| 07:46 | <zcorpan> | morning |
| 08:24 | <Hixie> | morning |
| 08:32 | <hallvors> | afternoon. whatever. :) |
| 08:32 | <Hixie> | anyone remember what the url for Lachy's live-dom-viewer-like tool is? |
| 08:34 | <Lachy> | http://html5.lachy.id.au/beta/ |
| 08:39 | <Hixie> | thanks |
| 08:39 | <Hixie> | what does the MIME type field do in that? |
| 08:40 | <Lachy> | nothing yet |
| 08:40 | <Hixie> | ah ok |
| 08:41 | <Lachy> | it will eventually serve the same purpose as the MIME type field on this page http://html5.lachy.id.au/ |
| 08:42 | <Hixie> | so we don't have a live dom viewer for xml yet, right? |
| 08:43 | <Lachy> | no. I tried to make one, but had difficulty getting it to function properly. I can't remember exactly what the problems were |
| 08:44 | <Hixie> | yeah i seem to recall the same |
| 08:46 | zcorpan | is confused why document.lastChild.lastChild isn't a text node in http://hsivonen.iki.fi/test/moz/xmlns-dom.xhtml |
| 08:47 | <zcorpan> | oh |
| 08:47 | <zcorpan> | the script runs during parsing |
| 08:50 | Lachy | is contemplating whether to get up and go skiing for the first time this season, or be lazy and stay home |
| 08:54 | <Lachy> | there's no real surprises in this list. It's basically the same stuff people have been asking for, for several years now http://fantasai.inkedblade.net/style/discuss/wasp-feedback-2008 |
| 08:56 | <Lachy> | this one is partially addressed by <style scoped> already, but the ability to reset styles wouldn't work too well http://fantasai.inkedblade.net/style/discuss/wasp-feedback-2008#scoped-style |
| 08:57 | <Lachy> | http://www.webstandards.org/2009/01/16/css-working-group-feeds-back-to-wasp/ |
| 09:13 | <Lachy> | oh dear. The trains to the ski slope don't run frequently enough today to be worth going :-( It'll be after midday before I arrive |
| 09:27 | <Hixie> | http://lists.w3.org/Archives/Member/w3c-html-cg/2009JanMar/0018.html are the xhtml2 wg even working on xhtml2 anymore? i haven't seen them mention it in status reports in months |
| 09:29 | <BenMillard> | Hixie, Member links for the lose. :( |
| 09:30 | <Lachy> | BenMillard, it looks like a status report about their specs, except that it omits XHTML2 |
| 09:32 | <zcorpan> | maybe xhtml2 is perfect so no more work needs to be done |
| 09:35 | <zcorpan> | discussion about html5 in general has gone up lately but the forums are dead |
| 09:35 | <zcorpan> | wonder why |
| 09:36 | <BenMillard> | Lachy, cheers. Do they currently have a charter? I could only find this: http://www.w3.org/2007/03/XHTML2-WG-charter |
| 09:37 | <BenMillard> | (linked to in the first paragraph of their homepage: http://www.w3.org/MarkUp/) |
| 09:37 | <Lachy> | that is their current charter |
| 09:38 | <BenMillard> | Lachy, oh I misread 2009 as 2008. :P |
| 09:40 | <BenMillard> | 2.3 Milestones seems out of date. "XHTML 2.0" to become "REC" in "Sep 2008"? |
| 09:41 | <Lachy> | LOL |
| 09:41 | <Lachy> | I wonder where they intend to find implementations? |
| 09:42 | <zcorpan> | Lachy: much of xhtml2 already works in browsers so maybe that's enough to label it as a rec |
| 10:15 | <hsivonen> | Lachy: do the implementations need to be browsers? can they be a schema that validates XHTML2 and an XSLT transform that produces XHTML2? |
| 10:40 | <Lachy> | hsivonen, I would hope there would be some implementations designed for end users as well, not just developer tools |
| 10:49 | <deane> | I can not understand how they were allowed to switch from the xhtml2 namespace to the same namespace (xhtml) that we are using, we were using it first. When html5 came to the W3C it was using the xhtml namespace, and xhtml2 had it's own namespace, I can't see how they can just change to the same one as us like that. |
