| 00:06 | <jruderman> | what is the difference between 'error' and 'failure' on http://files.myopera.com/blooberry/alexa/alexavalidation.htm |
| 00:09 | <Philip`> | I wish file.myopera.com would let me actually see the files, when I have Opera's referrer blocking switched on |
| 00:10 | <jgraham> | jruderman: It looks like errors might be fatal problems e.g. due to character encoding issues |
| 00:10 | <Philip`> | jruderman: Looks like 'failure' means like http://validator.w3.org/check?uri=http%3A%2F%2Fwww.mop.com |
| 00:11 | <jruderman> | ahh, that's pretty sad |
| 00:11 | <jgraham> | html5.validator.nu seems to cope fine :) |
| 00:13 | <Hixie> | did hsivonen ever release the numbers from his study? |
| 00:13 | <Hixie> | i forget |
| 00:13 | <Hixie> | i think he did |
| 00:31 | <gsnedders> | Hixie: yes |
| 00:32 | <Hixie> | do you remember where? |
| 00:32 | <gsnedders> | Nah |
| 00:32 | <gsnedders> | That would be useful. |
| 00:32 | <gsnedders> | public-html, I _think_. |
| 00:33 | <Philip`> | That's a little vague |
| 00:34 | Philip` | fails to remember whether he has or hasn't seen such numbers released |
| 00:34 | <Hixie> | http://lists.w3.org/Archives/Public/public-html/2008Aug/0923.html |
| 00:35 | <Hixie> | 0.87% valid pages |
| 00:35 | <Hixie> | that's what i was looking for |
| 00:36 | <Hixie> | (that's treating transitional stuff as invalid) |
| 00:46 | <gsnedders> | Now, I want sane IDs |
| 00:51 | <Hixie> | the validation errors are stopping me from migrating right now |
| 00:59 | <Hixie> | some reporter just e-mailed me a question about html5 |
| 01:00 | <Hixie> | which i replied to in detail, with links to various pages with more background information |
| 01:00 | <Hixie> | he replied back with "FWIW, I am hurrying a story and don't have time to read background materials. I want to know if this is a W3C technology." |
| 01:00 | <Hixie> | to which i felt like replying "well fuck you too" |
| 01:01 | <othermaciej> | Hixie: you need a PR agency |
| 01:01 | <Hixie> | i have one |
| 01:01 | <Hixie> | i'm sure google would be happy to do the PR for me if i let them |
| 01:10 | <Hixie> | same reporter just asked me if there was some sort of editor's draft released today |
| 01:10 | <Hixie> | i explained that the there have been eight so far today |
| 01:10 | <Hixie> | i think whatever article ends up coming out of this is going to be very confused |
| 02:06 | Hixie | ponders whether to bother keeping inputmode="" |
| 02:06 | <Hixie> | xforms is apparently not a stable reference point |
| 02:06 | <Hixie> | given that it changed since wf2 came out |
| 02:06 | <Hixie> | so much for persistent urls at the w3.org/TR/ space |
| 02:08 | <Hixie> | othermaciej: has there been any desire for inputmode in iPhone Safari? |
| 02:08 | <othermaciej> | what's inputmode? |
| 02:10 | <billyjack> | othermaciej: a means to specify what input mode to put the IME into when the user is in a particular input field |
| 02:10 | <billyjack> | e.g., kanji input, or numeric input |
| 02:10 | <othermaciej> | I see |
| 02:10 | <othermaciej> | I'll ask |
| 02:10 | <billyjack> | seems like it's really only useful on handsets that don't have a full keyboard |
| 02:12 | <billyjack> | there are a couple few non-standard markup ways now for doing the same thing which are in pretty wide use |
| 02:12 | <billyjack> | as far as I know, there are no production browsers that support inputmode |
| 02:13 | <Hixie> | yeah i don't know of any either |
| 02:13 | <Hixie> | do you have any pointers for the non-standard ways? |
| 02:13 | <billyjack> | Hixie: yeah, I'll try to get the URLs now |
| 02:13 | <Hixie> | seems like the iphone would be able to use it, since it has a fully programmable keyboard |
| 02:13 | <Hixie> | cool, thanks |
