| 01:38 | <webben> | Lachy: AT actually supports AXIS and authors actually use it. Just one example of how it can be used is to label rowgroup headers (http://uk.biz.yahoo.com/16/sector.html). |
| 01:38 | <webben> | Lachy: You might dispute whether it's very useful or not, but it's an exaggeration to say it's "totally useless for everyone". |
| 01:40 | <webben> | (It's also worth noting that axis might be rather handy if microformat parsing ever goes beyond nested classes (something that doesn't translate to HTML data tables that well anyhow.) |
| 02:07 | <theunscene> | I don't trust someone name after a member of 98 degrees anyways |
| 02:08 | <theunscene> | named* |
| 02:31 | <Philip`> | I wish cssutils didn't use an O(n^2) tokeniser |
| 02:32 | <Philip`> | (Removing characters from the head of a list in Python is an officially Bad Idea) |
| 07:15 | <othermaciej> | Philip`: are Python lists secretly arrays? |
| 07:33 | hsivonen | notes WordPress inserts <em> when I click the button labeled "i" |
| 07:38 | <hsivonen> | I tried to post to the whatwg blog but got error 500 upon saving my post |
| 07:40 | <Lachy> | hsivonen, try again or wait till Hixie gets back and fixes it. Hixie's server is having trouble. Sometimes it works, sometimes it doesn't. |
| 07:44 | marcosc | notes that WordPress' HTML editor is the work of the devil... |
| 07:45 | <Lachy> | marcosc, indeed! that's why I disable it |
| 07:46 | <hsivonen> | Lachy: ok |
| 07:46 | marcosc | goes searching for ways to turn off that darn editor... |
| 07:47 | <hsivonen> | I tried to post a short response to Jeremy Keith's post pointing out that the design of HTML5 hasn't been about 80% of users but about 80% of authoring cases |
| 07:47 | <Lachy> | oh good, I was going to respond to that eventually too. now I don't have to :-) |
| 07:49 | <Lachy> | I'll write a custom error document for the blog and wiki, pointing out that it's a known issue, since people keep asking about it in here. |
| 07:57 | <Lachy> | that's weird. The ErrorDocument directive isn't working in .htaccess on the blog |
| 07:59 | <hsivonen> | AllowOverride at work, I guess |
| 08:07 | <Lachy> | yeah, that's probably it. There's a few other directives that aren't working either |
| 08:17 | <Lachy> | I wonder how often people are going to keep suggesting IE's broken, pseudo-namespace syntax as a solution to various problems? |
| 08:21 | <virtuelv> | <q>The problem, as I see it, is that img is an empty element. Had it not been, both longdesc and alt could be replaced with inline content, as advanced as anyone could ever grok their HTML to be. </q> |
| 08:22 | <virtuelv> | I don't really think I've ever seen proper use of <object> fallback, except in tutorials explaining it |
| 08:26 | <hsivonen> | virtuelv: how do you know? if it worked perfectly, you wouldn't notice. |
| 08:26 | <virtuelv> | hsivonen: hence, the "think" |
| 08:28 | <jacobolus> | hmm, blog.whatwg.org is giving back “500 Internal Server Error” |
| 08:28 | <virtuelv> | this is entirely "gut feeling"-based. I'd like to see some evidence that it's used outside the standards community |
| 08:28 | <Lachy> | jacobolus, we know, we can't fix it yet |
| 08:28 | <jacobolus> | okay, fair enough :) |
| 08:30 | <hsivonen> | 500: hosting broken in an unforeseen way |
| 08:30 | <hsivonen> | 503: hosting broken in a foreseeable way |
| 08:31 | <Lachy> | it's interesting how some people claim that <img>fallback</img> would solve problems if it were possible, whereas some others claim that <object>fallback</object> doesn't even work like that |
| 08:31 | <Lachy> | (although I hope object fallback can be fixed) |
| 08:31 | <hsivonen> | fallback != alternative content |
| 08:32 | <hsivonen> | in the case of video, at least |
