00:40:22 gavin_ has joined #html-wg 00:50:34 zcorpan has left #html-wg 01:41:08 inimino has joined #html-wg 02:37:08 Zeros has joined #html-wg 02:54:52 myakura has joined #html-wg 02:57:04 marcos has joined #html-wg 03:34:06 Sander has joined #html-wg 03:48:27 hyatt has joined #html-wg 04:42:09 gavin_ has joined #html-wg 06:28:08 ROBOd has joined #html-wg 06:47:40 zdenko has joined #html-wg 06:49:42 gavin_ has joined #html-wg 06:56:08 loic has joined #html-wg 06:57:52 heycam has joined #html-wg 07:21:04 ( is a void tag) 07:22:45 anne: what are you referring to? what's treating it like applet? 07:24:25 loic has joined #html-wg 07:25:58 wow i should add applet button marquee 07:25:58 i should probably just add embed too 07:49:43 Yudai has joined #html-wg 07:49:48 gavin has joined #html-wg 07:49:51 Bob_le_Pointu has joined #html-wg 07:49:56 mw22_ has joined #html-wg 07:50:01 Philip` has joined #html-wg 07:50:02 myakura has joined #html-wg 07:50:02 Hixie has joined #html-wg 07:50:05 wilhelm has joined #html-wg 07:50:07 xower has joined #html-wg 07:50:09 Lachy has joined #html-wg 07:50:10 krijnh has joined #html-wg 07:50:11 citoyen has joined #html-wg 07:50:11 Zeros has joined #html-wg 07:50:13 jmb has joined #html-wg 07:50:13 jgraham has joined #html-wg 07:50:14 loic has joined #html-wg 07:50:16 martijn has joined #html-wg 07:50:20 hsivonen has joined #html-wg 07:50:25 heycam has joined #html-wg 07:50:33 ROBOd has joined #html-wg 07:50:47 schepers has joined #html-wg 07:50:50 inimino has joined #html-wg 07:50:51 nickshanks has joined #html-wg 07:53:58 gavin_ has joined #html-wg 07:54:21 Dashiva has joined #html-wg 07:54:28 sbuluf has joined #html-wg 07:57:28 hyatt has joined #html-wg 08:35:03 MikeSmith has joined #html-wg 08:46:52 glazou has joined #html-wg 08:46:55 hello 08:47:03 when is the BOF at Xtech ? 08:47:32 gsnedders has joined #html-wg 08:51:52 glazou - there was mention of Thursday night 08:57:25 hmmm 08:57:35 possibly Friday 08:57:36 apparently xtech BOFs are all scheduled on tuesday night 08:57:41 friday ? 08:57:46 xtech ends friday noon ! 08:58:01 see http://2007.xtech.org/public/content/2007/05/08-bofs-free 08:58:02 Friday was mentioned ... but I think many people are leaving early on Friday 08:58:05 right 08:58:09 glazou - yeah, saw that 08:58:38 I'm hosting a BoF myself on Tuesday ... on location-aware browsing 08:59:09 yep saw that 08:59:36 and Molly is having an all-day "browser summit" thing on Tuesday 08:59:44 maybe during that? 08:59:49 in the daytime 09:01:46 hmmm 09:02:00 you have a url for molly's thing ? 09:02:40 http://2007.xtech.org:80/public/news 09:02:59 http://www.molly.com/2007/05/10/blue-sky-web-browser-standards-and-interop-summit-xtech-paris/ 09:03:28 http://2007.xtech.org:80/public/content/2007/05/10-browser-summit 09:04:12 thanks 09:06:45 glazou - ask annevk when he's on ... he might know more 09:07:03 sure 09:07:11 I was expecting to see him here actually 09:07:52 I guess he is probably traveling today 09:09:10 k 10:01:00 gavin_ has joined #html-wg 10:28:48 zcorpan has joined #html-wg 11:14:02 zcorpan has joined #html-wg 11:19:31 beowulf has joined #html-wg 11:20:09 tH has joined #html-wg 11:20:16 zcorpan_ has joined #html-wg 11:35:47 dk has joined #html-wg 12:07:43 jdandrea has joined #html-wg 12:13:47 zcorpan_ has joined #html-wg 12:21:50 anne has joined #html-wg 12:23:10 for some reason I was disconnected 12:26:44 www-html is amusing 12:28:41 glazou, tomorrow is the day for discussion I think 12:28:50 glazou, and maybe during the rest of the conference too 12:28:53 Shunsuke has joined #html-wg 12:29:07 glazou, at this point it doesn't look like we'll have some specific meeting 12:29:22 glazou, apart from browser & standards & interop day tomorrow 12:55:41 ROBOd has joined #html-wg 13:03:40 anne - just glad the amusement is on www-html instead of public-html ... hopefully public-html will get a lot less amusing 13:06:27 Maybe it'd be alright if www-html were like public-html was a week ago (except with fewer people), and if public-html were like whatwg is now 13:13:18 public-html@w3.org like whatwg@whatwg.org would be most excellent 13:16:24 anne - yeah, let's work on getting it there 13:16:35 Philip` - yeah that too 13:24:49 No idea how long it'll take to get there - whatwg is at the stage of refining and fixing the features that have been added, but public-html will probably have to go through a stage of questioning all the basic decisions about every single feature 13:25:37 depending on how much effort "the new guys" put into it 13:25:50 Philip` - as long as it is a specific technical discussion of the spec itself, that would be big progress 13:25:58 (for the public-html list) 13:26:22 heh 13:26:34 i think some did discuss the spec though 13:28:22 anne - yeah, some. problem is the not-some other stuff ... 13:34:01 jdandrea_ has joined #html-wg 14:05:08 frippz has joined #html-wg 14:37:28 Shunsuke has joined #html-wg 14:40:49 billmason has joined #html-wg 14:40:51 billmason has left #html-wg 14:46:57 jdandrea has joined #html-wg 14:48:03 gavin_ has joined #html-wg 14:49:04 hasather has joined #html-wg 15:10:42 DanC_lap has joined #html-wg 15:23:54 anne: ok 15:25:00 frippz has joined #html-wg 15:25:41 myakura has joined #html-wg 15:28:47 h3h has joined #html-wg 15:57:19 Sander has joined #html-wg 16:02:13 zcorpan_ has joined #html-wg 16:04:58 dbaron has joined #html-wg 16:15:49 myakura has joined #html-wg 16:55:03 gavin_ has joined #html-wg 17:11:35 hmm.. more design principles discussion, not particularly productive 17:12:00 I suppose repeating my request to stop that isn't as useful as starting some other discusison 17:14:42 othermaciej has joined #html-wg 17:18:24 DanC_lap, the problem is, particluarly with pave the cowpaths, people don't really understand the scope of it, and so they're jumping to conclusions that it could be used to support a wide range of bad things 17:19:14 DanC_lap: sorry about sending email during the recess. I was too routinized to hit reply to all plus send. I was rather ashamed when I realized I had responded to the list that was in recess. 