| 02:14 | <billyjack> | one way is the Docomo "istyle" attribute, another is a WAP/WML thing that's used everywhere else |
| 02:14 | <billyjack> | http://www.nttdocomo.co.jp/english/service/imode/make/content/html/tag/istyle.html |
| 02:16 | <billyjack> | http://blog.trasatti.it/2005/06/formatting-input-fields-in-xhtml-mp.html |
| 02:17 | <billyjack> | http://www.developershome.com/wap/wcss/wcss_tutorial.asp?page=inputExtension2 |
| 02:17 | <billyjack> | <input type="text" style="-wap-input-format: 'N'"/> |
| 02:18 | <Hixie> | cool, thanks |
| 02:22 | <Hixie> | ok i guess i'll tell people to use -wap-input-format |
| 02:23 | <billyjack> | Hixie: fwiw, the actual spec of -wap-input-format is here: http://www.wapforum.org/tech/documents/WAP-239-WCSS-20011026-a.pdf |
| 02:23 | <billyjack> | only seems to be available in PDF |
| 02:23 | <billyjack> | Opera supports that |
| 02:24 | <Hixie> | cool |
| 02:24 | <billyjack> | I think most all mobile browsers support it. |
| 02:24 | <Hixie> | yeah, definitely no point us defining our own things then |
| 02:25 | <billyjack> | yeah, mobile developers seem to be pretty happy with it |
| 02:25 | <billyjack> | though it seems overengineered to me in that it's both a way of specifying the input mode and also the microsyntax of the input |
| 02:27 | <Hixie> | i love removing things |
| 02:28 | <Hixie> | readonly="" was suspiciously easy to define when i got around to it |
| 02:28 | <Hixie> | the entire spec's discussion of "readonly" is "The readonly attribute is a boolean attribute. When specified, the element is immutable." |
| 02:28 | <Hixie> | granted, "boolean attribute" and "immutable" are both terms with underlying definitions |
| 02:28 | <Hixie> | but still |
| 02:28 | <Hixie> | ok dinner |
| 02:29 | <Hixie> | bbl |
| 03:31 | <billyjack> | dglazkov|home: hei |
| 03:31 | <dglazkov|home> | hey MikeSmith |
| 03:31 | <dglazkov|home> | how's W3 treating ya? |
| 03:32 | <MikeSmith> | dglazkov|home: swimmingly |
| 03:32 | <dglazkov|home> | that's good |
| 03:32 | <MikeSmith> | dglazkov|home: big congrats on your move and new job |
| 03:32 | <dglazkov|home> | MikeSmith: thanks! |
| 03:32 | <dglazkov|home> | the job is great! |
| 03:33 | <dglazkov|home> | still reeling from the move |
| 03:33 | dglazkov|home | never in his dreams imagined he'd be hacking on WebKit full time |
| 03:33 | <MikeSmith> | dglazkov|home: what city were you in before the move? |
| 03:34 | <dglazkov|home> | Birmingham, AL |
| 03:34 | <MikeSmith> | ah |
| 03:35 | <MikeSmith> | I lived in Montgomery when I was young. I had some second cousins in Birmingham. Do they still have the stinking paper mill place outside the city? |
| 03:35 | MikeSmith | wonders if Hixie is subscribed to webkit-dev or knows Michael Nordman |
| 03:36 | <MikeSmith> | re: https://lists.webkit.org/pipermail/webkit-dev/2008-September/004954.html |
| 03:39 | <dglazkov|home> | I know Michael Nordman |
| 03:39 | <dglazkov|home> | :) |
| 03:49 | <dglazkov|home> | MikeSmith: cool, Montgomery, wow! |
| 03:49 | <dglazkov|home> | I don't think the paper mill is there |
| 04:04 | <MikeSmith> | dglazkov|home: seems like Michael might find it worthwhile to come on #whatwg to talk about the appcache stuff |
| 04:05 | <MikeSmith> | Though dunno of course if he's a IRC-liker or not |
| 04:06 | <MikeSmith> | oh |
| 04:06 | MikeSmith | sees michaeln on the channel |
| 04:06 | <MikeSmith> | geez, I'll shutup now and quit being such goddamn busybody |
| 04:07 | MikeSmith | kicks himself |
| 04:11 | MikeSmith | writes 937th note to himself to mind his own business |
| 05:28 | <Hixie> | so the article that the guy who was e-mailing earlier was writing is out now |
| 05:28 | <Hixie> | http://www.infoworld.com/article/08/09/18/Canvas_set_to_boost_AJAX_1.html |
| 05:28 | <Hixie> | it has fewer mistakes than i would have expected given the rush the guy was apparently in |
