| 00:00 | <Dashiva> | I wish people would ooooh over type="email" too |
| 00:02 | <Dashiva> | Hopefully it will save us from the plague of broken regular expressions |
| 00:03 | Dashiva | wonders if there's a corrolary to Godwin's Law for cheating by naming Mussolini and fascists instead |
| 00:05 | <Philip`> | You are assuming browsers won't implement type="email" with broken regular expressions |
| 00:06 | <Hixie> | and that the servers won't continue to do the checking in a broken way |
| 00:09 | <Dashiva> | Philip`: Luckily I know regular expressions, so I could make my own build ;) |
| 00:10 | <Dashiva> | Hixie: I suspect a good deal of people in the group making broken regexpes are also in the group thinking clientside validation is sufficient |
| 00:11 | <blooberry> | dashiva: I immediately think of http://xkcd.com/208/ when I see that. ;-} |
| 00:12 | <Hixie> | Dashiva: :-/ |
| 00:13 | <Dashiva> | Email validation is such a joke anyway. If the domain and mailbox don't actually exist, it doesn't matter how well-formed the address is. |
| 00:14 | <Philip`> | But it can help detect unintentional typos |
| 00:15 | <Dashiva> | I can't think of any common cases a regexp would detect, apart from forgetting the @ |
| 00:17 | <Dashiva> | (which is incidentally the regexp I usually recommend, .+@.+) |
| 00:53 | <Hixie> | <input type=file accept> seems problematic |
| 00:55 | <Hixie> | there don't seem to be any use cases that limit the type without also limiting other things, like image size, framerate, audio codec, or whatever |
| 00:55 | <Hixie> | so it seems pointless to keep it |
| 00:56 | <Hixie> | the main use case was to enable on-the-fly video/audio/photo upload, but maybe we're better off providing direct access to the camera/mic |
| 00:57 | <shepazu> | that seems like that should be part of the file-upload spec |
| 00:57 | <Hixie> | yeah, i was about to say that it seems like something to punt to the binary data work |
| 00:57 | <shepazu> | which arun will be editing, it seems |
| 00:58 | <shepazu> | I do agree we need to allow (controlled) access to camera & mic |
| 00:58 | <Hixie> | yeah that's been requested a lot from cutting-edge app authors |
| 00:59 | <shepazu> | well, Flash has had it for many years |
| 00:59 | <Hixie> | yup |
| 00:59 | <shepazu> | and I'm sure Silverlight has it (or will) |
| 01:00 | <shepazu> | hmmm... wonder how that plays into uploading RAW files from a USB camera/memory card? or would that just be a normal file upload? |
| 01:00 | <Hixie> | dunno |
| 01:01 | <othermaciej> | Hixie: accept is more useful as a hint than a limit |
| 01:01 | <othermaciej> | Hixie: so the UA can show a specialized media browser for audio, video or image types for example |
| 01:01 | <othermaciej> | instead of generic file open dialog |
| 01:01 | <shepazu> | those files are huge... but photo-editing webapps would do well to allow them |
| 01:01 | <Hixie> | othermaciej: does anyone do that? or have any plans to? |
| 01:02 | <othermaciej> | Apple does not comment on future product releases |
| 01:03 | <othermaciej> | but I think it would be a swell idea, given that both Mac OS X and the iPhone have specific media picker UIs |
| 01:03 | <Hixie> | hypothetically, should someone want this to remain in the spec, would they want it to just be image|video|audio|mime-type ? or just image|audio|video ? or just mime-type? or something more elaborate? |
| 01:03 | <Hixie> | comma separated list, any mime supertype? parameters? |
| 01:03 | <othermaciej> | I don't know if any of the other base MIME types are particularly useful |
| 01:03 | <othermaciej> | text and application don't tell you much |
| 01:04 | <othermaciej> | I guess font is a root type though? |
| 01:04 | <Hixie> | multipart either |
| 01:04 | <Hixie> | application audio example image message model multipart text video |
| 01:04 | <othermaciej> | anyway the semantics I describe are not really exactly what would be implied by "accept" |
| 01:05 | <Hixie> | html4 has accept="mime/type" iirc |
| 01:06 | <Hixie> | wf2 has accept="mime/type,mime/type,mime/type" (comma-separated list) where any of the mime/type entries can be mime/* or */* as well |