| 08:33 | <Lachy> | yeah, that's the problem with object. Gez Lemon ponited out on JuicyStudio recently that ATs don't read the fallback content if the object is rendered |
| 08:33 | <hsivonen> | makes sense if the rendered content is AT-enabled like Flash on Windows |
| 08:34 | <Lachy> | sometimes that's the desired behaviour (like embedded Flash or other multimedia with built in accessability), but other times it's not (like embedded images) |
| 08:34 | hsivonen | wonders if anything is happening at Adobe when it comes to Gnome & OS X accessibility of Flash |
| 08:35 | <hsivonen> | Lachy: one could use this as another argument of why <object> is too multipurpose |
| 08:35 | <Lachy> | indeed |
| 08:36 | <hsivonen> | as for <video>, the accessibility story for the deaf is that the video file itself should have captions |
| 08:37 | <Lachy> | yeah, that's why <embed> is good for plugins, because they should ideally be natively accessible |
| 08:37 | <hsivonen> | but I'm not sure what the accessibility story for the blind is considering that fallback is for legacy UAs |
| 08:37 | <Lachy> | audio descriptions are a possibility |
| 08:38 | <hsivonen> | Lachy: do Ogg and MP4 support flagging additional sound tracks as audio descriptions? |
| 08:39 | <Lachy> | sometimes, the normal audio may be sufficient if the visuals aren't that important. such as an interview or speech that isn't accompanied by slides |
| 08:39 | <hsivonen> | I wonder if Apple has already fixed the MP4 captioning perf bug markp found |
| 08:40 | <Lachy> | I don't know about flagging audio tracks |
| 08:40 | <Lachy> | (not really our problem though, we can't fix every format) |
| 08:43 | <hsivonen> | well, it is a bit disingenious to suggest that authors use the functionality of the video format if Mozilla/Opera baseline (Ogg) and the Apple/Nokia baseline (MP4) lack support |
| 08:43 | <hsivonen> | although I agree as far as theory goes |
| 08:53 | <jacobolus> | Lachy: so will the html5 spec itself include suggested best practices for image accessibility, etc? or is that the purview of some separate document? |
| 08:54 | <Lachy> | the HTML5 spec already includes some good accessibility suggestions for alt text and it will be improved where necessary |
| 08:54 | <Lachy> | although the spec isn't a tutorial and so there's a limit to how far it will go |
| 08:55 | <jacobolus> | rather, is there anything to be done to quiet the longdesc flamewar? |
| 08:55 | <hsivonen> | the accessibility advice in the spec draft is for HTML5-compliant UAs, though |
| 08:55 | <jacobolus> | or will that just die down by itself? |
| 08:55 | <Lachy> | sure, find a solution that works. longdesc="" doesn't. |
| 08:55 | <hsivonen> | not for current JAWS |
| 08:56 | <Lachy> | I think the best solution is to either include the alternative content within the same page or link to it with an ordinary link, possibly with the addition of rel=longdesc |
| 08:57 | <jacobolus> | css can presumably hide such links if it's undesirable for sighted people to see them |
| 08:58 | <Lachy> | sure, though the long description may include information that is useful for more than just the blind |
| 08:58 | <hsivonen> | Lachy: I think the WG needs to find out if mere juxtaposition is enough in practice for blind users to associate an image indicator and a link to a description |
| 08:58 | <hsivonen> | Lachy: or if explicit association is needed |
| 08:58 | <jacobolus> | so why is there such opposition to scrapping longdesc? |
| 08:59 | <Lachy> | <a href=desc.html rel=longdesc><img src=image alt=...></a> |
| 08:59 | <jacobolus> | if neither authors nor users ever look at it? |
| 08:59 | <Lachy> | because it works in theory, just not in practice |
| 09:00 | <hsivonen> | Lachy: well, yeah, if it turns out that the images that require long descriptions are extremely used as links to somewhere else |
