| 00:47 | <AryehGregor> | It's amazing how much smoother to use Gmail is on my tiny Nexus One than on my desktop. |
| 00:47 | <AryehGregor> | I wonder: is this some deficiency in the web platform, or just Gmail not caching things locally aggressively enough? |
| 00:48 | AryehGregor | bets the latter, wonders why |
| 00:50 | <AryehGregor> | Local Storage size on disk for mail.google.com: 3 KB for http, 3 KB for https. :/ |
| 01:32 | <karlcow> | hmm the discussions about editors drafts, working drafts, etc. is argueing if it's bad or not |
| 01:32 | <karlcow> | INSTEAD of trying to list the issues/needs and how we solve them. |
| 01:34 | <Dashiva> | Issue: working drafts are always out of date |
| 01:34 | <karlcow> | Dashiva: I would add a "for [here an audience]" |
| 01:34 | <karlcow> | for example in your sentence. |
| 01:35 | <AryehGregor> | Why would anyone want to use an out-of-date draft? |
| 01:35 | <AryehGregor> | For any purpose? |
| 01:35 | <AryehGregor> | It's obsolete. |
| 01:35 | <karlcow> | Issue: last drafts of a specification is useful for some implementers. |
| 01:35 | <Hixie> | the only audience for which out of date drafts is useful is spec archeologists, like when henri and i try to work out why something was done |
| 01:35 | <Hixie> | but with HTML5, the whatwg svn repo is far more useful for that now |
| 01:36 | <karlcow> | How do we solve it? What is needed in the document, etc. Editor drafts could be retitled Developers Edition |
| 01:36 | <Dashiva> | Resolution: Stop issuing working drafts |
| 01:36 | <Dashiva> | Editor's draft covers all the use cases |
| 01:37 | <karlcow> | AryehGregor: People who need stability: Book writers, documentation writers, etc. A lot easier to update your work big by big step than because on a moving target with micro-edits |
| 01:37 | <Hixie> | Dashiva: that's what i did at the whatwg when i realised this problem existed |
| 01:37 | <Dashiva> | Should we really be encouraging people to write outdated documentation? |
| 01:37 | <AryehGregor> | karlcow, why would you want to write a book based on a snapshot that's months out of date? If you really want that, you can pick a specific revision from SVN and only use that. |
| 01:38 | <AryehGregor> | At least that will be up-to-date as of the time you start writing. |
| 01:38 | <AryehGregor> | But I'll stick to stuff written by people who actually follow spec changes, for my part. |
| 01:38 | <karlcow> | AryehGregor: because not everyone is a geek aka pick up a version in SVN |
| 01:38 | <AryehGregor> | Then why the heck are they writing documentation for a spec? |
| 01:38 | <Dashiva> | People who write books and documentation should be competent enough to do so |
| 01:39 | <karlcow> | Dashiva: why do you hate freedom? ;) |
| 01:39 | <Hixie> | nobody's stopping someone from taking a snapshot of the PDF or HTML versions of the spec |
| 01:39 | <Dashiva> | I don't hate freedom, I hate freedumb |
| 01:39 | <Hixie> | doesn't mean we need to make explicit copies |
| 01:40 | AryehGregor | just wrote a 1377-word blog comment about HTML5 |
| 01:40 | <karlcow> | It is usually wiser for the peace of each community to give tools that these communities need. (if they need it :) ) |
| 01:40 | <othermaciej> | AryehGregor: 40 words too many! |
| 01:40 | <othermaciej> | which blog? |
| 01:40 | <AryehGregor> | othermaciej, <http://www.cssquirrel.com/2010/06/15/693comic-update-html5-unicorn-heuristics/comment-page-1/#comment-32000>;, but it has to pass moderation because of the links I put in for references. |
| 01:40 | <Dashiva> | (People are free to be dumb, but we are not obligated to help them) |
| 01:41 | <othermaciej> | there's been some interesting discussion in those comments |
| 01:41 | <AryehGregor> | karlcow, so people who actually want a snapshot can just save the page to their hard disk. What's the problem with that? |
| 01:41 | <AryehGregor> | WD would make sense if there were some meaningful procedure for approving it, so it were actually more likely to be correct than an ED, but that's not the case. |
| 01:41 | <AryehGregor> | othermaciej, I have hopes of converting Kyle Weems to the Light Side. |
| 01:42 | <karlcow> | AryehGregor: I have no issue with people saving a copy, they can do it already :) |
| 01:42 | <AryehGregor> | I think he's just misinformed. |
| 01:43 | karlcow | wishes that people could accept diversity in opinions ;) |
| 01:43 | <Dashiva> | Diversity in opinions is fine |
| 01:43 | <Dashiva> | Diversity in facts is not |
| 01:43 | <karlcow> | religious statement |
