| 00:06 | <Hixie> | i wonder how i should handle the relative order of events fired on a <video> element when it has a pending load() about to start |
| 00:07 | <Hixie> | i mean, if it takes 250ms to get around to calling load(), and timeupdate has fired since that load() was queued |
| 00:07 | <Hixie> | should i just discard all those events? |
| 00:08 | <Hixie> | we can't very well delay the load() until we've dealt with all the timeupdate events, we might just end up generating more |
| 02:01 | <Hixie> | wow, this alt discussion just gets more and more asinine |
| 02:02 | <Hixie> | luckily since none of the recent posts seem to actually be about the spec, i probably don't have to actually respond to any of the recent ones |
| 02:02 | <Hixie> | (the ones steven sent near the start of the thread are actual useful feedback that needs replies, obviously) |
| 05:10 | <Hixie> | roc: the idea is that the bounding box height is used to set the height, at which point it can't be scrolled, since all the content is visible |
| 05:12 | <roc> | even if, say, the iframe element has a CSS max-height? |
| 05:13 | <roc> | I have to admit I'm a bit confused about how to compute this bounding box |
| 05:17 | <roc> | it seems we're supposed to compute the bounding box of everything that's rendered, ignoring viewport clipping ... but what about completely transparent elements? elements with width 0 but nonzero height? |
| 05:17 | <roc> | Should we just use the area that we would permit scrolling to, assuming the viewport had scrollbars? |
| 05:18 | <roc> | that might work, although since we don't allow scrolling up to see content above the top of the viewport, it doesn't quite jive with your answer about having to include content that overflows vertically above in the bounding box |
| 05:19 | <Hixie> | i guess it is pretty complicated :-) |
| 05:19 | <roc> | using the height of the root element seems a lot simpler |
| 05:19 | <Hixie> | originally i went this way because it was the route you suggested in the mozilla bug on the topic :-) |
| 05:19 | <roc> | I guess I changed my mind :-) |
| 05:19 | <Hixie> | let's use the root element height then |
| 05:19 | <roc> | ok cool |
| 05:20 | <Hixie> | like i said in my last e-mail, i'm basically just waiting for your experience |
| 05:20 | <Hixie> | i don't know what the spec should say really |
| 05:20 | <Hixie> | i don't want it to be as complicated as sicking's "let's merge the DOMs" idea though |
| 05:20 | <Hixie> | that scares me |
| 05:20 | <roc> | me too |
| 05:20 | <roc> | we're not going to be doing anything except the intrinsic width/height stuff initially |
| 05:21 | <Hixie> | cool |
| 05:21 | <roc> | I don't think that level of incomplete implementation will break the feature going forward, but if that's a concern, we can make it mozSeamless or something |
| 05:21 | <Hixie> | what are you not doing? |
| 05:22 | <roc> | the stylesheet integration |
| 05:22 | <Hixie> | ah, right |
| 05:22 | <Hixie> | hmm |
| 05:22 | <Hixie> | yeah probably best to call it moz-seamless |
| 05:22 | <Hixie> | that's a pretty big difference |
| 05:23 | <roc> | ok |
| 07:58 | <jmb> | ~ |
| 08:21 | <BenMillard> | for the IRC records, if you haven't seen it on Public-HTML the current progress of my web page research for this year is here: http://projectcerbera.com/web/study/2008/collection |
| 08:34 | webben_ | finds the use by http://www.enableholidays.com/ (from BenMillard's list) outline-style: none; mildly depressing |
| 08:35 | <BenMillard> | that disables focus outlines, right? |
| 08:36 | <BenMillard> | I find it rather unfortunate, too |
| 08:37 | <webben_> | BenMillard: Right. |
| 08:37 | <Hixie> | http://www.hereticpress.com/Dogstar/Publishing/WebSurvey.html gives me an error message |
| 08:39 | <BenMillard> | works for me |
| 08:40 | <Hixie> | it says: |
| 08:40 | <Hixie> | Your permission to access Heretic Press has been denied. |
| 08:40 | <Hixie> | Your browser may be unidentified, you may be downloading too many files for offline viewing or are breaching copyright on digital images. Your IP address might be trying to access password protected files? |
| 08:40 | <Hixie> | You may have been unfairly excluded from access. Sorry. |
| 08:40 | <BenMillard> | I'm using Firefox 2.0.0.16 on Windows XP |
| 08:42 | <webben_> | Hixie: wfm in Lynx and OmniWeb |
| 08:42 | <Hixie> | oh well |
| 08:43 | <gDashiva> | At least they acknowledge false positives may happen :) |
| 08:44 | <BenMillard> | Hixie, do you have scripting disabled? |
| 08:44 | <BenMillard> | looking at the source code, the error you gave is placed in a <noscript> element near the start |
| 08:45 | <BenMillard> | the pages are built in a very weird way...maybe I should study them :P |
| 08:46 | <BenMillard> | it has 23 <link rel> elements, many seem to be using made-up rel values |
| 08:47 | <BenMillard> | 23 <link rel> elements pointing to related documents, I mean |
| 10:29 | <Hixie> | ok, the spec now defines an event loop mechanism |
| 10:29 | <Hixie> | and that is used throughout whenever asynchronous behaviour is involved |
| 10:33 | <Hixie> | i wonder how the es-future work is going to do opt-in with javascript: urls |
| 10:38 | <hsivonen> | Hixie: do I guess correctly that you are not a fan of opt-in? |
| 10:38 | <Hixie> | yes. |
| 10:38 | <Hixie> | well no |
| 10:39 | <Hixie> | i'm specifically not a fan of out-of-band opt-in |
| 10:39 | <Hixie> | that is, i have professional reservations about out-of-band opt-in |
