| 00:37 | <Hixie> | annevk2: thanks |
| 06:16 | <roc> | what should cloning a media element do? |
| 06:16 | <roc> | should it clone the current source, current time, and all other internal state? |
| 06:25 | <cardona507> | mpilgrim - diving into HTML 5 - great article |
| 06:31 | <cardona507> | ^book |
| 06:34 | <roc> | mpilgrim: in the canvas chapter, where you mention 3D, you might want to mention WebGL |
| 06:49 | <MikeSmith> | roc dropped off.. I was going to say that as far as mentioning WebGL it could help a lot of Mozilla and/or Khronos actually published some public information about it somewhere |
| 06:49 | <MikeSmith> | last I looked I could not find anything |
| 06:50 | <MikeSmith> | just a high-level overview page somewhere, and maybe some kind of status update about the progress of the discussions |
| 06:50 | <MikeSmith> | and some current list of organizations who are involved in the discussions |
| 06:50 | <MikeSmith> | of course what would really help is if the draft spec and the discussions themselves were actually public |
| 06:51 | <MikeSmith> | but so far there doesn't seem to be much indication that's actually going to happen any time soon |
| 06:51 | <MikeSmith> | (unless I missed some recent news) |
| 06:56 | <zcorpan> | MikeSmith: feel like fixing <footer> in v.nu? :) |
| 06:59 | <virtuelv> | I really, really, really wish Google stopped treating me as a lab rat on the search results page |
| 06:59 | <othermaciej> | MikeSmith: Khronos doesn't use a public process, sadly |
| 07:00 | <othermaciej> | the Apple folks in the relevant Khronos group seem to think it's going ok |
| 07:00 | <zcorpan> | http://bugzilla.validator.nu/show_bug.cgi?id=646 |
| 07:00 | <othermaciej> | I do wish we'd used a standards org with public process for this |
| 07:05 | <MikeSmith> | I'm glad to hear it's going OK at least. I think it will prove to be something extremely useful to developers |
| 07:11 | <MikeSmith> | othermaciej: I guess I would just hope that those involved who have worked to help for more openness at the W3C and elsewhere (instead of on member-only lists and such) would also make an genuine effort to push for similar change for real openness there |
| 07:12 | <othermaciej> | MikeSmith: Chris Marrin of Apple is the editor of the spec, and Oliver Hunt and a few others from Apple are also involved |
| 07:12 | <MikeSmith> | zcorpan: can you remind me what needs to be fixed around footer? is there an open bug? if it's an easy one I can try to get it fixed fast |
| 07:12 | <othermaciej> | MikeSmith: I expect changing Khronos would be hard - I wish a different venue had been chosen |
| 07:12 | <MikeSmith> | othermaciej: ah, cool. that's good to know |
| 07:13 | <othermaciej> | Khronos also makes it very likely that Microsoft will not participate or ever adopt the spec |
| 07:13 | <Hixie> | othermaciej: seems like it'd be pretty easy if opera, google, mozilla, and apple all said "ok let's make this a public process or we quit" |
| 07:14 | <othermaciej> | Hixie: Mozilla and Google were the main ones who pushed for the spec to be done using Khronos's known non-open process, IIRC |
| 07:14 | <othermaciej> | but I should probably join the list and get a handle on what's up |
| 07:14 | <Hixie> | i'd be mildly surprised if google was, but ok |
| 07:14 | <Hixie> | and i don't get why mozilla would want a closed process |
| 07:14 | <Hixie> | but i know they were pushing khronos |
| 07:15 | <MikeSmith> | othermaciej: if there wasn't open discussion about the choice of venue before that choice was made -- that is, if the choice was basically made unilaterally without much discussion among the potential implementors and other stakeholders... well, that is pretty disheartening on its own |
| 07:16 | <othermaciej> | MikeSmith: as far as I can tell, it was mainly decided by Mozilla (possibly also Google but not sure) |
| 07:16 | <othermaciej> | MikeSmith: by the time they talked to us about it, they'd already talked to Khronos and decided it was the way to go |
| 07:17 | <zcorpan> | MikeSmith: change content model to the same as <header>. don't know if there's a bug, but iirc hixie annotated the checkin as affecting validators |
| 07:17 | <othermaciej> | we'd have preferred open process, but Apple participates in Khronos a lot already so we can't really call it a bad venue |
| 07:17 | <MikeSmith> | othermaciej: hmm. that seems really unfortunate but I guess I should shut about it |
| 07:17 | <MikeSmith> | *shut up about it |
| 07:17 | <othermaciej> | the risk I'm afraid of is too many graphics experts and not enough web technology experts |
| 07:18 | zcorpan | embeds a youtube video in http://blog.whatwg.org/vim-checker with just <embed> |
| 07:18 | <MikeSmith> | othermaciej: that doesn't seem like so big a risk.. I guess the quality of the representation is more important then the quantity |
| 07:18 | <MikeSmith> | and sounds like you guys have the right people in their at least |
| 07:19 | <zcorpan> | Hixie: the markup youtube provides for embedding videos could be simplified |
| 07:20 | <Hixie> | not surprised |
| 07:21 | <hsivonen> | http://codinginparadise.org/weblog/2009/09/new-svg-web-release-owlbear.html says: "SVG itself stands for Scalable Vector Graphics, an open standard that is part of the HTML 5 family of technologies for interactive, search-engine friendly web vector graphics." |
| 07:21 | <othermaciej> | MikeSmith: I'm told that at times it's hard to get people to really understand that APIs that are part of the Web platform need to be safe - you can't let bad calls cause leaks or crashes |
| 07:21 | <othermaciej> | MikeSmith: classically, OpenGL counts on the application developer to get things right |
| 07:21 | <tantek> | HTML 5 "family of technologies" ? |
| 07:22 | <MikeSmith> | zcorpan: very cool video |
| 07:22 | <othermaciej> | HTML5 is a very flexible buzzword |
| 07:23 | <MikeSmith> | zcorpan: much better then the crappy screenshot I had there before that the dog ate |
| 07:23 | <MikeSmith> | the term HTML5 has market value.. other technologies should be happy to be associated with it |
| 07:24 | <zcorpan> | it seems that every five years or so, a new buzzword appears that means the same thing: DHTML, Ajax, HTML5 |
| 07:24 | <tantek> | MikeSmith, has "HTML5" superceded "W3C" as the buzzword of choice? |
| 07:24 | <zcorpan> | was W3C ever a buzzword? |
| 07:25 | <hsivonen> | I guess "the Open Web Platform family of technologies" would be a bit long |
| 07:25 | <hsivonen> | zcorpan: it was when the WaSP started |
| 07:25 | <tantek> | e.g. SVG was part of the W3C family of technologies for interactive search-engine friendly web vector graphics |
| 07:25 | MikeSmith | considers how to finesse wording for associating client-side XSLT with HTML5... |
| 07:26 | <tantek> | zcorpan W3C was a buzzword as long as most of the work that came out of W3C was for the Open Web Platform in practice |
| 07:27 | hsivonen | wonders if SVG/HTML5 is a bit like GNU/Linux from the SVG POV |
| 07:27 | <MikeSmith> | othermaciej: counting on application developers to get things right doesn't seem to be the safest approach for the Web.. I'm reminded of the current discussion about mutex-like locking and also Crockord's comment "most of the world's body of JavaScript programs is crap" |
| 07:27 | <tantek> | when the work became primarily about "enterprise/intranet" technologies (e.g. XML*** family) rather than the public world wide web, you might say the term W3C lost favor among those who worked hard for the actual open web. |