| 09:00 | <Lachy> | some people care more about the theory than the practice, and cling to the hope that longdesc="" will start to improve. However, there's no sign of that actually happening. |
| 09:01 | <hsivonen> | Lachy: moreover, sighted users tend to assume that, by convention, a non-iconic photo is a link to a larger version if hover shows the hand cursor |
| 09:01 | <hsivonen> | longdesc has been in public drafts/specs for over a decade |
| 09:01 | <Lachy> | you mean like in wikipedia? |
| 09:02 | <hsivonen> | Lachy: like almost anything, actually |
| 09:02 | <hsivonen> | Lachy: when iconic images are links, |
| 09:02 | <Lachy> | that works. just put the long description on the same page as the larger version and the thumbnail model still works |
| 09:02 | <hsivonen> | they take you to some place represented by the icon |
| 09:02 | <hsivonen> | when photos are links, they almost always take you to a larger version of the photo |
| 09:03 | <Lachy> | iconic images (as opposed to thumbnails) used as links rarely require long descriptions anyway |
| 09:03 | <hsivonen> | Lachy: yes. |
| 09:03 | <hsivonen> | Lachy: yeah, putting the large file and the description at the same URI would avoid breaking the pattern |
| 09:04 | <hsivonen> | in theory, image files themselves should contain the long description |
| 09:04 | <hsivonen> | of course, in practice, they won't |
| 09:05 | <Lachy> | maybe, though the content of the long description sometimes depends on the context in which the image is used |
| 09:12 | <hsivonen> | "The Flash Player itself does not care about file extensions, you can feed it .txt files for all it matters. The Flash Player always looks inside the file to determine what type of file it is." |
| 09:12 | <hsivonen> | http://www.kaourantin.net/2007/08/what-just-happened-to-video-on-web_20.html |
| 09:16 | <hsivonen> | whoa! Flash Player will support parts of MPEG-4 part 10 without supporting part 2. that's radical |
| 09:29 | <jacobolus> | hsivonen: well it looks like they were able to license a 100kb h.264 decoder, and wanted to be able to support all the h.264 m4v files people are now cranking out… doesn't seem too surprising that they'd skip mpeg-4 part 2. that would just add complexity for an unclear benefit |
| 09:30 | <othermaciej> | ok I'm starting to think Robert Burns is from a parallel universe where black is white and up is down |
| 09:31 | <jacobolus> | othermaciej: swapping black and white is part of the accessibility options of OS X |
| 09:31 | <othermaciej> | Lachy: in MP4 and many other audio formats it is possible to have an additional descriptive audio track |
| 09:31 | <othermaciej> | er |
| 09:32 | <othermaciej> | I meant that for hsivonen |
| 09:32 | <othermaciej> | it works mostly the same way as multi-language soundtracks |
| 09:32 | <othermaciej> | however even without descriptive audio tracks, just hearing the audio of a video with dialog is probably better than whatever fallback markup might be provided |
| 09:34 | <hsivonen> | othermaciej: any pointers for docs? I tried to google, but the results are full of noise. |
| 09:34 | <hsivonen> | s/for/to/ |
| 09:35 | <othermaciej> | hsivonen: I don't know offhand, but I can ask the QuickTime folks |
| 10:29 | <othermaciej> | http://www.programimi.com/2007/09/14/55-reasons-to-design-in-xhtml-css/ |
| 10:34 | <Whiskey_M> | 'lo |
| 10:35 | <Whiskey_M> | henry in regards to your e-mail to the list have you looked at SMIL - http://www.w3.org/TR/REC-smil/ |
| 10:36 | <hsivonen> | Whiskey_M: SMIL would add a huge layer of complexity compared to using the simpler features of video formats themselves |
| 10:37 | <hsivonen> | Whiskey_M: also, as far as closed captioning goes, it would break stuff across files instead of making captions always travel with the video |