| 05:29 | <Hixie> | but it still gets one of the key points completely backwards |
| 06:07 | MikeSmith | reads http://webkit.org/blog/214/introducing-squirrelfish-extreme/ |
| 06:09 | <MikeSmith> | polymorphic inline caching |
| 06:10 | <MikeSmith> | that sounds pretty useful, whatever that is |
| 06:10 | <othermaciej> | it is :-) |
| 06:11 | <MikeSmith> | JS learning from Self |
| 06:11 | <MikeSmith> | othermaciej: great writeup |
| 06:12 | <roc> | a text editor written in canvas??? |
| 06:12 | <othermaciej> | roc: what you say? |
| 06:12 | <roc> | sorry, reading Hixie's URL above |
| 06:12 | <roc> | "Galbraith touted a responsive text editor he co-developed that was implemented in JavaScript entirely in Canvas and features syntax highlighting." |
| 06:14 | <roc> | Hixie: actually ... technically, "retained image" is kinda accurate for canvas |
| 06:14 | <MikeSmith> | ah, Paul Krill |
| 06:14 | <roc> | I could see that being a reasonable way to describe the idea that the canvas remembers its contents, as opposed to WM_PAINT-style event-driven application painting |
| 06:15 | <othermaciej> | roc: wow, that's crazy |
| 06:15 | <MikeSmith> | the Paul Krill byline maybe explains a lot about Hixie's earlier comments |
| 06:15 | <roc> | I hope something was lost in translation |
| 06:16 | <MikeSmith> | roc: I very much suspect it was |
| 06:17 | <roc> | since you can't get text metrics from canvas, I think writing an editor using it is completely impossible |
| 06:17 | <othermaciej> | maybe he draws the text by hand |
| 06:19 | <MikeSmith> | heh.. WREC is a nice acronym |
| 06:25 | <Hixie> | MikeSmith: howcome likes to talk about "w3c wreck-ommendation stage" |
| 06:26 | <othermaciej> | I hope one day HTML5 can be a full W3C WREC |
| 06:26 | <Hixie> | 2022 |
| 06:26 | <Hixie> | :-) |
| 06:31 | BenMillard | is reminded of Hixie's "SVG Tiny 1.2 in Candidate Wreckommendation stage": http://ln.hixie.ch/?start=1155235213&count=1 |
| 06:36 | <MikeSmith> | Hixie: :p |
| 06:54 | <hsivonen> | for better PR engagements, we need a Scott McCloud comic about HTML5 |
| 07:24 | <MikeSmith> | I think we need Ralph Steadman for HTML5 |
| 07:24 | <MikeSmith> | hsivonen: ↑ |
| 07:25 | hsivonen | googles |
| 07:44 | <zcorpan_> | Hixie: from http://www.opera.com/docs/specs/#html |
| 07:44 | <zcorpan_> | The inputmode attribute is supported; however, |
| 07:44 | <zcorpan_> | its implementation is platform-dependant |
| 07:44 | <zcorpan_> | it is not enabled by default |
| 07:44 | <zcorpan_> | it is not in the Opera 9.5 Desktop version |
| 07:57 | <MikeSmith> | zcorpan_: do you know if inputmode support in Opera actually implemented on any particular devices? |
| 07:58 | <MikeSmith> | maybe on some S60 devices in Opera 9.5? |
| 08:01 | <MikeSmith> | it seems like it would be trivial to enable support for it on any browser/device combination that already supports -wap-input-format |
| 08:01 | hsivonen | wonders if Opera 9.5 for S60 is still coming or canceled |
| 08:01 | MikeSmith | thought it had already been released |
| 08:01 | <hsivonen> | wow |
| 08:02 | hsivonen | goes check |
| 08:03 | <MikeSmith> | hsivonen: http://www.opera.com/products/mobile/ |
| 08:03 | <hsivonen> | MikeSmith: as far as I can tell, S60 is stuck at Opera 8.65 |
| 08:04 | <MikeSmith> | ah |
| 08:04 | <MikeSmith> | I see the 9.5 beta is Windows Mobile only |
| 08:05 | <hsivonen> | the browser situation on S60 sucks ATM |
| 08:05 | <hsivonen> | both Nokia and Opera haven't updated their offerings in a long time |
| 08:05 | <MikeSmith> | yeah, no public word of any progress at all on an updated Webkit browser for S60 |
| 08:06 | <hsivonen> | someone seems to be working on Gecko for S60, but it's going to be a long, long road for one guy to port it over |
| 08:07 | <MikeSmith> | the Fennec product-dev team is not targeting S60 at all? |