17:19:53 but at least John Foliot's latest mail didn't really focus on the principle, but rather somewhat productively addressed my questions. 17:25:59 not a big deal, hsivonen 17:26:55 DanC_lap: thanks 17:35:36 deltab has joined #html-wg 17:36:03 Lachy: I'm glad he responded to your points, regrettably, a lot of his reply is in vague generalities 17:36:38 I'm particularly concerned about his remark that he can't prove more semantic features are needed, but he *knows* it's true 17:37:30 truthiness? 17:37:44 yeah, I'm also concerned by him pushing the burden of proof on to others to disprove him, rather than having him prove his own claims 17:38:21 let's not spend time arguing who has the burden of proof. 17:38:38 let's just put the design principles on the back burner 17:38:48 ok, fair enough 17:39:06 design principles or no, we'll have to make decisions about specific technical issues 17:39:35 so avoiding debate on the design principles will not actually let us avoid disagreement 17:40:14 I'm having a hard time taking this suggestion seriously, but trying to respond calmly and objectively http://lists.w3.org/Archives/Public/www-html/2007May/0462.html 17:40:15 I am disappointed that the outcome of the two positive reviews of the design principles document by the designated reviewers is to ignore it for the time being 17:40:15 Ryan King suggested we start with the parsing/tree generation part of the spec, rather than just breadth-first. I find the idea appealing. What do you think, hixie? 17:40:32 Maybe it needs time for people to see decisions made about specific technical issues, and then the points made in those decisions can be generalised into design principles and people will be happier since they've seen that they work 17:40:48 that's sorta what I'm hoping, Philip`... 17:41:05 I do think we should have a review agenda, but I don't think parsing is the most interesting place to start 17:41:35 mainly because it is all about very obscure technical details and most people do not have the technical expertise to meaningfully review it 17:41:57 Parsing is a good way to scare off people who don't like going into the messy details of how browsers work and how they ignore the previous standards and disagree with other browsers 17:42:00 maciej, I can see how that's disappointing... the two positive reviews were mixed with a lot of mixed reviews, and I haven't managed to review the discussion carefully. 17:42:33 I would suggest starting with a topic that is conceptually simpler and where there might be actual design work, not just reverse engineering 17:42:49 hmm... why? 17:43:14 because, as I said, I think the vast majority of people can't usefully contribute to a review or discussion of the parsing algorithm 17:44:03 I suppose in a certain light that could be seen as an advantage 17:44:29 but people may try to contribute anyway based on uninformed opinion, in an area where opinion is not really applicable 17:44:41 yeah, I would recommend *not* starting with Parsing 17:45:01 the Parsing section is what has caused 90% of the confusion thus far on the list 17:45:15 and is very technical (and non-debatable to a point, really) in nature 17:45:20 my attempts to get this group to *not* do X have had very little success. I need a suggestion of what to do instead. 17:45:32 I would suggest starting with the definition of some set of non-controversial elements that existed in HTML 4, or with 17:45:36 I'd say start with forms 17:45:43 they're the most established 17:46:08 if we start with reviwing elements and semantics, we have to be careful to avoid bikeshedding 17:46:24 I actually think might be a good place to start 17:46:32 works for me 17:46:43 and I seriously doubt a linear traversal of the spec will be realistic 17:46:46 it's new since HTML 4, relatively self-contained, and not insanely hard to understand 17:46:59 we need someone to run an agenda 17:47:01 yeah, until some argues that HTML isn't a graphics API, and cavnas shouldn't be included 17:47:06 I'd much rather have many active threads about different parts of the spec, so people can apply their expertise in the right places 17:47:24 instead of replying to discussions that they don't understand simply because they are the only discussions happening 17:47:27 Would it be sensible to wait a week or so until Hixie's finished integrating the current canvas feedback into the spec, before having new people start reviewing it? 17:47:40 h3h: many active threads is hard on editors, hard on chairs, hard on people responsible for issue tracking, and hard on people who might want to pay attention selectively to the list 17:47:48 havnig too many threads becomes difficult to follow, particlaly for those of use who have knowledge about all of it 17:47:58 yeah... 17:48:08 "and I seriously doubt a linear traversal of the spec will be realistic" <- what does that mean? surely it's feasible to go over the spec breadth-first and collect issues, no? 17:48:13 DanC_lap: I think before raising a specific review topic, we might want to set up a basic issue tracker and designate some people to be responsible for recording issues 17:48:24 othermaciej, of course 17:48:39 DanC_lap: it means only having discussions about one topic at a time in the spec 17:49:10 the breadth-first tactic doesn't involve (much) discussion. Just raising issues. 