| 10:40 | <hsivonen> | Whiskey_M: the latest SMIL is much larger: http://www.w3.org/TR/SMIL/ |
| 10:43 | <Whiskey_M> | cool - if it's been considered then that's fine. Nothing you've suggested as far as I can see would stop anyone who wants to also providing SMIL for their videos and by providing a halfway house route to improving accessibility that is easier to attain it can only improve matters |
| 10:48 | <Lachy> | hsivonen, did you mean to say "The spec already says that the fallback is not for old browsers."? did you mean s/is not/is/ ? |
| 10:49 | <hsivonen> | Lachy: oops |
| 10:50 | <hsivonen> | this is what happens when I edit a sentence instead of writing it from start to finish in one go. :-( |
| 10:54 | <hsivonen> | it is hard to proof-read one's own text |
| 10:54 | <hsivonen> | when you already know what it is supposed to say and think it says that |
| 10:56 | <hsivonen> | hmm. a phpbb exploit scanner is scanning validator.nu |
| 10:56 | <hsivonen> | I guess that means phpbb is popular enough for such scan to be worthwhile |
| 10:59 | <Lachy> | which forum software is forums.whatwg.org running? I wonder if it's been kept up to date. |
| 11:00 | <hsivonen> | Lachy: phpbb I think |
| 11:01 | <virtuelv> | has any thought gone in to providing external captioning files? |
| 11:02 | <virtuelv> | so that captioning is made independent of the video stream itself |
| 11:02 | <Lachy> | there are external caption file formats |
| 11:02 | <virtuelv> | Lachy: I seem to remember markp blogging about this |
| 11:02 | <Lachy> | it may be useful to provide a way to link to them from the HTML |
| 11:03 | <virtuelv> | http://diveintomark.org/archives/2007/06/26/piracy-lessons |
| 11:04 | <Lachy> | this is one such format http://en.wikipedia.org/wiki/SubRip |
| 11:04 | <Lachy> | oops, wrong link |
| 11:05 | <hsivonen> | virtuelv: I guess the crucial question is: will external files be better for users than muxed captions? |
| 11:06 | <virtuelv> | hsivonen: potentially, since you can probably make use of browser settings to get subtitles and captioning based on what language the UA is set to prefer |
| 11:06 | <virtuelv> | and it makes the subtitling separate from the data stream itself |
| 11:06 | <hsivonen> | virtuelv: do you mean content negotiation on the HTTP level? |
| 11:06 | <virtuelv> | hsivonen: yes |
| 11:06 | hsivonen | has severe doubts about conneg |
| 11:07 | <Lachy> | oh, that is the right link (the first paragraph of that wikipedia page is confusing) |
| 11:07 | <hsivonen> | conneg doesn't tell you what else was available |
| 11:07 | <virtuelv> | hsivonen: I have doubts as well, but there are other ways around this, using scripting |
| 11:07 | <hsivonen> | and conneg-based approaches assume that the site developer is able and competent to configure the HTTP layer |
| 11:10 | <hsivonen> | anyway, I'd expect a dedicated external timed text format to have better chances of implementation than SMIL |
| 11:12 | <hsivonen> | "Unfortunately there is no exact specification of the .SRT file format." -- Wikipedia |
| 11:12 | <hsivonen> | is "Unfortunately" NPOV? :-) |
| 11:15 | <hsivonen> | still about conneg: |
| 11:16 | <hsivonen> | virtuelv: with captioning for accessibility, the captions are expected to be in the same language as the soundtrack |
| 11:16 | <hsivonen> | virtuelv: when you start doing translations as subtitles, it isn't accessibility anymore |
| 11:16 | <hsivonen> | virtuelv: and for subtitles, automating the choice becomes much more difficult |
| 11:17 | <hsivonen> | virtuelv: with captions, if one is deaf, a permanent "Enable closed captions" pref makes sense |
| 11:17 | <hsivonen> | virtuelv: with subtitles, the choice of subtitles depends not only on the language of the subtitles but also on the language of the audio content |