| 08:07 | <MikeSmith> | seems like a decent business opportunity from any third-party company that wanted to come in a provide a better browser for S60 |
| 08:07 | <MikeSmith> | A Webkit-based browser to displace the Nokia one, I mean |
| 08:08 | <MikeSmith> | or really anything |
| 08:08 | <MikeSmith> | kind of surprised that Opera has exploited that more |
| 08:10 | <hsivonen> | even though S60 is behind the times, I find I'm using my phone more often than my N800 |
| 08:11 | <hsivonen> | the reason is that Fennec is so far unusable on N800 (no release with soft keyboard support) |
| 08:11 | <hsivonen> | and MicroB gives a bogus YSoD on Planet Intertwingly |
| 08:11 | <hsivonen> | and Planet Intertwingly is one of my favorite mobile destinations |
| 08:12 | <hsivonen> | it works in Opera 8.65 on S60. The Nokia browser gets the character encoding wrong |
| 08:12 | <virtuelv> | hsivonen: and you can't live with Opera Mini in the meantime? |
| 08:13 | <hsivonen> | what sucks about Opera 8.65 on Planet Intertwingly is the lack of domcontentloaded event |
| 08:13 | <hsivonen> | virtuelv: it doesn't let me switch between SSR and desktop mode with one button without reload |
| 08:13 | <hsivonen> | virtuelv: so no |
| 08:13 | <hsivonen> | virtuelv: Mini requires crawling through menus and reloading |
| 08:14 | <virtuelv> | I come to never ever use desktop mode |
| 08:15 | <hsivonen> | virtuelv: the few tables that are actually data tables instead of layout tables look better in the desktop mode |
| 08:16 | hsivonen | notices that Opera has 7 builds for Windows Mobile |
| 08:19 | <MikeSmith> | virtuelv: doesn't Opera 9.5 mobile have adaptive zooming in desktop mode? |
| 08:19 | <hsivonen> | the banner advertising Opera Mobile 9 coming soon was pulled about a year ago, it seems |
| 08:20 | <virtuelv> | MikeSmith: afaik, yes, http://www.opera.com/products/mobile/features/ |
| 08:21 | <virtuelv> | I detest smartphones for a number of reasons, and windows mobile even more for a bunch of different reasons, so I haven't actually tested 9.5 mobile much yet |
| 08:32 | <MikeSmith> | virtuelv: I think you might kinda like some of the HTC devices at least |
| 08:33 | <virtuelv> | MikeSmith: I have seen the devices, and although their outward physical appearance is nice I can't stand the horrid slowness, subpar usability, sucky battery life and the abomination that is WM |
| 08:35 | <virtuelv> | the usability problems being a consequence of windows |
| 08:35 | <MikeSmith> | yeah |
| 08:36 | <virtuelv> | besides, any phone that uses more than five seconds to boot == fail |
| 08:36 | <MikeSmith> | well, we will soon have another OS running on an HTC device that we can use for comparison purposes |
| 08:37 | <virtuelv> | MikeSmith: I'm not sure the Linux on phone situation is much better |
| 08:38 | <virtuelv> | nor do I believe that Android will save us all |
| 08:39 | <hsivonen> | on Android, all apps are equal but all rendering engines aren't |
| 08:40 | <MikeSmith> | Android will at least add some competition and help raise the bar on user expectations about what's possible/usable on a mobile handset |
| 08:40 | <virtuelv> | the failing here is a far bigger one than the browser engine |
| 08:40 | <virtuelv> | in their insane rush to make the phone a general-purpose computing design, everyone seems to have forgotten the primary objective of a phone: the ability to make phone calls |
| 08:40 | <MikeSmith> | nope |
| 08:41 | <MikeSmith> | that's not the primary purpose for me at all |
| 08:41 | <MikeSmith> | the primary purpose of my handset for me is mobile data services |
| 08:41 | <MikeSmith> | by a longshot |
| 08:41 | <virtuelv> | for me, it still is, together with the ability to send text messages |
