| 00:24 | <roc> | annevk: so currently the CSSOM spec doesn't have a way to get the size of the viewport excluding any scrollbars it may be showing |
| 00:48 | <Lachy> | Hixie, yt? |
| 00:48 | <Hixie> | hey |
| 00:48 | <Lachy> | http://html5.lachy.id.au/beta/ now has some functionality |
| 00:48 | <Lachy> | I disabled all the button cause they don't do anything yet |
| 00:49 | <Lachy> | but the basic live dom viewing feature works |
| 00:49 | <Hixie> | can you swap the dom tree and preview windows? |
| 00:49 | <Hixie> | the dom tree window is too small to see anything |
| 00:49 | <Hixie> | and almost always is bigger than the preview |
| 00:50 | <Lachy> | yeah, that's easy to do |
| 00:52 | <Lachy> | done |
| 00:53 | <Hixie> | neat |
| 00:54 | <Lachy> | I'll add the clipboard function now, that won't take long |
| 01:01 | <Hixie> | othermaciej, roc: i spoke to the people who are experimenting with sdch, asked them about numbers, and they said they're experimenting with that right now trying to find good ways to compress data and so forth, trying to see how it works with proxies, caches, etc, and should have more concrete numbers in the coming months |
| 01:01 | <roc> | ta |
| 01:04 | <othermaciej> | Hixie: ok; hard to evaluate the value of pursuing the technology as a standard without some data |
| 01:04 | <othermaciej> | I'll give it some review once there is more data available |
| 01:04 | <Hixie> | cool |
| 01:07 | <aboodman> | sicking: you rang? |
| 01:09 | <sicking> | aboodman, hey so we have a quick question |
| 01:09 | <sicking> | aboodman, does gears let you create workers from inside a worker? |
| 01:09 | <aboodman> | yes |
| 01:09 | <sicking> | ben wants to know why |
| 01:10 | <sicking> | ben is the guy implementing our workers |
| 01:10 | <aboodman> | ben who? |
| 01:10 | <aboodman> | ok |
| 01:10 | <sicking> | he is not happy about this idea :) |
| 01:10 | <Hixie> | what's difficult about it? |
| 01:10 | <sicking> | and that hooks up the the whole error chain thing? |
| 01:10 | <aboodman> | to be honest, it isn't something i'm passionate about |
| 01:10 | <sicking> | Hixie, nothing's difficult really |
| 01:10 | <Hixie> | ah ok |
| 01:11 | <aboodman> | i think that it is just something that made sense to generalize |
| 01:11 | <Lachy> | Hixie, the clipboard feature works now, though it has a known bug with not updating the DOM tree after using Retrieve |
| 01:11 | <sicking> | i think bens concern is basically that it allows a page to create very complicated things. I.e. trees of workers |
| 01:11 | <sicking> | me, i love it :) |
| 01:11 | <aboodman> | i think that a tree is not the right way to think about it |
| 01:11 | <sicking> | it lets me calculate fibanacci in crazy complicated ways :) |
| 01:12 | <aboodman> | i can't remember what the current state is in the proposed spec |
| 01:12 | <aboodman> | but at one time, workers had this elegant property that they weren't really trees |
| 01:12 | <sicking> | ah, right, you had these ids |
| 01:12 | <sicking> | so you could really communicate with anyone else |
| 01:13 | <aboodman> | right, not ids, but names |
| 01:13 | <sicking> | not just the workers that you created |
| 01:13 | <aboodman> | and you could even pass a port pointing to one worker to another, and make it stay alive longer |
| 01:13 | <sicking> | oh, the docs said 'id' |
| 01:13 | <aboodman> | i'm talking about hixie's original draft, not gears |
| 01:13 | <sicking> | you mean in current spec, or in gears |
| 01:13 | <sicking> | ah |
| 01:13 | <Hixie> | the spec still has that |
| 01:13 | <sicking> | i meant gears |
| 01:13 | <Hixie> | in that you can create a MessageChannel |
| 01:13 | <Hixie> | and hand the ends to two workers |
| 01:14 | <sicking> | right |
| 01:14 | <sicking> | so about this error thingy |
| 01:14 | <aboodman> | you're asking about onerror? |
| 01:14 | <sicking> | well, now i am :) |
| 01:14 | <Hixie> | right now onerror in the spec doesn't propagate more than one level out |
| 01:14 | <Hixie> | and in fact right now it doesn't even forward exceptions |
| 01:14 | <sicking> | initially i think i was just curious if gears (not spec) allowed workers to create workers |
| 01:15 | <aboodman> | yes |
| 01:15 | <aboodman> | and in gears, workers are arranged in trees |
| 01:15 | <aboodman> | and exceptions propagate up |
| 01:15 | <Hixie> | (though i think it would make sense to fire error events for exceptions, and make the default action of that event be to propagate to the higher level) |
| 01:15 | <sicking> | Hixie, yeah, i agree |
| 01:16 | <sicking> | the case for workers within workers that i've been thinking of is if you parse a word document which references external documents |
| 01:16 | <sicking> | you could start up a worker to parse the external (or embedded) resource |
