| 00:01 | <Hixie> | can anyone recall where we dicussed the idea of making Window have a prototype chain on which we put the named items? |
| 00:01 | <Hixie> | i can't find either bug or mail about it |
| 00:05 | <zewt> | gmail works well for most things, but yeah, random little things missing (which aren't so little when you need them) |
| 00:05 | <Hixie> | hober: i assume the copyright on http://www.w3.org/Bugs/Public/show_bug.cgi?id=12409 is a mistake? :-) |
| 00:05 | <zewt> | complete lack of sorting is pretty bad |
| 00:06 | <zewt> | actually the biggest is the inability to view non-list mails, eg. things with no labels; they just fall into the abyss of "all mail" |
| 00:07 | <hober> | Hixie: yes, indeed |
| 00:07 | <hober> | Hixie: forgot to strip that out before copying and pasting the proposal into the textarea |
| 00:07 | <hober> | Hixie: I can comment there to that effect if you'd like |
| 00:08 | <Hixie> | nah, no worries |
| 00:08 | <Hixie> | i was just amused :-) |
| 00:08 | <hober> | yeah, i noticed that like 2 minutes after hitting submit |
| 00:08 | <hober> | and was sadface |
| 00:32 | <TabAtkins> | Hixie: I think that's on our list of "possible Selectors 4" things. |
| 00:32 | <Hixie> | they've been on the list of possible things since 2002 or before |
| 00:32 | <Hixie> | but the interesting new thing is the idl annotations |
| 00:33 | <Hixie> | btw i've started http://wiki.whatwg.org/wiki/Dialogs to collect research on pages that use dialog boxes |
| 00:33 | <TabAtkins> | Ooh, kk. |
| 00:33 | <TabAtkins> | There's unescaped html on that page, btw. |
| 00:34 | <Hixie> | i hate wiki syntax |
| 00:34 | <Hixie> | it can't make its mind up about whether it's HTML or not |
| 00:34 | <TabAtkins> | Indeed. |
| 00:34 | <Hixie> | (thanks, fixed) |
| 00:35 | <hober> | Hixie: ahh, cool. I've been looking into dialogs lately, to determine what sort of CP (if any) to write for ISSUE-133 |
| 00:35 | <Hixie> | that's all we need, language design by issue process :-) |
| 00:35 | <hober> | heh |
| 00:36 | <Hixie> | i plan to be doing something along those lines in the coming weeks |
| 00:36 | <Hixie> | (though certainly not by the deadline for that issue) |
| 00:36 | <hober> | I'm leaning toward writing a zero-edit CP that says basically "there are valid use cases, but the other CP's proposal is woefully underspecified and broken, and to do it right would require a lot of work. let's punt until html.next" |
| 00:37 | <Hixie> | haha |
| 00:37 | Hixie | reads said CP |
| 00:37 | <Hixie> | you should totally just not post a CCP and let the chairs say this one is gold |
| 00:37 | <Hixie> | that would just be funny |
| 00:37 | <boogyman> | are you talking to yourself Ian? |
| 00:38 | <Hixie> | hober != Hixie |
| 00:38 | <hober> | If I don't do said CP, I'll do a proper one with a decent design, which would require a time extension |
| 00:39 | <Hixie> | hober: i expect we can probably do something pretty simple that hits the majority of use cases |
| 00:39 | <Hixie> | hober: but first i need to figure out what the use cases are :-) |
| 00:39 | <Hixie> | (hence the wiki page) |
| 00:40 | <hober> | Hixie: yeah, I'll be contributing to the wiki page |
| 00:40 | <Hixie> | basically whenever i come across a dialog i'll take a screenshot and paste the url or markup used to the page |
| 00:40 | <hober> | Hixie: I think stev'e content attribute is totally broken |
| 00:40 | <Hixie> | i kinda just want to see what the chairs would do if something that broken were to be the only proposal |
| 00:41 | <hober> | Hixie: there would be a call for consensus on the existing cp, and all the usual suspects would object in the poll because the cp is so bad |
| 00:41 | <Hixie> | they don't do polls if there's just one cp |
| 00:42 | <Hixie> | and honestly i'm not sure how many of the usual suspects would notice the call |
| 00:42 | <hober> | ok, so s/in the poll/on public-html/ |
| 00:42 | <hober> | fair enough |
| 04:04 | <Hixie> | AryehGregor: dude how do you link to a category page in a wikimedia page? |
| 04:05 | <Hixie> | just a regular link, not the thing that makes the category appear at the bottom |
| 05:01 | <Hixie> | wtf |
| 05:01 | <Hixie> | how did that spam get through |
| 05:03 | <MikeSmith> | Hixie: hey, you pinged me about something last week. I'm guessing you probably don't remember now why |
| 05:03 | <Hixie> | i would have sent you mail or cc'ed you on the relevant bug or whatnot |
| 05:03 | <MikeSmith> | ok |
| 05:03 | <Hixie> | when i ping people these days it's just to see if i can get some quick reaction rather than letting it sit another 3 months for the next time i see the reply :-) |
| 05:04 | <MikeSmith> | hai |
| 05:07 | <Hixie> | well it looks like this spam lost its payload, so i'm just going to ignore it. i still don't understand how it got through the list's filters though. |
| 05:18 | <MikeSmith> | Hixie: does it have multiple From addresses? |
| 05:19 | <Hixie> | dunno, i deleted it |
| 05:19 | <MikeSmith> | OK |
| 05:19 | <Hixie> | didn't see multiple From addresses, but it could have |
| 05:20 | MikeSmith | checks and sees that it doesn't |
| 07:22 | <mhausenblas> | foolip around? |
| 08:52 | <MikeSmith> | «In validation proper, or in "DTD based validation", as opposite to the mixed-technology heuristic checker called HTML5 mode of the validator» |
| 08:52 | <MikeSmith> | http://lists.w3.org/Archives/Public/www-validator/2011Jun/0047.html |
| 08:54 | <MikeSmith> | hsivonen: btw, I updated the W3C validator backends today with latest from upstream |
| 08:54 | <mhausenblas> | ping foolip |
| 08:55 | <foolip> | pong mhausenblas |
| 08:55 | <mhausenblas> | heya |
| 08:55 | <foolip> | what's up? |
| 08:55 | mhausenblas | understands that you're quite active in the Schema.org discussion group |
| 08:55 | <hsivonen> | MikeSmith: cool. thanks. Let's see how that affect meta name and link rel bug reports |
| 08:55 | <hsivonen> | MikeSmith: let me guess. the email above is from Jukka. |
| 08:55 | <mhausenblas> | now, I was wondering if you'd like to do some joint work in Schema.org evangelism, foolip? |
| 08:55 | hsivonen | opens the link |
| 08:55 | <MikeSmith> | heh |
| 08:56 | <MikeSmith> | hsivonen: bingo |
| 08:56 | <foolip> | mhausenblas, well, I spam them with bug I find a lot :) |
| 08:56 | <mhausenblas> | that is a good start, foolip |
| 08:56 | <hsivonen> | MikeSmith: my heuristics are working! |
| 08:56 | <mhausenblas> | but I was thinking rather long-termish |
| 08:56 | <MikeSmith> | :) |
| 08:56 | <foolip> | mhausenblas, do you represent schema.org in some way, or just a fanboy/flamerboy like me? |
| 08:56 | <mhausenblas> | the latter - and behind Schema.RDFS.org :P |
| 08:57 | <mhausenblas> | well fanboy rather than flamerboy |
| 08:57 | mhausenblas | has made his position clear in http://webofdata.wordpress.com/2011/06/08/towards-networked-data/ ... methinks ;) |
| 08:58 | <mhausenblas> | so, anyways, I know your good work from the MF WG were we, well at least partially happened both to work, foolip and hence I was wondering ... |
| 08:58 | <foolip> | now why haven't I seen that post before? perhaps it's not been mentioned on twitter together with "schema.org" :) |
| 08:58 | <mhausenblas> | :P |
| 08:58 | <foolip> | what kind of thing did you have in mind? |
| 08:58 | <mhausenblas> | well, we're using https://github.com/mhausenblas/schema-org-rdf |
| 08:58 | <foolip> | I just toy around with this stuff in my spare time |
| 08:59 | <mhausenblas> | and me currently developing http://omnidator.appspot.com/ |
| 08:59 | <mhausenblas> | which happens to exploit Schema.org terms for anything2anything |
| 08:59 | <foolip> | ok |
| 08:59 | <mhausenblas> | anyways, I was wondering if you'd like to chime in on the above github repo |
| 09:00 | <mhausenblas> | there are two new sections (directories) called examples and mappings |
| 09:01 | <mhausenblas> | the former is for collecting examples of Schema.org in various formats the latter for collecting vocab mappings (DBpedia, FOAF, etc - Schema.org terms) |
| 09:01 | <mhausenblas> | but what I find most thrilling is the challenges and opportunities re extending Schema.org terms |
| 09:01 | <mhausenblas> | and I think there we can assist the sponsors |
| 09:01 | <mhausenblas> | what do ya think foolip? |
