] <jgraham>
Sure :)

[05:11:03.0000] <jgraham>
But a testsuite makes it really obvious that they suck

[05:12:00.0000] <jgraham>
Marketing "we introduced a cool new HTML5 feature" is undermined somewhat if you are known to fail every test for that feature

[05:13:00.0000] <jgraham>
So if you want to dispatch the event on a different object for whatever reason then you have to involve the community in that

[05:13:01.0000] <jgraham>
Or at leeast there is some incentive to

[05:16:00.0000] <Dashiva>
Whoever made that YTSO mashup must love timpani

[05:21:00.0000] <annevk5>
Safari seems buggy too

[05:21:01.0000] <annevk5>
returns undefined rather than null for unknown keys; events don't seem to function

[05:22:00.0000] <annevk5>
"If there is one thing the world needs, it’s more testcases."

[05:23:00.0000] <jgraham>
Returning undefined for unknown keys does make more sense, to be fair

[05:23:01.0000] <jgraham>
(obviously if the spec says otherwise, that's what they should do)

[05:24:00.0000] <annevk5>
IE does undefined too

[05:24:01.0000] <jgraham>
We should change the spec than

[05:24:02.0000] <jgraham>
*then

[05:24:03.0000] <jgraham>
it makes sense since o = {}; o.p === undefined in normal js

[05:24:04.0000] <annevk5>
It's nice that we have to figure that out by reverse engineering them

[05:25:00.0000] <annevk5>
I thought clear unambigious specs were supposed to safe us from this mess?

[05:27:00.0000] <jgraham>
annevk5: I guess they will just make it better rather than solve all problems

[05:28:00.0000] <annevk5>
You know, I wonder what Firefox does here. I'm guessing they do it right.

[05:29:00.0000] <annevk5>
/me will know in few minutes

[05:29:01.0000] <annevk5>
a few, even

[05:33:00.0000] <annevk5>
and they do, someone buy me a beer

[05:34:00.0000] <Dashiva>
Sorry, the GMFMWG hasn't responded yet

[05:35:00.0000] <annevk5>
/me doesn't get the joke; also not after reading http://krijnhoetmer.nl/irc-logs/html-wg/20090326

[05:36:00.0000] <Dashiva>
http://krijnhoetmer.nl/irc-logs/html-wg/20090326#l-104

[05:37:00.0000] <annevk5>
heh

[05:38:00.0000] <annevk5>
Firefox doesn't do events either? Am I missing something?

[05:39:00.0000] <annevk5>
code:

[05:39:01.0000] <annevk5>
window.onstorage = function(e) { w(e) }

[05:39:02.0000] <annevk5>
localStorage.x = "y"

[05:40:00.0000] <Dashiva>
Maybe they have it on document?

[05:40:01.0000] <annevk5>
nope

[05:40:02.0000] <annevk5>
they also do not expose onstorage on document unlike WebKit

[05:56:00.0000] <heycam`>
qq: "center of rotation" or "centre of rotation" in en-US?

[05:56:01.0000] <jgraham>
I believe the former

[05:56:02.0000] <jgraham>
But I don't speak en-US

[05:56:03.0000] <hsivonen>
:-( https://html5lib.googlecode.com/ is 503

[05:57:00.0000] <heycam`>
jgraham, ok thanks

[05:57:01.0000] <heycam`>
i think it's the former too, since the latter is how i would spell it... but wanted to be sure

[05:57:02.0000] <jgraham>
hsivonen: http://html5lib.googlecode.com/svn/trunk/ seems to work

[06:01:00.0000] <remysharp>
qq: is a div still an appropriate element to wrap an arbitrary block of content for styling? Or is that wrong? this particular example doesn't add any semantic value, just required for JS effect

[06:05:00.0000] <beowulf>
remysharp: yes

[06:06:00.0000] <remysharp>
ta

[06:06:01.0000] <remysharp>
I've another question, but it's possible it's not appropriate for this chan - if so, I'll go looking elsewhere - 

[06:07:00.0000] <remysharp>
I'm using xhtml content type serving to get FF2 to render the html5 properly - which works just fine, but -

[06:07:01.0000] <remysharp>
when it encounters an html entity that it doesn't know about, i.e. &bull; it borks

[06:08:00.0000] <remysharp>
Obviously this would be fixed by specifying a DTD, but that's not the way html5 is done

[06:08:01.0000] <beowulf>
that's xml for you, define the entity in the dtd or use xml entites

[06:08:02.0000] <beowulf>
entities

[06:09:00.0000] <annevk5>
or don't use XML

[06:09:01.0000] <remysharp>
annevk5: sure, and I know FF2 has zip all usage, but it looks aweful - and the JS fix for FF2 needs a lot more work

[06:11:00.0000] <annevk5>
I wouldn't know, I'm running Firefox 3.6something

[06:11:01.0000] <jgraham>
remysharp: Seriously ignore FF2

[06:11:02.0000] <remysharp>
I would if I had justified ignoring IE6, but since the site works perfectly in that browser, I can't live with the idea that it's shot all over the place in FF2

[06:11:03.0000] <remysharp>
and the htaccess fix is easy

[06:12:00.0000] <remysharp>
and working - but it's just html entities that was tripping me up

[06:12:01.0000] <annevk5>
doesn't sound like working to me :p

[06:12:02.0000] <beowulf>
would it be better to have an #html5 room for html authors like me to ask q's or is here fine?

[06:12:03.0000] <jgraham>
remysharp: You know that is not a rational argument, right

[06:12:04.0000] <jgraham>
beowulf: Here is fine

[06:12:05.0000] <remysharp>
jgraham: absolutely!

[06:13:00.0000] <remysharp>
just a matter of pride if it doesn't blow my mind to fix it :-)

[06:14:00.0000] <jgraham>
remysharp: OK, well the other solution is to not use entities

[06:14:01.0000] <jgraham>
Ideally you should just use UTF-8 over the wire

[06:14:02.0000] <remysharp>
jgraham: I've just changed it to xml entities and it works perfectly

[06:14:03.0000] <annevk5>
jgraham, that's not an argument against entities

[06:14:04.0000] <remysharp>
oh - as in the raw bullet character?

[06:14:05.0000] <jgraham>
annevk5: What isn't?

[06:15:00.0000] <jgraham>
That they don't work?

[06:15:01.0000] <annevk5>
jgraham, "just use UTF-8" because it's still easier to type an entity than the actual character

[06:15:02.0000] <jgraham>
annevk5: I agree

[06:15:03.0000] <annevk5>
e.g. &hellip;

[06:15:04.0000] <jgraham>
But if you are producing XML you need a real serializer anyway so doing entity->UTF-8 conversion should be trivial

[06:16:00.0000] <annevk5>
in fact, when I want the character I usually construct a data URL and copy and paste

[06:16:01.0000] <jgraham>
annevk5: &hellip is not a great example (it is erasonably easy to type)

[06:16:02.0000] <jgraham>
But $cint; probably is

[06:17:00.0000] <annevk5>
the character the entity represents is not easy to type

[06:17:01.0000] <annevk5>
not for me anyway

[06:24:00.0000] <jgraham>
annevk5: It is … right? If you set a compose key under Linux, that is just Compose  then ..

[06:28:00.0000] <annevk5>
dunno, I'm using my Mac atm because I'm running Windows XP on the other machine to play with IE8

[06:28:01.0000] <annevk5>
also, I don't know of Compose or its shortcuts nor do I know for most other characters

[06:29:00.0000] <annevk5>
(even the euro character...)

[06:29:01.0000] <jgraham>
annevk5: I don't know most entities :)

[06:30:00.0000] <jgraham>
(nor do I know most compose shortcuts)

[06:30:01.0000] <annevk5>
I know a bunch for umlauts and such

[06:30:02.0000] <annevk5>
and for the euro :)

[06:32:00.0000] <remysharp>
if you want entities - there you go, a nice little dashboard widget too: http://leftlogic.com/lounge/articles/entity-lookup/

[06:52:00.0000] <beowulf>
what is wrong with <h1><p>Typeline</p><p>Mainline</p></h1>?

[06:53:00.0000] <annevk5>
you want <header>

[06:54:00.0000] <beowulf>
I do for sure, but when we then go onto describe what header means and how it's used it sounds very much like the above, so why not do that? i'm curious is all

[06:55:00.0000] <Dashiva>
Because those aren't paragraphs?

[06:56:00.0000] <annevk5>
because typeline and mainline can't be distinguished then

[06:56:01.0000] <annevk5>
would be one reason I can think of

[06:56:02.0000] <hsivonen>
Grr. Parallels refuses to virtualize non-Server Mac OS X

[06:57:00.0000] <beowulf>
Dashiva: are they not paragraphs because they are in a <h1> or because of the content?

[06:58:00.0000] <Dashiva>
Both?

[07:19:00.0000] <hsivonen>
whee! now that I worked around the Opera 10 @font-face bug, WebKit shows me all italics!

[07:20:00.0000] <annevk2>
fonts are a nightmare :/

[07:27:00.0000] <hsivonen>
filed https://bugs.webkit.org/show_bug.cgi?id=25207 and https://bugs.webkit.org/show_bug.cgi?id=25208

[07:27:01.0000] <hsivonen>
WebKit has such small bug ids :-)

[07:33:00.0000] <virtuelv>
hsivonen: details on our @font-face bug?

[07:33:01.0000] <Rik|work>
hsivonen: damn, i've filed https://bugs.webkit.org/show_bug.cgi?id=25209 :)

[07:33:02.0000] <Rik|work>
i'll close mine

[07:36:00.0000] <annevk2>
fwiw, I think dhyatt did that on purpose

[07:36:01.0000] <Rik|work>
annevk2: there's maybe a reason behind but the result is really ugly

[07:36:02.0000] <hsivonen>
virtuelv: DSK-251114: font-weight and font-style modifiers ignored on @font-face

[07:37:00.0000] <virtuelv>
hsivonen: thanks

[07:37:01.0000] <hsivonen>
/me wishes these face selection issues within a family get fixed before Opera 10 final and Safari 4 final

[07:37:02.0000] <hsivonen>
otherwise, multiface families will suck badly

[07:38:00.0000] <hsivonen>
/me wonders why OpenType models italics as a separate face but small caps as alternative glyphs

[07:49:00.0000] <takkaria>
jgraham: my test czar plan got taken over by real life

[07:50:00.0000] <jgraham>
takkaria: You should consider abandoning rel life then

[07:50:01.0000] <jgraham>
*real

[07:51:00.0000] <takkaria>
I have, many times, but it always seems to come back like some kind of weed between the cracks of the pavement

[08:00:00.0000] <jgraham>
annevk5: Are you going to mention the undefined vs null thing for stoarge somewhere?

[08:20:00.0000] <hsivonen>
how do I intercept a email that is waiting for sending when Mail.app is being stupid with its threads?

[08:22:00.0000] <hsivonen>
hmm. http://lists.w3.org/Archives/Public/public-webapps/2009AprJun/0188.html went through, so I can kill Mail.app

[08:24:00.0000] <annevk2>
jgraham, other than on my blog? not really :p

[08:33:00.0000] <jgraham>
annevk2: I'm pretty sure that your blog doesn't count

[08:34:00.0000] <annevk2>
watch and learn

[08:34:01.0000] <annevk2>
/me files a bug

[08:39:00.0000] <annevk2>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=6814

[08:39:01.0000] <jgraham>
/me wwonders what he is supposed to be learning

[08:40:00.0000] <jgraham>
In other, marginally related, news, I wonder if there is any point in filing web-compat bugs on the ES5rc

[08:42:00.0000] <annevk2>
seems useful

[08:42:01.0000] <annevk2>
you were supposed to learn that blogs work and then I gave in and filed the silly bug report

[08:42:02.0000] <jgraham>
Well, we shall see

[08:43:00.0000] <annevk2>
maybe also keep track of web-compat issues with the spec on a wiki page

[08:43:01.0000] <annevk2>
i made one for HTTP

[08:43:02.0000] <annevk2>
where they also don't care too much

[08:46:00.0000] <jgraham>
Yeah, it would be nice to do a "Web ECMAScript" spec that documented all the things that actually have to be supported in browsers

[08:47:00.0000] <hsivonen>
aargh. Google still hasn't a proper mechanism for filing bugs on Google Maps

[09:51:00.0000] <annevk2>
hsivonen, I think both your changes to Selectors make sense

[09:52:00.0000] <annevk2>
hsivonen, maybe it helps if we provide explicit change requests

[10:09:00.0000] <zcorpan>
hsivonen: "Downloadable fonts are not supported." - https://developer.mozilla.org/en/Mozilla_Web_Developer_FAQ

[10:11:00.0000] <zcorpan>
hsivonen: also, "According to the Accept header, Mozilla prefers application/xhtml+xml over text/html." - is that still the case?

[10:12:00.0000] <zcorpan>
hsivonen: also "Currently Mozilla does not catch character encoding errors"

[10:14:00.0000] <zcorpan>
hsivonen: "HTML-specific CSS exceptions do not apply. For example, the body element gets no special treatment."

[10:16:00.0000] <zcorpan>
hsivonen: "Simple XLink" - is that still supported?

[10:32:00.0000] <annevk2>
Is Steve arguing ATs trump authors?

[10:34:00.0000] <tantek>
would that just be a special case of user preferences trumping authors? e.g. user stylesheet etc.

[10:35:00.0000] <annevk2>
more like change the spec because AT might not be fixed

[10:35:01.0000] <jcranmer>
/me struggles to expand the acronym

[10:35:02.0000] <jcranmer>
Applachain Trail is funny though :-)

[12:14:00.0000] <zcorpan_>
annevk2: "The menu element is redefined to be useful for actual menus." - please elaborate on "menus" so people don't confuse it with <nav>

[12:14:01.0000] <zcorpan_>
annevk2: e.g. s/actual menus/toolbars and context menus/

[12:19:00.0000] <zcorpan_>
annevk2: "Defined that this in the global object returns a WindowProxy object rather than the Window object." - clarify that it's the javascript "this"

[12:20:00.0000] <zcorpan_>
annevk2: "The way media elements load resources has been clarified." - changed, even

[12:21:00.0000] <zcorpan_>
annevk2: "You are now allowed to specify the meta element with a charset attribute in XML documents if the value of that attribute matches the encoding of the document." - not the full truth

[12:24:00.0000] <zcorpan_>
annevk2: "A "storage mutex" concept has been added to deal with separate pages trying to change the document.cookie and localStorage object at the same time." - could be interpreted as the problem being changing document.cookie and localStorage at the same time

[12:41:00.0000] <zcorpan_>
annevk2: btw, i've made createElement() case-folding in xml in web dom core

[12:42:00.0000] <annevk2>
saw that, wondered whether I should mention it in the email, but decided not to

[12:49:00.0000] <dbaron>
annevk2, your latest public-html message has an mid URI that doesn't resolve for me

[13:33:00.0000] <annevk2>
dbaron, oh, it was meant to point to http://lists.w3.org/Archives/Public/www-style/2009Apr/0273.html

[13:33:01.0000] <annevk2>
dbaron, I just copied it from the email

[14:00:00.0000] <jgraham>
/me discovers dsypathy is a real word

[14:54:00.0000] <annevk2>
"HTML5 looks great but I think you should stick to page layout and leave protocols either to JavaScript or to some other extension mechanism."

[14:54:01.0000] <takkaria>
excellent, if only all comments were along those lines

[15:03:00.0000] <franksalim>
annevk2: where is that quote from?

[15:09:00.0000] <Hixie>
didn't we get rid of all the protocols already?

[15:09:01.0000] <Hixie>
oh that's the registerProtocolHandler() e-mail?

[15:12:00.0000] <takkaria>
http://blog.nihilogic.dk/2009/02/html5-canvas-cheat-sheet.html is nice

[15:33:00.0000] <annevk2>
it's the one about <keygen> and why we should drop it

[15:34:00.0000] <annevk2>
this is the closing line

[15:34:01.0000] <annevk2>
I think it's great

[15:37:00.0000] <mpilgrim_>
annevk2: 

[15:37:01.0000] <mpilgrim_>
annevk2: "representa control for key pair generation" --> "represents control for key pair generation"

[15:37:02.0000] <mpilgrim_>
in http://dev.w3.org/html5/html4-differences/

[15:38:00.0000] <mpilgrim_>
annevk2: "For the XML syntax authors" --> "For the XML syntax, authors"

[15:39:00.0000] <mpilgrim_>
annevk2: "The HTML 5 language has a "custom" HTML syntax" --> "The HTML 5 language has a "custom" SGML syntax" ? not sure what else you would mean here, the way it's phrased now doesn't make sense

[15:40:00.0000] <Hixie>
s/"custom" HTML/dedicated/, probably

[15:42:00.0000] <mpilgrim_>
"space separated list" --> "space-separated list"

[15:43:00.0000] <mpilgrim_>
also, @ping attribute description mentions URIs in one sentence, then URLs later in the same paragraph

[15:43:01.0000] <annevk2>
I'll just make it "in HTML syntax and in XML syntax"

[15:43:02.0000] <mpilgrim_>
but section 2.4 says HTML now has native support for IRIs.

[15:44:00.0000] <mpilgrim_>
"The area element, for consistency," <-- consistency with what?

[15:45:00.0000] <mpilgrim_>
(i assume the a and link elements)

[15:45:01.0000] <mpilgrim_>
"The base element can now have a target  attribute as well, mainly for consistency with the a  element (it is also widely supported)." --> "The base element can now have a target  attribute as well, mainly for consistency with the a  element. (This is already widely supported.)"

[15:46:00.0000] <annevk2>
changed all URI to URL and clarified the IRI paragraph

[15:49:00.0000] <mpilgrim_>
"Using this feature should enhance the user experience as the user can turn it off if he does not like it, for instance. " --> "This is a superior alternative to script-based solutions because browsers and assistive technologies can more easily override it."

[15:49:01.0000] <mpilgrim_>
bah, still not happy with that sentence

[15:49:02.0000] <mpilgrim_>
feel free to ignore some or all of this

[15:49:03.0000] <mpilgrim_>
especially that one

[15:49:04.0000] <annevk2>
i'll ignore that last one then :)

[15:50:00.0000] <mpilgrim_>
"(e.g. one they are not a descendant of)" --> "e.g. these elements can now be placed anywhere on a page, not just as descendants of the form element"

[15:52:00.0000] <mpilgrim_>
"They are equivalent to the attributes not prefixed with form on the form  element and override those." --> "If present, they override the action, enctype, method, novalidate, and target attributes on the form element."

[15:52:01.0000] <annevk2>
I guess I should e.g. to i.e. as well then

[15:54:00.0000] <mpilgrim_>
"because their effect is purely presentational and whose function" --> "because their effect is purely presentational and their function"

[16:02:00.0000] <mpilgrim_>
should mention somewhere that @alt is now optional if-and-only-if (whatever the spec says these days)

[16:06:00.0000] <mpilgrim_>
"This API has the necessary security restrictions in place" seems like a bold and unsubstantiated claim :)

[16:09:00.0000] <mpilgrim_>
rest looks great

[16:12:00.0000] <Philip`>
jgraham: About binding: Can't you just ask somebody in the remote location to do it for you?

[16:12:01.0000] <Philip`>
(like, a proper person, not a representative of a commercial service)

[16:12:02.0000] <Philip`>
annevk5: I thought IE8 final had changed (since beta) to return undefined instead of null, since I pointed it out in response to their submitted tests

[16:20:00.0000] <annevk2>
you said they should return undefined instead?

[16:23:00.0000] <Philip`>
I think I said they should do what the spec says, and I think they said they changed their tests to match that, and I think I might have tested the implementation and found it was fixed, but I might be hallucinating

[16:41:00.0000] <annevk2>
Philip`, I'm pretty sure I did w(localStorage.foobar) and it said undefined but if someone could verify that again that'd be cool

[16:47:00.0000] <Hixie>
why would it not say undefined

[16:54:00.0000] <Philip`>
annevk2: Oh - it looks like they changed localStorage.getItem('foobar'), but not localStorage.foobar

[16:55:00.0000] <Philip`>
(At least getItem('foobar') returns null in current IE8, while .foobar returns undefined)

[16:57:00.0000] <Hixie>
that's what the spec says should happen

[16:58:00.0000] <Philip`>
Really?

[16:59:00.0000] <Hixie>
localStorage.foobar doesn't exist on the object, and ECMA-262 says it should thus return undefined

[16:59:01.0000] <Hixie>
getItem() says to return null for unknown values

[16:59:02.0000] <Dashiva>
But doesn't property access map to get/setItem?

[16:59:03.0000] <Philip`>
That's quite non-obvious from the spec

[16:59:04.0000] <Hixie>
yes, for properties that are present

[16:59:05.0000] <Hixie>
Philip`: read WebIDL


2009-04-16
[17:00:00.0000] <Philip`>
It'd be nice if it linked to the relevant part of WebIDL :-)

[17:00:01.0000] <Hixie>
i'm waiting for gsnedders to give us cross-spec auto-xreffing :-)

[17:00:02.0000] <Philip`>
/me kind of assumed that NameGetter would be called for any name that wasn't already part of the interface

[17:00:03.0000] <Philip`>
gsnedders: It'd be nice if it linked to the relevant part of WebIDL :-)

[17:00:04.0000] <Hixie>
i've already set up the cross-references for this, it just isn't being generated automatically yet

[17:01:00.0000] <Hixie>
Philip`: if you had it defined that way, you couldn't enumerate over the object

[17:02:00.0000] <Philip`>
Hixie: Enumeration could be defined over some finite list of keys, but NameGetter could still be called for all name gets regardless of whether they're in that list

[17:02:01.0000] <Philip`>
but I guess it's not done that way

[17:02:02.0000] <Philip`>
in which case that's fine

[17:03:00.0000] <Philip`>
/me prefers it when he doesn't have to read specs to work out what behaviour is specified, but that might be asking for too much

[17:04:00.0000] <Hixie>
i'm just doign what heycam and the browser vendors tell me to do here

[17:04:01.0000] <Hixie>
if you have any suggestiosn on making the spec clearer, let me know :-)

[17:04:02.0000] <Dashiva>
Is this behavior described in WebIDL?

[17:05:00.0000] <Hixie>
yes

[17:05:01.0000] <Hixie>
web storage invokes all the webidl terms

[17:05:02.0000] <Hixie>
they re underlined in green

[17:06:00.0000] <Philip`>
By green, do you mean blue?

[17:06:01.0000] <Dashiva>
Yes, but I don't see anything about the undefined/null duality there

[17:06:02.0000] <Hixie>
er, blue in the w3c version, yes

[17:06:03.0000] <Philip`>
Is there a non-W3C version?

[17:06:04.0000] <Philip`>
/me has a hard time finding a copy of the spec via Google whenever he wants it

[17:06:05.0000] <Hixie>
i don't think i'm currently generating a whatwg version of the web storage spec, no

[17:07:00.0000] <Philip`>
Oh, neat, Google's 'promote' thing lets me make it the top search result for "web storage"

[17:07:01.0000] <Hixie>
the web storage spec is the 6th hit for [web storage] on google

[17:08:00.0000] <Dashiva>
Hmm

[17:08:01.0000] <Dashiva>
No NameCreator

[17:08:02.0000] <Philip`>
I'm afraid I don't acknowledge the existence of any search results beyond 2 from the top, except in exceptional circumstances

[17:08:03.0000] <Hixie>
heh

[17:08:04.0000] <Hixie>
/me just uses the links on http://☺.damowmow.com/

[17:09:00.0000] <Dashiva>
Okay, so reading WebIDL the "supported named properties" thing supports the null/undefined duality thing

[17:09:01.0000] <Dashiva>
But that also means you can't create new properties with .x = whatnot, since there's no NameCreator

[17:10:00.0000] <Hixie>
er yeah, i should put in a NameCreator

[17:11:00.0000] <Hixie>
file a bug or drop me a mail or something?

[17:11:01.0000] <Hixie>
i'm in the middle of massive edits for datagrid

[17:11:02.0000] <Philip`>
IE8 is buggy because it allows you to create new properties, in violation of the spec!

[17:11:03.0000] <Dashiva>
k

[17:11:04.0000] <Dashiva>
Philip`: Or just prescient!

[17:12:00.0000] <Dashiva>
Hixie: Can I send to whatwg list even though it's w3c spec? :)

[17:12:01.0000] <Hixie>
yes

[17:12:02.0000] <Hixie>
the web storage spec is a whatwg spec (it's in the same source document as html5), i'm just not currently generating a whatwg copy of the output

[17:13:00.0000] <Hixie>
i wonder why i don't have the NameCreator annotations

[17:13:01.0000] <Hixie>
that's odd

[17:15:00.0000] <Dashiva>
Maybe you wrote the IDL before the supported named properties thing was added to WebIDL, and forgot to update

[17:16:00.0000] <Hixie>
no, the rest of the spec carefully talks about the named properties stuff

[17:16:01.0000] <Hixie>
i had a big update day

[17:16:02.0000] <Hixie>
where i went through and redid all the interfaces to match WebIDL

[17:16:03.0000] <Hixie>
i just forgot, i guess

[17:16:04.0000] <Hixie>
i wonder how many more are missing NameCreator

[17:17:00.0000] <Hixie>
i guess it's pretty rare that you can add keys

[18:16:00.0000] <annevk2>
Hixie, that's ugly, they should do the same, imo

[20:13:00.0000] <myakura>
annevk2: html5-diff sec 3.1 says "ruby, rt and rb allow for marking up ruby annotations. " but i think it's not <rb> but <rp>

[00:44:00.0000] <MikeSmith>
are there any significant number of pages that use <basefont>?

[00:44:01.0000] <MikeSmith>
is it worth having a validator report it as obsolete?

[00:45:00.0000] <MikeSmith>
also <isindex>

[00:46:00.0000] <MikeSmith>
hmm, http://dev.opera.com/articles/view/mama-phrase-block-list/#misc

[00:46:01.0000] <MikeSmith>
so basefont appears in more pages than ins or del

[01:04:00.0000] <MikeSmith>
http://dev.opera.com/articles/view/mama-head-structure/#head

[01:05:00.0000] <MikeSmith>
"ISINDEX element was only found in 63 URLs" [out of 3,000,000+]

[01:06:00.0000] <MikeSmith>
so definitely not worth reporting isindex

[01:06:01.0000] <MikeSmith>
but <basefont> is also used in more pages than <acronym>, <s> and <strike> are -- and in around the same order of magnitude as many pages as <tt> is

[01:14:00.0000] <zcorpan>
what is the implementation status of <script async> and <script defer>?

[01:16:00.0000] <jwalden>
http://☺.damowmow.com/ hurts my eyes to read

[01:18:00.0000] <zcorpan>
ugh

[01:18:01.0000] <zcorpan>
Hixie: hire a designer :)

[01:25:00.0000] <jgraham>
Philip`: I couldn't really get anyone to do the binding for me since it has to be hard bound with cloth covers and all. I could maybe have got someone to do the printing for me but it seems like a big ask since there is a large volume and it would have to be taken to the binders and so on

[01:29:00.0000] <jgraham>
Hixie: What is the precedent for having localStorage.getItem("foo") != localStorage.foo when !('foo' in localStorage)

[01:40:00.0000] <zcorpan>
is spellcheck="" implemented in webkit and/or chrome?

[01:54:00.0000] <MikeSmith>
zcorpan: <script async> and <script defer> not implemented in WebKit yet

[01:54:01.0000] <MikeSmith>
https://bugs.webkit.org/show_bug.cgi?id=20710

[01:55:00.0000] <MikeSmith>
but that comment there says, "Firefox finished this. It's in the 3.1 betas now." .. with some caveats

[01:55:01.0000] <MikeSmith>
oh, resolution : "Firefox decided to fire DCL after deferred scripts to faciliate the load pattern mentioned in this bug"

[02:01:00.0000] <zcorpan>
MikeSmith: so firefox supports async and defer?

[02:02:00.0000] <Rik`>
zcorpan: I believe it only supports defer

[02:02:01.0000] <zcorpan>
Rik`: ok

[02:14:00.0000] <hsivonen>
http://groups.google.com/group/mozilla.dev.platform/browse_thread/thread/89eea1cb2d87999/83b917bfaba2b597

[02:31:00.0000] <zcorpan>
wow i didn't know about javascript "foo".sub() and .link() etc

[02:37:00.0000] <hsivonen>
does DanC's URL spec have a permalink yet?

[02:40:00.0000] <zcorpan>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/80

[02:41:00.0000] <hsivonen>
zcorpan: from Netscape 2.0, IIRC

[02:42:00.0000] <hsivonen>
zcorpan: great companions to document.write()

[02:43:00.0000] <hsivonen>
http://homer.w3.org/~connolly/projects/urlp/raw-file/008373680cae/wah5/draft.html is 403 for me

[02:44:00.0000] <zcorpan>
the need for Web EcmaScript increases

[02:47:00.0000] <olliej>
zcorpan: ?

[02:52:00.0000] <zcorpan>
olliej: i doubt the ecmascript committee are interested in specifying "foo".fontcolor("red")

[02:52:01.0000] <olliej>
zcorpan: yes because that would belong (if anywhere) in html5 sec

[02:52:02.0000] <olliej>
spec

[02:53:00.0000] <zcorpan>
maybe

[02:54:00.0000] <hsivonen>
do those methods on strings exist in all script contexts in browsers? do they exist in Rhino?

[02:54:01.0000] <zcorpan>
they exist in the futhark jsshell

[02:55:00.0000] <olliej>
zcorpan: but it also doesn't make sense for them to really belong on string anyway, as hsivonen said they hark back to netscape2

[02:55:01.0000] <olliej>
zcorpan: where i'm not even sure the dom would have actually existed

[02:56:00.0000] <roc>
it didn't

[02:56:01.0000] <hsivonen>
olliej: but if interop requires string objects to have those methods, should the spec that specs string object methods spec those methods?

[02:57:00.0000] <zcorpan>
are they required for web compat or are they just leftovers that noone use?

[02:57:01.0000] <olliej>
hsivonen: no, because a) the DOM spec could easily say "these additional functions must be present on String.prototype"

[02:57:02.0000] <olliej>
hsivonen: and b) i didn't even know they existed till now

[02:58:00.0000] <olliej>
hsivonen: so i doubt they're widely used

[02:58:01.0000] <jgraham>
They are implemeted in all the standalone shells I have

[02:58:02.0000] <olliej>
hsivonen: we could ask hixie to check for us :D

[02:58:03.0000] <jgraham>
So they are really part of ECMAsScript implementation wise

[02:58:04.0000] <hsivonen>
/me gets my first and only JS book from 1996 from the shelf

[02:59:00.0000] <annevk2>
does IE have them?

[03:00:00.0000] <jgraham>
/me is very happy  to drop them if they are shown to not be needed for web-compat

[03:00:01.0000] <jgraham>
But I suspect that will not be the case unfortunately

[03:02:00.0000] <hsivonen>
the JS book from 1996 has a section on string methods and it lists stuff like big() and fixed()

[03:02:01.0000] <hsivonen>
aside: the hello world equivalent in the book used document.write()

[03:03:00.0000] <olliej>
hsivonen: that's still stupid

[03:03:01.0000] <hsivonen>
my guess is that existing content from the Netscape 2.0 era might use them

[03:03:02.0000] <olliej>
hsivonen: and they can still be made into dom defined functions

[03:03:03.0000] <olliej>
hsivonen: yeah, but i don't think they realistically matter anymore

[03:04:00.0000] <olliej>
we can even have "4" in our user agent now

[03:05:00.0000] <olliej>
hsivonen: and back when safari 2.0.4 was release we discovered that 4 in the user agent string was bad

[03:05:01.0000] <hsivonen>
olliej: good times

[03:06:00.0000] <olliej>
because we all know that is your user agent includes "4" you support layers, right?

[03:06:01.0000] <jgraham>
"Javascript: The Definitive Guide" (O'Reilly, 1996) lists them (is that the same book that hsivonen looked at?)

[03:06:02.0000] <hsivonen>
jgraham: mine is teach yourself JavaScript in a week by Arman Danesh

[03:08:00.0000] <olliej>
mine is JavaScriptCore.xcodeproj ;)

[03:09:00.0000] <hsivonen>
olliej: clearly, you need more accurate docs :-)

[03:09:01.0000] <olliej>
hsivonen: heheh

[03:12:00.0000] <zcorpan>
so what's the complete list of these methods?

[03:13:00.0000] <hsivonen>
whoa. a .otf (postscript outlines) version of a font can be one third smaller than a .ttf (truetype outlines) version

[03:14:00.0000] <hsivonen>
I wonder how SVG fonts fare on file size

[03:14:01.0000] <annevk2>
has anyone checked if IE supports them?

[03:15:00.0000] <hsivonen>
zcorpan: anchor, big, blink, bold, fixed, fontcolor, fontsize, italics, link, small, strike, sub, sup

[03:15:01.0000] <zcorpan>
http://www.google.com/codesearch?hl=en&lr=&q=%28%22%7C%27%29%5Cs%2A%5C.%5Cs%2Afontcolor%5Cs%2A%5C%28+lang%3Ajs&sbtn=Search

[03:16:00.0000] <hsivonen>
annevk2: it does

[03:17:00.0000] <annevk2>
nasty

[03:17:01.0000] <hsivonen>
it seems to me that purging the Web of these methods isn't worth the effort compared to just writing them down in the corner of a spec

[03:17:02.0000] <annevk2>
true

[03:18:00.0000] <zcorpan>
http://www.google.com/codesearch?hl=en&lr=&q=%5C.%5Cs%2Abig%5Cs%2A%5C%28+lang%3Ajs&sbtn=Search

[03:18:01.0000] <jgraham>
Things with "Demo" in the title don't count :)

[03:19:00.0000] <jgraham>
Where is Philip` when you need him?

[03:19:01.0000] <zcorpan>
http://www.google.com/codesearch?hl=en&lr=&q=%5C.%5Cs%2Afontsize%5Cs%2A%5C%28+lang%3Ajs&sbtn=Search

[03:21:00.0000] <annevk2>
putting them in HTML5 makes some sense I suppose if the ECMAScript guys don't do it

[03:21:01.0000] <annevk2>
it's HTML stuff after all

[03:22:00.0000] <jgraham>
Shall I start a ECMAScript iki page?

[03:22:01.0000] <jgraham>
*wiki

[03:23:00.0000] <annevk2>
yeah, something similar to http://wiki.whatwg.org/wiki/HTTP_Issues

[03:29:00.0000] <olliej>
!!

[03:30:00.0000] <olliej>
"foo".blink() exists!!

[03:30:01.0000] <olliej>
noooooo

[03:33:00.0000] <hsivonen>
zcorpan: IIRC, simple XLink is supported on XML element nodes whose namespace isn't well-known to Gecko

[03:34:00.0000] <annevk2>
I wonder why they don't just kill XLink support

[03:35:00.0000] <hsivonen>
annevk2: my understanding is that killing it is the plan

[03:35:01.0000] <hsivonen>
annevk2: and restoring special support for MathML is the plan also

[03:36:00.0000] <annevk2>
might make more sense to get the MathML guys to introduce an href attribute

[03:36:01.0000] <annevk2>
they seemed willing

[03:37:00.0000] <hsivonen>
perhaps I should switch my site to OTF fonts and let the theoretical Price users upgrade to Price 7

[03:38:00.0000] <hsivonen>
it seems that the OTF rasterizer on Windows sucks less than the TTF rasterizer

[03:38:01.0000] <hsivonen>
and the OTF fonts are smaller

[03:40:00.0000] <hsivonen>
Grr. Opera 10 on Windows doesn't support OTF

[03:41:00.0000] <hsivonen>
oops. my test case is wrong

[03:48:00.0000] <zcorpan>
still wonder if it's practical to use html <a href> for linking in mathml

[04:03:00.0000] <annevk2>
so Safari/IE conform to the spec for nonexistent keys on localStorage and Firefox has a bug for localStorage.missing

[04:03:01.0000] <annevk2>
I guess it's not worth changing anymore then :/

[04:09:00.0000] <hsivonen>
zcorpan: I think I've now fixed the answers in the Mozilla Web Author FAQ to be up-to-date

[04:09:01.0000] <hsivonen>
zcorpan: however, the set of questions might no longer be the most frequent set of questions

[04:16:00.0000] <jgraham>
Interestingly although the enumeration order for user-defined properties on JS Objects seems to be well defined, the relative order between between host-object properties and user defined properties seems to vary

[04:17:00.0000] <jgraham>
Although perhaps "interestingly" is the wrong word

[04:17:01.0000] <olliej>
jgraham: hehehe

[04:19:00.0000] <jgraham>
/me wwould have expected user-defined properties to enumerate last

[04:19:01.0000] <hsivonen>
/me points to topic

[04:20:00.0000] <jgraham>
/me has been n00b'd

[04:26:00.0000] <jgraham>
http://wiki.whatwg.org/wiki/Web_ECMAScript

[04:26:01.0000] <jgraham>
Still lots to add of course :0

[04:26:02.0000] <jgraham>
:) even

[04:27:00.0000] <annevk2>
e.g. escaping of attribute values and such maybe?

[04:28:00.0000] <annevk2>
hmm, maybe not

[04:29:00.0000] <jgraham>
annevk2: I don't think there is anything clever at all

[04:29:01.0000] <annevk2>
this stuff is not safe at all

[04:29:02.0000] <jgraham>
I more meant other things where the ES5 spec and web browsers have disagreements e.g. things that are formally implementation defined but actually rather tightly constrained

[04:30:00.0000] <jgraham>
annevk2: No

[04:31:00.0000] <annevk2>
note btw that arguments are optional and infinite

[04:32:00.0000] <jgraham>
annevk2: What do you mean optional and infinite?

[04:32:01.0000] <annevk2>
e.g. .link("a","x","xd") gives "a" and .link() gives undefined

[04:32:02.0000] <annevk2>
"undefined"*

[04:32:03.0000] <jgraham>
annevk2: Yeah ToString will do that

[04:33:00.0000] <annevk2>
doesn't extra arguments normally throw?

[04:33:01.0000] <annevk2>
maybe not

[04:33:02.0000] <jgraham>
annevk2: I don't think so

[04:34:00.0000] <annevk2>
I guess I'm confusing ECMAScript and WebiDL

[04:34:01.0000] <annevk2>
though maybe in WebIDL they don't throw either

[04:53:00.0000] <Philip`>
jgraham: In bed

[05:03:00.0000] <jgraham>
Philip`: That's what you said last time I asked the same question

[05:04:00.0000] <Philip`>
jgraham: I sleep a lot

[05:04:01.0000] <Philip`>
Actually I don't, but I sleep a lot in mornings

[05:06:00.0000] <jgraham>
Whereas I actually do sleep a lot but am much less useful so people rarely notice

[05:11:00.0000] <Philip`>
<script language="javascript">var p1="mai";var p2="lto:";var r="johnny";var d="superbikeschool.co.uk";document.write(("Johnny Haynes").link(p1+p2+r+String.fromCharCode(64)+d));</script>

[05:11:01.0000] <Philip`>
this.value=this.value.fontcolor(c);

[05:11:02.0000] <Philip`>
strsearch +=' <div class="bodytext"><strong>The Keyword(s) you searched : </strong></div>'.big() ;

[05:12:00.0000] <Philip`>
People seem to use these things quite a bit

[07:06:00.0000] <annevk2>
hsivonen, I took the initiative of replying to adactio's Ariability post

[07:24:00.0000] <hsivonen>
annevk2: cool

[08:25:00.0000] <hsivonen>
annevk2: your comment system rejected http://pastebin.mozilla.org/643135 upon Post but not upon Preview

[08:27:00.0000] <hsivonen>
annevk2: in case you put it in via back door, typo fix: http://pastebin.mozilla.org/643136

[08:34:00.0000] <annevk2>
hsivonen, added

[08:34:01.0000] <hsivonen>
annevk2: thanks

[08:46:00.0000] <myakura>
/me wonders if any browser vendors would support "grid" media features

[10:05:00.0000] <zcorpan>
hsivonen: http://annevankesteren.nl/2009/04/html5-wai-aria#comment-6756 - the architectural forms link is broken

[10:06:00.0000] <zcorpan>
um or at least it was broken for me. but i could search for the url in waybackmachine and then it worked

[12:47:00.0000] <jwalden>
olliej, annevk2, hsivonen, et al: the HTML String.prototype methods are broken, at least in SpiderMonkey, wrt escaping; I don't know whether this is common across implementations or not

[12:53:00.0000] <jgraham>
jwalden: The behaviour is the same in Squirrelfish and Futhark

[12:54:00.0000] <jwalden>
win!

[12:54:01.0000] <jwalden>
:-\

[12:54:02.0000] <jgraham>
jwalden: At least it makes the spec and implementation simple :)

[13:02:00.0000] <annevk2>
jwalden, lets consider it a feature

[13:03:00.0000] <mpilgrim_>
/me wonders where the hell i'm logged in already

[13:04:00.0000] <mpilgrim>
aha

[13:05:00.0000] <annevk2>
we call this hell #whatwg

[13:05:01.0000] <mpilgrim>
found this while digging for link rel discussions

[13:05:02.0000] <mpilgrim>
http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2006-September/007301.html

[13:07:00.0000] <mpilgrim>
"I wonder how open the HTML WG will be with regards to working with the WHATWG and HTML 5, especially now that the 2 specs will share the same namespace.  If we don't resolve the incompatibilities, one of the specs will simply be doomed to failure."

[13:08:00.0000] <mpilgrim>
fast forward to today: http://www.w3.org/2009/04/16-html-wg-minutes.html#item05

[13:09:00.0000] <mpilgrim>
and http://www.w3.org/html/wg/tracker/actions/105

[13:09:01.0000] <mpilgrim>
"Current status: awaiting input from plh and Steve Pemberton as to next steps, will update the working group with status as it becomes available."

[13:11:00.0000] <jwalden>
I agree with "one of the specs will simply be doomed to failure"

[13:11:01.0000] <jwalden>
give you one guess which I mean

[13:11:02.0000] <jgraham>
Yeah well there is reality and there is W3C poliics

[13:12:00.0000] <jgraham>
The goal is to not let politics affect reality too much

[13:12:01.0000] <annevk2>
And there's me buying a round to the #whatwg channel if we give up now and let them have it :p

[13:13:00.0000] <annevk2>
Though that might not help as before #whatwg it didn't gain much traction either...

[13:15:00.0000] <Hixie>
if w3c politics could affect reality it wouldn't be such a big deal

[13:17:00.0000] <mpilgrim>
anyone know when rel=feed was added to the spec?

[13:18:00.0000] <mpilgrim>
earlier discussion i can find is http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2006-November/008080.html which references text already in the spec

[13:18:01.0000] <Hixie>
try an svn blame and then do spec archeology :-)

[13:22:00.0000] <annevk2>
mpilgrim, http://html5.org/tools/web-apps-tracker?from=334&to=335

[13:22:01.0000] <mpilgrim>
bless you

[13:22:02.0000] <annevk2>
mpilgrim, the frontpage of the tracker takes a hidden argument limit with a hidden value -1

[13:22:03.0000] <mpilgrim>
oooooh

[13:23:00.0000] <annevk2>
mpilgrim, it takes a while to complete the request but then you can do text search on the svn log

[13:25:00.0000] <mpilgrim>
what's the query parameter for that?

[13:26:00.0000] <mpilgrim>
oh

[13:26:01.0000] <mpilgrim>
is it "limit"?

[13:27:00.0000] <mpilgrim>
yes, yes it is

[13:27:01.0000] <mpilgrim>
/me enjoys talking to himself

[13:28:00.0000] <mpilgrim>
wow, my article just got a whole lot better

[13:28:01.0000] <mpilgrim>
thanks annevk2

[13:28:02.0000] <mpilgrim>
i found checkins for virtually every rel value i want to talk about

[13:29:00.0000] <annevk2>
:)

[13:52:00.0000] <tantek>
rel="feed" - interesting, first I've heard of this. mpilgrim, do you think it's a good idea? do you agree with Mark Baker's (theoretical) points he makes in http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2006-November/008080.html

[13:54:00.0000] <tantek>
mpilgrim, also if you're writing an article about many rel values, perhaps this resource may be of assistance: http://microformats.org/wiki/existing-rel-values

[13:57:00.0000] <Hixie>
rel=feed was added because of implementations, iirc

[13:57:01.0000] <Hixie>
(and mark's e-mail i believe got an answer, mostly along the lines of "i'm just doing what browsers do", i imagine)

[14:00:00.0000] <annevk2>
I thought we added rel=feed because the existing solution became a mess

[14:01:00.0000] <annevk2>
afaik only Firefox supports it

[15:22:00.0000] <mpilgrim>
test cases: http://diveintomark.org/tmp/relalternate.html and http://diveintomark.org/tmp/relfeed.html

[15:22:01.0000] <mpilgrim>
all modern browsers support the former (except google chrome, which has no feed autodiscovery at all)

[15:23:00.0000] <mpilgrim>
firefox 3 supports rel=feed

[15:23:01.0000] <Hixie>
sounds right

[15:23:02.0000] <mpilgrim>
firefox 2 does not support rel=feed

[15:24:00.0000] <mpilgrim>
opera 9.62 does not support rel=feed

[15:24:01.0000] <mpilgrim>
safari 4 beta does not support rel=feed

[15:24:02.0000] <mpilgrim>
IE8 final does not support rel=feed

[15:37:00.0000] <mpilgrim>
rel=feed originated in the Atom working group in 2005

[15:37:01.0000] <mpilgrim>
http://www.intertwingly.net/wiki/pie/PaceDifferentRelValue

[15:37:02.0000] <mpilgrim>
this appears to be the origin: http://www.imc.org/atom-syntax/mail-archive/msg15042.html

[15:38:00.0000] <mpilgrim>
my reaction to it at the time: http://www.imc.org/atom-syntax/mail-archive/msg15069.html "Part of my newfound personal definition of a life well-lived is to never again argue about semantics, markup, or the "correct" way to use them.  This Pace will break every aggregator on the planet, but then again, so will Atom 1.0 feeds, so... +0."

[15:40:00.0000] <mpilgrim>
here's annevk2 explaining the benefits of rel=feed to sayrer: http://www.imc.org/atom-syntax/mail-archive/msg15074.html

[15:41:00.0000] <mpilgrim>
to answer tantek's original question, i think rel=feed is an interesting idea that hasn't panned out

[15:41:01.0000] <mpilgrim>
only one browser vendor has implemented it in 4 years

[15:42:00.0000] <mpilgrim>
author awareness is zero

[15:42:01.0000] <mpilgrim>
AFAIK, rel=feed links are not included in any major CMS's default templates

[15:43:00.0000] <mpilgrim>
though perhaps Philip` has some concrete data on that

[15:45:00.0000] <annevk2>
the benefits I mention are very ahum compelling

[15:46:00.0000] <mpilgrim>
/me doesn't know whether "ahum" is a typo or a new word he should look up

[15:46:01.0000] <mpilgrim>
firefox 3 support appears to be here: http://mxr.mozilla.org/seamonkey/source/browser/base/content/browser.js#2818

[15:46:02.0000] <mpilgrim>
it looks like removing support for rel=feed would be a matter of deleting a single case statement in browser.js

[15:46:03.0000] <annevk2>
apparently it's a word in the language Twi?! -- http://www.websters-online-dictionary.org/translation/Twi/ahum

[15:46:04.0000] <mpilgrim>
and removing the associated tests cases

[15:47:00.0000] <Philip`>
I see four pages with rel=feed

[15:47:01.0000] <annevk2>
http://www.urbandictionary.com/define.php?term=ahum is also not that useful

[15:47:02.0000] <Philip`>
(based on token matches)

[15:47:03.0000] <mpilgrim>
4 out of ...?