| 01:44 | <karlcow> | too far for me |
| 01:44 | <karlcow> | :) |
| 01:44 | <othermaciej> | AryehGregor: I think your messages in those comments come across as fairly rational and balanced |
| 01:44 | <Dashiva> | I should probably cite properly |
| 01:44 | <AryehGregor> | othermaciej, hope so! |
| 01:44 | <othermaciej> | AryehGregor: even though you are kind of questioning the quality of my service as chair :-p |
| 01:44 | <AryehGregor> | The last sentence of my new one is "If the W3C is unwilling to accept that, it will become a victim of natural selection, like it or not", but I think I supported it fairly well. |
| 01:45 | <Dashiva> | "You’re entitled to your own opinions. You’re not entitled to your own facts." - Daniel Patrick Moynihan |
| 01:45 | <AryehGregor> | othermaciej, don't worry, I'm only comparing you unfavorably to Hixie. That's not an insult in absolute terms. |
| 01:46 | <AryehGregor> | Dashiva, can I be entitled to my own facts if I write a suitably advanced virtual reality program that implements them? |
| 01:48 | <karlcow> | AryehGregor: that is a lot more reasonable and basically what every society does. |
| 01:48 | <AryehGregor> | What, writing suitably advanced virtual reality programs? |
| 01:48 | <Dashiva> | No, you have to pass a Trial of Philosophy to justify why a virtual reality should be considered factual first |
| 01:49 | <Dashiva> | How come these discussions always start hours past midnight |
| 01:49 | <Dashiva> | Time to sleep |
| 01:49 | <karlcow> | hehe |
| 01:49 | <AryehGregor> | Because you live in the wrong time zone? |
| 01:49 | <karlcow> | Dashiva: because you are living in your own reality |
| 01:50 | <Dashiva> | CET is the best time zone ever |
| 01:51 | <AryehGregor> | Dashiva, reality is indistinguishable from a sufficiently good illusion. Ergo, there is no difference according to any empirical worldview, and empirical worldviews are the only ones that count because all others are stupid. QED. |
| 01:51 | <Dashiva> | Then you have to prove that your virtual reality is sufficiently good |
| 01:51 | <Dashiva> | A sort of a turing test, if you will |
| 01:51 | <AryehGregor> | That was postulated. |
| 01:51 | <AryehGregor> | "suitably advanced" |
| 01:52 | <karlcow> | http://en.wikipedia.org/wiki/Allegory_of_the_Cave |
| 01:52 | <Dashiva> | Besides, people outside would be able to notice those on the inside, and thereby be able to reject their statements about the warped reality |
| 01:53 | <AryehGregor> | Dashiva, yes, but that's exactly what we were talking about: everyone gets their own version of the facts. |
| 01:53 | <Dashiva> | No, some people get a warped reality and other people get facts |
| 01:54 | <Dashiva> | Surely you do not consider the worldview of a lunatic to be factual? |
| 01:54 | <Dashiva> | "There will have to be a couple of implementations before CR of HTML5, anyway." |
| 01:55 | <Dashiva> | It's like nobody even cares what the process really says anymore |
| 01:55 | <AryehGregor> | Dashiva, but if everyone uses VR adjusted to taste, then everyone's reality will be equally accurate. |
| 01:56 | <Dashiva> | You postulate a world in which a) everyone is in a virtual reality and b) nobody realizes they are in a virtual reality |
| 01:56 | <AryehGregor> | Sure. Maybe we're in one already. Can't say. |
| 01:56 | <AryehGregor> | Just as good as actual reality. |
| 01:56 | <karlcow> | hehe |
| 01:56 | <Dashiva> | Not quite |
| 01:56 | <Dashiva> | Even if we cannot say for sure if our reality is indeed reality, we can say for sure that virtual realities inside our own reality are not reality |
| 01:57 | <karlcow> | Dashiva: and it is not an issue |
| 01:57 | <AryehGregor> | Also, a lunatic's worldview is indistinguishable from a factual worldview, from the lunatic's perspective. Other people think the opposite, but in principle, that can be a perfectly symmetric disagreement, and we only win by majority vote. |
| 01:57 | <karlcow> | as long as you have developed a coherent system for you and maybe your social groups. |
| 01:57 | <AryehGregor> | We don't know that virtual realities inside our own reality are not reality if we make ourselves all forget. |
| 01:58 | <karlcow> | Physics (science) is all about *creating* a reality. |
| 01:58 | <AryehGregor> | I'm speaking from the viewpoint of us after we all go into the VR and forget everything. |
| 01:58 | <Dashiva> | If everyone in reality forgets about reality in favor of a virtual reality, then the reality is no longer part of their conceptual universe |