| 01:06 | <othermaciej> | single instance of mime/* (or just the base part) is the use case I would be most interested in |
| 01:08 | <Hixie> | HTML4 has a comma-separated list |
| 01:08 | <Hixie> | so i propose making it a comma-separated list _or_ one of "image", "video", "audio" |
| 01:09 | <Hixie> | and the selected file(s) must either have the given types or be of the given category |
| 01:09 | <Hixie> | sound good? |
| 01:16 | <blooberry> | would there be any use in allowing multiple types at once for the latter? |
| 01:17 | <blooberry> | say, "image,audio"? |
| 01:22 | <Hixie> | what's the use case? |
| 01:35 | <roc> | shepazu: Silverlight doesn't have it |
| 01:37 | <othermaciej> | flickr might want an "image,video" uploader |
| 01:37 | <othermaciej> | (if this results in a media browser with only image and video categories, say) |
| 01:38 | <Hixie> | hm, true |
| 02:03 | <blooberry> | othermaciej: yeah, that was my thought. A restriction to "sighted media" would encompass both image and video. I can't think of an easy case for mixing audio with the others. |
| 02:03 | <othermaciej> | I could imagine a site where you could post audio or video |
| 02:04 | <othermaciej> | or any of audio, video or image but not other kinds of files |
| 02:04 | <blooberry> | hmm. |
| 02:04 | <blooberry> | itunes now is currently sellers->consumers. It does both audio and video. If in the future such a service allowed posting of same... |
| 05:49 | <MikeSmith> | Hixie: wondering if I might steal some of your demos for my presentation at Web Directions South later this week |
| 08:10 | <annevk2> | Work on an HTML5 parser in Perl: http://suika.fam.cx/www/markup/html/whatpm/readme |
| 08:34 | <Hixie> | MikeSmith: go for it |
| 08:38 | <ginger> | MikeSmith: I see you're speaking at Sydney this week - are you from AU? |
| 10:10 | <zcorpan_> | am i missing something with dir="rtl" and the table header algorithm? |
| 10:27 | <Hixie> | zcorpan_: like what? |
| 10:30 | <zcorpan_> | Hixie: "HTML 5's proposed smart scope algorithm pretty much does this, except it overlooks dir="rtl"." -- http://www.w3.org/mid/e2a28a920809230152p707fb0ccoc4edc9c3947d9bf5⊙mgc |
| 10:31 | <zcorpan_> | i thought it would work exactly the same for rtl tables so i'm not sure what the algorithm is overlooking |
| 10:32 | <Hixie> | dunno |
| 10:32 | <Hixie> | maybe i forgot to swap it around for rtl? |
| 10:32 | <Hixie> | though i thought i handled it |
| 10:33 | <zcorpan_> | rtl tables are just mirrored in the rendering |
| 10:33 | <Hixie> | yeah the algorithms looks to be direction agnostic |
| 10:34 | <Hixie> | algorithm. or look. one of the two. |
| 10:34 | <Hixie> | oh well |
| 11:23 | <zcorpan_> | why is window[0].document undefined in webkit? |
| 11:24 | <zcorpan_> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3Ciframe%20src%3Ddata%3Atext%2Fhtml%2Cx%20onload%3Dw(window%5B0%5D.document)%3E |
| 11:28 | <zcorpan_> | is there a difference between importNode and cloneNode? |
| 11:28 | <zcorpan_> | in browsers i.e. |
| 11:31 | <zcorpan_> | i don't understand why people have been working so hard on replacing html with something "saner" |
| 11:31 | <zcorpan_> | when no-one has pushed for browsers do replace the dom |
| 11:39 | <zcorpan_> | seems the difference is only in setting ownerDocument |
| 11:56 | <zcorpan_> | hmm wonder how i should define cloning |
| 12:06 | <hsivonen> | zcorpan_: the DOM and HTML are not part of the W3C's Techology Stack... |
| 12:07 | <hsivonen> | I started writing a blog post about the technology stack yesterday but didn't have enough hours in a day to finish |
| 12:12 | <zcorpan_> | hsivonen: i don't follow but i guess i'll read your blog post WIR |
| 12:14 | <hsivonen> | zcorpan_: I mean: the W3C has not only dropped HTML but has also dropped the DOM from its powerpointware |
| 12:15 | <hsivonen> | so presumably the DOM doesn't count as "saner", either |
| 12:16 | <zcorpan_> | hsivonen: but does the w3c have a replacement for the dom? |
| 12:16 | <hsivonen> | zcorpan_: the SVG WG has a uDOM, but other than that, not that I'm aware |