| 07:28 | <tantek> | hsivonen - originally SVG was one of many efforts that was supposed to obsolete/kill HTML. |
| 07:28 | <hsivonen> | tantek: not just enterprise/intranet but also mobile not-the-real-Web and Semantic Web |
| 07:29 | <hsivonen> | tantek: interesting. SVG seems to be in a different category |
| 07:29 | <othermaciej> | MikeSmith: browsers simply can't offer APIs that let Web content cause crashes or hangs; but I'm sure it will all be worked out, it just takes a bunch of explaining at times |
| 07:29 | <hsivonen> | tantek: though the API stuff occasionally seems like ignoring what was already there |
| 07:29 | <tantek> | hsivonen - the mobile companies weren't fooled for long, and switched toward actual open web technologies quite quickly |
| 07:30 | <MikeSmith> | tantek: I guess I think in general that the name of any one particular organization is not the best term to use as a term for promoting a particular set of technologies |
| 07:30 | <tantek> | MikeSmith - it should have been, given that W3C's mission was/is to lead the web to its full potential. |
| 07:31 | <hsivonen> | tantek: it turns out the Web doesn't mean what it seems to mean |
| 07:31 | <MikeSmith> | othermaciej: I think the basic explanation to them should be, don't assume this stuff will be used safely and responsibly by thoughtful programmers who actually know what they are doing and won't shoot themselves in the foot |
| 07:32 | hsivonen | has discovered that the W3C's mission statement is very unhelpful as a frame of discussion on what the W3C should do |
| 07:32 | <tantek> | hsivonen - "the Web" means *the* Web, it doesn't mean a bunch of little closed webs (intranets), it also means the *existing* Web, not some future parallel universe all-cleaned-up Web. |
| 07:32 | <hsivonen> | tantek: oh, I agree with you |
| 07:34 | <othermaciej> | tantek: WAP and the rainbow spectrum of Mobile Profiles of things aren't quite dead yet... |
| 07:35 | <tantek> | othermaciej - the mobile profiles that diverge from actual open web technologies are dead enough. does iPhone support WAP 1.0? |
| 07:35 | <MikeSmith> | tantek: someone I know once said to me, "Standards bodies don't age well." ... it is perhaps to be expected that any given standards body that's around for long enough is going to eventually do some things to seriously disappoint some of the very people who helped to build it and make it successful |
| 07:35 | <othermaciej> | tantek: iPhone does not support that stuff (as far as I know) |
| 07:35 | <Hixie> | i think standards bodies are an outdated concept |
| 07:35 | <tantek> | othermaciej - right. and BlackBerry barely does. |
| 07:36 | <othermaciej> | we do have vendors who want to ship WebKit with stuff like WAP, WCSS, XHTML MP, ESMP, etc bolted on |
| 07:36 | <othermaciej> | apparently due to residual carrier demand in some markets |
| 07:36 | <hsivonen> | othermaciej: do you accept patches? |
| 07:36 | <MikeSmith> | othermaciej: the markets in question include mainland China |
| 07:36 | <othermaciej> | hsivonen: we have accepted patches for some of that stuff (as long as it's off by default and minimally intrusive to core code) |
| 07:36 | <hsivonen> | othermaciej: isn't XHTML MP support in a backwards-compatible way incompatible with the Web? |
| 07:37 | <othermaciej> | MikeSmith: indeed, that is the main one cited |
| 07:37 | <tantek> | MikeSmith - I think it's the difference in this case between those who helped build it and make it successful, and those who pay membership dues and thus expect a stronger impact upon the steering of goals, resources, often divergent from the former group. |
| 07:37 | <othermaciej> | hsivonen: I think that's true for ESMP as well (it's really a "profile" of both the DOM and ECMAScript) |
| 07:37 | <hsivonen> | IIRC, the mobile legacy puts pressure against parsing application/xhtml+xml as XML |
| 07:37 | <MikeSmith> | othermaciej: so I guess it could be argued it's a fairly pragmatic concern given the size of that market |
| 07:37 | <tantek> | Hixie, WHATWG is effectively a standards body - it just has different rules, procedures, culture, principles etc. than W3C. |
| 07:38 | <othermaciej> | hsivonen: I strongly urge people not to ship the stuff that incompatibly changes standard Web APIs |
| 07:38 | <tantek> | a W3C was/is different from IETF. |
| 07:38 | <Hixie> | tantek: whatwg is a mailing list. |
| 07:38 | <tantek> | Hixie, WHATWG *has a* mailing list. |
| 07:38 | <tantek> | it's still a *WG* |
| 07:38 | <Hixie> | tantek: sure, working groups aren't going anywhere |
| 07:38 | <othermaciej> | I think standards bodies may be necessary for competing companies to be able to legally collaborate without it being an antitrust violation |
| 07:39 | <Dashiva> | What does @role do that @class doesn't except assume support for lots of predefined values... |
| 07:39 | <tantek> | othermaciej - indeed, and the more open the standards body, the better. |
| 07:39 | <zcorpan> | hsivonen: indeed |
| 07:39 | <Hixie> | tantek: but it's not a standards body by any definition anyone would have given 5 years ago. i agree that it's what should happen to standards bodies, though |
| 07:39 | <tantek> | Hixie, neither is microformats.org |
| 07:39 | <Hixie> | tantek: indeed |
| 07:40 | <hsivonen> | Dashiva: @role is sensitive to accessibility API mapping. @class isn't. |
| 07:40 | <Dashiva> | hsivonen: But only because we declare it to be |
| 07:40 | <othermaciej> | I would say an organization creating a standard is a "standards body", no matter how informally constituted |
| 07:40 | <tantek> | othermaciej - indeed |
| 07:40 | <hsivonen> | Dashiva: well, initially everything is only because so declared to be |
| 07:40 | <othermaciej> | after all, for a long time the IETF had no legal existence whatsoever, and even now, its only formal existence is through association with other bodies like ISOC |
| 07:40 | <hsivonen> | Dashiva: then the legacy requirements kick in |
| 07:41 | <hsivonen> | Dashiva: with @role and ARIA, the legacy has already shipped |
| 07:41 | <tantek> | othermaciej - "formal existence" may not be necessary for success. |
| 07:41 | <Dashiva> | hsivonen: Indeed. But the talk about using the full XHTML2 @role with all its vague uses doesn't have legacy |
| 07:41 | <hsivonen> | tantek: does microformats.org exist formally? |
| 07:41 | <othermaciej> | tantek: that may be so, although patent law and the paranoia about it can complicate matters |
| 07:42 | <hsivonen> | Dashiva: correct |
| 07:42 | <Dashiva> | So I'm wondering what's the difference between XHTML2 @role and our previously rejected predefined class names |
| 07:42 | <tantek> | othermaciej - indeed, and it's good that other orgs have done work (CC PD/CC0, W3C/IETF RF) that can be re-used. |
| 07:43 | <othermaciej> | IETF does not have an RF IPR policy |
| 07:43 | <othermaciej> | technically they are under RAND with disclosure requirements (though I think individual Working Groups can choose to chart an RF course) |
| 07:43 | <tantek> | hsivonen microformats.org is an all volunteer community that uses a website (wiki), an IRC channel, and a few mailing lists. not sure if that fulfills anyone's definition of "formal" or not, nor whether it matters whether it does or not. |
| 07:43 | <hsivonen> | Dashiva: XHTML2 @role takes non-predefined CURIEs |
| 07:44 | <hsivonen> | Dashiva: so XHTML2 @role is like the old HTML5 class with predefined names |