| 09:02 | <foolip> | It's not entirely clear to me what the goal of the project is, except having a bit of fun or course |
| 09:02 | <hsivonen> | how can it be so hard to make a proper email client |
| 09:02 | <hsivonen> | webmail or native |
| 09:02 | <hsivonen> | maybe I should just give up and make all my data are belong to google |
| 09:02 | <mhausenblas> | foolip of which project? schema.org or schema.rdfs.org? :D |
| 09:02 | <foolip> | I've never seriously toyed with RDF, so looking at mappings isn't something I'd really be helpful in I think |
| 09:03 | <mhausenblas> | ok, fair enough |
| 09:03 | <foolip> | mhausenblas, whichever you're asking my opinion on :) |
| 09:03 | <mhausenblas> | he he |
| 09:03 | <foolip> | as for examples, are you not just copying them from schema.org? |
| 09:03 | <mhausenblas> | I rather thought of helping people with the extensions part |
| 09:03 | <mhausenblas> | re examples, yeah, sorta reformulating them in CSV, JSON, RDFa, etc. |
| 09:03 | <foolip> | I haven't quite understood how that's supposed to work, is it documented on schema.org? |
| 09:03 | <mhausenblas> | what? |
| 09:03 | <mhausenblas> | the examples? |
| 09:04 | <foolip> | extensions |
| 09:04 | <mhausenblas> | oh |
| 09:04 | <mhausenblas> | yes |
| 09:04 | <mhausenblas> | lemme check |
| 09:04 | <mhausenblas> | http://schema.org/docs/extension.html |
| 09:04 | <foolip> | that looks fairly icky :) |
| 09:05 | <mhausenblas> | the problem AFAICT (looking at the Schema.org Google group) is that most people that wanna extend are not exactly experts in vocab design |
| 09:05 | <foolip> | how could you ever transition an extension into official-dom? |
| 09:05 | <mhausenblas> | dunno. ask the sponsors :P |
| 09:06 | <foolip> | For example, http://schema.org/Person/Engineer/ElectricalEngineer would become http://schema.org/ElectricalEngineer if incorporated into the official vocab |
| 09:06 | <mhausenblas> | mhm |
| 09:06 | <jgraham> | AryehGregor: You could try something like WebDriver for automating boring manual tests although there are risks |
| 09:06 | <mhausenblas> | but it's also about properties and their domain/ranges |
| 09:06 | <mhausenblas> | there is a hell of a lot one can get wrong |
| 09:07 | <foolip> | seems to me that schema.org is too big to begin with, encouraging people to extend it even more seems... unwarranted |
| 09:07 | <mhausenblas> | so, gotta run soon but if you like to chime in, lemme know - either clone https://github.com/mhausenblas/schema-org-rdf and send in pull request or I can add you to the repo |
| 09:07 | <mhausenblas> | dunno what you mean w/ 'too big to begin with' |
| 09:07 | <foolip> | ok, I'll have a look at the examples tonight when I'm not supposed to be working :) |
| 09:07 | <mhausenblas> | it's there |
| 09:08 | <mhausenblas> | ahm, there are no examples yet ;) |
| 09:08 | <mhausenblas> | nor mappings |
| 09:08 | <mhausenblas> | jsut started y'day |
| 09:08 | <mhausenblas> | OK, I see, you gotta work as well - have fun foolip :D |
| 09:08 | <foolip> | I mean that they seem to have a way bigger vocabulary than they can probably make any good use of for search results |
| 09:08 | <mhausenblas> | 288 class and 182 props? yes, true |
| 09:09 | <foolip> | which means that a lot of it probably isn't terribly well thought out and will just end up being cruft forever |
| 09:09 | mhausenblas | sorta agrees |
| 09:09 | <foolip> | so, I'm still on the fence about fanboy/flamerboy :) |
| 09:09 | <mhausenblas> | I guess they'll assess what people really use and to which extent |
| 09:09 | <hsivonen> | http://lists.w3.org/Archives/Public/www-validator/2011Jun/0043.html makes it look like someone has managed to sell SEO to Illinois state. Yay for tax dollars. |
| 09:09 | <mhausenblas> | come on - that's a great thing, no doubt |
| 09:10 | <foolip> | I don't know, I can totally imagine people going to great lengths to use lots and very specific parts this vocabulary as a SEO incantation even though it has no effect whatsoever on search results |
| 09:10 | <mhausenblas> | btw, my current favourite script is curl http://omnidator.appspot.com/mdpretty?url=$1 |
| 09:11 | <foolip> | so it will be used incorrectly, and they will be unable to start using it for search results. |
| 09:11 | <mhausenblas> | that might well be, but think beyond SEO |
| 09:11 | <mhausenblas> | that the future will tell |
| 09:11 | <foolip> | if people use these things in completely the wrong way, it breaks it for everything, not just SEO, right? |
| 09:12 | <mhausenblas> | well, that's why I think guidance is top priority |
| 09:12 | <mhausenblas> | you do a great job already there foolip and it's good to see that guha personally is very responsive |
| 09:12 | <foolip> | I guess we'll see |
| 09:12 | <mhausenblas> | yup |
| 09:12 | <mhausenblas> | guess we both have to earn our bread now |
| 09:13 | <foolip> | right |
| 09:13 | <foolip> | talk to you later :) |
| 09:13 | mhausenblas | wondering, can you say that in English - in Austria we use this idiom |
| 09:13 | <mhausenblas> | cya |
| 09:13 | <foolip> | you can say it in Swedish at least :) |
| 09:13 | <mhausenblas> | and thanks for your time foolip - much appreciated |
| 09:13 | <foolip> | likewise! |
| 09:13 | <mhausenblas> | he he, good to learn that |
| 09:13 | <mhausenblas> | what would be the Swedish phrase? |
| 09:14 | <foolip> | "att tjäna sitt dagliga bröd" |
| 09:14 | <jgraham> | (ir is pretty much something you can say in English, although I can't think of the exact idiom right now) |
| 09:14 | <mhausenblas> | tx |
| 09:14 | <mhausenblas> | he he, tx to jgraham as well ;) |
| 09:16 | mhausenblas | thinks 'to earn one's crust' might be the right thing to say |
| 09:20 | <MikeSmith> | hsivonen: in the interest of closing out http://bugzilla.validator.nu/show_bug.cgi?id=842 (srcdoc attribute on iframe in HTML5 does not validate), is it OK with you if I add srcdoc to the schema, but with the datatype set to string for now, and add a FIXME noting that the real datatype checking for it needs to be added later? |
| 09:21 | <hsivonen> | MikeSmith: yeah, but please also add it to the "not implemented in browsers yet" warnings |
| 09:22 | <MikeSmith> | hsivonen: I'm pretty sure you added it there already |
| 09:22 | <MikeSmith> | so I think that's covered |
| 09:22 | <MikeSmith> | but I'll doublecheck |
| 09:29 | <MikeSmith> | hmm, incidentally, I don't see how srcdoc would work in XML documents |
| 09:30 | <MikeSmith> | without violating XML well-formedness constraints |
| 09:30 | <hsivonen> | MikeSmith: IIRC, you escape what XML requires but the payload is HTML |
| 09:30 | <MikeSmith> | ah |
| 09:41 | <david_carlisle> | hsivonen: Blurg that's what the spec says, but why doesn't it parse as xml in an xml context (like innerhtml)? |
| 09:41 | <david_carlisle> | channel topic. I suppose... |
| 09:42 | <hsivonen> | david_carlisle: In my opinion, it was a mistake to make innerHTML depend on how the DOM got created |
| 09:43 | <hsivonen> | david_carlisle: that is to say: If you change the serialization of the docement itself, I think it's a mistake that innerHTML starts wanting different input, too |
| 09:43 | <david_carlisle> | hsivonnen: actually for innerhtml I'd probably agree since it's a post parse dom thing, but srcdoc is (more or less) a parse time feature isn't it |
| 09:43 | <hsivonen> | david_carlisle: I think it would be a similar mistake if srcdoc behavior changed |
| 09:44 | <hsivonen> | david_carlisle: as I understand it, you could set the srcdoc attribute value from JS |
| 09:44 | <david_carlisle> | hsivonen: ah |
| 09:44 | <david_carlisle> | hsivonen: stop being right, makes arguing with you harder |
| 09:45 | <david_carlisle> | in a perfect world though, people of a same disposition ought to be able to just use the xhtml syntax and not know the html syntax at all. there's always data:text/xml i suppose... |
| 09:46 | <david_carlisle> | s/same/sane/ |
| 09:47 | <hsivonen> | david_carlisle: if there was even near-equal author interest in text/html and application/xhtml+xml, it would make sense to have an XML-enabled srcdoc equivalent without data: URLs |
| 09:48 | <hsivonen> | david_carlisle: but in practice, normal authors outside the W3C are probably not going to care |
