Options:
- # Session Start: Thu Sep 27 00:00:00 2007
- # Session Ident: #html-wg
- # [00:00] * Joins: mjs (mjs@17.255.111.173)
- # [00:02] <beowulf> was it mentioned on the html-wg list? i must have missed it
- # [00:03] <anne> just on IRC
- # [00:03] <anne> i think it was "spotted" because @html4all.org was cc'ed in some e-mail or something
- # [00:04] <Philip> http://krijnhoetmer.nl/irc-logs/whatwg/20070824#l-227 seems to be the first mention of the wiki, presumably via Referer
- # [00:06] <beowulf> so why is html4all trying to remove the editor of... no don't answer that
- # [00:11] <Philip> I expect they're not a unified monolithic group with a single set of principles that they all agree on, so it's probably better to wonder why certain people want something rather than why the group does
- # [00:17] * Quits: Sander (svl@86.87.68.167) (Quit: And back he spurred like a madman, shrieking a curse to the sky.)
- # [00:18] * Quits: heycam (cam@203.214.33.166) (Ping timeout)
- # [00:23] <anne> heh, http://www.royal-tunbridge-wells.org/ longdesc= points to a non-existing URL...
- # [00:23] <anne> maybe I should e-mail an open letter
- # [00:25] <beowulf> i looked at the html of that page a while ago, i think when we were being called lazy and stupid, it didn't inspire me ...
- # [00:25] <anne> if you add .org it points to a 404
- # [00:25] <Hixie> anne: as a community leader, i have to ask you to not be so derisive
- # [00:26] <beowulf> by we i don't mean you, fwiw
- # [00:26] <anne> Hixie, this is a serious accessibility breach, you have to understand
- # [00:27] * anne will stop now
- # [00:27] <Hixie> (i was kidding)
- # [00:27] <anne> (too)
- # [00:28] <Hixie> i was interested to hear that as editor i'm supposed to be doing leadership work in the html working group, though
- # [00:28] <Dashiva> I want to comment, but the joke would be trouble, and explaining why I don't make the joke would be trouble as well, so I might as well not write this line
- # [00:28] <Hixie> i wonder how i'm allowed to use this new authority
- # [00:28] <Dashiva> You are our secret google overlord, after all
- # [00:29] <Philip> anne: "I don't think this should be solved in Selectors [...] as it would require specific knowledge about which attributes contain qnames et cetera" - but there's already selectors for space-separated-list values and language values, which don't need UAs to have any specific knowledge about attributes
- # [00:31] <anne> Philip, given foo="x:x" how can I select x in namespace "bar" regardless of what its prefix is?
- # [00:31] <anne> afaict there's no such syntax
- # [00:32] <anne> maybe I should have mentioned that more explicitly...
- # [00:32] <Philip> You solve the problem in Selectors by making new syntax
- # [00:32] <anne> that would allow each attribute to take qnames
- # [00:32] <anne> ugh
- # [00:33] <Hixie> is | allowed on the RHS yet? it's not, is it?
- # [00:33] <Philip> You can already select elements that have a specific space-separated word in their alt text
- # [00:33] <Hixie> we could have [foo|bar~=xx|yy]
- # [00:33] <Hixie> and [foo|bar=xx|yy]
- # [00:33] <anne> aah, blasphemy!
- # [00:33] <Hixie> (like authors would understand _that_)
- # [00:34] <Hixie> "please sir, hit me harder, sir"
- # [00:38] * Joins: aaronlev (chatzilla@209.6.168.245)
- # [00:40] * Quits: mjs (mjs@17.255.111.173) (Quit: mjs)
- # [00:46] <beowulf> let's not be down on us authors
- # [00:46] <beowulf> we get a hard time as it is
- # [00:50] <Dashiva> Real authors write their HTML in machine language, and use browser detection to determine which architecture to serve for
- # [00:59] <beowulf> you don't realise how sensitive we are, all these hurtful comments about being lazy and stupid ...
- # [01:03] * Joins: mjs (mjs@17.255.111.173)
- # [01:05] <Philip> The Perl community is nice since laziness is a virtue there
- # [01:07] <Philip> ("It makes you write labor-saving programs that other people will find useful, and document what you wrote so you don't have to answer so many questions about it.")