| 10:39 | <Hixie> | i also have personal reservations about any kind of opt-in, but that's just a personal preference |
| 10:40 | <Hixie> | out-of-band opt-in as far as i am concerned is a complete non-starter. for example, it means that you can never update an existing widely deployed script to use |
| 10:40 | <Hixie> | the new features |
| 10:41 | <Hixie> | e.g. google couldn't on-the-fly use new features in the google analytics scripts |
| 10:45 | <Philip`> | Gecko supports most of the new JS features without any opt-in; you only need opt-in for new keywords |
| 10:46 | <Hixie> | yeah |
| 10:46 | <Hixie> | so silly |
| 10:46 | <Philip`> | (so it should only change the behaviour of scripts that didn't compile under previous versions of the language) |
| 10:47 | <Hixie> | anyway, bed time |
| 10:48 | <virtuelv_> | Hixie: any particular reasoning behind the odd camelcasing of navigator.onLine? |
| 10:53 | <annevk> | virtuelv_, ask Microsoft |
| 10:54 | <virtuelv_> | heh |
| 10:54 | <virtuelv_> | that was an answer at least |
| 11:18 | hsivonen | doesn't like it when in a public bug tracker fanboys try to explain how the product doesn't suck and the reporter should jump through hoops instead |
| 11:18 | <hsivonen> | (remark not related to any HTML bug tracker bug) |
| 11:27 | <gDashiva> | Don't you know they're doing you a service by letting you help improve the product? ;) |
| 11:28 | gDashiva | groans at Dr. Hoffman getting started on poetry markup again |
| 13:11 | <jcranmer> | "The W in WHATWG stands for..." shouldn't that be "the first W" ? |
| 14:16 | jgraham | wonders if it's worth dragging up the fact that WCAG is only chartered to address the needs of users with disabilities whereas HTML has to work for a wider range of people |
| 14:18 | <hsivonen> | jgraham: yes, if you think it's worthwhile to touch the subject matter in general |
| 14:19 | <jgraham> | hsivonen: That's what I can't decide. I don't think I will be very popular... |
| 14:19 | <jgraham> | (but perhaps it needs to be said) |
| 14:21 | <hsivonen> | PT(W) completely missed the point of my question |
| 14:22 | <Philip`> | Then people might say that alt is primarily an accessibility feature so it needs to address the needs of users with disabilities and therefore WCAG is the right place to do that |
| 14:22 | <gDashiva> | I am amazed, over and over, at how little faith they have in WCAG |
| 14:23 | <hsivonen> | gDashiva: have you read WCAG 2.0? If you have, do *you* have faith in it? |
| 14:23 | <gDashiva> | hsivonen: No, but then I don't consider it a good idea either |
| 14:24 | <gDashiva> | (My first introduction to it was Joe Clark's article, and I haven't seen anything to dispute it) |
| 14:29 | <jgraham> | Philip`: It's not just an accessibility feature though. It's also needed or e.g. text-only browsers |
| 14:29 | <jgraham> | s/or/for/ |
| 14:30 | <gDashiva> | Are balloons 'moored' when landed or is there a different verb? |
| 14:30 | <hsivonen> | gDashiva: it's not quite as bad as it used to be |
| 14:30 | <hsivonen> | gDashiva: it's still bad, though |
| 14:30 | <Lachy> | othermaciej, yt? |
| 14:31 | <gDashiva> | hsivonen: Bad enough that you might be better off with the samurai errata instead? |
| 14:32 | <Philip`> | gDashiva: http://www.balloonlife.com/publications/balloon_life/9608/images/tetherm.gif shows a "MOORED" balloon, so that sounds reasonable |
| 14:32 | <Philip`> | (assuming it is actually moored, not just sitting on the ground because it's too heavy to float) |
| 14:32 | <hsivonen> | gDashiva: bad enough that the WG chose to produce exegesis documents themselves, because the spec itself is not suitable for mortals to interpret on their own |
| 14:33 | <gDashiva> | Philip`: Good enough, thanks |
| 14:33 | <gDashiva> | I'm anticipating a new category of spec writers, the ones sitting in the balloon moored on top of the ivory tower |
| 14:37 | <Philip`> | jgraham: Indeed, but it sounds like that'd possibly just shift the arguments to go in a different direction, rather than actually resolving anything |
| 14:38 | <Lachy> | http://arstechnica.com/news.ars/post/20080819-mozilla-drags-ie-into-the-future-with-canvas-element-plugin.html |
| 14:38 | <Philip`> | How would you make a whole tower out of ivory, unless you poached one of the Discworld elephants? |
| 14:38 | <gDashiva> | Philip`: That's what we've been doing all along |
| 14:39 | <gDashiva> | One person agrees to one resolution, and then another one picks up a different end and starts it all over again |
| 14:39 | <jgraham> | Philip`: Well the point that I may or not make is that you end up with different conformance requirements if your only goal is to support the disabled compared to those you end up with if your goal is to support everyone |
| 14:39 | <gDashiva> | I'm sure every single person on the list has said "Yeah, that would work" to at least one suggestion |
| 14:39 | <gDashiva> | But they never do it to the same one |
| 14:40 | <jgraham> | e.g. WCAG seems to forbit publishing James Joyce's Uylsses in an unaltered form |
| 14:40 | <jgraham> | s/forbit/forbid/ |
| 14:42 | <Philip`> | Lachy: Why do people think Adobe releasing an ActionScript VM as open source has any relation to making it not totally unrealistic that Adobe would distribute Mozilla's canvas plugin with Flash? That part of the article seems totally irrelevant to me... |