17:49:44 that sounds great, but I think is terribly idealistic for a group of this size and varying expertise 17:49:57 everyone is going to try to throw in their 2 cents at every step of the way 17:50:09 (DanC_lap: http://lightyellow.bikeshed.org/ though you probably found it already) 17:51:51 Lachy: I should have specified that I don't necessarily think all those changes are good 17:52:01 I think this is where the group size is going to hurt the most 17:52:21 I have a big blind-spot around forms. back in 199x, I proposed finishing the HTML 1.x spec without forms. didn't fly. If we're to start with forms, I'll need Chris W. to do most of the chairing. 17:52:46 DanC_lap: raising issues ad-hoc about all sorts of areas of the spec means it is unlikely they'll be promptly addressed, compared to focusing on a particular area 17:53:03 kingryan has joined #html-wg 17:53:06 Philip`, you should have painted it blue! ;-) 17:53:45 it's hard to think of a section that would not have some sort of controversy around it 17:53:53 othermaciej, I'm not sure I see your point. It's tautological that a breadth-first approach is different from a depth-first approach. 17:54:24 DanC_lap: I think I mentioned a specific likely difference, rather than just asserting it will be different 17:54:32 DanC_lap: i think that a review of the parsing section can only be usefully done by comparing large numbers of pages as parsed by an html5 browser and as parsed by IE. It's not clear to me it's a section that can be reviewed without implementation experience. At least the 8.2 section. 8.1 (Writing HTML documents) would make sense to review. 17:54:48 I suspect these elements would have very little controversy http://www.whatwg.org/specs/web-apps/current-work/#sections (
, headings, etc.) 17:55:01 yeah, I was going to suggest 3.8 Sections 17:55:21 DanC_lap: I see that what we talked about the other day has already come up to the group at large :D 17:55:29 3.11 Lists, as well 17:55:51 and 3.15 tabular data 17:55:57 i would recommend against this week for the reason Philip` gave, actually 17:55:58 also 3.9 Prose, 3.11 Lists, or 3.14 Embedded Content (skipping and media elements as deserving their own review agenda topic) 17:56:04 (i'm still integrating past feedback) 17:56:12 sorry, *proposed* ordered agenda 17:57:01 DanC_lap, are we going to try and allocate a time limit for the review of each section, so that threads don't drag on forever? 17:57:17 4.11 Client-side session and persistent storage might be a good one too, or 4.3 Session history and navigation 17:57:31 Lachy: I like that idea in theory 17:57:34 lachy: we need a group to collect all the arguments made and summarise them 17:57:48 Lachy: it should become obvious when the comments start repeating themselves 17:57:53 Lachy: that's when to move on :-) 17:58:02 I agree with Hixie that 8.1. Writing HTML documents might also be a good early one 17:58:12 yeah, we just need someone to step in and say stop when that starts to happen 17:59:36 Lachy: well that's objectively measurable -- when the people collecting feedback have done nothing but say add "[argument also made here, here, here, here, here]" to the relevant wiki/issue page for a certain amount of time, it's done :-) 17:59:44 the certain amount of time would need establishing of course 18:00:27 Hixie, when you get a chance, can I please get whatever stats you have for code, samp, kbd, and var from you? They're asking on www-html. 18:01:35 Lachy: i'll be at work in about an hour 18:01:45 Lachy: if you're still up, remind me :-) 18:01:47 ok, I'll probably be asleep. email them to me 18:01:51 k 18:01:53 if i remember 18:01:53 I'll write you an email now 18:01:58 cool 18:02:20 anybody know Chasen Le Hara or Roman Kitainik? They offered to help with issue tracking, summarization, and clustering 18:03:25 anybody wanna hold my hand setting up bugzilla for our use? 18:03:59 we want Bugzilla? 18:04:01 ouch 18:04:52 I want roundup, but I also want something supported by the systems team. 18:05:17 W3C systems team supports bugzilla and tracker. tracker is not mature enough for my tastes. 18:05:32 what do you want, anne ? 18:06:51 you're not nervous that it hasn't been used by zillions of people? 18:07:11 it doesn't seem to keep a complete audit trail 18:07:16 dino has my faith 18:07:43 anyway, Hixie put down some requirements somewhere... 18:08:33 zcorpan_ has left #html-wg 18:08:34 I believe that (a) tracker is more hackable than Bugzilla and (b) more usable than Bugzilla 18:08:37 How about we just get something set up to get started, and then bug the systeam about new versions or other tools as needed? 18:08:40 zcorpan_ has joined #html-wg 18:09:13 picking an issue at random... http://www.w3.org/2005/06/tracker/webapi/issues/2 ... it's closed, but I don't see an audit trail event that set the state to closed 18:09:14 If that's the criteria a web page will work :) 18:09:56 DanC_lap, doesn't e-mail 4 tell you why it's closed? 18:10:50 I've been using a web page for a while. (http://www.w3.org/html/wg/il16 ). it only works if I personally am involved in every issue. It's time to move beyond that now. 18:11:02 Bugzilla doesn't track e-mails on the list and isn't very user friendly 18:11:31 Tracking all e-mails related to an issue is quite important, I think 18:11:49 did the machine somehow read e-mail 4 and automatically set the status to "closed"? 18:12:06 No, I closed it 18:12:19 tracking email related to an issue is not that hard, if people put the issue name/number in the subject 18:12:20 It does track the e-mails automatically 18:13:20 I think Bugzilla works ok for tracking bugs in software, but it doesn't seem very good issue tracking software to me... 18:13:35 one of hixie's requirements is a summary. I dunno if any of tracker/bugzilla/roundup supports an issue summary 18:13:54 http://lists.w3.org/Archives/Public/www-archive/2007May/0011.html is one of those e-mails from Hixie 18:14:21 DanC_lap: Last bullet on http://www.w3.org/2005/06/tracker/features ? 