| 07:44 | <othermaciej> | role was originally considered to have predefined values plus open-ended non-predefined values |
| 07:44 | <tantek> | othermaciej: IETF RFC3667 & RFC3668. |
| 07:44 | <hsivonen> | Dashiva: where the XHTML2 predefined roles don't have colons and the non-predefined roles are CURIEs |
| 07:44 | <othermaciej> | people have at times argued that even the non-predefined values of @role are "more semantic" than any @class value |
| 07:45 | <Dashiva> | Well, take JF's statement that @role could be used instead of accesskey |
| 07:45 | <tantek> | features and functionality are as semantic as you define them to be. it's fairly pointless to argue that use of one attribute is "more semantic" than another in a vacuum. |
| 07:45 | <hsivonen> | maybe a useful definition of formal existence is whether it's a non-natural person that can hold copyrights |
| 07:45 | <Dashiva> | Wouldn't that require predefined roles? |
| 07:45 | <hsivonen> | Dashiva: landmarks? |
| 07:46 | <Dashiva> | No, he was talking about it in the XHTML2 context |
| 07:46 | <hsivonen> | Dashiva: I haven't read any email from JF this morning yet |
| 07:46 | <othermaciej> | Dashiva: that would require predefined roles |
| 07:46 | <tantek> | once XHTML2 went down the CURIE/namespace path for the role attribute, I gave up on it. |
| 07:46 | <othermaciej> | Dashiva: it also doesn't make sense for most application-level keyboard shortcuts |
| 07:46 | <tantek> | it is good to see that ARIA did not adopt that mistake |
| 07:47 | <tantek> | keyboard shortcuts are an *exceptionally* difficult problem to solve |
| 07:47 | <hsivonen> | tantek: it took some effort to get ARIA out of that hole |
| 07:47 | <tantek> | hsivonen - thanks for the clarification, I did not realize that. |
| 07:47 | <Dashiva> | jgraham: Good job on finding that summary table, by the way. It is indeed quite absurd. |
| 07:48 | <tantek> | it's amazing how complex technologies that raise the barrier for ease of use could possibly be seriously considered for a technology that supposed to *increase* accessibility |
| 07:48 | <tantek> | no amount of handwaving that a "role" attribute could solve/replace keyboard shortcuts is going to make it true |
| 07:49 | <Dashiva> | A lot of specs would be gone if handwaving had been outlawed :) |
| 07:50 | tantek | can say this from having done *a lot* of work to replace/upgrade "accesskey" with a more flexible/platform-useful "key-equivalent" property. look for "key-equivalent" in this document: http://www.w3.org/Style/css3-updates/css3-ui-comments |
| 07:50 | <hsivonen> | "[UPDATE: 'key-equivalent' has been dropped from this version of CSS3-UI.]" |
| 07:51 | hsivonen | agrees it is a hard problem |
| 07:51 | <tantek> | precisely, hard enough to drop and not have it hold back other features. |
| 07:51 | <othermaciej> | tantek: I don't see anything in either of those RFCs requiring RF patent licensing - the latter specifically mentions that in your disclosure you can state an intent to license a patent under royalty-bearing RAND terms |
| 07:51 | <tantek> | othermaciej - said nothing about *requiring* |
| 07:52 | <tantek> | simply *having* an RF policy |
| 07:52 | <tantek> | so yes, IETF does *have* RF policy |
| 07:52 | <tantek> | they just don't *require* it |
| 07:52 | <tantek> | others can require it though |
| 07:52 | <othermaciej> | it has an IPR policy |
| 07:52 | <tantek> | which includes RF as an option |
| 07:53 | <othermaciej> | sure, but anyone is always free to license their IP under RF terms |
| 07:53 | <tantek> | that's undefined |
| 07:53 | <othermaciej> | or at least, I've never heard of a standards body that required you to charge royalties for any patents you hold |
| 07:53 | <tantek> | "RF terms" is specific to each standards body |
| 07:53 | <tantek> | that's bother to define it |
| 07:54 | tantek | wishes there was Creative Commons equivalent for patent IPR policy. |
| 07:54 | <hsivonen> | tantek: isn't that what the Open Web Foundation is for? |
| 07:54 | <Hixie> | open web foundation isn't going to be doing it as far as i can tell |
| 07:54 | <tantek> | so you could say rel="patent-policy" href="http://creativecommons.org/rf/1"; etc. |
| 07:55 | hsivonen | really wishes there were a reusable RF policy--not a Creative Commons catalog of zillions of different policies to choose from |
| 07:56 | <tantek> | hsivonen, yes OWF is trying to solve this problem with a contribution template/boilerplate for open web technologies. |
| 07:56 | <othermaciej> | the IETF's definition of royalty-free terms is much less precise than the W3C's |
| 07:56 | <hsivonen> | Hixie: what is the OWF doing these days, then? |
| 07:56 | <tantek> | hsivonen, Hixie might not be the right person to ask, he quit the OWF. |
| 07:56 | <othermaciej> | they do say there is a preference for royalty-free, which I guess is a policy |
| 07:56 | <tantek> | othermaciej - agreed their terms are different. |
| 07:56 | tantek | agrees with hsivonen in wishing for *a* reusable RF policy. |
| 08:01 | tantek | has asked Creative Commons for this (a reusable RF policy) publicly, several times in open Q&A. |
| 08:01 | <tantek> | (over the years) |
| 08:04 | <tantek> | speaking of open web - anyone hear speak French? there's quite a conversation going (in French) in #openweb |
| 08:04 | <tantek> | s/hear/here |
| 08:04 | <hsivonen> | It seems to me that CC isn't really about Royalty Free or about Freedom in the FSF sense. |
| 08:04 | <Dashiva> | karl does, but he might not be around |
| 08:04 | <Dashiva> | karlcow/karlushi |
| 08:04 | <hsivonen> | CC is pushing for NonCommercial and MoreRights |
| 08:05 | <hsivonen> | which are about easy royalty |
| 08:05 | <hsivonen> | not royalty-free |
| 08:05 | <tantek> | hsivonen - the Science Commons effort is working toward RF-like language |
| 08:05 | <tantek> | it's the closest of the CC efforts towards RF |
| 08:05 | <tantek> | but it doesn't go far enough (IMHO) |
| 08:05 | <hsivonen> | also, Lessig's talks lately have focused on the ease of clearing rights for remixes--not about making remixes simply allowed without rights clearing |
| 08:05 | <othermaciej> | IPR policies are different from license agreements though |
| 08:06 | <othermaciej> | you can't just slap one on your work unilaterally |
| 08:06 | <tantek> | othermaciej - why not? |
| 08:06 | <othermaciej> | participants have to actually agree to it as grounds for participation |
| 08:06 | <tantek> | othermaciej - as they do for copyright assignment as well |
| 08:06 | <tantek> | no different |
| 08:07 | <othermaciej> | copyright assignment is a different issue than license agreement |
| 08:07 | <tantek> | not really. both are IP assignment. |
| 08:07 | <othermaciej> | there again, you need active assent of the participant, not just a passive license statement |
| 08:07 | <othermaciej> | a license agreement is not an IP assignment at all |
| 08:08 | <othermaciej> | it grants you a license for a work under certain conditions |
| 08:08 | <tantek> | a license is a form of IP assignment. |
| 08:08 | <tantek> | it's not all, but it is some |
| 08:08 | <tantek> | the mechanisms are similar enough that a rel to external document solution should work for both |
| 08:09 | <Hixie> | hsivonen: they're making a patent policy for specs made in a few months by small groups with the patent holders being the people who wrote the spec |
| 08:09 | <othermaciej> | I don't think a rel link in a document would create a legally binding requirement to assign copyright or disclose or license patents |