[15:47:04.0000] <Philip`>
and two of them are bogus

[15:47:05.0000] <Philip`>
Out of 130K from dmoz.org

[15:48:00.0000] <Philip`>
Oh, actually, 7 pages, but 3 are from the same site

[15:48:01.0000] <mpilgrim>
in the 2 valid cases, does the page also specify rel=alternate links?

[15:48:02.0000] <Philip`>
s/7/6/

[15:49:00.0000] <sayrer>
I dunno

[15:49:01.0000] <sayrer>
I'm pretty lazy

[15:49:02.0000] <sayrer>
I think someone made me do rel=feed

[15:49:03.0000] <sayrer>
and I can at least find some ESPN feeds that use it

[15:50:00.0000] <mpilgrim>
my question stands: are there a statistically significant number of pages that use rel=feed to the exclusion of other feed autodiscovery mechanisms?

[15:50:01.0000] <Philip`>
http://www.dip-badajoz.es/municipios/municipio_dinamico/inicio/index_inicio.php?codigo=43 - <link rel="feed alternate" type="application/atom+xml" href="/canales/atom_xml_bop.php?c=1&amp;u=1" title="&Uacute;ltimo B.O.P."/>

[15:50:02.0000] <Philip`>
http://www.volkswagen-stiftung.de/ - <link title="News und Termine der VolkswagenStiftung (RSS 2)" href="http://volkswagenstiftung.de/index.php?id=129&amp;type=100" type="application/rss+xml" rel="alternate feed" />

[15:50:03.0000] <Philip`>
http://www.tourismcapetown.co.za/ - <a href="rss20.xml" rel="RSS 2.0 Feed" type="application/rss+xml" title="RSS Feed"><img border="0" src="img/template/rss20.gif" alt="RSS 20 Feed"></a>

[15:50:04.0000] <Philip`>
http://www.sevillagrande.com/ - <a href="rss.php" rel="rss feed" target="_blank"><img src="images/rss.gif" align="middle" alt="RSS" /></a>

[15:51:00.0000] <Philip`>
(plus two other pages from that first site)

[15:51:01.0000] <sayrer>
mpilgrim, it's a good question. I don't know.

[15:51:02.0000] <mpilgrim>
dip-badajoz.es would match on existing rel=alternate autodiscovery

[15:51:03.0000] <Philip`>
(This is from searching from parsed 'rel' attributes whose value matches (.*\s)?feed(\s|") (case-insensitive)

[15:51:04.0000] <Philip`>
)

[15:51:05.0000] <mpilgrim>
so would volkswagen-stiflung.de

[15:51:06.0000] <Philip`>
(Uh, where the " is a delimiter)

[15:52:00.0000] <Philip`>
(Anyway I think the regexp is about right)

[15:52:01.0000] <mpilgrim>
tourismcapetown appears to be a typo / misunderstanding that happens to match rel=feed by accident

[15:52:02.0000] <mpilgrim>
same with sevillagrande.com

[15:52:03.0000] <sayrer>
oh that's weird

[15:52:04.0000] <sayrer>
the ESPN ones are Apple <pcast> documents

[15:53:00.0000] <mpilgrim>
honk. snort.

[15:53:01.0000] <mpilgrim>
i don't think those are within our scope

[15:53:02.0000] <mpilgrim>
...yet

[15:53:03.0000] <sayrer>
yeah...

[15:54:00.0000] <mpilgrim>
if this is all the data we have, i'm going to send an email to the list recommending dropping rel=feed

[15:54:01.0000] <Philip`>
You could recommend getting more data

[15:55:00.0000] <sayrer>
I would be willing to turn it off in a beta

[15:55:01.0000] <mpilgrim>
i could

[15:55:02.0000] <Philip`>
There's non-zero usage of feed, so maybe there's still non-zero usage of feed without alternate, but my sample is far too small to find it :-p

[15:55:03.0000] <sayrer>
it doesn't seem to generate bug reports, so the cost seems low

[15:55:04.0000] <mpilgrim>
i could dust off some of hixie's mapreduce code and query against a larger sample

[15:56:00.0000] <mpilgrim>
sayrer: ok

[15:56:01.0000] <Philip`>
/me would like it if dotnetdotcom.org's data was more accessible (e.g. via EC2) so he could search more pages easily

[15:56:02.0000] <annevk2>
it looks a whole lot neater than all the cruft you need now, but I guess that's not sufficient

[15:57:00.0000] <mpilgrim>
annevk2: i'd say that was sufficient in 2005 to propose it

[15:57:01.0000] <sayrer>
yeah, worth a try, but IE and Firefox match on most other things

[15:57:02.0000] <mpilgrim>
but implementation has been... minimal

[15:57:03.0000] <sayrer>
so I can see how authors arrived somewhere else

[15:58:00.0000] <sayrer>
especially in a world with lower mac market share and no google browser

[15:58:01.0000] <mpilgrim>
we've cut larger features ;)

[15:58:02.0000] <mpilgrim>
based on non-implementation

[15:58:03.0000] <mpilgrim>
or near-zero implementation

[15:58:04.0000] <annevk2>
some of those we retain as well

[15:58:05.0000] <mpilgrim>
(no offense to sayrer's work)

[15:58:06.0000] <sayrer>
i think maybe philor heckled me

[15:58:07.0000] <mpilgrim>
lol

[15:59:00.0000] <sayrer>
or it was a way to short circuit autodiscovery

[15:59:01.0000] <sayrer>
er, sniffing

[15:59:02.0000] <annevk2>
i.e. <datagrid>, <menu>

[15:59:03.0000] <sayrer>
did datagrid get cut?

[15:59:04.0000] <mpilgrim>
annevk2: but it's not a new feature, it's a reformulation of an existing feature

[15:59:05.0000] <annevk2>
true

[16:00:00.0000] <mpilgrim>
into a form that virtually no one supports

[16:00:01.0000] <annevk2>
sayrer, no(t yet)

[16:00:02.0000] <mpilgrim>
when there's an existing form that virtually everyone supports

[16:01:00.0000] <mpilgrim>
IIRC, rel=alternate type=application(atom|rss)+xml links were present in 17% of all pages on the internet

[16:01:01.0000] <mpilgrim>
in 2005

[16:01:02.0000] <mpilgrim>
but i'd have to check with hixie on that

[16:02:00.0000] <sayrer>
things that generate feeds are probably generate lots of pages

[16:02:01.0000] <sayrer>
er, no "are" needed

[16:02:02.0000] <mpilgrim>
(actually in 2005, it was probably all rss ;) )

[16:02:03.0000] <mpilgrim>
true

[16:03:00.0000] <sayrer>
mpilgrim, did you see what your site did to AT&T?

[16:03:01.0000] <mpilgrim>
i wonder if i can do a mapreduce against the (pick a large number) most popular home pages

[16:03:02.0000] <mpilgrim>
sayrer: no

[16:03:03.0000] <mpilgrim>
reference?

[16:03:04.0000] <sayrer>
I had a chuckle, because that was a one of the old examples of broken RSS, and I thought it meant "Accessibility Technology" until I clicked

[16:04:00.0000] <Philip`>
I see type="application/(atom|rss)+xml" on 9942 out of ~130K pages on dmoz.org

[16:04:01.0000] <sayrer>
mpilgrim, you have that page that shows all of your internet activity

[16:04:02.0000] <sayrer>
it was on there

[16:04:03.0000] <mpilgrim>
ah

[16:04:04.0000] <mpilgrim>
yeah

[16:04:05.0000] <mpilgrim>
that's a bug

[16:04:06.0000] <mpilgrim>
i should fix that

[16:04:07.0000] <sayrer>
yeah, it's just funny that we still have it

[16:05:00.0000] <Philip`>
(2674 with atom, 9822 with rss)

[16:05:01.0000] <sayrer>
clearly ampersands are broken tech

[16:05:02.0000] <mpilgrim>
philip`: that's about 7%

[16:05:03.0000] <Philip`>
(But this data is a bit rubbish because it's all skewed by things like Wikipedia)

[16:05:04.0000] <mpilgrim>
damn, that whole firehose page is a complete hack

[16:05:05.0000] <mpilgrim>
yeah

[16:05:06.0000] <mpilgrim>
that's why i want to query just home pages

[16:05:07.0000] <sayrer>
the microformats religion supports that theory

[16:05:08.0000] <mpilgrim>
surely we can do that

[16:05:09.0000] <mpilgrim>
we're google

[16:06:00.0000] <sayrer>
if IE implements an HTML5ish parser, maybe that will work

[16:07:00.0000] <mpilgrim>
/me wanders off in search of food

[16:48:00.0000] <Rik`>
in what element should meta information about a blog post go (dates, tags, author name, number of comments, stuff like that) ? nav, aside, anything else ?

[16:49:00.0000] <Hixie>
<p>

[16:51:00.0000] <sayrer>
surely dates should go in <time>

[16:52:00.0000] <Hixie>
should? no. may, sure.

[16:52:01.0000] <Rik`>
sayrer: of course, I was just talking about the "meta block", not the individual elements inside

[16:52:02.0000] <Hixie>
<footer> is probably the right element for most of the stuff that nobody cares about

[16:52:03.0000] <sayrer>
oh, good luck with your meta block

[16:53:00.0000] <Rik`>
Hixie: so <aside> is better suited for footnotes maybe ?

[16:53:01.0000] <Hixie>
<aside> is for, well, asides. sidebars. side notes. footnotes too, though they wouldn't be actual footnotes if they were in an <aside>.

[16:54:00.0000] <Rik`>
ok thanks

[16:56:00.0000] <Rik`>
i bet will see a lot of different use for those new sectioning elements

[16:56:01.0000] <sayrer>
I wonder if they are meta crap


2009-04-17
[18:31:00.0000] <Hixie>
i mentioned the address hedral⊙dc on the html5 spec in december last year

[18:31:01.0000] <Hixie>
it's received hundreds of e-mails of spam

[05:24:00.0000] <gsnedders>
/me wonders who Bibbi Svärd is

[05:25:00.0000] <jgraham>
/me wonders how to most usefully answer that question

[05:25:01.0000] <jgraham>
She is the HR person and general enabler of all things in Sweden 

[05:25:02.0000] <gsnedders>
Ah

[05:26:00.0000] <gsnedders>
jgraham: Is she evil?

[05:26:01.0000] <jgraham>
gsnedders: No

[05:26:02.0000] <gsnedders>
That's good then.

[05:26:03.0000] <jgraham>
In other news: I value my job ;)

[05:26:04.0000] <gsnedders>
:)

[05:27:00.0000] <jgraham>
(but seriously Bibbi is pretty wonderful)

[05:27:01.0000] <gsnedders>
/me heads off to lunch, and ponders the prospect of his third ever interview

[05:31:00.0000] <jgraham>
gsnedders: Phone interview? Or f2f?

[05:32:00.0000] <Philip`>
/me thinks interviews should all be conducted on IRC

[05:35:00.0000] <mpt>
That would be much less useful for the interviewer

[05:36:00.0000] <mpt>
Er ... don't mind me, it's Friday and my sense of humor has gone on holiday

[05:38:00.0000] <jgraham>
mpt: I hope it has gone somewhere nice. My sense of humor once spent two weeks in a caravan in North Wales and it rained every day. On the upside, it was happy to come home.

[05:39:00.0000] <mpt>
So you ended up with a wealth of sheep jokes?

[05:45:00.0000] <Lachy>
gsnedders, does this mean you'll be doing your internship in the Swedish office if you get accepted?

[05:58:00.0000] <jcranmer>
Sveden

[07:01:00.0000] <gsnedders>
jgraham: Phone

[07:01:01.0000] <gsnedders>
Lachy: I assume so

[07:26:00.0000] <jgraham>
gsnedders: Cool

[09:08:00.0000] <kanchan_Tripathi>
hi, can anyone tell me why this simple POST via telnet is'nt working

[09:09:00.0000] <kanchan_Tripathi>
http://pastebin.com/d36e8e302

[09:12:00.0000] <Philip`>
Maybe it needs "Content-Type: application/x-www-form-urlencoded"?

[09:14:00.0000] <Philip`>
kanchan_Tripathi: Maybe it needs "Content-Type: application/x-www-form-urlencoded"?

[09:15:00.0000] <kanchan_Tripathi>
Philip can you elaborate it.

[09:16:00.0000] <Philip`>
kanchan_Tripathi: You might need to include that header in your telnet POST request

[09:24:00.0000] <kanchan_Tripathi>
Philip i did that its not working

[09:24:01.0000] <kanchan_Tripathi>
??

[10:04:00.0000] <gsnedders>
/me is way too tempted to collapse asleep in bed

[10:43:00.0000] <smedero>
Philip`: http://boston.lti.cs.cmu.edu/Data/clueweb09/ -- just in case you'd like a lot more web data and have someone willing to pay for it

[10:43:01.0000] <smedero>
if i recall, you're at a Uni... so you might have access to it already through a partnership.

[13:03:00.0000] <takkaria>
"I think that real openness is often much easier to get in a closed forum", hmm

[13:04:00.0000] <Dashiva>
real openness or apparent openness?

[13:04:01.0000] <gavin_>
presumably he means "openness" in the "honesty" sense

[13:51:00.0000] <jgraham>
/me wonders what takkaria was quoting

[13:55:00.0000] <takkaria>
http://annevankesteren.nl/2009/04/html5-wai-aria#comment-6762

[14:28:00.0000] <Hixie>
i'm baffled by the people who actually want working groups to work in a closed manner, ignoring external input

[14:29:00.0000] <Hixie>
(i've heard several people criticize the whatwg development model, but i've yet to hear anyone propose anything that actually results in a better spec)

[14:33:00.0000] <Philip`>
/me wonders how one would judge whether a proposal would result in a better spec

[14:34:00.0000] <gsnedders>
The score of "The Matrix Revolutions" is awesome.

[14:35:00.0000] <Hixie>
Philip`: i've yet to hear anyone propose anything that even they think results in a better spec

[14:35:01.0000] <Philip`>
gsnedders: I thought its score was about 5 out of 10, which doesn't sound that good

[14:35:02.0000] <jgraham>
Philip`: Presumably you do it like astrophysics and just look at lots of examples and try to infer the common proprties and so on

[14:36:00.0000] <gsnedders>
Philip`: The score has far better reviews than the film itself :)

[14:36:01.0000] <Philip`>
gsnedders: Oh, you mean the score scores highly?

[14:37:00.0000] <gsnedders>
Philip`: s/score/soundtrack/g if that makes you understand better

[14:37:01.0000] <Philip`>
gsnedders: That removes the fun of misunderstanding

[14:39:00.0000] <Philip`>
jgraham: In astrophysics there's like a trillion trillion examples you can look at, but for HTML specs there's like one (to the nearest couple of orders of magnitude)

[14:40:00.0000] <Philip`>
which I guess makes any conclusions much less reliable

[14:41:00.0000] <richb_>
just read the article about link-is rel=shorter the same idea as rev canonical?

[14:42:00.0000] <smedero>
rel=short|shorter|shortlink|shorturl|short_url

[14:42:01.0000] <smedero>
what fun!

[14:43:00.0000] <richb_>
great! thought there was meant to be some sort.of spec...

[14:44:00.0000] <richb_>
any ua implement any usage of any?

[15:03:00.0000] <takkaria>
richb_: no, no-one implements any of them as yet

[15:10:00.0000] <richb_>
heard lots more about html5 recently - what happened to xhtml2.0?

[15:11:00.0000] <Hixie>
nothing particularly, it still exists and is being developed:

[15:11:01.0000] <Hixie>
http://www.w3.org/TR/xhtml2/

[15:12:00.0000] <Hixie>
http://lists.w3.org/Archives/Public/public-xhtml2/latest/ is where development happens

[15:12:01.0000] <richb_>
is it still relevant? seems html5 is making more progress

[15:12:02.0000] <Hixie>
depends who you ask

[15:13:00.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/#relationships-to-other-specifications might help

[15:13:01.0000] <richb_>
guess that's just an outsiders view

[15:14:00.0000] <Hixie>
well the people working on xhtml2 presumably think html5 will go nowhere, an the people working on html5 by and large think the same about xhtml2 :-)

[15:14:01.0000] <Hixie>
hence why there are two groups :-)

[15:14:02.0000] <Hixie>
though i hear there's work underway to merge the two groups and just have one language

[15:14:03.0000] <Hixie>
dunno how that's gonna work

[15:18:00.0000] <takkaria>
Hixie: do you ever regret having large portions of Yes Minister on your website under /bible? :)

[15:18:01.0000] <Hixie>
no

[15:18:02.0000] <Hixie>
i use it the same way i used the text on ian.hixie.ch/'s main page

[15:18:03.0000] <Hixie>
it's a bogofilter

[15:19:00.0000] <takkaria>
fair play

[15:19:01.0000] <Hixie>
(people self-identify as feeling unable to argue their case by pointing to the page as evidence that i am not arguing fairly)

[15:21:00.0000] <takkaria>
it also seems blatanly unfair to assume that because something's on your website that's clearly quoted from somewhere else, that you agree with it and use it on a daily basis

[15:21:01.0000] <Hixie>
right -- people who claim that are self-identifying as candidates for my mail filters

[15:21:02.0000] <Hixie>
it's pretty awesome

[15:23:00.0000] <Hixie>
i had the sentence "I am always ready to reassess my core assumptions in the light of new evidence" on that site's home page before that i used in the same way

[15:23:01.0000] <Hixie>
i knew that if people quoted that sentence at me that arguing with them was likely a waste of time

[15:35:00.0000] <Dashiva>
So xhtml2 is slowly adding more and more of html4 back as deprecated? Or?

[15:39:00.0000] <Hixie>
no idea, you'd have to ask the xhtml2 wg

[15:40:00.0000] <Hixie>
the end is in sight with the <code>datagrid</code> rewrite

[15:40:01.0000] <annevk2>
is someone going to implement it that you're doing that now?

[15:43:00.0000] <Hixie>
chrome almost has someone to implement it but i had to tell them to stop given the big problems with the design

[15:43:01.0000] <Hixie>
that was in jan or feb

[15:44:00.0000] <Hixie>
it's pretty late to be doing changes like this at all; when would you want me to do it?

[15:44:01.0000] <annevk2>
i'm just curious :)

[15:45:00.0000] <Hixie>
rdfa is next, after a brief pause for websocket stuff and to fix the typos people found

[15:45:01.0000] <Hixie>
oh and making the specs ready to publish

[15:46:00.0000] <Lachy>
is rdfa actually going to make it into HTML5?

[15:46:01.0000] <Hixie>
depends what the use cases are and how well they are a match for them

[15:47:00.0000] <Lachy>
oh, ok. So you haven't made a final decision yet?

[15:47:01.0000] <Hixie>
"final"?

[15:47:02.0000] <Hixie>
i haven't made any decisions yet, if i had i wouldn't have to do anything :-)

[15:48:00.0000] <Hixie>
and decisions i make are hardly final, after all, the wg has to agree to them before the spec is done

[15:48:01.0000] <Lachy>
I meant, you haven't decided whether or not you will be adding any RDFa to the spec yet

[15:48:02.0000] <Hixie>
i decided not to at least once before

[15:49:00.0000] <Hixie>
but there are new use cases that have been introduced to the discussion

[15:49:01.0000] <takkaria>
I wish people would stop harping on about this "invalid code in HTML5 has no negative consequences" thing, since s/5/4/ and you're just as accurate

[15:50:00.0000] <Hixie>
"people"? i thought it was just one person

[15:50:01.0000] <Hixie>
and yeah i don't understand what they mean by "no negative consequences"

[15:50:02.0000] <Hixie>
isn't a page being inaccessible a pretty serious negative consequence?

[15:50:03.0000] <Lachy>
who said there's no negative consequences?

[15:50:04.0000] <takkaria>
maybe it is just John Foliot getting around

[15:50:05.0000] <Hixie>
john foliot keeps saying that making pages inaccessible has no serious negative consequences, or something along those lines. I don't fully understand what he's saying exactly.

[15:51:00.0000] <annevk2>
John Foliot has been touting it lately

[15:51:01.0000] <Lachy>
invalid code doesn't inherently mean inaccessible though.

[15:51:02.0000] <Lachy>
Hixie, I assume you mean s/inaccessible/invalid/

[15:51:03.0000] <Hixie>
Lachy: no but he's talking about invalid code in terms of not providing <canvas> fallback, and things along those lines, as i understand it

[15:51:04.0000] <annevk2>
his point is that since HTML5 defines error handling making errors is not a problem

[15:51:05.0000] <Hixie>
clearly i don't understand what he's saying

[15:52:00.0000] <Hixie>
but that's nothing new

[15:52:01.0000] <annevk2>
and that therefore using ARIA (which is invalid) is fine

[15:52:02.0000] <Hixie>
aria is invalid?

[15:52:03.0000] <Lachy>
ok

[15:52:04.0000] <othermaciej>
sicking: I'm glad I'm not the only one to fail to communicate with Anders Rundgren

[15:52:05.0000] <annevk2>
Hixie, the validator complaints about bits I believe

[15:52:06.0000] <Hixie>
is this a bug in the validator? or is he saying that violating aria's own rules is fine?

[15:53:00.0000] <Hixie>
i'm even more confused now than i was before!

[15:53:01.0000] <annevk2>
Hixie, I don't really understand what his aim is either; validation seems useful regardless of whether you use some extensions that not validate currently

[15:53:02.0000] <annevk2>
Hixie, hsivonen doesn't support e.g. landmark roles

[15:55:00.0000] <Hixie>
ok...?

[15:57:00.0000] <annevk2>
I had nothing else. (Could also be that it's not about validators and just about the HTML5 spec not "blessing" WAI-ARIA I suppose.)

[15:57:01.0000] <Hixie>
well yeah, we haven't done that yet

[15:57:02.0000] <Hixie>
we can't very well do that until aria is ready for it

[15:57:03.0000] <Hixie>
i even sent feedback on that very topic recently

[15:59:00.0000] <annevk2>
yeah, me too

[16:00:00.0000] <annevk2>
I suppose I should just stop reading rants from John Foliot because they are indeed confusing and not very nice. E.g. "I am dismissing many of your arguments based upon the process and procedure that Ian applies (deny, ignore, stall)."

[16:01:00.0000] <Hixie>
q.v. the earlier discussion of how to tell who's actually acting in good faith and who isn't. :-)

[16:01:01.0000] <annevk2>
Irrespective of whether what he says about you is true, it would be nice if he and I could converse normally, but I guess that's not possible.

[16:22:00.0000] <gsnedders>
/me realizes why backup has failed… the server is off.

[16:23:00.0000] <inimino>
/me notes that it's always better to realize why your backup failed than why your restore failed

[16:35:00.0000] <gsnedders>
Oh dear… Someone using PHP 4.3.3RC2


2009-04-18
[17:32:00.0000] <jwalden>
somebody in here should fix the note on <http://wiki.whatwg.org/wiki/Web_ECMAScript> about eval: a use of eval not called "eval" does not necessarily work according to ECMA-262 ed. 3; see section 16, and note that (mentally fixing borken spacing due to a half-broken PDF) "An i mpl ement at i on may t r eat any i nst ance of t he f ol l owi ng ki nds of r unt i me er r or s as a synt ax er r...

[17:32:01.0000] <jwalden>
...or and therefore r epor t i t ear l y: • Usi ng t he eval propert y ot her t han vi a a di rect cal l ."

[17:32:02.0000] <jwalden>
var x = eval; // not a direct call

[17:32:03.0000] <Hixie>
it's a wiki, you are welcome to edit it :-)

[17:32:04.0000] <jwalden>
I don't think I have a login yet, and I'm preoccupied right now :-)

[18:54:00.0000] <Hixie>
crap, i don't know how to do multi-row selection in <datagrid>

[19:03:00.0000] <karlcow>
I'm using style attributes most of the time for very local and specific style rule (which font element doesn't cover). Such as modifying the style of an image for a one time. <img style="float:right;margin-left:1em;width:350px;" …/> I haven't found a better way of doing it. 

[19:12:00.0000] <Hixie>
ok, <datagrid> mark II is now checked in

[19:12:01.0000] <Hixie>
i'm sure it has issues

[19:12:02.0000] <Hixie>
e.g. the content model makes no sense anymore

[02:00:00.0000] <jgraham>
jwalden: The point re: eval is that ES5 allows x = eval; x() to throw but browsers don't do that, they call something like eval but possibly in a different scope

[02:01:00.0000] <jgraham>
So it seems at least plausible that there are sites that rely on that not throwing

[02:23:00.0000] <zcorpan_>
"HTML 5 is at least as far off implementation as XHTML 2 is. There is nowhere near as much need to discuss XHTML 2 because it is a finished stable product just waiting to be implemented. It will probably be at least another 10 years (if ever) before HTML 5 reaches that same point." -- http://www.sitepoint.com/forums/showthread.php?p=4226031#post4226031

[02:47:00.0000] <zcorpan_>
Hixie: s/Either: Nothing./Either: Empty./

[02:49:00.0000] <zcorpan_>
/me wonders whether <datalist>foo<table></table></datalist> is allowed

[03:10:00.0000] <annevk2>
zcorpan_, that's a fun quote

[03:17:00.0000] <gsnedders>
/me is missing 5.85

[03:17:01.0000] <gsnedders>
(GBP)

[03:29:00.0000] <zcorpan_>
hmm, maybe the last argument to addColumn() should be *in*visible? it's weird when absent argument is the same as true

[04:10:00.0000] <takkaria>
zcorpan_: hmm, interesting thread

[04:10:01.0000] <takkaria>
apparently "it is impossible to write documents about Julius Cesar, Moses or dinosaurs using HTML 5" on teh third thread

[04:13:00.0000] <takkaria>
*third page

[04:41:00.0000] <gsnedders>
Hmm, breakfast is starting to be rather tempting

[07:17:00.0000] <gsnedders>
/me slaps last.fm for having no pause button

[08:04:00.0000] <takkaria>
if the HTTP WG care more about theoretical purity than having people actually implement the spec as written, they should rename the Referer header Referrer

[08:09:00.0000] <Philip`>
takkaria: But that would break non-browser UAs, which is unacceptable

[08:09:01.0000] <Philip`>
(Browsers deserve to be broken because they ignore the current specs anyway)

[13:34:00.0000] <annevk2>
hmm, can you post arbitrary HTML to www-archive?

[13:34:01.0000] <annevk2>
if you then slip such a link to a W3C Member you can presumably read all kinds of information from Member only pages

[13:34:02.0000] <annevk2>
seems dangerous

[13:36:00.0000] <Philip`>
annevk2: I guess it would be contained within lists.w3.org, and you couldn't access www.w3.org?

[13:37:00.0000] <Philip`>
(but it seems nobody has cared to fix http://lists.w3.org/Archives/Public/www-rdf-validator/2009Mar/0005.html so you can access Member-only pages on www.w3.org anyway)

[13:53:00.0000] <annevk2>
Philip`, there is http://lists.w3.org/Archives/Member/

[14:07:00.0000] <annevk2>
"Seven Things You Probably Don't Know About Me"... hmm

[14:08:00.0000] <annevk2>
1. I stopped doing memes. 2-7. See 1. Harharhar.

[14:09:00.0000] <gsnedders>
:P

[14:09:01.0000] <annevk2>
What would be the appropriate markup for such a list?

[14:10:00.0000] <gsnedders>
<ol><li value=1>I stopped doing memes.<li min=2 max=7>See 1.</ol><p>Harharhar.

[14:10:01.0000] <gsnedders>
(I know that doesn't work.)

[14:10:02.0000] <gsnedders>
For those who don't read what I blog within two minutes: http://gsnedders.com/seven-things-you-probably-dont-know-about-me

[14:10:03.0000] <Philip`>
<p>1. I stopped doing memes. 2-7. See 1. Harharhar.</p>

[14:15:00.0000] <annevk2>
http://bitworking.org/news/425/atompub-is-a-failure he's blaming browsers in a non-negative way; that's new

[14:29:00.0000] <Philip`>
/me might be more interested in AtomPub if he actually had any idea what it was

[14:29:01.0000] <Philip`>
I must have read dozens of blog posts talking about it, and I still have no idea

[14:33:00.0000] <annevk2>
it's a protocol to manipulate content over HTTP

[14:33:01.0000] <gsnedders>
/me needs to stop doing pointless Facebook quizzes

[14:36:00.0000] <annevk2>
Philip`, if that is too abstract you might just want to read the spec: http://tools.ietf.org/html/rfc5023

[14:36:01.0000] <Philip`>
HTTP is a protocol that lets me manipulate content over HTTP :-)

[14:37:00.0000] <Philip`>
annevk2: I think that's a bit too concrete

[14:38:00.0000] <annevk2>
Philip`, get over it

[14:44:00.0000] <sayrer>
Philip`: it's just a way to sequence certain kinds of HTTP requests. A protocol, if you will.

[14:50:00.0000] <gsnedders>
/me wonders how to not panic at the prospect of a phone interview

[14:50:01.0000] <Dashiva>
Talk to strangers on the phone in advance

[14:50:02.0000] <Dashiva>
Have your friends prank call you at random times of day to answer questions

[14:50:03.0000] <gsnedders>
But that's scary!

[14:51:00.0000] <Dashiva>
Realize that in reality, the phone is scared of you

[14:51:01.0000] <gsnedders>
Also, if I do that, they'll likely phone up at 2am, drunk, asking for pizza.

[15:22:00.0000] <gsnedders>
annevk2: I do like your second latest blog post.


2009-04-19
[03:35:00.0000] <erlehmann>
Hmmm — if I get no replies on my mail, does that mean no one is interested?

[05:39:00.0000] <Dashiva>
erlehmann: http://en.wikipedia.org/wiki/Warnock's_Dilemma

[09:17:00.0000] <hsivonen>
http://twitter.com/michaelhkay/status/1541670540

[09:34:00.0000] <Xanthir>
Heya, everyone.  I think this topic has been raised before, but has anyone in-room heard of the proposal for the @fullsize attribute on img?

[09:36:00.0000] <Xanthir>
I *really* like the idea, and believe that it addresses an important use-case, but can see that it fails badly at graceful degradation.

[09:37:00.0000] <Xanthir>
As such, what would the room think of a "_modal" value for @target of <a> instead?  When clicked, it loads the resource indicated in the @href, but in a lightbox-style page-modal window.

[09:38:00.0000] <Xanthir>
If the resource has a title which can be auto-discovered (such as if it is an html document with a <title>), it would use that in a title bar; otherwise it would take the title from the @title attribute on <a>

[09:38:01.0000] <jgraham>
Xanthir: Why is it an important enough use-case that is deserves native support in html? Why are custom css/js solutions not better? Is a single attribute enough expressiveness to do what people actually want?

[09:38:02.0000] <Xanthir>
Modal windows in general, and modal image popups in specific, are used *everywhere*.  Pretty much *every* image gallery that's been produced in the last two years or so uses it.

[09:39:00.0000] <Xanthir>
This is because it's honestly a very useful thing to do.  It lets you display large numbers of low-bandwidth resources on a page, and then allow people to load large-size versions on command.

[09:39:01.0000] <jgraham>
Well e.g. flickr and facebook don't really do that

[09:40:00.0000] <jgraham>
(they just use normal links)

[09:40:01.0000] <Xanthir>
Indeed, because they are explicitly focused on the image itself, and so that gets loaded into the main frame.  Many many personal and small-business sites use it, though.

[09:42:00.0000] <Xanthir>
CSS solutions don't really exist right now.  With :target, they *could*, but you'd be explicitly loading all the large-size images immediately.

[09:42:01.0000] <jgraham>
It is really unapparent to me that there is any need to standardise what is quite a sophisticated way of displaying images. 

[09:43:00.0000] <jgraham>
It is much more useful to standardise simple primitives in HTML and then let authors build more complex things on the platform using js + CSS

[09:44:00.0000] <jgraham>
Especially things that have such a limited scope for reuse

[09:45:00.0000] <jgraham>
(a lightbox-style image popup is really only useful if you want a lightbox style image popup; just because they are fashionable today doesn't mean that they won't be like a 2009-era mullet)

[09:45:01.0000] <Xanthir_>
Blar, sorry jgraham.  My wireless is flaky.

[09:45:02.0000] <jgraham>
Join the club :)

[09:46:00.0000] <Rik`>
and modal dialogs are not that cool

[09:46:01.0000] <Xanthir_>
I'm not quite sure what you mean by "sophisticated" in this context.  Lightbox implementations have some variance in exact display, but overall appearance and functionality is pretty standard across them.

[09:46:02.0000] <Rik`>
you can't use the page while viewing images

[09:47:00.0000] <Xanthir_>
Point, Rik`, but lightbox is still extremely common on the web, whether used to display images, videos, or text blocks.

[09:47:01.0000] <Rik`>
and functionnalities are not that standard

[09:48:00.0000] <Xanthir_>
Can you explain further on that?  In my experience, they are.

[09:48:01.0000] <Rik`>
some lightboxes allow you to go to the next or previous image, some don't

[09:48:02.0000] <Rik`>
some accept escape to close it, some don't

[09:49:00.0000] <Xanthir_>
next/previous image: indeed, but that's more of a gallery thing than a modal thing in and of itself.  Useful, yes, but only in limited contexts and requires more autodiscovery than would probably be possible.

[09:49:01.0000] <Rik`>
I still prefer popups for that kind of content but that's my point of view

[09:49:02.0000] <jgraham>
Some allow text some don't. Some style the text differently to others

[09:49:03.0000] <Xanthir_>
Allowing escape: that's a minor issue, and should be addressed by the browser itself.

[09:50:00.0000] <Xanthir_>
Allowing text: indeed, but there's no reason *not* to allow it.  Styling should be the domain of CSS (and is, in any major lightbox implementation).

[09:51:00.0000] <jgraham>
So standardising this would need a way of specifying the image, a way of specifying the text (icluding maybe a title, maybe allowing rich markup for lists of e.g. photo settings), allowing a way to set the background colour and opacity, and probably more

[09:51:01.0000] <Xanthir_>
Rik`: It would be perfectly appropriate for the browser to open an "_modal" link in a popup or new tab or whatever.  It's a display suggestion, not a requirement.

[09:52:00.0000] <jgraham>
So you need to work out how all that stuff is going to work and, at the end of the day, you end up with something that is rather trivial to implement in javascript+CSS

[09:52:01.0000] <Rik`>
Xanthir_: and :target doesn't loading all images

[09:52:02.0000] <Rik`>
*mean

[09:52:03.0000] <jgraham>
Except in a form that doesn't work in any existing browser

[09:52:04.0000] <Xanthir_>
jgraham: Yes.  My suggestion is @target="_modal" to specify the resource.  Text is specified in the resource.  Setting CSS should be a CSS issue, likely via a pseudoelement.

[09:53:00.0000] <jgraham>
and is unlikely to work in future browsers because browser makers are unlikely to see implementing it as a huge deal

[09:53:01.0000] <Xanthir_>
Rik`: ?_?  It would.  To simulate lightboxes in pure CSS, you have to put them in the document from the beginning.  A browser *could* possibly optimize by not loading the display:none images immediately, I suppose.

[09:53:02.0000] <Rik`>
I believe they already do that

[09:54:00.0000] <gsnedders>
They don't.

[09:54:01.0000] <gsnedders>
Opera used to, but nothing else ever did, AFAIK.

[09:55:00.0000] <Xanthir_>
jgraham:  I'll be writing it out explicitly.  These issues basically *are* already worked out, because most of them are trivial when you approach the problem from the right direction.

[09:56:00.0000] <Xanthir_>
The bonus is that it allows authors to do something they're already doing, but without a javascript library being required.

[09:56:01.0000] <Xanthir_>
Plus, it degrades to a perfectly useful new window in a non-supporting browser.

[09:56:02.0000] <Xanthir_>
(One named "_modal".)

[09:58:00.0000] <Xanthir_>
In a good browser, you'll just get a fancier display and better user interaction, similar to how many of the new input types work.

[10:01:00.0000] <jgraham>
Xanthir_: I don't really understand how your suggestion would work. You are suggesting creating an auxillary browsing context that im supporting browsers would be overlaid on the existing browsing context but with different default style rules? How would you get back to the first browsing context?

[10:03:00.0000] <Xanthir_>
Yes, that is what I'm suggesting.  The browser, as part of its default UI, would provide a way to close the new context (most likely, a simple X button in the upper-right, with some keyboard ability as well.)

[10:04:00.0000] <jgraham>
Having browser-added chrome on the page seems odd

[10:06:00.0000] <Xanthir>
Really?  For accessibility and discovery purposes, I'd prefer it to be browser-added.

[10:08:00.0000] <jgraham>
Xanthir: Right but there are very feww examples of browser added chtrome on a page. Usually keeping a clean sepaeration between browser-added things and site-added things is good for aesthetics, security, etc.

[10:10:00.0000] <Xanthir>
Crap, I just threw together a mock page for exploring CSS-driven modal dialogs.  http://www.xanthir.com/css-modal.html

[10:10:01.0000] <Xanthir>
:target just keeps getting more useful.

[10:11:00.0000] <Xanthir>
There's still a handful of problems with it, but they might need to be solved anyway in the future.

[10:12:00.0000] <Xanthir>
(better centering ability in CSS, better sizing (min-intrinsic should help), ability to defeer loading of non-displayed resources)

[10:12:01.0000] <jgraham>
Xanthir: That is quite cool but using js has the advantage that it will work in the absence of stylesheets (and maybe in the absence of js too, if you code it well)

[10:15:00.0000] <Xanthir_>
Mine works in the absence of stylesheets. ^_^ You then just have the content on the page, and the link shoots you over to it.  That's the beauty of using :target.

[10:15:01.0000] <Xanthir_>
Then the Back button returns you to where you clicked.

[10:16:00.0000] <Xanthir_>
It just fails in the absence of :target, and you can deploy some feature-testing js to then pick up the slack.

[10:18:00.0000] <jgraham>
Xanthir: For the kind of use case I imagine, having the content on the page is not the right fallback behaviour

[10:19:00.0000] <Xanthir_>
Elaborate?

[10:20:00.0000] <jgraham>
E.g. a popup with a larger verion of a photo in

[10:20:01.0000] <Xanthir_>
Because you want to defer loading, right?

[10:27:00.0000] <Xanthir_>
Does anyone in the room have the release version of IE8 on hand?  If so, would you be so kind as to visit http://www.xanthir.com/css-modal.html and tell me if it works?

[10:40:00.0000] <Niictar24>
Xanthir_: IE 8 doesn't like your modal pop-up

[10:40:01.0000] <Xanthir_>
Indeed.  T_T

[10:41:00.0000] <Xanthir_>
But now that I know that IE is the only browser that doesn't, I can just drop a conditional comment on it delivering the same functionality via jquery.

[10:41:01.0000] <Niictar24>
I'm really glad there is just the one single browser that is consistently annoying in that regard

[10:42:00.0000] <Niictar24>
Of course there is a meriad of different support for different things across browsers

[10:42:01.0000] <Xanthir>
Yeah, but as long as you're not on the *bleeding* edge, you're typically all right.

[10:42:02.0000] <Niictar24>
But imagine if Firefox and Safari were as different as IE is to either of them

[10:43:00.0000] <Niictar24>
different to each other*

[10:45:00.0000] <Niictar24>
Xanthir: Yea, but it sucks when using HTML younger than 10 years old is considered bleeding edge :P

[10:47:00.0000] <Xanthir_>
Niictar24: Yeah, that would be horrifying.

[10:53:00.0000] <Xanthir_>
Woo, with the help of :target and the magic of hidden :checked elements, I have a strong tabbed interface mockup in pure CSS.


2009-04-20
[17:24:00.0000] <[SMQ]Mrpsycholo>
hello

[17:24:01.0000] <[SMQ]Mrpsycholo>
any major differences from xhtml? will there be compatibility issues or differences?

[17:25:00.0000] <jcranmer>
/me snickers

[17:25:01.0000] <jcranmer>
http://wiki.whatwg.org/wiki/HTML_vs._XHTML

[17:28:00.0000] <[SMQ]Mrpsycholo>
so document.write can be used in regular without using script in html5?

[17:29:00.0000] <[SMQ]Mrpsycholo>
will there be compatibility issues with html5 or will xhtml be more compatible across devices or browsers?

[17:29:01.0000] <[SMQ]Mrpsycholo>
safari recognized <audio> tag but firefox and ie8 didnt

[17:29:02.0000] <Hixie>
document.write will work the same in html5 as it has in the past

[17:29:03.0000] <jcranmer>
HTML is designed, in part, to represent current browser semantics

[17:30:00.0000] <Hixie>
in fact html5 in general is supposed to work the same as in the past, yeah

[17:30:01.0000] <jcranmer>
besides

[17:30:02.0000] <jcranmer>
FF nightlies have <audio> support

[17:30:03.0000] <Hixie>
so html5 will likely be more compatible than xhtml in general

[17:30:04.0000] <Hixie>
(IE doesn't do XHTML, e.g.)

[17:30:05.0000] <[SMQ]Mrpsycholo>
i made my site in xhtml

[17:30:06.0000] <[SMQ]Mrpsycholo>
becuase i was told it was a standard based form

[17:31:00.0000] <[SMQ]Mrpsycholo>
so i decided to use it and keep strict validation

[17:31:01.0000] <[SMQ]Mrpsycholo>
as a noob though, but i was udner impression xhtml was the future

[17:32:00.0000] <Hixie>
it was, for a while

[17:32:01.0000] <Hixie>
then it wasn't

[17:32:02.0000] <[SMQ]Mrpsycholo>
html5 will be lowercase?

[17:32:03.0000] <[SMQ]Mrpsycholo>
will it have a validation feature like xhtml?

[17:32:04.0000] <Hixie>
html5's tags are case-insensitive

[17:32:05.0000] <[SMQ]Mrpsycholo>
so they have to be uppercase?

[17:32:06.0000] <[SMQ]Mrpsycholo>
i like lowercase xhtml

[17:32:07.0000] <[SMQ]Mrpsycholo>
lol

[17:32:08.0000] <Hixie>
http://html5.validator.nu/

[17:32:09.0000] <Hixie>
no, they can be whatever case you prefer

[17:32:10.0000] <Hixie>
lowercase, uppercase, both, whatever

[17:33:00.0000] <[SMQ]Mrpsycholo>
will html5 require all tags to be closed like xhtml?

[17:33:01.0000] <[SMQ]Mrpsycholo>
<br />

[17:33:02.0000] <Hixie>
no, but it allows it

[17:33:03.0000] <jcranmer>
<A hReF="#asdf">

[17:33:04.0000] <Hixie>
<br> and <br/> are equivalent in html5

[17:33:05.0000] <[SMQ]Mrpsycholo>
nice its good you can use both :)

[17:34:00.0000] <[SMQ]Mrpsycholo>
and in html5 y ou dont have to specify type for script?

[17:34:01.0000] <Hixie>
nope

[17:34:02.0000] <Hixie>
<script> is fine

[17:34:03.0000] <[SMQ]Mrpsycholo>
of course css is still good

[17:35:00.0000] <[SMQ]Mrpsycholo>
www.q4supremacymod.com  is my site i am using as a learning tool :P

[17:36:00.0000] <[SMQ]Mrpsycholo>
so in html5 y ou wont have to use object tag? for video?

[17:38:00.0000] <Hixie>
for video you'll be able to use <video>

[17:38:01.0000] <[SMQ]Mrpsycholo>
nice 

[17:39:00.0000] <[SMQ]Mrpsycholo>
came across this from youtube saw a demo on html5 

[17:39:01.0000] <Hixie>
cool

[17:40:00.0000] <Hixie>
is that the "features you desperately want but can't yet use" video?

[17:40:01.0000] <[SMQ]Mrpsycholo>
yeah

[17:40:02.0000] <Hixie>
i'm the guy giving that presentation :-)

[17:40:03.0000] <[SMQ]Mrpsycholo>
the car with the laptop on the hood :)

[17:40:04.0000] <[SMQ]Mrpsycholo>
nice

[17:41:00.0000] <[SMQ]Mrpsycholo>
i will put a link on my site to www.whatwg.org :)

[17:41:01.0000] <Hixie>
:-)

[17:41:02.0000] <[SMQ]Mrpsycholo>
my site is a community for quakers and doomers 

[17:41:03.0000] <[SMQ]Mrpsycholo>
:)

[17:42:00.0000] <jcranmer>
and my site is an intranet where my latest addition was a theming style that derives from an april fools' day prank derived from google

[17:42:01.0000] <[SMQ]Mrpsycholo>
i dont think i saw that april fools one

[17:43:00.0000] <jcranmer>
well, it's an intranet

[17:43:01.0000] <jcranmer>
I don't think you'd see intranet april fools' day pranks

[17:43:02.0000] <[SMQ]Mrpsycholo>
ahh

[17:44:00.0000] <[SMQ]Mrpsycholo>
intranet being your personal local net not the global one?

[17:44:01.0000] <[SMQ]Mrpsycholo>
i should have went and studied computer science instead of psychology

[17:44:02.0000] <[SMQ]Mrpsycholo>
but my eye did catch intra vs inter

[17:44:03.0000] <jcranmer>
intranet = intra-net; intra- = within

[17:45:00.0000] <[SMQ]Mrpsycholo>
yeah latin idea

[17:45:01.0000] <[SMQ]Mrpsycholo>
endocrine exocrine

[17:45:02.0000] <[SMQ]Mrpsycholo>
internal secrete versus external

[17:48:00.0000] <[SMQ]Mrpsycholo>
so for <video src="firefox.ogg" controls autoplay></video>

[17:48:01.0000] <[SMQ]Mrpsycholo>
you dont need to separate with commas?

[17:48:02.0000] <[SMQ]Mrpsycholo>
controls and autoplay you can list them after the src?

[17:49:00.0000] <Hixie>
yup

[17:49:01.0000] <Hixie>
it's like with <input type=text value="test" readonly disabled>

[17:50:00.0000] <[SMQ]Mrpsycholo>
and some habits i have with xhtml i can still keep in html5

[17:51:00.0000] <[SMQ]Mrpsycholo>
its good i dont have to have the doctype in 5

[17:51:01.0000] <Hixie>
yeah just the short one

[17:51:02.0000] <Hixie>
<!DOCTYPE HTML>

[17:51:03.0000] <Hixie>
it's actually possible to memorise it!

[17:52:00.0000] <[SMQ]Mrpsycholo>
hehehe yeah 

[17:52:01.0000] <[SMQ]Mrpsycholo>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"

[17:52:02.0000] <[SMQ]Mrpsycholo>
  "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">

[17:52:03.0000] <[SMQ]Mrpsycholo>
 <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">

[17:52:04.0000] <[SMQ]Mrpsycholo>
is a bit of a pain lol

[17:53:00.0000] <[SMQ]Mrpsycholo>
i havent memorized it yet

[17:54:00.0000] <Rik`>
Hixie: is the lang attribute required on <html> ?