| 14:43 | <jgraham> | also the WCAG people still seem to have the attitude that people should be prepared to put an unlimited amount of effort into accessibility despite the evidence that this is not true. Designing around this assumption doesn't help accessibility but does have negative consequences on people's perception of the value of conformance requirements |
| 14:43 | <Lachy> | Philip`, I have no idea |
| 14:44 | <Philip`> | (and the rest of the article is just rephrasing the three-week-old blog post) |
| 14:44 | <Lachy> | I didn't see the 3 week old blog post before |
| 14:44 | <Philip`> | You need to subscribe to more people's blog feeds :-) |
| 14:45 | <Lachy> | I'm probably on too many already. It's difficult to keep up with everything |
| 14:55 | hsivonen | wonders if the whole content area of IE could be taken over by Gecko-as-ActiveX |
| 14:59 | hsivonen | sees http://starkravingfinkle.org/blog/2006/12/xule-what-if/ |
| 15:04 | <Philip`> | Combining the web-compatibility of Gecko with the UI of IE? That sounds like a fantastic idea |
| 15:04 | <gDashiva> | How long does an internet argument have to last before you can stop assuming ignorance and start assuming malice? |
| 15:07 | <jgraham> | gDashiva: I thought most people assumed that from the start ;) |
| 15:12 | <Lachy> | is there anyone here who understands and can explain how DOMImplementation.getFeature() and Node.getFeature() in DOM3Core are supposed to work? |
| 15:14 | <takkaria> | gDashiva: given the length of the alt thread, I'd say at least a few years |
| 15:20 | <Lachy> | JohnResig, yt? |
| 15:21 | <JohnResig> | Lachy: what's up |
| 15:21 | <Lachy> | JohnResig, the selectors api test suite needs tests for hasFeature() and getFeature() |
| 15:22 | <Lachy> | I'm not exactly sure what getFeature() is supposed to do, besides return a copy of the object itself if the feature is implemented |
| 15:22 | <Lachy> | and if that's all it's meant to do, then it seems quite useless |
| 15:23 | <Lachy> | and Node.isSupported() too |
| 15:24 | <Lachy> | and probably getDOMImplementation() too |
| 15:24 | <Lachy> | and getDOMImplementationList() |
| 15:24 | Lachy | wishes he didn't add a feature string. I didn't realise it would complicate so many things :-( |
| 15:25 | <JohnResig> | Lachy: could you come up with a list of what exactly is supposed to be tested? |
| 15:28 | <Lachy> | JohnResig, I'm mailing public-webapps about it now |
| 15:33 | <Lachy> | JohnResig, http://lists.w3.org/Archives/Public/public-webapps/2008JulSep/0436.html |
| 15:35 | <JohnResig> | Lachy: what feature/version strings should I be checking for? |
| 15:37 | <Lachy> | see the spec |
| 15:38 | <Lachy> | "Selectors-API", "1.0" |
| 16:24 | <takkaria> | argh, someone talking about it being a moral duty to the web to produce valid markup |
| 16:26 | <Philip`> | We need to define some HyperText Morality Laws to enforce that |
| 16:28 | <takkaria> | I'm not sure there's enough demand for that spec :) |
| 16:29 | <Philip`> | Lack of demand doesn't mean we don't have a moral obligation to write it |
| 16:30 | <takkaria> | true enough |
| 16:31 | <hsivonen> | a moral duty to use an html5 validator could be good for me |
| 16:31 | <hsivonen> | if others had the duty that is |
| 16:31 | <MikeSmith> | hsivonen: fwiw, I notice that the v.nu schemas define ol.elem.phrase, ul.elem.phrase, and dl.elem.phrase patterns -- but those are never actually referenced by any defines |
| 16:32 | <takkaria> | hsivonen: it could be good for you too if you used AdSense :) |
| 16:33 | <hsivonen> | MikeSmith: I left those around expecting Hixie to allow a <span> hack for doing phrase-level lists |
| 16:36 | <MikeSmith> | hsivonen: ah, I see |
| 17:21 | <hober> | I wonder why Sunava always says "HTML 5.0". Where did the ".0" come from? |
| 17:23 | <gDashiva> | Maybe it's a microsoft thing? |
| 17:24 | <gDashiva> | If you're always doing minor releases, it might become a habit to attach .0 |
| 17:24 | <Lachy> | hober, 5.0 == 5. |
| 17:25 | <Lachy> | it's probably just an assumption based on the fact that it was HTML 2.0, 3.0, 3.2, 4.0 and 4.01 before. |
| 17:29 | gDashiva | wonders about the logic going on |
| 17:30 | <gDashiva> | Maybe something like "If it's easy, it's too easy. Too easy is bad. Accessibility isn't bad. Therefore easy means it's not accessible!" |
| 17:47 | <othermaciej> | Lachy: yes? |
| 18:00 | <Lachy> | othermaciej, someone at Opera was wondering what the licencing is for WebKit's testcases, like these. http://trac.webkit.org/browser/trunk/LayoutTests/fast/css/ Do you know? |
| 18:01 | <othermaciej> | I guess it's not always clearly stated in the test case |
| 18:01 | <othermaciej> | I think the intent is for the test suite to be BSD licensed, other than test cases from external sources |
| 18:02 | <Lachy> | ok, that makes sense, since the rest of webkit is BSD or LGPL |
| 19:05 | <anne_mibbit> | jgraham, where shall we meet? |
| 19:05 | <jgraham> | anne_mibbit: Dunno. |
| 19:05 | anne_mibbit | is in cambridge now, at "Cats" |
| 19:05 | <jgraham> | Did we decide when we were meeting? |
| 19:06 | <anne_mibbit> | I thought today :) |
| 19:06 | <jgraham> | Oh, I didn't realise that we had decided that. Does Philip` know? |
| 19:06 | <jgraham> | Philip`: ^^ that was aimed at you |
| 19:07 | <jgraham> | I can wonder into town, I guess |
| 19:07 | <anne_mibbit> | I'm not really sure where the rest of the CSS gang is and I need some food |
| 19:08 | <anne_mibbit> | quite a long trip to get here |
| 19:08 | <jgraham> | OK I will come in |