| 01:16 | <sicking> | in which case you'd want to know if that worker fails somehow |
| 01:17 | <Hixie> | yeah |
| 01:17 | <Hixie> | i've filed a bug to make me add that |
| 01:17 | <Hixie> | you can always implement it in the meantime using exception handlers of course |
| 01:17 | <Hixie> | i mean, an author can |
| 01:19 | <sicking> | right |
| 01:21 | Hixie | looks at the w3c's debates on membership fees |
| 01:21 | <sicking> | so there is one thing that is sort of unfortunate |
| 01:21 | <Hixie> | maybe we should charge membership fees for whatwg |
| 01:21 | <sicking> | with the way exception handlers would work |
| 01:21 | <Hixie> | i need a new mac mini for my living room |
| 01:22 | <sicking> | Hixie, excellent idea. And the money can go towards the people implementing the specs :) |
| 01:22 | <Hixie> | if we follow the w3c model, it's the people implementing the specs who pay :-) |
| 01:22 | <sicking> | ugh |
| 01:23 | <sicking> | so these error events are sort of like exceptions, right |
| 01:23 | <Hixie> | sort of |
| 01:23 | <Hixie> | we could even clone the actual exception and put it as a member of the event object |
| 01:23 | <Hixie> | event.exception |
| 01:23 | <sicking> | you can catch them by having a handler, otherwise it bubbles up to be handled by someone higher up |
| 01:23 | <sicking> | and ultimately, if unhandled everywhere, gets logged |
| 01:23 | <Hixie> | not technically bubbling, but yeah, effectively |
| 01:23 | <Hixie> | right |
| 01:24 | <sicking> | so one difference is that for exceptions, if you have a catch block, just having the catch prevents the error from propagating |
| 01:24 | <aboodman> | Hixie: is this only for non-shared workers? |
| 01:24 | <aboodman> | because i don't understand what "higher level" means for shared workers |
| 01:24 | <aboodman> | as they currently exist in the spec |
| 01:24 | <sicking> | aboodman, agreed |
| 01:25 | <Hixie> | aboodman: yes, SharedWorker objects don't have an onerror |
| 01:25 | <sicking> | but with error handlers, you would have to explicitly prevent "bubbling" by calling event.preventDefault() |
| 01:25 | <Hixie> | sicking: right |
| 01:25 | <sicking> | even if you have a handler |
| 01:25 | <sicking> | this seems unfortunate |
| 01:25 | <Hixie> | sicking: yes |
| 01:26 | <Hixie> | dunno that we can do anything about it though |
| 01:26 | <Hixie> | how does <body onerror=""> work? |
| 01:26 | <Hixie> | that's not an event |
| 01:26 | <Hixie> | does it not need to be canceled? |
| 01:27 | <sicking> | hmm.. i'm not really sure when that fires, it might even fire after the logging has taken place |
| 01:27 | <sicking> | (which would otherwise be the default action) |
| 01:27 | <Hixie> | apparently returning true prevents the default action |
| 01:28 | <Hixie> | so if you use worker.onerror, you'd just need to return true (or false or whatever) to prevent the default |
| 01:28 | <Hixie> | as opposed to calling preventDefault() |
| 01:28 | <Hixie> | like with any event handler |
| 01:28 | <Hixie> | in fact |
| 01:28 | <sicking> | true |
| 01:28 | <Hixie> | you only need preventDefault() when you use addEventListener() |
| 01:28 | sicking | proxies to ben |
| 01:28 | <Hixie> | heh |
| 01:28 | <Hixie> | get ben to come online :-) |
| 01:29 | <Hixie> | brb moving building |
| 01:29 | <aboodman> | sicking: are you guys implementing both shared and non-shared workers? |
| 01:29 | <aboodman> | also are you planning to implement messagechannels |
| 01:29 | <aboodman> | (as in new MessageChannel()) |
| 01:33 | <sicking> | probably no shared and no channels for now :( |
| 01:33 | <sicking> | due to time constraints |
| 01:33 | <sicking> | aboodman, ^ |
| 01:34 | <aboodman> | k |
| 01:34 | <sicking> | and possibly no localStore |
| 01:34 | <sicking> | but xhr is mostly working |
| 01:34 | <sicking> | should work fine with cross-site xhr even |
| 02:18 | <billyjack> | does anybody have appcache test cases online? |
| 08:20 | <aboodman> | sicking: in the house? |
| 10:03 | <zcorpan> | problem: i want to make people happier. solution: RDFa. |
| 10:03 | <annevk> | result: fail |
| 10:03 | <annevk> | . |
| 10:03 | <zcorpan> | dang |
| 10:05 | <zcorpan> | alternative solution: give people money. |
| 10:06 | <annevk> | I'm not sure that works long term |
| 10:06 | <annevk> | btw, Firefox <video> impl fails a bit on http://people.opera.com/howcome/2007/video/wikipedia/macaw.html (pressing stop, play, etc.) |
| 10:06 | <annevk> | though it works good enough for a demo |
| 10:07 | <zcorpan> | is working long term a requirement? |
| 10:07 | <annevk> | (it also doesn't determine the size correctly until the video starts playing) |
| 10:07 | <annevk> | I think so |
| 10:07 | <zcorpan> | dang |
| 10:14 | zcorpan | wonders why he is reading RDFa emails |