| 01:58 | <AryehGregor> | In that situation, it would be fair to say that everyone gets to have their own version of reality. |
| 01:59 | <AryehGregor> | Right, exactly. |
| 01:59 | <Hixie> | as a physicist by training, i have to take objection ot the idea that physics _creates_ reality |
| 01:59 | <AryehGregor> | Me too, as a physics minor! |
| 01:59 | <AryehGregor> | Although it would be cool if it did. |
| 01:59 | <Hixie> | unless you have a very odd definition of "reality" |
| 01:59 | <AryehGregor> | I always play wizards in fantasy games. :( |
| 01:59 | <Dashiva> | That's not very physicsy of you |
| 02:00 | <karlcow> | http://www.w3.org/2010/09/web-on-tv/ |
| 02:00 | <AryehGregor> | No, it's perfectly physicsy. I'm just utilizing the laws of physics that are present in a hypothetical alternate reality where they allow me to, say, cause you to spontaneously catch fire when I get annoyed at you. |
| 02:00 | <karlcow> | :D |
| 02:00 | <Dashiva> | Also known as wish fulfillment |
| 02:01 | <AryehGregor> | Exactly! |
| 02:01 | <karlcow> | Dashiva: and you are still not in bed. |
| 02:01 | <Dashiva> | I actually have both feet in bed |
| 02:01 | <karlcow> | TMI |
| 02:01 | <Dashiva> | The bed is my foot stool |
| 02:02 | <Dashiva> | The bookshelf is my monitor stand |
| 02:07 | <AryehGregor> | Awesome, this is how web PDF support should have always been done: http://blog.chromium.org/2010/06/bringing-improved-pdf-support-to-google.html |
| 02:10 | <karlcow> | hmmm, if I can still download the pdf locally or save the generated html on my computer I'm ok. |
| 02:10 | <karlcow> | Browser is not the best tool for me to read a pdf. |
| 02:10 | <karlcow> | I have seen this today too http://simonwillison.net/2010/Jun/17/pdfjs/ |
| 02:11 | <AryehGregor> | I'd way prefer to get a good PDF reader implementation in a browser than to have to use external programs. |
| 02:11 | <karlcow> | AryehGregor: and me the opposite. every taste in nature ;) |
| 02:12 | <AryehGregor> | Anyway, things I was supposed to do tonight: file health insurance claim, buy several things, study some religious texts, pack for trip, go to bed early. Things I actually did: chat on IRC. |
| 02:12 | <AryehGregor> | Sigh. |
| 02:12 | <AryehGregor> | Good night. |
| 02:12 | <roc> | roc |
| 02:13 | <AryehGregor> | o_O |
| 02:14 | <karlcow> | AryehGregor: nite nite |
| 02:16 | <roc> | oops |
| 02:22 | <karlcow> | http://lists.w3.org/Archives/Public/public-web-perf/ Web Browser Performance mailing List |
| 03:59 | <abarth> | Hixie: no complaints from day 1 of the HTML5 tokenizer in WebKit nightly |
| 04:00 | <abarth> | Hixie: that's pretty good :) |
| 04:13 | <roc> | karlcow: what is that mailing list actually for? |
| 04:23 | <karlcow> | roc: I guess what it says, but we might see soon. |
| 04:23 | <roc> | it doesn't say anything |
| 04:24 | <roc> | why do we need a mailing list to discuss browser performance? |
| 04:25 | <karlcow> | roc: I think you should ask plh. I found out through twitter |
| 04:25 | <roc> | oh, I thought you W3C people had some sort of cabal |
| 04:25 | karlcow | notes to roc that he is not working at W3C |
| 04:25 | <roc> | hum |
| 04:26 | <roc> | oh well |
| 04:27 | <karlcow> | roc: http://twitter.com/w3c/statuses/16421966743 |
| 04:27 | <karlcow> | and the page says "Discussion related to Web browser performance " |
| 04:41 | <MikeSmith> | I forget -- is there a way to express negation with CSS selectors? |
| 04:41 | <MikeSmith> | e.g., to express "the first child that is not [something]"? |
| 04:55 | <MikeSmith> | roc, karlcow - that list is essentially just for discussion of the Web Timing spec |
| 04:55 | <MikeSmith> | at this point |
| 04:56 | <karlcow> | ah |
| 04:56 | <karlcow> | MikeSmith: thanks. We always need a Mike in one's pocket… hmmm did I write that. I guess so ;) |
| 04:58 | <MikeSmith> | http://dev.w3.org/2006/webapi/WebTiming/ |
| 05:01 | <MikeSmith> | riding on a bus from Tokyo to Nagano is not the best way to get stable Internet connectivity |
| 05:01 | <wirepair> | i love nagano |
| 05:02 | <MikeSmith> | karlcow: btw, have you been following the current problems with the W3C markup validator |
| 05:02 | <MikeSmith> | wirepair: I never been there |
| 05:02 | <MikeSmith> | this will be my first visit |
| 05:02 | <wirepair> | ah nice, it's very very clean air |
| 05:02 | <MikeSmith> | to Matsumoto, actually |
| 05:02 | <wirepair> | ahh... |
| 05:02 | <wirepair> | i was out in the woods hehe |
| 05:02 | <karlcow> | MikeSmith: about LibXML? I agree with your suggestion |
