| 00:00 | <Dashiva> | Yeah, probably local config done similar to google's urchin thing |
| 00:01 | <Philip`> | http://www.516545.com/bbs/archiver/?tid-280.html |
| 00:02 | <Hixie> | i just called it Baidu. |
| 00:03 | <Hixie> | wow, a lot of pages refer directly to http://s.wordpress.com/wp-content/plugins/highlight/shCore.js |
| 00:06 | <Philip`> | I imagine Baidu has plenty of other scripts, so maybe it's better to call it "Baidu contextual advertising" or something |
| 00:07 | <Hixie> | that's ok, i have lots of things that share the same name |
| 00:09 | Philip` | wonders if Baidu offers language translation services too |
| 00:10 | <othermaciej> | what does shCore.js do? |
| 00:12 | <Philip`> | It highlights |
| 00:12 | <Hixie> | syntax highlighter plugin for wordpress |
| 00:13 | <Hixie> | oddly, that url seems to be addressed more often than any other wordpress urls |
| 00:14 | <Hixie> | i'm amazed at how many of these scripts are for stats collection |
| 00:49 | <Hixie> | i wish i spoke russian |
| 00:49 | <Hixie> | anyone know anything about http://autocontext.begun.ru/autocontext.js ? |
| 00:51 | Lachy | is disabling comments on old whatwg blog posts. |
| 00:51 | <Hixie> | too much spam? |
| 00:52 | <Lachy> | I'm getting annoyed with the constant moderation requests slipping through the spam filter |
| 00:52 | <Hixie> | can we point people to help⊙wo or something in the "comments are closed" message? |
| 00:52 | <Lachy> | it's only about 5 per day that get through (4910 sitting in the Akismet spam filter, though) |
| 00:53 | <Lachy> | I supppose I could update the template |
| 00:54 | <Hixie> | i think it'd be helpful, so that we don't lose potential contributors |
| 00:58 | <Hixie> | man, this is mind numbing work |
| 00:59 | <Lachy> | I also need to delete the bogus users that have registered. I cleared about 50 of them a few weeks ago and about a hundred have registered :-) |
| 00:59 | <Hixie> | i have seen so much bad JS today |
| 00:59 | <Hixie> | heh |
| 00:59 | <Hixie> | i wouldn't worry too much about bogus users |
| 00:59 | <Hixie> | do they cause any harm? |
| 00:59 | <Hixie> | not that i want to discourage you from maintaining the blog, i just don't want you to burn out doing it :-) |
| 01:00 | <Lachy> | although it's difficult to tell if some are legit, the obvious ones like "alltickets" and "handbagcollection" can go :-) |
| 01:00 | Hixie | stares at http://www230.clickeye.cn/common/clickeye.js |
| 01:00 | <Hixie> | Lachy: hehe |
| 01:01 | <Lachy> | not much effort, it only took about 20 min to disable comments and should only take about 5 to sort out the bogus users |
| 01:01 | <Hixie> | k :-) |
| 01:01 | <Hixie> | i went it a few months ago and promoted those names i recognised |
| 01:04 | <Hixie> | does "export/sites/www/shared/scripts" mean anything to anyone? |
| 01:08 | <Lachy> | all done |
| 01:10 | <Hixie> | i have labelled 435 src="" attribute values |
| 01:11 | <Hixie> | and am about ready to shoot myself |
| 01:11 | <Hixie> | maybe i should take a break! |
| 01:13 | <Lachy> | I updated the authoring document with sections about the body and section elements |
| 01:13 | <Lachy> | http://lachy.id.au/temp/html5-authors.html |
| 01:14 | <Lachy> | it'll be similar to the spec, but with extra sections for talking about start/end tags, i18n and accessibility |
| 01:16 | <Philip`> | "XHTML Example: <img src="image" alt="example">" - s/X// |
| 01:17 | <Philip`> | Also maybe s/image/image.png/ so it's more obvious what's going on |
| 01:17 | <Lachy> | ok. Then I should be consistent and use movie.ogv |
| 01:18 | <Dashiva> | Could always use file.ext to avoid specifics |
| 01:18 | <Philip`> | Why avoid specifics? |
| 01:18 | <Hixie> | oo, only two more days until the two questionnaires are up |
| 01:18 | <Philip`> | Incidentally, .ogv seems an odd choice since I mentally expand it to "Ogg Vorbis" which is totally wrong |
| 01:19 | <Dashiva> | Because some people might complain about the spec endorsing specific formats |
| 01:19 | <Lachy> | realistic examples using real file extensions is better |
| 01:19 | <Hixie> | people will complain regardless of what you do |
| 01:19 | <Lachy> | ogv is ogg vorbis |
| 01:19 | <Hixie> | might as well be helpful :-) |
| 01:19 | <Philip`> | Lachy: It seems to be Ogg Video |
| 01:19 | <Lachy> | (I think that's what I saw in that draft RFC for video/ogg) |
| 01:19 | <Philip`> | which is Theora more than Vorbis |
| 01:19 | <Lachy> | oh right, I meant ogg video |