| 19:09 | <anne_mibbit> | wait, where are you going to? |
| 19:09 | <anne_mibbit> | hmm |
| 19:10 | <jgraham> | I could meet you outside cats I guess. Oroutside Kings which is a bit closer to where I am going and I guess better for Phillip |
| 19:10 | <jgraham> | Only one l ... |
| 19:10 | <anne_mibbit> | lets see |
| 19:10 | <jgraham> | It will take me about 30 minutes to get there |
| 19:11 | <anne_mibbit> | where is kings? |
| 19:12 | <anne_mibbit> | oh, Kings College that is a bit further up the road? |
| 19:12 | <jgraham> | http://maps.google.co.uk/maps?f=q&hl=en&geocode=&q=kings+cambridge&ie=UTF8&ll=52.207764,0.132179&spn=0.022617,0.06815&z=15&iwloc=A |
| 19:13 | <anne_mibbit> | where are looking at the same building :) |
| 19:13 | <anne_mibbit> | s/where/we/ |
| 19:13 | <anne_mibbit> | ok, I'll go there in 10 minutes |
| 19:14 | <jgraham> | OK, I'll meet you outside the front gate, by the odd-shaped post box |
| 19:14 | jgraham | leaves |
| 19:15 | <anne_mibbit> | Philip`, ^^ |
| 19:22 | <gsnedders> | anne_mibbit: I'd say it'd take maybe two or three minutes to walk there from Cat's, FYI |
| 19:24 | <anne_mibbit> | anyway, got to go |
| 20:08 | <zcorpan> | Hixie: #seekUpdate says The user agent must queue a task to queue a task to ... |
| 21:05 | <Philip`> | anne, jgraham: Oh, I thought people had said Thursday or Friday or something, not today |
| 21:06 | <Philip`> | and I can't make it now |
| 21:08 | <Philip`> | (jgraham: By the way, I've moved house since we last met and now I live on pretty much exactly the opposite side of King's, though still about the same distance away so it wouldn't actually make much difference)C[C[C[C[C[C[C[C[C |
| 21:09 | <Philip`> | (Um, that's a peculiar sequence of characters on the end there) |
| 21:09 | Philip` | blames his intermittent wireless disconnections |
| 22:32 | Hixie | replies to http://lists.w3.org/Archives/Public/wai-xtech/2008Aug/0114.html |
| 22:33 | <Hixie> | i wish i understood more of what Al said |
| 22:33 | <Hixie> | http://lists.w3.org/Archives/Public/wai-xtech/2008Aug/0118.html is my reply |
| 22:47 | <webben> | Hixie: Isn't Al's point about not bothering referring to mass photo uploaders who don't bother to provide text equivalents? His stuff about the two agents is simply that it's only by considering the mass upload photo site web developers are agents in isolation that one can shunt this not bothering outside the remit of the spec. |
| 22:48 | <webben> | Hixie: Seems to me people regular pop up to say they can't be bothered to provide text equivalents for all their photos on Flickr. |
| 22:48 | <webben> | (in the various threads about alt) |
| 22:49 | <Hixie> | but the author in that case is flickr, not the uploader |
| 22:50 | <Hixie> | the uploader is just a user |
| 22:50 | <webben> | Hixie: I think that's a distinction Al's questioning. |
| 22:51 | <Hixie> | i doubt it, that would be crazy |
| 22:51 | <webben> | how would it be crazy |
| 22:51 | <webben> | let's say you have an interface that provides a box in which i enter text |
| 22:51 | <webben> | and then that text is dumped out in some HTML |
| 22:51 | <webben> | are you really saying I'm not the author of the text? |
| 22:52 | <webben> | conversely, if the box is left blank, am I not the author of the blankness |
| 22:52 | <Hixie> | you're not the author of the web page, certainly. you don't even know what html is, in all likelihood. |
| 22:52 | <webben> | I think that's like arguing a publisher is an author of the book. |
| 22:52 | <Hixie> | nobody is seriously going to think that flickr users are going to give two hoots about what html5 says about their usage of flickr |
| 22:52 | <webben> | Hixie: yes. but that's not the point contended. |
| 22:53 | <Hixie> | no, it's like arguing that the person who sends a letter to an editor of a newspaper is responsible for the typography in the newspaper |
| 22:53 | <webben> | Hixie: I don't think the absence of a text equivalent is equivalent to the typography of the newspaper. |
| 22:53 | <Hixie> | or that a book author is responsible for the placement of the book on the bookshelf in the book shops |
| 22:53 | <webben> | I think that's more like leaving your name off the letter. |
| 22:54 | <webben> | what the software does with the absence of the text equivalent is perhaps like the typography; but not "authorship". |
| 22:55 | <Hixie> | *shrug* |
| 22:55 | <Hixie> | it's a moot argument |
| 22:55 | <webben> | it might help if "author" wasn't used in isolation but in a phrase like "author of" |
| 22:55 | <Hixie> | flickr users aren't going to provide useful text replacements |
| 22:55 | <Hixie> | there's really no point discussing whose fault that is |
| 22:55 | <webben> | e.g. "author of code to handle some user input" vs. "author of an image and its text equivalent, or lack thereof" |
| 22:55 | <Hixie> | the html5 spec talks about the conformance of the web page |
| 22:55 | <webben> | Hixie: Al isn't discussing whose fault it is. |
| 22:56 | <Hixie> | (regardless of who wrote it) |
| 22:56 | <Hixie> | so what is he discussing? |
| 22:57 | <webben> | He's saying there's some not bothering going on, and that's clearly true: indeed it's precisely what you're saying. |
| 22:57 | <Hixie> | ah |
| 22:57 | <Hixie> | well that seems like a moot point too |
| 22:58 | <webben> | he's saying that "According to HTML5, if the human didn't bother, the page isn't compliant" isn't true, because it's meaningless to exclude the most important human - the one who actually authors the meaning and would normally provide the text equivalent. |