[17:54:01.0000] <Rik`>
oh, if i remember, <html> is not even required

[17:54:02.0000] <Hixie>
lang="" isn't required, but it's a good idea to put it in to declare the language, that way things like :lang() in CSS work

[17:55:00.0000] <Rik`>
or assistive technology can use it

[17:57:00.0000] <Hixie>
indeed

[17:57:01.0000] <[SMQ]Mrpsycholo>
php will be usable to? i am looking at the file open and save

[17:57:02.0000] <Rik`>
I'd like to see it required but it's maybe more a WCAG concern

[17:57:03.0000] <Hixie>
[SMQ]Mrpsycholo: html5 doesn't really affect php

[17:58:00.0000] <[SMQ]Mrpsycholo>
cause i wanted to have users to be able to upload files

[17:58:01.0000] <[SMQ]Mrpsycholo>
but my site doesnt do php with windows i guess

[17:58:02.0000] <Hixie>
Rik`: i don't think in practice there's a big enough disadvantage to not including it to warrant requiring it

[17:58:03.0000] <[SMQ]Mrpsycholo>
people in the community who have content could upload the content is what i wanted to do originally

[18:01:00.0000] <[SMQ]Mrpsycholo>
will html ever use // for comment lines? or /* */ ?

[18:02:00.0000] <[SMQ]Mrpsycholo>
that would interfere wouldnt it with what could be on the page?

[18:02:01.0000] <Rik`>
Hixie: accessibility concerns ? good practices ? but like I said, it's maybe not the role of html5

[18:03:00.0000] <Hixie>
Rik`: in practice, ATs don't need the language, they'll assume the user's language, which will almost always be right

[18:03:01.0000] <Hixie>
[SMQ]Mrpsycholo: too many pages already include those for us to do that i think

[18:04:00.0000] <[SMQ]Mrpsycholo>
do you guys have an icon?

[18:04:01.0000] <Rik`>
"almost always" ? i go to french and english websites, in the same proportions

[18:04:02.0000] <[SMQ]Mrpsycholo>
when i put the link on the page i could put an icon 

[18:05:00.0000] <Hixie>
[SMQ]Mrpsycholo: http://whatwg.org/images/

[18:05:01.0000] <Rik`>
so if i used ATs, half the time it would be wrong

[18:06:00.0000] <Hixie>
Rik`: true. though in those cases, basic heuristics could guess the language anyway.

[18:07:00.0000] <Rik`>
I don't know if they have this mechanism, I believe not from the last demo I had

[18:09:00.0000] <Rik`>
has this subject been discussed already ?

[18:12:00.0000] <Hixie>
Rik`: which, requiring language declarations?

[18:12:01.0000] <Rik`>
yep

[18:13:00.0000] <[SMQ]Mrpsycholo>
i appreciate the help i will defintiely read more on that demo site

[18:14:00.0000] <[SMQ]Mrpsycholo>
its good you also provided the text too :)

[18:15:00.0000] <[SMQ]Mrpsycholo>
any of you guys play quake? lol

[18:19:00.0000] <[SMQ]Mrpsycholo>
when you guys need a break from coding play this :P  http://www.youtube.com/watch?v=2hqWgRx7QgM :)

[18:25:00.0000] <Hixie>
Rik`: don't think so

[18:29:00.0000] <[SMQ]Mrpsycholo>
quakelive is the new version ported into a browser actually 

[18:29:01.0000] <[SMQ]Mrpsycholo>
i was impressed with how unreal tournament used web for control of a server in unreal tournament 3

[21:15:00.0000] <[SMQ]Mrpsycholo>
is it possible to make your own tags yet?

[21:16:00.0000] <[SMQ]Mrpsycholo>
document.getElementByTagName('div')   could you possibly define your own?

[21:23:00.0000] <Hixie>
[SMQ]Mrpsycholo: to make up your own tag, you pick a tag that is closest to what you want your tag to mean, and then you put your tag's name in the "class" attribute

[22:32:00.0000] <[SMQ]Mrpsycholo>
oh yeah that makes sense :)

[22:33:00.0000] <[SMQ]Mrpsycholo>
under type in css you would define it as a block right?

[22:33:01.0000] <Hixie>
depends what kind of presentation you want, but yes

[22:34:00.0000] <[SMQ]Mrpsycholo>
the most difficulty i have had so far is the positioning

[22:34:01.0000] <[SMQ]Mrpsycholo>
thats why i went with frameset

[22:35:00.0000] <[SMQ]Mrpsycholo>
trying that layout wiht css without using frames was such a pain for me

[22:35:01.0000] <[SMQ]Mrpsycholo>
it was easier just to define columsand rows 

[22:35:02.0000] <[SMQ]Mrpsycholo>
make several pages to accomplish it

[22:46:00.0000] <kanchan_Tripathi>
Hello to all can anybody please tell me how to post a form value using telnet

[22:46:01.0000] <kanchan_Tripathi>
??

[22:59:00.0000] <kanchan_Tripathi>
Hello to all can anybody please tell me how to post a form value using telnet??

[00:45:00.0000] <hsivonen>
I had dinner with friends who make software

[00:45:01.0000] <hsivonen>
two of them have done XML signatures

[00:45:02.0000] <hsivonen>
and said that one should avoid XML signatures if one wants interop

[00:48:00.0000] <MikeSmith>
hsivonen: why is that? because not enough implementations handle XML signatures, or because they don't handle them interoperably?

[00:51:00.0000] <hsivonen>
MikeSmith: because XML signatures depend on canonicalized serialization of XML and everyone has bugs in the canonicalizing serializer

[00:51:01.0000] <MikeSmith>
ah

[00:52:00.0000] <MikeSmith>
yeah, that seems like a weak design

[00:53:00.0000] <MikeSmith>
I wonder how many of the bugs are due to ambiguities in the Canonicalization spec?

[00:54:00.0000] <hsivonen>
as I understand it, it's mainly that UTF-8 is tough, Namespaces are tough and string interning is tough

[00:54:01.0000] <hsivonen>
all in all, serializing XML is tough

[00:57:00.0000] <MikeSmith>
so in the case of the Widgets work, it really doesn't seem like it's necessary to use XML signatures if something less complex and more robust will work instead

[00:58:00.0000] <hsivonen>
seems that way to me

[00:59:00.0000] <MikeSmith>
hsivonen: btw, I notice that the xhtml5core.rnc whattf driver doesn't toggle off the <mark> element the way that the html5core.rnc driver does

[00:59:01.0000] <MikeSmith>
do you recall of that's by design for some reason?

[01:00:00.0000] <hsivonen>
MikeSmith: seems like a bug

[01:00:01.0000] <MikeSmith>
OK

[01:01:00.0000] <MikeSmith>
I'll try to fix it in the patch I'm working on for generating the driver files

[01:14:00.0000] <MikeSmith>
hsivonen: I notice also that xhtml5core.rnc has "v5only = notAllowed" but html5core.rnc does not

[01:16:00.0000] <MikeSmith>
hsivonen: ah, I see that v5only is, among other things, used to toggle the <mark> element

[01:18:00.0000] <MikeSmith>
so seems like the 'include "phrase.rnc" { mark.elem = empty }' part is maybe not necessary

[01:18:01.0000] <MikeSmith>
if  html5core.rnc uses "v5only = notAllowed" instead too

[01:19:00.0000] <hsivonen>
MikeSmith: yeah, the bug is probably the other way round than first thought

[01:20:00.0000] <MikeSmith>
hsivonen: OK

[01:47:00.0000] <hsivonen>
annevk2: why do CSS namespaces allow @namespace url(http://www.w3.org/1999/xhtml); ?

[01:47:01.0000] <hsivonen>
annevk2: after all, the NS "URI" is a string and not an address

[01:47:02.0000] <hsivonen>
(probably too late to change now. I'm just curious.)

[01:50:00.0000] <annevk2>
because it's too late to change now

[01:51:00.0000] <annevk2>
CSS namespaces is from 99

[01:51:01.0000] <hsivonen>
annevk2: ok

[01:53:00.0000] <hsivonen>
http://www.w3.org/2005/Incubator/socialweb/charter

[01:53:01.0000] <hsivonen>
how come W3C Widgets are now a social Web topic?

[01:54:00.0000] <MikeSmith>
hsivonen: no clue from me

[01:54:01.0000] <MikeSmith>
I guess I should actually take time to read charters

[01:56:00.0000] <MikeSmith>
"The scope includes issues such as widget platforms (such as OpenSocial, Facebook and W3C Widgets).."

[01:58:00.0000] <MikeSmith>
I didn't even know that widget platforms were an "issue", let alone one related to the idea of social Web

[01:58:01.0000] <hsivonen>
I guess the kind of widgets that are embedded on Web pages are commonly used on Social Media Web 2.0 sites

[01:59:00.0000] <hsivonen>
but W3C Widgets aren't that kind of widgets

[02:11:00.0000] <jgraham>
There's discussion going on of ecmascript versioning on es-discuss, for those not following along at home

[02:18:00.0000] <hsivonen>
/me predicts that out of band version won't be adopted by authors.

[02:33:00.0000] <hendry>
/me discovers http://syntax.whattf.org/relaxng/

[02:34:00.0000] <MikeSmith>
hendry: don't get too used to that

[02:34:01.0000] <MikeSmith>
the schema drivers are likely going to disappear

[02:34:02.0000] <hendry>
MikeSmith: drivers?

[02:35:00.0000] <MikeSmith>
html5full.rnc and others

[02:35:01.0000] <MikeSmith>
those drivers are really intended to be used within the context of validator.nu

[02:36:00.0000] <MikeSmith>
and the legacy.rnc file in particular is intended to be used within the validator.nu context and not for schema-only checking

[02:36:01.0000] <hendry>
just thinking it would be nice to validate html5 locally with rnv and one of these schemas. some Debian Web apps (ikiwiki) want to validate generated HTML offline and rnv would be lightweight. validator.nu is too heavy to get running on a buld machine.

[02:37:00.0000] <zcorpan>
/me notes that 'text-shadow' does not support the effect used at http://syntax.whattf.org/

[02:37:01.0000] <MikeSmith>
hendry: yeah, I can understand that

[02:37:02.0000] <hsivonen>
hendry: if you come up with a regexp hack to replace the proper datatypes with rnv-compatible approximations, please share it

[02:37:03.0000] <hsivonen>
hendry: does rnv support XSD regexps?

[02:38:00.0000] <zcorpan>
oh the shadow has different text

[02:38:01.0000] <MikeSmith>
I don't think it does

[02:38:02.0000] <hendry>
XSD? i thought we don't use XSD :)

[02:38:03.0000] <MikeSmith>
I mean I don't think RNV supports pluggable datatypes

[02:39:00.0000] <MikeSmith>
hendry: it's debatable whether we want to be doing anything to facilitate schema-only validation outside of the context of a complete conformance checker

[02:41:00.0000] <hsivonen>
hendry: we don't use XSD. However, the XSD regexps aren't all bad, although they have surprising quirks.

[02:41:01.0000] <hendry>
i prefer a 'less is more' approach. or i want a non-complete lightweight conformance checker.

[02:42:00.0000] <hsivonen>
hendry: so if you don't port the HTML5 datatype library for rnv and rnv supports XSD regexps, you are better off using regexps than no attribute value checking at all

[02:43:00.0000] <hsivonen>
hendry: I'd expect it to be feasible to run a perl or python script over the schemas so that the proper datatypes are trashed and the commented regexps uncommented

[02:44:00.0000] <MikeSmith>
hendry: outside of the datatype issue, you'll still be able to do that kind of checking with the schemas, by taking the schema modules and writing your own drivers

[02:47:00.0000] <annevk2>
One problem with ECMAScript versions is that Mozilla uses its own versioning stuff for JavaScript. E.g. JavaScript 1.7. (Not too different from UA=IE8...) The other of course that the specification does not actually document how to react to different version identifiers and such.

[02:47:01.0000] <hendry>
MikeSmith: sorry, bit confused by terminology; so a driver = .rnc = RelaxNG Schema ?

[02:48:00.0000] <hsivonen>
hendry: driver as in XHTML2 WG Modularization speak :-)

[02:48:01.0000] <hsivonen>
hendry: a schema that just imports other schema files

[02:49:00.0000] <hendry>
hsivonen: and i'm confused about you talk of regexps. were you saying you could not to use rnv and use regexps to validate?

[02:49:01.0000] <MikeSmith>
hendry: the drivers are the files there that don't define any content models themselves, but just import the other schema files that have the actual definitions

[02:49:02.0000] <hendry>
MikeSmith: gotcha

[02:50:00.0000] <MikeSmith>
http://www.davidashen.net/rnv.html says:

[02:50:01.0000] <MikeSmith>
"Besides built-in support for XML Schema datatypes, RNV provides two different ways to add user-defined datatype libraries."

[02:51:00.0000] <hsivonen>
hendry: I'm saying that it'll be easier for you to approximate the HTML5 datatype with XSD regexps than to write a an implementation of the HTML5 Datatype Library in such a way that it can be used with RNV

[02:52:00.0000] <annevk2>
you in the US hsivonen?

[02:53:00.0000] <hsivonen>
annevk2: not yet

[02:53:01.0000] <hsivonen>
annevk2: flying on Saturday

[02:53:02.0000] <hendry>
hsivonen: ok, i grok .rncs, though i have no real exp with 'XSD regexps'. do you have XSD examples i could peer at? and what tool uses them? xmlstarlet?

[02:53:03.0000] <hendry>
hsivonen: San Francisco? :-)

[02:53:04.0000] <hsivonen>
hendry: there are commented-out examples in common.rnc

[02:53:05.0000] <hsivonen>
hendry: first Mountain View, then Alexandria, VA

[02:56:00.0000] <hendry>
MikeSmith: so you think rnv can support XDF with an external command or something? what tool grok XSD regexps then?

[02:56:01.0000] <hendry>
s/XDF/XSD/

[02:56:02.0000] <hendry>
hsivonen: what's in Alexandria?

[02:57:00.0000] <MikeSmith>
hendry: to use that feature in rnv, I think you have to roll the external command yourself

[02:57:01.0000] <MikeSmith>
or use RNV's alternate support for expressing the datatype-checking in scheme

[02:57:02.0000] <MikeSmith>
if you know scheme or lisp already, that might be a good way to go

[02:58:00.0000] <MikeSmith>
or a useful project for somebody who does know lisp or scheme well

[02:58:01.0000] <hsivonen>
hendry: friend's wedding

[03:01:00.0000] <hendry>
MikeSmith: i don't know lisp well. i didn't like my dr scheme class and University. or emacs. drat.

[03:01:01.0000] <hendry>
s/and/at :-) University was great

[03:02:00.0000] <MikeSmith>
hendry: anyway, there are some examples and more info in the README that's distributed with rnv

[03:04:00.0000] <hendry>
MikeSmith: yup. I see them.

[03:41:00.0000] <hsivonen>
Philip`: I was able to find and fix the bug you reported about parsing http://www.giftology.co.uk

[03:42:00.0000] <hsivonen>
Philip`: I'm unable to reproduce the bugs about http://www.villatraining.ca/ and http://www.autobanga.lt/

[04:20:00.0000] <hsivonen>
Philip`: do you have additional clues on how to track down http://krijnhoetmer.nl/irc-logs/whatwg/20090401#l-410 ?

[04:30:00.0000] <hsivonen>
"Use aria-label only if the interface is such that it is not possible to have a visible label on the screen"

[04:30:01.0000] <hsivonen>
I wonder how strict "possible" is there

[04:33:00.0000] <hsivonen>
/me wonders how aria-label use cases differ from title attribute use cases

[04:34:00.0000] <zcorpan>
/me wonders why one can't use css positioning to hide labels off-screen

[04:35:00.0000] <zcorpan>
hsivonen: have you had a look at http://philip.html5.org/data/table-implicitly-closed-p.txt ?

[04:36:00.0000] <hsivonen>
zcorpan: only took a look at a tiny sample so far

[04:36:01.0000] <hsivonen>
zcorpan: tiny as in 10

[04:37:00.0000] <zcorpan>
hsivonen: maybe it's possible to set up a script that takes a screenshot using the old moz parser and using the html5 parser and then comparing the screenshots

[04:43:00.0000] <hsivonen>
zcorpan: I guess the reftest framework could be changed to do something like that

[04:44:00.0000] <hsivonen>
why did ARIA hois aria-invalid to a global state???

[04:44:01.0000] <hsivonen>
*hoist

[04:44:02.0000] <hsivonen>
and required

[05:22:00.0000] <Philip`>
hsivonen: Hmm, that might just be because I was using a really old version

[05:22:01.0000] <Philip`>
(1.0.7 apparently)

[05:22:02.0000] <hsivonen>
Philip`: oh. I thought you were using my special Philip` build

[05:23:00.0000] <Philip`>
hsivonen: Oh, maybe I was

[05:24:00.0000] <Philip`>
hsivonen: I can try testing it some time later, if I don't forget

[05:34:00.0000] <hsivonen>
Philip`: thanks

[06:00:00.0000] <hsivonen>
how does JAWS 10 get the landmark roles from a browser?

[06:11:00.0000] <annevk2>
if they don't map to an accessibility API direcltly I believe accessibility APIs (maybe not all) have some way to store unknown roles

[06:12:00.0000] <annevk2>
are there no posts on the benefits of e.g. using role=navigation in current screen readers?

[06:12:01.0000] <annevk2>
seems quite logical that some would have toyed with it already and posted results

[06:17:00.0000] <jgraham>
http://www.paciellogroup.com/blog/misc/landmarks.html

[06:20:00.0000] <jgraham>
/me makes a note to self that evolving the language seperatly from the implementation requirements is bad beacuse feedback from implementation can't affect the language design

[06:21:00.0000] <annevk2>
I'm thinking that dealing with screen readers must suck. I hate learning shortcuts but there you'd have to learn whole sets of words and their meaning to figure out what the screen reader is talking about.

[06:21:01.0000] <Philip`>
annevk2: That's probably why most people aren't voluntarily blind

[06:23:00.0000] <annevk2>
You think better software is holding them back?

[06:23:01.0000] <Philip`>
(Well, not specifically because of that issue with screen readers, but the general issue of it making life much harder)

[06:23:02.0000] <annevk2>
So only JAWS supports them.

[06:24:00.0000] <jgraham>
I guess it is interesting that there is no common use of audio to supplement the browsing experience

[06:24:01.0000] <jgraham>
Presumably that means that visuasl input is so high bandwidth that trying to add audio would just cause blocking

[06:25:00.0000] <Philip`>
jgraham: That's not true - a large number of pages use audio to provide information such as "this page was designed in the 90s" to their readers

[06:26:00.0000] <Philip`>
Also there's the click sound when you click links (at least in IE, at some point in the past)

[06:27:00.0000] <annevk2>
hsivonen, "IO Error: <http://annevankesteren.nl/2009/04/html5-wai-aria#comment-6768> Code: 59/PROHIBITED_COMPONENT_PRESENT in FRAGMENT: A component that is prohibited by the scheme is present." is annoying

[06:27:01.0000] <annevk2>
hsivonen, just ignore the fragment (maybe say you did if you really want to)

[06:27:02.0000] <jgraham>
Philip`: I'm not sure author-supplied midi files count :)

[06:33:00.0000] <annevk2>
"I worked hard to get a good list of landmarks from the PF group, but what it ended up as was spec writers deciding what users should want." -- https://bugzilla.mozilla.org/show_bug.cgi?id=459395

[06:34:00.0000] <annevk2>
What is it again they criticize the WHATWG for?

[06:41:00.0000] <annevk2>
http://wiki.whatwg.org/index.php?title=MetaExtensions&diff=3626&oldid=3467 o_O

[06:44:00.0000] <Lachy>
woah. That's a lot of proposals

[06:45:00.0000] <Lachy>
I thought he only added the useless addmark and addmarklocal extensions to address the non-existent problem he imagined in bug 6774

[06:45:01.0000] <annevk2>
it's from the guy "spamming" W3C bugzilla

[06:46:00.0000] <Lachy>
yeah, I realised that

[06:47:00.0000] <annevk2>
hmm, per https://bugzilla.mozilla.org/show_bug.cgi?id=459395 browsers don't support ARIA roles. It's the AT being creative

[06:48:00.0000] <Lachy>
dammit, my email on dreamhost is down. That explains why I haven't received anything for a few hours :-(

[07:07:00.0000] <zcorpan>
annevk2: s/roles/landmarks/

[07:08:00.0000] <zcorpan>
annevk2: the value of the role="" attribute is, in firefox, always exposed as-is via xml-roles

[08:10:00.0000] <takkaria>
I think that Nick guy thinks the mark element is going to be used like MS' smart tags were going to be

[08:11:00.0000] <takkaria>
but I'm not quite sure where he thinks the third-party servers are going to come in to it from

[08:26:00.0000] <Philip`>
takkaria: Maybe the third-party servers are like Google's cache servers, that serve pages and insert extra markup to highlight search term matches?

[08:27:00.0000] <Philip`>
/me assumes this is related to bug 6774, but he only skimmed the first half of the first comment before getting terminally bored

[08:31:00.0000] <jgraham>
If you got terminally bored, are you posting from beyond the grave?

[08:36:00.0000] <Philip`>
It was terminal to my reading of the bug, not to my life

[08:38:00.0000] <annevk2>
zcorpan, yeah, that's what I meant; hoped it was clear that the statement was scoped by the bug; guess not :)

[08:45:00.0000] <jgraham>
http://socialcollider.net/ is a kind of interesting way of killing browsers

[08:45:01.0000] <jgraham>
(because it's so slow)

[08:46:00.0000] <jgraham>
(especially in gecko)

[08:48:00.0000] <annevk2>
hsivonen, your HTML parser is not used for everything in Gecko?

[09:37:00.0000] <zcorpan>
hsivonen: still appears to be crashy for me

[09:52:00.0000] <zcorpan>
hsivonen: http://software.hixie.ch/utilities/js/live-dom-viewer/saved/87

[09:53:00.0000] <annevk2>
whoa, Firefox 3.6something uppercases ö to Ö

[09:56:00.0000] <gsnedders>
WebKit does too

[09:56:01.0000] <gsnedders>
(Saf4b, this is)

[10:29:00.0000] <gsnedders>
/me is trying to get lyrics from music

[10:30:00.0000] <gsnedders>
/me supposes he could just ask the vocalist

[10:34:00.0000] <gsnedders>
/me is listening to Epic Fail by Hit the Breaks

[10:34:01.0000] <Philip`>
You could try listening to it and writing down the words

[10:34:02.0000] <gsnedders>
I expect a lot of you would find this amusing.

[10:34:03.0000] <gsnedders>
http://www.myspace.com/hitthebrakesno5

[10:48:00.0000] <gsnedders>
Philip`: I was trying to listen to it and write down the world

[10:48:01.0000] <gsnedders>
*words

[12:30:00.0000] <nitroball>
Hey, wassup? Anyone try using the canvas tag with the Wii brower? It works very well except the color is missing for me on the lines.

[12:31:00.0000] <nitroball>
Color does show up in the IE, Opera, Firefox, etc.

[12:46:00.0000] <weinig>
annevk2: ping

[13:01:00.0000] <nitroball>
I think I figured it out... ctx.lineStyle = "rgba(255,165,0,1)"; doesn't work on the Wii browser. ctx.lineStyle = "rgb(255,165,0)"; does

[13:01:01.0000] <nitroball>
its the alpha value that was turning everything black

[13:51:00.0000] <Hixie>
so i'm going through the rdf folder looking for actual use cases and requirements

[13:51:01.0000] <Hixie>
99% of the e-mails have absolutely nothing to do with use cases, despite the e-mail i sent saying that that's all i was looking for

[13:51:02.0000] <Hixie>
of the remainder, there are use cases like "provide information useful for improving the accessibility of Web content"

[13:52:00.0000] <Hixie>
which are the most vague, under-defined, and useless use case descriptions in the history of mankind

[15:06:00.0000] <gsnedders>
Hixie: Time to beat that: "to improve the long-term direction of the human race"

[15:08:00.0000] <Dashiva>
gsnedders: Can you prove that axiomatically?

[15:08:01.0000] <gsnedders>
Proof? Peh.

[15:30:00.0000] <gsnedders>
http://www.tesco.com/books/search.aspx?Ntt=%EF%BF%BF&Ntk=primary&VSI=1&Ntx=mode%2Bmatchall&Nty=1&N=0

[15:43:00.0000] <gsnedders>
/me has interview 8am tomorrow morning, yay


2009-04-21
[19:26:00.0000] <Hixie>
/me is half-way through 14000 lines' worth of text on microdata in html5, and has so far found 296 lines' worth of use cases, requirements, scenarios, and problem descriptions

[19:26:01.0000] <Hixie>
(of which i wrote about half myself)

[19:27:00.0000] <Hixie>
(and 71 of which are blank lines)

[00:57:00.0000] <gsnedders>
/me has interview any minute…

[00:59:00.0000] <MikeSmith>
NicolasRaoul: http://hsivonen.iki.fi/testing-html5-parsing/

[01:02:00.0000] <MikeSmith>
hsivonen: is there any way I can tell Minefield to automatically look for updates from the try server instead of wherever it normally looks?

[01:04:00.0000] <hsivonen>
MikeSmith: I don't know. Does it look for updates at all?

[01:04:01.0000] <hsivonen>
AFAIK, tryserver builds don't look for updates on the tryserver.

[01:04:02.0000] <MikeSmith>
OK

[01:05:00.0000] <MikeSmith>
I think Minefield itself normally checks for updates

[01:06:00.0000] <hsivonen>
nightly builds do update to later nightly builds

[01:12:00.0000] <MikeSmith>
hsivonen: about the statement "Other HTML parsing operations still use the old parser unconditionally."

[01:12:01.0000] <MikeSmith>
what other parsing operations are there?

[01:12:02.0000] <hsivonen>
off the top of my head:

[01:13:00.0000] <hsivonen>
feed payload sanitizing

[01:13:01.0000] <hsivonen>
Netscape bookmarks import

[01:13:02.0000] <hsivonen>
HTML clipboard import

[01:13:03.0000] <hsivonen>
text/plain loading

[01:13:04.0000] <hsivonen>
view source syntax highlighting

[01:14:00.0000] <hsivonen>
I wouldn't be at all surprised if there were other cases lurking somewhere

[01:14:01.0000] <hsivonen>
DOMParser and XHR2 will introduce new cases

[01:15:00.0000] <MikeSmith>
OK, I see

[01:17:00.0000] <MikeSmith>
hsivonen: so in my about:config, html5.enable is not showing up as an option

[01:18:00.0000] <MikeSmith>
I just installed from hsivonen⊙ifd

[01:19:00.0000] <MikeSmith>
hsivonen: is there any test I can do to make sure it's actually using the HTML5 parse?

[01:19:01.0000] <MikeSmith>
parser

[01:19:02.0000] <gsnedders>
/me takes a deep breath

[01:19:03.0000] <gsnedders>
My hand is sore from holding the phone so tight during the interview…

[01:19:04.0000] <MikeSmith>
gsnedders: interview?

[01:20:00.0000] <gsnedders>
MikeSmith: Summer internship, Opera

[01:20:01.0000] <hsivonen>
MikeSmith: http://hsivonen.iki.fi/test/moz/html5-parsing.html

[01:20:02.0000] <hsivonen>
MikeSmith: if you see proper math and the video has a clip path

[01:22:00.0000] <MikeSmith>
/me checks

[01:22:01.0000] <MikeSmith>
wow, cool

[01:23:00.0000] <MikeSmith>
hsivonen: thanks

[01:23:01.0000] <MikeSmith>
btw, I notice that the blur also blurs the default video controls

[01:24:00.0000] <hsivonen>
I guess it's a feature :-)

[01:24:01.0000] <MikeSmith>
heh

[01:24:02.0000] <MikeSmith>
I wonder if it's also a feature that it doesn't let me actually use the controls

[01:25:00.0000] <MikeSmith>
if I mouse over the video, the controls pop up, but then I when I most down to try to click, they disappear again

[01:26:00.0000] <hsivonen>
hmm. that used to work some builds ago

[01:26:01.0000] <MikeSmith>
hsivonen: anyway, fwiw, html5.enable pref is definitely not showing up in the OSX environment I'm trying it on

[01:26:02.0000] <hsivonen>
I wonder if it is a bug or a clickjacking prevention feature

[01:27:00.0000] <hsivonen>
MikeSmith: you need to add the pref

[01:27:01.0000] <MikeSmith>
d'oh

[01:27:02.0000] <MikeSmith>
OK

[01:27:03.0000] <MikeSmith>
gsnedders: about my tweet, I don7t think thy are using Unicode at all, but whatever the normal mainland Chinese encoding is

[01:28:00.0000] <MikeSmith>
or some subset of it

[01:28:01.0000] <MikeSmith>
or some font that doesn't have all the glyphs

[01:28:02.0000] <MikeSmith>
or something

[01:28:03.0000] <hsivonen>
MikeSmith: PRC national standard maps to Unicode, IIRC

[01:28:04.0000] <hsivonen>
GB18030 IIRC

[01:28:05.0000] <MikeSmith>
OK

[01:29:00.0000] <hsivonen>
it would have been madness to make it not fully mappable

[01:29:01.0000] <hsivonen>
fortunately that was situation was avoided

[01:30:00.0000] <MikeSmith>
it may be that the characters in question are neither in Unicode nor in GB18030

[01:31:00.0000] <MikeSmith>
that's probably it, actually

[01:31:01.0000] <MikeSmith>
I know that's the case for some name characters in use in Japan

[01:31:02.0000] <MikeSmith>
variant kanji

[01:31:03.0000] <MikeSmith>
so-called gaiji

[01:32:00.0000] <hsivonen>
"Like UTF-8, GB18030 is a superset of ASCII and can represent the whole range of Unicode code points; in addition, it is also a superset of GB2312." says Wikipedia

[01:32:01.0000] <MikeSmith>
OK

[01:34:00.0000] <hsivonen>
/me still hasn't found out the technical specs for personal names in Finland

[01:35:00.0000] <MikeSmith>
I've tried to find an estimated of how many not-in-Unicode gaiji for personal names there are in use in Japan, but nobody seems to know

[01:36:00.0000] <MikeSmith>
estimate of

[01:36:01.0000] <MikeSmith>
it's sort of a long tail, I guess

[01:39:00.0000] <hsivonen>
I think the tone of the NYT article is a bit hypocritical considering that the U.S. requires European to coerce their names to ASCII on DHS forms

[01:39:01.0000] <hsivonen>
*Europeans

[01:41:00.0000] <MikeSmith>
I didn't know the US required that

[01:41:01.0000] <MikeSmith>
can't say that I'm surprised, though

[01:45:00.0000] <MikeSmith>
hsivonen: the problem with that is, I seem to remember reading something a while back that said US was requiring people to also have boarding passes that exactly match the name as it it on their passports

[01:46:00.0000] <hsivonen>
fortunately, I have an ASCII-compatible name

[01:47:00.0000] <MikeSmith>
me too

[01:47:01.0000] <MikeSmith>
even including the parenthesis

[01:52:00.0000] <MikeSmith>
hsivonen: so are there any pages that currently have problem behavior if I use the HTML5-enabled build with them?

[01:52:01.0000] <hsivonen>
MikeSmith: I think there may be problems around meta refresh and HTML5 application manifests

[01:53:00.0000] <MikeSmith>
OK

[02:19:00.0000] <hsivonen>
sigh. aria-dropeffect now has even more possible tokens

[02:20:00.0000] <hsivonen>
one would think the tokens were mutually exclusive...

[02:34:00.0000] <jgraham>
hsivonen: AFAICT the only choice that makes sense to combine with anything else is execute since that seems to be a different way of saying "other"

[02:37:00.0000] <hsivonen>
it seems to me that the principle applied to ARIA is the exact reverse of what the PFWG applied to alt:

[02:37:01.0000] <hsivonen>
if there is a semiconceivable possibility of a combination maybe making sense, allow it

[02:45:00.0000] <Hixie>
hsivonen: it's different people, isn't it?

[02:46:00.0000] <hsivonen>
Hixie: I don't know. I haven't done Member-only mailing list archeology to find out who wanted what.

[02:46:01.0000] <Hixie>
ah ok

[02:48:00.0000] <hsivonen>
what's the point of having values "true", "false", "undefined" where "undefined" is equivalent to not specifying the attribute?

[02:53:00.0000] <annevk2>
that's what you get when you define things in terms of markup rather than states I suppose

[02:54:00.0000] <annevk2>
it would be nice if for consistency the empty string mapped to true as well

[02:56:00.0000] <hsivonen>
as far as I can tell, the empty string is equivalent to "undefined"

[03:00:00.0000] <jgraham>
<span aria-checked> == <span aria-checked="undefined">? Nice...

[03:05:00.0000] <annevk2>
aria-checked=false actually being false is somewhat inconsistent already

[03:07:00.0000] <hsivonen>
whoa! the new values for aria-invalid are, umm, interesting

[03:07:01.0000] <hsivonen>
"true", "false", "grammar", "spelling"

[03:07:02.0000] <zcorpan>
o_O

[03:08:00.0000] <MikeSmith>
beautiful

[03:08:01.0000] <MikeSmith>
I'm pretty sure it's some kind of surrealist put-on

[03:09:00.0000] <jgraham>
Submit it to the Daily WTF?

[03:11:00.0000] <hsivonen>
hmm. "rude" is no longer supported for aria-live :-(

[03:11:01.0000] <annevk2>
I was hoping I did not have to review the actual vocabulary

[03:13:00.0000] <annevk2>
"User agents may refuse to submit the form as long as there is an element for which aria-invalid is true."

[03:13:01.0000] <hsivonen>
the authoring requirements of aria-posinset and aria-level still depend on things that aren't well-defined

[03:13:02.0000] <annevk2>
Same for aria-labelledby

[03:13:03.0000] <annevk2>
I thought the idea was that there would be no user agent requirements other than exposing information correctly to assistive technology?

[03:14:00.0000] <hsivonen>
annevk2: are you going to send feedback for that form submission thing?

[03:14:01.0000] <annevk2>
/me sighs

[03:15:00.0000] <jgraham>
/me adores the note for aria-labelledby that explains that they have chosen a word with known ambiguous spelling to "minimize the difficulty for developers"

[03:16:00.0000] <annevk2>
ouch

[03:18:00.0000] <annevk2>
hsivonen, the deadline for comments was last Friday I believe; I wonder if any new comments would be in vain

[03:20:00.0000] <hsivonen>
annevk2: depends on whether they want to make to REC fast or whether they want a proper spec.

[03:21:00.0000] <hsivonen>
annevk2: also, does Process require anything about a DL or just about things happening before the next transition?

[03:21:01.0000] <hsivonen>
it's still unclear to me, how ARIA can make it to REC with two interoperable implementations

[03:22:00.0000] <hsivonen>
is there another criterion for transitioning to REC?

[03:22:01.0000] <hsivonen>
or do desks count as implementations

[03:22:02.0000] <annevk2>
Looking at PF minutes I'm not sure they're addressing my comments in a sensible way. Mu

[03:23:00.0000] <annevk2>
The XHTML2 WG uses schema "implementations" as implementations so I'm pretty sure anything goes.

[03:23:01.0000] <annevk2>
Of course, XHTML2 is not something that makes it into browsers so that is not really a concern, but WAI-ARIA is somewhat. (Although it remains to be seen how good it will work.)

[03:24:00.0000] <jgraham>
This is why I have given up taking the time to review aria. The process seems set up to produce bad results

[03:24:01.0000] <hsivonen>
huh? does Google Groups intentionally require login to browse?

[03:26:00.0000] <zcorpan>
at some point, the aria spec referenced html5 for the definition of "whitespace"

[03:26:01.0000] <zcorpan>
then, the reference was removed because aria was going to rec before html5 was

[03:26:02.0000] <zcorpan>
so whitespace is now undefined

[03:26:03.0000] <MikeSmith>
I think if the ARIA spec includes testable assertions about UA requirements, the normal criteria for it transitioning beyond CR would need to be that there are at least two interoperable implementations that meet those requirements

[03:27:00.0000] <MikeSmith>
but I don't think the process doc has any specific language stating that

[03:27:01.0000] <hsivonen>
MikeSmith: I think any browser since circa IE4 or Netscape 6 is a conforming implmentation, then

[03:29:00.0000] <MikeSmith>
/me looks back and sees annevk2's quote "User agents may refuse to submit the form as long as there is an element for which aria-invalid is true."

[03:29:01.0000] <MikeSmith>
but I see also that's a "may"

[03:31:00.0000] <jgraham>
/me wonders if the requirement "User agents MUST implement all supported states and properties for the role" is as silly as it sounds

[03:32:00.0000] <hsivonen>
MikeSmith: even exposure to AT is only a "should"!

[03:32:01.0000] <zcorpan>
http://lists.w3.org/Archives/Public/public-xml-core-wg/2009Apr/0031.html

[03:32:02.0000] <jgraham>
"Assistive technology must treat an article like a document in that article must be processed like an application" makes no sense

[03:34:00.0000] <jgraham>
"When a value is indicated as the default, the behavior prescribed by this value MUST be followed when the state or property is not provided" - sounds almost like a conformance criterion

[03:37:00.0000] <jgraham>
I wonder if I can claim that section 6.1 requires two host langauges to implement @role, both of which must presumably reach REC before aria :)

[03:40:00.0000] <zcorpan>
jgraham: svg 1.2 tiny could be one of them

[03:41:00.0000] <hsivonen>
can SVG 1.2 tiny be one of them the same way HTML 4.01 can be one of them?

[03:42:00.0000] <zcorpan>
hsivonen: no, i think svg 1.2 tiny actually includes role=""

[03:42:01.0000] <hsivonen>
oh

[03:42:02.0000] <zcorpan>
but not the other aria-*

[03:42:03.0000] <hsivonen>
also, IIRC, SVG 1.2 Tiny can time-warp the REC track :-)

[03:45:00.0000] <annevk2>
jgraham, are you sending in comments?

[03:46:00.0000] <hsivonen>
annevk2: I sent a comment about form submission

[03:46:01.0000] <jgraham>
annevk2: No. Should I? I don't have any faith that they will be dealt with in a sensible way but if you think it is a good idea I will

[03:47:00.0000] <annevk2>
jgraham, yeah, lets give it the benefit of the doubt

[03:47:01.0000] <jgraham>
OK

[03:51:00.0000] <jgraham>
Has someone already commented on the use of must vs MUST?

[03:53:00.0000] <MikeSmith>
zcorpan: that message from Paul Grosso about avoiding "must" is disappointing

[04:02:00.0000] <annevk2>
/me isn't even sure whether the behavior of aria-invalid="false spelling" is defined or aria-invalid="false x" for that matter

[04:03:00.0000] <zcorpan>
is the behavior of aria-invalid="false" defined? :)

[04:03:01.0000] <annevk2>
yes

[04:03:02.0000] <zcorpan>
ok

[04:04:00.0000] <annevk2>
"If the attribute is not present, or its value is false, or its value is an empty string, the default value of false applies."

[04:04:01.0000] <jgraham>
Does section 4.2.2 make any sense to anyone?

[04:05:00.0000] <annevk2>
jgraham, not a whole lot

[04:06:00.0000] <jgraham>
/me is starting to understand little

[04:07:00.0000] <jgraham>
s//a/

[04:11:00.0000] <zcorpan>
hsivonen: http://alfonsopina.eresmas.net/ looks weird in minefield-html5

[04:24:00.0000] <hsivonen>
zcorpan: just that the heading isn't centered?

[04:24:01.0000] <zcorpan>
hsivonen: no, i get text like "�������t hesitate to contact the author (see end of page).', TEXTFONT, 'Comic Sans MS', TEXTCOLOR, '#FFFFFF', FGCOLOR, '#3366ff') " onMouseOut="nd();">"

[04:25:00.0000] <hsivonen>
hmm. that's bad

[04:26:00.0000] <zcorpan>
and "Libro de visitas Ver   &n�����a) "

[04:26:01.0000] <hsivonen>
zcorpan: what's your character encoding setting?

[04:27:00.0000] <zcorpan>
hsivonen: UTF-8

[04:27:01.0000] <hsivonen>
if I switch to UTF-8, the result is totally bogus, indeed

[04:27:02.0000] <hsivonen>
I think I don't understand the Gecko unicode converter API contract right

[04:27:03.0000] <hsivonen>
zcorpan: thanks

[04:28:00.0000] <hsivonen>
I still wonder why the heading/header isn't centered

[04:28:01.0000] <zcorpan>
hsivonen: maybe because of <p><table>? i tested a random url from the list

[04:29:00.0000] <zcorpan>
hsivonen: if i select another encoding or check autodetect, it crashes

[04:29:01.0000] <hsivonen>
zcorpan: indeed, it's a <p><table> issue.

[04:29:02.0000] <hsivonen>
<p align=center>

[04:41:00.0000] <zcorpan>
/me wonders if it's possible to fix the styling issues in the parser by copying over attributes to the table

[04:42:00.0000] <hsivonen>
I wonder if it's possible not to close <p> if it has the align attribute

[04:42:01.0000] <hsivonen>
and whether that's even worse than having one quirk

[04:42:02.0000] <zcorpan>
oh yeah that's probably better than copying attributes

[04:46:00.0000] <zcorpan>
<P ALIGN=RIGHT><TABLE - http://123relax.holisticwebdirectory.com/drhypnosis/

[04:51:00.0000] <zcorpan>
thinking about it, i wonder whether <p><form> would benefit from the same treatment

[05:00:00.0000] <zcorpan>
http://64.8.148.68/webpage/highschool.htm - <p align=center>...<table

[05:01:00.0000] <annevk2>
is it just <p align>?

[05:02:00.0000] <zcorpan>
from what i've looked at so far, yeah

[05:03:00.0000] <zcorpan>
<p align="center"> http://www.shopin.co.il/

[05:06:00.0000] <zcorpan>
<p class="topbana"><table border="0" width="" cellspacing="0" cellpadding="3" style="width: 583px"> http://www.ace-company.net/~coolcat/

[05:07:00.0000] <zcorpan>
.topbana   { margin-top: 5; margin-bottom: 12 }

[05:07:01.0000] <zcorpan>
nothing too serious i guess

[05:08:00.0000] <annevk2>
it would regress layout of the page

[05:08:01.0000] <hsivonen>
maybe we should just fix Acid2 instead

[05:10:00.0000] <annevk2>
we could do that I suppose, but all UAs interoperate on this now

[05:10:01.0000] <hsivonen>
annevk2: yeah. I blame Hixie.

[05:12:00.0000] <annevk2>
we almost had this with comments too

[05:12:01.0000] <annevk2>
http://ln.hixie.ch/?start=1137799947&count=1

[05:13:00.0000] <hsivonen>
I was thinking of that blog post, too

[05:17:00.0000] <zcorpan>
<p>&nbsp; <table> http://adsg.syix.com/ - would get more space above the table

[05:18:00.0000] <zcorpan>
http://agri.astate.edu/Ag%20Education/education.htm too

[05:21:00.0000] <zcorpan>
http://alexey-savrasov.ru/ has <p align=justify> blah blah <table align=right>...</table> blah blah

[05:22:00.0000] <hsivonen>
ooh. I now see my character decoding bug

[05:29:00.0000] <zcorpan>
ok i've looked at the 50 first urls

[05:30:00.0000] <zcorpan>
i think our options are (1) having the quirk, and (2) changing acid2

[05:30:01.0000] <hsivonen>
zcorpan: I agree

[05:31:00.0000] <zcorpan>
looking at align="" is not enough

[05:33:00.0000] <hsivonen>
I wonder if we could get rid of the quirks mode these days if CSS 1) defaulted to px for unitless numbers and 2) defaulted to almost standards cell height

[05:33:01.0000] <hsivonen>
perhaps inheriting into tables should have been author opt-in via CSS itself way back when

[05:34:00.0000] <annevk2>
case-insensitive class names are important too I think and there are probably a few more

[05:35:00.0000] <zcorpan>
hsivonen: i'm pretty sure people use unitlessness as a css filter

[05:35:01.0000] <annevk2>
case-insensitive class names are not scoped to HTML by the way at this point

[05:35:02.0000] <zcorpan>
hsivonen: to fix the ie5 box model

[05:35:03.0000] <annevk2>
though actually getting rid of quirks mode would be awesome

[05:36:00.0000] <hsivonen>
it seems to me that after IE3 none of the CSS quirks have been so horrible that CSS2 developed according to HTML5 principles couldn't have had backwards-compatible default behavior

[05:36:01.0000] <hsivonen>
too late now, I guess

[05:36:02.0000] <hsivonen>
hindsight 20/20

[05:36:03.0000] <hsivonen>
the attitude towards 'standards' was very different back then

[05:37:00.0000] <annevk2>
https://twitter.com/erikdahlstrom/status/1574170590 sweet

[05:37:01.0000] <annevk2>
guess that's next on my list of things to change on my site

[05:38:00.0000] <hsivonen>
/me notes that ed doesn't even mention IE

[06:07:00.0000] <hsivonen>
zcorpan, Philip`: thanks for doing the <p><table> research

[06:14:00.0000] <Philip`>
annevk2: It's a shame that Opera doesn't support APNG favicons properly, and only Mozilla does

[06:17:00.0000] <zcorpan>
Philip`: have you filed a bug?

[06:19:00.0000] <Philip`>
zcorpan: No

[06:19:01.0000] <annevk2>
whoa, air greeland has instant customer service via email, what a treat

[06:19:02.0000] <Philip`>
since I'd probably just be told that it was a design decision to not animate favicons, or something

[06:19:03.0000] <annevk2>
what do you mean with "properly"?

[06:21:00.0000] <zcorpan>
Philip`: we animate gif favicons i think

[06:21:01.0000] <jgraham>
annevk2: You're going to greenland?

[06:22:00.0000] <annevk2>
yup

[06:22:01.0000] <jgraham>
Nice

[06:22:02.0000] <annevk2>
yup :)

[06:24:00.0000] <Philip`>
annevk2: Animatedly

[06:25:00.0000] <annevk2>
i doubt it's a design decision since we let you animate them through JavaScript

[06:29:00.0000] <zcorpan>
annevk2: we let pages animate things through javascript when the "disable animated images" option is set, too

[06:31:00.0000] <Philip`>
/me attempts to upload a less-subtly-animated test

[06:36:00.0000] <hsivonen>
/me predicts that in future studies, Philip` will find role=complimentary on the Web

[06:37:00.0000] <Philip`>
http://philip.html5.org/tests/apng/favicon.html

[06:38:00.0000] <Philip`>
hsivonen: That seems to be presupposing that people will use role at all

[06:38:01.0000] <Philip`>
Oh, some people do use it already

[06:38:02.0000] <Philip`>
Especially http://www.tee-rail-alliance.de/

[06:39:00.0000] <Philip`>
which uses loads of different roles

[06:39:01.0000] <hsivonen>
Philip`: that one uses role="complementary" 

[06:40:00.0000] <Philip`>
and then there's one site with a role="navigation" on each page, and two with a role=""

[06:40:01.0000] <hsivonen>
Philip`: I meant that complementary is one of the English words that people can't spell

[06:42:00.0000] <Philip`>
hsivonen: Indeed, I was just wondering whether anybody would bother with role at all, because that's necessary before they'll have an opportunity to mis-spell it

[06:44:00.0000] <zcorpan>
does a label in javascript need to be followed by anything? i.e. is onclick="http://example.org" a valid script?

[06:46:00.0000] <jgraham>
zcorpan: a label must be followed by a statement

[06:46:01.0000] <jgraham>
So no, that's not valid

[06:46:02.0000] <zcorpan>
ok

[06:46:03.0000] <zcorpan>
i don't seem to get anything in the error console though

[06:47:00.0000] <zcorpan>
i get an error in firefox

[06:47:01.0000] <zcorpan>
bug in opera?