| 05:05 | <karlcow> | checking the page served as text/html with an html validator. The issue is to give the possibility for people who are working with local document. Upload of an XHTML document. You do not know how it will be served. So in the UI it has to be super clear what is happening |
| 05:16 | <MikeSmith> | karlcow: I suppose I would not object to having it do XML WF-ness checking on file uploads by default |
| 05:17 | karlcow | replied |
| 05:17 | <karlcow> | MikeSmith: have you seen Iwate's Matsuo. Quite impressive. http://www.michaeljohngrist.com/2009/08/the-mist-wreathed-apartments-of-matsuo-mine/ |
| 05:17 | <karlcow> | never seen it myself |
| 05:17 | <MikeSmith> | but I really don't think it's productive for it to be doing XML WF checking on remote docs served as text/html |
| 05:18 | <MikeSmith> | and iirc the performance problem is for checking of remote docs |
| 05:18 | <karlcow> | yep agreed |
| 05:20 | <MikeSmith> | I seem to remembers somebody commenting on the validator list that they had no problems if they checked files by upload instead of by HTTP |
| 05:20 | <MikeSmith> | have not seen |
| 05:20 | <MikeSmith> | wow, looks great |
| 05:21 | <MikeSmith> | beautiful decay |
| 05:21 | <karlcow> | yep |
| 05:21 | <MikeSmith> | "dilapidation's progresses are organized decays" |
| 05:57 | <hsivonen> | what's the web on tv workshop trying too accomplish? |
| 05:57 | <othermaciej> | hello all |
| 05:57 | <hsivonen> | hi |
| 05:58 | <hsivonen> | s/too/to/ |
| 05:58 | <othermaciej> | there's a "web on tv" workshop? |
| 05:59 | <hsivonen> | according to the scrollback |
| 06:01 | <hsivonen> | what pdf implementation does chrome use? |
| 06:02 | <hsivonen> | and will it be in chromium, too? |
| 06:04 | <hsivonen> | I wish Gecko came with Poppler built-in, but probably not gonna happen due to licensing |
| 06:12 | <hsivonen> | AryehGregor: when you say "Firefox currently", do you mean 3.6 and not the trunk? |
| 06:12 | <hsivonen> | my notion of current is the trunk |
| 08:28 | <annevk> | sigh |
| 08:28 | <annevk> | PFWG persists in their thinking that having three value syntaxes for WAI-ARIA is better than one |
| 08:29 | <annevk> | that seems like something an overarching body like the TAG could tackle |
| 08:29 | <annevk> | instead of all this polyglot bullshit |
| 08:30 | <annevk> | (and with three I mean potentially infinite) |
| 09:16 | <Dashiva> | http://twitter.com/simonstl/status/16439272227 |
| 09:19 | <annevk> | yeah, and some part doesn't work in Chrome, and some not in Opera, and some not in Firefox... ooh |
| 09:19 | <Dashiva> | And some parts do work in IE6 :P |
| 09:20 | <annevk> | heh |
| 09:24 | jgraham | is disturbed that he seems to have had a dream about breaking and entering Hixie's house |
| 09:25 | <jgraham> | Although I think I was looking for him at the time rather than wanting to steal all his stuff |
| 09:26 | <jgraham> | And I think (although this part is really hazy) I eventually found him but he looked a lot older than in real life |
| 09:36 | <annevk> | sounds like vacation time :) |
| 09:38 | <annevk> | also that the PFWG puts DAISY on equal footing with HTML and SVG ... oh well |
| 09:39 | <annevk> | maybe we should just define the "host blah blah" mapping in HTML5 and use that everywhere and fix the specification to match reality later |
| 09:39 | <annevk> | that's probably the easier way to fix this |
| 09:43 | <hsivonen> | annevk: I think the PF really should focus on Web languages |
| 09:45 | <annevk> | meanwhile the WOFF people have not read the CORS specification which clearly says the API (@font-face in their case) is in charge of the request semantics whereas they make it dependent on the format |
| 09:46 | <annevk> | it is becoming quite clear that the reason the web platform is so inconsistent is because nobody really knows how it works but is extending it in their own vision anyway without regards to precedence and design patterns that came before |
| 09:47 | <annevk> | and in this case not even with regards to things that came before, as CORS is being developed pretty much in line with @font-face / WOFF and is pretty clear on how CORS API specifications ought to work and all |
| 09:50 | <annevk> | (not that I think it is at all justified that something like @font-face would use CORS, but that is a different matter) |
| 10:06 | <hsivonen> | annevk: indeed. |
| 10:07 | <hsivonen> | annevk: There's some old stuff that came about by Netscape Netcenter or Bugzilla devs just asking for something and Gecko developers simply said "OK". |