| 10:15 | <annevk> | dom |
| 10:16 | <annevk> | (dubbele betekenis bedoelt :) ) |
| 10:17 | <hsivonen> | I'd be more happy if people gave me (notable) money than RDF |
| 10:18 | <jgraham> | hsivonen: Presumably the RDF equivalent would be allowing everyone to mint their own currency |
| 10:19 | <hsivonen> | jgraham: I think there are bad historical precedents with that approach |
| 10:19 | <annevk> | and DNS is like gold? |
| 10:19 | <annevk> | (like gold used to be, that is) |
| 10:20 | <hsivonen> | annevk: but if you use URIs only for identification without resolving them, DNS becomes like gold now :-) |
| 10:20 | <annevk> | haha |
| 10:20 | <Hixie> | was there much RDF talk |
| 10:20 | <Hixie> | ? |
| 10:20 | <Hixie> | i thought it was just a brief bit of process talk |
| 10:22 | <annevk> | well, a week ago or so there was |
| 10:22 | <Hixie> | oh |
| 10:22 | <Hixie> | i thought people were saying they were reading abotu rdfa today |
| 10:23 | <zcorpan> | Hixie: i was |
| 10:23 | zcorpan | is catching up on email |
| 10:45 | Hixie | updates the rdfa wiki page to indicate the problems more obviously |
| 10:54 | <hsivonen> | are developers supposed to build federated facebook-like apps with FOAF? |
| 10:58 | <hsivonen> | I wonder how much overlap there is between Lips folks and RDF folks |
| 10:58 | <hsivonen> | Lisp |
| 10:59 | <hsivonen> | Lisp and RDF both have very simple core concepts |
| 10:59 | <hsivonen> | and both aren't mainstream |
| 10:59 | <hsivonen> | both are harder to grok than mainstream stuff |
| 10:59 | <hsivonen> | and the mainstream stuff keep reinventing things that Lisp and RDF had years ago |
| 11:00 | <Hixie> | foaf has near zero uptake so far |
| 11:00 | <Hixie> | xfn has quite a lot, though that's somewhat biased by wordpress |
| 11:00 | <Hixie> | (foaf's only notable user is livejournal, as i understand it, which also does xfn) |
| 11:01 | <hsivonen> | Hixie: does livejournal consume FOAF? |
| 11:01 | <hsivonen> | OTOH, Lisp is very much about minimalist syntax |
| 11:02 | <hsivonen> | whereas RDF's first attempt at syntax wasn't minimalist at all and RDF tries to pretend it isn't about syntax |
| 11:02 | <Hixie> | i don't think lj consumes foaf no |
| 11:02 | <hsivonen> | what kind of software consumes FOAF? |
| 11:03 | <Hixie> | google's social stuff consumes foaf |
| 11:03 | <hsivonen> | and lisp has short identifiers |
| 11:03 | <hsivonen> | like cons |
| 11:03 | <Hixie> | (for now, anyway) |
| 11:03 | <hsivonen> | instead of http://ai.mit.edu/lisp/namespace/cons |
| 11:03 | <Hixie> | i wonder if we should fire onsubmit before or after validating the form |
| 11:03 | <Hixie> | i guess after makes sense |
| 11:04 | <Hixie> | if onsubmit does something like disable the submit button |
| 11:04 | <hsivonen> | what does Opera do? |
| 11:04 | <Hixie> | wf2 says after |
| 11:04 | <Hixie> | i don't know what opera does |
| 11:06 | <Lachy> | I like how, in the Generic_Metadata_Mechanisms wiki page, one of the requirements is "Inlinability", yet the proposal with the most pro's (using an external resource) doesn't meet that requirement :-) |
| 11:07 | <Hixie> | the proposals were supposed to be "and"ed, i believe |
| 11:08 | <Lachy> | oh |
| 11:08 | <Lachy> | then a con for the whole lot is that it is overly complex to have them all |
| 11:08 | <Hixie> | quite possibly |
| 11:08 | <Hixie> | i don't really care what the proposals are |
| 11:08 | <Lachy> | and redundant to have "Inline (as multiple attributes)" and "Inline (in a single attribute)" |
| 11:09 | <Hixie> | it's the actual problem and the resulting requirements that matter |
| 11:09 | <Hixie> | feel free to edit |
| 11:09 | <Lachy> | yeah, but they haven't filled out the problem yet |
| 11:09 | <Hixie> | i know |
| 11:09 | <Lachy> | I would if I knew what the problem was |
| 11:09 | <annevk> | the primary RDF advocates haven't touched that page |
| 11:10 | <Hixie> | so it seems |
| 11:10 | <Hixie> | i wish they would |
| 11:11 | <Hixie> | the 'submit' event bubbles? wow |
| 11:11 | <Hixie> | who'da thought |
| 11:59 | <Hixie> | nn |
| 12:00 | <hsivonen> | nn |
| 12:18 | <hsivonen> | should jslint be considered a legitimate general JS validator or a manifestation of Douglas Crockfords personal opinions? |
| 12:20 | <hsivonen> | Problem at line 59 character 25: 'installHandlers' was used before it was defined. |
| 12:20 | <hsivonen> | are forward references a real problem in JS? |
| 12:20 | <Lachy> | hsivonen, not for functions |
| 12:21 | <Lachy> | only variables have to be declared before you can use them |
| 12:21 | <hsivonen> | So Doug Crockford doesn't like semicolon omission |
| 12:22 | <hsivonen> | does semicolon omission now suck officially, or is it just one person's opinion? |