| 08:42 | <MikeSmith> | well, I'm blessed to live in a part of the world where we have real SMTP/MIME e-mail clients on handsets |
| 08:42 | <hsivonen> | hmm. I need some kind of Jetty thread pools for dummies guide... |
| 08:42 | <MikeSmith> | so I couldn't care less about text messages |
| 08:42 | <doublec> | email is too slow, text is instant. that's the attraction. |
| 08:42 | <virtuelv> | MikeSmith: so you never feel the need to send just "Ok", or "I'm there in fifteen minutes"? |
| 08:43 | <MikeSmith> | doublec: email on handsets in Japanese is instant |
| 08:43 | <virtuelv> | last phone I bought was a non-smartphone from Sony Ericsson |
| 08:43 | <MikeSmith> | we use email the same way most people use text messages |
| 08:43 | <virtuelv> | except that it tries to be a smartphone-in-disguise |
| 08:43 | <doublec> | interesting |
| 08:43 | <virtuelv> | booting it takes 20 seconds |
| 08:43 | <doublec> | do they have some non-polling method of getting it? |
| 08:43 | <MikeSmith> | virtuelv: yeah, I send email |
| 08:43 | <virtuelv> | MikeSmith: so basically empty e-mails with a subject? |
| 08:44 | <MikeSmith> | empty subject with a body usually |
| 08:44 | <MikeSmith> | and I use it when I want to say, I'll be there in 2 minutes |
| 08:44 | <MikeSmith> | or 1 minute |
| 08:45 | <MikeSmith> | or when I want to send somebody a URL with my exact GPS location |
| 08:45 | <MikeSmith> | so they know where to meet me |
| 08:45 | <hsivonen> | MikeSmith: what's the spam situation with received email? |
| 08:45 | <hsivonen> | my IMAP inbox is too big and spammy for Nokia's S60 email client |
| 08:46 | <hsivonen> | so my only mobile email use case is sending photos as attachments to an IRC bot |
| 08:46 | <hsivonen> | SMS still going strong |
| 08:46 | <MikeSmith> | I used to get very little spam on my handset e-mail |
| 08:46 | <hsivonen> | MMS: FAIL |
| 08:46 | <doublec> | same here, my s60 device can't handle my imap at all |
| 08:46 | <MikeSmith> | I'm getting more lately |
| 08:47 | <MikeSmith> | but I'm sure in general spam will become as big a problem with mobile e-mail as it everywhere else |
| 08:47 | <MikeSmith> | doublec: yeah, non-polling |
| 08:48 | <MikeSmith> | carries push out notifications based on mailbox events |
| 08:48 | <MikeSmith> | to set the "message waiting" indicator on the handset |
| 08:48 | <virtuelv> | and with SMS spam is an almost inexistent problem |
| 08:51 | <MikeSmith> | anyway, the real important this as far as user experience on handsets is integration among the apps |
| 08:52 | <MikeSmith> | e.g., the ability to open a browser while you're in the messaging client |
| 08:53 | <MikeSmith> | or the ability to invoke an app to get your lat/lon location and turn it into a URL and automatically put it in a message and send it to somebody |
| 08:53 | <hsivonen> | having a Chrome-like process per browser view thing going on mobile would be good too |
| 08:53 | <hsivonen> | the browsers are unstable |
| 08:53 | <hsivonen> | and Google turns to Java for stuff like GMM |
| 08:54 | <MikeSmith> | in general apps on handsets are unstable |
| 08:54 | <MikeSmith> | often, if one app crashes the whole handset has to reset |
| 08:56 | <hsivonen> | at least S60 is pretty good at recoving from browser OOM crashes |
| 08:56 | <hsivonen> | recovering |
| 09:10 | <hendry> | hsivonen: are you just using the regular 3 year old S60 browser on your device? |
| 09:10 | <hsivonen> | hendry: yes. |
| 09:11 | <hsivonen> | hendry: are there alternatives? |
| 09:11 | <hendry> | hsivonen: opera :) |
| 09:11 | <hsivonen> | hendry: oh, I thought you asked what "S60 Browser" I used. |
| 09:12 | <hsivonen> | hendry: yeah, I use Opera for stuff other than the logs of this channel |
| 09:13 | <hsivonen> | (Nokia's S60 Browser has an annoyingly generic name) |
| 09:13 | <hsivonen> | probably on purpose |