| 08:10 | <hsivonen> | Hixie: are comments by people who aren't spec editors left as an open IPR injection vector? |
| 08:10 | <othermaciej> | organizations that use copyright assignment make contributors actually sign a contract |
| 08:10 | <Hixie> | hsivonen: i don't think any proposals by anyone has any effect on that |
| 08:10 | <hsivonen> | Hixie: that sucks |
| 08:11 | <Hixie> | hsivonen: not really |
| 08:11 | <tantek> | othermaciej - I'm not sure either, but IANAL, nor are you, so all I'm saying is that it *seems* possible - and that therefore it is worth some CC/open-content/standards lawyer's time to try to make it so. |
| 08:11 | <Hixie> | hsivonen: that's a non-issue in practice |
| 08:11 | <tantek> | hsivonen - it's definitely a problem with mailing lists |
| 08:11 | <Hixie> | hsivonen: patent violations come from primary participants and from people who had patents you didn't know about |
| 08:11 | <tantek> | where there is no IP requirement for contribution |
| 08:12 | <Hixie> | hsivonen: not from people who have patents and try to inject them into random specs |
| 08:12 | <tantek> | Hixie - IIRC the Sun/XLink/XPointer problem could be characterized as injection from holder. |
| 08:13 | <Hixie> | a non-wg member sent comments that caused the spec to change to intentionally infringe the commentor's patents? |
| 08:16 | <MikeSmith> | zcorpan: I checked in the footer change to align with whatwg r3751 and tested it and I just now synced up qa-dev.w3.org:8888 to it for testing. Please check and let me know if anything appears broken. |
| 08:20 | <Hixie> | holy crap, paul posted to the list |
| 08:21 | <tantek> | Paul does good work. |
| 08:56 | <hsivonen> | sigh. http://en.wikipedia.org/w/index.php?title=HTML&diff=312446147&oldid=prev |
| 08:57 | <erlehmann> | hsivonen, but HTML 5 is official lingo, isnt it? |
| 08:59 | <hsivonen> | erlehmann: the vote said "HTML 5" but now even the W3C Editor's Draft says "HTML5" |
| 09:00 | <hsivonen> | othermaciej: is losing the space something that can be made official by lazy consensus? |
| 09:01 | <erlehmann> | hsivonen, impeach the editor ;) |
| 09:01 | <hsivonen> | Super Friends asked for the space to be dropped, too. |
| 09:03 | <othermaciej> | hsivonen: the whatwg draft seems to be consistently spaceless, as is the latest Editor's Draft |
| 09:04 | <othermaciej> | hsivonen: I don't think anyone is asking for the space back, so we could make it a WG decision, but I'm also not sure it is worth the WG's time |
| 09:04 | <othermaciej> | hsivonen: I don't know if a WG vote would be binding on Wikipedia in any case |
| 09:04 | <othermaciej> | er, WG consensus resolution |
| 09:06 | <annevk2> | I think there will be a lot of confusion about microformats being layered on top and not being directly associated with the markup. |
| 09:07 | <tantekc> | annevk2 why? |
| 09:08 | <Hixie> | are we seriously discussing the consensus on whether there should be a space between the L and the 5 in "HTML5"? |
| 09:08 | <tantekc> | existing success of microformats would seem to refute that hypothesis |
| 09:08 | <annevk2> | sorry, microdata |
| 09:08 | <annevk2> | such similar names :) |
| 09:08 | <hsivonen> | annevk2: the spec needs to say something to make that clear |
| 09:09 | <tantekc> | annevk2 - I think microformats have proven out that simple microdata will likely be a success as well. |
| 09:09 | <annevk2> | e.g. on the list people talk about adding semantics by using microdata |
| 09:09 | <annevk2> | but things like <dialog> cannot be introduced with microdata |
| 09:10 | <tantekc> | annevk2 - not true, in fact, there has been work on microformats for conversations |
| 09:11 | <tantekc> | which I expect to do a better job than <dialog> |
| 09:11 | <annevk2> | how do you handle arbitrary markup? |
| 09:11 | <annevk2> | e.g. images, boldening, links, etc. in a phrase a person just typed |
| 09:12 | <erlehmann> | othermaciej, nothing is binding for wikipedia. |
| 09:12 | <tantekc> | annevk2 - based on what real world evidence are you asking/implying that handling arbitrary markup is necessary? |
| 09:12 | <othermaciej> | presumably ability to cite an authoritative reference is binding, to the extent anything is |
| 09:12 | <tantekc> | see here for current research into conversation/chat examples: http://microformats.org/wiki/chat-examples |
| 09:13 | <tantekc> | feel free to add any more real world evidence you have to that page |
| 09:13 | hsivonen | wonders if othermaciej's statements in an IRC log are authoritative for Wikipedia purposes |
| 09:13 | <erlehmann> | regarding microformats … Hixie, didn't you do a markup study (or planned to do) ? |
| 09:14 | <othermaciej> | hsivonen: I would guess next time HTML5 is published by the HTML WG, you could cite the TR link |
| 09:14 | <othermaciej> | I'm not sure if they would find the Editor's Draft authoritative |
| 09:14 | <tantekc> | erlehmann - the latest broad data on microformats is available via Yahoo Search monkey |
| 09:14 | <Hixie> | erlehmann: we're recruiting participants now |
| 09:14 | <erlehmann> | Hixie, fine to know. :) |
| 09:16 | <annevk2> | tantekc, the logs for this channel often have links at least, in #html-wg pimpbot utters things in colors but the details are not logged, I expect that if Google Wave takes off a little you get more conversations with images, maybe video, etc. |
| 09:16 | <erlehmann> | tantekc, if you couly put a time element / attribute in a dialog, would it help ? |
| 09:16 | <othermaciej> | I wish more albums that I'm actually interested in had iTunes LP available |
| 09:16 | <annevk2> | tantekc, though simple images is already done today in most IM clients |
| 09:17 | <tantekc> | erlehmann - <dialog> is too broken to be worth discussing what might be necessary to fix it. |
| 09:17 | <erlehmann> | tantekc, can you elaborate on that ? |
| 09:17 | <annevk2> | tantekc, I agree we should drop <dialog> |
| 09:18 | <tantekc> | annevk2 - if you can add URLs to illustrative examples of the phenomena you mention to the /wiki/chat-examples page - it would be appreciated. |
| 09:18 | annevk2 | filed a bug to that effect |
| 09:18 | <erlehmann> | hmm, does the xmpp xhtml-im need an overhaul then to keep up with waves and HTML 5 (ha, spaces) ? |
| 09:18 | <tantekc> | annevk2 - feel free to add a +1 to that bug from me. |
| 09:18 | <erlehmann> | annevk2, and how do i markup natural conversations then ? going back to <dl>s ? |
| 09:20 | <tantek> | erlehmann - see the markup here: http://rbach.priv.at/Microformats/IRC/2009-08-29 |
| 09:21 | <tantek> | it makes quite good use of existing HTML 4.01 semantic elements |
| 09:21 | <erlehmann> | tantek, and it uses <cite> in a way it was agreed not to in HTML5 |
| 09:22 | <tantek> | erlehmann - HTML5 is broken in that regard |
| 09:22 | <tantek> | as there is numerous existing uses of <cite> with that semantic |
| 09:22 | <tantek> | thus HTML5 should be amended to match existing semantic use of <cite> |
| 09:22 | <annevk2> | erlehmann, whatever markup suits your usecase best |
| 09:23 | <tantek> | that use of <cite> has been well documented and implemented in markup since at least 2005 |
| 09:23 | <annevk2> | erlehmann, HTML5 does not have dedicated markup for all use cases, see e.g. the chapters on tag clouds and footnotes |
| 09:23 | <takkaria> | how about making i and cite semantically identical? |