| 10:51 | <Philip`> | I thought XHTML2 had switched to the xhtml namespace (in (unpublished?) drafts) before HTML5 was in the W3C |
| 10:51 | <hsivonen> | deane: the same way they can write a de facto text/html-targeted spec that relies on stuff like xmlns:dc |
| 11:00 | <deane> | hsivonen: And the same way that they can publish a document about what mime types are ligit for xhtml even though they are not the only ones working on xhtml specs and this bogus document contradicts the HTML5 spec. Then when someone complains about it, they just go ahead and publish the document anyway. |
| 11:02 | <hsivonen> | deane: if you have the cycles, you could file Formal Objections, I suppose |
| 11:14 | <deane> | hsivonen: "...stuff like xmlns:dc" yeah, but instead of acknowledging that these people aren't writting specs for the real web, it appears that Shelley and others think it is simply the html5 communitie's fault for not having the forsight/insight/intelligence to put distributed extensibility into text/html |
| 11:16 | <Lachy> | deane, where/when did the XHTML2 WG announce that they have actually changed the namespace? The current draft, published in 2006, still uses the other ns |
| 11:17 | <BenMillard> | Lachy, I noticed it in a diff-marked Wikipedia page hsivonen linked this channel to a while back...not sure of official documents, though. |
| 11:18 | <Lachy> | the last I heard about the NS issue was that they were intending to do so, but that was well before we had all the discussion about it on public-html |
| 11:19 | <Philip`> | http://www.w3.org/MarkUp/2007/ED-xhtml2-20071024/conformance.html |
| 11:19 | <Philip`> | <html xmlns="http://www.w3.org/1999/xhtml"; xml:lang="en" |
| 11:19 | <Philip`> | xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"; |
| 11:19 | <Philip`> | xsi:schemaLocation="http://www.w3.org/1999/xhtml |
| 11:19 | <Philip`> | http://www.w3.org/MarkUp/SCHEMA/xhtml2.xsd"; |
| 11:20 | <Philip`> | (It's so pretty) |
| 11:20 | <hsivonen> | http://ascii.textfiles.com/archives/1649 |
| 11:21 | <deane> | Lachy, http://www.w3.org/2007/10/03-xhtml-minutes.html |
| 11:26 | <Lachy> | hsivonen, re the RDFa namespace and DOM discussion, that seems to be roughly the same issues that aria encountered when they wanted to add in namespace prefixed attributes and values |
| 11:26 | <deane> | Lachy: Shane's latest "Editor's Draft" http://www.w3.org/MarkUp/2009/ED-xhtml2-20090109/ the "public draft" hasn't been updated since 2006 |
| 11:29 | <Lachy> | I wish there were a latest editors draft link for XHTML2, instead of just dated links |
| 11:29 | <Lachy> | where does the xhtml2 group post the links to the editors drafts? |
| 11:31 | zcorpan | wonders what happened with this week in html5 |
| 11:35 | <deane> | Lachy: in the "XHTML Document Development Area" http://www.w3.org/MarkUp/Drafts/ |
| 11:39 | <BenMillard> | hsivonen, I had worries along those lines about Geocities in 1998 and freewebs in 2000. Those are what eventually compelled me to buy hosting from someone whose home phone number I have and who sends me an Xmas card each year. :) |
| 11:40 | <BenMillard> | hsivonen, it's also one reason my has no comments system; the responsibility of looking after other people's content is daunting. |
| 11:40 | <BenMillard> | s/my has/my website has/ |
| 11:43 | <Lachy> | that's why I keep a copy of all my data locally and have a weekly backup of my blog database emailed to me |
| 11:44 | <Lachy> | although there was one occasion that I lost a number of irreplacable files |
| 11:45 | <Lachy> | not too many though |
| 12:43 | <gsnedders> | My hands are cold :( |
| 16:44 | <Philip`> | Incrementally parsing XML doesn't seem a particularly weird thing, since that's kind of what SAX is for |