| 09:14 | <hendry> | hsivonen: you use opera mini? |
| 09:14 | <hsivonen> | hendry: I don't |
| 09:17 | <hsivonen> | ah. the paradox of choice with Jetty thread pools |
| 09:18 | <hendry> | hsivonen: we're running JS tests on mobiles over at http://lists.w3.org/Archives/Public/public-mwts/2008Sep/0003.html |
| 09:18 | <hendry> | http://flickr.com/photos/hendry/2867376431/ my Thinkpad is 10x faster than my 'ipod touch' |
| 09:23 | hsivonen | finds http://www-sop.inria.fr/mimosa/scheme2js/ |
| 10:09 | Hixie | checks in r2222 |
| 10:13 | <annevk2> | Hixie, does it have the <magic-fairy> tag? |
| 10:13 | <annevk2> | seems to be something more boring :) |
| 10:18 | <Hixie> | nn |
| 10:19 | <zcorpan_> | Hixie: we'd like to extend inputmode with e.g. "number" |
| 10:20 | <zcorpan_> | Hixie: so that on a mobile you can use the number input by default instead of tripletap |
| 10:20 | <annevk2> | Hixie, image should also be barred from contraint validation |
| 10:20 | <annevk2> | zcorpan_, doesn't it already have that? |
| 10:21 | <zcorpan_> | annevk2: does it? |
| 10:21 | annevk2 | thought it did |
| 10:21 | <zcorpan_> | not afaict |
| 10:23 | <annevk2> | "latin digits" |
| 10:26 | <zcorpan_> | aha |
| 10:26 | <zcorpan_> | ok |
| 10:30 | <hsivonen> | I wonder if I'm just bad at googling or if Jetty thread pool configuration is a domain with no canned recipe |
| 10:31 | <hsivonen> | I'd expect there to be a function that yields to optimal max thread pool size given some measurable inputs |
| 10:32 | <annevk2> | zcorpan_, if nobody implements this though something like "number" might be better |
| 10:32 | <annevk2> | hsivonen, you could try asking questions on the thing from Joel Spolsky |
| 10:32 | <hsivonen> | annevk2: stack overflow? |
| 10:33 | <virtuelv> | hsivonen: yes |
| 10:33 | <virtuelv> | registration's a breeze, openid all the way |
| 10:34 | <hsivonen> | I like the idea of OpenID |
| 10:34 | <hsivonen> | the problem is the paradox of choice with providers |
| 10:34 | <virtuelv> | I am my own |
| 10:35 | <hendry> | openid2 perl implementation sucks too |
| 10:36 | <hsivonen> | I want to make http://hsivonen.iki.fi/ my OpenID, but I need to move it to a server that is better isolated from other users to feel secure about doing that |
| 10:40 | <annevk2> | you won't do banking with it ;) |
| 10:46 | <virtuelv> | hsivonen: so delegate the ID to some other provider |
| 10:48 | <hsivonen> | virtuelv: the threat model is that another user manages to run some utterly broken PHP that allows a cross-user exploit to bootstrap and a blackhat changes hsivonen.iki.fi to delegate to a bogus provider |
| 10:48 | <virtuelv> | yeah, I see your point |
| 10:48 | <virtuelv> | slicehost has vps hosting |
| 10:49 | <hsivonen> | I already have a couple of VMs that are private to me |
| 10:49 | <hsivonen> | it's just that hsivonen.iki.fi has legacy scripts that are hard to move |
| 10:54 | <hsivonen> | as for using an external provider (regardless of delegation) bothers me, because I don't understand their business model |
| 10:54 | <hsivonen> | and the obvious business model is selling data obtained by observing user behavior |
| 10:55 | <hsivonen> | I wish iki.fi became an OpenID provider |
| 10:55 | <hsivonen> | (and an XMPP provider too) |
| 10:55 | <hsivonen> | in fact, becoming an OpenID provider would fit the mission of iki.fi perfectly |
| 10:56 | <hsivonen> | the trouble is getting someone to implement and maintain stuff |
| 10:56 | <hsivonen> | (the general trouble with associations without paid staff) |
| 11:55 | <annevk2> | hsivonen, seems <input size> is conforming again? |
| 11:56 | <annevk2> | hsivonen, <input inputmode="latin digits x"> gives "Bad value latin digits x for attribute inputmode on element input: Not an absolute IRI." |