- # [01:07] <Philip> Sadly stupidity is not considered a good thing
- # [01:08] <beowulf> *sniff*
- # [01:09] * Joins: sbuluf (ldkkmg@200.49.140.144)
- # [01:10] * Joins: heycam (cam@130.194.72.84)
- # [01:12] * Quits: jmb (jmb@152.78.71.152) (Ping timeout)
- # [01:13] <Philip> http://html4all.org/pipermail/list_html4all.org/2007-September/000429.html - hmm, they deleted the message?
- # [01:13] * Joins: jmb (jmb@152.78.71.152)
- # [01:13] <anne> I suppose that's because DanC didn't give them permission to forward it and they did so anyway
- # [01:14] <anne> not the first time btw: http://html4all.org/pipermail/list_html4all.org/2007-September/000381.html
- # [01:14] <anne> (although I'm not sure Rich knows about it)
- # [01:18] * Quits: tH (Rob@87.102.114.133) (Quit: ChatZilla 0.9.78.1-rdmsoft [XULRunner 1.8.0.9/2006120508])
- # [02:06] * Quits: hasather (hasather@90.227.221.48) (Quit: Lost terminal)
- # [02:10] * Quits: Yudai (Yudai@211.128.41.224) (Ping timeout)
- # [02:10] * Joins: Yudai (Yudai@211.128.41.224)
- # [02:14] * Joins: karl (karlcow@128.30.52.30)
- # [02:29] * Quits: mjs (mjs@17.255.111.173) (Quit: mjs)
- # [02:52] * Quits: kingryan (rking3@208.66.64.47) (Quit: kingryan)
- # [03:13] * Quits: schepers (schepers@128.30.52.30) (Ping timeout)
- # [03:13] * Joins: DougJ (djones4@74.76.28.112)
- # [03:17] * Quits: aaronlev (chatzilla@209.6.168.245) (Ping timeout)
- # [03:41] * Joins: olivier (ot@128.30.52.30)
- # [04:10] * Parts: DougJ (djones4@74.76.28.112)
- # [04:27] * Quits: heycam (cam@130.194.72.84) (Quit: bye)
- # [04:28] * Joins: mjs (mjs@64.81.48.145)
- # [04:34] * Joins: aaronlev (chatzilla@209.6.168.245)
- # [05:22] * Quits: mjs (mjs@64.81.48.145) (Connection reset by peer)
- # [05:26] * Joins: hyatt (hyatt@209.173.92.142)
- # [05:29] * Joins: heycam (cam@203.214.33.166)
- # [05:30] * Quits: aaronlev (chatzilla@209.6.168.245) (Ping timeout)
- # [06:17] * Joins: mjs (mjs@64.81.48.145)
- # [07:01] * Joins: schepers (schepers@128.30.52.30)
- # [07:10] * Quits: hyatt (hyatt@209.173.92.142) (Quit: hyatt)
- # [07:44] * Quits: karl (karlcow@128.30.52.30) (Quit: This computer has gone to sleep)
- # [08:03] * Joins: karl (karlcow@128.30.52.30)
- # [08:27] * Joins: hyatt (hyatt@209.173.92.142)
- # [08:45] * Quits: schepers (schepers@128.30.52.30) (Quit: Trillian (http://www.ceruleanstudios.com)
- # [09:09] * Quits: Hixie (ianh@129.241.93.37) (Ping timeout)
- # [09:28] * Joins: schepers (schepers@128.30.52.30)
- # [09:29] * Joins: Hixie (ianh@129.241.93.37)
- # [09:36] * Quits: karl (karlcow@128.30.52.30) (Quit: Where dwelt Ymir, or wherein did he find sustenance?)
- # [09:36] * Quits: olivier (ot@128.30.52.30) (Quit: Leaving)
- # [09:37] * Joins: Steve (chatzilla@82.44.69.8)
- # [09:38] <Steve> html4all man here, morning all
- # [09:39] <Hixie> hey steve
- # [09:39] * Hixie is knee-deep in offline web app edge cases trying to define the cache updating algorithm based on the most recent proposal
- # [09:40] <Steve> <Philip> said I expect they're not a unified monolithic group with a single set of principles that they all agree on, so it's probably better to wonder why certain people want something rather than why the group does. philip is spot on here most of the people i hardly know and have widely differing views from
- # [09:41] * anne waves
- # [09:42] <Steve> but am happy to answer any queries where i can, but would rather observe and talk about relevant issues that i can give some constructive input to
- # [09:43] <Hixie> it's been pretty quiet here tonight
- # [09:43] <Hixie> i think most people are off doing other things
- # [09:43] <Hixie> i'm trying to work out where i should put the offline web apps api
- # [09:43] <Hixie> on Window, on Window.navigator, or elsewhere
- # [09:44] <Hixie> the onLine attribute is on window.navigator
- # [09:44] <Hixie> which is why i'm considering putting the rest there
- # [09:44] <Hixie> but it doesn't really feel right
- # [09:44] <Steve> hixie: i can be of no help with that one :-(
- # [09:44] <Hixie> Steve: no worries
- # [09:45] <Hixie> yesterday i was considering introducing a new object window.cache for this stuff
- # [09:45] <Hixie> window.cache.add(), window.cache.remove(), window.cache.update(), window.cache.status...