| 09:49 | <hsivonen> | aside: this recent idea that authors should have Choice between RDFa, Microformats and Microdata, completely ignores the implementor side. |
| 09:49 | <david_carlisle> | hsivonen: I suppose a world view where most people use xml and have as much mathematics as text in their documents counts me out of the "normal user" category |
| 09:49 | <hsivonen> | it's bad enough to have the Choice between application/xhtml+xml and text/html |
| 09:50 | <david_carlisle> | :-) |
| 09:50 | <hsivonen> | david_carlisle: you can now have math in text/html! |
| 09:50 | <david_carlisle> | hsivonen: Yeh! |
| 09:51 | <MikeSmith> | users should be able to create their own arbitrary syntaxes, and browsers implementors should have to provide a mechanism to let them do that |
| 09:51 | <MikeSmith> | that's true extensibility |
| 09:51 | <MikeSmith> | I want to write my HTML using s-expressions |
| 09:52 | <jgraham> | MikeSmith: I believe several browsers do that |
| 09:52 | <MikeSmith> | and I should just be able to do that |
| 09:52 | <david_carlisle> | hsivonen: Actually if you could get a few of your competitors to implement that, I'd dump the application/xhtml+xml version of our documentation base in a flash, I want xml everywhere inside our firewall but I'd be perfectly happy for it to be all text/html as published, if it worked |
| 09:52 | <jgraham> | It's called clone+fork |
| 09:52 | <MikeSmith> | no, the browser vendors should do it for me |
| 09:52 | <MikeSmith> | the browser vendors are all holding me back |
| 09:52 | <hsivonen> | david_carlisle: didn't Apple announce it for Safari 5? |
| 09:52 | <david_carlisle> | yes |
| 09:52 | <MikeSmith> | jgraham: they have all the power and I have none, and that's just not fair |
| 09:53 | <hsivonen> | david_carlisle: and Opera has Ragnarök in the pipeline though the math rendering isn't quite up to the quality one would want |
| 09:53 | <jgraham> | MikeSmith: They also have a good line in evil laughs |
| 09:53 | <hsivonen> | MikeSmith: search engines are now holding you back, too |
| 09:53 | <MikeSmith> | hsivonen: exactly |
| 09:54 | <david_carlisle> | anyone from Chrome: speak up now:-) |
| 09:55 | <MikeSmith> | "The typemustmatch attribute is a boolean attribute whose precense indicates..." |
| 09:55 | <jgraham> | hsivonen: Presumably david_carlisle is partially responsible for the MathML CSS Profile stuff :) |
| 09:55 | <MikeSmith> | I guess that's kind of like incense, but before you light it |
| 09:56 | <MikeSmith> | speaking of incense, time for some j |
| 09:57 | <david_carlisle> | jgraham: Well I partially have my name on the front, but really Giorgi from Opera did all the work |
| 09:59 | <jgraham> | david_carlisle: Like I said you are partially responsible for it :) |
| 09:59 | jgraham | is on the record as thinking it is not the greatest idea in the world |
| 09:59 | <Ms2ger> | Opera? |
| 10:01 | <jgraham> | Ms2ger: :p |
| 10:01 | <Ms2ger> | :) |
| 10:04 | jgraham | wonders if he should ignore the fact that Janina accused someone of lying and then followed up with what amounts to "what you said is basically true, but the situation is more complex than you realised" |
| 10:05 | <Ms2ger> | My approach is to ignore everything that happens in the HTML WG |
| 10:05 | <Ms2ger> | Works well, IMO :) |
| 10:07 | <jgraham> | Yes, I see the merit in that approach for sure |
| 10:16 | <MikeSmith> | hmm, I had been hoping CSSLint did actual validation of property names |
| 10:17 | <MikeSmith> | but I see it doesn't |
| 10:17 | <MikeSmith> | by design |
| 10:17 | <MikeSmith> | it would be nice to have an up-to-date CSS validator |
| 10:27 | <zcorpan> | anyone care about whether broken utf-8 in websockets should be U+FFFDd or fail the connection? |
| 11:10 | <hsivonen> | http://www.brucelawson.co.uk/2011/modal-dialogues-in-html5/ I thought dialogs in the UI sense were dialogs rather than dialogues even in en-GB. have I been mistaken for all these years? |
| 11:11 | <gsnedders> | hsivonen: Different style-guides say different things |
| 11:11 | <hsivonen> | gsnedders: ok |
| 11:12 | <gsnedders> | hsivonen: dialog is probably more common, though |
| 11:13 | <david_carlisle> | hsivonen: en-gb is whatever you want, for most computing things i'd follow us usage though (a fortran program but a tv programme, a css font but a church fount, ...) |
| 11:13 | gsnedders | has never come across that spelling of fount |
| 11:14 | <david_carlisle> | en-gb is whatever you want:-) |
| 11:14 | <hsivonen> | david_carlisle: is a church fount etymologically related to typesetting font? |
| 11:15 | <david_carlisle> | yes I think so, and old english typesetting books spell it as fount |
| 11:15 | <hsivonen> | ok |
| 11:15 | gsnedders | looks up in OED… |
| 11:16 | <gsnedders> | fount (for type) comes later than font (for type). |
| 11:17 | <gsnedders> | From the French fonte |
| 11:18 | <gsnedders> | The church font (always font), frm Latin font-em |
| 11:19 | <hsivonen> | the "CSS should support 'colour' folks" never seem to ask for "fount", too |
| 11:20 | <gsnedders> | The OED notes for that sense fount is more common in England than font. |
| 11:22 | <jgraham> | I had a "computing" practial in first year undergard |
| 11:22 | <jgraham> | Let's try that again |
| 11:23 | <jgraham> | I had a "computing" practical in first year undergrad |
| 11:23 | <jgraham> | Which iirc was basically using Office |
| 11:23 | <jgraham> | The hardest part was doing some numberial model in Excel |
| 11:24 | <jgraham> | *numerical |
| 11:24 | <gsnedders> | jgraham: I guess this wasn't the CS option. :) |
| 11:24 | <jgraham> | Anyway at one point you were required to type something out about typefaces involving the word "fount" |
| 11:25 | <jgraham> | And you were marked down for excluding the "u" |
| 11:25 | <jgraham> | In other news, Frank King is quite nuts |
| 11:28 | <jgraham> | (disclaimer: I'm sure he isn't nutes really. But he does insist on the use of archaic spelling) |
| 11:30 | <gsnedders> | Only recently archaic, though, as far as I can tell. |
| 12:42 | hsivonen | is surprised (in a positive way) to see someone other than the usual suspects say something (and reasonable!) in the longdesc debate |
| 12:44 | <jgraham> | hsivonen:? |
| 12:44 | <hsivonen> | the "stop right there" response wasn't so positive |
| 12:45 | <hsivonen> | jgraham: I'm catching up with public-html |
| 12:45 | <jgraham> | Ah |
| 12:45 | <hsivonen> | and just read Matthew Turvey's email |
| 12:45 | hsivonen | spent yesterday traveling and settling in the new location |
| 12:51 | <jgraham> | The PFWG list of "requirements" seems like a clear attempt to shift the goalposts to me. I get the stong impression that they were worked out by writing down all the positive properties of longdesc, writing down all the positive properties of other solutions, and eliding the intersection |
| 12:51 | <jgraham> | But I guess no good will come of getting more involved |
| 12:52 | <jgraham> | eliding the intersection... from the longdesc side and billing the remainder as "requirements" |
| 13:08 | <zcorpan> | hsivonen: a common authoring problem apparently is using <a href><div> and not setting the a to display:block which causes rendering issues in ie |
| 13:08 | <zcorpan> | hsivonen: not that v.nu checks stylesheets, but if it did... |
| 13:11 | <MikeSmith> | somebody should write a modern CSS validator |
| 13:16 | <MikeSmith> | zcorpan: on another topic, didn't you raise a valiator.nu bug a while back proposing a way to trim down the allowed-attribute spec fragments based on what the type of the attribute is? |
| 13:16 | <hsivonen> | zcorpan: might be worthwhile to warn about that |
| 13:16 | <MikeSmith> | *based on what the value of the type attribute is |
| 13:17 | <MikeSmith> | the allowed-attributes spec fragment that gets emitted for input continues to confuse people - |
| 13:17 | <MikeSmith> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=11048 |
| 13:17 | <MikeSmith> | we really ought to do something to fix that |
| 13:18 | <zcorpan> | ah yeah |
| 13:18 | <zcorpan> | at first i had no idea what you were talking about :) |
| 13:18 | <MikeSmith> | heh |
| 13:18 | <MikeSmith> | if you can find the bug number, I'd appreciate it |
| 13:20 | <MikeSmith> | oh |
| 13:20 | <MikeSmith> | 339 |