| 22:58 | <webben> | whether it's important whether that's true or not doesn't make it true or not. |
| 22:58 | <Hixie> | hm |
| 23:00 | <Hixie> | given that flickr can't ever get suitable replacement text for most of these images, and given that we want to cater for flickr since that kind of site is a big part of the web, it's not clear to me how to address that feedback in the spec. |
| 23:01 | <webben> | well, Flickr is a largely moot case, because so far neither conformance not text equivalents have proved of any importance to flickr |
| 23:02 | <Hixie> | i use flickr as a symbol for any media sharing site |
| 23:02 | <webben> | one kind of has to imagine an idealized flickr that cares about conformance (why?) and text equivalents (SEO? accessibility? interoperability?) for it to be even useful as a way to think about this problem |
| 23:03 | <Hixie> | we want it to be technically possible to write a media sharing site using conforming html5 |
| 23:03 | <Hixie> | even if the users don't provider alternative text (since they won't) |
| 23:03 | <Hixie> | provide, even |
| 23:04 | webben | isn't really clear on why. |
| 23:04 | <Hixie> | why what? |
| 23:04 | <gavin_> | why you want that |
| 23:04 | <webben> | why we want it |
| 23:04 | <Hixie> | media sharing sites are amongst the most important (by usage) sites on the web |
| 23:04 | <Hixie> | it would be a pretty big failure if we didn't address them |
| 23:04 | <webben> | would it? why? |
| 23:05 | <webben> | why isn't enough to define how UAs should interoperate with them? |
| 23:05 | <gavin_> | what practical effects would "not addressing them" have? |
| 23:05 | <Hixie> | because we're trying to write a language for the web, and they're a big part of the web? |
| 23:05 | <Hixie> | i don't understand the question |
| 23:06 | <webben> | yes, but the reality is not every document will "conform" (indeed, if history is any guide at all, most won't) and that UAs need to cope with that. |
| 23:06 | <Hixie> | sure, but so what? |
| 23:06 | <Hixie> | that's already dealt with |
| 23:06 | <Hixie> | that's the easy part of this problem, frankly |
| 23:06 | <webben> | indeed. |
| 23:07 | <gavin_> | you want flickr to be conformant, and accessibility people want conformance to mean accessibility, and flickr will not be accessible |
| 23:07 | <gavin_> | right? |
| 23:07 | <webben> | I guess it depends on what the point of "conformance" is . |
| 23:07 | <Hixie> | i don't know what the accessibility people want, but if they want conformance to mean accessibility, they're doomed. |
| 23:07 | <Philip`> | Flickr's developers don't care about conformance, so it's not going to have any practical effect on them at all |
| 23:07 | <webben> | I mean, I completely agree one wants authors to write markup with defined effects. |
| 23:08 | <webben> | but afaict the only point of conformance is saying: here's a good way to do X. |
| 23:08 | <gavin_> | assuming what I said is true, the easy way to resolve the conflict is to stop wanting flickr to be conformant |
| 23:08 | <gavin_> | I don't see offhand what the problem with that would be |
| 23:08 | <gavin_> | (in practical terms) |
| 23:08 | <webben> | and there may be things where one can write markup with defined effects but there's fundamentally no good way to do it. |
| 23:09 | <Hixie> | conformance has many points, in particular promoting best practices and trying to keep people away from things that will make it harder to extend the language later |
| 23:09 | <webben> | maybe it's worth _thinking_ about seperating the two. |
| 23:10 | <Hixie> | however, to be able to convince people to care about conformance, we have to make it possible to be conformant |
| 23:10 | <Philip`> | What's the point in trying to keep people away from those things, when most people are going to completely ignore that attempt and will do those things regardless? |
| 23:10 | <Hixie> | it's not black and white, it's a matter of degrees |
| 23:10 | <webben> | Hixie: The funny thing about that statement, from what I've seen, is that the inability to extend is one of the /drivers/ of non-conformance. |
| 23:11 | <webben> | e.g. people wanting to make up their own elements and attributes |
| 23:11 | <Hixie> | conformance is like an attractor, by having it as a goal we get people closer to the goal, if not at the goal |
| 23:11 | <webben> | you've gone some way to head that off with data-* |
| 23:11 | <Hixie> | whereas without it, we'd have people distributed across all of phase space |
| 23:11 | <Hixie> | webben: i don't think that extending html is a driver of non-conformance |
| 23:12 | <Hixie> | webben: most non-conformance isn't to do things that html can't do |
| 23:12 | <Hixie> | webben: it's just mistakes and not understanding html |
| 23:12 | <webben> | Hixie: yeah that's accidental non-conformance. I'm talking about deliberate non-conformance. And not making a claim about "most". |
| 23:13 | <Hixie> | "most" is what matters |
| 23:13 | <Hixie> | it's all a matter of degrees |
| 23:13 | <Hixie> | anyway |
| 23:14 | <Hixie> | not much i can do to satisfy what al wants if what he wants is what you say he wants |
| 23:14 | <webben> | not with a single profile of conformance if you want publishing stuff without text equivalents to be promoted as a good thing to do. |
| 23:15 | <Hixie> | the spec says it's a bad thing to do |