| 16:44 | <Philip`> | and it's kind of obvious how it should work even if nobody has written a proper detailed spec |
| 16:45 | <Philip`> | (and the problem is just that implementors don't understand XML and introduce bugs by e.g. failing to use a proper namespace-aware serialiser) |
| 16:49 | <Lachy> | oh, this comment from Travis is a relief http://annevankesteren.nl/2009/01/gettters-setters#comment-6694 |
| 16:49 | <Lachy> | but there's still no mention of __defineGetter__ and __defineSetter__ in IE, which could be a problem |
| 16:52 | <zcorpan> | Lachy: maybe they still aren't used enough in the wild that one could get away with not supporting them |
| 16:54 | myakura | heard that javascript:alert(document.body.__defineGetter__); returns undefined in the pre-RC build (and it was [native code] in beta2) |
| 16:54 | <zcorpan> | myakura: seems about right |
| 16:58 | <Lachy> | myakura, it's returning undefined for me in RC1 |
| 16:58 | <myakura> | :( |
| 17:00 | <Lachy> | it really sucks that the ECMA spec is in PDF instead of HTML :-( |
| 17:07 | <annevk> | Lachy, no kidding |
| 17:10 | <Lachy> | it sucks even more that they've produced the PDF with revision tracking turned on, making it harder to read |
| 17:10 | <zcorpan> | annevk: the logo seems hard to redo in svg - it just becomes smudgy |
| 17:11 | <annevk> | it's not really vectorlike |
| 17:11 | <Philip`> | Combine a vector outline with a bitmap texture? |
| 17:11 | <annevk> | unless you use a filter or something I suppose |
| 17:12 | <annevk> | what Philip` said |
| 17:13 | <annevk> | hehe, "SVG SUCKS", so true |
| 17:17 | <zcorpan> | annevk: you could use the webfont for the "WEBLOG 4.2" part |
| 17:19 | <annevk> | would that be an improvement? |
| 17:19 | <annevk> | i guess it would |
| 17:21 | <annevk> | I wonder if the SVG can be further simplified |
| 17:21 | <Philip`> | Shouldn't it be at least WEBLOG 4.3 by now? |
| 17:22 | <annevk> | it's not an actual version number |
| 17:31 | <jcranmer> | o_O |
| 17:32 | <jcranmer> | what is so hard about creating viable use cases for RDFa? |
| 17:35 | <Philip`> | jcranmer: It's designed to satisfy a large number of classes of use cases, and if you only focus on a dozen individual cases then you'll probably come up with a dozen solutions that are simpler than RDFa for those use cases but that are useless for the million other use cases |
| 17:36 | <gsnedders> | annevk: Shouldn't it be WEBLOG 5? |
| 17:37 | <jcranmer> | Philip`: even so, everyone always acts as if there's no reason why RDFa shouldn't just be included |
| 17:38 | <Philip`> | so the concept of analysing use cases is fundamentally flawed, as it can't come up with globally optimal solutions that have the best overall cost-per-use-case |
| 17:43 | <takkaria> | RDFa is a globally optimal solution? :) |
| 17:44 | <annevk> | gsnedders, not a version number |
| 17:44 | <gsnedders> | annevk: But 5 > *. |
| 17:44 | <Philip`> | takkaria: It may or may not be, but we won't be able to determine that if we only look at individual use cases :-) |
| 17:58 | <zcorpan> | annevk: http://simon.html5.org/dump/logo.svg |
| 17:59 | <zcorpan> | it looks worse, double the size, but at least it's scalable |
| 18:01 | <annevk> | hmm, the non-SVG version is still nicer |
| 18:01 | <annevk> | thanks though :) |
| 18:01 | <zcorpan> | welcome :) |
| 18:02 | <zcorpan> | i could make it as nice, but then the file size would be ridiculuos |
| 18:03 | <zcorpan> | (like 50kb) |
| 18:04 | <zcorpan> | but daddy.svg is good enough, no? |
| 18:06 | <rubys> | how was the AvK svg produced? |
| 18:07 | <zcorpan> | rubys: livetrace in illustrator |
| 18:07 | <rubys> | I'm impressed... most tools like that produce 8 digits of (quite unnecessary) precision and a lot of style crap. |