18:14:29 He lists 3 things. Bugzilla gives you about a hundred but not really a good way to do those 3 things 18:14:48 Actually, Bugzilla is pretty good at showing the current status 18:15:37 I could maybe live with tracker if several people are happy to use it. 18:15:57 Where's the source for tracker? 18:16:22 a few clicks from http://www.w3.org/2005/06/tracker/ , I assume 18:16:51 anne, when you closed that issue, did it generate an audit trail entry? 18:17:24 Hmm. Can't seem to find it... 18:17:38 nor can I... 18:17:46 And the links there keep giving me permission deniet / auth dialogs. 18:17:52 I note that the tracker todo list doesn't use tracker. http://www.w3.org/2005/06/tracker/todo 18:18:17 DanC_lap, no 18:18:28 I don't think the source code is publicly available 18:18:46 ?! 18:18:51 So “hackable” means...? :-) 18:19:21 That I got from dino 18:19:58 (You should write it in HTML5 and add ;-) ) 18:20:20 I can't at most of the pages linked here or from the tracker pages; they're all not liking my user's privileges. 18:20:34 xover: ditto for me 18:20:45 Oh, sorry, it's UTF-8 already, though some of the content is double-encoded UTF-8 18:20:49 DanC_lap: I think tracker is pretty weak - mailing list integration is the only strong point 18:21:10 hmm... I don't see the encoding bug, Philip` . clue me in? 18:21:56 I see "Asbjrn Ulsberg" 18:22:06 it doesn't have a good way to record duplicates, doesn't support milestones very well (so we couldn't clearly defer an issue to HTML6), does not separate issue description from discussion, doesn't make it easy to organize things into categories (it only has the "product" level of category).... 18:22:32 (and it becomes four characters if I switch the browser to iso-8859-1) 18:22:49 phpht. nxml-mode garbled Ulsberg's name. 18:23:09 my kingdom for consistent copy-and-paste! 18:23:17 I think it might be worth having a tracker discussion on the list, since I don't think most people read www-archive regularly 18:23:28 Character should be: ø / ø for reference. 18:23:50 zcorpan_, maybe we should make quirks mode the web and be done with it :) 18:24:13 yeah 18:24:15 zcorpan_, no doctypes, no rendering differences because of a doctype and all works sort of... 18:24:21 But then you'd break the 50% of sites that rely on standards mode 18:24:41 we'll have to spec standards mode too 18:24:45 Some other survey indicated that number to be 10%... 18:24:54 It'd be the new quirks mode :) 18:24:56 unless you redefine quirks to be standards and standards to be quirks, and require HTML5 documents to have no doctype, and say all the old HTML4 Strict and XHTML pages are legacy content 18:25:03 hopefully we might be able to merge almost with full 18:25:26 but the standards-liking people on the web wouldn't be terribly happy with that, I'd expect 18:25:37 I like standards 18:25:41 name, quirksmode fixed in 1.22 18:26:02 Interesting autocomplete. 18:26:48 halt on the first error, hah! 18:27:20 DanC_lap: Looks good to me now :-) 18:27:20 I consider it a "he who does the work makes the rules" sort of thing. But I'm low on volunteers. 18:27:23 I'd rather discuss requirements than which to use per se 18:27:34 did anyone volunteer for issue tracking? 18:27:43 yes, 6 of us... or so... 18:27:59 see http://www.w3.org/2002/09/wbs/40318/tasks83/results 18:28:22 or "simplified results" under http://www.w3.org/html/wg/#current 18:29:31 David Dailey has made some progress investigating issue tracking stuff, but I think his target is too high. 18:30:08 dbaron has joined #html-wg 18:30:10 Set up Bugzilla to get started, then start bugging systeam about making tracker public and making whatever changes we need to it for possible later migration? 18:30:54 oops; I used up the time I have. I have another meeting now. xover, you're welcome to set up bugzilla for us 18:30:55 What exactly are the goals of the tracking system? 18:31:24 ... on a sort of trial basis, xo 18:31:25 xover 18:31:28 we need to discuss goals / requirements before picking one 18:31:36 hehe, will have a look at it. 18:32:58 I strongly agree with othermaciej 18:33:22 In related news, html5lib has been partially ported to Ruby 18:33:29 http://intertwingly.net/blog/2007/05/14/Ruby-HTML5-Tokenizer 18:33:38 Anyone working on a Perl version? 18:34:06 othermaciej: are there requirmeents beyond those i posted? 18:34:21 xover: There's someone working on a C version 18:34:25 xover, not that I know. I believe someone is doing a C version though which will rule them all :) 18:34:38 It'd be interesting to do some benchmark comparisons :-) 18:34:45 Hmm. SHould be feasible to do Perl bindings on top of that, yeah. 18:35:17 It's nice how HTML5 defines the parsing algorithm and also gives a large test case to feed through a parser for benchmarks 18:35:37 Do we want access control for our issue tracker? 18:35:43 Hixie, your e-mail is not elaborate on requirements on states and doesn't detail on the goals as far as I can tell 18:36:57 anne: I'm working on a ruby version as well. I guess I should go ahead and get my working version out there 18:39:23 anne: true 18:39:25 Hixie: well, it depends on how broadly you interpret "It should allow for a clear statement of the current state of the issue" 18:39:52 would "duplicate of issue XXX", or "deferred to the next version of HTML" be possible values of current state? 18:39:59 i just meant like "open" vs "closed, no objections" vs "closed over objections", i guess 18:40:09 i guess those would be good states too 18:40:20 is there a wiki page yet with these requirements? 18:40:30 I think it is also useful to have categories 18:40:34 and queries 18:40:39 so I can check for "all open canvas issues" 18:41:02 we could start a wiki page 18:41:07 that might be more useful than starting with email 18:41:12 Hixie: would out of scope issues be considered "closed"? or should that be a different state? 