| 11:57 | <hsivonen> | annevk2: that's because x isn't an absolute IRI |
| 11:57 | <hsivonen> | and inputmode didn't seem to be worth optmizing UI for |
| 11:58 | <zcorpan_> | the pubnotes for the next WD will be interesting |
| 11:58 | <hsivonen> | in what way? |
| 11:58 | <zcorpan_> | i guess it'll list differences between wf2 and html5 |
| 12:11 | <hsivonen> | and inputmode didn't seem to be worth optmizing UI for |
| 12:14 | <annevk2> | hsivonen, inputmode doesn't take IRIs? |
| 12:14 | <annevk2> | or does it? hmm |
| 12:15 | <annevk2> | but yeah, don't bother with it |
| 12:15 | <hsivonen> | annevk2: it does (or did at least) |
| 12:16 | <hsivonen> | annevk2: it was defined by reference to XForms |
| 12:16 | <hsivonen> | annevk2: in XForms, if a token does not contain a colon, it is a built-in token |
| 12:16 | <hsivonen> | annevk2: if a token does contain a colon, it is an absolute IRI identifying a custom input method |
| 12:28 | <hdh> | hsivonen: using v.nu, how do I override the doctype like http://qa-dev.w3.org/wmvs/HEAD/ can? |
| 12:29 | <zcorpan_> | hdh: change the preset |
| 12:30 | <hdh> | it still report Error: Legacy doctype |
| 12:30 | <hsivonen> | hdh: use the parser pop-up on the generic facet |
| 12:31 | <hdh> | same |
| 12:31 | <hsivonen> | hdh: which doctype? |
| 12:31 | <zcorpan_> | what's your doctype? |
| 12:32 | <hdh> | <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">; |
| 12:32 | <hsivonen> | hdh: it supports only HTML5, HTML 4.01 Strict and Transitional doctypes |
| 12:32 | <hsivonen> | hdh: that's not allowed in text/html |
| 12:32 | <hdh> | I want to replicate http://www.w3.org/QA/2008/09/top-500-html5-validity.html |
| 12:33 | <hsivonen> | hdh: the public validator.nu instance doesn't support legacy doctypes other than HTML 4.01 Strict/Transitional |
| 12:34 | <hsivonen> | hdh: however, the parser has a survey mode that suppresses doctype-related tree builder-level errors |
| 15:15 | <hsivonen> | it tooks me quite a while to figure out that the spec now has two h4 elements saying "The body element" |
| 15:15 | <hsivonen> | which broke my spec scraper |
| 15:19 | <zcorpan_> | the second heading is pretty confusing too |
| 15:20 | <zcorpan_> | it suggests that the body element is obsolete |
| 15:22 | <zcorpan_> | Hixie: would be clearer if it said "Attributes on the body element" |
| 15:25 | <hsivonen> | Hixie: is there any chance of putting presence constraints of elements-specific attributes in parentheses after the attribute |
| 15:25 | <hsivonen> | ? |
| 15:26 | <hsivonen> | Hixie: there used to be "(required)" |
| 15:26 | <hsivonen> | Hixie: and then you took that out instead of putting the more complex requirements in |
| 15:29 | <zcorpan_> | hsivonen: how would that look for the input element? |
| 15:30 | <hsivonen> | zcorpan_: good point. but for other elements it should be feasible |
| 15:31 | <hsivonen> | in the case of input, it might be a good idea to have an element-level h4 for each input type=foo |
| 15:32 | <zcorpan_> | i guess that makes sense |
| 15:33 | <annevk2> | they are now h6 |
| 15:34 | hsivonen | files a spec RFE |
| 15:34 | <zcorpan_> | though it would look weird in the toc |
| 15:35 | <Philip`> | You could fix that in the TOC generator |
| 15:38 | <zcorpan_> | hsivonen: i think you should instead write code that figures out exactly which attributes are missing :P |
| 15:38 | <hsivonen> | zcorpan_: I'm still waiting for oNVDL upstream to release that code |
| 15:39 | <zcorpan_> | ok |
| 15:40 | <hsivonen> | I wish oNVDL upstream had a publicly readable version control system so that I could go get code even when the maintainer is too busy to create a release |
| 15:42 | <MikeSmith> | hsivonen: seems pretty silly that they don't |
| 15:42 | <MikeSmith> | I'd wonder what version-control system they're using |