[06:48:00.0000] <Dashiva>
hmm

[06:58:00.0000] <Dashiva>
I've seen similar cases before. Error in sort comparison function didn't cause error in sorting, error in a statement that would've been a nop was optimized away.

[07:02:00.0000] <hsivonen>
why would an user want to read an accessibility conformance claim? http://www.brucelawson.co.uk/2009/rel-accessibility/

[07:02:01.0000] <hsivonen>
or a conformance claim of any kind?

[07:12:00.0000] <jgraham>
hsivonen: http://diveintoaccessibility.org/day_30_creating_an_accessibility_statement.html although it seems rather tenuous to me 

[07:14:00.0000] <jgraham>
In particular it seems to be predicated on the assumption that you are using a UA that does not provide clues about the avaliable functionality

[10:40:00.0000] <sicking>
Hixie, ping

[10:43:00.0000] <annevk5>
weinig, did you ping me the other day? still needed?

[10:55:00.0000] <annevk5>
hmm, OWL 2

[13:09:00.0000] <gsnedders>
Ow. In Saf4 for a display:table the 'width' property isn't the content width: the content width is 'width' - 'border-left'/'border-right' when it has borders (padding/margin is untested…)

[13:13:00.0000] <gsnedders>
/me stops going through bugs that need a reduction

[13:14:00.0000] <annevk5>
sounds like box-sizing is applied to table frames rather than HTML table elements or some such

[13:15:00.0000] <gsnedders>
Yeah, that's what I was guessing

[13:52:00.0000] <Hixie>
gsnedders: sounds vaguely like what the spec says to do

[13:52:01.0000] <gsnedders>
Hixie: Really? I had a look and it seemed wrong. Sites break if you implement the spec then, and Firefox and Opera don't follow it.

[13:53:00.0000] <gsnedders>
https://bugs.webkit.org/show_bug.cgi?id=13519

[13:53:01.0000] <Hixie>
i forget exactly what the spec says

[13:54:00.0000] <gsnedders>
I didn't see any special case for a table itself, only table cells

[14:44:00.0000] <jgraham>
gsnedders: BTW if you put 200px in the pass condition for the test it is really annoying because you have to verify that the shapre is really 200px to confirm if the test has passed

[14:45:00.0000] <jgraham>
Which doesn't really seem to be necessary here

[14:45:01.0000] <gsnedders>
jgraham: The test is the widths are calculated correctly, so it does seem relevant

[14:46:00.0000] <a-ja>
hi folks.  i may be missing something something, but is anyone aware of a reason placeholder shouldn't be supported on input type=search ?

[14:47:00.0000] <jgraham>
gsnedders: If you really care you should put a 200px green image and a 200px green box and make them line up and say something like "you should see a green rectangle below"

[14:48:00.0000] <jgraham>
Basically you don't want the pass condition to be anything that you can't verify in a second or two unless it is really super necessary

[14:49:00.0000] <jgraham>
(or hope that you have reftests so that you can say "this box should look exacly like this other box")

[14:50:00.0000] <gsnedders>
/me shrugs, and goes back to trying to shorten his English dissertation, as he thinks it is a bit more important

[14:50:01.0000] <jgraham>
a-ja: The only reason I can think of is some convention that input type=search should always have "search" as a placeholder

[14:51:00.0000] <jgraham>
gsnedders: Yeah that is more important :) Just a tip that may make some poor QA's life easier someday

[14:51:01.0000] <gsnedders>
Oh wow. I don't have the spec in my history

[14:52:00.0000] <gsnedders>
a-ja: placeholder is supported on input@type='search'

[14:53:00.0000] <a-ja>
jgraham: possible, i suppose...guess i'll followup with hixie, in case it's just  an oversight

[14:53:01.0000] <a-ja>
oh, guess i *did* miss it, then

[14:53:02.0000] <gsnedders>
a-ja: search is grouped with text in the spec. What makes you think it isn't supported?

[14:56:00.0000] <a-ja>
haven't read spec section lately....just recent change log comments / diffs re:placeholder

[14:57:00.0000] <a-ja>
makes sense if search is just subtype of text

[15:15:00.0000] <gsnedders>
\emph{That's} whom you should have married, […], instead of all that we \emph{teased} her to death!

[15:15:01.0000] <gsnedders>
What should the position of comma be around the […]?


2009-04-22
[17:39:00.0000] <cgriego>
Say HTML 6 adds a new void element. How would an HTML 5 parser know that the rest of the document isn't a child of that element?

[17:40:00.0000] <Hixie>
it wouldn't

[17:42:00.0000] <cgriego>
That seems problematic.

[17:43:00.0000] <Hixie>
why?

[17:47:00.0000] <cgriego>
Any new void element then creates a chicken & egg problem. Authors won't adopt it because they can't use it in a progressive enhancement safe way since it corrupts the DOM. Agents won't add it because authors don't use it. Similar situation to XHTML 2.

[17:50:00.0000] <Hixie>
doesn

[17:50:01.0000] <Hixie>
er

[17:50:02.0000] <Hixie>
doesn't seem to have caused particular problems in the past

[17:50:03.0000] <Hixie>
people don't really want to use new elements until the browsers use them anyway

[17:51:00.0000] <Hixie>
er, until they implement them i mean

[17:51:01.0000] <Hixie>
browsers add lots of stuff before authors use them

[17:51:02.0000] <Hixie>
in fact, they add everything before authors use them

[17:51:03.0000] <cgriego>
People are jumping in and using section/header/footer/etc now without proper browser support.

[17:51:04.0000] <Hixie>
not many people :-)

[17:52:00.0000] <Hixie>
and those people will get burnt anyway, e.g. when we rename <headeR>

[17:52:01.0000] <Hixie>
(or remove it, or whatever we end up doing with it)

[17:52:02.0000] <cgriego>
Not if they're serving as text/html ;)

[17:52:03.0000] <Hixie>
well i mean their pages won't be correct html5 anymore

[17:52:04.0000] <Hixie>
i don't mean that their site will break or anything

[17:53:00.0000] <cgriego>
As for the past, there's never been a time where HTML was actively evolving with so many authors and authoring programs.

[17:53:01.0000] <cgriego>
…as we have today

[17:53:02.0000] <Hixie>
yup, it'll keep on increasing until we have everyone on the planet online

[17:54:00.0000] <Hixie>
(and then it'll keep increasing until the population stops growing)

[17:55:00.0000] <cgriego>
I'd propose that HTML 5 parsing treat unrecognized elements (and only unrecognized elements) similar to foreign elements, where the SOLIDUS marks the start tag as self-closing.

[17:56:00.0000] <Hixie>
what problem does it solve?

[17:56:01.0000] <cgriego>
Author adoption of new void elements.

[17:57:00.0000] <Hixie>
wouldn't the fact that the browsers don't support those elements be a bigger barrier to adoption than the fact that the author has to jump through some minor hoops to get teh styling right?

[17:57:01.0000] <Hixie>
(if we introduce <foo>, you can always just do <span><foo></span> to get the parsing right in older browsers)

[17:58:00.0000] <Hixie>
put it this way: historically, HTML has never had a void element that was useful without the browser actually supporting the element directly somehow

[18:01:00.0000] <Philip`>
What about <meta>? That's useful for some UAs (e.g. search engines) even if browsers don't do anything with it

[18:01:01.0000] <Hixie>
elements in <head> have much bigger problems in the parser than whether they're void or not

[18:01:02.0000] <cgriego>
When <area> was first introduce, authors would add an image map to their site but also provide supplemental navigation below the image. Authors were able to adopt the element before all agents supported it.

[18:02:00.0000] <Hixie>
cgriego: but with <area>, the elements all get closed by the <map>, so it doesn't matter if the element gets closed or not

[18:03:00.0000] <cgriego>
You can't know that all new void elements will have a mandatory parent element.

[18:05:00.0000] <Hixie>
no, but i can predict the likelihood of there being a problem if unknown elements can't be closed by />, based on past experience

[18:06:00.0000] <Hixie>
(note that "span" is an unknown element per the parser, and we definitely don't want to support <span/>)

[18:06:01.0000] <Hixie>
(so actually what we would need to do is more complex than what you describe)

[18:09:00.0000] <cgriego>
Didn't realize span was unknown to the parser. Why is that?

[18:09:01.0000] <Hixie>
because only the elements with weird rules have to be known

[18:10:00.0000] <cgriego>
ah, like void elements :)

[18:10:01.0000] <Hixie>
yeah

[18:11:00.0000] <cgriego>
One of the solutions offered on the mailing list back in August was to have not add any new void elements.

[18:11:01.0000] <Hixie>
i maintain that there's no problem, so discussion of solutions seems pointless to me :-)

[18:11:02.0000] <Hixie>
but sure, we could just not add more void elements if that makes people happy :-)

[18:13:00.0000] <Hixie>
(at least until we have a need for one!)

[18:14:00.0000] <cgriego>
Not a fan of the idea, would just lead to more cases like <script src=# />

[18:15:00.0000] <Hixie>
hm?

[18:17:00.0000] <cgriego>
I can see how not allowing new void elements would lead to elements with empty content models instead, which then authors would just try to self-close (like with script src) causing more problems in the end.

[18:19:00.0000] <Hixie>
well then we'd just use a void element

[18:20:00.0000] <Hixie>
or we could just not introduce elements with empty content models as well

[18:23:00.0000] <Hixie>
wooot! i've finally gotten through all the feedback on microdata stuff

[18:23:01.0000] <Hixie>
ok now i have to go through all the notes i collected

[18:23:02.0000] <Hixie>
and actually come up with a set of use cases, requirements, and scenarios

[18:23:03.0000] <Hixie>
so that i can evaluate solutions

[18:25:00.0000] <Hixie>
so far i've boiled down 15000+ lines of e-mail and dozens of pages of wiki and blog comments into 600 lines or so of notes

[18:38:00.0000] <Hixie>
here are my notes http://wiki.whatwg.org/wiki/Microdata_Problem_Descriptions

[18:38:01.0000] <Hixie>
i'll clean it up later

[19:07:00.0000] <Dashiva>
The tomato video comment thing seems painful

[20:50:00.0000] <MikeSmith>
Hixie: the two tables in 4.10.4 "the input element" with info about states and attributes 

[20:51:00.0000] <MikeSmith>
it would be nice if those had IDs so the could be more easily referenced with fragment IDs

[20:51:01.0000] <MikeSmith>
and maybe table titles on them

[21:58:00.0000] <Hixie>
MikeSmith: file a bug

[22:23:00.0000] <MikeSmith>
Hixie: OK

[22:23:01.0000] <MikeSmith>
http://twitter.com/_masaka/statuses/1581828896

[22:23:02.0000] <MikeSmith>
"hmm, **HTML5** web addr says relative path to be % encoded with UTF-8, but query with document encoding. What if path+query in non UTF-8 doc?"

[22:26:00.0000] <Hixie>
i don't understand the question

[22:36:00.0000] <MikeSmith>
Hixie: I don't either. was hoping you would

[23:14:00.0000] <MikeSmith>
Hixie: 'wondering if a relative URI "パス?クエリ" in EUC-JP doc, "パス" to be % encoded as UTF-8 octets while "クエリ" as EUC-JP octets?'

[23:18:00.0000] <MikeSmith>
http://twitter.com/_masaka/status/1582345343

[23:28:00.0000] <Hixie>
MikeSmith: yes, that is indeed what browsers do (and what the spec says to do) as far as i'm aware

[23:29:00.0000] <MikeSmith>
Hixie: OK

[23:58:00.0000] <hsivonen>
the use case starting "Getting data out of poorly written Web pages" is clearly a winner

[23:59:00.0000] <MikeSmith>
hsivonen: where?

[00:02:00.0000] <Hixie>
http://wiki.whatwg.org/wiki/Microdata_Problem_Descriptions

[00:02:01.0000] <Hixie>
extracted out of 15,000 lines' worth of use case descriptions

[00:02:02.0000] <Hixie>
or what should have been 15,000 lines' worth

[00:02:03.0000] <Hixie>
but was actually 600 lines' worth about about 10,000 lines of stuff that had nothing to do with use cases

[00:04:00.0000] <MikeSmith>
Hixie: I see

[00:04:01.0000] <MikeSmith>
hsivonen: were you being sarcastic? "Getting data out of poorly written Web pages" does seem to be at least like a pretty decent use case

[00:05:00.0000] <hsivonen>
MikeSmith: I was being sarcastic. if the use case is about poorly written pages, a solution that requires the cooperation of the poorly-writing authors is likely to fail

[00:06:00.0000] <MikeSmith>
ah, yeah

[00:06:01.0000] <MikeSmith>
wasn't clear to me that it would require that cooperation

[00:06:02.0000] <Hixie>
hsivonen: you're assuming that that came from something to do with rdfa, which i don't think it did

[00:07:00.0000] <hsivonen>
Hixie: I'm not assuming anything about the origin of the use case

[00:07:01.0000] <MikeSmith>
"Getting data out of poorly written Web pages" seems similar to what browsers are already designed to do as far as rendering

[00:07:02.0000] <olliej>
MikeSmith: hehe

[00:08:00.0000] <hsivonen>
Hixie: I am assuming it is about new syntax rather than new algorithms, though

[00:08:01.0000] <Hixie>
hsivonen: use cases aren't about any kind of solution

[00:08:02.0000] <Hixie>
hsivonen: that's rather the point :-)

[00:08:03.0000] <Hixie>
we come up with problems then try and solve them, instead of coming up with solutions then trying to find problems for them

[00:13:00.0000] <hsivonen>
all the use cases that contain the word "commercial" seem like problems of CC's own creation

[00:15:00.0000] <hsivonen>
see also http://twitter.com/diveintomark/status/1256490748

[02:42:00.0000] <hsivonen>
I'll implement the <p><table> quirk...

[02:56:00.0000] <annevk2>
you could write a little story around it how Hixie infected HTML parsers with a quirks mode switch

[03:00:00.0000] <hsivonen>
annevk2: yeah, I was thinking I'd blog about this

[03:01:00.0000] <annevk2>
I was thinking about a comment in the source code to be laughed about by people who know, but I suppose that works too :)

[03:14:00.0000] <hsivonen>
made innerHTML propagate the document quirkiness properly, too

[03:15:00.0000] <hsivonen>
which means this thing now has API surface!

[03:24:00.0000] <annevk2>
/me finds it weird that only Gecko has this quirk: https://bugzilla.mozilla.org/show_bug.cgi?id=489532

[03:24:01.0000] <annevk2>
usually it's the other browsers that are less strict

[03:28:00.0000] <hsivonen>
annevk2: IIRC, someone behind a firewall complained

[03:29:00.0000] <annevk2>
if we really need to have it zcorpan better documents how it works, but hopefully Gecko people are willing to fix the bug

[03:30:00.0000] <annevk2>
e.g. afaict it doesn't match what IE does

[03:30:01.0000] <hsivonen>
oh?

[03:31:00.0000] <annevk2>
I read that IE supports attributes as well, createElement("<x foo=bar>")

[03:34:00.0000] <zcorpan>
check the source of web dom core

[03:44:00.0000] <hsivonen>
anyone got Netscape 1.1 around to test how it laid out <p align=center><table>?

[03:45:00.0000] <hsivonen>
out of historical interest

[03:51:00.0000] <zcorpan>
hsivonen: netscape 1.1 seems to not support the align attribute

[03:54:00.0000] <Philip`>
zcorpan: <p align=center> works in Netscape 1.1

[03:55:00.0000] <Philip`>
But <p align=center><table> looks left-aligned

[03:55:01.0000] <Philip`>
/me goes away for a bit

[04:00:00.0000] <zcorpan>
hmm it didn't want to center for me at first but then suddenly it did

[04:00:01.0000] <zcorpan>
netscape 1.1 seems to close the p

[04:04:00.0000] <zcorpan>
hsivonen: http://simon.html5.org/dump/p-table-nn11.PNG

[04:05:00.0000] <zcorpan>
same in netscape 2

[04:13:00.0000] <hsivonen>
I wonder how the non-closing behavior was introduced

[04:33:00.0000] <zcorpan>
even netscape 4 doesn't center the table

[04:37:00.0000] <zcorpan>
when did ie first support tables?

[04:39:00.0000] <zcorpan>
ie3 is the earliest i could get running, and it centers the table

[04:42:00.0000] <zcorpan>
netscape 6 does not center

[04:52:00.0000] <Philip`>
Netscape 1.1 would be more usable if it used HTTP 1.1

[04:52:01.0000] <Philip`>
(particularly because of Host headers, I guess)

[04:52:02.0000] <Philip`>
Currently even Google doesn't work in it

[04:58:00.0000] <Philip`>
(...where "doesn't work" means it always returns a 302 response saying it's moved to exactly the location I was looking at)

[05:05:00.0000] <Philip`>
IE 2.0 centers the table

[05:08:00.0000] <Philip`>
IE 1 (which calls itself "Microsoft Internet Explorer/4.40.308" but I think really is version 1) also centers the table

[05:12:00.0000] <Philip`>
zcorpan: (It seems easy enough to run IE1 and IE2 (at least on Linux) - download from http://browsers.evolt.org/?ie/32bit , make Wine pretend to be Windows 95, start the installer, copy the .cab file out of c:\windows\temp, cabextract it, then run iexplore.exe)

[05:39:00.0000] <jgraham>
/me wonders if anyone has got timbl's broser running on anything other than NeXT

[06:16:00.0000] <hsivonen>
zcorpan, Philip`: thank you for the testing. (the Netscape 4 and 6 results surprise me)

[06:33:00.0000] <annevk2>
in HTML4 <link> has a charset attribute

[06:33:01.0000] <annevk2>
I think browsers use that to determine e.g. the encoding of an external style sheet

[06:35:00.0000] <Philip`>
jgraham: Surely it wouldn't be too hard to recompile on modern Linux

[06:37:00.0000] <Philip`>
It's nice how its documentation still works correctly in modern web browsers

[06:38:00.0000] <Philip`>
though it looks a bit outdated since it says <HEADER> <TITLE>WorldWideWeb source files</TITLE> <NEXTID N="19"> </HEADER>

[10:04:00.0000] <myakura>
"HTML5 poses to make web development less secure, more complicated and more expensive" http://loud.anotherquietday.com/post/98616203/the-threshold-of-ebook-progress

[10:04:01.0000] <Philip`>
Sounds like technological progress to me

[10:06:00.0000] <Philip`>
since adding features will cause those effects, and it's impossible to remove features, so the only alternative is to stagnate

[10:09:00.0000] <Philip`>
("HTML5 allows more flexibility in the markup format, which will lead to more XSS vulnerabilities" just seems incorrect, since it's no more flexible than what browsers already accept, and at least it'll be more consistent and better defined)

[10:10:00.0000] <Philip`>
(but "adds a layer of the most maliciously exploited software layer in mankind to every page (SQL)" seems a reasonable criticism, because SQL is easy to misuse and so people will misuse it)

[10:14:00.0000] <mpt>
It's not reasonable to say it adds it to "every page", though

[10:16:00.0000] <Philip`>
Ah, yes

[10:16:01.0000] <Philip`>
/me missed that bit

[10:18:00.0000] <Philip`>
I wonder how long it'll be before some writes a sqlite_real_escape_string function so they can safely concatenate values into their query

[11:02:00.0000] <Hixie>
hsivonen: in my defence, i was just trying to get the browsers to follow the specs. the real problem is html 4.01 didn't go through a real CR phase.

[12:43:00.0000] <tantek>
Hixie - can you name any W3C RECs that went through a real CR phase?

[12:49:00.0000] <Hixie>
tantek: not yet. Selectors is close.

[12:50:00.0000] <tantek>
I would tend to agree with that assessment.


2009-04-23
[18:45:00.0000] <Hixie>
any opera people here? i'm trying to work out how to complete this sentence:

[18:45:01.0000] <Hixie>
Chaals could improve the Opera intranet if he had a mechanism for identifying the original source of various parts of a page, because...

[18:46:00.0000] <Hixie>
it's based on something chaals said but i don't seem to be able to find anything about how knowing the original source of a page would actually help improve the intranet

[21:51:00.0000] <Hixie>
i love this use case (paraphrasing): "how come if a tv guide site doesn't include a link to imdb my browser can't detect that it's a tv show from the microdata the author puts in the page and create an implied link to imdb?"

[21:53:00.0000] <Hixie>
if the tv guide author isn't including the link you want... what makes you think he'll include the microdata you want?

[21:59:00.0000] <MikeSmith>
Hixie: because experience has shown us that right-thinking authors will always go out of their way to embed hidden content that's only intended to be exposed to users of specialized tools, but not to normal users

[21:59:01.0000] <MikeSmith>
and they will never make any mistakes when they do it

[22:00:00.0000] <Hixie>
well actually some of the summary="" data did show that that does hapen sometimes!

[22:00:01.0000] <Hixie>
they have to be pretty motivated authors though

[23:47:00.0000] <MikeSmith>
Hixie: am I'm reading it incorrectly, or does abarth's content-sniffing data indicate that 0x00000100 signature almost never actually triggers in practice?

[23:47:01.0000] <MikeSmith>
http://webblaze.cs.berkeley.edu/2009/content-sniffing/metrics.html

[23:48:00.0000] <Hixie>
the metric counter is probably not being triggered for favicons

[23:48:01.0000] <Hixie>
but you'd have to speak to abarth for details

[23:49:00.0000] <MikeSmith>
I see

[23:49:01.0000] <MikeSmith>
just asking because that seems the be the only remaining case where Gecko content-sniffing behavior is different from the spec

[23:50:00.0000] <MikeSmith>
and abarth recommendation in the related Mozilla bug is that signature just be removed

[23:51:00.0000] <MikeSmith>
https://bugzilla.mozilla.org/show_bug.cgi?id=465007

[00:13:00.0000] <MikeSmith>
http://twitter.com/200ok/status/1592050739

[00:13:01.0000] <MikeSmith>
"why isn't NOSCRIPT valid inside HEAD? there are perfectly legit uses..."

[00:14:00.0000] <MikeSmith>
noscript is allowed in head, right?

[00:16:00.0000] <MikeSmith>
ah, I guess it's not allowed in HTML4

[01:02:00.0000] <Hixie>
Lachy: have you heard any news of a a requiem that works with the latest iTunes?

[01:33:00.0000] <Lachy>
Hixie, no

[01:34:00.0000] <Lachy>
Hixie, my only source of information is that forum thread. So if you checked there, I'm unlikely to know anything more than you

[01:35:00.0000] <Lachy>
but I just started freenet to see if the page has been updated since the last version was released

[01:46:00.0000] <Philip`>
/me wonders why they don't just publish all the news and updates on, like, a web page

[01:50:00.0000] <Lachy>
Philip`, because Requiem is techically illegal to distribute, and so an ordinary web page is subjected to DMCA take down notices and subsequent legal issues for failing to comply

[01:50:01.0000] <Lachy>
Using Freenet helps to avoid that by keeping anonymous

[01:51:00.0000] <olliej>
Lachy: requiem?

[01:51:01.0000] <Lachy>
it's illegal because of the anti-circumvention clauses in the DMCA and similar laws in other countries

[01:51:02.0000] <Lachy>
olliej, an iTunes DRM removal utility

[01:53:00.0000] <olliej>
happily at least some of that drm is gone

[01:57:00.0000] <Philip`>
/me recently discovered get_iplayer which can download BBC programmes via the streaming Flash RTMP protocol, including a few in HD (1280x720 3Mbps H.264), which is pretty nice, and much more useful than the official Flash interface

[01:58:00.0000] <Philip`>
(They have a cross-platform programme downloader built on Adobe Air, but Air only installs onto deb/RPM-based Linuxes, not Gentoo, so I can't use it)

[01:59:00.0000] <Philip`>
/me would prefer it if they didn't have to make things so unnecessarily complex (rather than e.g. providing a download link to an .mp4 file)

[03:04:00.0000] <jgraham>
/me realises that one of his scripts has broken because he installed a namespace-aware html5lib

[03:05:00.0000] <jgraham>
except s/one/many/, just that I haven't tried to use the others yet

[03:27:00.0000] <annevk2>
lol

[03:27:01.0000] <annevk2>
check http://www.stockholmgamlastan.se/ in Opera

[03:28:00.0000] <jgraham>
It would be much less funny if the photographs on the site weren't so crappy

[03:29:00.0000] <krijnh>
Nice, simpel protection :)

[03:29:01.0000] <svl>
heh

[03:29:02.0000] <svl>
/me toggles dom.event.contextmenu.enabled and happily rightclicks a couple of times

[03:29:03.0000] <krijnh>
That'll teach them!

[03:30:00.0000] <olliej>
wow, that's stupid

[03:30:01.0000] <olliej>
errr

[03:30:02.0000] <MikeSmith>
annevk2: hallvord should write a patch for browser.js that unblocks opera and then replaces all the images on the pages with lolcats or something

[03:31:00.0000] <Philip`>
/me left-clicks and drags the image into Firefox's location bar, and then uses Ctrl+S

[03:32:00.0000] <MikeSmith>
speaking of stupid: http://www.torchmobile.com/blog/?p=29

[03:33:00.0000] <jgraham>
Philip`: Why not jut drag the image to your folder of choice?

[03:40:00.0000] <Philip`>
jgraham: Because I don't trust drag-and-drop to work in any scenario more complex than from one application to itself

[03:41:00.0000] <Philip`>
(Actually it often works between native KDE applications too, but Firefox isn't one of those so I still wouldn't trust it)

[03:43:00.0000] <annevk2>
http://sourcefrog.net/projects/meantime/ is interesting

[03:43:01.0000] <annevk2>
basically cookies through HTTP caching

[03:44:00.0000] <annevk2>
so you'd have to disable your cache altogether if you want to prevent tracking

[04:11:00.0000] <annevk2>
it might also explain why Hixie got such strange values back for Last-Modified although somehow I doubt a lot of sites would be using the technique

[04:16:00.0000] <svl>
/me set his cache to 2MB a couple of years ago as a compromise between mostly preventing this and still having some use of the cache while staying within a single website.

[04:18:00.0000] <Philip`>
/me enables cookies and uses a mostly-static IP, so people don't have to bother wasting time on such tricks to track people

[06:28:00.0000] <Lachy>
re the right clicking on stockholmgamlastan.se , I guess they're not aware that Firefox has had the option to prevent context menu blocking for years

[06:32:00.0000] <annevk2>
or aware of browser caches...

[06:34:00.0000] <Lachy>
generally, people who try to prevent right clicking just want to use it as a deterrent

[06:34:01.0000] <jgraham>
Or drag and drop of view-page info or any of the other hundreds of ways of getting at those images

[06:34:02.0000] <Lachy>
besides, simply dragging the image from the browser to the desktop is the quickest way to save images now anyway

[06:35:00.0000] <Philip`>
The illusion of security is more valuable than security itself

[06:36:00.0000] <jgraham>
Philip`: To whom? In this case I would prefer the ability to access the site and use the context menu than the ability to save the files 

[06:37:00.0000] <Philip`>
jgraham: To the people who are choosing to implement illusory security, I guess

[06:38:00.0000] <Philip`>
(I am, of course, entirely wrong)

[06:38:01.0000] <Lachy>
it would be interesting to see some sort of usability study that analysed common techniques real users used when saving images, to find out if using the context menu is still a significantly relevant technique compared with others

[06:38:02.0000] <Philip`>
(because they'd prefer real security, but the cost is too high (since it's impossible), so they make do with the best can they easily implement)

[06:39:00.0000] <Philip`>
s/can they/they can/

[09:04:00.0000] <annevk2>
hurray: http://www.w3.org/TR/css3-mediaqueries/

[09:09:00.0000] <Philip`>
I assume someone has already pointed out that the brackets in media="print and (color)" look weird and stupid and unintuitive? :-)

[09:10:00.0000] <annevk2>
not sure, but at this point the comment comes about seven years too late

[09:11:00.0000] <Philip`>
/me wonders why it's restricted to disjunctive normal form

[09:13:00.0000] <mpilgrim>
recommendations for "this week" happenings?

[09:13:01.0000] <mpilgrim>
i'm covering the datagrid and the keygen elements

[09:13:02.0000] <annevk2>
HTML5 was just published

[09:13:03.0000] <mpilgrim>
ah, indeed it was

[09:13:04.0000] <mpilgrim>
and presumably your html-differences document as well?

[09:13:05.0000] <annevk2>
also the drafts that were splitted out from HTML5 as well as Web Workers have been published too

[09:13:06.0000] <annevk2>
yes it was

[09:13:07.0000] <mpilgrim>
ooh

[09:13:08.0000] <mpilgrim>
excellent

[09:13:09.0000] <hsivonen>
mpilgrim: it looks a lot like there's going to be a parsing quirk (unless Hixie changes Acid2 and all browsers agree)

[09:14:00.0000] <mpilgrim>
is this the <p><table> thing?

[09:14:01.0000] <hsivonen>
mpilgrim: yeah

[09:14:02.0000] <mpilgrim>
yes, i saw your blog post on it

[09:14:03.0000] <mpilgrim>
/me adds that link to the list of "around the web"

[09:14:04.0000] <hsivonen>
mpilgrim: zcorpan's public-html post has the conclusion, though

[09:15:00.0000] <mpilgrim>
checking

[09:15:01.0000] <Philip`>
annevk2: It may have been seven years but pretty much nobody uses it, so it'd be easy to change without breaking compatibility much :-)

[09:16:00.0000] <mpilgrim>
/me can't tell if Philip` is joking

[09:16:01.0000] <billyjackass>
mpilgrim: dunno if you'd find this newsworthy

[09:16:02.0000] <mpilgrim>
it's used quite often for mobile stylesheets for iphone, ipod, android

[09:16:03.0000] <billyjackass>
http://hg.mozilla.org/mozilla-central/rev/e6812f79d2ba

[09:16:04.0000] <billyjackass>
"Harmonize content sniffing in HTML5 and Firefox"

[09:16:05.0000] <billyjackass>
abarth checking in on April 5

[09:16:06.0000] <mpilgrim>
billyjackass: absolutely!

[09:17:00.0000] <Philip`>
As far as I can see, the only examples more complex than comma-separated tokens are (I guess) trying to select mobile devices based on screen width

[09:17:01.0000] <Philip`>
(and they're not using the weird single-token-in-parentheses syntax)

[09:17:02.0000] <billyjackass>
mpilgrim: https://bugzilla.mozilla.org/show_bug.cgi?id=465007 is the related bug

[09:17:03.0000] <mpilgrim>
/me first learned about css media queries from an article on iphone-specific stylesheets

[09:17:04.0000] <Philip`>
      1 only screen and (max-device width:480px)

[09:17:05.0000] <Philip`>
     31 only screen and (max-device-width: 480px)

[09:17:06.0000] <Philip`>
      1 screen and (min-device-width: 481px)

[09:17:07.0000] <Philip`>
      1 screen and (-webkit-min-device-pixel-ratio:0)

[09:18:00.0000] <Philip`>
(Total occurrence counts, and values, from 130K pages)

[09:18:01.0000] <mpilgrim>
i was ready to rant about how apple was making shit up again and polluting the web, until i checked discovered that it was a standard that nobody else was using

[09:18:02.0000] <annevk2>
:)

[09:18:03.0000] <annevk2>
it's been in Opera for a long time

[09:18:04.0000] <mpilgrim>
Philip`: that's interesting

[09:19:00.0000] <annevk2>
Firefox 3.5 has it too

[09:19:01.0000] <Philip`>
Polluting the web with standards isn't really any better than polluting the web with proprietary technologies

[09:19:02.0000] <hsivonen>
mpilgrim: media queries are nice with Opera Mini, too

[09:19:03.0000] <mpilgrim>
Philip`: blasphemy!

[09:19:04.0000] <mpilgrim>
;)

[09:19:05.0000] <Philip`>
(They might be a bit better documented but that's the only real difference, and they equally add to complexity and not-working-in-everyone's-browser-ity and so on)

[09:22:00.0000] <mpilgrim>
where is the new html5 draft published?

[09:22:01.0000] <mpilgrim>
http://www.w3.org/TR/html5/ still lists the feb 12 draft

[09:23:00.0000] <Rik|work>
http://www.w3.org/TR/2009/WD-html5-20090423/

[09:23:01.0000] <annevk2>
seems the latest version links aren't updated yet

[09:23:02.0000] <annevk2>
MikeSmith, ^^

[09:23:03.0000] <annevk2>
it is however announced on http://www.w3.org/

[09:23:04.0000] <Philip`>
Hmph, the multipage splits are all stupid now

[09:24:00.0000] <MikeSmith>
yeah, TR symlinks are not set up yet

[09:24:01.0000] <MikeSmith>
webmaster is working on it now

[09:24:02.0000] <annevk2>
hmm, webstorage, websockets, and workers

[09:24:03.0000] <annevk2>
nobody did a sanity check on the shortnames? :)

[09:24:04.0000] <MikeSmith>
annevk2: what's wrong with the shortnames?

[09:25:00.0000] <mpilgrim>
Rik|work: thanks

[09:25:01.0000] <annevk2>
MikeSmith, nothing much, but if either all were prefixed with web or none would've been better (preferably all, given webidl)

[09:26:00.0000] <MikeSmith>
annevk2: point taken

[09:27:00.0000] <Philip`>
Hmm, www.w3.org actually links to http://www.w3.org/TR/2009/WD-html5-20090421/ which redirects to ...23

[09:39:00.0000] <MikeSmith>
the /TR versions are up now

[09:39:01.0000] <MikeSmith>
Philip`: I'll get that fixed

[09:55:00.0000] <remysharp>
Is there any way, or current discussion on the audio and video tags to prevent the browser from downloading the content?

[09:56:00.0000] <gsnedders>
No

[09:56:01.0000] <remysharp>
I'm pretty new to how this works, so please humour me - but how do I start that discussion (or is it just via here!)?

[09:57:00.0000] <gsnedders>
remysharp: http://wiki.whatwg.org/wiki/FAQ#Is_there_a_process_for_adding_new_features_to_the_spec.3F

[09:57:01.0000] <annevk5>
yeah, don't like to any resource

[09:59:00.0000] <remysharp>
annevk2: I mean to prevent the tag from saturating the bandwidth when the tag's in heavy used on the page

[09:59:01.0000] <gsnedders>
remysharp: Why are you using it without wanting it to be downloaded?

[09:59:02.0000] <remysharp>
but then, thinking out loud, there's no way to say start loading without explicitly using some JS function

[09:59:03.0000] <remysharp>
gsnedders: if there's 50 audio tags on the page

[09:59:04.0000] <remysharp>
and they all come down at once

[09:59:05.0000] <gsnedders>
remysharp: But why would there be?

[09:59:06.0000] <remysharp>
then my bandwidth is going to go loopy

[09:59:07.0000] <remysharp>
one second - I'll give you a link

[10:00:00.0000] <remysharp>
http://huffduffer.com/tags/sxswi2009

[10:00:01.0000] <remysharp>
so this page - not 50, but a lot right?

[10:00:02.0000] <remysharp>
these used to be audio tags

[10:00:03.0000] <remysharp>
and it would fail back to flash for the play buttons

[10:01:00.0000] <annevk5>
that seems like something the browser should optimize

[10:01:01.0000] <Rik|work>
http://www.whatwg.org/specs/web-apps/current-work/multipage/video.html#attr-media-autobuffer

[10:01:02.0000] <remysharp>
as in, don't download them in parallel?

[10:01:03.0000] <Rik|work>
remysharp: isn't it the purpose of autobuffer ?

[10:01:04.0000] <annevk5>
remysharp, as in, do whatever is best for the user

[10:01:05.0000] <remysharp>
Rik|work: just reading that now...

[10:02:00.0000] <remysharp>
so would I set this to false to tell the UA to not go and get all this data?

[10:02:01.0000] <remysharp>
(for example)

[10:03:00.0000] <Rik|work>
from what I read, it is just a hint

[10:03:01.0000] <Rik|work>
maybe a browser could first download media with autobuffer=true and then those with false

[10:04:00.0000] <remysharp>
again, just thinking out loud, but what about when mobile devices support html5 - 

[10:04:01.0000] <remysharp>
bandwidth is a cost for some plans

[10:04:02.0000] <remysharp>
wouldn't we want to be able to prevent this?

[10:04:03.0000] <remysharp>
I guess this could boil down to the UA having to have a setting in the phone I guess.

[10:04:04.0000] <jmb>
is this not just something which should be left to UAs to sort out?

[10:04:05.0000] <remysharp>
(I guess...I guess)

[10:05:00.0000] <remysharp>
I'm starting to see that for the mobile case, yes.

[10:07:00.0000] <gsnedders>
remysharp: It's a problem for any number of images, videos, and audio files

[10:08:00.0000] <remysharp>
gsnedders: yeah, you're definitely right, and putting images in that context, kind of makes me see that it's definitely down to the browser.

[10:08:01.0000] <remysharp>
that said... 

[10:09:00.0000] <gsnedders>
remysharp: Limited bandwidth is not a new problem, we already had it with img (esp. in the days of dial-up being normal)

[10:09:01.0000] <remysharp>
didn't the image tag used to have a low-rez attribute or something...

[10:09:02.0000] <remysharp>
like a preview on the image tag...am I making this up??

[10:09:03.0000] <gsnedders>
remysharp: Um, maybe HTML 3.0? It did sometime. Nobody ever implemented it. Nobody cares about it.

[10:09:04.0000] <remysharp>
:D

[10:09:05.0000] <remysharp>
okay, I was harking back a bit then!

[10:10:00.0000] <gsnedders>
Before I had ever used the web, I think :)

[10:10:01.0000] <Philip`>
Nobody implemented it?

[10:11:00.0000] <Philip`>
I'm sure I remember it as actually working

[10:11:01.0000] <Philip`>
(a long long time ago)

[10:11:02.0000] <gsnedders>
Well, nobody somebody. Nobody relevant with enough influence, though.

[10:11:03.0000] <gsnedders>
s/nobody/maybe/

[10:11:04.0000] <remysharp>
Philip`: I'm pretty sure it did, I'm sure I used it back then

[10:11:05.0000] <Philip`>
Apparently it worked in Netscape

[10:11:06.0000] <remysharp>
nobody relevant == me :)

[10:11:07.0000] <Philip`>
who had influence

[10:11:08.0000] <Philip`>
http://www6.uniovi.es/gifanim/gifhtml.htm

[10:12:00.0000] <Philip`>
"Only Netscape (that I know of) supports the Netscape extension of LOWSRC"

[10:12:01.0000] <remysharp>
that was the one lowsrc

[10:13:00.0000] <Philip`>
I guess that technique for GIF animation didn't take off

[10:13:01.0000] <remysharp>
I just wonder whether there's a usecase for some attribute like 'loadondemand' 

[10:13:02.0000] <remysharp>
(for video + audio)

[10:47:00.0000] <annevk5>
remysharp, you start with the use case, not the solution

[10:47:01.0000] <annevk5>
remysharp, i.e. you're doing it wrong :)

[10:47:02.0000] <remysharp>
annevk5: okay, so I should start with my own experience when I visited the page, and it tried to load all the audio thus slowing my connection and killing my experience

[10:48:00.0000] <remysharp>
it's a bit over the top - but that's the point you're making right

[10:48:01.0000] <remysharp>
how do we make sure the experience is still fast and not hammered by lots of things downloading that I don't want.

[10:48:02.0000] <annevk5>
by getting a better browser

[10:49:00.0000] <remysharp>
it's not the browser - it was my connection

[10:49:01.0000] <Philip`>
remysharp: That seems like a reasonable statement of the problem

[10:49:02.0000] <remysharp>
Philip`: okay, cheers.

[10:49:03.0000] <remysharp>
at least I post together my thoughts on it, *then* it can be poo-pooed :-)

[10:50:00.0000] <remysharp>
I take back what I said about connection - you could be right, maybe it's the browser, perhaps not. That I can test myself fairly easily.

[10:51:00.0000] <annevk5>
my point is that the browser is more likely aware of the user's available bandwidth than the website owner

[10:51:01.0000] <Philip`>
I guess the solutions are something like (1) require page authors to add an attribute to most of their audio elements when they've got quite a few and they think it might be bad for most of their users; (2) make browsers manage resources sensibly, by not opening a hundred socket connections at once or downloading dozens of megabytes automatically or whatever; (3) something else

[10:51:02.0000] <annevk5>
so if someone (between the browser and website owner) needs to be in control of bandwidth usage (and disk usage, and memory usage, etc.) it's the browser

[10:52:00.0000] <Philip`>
and (2) seems a solution that's likely to work better for users

[10:52:01.0000] <remysharp>
Sure, but as a developer of page, I'm creating a listing page with video tags, I know the user doesn't want them all at once down the wire though.

[10:52:02.0000] <remysharp>
(in response to annevk5)

[10:52:03.0000] <annevk5>
yeah, and the browser can know that too

[10:53:00.0000] <remysharp>
Philip`: I definitely agree that browsers should need to be wary of concurrent downloads for audio and video tags - 

[10:53:01.0000] <Philip`>
I assume browsers already apply their 6-connections-per-domain-name limit to media downloads?

[10:53:02.0000] <remysharp>
but, right now, that's limited to 2 for IE and 4(?) for FF (I think?)

[10:54:00.0000] <remysharp>
I thought FF was on 4.

[10:54:01.0000] <Philip`>
remysharp: I think it's 6 in IE8 and recent FFs etc

[10:54:02.0000] <remysharp>
Right, cool.

[10:54:03.0000] <Philip`>
at least for normal page content

[10:54:04.0000] <Philip`>
("recent" might mean "not yet released"; I'm not at all sure)

[10:54:05.0000] <remysharp>
I guess that would actually be a big problem point then too

[10:55:00.0000] <remysharp>
big-ish

[10:55:01.0000] <remysharp>
because if there's images or scripts below the 6th video, they're all going to hang if they're the same domain

[10:55:02.0000] <remysharp>
obviously solved by splitting your content - but I don't think that's the point

[10:55:03.0000] <Rik|work>
http://stevesouders.com/ua/

[10:55:04.0000] <Rik|work>
check the conns/host column

[10:56:00.0000] <remysharp>
/me bookmarks - cheers

[10:56:01.0000] <Philip`>
Rik|work: Nice, thanks

[10:56:02.0000] <Rik|work>
btw, webkit will be 6 conns/host soon

[10:57:00.0000] <Philip`>
remysharp: Only while the videos are actively downloading - the connections should be released when it's stopped buffering, and it probably shouldn't be buffering the entire video before you've started playing it

[10:58:00.0000] <remysharp>
are you saying that they would buffer a bit, then move on to the next video then?

[10:58:01.0000] <Philip`>
I suppose I don't see why they'd buffer anything at all

[10:58:02.0000] <Philip`>
other than the poster image

[10:59:00.0000] <remysharp>
Yeah, that's the problem though - in WebKit, last I saw, the just download the whole shebang

[11:00:00.0000] <remysharp>
which is why Jeremy's site (the link I posted above) was changed from the audio tag to just flash

[11:15:00.0000] <Philip`>
mpilgrim: "Adam Barth's [PDF] whitepaper" seems misleading, since I interpret 'whitepaper' as being a marketing document (and Wikipedia seems to agree so I must be right), whereas this thing is an actual proper academic paper (so I'd just call it a 'paper')

[11:16:00.0000] <annevk5>
proper response :)

[11:17:00.0000] <Philip`>
(?)

[11:17:01.0000] <annevk5>
joking

[11:17:02.0000] <annevk5>
<datagrid> is also not quite as new as mark said

[11:18:00.0000] <Philip`>
/me is confuzzled

[11:18:01.0000] <annevk5>
you'll get over it

[11:18:02.0000] <Philip`>
I may

[12:35:00.0000] <gmiernicki>
<audio> +1 <video> +1 

[12:35:01.0000] <gmiernicki>
flash -1

[14:47:00.0000] <gsnedders>
Hixie: You _would_ just reference Avenue Q in your email.

[14:47:01.0000] <gsnedders>
/me now really wants to see Avenue Q after all your references to it

[14:47:02.0000] <Hixie>
i've seen it THREE times

[14:47:03.0000] <Hixie>
it's that good

[14:47:04.0000] <zcorpan_>
http://www.w3.org/TR/2009/CR-CSS2-20090423/ still hasn't fixed "BODY element in XHTML" :-(

[14:48:00.0000] <gsnedders>
Hixie: I mean, I was thinking about going to London in June, and there's a production there…

[14:53:00.0000] <annevk5>
zcorpan_, really?

[14:53:01.0000] <annevk5>
zcorpan_, might have been a resolved issue without edits, but I actually thought the edits were made too

[14:54:00.0000] <gsnedders>
/me realizes how crazy his June will be

[14:54:01.0000] <zcorpan_>
annevk5: if an issue is resolved but not changed in the spec, could the spec go to rec without the change on the basis that there are no unresolved issues?

[14:54:02.0000] <annevk5>
zcorpan_, the actual text is updated I think

[14:54:03.0000] <gsnedders>
I have an exam on the 4th, I am away 12th to 14th at leadership training thing, and Hadrian's Wall trip is 19th to 21st, then I start internship on the 29th… And I was going to go to London too…

[14:55:00.0000] <zcorpan_>
oh i didn't check further than the "features at risk" section

[14:55:01.0000] <annevk5>
zcorpan_, "For documents whose root element is an HTML "HTML" element or an XHTML "html" element [...]"

[14:55:02.0000] <annevk5>
zcorpan_, I guess somebody forgot to remove that text there

[14:55:03.0000] <zcorpan_>
ok

[14:55:04.0000] <zcorpan_>
i guess they'll notice

[14:58:00.0000] <annevk5>
zcorpan_, I guess the spec could go to rec btw if nobody noticed

[15:07:00.0000] <Hixie>
i wonder if the e-mail i sent will trigger 15,000 replies, or zero.

[15:07:01.0000] <Hixie>
i guess i'll poke at websocket then get back to microdata

[15:27:00.0000] <gsnedders>
Weee… More bug reports of XML parse errors!

[15:27:01.0000] <gsnedders>
annevk2: Finish XML5, kthxbai.

[15:32:00.0000] <Philip`>
XML parse errors should be rebranded as "XML parse bonuses", so people don't think it's so bad to encounter one and report them as bugs

[15:33:00.0000] <jmb>
that'll lead to competition to get the highest bonus score, though

[15:34:00.0000] <annevk5>
gsnedders, it seems hardly anyone cares anymore about XML these days

[15:34:01.0000] <annevk5>
gsnedders, I mean sure, there's the occasional issue, but nothing compared to CSS/HTML/DOM/ECMAScript

[15:34:02.0000] <gsnedders>
annevk5: Should we just use Aaron Swartz's suggestion of RSS 3.0?

[15:34:03.0000] <gsnedders>
http://www.aaronsw.com/weblog/000574

[15:34:04.0000] <annevk5>
gsnedders, I'd rather we just give up on feeds and use HTML; HTML files are not that much larger

[15:35:00.0000] <gsnedders>
RSS 3.0 is simpler.

[15:35:01.0000] <annevk5>
gsnedders, though meanwhile I guess we're stuck with several flavors of RSS and Atom

[15:35:02.0000] <gsnedders>
(Read that page.)

[15:36:00.0000] <annevk5>
(I know RSS 3.0.)

[15:36:01.0000] <gsnedders>
(Ah, OK.)

[15:36:02.0000] <gsnedders>
(It wasn't obvious.)

[15:38:00.0000] <jwalden>
(I wonder why all recent lines of discussion have been parenthesized.)