- # [09:46] <Hixie> i guess that would be nice and conceptually easy
- # [09:46] <Hixie> maybe the events could even fire on that object
- # [09:46] <Hixie> hmmm
- # [09:48] <anne> navigator isn't that bad, it's after all a UA object
- # [09:48] <Hixie> yeah but it's wrong... you're not futzing with the UA
- # [09:49] <anne> you sort of are, if you switch to a different UA later it won't have the same offline cache
- # [09:49] <anne> s/you/the user/
- # [09:49] <Hixie> by that argument, everything should be in window.navigator
- # [09:49] <anne> that's sort of convincing to not do it, yes
- # [09:50] <anne> :)
- # [09:50] <Hixie> i wish JS/DOM had strongly typed enums, sigh
- # [09:50] * Hixie defines a new set of constants
- # [09:53] <Steve> off topic but something you may be able to help with Ian: i want to do a small scale study of images and their alts, but in the test design i want to get a truly random sample of say 100 urls, any ideas on how to acheive this? or would it be btter to focus on pages from the most popular sites?
- # [09:55] <anne> I'm pretty sure Philip would be able to help you with some URIs which are sort of random
- # [09:56] <Hixie> i can get you a random sample of 100 urls if you like, but sadly a representative random sample of 100 pages will probably get you about 45 porn pages, 45 spam pages, and 10 pages of dead forums, or something like that
- # [09:56] <Hixie> you're probably best doing a crawl seeded at some more reputable site (though i recommend not picking dmoz, just because that's what everyone usually picks, and variety in these studies would be good)
- # [09:57] <Hixie> and just picking 100 of the first 100,000 pages you crawl, at random
- # [09:57] <Hixie> by following every link
- # [09:57] <Hixie> (and avoiding dupes/loops)
- # [09:57] <Hixie> Philip has been doing a bunch of work on small scale studies like this, he might be able to give better advice
- # [09:59] * anne believes he's using dmoz though
- # [09:59] <Steve> cool thanks; i just want to ensure that my sample is not tainted by me unconsciuosly, btw i have looked for other studies of this nature but have not found any, anybody else seen anything?
- # [10:00] <anne> lots of informal data gathering
- # [10:00] <Steve> what i want to do is classify images (decorative, functional, spacer etc) and then compare their alts (or lack of)
- # [10:02] <anne> for instance, http://wiki.whatwg.org/wiki/Longdesc_usage documents some usage of longdesc (different from the stats Hixie gave, based on a much smaller sample)
- # [10:02] <Steve> anne: have seen that thanks
- # [10:03] <Hixie> Steve: of alt attributes in particular?
- # [10:04] <anne> there was something similar for <input usemap> but that was just done on #whatwg and not documented anywhere else
- # [10:04] <Hixie> i don't know of any research where people have studied alt attributes by hand
- # [10:04] <Steve> hixie: yes, perhaps should include title attributes as well
- # [10:05] <Hixie> the studies i do myself are all on such large datasets that manual examination only makes sense once you have pruned the data somehow (as we did with longdesc, e.g.)
- # [10:05] <Hixie> i think it would be a great thing to look at though
- # [10:06] <Steve> i want to get a picture of relative usage patterns for different types of images and how/if the the alt is provided and how useful it is, i know the results are going to be depressing (for me) but..
- # [10:07] <Hixie> the results are sadly always depressing when you look at real usage
- # [10:07] <Hixie> it's so sad
- # [10:07] <Hixie> 93% of pages i looked at in one study (of several billion pages) had at least one major syntax error
- # [10:07] <Hixie> that's not even looking at things like duplicate IDs and other errors
- # [10:09] * Joins: krijnh2 (krijnhoetm@213.84.148.98)
- # [10:09] * Parts: krijnh2 (krijnhoetm@213.84.148.98)
- # [10:09] <mjs> Hixie: window.application?