18:42:15 i can make a page 18:43:24 kingryan: "closed" should presumably have reason it was resolved, such as indicating that the spec was changed, that the spec was not changed because the issue was out of scope, spec not changed because spec text was determined to be correct, etc 18:43:27 kingryan: if the target audience is the editors, then 'closed' is fine; if the target audience includes those raising the issues, then that would be useful 18:43:52 working out who the target audience is is the first step to working out the requirements 18:44:14 Hixie: that's close to what I was thinking. Managing incoming feedback by helping people find answers for themselves would be a benefit to the WG. 18:44:20 hyatt has joined #html-wg 18:44:31 I think the audience is both editors and reviewers 18:45:11 hyatt has joined #html-wg 18:46:49 http://esw.w3.org/topic/HTML/IssueTrackerRequirements 18:46:55 feel free to update 18:47:48 anne: Was working on a page in paralell, so I cut&pasted over to yours. 18:48:27 (So it probably repeats what's already there) 18:51:54 updated 18:58:16 hyatt: Do you have a W3C Bugzilla account? 19:03:04 gavin_ has joined #html-wg 19:03:06 heh 19:17:49 DanC_lap: Did a quick and dirty Bugzilla setup for this. Details on the Issue Tracker Requirements Wiki . 19:19:12 Nothing fancy, just the basics so we can have a look at it. 19:19:53 Yeah. And we may want to be more fine-grained then that in setting it up. 19:22:14 "what is this web apps 1.0 thing?" 19:22:29 :-P 19:22:38 (as anne says, there is no "Web Apps 1.0" anymore) 19:23:13 Give me a list of components and I'll set it up. 19:24:06 "The DOM", "The semantics", "Browsing Contexts", "APIs", "The Language Syntax" 19:24:11 Actually, that might as well go in the Wiki too. 19:24:12 (from the HTML 5 introduction) 19:25:34 BTW, http://www.w3.org/Bugs/Public/editkeywords.cgi lists existing keywords that may cover some of the "closed, no objection" status stuff. 19:26:33 can't access that 19:26:48 I think I'll go home 19:26:55 Hmm. Ah, it's a config page so you need extra right to access it. Pity. 19:26:58 I have to get up at 4:30AM tomorrow or something 19:27:10 Ugh. Rather you then me... :-) 19:27:20 `night 19:28:44 hmm.. IssueTrackerRequirements ... says "should". a requirement is a *must*, no? and... where's the other side of the negotiation? i.e. who proposes to meet these requirements? 19:29:01 this is not a spec 19:29:06 so no RFC2119 terminology 19:29:32 although you're free to change stuff as you see fit :) 19:29:57 WikiTitle should prolly say "Brainstorm" or summat... :-) 19:31:58 kwijibo has joined #html-wg 19:33:16 anne: I added the keywords to the Wiki if you want to have a look. 19:33:59 (and all these fields can be queried against, obviously) 19:34:54 is @profile really being dropped in html5? What will it be replaced by? 19:37:34 I sure hope @profile isn't dropped. GRDDL depends on it. 19:40:05 Lachy has joined #html-wg 19:42:13 me too 19:45:06 why was it dropped in the whatwg? http://wiki.whatwg.org/wiki/Changes_from_HTML4#Dropped_Attributes 19:45:31 As far as I'm aware, the reason for profile not being included is that microformats are meant to use it but in practice they don't (but they work alright without it anyway, and content authors aren't going to change) and presumably there hasn't been a compelling reason to include it so far 19:45:55 (I could be wrong, though) 19:46:15 eRDF and GRDDL use it though 19:46:44 and embedded semantics in html is just starting to take off 19:47:16 @profile's use is minimal, but growing, so it doesn't seem very forward looking to drop it 19:47:18 microformats only sorta work without @profile. it's more clear how they work with @profile. 19:48:00 and especially unjoined up if w3c drops @profile in x/html5 but demands it for GRDDL 19:48:52 I'd guess those would be good points to bring up in favour of adding profile and defining what it means 20:00:35 DanC_lap, re: GRDDL and microformats, with hCard at least, do you think perhaps what is needed is not a canonical @profile, but a choice of profiles? 20:01:18 or perhaps just an hCard ontology with loose semantics 20:02:59 dbaron has joined #html-wg 20:04:56 DanC_lap: are there known microformat consumers that ignore microformat data in the absence of profile=''? 20:05:08 if they only "sort of" work without profile="", someone should tell the microformats community 20:05:15 because currently, basically nobody uses profile="" 20:05:22 and those few that do, usually use it incorrectly 20:06:08 hsivonen, http://esw.w3.org/topic/GrddlImplementations is a list of bits of software that grok hCard given a suitable profile but not otherwise. 20:07:15 hsivonen: AFAIK, *all* microformat consumers will consume the mf even without the profile 20:07:30 basically, any software that assumes hCard based on class names is squatting. HTML authors are supposed to be able to use any class names they like, for whatever purpose they like. 20:07:37 fwiw, I've been working on getting proper profiles published on mf.org for hcard, hcalendar, hreview and xfolk 20:08:09 I suppose the HTML WG has the right to take back class names, but it bothers me a bit. 20:08:49 the central registry of class names seems... I dunno... boring, somehow. 20:09:36 HTML4 says class is to be used "For general purpose processing by user agents" (as well as for stylesheets) - I'm not entirely sure what means - would it include UAs processing certain class values as e.g. calendars? 20:09:38 DanC_lap: it seems that the microformat community sees more benefit in assuming that anything that looks like a hcard can be treated as such 20:10:03 yes, they do. but I don't think W3C should endorse that. 20:10:15 DanC_lap: do the Grddl implementations work with real-world microformat data out there? 