[15:41:00.0000] <Philip`>
(It helps prevent the characters from falling off the edges of the sentence and making a terrible mess on the channel's carpet)

[15:43:00.0000] <Philip`>
(Well, it's either a carpet or a thick layer of hair and biscuit crumbs, and I'm not quite sure which)

[16:23:00.0000] <annevk5>
http://sharovatov.wordpress.com/2009/04/22/relcanonical/ more evidence as to why rel= fails

[16:29:00.0000] <Dashiva>
annevk5: Heh, I didn't even notice the error until I went to read the comments on your post

[16:29:01.0000] <annevk5>
and I made a mistake too, I meant to say rev=

[16:30:00.0000] <annevk5>
/me blinks


2009-04-24
[19:08:00.0000] <Hixie>
http://www.w3.org/mid/D11E7D60-62A0-426A-B379-62A950A9C487⊙nc

[23:26:00.0000] <hsivonen>
http://blogs.adobe.com/jd/2009/04/thoughts_on_rubys_html_reunifi.html

[01:29:00.0000] <jgraham>
I am unconvinced that the bibliographic data use case is really strong. I mean it is very useful if there is a list of papers to be able to get bibliographic data for them but the main problem I had when using such systems was a) the bibliographic data was incorrect or b) the bibliographic data was in the wrong format and could not be trivially converted to the correct fprmat

[01:30:00.0000] <jgraham>
Neither of which seems like it can be solved at the HTML level

[01:32:00.0000] <hsivonen>
its pretty useful that ACM Portal gives you snippets you can paste into .bib

[01:32:01.0000] <hsivonen>
I used .bib with HTML in oder to avoid reinventing the wheel

[01:32:02.0000] <hsivonen>
*order

[01:33:00.0000] <jgraham>
My experience is mostly with NASA's ADS which can provide .bib

[01:33:01.0000] <jgraham>
A link to a bib file is a fine solution to the problem as long as a) you are using TeX and b) the information in the .bib is correct

[01:34:00.0000] <jgraham>
(my understanding is that in life sciences research, LaTeX is very uncommon and citations are usually in some other format)

[01:36:00.0000] <hsivonen>
jgraham: I used .bib without TeX

[01:36:01.0000] <hsivonen>
jgraham: off-the-shelf .bib parsers FTW

[01:37:00.0000] <Philip`>
Google Scholar provides BibTeX entries too

[01:37:01.0000] <Philip`>
(among other formats)

[01:37:02.0000] <jgraham>
hsivonen: What exactly did you do? Preprocess a HTML document adding bibtex-extracted information?

[01:38:00.0000] <hsivonen>
jgraham: preprocess an XHTML document to add a bibliography outputting both HTML and XHTML

[01:39:00.0000] <jgraham>
hsivonen: Code?

[01:39:01.0000] <jgraham>
(it's not important, I'm just curious)

[01:39:02.0000] <hsivonen>
jgraham: http://hsivonen.iki.fi/thesis/ has a link

[01:39:03.0000] <hsivonen>
http://hsivonen.iki.fi/thesis/bib4ht-0.9.tar.gz

[01:40:00.0000] <jgraham>
hsivonen: Thanks

[01:45:00.0000] <Philip`>
/me notes that the ACM people ask for the .tex or .doc source of submitted papers, and in the .tex case it must include the post-processed bibliography section (i.e. the .bbl file, not the .bib), so they can extract the metadata for the ACM digital library

[01:45:01.0000] <Philip`>
but I guess in the .doc case they'll have to extract it manually

[01:45:02.0000] <jgraham>
Philip`: The same is true on arXiv.org although I'm not sure they do anything useful with the bbl file

[01:48:00.0000] <jgraham>
It seems like "annotate structured data for private/small group use" and "using the data should not involve learing new vocabularies" are rather mutually contradictory requirements

[01:49:00.0000] <jgraham>
I'm also not sure what Amazon has to do with that use case since Amazon marking up its prices would presumably not be for small group use

[01:50:00.0000] <hsivonen>
jgraham: what if the small group already has a vocabulary

[01:51:00.0000] <hsivonen>
amazon is a group of one, I guess

[01:52:00.0000] <jgraham>
hsivonen: Then class values - which are specifically excluded as a solution seem fine

[01:52:01.0000] <jgraham>
I'm not really sure why the use case document is discussing solutions there at all to be honest

[02:04:00.0000] <Philip`>
http://www.charlespetzold.com/etc/CSAML.html

[02:07:00.0000] <Hixie>
class values are excluded as a solution?

[02:09:00.0000] <jgraham>
Hixie: Using the data should not involve learning a plethora of new APIs, formats or vocabularies [...] it's possible to get information from sites consistently using 'class' values in a documented way, but doing so requires learning a new vocabulary

[02:09:01.0000] <jgraham>
(that should have been in quotes)

[02:10:00.0000] <Hixie>
ah, yes. well, that's the requirement someone had.

[02:10:01.0000] <jgraham>
That seems to at least suggest that class values are not an acceptable solution

[02:10:02.0000] <Hixie>
i don't know if it makes sense particularly.

[02:11:00.0000] <jgraham>
The whole requirement makes no sense to me because I don't think it is possible to get at novel information without learning the associated vocabulary

[02:11:01.0000] <Hixie>
well this is why we right down the requirements without thinking about whether they are possible or not

[02:11:02.0000] <Hixie>
if it turns out we can't find a solution, then we see which requirement we'd be first willing to drop

[02:12:00.0000] <jgraham>
Oh, I didn't realise that reasonableness and logical consistency weren't requirements on requirements

[02:12:01.0000] <Hixie>
well it's hard to tell what's reasonable until you try to look for solutions

[02:13:00.0000] <jgraham>
This seems to fall more under logical consistency

[02:15:00.0000] <jgraham>
/me gets to a requirement with the words "generate custom UI" in, wishes the alarm bells and sirens would stop

[02:16:00.0000] <Hixie>
jgraham: it would be possible to come up with an API that solved all problems without the author having to learn any new languages

[02:17:00.0000] <Hixie>
jgraham: window.doTheRightThing(freeformTextInAnyNaturalLanguage);

[02:17:01.0000] <Hixie>
it might require a lot of work from implementations

[02:18:00.0000] <jgraham>
Hixie: Still doesn't work because the user would have to learn the right natural-language terms to use

[02:18:01.0000] <Hixie>
i could spec it so the method does the right thing regardless of which words are used :-)

[02:19:00.0000] <jgraham>
well yes, I guess window.readMyMind() would work for some definition of "work" which doesn't mind the fact that it may be impossible

[02:20:00.0000] <Hixie>
there's already research into mind-computer interaction

[02:20:01.0000] <Hixie>
it might not be as impossible as you think

[02:20:02.0000] <Hixie>
admittedly, unlikely to be suitable for a spec today

[02:20:03.0000] <jgraham>
Hixie: It's not the neural interface that I think is the hard part

[02:21:00.0000] <jgraham>
It's providing the answer to an ill-defined question provided by the neural interface

[02:21:01.0000] <Hixie>
fair enough

[02:21:02.0000] <Hixie>
anyway

[02:21:03.0000] <jgraham>
:)

[02:21:04.0000] <Philip`>
/me thought mind-computer interaction was currently limited to distinguishing two states and therefore playing Pong

[02:21:05.0000] <Hixie>
i don't want to exclude requirements or use cases until i consider them seriously

[02:22:00.0000] <Hixie>
and i don't want to do that until i have a complete list

[02:44:00.0000] <Philip`>
Someone should write a tool to generate names to use in use cases

[02:44:01.0000] <Philip`>
(Paul, David, Fred, etc)

[02:45:00.0000] <Philip`>
Ideally it would have a checkbox to make the names follow the pattern A*, B*, C*, etc, and sliders for gender diversity and racial diversity

[02:51:00.0000] <annevk2>
Hixie, the sending steps specify WebSocket-Protocol but the receiving steps specify websocket-protocol

[02:52:00.0000] <Hixie>
correct

[02:52:01.0000] <Hixie>
the receiving steps lowercase the header.

[02:54:00.0000] <annevk2>
hmm ok

[02:56:00.0000] <Hixie>
i was reading something recently about someone saying they liked minimal specs and comparing SVG and SMIL to canvas and other specs

[02:56:01.0000] <Hixie>
showing how the simpler ones were more effectively deployed and implemented

[02:56:02.0000] <Hixie>
but i can't work out where that is anymore

[02:56:03.0000] <Hixie>
anyone recognise what i'm talking about?

[02:57:00.0000] <Hixie>
might have been in the context of o3d...

[02:57:01.0000] <annevk2>
yes, it was in response to O3D by some Mozilla guy

[02:57:02.0000] <annevk2>
right

[02:57:03.0000] <annevk2>
hmm

[02:57:04.0000] <Hixie>
aha

[02:57:05.0000] <Hixie>
http://www.0xdeadbeef.com/weblog/?p=1223

[02:57:06.0000] <Hixie>
google rocks

[02:58:00.0000] <Hixie>
first hit for [o3d svg smil canvas]

[02:59:00.0000] <Philip`>
He seems to be overstating how high-level O3D is

[03:01:00.0000] <Philip`>
Something like X3D, where you can declare a sphere and it will be drawn, is much higher level than O3D, where you have to create a render graph that says to clear the buffer and traverse the object tree and execute a draw-list which is a lot of draw-elements, etc

[03:02:00.0000] <Hixie>
luckily for me, 3d is far outside my area of expertise

[03:04:00.0000] <Philip`>
(http://coderhump.com/archives/427 says some possibly interesting things)

[03:21:00.0000] <hsivonen>
I guess 3D solutions depend a lot on use cases

[03:21:01.0000] <hsivonen>
declarative scene graphs probably aren't important to have in browsers if you want to address gaming

[03:22:00.0000] <hsivonen>
but if you want to address rotatable e-commerce product samples, declarative scene graphs would be nice

[03:24:00.0000] <Philip`>
I don't think they help much with rotatable e-commerce product samples - you'd get a mesh from a 3D modelling program, use a standard library that renders it and rotates it using whatever underlying API there is, and that's that

[03:24:01.0000] <Philip`>
Declarative scene graphs (like in X3D) seem much more useful when people are creating complex 3D scenes, e.g. virtual worlds

[03:25:00.0000] <hsivonen>
Philip`: I guess that depends on one's philosophical outlook on the Rule of Least Power

[03:25:01.0000] <Philip`>
(either by hand or using a special editing program)

[03:26:00.0000] <Philip`>
(People who I've heard using X3D seem to generally claim they like it because they don't have to be programmers to use it)

[03:26:01.0000] <hsivonen>
Philip`: there were an open multivendor declarative format with quality renderers in all browsers, wouldn't you want to use that on an ecommerce site?

[03:27:00.0000] <Philip`>
hsivonen: I don't see any reason why I'd prefer it to an open multivendor imperative format with quality renderers in all browsers

[03:27:01.0000] <Philip`>
s/format/API/

[03:27:02.0000] <jgraham>
Are there any real-life examples of non-games where 3D is adding significant value?

[03:27:03.0000] <jgraham>
(and I guess non-specialised application i.e. not medical imaging)

[03:27:04.0000] <hsivonen>
(I very well do see the point that the game-oriented API is needed and has faster time-to-market, so it may be that the scene graph approach never takes off)

[03:27:05.0000] <Philip`>
jgraham: Modern desktop UIs

[03:28:00.0000] <Philip`>
with their fancy zooms and rotating windows and stuff

[03:28:01.0000] <jgraham>
Philip`: That's not really 3D is it?

[03:28:02.0000] <jgraham>
I mean it's using the 3D card but most of the useful affects are really 2D

[03:28:03.0000] <jgraham>
/a/e/

[03:28:04.0000] <hsivonen>
jgraham: pseudo-3D (QuickTime VR) is used for showing shiny Apple products after Stevenotes

[03:29:00.0000] <MikeSmith>
jgraham: is 2nd life a non-game? does it use real 3D?

[03:29:01.0000] <hsivonen>
/me doesn't understand the use cases of 2nd life

[03:29:02.0000] <hsivonen>
(I've never used it)

[03:29:03.0000] <jgraham>
MikeSmith: I don't know if 2nd life is a non-game. I always saw it as a game

[03:29:04.0000] <jgraham>
(but I have also never used it)

[03:30:00.0000] <Philip`>
jgraham: They're generally mostly-2D-ish effects, but sometimes they're really 3D, like the window-switcher in Vista

[03:30:01.0000] <Philip`>
and other effects like the ripples when dropping widgets in OS X are only really possible with 3D technology like pixel shaders

[03:30:02.0000] <hsivonen>
Philip`: those are for CSS 3D transitions--not the 3D scene rendering stuff whatever that one ends up being

[03:31:00.0000] <Philip`>
Also there's stuff like the Cover Flow UI

[03:31:01.0000] <jgraham>
Philip`: Fair enough. That sounds like it needs a low-level API though

[03:31:02.0000] <hsivonen>
how does O3D compare to Opera's canvas 3D from 2006?

[03:32:00.0000] <Philip`>
O3D differs in that it has features

[03:32:01.0000] <Philip`>
and a scene graph

[03:32:02.0000] <MikeSmith>
I would hope that some of the use cases of 2nd life are the same as real life (and not just the same as games)

[03:32:03.0000] <hsivonen>
oh, I thought Opera had a scene graph

[03:32:04.0000] <Philip`>
Opera's thing just had some functions for transforming the viewpoint, and global state parameters for colours and textures, and then functions for drawing triangles and for drawing models (which are collections of triangles), and that was pretty much it

[03:34:00.0000] <Philip`>
Google Earth and Google Street View are obvious uses of 3D

[03:34:01.0000] <jgraham>
Ah, Google street view is quite a nice example I guess

[03:34:02.0000] <Philip`>
And if you have Google Earth in a browser, you might also want an editing program like SketchUp in there too

[03:34:03.0000] <hsivonen>
would one want to hand a declarative model of the Earch including building to a browser and let it figure out LOD and stuff?

[03:35:00.0000] <hsivonen>
*Earth

[03:35:01.0000] <hsivonen>
I do want browsers to obsolete the dedicated Google Earth desktop app

[03:36:00.0000] <Philip`>
hsivonen: You could do something like http://www.web3d.org/x3d/specifications/ISO-IEC-19775-1.2-X3D-AbstractSpecification/Part01/components/geodata.html#GeoLOD

[03:36:01.0000] <hsivonen>
anyway, moving Google Earth into a browser seems like a use case for a low-level rendering API--not for a standard scene graph

[03:36:02.0000] <Philip`>
where your declarative model of the Earth is a low-LOD version, and it declares the URLs of recursively higher-LOD versions

[03:37:00.0000] <hsivonen>
Philip`: ok. makes sense

[03:37:01.0000] <hsivonen>
/me wonders how many polygons a URL is worth in terms of bytes

[03:37:02.0000] <Philip`>
Not much fun if you're the first person in the world to want to do a 3D model of the Earth and the relevant features haven't been added to the declarative spec and all the implementations yet

[03:38:00.0000] <Philip`>
unless it provides enough low-level primitives that you can emulate the right effect with decent performance

[03:39:00.0000] <Philip`>
hsivonen: Use TinyURL!

[03:41:00.0000] <hsivonen>
It would be cool to have a .pov scene renderer on top of canvas 3D

[03:41:01.0000] <MikeSmith>
so Google Street View seems potentially related to the Social Web to me -- at least more so than widgets (which are mentioned specifically in the Social Web WG charter are)

[03:41:02.0000] <hsivonen>
procedural textures would be hard to convert, though

[03:42:00.0000] <hsivonen>
in 2002, I wrote a simple generator that took OpenGL calls and emitted .pov

[03:42:01.0000] <hsivonen>
when our team found that we wanted a scene graph for pre-rendering textures in POV-Ray when we already had the data entangled in the source of a C program

[03:44:00.0000] <Philip`>
hsivonen: Convert procedural textures into GLSL :-)

[03:47:00.0000] <Philip`>
Oh, another use case for 3D is to take the Web 2.0 idea of logos-on-a-reflective-surface and extend it to logos-on-a-reflective-surface-that-is-animated-like-rippling-water

[03:48:00.0000] <Philip`>
Also: adverts

[03:48:01.0000] <Philip`>
which need to be as visually stimulating as possible, while not taking so long to download that the user has left the page before it's shown

[03:51:00.0000] <hsivonen>
/me notes that Google became successful in the ad business by having ads that aren't visually stimulating

[03:55:00.0000] <Philip`>
But obnoxious adverts still exist, so they're a valid use case

[03:55:01.0000] <Hixie>
hsivonen: wrong lesson. :-)

[03:55:02.0000] <Hixie>
hsivonen: we became successful by having them be useful

[03:58:00.0000] <Philip`>
"rel=next failed because UAs didn't expose it because authors didn't use it because UAs didn't expose it" - Opera's approach seems like a solution to that, since the 'forward' button will go to the rel=next page (if there is one) else it'll use some heuristics to guess where to go next

[03:59:00.0000] <Philip`>
and the heuristics are the way to bootstrap the process, so users will use the feature, and then authors will find it useful to improve the feature by using rel=next

[03:59:01.0000] <Hixie>
ok, rel=up

[03:59:02.0000] <jgraham>
/me disagrees with Philip` 

[03:59:03.0000] <Hixie>
but in general yes, if the browsers did something, it would be more useful

[03:59:04.0000] <Hixie>
rel=chapter might be an even better example

[03:59:05.0000] <jgraham>
Because I guess users don't use features that have unpredicatable behaviour

[04:00:00.0000] <Philip`>
jgraham: It might not actually work in practice, but I still think it seems like a solution :-)

[04:00:01.0000] <Philip`>
It'd be okay if the heuristics were generally reliable

[04:00:02.0000] <jgraham>
That's the problem with things that sound like they ought to work

[04:00:03.0000] <Philip`>
but the next-page heuristics in Opera only seem to work about half the time I try to use it

[04:01:00.0000] <hsivonen>
Fast Forward is the #1 Opera feature I miss in Firefox

[04:01:01.0000] <Philip`>
/me likes how it works at navigating Apache directory listings, automatically knowing the next file in the sequence

[04:04:00.0000] <hsivonen>
/me notes that Opera Mobile is pretty good at locating the start of real content after boilerplate cruft even without HTML5 structural elements and without ARIA

[06:05:00.0000] <Philip`>
Hmm, did Sourceforge recently change its bug tracker to be even less usable?

[06:05:01.0000] <Philip`>
In particular it seems to hide all the comments by default

[06:06:00.0000] <Philip`>
(And the comments are in reverse chronological order, so you have to read up the page to understand the conversation)

[06:51:00.0000] <Lachy>
Geocities has finally collapsed. http://tech.slashdot.org/article.pl?sid=09/04/23/2339224

[06:51:01.0000] <Lachy>
"So Long, GeoCities: We Forgot You Still Existed" - LOL :-) http://www.pcworld.com/article/163765/so_long_geocities_we_forgot_you_still_existed.html

[06:52:00.0000] <Philip`>
/me hopes the content will be archived somewhere

[06:54:00.0000] <Philip`>
I remember having fun searching for a nice-looking four-digit number in TimesSquare to create my site

[06:56:00.0000] <Lachy>
I guess the archive team might have a go at saving as much as they can. http://archiveteam.org/

[06:57:00.0000] <Lachy>
it's already on their Deathwatch list http://archiveteam.org/index.php?title=Deathwatch

[07:41:00.0000] <beowulf>
if i had a gecko based html5 rendering bug, where'd be the best place to get help with defining it?

[07:43:00.0000] <jgraham>
beowulf: That might depend on exatly what you mean by a "gecko based HTML5 rendering bug"

[07:45:00.0000] <beowulf>
jgraham: yeah, i was pretty sure that I'd get the description of what i want wrong too :)

[07:46:00.0000] <Philip`>
This is a good place to get help with defining the description of what your problem is :-)

[07:46:01.0000] <beowulf>
i have nested lists which contain <a href=""><h2>foo</h2><p>bar</p></a> 

[07:48:00.0000] <Philip`>
That sounds potentially problematic

[07:48:01.0000] <beowulf>
when this list gets repeated a few times the dom on one of these is different for one of the li's near the end on ff & camino

[07:48:02.0000] <beowulf>
i've tried reducing it here http://qa.getexceptional.com/test.html

[07:48:03.0000] <jgraham>
/me guesses that camino is using an older version of gecko or something

[07:49:00.0000] <beowulf>
what happens is the <a> closes and reopens beofre and after each of the elements it wraps

[07:49:01.0000] <beowulf>
but only on one, and only when i make the list over a certain size

[07:50:00.0000] <beowulf>
in that test you'll see one set of elements bordered, this is the problem

[07:50:01.0000] <Philip`>
Hmm, they all look the same to me

[07:50:02.0000] <beowulf>
i hope you see it anyway

[07:50:03.0000] <Philip`>
(in Firefox 2)

[07:51:00.0000] <beowulf>
we're using ff3.02

[07:51:01.0000] <beowulf>
and camino 1.6.7

[07:51:02.0000] <Philip`>
/me hypothesises that it's a packet-boundary issue

[07:52:00.0000] <beowulf>
Philip`: explain more, as it does seem limited to an ec2 instance

[07:52:01.0000] <Philip`>
Ah, I see the bordering in FF3

[07:54:00.0000] <Philip`>
/me checks in Wireshark

[07:54:01.0000] <Philip`>
Definitely looks like a packet-boundary issue :-)

[07:55:00.0000] <Philip`>
The HTTP response is split into two packets, and the boundary happens to be inside one of the <a>s

[07:56:00.0000] <Philip`>
and Firefox's parser does something weird which means it ends up closing various elements, before receiving the rest of the HTML in the next packet

[07:56:01.0000] <Philip`>
If you reload from cache, or if the packet boundary is outside the <a>, then you won't experience that bug

[07:57:00.0000] <Philip`>
(By "inside the <a>", I mean somewhere inside the content of the element, not inside the actual tag itself)

[07:57:01.0000] <beowulf>
which is fine if you're not putting a background on the <a>

[07:58:00.0000] <Philip`>
Putting <a> around block elements is probably a bad idea in general, given how Firefox parses such things

[07:58:01.0000] <Philip`>
but I think(?) that doing <a><span>...stuff with block elements...</span></a> will trick it into working sensibly

[07:59:00.0000] <beowulf>
Philip`: cool stuff, thanks for that

[08:00:00.0000] <Philip`>
hsivonen: Has someone already suggested that the validator should warn when you put block elements in <a> because it confuses things like Firefox?

[08:00:01.0000] <Philip`>
beowulf: Oh, and this is probably the best place to ask such questions :-)

[08:01:00.0000] <beowulf>
http://meyerweb.com/eric/thoughts/2009/01/02/an-event-apart-and-html-5/ # tell meyer too :)

[08:01:01.0000] <Rik|work>
Philip`: why the validator should warn ? If it's allowed by the spec then it shouldn't warn

[08:01:02.0000] <beowulf>
Philip`: noted :)

[08:02:00.0000] <Philip`>
beowulf: Hmm, I read that when it was posted but didn't realise it was probably this packet-boundary issue

[08:02:01.0000] <hsivonen>
Philip`: I don't remember. However, there's this new parser thing in the works for Firefox. :-)

[08:02:02.0000] <jgraham>
/me maintains that the <a>-around-elements thing is a silly spec hack

[08:02:03.0000] <Philip`>
hsivonen: Doesn't help people like beowulf or meyer who are trying to write web pages that work today, and encounter crazy impossible-to-debug bugs

[08:03:00.0000] <jgraham>
s/elements/block-level-elements

[08:03:01.0000] <beowulf>
Philip`: what meyer is doing wouldn't neccessarily expose it, i don't think

[08:03:02.0000] <beowulf>
it was obvious to me because the <a>'s were styled as blocks with height and a background image

[08:03:03.0000] <beowulf>
(in my work)

[08:04:00.0000] <Philip`>
Rik|work: Because it helps inform people who try using the new HTML5 feature/hack of putting <a> around block elements and discover that it doesn't quite work as expected

[08:04:01.0000] <Rik|work>
Philip`: then you'll have to warn for all the features that are not yet implemented

[08:04:02.0000] <Rik|work>
or badly implemented

[08:05:00.0000] <Philip`>
Rik|work: They're usually pretty obvious and easy to understand, when a feature is just missing or really broken; whereas the <a> parsing thing is highly non-obvious to anybody who hasn't read Hixie's blog post from several years ago discussing the issue :-)

[08:05:01.0000] <Philip`>
beowulf: He says "What I didn’t want, though, was the randomized layout weirdness that resulted once I started styling the descendants of the link. Sometimes everything would lay out properly, and other times the bits and pieces were all over the place. I could (randomly) flip back and forth between the two just by repeatedly hitting reload." so it sounds like he exposed that bug

[08:06:00.0000] <beowulf>
Philip`: i should start reading the prose as well as the html :)

[08:06:01.0000] <jgraham>
We should really avoid new features that require validator warnings

[08:14:00.0000] <Philip`>
Oh, the <a><span>block</span></a> trick is invalid HTML5 :-(

[08:17:00.0000] <Philip`>
jgraham: But the warnings could be removed in maybe five years from now, and then we'd have a useful widely-deployed feature that people could use, whereas five years from now we won't want to still be changing the definition of valid HTML5

[08:18:00.0000] <Philip`>
(and we'll all have got bored with HTML and moved on to other areas, so there won't be an HTML6 that we could add the feature to)

[09:32:00.0000] <Philip`>
Hmm, http://www.dotnetdotcom.org/ is now providing 600,000 pages via Amazon S3

[09:33:00.0000] <Philip`>
and transferring data between S3 and EC2 is free, so it could be sensible to run analyses on EC2

[09:34:00.0000] <Philip`>
but if it's only 3.2GB I could just download it onto my own computer, which has much more CPU power than sensibly-priced EC2 instances

[09:36:00.0000] <Philip`>
Not much point using the torrent since there's only one seed...

[09:37:00.0000] <Philip`>
(and no other peers)

[09:40:00.0000] <Philip`>
Looks like it's restricted to text/html files

[09:46:00.0000] <beowulf>
Philip`: any light reading you could recommend that might help me understand packet boundary issues?

[09:49:00.0000] <Philip`>
beowulf: Not quite sure what you're looking for - packets in general, or their effect on Firefox parsing of HTML, or something?

[09:51:00.0000] <beowulf>
Philip`: yeah, maybe I mean a "how a browser works" kind of read

[09:54:00.0000] <Philip`>
beowulf: Hmm, still not quite sure what you're looking for :-)

[09:55:00.0000] <Philip`>
The weirdness in Firefox isn't documented anywhere as far as I'm aware, but it's mentioned in http://ln.hixie.ch/?start=1138169545&count=1

[09:55:01.0000] <Philip`>
Packets in general are just like http://en.wikipedia.org/wiki/Packet_(information_technology) (there are probably much better references but I don't know of any in particular)

[10:00:00.0000] <beowulf>
Philip`: thanks for the hixie.ch link (packets in general is fine)

[10:04:00.0000] <Philip`>
gzip is slow :-(

[10:04:01.0000] <Philip`>
particularly when it's a 15GB file that's been gzipped

[10:04:02.0000] <Philip`>
/me supposes he should split it into smaller chunks for future processing

[10:05:00.0000] <Philip`>
(to benefit from multi-core goodness)

[10:16:00.0000] <gsnedders>
w00t

[10:17:00.0000] <gsnedders>
just had Flash crash my browser for the second time in as many days

[10:23:00.0000] <Philip`>
/me tries grepping the dotnetdotcom data

[10:23:01.0000] <Philip`>
Hmm, quite a lot of rel="canonical"

[10:56:00.0000] <gmiernicki>
fuck flash... <audio> <video> is teh future!

[10:56:01.0000] <gmiernicki>
uh, <canvas> too :)

[10:57:00.0000] <gmiernicki>
actually kinda wondering lately if video and canvas tags can be combined

[10:57:01.0000] <gmiernicki>
perhaps embedding a <video> in a <canvas>

[10:57:02.0000] <Rik|work>
gmiernicki: you can in firefox

[10:57:03.0000] <Rik|work>
paul rouget did some nice demos

[10:58:00.0000] <gmiernicki>
interesting, ill see if i can find em 

[11:04:00.0000] <gmiernicki>
amazing

[11:04:01.0000] <gmiernicki>
http://people.mozilla.com/~prouget/demos/DynamicContentInjection/play.xhtml

[11:04:02.0000] <gmiernicki>
just what i was wondering ;D

[11:04:03.0000] <gmiernicki>
drawImage() is more powerful than i thought

[11:31:00.0000] <Philip`>
Based on data from presumably quite recently, rel="canonical" is approximately the 33rd most common rel/rev value

[11:33:00.0000] <Philip`>
hsivonen: Hmm, I guess it's bad if I get java.nio.BufferUnderflowException when calling nu.validator.htmlparser.sax.HtmlParser.parse?

[11:35:00.0000] <Philip`>
/me hopes he's not just doing something stupid himself

[11:52:00.0000] <Philip`>
hsivonen: http://philip.html5.org/misc/gumpu-rundiary.txt

[11:52:01.0000] <Philip`>
using htmlparser-1.2.0

[11:52:02.0000] <Philip`>
and new HtmlParser(XmlViolationPolicy.ALLOW);

[11:53:00.0000] <Philip`>
I get:

[11:53:01.0000] <Philip`>
Exception in thread "main" java.nio.BufferUnderflowException at java.nio.Buffer.nextGetIndex(Buffer.java:474) at java.nio.HeapByteBuffer.get(HeapByteBuffer.java:117) at nu.validator.htmlparser.io.HtmlInputStreamReader.read(HtmlInputStreamReader.java:297)

[11:53:02.0000] <Philip`>
etc

[11:53:03.0000] <Philip`>
(If I make the file any smaller then the problem seems to go away)

[12:09:00.0000] <Philip`>
(That error came from gumpu.rundiary.co.kr/content.asp?menu=550&eid=6136&title= )

[12:09:01.0000] <Philip`>
(I get the same on www.aladdin.co.kr/shop/wproduct.aspx?ISBN=8956743339 )

[12:09:02.0000] <Philip`>
(I blame Korea)

[12:26:00.0000] <aboodman>
slightlyoff: interesting thread about EventTarget

[12:26:01.0000] <aboodman>
I'm not sure I get the value of doing this natively though

[12:26:02.0000] <aboodman>
poo

[12:26:03.0000] <slightlyoff_>
aboodman: eh, the proposal that arv made doesn't go as far as I'd lke

[12:26:04.0000] <slightlyoff_>
still here = )

[12:27:00.0000] <slightlyoff_>
basically I'd like all methods on objects to be listenable by default

[12:27:01.0000] <aboodman>
is it just another example of something like bind() that should have been there from the beginning,or is there mroe?

[12:27:02.0000] <aboodman>
basically, is there something here that isn't possible in js

[12:27:03.0000] <slightlyoff_>
instead of needing to create/implement EventTarget objects/interfaces

[12:27:04.0000] <slightlyoff_>
aboodman: no, but I think that's the wrong test these days

[12:27:05.0000] <aboodman>
sure, just trying to understand the proposal :)

[12:28:00.0000] <aboodman>
not arguing against

[12:28:01.0000] <slightlyoff_>
aboodman: we've gone a decade on "the minimum you could theoretically get away with"

[12:28:02.0000] <slightlyoff_>
I'd like us to start working up the food chain to "something you might actually want to use"

[12:28:03.0000] <aboodman>
it helps for the proposal to clarify whether this is a new fundamental capability though, think.

[12:28:04.0000] <slightlyoff_>
it's a unification

[12:28:05.0000] <aboodman>
one new fundamental capability i think ... someobdy .. should think about

[12:28:06.0000] <aboodman>
weak references

[12:28:07.0000] <slightlyoff_>
yes

[12:28:08.0000] <aboodman>
an event system w/o weak references kinda sucks

[12:28:09.0000] <slightlyoff_>
agreed

[12:28:10.0000] <slightlyoff_>
I'd like that a lot

[12:29:00.0000] <slightlyoff_>
and JS needs weakrefs anyway

[12:29:01.0000] <slightlyoff_>
else you leak badly

[12:29:02.0000] <aboodman>
yes, that's what i mean

[12:29:03.0000] <slightlyoff_>
just by dint of doing stuff that allocates

[12:29:04.0000] <aboodman>
separately from this proposal, a weak ref proposal would be useful

[12:29:05.0000] <aboodman>
we can already implement this very easily in v8

[12:29:06.0000] <slightlyoff_>
I was thinking that = )

[12:29:07.0000] <slightlyoff_>
just a different kind of pointer wrapper = )

[12:29:08.0000] <aboodman>
i mean v8 already has native weak refernces

[12:29:09.0000] <aboodman>
i don't know about the other engines

[12:30:00.0000] <aboodman>
i assume they must have something similar too

[12:30:01.0000] <Philip`>
http://philip.html5.org/data/rel-rev-200904.txt - now with some rel=canonical

[12:31:00.0000] <jgraham>
Google code is getting mercurial! Awesome

[12:31:01.0000] <takkaria>
orly?

[12:31:02.0000] <takkaria>
that's interesting

[12:31:03.0000] <Philip`>
Not Git?

[12:31:04.0000] <Philip`>
/me vaguely remembers that Sourceforge added both

[12:32:00.0000] <Philip`>
(which is strange because it took them five years to add Subversion support)

[12:32:01.0000] <Philip`>
(Warning: "five" is a totally made up number and may not reflect reality)

[12:33:00.0000] <slightlyoff_>
Philip`: didn't Google Code *start* on SVN?

[12:33:01.0000] <aboodman>
yes.

[12:33:02.0000] <slightlyoff_>
oh, right, misread

[12:33:03.0000] <aboodman>
the backend wasn't svn though

[12:34:00.0000] <Philip`>
slightlyoff_: I meant Sourceforge which was stuck with CVS until long after it had gone out of fashion :-)

[12:34:01.0000] <slightlyoff_>
Philip`: yeah, my bad

[12:35:00.0000] <Philip`>
http://google-code-updates.blogspot.com/2009/04/mercurial-support-for-project-hosting.html - "Our implementation of Mercurial is built on top of Bigtable, making it extremely scalable and reliable just like our Subversion on Bigtable implementation." - they wrote their own Mercurial and Subversion servers?

[12:36:00.0000] <Philip`>
(or new storage backends for the existing server code or something?)

[12:36:01.0000] <aboodman>
Philip`: basically; they implement the protocols

[12:36:02.0000] <slightlyoff_>
Philip`: you didn't think they were going to get to high availability on the back of fsfs, did you?

[12:38:00.0000] <Philip`>
/me has only had FSFS get horribly corrupted once in several thousand commits

[12:57:00.0000] <gsnedders>
http://www.flickr.com/photos/gsnedders/3470720119/ — oh noes! it's mes!

[13:04:00.0000] <gsnedders>
Is it bad that just having annevk link to my site caused a notable spike in traffic?

[13:31:00.0000] <Philip`>
gsnedders: I get notable spikes in traffic just from being linked from comments in blogs


2009-04-25
[01:27:00.0000] <yorick>
the html5 boolean attributes seem to be inconsistent with the html4 ones

[01:28:00.0000] <yorick>
html4: <option selected="selected">contents</option>

[01:28:01.0000] <yorick>
html5: <div draggable="true">contents</div>

[01:34:00.0000] <yorick>
also, is there a possibility to set DataTransfer on dragstart to insensitive, so it can be accessed when dragging over something?

[01:41:00.0000] <jgraham>
Philip`: Can you get data on things that start <!-- but with a > and no --> before the end of the document

[01:41:01.0000] <jgraham>
/me isn't quite sure how to express that as a regexp

[02:11:00.0000] <annevk5>
yorick, draggable is not a boolean attribute per HTML5

[02:24:00.0000] <yorick>
annevk5: then what is it?

[02:36:00.0000] <zcorpan>
yorick: an enumerated attribute

[02:52:00.0000] <zcorpan>
/me adds another entry to http://wiki.whatwg.org/wiki/HTML5_Presentations

[04:37:00.0000] <gsnedders>
Ergh…

[04:37:01.0000] <gsnedders>
/me doesn't want to sign up to another bug tracker to report a bug in Validator.nu

[04:54:00.0000] <Philip`>
jgraham: You mean something like, um, /<!--[^>]*(?<!-->)>([^>]|(?<!-->)>)*$/ perhaps?

[04:55:00.0000] <Philip`>
Whoops, not that one

[04:55:01.0000] <Philip`>
/<!--[^>]*(?<!--)>([^>]|(?<!--)>)*$/

[04:58:00.0000] <jgraham>
Philip`: Perhaps

[04:59:00.0000] <jgraham>
If that matches "<!-- foo > bar" and <!-- foo>" but not "<!-- foo > bar -->"

[05:00:00.0000] <Philip`>
It does

[05:00:01.0000] <Philip`>
/me should take this opportunity to hook his grep tool up to his new set of pages...

[05:00:02.0000] <jgraham>
Ah, well that sounds like what I want then

[05:01:00.0000] <Philip`>
/me notes that "(?<!--)" confusingly has nothing to do with the string "<!--", it's just a negative lookbehind assertion on the string "--"

[05:02:00.0000] <jgraham>
Ah, that makes a little more sense

[05:02:01.0000] <gsnedders>
And this is why you shouldn't use regex to parse HTML P

[05:02:02.0000] <gsnedders>
* :P

[05:03:00.0000] <jgraham>
gsnedders: No this is why should shouldn't have ultra-weird comment parsing

[05:03:01.0000] <gsnedders>
jgraham: It's saner than SGML.

[05:03:02.0000] <jgraham>
which requires lookhead

[05:04:00.0000] <jgraham>
s/should/you/

[05:05:00.0000] <gsnedders>
I don't. I have perfectly sane comment parsing, thank you very much.

[05:06:00.0000] <jgraham>
gsnedders: You use a sophisticated biological neural network to parse comments and it's not even that reliable. How can you describe that as sane?

[05:06:01.0000] <gsnedders>
jgraham: Through my own insaity.

[05:06:02.0000] <gsnedders>
*insanity.

[05:10:00.0000] <Philip`>
gsnedders: Parsing it's easy, it's just like /<!(-?>|--.*?-->)/

[05:10:01.0000] <Philip`>
s/it's/is/

[05:11:00.0000] <Philip`>
Uh

[05:11:01.0000] <Philip`>
gsnedders: Parsing it's easy, it's just like /<!(-?>|([^-]|--).*?-->)/

[05:11:02.0000] <Philip`>
s/it's/is/

[05:11:03.0000] <Philip`>
or something like that

[05:12:00.0000] <Philip`>
but anyway it's easy

[05:12:01.0000] <MikeSmith>
gsnedders: I'm looking at the class="" bug now

[05:12:02.0000] <Philip`>
The difficulty is trying to match things that are *not* matched by the normal state machine

[05:24:00.0000] <gsnedders>
How the hell do you get script to validate in HTML 4.01?

[05:24:01.0000] <gsnedders>
/me stabs SGML

[05:27:00.0000] <Philip`>
Use <script src>

[05:27:01.0000] <Philip`>
If you want inline scripts, use <script src="data:text/javascript,...">

[05:30:00.0000] <Dashiva>
gsnedders: Just rephrase all your less-than tests :)

[05:39:00.0000] <MikeSmith>
from the HTML4 spec, I can't tell whether id and class are allowed to be empty or not

[05:43:00.0000] <Dashiva>
It says "must begin with a letter" for id

[05:44:00.0000] <Dashiva>
But it's taken from the SGML spec, so would probably have to look there

[05:46:00.0000] <MikeSmith>
hmm, the XHTML DTD defines the value of class as NMTOKENS

[05:46:01.0000] <gsnedders>
Yay! More undocumented differences between XHTML 1.0 and HTML 4.01!

[05:48:00.0000] <MikeSmith>
I think as far as the HTML4 and XHTML1 specs are concerned, the value of class can't be empty

[05:48:01.0000] <Dashiva>
Can't class be a list of zero class names?

[05:49:00.0000] <MikeSmith>
Dashiva: not as far as I can see, as far as XML goes

[05:50:00.0000] <MikeSmith>
http://www.w3.org/TR/REC-xml/#NT-Nmtoken

[05:50:01.0000] <MikeSmith>
Nmtokens    ::=    Nmtoken (#x20 Nmtoken)*

[05:50:02.0000] <MikeSmith>
Nmtoken    ::=    (NameChar)+

[05:50:03.0000] <MikeSmith>
and XHTML1 DTD says:

[05:51:00.0000] <Dashiva>
Fair enough. Where does it say it's NMTOKENS? xhtml1 says CDATA where I'm looking

[05:51:01.0000] <MikeSmith>
"class        NMTOKENS                 #IMPLIED

[05:51:02.0000] <MikeSmith>
I'm looking at a DTD on my local system

[05:52:00.0000] <MikeSmith>
XHTML Modularization

[05:52:01.0000] <gsnedders>
Well, it claims to be an XHTML 1.0 schema, not an XHTML Mod. one

[05:53:00.0000] <Dashiva>
What I found: http://www.w3.org/TR/xhtml1/dtds.html#dtdentry_xhtml1-strict.dtd_coreattrs

[05:53:01.0000] <Dashiva>
I guess it's obsolete

[05:53:02.0000] <MikeSmith>
Dashiva: yeah, that's what I'm looking at now

[05:53:03.0000] <MikeSmith>
well, it's right, even if it's obsolete

[05:54:00.0000] <MikeSmith>
and XHTML modularization is wrong

[05:54:01.0000] <MikeSmith>
I mean, in practice at least

[05:54:02.0000] <Dashiva>
YSOD on empty class attribute? :)

[05:54:03.0000] <MikeSmith>
YSOD?

[05:55:00.0000] <Dashiva>
Yellow screen of death

[05:55:01.0000] <MikeSmith>
heh

[05:58:00.0000] <MikeSmith>
the HTML 4.01 DTD used by the W3C validator says "class       CDATA          #IMPLIED"

[05:59:00.0000] <MikeSmith>
and all the XHTML1 DTDs it uses says the same

[06:00:00.0000] <MikeSmith>
but the XHTML 1.1 DTD it uses says "class        NMTOKENS                 #IMPLIED"

[06:00:01.0000] <MikeSmith>
so those mooncalfs apparently redefined it in XHTML 1.1

[06:01:00.0000] <MikeSmith>
I think in in validator.nu we should preserve that brokenness 

[06:02:00.0000] <MikeSmith>
to learn people not to bother validating against XHTML 1.1

[06:03:00.0000] <MikeSmith>
but I see hsivonen had the foresight to not even include an XHTML1.1-checking option in validator.nu

[06:04:00.0000] <MikeSmith>
I wonder if they bothered to document this in the XHTML 1.1 spec

[06:05:00.0000] <MikeSmith>
nope

[06:05:01.0000] <MikeSmith>
http://www.w3.org/TR/xhtml11/changes.html#a_changes

[06:14:00.0000] <Philip`>
jgraham: With that search thing, how many results do you want?

[06:18:00.0000] <Philip`>
jgraham: Actually, you can have them all

[06:19:00.0000] <Philip`>
jgraham: http://philip.html5.org/data/comments-not-closed-but-with-a-gt-after-them.txt

[06:22:00.0000] <jgraham>
Philip`: Great, thanks

[06:30:00.0000] <Philip`>
/me hopes his regexp isn't wrong

[06:38:00.0000] <jgraham>
/me hasn't checked yet

[06:39:00.0000] <Philip`>
<!doctype html> seems to be one of those most popular HTML5 features

[06:39:01.0000] <Philip`>
/me sees it on 64 distinct domains

[06:40:00.0000] <Philip`>
like last.fm and pear.php.net and maps.google.com and edward.oconnor.cx and help.godaddy.com

[06:41:00.0000] <Philip`>
s/those most/the most/

[07:22:00.0000] <MikeSmith>
gsnedders: http://bugzilla.validator.nu/attachment.cgi?id=85

[07:26:00.0000] <MikeSmith>
hsivonen: ↑

[07:30:00.0000] <MikeSmith>
http://bugzilla.validator.nu/attachment.cgi?id=86

[09:11:00.0000] <zcorpan>
jgraham: have you documented <!-- and --> in web ecmascript yet?

[09:13:00.0000] <Philip`>
Hmm, I've seeded that dotnetdotcom index of web pages for almost 24 hours, and there's been zero connections

[09:13:01.0000] <Philip`>
Torrents aren't so useful for files that nobody wants to download anyway

[09:36:00.0000] <jgraham>
zcorpan: No, good point

[09:37:00.0000] <jgraham>
Philip`: I would quite like to download it but I'm not sure that it's a good idea over my crappy/bandwidth limited home connection

[09:38:00.0000] <Philip`>
jgraham: They provide plain HTTP download too, which might be more compatible with your connection

[09:39:00.0000] <Philip`>
and it's only 2.5GB, which is smaller than they claim

[09:41:00.0000] <Philip`>
And if you don't want all of it, you could just download the first N megabytes and discard the last entry

[09:41:01.0000] <Philip`>
where N is the largest value that still is considered a good idea to download

[13:46:00.0000] <jgraham>
Hmm, unless I am missing something, it looks like (almost?) all the cases of <!-- not followed by --> are in <script> blocks

[13:46:01.0000] <Philip`>
I looked at three, and saw one which wasn't

[13:46:02.0000] <Philip`>
but I don't know which one that was

[13:48:00.0000] <jgraham>
Yeah, my script is a bit buggy

[13:49:00.0000] <Philip`>
I hope mine wasn't

[13:49:01.0000] <Philip`>
/me wishes people would independently verify his data :-)

[13:49:02.0000] <jgraham>
/me would like to do that

[13:50:00.0000] <jgraham>
I will get some or more of that dotcomdotnet data at some point

[13:50:01.0000] <Philip`>
Also I hope their data isn't buggy

[13:51:00.0000] <jgraham>
That is, of course, quite possible

[13:51:01.0000] <Philip`>
I haven't seen any problems though

[13:52:00.0000] <Philip`>
except a few pages which seemingly return bogus HTTP responses that the HttpClient parser dies on

[13:52:01.0000] <Philip`>
but those could be legitimately broken servers

[13:53:00.0000] <Philip`>
I guess the main question is what direction their sample is most biased in

[13:54:00.0000] <Philip`>
(The sample of the web that they crawl, not the sample of their crawled data that they published (which they claim is uniform, and seemingly is restricted to 200s and text/html))

[13:54:01.0000] <jgraham>
Philip`: How is your sample biased? html5lib is telling me that all the sites are in no-quirks mode which seems unreasonable

[13:55:00.0000] <jgraham>
(the small selection I looked at were compatible with that hypothesis but that was like 2 sites)

[13:57:00.0000] <Philip`>
jgraham: I didn't do any sampling myself

[13:57:01.0000] <gsnedders>
Is html5lib reliable?

[13:57:02.0000] <Philip`>
Looking at a few random pages from my list, www.articlear.com/profile/Jason-Uvios/994 looks quirky

[13:58:00.0000] <jgraham>
gsnedders: No

[13:59:00.0000] <jgraham>
This is a good way of finding bugs in it though :)

[16:18:00.0000] <Hixie>
hsivonen: do you know if the spec reflects what you want wrt foster parenting?

[16:18:01.0000] <Hixie>
i have a comment suggesting you want something reverted but it seems already reverted


2009-04-26
[18:00:00.0000] <Scriber>
hello all

[18:00:01.0000] <Scriber>
silly question perhaps? does html5lib run under windows?

[18:01:00.0000] <Scriber>
and if so how do I install it?