- # [10:09] <mjs> or is that better reserved for something else
- # [10:09] <mjs> (re your earlier point of wonderment)
- # [10:11] <Hixie> probably better kept for something bigger
- # [10:11] <Hixie> though who knows
- # [10:11] <Hixie> i'm wondering where the events should go, too
- # [10:12] <Hixie> we already are planning on having window.ononline, should we have window.onupdating, or window.cache.onupdating?
- # [10:12] <Hixie> or should we only use addEventListener
- # [10:12] <Hixie> aah, so many decisions
- # [10:13] <anne> move them to .cache I'd say
- # [10:13] <anne> window is already _so_ overloaded
- # [10:14] <anne> (maybe so badly it can't hurt, but still)
- # [10:14] <anne> people can type cache.onupdating which isn't that bad either
- # [10:25] * Quits: hyatt (hyatt@209.173.92.142) (Quit: hyatt)
- # [10:30] * Joins: hyatt (hyatt@209.173.92.142)
- # [10:30] * Quits: hyatt (hyatt@209.173.92.142) (Quit: hyatt)
- # [10:38] <anne> Hixie, what happens btw if someone has var cache = foo in his code, that simply overwrites it?
- # [10:38] <mjs> anne: new Window properties in general need to be replaceable to avoid compat issues
- # [10:39] <anne> can you find out from the IDL whether a property is replaceable in ES?
- # [10:40] <mjs> I don't know if you can in the flavor of IDL that HTML5 uses, but I believe Bindings For DOM will cover this
- # [10:40] <Hixie> yeah
- # [10:40] <anne> k, cool
- # [10:40] <Hixie> it's on heycam's list, at least
- # [10:40] <Hixie> i put it there :-)
- # [10:41] <Hixie> offtopic, but:
- # [10:41] * Hixie is amused by http://searchengineland.com/070927-000001.php
- # [10:41] <Hixie> search for the paragraph starting "The improved ranking algorithms"
- # [10:41] <Hixie> and do the sample query on google and the new msn live search
- # [10:42] <Hixie> neither really answer the question asked, but it's hard to say that msn live search's results are more relevant than google's...
- # [10:44] <mjs> indeed - poorly chosen example
- # [10:45] <anne> for that example I get his site back on google btw :)
- # [10:45] <mjs> ditto
- # [10:46] <anne> but neither gives "useful" results
- # [10:46] <mjs> but on live search, you get porn, a radio station, and some band links
- # [10:46] <Hixie> yeah i love that google is so fresh that we can serve back that site even though it was only posted a few hours ago
- # [10:46] * Joins: ROBOd (robod@89.123.30.188)
- # [10:47] <Hixie> but yeah, the live search results do seem pretty much not at all relevant, whereas at least the google ones are about arizona, or the weather, or about the query itself
- # [10:52] * Joins: zcorpan_ (zcorpan@88.131.66.80)
- # [10:53] <mjs> looks like the kind of quality effort you would expect from the company that brought you Zune and Vista
- # [10:56] <Hixie> well in all fairness their update does seem like a major step forward
- # [11:01] * Joins: Lachy_ (chatzilla@124.171.3.159)
- # [11:04] * Quits: Lachy (chatzilla@124.171.0.33) (Ping timeout)
- # [11:04] * Lachy_ is now known as Lachy
- # [11:06] <zcorpan_> hmm, i kinda like matthew's proposal. it doesn't help with roles in other vocabs, but that might not really matter
- # [11:07] <zcorpan_> it also means we don't have to step on the toes of the xhtml role module
- # [11:07] <anne> what about all the existing role= usage?
- # [11:07] <zcorpan_> that would require the wairole: prefix
- # [11:07] <anne> well, "all"
- # [11:08] <anne> oh, you want both now? ouch
- # [11:08] * Joins: Sander (svl@86.87.68.167)
- # [11:08] <zcorpan_> well, i don't think firefox wants to drop support for role="wairole:foo"
- # [11:08] <zcorpan_> that's not to say we'll implement it in opera
- # [11:09] <zcorpan_> i dunno. having another attribute might make it more messy
- # [11:10] <anne> what would the solution for SVG be?