| 09:23 | <erlehmann> | annevk2, i use dialog currently, mainly because it is a) simple and b) enables me to state that several people said something simultaneously (am i understanding it right ?) |
| 09:24 | <erlehmann> | takkaria, <i> ? sure you are joking |
| 09:24 | <tantek> | takkaria - it's a smaller change to simply allow <cite> to represent a speaking person, as allowable per HTML 4.01 and in use given the examples. |
| 09:24 | <erlehmann> | tantek, i would prefer it too, but to markup quotes i heard. |
| 09:25 | <tantek> | quotes heard or seen |
| 09:25 | <tantek> | e.g. in an IRC channel |
| 09:25 | <takkaria> | erlehmann: not being entirely serious |
| 09:26 | <annevk2> | erlehmann, I think you might be using it wrong then |
| 09:27 | <annevk2> | erlehmann, yeah, if you use multiple <dt> followed by a single <dd> you're doing it wrong |
| 09:27 | <annevk2> | erlehmann, so maybe it's not simple :) |
| 09:31 | <erlehmann_> | annevk2, its not allowed ? :( |
| 09:34 | <annevk2> | erlehmann_, I'm not sure how much clearer I can make it |
| 09:34 | <annevk2> | erlehmann_, you could read the spec |
| 09:35 | <annevk2> | erlehmann, the validator flags it too... |
| 09:35 | <erlehmann> | :/ |
| 10:04 | <zcorpan> | polyglot++ |
| 10:04 | <annevk2> | seriously? |
| 10:04 | <jgraham> | It's OK, polyglot is idempotent under addition |
| 10:10 | <zcorpan> | http://intertwingly.net/blog/2009/09/08/First-Polyglot-Validator-Check-Deployed#c1252572959 |
| 10:17 | <Philip`> | Hixie: "A news organisation's front page ..." - shouldn't that be 'organization' in en-US? |
| 10:20 | <Lachy> | why is <meta http-equiv="Refresh"> considered conforming? |
| 10:22 | <zcorpan> | Hixie: was the PDF <link> example intended to link to /en/html ? |
| 10:22 | <annevk2> | Lachy, there's no equivalent |
| 10:22 | <zcorpan> | Refresh: ... is equivalent :) |
| 10:23 | <annevk2> | is that implemented? |
| 10:23 | <zcorpan> | iirc yes |
| 10:23 | <Lachy> | annevk2, HTTP 302 for most cases |
| 10:23 | annevk2 | thought only one or two browsers did that |
| 10:23 | <Lachy> | and for the case of a delayed refresh, that should never ever be used |
| 10:23 | <annevk2> | Lachy, disagreed, but I forgot the reason |
| 10:23 | <zcorpan> | Lachy: if it's not conforming, people will use setTimeout and location.href instead |
| 10:24 | <Lachy> | but the most common case is using it as <meta http-equiv="refresh" content="0;http://..."> |
| 10:24 | <zcorpan> | Lachy: the meta is easier to block for users |
| 10:25 | <Lachy> | we should at least warn about it, especially when it's being used for rediection |
| 10:25 | <Lachy> | *redirection |
| 10:27 | <zcorpan> | lol Frontweaver |
| 10:28 | <Philip`> | Lachy: That "most common case" seems to be about half of all Refreshes |
| 10:29 | <Philip`> | and the other half are mostly small numbers like 1 and 5 and 3 and 10 etc |
| 10:29 | <Philip`> | and a handful like 1800 which are presumably for auto-refreshing pages, not for redirects |
| 11:06 | <Hixie> | othermaciej: hah, iTunes 9 just crashed on me in WebCore::HTMLTokenizer::parseTag() :-P |
| 11:06 | <Hixie> | (well, actually in the hashtable code, but frame 4 was the tokeniser) |
| 11:08 | <Hixie> | wow, snow leopard's crash reporting is more advanced than leopard's, even |
| 11:09 | <hsivonen> | Hixie: what does iTunes 9 use WebCore for? |
| 11:10 | <Hixie> | itunes store, it seems |
| 11:10 | <hsivonen> | interesting |
| 11:10 | <hsivonen> | I thought iTunes store was using a product-specific markup language without WebKit |
| 11:10 | <hsivonen> | has it always been like that or does 9 change things substantially? |
| 11:11 | <annevk2> | per gruber it's 9 |
| 11:11 | <hsivonen> | ah |
| 11:12 | jgraham | assumes that Apple will nevertheless prevent people accessing iTunes Web Store using an ordinary web browser |
| 11:12 | <hsivonen> | it's quite sad, really. |
| 11:12 | <hsivonen> | I would have made purchases in the iTunes store using Ubuntu if they had a pure Web UI |
| 11:14 | <zcorpan> | hsivonen: but since you want iTunes store so badly, you don't use Ubuntu, right? |
| 11:14 | <beowulf> | the css for the itunes store is at http://ax.itunes.apple.com/htmlResources/CFA7/dt-storefront.cssz |
| 11:15 | <beowulf> | so the internet says |
| 11:15 | <hsivonen> | zcorpan: well, now we both an Ubuntu box and a Mac at home but the Mac came for a different reason |
| 11:15 | <Hixie> | HOLY SWEET KITTENS |
| 11:16 | <Hixie> | itunes9 has boolean search queries in the smart playlists finally |
| 11:16 | <jgraham> | Ah, I thought you had just discovered a tasty snack |
| 11:17 | <Hixie> | i can reconstruct my playlists so that i just have one "Music Mix" playlist instead of having to construct it using about 9 playlists that are slowly mixed together to do all the "and"ing and "or"ing that i wanted |
| 11:19 | jgraham | wonders if he is abnormal because he hardly ever uses playlist features |
| 11:19 | <hsivonen> | jgraham: you aren't |
| 11:20 | <Hixie> | i need playlists so that i can get itunes to sync the right podcasts to my ipod |
| 11:21 | Philip` | just runs Audacious with a flat list of 247 hours of audio files, and has never had any desire for more organisation |
| 11:21 | <hsivonen> | I wish people sarted doing QA with 1.0 UI cale factor |
| 11:22 | <hsivonen> | *started |
| 11:23 | <Philip`> | *scale? |
| 11:23 | <jgraham> | Needing to have iTunes is one of the reasons I don't have an iPod. |
| 11:23 | <Philip`> | Isn't 1.0 the default? |
| 11:23 | <hsivonen> | Philip`: yes |
| 11:23 | <hsivonen> | Philip`: non-1.0 |
| 11:23 | <hsivonen> | at least 3 bugs in one sentence so far |
| 11:23 | <Philip`> | Ah |
| 11:23 | <hsivonen> | I lose |
| 11:28 | <hsivonen> | Does Mac OS X, in principle, support differnt scale factors for different screens? |
| 11:29 | <Hixie> | Philip`: i'm up to 427 hours of just (legally-acquired) music, and that's not counting tv shows, movies, podcasts, poetry tracks, audiobooks, etc |
| 11:29 | <Hixie> | Philip`: so i need to narrow it down to sync to my music player |
| 11:31 | <Philip`> | Hixie: You just need a higher-capacity music player |
| 11:31 | <Hixie> | i was going to get the new ipod touch, but it's not really compelling enough to justify $400 or whatever it was |
| 11:35 | <Lachy> | Hixie, in webdatabase, when a database is first opened with "" as the version using: openDatabase("name", "", "Display Name", "1024"); and then next it's opened with "foo" as the version, should it throw an exception? |
| 11:35 | <Hixie> | off the top of my head, yes |
| 11:35 | <Lachy> | Our developer thinks the spec is ambiguous about this because the empty string is used to mean any version is fine |
| 11:35 | <Hixie> | i thought i made this very clear recently |
| 11:35 | <Hixie> | i changed it to have an explicit algorithm |
| 11:35 | <Lachy> | that's what I thought too |
| 11:45 | <Lachy> | Hixie, this requirement makes no sense "If a callback was passed to the method, then let the database's version be the empty string. Otherwise, let its version be the given database version" |
| 11:46 | <Hixie> | it certainly seems odd |
| 11:46 | <Lachy> | I assume it should be saying if the version passed was not the empty string or something |
| 11:46 | <Hixie> | send mail? |
| 11:46 | <Lachy> | ok, will do when I sort out what the spec is supposed to be saying for steps 3 and 4 |
| 11:46 | Hixie | looks |
| 11:48 | <Hixie> | actually it means exactly what it says |
| 11:48 | <Lachy> | Hixie, it should say something like this: |
| 11:48 | <Lachy> | * Database exists? |