| 10:08 | hsivonen | wonders if Apple and Opera are taking action on fingerprinting. (see https://panopticlick.eff.org/ ) |
| 10:10 | <abarth> | hsivonen: that's a hard problem |
| 10:12 | <hsivonen> | abarth: I think there are easier parts and harder parts. |
| 10:12 | <hsivonen> | there are 4 zones: HTTP request, CSS, JS and plug-ins |
| 10:12 | <abarth> | does it help to do some but not all of them? |
| 10:13 | <hsivonen> | abarth: I think it does |
| 10:13 | <jgraham> | presumably that depends on how much information you need to make reliable guesses about identity |
| 10:13 | <hsivonen> | abarth: if you do HTTP, you force fingerprinters to expose evidence of their activities to the client |
| 10:13 | <abarth> | ic |
| 10:14 | <hsivonen> | abarth: which means that big players with the widest ability to track people across the Web would be exposed to regulatory solutions |
| 10:14 | <abarth> | i have a link to the Fingerprint moz wiki page |
| 10:14 | <abarth> | but i haven't read it yet |
| 10:15 | <hsivonen> | my wishful thinking is that the plug-in part can be mitigated by treating everything but Flash as obsolete and by having .swf as click-to-play by default |
| 10:15 | <hsivonen> | but I realize this is wishful thinking |
| 10:16 | <hsivonen> | also, I have a naive Mac/Linux user POV |
| 10:17 | <hsivonen> | I'm saddened to see what kind of NPAPI cruft gets installed on Windows machines |
| 10:22 | <hsivonen> | whoa! after a couple of days of having Windows 7 installed and only installing other browsers, security patches, virus scanner and trying to install XP Mode but failing, IE8 running in full screen is already unique |
| 10:23 | <hsivonen> | maybe that's because I haven't installed Flash or Java, and not having those is rare |
| 10:23 | <jochen__> | hsivonen: i heard you think about reducing fingerprinting possibilities with http? |
| 10:23 | <jochen__> | hsivonen: any plan what to do there without making half of the webpages unusable? :) |
| 10:24 | <hsivonen> | jochen__: well, I filed all the obvious pieces of the HTTP part of the puzzle as bugs yesterday, but I'm not sure where it's going to go |
| 10:24 | <hsivonen> | jochen__: as far as I can tell, the only hard part about HTTP request is Accept-Language |
| 10:24 | <jochen__> | mind linking them from wiki.mozzila.org/Fingerprinting? |
| 10:25 | <hsivonen> | jochen__: the tracking bug is https://bugzilla.mozilla.org/show_bug.cgi?id=http-fingerprint |
| 10:25 | <jochen__> | ta |
| 10:25 | <hsivonen> | jochen__: I'll add it to teh wiki |
| 10:28 | <hsivonen> | oh yeah, OS/CPU is probably politically hard, even though is terms of actual Web compat you get far without them |
| 10:30 | <volkmar> | i see some css selectors are defined in html5 specs, does that mean a request/issues related to that should be expressed in whatwg ml? |
| 10:30 | <volkmar> | or i should use www-style |
| 10:33 | <Matjas> | tabindex=0 enables keyboard navigation for any element. However, I would expect the click event to fire for those elements when pressing Enter/Space while they’re focused. Any reason why that’s not the case? |
| 10:34 | <Matjas> | Enter/Space fires click for every element that’s keyboard-focusable by default |
| 10:36 | <annevk> | I argued that we should do that |
| 10:36 | <annevk> | and make onclick="" elements focusable |
| 10:36 | <annevk> | but not enough browsers did it to begin with so they would first have to be changed as the reasoning, I believe |
| 10:37 | <hsivonen> | annevk: Hixie's reasoning, PF reasoning or someone else's reasoning? |
| 10:43 | <annevk> | Hixie; he just wanted to match reality for now |
| 10:43 | <annevk> | (though Opera has the nicer behavior afaik) |
| 10:46 | <jgraham> | hsivonen: <b><button></b>. AFAICT that should reopen the <b> element inside the <button>, but it seems that gecko does not |
| 10:48 | <jgraham> | hsivonen: Am I missing something? |
| 10:57 | <Matjas> | annevk: Exactly. What Opera does feels like "the right thing" |
| 11:01 | <annevk> | hsivonen, interesting work on fingerprinting |
| 11:02 | <annevk> | hsivonen, I think I'll copy some of your bugs and file them on Opera; see what happens |
| 11:07 | <annevk> | nuking Accept-Charset makes a lot of sense |
| 11:17 | <hsivonen> | annevk: the sad part is that there's so much fear of making Firefox lead on this issue. The parts that are the easiest to argue are the ones where IE is already leading... Like Accept-Charset. |
| 11:18 | <hsivonen> | jgraham: the <button> situation sucks. I haven't properly though about <b><button></b>. |