| 01:20 | <Dashiva> | Hixie: Then I vote for mkv instead :P |
| 01:21 | <Hixie> | i vote for mpg :-) |
| 01:21 | <Philip`> | I think we should put videos in a PNG container, to get backward-compatibility with old UAs - they can still load the 600MB movie file and will just display the first frame |
| 01:21 | <Hixie> | Philip`: :-P |
| 01:21 | <Philip`> | (i.e. put the poster frame in the IDAT, and use decent compression in custom chunks for the actual video) |
| 01:22 | <Philip`> | It's not that insane, really |
| 01:22 | <Lachy> | Philip`, we could just use APNG for the video |
| 01:22 | <Lachy> | or GIF |
| 01:23 | <Philip`> | Lachy: That would be insane, since it's even worse than MJPEG compression :-p |
| 01:23 | <Lachy> | yeah but PNG is lossless compression, so it's better quality |
| 01:24 | <webben_> | Would be possible (I'm not talking about useful) to declare the HTML serialization of HTML5 in SGML terms? And if not, is there a handy list anywhere of fundamental incompatibilities that make that impossible? |
| 01:24 | <Philip`> | Good point - and you could use 16-bit colour channels if you want really excellent quality |
| 01:25 | <Hixie> | webben_: yes, you could describe a serialisation of html using any tree-based syntax |
| 01:25 | <Hixie> | webben_: you could describe a JSON serialisation, e.g. |
| 01:25 | <webben_> | no |
| 01:26 | <webben_> | I mean could you describe the HTML serialization (not fully, but compatibly) using an SGML declaration and DTD? |
| 01:26 | <Hixie> | oh, i see |
| 01:26 | <Hixie> | hmm |
| 01:26 | <Hixie> | probably not |
| 01:26 | <Philip`> | webben_: Does that include error handling, or only parsing conformant HTML-serialised character streams into the same tree as the HTML5 algorithm? |
| 01:26 | <Hixie> | it has some weirdass things |
| 01:26 | <Lachy> | Philip`, what about audio? AFAIK, PNG doesn't support audio tracks :-) |
| 01:26 | <Hixie> | you could probably get quite close |
| 01:26 | <Hixie> | but i doubt you could exactly define it |
| 01:27 | <webben_> | Philip`: I don't know enough about SGML-parsing conformance requirements to say. |
| 01:27 | <Philip`> | Lachy: Just define an Audio PNG extension |
| 01:27 | <webben_> | I'm not talking about expressing everything that html5 forbids necessarily, but perhaps everything it allows? |
| 01:27 | <Philip`> | Encode the audio track as a greyscale image, and stick that into the movie somewhere |
| 01:28 | <Lachy> | http://lists.w3.org/Archives/Public/www-archive/2007Nov/0080.html |
| 01:32 | <Hixie> | webben_: there are weird things like the way we treat entities in <textarea> that i don't think you can describe |
| 01:32 | <Hixie> | in sgml |
| 01:32 | <Hixie> | but i could be wrong |
| 01:32 | <Hixie> | so i would answer "no" to your question |
| 01:32 | <webben_> | Hixie: OK. Thanks. :) |
| 02:02 | <Lachy> | Hixie, re your post to public-html, does google really care about keeping documents small? I assume they realise that if it was really that really that important, the could save a whole heap just by cleaning up their homepage |
| 05:30 | <othermaciej> | wow, I hadn't heard that Opera is opening a Mountain View office |
| 05:30 | othermaciej | hopes he will get to meet Opera people more often |
| 06:01 | <Hixie> | http://slashdot.org/article.pl?sid=07/11/22/001228 - " |
| 06:01 | <Hixie> | http://slashdot.org/article.pl?sid=07/11/22/001228 - "Ecma Receives 3,522 Comments on Open XML Standards |
| 06:01 | <Hixie> | pah |
| 06:01 | <Hixie> | html5 has received more comments than that |
| 06:04 | <othermaciej> | clearly there's a need for better marketing |
| 09:41 | <hsivonen> | Hixie: OK if I allow any space characters in the space production for file upload accept: http://www.whatwg.org/specs/web-forms/current-work/#accept0 ? |
| 09:57 | <hsivonen> | email sent |
| 09:57 | <hsivonen> | to public-html this time |
| 10:57 | <hsivonen> | charset list and map coordinates to go... |
| 10:58 | <hsivonen> | aargh. a dependency on HTML4! |
| 10:59 | <hsivonen> | "Authors must not specify an encoding other than UTF-8 or US-ASCII in the accept-charset attribute when the method used is get and the action indicates an HTTP resource." yay. that's a fun requirement |
| 10:59 | <hsivonen> | Hixie: the requirement depends on the document base URI |
| 11:00 | <hsivonen> | Hixie: hence, the document may cease to be conforming if you move it |
| 12:01 | <hsivonen> | Hixie: regarding my email to public-html about accept-charset: it would help to get your current guess about where the requirements are headed |