20:10:29 some. 20:10:33 e.g. the XTech web site 20:11:14 DanC_lap: I don't think it would be particularly useful to not endorse how microformats are practiced 20:11:42 jdandrea has joined #html-wg 20:11:57 surely crawlers and such can have local policies that say "smells like a duck to me; close enough." but it's still useful to be able to have an explicit "I am using hCard per http://www.w3.org/2006/03/hcard " mechanism. 20:12:07 the question mark I have regarding µf and GRDDL is, if there is a canonical profile, what ontologies should be used in the mapping? 20:12:11 well, we disagree, hsivonen 20:12:29 DanC_lap: if the stuff works well enough with duck typing, what's the problem? 20:12:35 yes, that has to be answered for each microformat, kwijibo 20:12:58 DanC_lap, it's useless to try and mandate something like profile="" given the overwhelming evidence that it will fail in reality 20:13:10 DanC_lap: but some µf (well, hCard at least), are used in quite different ways 20:13:37 it's not useless to allow profile="". 20:13:38 is an hCard a vCard, a Person, an Address, a Place? 20:14:05 profile="" is already succeeding, for my purposes. 20:14:28 hsivonen: why should w3c endorse bad practice? 20:14:34 DanC_lap: I'm rather indifferent about allowing it, but I'm not keen on requiring it if I'm at the receiving end of conformance checker user complaints 20:14:57 kwijibo: if the practice works, classification as "bad" may be the thing to fix 20:15:07 DanC_lap, what use is profile if you're the only one using it? 20:15:22 hsivonen: it only works so far 20:15:40 Lachy, I'm not the only one using it, and I don't think it's a good use of my time to read the web to you. 20:16:11 kwijibo: let's fix it when it no longer works 20:16:12 what the? 20:16:29 you're jumping to conclusions, Lachy. 20:16:46 conclusions where the evidence to the contrary is available in the web. 20:16:53 DanC_lap, Hixie gave evidence to show that it's rarely used on the web, and rarely used correctly when it is 20:17:11 I'm not surprised. 20:17:45 zcorpan_ has joined #html-wg 20:17:45 s/pragmatic/'pragmatic' :) 20:18:12 DanC_lap: this is why we should agree on the principles first. not having profile="" is very much an application of the Pave The Cowpaths principle. 20:18:25 well, I expect to see some evidence to support your case, including sufficient real world examples on the web that use it correctly and implementations that require it to be used 20:18:39 (a) profiles for microformats are somewhat new, as is GRDDL, and (b) even if only a small fraction of the web uses profiles/GRDDL, it's still useful. There will always be more sloppy/unstructured stuff than formalized stuff, but the formalized stuff is still valuable. 20:18:57 profiles for microformats have existed ever since the dawn of microformats 20:19:09 the very first microformat, XFN, was introduced with a formal profile description grammar 20:19:18 Lachy, I'll repeat myself just once more: XTech and http://esw.w3.org/topic/GrddlImplementations 20:19:27 inimino has joined #html-wg 20:19:40 and profile="" failed almost overnight, with no XFN processors looking at the profile="" attribute after about a week 20:20:03 kwijibo: profile="" is not a future spec, it comes from HTML4 (198) 20:20:05 1998 even 20:20:16 so? 20:20:29 html 5 is a future specification 20:20:44 kwijibo: profile has had ample time to succeed 20:20:49 kwijibo: instead, it has failed 20:20:56 so given evidence from the past that shows little use, there's no reason that will change much in the future 20:21:02 define failed? 20:21:22 content doesn't use it and processors don't look at it 20:21:24 you could have said that of semantic markup as a whole in 2001 20:21:33 kwijibo, an abundance of pages that use microformats on the web successfully, without using profile 20:21:34 kwijibo: markup consumers seeing more value in ignoring it and producers not bothering to use it "right" 20:22:37 microformats is where the real world evidence comes from, what do you expect? 20:22:51 xfn is the most used value for profile="" 20:23:33 if we ignore microformats, we can conclude that profile="" isn't used at all in the wild :) 20:23:37 zcorpan_, that's also the default value in WordPress 20:24:15 Lachy, who gets to choose class names, authors or readers? Who gets to choose HTTP URI paths? (e.g. /robots.txt, /favico). Do you see an issue around taking choices away from content providers? 20:24:41 DanC_lap: tantek chooses :-) 20:24:49 lol 20:24:59 we like it when tantek chooses, but not when microsoft chooses. go figure. 20:25:08 we do? 20:25:18 half the features in HTML 5 are straight from microfost 20:25:20 microsoft 20:25:33 (well maybe a bit less than half. still a bit chunk) 20:26:08 DanC_lap, I'm aware that fixing the robots.txt and favico URIs was a mistake, and the favicon one has been fixed (mostly) 20:26:34 that's why we have rel=icon now 20:26:51 Lachy: but sadly from a UA POV you need to try /favicon.ico regardless :( 20:27:09 yes, rel="icon" is an improvement, but I'm still gettting zillions of /favico requests for hardly any good reason. 20:27:11 why is /favicon.ico frowned upon? 20:27:19 only for IE, that need will fade once IE fixes that bug and current versions of IE fade 20:27:22 zcorpan_: needless requests 20:27:34 ok 20:27:49 you don't need favicon.ico for IE 20:27:59 IE supports rel="SHORTCUT ICON" 20:28:01 how will IE fix that bug? they'll stop paying attention to anything but rel="icon"? they'll stop supporting /favicon.ico ? that would be a pleasant surprise. 20:28:03 Lachy: but it's needed for backwards compat, I doubt they'll drop it 20:28:23 we'll always support /favicon.ico 20:28:29 IIRC by someone at Mozilla just implementing support 20:28:53 hyatt could probably tell more. IIRC, he had something to do with the ico/bmp decoder 20:29:29 nn. plane to catch in the morning 20:29:54 zcorpan_, /favicon.ico is frowned on because the HTTP server operator is supposed to be the one who chooses what various paths mean, not the browser. 