| 12:22 | <Lachy> | semi-colon omission is considered bad practice |
| 12:22 | <hsivonen> | :-( |
| 12:22 | <Lachy> | (although, that's just my opinion too) |
| 12:22 | <hsivonen> | and I made an effort to try to be pythonic and avoid the silly semicolons |
| 12:22 | <Lachy> | javascript is not pything |
| 12:23 | <othermaciej> | I'd say in a real script, I would include the semicolons |
| 12:23 | <Lachy> | python* |
| 12:23 | <hsivonen> | Crockford did have some good reasons for using semicolons though |
| 12:23 | <othermaciej> | there's real pitfalls to skipping them |
| 12:23 | <hdh> | fwiw js2-mode also flag semicollon omission |
| 12:23 | <othermaciej> | but for a one-liner event listener they are ok |
| 12:24 | <Lachy> | what are the conditions under which semi colons are allowed to be omitted? |
| 12:24 | <othermaciej> | onclick="callSomething(event)" |
| 12:24 | <hsivonen> | can Aptana or something insert semicolons for me? |
| 12:24 | <othermaciej> | if there's a line break you can generally omit the semicolon unless it might look like the next line could continue the statement |
| 12:24 | <othermaciej> | so ++ and -- are dangerous |
| 12:24 | hsivonen | uninstalled Aptana because it seemed to leak memory |
| 12:25 | <othermaciej> | since they can be either prefix or postfix operators |
| 12:26 | annevk | also writes pythonic JavaScript |
| 12:27 | <hsivonen> | I guess it's less controversial if I implement links to the CSS Validator and leave JSLint out for now |
| 12:29 | <Lachy> | when I write python, I tend to write it more like JavaScript :-) |
| 12:31 | <Lachy> | hey, with CSS3 Media Queries, the spec has this BNF: media_query: [[only | not]? <media_type> [ and <expression> ]*] | <expression> [ and <expression> ]* |
| 12:31 | <Lachy> | I'm wondering what the OR (|) operator applies to. Is it equivalent to: media_query: [[only | not]? <media_type> ([ and <expression> ]*]) | (<expression> [ and <expression> ]*) |
| 12:31 | <Lachy> | Or ... |
| 12:32 | <Lachy> | media_query: ([[only | not]? <media_type> [ and <expression> ]*]) | (<expression> [ and <expression> ]*) |
| 12:33 | <othermaciej> | Lachy: or has looser binding than sequence in BNF |
| 12:33 | <annevk> | hmm, the first of your equivalents is malformed |
| 12:33 | <Lachy> | if it's the latter, I think that would make this legal: media="(min-width: 100px) and (min-height: 100px)" |
| 12:33 | <annevk> | that is legal |
| 12:33 | <Lachy> | annevk, probably not. I was only using the paretheses to mark the sections that the OR was applying to |
| 12:34 | <annevk> | well, you did it wrong then :) |
| 12:34 | <othermaciej> | in other words, you can either start with a media type (optionally prefixed by only or not) |
| 12:34 | <othermaciej> | or you can start with an expression |
| 12:34 | <annevk> | even if it was not malformed, it would not be less ambiguous |
| 12:34 | <othermaciej> | any number of and <expression> can follow in either case |
| 12:34 | <othermaciej> | that is what the BNF sys anyway |
| 12:35 | <Lachy> | ok. I didn't realise that before. I thought all expressions had to start with a media type |
| 12:35 | <othermaciej> | I am not sure if that was intentional but that appears to be the effect |
| 12:35 | <annevk> | as editor of said spec, i can say it's intentional |
| 12:35 | <Lachy> | there are no examples in the spec that don't start with a media type |
| 12:36 | <Lachy> | annevk, you're not listed as an editor in the /TR/ copy. When did you take it over? |
| 12:37 | <annevk> | later |
| 12:37 | <annevk> | you should really read http://dev.w3.org/csswg/css3-mediaqueries/ |
| 12:40 | <Lachy> | ok. I didn't realise it had a public editors draft |
| 12:42 | zcorpan | thinks specs at TR/ should point to the editor's draft |
| 12:42 | <zcorpan> | editor's drafts are pretty hard to find |
| 12:43 | <hsivonen> | is there ever a good reason to read TR/ instead of dev.w3.org? |
| 12:44 | <Lachy> | hsivonen, not if the spec is a draft and there is a copy of it on dev.w3.org. But if it's a REC, then reading the TR is reasonable |
| 12:45 | <hsivonen> | Lachy: well, you wouldn't want to read CSS2 instead of CSS 2.1 ED or HTML 4.01 instead of HTML5 ED |
| 12:45 | <Lachy> | but for all the CSS specs, all my bookmarks point to the TR drafts and I'm too lazy to look up the dev.w3.org versions and change them |
| 12:46 | <Lachy> | it depends on the reason. Sometimes it's useful to look up how some things were defined in HTML4, especially for web developers who are still only writing HTML4 in practice |
| 12:47 | <Lachy> | although, I tend to use both. HTML4 tells me what's valid, HTML5 tells me what it means and how browsers will handle it. |