| 11:19 | <hsivonen> | my personal subtitle preferences are complex enough that I wouldn't think it'd be reasonable for UAs to model them and try to relieve me of choosing manually |
| 11:22 | <hsivonen> | oh, btw, if one applies the Design Principles, one could argue that translations as subtitles as well as alternative language sound tracks are out of scope |
| 11:32 | <Lachy> | I've been asked to write an article for a popular online magazine, called "Developing with HTML5" (somewhat based on the presentation I did in August)... |
| 11:32 | <Lachy> | I'm trying to decide which topics to discuss. |
| 11:33 | <Lachy> | So far, I've written about choosing HTML or XHTML and using both together, and about the sectioning elements and heading algorithm |
| 11:33 | <Lachy> | I want to include one more topic and would like some feedback about what to choose: |
| 11:34 | <Lachy> | 1. new semantics: time/meter/progress |
| 11:34 | <Lachy> | 2. Embedded content and multimedia (video, audio and canvas) |
| 11:34 | <Lachy> | 3. Form controls |
| 11:34 | <Lachy> | any preferences? |
| 11:35 | <hsivonen> | Lachy: does the title imply developing today? or is it about developing in the future? |
| 11:37 | <hsivonen> | Lachy: form controls and video are perhaps the most interesting things to your audience |
| 11:37 | <Lachy> | a little of both. I'd prefer to avoid things that won't be available for a long time, such as <datagrid>, but also want to give some insight into the future |
| 11:37 | <hsivonen> | Lachy: you could use <progress> as an example of built-in accessibility semantics |
| 11:38 | <Lachy> | yeah, same with time and meter |
| 11:38 | <hsivonen> | Lachy: also, the good thing about form controls and video is that they seem to be getting attention from browser vendors and might even get deployed sooner than later |
| 11:39 | <Lachy> | I could do video and audio. There's already so much about canvas and it's already widely deployed |
| 11:39 | <zcorpan> | Lachy: will you only cover front-end development? |
| 11:39 | <Lachy> | (I'm also not an expert on canvas, so I'm probably not the best person to write about it) |
| 11:40 | <Lachy> | mostly front end, but the HTML/XHTML bit talks about using XHTML on the back end and serialsing as HTML |
| 11:40 | <zcorpan> | ok |
| 11:40 | <hsivonen> | Lachy: of the HTML5 features that weren't in HTML 4.01, canvas probably has the best market penetration |
| 11:41 | <hsivonen> | Lachy: to the point that it doesn't feel like a future feature anymore |
| 11:41 | <Lachy> | yeah, that's basically what I meant by what I said above |
| 11:42 | <Lachy> | I suppose further discussion of forms should really wait till the we get results from the Forms TF |
| 11:44 | <Lachy> | alright, I'll do video and audio |
| 11:50 | <Lachy> | if I do video, I should probably create a demo page with a sample video. |
| 11:50 | <Lachy> | I wonder if I should create a new one or just reuse one of my existing videos http://lachy.id.au/lib/media/2007/ |
| 11:56 | <virtuelv> | hixie's site is down? |
| 11:56 | <Lachy> | virtuelv, yes |
| 11:57 | <Lachy> | sometimes. It's having the same problems as the whatwg blog, wiki and forum |
| 11:57 | <Lachy> | (they're all on the same server) |
| 11:59 | <virtuelv> | Hm. And the server is not crumbling under the weight of the traffic? |
| 12:15 | <othermaciej> | hsivonen: Apple has a proposal for how to have multiple media sources w/ different accessibility features, that integrates with CSS media queries |
| 12:15 | <othermaciej> | (since other likely reasons for choosing one of several videos include screen size and bandwidth) |
| 12:16 | <othermaciej> | hsivonen: I need to get that sent out |
| 12:17 | <othermaciej> | hsivonen: this will make it possible to either use a video with a separate caption track, or if preferable a whole separate video with burned-in captions |