| 13:20 | <MikeSmith> | http://bugzilla.validator.nu/show_bug.cgi?id=339 |
| 13:20 | <MikeSmith> | I thought it was a new one but I guess it was just that I re-read it recently |
| 13:21 | <zcorpan> | are v.nu bugs that are autosubmitted from spec checkins annotated in some way so they can be excluded from searches? |
| 13:21 | <MikeSmith> | zcorpan: no, unfortunately :( |
| 13:21 | <MikeSmith> | not as far as I know at least |
| 13:22 | <MikeSmith> | maybe hsivonen knows |
| 13:22 | <zcorpan> | is that thing still enabled? |
| 13:22 | <hsivonen> | zcorpan: you could try to to filter by the URL field |
| 13:22 | <hsivonen> | zcorpan: no longer enabled |
| 13:23 | <zcorpan> | ah |
| 13:23 | <zcorpan> | ok |
| 13:24 | <hsivonen> | http://lists.w3.org/Archives/Member/tag/2011Jun/0040.html (Member-only; the To field is worth checking out in the light of the Subject field) |
| 13:26 | hsivonen | also notes http://lists.w3.org/Archives/Member/tag/2011Jun/0050.html |
| 13:30 | <MikeSmith> | zcorpan: about the input error-message thing, your suggestion in http://bugzilla.validator.nu/show_bug.cgi?id=339#c0 seems like the best way to handle it, so I guess I'll try to figure out how to implement that |
| 14:01 | zcorpan | learns about http://csslint.net/ |
| 14:24 | <rimantas> | not a good thing to learn about :( |
| 14:25 | <hsivonen> | rimantas: why not? did the lint hurt your feelings? |
| 14:26 | <rimantas> | hsivonen, yup, along with oocss |
| 14:26 | <hsivonen> | what's oocss? |
| 14:27 | <rimantas> | they should have called it oocss lint, that would make more sense then |
| 14:27 | <rimantas> | oocss is Nicole's brainchild: https://github.com/stubbornella/oocss/wiki/faq |
| 14:29 | MikeSmith | wishes somebody would brainchild a better css validator |
| 14:30 | jgraham | wishes the first FAQ was "WTF is this?" |
| 14:30 | <nlogax> | how hard is it to get patches into the w3 validator? |
| 14:31 | <MikeSmith> | nlogax: the html4 one? |
| 14:31 | <nlogax> | MikeSmith: the CSS validator i mean |
| 14:31 | <MikeSmith> | dunno |
| 14:32 | <mpilgrim> | hsivonen: rel=prerender, but apparently it's already listed on the microformats wiki |
| 14:32 | <MikeSmith> | nlogax: nobody is actually maintaining it actively any more |
| 14:32 | <hsivonen> | mpilgrim: it was already in the old registry |
| 14:33 | <nlogax> | MikeSmith: oh i see |
| 14:33 | <hsivonen> | mpilgrim: it was one of the very few proper registations in the old wiki, so I registered it in the new registry |
| 14:33 | <MikeSmith> | nlogax: it's kind of past the point of being worth patching |
| 14:33 | <mpilgrim> | when was it registered? did google put out a whitepaper on it a while ago or something? |
| 14:33 | <mpilgrim> | the feature was just announced like 2 days ago |
| 14:35 | <MikeSmith> | the oocss FAQ question "I need more than six (h1-h6) headings on my site. How do I add more?" needs a better answer |
| 14:36 | <hsivonen> | MikeSmith: registered in the old registry on May 17th |
| 14:36 | <hsivonen> | oops |
| 14:36 | <hsivonen> | mpilgrim: ^ |
| 14:36 | <mpilgrim> | wow |
| 14:36 | <hsivonen> | I guess this was a failure for the old registry to point properly to the new one |
| 14:38 | <hsivonen> | mpilgrim: it's nice to see a rel extension that comes with documentation with the words "Corner Cases" in a section heading |
| 14:38 | <hsivonen> | mpilgrim: quite refreshing compared to all the cargo-cult stuff out there |
| 14:40 | <hsivonen> | I'm still amazed that Dublin Core stuff hasn't been properly registered |
| 14:40 | <hsivonen> | are DC users boycotting the registry? |
| 14:40 | <hsivonen> | or are the specs too tedious to navigate even for people who put DC stuff on their pages |
| 14:49 | <Philip`> | jgraham: Did Frank King's numerical model happen to involve a bell and/or a sundial? |
| 14:50 | <jgraham> | Philip`: You would have thought so wouldn't you? |
| 14:50 | <jgraham> | As far as I recall they didn't, however |
| 14:52 | Philip` | has had seemingly nuttier lecturers, like one whose introduction-to-Java lecture notes were full of full-page photos of him wearing silly hats |
| 14:55 | jgraham | had one that decided it would be a good idea to sing an end-of-course song |
| 14:55 | <jgraham> | But I don't think he can hold a candle to Frank King |
| 14:56 | <Philip`> | Did he/she ask everyone to join in the singing? |
| 14:57 | <jgraham> | I don't remember. Terrifyingly this was almost a decade ago |
| 14:57 | <jgraham> | I have some dim suggestion that maybe we were expected to koin in for the final chorus |
| 14:57 | <jgraham> | *join |
| 14:58 | <jgraham> | But that could be a false memory |
| 14:58 | <Philip`> | That sounds potentially excruciating |
| 14:58 | <jgraham> | I do, however, remember the couplet "If you're feeling like a cabbage / sitting in the Babbage" quite clearly |
| 15:47 | <hsivonen> | so if crossorigin and typemustmatch get reverted from W3C HTML5, will they just get the same preprocessor treatment that <a ping> gets now? |
| 15:54 | <erlehmann> | preprocessor treatment? |
| 15:56 | <hsivonen> | erlehmann: there's a preprocessor that removes and splits stuff when generating the W3C specs and removes other stuff when generating the WHATWG spec |
| 15:57 | <erlehmann> | ah. |
| 17:31 | <TabAtkins> | hsivonen: Presumably, yes. |
| 17:59 | <Hixie> | hsivonen: definitely -- even more so than ping, since both of those are important security features |
| 17:59 | <Hixie> | btw re srcdoc="", the reason i didn't make it work usefully in xml is that xml's escaping rules make it not really any more useful than data: URLs |
| 18:04 | <Hixie> | could someone explain to Jukka that HTML is not related to SGML anymore so the term "DTD" is meaningless in this context? |
| 18:05 | <Ms2ger> | Would it help? |
| 18:38 | <jarib> | i think i've found a bunch of invalid WebIDL in the HTML spec, at least my parser complains about "SomeInterface?" and "object?" |
| 18:38 | <jarib> | since those types are already nullable, the ? isn't right |
| 18:38 | <Ms2ger> | jarib, your parser is out of date |
| 18:38 | <Ms2ger> | At least a week |
| 18:38 | <jarib> | Ms2ger: so this isn't up to date http://www.w3.org/TR/WebIDL/#idl-nullable-type ? |
| 18:39 | <Ms2ger> | No |
| 18:39 | <jarib> | aha |
| 18:39 | <Ms2ger> | Try the dev.w3.org draft |
| 18:39 | <Hixie> | TR/ page strikes again! |
| 18:39 | <Hixie> | you want http://dev.w3.org/2006/webapi/WebIDL/ |
| 18:39 | <jarib> | i always get confused :) |
| 18:39 | <Ms2ger> | So how about we solve that by not publishing new drafts on TR/... |
| 18:39 | <jarib> | thanks, i'll fix the parser |
| 18:41 | Ms2ger | hopes Hixie keeps up the bug fixing rate |
| 18:44 | <MikeSmith> | somebody here said earlier, "/TR stands for Trash" |
| 18:45 | <MikeSmith> | which is a dumb thing to say |
| 18:45 | <MikeSmith> | whoever said that |
| 18:48 | <TabAtkins> | It is, unfortunately, often accurate. |
| 18:48 | <Hixie> | any IE users around? i need a test of http://junkyard.damowmow.com/471 |
| 18:48 | <MikeSmith> | btw, that 2006 in the dev URL al |
| 18:48 | <MikeSmith> | *also is really helpful for readers |
| 18:49 | <Hixie> | i stopped being impressed with the bogus years in urls when i started seeing TWO bogus dates in some w3c urls |
| 18:49 | <Hixie> | that's where it's at |
| 18:50 | <TabAtkins> | Hixie: After hitting "Test", I get "script" and "new-original". After hitting "Reset" I get "new-original" and "original". |
| 18:50 | <TabAtkins> | ie9 |
| 18:51 | <Hixie> | thanks |
| 18:51 | <Hixie> | and also, what? |
| 18:51 | <Hixie> | that's even more crazy than what opera does |
| 18:51 | <Hixie> | fucking browsers |
| 18:55 | <clair> | Hixie: I noticed you've written up some stuff on <dialog> - sorry I've not done anything yet, promise I'll do some research on JS libraries by the end of the weekend! |
| 18:55 | <Hixie> | no worries! |
| 18:55 | <Hixie> | the more the merrier |
| 18:56 | <clair> | Yeah, just didn't want you thinking I'd cleared off and decided not to do anything :) |
| 18:56 | <Hixie> | :-) |
| 18:56 | <clair> | I was going to do stuff last weekend but ended up, er, clothes shopping |
| 18:56 | <clair> | (As you do) |
| 18:56 | <Hixie> | (http://wiki.whatwg.org/wiki/Dialogs if anyone else wants to contribute, btw) |
| 18:57 | <Hixie> | clair: cool |