| 12:47 | <hsivonen> | does anyone have an opinion on whether the usability of http://html5.validator.nu/ would go up or down if the text area and file upload selections had two submit buttons: Validate as HTML and Validate as XML? |
| 12:47 | Philip` | wants to read HTML 4.01 instead of HTML5 ED, because it's much easier to read and understand |
| 12:48 | <Philip`> | (at least when I'm just trying to write HTML pages) |
| 12:48 | <hsivonen> | it would be much easier to remove HTML 4.01 from Validator.nu if HTML5 allowed the most popular Transitionalisms |
| 12:49 | <Philip`> | hsivonen: Just make it validate as both and report both results - it's quicker for the computer to do that than for the human to choose which button to click :-) |
| 12:50 | <hsivonen> | hmm. Am I just being stupid by having a name attribute on the submit button on Validator.nu and going through all the trouble of disabling and enabling it? |
| 12:50 | <Philip`> | (P.S. That would be a bad idea) |
| 12:50 | <hsivonen> | Philip`: Hixie has suggested something like that |
| 12:51 | <Philip`> | It would just confuse users if the validator said their document was both valid (HTML) and invalid (XHTML) |
| 12:51 | <Lachy> | hsivonen, 2 buttons seem reasonable |
| 12:51 | <hsivonen> | whoa. I've forgotten to implement ruby in the HTML5 schema |
| 12:51 | <Lachy> | although, perhaps it will be a problem for authors who think they're writing XHTML, but still serve it as text/html |
| 12:51 | <hsivonen> | and no one has pointed it out to me |
| 12:52 | <hsivonen> | (unless I've missed the logs or something) |
| 12:52 | <Philip`> | If you have have two buttons, be sure to record stats on how many people use each one, and if almost nobody uses XHTML then you might as well remove it from the default UI |
| 12:53 | <hsivonen> | having 2 buttons poses the problem of which one which browser will pick as the default |
| 12:53 | <Philip`> | Also you should use onmouseover to record stats of how many times people wave their mouse indecisively between the two buttons, to get some idea of how much time it wastes |
| 12:54 | <hsivonen> | and whether I should put HTML on the right (Mac) or on the left (Windows) |
| 12:54 | <Philip`> | Use UA sniffing to decide |
| 12:55 | <Philip`> | Or put one below the other |
| 12:56 | <Philip`> | Or have a <select> to choose HTML/XHTML, since few people will use XHTML and so the extra clicks are not significant |
| 12:56 | <hsivonen> | does anyone happen to know if getResourceAsStream stuff in jars is subject to Java package/class naming restrictions? |
| 13:01 | <deane> | hsivonen: put the html button on the left, people usually say "html and xhtml", not "xhtml and html", and most people read from left to right, so have html first, then xhtml. Kinda makes sense to me anyway :) |
| 13:01 | <zcorpan> | hsivonen: just support HTML and let people use the generic interface for XHTML :) |
| 13:03 | <hsivonen> | deane: for LTR text, putting the default on the right (as on Mac) actually makes sense, since that's where the eye is going |
| 13:03 | <deane> | hsivonen: OK. But maybe it's best just to have the radio buttons instead |
| 13:05 | <hsivonen> | zcorpan: I want to add SVG 1.1 support to the HTML5 facet, though. |
| 13:07 | <hsivonen> | looks like the validator behavior isn't graceful if a blank file upload field is submitted |
| 13:07 | <annevk> | can't you just have two small links at the bottom: "Also from Validator.nu: XHTML Validator & Advanced Validator" |
| 13:08 | <hsivonen> | annevk: yes, I intend to do that RSN |
| 13:08 | <annevk> | define:RSN |
| 13:08 | <hsivonen> | Real Soon Now |
| 13:09 | <zcorpan> | hsivonen: why? most people don't use svg |
| 13:09 | <deane> | hsivonen: you wouldn't need an "HTML/XHTML" option for file upload right? Doesn't V.nu just decide between HTML and XHTML based on the file extension? |
| 13:10 | <hsivonen> | zcorpan: SVG is a real Web technology supported by 2 or more of the top 4 engines, and detecting it on root ns is easy for the URI validation case |
| 13:10 | <zcorpan> | hsivonen: yes but for textarea |
| 13:11 | <zcorpan> | hsivonen: perhaps if the first token is "svg" you switch to xml mode :) |
| 13:11 | <zcorpan> | (first start tag token) |
| 13:11 | <Philip`> | What you need is doctypes |
| 13:11 | <hsivonen> | deane: good point. It already looks at the filename extension. |
| 13:11 | <hsivonen> | I forgot :-) |
| 13:12 | <Philip`> | Inline metadata is far more reliable than external metadata that's bound to get separated from the data and will force the user to re-enter it multiple times whenever they try to process the data |
| 13:13 | <hsivonen> | Validator.nu now doesn't disable the submit button on submit |
| 13:13 | <hsivonen> | better? worse? |
| 13:14 | <zcorpan> | hsivonen: better because you can submit several times when it's too slow :) |