| 12:17 | <othermaciej> | and so forth |
| 12:22 | <hsivonen> | othermaciej: cool. |
| 13:21 | <Philip`> | othermaciej: It seems that lists are arrays or something similar - http://docs.python.org/lib/deque-objects.html says "list objects ... are optimized for fast fixed-length operations and incur O(n) memory movement costs for "pop(0)" and "insert(0, v)" operations which change both the size and position of the underlying data representation." |
| 13:22 | <othermaciej> | I guess that would only be confusing to someone who would expect a list to be a linked list |
| 13:28 | <Philip`> | Someone who expected that would be confused as soon as they accessed items[1000000] and got a quick answer |
| 15:18 | <othermaciej> | does HTML 4.01 have any requirements for the handling of unknown elements? |
| 15:18 | <othermaciej> | I thought it did, but I can't find it in the spec |
| 15:19 | <zcorpan> | not normatively |
| 15:19 | <zcorpan> | http://www.w3.org/TR/html401/appendix/notes.html#h-B.1 |
| 15:20 | <zcorpan> | (though close to nothing is normative in html4, but anyway) |
| 15:30 | <zcorpan> | "Since user agents may vary in how they handle error conditions, authors and users must not rely on specific error recovery behavior." -- so users are non-conforming by just browsing the web! (or was that sentence not normative?) |
| 15:49 | <hsivonen> | whoa! Rob characterizes my email as flame bait. |
| 15:50 | <Philip`> | Maybe that's because you were discussing accessibility, and accessibility discussions inevitably draw flames |
| 15:53 | <hsivonen> | well, it was more like a point came up on IRC so better post it on list so it gets in Hixie's IMAP box instead of getting lost |
| 16:04 | <othermaciej> | he's such a timesink |
| 16:19 | <zcorpan> | Philip`: is there a way to browse http://canvex.lazyilluminati.com/ other than using google? |
| 16:26 | <Philip`> | zcorpan: No - I always use Google to find things on it too :-p |
| 16:26 | <zcorpan> | Philip`: heh, ok |
| 16:26 | Philip` | probably should do some kind of organisation at some point |
| 16:27 | <Philip`> | but I always end up dumping everything into "misc" |
| 16:27 | <zcorpan> | could you enable indexes on that directory, maybe? |
| 16:30 | Philip` | checks for any embarrassing files in that directory |
| 16:31 | zcorpan | finds http://canvex.lazyilluminati.com/misc/htmldiffs.txt -- useful :) |
| 16:32 | <Philip`> | That's a tiny bit wrong - it misses at least one HTML5 attribute |
| 16:33 | <zcorpan> | which one? |
| 16:34 | <Philip`> | Indexes enabled now |
| 16:34 | <zcorpan> | thanks |
| 16:36 | <zcorpan> | font: Attributes only in HTML5 and HTML4 Transitional (deprecated by HTML4, restored by HTML5): color face size style |
| 16:37 | <zcorpan> | seems incorrect |
| 16:37 | <Philip`> | li@value is the only problem I can find |
| 16:37 | <Philip`> | Hmm |
| 16:38 | <Philip`> | Oh |
| 16:38 | <Philip`> | </dd><dt>Element-specific attributes:</dt> |
| 16:38 | <Philip`> | <!-- |
| 16:38 | <Philip`> | <dd><code title="attr-font-color">color</code></dd> |
| 16:38 | <Philip`> | <dd><code title="attr-font-face">face</code></dd> |
| 16:38 | <Philip`> | <dd><code title="attr-font-size">size</code></dd>--> |
| 16:38 | <Philip`> | My HTML5 spec parser didn't understand comments |
| 16:39 | <zcorpan> | hmm, <form onreset> is gone in html5? |
| 16:41 | <Philip`> | I didn't count anything that was defined in WF2 |
| 16:42 | <zcorpan> | perhaps it should be a global attribute like the rest of event handler attributes |
| 16:43 | <zcorpan> | reset bubbles so that could be useful |
| 16:43 | <Philip`> | (This was mostly meant for checking the contents of anne's differences document, so there are various areas where human judgement is needed to decide whether something's a real difference) |
| 16:46 | Philip` | is away for a while |