| 18:57 | <Hixie> | clair: if you do have time to help out, one thing that would help that hasn't been done at all yet is screenshots and snippets of markup and script showing how the pages are doing it now |
| 18:59 | <TabAtkins> | Dammit, why do list markers have to be so damn magical?!? >_< |
| 18:59 | <Ms2ger> | TabAtkins, it's the web |
| 18:59 | <clair> | I wasn't sure if code snippets would've been too much at this stage, but sure I can do that |
| 19:04 | <TabAtkins> | Ms2ger: But now I have no idea where to place ::marker pseudos in the element-tree, unless I punt and make position:marker *even more* magical. ;_; |
| 19:05 | <Ms2ger> | It isn't in the element tree |
| 19:05 | <Ms2ger> | When are you defining the box tree again? |
| 19:08 | <TabAtkins> | Pseudos live in the element-tree. |
| 19:09 | <TabAtkins> | (Thus the pseudo-*element* part.) |
| 19:09 | <TabAtkins> | And probably sometime next year. |
| 19:09 | <TabAtkins> | If fantasai doesn't beat me to it. |
| 19:18 | <AryehGregor> | See, if apple.com had enabled SPF with hard fail instead of soft fail, that spam mail to whatwg would have bounced. |
| 19:18 | <AryehGregor> | And also all forwarding would break for Apple employees. |
| 19:18 | <AryehGregor> | Sigh. |
| 19:18 | <AryehGregor> | I hate the Internet sometimes. |
| 19:20 | <Hixie> | i don't understand why that spam mail went through in the first place |
| 19:20 | <AryehGregor> | Why shouldn't it? |
| 19:20 | <Hixie> | none of the addresses are subscribed |
| 19:20 | <AryehGregor> | Hmm. |
| 19:21 | <AryehGregor> | Maybe it subscribed and then unsubscribed for some reason? |
| 19:21 | <Hixie> | maybe |
| 19:21 | <Hixie> | though how? |
| 19:21 | <Hixie> | they'd need access to the account |
| 19:22 | <Hixie> | who filed http://www.w3.org/Bugs/Public/show_bug.cgi?id=12470 and could they explain to me how you can have horizontal top and bottom margins? |
| 19:22 | <AryehGregor> | Oh, right . . . |
| 19:22 | <AryehGregor> | Good question, then. |
| 19:24 | <Ms2ger> | Hixie, wasn't me, but I don't think the spec is clear |
| 19:24 | <Hixie> | well i added the word vertical to make sure |
| 19:24 | <Hixie> | but i don't understand how else it could be interpreted |
| 19:25 | <Ms2ger> | You could read it as "all the margins of elements at the top..." |
| 19:25 | <Hixie> | i guess |
| 19:25 | Ms2ger | shuts up and lets Hixie work :) |
| 19:26 | <Hixie> | i guess i should go to the office |
| 19:26 | <Hixie> | bbiab. |
| 19:50 | <AryehGregor> | Wow. initKeyEvent is horrifying. Ten arguments, of which six are booleans. |
| 19:51 | <Ms2ger> | Yes. |
| 19:55 | <AryehGregor> | Argh, drat it, another bounce. |
| 20:38 | <TabAtkins> | AryehGregor: I wonder who came up with initKeyEvent first? Also, I wonder if that person hates the entire world, or just web devs in particular. |
| 21:19 | <jgraham> | TabAtkins: Damn you nerd sniped me |
| 21:19 | <jgraham> | +, |
| 21:19 | <jgraham> | But I can't find who is responsible for initKeyEvent |
| 21:19 | <jgraham> | I bet it was made up in a telecon |
| 21:20 | <jgraham> | However I did find http://lists.w3.org/Archives/Public/www-dom/1997JulSep/0007.html |
| 21:20 | <jgraham> | Which is very interesting |
| 21:29 | <TabAtkins> | jgraham: Interesting indeed. |
| 21:30 | <jgraham> | And makes for a fun human-interest angle since Tim Bray "invented" draconian XML and it seems his wife narrowly missed standardising HTML error recovery 10 years before Hixie finally succeeded |
| 21:30 | <TabAtkins> | Ah, didn't realize Lauren Wood was his wife. |
| 21:31 | <jgraham> | I believe so, although I have been mistaken about bigger things |
| 21:32 | <TabAtkins> | Hm, the newest "let's hash <input type=password>" suggestion seems pretty decent. |
| 21:32 | <jgraham> | (also, I guess Tom Pixley is the most likely candidate for initEvent and friends) |
| 21:32 | <TabAtkins> | (Except for suggesting sending the extra information in a header instead of an extra form input.) |
| 21:36 | jgraham | wonder why "Host cannot validate password requirements" is listed as a disadvantage |
| 21:37 | <TabAtkins> | Heh, indeed. |
| 21:37 | TabAtkins | will put that in his response email, as it's good snark (and a good point). |
| 21:41 | <TabAtkins> | Possible disadvantage - a good server uses a good crypto hash, which is slow to evaluate, forcing each login attempt to take a non-trivial amount of time. If the attacker can use their own resources to compute the cryptographic hash quickly relative to the attempts (frex, putting several machines to work hashing for each one making requests), they can bypass that. |
| 21:41 | <TabAtkins> | But then, those are machines not being spent making requests, so that may end up being a wash. |
| 21:43 | <AryehGregor> | It's not sane to distribute the actual hash computation, when you're doing brute-force hashing. |
| 21:43 | <AryehGregor> | Just have each node compute its own hash, it's embarrassingly parallel. |
| 21:43 | <TabAtkins> | I wasn't meaning to distribute the individual hashings, but to have several machines hashing for each machine making requests. |
| 21:44 | <AryehGregor> | Doesn't each request only submit one hash? |
| 21:44 | <TabAtkins> | Yes. |
| 21:44 | <TabAtkins> | But requests are faster than hashing with a good hash. |
| 21:44 | <AryehGregor> | Well, "good" for password storage, anyway. |
| 21:44 | <TabAtkins> | (The full rtt may not be, but you don't have to worry about that.) |
| 21:44 | <TabAtkins> | Yes, that's the contexty. |
| 21:45 | <TabAtkins> | s/y// |
| 21:45 | <AryehGregor> | Why wouldn't you just have each node both hash and submit at the same time? |
| 21:45 | <AryehGregor> | Submitting should take negligible CPU time. |
| 21:45 | <AryehGregor> | Compared to hashing. |
| 21:45 | <AryehGregor> | It doesn't make sense to forward the hashes to other nodes to submit. |
| 21:45 | <TabAtkins> | Then you've dropped down to the same rate that you'd get if the server was computing the hash itself. |
| 21:46 | <TabAtkins> | But like I said, it sounds like it'd be a wash anyway. |
| 21:46 | AryehGregor | looks at the thread |
| 21:46 | <TabAtkins> | Or rather, it sounds like it's exactly identical, except with useless message-passing. |
| 21:47 | <AryehGregor> | Right, that's what I meant. |
| 21:47 | <TabAtkins> | Yes. |
| 22:34 | <AryehGregor> | Wow, I spent ridiculously long replying to that post. |
| 22:34 | <AryehGregor> | Like almost an hour. Definitely not worthwhile. |
| 22:34 | <AryehGregor> | (what with me not being paid for it, I mean) |
| 22:34 | AryehGregor | goes back to work for the next not very many minutes |
| 22:35 | <Hixie> | which post? i want to waste some time too! |
| 22:35 | <AryehGregor> | The one about password hashing. |
| 22:35 | <Hixie> | oh, bummer, it's a whatwg mail |
| 22:35 | <AryehGregor> | Yeah. |
| 22:35 | <AryehGregor> | This is actually a fairly good proposal. |
| 22:35 | <Hixie> | :-P |
| 22:35 | <AryehGregor> | I just wish I could see some way to work proper salting into it. |
| 22:37 | <Hixie> | biggest problem with it is nobody will understand it |
| 22:37 | <Hixie> | but they'll get a false sense of security |
| 22:37 | <Hixie> | the proposal itself could be used securely |
| 22:37 | <Hixie> | it's like postMEssage() -- it can be used securely, but the API is widely misunderstood and misused |
| 22:38 | <clair> | Thing is, those browsers that don't do the hashing will send over plain text passwords and naive coders will just store that |
| 22:38 | <clair> | If I understand the post correctly... |
| 22:40 | <AryehGregor> | Hixie, if I know that my password is being hashed with SHA256 100,000 times before submission, that gives me a fairly substantial security guarantee as a user. |
| 22:40 | <AryehGregor> | If I could somehow also know that a long random per-user salt was being used, that's basically all the security guarantee I need. |
| 22:40 | <AryehGregor> | The latter is what's tricky. |
| 22:40 | <AryehGregor> | But at least the former is better than nothing. |
| 22:41 | <AryehGregor> | Also, it's better if the client transmits the hash instead of the plaintext password, to make life harder for eavesdroppers. |
| 22:41 | <TabAtkins> | Hixie: The benefit is that you don't need to understand it. Right now, a lot of devs just store the plaintext password directly. With @hash, they'd just store the hash directly. No change in their life. |