| 13:17 | <annevk> | and because when you hit back it might not be undone |
| 13:18 | <annevk> | and because then you can quickly fix a typo and hit submit again, etc. |
| 13:18 | <hsivonen> | I still disable the other fields that I want to omit from the dataset |
| 13:19 | <hsivonen> | because it would be confusing if those weren't fixed on back button but there was no visual feedback about it |
| 13:27 | <deane> | hsivonen: Although, that brings up another problem that I've been wondering about lately. What if some one is editing a XHTML5 document that they don't want validated as XHTML5, but want validated as HTML5 because they are going to serialize it to text/html before sending over the wire. So I wonder if V.nu needs to cater for someone who says: Validate my XHTML5 document in HTML5 (text/html)... |
| 13:27 | <deane> | ...compatible mode please, if you know what I mean. |
| 13:27 | <hsivonen> | deane: V.nu should support that via the complex interface |
| 13:27 | <hsivonen> | deane: I don't want that case complicating the simple interface |
| 13:28 | <hsivonen> | deane: check the lax type box and select the parser manually |
| 13:29 | <deane> | hsivonen: cool, I just wondered if it had that feature, that's all |
| 13:30 | <Lachy> | I started adding use cases for accessibility to the wiki http://wiki.whatwg.org/wiki/Video_accessibility#Selection_Mechanisms |
| 13:45 | <hsivonen> | hmm. http://code.google.com/p/concentre-xforms/ |
| 14:05 | <hsivonen> | so... when XForms JS libraries start using colon in names in text/html, does it mean does colon-based NS processing in the browser level in text/html is even less of an option than it used to be? |
| 14:12 | <Lachy> | hsivonen, it was already not an option at all. If it's possible to be less of an option than that, then yes |
| 14:15 | <JohnResig> | Lachy: does Opera 9.6 have querySelectorAll? |
| 14:15 | <JohnResig> | Lachy: does Opera 9.6 have querySelectorAll? |
| 14:15 | <JohnResig> | oops |
| 14:17 | <JohnResig> | hmm, doesn't seem to in the release notes, at least: http://www.opera.com/docs/changelogs/windows/960b1/ |
| 14:17 | <hsivonen> | no video it seems... |
| 14:18 | <JohnResig> | it seems like a bug fix release |
| 14:18 | hsivonen | sees Added Core version (currently "Presto/2.1.1") to the User Agent header |
| 14:18 | <Lachy> | JohnResig, I cannot comment on future product releases |
| 14:18 | <JohnResig> | Lachy: it's not really future - I was just wondering about the beta release |
| 14:18 | <Lachy> | oh, I didn't know the beta was released |
| 14:18 | <Lachy> | that's a desktop thing, which I don't work with |
| 14:19 | <hsivonen> | Lachy: what do you work with? |
| 14:19 | <virtuelv> | JohnResig: 9.60 does not support querySelectorAll, no |
| 14:19 | <JohnResig> | virtuelv: k |
| 14:19 | <Lachy> | I work in QA Core, using gogi |
| 14:20 | <Lachy> | like the Acid 3 build we released a while back. it's basically just the core of the browser with a minimal UI |
| 14:20 | <hsivonen> | I didin't realize QA Core didn't see the desktop UI |
| 14:21 | <hsivonen> | (even though I knew about gogi) |
| 14:21 | <virtuelv> | JohnResig: however, fwiw, the acid 3 build Lachy is talking about does support querySelectorAll |
| 14:21 | <JohnResig> | virtuelv: that's the one that I tested on, which is why I was wondering |
| 14:23 | <Lachy> | hsivonen, I use the desktop build on when I need to on Mac, since gogi is only build for windows and linux. But I generally try to avoid using it |
| 14:29 | <virtuelv> | JohnResig: useful read: http://dev.opera.com/articles/view/presto-2-1-web-standards-supported-by/ |
| 15:03 | <Philip`> | The text-shadow example on that page would be more impressive if it didn't make Opera (9.5) slow down to ~5fps while it's scrolling into view |
| 15:10 | <virtuelv> | Philip`: heh |
| 15:10 | <virtuelv> | I have an animated example with text-shadow lying around that easily manages a lot more |
| 15:12 | <virtuelv> | does anyone know if we've release anything to the public on w.o.c. with Rune's animation lib? |
| 15:13 | <virtuelv> | err, wrong channel, obviously |
| 15:13 | <virtuelv> | (I just need to sort out rights for some code before publishing the example) |
| 15:33 | <Lachy> | I added more use cases to http://wiki.whatwg.org/wiki/Video_accessibility#Selection_Mechanisms |
| 15:33 | <Lachy> | hsivonen, can you think of any that I've missed? |
| 15:34 | <Lachy> | Hixie, is that the quality of use cases you need? |
| 16:09 | <virtuelv> | Philip`: http://virtuelvis.com/gallery/text-shadow/ |
| 16:09 | <virtuelv> | is that just as slow for you? |
| 16:12 | <Philip`> | virtuelv: That seems much faster, though it still looks a bit jerky |
| 16:14 | <virtuelv> | hm, that has acceptable performance here, and doesn't tickle the CPU enough to upstep from 800MHz to 1.4GHz |