- # [11:10] <anne> introduce aria-role there too? :)
- # [11:10] <zcorpan_> no, use the "namespaced" role="wairole:foo"
- # [11:10] <zcorpan_> er
- # [11:11] <zcorpan_> html:role=..
- # [11:11] <zcorpan_> even
- # [11:11] <zcorpan_> hmm, it seems messy
- # [11:12] <anne> that's ugly
- # [11:13] <zcorpan_> yeah
- # [11:13] <anne> you'd want html:aria-role there then or something
- # [11:13] <anne> or simply aria-role
- # [11:13] <zcorpan_> yeah, html:aria-role might work
- # [11:13] <anne> hijacking role= seems like a more straightforward solution
- # [11:13] <zcorpan_> agree
- # [11:14] <anne> (if you actually want people to use it you should just introduce the attribute in no namespace on SVG)
- # [11:14] <zcorpan_> that's up to the svg wg
- # [11:14] <anne> although maybe they're used to it with the ugly xlink stuff
- # [11:14] <zcorpan_> yeah
- # [11:14] <zcorpan_> they might even like that it is namespaced :)
- # [11:14] <zcorpan_> ooh, namespaces! more!
- # [11:15] <anne> soon you'll get <a:b c:d="e:f=g:h">
- # [11:17] <anne> and authors will rejoice
- # [11:19] <zcorpan_> ha
- # [11:21] <anne> <a:b c:d="e:f='g:h'"> of course, unquoted inner attribute values are bad practice
- # [11:23] <zcorpan_> not well-formed! well-formedness is an xml feature.
- # [11:23] <anne> like qnames in content?
- # [11:23] <zcorpan_> yes :)
- # [11:42] <Dashiva> <a:b=c:d=apple>
- # [12:27] * anne summons MarkB
- # [12:47] * Quits: sbuluf (ldkkmg@200.49.140.144) (Ping timeout)
- # [13:05] * Joins: karl (karlcow@128.30.52.30)
- # [13:23] <Philip> Steve: I did http://canvex.lazyilluminati.com/misc/imgs.xhtml a while ago, extracting all the <img>s from some (32?) random pages from dmoz.org
- # [13:25] <Steve> philip: great thanks will have a look at it
- # [13:27] <Steve> philip: do you have a list of the pages?
- # [13:28] <Philip> Of just those 32, or of all of dmoz.org?
- # [13:28] <Steve> just the 32 that you used for the sample, i need to see the images in context
- # [13:29] <Philip> The images on that page link to the page they came from
- # [13:29] * Joins: DougJ (djones4@74.76.28.112)
- # [13:31] <anne> that's pretty terrible
- # [13:32] <Philip> http://tinyurl.com/2tqcyr is the pages
- # [13:33] <Philip> and seven others had no <img>s
- # [13:35] <Philip> (No images: http://members.aol.com/AdbastonWI/ http://www.laktrans.pl/ http://web.mit.edu/mitbac/www/ http://monikapyrek.pl/ http://www.kurashiki-oky.ed.jp/school/midori-e/ http://democracy.stanford.edu/ http://amalgamatedservices.com/ )
- # [13:42] <Steve> Ok thanks
- # [13:47] * Joins: olivier (ot@128.30.52.30)
- # [13:56] <Sander> yay, looks like I'm all caught up with the mailinglist again. *accidentally hit "mark all as read"* :/
- # [14:05] <krijnh> That works well :)
- # [14:14] * Quits: jmb (jmb@152.78.71.152) (Ping timeout)
- # [14:14] * Joins: aaronlev (chatzilla@209.6.168.245)
- # [14:21] <zcorpan_> karl: re http://www.w3.org/QA/2007/09/setting-default-style-sheet.html , that is a non-issue, because all browsers are forced to treat style="" as css
- # [14:22] <zcorpan_> karl: saying that authors should add content-style-language to declare text/css is just wasting their time (even though html4 says it's required)
- # [14:22] <karl> zcorpan_: read the article I say so
- # [14:22] * Quits: aaronlev (chatzilla@209.6.168.245) (Ping timeout)
- # [14:23] * Joins: tH_ (Rob@87.102.114.133)
- # [14:23] <zcorpan_> karl: right, so why do you suggest that authors should add the declaration?
- # [14:23] * tH_ is now known as tH
- # [14:23] <karl> because it is still a requirement for now. HTML 5 is far to be a spec *yet*.