| 22:42 | <TabAtkins> | As opposed to making people understand that they should throw away the plaintext password as soon as possible. |
| 22:42 | <TabAtkins> | AryehGregor: One downside of a per-user salt is that fingerprinting suddenly gets trivial. |
| 22:42 | <AryehGregor> | TabAtkins, huh? What do you mean? |
| 22:42 | <AryehGregor> | The per-user salt has to be supplied by the site, obviously. |
| 22:43 | <TabAtkins> | Oh, okay. |
| 22:43 | <AryehGregor> | It can't be supplied by the browser, because the user might use different browsers. |
| 22:43 | <Hixie> | AryehGregor: your browser could do that today (indeed that's what systems like lastpass do, no?) |
| 22:44 | <Hixie> | AryehGregor: "this" being the "hashed with SHA256 100,000 times before submission" stuff |
| 22:44 | <TabAtkins> | querySelector("input[name=username]").onchange = function() { querySelector("input[name=password]").salt = this.value; }; |
| 22:44 | <AryehGregor> | Hixie, yes, but then you have to use the same browser all the time, because the site doesn't know about it. |
| 22:44 | <AryehGregor> | No logging into sites from a friend's computer. |
| 22:44 | <Hixie> | AryehGregor: or the same plugin |
| 22:45 | <AryehGregor> | Generally my friends would be annoyed if I tried to install a plugin in their browser, I'd think. Operators of public computers would be even more annoyed. |
| 22:45 | <AryehGregor> | There's clear benefit to having the site involved in the process. |
| 22:45 | <Hixie> | anyway, it'll end up on my WF3 pile and will sit there until browsers are interoperable and pretty on the wf2 stuff :-) |
| 22:45 | <Hixie> | AryehGregor: such systems have web sites you can use and copy and paste the password, too |
| 22:46 | <Hixie> | AryehGregor: i'm not saying it's not better to have a standard way to do it, but i'm not sure it's better to have the site do it |
| 22:47 | <AryehGregor> | Bigger problem: your way requires user opt-in and awareness. |
| 22:47 | <AryehGregor> | If it's done by the author, all users benefit. |
| 22:48 | <jgraham> | Yes, somehing that requires users to use lastpass or whatever is hopeless |
| 22:48 | <jgraham> | That protects the 0.1% of users who were probably the best protected anyway |
| 22:48 | <AryehGregor> | If any significant number of savvy users come to expect it, and they can easily tell whether it's happening (like if the input displays slightly differently and it's hard for authors to forge the appearance), they'll pressure authors of large sites to use it. |
| 22:49 | <zewt> | sort of wondering if it could reuse something like SCRAM, instead of reinventing the wheel, but that's a real challenge-response mechanism so it's not quite the same |
| 22:50 | <AryehGregor> | What everyone really should use is SRP. |
| 22:50 | <AryehGregor> | TLS over SRP is the cure to all phishing problems. |
| 22:50 | <AryehGregor> | . . . As long as everyone uses it. |
| 22:50 | <zewt> | don't know anything about it |
| 22:51 | <AryehGregor> | SRP is a type of PAKE. It's basically like symmetric-key encryption, where the password is the key, and executing the protocol gives the client or server or any eavesdropper close to zero information about the key if they don't know it already. |
| 22:52 | <AryehGregor> | So you type in your password, and if the server doesn't know the password already, the encryption fails and the server gains no information about the password -- except that it gets one guess per connection attempt, obviously. |
| 22:52 | <zewt> | well, any competently-designed challenge-response mechanism should do that |
| 22:52 | <AryehGregor> | Really? How? |
| 22:52 | <AryehGregor> | It's quite nontrivial. |
| 22:53 | <Hixie> | AryehGregor: i wasn't suggesting using lastpass, i was suggestion that browsers should just do it |
| 22:53 | <AryehGregor> | Requires a bunch of modular arithmetic and stuff. |
| 22:53 | <AryehGregor> | Hixie, like Mozilla's Account Manager, basically? But they can't do it without the user opting in and understanding the system somewhat, because otherwise the same password won't work cross-browser. |
| 22:53 | <AryehGregor> | If it's opt-in, you lose 95% of the benefit. |
| 22:53 | <zewt> | what information can you get out of a SCRAM response, other than "it matches or it doesn't"? |
| 22:53 | <TabAtkins> | AryehGregor: That doesnt' seem to solve the problem of "servers shouldn't ever remember the password". |
| 22:54 | <zewt> | (note: not claiming to be a huge fan of SCRAM in particular, it's just one I happened to implement recently) |
| 22:54 | <AryehGregor> | TabAtkins, more sophisticated variants don't require the server to know the password. I think SRP falls into this category. |
| 22:54 | <TabAtkins> | kk |
| 22:54 | <AryehGregor> | zewt, Wikipedia doesn't know about SCRAM. |
| 22:54 | <AryehGregor> | I question its existence. |
| 22:54 | <zewt> | http://tools.ietf.org/html/rfc5802 |
| 22:56 | AryehGregor | doesn't have time to research it, but maybe it's similar to SRP |
| 22:56 | <zewt> | (sorry if that's not a quick read; IETF and all) |
| 22:57 | <David_Bradbury> | Any ideas if Canvas will ever support 3D transformations? |
| 22:57 | <zewt> | gah, why does gmail randomly log me out; I'm using it, thanks |
| 22:59 | <David_Bradbury> | 3D transformations in the 2D Context* |
| 23:02 | <AryehGregor> | "Although Google.com is the most high-profile site to use this new prerendering technology, it can be used by other sites since it’s been designed as a web standard." A.k.a. "We made a new link relation without talking it over with other implementers and we published some documentation on our website, therefore it's a web standard." |
| 23:02 | <AryehGregor> | Was it ever proposed to a standards list anywhere? |
| 23:02 | <AryehGregor> | (Or maybe it is in a standard, but they didn't seem to mention that anywhere . . .) |
| 23:03 | <TabAtkins> | Don't believe so, no. |
| 23:05 | <jgraham> | AryehGregor: Pretty sure Google have a history of using "web standard" to mean "propriatary invention with documentation" |
| 23:05 | <jgraham> | But I can't think of the specific examples so don't ask |
| 23:05 | <jgraham> | :) |
| 23:05 | <Lachy> | dammit. I didn't want the whole scope selector debate to start up again. :-(. I've been over it many times already in designing selectors api, and all the suggestions so far are the same as those that were rejected for very good reasons. |
| 23:08 | <Lachy> | TabAtkins, I thought you were mostly familiar with those discussions, and understood the reasons why selectors api, and scoped stylesheets, work the way they do. |
| 23:08 | <AryehGregor> | jgraham, I've seen it at least once before, yeah. |
| 23:08 | <AryehGregor> | At least this feature looks well-designed. |
| 23:09 | <TabAtkins> | Lachy: If I was, I've forgotten. ;_; |
| 23:09 | <TabAtkins> | All I recall is being against it for a long time. I don't remember what my initial reaction was during the design phase, but afterwards I think I've been consistent. |
| 23:11 | <Lachy> | well, it seems to be an issue that comes up frequently, so I guess I should write up a long and thourough explanation for why they must work the way they are currently defined, and why every alternative suggest has flaws. |
| 23:11 | <TabAtkins> | That sounds like a good idea. ^_^ |
| 23:12 | <Lachy> | I don't have time right now. It's midnight. Maybe I'll do it on Sunday, while recovering from this weekend's summer parties. |
| 23:15 | <hober> | Hixie: we'd appreciate an expedited look at http://www.w3.org/Bugs/Public/show_bug.cgi?id=12974 (which I've just filed) if possible. |
| 23:18 | AryehGregor | somehow never connected hober with Edward O'Conner the Apple employee |
| 23:19 | <Hixie> | hober: looking |
| 23:19 | <hober> | AryehGregor: that's me! :) |
| 23:20 | <hober> | Hixie: thanks. this is a really small change, though at this point i feel like every very small change will get a revert request on the other side. grumble. |
| 23:20 | <AryehGregor> | hober, if more implementers objected to the revert policy, maybe that would be helpful. |
| 23:20 | <Hixie> | hm, for type=number placeholder would actually make sense |