| 11:18 | <hsivonen> | one option is to just revert it into scoping and let the German Figure Collector site break |
| 11:18 | <daedb> | woah, .Net framework adds itself to Firefox's ua string now too, not just IE :O |
| 11:20 | <hsivonen> | daedb: using what mechanism? Whatever ships with Windows 7 doesn't tamper with the UA string of Minefield |
| 11:20 | <hsivonen> | everyone wants to add their mark to the UA string :-( |
| 11:20 | <daedb> | I don't know, but I have "(.NET CLR 3.5.30729)" at the end of my Firefox ua string right now (and Firefox was installed long after .Net 3.5) |
| 11:21 | <hsivonen> | daedb: do you see a .Net-specific add-on? |
| 11:21 | <daedb> | The :net framework assistant, yes, but it's disabled. |
| 11:22 | <annevk> | hsivonen, the others are more tricky I guess, yes |
| 11:23 | <hsivonen> | speaking of HTTP request headers, Opera Mobile on Maemo sure has a long Accept header |
| 11:31 | <karlcow> | hsivonen: worse than everyone wants to add their mark, nobody wants to remove the old ones. the pile is going up, up, up. :) I wonder if there is a physical limit to the size of UA string (characters number) |
| 11:32 | <annevk> | we cleaned it up a little bit |
| 11:32 | <annevk> | but then it exploded again ffs |
| 11:33 | <hsivonen> | annevk: Opera has cruft in the various Accept headers that clearly isn't critical for Web compat |
| 11:33 | <hsivonen> | Accept-Encoding is longer than in IE and Gecko, for example |
| 11:33 | <hsivonen> | Accept-Charset advertises UTF-16!!!! |
| 11:34 | <hsivonen> | those aren't fingerprinting-relevant, but just saying |
| 11:34 | <karlcow> | hehe |
| 11:34 | <annevk> | hsivonen, bug on accept-charset is filed |
| 11:35 | <hsivonen> | Mozilla has a "guardian of the Accept header", but clearly, there should be "guardian of the UA string, too" |
| 11:35 | <hsivonen> | s/, too"/", too/ |
| 11:35 | <annevk> | what is "identity" in A-E? or "sdch" (Chrome has that) |
| 11:36 | <annevk> | we even have spaces between the values |
| 11:36 | <annevk> | very generous |
| 11:37 | <hsivonen> | annevk: Shared Dictionary Compression over HTTP says the first hit on Google |
| 11:37 | <hsivonen> | IIRC, identity is just sending the plain bytes |
| 11:37 | <hsivonen> | without compression |
| 11:37 | <annevk> | http://groups.google.com/group/SDCH |
| 11:39 | <karlcow> | http://www.w3.org/TR/ct-guidelines/#sec-Mobile-DOCTYPES |
| 11:40 | <hsivonen> | karlcow: that's sad |
| 11:40 | <karlcow> | reality is sad ;) |
| 11:41 | <annevk> | filed a bug on Accept-Encoding |
| 11:41 | <annevk> | "This list is not exhaustive and is likely to change." jaja |
| 11:41 | <hsivonen> | karlcow: fortunately, Fennec, Opera Mobile, Opera Mini and Mobile Safari work despite "reality" |
| 11:42 | <annevk> | 2.0 of "Web Content Transformation Proxies" will say they realize now the whole thing is sort of dead |
| 11:42 | <karlcow> | hsivonen: do you mean they are happy? :) |
| 11:42 | <annevk> | (I'm making that up, but that appears to be reality) |
| 11:42 | <hsivonen> | karlcow: they seem happy to me :-) |
| 11:43 | karlcow | is making apologizes for his morning syllogism :) still only one eye opened. :) |
| 11:44 | <annevk> | another thing that is sad is that Accept: is still not */* |
| 11:45 | <hsivonen> | annevk: yeah, though it might be wise to wait for IE9 to have been in the market for a while before trying to change that |
| 11:45 | <annevk> | I thought IE had that since IE6 or so |
| 11:46 | <hsivonen> | annevk: I meant the reason not to have Accept: */* is negotiating application/xhtml+xml |
| 11:46 | <annevk> | but we don't want that :) |
| 11:47 | <hsivonen> | if IE9 says */* but support XHTML+SVG *maybe* it'll make Accept-based negotiation less popular |
| 11:47 | <annevk> | more often than not XML does harm to our end users |
| 11:49 | <karlcow> | what would be the benefit of Accept: */*, because I had a use case in the past where a precise Accept would have helped filtering a DDOS |
| 11:50 | <hsivonen> | karlcow: fewer bytes to send, fewer problems for non-IE browsers when webmasters goof |
| 11:50 | <annevk> | omitting it entirely breaks |
| 11:51 | <karlcow> | my issue was the Accept for GET /blabla resulting from an IMG |
| 11:51 | <hsivonen> | karlcow: you should just fix your site :-) |
| 11:51 | <karlcow> | I had no control on it |
| 11:52 | <karlcow> | the ddos was generated by a site with a very high trafic and a bad link. |