20:30:23 but I guess it's now an open marketplace; short strings are for sale to the highest bidder. 20:30:45 or the first one to get critical mass. or something. 20:31:07 even if microformats 'work' without @profile, nothing else does - get rid of @profile, and you might as well build microformats into the html5 spec 20:31:43 that seems to be the trend 20:32:07 and I'm pretty sceptical about the extent to which microformats are 'practical', and 'work' anyway. 20:32:10 I suppose it's a reasonable thing to do. But I still think @profile is worthwhile. It doesn't cost much, and it does have some value. 20:32:21 hyatt has joined #html-wg 20:33:34 despite having used µf as an html author for maybe 18 months or so, I still have never actually used them as a consumer 20:34:09 I've got zero benefit from them, and haven't seen any compelling examples of people that have 20:38:20 zcorpan_ has joined #html-wg 20:38:40 DanC_lap: tell that to the p3p folk 20:46:30 so you're inclined to start review with section 8.1. Writing HTML documents, Hixie ? I just took a quick look. it seems to invite a variety of editorial comments... 20:48:01 hyatt has joined #html-wg 20:48:34 ah... "this is why we should agree on the principles first" so you're inclined to continue that discussion. hmm. 20:49:11 I think we need more chairs if we're going to continue that discussion. or something. 20:49:33 The problem with reviewing 8.1 is that the comment about void-element slashes probably won't survive for long :-( 20:50:25 a useful property about 8.1 is that it doesn't contain any conformance requirements for UAs 20:50:55 DanC_lap: either would be fine imho. mjs also made some other suggestions of good starting points. 20:51:16 DanC_lap: at the moment i'm going through the feedback from 2005/2006 20:51:29 ...so it's easier to read for authors who can't distinguish between whether a requirement applies to authors or UAs 20:54:10 whether or not HTML5 includes profile="", I'm wonder how GRDDL will compete with other methods of extracting microformat data if it requires profile to be present, given the large amount of microformat data that exists in documents with no profile, particularly since the 2003/g/data-view profile is not the one defined by the microformats themselves 20:54:21 s/I'm wonder/I wonder/ 20:55:24 it's pretty weird that to parse XFN, it requires the GRDDL profile URI, not the XFN profile URI 20:56:14 that would be wierd. it's not the case, though. 20:56:28 really? I don't see any mention of the XFN profile URI in the xfn example 20:56:42 maybe I'm looking at an old draft 20:57:16 http://www.w3.org/TR/grddl/ 20:57:20 that seems like the current version 20:58:01 yes, http://www.w3.org/TR/grddl/#profile-bind . friends.html only bears the xfn-workalike profile 20:58:07 kingryan has joined #html-wg 20:59:01 the test case is http://www.w3.org/2003/g/td/friends.html 20:59:19 the only profile I actually see mentioned is http://www.w3.org/2003/g/data-view 20:59:27 I guess I'm confused about how this works 20:59:29 (in the spec) 20:59:44 yeah, it's kinda confusing. 21:00:41 hasather has joined #html-wg 21:00:56 a document can refer to a GRDDL transformation either directly (using the http://www.w3.org/2003/g/data-view profile) or indirectly, a la the XFN-workalike case. 21:01:19 would GDDDL work with the standard XFN profile that would be used by existing XFN content? 21:01:28 (i.e. http://gmpg.org/xfn/11 ) 21:01:39 it would if http://gmpg.org/xfn/11 were modified a bit 21:04:09 so it couldn't parse existing microformat content, even if it used the right profile URI, unless the content at the relevant profile URI were changed to have a transformation link (which is admittedly much more reasonable than requiring the content to be changed) 21:04:18 right. 21:04:57 GRDDL won't get any RDF data out unless somebody explicitly put it there, whether directly or indirectly. 21:05:12 http://www.456bereastreet.com/archive/200705/another_look_at_html_5/ # Roger Johansson sort of retracts previous sort of rant on html5 21:07:09 the whole GRDDL/microformats thing is not really my top priority today; I don't have enough discipline to leave it alone, though ;-) 21:07:33 I'm owe the WG something about what we *should* talk about this week 21:07:41 I think that is a reasonable technical approach, although it limits the applicability of GRDDL tools as current examples of microformat parsers 21:07:53 yes, the applicability is limited. 21:07:54 I'm not that interested in the profile attribute one way or another 21:09:23 DanC_lap: maybe a wiki page of agenda topics? would be useful even if you already have a good choice in mind 21:09:44 I think most review topics raised today would be good, except parsing generally (including the UA requirements) 21:11:03 gavin_ has joined #html-wg 21:11:27 I didn't make a new wiki page, but I dumped some agenda topics into http://esw.w3.org/topic/HtmlTaskBrainstorm 21:12:25 ah, cool 21:13:33 I still prefer to focus on parsing. I think it's time to explore the nitty-gritty detail end of the spectrum. But not many others seem to agree. 21:14:00 well, it depends what you mean by "explore" 21:14:08 how would you envision a review of section 8.2 going? 21:15:17 my favorite idea is to ask people to raise issues a la "I disagree with what section X.Y says about the attached document, because..." 21:16:27 maybe it would help me understand what you had in mind if you could do an example? like, what kind of issue in that section would you yourself disagree with and how would you raise the issue? 