| 23:20 | <gsnedders> | hober: You haven't worked there for long, have you? |
| 23:20 | <Hixie> | i wonder why i didn't include it before |
| 23:20 | <hober> | Hixie: I checked; it didn't get explicitly removed, it was just not added when placeholder="" got initially added |
| 23:21 | <TabAtkins> | Oh god, I actually wrote "are" for "our" in an email. >_< |
| 23:21 | <hober> | gsnedders: yeah, just since february |
| 23:21 | <AryehGregor> | I feel like I'm the only one really objecting to the idea that anyone can randomly decide to get an uncontroversial feature addition reverted. |
| 23:21 | <Hixie> | hober: yeah, i wonder why. i looked at all the types when adding it and tried to only add it where it maeks sense (e.g. not color or date) |
| 23:21 | <AryehGregor> | Well, we'll have to see how various pending revert requests are handled, I guess. |
| 23:21 | <Lachy> | TabAtkins, I'm pretty sure Dmitry's proposal was considered and rejected before. I can't remember why though. I'll have to spend some time scouring the archives. |
| 23:21 | <AryehGregor> | TabAtkins, that kind of thing just proves that language is really auditory, and writing is a hack. |
| 23:21 | <TabAtkins> | AryehGregor: I think a lot of us just realize that arguing the issue is useless. |
| 23:22 | <AryehGregor> | TabAtkins, I don't think it is. The chairs are supposed to be acting based on consensus, and if they face strong opposition to something they're likely to at least try to compromise. |
| 23:22 | <hober> | Hixie: I think placeholder="" makes sense for <input type=number> even when <input type=number> is rendered as a spinbutton that takes up most of the control |
| 23:22 | <Hixie> | hober: yeah, that's what i'm saying. it makes sense, so why didn't i add it before? :-) |
| 23:22 | <hober> | Hixie: the UA can always decide to punt on rendering the placeholder when it can't be done sensibly |
| 23:22 | <Hixie> | crap, doing this means i have to split another column in the summary table |
| 23:22 | <Hixie> | hate doign that |
| 23:23 | <hober> | Hixie: yeah, no idea. :) sorry about the extra work |
| 23:23 | <Hixie> | oh no worries |
| 23:23 | <TabAtkins> | AryehGregor: When consensus means two people, and two people arguing against doesn't form counter-consensus, I don't think you're accurately describing the way the chairs work. |
| 23:23 | <AryehGregor> | TabAtkins, it wasn't widely discussed at all. |
| 23:23 | <TabAtkins> | I mean something like the change in canvas content from a bit ago. |
| 23:24 | <AryehGregor> | No, I'm saying that the revert policy itself is supposed to be subject to consensus. |
| 23:24 | <AryehGregor> | If we get strong objections from a bunch of people, particularly implementers, against the policy itself, that might be effective. |
| 23:24 | <jgraham> | AryehGregor: I get the impression that most people don't want to get dragged into policy discussions |
| 23:25 | <TabAtkins> | Where I was told that revert requests are trivial to get issued, but counter arguments should be in the form of a bug, which is then turned into an issue... |
| 23:25 | <TabAtkins> | Oh, yeah, I don't even know where the revert policy itself came from. |
| 23:25 | <AryehGregor> | jgraham, except the people who have nothing better to do with their lives, which does not include implementers or other parties with a real stake. |
| 23:25 | <TabAtkins> | I don't think it's in The Process. |
| 23:26 | <jgraham> | AryehGregor: Yes, and history suggests that those people will always whine more than people with real work to do |
| 23:26 | <jgraham> | After a while it gets tiresome |
| 23:26 | <AryehGregor> | jgraham, so the obvious conclusion is that the implementers should officially tell the HTMLWG that if it doesn't make its process sane, they'll all just use the WHATWG copy of the spec exclusively and ignore the W3C copy. |
| 23:26 | <jgraham> | So mostly people just pretend that HTMLWG doesn't exist except for the patent policy |
| 23:27 | <AryehGregor> | Better yet, have a major implementer threaten to actually leave the group. That's not going to be ignored. |
| 23:27 | <jgraham> | I think getting an official position on that would be hard |
| 23:27 | <AryehGregor> | But no one seems to want to. |
| 23:27 | <jgraham> | Because of a) Microsoft and b) patent policy |
| 23:28 | <AryehGregor> | You could push through changes to make things saner without actually leaving the group. |
| 23:28 | <Hixie> | http://www.whatwg.org/specs/web-apps/current-work/complete.html#input-type-attr-summary has become quite crazily complicated |
| 23:28 | <AryehGregor> | But implementers aren't pulling their weight in the HTMLWG. |
| 23:28 | <Hixie> | that table is one of the first tables i did, back in the WF2 days |
| 23:28 | <jgraham> | But if the W3C were trying to actively make things better they would notice all the people recommending reading the WHATWG version of the spec and treat it as a problem at the W3C end |
| 23:28 | <Hixie> | someone should make some sort of video showing that table evolve over time |
| 23:29 | <jgraham> | Not as something that can be ignored or, depending on the individual, a conspiracy to disenfranchise developers/users with disabilities/non-browser vendors/other |
| 23:30 | <AryehGregor> | jgraham, nobody is trying to do anything at the W3C. Various individuals and groups are trying to impose their opinions on the spec, and the chairs are trying to juggle them so as to generate as little overall dissatisfaction as possible within the Process. |
| 23:30 | <AryehGregor> | The result winds up being largely incoherent. |
| 23:30 | <AryehGregor> | Design by committee. |
| 23:30 | <jgraham> | AryehGregor: I'm not entirely sure I believe that |
| 23:31 | <AryehGregor> | Not much of what the HTMLWG does or decides is actually controlled by the W3C administration. |
| 23:31 | <TabAtkins> | I'd believe that, if you rate "dissatisfaction" as "amount of public complaining on the list". |
| 23:31 | <jgraham> | I mean that W3C Process is enirely blameless here and that it is all the WG |
| 23:31 | <gsnedders> | AryehGregor: Arguably that's a problem of being it open — when it was just Members with an interest in the spec, there was far less of that. Sure, it still happened, but not like today. |
| 23:31 | <AryehGregor> | The Process is part of the problem, but it's not something I'd ascribe motives to. |
| 23:32 | <David_Bradbury> | Is there a way I can see ideas that have been previously proposed? |
| 23:32 | <AryehGregor> | David_Bradbury, read the entire archives of the whatwg list and the bug tracker. I'm fairly sure it would take less than a year if you did it full-time. |
| 23:32 | <David_Bradbury> | Haha |
| 23:32 | <jgraham> | The Process is just the manifestation of W3C goals and priorities |
| 23:32 | <TabAtkins> | David_Bradbury: What Aryeh said. He's not kidding. That's the record, and it's far too large to curate effectively anyway. |
| 23:33 | <AryehGregor> | gsnedders, more precisely, it's a problem with non-implementers having much power. Even in Member-only land, you can see lots of crazy stuff when it's random uninvolved organizations making the decisions. |
| 23:33 | <David_Bradbury> | Would you happen to know if 3D transformations for the 2D Canvas Context have been suggested? |
| 23:33 | <jgraham> | If somthing is in The Process it's because people at the W3C wanted it there |
| 23:33 | <AryehGregor> | Like look at the AC decisions on the HTML5 license. |
| 23:33 | <Hixie> | David_Bradbury: they've been floated |
| 23:34 | <Hixie> | David_Bradbury: there's also a 3d api for canvas, webgl |
| 23:34 | <gsnedders> | AryehGregor: It would set a precedent for the whole W3C — it would affect every member, and the entire W3C. |
| 23:34 | <AryehGregor> | gsnedders, yes. And? |
| 23:34 | <gsnedders> | AryehGregor: That gives them a vetted interest in it. |
| 23:34 | <David_Bradbury> | Hixie: Ah. My main concern for webgl is I mainly code for mobile devices right now, and that won't be supported for a long while |
| 23:34 | <AryehGregor> | What gives them a vested interest in it is solely that they're paying membership dues. |
| 23:34 | <zewt> | anything proposed today won't be supported for a long while anyway :) |
| 23:34 | <AryehGregor> | They have no direct interest in the *web*. |