- # [14:30] * Joins: jmb (jmb@152.78.71.152)
- # [14:41] * Parts: DougJ (djones4@74.76.28.112)
- # [14:48] * Quits: Thezilch (fuz007@68.52.119.203) (Ping timeout)
- # [14:55] * Joins: aaronlev (chatzilla@66.31.86.217)
- # [15:06] * Quits: karl (karlcow@128.30.52.30) (Quit: Where dwelt Ymir, or wherein did he find sustenance?)
- # [15:12] <zcorpan_> aaronlev: what are your thoughts about aria-role?
- # [15:14] * anne just replied
- # [15:24] <aaronlev> zcorpan_: basicaly, not backwards compatible
- # [15:24] <aaronlev> zcorpan_: will break dojo a11y
- # [15:24] <aaronlev> i understand why people would like that for consistency
- # [15:25] <anne> i think we should just fork role= and be done with it
- # [15:25] <aaronlev> i didn't read his whole mail, he wants to keep both role and have aria-role?
- # [15:26] <anne> implement the simple solution, have people bitch about theoretical purity until they realize that reality is different...
- # [15:26] <anne> aaronlev, yeah, afaict
- # [15:27] <aaronlev> anne: he says replace role with aria-role
- # [15:28] <anne> well yes, but he also argues: "3) People who use |role| aren't confused by the lack of a namespace for ARIA roles and have a clear understanding that the roles are ARIA-related."
- # [15:28] <aaronlev> chris wilson wanted aria-role as well
- # [15:29] <anne> advantage being?
- # [15:29] <anne> if we do aria-role, i'd like for Opera and other browsers to not implement role= at all...
- # [15:29] <anne> and also have html:aria-role
- # [15:29] <anne> and just keep it simple, with a single value, etc.
- # [15:31] <aaronlev> advantage is consistency for html authors
- # [15:32] <aaronlev> but i'm not for it right now, mainly because of dojo+ff2 i guess
- # [15:32] <anne> i suppose, though it's a matter of perspective
- # [15:32] <anne> role= indicates the widget where aria-* indicates the state
- # [15:32] <anne> there's some consistency in that too
- # [15:32] <anne> it just means forking role=...
- # [15:37] <aaronlev> anne: it won't kill me either way but it will be a pain for the dojo folks who are the main user of ARIA right now
- # [15:38] <anne> pain with distributing the new code or updating existing code?
- # [15:39] <anne> seems like updating won't be that hard as the code becomes simpler and it's only a few strings here and there...
- # [15:39] <aaronlev> well w/o backwards compat they have to choose ff2 or ff3
- # [15:39] <aaronlev> if we move from |role| to |aria-role| in ff3
- # [15:39] <aaronlev> or they have to user agent sniff before setting the attribute
- # [15:40] <zcorpan_> yeah, with fx2 compat in mind, i don't like aria-role
- # [15:41] <anne> seems like a good motivator for people to upgrade :)
- # [15:43] <aaronlev> anne: well dojo is coming out before ff3
- # [15:43] <aaronlev> i'm told that they must have a11y for ff2, i tried to argue just for targetting ff3
- # [15:43] <aaronlev> it does seem like a silly reason
- # [15:46] <Philip> They could set both role="foo" and aria-role="foo" to be compatible with both versions
- # [15:46] <anne> good point
- # [15:46] <anne> it's indeed not one or the other
- # [15:47] <anne> (as far as authoring is concerned, I hope it is for UAs!)
- # [15:50] <zcorpan_> in the proposal, role="wairole:foo" will work in both
- # [15:50] <zcorpan_> (fwiw)
- # [15:50] <aaronlev> Philip: yeah they could, true
- # [15:51] <aaronlev> zcorpan_: yes, so as for now they only have to set role
- # [15:52] * Quits: olivier (ot@128.30.52.30) (Quit: Leaving)
- # [15:53] <zcorpan_> indeed
- # [15:54] * Joins: Lachy_ (chatzilla@124.171.3.159)
- # [15:56] * Quits: Lachy (chatzilla@124.171.3.159) (Ping timeout)
- # [15:56] * Lachy_ is now known as Lachy
- # [16:39] * Joins: billmason (billmason@69.30.57.156)
- # [16:41] <hsivonen> what good would aria-role do? compared to role?