[18:12:00.0000] <annevk5>
if you get Python to run under Windows it should work

[18:19:00.0000] <Scriber>
annevk5, sorry... how exactly? do i run 'setup.py' or copy files?

[01:35:00.0000] <Hixie>
http://www.reddit.com/r/web_design/comments/8fhdt/post_some_examples_of_semantically_well_markedup/c0947s0?context=2

[06:22:00.0000] <Scriber>
Hello all

[06:22:01.0000] <Scriber>
could someone possibly give me some instructions on installing html5lib under windows?

[06:23:00.0000] <Scriber>
when I run setup.py i get the error: No module named setuptools

[07:29:00.0000] <krijnh>
Are there any <datagrid> implementations already?

[07:37:00.0000] <Philip`>
Has anybody even read the <datagrid> specification yet?

[07:37:01.0000] <Philip`>
/me doesn't remember having seen anyone commenting on it

[07:40:00.0000] <krijnh>
Before the big (re)write last weeks, did anybody start with something in JS?

[07:40:01.0000] <krijnh>
Or wasn't that really possible?

[07:41:00.0000] <Philip`>
/me is unaware

[07:41:01.0000] <Philip`>
(though I think Hixie said he was doing a prototype implementation of the new datagrid, presumably in JS)

[07:42:00.0000] <krijnh>
Cool

[07:43:00.0000] <krijnh>
Is it possible to help in any way?

[07:43:01.0000] <Philip`>
I'd guess comments on the spec would be the most useful thing at this stage

[07:45:00.0000] <Philip`>
and perhaps attempts to use it to implement the purported use cases would be informative

[07:52:00.0000] <Philip`>
http://www.answers.com/ - <html class="Firefox"> or <html class="Opera"> etc (depending on UA) - I don't think I've seen that kind of browser sniffing before

[07:53:00.0000] <krijnh>
http://rafael.adm.br/css_browser_selector/ does something like it, although with JS

[08:28:00.0000] <ap>
hi fishd 

[08:29:00.0000] <ap>
wrong channel

[09:32:00.0000] <zcorpan>
Hixie: i think browsers do something with -->

[10:13:00.0000] <jgraham>
/me wonders what zcorpan ws referring to

[10:14:00.0000] <jgraham>
/me finds the answer

[10:17:00.0000] <gsnedders>
Warning: I may soon start asking maths questions every two minutes :P

[10:20:00.0000] <jgraham>
/me readies his stopwatch

[10:21:00.0000] <gsnedders>
/me finds he no longer has the 2008 second prelim paper

[10:21:01.0000] <gsnedders>
/me concludes someone in his class must have two copies of it

[10:21:02.0000] <gsnedders>
/me ponders, and concludes he'll do real final past papers instead

[10:38:00.0000] <gsnedders>
/me doesn't understand any of this

[15:04:00.0000] <gsnedders>
/me wonders if HAVE_FUTURE_DATA predicts the future

[15:05:00.0000] <gsnedders>
/me guesses he could look at the spec and find out it didn't, but that'd be boring

[15:57:00.0000] <Philip`>
http://philip.html5.org/data/attr-count-pages-dotbot.txt

[16:06:00.0000] <Hixie>
dotbot?

[16:08:00.0000] <annevk5>
follow the pointers

[16:08:01.0000] <annevk5>
http://philip.html5.org/data/dotbot-20090424.txt

[16:09:00.0000] <Philip`>
It's a shame nobody's invented a technology that lets me embed clickable links, so I'll have to stick with text/plain for now

[16:10:00.0000] <Philip`>
s/with/with the limitations of/

[16:16:00.0000] <annevk5>
i agree that HTML is a bit too complex for the requirements; it'll never fly

[16:20:00.0000] <Philip`>
Ooh, HTML? Didn't think of that

[16:20:01.0000] <krijnh>
Isn't HTML dead by now?

[16:20:02.0000] <annevk5>
pretty much, XML took over

[16:21:00.0000] <krijnh>
Yeah, it was inevitable

[16:21:01.0000] <annevk5>
I read this book the other day, all Web development will switch to XSLT+XML+XML Schema+XLink+XPointer pretty soonish

[16:21:02.0000] <annevk5>
much cleaner and all

[16:21:03.0000] <annevk5>
you know, you actually close the fricking tags

[16:22:00.0000] <krijnh>
I already use XML strict 1.2, I don't like going back to ugly tag soup!

[16:23:00.0000] <annevk5>
i'm more a transitional kind of guy myself

[16:25:00.0000] <krijnh>
Jij ook naar Eindhoven 8 mei?

[16:25:01.0000] <annevk5>
New York

[16:25:02.0000] <krijnh>
Grmbl :)

[16:26:00.0000] <annevk5>
kan het ook niet helpen dat jullie alles door m'n vakanties en werkdingen doorplannen :D

[16:26:01.0000] <krijnh>
Sorry :$

[16:27:00.0000] <annevk5>
niet dat m'n agenda publikelijk is

[16:27:01.0000] <krijnh>
Nee idd, wilde het net vragen ;)

[16:28:00.0000] <annevk5>
Lachy is bang dat we plannen aan het smeden zijn om de wereld over te nemen

[16:28:01.0000] <annevk5>
zeg iets verdachts

[16:28:02.0000] <krijnh>
Heb je die bom nog gecontroleerd?

[16:29:00.0000] <Hixie>
hsivonen?

[16:29:01.0000] <krijnh>
En heb je je Hixie masker nou eindelijk eens af? :)

[16:29:02.0000] <Hixie>
/me doesn't understand the difference between IE and Firefox on http://www.w3.org/mid/8EFCF205-8F9D-4098-B9DB-703A5DA7A680⊙if

[16:29:03.0000] <jcranmer>
. . .

[16:29:04.0000] <annevk5>
krijnh, lol ja, binnenkort op shop.whatwg.org

[16:30:00.0000] <krijnh>
Zeker alleen te betalen met Google Checkout?

[16:30:01.0000] <Hixie>
/me ce demande ce que krijnh in annevk5 sont en train de se dire

[16:30:02.0000] <Hixie>
s/in/et/

[16:31:00.0000] <Rik`>
Hixie: se demande ;)

[16:31:01.0000] <Lachy>
Als je van plan bent de wereld over te nemen, doe ik mee!

[16:31:02.0000] <annevk5>
/me thinks Lachy is cheating

[16:31:03.0000] <krijnh>
Hixie: ce n'est pas la langue de l'amour :)

[16:32:00.0000] <krijnh>
Lachy: ons eerste doel is jou dumpen, dus dat wordt nog best lastig ;]

[16:33:00.0000] <Hixie>
crap, it doesn't work if krijnh understands what i say!

[16:33:01.0000] <krijnh>
Of all languages you could have picked..

[16:34:00.0000] <Hixie>
unfortunately i really can only pick from two. :-)

[16:34:01.0000] <Lachy>
Jeg vil ta over verden uten deg

[16:35:00.0000] <krijnh>
That's why the Dutch will conquer the world!

[16:35:01.0000] <annevk5>
inderdaad; eerst met ons en nu zonder

[16:35:02.0000] <annevk5>
kun je niet echt op bouwen

[16:36:00.0000] <krijnh>
Lijkt redelijk op Nederlands

[16:37:00.0000] <annevk5>
Noors? ja

[16:37:01.0000] <annevk5>
ik begrijp ongeveer 50% en de rest verzin ik

[16:37:02.0000] <krijnh>
:)

[16:38:00.0000] <jcranmer>
Je declare que tous ici sont bête... la declaration m'inclut.

[16:38:01.0000] <jcranmer>
er, bêtes

[16:41:00.0000] <krijnh>
C'est ce que vous obtenez lorsque vous quittez votre sens de la logique a la porte :)

[16:41:01.0000] <jcranmer>
evidement

[16:42:00.0000] <Rik`>
s/quittez/laissez/

[16:42:01.0000] <krijnh>
Baggertaal

[16:42:02.0000] <jcranmer>
quelqu'un a utilisé l'ordinatuer pour sa traduction? Quel horreur!

[16:42:03.0000] <krijnh>
Mais non!

[16:44:00.0000] <krijnh>
/me hopes Hixie doesn't have access to Google Translate logs :>

[16:44:01.0000] <Rik`>
Le français est donc la langue officielle de HTML 5 ? La version anglaise ne sera pas normative ?

[16:45:00.0000] <jcranmer>
Emporte les fourches! Et les torches!

[16:45:01.0000] <jcranmer>
Nous devons demander la tête de krijnh!

[16:46:00.0000] <krijnh>
:o

[16:46:01.0000] <krijnh>
Pourquoi?

[16:47:00.0000] <jcranmer>
Vous avez torturé une langue! Par utiliser la traduction d'un ordinateur...

[16:48:00.0000] <Rik`>
jcranmer: then we need your head too for the last one :)

[16:48:01.0000] <krijnh>
Seulement pour un ligne des mots :)

[16:48:02.0000] <gavin_>
"Par utiliser"?

[16:48:03.0000] <jcranmer>
by using

[16:48:04.0000] <Rik`>
En utilisant

[16:48:05.0000] <krijnh>
Heh

[16:48:06.0000] <gavin_>
/me agrees with Rik` 

[16:49:00.0000] <jcranmer>
krijnh: deux lignes...

[16:49:01.0000] <jcranmer>
En tout cas, je peux vivre sans ma tête...

[16:50:00.0000] <krijnh>
Comme un poulet?

[16:50:01.0000] <jcranmer>
... ばか

[16:52:00.0000] <Rik`>
/me shouldn't have missed so many japanese lessons

[16:52:01.0000] <jcranmer>
I've taken less than you missed

[16:52:02.0000] <krijnh>
/me shouldn't use a non-unicode IRC client :)

[16:52:03.0000] <jcranmer>
FAIL

[16:52:04.0000] <krijnh>
I know

[16:53:00.0000] <krijnh>
Luckily the logs are okay

[16:53:01.0000] <krijnh>
Still doesn't make any sense, but whatever

[16:53:02.0000] <jcranmer>
/me attempts to translate that into Frech

[16:53:03.0000] <jcranmer>
er, French

[16:53:04.0000] <krijnh>
Translation: Japanese � English: Stupid

[16:54:00.0000] <jcranmer>
I'm betting that RATE wouldn't do it...


2009-04-27
[18:11:00.0000] <hsivonen>
Hixie: Re: http://lists.w3.org/Archives/Public/public-html/2009Mar/0503.html

[18:12:00.0000] <hsivonen>
Hixie: Firefox appends more text to the last text node it was appending to

[18:12:01.0000] <hsivonen>
Hixie: IE appends more text as child of the element it previously appended text as a child

[18:13:00.0000] <hsivonen>
Hixie: coalescing the text nodes if the existing last child is a text node

[18:14:00.0000] <hsivonen>
Hixie: so Firefox and IE are different if you insert an element from script between parser text flushes

[18:41:00.0000] <Hixie>
hsivonen: can you give an example of what you mean?

[18:41:01.0000] <hsivonen>
Hixie: I had an example in the email

[18:41:02.0000] <hsivonen>
looking it up

[18:43:00.0000] <hsivonen>
Hixie: http://hsivonen.iki.fi/test/moz/flushing-document-written-text-no-alert.html is different for you in Firefox 3.5 and IE8, right?

[18:43:01.0000] <Hixie>
i meant like a tiny example :-)

[18:44:00.0000] <Hixie>
is your example basically  <script>append a text node</script>x -- are there two text nodes or one?

[18:45:00.0000] <hsivonen>
Hixie: no, the crux is the line document.getElementById("foo").appendChild(document.createElement("span"))

[18:45:01.0000] <Hixie>
my confusion is with "Firefox appends more text to the last text node it was appending to", since if a script is running, the last text node it inserted into was the <script>'s child node

[18:46:00.0000] <Hixie>
you're saying you want     <script>write 'aaa'; append element node</script>bbb  to result in "aaabbb element node" rather than "aaa element node bbb"?

[18:46:01.0000] <Hixie>
i'm very confused

[18:47:00.0000] <hsivonen>
Hixie: former in Firefox, latter in IE

[18:47:01.0000] <Hixie>
well clearly imho aaabbb element is outright incorrect.

[18:48:00.0000] <hsivonen>
Hixie: you want to spec what IE does?

[18:48:01.0000] <hsivonen>
WebKit is element aaabbb, IIRC

[18:48:02.0000] <hsivonen>
and Opera

[18:48:03.0000] <Hixie>
the spec already requires the text after the <script> to come after whatever state the DOM is in after the script executes.

[18:50:00.0000] <Hixie>
element aaa bbb is definitely wrong per spec, it means the parser isn't re-entrant.

[18:50:01.0000] <Hixie>
(or acts as if it's not, which is worse)

[18:50:02.0000] <Hixie>
http://www.hixie.ch/tests/adhoc/html/parsing/026-demo.html

[18:51:00.0000] <hsivonen>
Hixie: the webkit/opera behavior flows naturally from a late-flushed accumulation buffer

[18:51:01.0000] <Hixie>
webkit's behaviour is clearly wrong.

[18:51:02.0000] <Hixie>
then a late-flushed accumulation buffer is a bad dea.

[18:51:03.0000] <Hixie>
idea

[18:51:04.0000] <hsivonen>
Hixie: it's 2 out of 4 :-)

[18:52:00.0000] <hsivonen>
I'll have to get back on this when I have a decent text input method

[18:52:01.0000] <Hixie>
IE's behaviour in http://www.hixie.ch/tests/adhoc/html/parsing/026-demo.html is what the spec requires

[18:52:02.0000] <Hixie>
and is the only behaviour i've found in the four browsers i tested that is even remotely sane.

[18:53:00.0000] <Hixie>
anyway, as i noted in my e-mail reply, just use the text node's own buffer as the buffer

[18:53:01.0000] <hsivonen>
Hixie: you have an interesting notion of sane

[18:53:02.0000] <Hixie>
and then you get IE's behaviour for free

[18:53:03.0000] <Hixie>
really? you don't think having things come out in the order they were written is sane?

[18:53:04.0000] <hsivonen>
Hixie: nope, then you get the Firefox behavior

[18:54:00.0000] <Hixie>
you only get firefox's behaviour if you don't check to make sure the text node is still the last child each time a script runs

[18:54:01.0000] <Hixie>
(which is a trivial pointer test)

[18:54:02.0000] <hsivonen>
Hixie: my independent implementation matched webkit and opera, and I like to think I'm sane

[18:54:03.0000] <Hixie>
i'm not saying the implementor isn't sane

[18:55:00.0000] <Hixie>
i'm saying that having nodes come out in an order different than the order they went in is unintuitive

[18:55:01.0000] <hsivonen>
depends on whether your intuition is buffered :-)

[18:56:00.0000] <Hixie>
reload http://www.hixie.ch/tests/adhoc/html/parsing/026-demo.html

[18:56:01.0000] <Hixie>
i really think that only IE's behaviour here matches author expectations

[18:56:02.0000] <Hixie>
and i don't buy that it's hard to implement.

[18:57:00.0000] <hsivonen>
I guess I'll have to implement it then, even though WebKit and Opera have gotten away with the simple behavior

[19:07:00.0000] <Hixie>
hsivonen: simpler for authors is more important than simpler for implementors :-)

[19:50:00.0000] <hsivonen>
http://lists.xml.org/archives/xml-dev/200904/msg00062.html

[19:51:00.0000] <hsivonen>
http://lists.xml.org/archives/xml-dev/200904/msg00061.html

[21:03:00.0000] <Hixie>
heycam: be snarky here instead :-)

[21:04:00.0000] <heycam>
#whatwg are indeed the go-to guys for snark!

[21:05:00.0000] <heycam>
if i'm going to spend some time doing something, it may as well be to fix things rather than snark about them, anyway

[21:08:00.0000] <Hixie>
boring!

[21:10:00.0000] <heycam>
:)

[23:39:00.0000] <heycam>
http://www.unicode.org/charts/PDF/UBOOP.pdf

[00:03:00.0000] <Hixie>
http://lists.w3.org/Archives/Member/chairs/2009AprJun/0035.html

[00:03:01.0000] <Hixie>
i wonder if at our next rechartering we can get the htmlwg chartered to 2022 on the same basis

[00:04:00.0000] <Hixie>
"here are some realistic times"

[02:57:00.0000] <annevk2>
Hixie, I e-mailed some comments on webstorage last Friday and it seems they had some delay while getting to the list

[02:57:01.0000] <annevk2>
Hixie, do you have them?

[05:07:00.0000] <zcorpan>
yay http://wiki.whatwg.org/wiki/Web_ECMAScript#HTML_comments

[05:09:00.0000] <annevk2>
yay as in mess?

[05:17:00.0000] <jgraham>
annevk2: What do you mean. The IE behaviour is toatlly sane and I can't see why everyone doesn't do that

[05:18:00.0000] <jgraham>
(beware: low-flying sarcasm)

[05:21:00.0000] <Philip`>
SpiderMonkey's behaviour is different if you use ;version=something-large-enough or ;e4x=1

[05:22:00.0000] <jgraham>
The e4x thing makes sense

[05:22:01.0000] <jgraham>
But the version thng?

[05:22:02.0000] <Philip`>
I might be lying about that

[05:22:03.0000] <zcorpan>
i don't find any differences with version=1.8

[05:23:00.0000] <Philip`>
Oh, I'm not lying

[05:23:01.0000] <Philip`>
http://philip.html5.org/demos/js/jsversion.html - see the "<!-- 1 -->" line

[05:23:02.0000] <Philip`>
(in Firefox)

[05:24:00.0000] <zcorpan>
but <!-- 1 --> is not "line that starts with just whitespace and comments is treated as a line comment" is it?

[05:25:00.0000] <zcorpan>
s//a/

[05:25:01.0000] <Philip`>
No, but it's an HTML comment in the script, and the behaviour is different between ;versions

[05:47:00.0000] <hendry>
what's the issue with DOMTimeStamp? http://lists.whatwg.org/pipermail/commit-watchers-whatwg.org/2009/001994.html

[05:50:00.0000] <hendry>
http://www.w3.org/TR/DOM-Level-3-Core/core.html#Core-DOMTimeStamp DOMTimeStamp is long for non-JS

[05:50:01.0000] <Philip`>
hendry: http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-February/018487.html and associated thread seems relevant

[05:51:00.0000] <hendry>
urgh

[06:12:00.0000] <zcorpan>
Hixie: "The rules right now are really simple -- you only need quotes if your value includes spaces, quotes, or equal signs." ...or greater than

[06:15:00.0000] <Philip`>
or if your value is the empty string

[06:23:00.0000] <zcorpan>
there you go, it's not actually really simple :)

[06:24:00.0000] <zcorpan>
and if you care about compat with firefox or safari or ie or opera then it becomes more complicated still

[06:25:00.0000] <zcorpan>
which is not documented in the spec

[06:25:01.0000] <zcorpan>
s/opera/old versions of opera/

[06:26:00.0000] <jgraham>
So it's only simple if you don't learn any rules except to always quote attribute values

[06:27:00.0000] <jgraham>
Which conincidentially is the fashion amonst hip web developers

[06:28:00.0000] <Philip`>
It's the fashion amongst pretty much all web developers, not just hip ones, as far as I can see

[06:29:00.0000] <Philip`>
(judging by http://canvex.lazyilluminati.com/misc/stats/tokeniser.html having 6902804 double-quoted characters, vs 387236 single-quoted and 376258 unquoted

[06:29:01.0000] <Philip`>
)

[06:30:00.0000] <jgraham>
Oh well maybe it's only the hip web developers who go around telling other people how uncool they are for not using quotes

[06:31:00.0000] <zcorpan>
so unquoted attributes are as common as single quoted attributes?

[06:33:00.0000] <zcorpan>
i.e. about 0.05%

[06:33:01.0000] <Philip`>
jgraham: It's the web developers who have gone *beyond* hip, who omit needless tags and quotes and everything

[06:33:02.0000] <Philip`>
and often omit compatibility with real web browsers too

[06:33:03.0000] <Philip`>
zcorpan: Only if you trust my data, which you shouldn't

[06:34:00.0000] <Philip`>
because it's counting numbers of tokeniser state transitions that are not uppercase letters or ampersands etc

[06:34:01.0000] <zcorpan>
Philip`: it was the only data i had at hand :)

[06:34:02.0000] <Philip`>
rather than counting something useful like number of attributes

[06:35:00.0000] <Philip`>
(But that's the only data I've acquired from an instrumented tokeniser, so I don't have anything better)

[06:35:01.0000] <Philip`>
/me tries to find the quote he's thinking of

[06:36:00.0000] <zcorpan>
Philip`: can unquoted attributes go through more state transitions than a single quoted attribute because of entities?

[06:36:01.0000] <Philip`>
jgraham: It's only the web developers who are so hip they can hardly see over their pelvises

[06:37:00.0000] <Philip`>
zcorpan: I don't think entities make a difference; I was just counting the number of times the "Anything else" transition is encountered

[06:37:01.0000] <Philip`>
Uh

[06:37:02.0000] <Philip`>
Did I say something about uppercase letters?

[06:37:03.0000] <Philip`>
Please ignore that entirely

[06:39:00.0000] <Philip`>
Ooh, actually my data probably does say how many attributes there were

[06:40:00.0000] <Philip`>
BeforeAttributeValueState: currentCharacter == 34 -- 366746

[06:40:01.0000] <Philip`>
BeforeAttributeValueState: true -- 52451

[06:40:02.0000] <Philip`>
BeforeAttributeValueState: currentCharacter == 39 -- 17591

[06:40:03.0000] <Philip`>
which are (respectively) the starts of double-quoted, unquoted and single-quoted attribute values

[06:40:04.0000] <Philip`>
("true" means the "Anything else" case)

[06:41:00.0000] <Philip`>
One can thus conclude that double-quoted attribute values have average length 19 characters, single-quoted 22, and unquoted 7

[06:44:00.0000] <Philip`>
jgraham: Oh, actually it might be pelves

[06:49:00.0000] <zcorpan>
so 84% double-quoted, 12% unquoted and 4% single-quoted

[06:50:00.0000] <Philip`>
Something like that

[06:50:01.0000] <Philip`>
based on a small, unspecified, biased sample of pages

[07:00:00.0000] <Lachy>
The rules for unquoted attributes are relatively simple because most of the special characters that require quotes are rather intuitive becuase they're the characters that have some other special meaning in the tag

[07:03:00.0000] <Philip`>
Is it intuitive that you have to include quotes if the value contains U+000C?

[07:04:00.0000] <Philip`>
where "intuitive" means you don't have to read the spec in detail to discover it

[07:10:00.0000] <Philip`>
(And if you're implementing an HTML5 serialiser, and even if you do read the spec in detail, is it intuitive that you also have to include quotes when the value contains ` or < ?)

[07:15:00.0000] <Lachy>
Philip`, yes, because '>' is the character used to end the tag, so it seems fairly obvious that in <span class=foo>bar..., the attribute only contains "foo"

[07:33:00.0000] <zcorpan>
Lachy: he didn't ask about ">"

[07:49:00.0000] <drostie>
so, an idea that came up in developer.mozilla.org/Talk:En/JavaScript intrigued me, although the way that the poster presented it was ill-defined.

[07:51:00.0000] <drostie>
A lot of people spend a lot of time with bad or generic parsers against XSS. HTML 5 could really reduce this problem if it included a <sandbox> element or so, which restricted the sorts of things you find in XSS exploits anyway. The parser would only have to check for <sandbox> and </sandbox> tags instead of *everything* that could launch a script.

[07:51:01.0000] <drostie>
The browser would make sure, e.g., that a <b> tag started inside of a sandbox didn't leave it, and that scripts didn't execute within the sandbox, and so forth.

[07:53:00.0000] <jgraham>
Philip`: What was the sample size for the comment data that you got for me (i.e. the total number of pages)

[07:53:01.0000] <drostie>
The only thing I'm thinking is that it would probably play not-nice with DOM methods.

[07:55:00.0000] <jgraham>
drostie: That has been discussed before and it turns out to not be quite as simple as you would hope. Ill let someone else explain though (or you can check the mail archives)

[07:58:00.0000] <drostie>
/me goes a-searching ^_^

[07:58:01.0000] <jgraham>
Philip`: Oh it says. I am blind :)

[07:59:00.0000] <Philip`>
jgraham: I guess it was the 425K pages?

[08:00:00.0000] <Philip`>
jgraham: Being blind is not an excuse for failing to read the plain text that says where the data came from :-p

[08:01:00.0000] <Philip`>
I suppose one problem with this dataset is it's going to be heavily biased towards sites with many pages

[08:01:01.0000] <Philip`>
e.g. a forum which exposes a million URLs will be a million times more likely to be included in the sample than a site that's just a single page

[08:02:00.0000] <Philip`>
(though the population is large enough that a single site doesn't occur more than dozens of times, if I remember correctly)

[08:02:01.0000] <zcorpan>
drostie: if those pages would use HTML5 parsers instead of "generic parsers", wouldn't the XSS problems go away?

[08:04:00.0000] <drostie>
zcorpan: that would depend -- but it's much easier to use a generic parser that replaces < and > with character entities than it is to write an HTML5 parser. 

[08:05:00.0000] <annevk2>
drostie, there's a plan to address that through <iframe doc="string of html" seamless sandbox></iframe> iirc

[08:05:01.0000] <drostie>
But jgraham is right, there's a *lot* of discussion about this idea on the mailing list; and one good reason it will never work is because you will always need to parse the content anyway, for legacy clients.

[08:06:00.0000] <drostie>
anne: I'll look at that as well. :D

[08:06:01.0000] <annevk2>
drostie, the fallback would be a src=cross-origin

[08:06:02.0000] <annevk2>
drostie, note that doc= is not defined currently

[08:08:00.0000] <Philip`>
zcorpan: They would have to use HTML5 parsers plus sanitisers that have zero bugs

[08:08:01.0000] <Philip`>
plus serialisers that have zero bugs

[08:08:02.0000] <jgraham>
Philip`: Only zero bugs that result in exploitable holes

[08:08:03.0000] <Philip`>
and also browsers must not be buggy or non-standard

[08:08:04.0000] <Philip`>
Then you'll be secure

[08:10:00.0000] <Philip`>
In the meantime, the 'defence in depth' thing seems like a good idea

[09:24:00.0000] <gsnedders>
Oh dear

[09:24:01.0000] <gsnedders>
/me finds photos of him

[13:37:00.0000] <jgraham>
Note to the people who designed github: I am really stupendeously uninterested in whether I get a zip file or a tar.gz file

[14:14:00.0000] <Philip`>
jgraham: I imagine other people might care, since Windows users can only read zip, and users of more functional OSes can read both but .tar.gz is often much more efficient

[14:19:00.0000] <jgraham>
Philip`: There is no excuse for making the UI "click a link marked download, get a modal overlay thing sking what type of file I want". In almost all cases the extra efficiency of tar.gz will save much less download time than the extra time taken to deal with unexpected popup

[14:20:00.0000] <jgraham>
So just going with zip would be fine

[14:20:01.0000] <jgraham>
Or if they really think people need a choice it should be two different links

[14:23:00.0000] <Philip`>
jgraham: Oh, that does sound a bit annoying

[14:33:00.0000] <gsnedders>
Somebody teach me how to solve first and second order differential equations.

[14:33:01.0000] <Philip`>
/me forgets what first and second order differential equations are

[14:34:00.0000] <gsnedders>
Things in the Advanced Higher maths course :)

[14:35:00.0000] <Philip`>
That's not an entirely helpful description

[14:44:00.0000] <jgraham>
gsnedders: In general it's not possible or at least not easy, I think. But you only need to know special forms, not in general

[14:47:00.0000] <jgraham>
gsnedders: So if you really want help you'll need to be more specific about what you want to know

[14:47:01.0000] <gsnedders>
jgraham: What is required in the AH course.

[14:47:02.0000] <gsnedders>
:P

[14:53:00.0000] <jgraham>
gsnedders: The bit of the syllabus that I found so far talks about equations in the form dy/dx = F(x)/G(y)

[14:54:00.0000] <gsnedders>
No, not them

[14:54:01.0000] <gsnedders>
/me googles for ah maths and finds his school's website

[14:56:00.0000] <jgraham>
gsnedders: you mean ones like a(x)dy/dx + b(x)y = f(x)?

[14:56:01.0000] <gsnedders>
yeah

[14:57:00.0000] <gsnedders>
Search http://www.sqa.org.uk/files_ccc/Maths_AH_5th_ed.pdf for differential equations

[14:57:01.0000] <jgraham>
Ah, well you do what it say then. You rearrange it to the form dy/dx + F(x)y = G(x)

[14:58:00.0000] <jgraham>
and notice that multiplyin both sides by exp(int(F(x), dx)) gives you something that you can write as an exact differential

[15:00:00.0000] <jgraham>
since d/dx(y*exp(int(F(x, dx)))) = F(x)exp(int(F(x), dx)))*y + dy/dx*exp(int(F(x),dx))

[15:01:00.0000] <jgraham>
(excude the probably mismtched brackets)

[15:02:00.0000] <jgraham>
That's applying the chain rule to d/dx(exp(g(x)) where g(x) happens to be an integral

[15:02:01.0000] <gsnedders>
/me badly needs to learn maths

[15:03:00.0000] <jgraham>
gsnedders: The important thing to realise is that a) you can't do everyhing "forwards"; someimes you need to know/intuit the answer

[15:03:01.0000] <jgraham>
and b) you can get a long way with practice

[15:03:02.0000] <gsnedders>
Well, for b) I have four weeks until the exam

[15:04:00.0000] <jgraham>
gsnedders: Well make good use of it :)

[15:05:00.0000] <gsnedders>
This week I'm kinda stuck with second prelims and ball, so only three weeks really

[15:05:01.0000] <gsnedders>
ahhhh

[15:06:00.0000] <Philip`>
/me notes that balls aren't really very productive, so you could spend that time doing fun maths instead

[15:06:01.0000] <gsnedders>
And ditch the gorgeous girl I'm meant to be going with now? :P

[15:06:02.0000] <Philip`>
Sure

[15:07:00.0000] <Philip`>
It's statistically unlikely that you'll marry her, so it's just a waste of time in the long run

[15:07:01.0000] <gsnedders>
LOL

[15:09:00.0000] <jgraham>
/me sees a flaw in that argument

[15:09:01.0000] <jgraham>
Well I say one, but that's really an insult to arguments with just one flaw

[15:11:00.0000] <gsnedders>
Philip`: I am, however, avoiding being overly unproductive by not going to the afterball

[15:11:01.0000] <Philip`>
I suppose it depends on how much value you place on fleeting happiness

[15:14:00.0000] <jgraham>
Philip`: Not at all. It assumes that the experience will have a negligible impact on your future life

[15:14:01.0000] <jgraham>
Which seems like an entirely unsupportable hypothesis

[15:15:00.0000] <jgraham>
(it also assumes that marriage is a end unto something which is rather sily)

[15:32:00.0000] <jgraham>
/me notes that gsnedders should in no way use his previous argument to prioritise the ball over maths revision :)

[16:02:00.0000] <gsnedders>
jgraham: Sorry, but, girl I'm going with > you.

[16:08:00.0000] <jgraham>
/me has no idea what that has to do with anything

[16:09:00.0000] <jgraham>
/me decides it is bed time


2009-04-28
[18:02:00.0000] <Hixie>
hey i passed r3000 earlier without noticing

[00:28:00.0000] <jgraham>
html5lib now with a hg repo

[00:29:00.0000] <MikeSmith>
jgraham: cool

[00:29:01.0000] <MikeSmith>
jgraham: google code provides hg as an option?

[00:36:00.0000] <jgraham>
MikeSmith: Yeah, it is experimental at the moment

[00:36:01.0000] <jgraham>
I haven't actually tried yet to se if it works :)

[00:38:00.0000] <MikeSmith>
jgraham: I notice on Google Code blog:

[00:38:01.0000] <MikeSmith>
"given that Google Code's infrastructure is built for HTTP-based services, we found that Mercurial had the best protocol and performance characteristics for HTTP support. For more information, see our analysis."

[00:38:02.0000] <MikeSmith>
that's an interesting statement

[00:38:03.0000] <MikeSmith>
I wonder what the git people would say about that

[00:43:00.0000] <jgraham>
MikeSmith: Maybe they would say "In a market with two rather similar, both popular, products they did  careful analysis and concluded that the other product was a better fir for their needs. Bravo on such dilligence in assessing the best solution!". But I doubt it somehow

[00:43:01.0000] <MikeSmith>
jgraham: their http://code.google.com/p/support/wiki/DVCSAnalysis doc is worth reading

[00:45:00.0000] <MikeSmith>
I pointed W3C system team at it; maybe will help inform decisions about what DCVS to use at W3C

[00:50:00.0000] <jgraham>
MikeSmith: Yeah, I saw that.

[00:52:00.0000] <MikeSmith>
jgraham: but hey, no reason why that should let us choose to favor a "custom stateful protocol" rather than a stateless HTTP one

[00:53:00.0000] <MikeSmith>
I mean because, what evidence do we have to suggest that this stateless HTTP approach to things really works?

[00:53:01.0000] <MikeSmith>
we need to remain skeptical until we see some proof of that

[00:54:00.0000] <MikeSmith>
in the mean time, we can do like those guys did: Assume we know better, and invent more custom stateless stuff.

[01:32:00.0000] <zcorpan>
/me does not understand role="math"

[01:46:00.0000] <jgraham>
zcorpan: In what sense?

[01:49:00.0000] <jgraham>
The use case seems to be <img src="equation.gif" role="math" alt="\int_0_y x^{2} \d x" aria-describedBy="a"><p id="a">The integral between 0 and y of x squared dx</p>

[01:51:00.0000] <jgraham>
Of which only the last part is really silly because e.g. a LaTeX to HTML convertor could produce both equaton.gif and the alt text, and there is AT that understands some LaTeX

[01:51:01.0000] <jgraham>
Although it might not do much for non AT users and would probably not work in any reasonably complex LaTeX scenario with e.g. custom macros

[02:28:00.0000] <zcorpan_>
jgraham: will you post the --!> and --\s*> findings to the list?

[02:34:00.0000] <zcorpan_>
jgraham: apparently role="math" is appropriate for MathML, too

[02:35:00.0000] <annevk2>
/me has the feeling aria-describedby is way too generic to succeed

[02:46:00.0000] <MikeSmith>
annevk2: please convey my appreciation to whomever it was that made the opera retro page happen

[02:46:01.0000] <MikeSmith>
that thing is great

[02:46:02.0000] <MikeSmith>
they just should have kept it there for longer

[02:47:00.0000] <MikeSmith>
but glad that I caught it when it was live

[02:47:01.0000] <annevk2>
it's still there for me

[02:48:00.0000] <MikeSmith>
hmm, it's showing the normal start page for me now

[02:48:01.0000] <annevk2>
if you tweak the URL you should get it back

[02:49:00.0000] <MikeSmith>
OK

[02:50:00.0000] <MikeSmith>
anyway, the Opera spirit is great

[02:50:01.0000] <takkaria>
I like Opera

[02:50:02.0000] <MikeSmith>
the humor, not taking itself deadly seriously all the time

[02:50:03.0000] <takkaria>
they offered me an internship

[02:51:00.0000] <MikeSmith>
takkaria: not surprised. Opera definitely has a knack for finding and recruiting smart people

[02:53:00.0000] <Philip`>
MikeSmith: Or they just offer internships to everybody :-)

[02:53:01.0000] <annevk2>
congrats takkaria

[02:54:00.0000] <MikeSmith>
Philip`: hmm, they never offered me an internship

[02:54:01.0000] <MikeSmith>
I feel left out

[02:56:00.0000] <zcorpan_>
MikeSmith: you can apply for one

[02:56:01.0000] <MikeSmith>
cool, I didn't know I still had time

[02:57:00.0000] <MikeSmith>
zcorpan_: how much does the gig pay?

[02:57:01.0000] <takkaria>
annevk2: cheers

[02:57:02.0000] <zcorpan_>
MikeSmith: no idea

[02:57:03.0000] <takkaria>
does anyone here work in the Linköping office?

[02:58:00.0000] <zcorpan_>
me and jgraham

[02:59:00.0000] <takkaria>
oh, great :)

[03:01:00.0000] <MikeSmith>
zcorpan_: do people in Linköping office still use that Emacs-based messaging thing instead of IRC?

[03:01:01.0000] <MikeSmith>
something-COM or COM-something

[03:02:00.0000] <zcorpan_>
MikeSmith: yes, LysKOM

[03:02:01.0000] <MikeSmith>
zcorpan_: do you use it?

[03:02:02.0000] <zcorpan_>
MikeSmith: yes i have to :)

[03:03:00.0000] <jgraham>
Yay LysKOM

[03:03:01.0000] <MikeSmith>
I guess I should instead ask, do you like using it?

[03:03:02.0000] <zcorpan_>
although i read it just once a week or so

[03:03:03.0000] <zcorpan_>
no i hate it

[03:03:04.0000] <jgraham>
Reinventing the wheel but making it triangular and spiky

[03:03:05.0000] <MikeSmith>
heh

[03:03:06.0000] <MikeSmith>
good description

[03:03:07.0000] <MikeSmith>
I don't know what they like about that thing

[03:04:00.0000] <MikeSmith>
I mean, I can see it being having some use

[03:04:01.0000] <MikeSmith>
but not as an IRC replacement

[03:04:02.0000] <zcorpan_>
it's something in between irc and email

[03:05:00.0000] <MikeSmith>
zcorpan_: right, but more of a worst-of-both-worlds combination of the two

[03:05:01.0000] <zcorpan_>
don't disagree

[03:06:00.0000] <jgraham>
The choice of clients seems to be a) a webbased client that breaks the back button and intermitttently gets confused about whether there are new messages or b) an emacs-based client

[03:07:00.0000] <MikeSmith>
jgraham: there's a curses-based client too, I seem to remember

[03:07:01.0000] <MikeSmith>
I think that's what I used, when I needed to

[03:07:02.0000] <jgraham>
Really? 

[03:07:03.0000] <MikeSmith>
yeah, somewhere

[03:08:00.0000] <MikeSmith>
I think I found it less painful than Emacs

[03:09:00.0000] <MikeSmith>
http://www.lysator.liu.se/lyskom/klienter/ttyklient/

[03:09:01.0000] <MikeSmith>
I guess

[03:11:00.0000] <MikeSmith>
http://packages.ubuntu.com/dapper/net/lyskom-tty-client

[03:13:00.0000] <MikeSmith>
is hsivonen site using downloadable fonts?

[03:13:01.0000] <MikeSmith>
3MB and counting...

[03:14:00.0000] <MikeSmith>
4MB+

[03:16:00.0000] <MikeSmith>
hsivonen: anyway, you updated http://hsivonen.iki.fi/doctype/ today, I guess?

[03:17:00.0000] <MikeSmith>
/me sees <!DOCTYPE html> among the recommendations there now

[03:17:01.0000] <Philip`>
It sounds somewhat ironically like the Emacs client elicits more curses than the curses-based client

[03:21:00.0000] <MikeSmith>
heh

[03:21:01.0000] <MikeSmith>
actually, I see now it just says "TTY-based client"

[03:22:00.0000] <MikeSmith>
and package seems to depend only on libc6 package

[03:22:01.0000] <MikeSmith>
man, hsivonen got a truckload of downloadable fonts on his pages

[03:23:00.0000] <Philip`>
Not as many as on my page, I'm sure

[03:24:00.0000] <jgraham>
MikeSmith: It also says "With lyskom you can connect to your favourite LysKOM"

[03:24:01.0000] <MikeSmith>
well, your page has a specific demonstration purpose

[03:24:02.0000] <MikeSmith>
heh

[03:24:03.0000] <MikeSmith>
"your favourite LysKOM"

[03:24:04.0000] <MikeSmith>
that kind of limits things

[03:24:05.0000] <MikeSmith>
to zero

[03:24:06.0000] <MikeSmith>
zero choices there

[03:25:00.0000] <jgraham>
Indeed. It seems entirely unsutiable for my purposes

[03:25:01.0000] <Philip`>
"favourite" is relative

[03:25:02.0000] <Philip`>
like how you could have a favourite evil dictator

[03:25:03.0000] <MikeSmith>
"With lyskom you can connect to your least-despised LysKOM"

[03:26:00.0000] <MikeSmith>
"With lyskom you can connect to the LysKOM you have a need to use even though you hate it"

[03:27:00.0000] <MikeSmith>
so I'm wondering if the size/number of fonts at hsivonen site is atypical, or if a typical site that uses downloadable is going to be in similar range

[03:27:01.0000] <MikeSmith>
~5.1MB

[03:28:00.0000] <zcorpan_>
depends on whether they use Philip`'s service or not

[03:28:01.0000] <MikeSmith>
or 6.4MB actually, on my Mac

[03:28:02.0000] <Philip`>
/me wonders how he could make his service more usable

[03:29:00.0000] <Philip`>
I suppose it'd be nice if the TTF+EOT thing wasn't such a horrid hack that results in IE making a 404 request for every font

[03:29:01.0000] <MikeSmith>
speaking of usable, Web Inspector UI is great for getting at this info

[03:29:02.0000] <MikeSmith>
download time/speed, I mean

[03:30:00.0000] <Philip`>
Also I suppose it'd be nice if you could ask for e.g. all ASCII letters, or all European letters, rather than having to type in exactly the characters you want

[03:30:01.0000] <Philip`>
MikeSmith: Better than Firebug?

[03:30:02.0000] <MikeSmith>
Philip`: I like it better at least

[03:31:00.0000] <Philip`>
Firebug handily alerted me to the issue that my pages were all taking 15 seconds to finish loading

[03:31:01.0000] <MikeSmith>
I really wish mozilla would just make Firebug a standard part of Firefox

[03:32:00.0000] <MikeSmith>
or at least of Minefield

[03:32:01.0000] <Philip`>
(since I had a (Fast)CGI script outputting some text/html with a Content-Length, and Apache was set to automatically compress text/html, but it doesn't change the Content-Length when compressing, so Squid was spending 15 seconds waiting for more content before timing out, I think)

[03:33:00.0000] <jgraham>
/me got a couple of crashes when he trie to use WI with Safari 4

[03:34:00.0000] <Philip`>
MikeSmith: That sounds like needless bloat for a hundred million users who are never going to do any web development

[03:34:01.0000] <MikeSmith>
Philip`: yeah, I suppose so, for the FF case at least

[03:35:00.0000] <MikeSmith>
but seems it has some obvious practical use for bundling in Minefield

[03:35:01.0000] <Philip`>
MikeSmith: People who willingly install Minefield ought to know how to install an extension - it only takes a minute when you're first setting up your profile

[03:36:00.0000] <jgraham>
The big problem is when it is incompatible with minefield.latest

[03:36:01.0000] <MikeSmith>
right

[03:36:02.0000] <Philip`>
Then use some other extension that makes extensions compatible

[03:36:03.0000] <MikeSmith>
and I bet there are still people who don't install Firebug with Minefield anyway

[03:36:04.0000] <Philip`>
like https://addons.mozilla.org/en-US/firefox/addon/6543

[03:36:05.0000] <jgraham>
Philip`: No, when it is actually incompatible

[03:38:00.0000] <Philip`>
jgraham: In that case bundling it with Minefield wouldn't help, because it would still be incompatible until somebody fixed it

[03:38:01.0000] <zcorpan_>
Philip`: you should make your service more discoverable using google, too

[03:38:02.0000] <MikeSmith>
Philip`: exactly what you said

[03:39:00.0000] <Philip`>
zcorpan_: But that might result in people using it, which would be bad because it's inefficient and buggy and ugly :-)

[03:39:01.0000] <MikeSmith>
a change that breaks Firebug is a change that should be backed out before the nightly is built

[03:39:02.0000] <jgraham>
Philip`: It would help because it wwould have to be treated at the same priority level as other features

[03:40:00.0000] <Philip`>
jgraham: It could be treated at the same priority level as other features without being bundled with Minefield

[03:41:00.0000] <Philip`>
by appropriate changes to development policy and to the automatic tests

[03:41:01.0000] <jgraham>
Philip`: Sure. But it is less likely to actually happen

[03:41:02.0000] <zcorpan_>
Philip`: maybe you could use spellchecking dictionaries to determine which characters to include for different languages

[03:41:03.0000] <jgraham>
zcorpan_: Punctuation

[03:41:04.0000] <Philip`>
jgraham: But more likely to actually happen than bundling with Minefield

[03:42:00.0000] <Philip`>
since it's a necessary step before bundling with Minefield, and therefore a subset of the challenge

[03:43:00.0000] <Philip`>
zcorpan_: Or look through the textual content of web pages that use different languages, to see what characters they use

[03:43:01.0000] <Philip`>
(which would be better at finding punctuation and digits and ASCII and copyright symbols, I guess)

[03:45:00.0000] <Philip`>
But it'd probably be easier to just go by Unicode blocks

[03:47:00.0000] <zcorpan_>
i guess the copyright symbol should be present for all languages

[03:47:01.0000] <Philip`>
Not if you only want to use the font for your logo, though

[03:47:02.0000] <zcorpan_>
right

[03:48:00.0000] <Philip`>
Perhaps what we should do is invent a new dynamic font format, so the web browser can download glyphs automatically on-demand without having to download irrelevant junk

[03:50:00.0000] <Philip`>
(EOT seems to have a primitive form of that idea, by copying various font properties to a header at the start of the file, so the browser can simply download the first chunk and decide whether it can use this font (because it covers the right character range, or has the right font style, etc) instead of having to download the entire file and then search through it for the same information in arbitrary unpredictable locations)

[03:59:00.0000] <Philip`>
Hmm, it looks like people have actually used my font thing

[03:59:01.0000] <Philip`>
since there's ~500 fonts in the output cache

[04:00:00.0000] <zcorpan_>
Philip`: could be people just trying it without actually using the fonts

[04:00:01.0000] <zcorpan_>
Philip`: btw the output page doesn't have a <title>

[04:00:02.0000] <zcorpan_>
last i checked

[04:03:00.0000] <Philip`>
zcorpan_: Oops

[04:05:00.0000] <Philip`>
/me fixes

[06:04:00.0000] <annevk2>
MikeSmith: http://www.opera.com/?flashed=0 lets you revisit the home page

[06:04:01.0000] <annevk2>
/me wonders if home pages are archived somewhere

[06:24:00.0000] <hendry>
Hixie: noticed a typo in html5 index, %s/initalize/initialize/

[06:39:00.0000] <zcorpan_>
<div class="float-left padding-top-5 padding-bottom-9 width-385"> -- http://www.thetimes.co.uk 

[06:43:00.0000] <Philip`>
zcorpan_: They're not the only ones - e.g. www.grownups.co.nz/profile/viewPublicProfile/id/8209 has a class="border-left-blue border-right-blue padding-left-10 padding-right-10 padding-bottom-5"

[06:43:01.0000] <Philip`>
www.paintingmax.com/shop-style-ashcan-school-c-3_304.html?sort=2d - class="padding-10px"

[06:44:00.0000] <Philip`>
surfingshot.com/index.php?main_page=product_info&products_id=4432 - class="padding-2px"

[06:44:01.0000] <Philip`>
etc

[06:44:02.0000] <Philip`>
/me wonders if these sites picked up that trick from a common source

[06:50:00.0000] <Philip`>
hsivonen: http://hsivonen.iki.fi/doctype/ says "emumerate"

[06:51:00.0000] <Philip`>
hsivonen: "Testing image alignment with Internet Explorer is inadequate however be sure to test in IE8, too." is confusing to read, and maybe could do with some punctuation around the 'however'

[06:58:00.0000] <Philip`>
hsivonen: "To put it another way, if you have #include "foo.h", you should not bind any black magic to the name foo.h, because it should be permissible to paste the contents of foo.h inline or copy the contents of foo.h to bar.h and say #include "bar.h"."