| 11:48 | <Lachy> | - Version passed is empty, or version passed matches existing version? Open database. |
| 11:48 | <Lachy> | - Version passed is not empty and does not match, throw INVALID_STATE_ERR |
| 11:48 | <Lachy> | * Otherwise, database does not exists: |
| 11:48 | <Lachy> | - Create and open the database, let "created" be true. Set the version to the version passed. |
| 11:49 | <Hixie> | (except for missing the "." at the end of the sentence) |
| 11:50 | <Lachy> | but it doesn't make any sense for the value of the version to be set based on whether or not a callback was passed. |
| 11:50 | <Hixie> | yeah, it does |
| 11:50 | <Lachy> | how does it? |
| 11:50 | <Hixie> | because the callback is meant to call changeVersion() |
| 11:50 | <Lachy> | what? |
| 11:51 | <Hixie> | otherwise, two calls to the method at the same time would both call the callback |
| 11:51 | <Hixie> | or, one would call the callback and the other would succeed, but before the callback in the other thread had finished |
| 11:53 | <Lachy> | oh. That just makes things confusing. So if I call openDatabase("db", "foo", "Whatever", 1024, callback); then the version is meant to be set to "" regardless? |
| 11:53 | <Hixie> | only if the database doesn't exist |
| 11:55 | <Lachy> | ok, well we don't support the creation callback yet anyway, so we can ignore that statement for now |
| 11:55 | <Lachy> | so just considering the cases where there is no callback |
| 11:56 | <Lachy> | the steps as written are still not particularly clear |
| 11:56 | <Hixie> | yeah |
| 11:56 | <Hixie> | remind me in 15 minutes |
| 11:56 | <Lachy> | ok |
| 12:02 | <karlcow> | [03:06] <Dashiva> karl does, but he might not be around |
| 12:02 | <karlcow> | even a better person to ask is people from CreativeCommons who are part of the htmlwg. |
| 12:04 | <karlcow> | indeed CreativeCommons is a way to fill the space in between two extremes: Copyrights to Public Domain, with a variety of options. |
| 12:11 | <Hixie> | man, even since i put my cat's e-mail address in the spec, he gets more spam than i do |
| 12:12 | <jgraham> | I guess if you didn't spy on your cat's inbox it might not bother you so much |
| 12:12 | <beowulf> | if you can train him to filter that spam you could make money from it i'm sure |
| 12:12 | <Hixie> | umm... would you believe "he asked me to redirect his mail to my inbox"? |
| 12:12 | <Hixie> | so that i can read his mail for him, since, you know, he's a cat |
| 12:13 | <annevk2> | you know, most cats don't have email addresses |
| 12:13 | <jgraham> | It sounds like you're jsut checking that he doesn't get email from other would-be owners |
| 12:14 | <Hixie> | he needed an e-mail address to get his livejournal account... |
| 12:15 | <Hixie> | Lachy: ok, updated the steps a bit. |
| 12:15 | <Hixie> | give it a few secs to upload |
| 12:15 | <jgraham> | (interesting aside: my landlady did actually steal someone else's cat by placing a continual supply of food outside the door so that it preferred living in our building to his actual home. Then, when he needed to go to the vet she went and talked to the real owners and somehow convinced them that she should take it to the vet and keep it thereafter) |
| 12:16 | <hsivonen> | https://twitter.com/ppk/status/3867647272 |
| 12:16 | <Hixie> | hsivonen: so sad |
| 12:16 | <Hixie> | (the "normal" part) |
| 12:20 | <Hixie> | ok i'm starting to notice a bad trend |
| 12:20 | <Hixie> | when i work on e-mails, bug count goes up |
| 12:20 | <Hixie> | when i work on bugs, e-mail count goes up |
| 12:20 | <Hixie> | let's see what happens when i sleep |
| 12:20 | <Hixie> | nn |
| 12:25 | <Lachy> | Hixie, the spec is clearer now. Thanks |
| 12:27 | <othermaciej> | Hixie: we've had other reports of iTunes crashes while tokenizing - oddly it doesn't seem to happen in Safari |
| 12:28 | <annevk2> | doesn't your parser have some non-Web flags? |
| 12:29 | <hsivonen> | looks like iTunes broke with HIG again and has an almost-but-not-quite unified window look |
| 12:36 | <hsivonen> | was the it ever considered if scripts should have an attribute that does defer them (as opposed to defer and async) but promises that the scripts won't do document.write? |
| 12:36 | <hsivonen> | with some obvious penalty for doing document.write anyway |
| 12:36 | hsivonen | is tempted to prototype and benchmark such a feature in the future |
| 12:47 | <zcorpan> | "limit when <!-- takes escaping effect so that it only takes escaping effect if there has been either nothing or only whitespace on the same line before it" -- http://wiki.whatwg.org/wiki/CDATA_Escapes |
| 12:48 | <zcorpan> | any research in this area? |
| 12:48 | <hsivonen> | zcorpan: not yet |
| 12:49 | <hsivonen> | zcorpan: it seems Hixie isn't going to do it and I don't have enough time in Q3 |
| 12:49 | <zcorpan> | i don't think it'll help much, as from what i remember when studying this before, pages have <!-- in the beginning of the script but no matching --> in the end |
| 12:51 | <zcorpan> | i don't see why we'd remove it from title, textarea and xmp |
| 12:52 | <zcorpan> | although i think it's only required for script |
| 12:52 | <zcorpan> | i.e. not style |
| 12:52 | <zcorpan> | style might even benefit from not having it, considering pages that have <style><!-- blah </style> |
| 12:54 | <annevk3> | hsivonen, in your question it might help if you tell what it is for |
| 12:54 | <annevk3> | hsivonen, I guess it is the speculative parsing stuff |
| 12:56 | <zcorpan> | hmm |
| 12:57 | <zcorpan> | maybe we should require to have seen "<script" also |
| 12:58 | <zcorpan> | i.e. <script><!--</script> is closed but <script><!--<script></script></script> closes on the second end tag |
| 13:00 | <hsivonen> | annevk3: it's for avoiding a speculative future |
| 13:00 | <hsivonen> | annevk3: i.e. parsing past </script> ahead of time but committing to that future |
| 13:01 | <annevk3> | is that really a use case? |
| 13:01 | <annevk3> | if authors become aware of these issues they'll prolly just use <script defer> and such |
| 13:02 | <hsivonen> | annevk3: it's as much a use case as async has a use case if you formulate it from the author-perceived performace POV instead |
| 13:02 | <hsivonen> | annevk3: defer and async make the script execute later in the parse |
| 13:02 | <annevk3> | i thought async was for things that analyze the page and visitors and such |
| 13:03 | <hsivonen> | annevk3: I mean a promise that makes things faster without changing the detectable order of parsing and script execution |
| 13:03 | <annevk3> | i.e. execute asap but do not halt parsing |
| 13:03 | <hsivonen> | annevk3: it is |
| 13:03 | <annevk3> | i understand what you're proposing, i just don't see the use case |
| 13:04 | <jgraham> | I thought the use case was "be faster" |
| 13:04 | <jgraham> | Although I don't know if it is fater enough to make people actually care |
| 13:04 | <othermaciej> | it could be a minor performance benefit but I'd be surprised if it made much difference |
| 13:04 | <hsivonen> | annevk3: the use case is having a script set up event handler-triggered code (that the user can trigger during page load) near the top of the page without taking the perf hit of a speculative future |
| 13:04 | <hsivonen> | othermaciej: that may be |
| 13:04 | zcorpan | adds a proposal to cdata escapes |
| 13:05 | <hsivonen> | othermaciej: which is why I'm asking if it was considered and I'm not proposing anything at this point |