| 11:52 | <hsivonen> | karlcow: 99.99999% of webmasters wouldn't apply your mitigation method anyway, so it's not useful to support your mitigation method on the browser side |
| 11:52 | <annevk> | karlcow, you could have done something based on Referer |
| 11:53 | <hsivonen> | karlcow: if it was cross-origin, the Origin header will save us |
| 11:54 | <karlcow> | annevk: nope. because it would have removed valid request too :) but I'm more sympathetic to the argument of hsivonen. In the end, we just block the query on that page until we found out where the problem was :) |
| 11:55 | <karlcow> | s/block/blocked/ |
| 11:55 | <annevk> | no it wouldn't |
| 11:55 | <annevk> | if you just blocked that specific referer for that resource it would've been fine |
| 12:07 | <Dashiva> | karlcow: I don't see how it makes sense to talk about a stable snapshot or milestone of an incomplete in-development specification |
| 12:07 | <Dashiva> | You can make snapshots and you can assign milestones, but they are arbitrary, not stable |
| 12:08 | <jgraham> | My pet example of why /TR is evil is WebIDL |
| 12:09 | <jgraham> | I have eventually learnt that you can find the actual latest version by being careful about whether you insert a space between Web and IDL in google |
| 12:09 | <Dashiva> | karlcow: If you agree about the stable part, wouldn't a permalink in each editor's draft satisfy the snapshot use case? |
| 12:10 | <annevk> | great, Jonny was already on Accept-Encoding simplification |
| 12:23 | <zcorpan_> | shouldn't the web timing spec put its stuff on the navigator object? |
| 12:24 | zcorpan_ | imagines window.performance is already taken by web authors |
| 12:26 | <annevk> | yeah, you should say that |
| 12:26 | <annevk> | putting new stuff on navigator when it is somewhat sensible is better than bloating window even more |
| 12:30 | <hsivonen> | there should be a standing memo for spec editors about that |
| 12:30 | <jgraham> | But everyone loves everything being in the global scope, right? |
| 12:34 | <Dashiva> | Maybe we need a new parent object for stuff that doesn't really fit on navigator :) |
| 12:36 | <jgraham> | Personally I want more properties on the window object that just refer to the window object |
| 12:37 | <jgraham> | window.Hofstadter for example |
| 12:38 | <annevk> | Dashiva, the bar for navigator is low enough |
| 13:13 | <annevk> | god |
| 13:13 | <annevk> | why would http://html5.org/tools/web-apps-tracker?from=5105&to=5106 fail but http://html5.org/tools/web-apps-tracker?from=5104&to=5105 would work? |
| 13:18 | hsivonen | wonders why Opera still ships with both the "Opera Standard" and "Windows Native" skins when the "Opera Standard" skin is *more* Windows 7 -like of the two very similar skins |
| 13:29 | <annevk> | hsivonen, we'll soonish no longer do Accept-Language |
| 13:30 | <annevk> | maybe I should file some more bugs on our network module; he's being pretty active :) |
| 13:30 | <annevk> | euh, Accept-Charset |
| 13:30 | <annevk> | my bad |
| 13:30 | <hsivonen> | annevk: cool anyway |
| 13:37 | <hsivonen> | aargh. Radiance and Ambiance are FAIL with Eclipse |
| 13:38 | <hsivonen> | still wondering how Canonical didn't have some kind of organizational function to stop the Lucid default theme |
| 13:38 | <annevk> | btw, I'm not really part of Opera QA, but then I don't really have a better description available either :) |
| 13:38 | <jgraham> | It is pretty awful |
| 13:38 | <hsivonen> | annevk: sorry |
| 13:38 | <jgraham> | Not anne's job, the Lucid theme |
| 13:39 | <annevk> | heh |
| 13:39 | <annevk> | how do you know? :p |
| 13:39 | <hsivonen> | jgraham: they should hire whoever is developing Elemental |
| 13:40 | <hsivonen> | (but having him/her/they make a theme that doesn't feel like too much of an Apple rip-opp) |
| 13:40 | <hsivonen> | *off |
| 13:41 | <hsivonen> | also, Radiance could be tweaked not to suck |
| 13:41 | <hsivonen> | Ambiance is just massive FAIL |
| 13:42 | <jgraham> | It's hard to believe that the people who designed it actually use it |
| 13:48 | <hsivonen> | s/Elemental/Elementary/ |
| 14:44 | <karlushi> | Dashiva, it could be. :) we just have to find a model which makes it usable for different types of communities and not too burdensome to manage. Often we know our current burden, it's rare we understand the issues we create when changing. But yes it could be a good idea. |
| 14:51 | <jgraham> | karlushi: I am really not sure who the community is that wants a version of the spec with fewer bugfixes in than the current version |