| 23:15 | <webben> | what's the point of making a bad thing to do conformant again? |
| 23:15 | <webben> | to encourage people do other things right? |
| 23:16 | <webben> | *to do |
| 23:16 | <Hixie> | the point is to get people to do the best thing given their situation |
| 23:17 | <webben> | where "people" is basically software developers? |
| 23:17 | <Hixie> | anyone writing HTML documents |
| 23:18 | <webben> | yeah, the notion "writing" there seems rather problematic |
| 23:18 | <Hixie> | only if one is looking for it to be problematic, imho |
| 23:18 | <webben> | I'm a web developer on a movies website, but I don't "write" the documents |
| 23:19 | <webben> | I guess the point is clearly conformance here is not aimed at getting people to write text equivalents for all their photos: which would be the "best" thing to do. |
| 23:20 | <webben> | the conformance is rather aimed at getting software developers to handle the lack of text equivalents sanely. |
| 23:20 | <webben> | that goes back to the split Al identifies between two agents. |
| 23:20 | <Hixie> | the "best" thing to do would be for the blind people to be able to see. |
| 23:22 | <Hixie> | the spec is just giving advice to the people outputting the HTML page in what the best they might be able to do given their constraints is |
| 23:22 | <webben> | writing text equivalents is possible with current technology; allowing the blind to see doesn't seem to be (yet) and wouldn't help some of the other use-cases of text equivalents |
| 23:22 | <Hixie> | even suggesting that the users uploading their images are html authors in this context is laughable, imho |
| 23:23 | <Hixie> | whether writing text equivalents is possible or not is moot if the text equivalents aren't going to be forthcoming |
| 23:23 | <Hixie> | it is just as impossible to convince flickr users to provide suitable replacement text as it is to make the replacement text unnecessary |
| 23:23 | <webben> | doesn't really see where the laughing matter is actually. I think you're reversing the normal sense of authorship in publishing. |
| 23:23 | <Hixie> | (if anything, the latter is likely easier on the long term) |
| 23:24 | <webben> | that's not necessarily a big problem, it just makes the whole thing more obfuscated |
| 23:24 | <Hixie> | i'm not sure what the antecedent for "that" was |
| 23:24 | <webben> | that => "reversing" |
| 23:25 | <Hixie> | ah |
| 23:25 | <Hixie> | the user has no control over |
| 23:25 | <Hixie> | what flickr does |
| 23:25 | <Hixie> | flickr has no control over what the user does |
| 23:25 | <Hixie> | the html page is under the control of flickr |
| 23:26 | <Hixie> | so flickr is the author |
| 23:26 | <webben> | I don't think the "control" thing has much to do with authorship. |
| 23:26 | <Hixie> | seems pretty cut and dry to me |
| 23:26 | <webben> | I don't control what Penguin does. Penguin doesn't control what I do. |
| 23:26 | <svl> | "I wear the cheese. The cheese does not wear me." |
| 23:27 | <Hixie> | i just saw that episode |
| 23:27 | <svl> | :) |
| 23:27 | <Hixie> | webben: control over the html page has everything to do with whose responsibility the conformance of the page is |
| 23:28 | <webben> | Hixie: possibly yes. I think "author" is a poor choice of word for the person "whose responsibility the conformance of the page is" |
| 23:28 | <webben> | (yes, in rare circumstances, they are the same person) |
| 23:29 | <Hixie> | the spec doesn't use the term "author" |
| 23:29 | <Hixie> | (for this requirement) |
| 23:29 | <webben> | I don't think photo sharing sites can consistently say the lack of text equivalents is both the user's fault and their responsibility however |
| 23:29 | <Hixie> | it doesn't matter whose blame it is |
| 23:30 | <Hixie> | it is their responsibility because they generate the html |
| 23:30 | <Hixie> | maybe a better example would be a webcam |
| 23:30 | <Hixie> | or a blind photographer's photo site |
| 23:30 | <webben> | I doubt it. |
| 23:30 | <webben> | webcam's don't have authors |
| 23:30 | <Hixie> | that's the point |
| 23:30 | <Hixie> | there is html |
| 23:30 | <webben> | flickr does |
| 23:31 | <Hixie> | ok let's talk about the webcame case then |
| 23:31 | <Hixie> | there is only one human |
| 23:31 | <Hixie> | the one who generates the HTML with the <img> |
| 23:31 | <Hixie> | he has no idea what the image is of |
| 23:31 | <Hixie> | does that make the issue clearer? |
| 23:31 | <webben> | is this an important use-case though? |
| 23:32 | <Dashiva> | webben: Tried using the internet lately? There are a lot of webcams |
| 23:32 | <webben> | Dashiva: aren't they broadcasting in Flash players? |
| 23:32 | <Hixie> | it is a use case that i want to make sure we cover. |
| 23:33 | <Dashiva> | You'd be surprised how clever people are |
| 23:33 | <Hixie> | because i see no reason to not cover it. |
| 23:33 | <jruderman> | how about making HTML validators warn when they encounter alt="{User-uploaded image}", saying "it would be great if you could supply real alt text there"? |
| 23:33 | <Dashiva> | multipart jpeg, motion jpg, refreshing with script... |
| 23:33 | <Hixie> | jruderman: doesn't the validator.nu validator already do that? |
| 23:33 | <jruderman> | ahh :) |
| 23:33 | <webben> | jruderman: Doesn't Flickr already know that... ? |
| 23:34 | <webben> | when I deal with photos coming in from a feed with failsome or missing text equivalents and need to pump them into a gallery, I certainly know the alts could be better! |