| 12:08 | <Lachy> | hsivonen, allowing commas as separators in accept-charset makes sense because the value can then be consistent with the HTTP AcceptCharset header value |
| 12:14 | <hsivonen> | Lachy: it still sucks from the SGML/XML/HTML design patterns point of view |
| 12:16 | <Lachy> | have you tested to see what browsers actually support? |
| 12:16 | <hsivonen> | Lachy: I have not |
| 12:16 | <zcorpan> | hsivonen: same argument with <area coords>? |
| 12:16 | <hsivonen> | zcorpan: yeah |
| 12:16 | <hsivonen> | zcorpan: but too late |
| 12:17 | <hsivonen> | Lachy: I'm always kinda hoping that someone else would reply to my messages with browser testing results :-) |
| 12:22 | <hsivonen> | while I'm at it: it also sucks that SVG allows commas in its coordinate lists |
| 12:30 | <hsivonen> | I wonder whose idea it was to allow single quotes as part of the IANA charset token syntax... |
| 12:32 | <hsivonen> | how do I output <var> or <samp> from mediawiki? should I just settle for <i> and <code>? |
| 12:35 | <hsivonen> | ISO reaches out with advocacy FAQs: http://www.iso.org/iso/support/faqs/faqs_widely_used_standards/widely_used_standards_other/date_and_time_format.htm |
| 12:35 | <hsivonen> | now if they could serve their PDFs without zip wrappers... |
| 13:51 | <Philip`> | http://www.iso.org/iso/date.jpg - it's weird how my brain can decode the tiny letters there, but if I zoom in then it turns into a meaningless lump of grey pixels |
| 13:51 | <Philip`> | (It's also weird how they don't link to a larger version of that image) |
| 13:52 | <gsnedders> | Philip`: damn you and your eye-sight! |
| 13:55 | <virtuelv> | Philip`: heh. Same experience here |
| 14:08 | <Philip`> | X3D appears to have a lot more undefined behaviour than what web browsers seem to appreciate :-( |
| 15:28 | <Lachy> | Philip`, I have seen a larger version of that image previously. Though I have no idea where to find it |
| 19:39 | <jruderman> | Philip`: zooming with smoothing or with nearest-neighbor? |
| 19:39 | <jruderman> | Philip`: with smoothing, zooming doesn't seem to hurt the readability for me |
| 19:40 | <hsivonen> | I could use examples of invalid IDNs, but the obvious google terms don't help me |
| 19:46 | <hsivonen> | Do I understand correctly that there is no proper spec for IDN-aware mailto: IRI scheme? |
| 19:57 | <hsivonen> | am I right that "foo@" is an invalid email address? or is there something I'm missing? |
| 19:57 | <hsivonen> | hmm. I think I have established that the Jena IRI library doesn't do proper scheme-aware mailto: IRI validation |
| 20:07 | <Philip`> | jruderman: Nearest-neighbour (i.e. what Opera and FF3 do) |
| 20:16 | <hsivonen> | does anyone remember off-hand whether XMPP addresses are defined by reference to RFC 2822 or whether it is a similar-looking standalone syntax? |
| 20:17 | <gsnedders> | I want to say a subset of RFC 2822, but I might be wrong |
| 20:17 | <hsivonen> | http://www.xmpp.org/rfcs/rfc3920.html section 3.1. |
| 20:17 | <hsivonen> | looks like they defined their own |
| 20:18 | <hsivonen> | in a remarkably sane way I might add |
| 20:18 | <gsnedders> | hsivonen: doesn't 2822 make anything look sane? :) |
| 21:36 | Philip` | wonders why X3D uses the "DEF" attribute for IDs, instead of "id" or "xml:id" or something |
| 21:50 | <hsivonen> | Philip`: are they really IDs? |
| 21:50 | <hsivonen> | Philip`: CML makes id non-ID |
| 21:51 | <Philip`> | hsivonen: The schema says <xs:attribute name="DEF" type="xs:ID"> so I assume that means they're IDs |
| 21:52 | <hsivonen> | Philip`: ok |
| 21:52 | <Philip`> | I guess getElementById won't work, which is a pain |
| 22:20 | <Philip`> | Is XPath like //*[DEF="foo"] likely to be horribly slow in Firefox/Opera? |
| 22:25 | <Dashiva> | Compared to what? gEBTN and attribute checking? |
| 22:26 | <Philip`> | Compared to precomputing a hash of DEF->object, I guess |
| 22:27 | <Philip`> | I suppose it depends on how many lookups I'm doing and how big the document is, and I've got no idea about those at the moment, so I should forget about it unless it becomes a problem in practice |
| 22:28 | <Dashiva> | Sounds like a plan. No need to overoptimize something that might not even be a bottleneck |
| 22:28 | <Philip`> | Also no need to optimise something that will probably never get anywhere near completion either :-) |
| 22:58 | <gsnedders> | Hixie: relating to what we were talking about yesterday: http://lists.w3.org/Archives/Public/ietf-http-wg/2007OctDec/0270.html |