[06:59:00.0000] <Philip`>
Modern compilers do do black magic when you have #include "foo.h", e.g. GCC will look for a precompiled header named foo.gch

[07:00:00.0000] <Philip`>
and if foo.h contains "#pragma once" then MSVC will treat #include "foo.h"; #include "foo.h" differently to how it treats #include "foo.h"; #include "differently-named-copy-of-foo.h"

[07:01:00.0000] <Philip`>
I suppose #pragma once is generally considered a bad idea, but precompiled headers a good idea, and they're definitely black magic

[07:01:01.0000] <Philip`>
(but they're magic which only affects performance, not functionality, which seems like a significant difference)

[07:11:00.0000] <zcorpan_>
/me wonders why testing image alignment in IE8 is inadequate

[07:45:00.0000] <krijnh>
Philip`: sup? :)

[09:16:00.0000] <annevk2>
hsivonen, there's <table><tr><td><p><\/tbody> <\/table> in a document.write in Acid3 that assumes the space is still there after parsing

[09:16:01.0000] <annevk2>
hsivonen, Opera initially failed that

[09:16:02.0000] <hsivonen>
grrr!

[09:16:03.0000] <annevk2>
hsivonen, not a 100% sure it affects what you're discussing but I thought I'd bring it up

[09:17:00.0000] <hsivonen>
annevk2: what's "there"

[09:18:00.0000] <annevk2>
hsivonen, bottom of the page has the document.write

[09:18:01.0000] <hsivonen>
inside table?

[09:18:02.0000] <annevk2>
hsivonen, it writes a <table> including a space

[09:20:00.0000] <gsnedders>
/me has his internet get really slow

[09:20:01.0000] <hsivonen>
http://software.hixie.ch/utilities/js/live-dom-viewer/saved/98 is the same for me in Gecko/HTML5, WebKit and Opera

[09:22:00.0000] <annevk2>
since Opera 10, yes

[09:24:00.0000] <annevk2>
we used to ignore whitespace at certain places in tables during parsing

[09:36:00.0000] <hendry>
is there an EU mirror of whatwg.org/html5? :)

[09:38:00.0000] <Philip`>
hendry: What's wrong with America? It's only a hundred milliseconds away :-)

[09:41:00.0000] <hendry>
/me wonders how http://svn.whatwg.org/webapps/index is split up to multipage

[09:42:00.0000] <MikeSmith>
hendry: Philip` does it manually every time Hixie commits

[09:44:00.0000] <hendry>
hah

[09:45:00.0000] <MikeSmith>
hendry: 

[09:45:01.0000] <MikeSmith>
http://code.google.com/p/html5/source/browse/trunk/spec-splitter/spec-splitter.py

[09:46:00.0000] <MikeSmith>
http://html5.googlecode.com/svn/trunk/spec-splitter/spec-splitter.py

[09:48:00.0000] <Philip`>
hendry: That spec-splitter script runs on my server, and Hixie executes it and then downloads the output

[09:50:00.0000] <jgraham>
Don't believe Philip`; that script is just for show. He actually has a highly trained team of bees that read the spec and output the html for the split spec as a pattern of filled an unfilled cells in their hive

[09:50:01.0000] <Philip`>
I love bees

[09:55:00.0000] <hendry>
Philip`: don't laugh now, did you try xsl?

[09:56:00.0000] <jgraham>
/me wonders wwhen he should laugh

[09:59:00.0000] <jgraham>
How can xsl possibly compete with the massive concurrency of a whole swarm of bees

[09:59:01.0000] <jgraham>
?

[10:04:00.0000] <hendry>
i ran xsltproc on html5's index today to extract pre[@class='idl'] and I was impressed how fast it was.

[10:05:00.0000] <Philip`>
hendry: The difficulty with XSL is that I don't understand it

[10:06:00.0000] <Philip`>
particularly for complex things like splitting a document in multiple parts and gluing the right headers and inter-page navigation and fixing hashed links and stuff

[10:06:01.0000] <Philip`>
(and particularly since I've never cared to try learning XSL)

[10:07:00.0000] <Philip`>
whereas I had already learned Python and knew how to use it

[10:08:00.0000] <gsnedders>
Philip`: But parsing the spec is slow in Python.

[10:08:01.0000] <Philip`>
gsnedders: No it's not, since I use lxml's parser

[10:08:02.0000] <gsnedders>
Cheat.

[10:08:03.0000] <Philip`>
It's all your fault for not making html5lib fast enough

[10:09:00.0000] <gsnedders>
I'm not the one who was optimizing it.

[10:09:01.0000] <jgraham>
Philip`: Isn't that your fault? I thought thte deal was that I worked on making it slower, you worked on making it faster, gsnedders hecked from the sidelines?

[10:09:02.0000] <Philip`>
gsnedders: That's why it's your fault

[10:10:00.0000] <gsnedders>
jgraham: s/hecked/hacked/?

[10:10:01.0000] <Philip`>
heckled

[10:10:02.0000] <jgraham>
heckled

[10:10:03.0000] <gsnedders>
Ah, I wasn't sure which

[10:10:04.0000] <gsnedders>
I guessed it was one of them

[10:10:05.0000] <jgraham>
Raeding what you write on irc is way overrated

[10:10:06.0000] <gsnedders>
/me turns up the volume of Life in Cartoon Motion

[10:11:00.0000] <Philip`>
/me feels like a proper researcher now since he's reading archeologically ancient papers - some are from as far back as 1999, and published as PostScript rather than PDF

[10:12:00.0000] <gsnedders>
/me is listening to Lollipop by Mika from Life In Cartoon Motion

[10:12:01.0000] <gsnedders>
/me is blatantly cool

[10:12:02.0000] <MikeSmith>
hendry: I've been planning to write an XSLT-based splitter for the H:TML draft

[10:12:03.0000] <jgraham>
Philip`: Isn't ps irritating though?

[10:13:00.0000] <Philip`>
jgraham: Yes

[10:13:01.0000] <gsnedders>
/me goes back to trying to understand solving first order linear differential equations

[10:13:02.0000] <Philip`>
jgraham: mainly since I can't read them in KPDF, and have to use something weird and crazy like KGhostView which doesn't handle page-breaks nicely

[10:14:00.0000] <hendry>
MikeSmith: get anywhere with that? ;)

[10:15:00.0000] <hendry>
Philip`: i'm trying now to minimize your alpha canvas test for inclusion to the MCTMB. http://static.webvm.net/canvas/2d.drawImage.alpha.html

[10:16:00.0000] <MikeSmith>
hendry: not yet, but I suppose I should turn my attention to it now. I would need to eventually anyway

[10:16:01.0000] <hendry>
Philip`: going to give up for now as I need to catch a train. Need to understand your assert thingie. Just wanted to update http://www.w3.org/2008/06/mobile-test/canvas.js

[10:17:00.0000] <Philip`>
hendry: The _assertPixelApprox is just requiring that the output is green

[10:18:00.0000] <MikeSmith>
only troublesome part about splitting with XSLT is how to figure out which fragment refs are going to end up being intra-file ones, and dealing with those

[10:18:01.0000] <Philip`>
(with some tolerance, because implementations might not be precise)

[10:18:02.0000] <MikeSmith>
might be a good case for using XSL kyes

[10:18:03.0000] <Philip`>
(and only checking one pixel of the image, because it's too slow and pointless to check every pixel)

[10:19:00.0000] <Philip`>
hendry: If your test output is checked by humans then you shouldn't need to do anything fancy like that - just tell people the output should look green

[10:20:00.0000] <hendry>
Philip`: yes, that's what I am trying to do. That's haot the MCTMB test already works: http://www.w3.org/2008/06/mobile-test

[10:42:00.0000] <gsnedders>
Oh, wow. This isn't as hard as I thought.

[10:43:00.0000] <gsnedders>
/me is having vague understanding creep into his head

[10:48:00.0000] <Philip`>
/me wonders why none of his PS viewers have a search function

[10:49:00.0000] <Philip`>
(I'm guessing it's just because PS is too stupid a document format for such things)

[13:08:00.0000] <cryzed>
Hey

[13:08:01.0000] <cryzed>
just wanted to drop by

[13:08:02.0000] <cryzed>
and tell your that's something wrong with the Google

[13:09:00.0000] <cryzed>
mercurial

[13:09:01.0000] <cryzed>
repository

[13:09:02.0000] <cryzed>
If you haven't already noticed

[13:09:03.0000] <jgraham>
Which

[13:09:04.0000] <jgraham>
one

[13:09:05.0000] <jgraham>
?

[13:09:06.0000] <cryzed>
http://code.google.com/p/html5lib/

[13:09:07.0000] <jgraham>
Ah

[13:09:08.0000] <cryzed>
You can't even browse the source

[13:09:09.0000] <cryzed>
via the web interface

[13:09:10.0000] <jgraham>
/me should investigate

[13:09:11.0000] <cryzed>
http://code.google.com/p/html5lib/source/browse/ -> Gives you an error

[13:09:12.0000] <cryzed>
"Error retrieving directory contents."

[13:09:13.0000] <Philip`>
/me wonders if it was wise to switch to such an experimental system :-)

[13:10:00.0000] <jgraham>
Philip`: Well we have other source trres around, right?

[13:10:01.0000] <jgraham>
:)

[13:11:00.0000] <Philip`>
jgraham: Do we?

[13:12:00.0000] <jgraham>
Philip`: Sure. Just not with revision history and stuff

[13:12:01.0000] <jgraham>
Which is useful but sort-of non essential

[13:14:00.0000] <jgraham>
It might be that I need to do something

[13:14:01.0000] <jgraham>
I am investigating now

[13:15:00.0000] <gsnedders>
jgraham: Are we meant to all be moving to hg now for html5lib?

[13:15:01.0000] <gsnedders>
http://html5lib.googlecode.com/hg/ looks sorta wrong

[13:16:00.0000] <jgraham>
gsnedders: That is sort of the plan

[13:17:00.0000] <jgraham>
cryzed, Philip`, gsnedders The svn repo is still avaliable at https://html5lib.googlecode.com/svn/trunk/

[13:18:00.0000] <jgraham>
It seems like I need to import the svn repositoy into hg

[13:19:00.0000] <cryzed>
thanks

[13:19:01.0000] <jgraham>
/me wishes OSX had sane package mangement, decides to move to a linux computer

[13:21:00.0000] <cryzed>
jgraham, high five!

[13:21:01.0000] <cryzed>
Ubuntu <3

[13:21:02.0000] <cryzed>
Well not necessarily Ubuntu, any Linux will do :D

[13:21:03.0000] <Philip`>
Except Slackware

[13:21:04.0000] <cryzed>
Except Gentoo

[13:22:00.0000] <Philip`>
Gentoo has sane package management

[13:22:01.0000] <cryzed>
Gentoo's compile-time is just too fucking long

[13:22:02.0000] <cryzed>
Why would I want to waste hours compiling X11 and GNOME?

[13:23:00.0000] <Philip`>
It's not wasting time - you can multitask

[13:23:01.0000] <cryzed>
imho it is

[13:23:02.0000] <cryzed>
You can simply use pre-compiled binaries

[13:23:03.0000] <Philip`>
You can write "emerge -u xorg-server" and then it'll be finished long before you next reboot

[13:24:00.0000] <cryzed>
yeah, and what do you do without a GUI?

[13:24:01.0000] <cryzed>
Editing text-files?

[13:24:02.0000] <cryzed>
Lynx?

[13:24:03.0000] <cryzed>
:D

[13:24:04.0000] <Philip`>
You install from the precompiled binaries on the installation CD

[13:24:05.0000] <Philip`>
and then once you've got a running system you can keep it up-to-date in the background

[13:25:00.0000] <cryzed>
there are precompiled xorg binaries?

[13:25:01.0000] <cryzed>
(and gnome?)

[13:25:02.0000] <Philip`>
There's a whole Live CD with GUI that you can install from

[13:25:03.0000] <cryzed>
Well, yes the LiveCD hast got a GUI

[13:26:00.0000] <cryzed>
But the installed operating system hasn't

[13:26:01.0000] <cryzed>
right?

[13:26:02.0000] <gsnedders>
It depends if you install it.

[13:26:03.0000] <gsnedders>
There are binaries available for a lot of packages.

[13:26:04.0000] <cryzed>
gsnedders, So where exactly is the advantage

[13:26:05.0000] <cryzed>
over Arch and Ubuntu?

[13:26:06.0000] <cryzed>
You can compile your programs

[13:26:07.0000] <cryzed>
ok cool

[13:27:00.0000] <cryzed>
1-2 seconds execution time boost

[13:27:01.0000] <cryzed>
I guess

[13:27:02.0000] <Philip`>
You can get continually up-to-date software, rather than waiting six months for the next big release

[13:27:03.0000] <Philip`>
You can customise the build options and compile-time features of all the software

[13:27:04.0000] <cryzed>
Ubuntu's updating regularly aswell

[13:27:05.0000] <gsnedders>
Better way of speeding up start up time: switch to FreeBSD, which has always been more responsive for me than any Linux distro.

[13:28:00.0000] <cryzed>
gsnedders, Ah I'm curious

[13:28:01.0000] <cryzed>
I never used FreeBSD

[13:28:02.0000] <cryzed>
FreeBSD is basically

[13:28:03.0000] <cryzed>
another kernel

[13:28:04.0000] <cryzed>
right?

[13:28:05.0000] <gsnedders>
No, it's a complete OS

[13:28:06.0000] <cryzed>
Does it use GNU?

[13:28:07.0000] <gsnedders>
No

[13:28:08.0000] <cryzed>
So it isn't compatible

[13:28:09.0000] <cryzed>
with Linux binaries?

[13:28:10.0000] <gsnedders>
It has a binary compatibility module

[13:29:00.0000] <gsnedders>
(kernel module, that is)

[13:29:01.0000] <cryzed>
but still

[13:29:02.0000] <cryzed>
I don't know

[13:29:03.0000] <cryzed>
I think I prefer Ubuntu, seems easier and I'm by nature a lazy person

[13:30:00.0000] <cryzed>
Not stupid, I probably could switch if I wanted, but I don't

[13:30:01.0000] <cryzed>
I'm rather content with Ubuntu and the boot-time

[13:30:02.0000] <cryzed>
(about 18 seconds)

[13:30:03.0000] <gsnedders>
/me has plenty of problems with most DEs' usability

[13:30:04.0000] <gsnedders>
(where most is defined to be everyone I've tried)

[13:30:05.0000] <gsnedders>
(this is X DEs)

[13:31:00.0000] <cryzed>
You are using X DE?

[13:31:01.0000] <gsnedders>
?

[13:31:02.0000] <cryzed>
nevermind

[13:31:03.0000] <cryzed>
I didn't understand what you said

[13:31:04.0000] <gsnedders>
an X desktop enviroment

[13:31:05.0000] <gsnedders>
a desktop environment for an X window server

[13:32:00.0000] <cryzed>
ah okay

[13:32:01.0000] <cryzed>
Well I like GNOME

[13:32:02.0000] <cryzed>
I hate KDE

[13:32:03.0000] <cryzed>
XFCE is not my thingy

[13:32:04.0000] <cryzed>
never tried any *box

[13:34:00.0000] <gsnedders>
/me sticks to OS X because it has a better desktop environment and still has a proper POSIX command line

[13:36:00.0000] <jgraham>
/me wonders what problems gsnedders has

[13:36:01.0000] <gsnedders>
/me has a lot of problems

[13:37:00.0000] <jgraham>
with X desktop environments in particula

[13:37:01.0000] <jgraham>
r

[13:37:02.0000] <gsnedders>
Then tend to just suck once you get to the small details

[13:38:00.0000] <gsnedders>
Now, unless this helps me with maths, this is about as productive as a ball.

[13:39:00.0000] <Philip`>
Why don't you think this is productive? It's an opportunity to end the OS/desktop debate once and for all, by coming to a decision on which is best - you can't just give up now

[13:44:00.0000] <gsnedders>
Philip`: I don't have an exam on this.

[13:45:00.0000] <gsnedders>
/me needs to get a buttonhole

[13:45:01.0000] <gsnedders>
(as in, the floral thing, not the actual hole)

[13:45:02.0000] <Philip`>
gsnedders: Sure, but think of how much humanity will benefit once we've decided which OS is best and so nobody else will have to spend time debating it

[13:46:00.0000] <gsnedders>
"There's many battles lost/But tell me who has won"

[13:51:00.0000] <jgraham>
Plenty of OS X details suck too

[13:51:01.0000] <jgraham>
:)

[13:51:02.0000] <cryzed>
Ehrm

[13:51:03.0000] <cryzed>
Guys

[13:51:04.0000] <cryzed>
Linux is best

[13:51:05.0000] <cryzed>
/me is running

[13:52:00.0000] <cryzed>
btw, thanks for the great library

[13:52:01.0000] <cryzed>
using it daily

[13:52:02.0000] <cryzed>
to do my webscraping with python

[13:57:00.0000] <gsnedders>
/me uses what he thinks is best for what he needs to do

[13:57:01.0000] <gsnedders>
Which tends to be for any computer I have a GUI on OS X

[13:58:00.0000] <Philip`>
/me does the same, and ends up using Linux, and the difference is that he is right and gsnedders is wrong

[14:01:00.0000] <jgraham>
/me tendds to use anything that isn't Windows as he finds that too annoying and everything else merely differntly deficient

[14:02:00.0000] <gsnedders>
/me is getting to the point where he can actually solve first order linear differential equations without looking at how to do it and get the answer corret

[14:02:01.0000] <gsnedders>
*correct

[14:03:00.0000] <jgraham>
Examples of things that are broken on OS X: The lack of control over window positioning. The applications folder. The broken virtual desktops. The finder.

[14:04:00.0000] <jgraham>
gsnedders: good :)

[14:05:00.0000] <Philip`>
jgraham: Also, it looks too pretty

[14:05:01.0000] <gsnedders>
Virtual desktops have been far less broken since 10.5.3, though

[14:05:02.0000] <gsnedders>
Like, they are actually bearable and useful now

[14:05:03.0000] <Philip`>
I like to be reminded that I'm using a deterministic machine and I have dominion over it

[14:06:00.0000] <jgraham>
gsnedders: Not with Lightroom

[14:06:01.0000] <gsnedders>
How so?

[14:06:02.0000] <jgraham>
(this may not be Apple's fault. I don't care)

[14:07:00.0000] <jgraham>
gsnedders: Click LR icon in the dock. Dock dissapears. Desktop containing LR is not focused

[14:08:00.0000] <jgraham>
(assuming i is open and on a different virtual desktop)

[14:08:01.0000] <gsnedders>
/me assumes that LR creates its own windows and doesn't use any standard API for it

[14:09:00.0000] <gsnedders>
But yeah, that sounds b0rked

[14:17:00.0000] <Philip`>
/me got confused when switching from desktop A to desktop B resulted in all the terminal windows on desktop A jumping into the foreground, and then gave up

[16:48:00.0000] <mathuin>
What is the best way to normalize all the tags in a BeautifulSoup tree generated by html5lib?  Some files I use have tags like 'dc:Title', others 'dc:title' so I'd like to force all tags to lower case.  Do I use a sanitizer or something else?


2009-04-29
[17:15:00.0000] <Hixie>
i really wish Kristof could quote context

[17:15:01.0000] <Hixie>
i never have any idea wtf his e-mails are about

[22:14:00.0000] <Hixie>
heycam: yt?

[22:14:01.0000] <heycam>
Hixie, yep

[22:14:02.0000] <Hixie>
heycam: in webidl you have the following example:

[22:14:03.0000] <Hixie>
map[1] = false;

[22:14:04.0000] <Hixie>
...which sets a property named "1"

[22:15:00.0000] <Hixie>
what happens if the indexed properties are automatically set from 0..length-1

[22:15:01.0000] <Hixie>
and you set map[0] = true;

[22:15:02.0000] <Hixie>
and then fetch map[0]

[22:15:03.0000] <Hixie>
if you see what i mean

[22:15:04.0000] <Hixie>
sorry, i'm not expressing myself well

[22:16:00.0000] <heycam>
sorry i didn't understand :)

[22:16:01.0000] <Hixie>
heycam: the concern comes from the Storage object, where you can get the keys by index, and then index into the keys to get the values

[22:16:02.0000] <Hixie>
heycam: there is a clash if someone sets the key "0" and then fetches the property "0" -- it'll return "0" because that's the name of the key

[22:16:03.0000] <Hixie>
but then if you fetch the property "0", you'll still just get "0", not the value of the key "0"

[22:17:00.0000] <Hixie>
because the keys and the indexes are in the same namespace

[22:17:01.0000] <Hixie>
i'm wondering whether to ban numeric keys to get around this

[22:17:02.0000] <Hixie>
or to just say "tough"

[22:17:03.0000] <heycam>
Hixie, i see what you mean now

[22:17:04.0000] <heycam>
perhaps that isn't good design to have numeric values also be valid key names

[22:18:00.0000] <Hixie>
yeah

[22:18:01.0000] <heycam>
and to have the property getting be the way to access both key name and key value

[22:18:02.0000] <Hixie>
what syntax should i disallow?

[22:18:03.0000] <Hixie>
[1-9][0-9]* ?

[22:18:04.0000] <heycam>
um

[22:19:00.0000] <heycam>
i *think*

[22:19:01.0000] <Hixie>
as in, indexes matching that regexp

[22:19:02.0000] <heycam>
you mean keys?

[22:19:03.0000] <Hixie>
right sorry

[22:19:04.0000] <heycam>
whatever es232 says for how it distinguishes indexes vs names for Array objects

[22:20:00.0000] <heycam>
which i think works out to [1-9][0-9]*, where the value is < 2**32 - 1

[22:20:01.0000] <heycam>
(and note the < there rather than <=)

[22:21:00.0000] <heycam>
that same distinguishing is done in the webidl [[Put]], to determine whether to invoke the index or name getters/setters

[22:21:01.0000] <heycam>
actually no that last statement is false

[22:21:02.0000] <heycam>
it uses the "names of named properties" etc. for that

[22:22:00.0000] <Hixie>
yeah i had looked there

[22:22:01.0000] <heycam>
er s/232/262/ earlier

[22:23:00.0000] <heycam>
http://bclary.com/2004/11/07/#a-15.4.5.1

[22:23:01.0000] <heycam>
oh, it just says "If P is not an array index"...

[22:23:02.0000] <heycam>
http://bclary.com/2004/11/07/#a-15.4

[22:23:03.0000] <heycam>
A property name P (in the form of a string value) is an array index if and only if ToString(ToUint32(P)) is equal to P and ToUint32(P) is not equal to 232 - 1.

[22:24:00.0000] <heycam>
(thats 2**32 - 1 at the end)

[22:24:01.0000] <heycam>
so i think that's equivalent to matching the string against [1-9][0-9]* and ensuring it's < 2**32 - 1

[22:25:00.0000] <Hixie>
yeah i found that, i thought you had already seen it when you said the 2**32 - 1 thing :-)

[22:25:01.0000] <heycam>
yeah i had, but ages ago...

[22:25:02.0000] <heycam>
:)

[22:27:00.0000] <heycam>
/me brb coffee

[22:32:00.0000] <Hixie>
heycam: [[Put]] is broken, as far as i can tell.

[22:33:00.0000] <Hixie>
oh wait

[22:33:01.0000] <Hixie>
no it isn't

[22:33:02.0000] <Hixie>
i've just been really confused about how named properties should be defined

[22:34:00.0000] <Hixie>
maybe not

[22:34:01.0000] <heycam>
it's possible it's broken

[22:35:00.0000] <heycam>
i really should make it a structured algorithm instead of a big 40-step flat one!

[22:35:01.0000] <Hixie>
heycam: if i have an empty Storage object, are there any "names of the supported named properties"?

[22:35:02.0000] <heycam>
it is just up to how you define the "names of the supported named properties"

[22:35:03.0000] <heycam>
btw, if you can come up with better terminology for those things, i'm open to suggestions :)

[22:36:00.0000] <Hixie>
heycam: i don't understand "it is just up to how you define the "names of the supported named properties""

[22:37:00.0000] <heycam>
got a pointer to the spec?

[22:37:01.0000] <Hixie>
which?

[22:37:02.0000] <heycam>
Storage?

[22:38:00.0000] <Hixie>
http://dev.w3.org/html5/webstorage/ but i'm more concerned about what the spec _should_ say

[22:38:01.0000] <heycam>
right, i just want to work out what it's trying to say first

[22:38:02.0000] <heycam>
so you have this statement in there at the moment: "The names of the supported named properties on a Storage  object are the keys of each key/value pair currently present in the list associated with the object."

[22:38:03.0000] <heycam>
so if the list associated with the object is empty, then there'll be no named properties

[22:39:00.0000] <heycam>
oh sorry i think i misinterpreted your question earlier

[22:39:01.0000] <heycam>
i thought you were saying "how do you say that there are no names of the supported named properties"

[22:39:02.0000] <heycam>
so the answer to your actual question is: no

[22:40:00.0000] <heycam>
/me back in a couplea minutes, coffee is brewed!

[22:41:00.0000] <Hixie>
ok if the answer is no then: How do I ever get to [NameCreator] given the definition of [[Put]]?

[22:45:00.0000] <heycam>
the answer (it seems) is that you can't

[22:45:01.0000] <heycam>
so you're right, it's buggy

[22:45:02.0000] <Hixie>
woo, i win

[22:45:03.0000] <heycam>
^_^

[22:45:04.0000] <heycam>
similarly for indexes too

[22:46:00.0000] <heycam>
i'll make a note in the spec to fix later

[22:46:01.0000] <Hixie>
so the reason i was looking there is i was hoping that you would automatically dispatch to [NameCreator] or [IndexCreator] based on what we discussed earlier

[22:46:02.0000] <Hixie>
and _not_ have them fallback to each other

[22:46:03.0000] <heycam>
ah i see

[22:46:04.0000] <Hixie>
so that all i have to do to fix Storage is only add [NameCreator] and not add [IndexCreator]

[22:47:00.0000] <heycam>
ok. so that might be better, to distinguish between indexes and names like that.

[22:47:01.0000] <Hixie>
cool

[22:47:02.0000] <heycam>
i wonder about things like the data-* attribute interface tho

[22:47:03.0000] <heycam>
where there are only names, effectively

[22:47:04.0000] <Hixie>
i was about to add IndexCreator to that

[22:47:05.0000] <heycam>
and you want numbers to work there too

[22:47:06.0000] <heycam>
ok

[22:47:07.0000] <Hixie>
NameGetter and NameSetter can still work as now

[22:48:00.0000] <Hixie>
since they work from a fixed list

[22:48:01.0000] <heycam>
yeah you could just have both index properties and name properties defer to the same abstract list of items

[22:48:02.0000] <Hixie>
well as written now the spec already handles name properties that are numbers and already avoids clashes if you have both

[22:48:03.0000] <heycam>
ok i think that makes more sense, rather than making other spec writers worry about confusion of indexes/names

[22:49:00.0000] <Hixie>
so for DOMStringMap we're agreed that all I need is [NameCreator, IndexCreator, NameDeleter, NameGetter, NameSetter] interface DOMStringMap {}

[22:49:01.0000] <Hixie>
and some prose

[22:49:02.0000] <Hixie>
right?

[22:49:03.0000] <heycam>
hum, no

[22:50:00.0000] <heycam>
not if i changed it to distinguish between indexes and names more strongly

[22:50:01.0000] <Hixie>
i would only have it distinguish for creation

[22:50:02.0000] <heycam>
you mean, as written?

[22:50:03.0000] <heycam>
(assuming the bugs for getting to the creators were fixed?)

[22:50:04.0000] <Hixie>
yes

[22:50:05.0000] <heycam>
ok so you don't want to force indexes and names apart?

[22:51:00.0000] <heycam>
except for the creators?

[22:51:01.0000] <Hixie>
right

[22:51:02.0000] <Hixie>
exactly

[22:51:03.0000] <heycam>
ok

[22:51:04.0000] <Hixie>
i think that's what's easiest for spec authors, anyway

[22:51:05.0000] <Hixie>
probably easiest for you too :-)

[22:51:06.0000] <Hixie>
you already do the work of making sure that indexed properties win over named properties in case both are there

[22:51:07.0000] <Hixie>
so there's no clash other than with creation, which right now is buggy adn makes no sense

[22:51:08.0000] <heycam>
do i?

[22:51:09.0000] <heycam>
i thought it was the other way around

[22:51:10.0000] <Hixie>
yeah

[22:52:00.0000] <Hixie>
oh i didn't check which wins

[22:52:01.0000] <Hixie>
yeah i guess named wins

[22:52:02.0000] <heycam>
right, named wins

[22:52:03.0000] <heycam>
and i think i chose that because of how some html interfaces happened to work

[22:52:04.0000] <heycam>
collections or something

[22:53:00.0000] <Hixie>
hmmm

[22:54:00.0000] <Hixie>
i wonder whether i care for Storage

[22:54:01.0000] <Hixie>
because if it's that way, then it doesn't really matter if we distinguish namecreator and indexcreator

[22:54:02.0000] <Hixie>
we can just always have namecreator run, and i don't need to worry about preventing numeric keys being added

[22:54:03.0000] <Hixie>
since it'll still work

[22:55:00.0000] <heycam>
so: if there's no indexcreator, then a property named "1" would still invoke the namecreator?

[22:55:01.0000] <Hixie>
i guess

[22:55:02.0000] <Hixie>
yeah

[22:55:03.0000] <Hixie>
i guess you'd give numbers to the indexcreator if there is one

[22:55:04.0000] <heycam>
mm

[22:55:05.0000] <Hixie>
and if there isn't, just fall back to namecreator

[22:55:06.0000] <heycam>
i guess that'd work

[22:55:07.0000] <heycam>
it's not *too* confusing

[22:56:00.0000] <heycam>
*ahem* do you mind if you mail public-webapps with exactly what you want?

[22:57:00.0000] <Hixie>
sure

[22:57:01.0000] <heycam>
i fear i'll get confused and do something else, otherwise :)

[22:59:00.0000] <heycam>
thanks

[23:01:00.0000] <Hixie>
sent

[23:02:00.0000] <heycam>
cool

[23:58:00.0000] <Hixie>
Philip`: yt?

[01:09:00.0000] <Philip`>
Hixie: Yes

[01:10:00.0000] <Hixie>
it was about the shadows

[01:10:01.0000] <Hixie>
i sent e-mail instead

[01:20:00.0000] <annevk5>
Hixie, any chance you can get to my webstorage feedback soonish?

[01:30:00.0000] <Philip`>
Hixie: I hope there wasn't anything in that e-mail I was meant to respond to

[01:34:00.0000] <Hixie>
annevk5: i can reply to them now if you want

[01:34:01.0000] <Hixie>
annevk5: are they all [webstorage]?

[01:34:02.0000] <annevk5>
Hixie, yes

[01:47:00.0000] <Hixie>
annevk5: 'storage' doesn't really enable fork bombing any more than postMessage() does

[01:47:01.0000] <Hixie>
annevk5: it's just an amplification effect

[01:47:02.0000] <Hixie>
annevk5: you could do the same thing with, say, mutation events (do two mutations per event)

[01:47:03.0000] <Hixie>
unless i misunderstand something

[02:25:00.0000] <Hixie>
annevk2: done

[02:26:00.0000] <Hixie>
Philip`: not really

[02:29:00.0000] <Hixie>
zcorpan_: it seems like the best solution might be to get rid of MEDIA_ERR_NONE_SUPPORTED altogether and _only_ use the error events on the source elements

[02:35:00.0000] <annevk2>
Hixie, fair enough I suppose

[02:35:01.0000] <annevk2>
Hixie, and thanks

[02:39:00.0000] <zcorpan_>
Hixie: hmm. yeah, maybe. for src="" you would fire error on the <video>?

[02:39:01.0000] <Hixie>
yeah, seems reasonable

[02:39:02.0000] <zcorpan_>
Hixie: authors would then have to listen on error on the last <source>, which seems ok

[02:39:03.0000] <Hixie>
yeah

[02:40:00.0000] <zcorpan_>
ok. philipj agreed

[02:43:00.0000] <zcorpan_>
there'd be asymmetry with 'load', but maybe that's ok

[02:44:00.0000] <zcorpan_>
Hixie: should i send email?

[02:44:01.0000] <Hixie>
nah s'ok i got it

[02:44:02.0000] <zcorpan_>
ok

[03:02:00.0000] <Hixie>
zcorpan_: i'll do it tomorrow, getting too tired to work out how to solve the race condition problem he brought up in that other mail

[06:22:00.0000] <Philip`>
jgraham: Your recent email doesn't link to the right URL for data

[06:28:00.0000] <jgraham>
Philip`: Oh. Oops

[06:54:00.0000] <Philip`>
/me wonders if jgraham is planning to post an erratum for that issue

[07:30:00.0000] <jgraham>
/me didn't think that anyone would be interested enough but he can post an erratum if Philip` likes

[07:43:00.0000] <Philip`>
jgraham: Presumably you thought it was worthwhile including the link with your original mail, so I guess it would be similarly worthwhile to include the correct link in a followup

[07:45:00.0000] <Philip`>
/me discovers PGF/TikZ, which is surprisingly nice

[07:45:01.0000] <Philip`>
...as a declarative way of constructing diagrams in LaTeX

[07:49:00.0000] <cryzed>
Hey

[07:49:01.0000] <cryzed>
I'm just here to ask if you could fix the following

[07:49:02.0000] <cryzed>
html5lib/inputstream.py:448: DeprecationWarning: object.__init__() takes no parameters

[07:49:03.0000] <cryzed>
  str.__init__(self, value)

[07:49:04.0000] <cryzed>
Or would that brack compatibility with python 2.5.4?

[07:49:05.0000] <cryzed>
I'm using 2.6.2

[07:59:00.0000] <Philip`>
/me isn't sure whether it's currently possible to fix anything in html5lib, or if its source repository is too much in flux to risk committing any changes

[08:13:00.0000] <jgraham>
cryzed: I remember fixing that for the python 3 branch. So I guess it would be good to fix it on trunk now :)

[08:13:01.0000] <jgraham>
Philip`: It should be stable now I think

[08:15:00.0000] <jgraham>
But I guess you should sithc to using hg

[08:15:01.0000] <jgraham>
*switch

[08:17:00.0000] <cryzed>
So basically

[08:17:01.0000] <cryzed>
a hg source grab

[08:17:02.0000] <cryzed>
should do the trick now?

[08:19:00.0000] <jgraham>
cryzed: In theory :)

[08:19:01.0000] <cryzed>
okay, thanks :)

[08:19:02.0000] <cryzed>
When are you going to make the next "official" release?

[08:20:00.0000] <cryzed>
And on the google-code-homepage the following is mentioned: Support for minidom, ElementTree (including cElementTree and lxml.etree), BeautifulSoup and custom simpletree output formats 

[08:20:01.0000] <cryzed>
I don't know minidom, and simpletree

[08:20:02.0000] <cryzed>
Is there any info about them?

[08:22:00.0000] <ossud>
hi! if i write you guys an email everytime i find an error in the html5 page ... wouldn't that annoy you ?

[08:22:01.0000] <jgraham>
cryzed: minidom is the python stdlib implementation of DOM Core

[08:23:00.0000] <jgraham>
It kinda sucks

[08:23:01.0000] <jgraham>
simpletree was a throaway tree implemntation I wrote to test html5lib

[08:23:02.0000] <cryzed>
So it's shitty? :D

[08:23:03.0000] <ossud>
btw... is there a better way of letting you know about this ?

[08:23:04.0000] <cryzed>
ossud, you can report bugs

[08:23:05.0000] <cryzed>
ossud, @ the google code page

[08:23:06.0000] <cryzed>
ossud, the "Bug Tracker"

[08:24:00.0000] <cryzed>
jgraham, do you know what I would absolutetely love? :D

[08:24:01.0000] <cryzed>
jgraham, A mix of BeautifulSoup and lxml.html etree

[08:24:02.0000] <cryzed>
jgraham, Allowing you the flexibility of BeautifulSoup while Supporting more advanced stuff

[08:24:03.0000] <cryzed>
like XPath

[08:25:00.0000] <ossud>
there are only like 2 issues in html5 O_o

[08:25:01.0000] <ossud>
(on google code)

[08:29:00.0000] <ossud>
how can i get into html5.org/tools/web-apps-tracker ?

[08:44:00.0000] <jgraham>
cryzed: Yeah, you don't want to use simpletreee for anything. Even though it is the default tree type

[08:44:01.0000] <jgraham>
Dunno what to do about that though because you don't want to use minidom for anything either

[08:44:02.0000] <jgraham>
and nothing else comes with the stdlib

[08:45:00.0000] <jgraham>
Actually I guess I could make ElementTree the default. I wonder how many people would complain

[08:45:01.0000] <jgraham>
ossud: What bugs do you want to report? Bugs in the spec?

[08:48:00.0000] <ossud>
mostly typos ^^

[08:49:00.0000] <ossud>
*in the spec

[08:50:00.0000] <jgraham>
ossud: You can either use the w3c.org bugzilla or use the mailing list. I guess one typo per email is not very friendly though :)

[08:51:00.0000] <ossud>
the w3c bugzilla?? how cool is that! :)

[08:51:01.0000] <cryzed>
jgraham, I really like the BeautifulSoup tree though :P

[08:52:00.0000] <cryzed>
jgraham, It's just very hard to correctly extract whole paragraphs of text out of webapges with the ElementTree API, while preserving the formatting and stripping the html tags

[08:55:00.0000] <gsnedders>
Oh stupid slow intarwebs!

[08:56:00.0000] <jgraham>
cryzed: html5lib really needs a maintainer for the BeautifulSoup api

[08:57:00.0000] <cryzed>
jgraham, I don't really have any clue how BS works internally

[08:57:01.0000] <cryzed>
jgraham, I just like to use it :D

[08:58:00.0000] <jgraham>
cryzed: So on a list of ideal candidates you score 1/2 whereas I score 0/2 and I wrote the treebuilder

[08:58:01.0000] <cryzed>
jgraham, haha ^^

[08:58:02.0000] <jgraham>
(does BS support namespaces?)

[08:58:03.0000] <cryzed>
what do you mean exactly

[08:58:04.0000] <cryzed>
with namespaces?

[08:58:05.0000] <jgraham>
The ability to have elements in differnt namespaces

[09:01:00.0000] <jgraham>
like to put an element called svg in a namespace identified by the URI http://www.w3.org/2000/svg

[09:04:00.0000] <cryzed>
I've got no clue actually

[09:04:01.0000] <cryzed>
oO

[09:05:00.0000] <jgraham>
If not, it will be "fun" to get it to work with MathML/SVG. Can probably use Clarke notation and just live with the suckiness or something though

[09:06:00.0000] <cryzed>
jgraham, btw, I'm still getting

[09:06:01.0000] <cryzed>
the error

[09:06:02.0000] <cryzed>
with the newest hg checkout

[09:06:03.0000] <cryzed>
DeprecationWarning: object.__init__() takes no parameters

[09:06:04.0000] <cryzed>
  str.__init__(self, value)

[09:06:05.0000] <cryzed>
Or rather warning

[09:06:06.0000] <jgraham>
cryzed: I didn't apply the patch yet

[09:07:00.0000] <cryzed>
oh

[09:07:01.0000] <cryzed>
shoot

[09:07:02.0000] <cryzed>
Could you? ;D?

[09:11:00.0000] <Philip`>
jgraham: http://www.crummy.com/2009/04/09/0 - "I'll be writing an html5lib tree builder and packaging it and the lxml builder in Beautiful Soup" - he's probably a good person to be maintainer

[09:13:00.0000] <Philip`>
gsnedders: Better than the intarwebs here, which are currently lacking all the normal intar functionality because of power cuts and not-very-good UPSes

[09:13:01.0000] <Philip`>
(unless they've fixed it already)

[09:14:00.0000] <Philip`>
(which I guess they have, since I can access the box sitting at my feet via London)

[09:15:00.0000] <jgraham>
cryzed: Try now

[09:16:00.0000] <cryzed>
kk

[09:16:01.0000] <Philip`>
The commit mail's subject line is horrid

[09:17:00.0000] <Philip`>
[html5lib push] 77f13a977be6f47a9724ee5dbfee756eaf12317e - Fix deprecation warning

[09:17:01.0000] <gsnedders>
Philip`: We had no network/internet at school yesterday because the other side of the split-site school had no power.

[09:17:02.0000] <Philip`>
I want to see the actual commit message in my mail client's little pop-up alert box, not forty hex digits

[09:18:00.0000] <gsnedders>
But how else will you know what revision it is!?

[09:18:01.0000] <Philip`>
It should make up a monotonically increasing integer and use that as the revision number

[09:19:00.0000] <jgraham>
Forty hex digits sitting on the wall, forty hex digits sitting on the wall, and if one hex digit should accidentially fall...

[09:20:00.0000] <gsnedders>
hg won't care.

[09:20:01.0000] <Philip`>
Hopefully the hex digit would pull its partner off the wall too

[09:21:00.0000] <Philip`>
because hex digits are aesthetically displeasing if they don't always come in pairs

[09:30:00.0000] <cryzed>
jgraham, works

[09:30:01.0000] <cryzed>
jgraham, thanks

[09:30:02.0000] <cryzed>
jgraham, It's probably fully 2.6.2 compatible now

[09:30:03.0000] <cryzed>
:D

[10:28:00.0000] <Philip`>
/me receives "508 messages from Google Code"

[10:29:00.0000] <Philip`>
It's quite nice that they handle that situation sensibly, not by sending 508 emails or dropping them all entirely

[12:36:00.0000] <Hixie>
ossud: send as much feedback as you like, one per e-mail or one big e-mail or anything in between, it doesn't matter it'll get treated the same in the end :-)

[14:03:00.0000] <Hixie>
hsivonen: any particular examples you want in the parser section?

[14:07:00.0000] <jgraham>
Unicode or Death?

[14:10:00.0000] <Philip`>
That's a tough choice

[15:31:00.0000] <gsnedders>
/me wonders what anne and Lachy are tweeting

[15:33:00.0000] <Lachy>
gsnedders, 010101110110100001100001011101000011111100100000010110010110111101110101001000000110010001101111011011100010011101110100001000000111001101110000011001010110000101101011001000000110001001101001011011100110000101110010011110010011111100100001

[15:33:01.0000] <gsnedders>
Lachy: I got that. But why?

[15:39:00.0000] <Lachy>
gsnedders, you gotta be here with us in Oslo to understand why.

[15:39:01.0000] <jgraham>
That was much less funny than I had hoped. I guess I don't get the context

[15:39:02.0000] <jgraham>
/me really should be asleep


2009-04-30
[18:05:00.0000] <jcranmer>
where is the character encoding sniffing described?

[18:06:00.0000] <Philip`>
jcranmer: http://www.whatwg.org/specs/web-apps/current-work/multipage/syntax.html#determining-the-character-encoding ?

[18:07:00.0000] <jcranmer>
ども!

[18:07:01.0000] <Philip`>
Pardon?

[18:08:00.0000] <jcranmer>
thanks

[18:51:00.0000] <billyjackass>
wakaba: you there?

[22:56:00.0000] <boblet>
Phillip` are you the Phillip of http://fonts.philip.html5.org/ fame?

[22:58:00.0000] <boblet>
I’m interested in whether this tool can include the modifiable M+ fonts http://bit.ly/mplus as font subsetting for 350KB Japanese fonts would make them actually usable

[23:01:00.0000] <MikeSmith>
it's a boblet!

[23:01:01.0000] <MikeSmith>
boblet: こんちは

[23:01:02.0000] <boblet>
Hey Mike™

[23:02:00.0000] <MikeSmith>
boblet, yeah, that's the same Philip` 

[23:02:01.0000] <boblet>
ワサップ? :)

[23:02:02.0000] <boblet>
Found his email and sending that way just in case

[23:03:00.0000] <MikeSmith>
boblet: I'm enjoying the weather today

[23:03:01.0000] <MikeSmith>
you at home in Osaka?

[23:03:02.0000] <boblet>
350KB Japanese font without subsetting plus Safari not displaying anything on load = noticeable render delay even on hikari

[23:03:03.0000] <MikeSmith>
Philip` is in the UK, so might not be on until 2-3 hours from now

[23:04:00.0000] <MikeSmith>
boblet: yeah, the Safari behavior for rendering of pages with downloadable fonts is suboptimal

[23:04:01.0000] <boblet>
yeah it’s fantastic down here too. I’ll be up May 15-16th for John Allsopp’s workshop tho. If you’re not busy it’d be great to catch up then

[23:05:00.0000] <MikeSmith>
boblet: try http://hsivonen.iki.fi/doctype

[23:05:01.0000] <boblet>
I plan to email the webkit list about it

[23:05:02.0000] <MikeSmith>
boblet: I'll be in Australia from May 8 to May 22

[23:06:00.0000] <MikeSmith>
I think I'll see John there on the 13th before he comes ehre

[23:06:01.0000] <boblet>
aah that sucks (well, for me ;-)

[23:06:02.0000] <MikeSmith>
ap: me and boblet was just talking about webkit rendering of pages with downloadable fonts

[23:06:03.0000] <MikeSmith>
e.g., page like http://hsivonen.iki.fi/doctype

[23:07:00.0000] <ap>
MikeSmith: do we do it wrong?

[23:08:00.0000] <MikeSmith>
ap: just that you basically get no next at all while the font is downloading

[23:09:00.0000] <boblet>
I know the current HTML5 spec doesn’t recommend browser behavoiur atm, and laying out with local font then redrawing is kinda FOUCy for small fonts, but for double-byte (350KB+) fonts no text until download finishes looks like the page is broken

[23:09:01.0000] <MikeSmith>
so on hsivonen's site, which uses some relatively big fonts, there's no text content for quite a while

[23:10:00.0000] <ap>
MikeSmith: hmm, is http://hsivonen.iki.fi/doctype the right url? I don't see any references to downloadable fonts there

[23:10:01.0000] <MikeSmith>
boblet: right -- that's the problem. casual users are going to assume its's a browser bug

[23:10:02.0000] <boblet>
I only used @font-face for the <h1> here (dark green band) http://oli-studio.com/work/wde/200905-roadshow/ but there’s still a noticeable delay

[23:10:03.0000] <ap>
MikeSmith: but yeah, I saw someone complain about that - not sure if there's a bug filed

[23:11:00.0000] <MikeSmith>
ap: open Web Inspector on http://hsivonen.iki.fi/doctype and look at the Resources tab

[23:11:01.0000] <boblet>
FF 3.5 beta renders the text using a local font first then redraws. This would also be much safer if the font doesn’t finish downloading (eg server times out)

[23:11:02.0000] <MikeSmith>
5MB+ of fonts there

[23:11:03.0000] <ap>
MikeSmith: test-quirks.php and spacer.gif

[23:11:04.0000] <boblet>
5MB? I feel like small fry :D

[23:12:00.0000] <ap>
MikeSmith: ah, I see - I didn't notice that I got auto-completion of the url, and was looking at another page

[23:13:00.0000] <MikeSmith>
ap: o_0

[23:13:01.0000] <MikeSmith>
boblet: ap is a webkit developer

[23:13:02.0000] <boblet>
MikeSmith: he’s waiting for something to display… ;-)

[23:13:03.0000] <ap>
MikeSmith: not one having anything to do with fonts though

[23:15:00.0000] <boblet>
ap: props to the Webkit team for all the fun goodies you’re giving us! and where do I go to complain about tabs not being maintained automatically between sessions? ;-)

[23:15:01.0000] <MikeSmith>
boblet: there's a way you can save them manually at least

[23:15:02.0000] <MikeSmith>
"Reopen all windows from last session"

[23:16:00.0000] <ap>
boblet: as MikeSmith says, it's a feature - but it's in Safari, not in WebKit, of course

[23:16:01.0000] <boblet>
MikeSmith: yeah I know, but I’ve had to pull previous versions of LastSession.plist out of git several times so far

[23:17:00.0000] <boblet>
ap: damn, foiled again!