- # [16:41] <hsivonen> now at least the aria-* naming schema is the minimal change to ARIA as opposed to second-guessing the names some more
- # [16:46] <anne> it would not conflict with the xhtml2 wg or any other ideas people might about role
- # [16:46] <anne> i suppose it also makes it a little bit more clear it's for accessibility, but apart from that, dunno
- # [16:51] <hsivonen> anne: I thought namespaces were supposed to solve any conflicts between XHTML2 and anything else. :-)
- # [16:52] <zcorpan_> hsivonen: the xhtml2 *wg*; the xhtml role attribute module is their spec
- # [16:54] <anne> i'm fine with either forking role= or creating aria-role but I rather not have both
- # [16:54] <zcorpan_> agree
- # [16:54] <zcorpan_> i don't have an opinion about which one
- # [16:54] <hsivonen> yeah.
- # [16:54] <hsivonen> intuitively, I'd go with role=.
- # [16:57] * Joins: myakura (myakura@125.207.244.50)
- # [17:09] * Parts: schepers (schepers@128.30.52.30)
- # [17:14] * Quits: gsnedders (gsnedders@86.137.237.196) (Quit: gsnedders)
- # [17:20] * Joins: gsnedders (gsnedders@86.137.237.196)
- # [17:34] <beowulf> i've just had a discussion were a certain company wouldn't allow background-images in the css because you can't add alt tags, and all images should have alt tags; is this sane?
- # [17:35] <hsivonen> beowulf: what were the images like?
- # [17:35] <Philip> Maybe you could argue that background-images in CSS have an implicit empty alt="", so it's fine to use them iff <img src=... alt=""> would be acceptable for that image
- # [17:35] <beowulf> hsivonen: eye candy
- # [17:35] * Parts: zcorpan_ (zcorpan@88.131.66.80)
- # [17:36] * Joins: zcorpan_ (zcorpan@88.131.66.80)
- # [17:36] <beowulf> lifestyle images
- # [17:36] <beowulf> if you like
- # [17:36] <beowulf> but this was a black and white policy, images must have alt tags, therefore no images in css
- # [17:37] <hsivonen> beowulf: does lifestyle mean alt='group of smiling multiethnic people equally gender-divided doing sports' ?
- # [17:37] * Joins: SBratt (steve@128.30.52.30)
- # [17:37] <beowulf> hsivonen: yes
- # [17:38] * Parts: SBratt (steve@128.30.52.30) (Ciao ...)
- # [17:38] * Parts: zcorpan_ (zcorpan@88.131.66.80)
- # [17:38] * Joins: zcorpan_ (zcorpan@88.131.66.80)
- # [17:48] * Joins: hyatt (hyatt@209.173.92.142)
- # [17:55] * Joins: schepers (schepers@128.30.52.30)
- # [17:59] * Quits: myakura (myakura@125.207.244.50) (Quit: Leaving...)
- # [18:16] * Quits: Lachy (chatzilla@124.171.3.159) (Quit: ChatZilla 0.9.78.1 [Firefox 2.0.0.6/2007072518])
- # [18:16] * Joins: Lachy (chatzilla@124.171.3.159)
- # [18:18] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [18:26] <aaronlev> hsivonen: lol
- # [18:27] <aaronlev> beowulf: tell them that background images are fine for images that are not interactive and are only for looks, adding no meaning
- # [18:27] <aaronlev> becuase the purpose is to make things better for blind users, and sometimes adding more alt text is just redundant clutter
- # [18:33] <Philip> That's far harder to explain than "images must have an alt attribute" :-(
- # [18:34] <billmason> "image tags must have an alt attribute". The explanation works fine paired with that.
- # [18:35] <Philip> What about CSS background image tags?
- # [18:35] <billmason> That's not a "tag".
- # [18:35] <Philip> That makes it harder to explain :-)
- # [18:36] <zcorpan_> "foreground images", "content images", "non-decorative images", ...?