| 12:19 | <Philip`> | Just use XPath and declarative expressions |
| 12:20 | <hsivonen> | oops. I take that back. DOM is in the stack but not where I expected to see it. Sorry. |
| 13:22 | <zcorpan_> | gsnedders: i need automatic xrefs to html5 for web dom core :) |
| 13:28 | <zcorpan_> | ok i've specced importNode and cloneNode, except i've not covered attributes yes http://simon.html5.org/specs/web-dom-core |
| 13:32 | <zcorpan_> | s/yes/yet/ |
| 14:31 | <hsivonen> | http://www.alistapart.com/articles/webstandards2008 |
| 14:58 | <csarven> | Ahh, is there a way to use unicode values in input[type=value] ? |
| 14:59 | <hsivonen> | csarven: yes. UTF-8 of backslash escapes |
| 14:59 | <csarven> | Doh! Thanks. |
| 15:00 | <hsivonen> | s/of/or/ |
| 15:01 | <csarven> | hsivonen Sorry, my mistake, I've meant the @value for input type submit. |
| 15:02 | <hsivonen> | csarven: in HTML? |
| 15:02 | <csarven> | Yes. |
| 15:02 | <hsivonen> | csarven: doesn't Unicode Just Work there? |
| 15:03 | <zcorpan_> | it should... |
| 15:03 | <csarven> | Trying to show the heart character ? ♥ |
| 15:03 | <zcorpan_> | wfm in opera |
| 15:03 | <hsivonen> | csarven: might be a font problem |
| 15:04 | <csarven> | Well the character does display if I use the symbol directly as opposed to the char value. |
| 15:05 | <hsivonen> | csarven: which browser and OS? |
| 15:05 | <hsivonen> | WFM in Firefox 3 and Opera 9.5 on Leopard |
| 15:05 | <csarven> | Firefox/Vista. I must be doing this wrong then. |
| 15:08 | <csarven> | Do I have to use the octal value? |
| 15:11 | <Philip`> | You can't use octal - you should use decimal (♥) or hex (♥) |
| 15:12 | <csarven> | I swear that is not working for me. |
| 15:12 | <Philip`> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cinput%20type%3D%22submit%22%20value%3D%22%26%239829%3B%20%26%23x2665%3B%22%3E%0D%0A |
| 15:13 | <Philip`> | shows me a hearty button |
| 15:13 | <csarven> | I'm guessing that Firebug is fooling me. |
| 15:16 | <zcorpan_> | csarven: are you testing with a data url? |
| 15:19 | <csarven> | zcorpan_ No. |
| 15:31 | <hsivonen> | sorry about html5.validator.nu being down... |
| 15:31 | <hsivonen> | fixing it... |
| 17:10 | Philip` | wonders how you could possibly do feature-sniffing of something like APNG, e.g. if you wanted to emulate it with scripts in less able browsers but didn't want to hardcode a list of UA strings |
| 17:11 | <annevk2> | with setTimeout and <canvas> or does it always draw the first frame? |
| 17:11 | <Dashiva> | You could use a APNG wih the dead frame showing "Click here to win a million dollars" and have onclick to do script animation |
| 17:11 | <Philip`> | <canvas> is defined (though not implemented) to always draw the poster frame or first frame |
| 17:12 | <hsivonen> | Philip`: spec bug |
| 17:12 | <annevk2> | can the first frame differ depending on APNG support? |
| 17:12 | <Philip`> | ("When the drawImage() method is passed, as its image argument, an animated image, the poster frame of the animation, or the first frame of the animation if there is no poster frame, must be used.") |
| 17:14 | <Philip`> | annevk2: The first frame can, but the poster-frame-or-else-first-frame can't |
| 17:14 | <Philip`> | (If an APNG has a poster frame, then that's what non-APNG implementations will show; otherwise the APNG only has a first frame, which is what non-APNG implementations will show) |
| 17:16 | <Philip`> | Opera doesn't appear to even bother animating images unless they're visible on the screen, so it'd be kind of hard to do any reliable timer-based testing anyway |
| 17:18 | <Philip`> | (See e.g. http://philip.html5.org/tests/apng/tests.html in Opera, like how the "dispose ops" images change colour briefly if you scroll them off then back on the screen, at least in my copies of Opera) |
| 18:25 | <csarven> | What's more accessible: ♥ or <img> with @alt. |
| 19:38 | <takkaria> | I fear public-html |
| 19:44 | <gsnedders> | csarven: Probably <img> with @alt. |
| 19:52 | <webben> | csarven: I suspect gsnedders right (at least in terms of user-agent support - e.g. encodings, fonts, screen readers pronouncing things right) |