[23:18:00.0000] <boblet>
ap: is Safari 4b feature complete?

[23:18:01.0000] <ap>
boblet: no announcements were made, as far as I know

[23:19:00.0000] <boblet>
ap: I guess I shouldn’t assume you have the inside track on what webkit build the Safari team will roll with huh

[23:23:00.0000] <olliej>
boblet: apple engineers cannot discuss anything like that :D

[23:25:00.0000] <olliej>
boblet: in general if apple has not announced it, we can't discuss it

[23:25:01.0000] <boblet>
ap: actually I came across something else a while back that I didn’t know what to make of in Webkit: Japanese text appearing to break out of it’s inline box when Hoefler Text is assigned: http://oli-studio.com/code/font-set-inline-box.html

[23:25:02.0000] <olliej>
boblet: filing a bug at bugs.webkit.org would be great :D

[23:25:03.0000] <boblet>
olliej: heh, indeed. It’s part of Apple’s charm

[23:26:00.0000] <olliej>
boblet: hey, at least we do our development publically

[23:26:01.0000] <olliej>
boblet: for the purposes of engine development the "webkit team" == Apple :D

[23:26:02.0000] <boblet>
olliej: yeah I think that was next on the list of things to do, right before Christmas attacked

[23:29:00.0000] <olliej>
boblet: ah, you're talking about the top overflow?

[23:30:00.0000] <boblet>
olliej: I may well be.…?

[23:30:01.0000] <ap>
boblet: try Apple Chancery - I think it may break out even more. likely a known issue, but filing a bug is the best way to reach people who work on text rendering

[23:31:00.0000] <boblet>
ap: ok. I’ll do that for both issues

[23:31:01.0000] <olliej>
boblet: and link in your testcase -- we love testcases

[23:31:02.0000] <boblet>
I kinda went overboard on that one huh

[23:31:03.0000] <olliej>
boblet: bugs are vastly more likely to be fixed with a reduced testcase

[23:32:00.0000] <olliej>
boblet: vs. entire bug content being "japanese text overflows with the Hoefler Text font"

[23:32:01.0000] <olliej>
boblet: i'm guessing the font is giving "interesting" metrics to us

[23:32:02.0000] <boblet>
olliej: ok. I’ll make a reduced version and link it from the fat version

[23:33:00.0000] <olliej>
boblet: even your current version is vastly better than what we normally get

[23:33:01.0000] <olliej>
so i'd initially file a bug that links to that url

[23:47:00.0000] <hsivonen>
ap: already filed https://bugs.webkit.org/show_bug.cgi?id=25207

[23:48:00.0000] <ap>
hsivonen: thanks

[00:08:00.0000] <zcorpan_>
Hixie: "if the user agent has not fired a timeupdate event at the element in the past 15 to 250ms" -- s/15 to // ?

[00:09:00.0000] <Hixie>
zcorpan_: UAs are allowed to do it every 15ms

[00:09:01.0000] <zcorpan_>
Hixie: yes, but if it hasn't fired it in hte past 250ms, then it won't have fired it in the past 15ms either

[00:10:00.0000] <zcorpan_>
oh wait

[00:10:01.0000] <zcorpan_>
nm

[00:11:00.0000] <Hixie>
zcorpan_: :-)

[00:21:00.0000] <weinig>
hi olliej 

[01:01:00.0000] <jgraham>
boblet: Note that afaict Philip`'s natural biological rhythms seem to place him somewhere in the mid atlantic timezone wise so it may yet be a while before he is around

[01:02:00.0000] <boblet>
jgraham hehe, thanks. I’m gonna email him just in case

[01:03:00.0000] <MikeSmith>
jgraham: you up early. 8am now?

[01:04:00.0000] <jgraham>
MikeSmith: 9am here

[01:04:01.0000] <jgraham>
/me wonders what kind of weird algorithm Apple are using where 0% battery life = 15 minutes of reamining time

[01:05:00.0000] <jgraham>
(and it really was 0% because the power suddenly died)

[01:06:00.0000] <Hixie>
jgraham: i have the opposite experience

[02:03:00.0000] <Philip`>
boblet: That's me :-)

[02:03:01.0000] <Philip`>
Oh, you're not here

[02:23:00.0000] <a-ja>
!seen hsivonen

[02:25:00.0000] <MikeSmith>
a-ja: no bots here, only peoples

[02:25:01.0000] <MikeSmith>
I think hsivonen is awake but busy

[02:25:02.0000] <a-ja>
tks Mike

[03:27:00.0000] <MikeSmith>
hendry : I've put together part of a HTML document splitter in XSLT

[03:28:00.0000] <MikeSmith>
mostly just by borrowing stuff from that Docbook XSL stylesheets distro

[03:28:01.0000] <MikeSmith>
I still need to get the in page next/prev/up/down navigation stuff working

[03:30:00.0000] <Philip`>
You get bonus points for including the relevant TOC extract at the top of each page

[03:30:01.0000] <MikeSmith>
Docbook stylesheets have a way to do that, but I'm not using it

[03:31:00.0000] <MikeSmith>
default output from DocBook stylesheets looks like this:

[03:32:00.0000] <MikeSmith>
http://docbook.sourceforge.net/release/xsl/current/doc/html/chunker.output.cdata-section-elements.html

[03:32:01.0000] <MikeSmith>
includes the major section name in the page header

[03:32:02.0000] <MikeSmith>
title

[03:32:03.0000] <MikeSmith>
and the titles of the next page and previous page in teh footer

[04:37:00.0000] <Philip`>
http://www.tbray.org/ongoing/When/200x/2009/04/29/Model-and-Syntax - "I have long believed and repeatedly written that bits on the wire are the only serious reliable basis for interoperability, and worried in public about the feasibility of shared models"

[04:38:00.0000] <Philip`>
That seems kind of completely opposite to the HTML 5 philosophy, which is that the document model is the important thing, and the bits on the wire are just one of many ways of serialising the model

[04:39:00.0000] <MikeSmith>
Philip`: interesting

[04:40:00.0000] <Philip`>
/me wonders how the connect the different viewpoints

[04:40:01.0000] <MikeSmith>
I wonder if that's a key difference between some of the people who come from same background and Tim, and, say, browser implementors or client-side web-apps developers

[04:41:00.0000] <jgraham>
Philip`: Since HTML5 defines exactly how you get from bits on the wire to the internal model and back again, I'm not so sure that's true

[04:41:01.0000] <MikeSmith>
hmm, "I believe the existence and success of the Internet and the Web are strong evidence in my favor. They have no object models or APIs; nor could they: they are just a collection of agreements what bits we send each other, with accompanying narrative about what those bits are supposed to mean."

[04:42:00.0000] <MikeSmith>
is that statement accurate?

[04:43:00.0000] <Philip`>
jgraham: But the main basis for interoperability is the rules for how UAs process the document model, and isn't just about the syntax

[04:43:01.0000] <Philip`>
unlike XML, which is defined as purely syntax

[04:44:00.0000] <MikeSmith>
Philip`: yeah, that seems to me too like an accurate way to describe it

[04:44:01.0000] <zcorpan>
MikeSmith: the Web uses script, and the scripts assume the DOM as the model

[04:44:02.0000] <jgraham>
What zcorpan said

[04:44:03.0000] <Philip`>
(and that purely syntactic definition causes problems with e.g. when do you execute <script>s while parsing XHTML, so the model is valuable too)

[04:45:00.0000] <jgraham>
I don't see how the idea of client side scripting fits with the "it's all just bits" viewpoint

[04:45:01.0000] <jgraham>
(that applies to XML too of course)

[04:47:00.0000] <jgraham>
"""I’ll tell you this: As a TDD disciple, if I read the initial spec language, I’d have cooked up a bunch of test cases that would have explored the value space pretty thoroughly before I shipped any code. Tighten the spec and I might have gotten lazy."""

[04:47:01.0000] <jgraham>
That is the single worst defence of bad specs I have ever read

[04:48:00.0000] <MikeSmith>
this makes me think of the discussion about the whether it's a mistake to attempt to separately spec a particular language without in the same document also specifying how it is meant to be processed

[04:49:00.0000] <MikeSmith>
jgraham: yeah, on the face of it, that seems not particularly compelling argument

[04:49:01.0000] <Philip`>
It still seems true that bits-on-the-wire are important by themselves - people think of HTML and XML as character strings, and that's a useful and successful view, and we'd lose that if we just had the DOM API and some unreadable custom binary protocol for transferring DOMs

[04:49:02.0000] <MikeSmith>
btw, what's TDD?

[04:49:03.0000] <Philip`>
Test-driven development

[04:49:04.0000] <MikeSmith>
ah

[04:51:00.0000] <Philip`>
I suppose the focus on bits-on-the-wire vs document models could also be blamed for e.g. RSS readers that just use regexps to interpret those bits

[04:53:00.0000] <MikeSmith>
Philip`: yeah, good example i guess

[04:53:01.0000] <MikeSmith>
that's a genuine problem

[04:54:00.0000] <Philip`>
Maybe the bits-on-the-wire view is more useful during the early stages of a protocol, when we want to encourage competition between widely varying ways of implementing and using the protocol; but as the technology matures, the dominant document model should be specified and standardised, to improve interoperability

[04:54:01.0000] <MikeSmith>
mooncalves writing parsers in regex

[04:54:02.0000] <MikeSmith>
Philip`: seems like that's the way it evolved with HTML anyway

[04:55:00.0000] <Philip`>
Maybe RSS->Atom could be seen the same way?

[04:55:01.0000] <MikeSmith>
yeah

[04:56:00.0000] <Philip`>
/me can't think of any other examples...

[04:56:01.0000] <MikeSmith>
well, maybe XML processing in general could not evolve that way because of you-know-what

[04:57:00.0000] <MikeSmith>
spec'ing catch-fire-and-fail error handling seems like a pretty good way to stop natural evolution

[04:57:01.0000] <Philip`>
I don't quite see why that's relevant

[04:58:00.0000] <Philip`>
Someone could define a processing model for XML that turns a stream of bytes into a DOM or Infoset or SAX stream or whatever, and that'd solve the questions about when scripts execute and how document.write could be fitted in etc, but they don't seem to have done that yet

[04:58:01.0000] <Philip`>
(and draconity doesn't have anything to do with that, as far as I can tell)

[04:58:02.0000] <MikeSmith>
Philip`: if you want processing applications to be able do deal with content that people actually produce, rather than the content you assume ahead of time they will produce

[04:59:00.0000] <Philip`>
(There's also things like Jabber which rely somewhat on the processing model of XML parsers, e.g. elements should be reported as soon as their close tag's '>' is seen)

[05:01:00.0000] <MikeSmith>
as far as the mention that Tim makes of agreements: the agreements in the case of browsers have been a moving target

[05:02:00.0000] <MikeSmith>
"unstable equilibrium" I think is how hsivonen or someone referred to i

[05:02:01.0000] <MikeSmith>
it

[05:03:00.0000] <Philip`>
/me goes away for a bit

[06:38:00.0000] <Philip`>
http://img.thedailywtf.com/images/200904/mandelbrot.xml - ouch

[06:46:00.0000] <annevk42>
respect

[06:51:00.0000] <heycam>
very nice!

[06:54:00.0000] <heycam>
zot:~ $ lynx -dump -nolist http://www.w3.org/TR/SVG/refs.html | grep -B1 -A3 fixme # whoops...

[06:54:01.0000] <heycam>
   [XHTMLplusMathMLplusSVG]

[06:54:02.0000] <heycam>
          @@fixme "An XHTML + MathML + SVG Profile", M. Ishikawa editor, 9

[06:54:03.0000] <heycam>
          August 2002.

[06:54:04.0000] <heycam>
          Available at:

[06:54:05.0000] <heycam>
          http://www.w3.org/TR/2002/WD-XHTMLplusMathMLplusSVG-20020809/

[06:54:06.0000] <heycam>
actually i wonder what the fixme is for

[06:55:00.0000] <heycam>
doesn't look like that document got past WD anyway

[06:57:00.0000] <annevk42>
hey heycam

[06:57:01.0000] <heycam>
hey annevk42

[06:57:02.0000] <annevk42>
heycam, do you know when you're going to edit webidl again?

[06:58:00.0000] <annevk42>
I was wondering if it could provide some hook for when an interface object is accessed

[06:58:01.0000] <heycam>
ar, when i get some time. i can certainly add some notes for features, with particular syntax, so you can start using it..

[06:58:02.0000] <annevk42>
e.g. if a script does var something = somewindow.XMLHttpRequest or var x = new XMLHttpRequest() I need an implementation to set some internal variables on the interface object (later copied to the constructed object)

[06:59:00.0000] <heycam>
sounds weird :)

[06:59:01.0000] <heycam>
what needs it?

[07:00:00.0000] <heycam>
(the latter you could just define things to happen in the constructor)

[07:00:01.0000] <annevk42>
xhr needs some kind of pointer to the Document object

[07:01:00.0000] <heycam>
and it needs to grab it at exactly the time somewindow.[[Get]]("XMLHttpRequest") is done?

[07:01:01.0000] <annevk42>
if it grabs it later the current document could be away

[07:02:00.0000] <Philip`>
What if somewindow does "window.foo = window.XMLHttpRequest" and then someone else does somewindow.[[Get]]("foo")?

[07:03:00.0000] <annevk42>
Philip`, it should be bound to the initial window

[07:03:01.0000] <annevk42>
I wonder how localStorage/sessionStorage solves this

[07:04:00.0000] <heycam>
i think i'm not understanding excatly what needs to happen

[07:04:01.0000] <heycam>
...and i'm being called to go to bed. can you mail public-webapps annevk42?

[07:04:02.0000] <annevk42>
k

[07:04:03.0000] <heycam>
thanks

[07:26:00.0000] <annevk42>
Philip`, afaict it is bound the initial time it is accessed

[07:34:00.0000] <MikeSmith>
"If you were to use The Daily WTF as a guide, your impression of XSL Transformations (XSLT) would probably be fairly low. I mean, seeing article after article after article might have given the impression that XSLT is often not the right tool for the job... or, perhaps, maybe not even a right tool. Period."

[07:35:00.0000] <MikeSmith>
http://thedailywtf.com/Articles/Where-the-Wild-Web-Things-Are.aspx is nice too

[07:35:01.0000] <MikeSmith>
(great find by Hallvord R. M. Steen)

[07:37:00.0000] <Philip`>
I wish free fonts were more discoverable

[07:37:01.0000] <Philip`>
/me found some GFDL ones for the Khmer language, wrapped inside an .exe installer on a Yellow Pages site

[07:38:00.0000] <Philip`>
(The comments inside the font point at a Geocities site which shows the default placeholder page)

[07:38:01.0000] <Philip`>
s/GFDL/LGPL/

[07:52:00.0000] <wakaba>
MikeSmith: ?

[07:53:00.0000] <MikeSmith>
wakaba: was going to ask you off-topic question

[07:54:00.0000] <MikeSmith>
which is, I see you're using a e-mobile modem

[07:54:01.0000] <wakaba>
yeah

[07:54:02.0000] <MikeSmith>
and was wondering if there's a way to get around the arbitrary time limit

[07:55:00.0000] <MikeSmith>
or maybe if they have changed the time-limit thing

[07:55:01.0000] <MikeSmith>
specifically, the driver I have on my Mac OSX machine drops that connection after 6 hours or use

[07:56:00.0000] <wakaba>
my connection seems also time limited

[07:56:01.0000] <wakaba>
i don't know how to avoid it...

[07:59:00.0000] <wakaba>
i was offline today because and redial feature did not work for some reason...

[08:00:00.0000] <MikeSmith>
wakaba: I see

[08:01:00.0000] <MikeSmith>
I seem to rembember somebody telling me that there is a newer driver that doesn't have the time limit, but maybe I imagined that

[08:03:00.0000] <annevk42>
/me hopes his email is clear enough for heycam

[09:42:00.0000] <gsnedders>
Why do I think having a sprained ankle the day before the ball is not a good idea?

[09:42:01.0000] <Philip`>
It provides you with a good excuse to not dance

[09:43:00.0000] <Philip`>
On the other hand, it will be harder for you to run away from any nearby bombs or aliens

[10:29:00.0000] <mpilgrim>
grr http://lists.w3.org/Archives/Public/public-html/2009Apr/0260.html

[10:31:00.0000] <mpilgrim>
"W3C HTML5 spec"

[10:31:01.0000] <mpilgrim>
i can't decide if this is intentionally divisive or just stupid

[10:32:00.0000] <mpilgrim>
if it had come from sam, i would guess "intentionally divisive," but here i'm gonna go with "stupid"

[10:35:00.0000] <Philip`>
Sounds quite reasonable to me, as a way of referring to the document that's published by the W3C and is called "HTML5"

[10:35:01.0000] <Philip`>
...and is a spec

[10:58:00.0000] <Hixie>
http://www.w3.org/mid/FAA1D89C5BAF1142A74AF116630A9F2C0A26ED872D⊙Ooac - money first, principles after

[11:02:00.0000] <Hixie>
/me wonders what to do about <header>

[11:04:00.0000] <Philip`>
Rename it to <hgroup>

[11:04:01.0000] <Philip`>
and then create a new element called <header>

[11:04:02.0000] <Philip`>
which is basically equivalent to <div class=header>, because people want semantic elements

[11:05:00.0000] <Philip`>
Easy :-)

[11:05:01.0000] <Hixie>
yeah that was basically what i was thinking of doing

[11:05:02.0000] <Hixie>
if i do that i'll change <hgroup>'s content model to just be <hx>+, i think

[11:05:03.0000] <Hixie>
and make <header> not participate in any special stuff like sectioning

[11:06:00.0000] <Hixie>
does that make sense?

[11:07:00.0000] <Philip`>
I think it makes sense to me, but I don't think there's much correlation between that and whether it's actually a good idea or not

[11:08:00.0000] <Philip`>
But if you do those changes, then you could drop <hgroup> from HTML5, and add it back into HTML6 if it turns out that people actually use the outlining algorithm and want to avoid the minor unsightliness of empty sections

[11:09:00.0000] <Hixie>
given that every w3c spec does the h1-h2 thing at the start, i think it's clear that either the outlining algorithm is wrong without hgroup, or we need hgroup

[11:22:00.0000] <a-ja>
hmmm...when did header element morph into hgroup ?

[11:22:01.0000] <Hixie>
about 5 minutes ago

[11:22:02.0000] <Hixie>
i'm about to check in the change and then introduce a new element to do more of a site-wide heading thing

[11:23:00.0000] <Hixie>
why, is it a bad idea? :-)

[11:23:01.0000] <a-ja>
don't forget the examples               :)

[11:26:00.0000] <Hixie>
a-ja: i think i got them all

[11:26:01.0000] <a-ja>
not a  bad idea as   far as i'm concerned.  i'll just need to update  soame pages/templates/stylesheets

[11:28:00.0000] <Philip`>
Because you're currently using <header> on those pages?

[11:28:01.0000] <a-ja>
yep

[11:28:02.0000] <Philip`>
What purpose are you using it for?

[11:28:03.0000] <Philip`>
(like, is it actually to group multiple <hX> elements?)

[11:29:00.0000] <krijnh>
Wasn't <header><hx>..</hx><p>Tagline</p></header> correct usage as well?

[11:30:00.0000] <a-ja>
was using  it per spec....h1's & h2's, sometimes with a p in  it too

[11:30:01.0000] <Philip`>
krijnh: It used to be, but now you can't do that with <hgroup>

[11:30:02.0000] <a-ja>
what krijnh    said :)

[11:30:03.0000] <krijnh>
With <hgroup> that's a bit weird indeed

[11:30:04.0000] <krijnh>
<heading> would be better, I think

[11:30:05.0000] <Philip`>
Why do you want to group the <p> with the <hx>?

[11:30:06.0000] <a-ja>
no more flow content?

[11:31:00.0000] <krijnh>
<header> sounds too much like it replaces <div id="header">

[11:32:00.0000] <Hixie>
i plan to introduce a new element called <header> that can act as a more generic container for header-like content

[11:32:01.0000] <a-ja>
Philip`: tagline/slogan sorta thing, as opposed to a subtitle

[11:32:02.0000] <Hixie>
including <nav>

[11:32:03.0000] <Hixie>
more on par with "article" than what <header> used to be

[11:32:04.0000] <Hixie>
and <hgroup> now only accepts h1-h6

[11:32:05.0000] <a-ja>
/me like addition of nav

[11:33:00.0000] <krijnh>
Hixie: a document-wide header thingy?

[11:33:01.0000] <a-ja>
search form, too?

[11:34:00.0000] <Hixie>
i could go either way, document-wide or something else

[11:34:01.0000] <Hixie>
what do people want?

[11:34:02.0000] <Hixie>
i could make it that it can't be in a section other than <body>

[11:34:03.0000] <krijnh>
I use <div id="header"> a lot

[11:34:04.0000] <krijnh>
With navigation, search, logo, that stuff

[11:34:05.0000] <Philip`>
It's quite common to have >1 <div class=header> on a page

[11:35:00.0000] <Hixie>
probably just make it equivalent to a <div> then, except with slightly more semantic juice

[11:35:01.0000] <Philip`>
Uh, I think I mean <* class=header>

[11:36:00.0000] <krijnh>
Hixie: is that enough for a new element then?

[11:36:01.0000] <Philip`>
http://lists.w3.org/Archives/Public/public-html/2009Mar/0679.html - oh, good, someone already counted it

[11:36:02.0000] <a-ja>
implicit role=banner

[11:36:03.0000] <a-ja>
w/o                  override ?

[11:36:04.0000] <krijnh>
Philip`: also have data for id="header" ?

[11:37:00.0000] <Philip`>
krijnh: I doubt many people have >1 id="header" on a page :-p

[11:37:01.0000] <krijnh>
That's not what I mean :)

[11:38:00.0000] <Philip`>
What do you mean, then? :-)

[11:39:00.0000] <krijnh>
How many pages uses id="header" as a page-wide thingy

[11:39:01.0000] <krijnh>
Also, the class="header" numbers also include <hx> and <p>, I guess?

[11:40:00.0000] <Philip`>
Out of ~130K, I see 11754 pages with one id=header

[11:40:01.0000] <Philip`>
and 71 with 2

[11:40:02.0000] <Philip`>
and 8 with 3

[11:40:03.0000] <Philip`>
and 15 with 4

[11:40:04.0000] <Philip`>
and 2 with 5, and 2 with 6

[11:40:05.0000] <Philip`>
and 1 with 8, and 1 with 17

[11:40:06.0000] <Philip`>
Someone should tell them that id is meant to be unique :-/

[11:41:00.0000] <krijnh>
Heh :)

[11:41:01.0000] <Philip`>
The class=header numbers include all elements, I believe

[11:41:02.0000] <krijnh>
So more id="header" usage than class="header"

[11:41:03.0000] <Philip`>
Yes, by approximately a factor of 4

[11:41:04.0000] <krijnh>
Yeh

[11:42:00.0000] <krijnh>
And now the cases where class is used only on <div> and <td>? :)

[11:42:01.0000] <krijnh>
That's probably what something like <hgroup> represents

[11:42:02.0000] <Hixie>
when i did some studies on id="" i found some pages with thousands of duplicate IDs

[11:43:00.0000] <krijnh>
Glad I'm not a browser developer

[11:45:00.0000] <Philip`>
I see 896 id="menu_categories" on one page

[11:56:00.0000] <Hixie>
ok new text is in

[11:56:01.0000] <Hixie>
what do people think

[11:57:00.0000] <krijnh>
Hixie: to mark up a page's title?

[11:58:00.0000] <Hixie>
i mean, what do you think of the next text :-)

[11:58:01.0000] <Hixie>
http://www.whatwg.org/specs/web-apps/current-work/#the-hgroup-element http://www.whatwg.org/specs/web-apps/current-work/#the-header-element

[12:00:00.0000] <gsnedders>
/me dislikes the change, as it means his implementation of the outlining algorithm has to change

[12:00:01.0000] <gsnedders>
(And that requires effort)

[12:00:02.0000] <a-ja>
heh

[12:00:03.0000] <a-ja>
henriwill curse you, too 

[12:01:00.0000] <gsnedders>
I will strike a curse in the name of the Lord, and two she-bears will appear from the woods and maul the annoying youths to death!

[12:01:01.0000] <krijnh>
Hixie: yeah, isn't it to mark up a section's title with its subtitle or tagline?

[12:01:02.0000] <Hixie>
gsnedders: s/header/hgroup/ is the only change you have to make i think

[12:01:03.0000] <gsnedders>
Hixie: So do I.

[12:01:04.0000] <gsnedders>
Hixie: Then hg commit; hg push, and then manually update the copy of the outliner on my server

[12:01:05.0000] <Hixie>
krijnh: i don't understand your question

[12:02:00.0000] <krijnh>
I don't understand the line 'The hgroup element represents the header of a section. The element is used to group a set of h1�h6 elements to mark up a page's title with its subtitle or tagline.'

[12:02:01.0000] <Hixie>
ohhhh

[12:02:02.0000] <Hixie>
hm

[12:02:03.0000] <krijnh>
:)

[12:02:04.0000] <Hixie>
that was the text that was there before

[12:02:05.0000] <Hixie>
let's see

[12:02:06.0000] <Hixie>
i'm sure it can be improved

[12:02:07.0000] <Hixie>
/me pokes

[12:02:08.0000] <gsnedders>
http://www.biblegateway.com/passage/?search=2%20Kings%202:23-25;&version=31; is what I was referencing above

[12:03:00.0000] <gsnedders>
Hixie: Also see the bug in bugzilla about header

[12:03:01.0000] <gsnedders>
(and sectioning roots)

[12:03:02.0000] <gsnedders>
Or have you looked at that?

[12:03:03.0000] <gsnedders>
/me RTFS

[12:04:00.0000] <Hixie>
which one?

[12:04:01.0000] <gsnedders>
The one I was meaning.\

[12:05:00.0000] <Hixie>
-_-

[12:05:01.0000] <gsnedders>
/me looks

[12:05:02.0000] <Hixie>
krijnh: look in about 90 seconds and tell me if the new text is clearer

[12:05:03.0000] <krijnh>
Oki

[12:06:00.0000] <gsnedders>
Hixie: http://www.w3.org/Bugs/Public/show_bug.cgi?id=6750

[12:06:01.0000] <gsnedders>
"The hgroup element represents the header of a section." v. "The header element represents a header for the section it applies to."

[12:06:02.0000] <gsnedders>
Um, yeah.

[12:07:00.0000] <krijnh>
Yeah, I don't really get it either :)

[12:07:01.0000] <gsnedders>
Hixie: Write a better pec.

[12:07:02.0000] <gsnedders>
*spec

[12:08:00.0000] <Hixie>
hmm

[12:08:01.0000] <Hixie>
wonder how to distinguish these

[12:08:02.0000] <Hixie>
h1-h6 also "define headers for their sections"

[12:08:03.0000] <krijnh>
Why do we need hgroup again?

[12:09:00.0000] <Hixie>
subheadings

[12:09:01.0000] <Hixie>
as in every w3c spec for instance

[12:09:02.0000] <Hixie>
and taglines, as seen on most blogs

[12:09:03.0000] <krijnh>
Yeah, why not just use <header><hx>..</hx><p>Subheading</p></header> ?

[12:09:04.0000] <Hixie>
doesn't seem as right to me :-)

[12:10:00.0000] <krijnh>
And say that a p element in a header should be seen as a subheading :)

[12:10:01.0000] <Hixie>
i mean, what's the difference between that <p> and a <p>Last Modified: ...</p> or <p>Site links: ...</p> ?

[12:10:02.0000] <krijnh>
Why should there be a difference?

[12:11:00.0000] <krijnh>
For outliners?

[12:12:00.0000] <Hixie>
krijnh: amongst other things, yeah

[12:17:00.0000] <krijnh>
It's better now, and I like the new <header>

[12:17:01.0000] <a-ja>
fwiw, only  reason i'd been using p instead  of h2 for tagline  was cuz it played better with CITA's toolbar

[12:19:00.0000] <krijnh>
Hixie: does a <header> require a <hx> ?

[12:21:00.0000] <Hixie>
krijnh: no

[12:22:00.0000] <Hixie>
a-ja: hopefully when they support <hgroup> that will no longer be n issue

[12:22:01.0000] <a-ja>
gives em something to do next semester:)

[12:23:00.0000] <tantek>
Hixie, what is <hgroup> ?

[12:24:00.0000] <Hixie>
groups <h1>-<h6> elements so you can have subheadings without implying a subsection

[12:24:01.0000] <Hixie>
the way, e.g., most w3c specs do

[12:24:02.0000] <krijnh>
Formerly known as <header>

[12:27:00.0000] <tantek>
so <hgroup> is a renaming of <header> ? if so, curious about the reasons for the renaming.

[12:27:01.0000] <Hixie>
we wanted to introduce a new element that was more of a generic header element

[12:28:00.0000] <Hixie>
since that's what people used the old <header> for

[12:29:00.0000] <Hixie>
oh hey while you're here tantek it would be good to have your feedback on the use cases listed in this e-mail: http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-April/019374.html

[12:29:01.0000] <krijnh>
The <header> example now contains text which should be in the <footer> :)

[12:29:02.0000] <krijnh>
Isn't <header> / <footer> really a presentational issue now?

[12:29:03.0000] <Hixie>
tantek: in particular, which requirements or use cases you think we shouldn't address, and which you think are already addressed

[12:30:00.0000] <Hixie>
tantek: and if you have any ideas on addressing the ones that aren't yet addressed

[12:30:01.0000] <krijnh>
"A footer typically contains information about its section such as who wrote it, links to related documents, copyright data, and the like."

[12:30:02.0000] <Hixie>
krijnh: it's not really presentational, but it's certainly easy to see how an Architect might consider them the same thing and merge them into one element

[12:31:00.0000] <tantek>
Hixie, general comment, the scenarios seem fairly reasonable overall, and yet the requirements appear to be fairly arbitrary, in many cases unnecessary, and disconnected from the scenarios.

[12:31:01.0000] <Hixie>
krijnh: updated the intro paragraphs for <hgroup> and <header>, let me know if they're more understandable now

[12:31:02.0000] <tantek>
In my experience, whenever requirements are disconnected from scenarios, the requirements have been artificially written to pre-suppose a specific solution.

[12:32:00.0000] <Hixie>
yeah that was my general impression too

[12:32:01.0000] <tantek>
In otherwords, the requirements should be derived from the scenarios

[12:32:02.0000] <tantek>
and scenarios really should be derived from real world web publishing examples (with URLs provided)

[12:33:00.0000] <tantek>
any requirements that are not derived from a scenario should be labeled hypothetical/a-priori and dropped

[12:33:01.0000] <Hixie>
okie dokie. sounsd basically like what i figured. if you have any specific comments on any of the issues, drop me a mail, i'm planning on looking at these in more detail relatively soon.

[12:34:00.0000] <krijnh>
Hixie: s/(An/(an/ but other than that I think it's better

[12:34:01.0000] <tantek>
any scenarios which do not withstand the test/request for basis in real world web publishing examples (with URL) should be considered "second class" / optional in comparison to scenarios bounded in real world web publishing examples.

[12:35:00.0000] <tantek>
in otherwords, I am not against "new"/theoretical scenarios, because that's often how new problems are solved. however, solutions should solve existing real world problems first, and new/theoretical problems second IMHO.

[12:35:01.0000] <tantek>
it's a matter of prioritization

[12:39:00.0000] <tantek>
one exception to the requirements feedback - it's reasonable to consider requirements that are based on well known/practiced/tested principles of data format design. this of course depends on agreeing/disagreeing on principles.

[12:39:01.0000] <tantek>
for example this requirement:

[12:39:02.0000] <tantek>
"Machine-readable event data shouldn't be on a separate page than

[12:39:03.0000] <tantek>
       human-readable dates."

[12:40:00.0000] <tantek>
could be expressed as a specific instance of the DRY (don't repeat yourself) principle.

[12:41:00.0000] <tantek>
and perhaps even a specific subrequirement of "Should be unlikely to get out of sync with prose on the page."

[12:42:00.0000] <tantek>
Just to point out one egregious requirement (which I think is quite questionable).

[12:42:01.0000] <tantek>
This one sounds highly theoretical until/unless sufficient real world examples (with URLs) are provided that demonstrate the need: "Should be possible for different parts of an event to be given in

[12:42:02.0000] <tantek>
       different parts of the page. For example, a page with contact details

[12:42:03.0000] <tantek>
       for people in columns (with each row giving the name, telephone

[12:42:04.0000] <tantek>
       number, etc) should still have unambiguous grouped contact details

[12:42:05.0000] <tantek>
       parseable from it."

[12:43:00.0000] <Hixie>
tantek: that particular one has been demonstrated to me several times, though i don't have URIs handy right now

[12:43:01.0000] <tantek>
(note the typo "an event" in that requirement, as it was obviously copy-pasted from  the previous version which has the same theoreticalness flaw

[12:43:02.0000] <tantek>
"Should be possible for different parts of an event to be given in

[12:43:03.0000] <tantek>
       different parts of the page. For example, a page with calendar events

[12:43:04.0000] <tantek>
       in columns (with each row giving the time, date, place, etc) should

[12:43:05.0000] <tantek>
       still have unambiguous calendar events parseable from it."

[12:44:00.0000] <Hixie>
but basically there are people who want to add events, contacts, etc, to pages that aren't conforming, and won't be conforming

[12:44:01.0000] <Hixie>
and they're willing to use rdfa rather than microformats because of this requirement (!)

[12:45:00.0000] <Hixie>
myspace was the main example of this, but i've been shown others

[12:45:01.0000] <tantek>
without the URIs to specific examples, arguments about what solution to use are then based in theory/religious reasoning rather than actual practicality

[12:45:02.0000] <tantek>
which is ironic for an requirement based on "pages that aren't conforming, and won't be conforming" - which is obviously an appeal to practicality

[12:46:00.0000] <Hixie>
one example would google's internal contacts app

[12:46:01.0000] <Hixie>
which has the information about the main person intermixed with manager information and reportees information

[12:47:00.0000] <tantek>
it seems to be against reason to suggest a set of complex solutions/changes (rdfa) to a page that won't accept simple solutions/changes (conformance)

[12:48:00.0000] <Hixie>
yeah i don't understand why they don't just hack hcard to support what they want in the myspace case

[12:48:01.0000] <tantek>
also, regarding microformats for relating pieces of data across a document, this has been solved for quite some time with the "include-pattern". http://microformats.org/wiki/include-pattern

[12:48:02.0000] <tantek>
hResume uses it for example, to avoid duplicating a person's name in every job experience entry

[12:49:00.0000] <tantek>
Hixie, what often happens in many more complex markup cases, is the authors of those complex markup cases *assume* they can't use a simple solution, and immediately leap to a more complex solution because they think they are sufficiently knowledgeable about simple solutions to come to that conclusion.

[12:49:01.0000] <tantek>
They're almost always wrong.

[12:50:00.0000] <Hixie>
frankly microformats' lack of a clear parsing specs nor any apparent movement towards creating any has hurt its credibility significantly amongst a lot of the people who contributed feedback here

[12:50:01.0000] <tantek>
What such authors should do is first ask: "how do I markup example XYZ with microformats"

[12:50:02.0000] <tantek>
rather than assuming it's not possible

[12:50:03.0000] <a-ja>
they working on fixing date/time abbr-abuse? 

[12:51:00.0000] <Hixie>
tantek: no argument from me there

[12:51:01.0000] <Hixie>
tantek: but for some reason, they are making that leap anyway

[12:51:02.0000] <tantek>
a-ja - "abuse" is not really a proper characterization of the problem. in particular, the real problems have been with accessibility and localization, and yes, working on that problem is what Ben Ward and I have been focusing on for perhaps the past 6-9 months

[12:52:00.0000] <tantek>
we have an alpha draft (feature complete) of a solution and are tweaking it: http://microformats.org/wiki/value-excerption-pattern

[12:52:01.0000] <tantek>
Hixie, such leaps typically indicate some degree of arrogance

[12:52:02.0000] <tantek>
a-ja - sorry this is the proper URL: http://microformats.org/wiki/value-class-pattern

[12:52:03.0000] <Hixie>
tantek: clearly, but what do we do about it?

[12:53:00.0000] <tantek>
a-ja, if you want to explore / discuss the value-class-pattern further, feel free to bring it up in #microformats

[12:53:01.0000] <a-ja>
will do.....reading

[12:54:00.0000] <tantek>
Hixie, I typically do http://microformats.org/wiki/mailing-lists#Point_out_logical_flaws

[12:54:01.0000] <krijnh>
Isn't #microformats busy enough already? ;)

[12:55:00.0000] <Hixie>
tantek: that doesn't scale; people are making this mistake without us knowing about it.

[12:55:01.0000] <tantek>
Hixie, there are two cases

[12:56:00.0000] <tantek>
either they bring the examples to your attention (as you have stated), at which point, you point out the logical flaw(s) in their reasoning/assumptions, preferably with URLs to the logical flaws explaining them further

[12:57:00.0000] <tantek>
OR they don't bring the examples to you, and you don't know about, and they go off and invent their own random XML/RDF/JSON/tab-delimited/comma-separated format. in which case, let them perform such a science experiment, from which everyone can learn.

[12:57:01.0000] <Hixie>
tantek: it stops being a science experiment when companies like Yahoo! actively evangelise the use of RDFa in text/html to solve actual problems they perceive authors want solving

[12:58:00.0000] <tantek>
a-ja, thanks and appreciated. Ben and I are iterating rapidly on that document and so now is a very good time for feedback and improvements.

[12:59:00.0000] <tantek>
Hixie, most authors glaze over at the very first mention of namespaces and the complexities that brings with it. Nevermind the historically bad (more than deserved) marketing that RDF has gotten. So this is not really a problem.  Big companies evangelize all sorts of science experiments that go nowhere. I can't count the number of XML formats Microsoft has evangelized that have been abandoned. And for Yahoo - how well is MediaRSS

[13:00:00.0000] <Hixie>
there's plenty of random ccREL crap on the web now

[13:00:01.0000] <tantek>
and plus, as soon as a big company evanglizes something, then it brings it back to the first case, you know about it, can ask for real world examples, and can point out the logical flaw(s) in their reasoning/assumptions

[13:01:00.0000] <a-ja>
tantek: maybe date/time examples using html5 time element?                            

[13:01:01.0000] <tantek>
Hixie, before ccREL, there was the creative commons RDF vocabulary that they embedded in HTML comments - that failed as well.

[13:01:02.0000] <Hixie>
when i did that with the yahoo! case, they said their customers (including e.g. myspace) found microformats didn't solve their problems and that's why they supported rdfa as well.

[13:02:00.0000] <Hixie>
tantek: depends how you define "failed"

[13:02:01.0000] <Hixie>
tantek: it's all over the web

[13:02:02.0000] <Hixie>
tantek: it's wasted countless man hours

[13:02:03.0000] <tantek>
Hixie, authors are no longer actively publishing it, and no one is building tools to depend on it.

[13:02:04.0000] <Hixie>
no, they're publishing the new ccREL crazyness instead

[13:03:00.0000] <tantek>
Hixie, it's not really wasted, in the same way as testing flawed scientific theories is not wasted. Either way science makes progress, by documenting what doesn't work as well as what does.

[13:03:01.0000] <Philip`>
rel="license" is the 39th most common rel value, in some set of data

[13:03:02.0000] <Philip`>
on about 0.4% of pages

[13:04:00.0000] <Philip`>
rel="cc:attributionURL" is 184th, on about 0.02%

[13:04:01.0000] <tantek>
Philip`, right. rel="license" was introduced as a microformat back in 2004, and took off rapidly, in comparison to the "official" solution of creative commons RDF in HTML comments.

[13:04:02.0000] <tantek>
the simpler unofficial solution far outdid the complex official solution

[13:04:03.0000] <tantek>
this is not the first time this has happened, and won't be the last

[13:05:00.0000] <Philip`>
(xmlns:cc is on about 0.03% of the pages)

[13:05:01.0000] <tantek>
a-ja, current microformats documentation of HTML5 is here: http://microformats.org/wiki/html5 and yes, more HTML5 examples are a good idea / would be welcome.

[13:07:00.0000] <Hixie>
tantek: does rel-license have a solution to the problem of setting the license on subparts of the page btw? e.g. on images without affecting the license of the whole page

[13:07:01.0000] <Philip`>
http://philip.html5.org/data/cc-errors.txt has some data on the CC-RDF-in-comments thing

[13:11:00.0000] <a-ja>
<img src='' rel="license"/> ???

[13:11:01.0000] <a-ja>
nah

[13:12:00.0000] <tantek>
Hixie, rel-license itself does not, but Mike Linksvayer and I worked on developing a licensing microformat which would work for inclusion on parts of a page, e.g. for images. http://microformats.org/wiki/licensing

[13:12:01.0000] <tantek>
(Mike was former CTO of CC I believe)

[13:13:00.0000] <Hixie>
interesting

[13:14:00.0000] <Philip`>
/me remembers how Encarta would automatically add a copyright notice when you copied-and-pasted more than a few words into a text editor

[13:14:01.0000] <tantek>
Hixie, currently however, I am going to state that with respect to Creative Commons, this is more of a political problem rather than a technical problem.

[13:14:02.0000] <Hixie>
personally i think it's wrong to have copyright information be machine-readable at all

[13:14:03.0000] <tantek>
Ben Adida, is both CC's rep to the W3C and chairs the task force on RDFa : http://creativecommons.org/about/people/#13

[13:15:00.0000] <Philip`>
When I'm scanning a load of .ttf files to find ones with licenses that allow modification, I really wish they contained machine-readable copyright information

[13:15:01.0000] <Hixie>
and i strongly object to creative commons' repeatedly causing license proliferation

[13:15:02.0000] <Hixie>
we have too many damn licenses as it is

[13:15:03.0000] <Philip`>
(Instead I've got a dozen regexps like qr/This (file|software|truetype software) is public domain\./ to find them)

[13:16:00.0000] <tantek>
Hixie, I tend to agree that license proliferation is a problem.

[13:17:00.0000] <Philip`>
/me 's regexp list includes qr/This font is inspired by Polish-Lithuanian Constitution of 3 Mai 1791 and is released to the Public Domain\./

[13:18:00.0000] <tantek>
in conversations in the past, Ben Adida and I have chosen to agree to disagree with respect to namespaces. I think it is a fair characterization to say that he believes they are both essential and not a difficult hurdle for authors, and I believe they are undesirable, and sufficiently ugly/complex for vast majority of authors to reject any form of authoring that requires them.

[13:19:00.0000] <Hixie>
well i think it's clear where i stand on that particular issue

[13:19:01.0000] <tantek>
Thus I don't hold much hope for changing ccREL. 

[13:20:00.0000] <tantek>
However I have hope that a licensing microformat will handling 80%+ of the use cases that ccREL claims to handle, and will be adopted by authors, blogging tools etc. and hopefully eventually search engines (as they adopted rel="license" when that was shown to be a success).

[13:20:01.0000] <tantek>
s/handling/handle

[13:28:00.0000] <tantek>
Hixie, note btw that ccREL stands for "Creative Commons Rights Expression Language" which, upon a brief reading of the vocabulary description ( http://creativecommons.org/ns - which itself does not validate - always makes me suspicious of anyone proposing new standards if they can't conform to current ones) seems to actually *encourage* more variants of licenses/rights/requirements.

[13:28:01.0000] <Hixie>
yes

[13:28:02.0000] <Hixie>
hence my comments above

[13:30:00.0000] <tantek>
yes, I think that attempting to capture/express that fine a detail is a big mistake

[13:30:01.0000] <Hixie>
yes, it's a terrible mistake

[13:30:02.0000] <Hixie>
and it doesn't work, either

[13:30:03.0000] <Hixie>
you can't properly encode legal practices in a finite vocabulary

[13:30:04.0000] <Hixie>
that's ridiculous

[13:31:00.0000] <tantek>
in fact, I'll go further than that, any attempt to express specific rights (other than just referring to a license URL) or any other kind of legality in supposedly machine readable/understandable form is a big mistake

[13:31:01.0000] <tantek>
Hixie, I agree, quite ridiculous. And undesirable.

[13:42:00.0000] <tantek>
/me is continuing the licensing microformat discussion in #microformats, for anyone who would like to view and or participate.

[15:52:00.0000] <Hixie>
Philip`: yt?

[15:52:01.0000] <Hixie>
Philip`: http://philip.html5.org/tests/canvas/suite/tests/2d.path.stroke.prune.arc.html

[15:52:02.0000] <Hixie>
Philip`: what would it take to change the spec to match what browsers do there?

[16:18:00.0000] <olliej>
Hixie++

[16:19:00.0000] <cgriego>
I saw the <header> is now <hgroup> @whatwg tweet and I was all "nooo!" then I saw the new <header> tweet and I was all "whaa?" then I read the IRC log and I was like "oh, good stuff."

[16:23:00.0000] <hsivonen>
where should I look for offline cache manifest test cases?

[16:25:00.0000] <gavin_>
http://mxr.mozilla.org/mozilla-central/source/dom/tests/mochitest/ajax/offline/ ?

[16:27:00.0000] <hsivonen>
gavin_: thanks (I was looking for something that I can just browse to, though)

[16:31:00.0000] <Hixie>
cgriego: hehe

[16:31:01.0000] <Hixie>
olliej: what did i do to get incremented? :-)

[16:32:00.0000] <olliej>
Hixie: epic html5 work :D

[16:32:01.0000] <Hixie>
oh just the volume? :-)

[16:36:00.0000] <olliej>
Hixie: well and drawImage(<video>)

[16:36:01.0000] <Hixie>
ah

[16:39:00.0000] <Philip`>
Wasn't that months ago?

[16:39:01.0000] <Hixie>
yeah but i just sent mail about it again

[16:40:00.0000] <Philip`>
Oh

[16:40:01.0000] <Hixie>
so i'm willing to take double credit

[16:40:02.0000] <Hixie>
:-)

[16:41:00.0000] <hsivonen>
yay for hgroup!

[16:41:01.0000] <hsivonen>
Hixie: thanks

[16:42:00.0000] <hsivonen>
also, with the new <header>, HTML5 covers all of ARIA landmarks natively

[16:42:01.0000] <Hixie>
all of them?

[16:42:02.0000] <Hixie>
wow

[16:42:03.0000] <Hixie>
that's unexpectedly fortuitous

[16:44:00.0000] <hsivonen>
Hixie: yes, if you infer <form role=search> from <form> ... <input type=search>

[16:45:00.0000] <Hixie>
ew, they put the role no the <form>? that seems problematic

[16:45:01.0000] <Hixie>
surely you want it on the input control