- # [18:36] <billmason> Fine, 'HTML image tags must have an alt attribute.' You know, if the target audience is stuck on that concept, they shouldn't be trying to mandate anything from the code author. :)
- # [18:37] <zcorpan_> people who mandate things often don't have a clue :)
- # [18:37] <billmason> funny how that happens
- # [18:39] <Philip> It seems better if they can mandate some guidelines / checking tools which were written by people who do have a clue
- # [18:39] * Quits: schepers (schepers@128.30.52.30) (Client exited)
- # [18:39] <Philip> so there's less chance for the mandaters to mess it all up
- # [18:39] <Philip> s/it all up/up all the details/
- # [18:42] <zcorpan_> Philip: where's the ` ? :)
- # [18:44] <Philip> zcorpan_: This network is small enough that nobody else has taken the non-suffixed name :-)
- # [18:45] <zcorpan_> Philip: ok :)
- # [18:48] * Joins: schepers (schepers@128.30.52.30)
- # [18:58] * Joins: timbl_ (timbl@128.30.7.41)
- # [19:17] * Joins: Steve_ (chatzilla@82.44.69.8)
- # [19:18] * Quits: Steve (chatzilla@82.44.69.8) (Ping timeout)
- # [19:18] * Steve_ is now known as Steve
- # [19:39] * Quits: hyatt (hyatt@209.173.92.142) (Quit: hyatt)
- # [20:06] * Joins: hyatt (hyatt@209.173.92.142)
- # [20:09] * Quits: hyatt (hyatt@209.173.92.142) (Quit: hyatt)
- # [20:32] * Quits: Steve (chatzilla@82.44.69.8) (Ping timeout)
- # [20:44] <zcorpan_> aaronlev: what do you think about al's suggestion to use the first "aria" token, if any, instead of just the first token?
- # [20:45] <aaronlev> i didn't read it
- # [20:45] <aaronlev> sometimes i need to concentrate pretty hard to understand his emails
- # [20:45] <zcorpan_> yeah, it was a bit confusing
- # [20:47] <aaronlev> i give al a hard time
- # [20:47] <aaronlev> he uses some odd terminology
- # [20:47] <aaronlev> on our phone calls i need other people to translate his text into plain speak
- # [20:48] <aaronlev> i think he used the phrase "self-nucleating snowball" recently
- # [20:51] * Joins: Steve_ (chatzilla@82.44.69.8)
- # [20:51] * Steve_ is now known as Steve
- # [20:51] * Joins: polin8 (polin8@75.71.72.175)
- # [20:54] * Joins: kingryan (rking3@208.66.64.47)
- # [20:54] * Joins: hyatt (hyatt@209.173.92.142)
- # [20:55] * Quits: Steve (chatzilla@82.44.69.8) (Ping timeout)
- # [21:08] * Joins: dbaron (dbaron@63.245.220.241)
- # [21:20] * Joins: hober (ted@68.107.112.172)
- # [21:27] * Quits: zcorpan_ (zcorpan@88.131.66.80) (Ping timeout)
- # [21:51] * Joins: Roger (roger@213.64.74.230)
- # [21:54] <anne> seems that Al doesn't understand how the DOM works...
- # [21:59] * Quits: laplink (link@193.157.66.199) (Quit: This computer has gone to sleep)
- # [22:14] <anne> hmm, we need <noflash> :)
- # [22:14] <anne> http://www.joshuabudich.com/SWCollection/
- # [22:14] <anne> people start using <noscript> for it, can't be good
- # [22:15] <kingryan> why not just object fallback?
- # [22:15] * anne was joking
- # [22:15] <anne> it's sad though that flash support is assumed
- # [22:15] <kingryan> <noobject> ?
- # [22:16] <anne> <i> <object>
- # [22:26] * Quits: hyatt (hyatt@209.173.92.142) (Quit: hyatt)
- # [22:35] * Quits: ROBOd (robod@89.123.30.188) (Quit: http://www.robodesign.ro )
- # [22:35] * Quits: Roger (roger@213.64.74.230) (Quit: Roger)
- # [22:55] * Quits: timbl_ (timbl@128.30.7.41) (Quit: timbl_)
- # [23:07] * Joins: hyatt (hyatt@209.173.92.142)
- # [23:08] * Quits: Sander (svl@86.87.68.167) (Ping timeout)
- # [23:11] * Joins: Sander (svl@86.87.68.167)
- # [23:17] * Joins: Steve_ (chatzilla@82.44.69.8)
- # [23:17] * Steve_ is now known as Steve
- # [23:21] * Quits: jgraham (jgraham@81.86.223.139) (Quit: Ex-Chat)
- # [23:25] * Quits: Steve (chatzilla@82.44.69.8) (Ping timeout)
- # [23:29] * Joins: laplink (link@193.157.66.199)
- # [23:40] * Joins: jgraham (jgraham@81.86.212.61)
- # [23:43] * Quits: hyatt (hyatt@209.173.92.142) (Quit: hyatt)
- # [23:56] * Quits: mjs (mjs@64.81.48.145) (Quit: mjs)
- # Session Close: Fri Sep 28 00:00:00 2007
The end :)