| 23:35 | <AryehGregor> | Or if they do, they don't have the expertise to effectively advance it. |
| 23:35 | <Hixie> | David_Bradbury: i would expect webgl in browsers on mobile before 3d transforms in 2d canvas :-) |
| 23:35 | <AryehGregor> | The problem with member-only things is that the W3C tries to get everyone to be a member so it can get all the membership dues, so you wind up having decisions of things like the AC based on a bunch of random tech organizations with no connection to the web. |
| 23:35 | <jgraham> | AryehGregor: (A good example of The Process causing weird effects is all the artifical deadlines that the chairs keep imposing) |
| 23:35 | <AryehGregor> | jgraham, I never said the Process wasn't a problem, it's a large part of the problem. |
| 23:35 | <David_Bradbury> | Hixie: Fair enough :p I'll likely just have to simulate 3D transformations until WebGL goes mobile |
| 23:36 | <AryehGregor> | But the broader problem is that things are decided by the wrong parties. |
| 23:36 | <AryehGregor> | I don't really care if authoring requirements are left to popular vote or whatever, but giving anyone but implementers direct say over implementation requirements is just stupid. |
| 23:37 | <AryehGregor> | That includes Public Invited Experts and it includes non-implementer Members. |
| 23:37 | <AryehGregor> | But nothing's going to happen to solve it unless implementers are willing to put their foot down, which it currently seems they aren't. |
| 23:38 | <David_Bradbury> | http://hacks.mozilla.org/2009/12/webgl-goes-mobile/ > *drool* |
| 23:38 | <AryehGregor> | You won't ever really solve the decision-making problem unless you solve the funding problem, though. |
| 23:39 | <AryehGregor> | Which is why I think salvaging the W3C is hopeless. To make appropriately implementer-friendly decisions, implementers need to be in sole control, which means they need to bankroll it. |
| 23:40 | <AryehGregor> | Which is tenable if we had a low-bureaucracy group with no process or administration. |
| 23:40 | <TabAtkins> | Which, surprise surprise, we do. |
| 23:41 | <Hixie> | still need a patent policy |
| 23:41 | <AryehGregor> | Except that the WHATWG doesn't really host anything but one spec, and that isn't set up to change anytime soon. |
| 23:41 | <AryehGregor> | Yeah, and the patent policy. |
| 23:41 | <Hixie> | we could easily host more specs |
| 23:41 | <AryehGregor> | The patent policy is an issue if implementers are worried about non-implementers suing them, because how will you get non-implementers to agree to the patent policy if they don't get anything in return? |
| 23:41 | <Hixie> | in fact yesterday we hosted one more than we do today :-) |
| 23:42 | <Hixie> | AryehGregor: that problem isn't solved in w3c yet either. First solve the problem the W3C solves, then worry about making it better. |
| 23:42 | <AryehGregor> | Don't let people sign up to the mailing list unless their employer agrees to the patent policy? |
| 23:42 | jgraham | -> bed |
| 23:42 | <AryehGregor> | Hixie, HTML5 has patent protection from every organization that's a member of the HTMLWG. |
| 23:43 | <Hixie> | AryehGregor: how many of them own patents? |
| 23:43 | <Hixie> | jgraham: nn |
| 23:43 | <AryehGregor> | It includes companies like Adobe, IBM, Intel, Nokia, Samsung. |
| 23:43 | <AryehGregor> | That's a heck of a lot of patents between them. |
| 23:44 | <Hixie> | sure, those are all implementors |
| 23:44 | <AryehGregor> | Of HTML? |
| 23:44 | <Hixie> | yup |
| 23:44 | <Hixie> | well, maybe not intel |
| 23:44 | <AryehGregor> | Well, Nokia and Samsung probably contribute to WebKit, I guess. |
| 23:44 | <AryehGregor> | What do Adobe and IBM do? |
| 23:44 | <Hixie> | adobe does a ton of tools, and ibm has fingers everywhere |
| 23:45 | <AryehGregor> | But realistically, none of them contribute enough to web stuff to have much say over WHATWG specs. |
| 23:46 | <AryehGregor> | Actually, even if they did -- why should they agree to the patent policy? What are they getting in return? |
| 23:46 | <AryehGregor> | They get the influence anyway, if they're big enough and you want to match reality. |
| 23:46 | <Hixie> | they don't have much say over the w3c spec either, in practice, except for ibm (who has a chair who happens to not recuse himself when it comes to deciding on ibm proposals) |
| 23:46 | <TabAtkins> | AryehGregor: Adobe contributes to webkit. |
| 23:46 | <AryehGregor> | Hixie, they aren't allowed to even sign up to the mailing list unless they agree to the patent policy. |
| 23:47 | <Hixie> | AryehGregor: they presumably want apple not to sue them on html stuff, just like apple don't want them to sue them :-) |
| 23:47 | <Hixie> | (and s/apple/any vendor/) |
| 23:47 | <Hixie> | (but apple happens to have history with several of those you listed) |
| 23:47 | <AryehGregor> | Hixie, the W3C patent policy gives a license to everyone, not just others who agree to it. |
| 23:47 | <AryehGregor> | Presumably a WHATWG patent policy would do the same. |
| 23:47 | <AryehGregor> | So they aren't gaining any immunity. |
| 23:47 | <Hixie> | AryehGregor: maybe making it reciprocal is the solution then |
| 23:48 | <Hixie> | AryehGregor: right now they're not getting any protection at all since the spec isn't in REC |
| 23:48 | <AryehGregor> | Anyway, back to the other point: if you want more specs hosted at the WHATWG, maybe you should make it clear that that's the case, and post instructions for submitting a new spec. |
| 23:48 | <Hixie> | i have no desires one way or the other |
| 23:48 | <Hixie> | i'm just saying it would be easy to set up |
| 23:48 | <AryehGregor> | There are various specs edited by me and Ms2ger and whatever at random places like aryeh.name and html5.org that could be better situated at whatwg.org. |
| 23:48 | <AryehGregor> | The more standards there, the more credible it becomes. |
| 23:48 | <Hixie> | (just like we've set up wikis, blogs, etc) |
| 23:49 | <AryehGregor> | A reciprocal patent policy is a very interesting idea. |
| 23:49 | <AryehGregor> | It could work quite well, if anyone can easily sign up to it for free even after someone else has already tried to sue them and still get the immunity. |
| 23:50 | <Hixie> | if you would like to host the specs you work on on the whatwg.org domain, send me a mail with the current urls of the specs in question and i'll see what i can do |
| 23:50 | <AryehGregor> | In practice it would be the same as a universal license, since anyone could get it, except they have to agree to it as well. Copyleft-style. |
| 23:50 | Hixie | isn't a patent lawyer, so really isn't in a good place to have an educated opinion on the topic |
| 23:51 | <AryehGregor> | I don't particularly want to host at the WHATWG. I'd do it if we were trying to make it a credible competitor to the W3C, but it seems we aren't doing that anyway. |
| 23:51 | <Hixie> | personally i certainly have no interest in competing with the w3c |
| 23:51 | <Hixie> | my interest lies purely in making the web better |
| 23:52 | <Hixie> | whether that involves the w3c or not is not really important to me |
| 23:52 | <Hixie> | if people want to host specs on the whatwg.org domain, i'm happy to look into setting that up; if people want to set up a whatwg.org patent policy, i'm happy to contribute whatever resources i can bring to bear on the topic |
| 23:53 | <Hixie> | not particularly interested in spearheading the latter effort though :-) |
| 23:55 | <AryehGregor> | The W3C seems to me an incorrigible obstruction to making the web better. |
| 23:56 | <boogyman> | s/W3C/Internet Explorer/ |
| 23:56 | <AryehGregor> | It just makes its obstruction small enough that it's not worth the effort to abandon it, every time there's any threat to it. |
| 23:56 | <AryehGregor> | Make the minimal concessions to get implementers to use it. |
| 23:56 | <AryehGregor> | Oh well, nothing I can do. |
| 23:56 | <AryehGregor> | boogyman, IE9 is actually quite good. |
| 23:58 | <boogyman> | Yeah, but it will be nearly a decade before it's wide use... eg, when people who bought "vista" get away from the OS. But yes, I'm very happy about the strides the MSIE have made in developing v9 |
| 23:58 | <boogyman> | widely* |
| 23:58 | <Hixie> | competition in the browser space does seem to have done wonders to the IE situation |