| 23:34 | <Hixie> | jruderman: hsivonen certainly has a great image report feature, dunno whether it's been updated to the new text. |
| 23:34 | <Hixie> | jruderman: but also... his image report feature itself is a great example of a page where there is an image and no alternative text |
| 23:34 | <Philip`> | jruderman: Then people will write alt="User-uploaded image" to get rid of the ugly warning |
| 23:34 | <Hixie> | jruderman: and i'm pretty sure we want to make sure the html5 conformance checker can be conforming |
| 23:34 | <jruderman> | Philip`: hmm, good point |
| 23:35 | <Dashiva> | Hixie: It's an edge case ;) |
| 23:35 | <Hixie> | pretty damn important one |
| 23:35 | <jruderman> | lol |
| 23:35 | <Dashiva> | He can just run it on ftp, can't he? |
| 23:36 | <Hixie> | it'd still be html :-) |
| 23:36 | <Philip`> | He could remove the HTML output, and just have XML/JSON/text/etc |
| 23:37 | <Hixie> | oh yeah, that'd be awesome. html5 is so great, even its conformance checker can't use it. |
| 23:37 | <webben> | Hixie: Sure. But I think what's "best" for a validator may be different than what's "best" for a photo uploader. |
| 23:37 | <Hixie> | webben: then let's cover both in the spec. what is the best for both? |
| 23:37 | <Hixie> | webben: as in, what should the spec recommend in each case? |
| 23:37 | <Philip`> | I hope the validator's text output will use AAlib for the image report feature |
| 23:37 | <Hixie> | Philip`: doesn't help a blind user. :-) |
| 23:38 | <Philip`> | Does anything like WCAG require that text/plain documents be accessible? |
| 23:39 | <webben> | Hixie: I think you've already covered that what's best for the person uploading images to do is provide text equivalents. |
| 23:39 | <Dashiva> | Imagine if all the energy being used on alt was used to make people consider WCAG conformance equally (not so) important as HTML conformance |
| 23:39 | <webben> | and what's best for the person designing a system to house content images is to require text equivalents. |
| 23:39 | <webben> | that doesn't mean those systems will require text equivalents |
| 23:39 | <jruderman> | Philip`: ASCII art ftw |
| 23:39 | <webben> | (and relaxing that recommendation wouldn't _mean_ that those systems would in fact conform) |
| 23:40 | <Dashiva> | Philip`: You have to use as simple language as possible |
| 23:40 | <Hixie> | webben: but what should happen when they don't have text equivalents? |
| 23:40 | <Hixie> | Philip`: common sense does, but maybe that's not "like WCAG" |
| 23:40 | Hixie | ducks |
| 23:40 | <webben> | Hixie: but they get into that position by choice |
| 23:40 | <webben> | Hixie: i.e. they create a system that doesn't gather the data. |
| 23:41 | <webben> | Hixie: whereas a validator cannot gather the data since there is none |
| 23:41 | <Hixie> | webben: hsivonen's validator doesn't, nor does a webcam page, nor does a blind photographer, nor does a fractal generator, nor does a ... |
| 23:41 | <Hixie> | webben: so what should those cases do? |
| 23:41 | <Hixie> | webben: and what should happen if the user doesn't provide alternative text? |
| 23:42 | <webben> | Hixie: You can't get into a situation where the user doesn't provide alternative text without choosing that to allow that situation; in which case you've already chosen not to produce the best possible consumer experience. |
| 23:42 | <webben> | it might well be a reasonable business tradeoff |
| 23:43 | <Dashiva> | webben: What if the user chooses to provide nothing? |
| 23:43 | <webben> | Dashiva: Again, it's the business that has chosen to allow the user to provide nothing and still publish. |
| 23:43 | <Dashiva> | No, nothing as in nothing useful |
| 23:44 | <Dashiva> | A space, asdf, something like that |
| 23:44 | <Hixie> | webben: but assuming one has made the "reasonable business tradeoff" of not rejecting 99% of users, what should one do? |
| 23:44 | <webben> | Dashiva: Is "nothing useful" machine detectable? |
| 23:44 | <Dashiva> | No |
| 23:44 | <webben> | Hixie: conform on other points? |
| 23:44 | <Dashiva> | If it were, we wouldn't be having this discussion |
| 23:44 | <webben> | (if one care's about conformance at all) |
| 23:44 | <Hixie> | webben: not good enough |
| 23:44 | <Dashiva> | We'd just generate alt text |
| 23:45 | <Hixie> | webben: what should the spec say? if you don't have alternative text you... what? omit alt=""? include alt="{image}"? what? |
| 23:45 | <webben> | Hixie: what's not good enough? conforming to best practice on some points and not others ? you can't avoid that, since best practice will be avoided on business grounds. |
| 23:46 | <webben> | e.g. opening new windows sucks, but the business chooses to do it anyway, or whatever. |
| 23:46 | <Dashiva> | webben: So you'd rather have a photo site filled with bogus alt values, than a photo site with some useful and some missing alt values? |
| 23:46 | <Hixie> | webben: the spec not giving advice for the case where the html generator doesn't have suitable alternative text is not good enough |
| 23:46 | <webben> | Dashiva: Nothing I said implies that. |
| 23:47 | <Dashiva> | webben: Sure it did. You said a site has to force its users to provide alt to be conforming. |
| 23:47 | <Hixie> | webben: the spec _will_ give advice for the case where suitable alternative text is missing. the question is, what should that advice be? |
| 23:47 | <Dashiva> | Forcing users leads to all kinds of "interesting" input |