| 16:05 | <gsnedders> | I'm sleepy. |
| 16:06 | <gsnedders> | But a bit less worried. |
| 17:34 | <takkaria> | all I really want to know if when all these clever JS speedups will make it to other languages |
| 17:41 | <hsivonen> | takkaria: just compile your stuff into JS :-) |
| 18:20 | <Philip`> | takkaria: Maybe it'll happen when those other languages have half a dozen competing implementations, each critically relied on by major software projects, with very little evolution of the language itself to take up the developers' time |
| 18:21 | <Philip`> | and when there's a billion users who are going to be directly affected by the language's performance |
| 18:46 | <jgraham> | Philip`: I wonder if I can convince anyone that javascript is the new C on the basis that it is becoming fast (for a dynamic language) but requires difficult and not entirely effective hacks to get the class-based object model that mainstream programmers like |
| 19:54 | <Hixie> | hsivonen: no, there's no chance because the constraints are far too complicated to put normatively into the summary |
| 20:10 | <hsivonen> | Hixie: :-( |
| 20:10 | <Hixie> | there are very few attributes for which the requirements, if different than "optional", are anywhere near as simple as "required". |
| 20:12 | <hsivonen> | in the machine-detectable cases the condition is rather simple, though |
| 20:15 | <Hixie> | do you have a list of the attributes whose requirement is not "optional"? |
| 20:18 | <hsivonen> | Hixie: I can produce a list |
| 20:19 | <Hixie> | i'd be curious to see it, maybe i'm wrong about how easy it would be |
| 20:19 | <Hixie> | if you're producing a list, the ideal list would be a list of all attributes and their requirement constraint |
| 20:20 | <hsivonen> | do you mean listing optional attributes as well? |
| 20:22 | <Hixie> | if that's easy |
| 20:22 | <Hixie> | otherwise don't worry |
| 20:22 | <Hixie> | whatever you can generate |
| 20:22 | <Hixie> | i'm just looking for an overview |
| 20:22 | <Hixie> | to get a feel of the place, as it were |
| 20:40 | <Hixie> | hsivonen: if i move type="" attribute values up to the element level, what happens to the many commen attribute sections? |
| 20:43 | <Hixie> | "Most specs get grossly outdated within a few days of publication on the TR page." |
| 20:43 | <Hixie> | some are outdated even before they get published |
| 20:55 | <hsivonen> | Hixie: The common things aren't that common if you treat value/min/max as distinct when they differ in datatype |
| 20:56 | <annevk2> | shepazu, you're pointing to a member only thread on a public list |
| 21:06 | <hsivonen> | host html5.validator.nu |
| 21:06 | <hsivonen> | doh |
| 21:08 | <hsivonen> | html5.validator.nu is now isolated from the complex interface and parse tree |
| 21:09 | <hsivonen> | It is now on a VM that can be scaled up and down |
| 21:21 | <hsivonen> | I'm curious about the implementation of xen upflexing |
| 21:22 | <hsivonen> | one would think they need to be able to migrate live VMs across real servers |
| 21:57 | hsivonen | reads a paper about live Xen migration. cool stuff. |
| 22:29 | <Hixie> | hsivonen: value/min/max? |
| 22:29 | <Hixie> | hsivonen: there's like a dozen common attributes. value is one of them (and is already defined as such) |
| 22:29 | <Hixie> | hsivonen: not to mention the DOM attributes most of which are only defined once |
| 22:31 | <hsivonen> | I suppose those could go into a common definition section |
| 22:32 | <hsivonen> | anyway, seeing the applicable attributes per type sate would be more readable than the table |
| 22:34 | <Hixie> | the applicable attributes are already listed per type state |
| 22:34 | <Hixie> | the table is non-normative |
| 22:34 | <Hixie> | i only added it to please anne |
| 22:58 | <krijnh> | Pleasing anne does pay off sometimes :) |