| 14:52 | <jgraham> | I think the stable/editor's draft model is designed around an idea of spec development that doesn't really match reality, at least for the way HTML5 is developed |
| 14:54 | hsivonen | observes that when document.write() implies a call to document.open() in Gecko, the call is always to the built-in open() and not to whatever the page has replaced open() with |
| 14:54 | <karlushi> | I do not claim reality. ;) Just trying to identify use cases. |
| 14:55 | <hsivonen> | karlushi: so what's the use case of reading a spec you know is an already-obsolete draft when you know you could read the latest draft? |
| 14:56 | <jgraham> | That is, if you have stable versions, N, N+1, etc. and unstable drafts at all times t, it only makes sense to promote version N if, on averge, version N is closer to version N+1 than unstable version at time t(N)<t<t(N+1) |
| 14:56 | <karlushi> | hsivonen, will there be at a point a stable version? |
| 14:57 | <hsivonen> | karlushi: I expect there will be, but not one that's useful to read |
| 14:57 | <karlushi> | hsivonen, for you and implementers, not useful indeed. |
| 14:58 | <Lachy> | hsivonen, the use case for snapshots is so you can compare what was written at the time of the snapshot with what is now written, which is easier than digging through CVS/SVN logs. |
| 14:59 | <jgraham> | karlushi: It is possible to be unstable, but to be convering on a stable point |
| 14:59 | <hsivonen> | Lachy: that's a reasonable use case |
| 14:59 | <karlushi> | translation is another use case. Translating a document which is changing all the time is hell. |
| 15:00 | <hsivonen> | karlushi: you just pick the draft that's the most current when you start translating |
| 15:00 | <karlushi> | s/you/one/ who's not necessary a geek. |
| 15:02 | gsnedders | likes what jgraham is typing now |
| 15:02 | <jgraham> | Heh, putting <plaintext> in <table> so it gets foster parented is amusing |
| 15:02 | <jgraham> | gsnedders: tinfb |
| 15:02 | gsnedders | is psycic |
| 16:12 | <annevk> | hmm, guess I'll look into the html5.org issue tomorrow or so |
| 16:12 | <annevk> | not gonna happen now |
| 17:14 | <AryehGregor> | hsivonen, my notion of "current" for most software is "current version I'm using", which might be current stable or trunk or something in between or maybe slightly older than current stable. In this case, 3.6; didn't bother to test on trunk because it didn't matter. I should have specified, though, since it's ambiguous. |
| 19:51 | <Gurpartap> | HTML6 http://www.artisopensource.net/hacks/deliciouspoetry.php?cshow=5217210 |
| 20:22 | <jrgarrison> | I'm confused about something mentioned in the last paragraph of the message at http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2010-June/026728.html |
| 20:22 | <jrgarrison> | "I should have written input type=file accept='audio/x-speex;quality=(0-10);bitrate=...;source=microphone' to reflect the newly-announced Android/Firefox placement of the source= parameter." |
| 20:23 | <jrgarrison> | does anybody know what announcement this refers to? |
| 20:38 | AryehGregor | has no idea. |
| 20:48 | <variable> | question: is there a slightly more offical method of tracking proposals than the mailing list? I know about the bug list for the W3C. Does the WHATWG have anything similar? |
| 21:08 | <Dashiva> | The issue tracker |
| 21:09 | <Dashiva> | http://www.whatwg.org/issues/ |
| 21:11 | <variable> | Dashiva, thank |
| 21:33 | <AryehGregor> | My giant comment got accepted. Everyone feel free to complain about how terribly wrong I am, particularly since I don't have such a good grasp on the history here. http://www.cssquirrel.com/2010/06/15/693comic-update-html5-unicorn-heuristics/comment-page-1/#comment-32000 |
| 21:44 | <johnst> | tl;dr |
| 21:44 | <AryehGregor> | :) |
| 23:19 | <Hixie> | annevk: index didn't change in r5106 but did in the other two |
| 23:46 | <karlcow> | AryehGregor: only one wrong fact in the comment |
| 23:47 | <karlcow> | "this is directly due to completely wrong-headed thinking by the W3C administration." |
| 23:47 | <karlcow> | s/W3C Administration/W3C Members/ |
| 23:50 | <karlcow> | Also "The W3C has to this day contributed almost nothing to the spec for practical purposes." makes me huh? |
| 23:50 | <karlcow> | W3C = Members of W3C |
| 23:50 | <karlcow> | if you meant W3C employees have not contributed. Then 1) it's not their primary role. 2) They did contributed in terms of handling all the administrivia of WGs life, liaison, setting mailing lists, web spaces. I would be surprised to consider MikeSmith work = nothing |