| 13:05 | <othermaciej> | the main useful aspect of async/defer is not having to block parsing on a network load |
| 13:05 | <mpilgrim> | roc: i'm not planning on mentioning any 3d canvas proposals in this book |
| 13:05 | <mpilgrim> | you'll have to wait for "dive into html6" |
| 13:06 | <othermaciej> | with this proposed feature you have to block at least some aspects of parsing on the network load if the script is external, since the script can use DOM APIs to observe the DOM |
| 13:06 | <zcorpan> | my proposal would work for <script><!-- d.w('<script></script>'); d.w('<script></script>'); </script> |
| 13:07 | <zcorpan> | it would break for <script><!-- d.w('<scr'+'ipt></script>'); </script> |
| 13:07 | <zcorpan> | even <script><!-- d.w('<scr'+'ipt></script>'); --></script> |
| 13:10 | <hsivonen> | othermaciej: you'd block the actuation of the DOM ops on the main thread, but the parser thread could continue to produce a non-speculative stream of tree ops |
| 13:11 | <annevk3> | hsivonen, what would you do different if <script> didn't include the hint? |
| 13:11 | <hsivonen> | annevk3: create the bookkeeping for speculation |
| 13:12 | <hsivonen> | annevk3: hopefully not a big deal |
| 13:12 | <hsivonen> | like I said, I'm not proposing anything without data |
| 13:12 | <hsivonen> | just asking if Hixie has already considered this |
| 13:13 | <annevk3> | fair enough |
| 13:13 | <zcorpan> | hsivonen: what do you think about proposal #3? |
| 13:13 | <othermaciej> | hsivonen: our tokenizer is so cheap (relative to other things that happen during page loading) that I can't imagine even expending the effort to put it on a separate thread |
| 13:13 | <othermaciej> | hsivonen: maybe Gecko has wildly different performance characteristics |
| 13:14 | <hsivonen> | zcorpan: what proposal are you referring to? |
| 13:14 | <zcorpan> | in http://wiki.whatwg.org/wiki/CDATA_Escapes |
| 13:14 | <annevk3> | hmm, the tokenizing ahead will also fail if you do <select> <link> </select> or some such |
| 13:15 | <annevk3> | zcorpan, no chance you can work on Web DOM Core? |
| 13:16 | <zcorpan> | annevk3: not now anyway |
| 13:16 | <hsivonen> | annevk3: what do you mean tokenizing ahead will fail? |
| 13:17 | <hsivonen> | zcorpan: proposal #3 seems more magic than proposal #2. I am unable to tell without empirical data which one is better. |
| 13:18 | <hsivonen> | zcorpan: more magic scares me on the knee-jerk level |
| 13:18 | <annevk3> | hsivonen, you'd find links that get dropped |
| 13:18 | <zcorpan> | Philip`: which is better? :) |
| 13:19 | <hsivonen> | annevk3: starting speculative GETs that don't get used is fine |
| 13:19 | <hsivonen> | GET idempotence FTW |
| 13:19 | <zcorpan> | hsivonen: my knee-jerk reaction to #2 is that it wouldn't do much of a difference for compat |
| 13:19 | <hsivonen> | zcorpan: we need data |
| 13:19 | <zcorpan> | indeed |
| 13:53 | <karlushi> | pigeons are fast http://www.engadget.com/2009/09/10/south-african-pigeon-transmits-data-faster-than-local-dsl/ |
| 13:58 | <hsivonen> | where's the latest real-world URL spec that wasn't damaged to uselessness? |
| 14:00 | <jcranmer> | RFC 11-something? |
| 14:00 | <hsivonen> | I mean the one that was Web Addresses before getting edited too much |
| 14:00 | <hsivonen> | and that was the HTML5 URL section before that |
| 14:00 | <jcranmer> | ah, 1738 |
| 14:01 | <Philip`> | zcorpan: I'm afraid that question requires too much thinking and interpretation from me, and I'm lazy :-p |
| 14:02 | <hsivonen> | jcranmer: I mean an non-fiction version--not the RFCs |
| 14:03 | <hsivonen> | is this the version that hasn't been edited too much? http://www.w3.org/html/wg/href/draft.html |
| 14:09 | annevk3 | used http://www.w3.org/TR/2009/WD-html5-20090423/infrastructure.html#urls |
| 14:10 | <Lachy> | karlushi, I've seen similar comparisons made in the past, but find those tests to be fairly meaningless. It will always be possible to send a physical disk with more data than can be transmitted over a network in the same time period, simply by increasing the required amount of data |
| 14:11 | <annevk3> | karlushi, hehe |
| 14:12 | <jgraham> | (or decreasing the distance) |
| 14:12 | <jgraham> | But it seems like their network speed was 11 kB/s which is pretty poor |
| 14:12 | <Lachy> | unless the test has been calibrated by calculating how much data would be expected to be transmitted in the time |
| 14:13 | <annevk3> | hsivonen, that version works too though it misses all the xrefs |
| 14:17 | <Lachy> | jgraham, 23kBps, http://www.google.com/search?q=4%25+of+4GB+%2F+2+hours although the article isn't clear about exactly how much data it is. It only says it was a 4GB memory card. |
| 14:18 | <jgraham> | Oh I remembered 2% somehow |
| 14:18 | <hsivonen> | annevk3: thanks |
| 15:16 | hsivonen | minted an authoritative source for Wikipedia to cite: http://blog.whatwg.org/spelling-html5 |
| 15:19 | <svl> | hsivonen: now for someone to update http://www.w3.org/TR/html5-diff/ http://dev.w3.org/html5/html-author/ etc |
| 15:20 | <hsivonen> | http://html5.org/ too |
| 15:21 | Mrmil | likes the new whatwg blog design, much better than the default :) |
| 15:24 | <jgraham> | The green bar still clashes with the other greens |
| 15:25 | <hsivonen> | the border around the whatwg logo sucks in Firefox |
| 15:26 | <annevk2> | html5.org fixed |
| 15:26 | <hsivonen> | annevk3: thanks |
| 15:26 | <annevk2> | svl, html5-diff is already fixed, see editor's draft |
| 15:27 | <svl> | annevk2: ah, indeed. |
| 16:05 | <annevk2> | and, if you're around: http://wiki.whatwg.org/wiki/Web_Encodings#Encodings_3 |
| 16:15 | <Lachy> | svl, when I resume editing the HTML5 Reference soon, I'll convert everything from "HTML 5" to "HTML5" |
| 16:16 | <Lachy> | when did Hixie remove the distinction between "HTML 5" and "HTML5" that was once in the spec? |
| 16:16 | <Philip`> | A small number of days ago |
| 16:17 | <Philip`> | http://html5.org/tools/web-apps-tracker?from=3702&to=3703 |
| 16:21 | <Lachy> | ok, so we no longer have a way to refer to HTML 5, the vocabulary, and instead only the ability to refer to the serialisations. |
| 16:22 | <Philip`> | You can refer to it as "HTML5", and if that's ambiguous use "the HTML5 vocabulary" |
| 16:22 | <Lachy> | yeah, I guess that's a lot less confusing than making it dependent upon a space |
| 16:22 | <Lachy> | which no-one used correctly anyway |
| 17:04 | <mpilgrim> | <annevk2> you know, most cats don't have email addresses |
| 17:04 | <mpilgrim> | my dog had an email address in 1995 |
| 17:05 | <takkaria> | those were the days |
| 17:05 | <mpilgrim> | i had to fight for it with the ISP |
| 17:05 | <takkaria> | do you also have a fish licence? |
| 17:05 | <mpilgrim> | then the ISP died |
| 17:05 | <mpilgrim> | then the dog died |
| 17:05 | <mpilgrim> | that story was more depressing than i intended |
| 17:27 | <Lachy> | mpilgrim, why did your dog have an email address? |
| 17:29 | <Philip`> | How else would anyone send email to it? |
| 17:30 | <Lachy> | Philip`, I'm curious who would be sending e-mail to a dog? |
| 17:31 | <beowulf> | there was a dog could use an atm |
| 17:31 | <Philip`> | Maybe one of the dog's friends |
| 17:31 | <beowulf> | http://en.wikipedia.org/wiki/Endal |
| 17:31 | <Lachy> | most other dogs can't type, and I'm sure all of his friends lived in the same neighbourhood and could communicate easily by barking. |
| 17:36 | <Philip`> | Maybe they are cyborg rat-things |
| 21:39 | <and> | annevk2: Thanks. |