| 16:15 | <zcorpan> | hmm. should </x><!doctype html> trigger standards mode? |
| 16:17 | <zcorpan> | no quirks mode |
| 16:18 | <Philip`> | virtuelv: Doesn't use much CPU here either, but it's not fading smoothly - it seems to kind of get very briefly stuck occasionally, so it looks like it's going in steps rather than a single continuous transition |
| 16:22 | <Lachy> | virtuelv, the page works in Safari now |
| 16:24 | <virtuelv> | Lachy: yeah, seems opera's parsing of arrays is a little on the lenient side, so I never caught the error |
| 16:24 | <Lachy> | file a bug about that array parsing then, I don't think we should be lenient |
| 16:25 | <Lachy> | Firefox and Safari both throw the same error, so I don't see a reason for us to ignore it |
| 16:29 | <hallvors> | what array parsing issue is it? I think it's known. |
| 16:33 | <virtuelv> | hallvors: [ "foo" "bar"] produces an array with length 2 |
| 16:34 | <Lachy> | hallvors, http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cscript%3E%0D%0Avar%20x%20%3D%20%5B%22a%22%20%22b%22%5D%0D%0Aw(x)%3B%0D%0A%3C%2Fscript%3E |
| 16:35 | <virtuelv> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cscript%3E%0D%0Adocument.write(%5B%22a%22%20%22b%22%5D.length)%0D%0A%3C%2Fscript%3E |
| 16:35 | <hallvors> | known |
| 16:36 | <hallvors> | bug 298967 if you're interested |
| 16:44 | <zcorpan> | ["a""b"] |
| 17:09 | <sicking> | aboodman, in the house now :) |
| 17:29 | <sicking> | aboodman, you pung |
| 17:30 | <takkaria> | is that the past tense of "smell"? :) |
| 17:31 | <aboodman> | sicking - yeah, wanted to talk about worker stuff, but i will write up coherent thoughts. |
| 17:31 | <sicking> | Lachy, virtuelv: That gives me a parse error in FF3 |
| 17:32 | <sicking> | aboodman, cool |
| 17:43 | <Lachy> | sicking, yes, we know. It's a bug in Opera that makes it not throw a parse error |
| 17:46 | <sicking> | Lachy, ah, ok |
| 17:57 | <gavin> | 23 |
| 19:25 | <hsivonen> | interesting accessibility hack: <img src="http://images.apple.com/main/elements/spacer.gif"; border="0" height="5" width="20" alt="Voiceover users click here to listen to keynote address."> |
| 19:42 | <webben> | alts in spacers is old school. |
| 20:09 | <gsnedders> | http://tech.groups.yahoo.com/group/rss-public/message/1778 — that is lovely. |
| 21:07 | <Lachy> | how do I make VoiceOver work with Firefox? All I can make it do is make it tell me that the close button is selected |
| 21:09 | <webben> | Lachy: VO doesn't work with Fx. |
| 21:09 | <Lachy> | ok |
| 21:09 | <webben> | Lachy: Mozilla haven't implemented the Apple Accessibility API for Gecko. |
| 21:10 | <webben> | Lachy: best current option for speech on Firefox Mac is Fire Vox. |
| 21:10 | <webben> | Lachy: see also aaronlev 's article: http://accessgarage.wordpress.com/2008/08/21/firefox-and-os-xs-voiceover-reading-the-magic-8-ball/ |
| 21:11 | <webben> | Lachy: oh, or WebAnywhere. |
| 21:11 | <aaronlev> | we have only very basic support |
| 21:12 | <aaronlev> | needs work |
| 21:13 | <Lachy> | webben, I just wanted to try using voiceover to get to that audio only copy of the apple keynote that hsivonen mentioned |
| 21:13 | <Lachy> | I can't get it to navigate the page using safari either. It's really not intuitive for me |
| 21:16 | <Lachy> | anyway, If I copy the link manually from the source, I get 404 |
| 21:17 | <hsivonen> | it seems that the stevenote video track is text/html |
| 21:18 | <hsivonen> | I had problem with getting non-Apple software to show it |
| 21:20 | <webben> | Lachy: which URL is the link on? |
| 21:20 | <Lachy> | http://events.apple.com.edgesuite.net/0809dt4bs89/event/index.html |
| 21:20 | <webben> | ta |
| 21:21 | <webben> | I see the img hack is in a link. |
| 21:21 | <webben> | nasty |
| 21:21 | <Lachy> | I don't get why they don't make that audio more easily available to everyone |
| 21:22 | <webben> | agreed |
| 21:23 | <webben> | Yeah, the link 404s for me. |
| 21:24 | <webben> | Also, in Safari, when I tab to the link it reads the href not the alt. |
| 21:25 | <smedero> | webben: that happens when using VoiceOver on OS X? are you using Safari on Windows.. or? |
| 21:26 | <webben> | VoiceOver, Safari, Leopard. |
| 21:26 | <webben> | s/tab to/focus on/ |
| 21:27 | <webben> | If I had to guess, maybe WebKit/VO is specifically excluding pixel gifs from consideration as content. |
| 21:29 | <webben> | Lachy: reads correctly in Opera :) |
| 21:30 | <webben> | also works in latest WebKit |
| 21:35 | <smedero> | webben: hrm, yeah with respect to the latest WebKit - the Safari 4 developer preview reads the @alt text as well. |
| 22:42 | <Lachy> | wow, this is cool http://www.youtube.com/blog?entry=mi8D3ntPgFQ |
| 22:43 | <Lachy> | this is a video with subtitles http://www.youtube.com/watch?v=XraeBDMm2PM |