| 19:52 | <webben> | csarven: e.g. JAWS confuses mathematical comparison symbols. |
| 19:55 | <csarven> | gsnedders webben Thanks. |
| 19:55 | <gsnedders> | (I don't claim to be an accessibility expert, so maybe I'm right) |
| 19:56 | <csarven> | They are both fairly close IMO. |
| 19:56 | <Dashiva> | Having a list of symbol names is pretty simple, so it seems likely it wouldn't be supported :P |
| 19:59 | <hsivonen> | Dashiva: IIRC, VoiceOver (in English) can fall back onto the unicode character name |
| 19:59 | <webben> | I think all pop. screen readers allow users to customize their dictionaries too |
| 20:00 | <webben> | of course, they'd need to know to do that in advance. |
| 20:01 | <webben> | Perhaps worth thinking about whether the unicode char name is what you really _mean_ however. e.g. it might be "black heart". "I black heart HTML5" for example doesn't necessarily convey the sense well. |
| 20:01 | webben | checks |
| 20:02 | <webben> | apparently the name is "BLACK HEART SUIT" |
| 20:06 | <Dashiva> | So you want text equivalents for text |
| 20:07 | <hdh> | just like smileys, abbr[title] |
| 20:09 | <webben> | Dashiva: is it text? |
| 20:09 | <webben> | (seems a bit inbetweeny) |
| 20:12 | <Dashiva> | Well, is a metaphor text? |
| 20:16 | <webben> | Not necessarily. |
| 20:16 | <webben> | You can speak a metaphor, for starters :) |
| 20:17 | <webben> | I'm not sure I (heart symbol) X is a strictly a metaphor anyways. |
| 20:57 | <Philip`> | It seems less useful to ponder whether ♥ is text, than to ponder concrete use cases like how a non-visual UA could interpret that character |
| 20:59 | <gsnedders> | e.g., I <abbr title="love">♥</abbr> Philip`. |
| 20:59 | <gsnedders> | As opposed to, I picked up a <abbr title="suite of hearts">♥</abbr> card. |
| 20:59 | <Philip`> | It'd be nice if AI could work out whether it's an abbreviation for the word "love" or is representing a card suit or is part of a manual for medical students or whatever |
| 20:59 | <gsnedders> | How about >3? |
| 21:00 | <hsivonen> | ☙ |
| 21:00 | <gsnedders> | s/g/l/ |
| 21:01 | <Philip`> | I wonder how well you could do by just determining whether ♥ is being used as a verb, in which case it probably means "love" |
| 21:01 | <hsivonen> | what about ❤ |
| 22:28 | <Hixie> | the criticism of whatwg in http://www.alistapart.com/articles/webstandards2008 misses the mark by a wide margin |
| 22:28 | <Hixie> | there are plenty of things to criticise the whatwg for |
| 22:28 | <Hixie> | those aren't it :-) |
| 22:29 | <Hixie> | e.g. we're anonymous?? |
| 22:29 | <Hixie> | and lack of leadership?? i thought the whole problem was that my being a dictator. |
| 22:29 | <Hixie> | s/that// |
| 22:29 | <Hixie> | and "No economic support" is a con? |
| 22:32 | <Dashiva> | I like the repeated use of "Nothing demonstrates this more than..." |
| 22:40 | <Hixie> | the second one of those is wrong too (webkit is always careful to prefix their extensions) |
| 22:40 | <Hixie> | and "Nothing demonstrates this more than Mozilla’s and Opera’s inability to grow a user base beyond a certain point." is totaly wrong |
| 22:41 | <Hixie> | first of all firefox market share hasn't peaked |
| 22:41 | <Hixie> | and secondly, opera's market share being flat is clearly not anything to do with proprietary stuff since firefox's market share is bigger than opera's |
| 22:45 | <Hixie> | wow |
| 22:46 | <Hixie> | the entirety of the "Is "breaking the Web" with HTML 5 a non issue" thread got caught in my "AAA-Productivity" folder |
| 22:46 | <Hixie> | my filters work remarkably well |
| 22:47 | <roc> | I think I'd better not read this or I will be unnecessarily enraged |
| 22:48 | <roc> | why is it so hard to resist the temptation |
| 22:48 | <Dashiva> | Rage, rage against the thread of the email |
| 22:55 | <gavin__> | Hixie: I think you missed the "rather than" in front of "anonymous" |
| 22:58 | <Hixie> | ah, yes, good call |
| 22:58 | <Hixie> | ok then i'll complain that "not enough time for research" is ridiculous given that we do more research and have more time in our timetable than any w3c project to date |