| 21:40 | <and> | annevk2: Does (???) mean that the encoding is not supported (any more)? |
| 21:41 | <and> | annevk2: Are you at all interested in EBCDIC-based and old European 7-bit encodings? |
| 21:44 | <Hixie> | hsivonen: async="" penalises document.write(), no? |
| 21:45 | <and> | annevk2: (???) seems to correspond to long names in the text file, actually. I shall update the wiki. |
| 21:54 | <and> | annevk2: No, that does not seem plausible after all. I reiterate the question. |
| 22:15 | <and> | annevk2: gsnedders: Firefox seems to understand the HTTP header 'Content-Type: text/plain; charset= "iso-8859-16" ', but does not accept whitespace before the equals sign or inside the (double) quotes. The spec annevk2 mentioned yesterday allows whitespace before the equals sign, if I remember correctly. |
| 22:17 | <and> | `svn co http://svn.versiondude.net/whattf/build/trunk/ build` does not seem to work. Should it, or is there a better way to get hsivonen's parser? |
| 22:23 | <gsnedders> | and: Whitespace inside double quotes should matter, so it is right. |
| 22:24 | <gsnedders> | and: I still maintain my POV that the HTTP spec is what happens for HTTP headers, and the spec annevk2 linked to only deals with meta[@http-equiv="content-type"] |
| 22:26 | <and> | gsnedders: The spec annevk2 seems to be more ambitious. |
| 22:26 | <and> | gsnedders: Does the HTTP spec allow whitespace between charset and =? |
| 22:26 | <gsnedders> | and: I think so. |
| 22:27 | <gsnedders> | It's all a bit horrible with implied LWS (which isn't that uncommon in older specs) |
| 22:28 | <and> | gsnedders: Yes, I guess it means that whitespace is allowed if charset and = are to be regarded as two separate tokens. |
| 22:29 | <gsnedders> | Oh., no, you can't have it between the = |
| 22:29 | <gsnedders> | On either side. |
| 22:29 | <and> | s/annevk2/& linked to/ |
| 22:29 | <gsnedders> | It's explicitly prohibited |
| 22:30 | <gsnedders> | LWS is allowed either side of the ; though |
| 22:32 | <and> | Two specifications and one implementation; three interpretations. |
| 22:35 | <gsnedders> | Well, the HTTP spec doesn't define what to do for non-conforming messages |
| 22:41 | <Lachy> | Hixie, this section seems rather difficult to comprhend. http://www.whatwg.org/specs/web-apps/current-work/#distinguishing-site-wide-headings-from-page-headings |
| 22:41 | <Lachy> | But I get the impression that it's trying to tell authors to do this: |
| 22:42 | <Lachy> | <body><header><h1>Site Wide Heading></h1></header><article><h1>Page Heading</h1> ... </article><footer>...</footer></body> |
| 22:43 | <Lachy> | and if that's right, then it does seem to be encouraging authors to use <article> as a content wrapper. |
| 22:45 | <Lachy> | in the last paragraph of that section, what does it mean to have "... more than one section at the root of the outline."? |
| 22:46 | <Lachy> | is that like <body><section><h1>Heading Level 1</h1></section><section><h1>Another Level 1 Heading</h1></section></body> |
| 22:50 | <Lachy> | gsnedders, in your outliner, can you make it so that it gives some kind of indication about which element created the section? Especially in the case of "Untitled Section" sections. |
| 22:50 | <Lachy> | e.g. http://gsnedders.html5.org/outliner/process.py?url=http://html5doctor.com/about/ |
| 22:51 | <gsnedders> | Lachy: No, because I keep saying I'll do things and never have the time to do them. |
| 22:52 | <Lachy> | ok |
| 22:52 | <Lachy> | let me know when you'll do it :-) |
| 22:52 | <gsnedders> | Once my to-do is a bit emptier, which won't be for a while |
| 22:53 | gsnedders | writes a to-do list purely of things that _must_ be done tomorrow. |
| 22:53 | <gsnedders> | (Honestly, it only involves going to another city for a bit…) |
| 22:53 | <Lachy> | I suggest you cross out items 2, 4, 5 and 7 from your todo list. That will shorten it a bit. |
| 22:53 | <gsnedders> | (and I think everything on it has to be done from 9–5) |
| 22:54 | <gsnedders> | Lachy: But #2 is a dependancy on #1, #4 I need for Saturday, #5 I guess I could (but once I'm in the other city for #4 it's not that much), and #7 throws an IndexError |
| 22:54 | <Lachy> | right, so that's one off your list already! |
| 23:24 | <jamesr> | michaeln: the "they wrote a bug and should fix it if they wanted mutex semantics." part is what i think folks are objecting to |
| 23:24 | <jamesr> | because that's been tried many times and has never really worked well |
| 23:29 | <jamesr> | given that there's very little shared state in html5 currently it seems better to try to avoid the mess altogether |
| 23:31 | <jamesr> | what about restricting access to to shared resources (localStorage, databases) to a single event loop the same way DOM access is? |
| 23:39 | <annevk2> | and, sorry, ??? means that some other table I found that mapped the code pages to a definitive name misses that code page |
| 23:41 | <annevk2> | and, I sort of finished that table when I had to run so I couldn't add explanations |
| 23:42 | <and> | annevk2: No problem. |
| 23:42 | <and> | annevk2: The cp* names should probably be taken as canonical. Not sure about the remaining three. |
| 23:43 | <annevk2> | I hope we can make supporting UTF-7, UTF-32 and EBCDIC to be non-conforming |
| 23:43 | <annevk2> | especially for EBCDIC that should get rid of a gazillion encodings |
| 23:44 | <and> | Yes, I'll ignore those for the moment (and add a comment to the table). |
| 23:44 | <and> | Old 7-bit European encodings as well? |
| 23:45 | <annevk2> | are those 7-bit turned into 8-bit by Windows? |
| 23:45 | <annevk2> | for all of them it kind of depends on the amount of implementations |
| 23:46 | <and> | I am not sure. IE used to ignore the high bit for ASCII, but not in IE8. |
| 23:46 | <annevk2> | it seems implementations can get away with not supporting UTF-7 and UTF-32 and EBCDIC |
| 23:46 | <annevk2> | and, Sylvain (IE) will be looking into IE-specific encoding hacks as well |
| 23:47 | <and> | 7-bit European encodings were effectively legacy before Internet (though they remained for a long time in certain system, including Norwegian library catalogues). |
| 23:47 | <annevk2> | and, he has not done that yet |
| 23:47 | <annevk2> | and, this the backend for IE/Windows (apparently a little bit different from .NET/Windows |
| 23:47 | <annevk2> | ) |
| 23:47 | <and> | s/system/&s/ |
| 23:48 | <annevk2> | is windows-sami-2 one of those? |
| 23:48 | <and> | Do you mean that the table is based on data for IE, and that .NET differs? |
| 23:48 | <annevk2> | yes |
| 23:48 | <annevk2> | from what I heard anyway |
| 23:48 | <and> | No, that one is 8-bit and a fairly recent innovation. |
| 23:49 | <and> | Neither Opera nor Firefox supports any European 7-bit encoding. |
| 23:49 | <annevk2> | to be sure you can compare my (or sylvain's csv file) with http://blogs.msdn.com/shawnste/archive/2009/08/18/alternate-encoding-names-recognized-by-net-ie.aspx |
| 23:49 | <annevk2> | and, then they're obsolete imo |
| 23:51 | <annevk2> | maybe we should put down some principles somewhere to determine whether an encoding is in or out so people can backtrack the reasoning |
| 23:55 | <and> | I'll go through the list and add a note for EBCDIC-based and legacy 7-bit encodings to start with. |
| 23:56 | <and> | I was a bit surprised to see support for all the Indian ISCII encodings as well. |
| 23:58 | <annevk2> | i was surprised the number of code pages surpassed 200 |
| 23:59 | <annevk2> | or maybe shocked, dunno :) |