| 22:45 | <jgraham> | Their UI won't scale to 120 languages... |
| 22:46 | <jgraham> | Er, maybe I should have started with "cool!" |
| 22:46 | <Lachy> | I don't think anyone will provide subtitles in 120 languages for one video |
| 22:46 | <jgraham> | Ironic the BBC don't capion in English |
| 22:46 | <Lachy> | I think they just meant that there are 120 langauges to choose from |
| 22:47 | <jgraham> | Right, so they should design a system that will work for 120 languages :) |
| 22:47 | <Lachy> | this one has english captions http://www.youtube.com/watch?v=ullmDNjb_ec |
| 22:47 | <jgraham> | (Yes I realise that would be really hard) |
| 22:47 | <jgraham> | (and almost certianly not worth the effort) |
| 22:50 | <gsnedders> | Is a script executed when added to the DOM using createElement() and appendElement()? |
| 22:58 | <jgraham> | gsnedders: I think the answer is yes but I don't know if/where it is defined |
| 22:58 | <gsnedders> | it isn't, as far as I can see |
| 22:59 | <jgraham> | When a script element that is marked as neither having "already executed" nor being "parser-inserted" is inserted into a document, the user agent must run the script element. |
| 23:00 | <jgraham> | http://www.whatwg.org/specs/web-apps/current-work/#script |
| 23:00 | <jgraham> | So all you need is to verify that DOM calls neither set the "parser inserted" nor "already executed" flags |
| 23:01 | <gsnedders> | ah, ok |
| 23:04 | <Lachy> | gsnedders, have you got those photos yet? |
| 23:05 | <gsnedders> | Lachy: Oh, let me nag Ross to see if I can use his photo |
| 23:05 | <jgraham> | gsnedders: You have a camera, right? You could have a picture on the tubes within minutes |
| 23:06 | <gsnedders> | jgraham: Temp. dye. |
| 23:06 | <gsnedders> | jgraham: I no longer have green hair |
| 23:06 | <jgraham> | lame |
| 23:06 | <Lachy> | oh, what?! That wasn't the deal! |
| 23:06 | <jgraham> | ;) |
| 23:06 | <gsnedders> | Lachy: You didn't specify :) |
| 23:07 | <Lachy> | I was expecting you to turn up at TPAC with it |
| 23:08 | <gsnedders> | Lachy: OK, maybe I'll let someone perm. dye it there |
| 23:12 | <gsnedders> | I have next to no money, so maybe I'll have to get someone else to buy the dye |
| 23:12 | <gsnedders> | Oh, my hair will certainly need bleaching to get it to work at all |
| 23:18 | <gsnedders> | Lachy: Give me a bit and I'll email you a photo |
| 23:19 | <jgraham> | flickr or it didn't happen :) |
| 23:19 | <gsnedders> | jgraham: When I put up the rest of my photos, sure |
| 23:19 | <gsnedders> | jgraham: Till then, be patient :) |
| 23:19 | <Dashiva> | Don't forget to add descriptions |
| 23:19 | <gsnedders> | Should upload them this week, though |
| 23:19 | <gsnedders> | Dashiva: I don't know who half the photos are of! |
| 23:19 | <gsnedders> | I'm terrible with names! |
| 23:19 | <gsnedders> | I can't label them! |
| 23:20 | <jgraham> | gsnedders: Surely it isn't much harder to upload 1 photo to flickr than to email it |
| 23:20 | <gsnedders> | jgraham: I want all my photos going up at once! :P |
| 23:20 | gsnedders | needs to go to sleep |
| 23:20 | <jgraham> | ?!? |
| 23:20 | <jgraham> | (the sleep bit I got) |
| 23:20 | <Dashiva> | If he doesn't upload them all at once, the alt defense fails ;) |
| 23:21 | <gsnedders> | Exactly :P |
| 23:21 | <jgraham> | I almost never upload more than one photo at a time so I have a different alt defence |
| 23:21 | <jgraham> | :) |
| 23:27 | <gsnedders> | Oh, there are so many nice search strings for my site |
| 23:27 | <gsnedders> | "what causes girls feelings to change" |
| 23:28 | <Dashiva> | As long as you don't get the ones involving donkeys and neon bathtubs, it's all good |
| 23:28 | <gsnedders> | I could quote the bible on donkeys penises… |
| 23:29 | <smedero> | the what on the who? |
| 23:29 | <smedero> | gotta love search query data |
| 23:29 | <Dashiva> | Just remember that once you've posted it, you can't unpost it |
| 23:29 | <gsnedders> | "There she lusted after her lovers, whose genitals were like those of donkeys and whose emission was like that of horses." — Ezekiel 23:20 |
| 23:31 | <smedero> | is this why you've been writing a habari fortune plug-in? |
| 23:32 | <gsnedders> | I have? |
| 23:32 | <smedero> | http://hg.gsnedders.com/hgwebdir.cgi/sneddy_quote/ |
| 23:32 | <smedero> | ;-) |
| 23:32 | <gsnedders> | smedero: Oh, that's just random quotes of me |
| 23:33 | <smedero> | oh good, so no donkeys then? |
| 23:33 | <gsnedders> | No, I'm not Ezekiel |
| 23:33 | <smedero> | most excellent |
| 23:33 | <gsnedders> | They're all totally out of context though |
| 23:35 | <gsnedders> | There again, that's most of the fun |
![http://images.apple.com/main/elements/spacer.gif"](http://images.apple.com/main/elements/spacer.gif"){kind=link}