| 23:47 | <webben> | Hixie: I agree the spec probably should give advice. don't see why it needs to be a conformance point. |
| 23:47 | <Hixie> | webben: the spec saying "when you are in case X, do Y. Never do Y." is the height of idiocy in a spec. |
| 23:47 | <Hixie> | webben: it's what we call a contradiction. |
| 23:48 | <webben> | Dashiva: No. I said a site has to require text equivalents in order to provide the best experience. |
| 23:48 | <Hixie> | webben: it makes people ignore the spec as a whole |
| 23:48 | <Dashiva> | webben: Regardless of the reason, 'require' is the problem |
| 23:48 | <webben> | Dashiva: it may be good enough, for a given site, to provide less than the best experience. |
| 23:50 | <Dashiva> | webben: So you're saying there's no point in striving for the best |
| 23:50 | <webben> | Hixie: No, the spec would be saying if you put yourself in case X, you cannot expect the best experience. |
| 23:50 | <webben> | Dashiva: of course there's a point in striving for the best. |
| 23:51 | <Hixie> | webben: if the spec says to do X, it can't also say that doing X is not allowed. That's a contradiction. |
| 23:51 | <Dashiva> | webben: You said yourself that less than best is good enough |
| 23:51 | <jgraham> | webben: I dispute that prviding text equivalents for my flickr images would provide an improved user experience; the point of the images is their aesthetic value as images. Any replacement text (other than something banal like "phtograph") would just highlight my (lack of) writing skills |
| 23:51 | <webben> | Hixie: Which is simply true. |
| 23:51 | <Hixie> | as far as i can tell, you're just sticking your head in the sand here |
| 23:51 | <Hixie> | your position is logically contradictory and you're just denying that that is the case |
| 23:52 | <Hixie> | which doesn't make it any less the case |
| 23:52 | <Hixie> | if anything, it just makes it metacontradictory as well |
| 23:52 | <webben> | jgraham: Your inability to transmute the aesthetic experience into words (and I might well be just as bad) does not mean that such a transmutation would not provide a better user experience. |
| 23:53 | <webben> | jgraham: There do seem to be whole reams of writing about aesthetic experiences. |
| 23:53 | <Hixie> | webben: what should the spec say the author should do in the case of a page with a webcam? |
| 23:53 | <Dashiva> | webben: You can't have it both ways. Either striving for the best is worthwhile, and then people WILL strive for it. Or it's a waste of time, in which case people won't try, and we're wasting time defining it. |
| 23:54 | <webben> | Dashiva: Can't see where I'm trying to have it both ways. |
| 23:54 | <webben> | Dashiva: I think there are different "bests" in operation. |
| 23:54 | <jgraham> | webben: So given that I would be unable to communicate the aesthetic experience, even if I believed it was possible to achieve at all (I don't) who should I get to make my pages containing photos conforming? |
| 23:54 | <Dashiva> | webben: Two legs bad |
| 23:54 | <webben> | And it's important to clarify which of those "bests" conformance is about (if any). |
| 23:54 | <jruderman> | stop using a webcam because webcams are omg inaccessible. porn cam models, in particular, should use text only for cybersex. |
| 23:55 | <jgraham> | Or should publishing anything that cannot be made accessible to everyone be non-conforming? |
| 23:55 | <Hixie> | webben: sorry to keep hammering on this, but you haven't replied -- what should the spec say the author should do in the case of a page with a webcam? |
| 23:56 | <webben> | jgraham: I'm still trying to work out what HTML5 conformance is for, so it's difficult for me to answer. I'm not sure the way lines are being drawn in the sand makes much sense to me. |
| 23:56 | <Hixie> | conformance defines what you're allowed to do |
| 23:56 | <Dashiva> | webben: Conformance is what we should strive for |
| 23:56 | <webben> | Hixie: "allowed" as far as the validator is concerned |
| 23:56 | <webben> | Hixie: not "allowed" in terms of a) what works or b) what makes money |
| 23:57 | <Hixie> | conformance is the html5's spec way of saying "this is what is ok, and this is what you are not allowed to do" |
| 23:57 | <jgraham> | webben: I think the problem is that many people see conformance as a poliical tool that they can use o leverage some ideology |
| 23:57 | <Dashiva> | If conformance doesn't allow for making money, it's a pretty useless concept :) |
| 23:57 | <jgraham> | webben: for example universal access |
| 23:57 | <Dashiva> | jgraham: Indeed. |
| 23:57 | <Dashiva> | It's been all but admitted that the WCAG is useless without a law to enforce it, so as much as possible is being piggybacked onto HTML5 |
| 23:58 | <jruderman> | i wouldn't want a law enforcing HTML5 conformance either... |
| 23:58 | <Hixie> | webben still hasn't actually replied to my question |
| 23:58 | <Hixie> | jruderman: i would! |
| 23:58 | <jgraham> | webben: The problem is that many of these ideologies are not practical or even, frankly, desirable |
| 23:58 | <webben> | jgraham: That doesn't really get at the nub of what conformance is for. It just says: conformance is not for universal access. |
| 23:59 | <Hixie> | jruderman: i could make the spec say "a page is conforming only if you have paid hixie $5"! :-D |
| 23:59 | <webben> | Hixie: I'm not sure why the syntax used is crucial to the discussion of what conformance is for. |
| 23:59 | <Dashiva> | Hixie: The google tax! |