21:16:30 I would have a hard time reviewing 8.2 other than to try to implement and test on existing content 21:17:20 (though I do know of a tiny handful of issues based on experience w/ parsing issues on existing content, like the fact that the meta charset preparser probably needs to scan the whole document, not a fixed number of bytes) 21:17:53 yeah, my head is in the same about that issue for now. i think someone already raised it though so eventually we'll have to fix it. 21:18:19 so that's one example... something where the charset isn't in the 1st 512 bytes, I guess. 21:18:35 I was thinking about the cases where IE doesn't make a tree. 21:18:58 DanC_lap: that's easy to respond to -- "there's no evidence that charsets are provided later than the 512th byte on a statistically significant number of pages" 21:19:17 my notes also say "Gecko coerces unknown tags to upper case, Opera doesn't" 21:20:18 ok, that response makes sense; i.e. it's easy enough to imagine chairing that discussion. 21:20:18 gecko doesn't coerce tags to uppercase. that's a DOM thing, not parser. 21:20:22 (and already covered by the spec) 21:20:57 well, my point is i would hate for lots of issues to be raised without supporting evidence 21:21:02 in that way 21:21:23 because then when the editors get around to going through them they'd all just get closed with "no evidence" 21:22:43 I don't know either of these issues in detail, but... 21:23:15 you hand othermaciej both said the charset parsing issue needs work. 21:23:18 I don't necessarily mind that, but I think it would be a more positive experience for the group to start out reviewing something less technically demanding 21:23:37 we've done plenty of "less technically demanding" for a while, for my tastes. 21:24:32 there are only a handful of people who can make useful comments on the parsing section 21:24:33 honestly the details of parser error handling are extremely hard, extremely boring, and mostly only deeply relevant to implementors 21:24:35 well for the charset thing i happen to know that browsers don't do what the spec says and i need to do some research to work out if they actually need to not do what the spec says 21:24:37 what is it that has you saying "eventually we'll have to fix it" about charset stuff, Hixie ? 21:24:38 but it's hard 21:25:07 extremely hard, extremely boring, and our job, IMO. 21:25:17 well i'm all in favour of people reviewing this section 21:25:34 but i have no patience for uninformed feedback on this section 21:25:37 DanC_lap: most people on the list aren't implementors, and aren't test case writers either, unless i'm mistaken 21:25:39 just saying :-) 21:25:57 Hixie, if you can just ignore stuff you have no patience for, that would probably be OK. 21:26:07 fair enough 21:26:41 i guess in practice it wouldn't really affect me and hyatt anyway, if we're just going to go through the issues once they've been distilled into bug reports in bugzilla 21:26:50 the parsing section does have the advantage that it's relatively objective compared to other things, but that's only helpful if people respect that 21:27:16 DanC_lap: if you want people to make useful comments about the parsing section, you would have to educate them to write test cases first (which might be a good thing), or have them implement it. just commenting on it without doing any of those things is a waste 21:27:21 i expect to put in place some mechanism where people can escalate to the chair if they expected to hear from the editor and didn't. So it'll be up to me to explain why the editor didn't answer (or to prompt the editor to answer) 21:27:52 certainly i'd be happy to answer to any e-mail that people think i should have replied to but didn't 21:27:57 indeed, I'd expect people to do test cases or write code. I'm kinda bored with anything else. 21:28:44 oh me too, but i don't expect everyone to :) 21:29:00 some just want to write tutorials 21:29:05 for instance 21:29:32 a tutorial would be fine. lots of people have indicated an interest in doing it, but nobody has done it. (or at least: nobody has showed it to me if they have) 21:30:02 right 21:30:30 what would you write a tutorial on, at this stage? writing tests? 21:30:53 canvas, wf2 21:31:02 writing html in general 21:31:27 ah 21:31:46 dbaron has joined #html-wg 21:33:46 kingryan has joined #html-wg 21:34:00 DanC_lap: tutorials that i know of: http://dev.opera.com/articles/view/improve-your-forms-using-html5/ http://developer.mozilla.org/en/docs/Canvas_tutorial 21:35:15 I was thinking it'd be useful to have a tutorial for making animated interactive things (mainly games, I guess) with canvas, since it's not covered by what's on MDC 21:35:42 and the point of canvas is that it's dynamic - it's not much fun if all you're doing is drawing some lines and gradients 21:36:07 zcorpan_: this was the first tutorial I read on wf2 http://olav.dk/wf2/demo/default.asp 21:36:46 But then I was thinking I've got to write about another three thousand words for some project I should have finished weeks ago, so the thought left my mind... 21:36:50 beowulf: ah, right. more like a demo though 21:37:36 it worked for me :) 21:38:54 yeah 21:39:38 i guess opera/dashboard widget tutorials might cover some html5 things too 21:39:43 34 volutneers for the tutorial task. that's maybe enough for its own mailing list. 21:40:18 tutorials for canvas, forms would be interesting candidates for June WDs 21:41:37 http://wiki.whatwg.org/wiki/HTML5_Tutorial 21:46:01 which would really be an xml + namepsaces tutorial 21:53:58 http://www.accessifyforum.com/viewtopic.php?t=8083 22:06:40 Zeros has joined #html-wg 22:14:20 [13:49] DanC, seems some if not all is at http://www.w3.org/2005/06/tracker/src/ . 22:15:09 [13:51] DanC, trackbot is in a different location. 22:44:14 DanC_lap: That's also ACL'ed. 22:55:37 tH_ has joined #html-wg 23:08:14 hasather has left #html-wg 23:17:37 gavin_ has joined #html-wg 23:31:30 kingryan has joined #html-wg 23:52:10 <[TCP-IP]> [TCP-IP] has joined #html-wg 23:52:16 <[TCP-IP]> [TCP-IP] has left #html-wg 23:53:57 kingryan has joined #html-wg 23:56:28 heycam has joined #html-wg