| 18:08 | <zcorpan> | rubys: i did save for web & devices, set digits to 1 and cleaned it up afterwards |
| 18:09 | <jcranmer> | I'd say 2, maybe 3, digist is all you need |
| 18:09 | <zcorpan> | also i traced twice, once with high threshold (the 0.5 opacity group) and once with low threshold (foreground) |
| 18:10 | <zcorpan> | maybe it could be tweaked to look better |
| 18:11 | <zcorpan> | i wonder why they emit digits at all - if most coordinates have digits you're wasting the file size with dots |
| 18:12 | <zcorpan> | instead you could bump up viewBox and work with integers |
| 18:13 | <rubys> | I try to do mine all in integers -- and with a viewBox of about 100x100 |
| 18:13 | <Philip`> | If you care about filesize then you'll gzip it, and the decimal points add very little extra entropy so they shouldn't take much space |
| 18:14 | <zcorpan> | Philip`: it's still an optimization that could be made |
| 18:14 | <rubys> | in this svg, consecutive paths could be collapsed, and the translations factored in. |
| 18:14 | <annevk> | as long as there are enough dots :) |
| 18:15 | <zcorpan> | and it allows for more fine-tuned setting of the accuracy (instead of 1..7 digits you can have say 10..10000000 wide viewBox) |
| 18:17 | <zcorpan> | rubys: feel free to modify it :) |
| 18:18 | <rubys> | another example of unnecessary precision: c-1,0-0.8,1.2-1,2 ... unless you zoom in *really* strong, that's a straight line: l-1,2 |
| 18:19 | <rubys> | I have some scripts that can optimize individual paths. Annevk: is this something you intend on using? |
| 18:21 | <annevk> | I plan on using http://simon.html5.org/dump/daddy.svg at least |
| 18:21 | <annevk> | I would like to use the other as well though currently it looks a lot worse imo than what I have now |
| 18:22 | <zcorpan> | annevk: would you use it if it looked better but was maybe 20kb? |
| 18:23 | <rubys> | I like that one! (daddy.svg). Especially how the text is, well, text. So you can easily rotate in different slogans... |
| 18:30 | <annevk> | yeah, and the font Arthur used is free, so we can make it identical |
| 18:30 | <annevk> | zcorpan, I suppose |
| 18:32 | <annevk> | my weblog uses less bandwidth than html5.org |
| 18:32 | <annevk> | and 3 times as much as philip.html5.org |
| 18:35 | <rubys> | looks like I can cut daddy.svg in half, and that's without combining the paths |
| 18:38 | <annevk> | wow |
| 18:40 | <annevk> | html5.org now uses 9.37 GB of disk space for diff caches |
| 18:40 | <annevk> | that's 3 GB more than a month ago |
| 18:42 | <rubys> | early results: http://intertwingly.net/tmp/daddy.svg; clearly the text is misplaced and the bottom half is a single path that was mangled by my script. But this should be an indication of some of the savings that can be made. |
| 18:42 | <rubys> | And I can do more - given the mangling, I turned off a portion of my script that optimizes. |
| 18:46 | <rubys> | original image is 14KB :-) |
| 18:49 | <zcorpan> | http://simon.html5.org/dump/logo2.svg might be slightly better (and to my surprise no larger) |
| 18:54 | <hsivonen> | jcranmer: the usual demo use cases of RDFa take a microformat use case and make the solution more complex |
| 19:10 | <annevk> | except that they hide the namespace declarations from the audience so it looks more acceptable |
| 19:11 | <annevk> | hmm, I know xmlns="" can't be put in a namespace due to authors using [xmlns] as a style hook, dunno about xmlns:<str>="" |
| 19:11 | <rubys> | seems less likely |
| 19:14 | <annevk> | yeah |
| 19:14 | <Lachy> | rubys, what benefit would it be to have xmlns:foo put into the DOM in the same way it is in XML, without having to introduce full namespace syntax for elements and attributes? |
| 19:14 | <annevk> | though I suspect script libraries (might even be RDFa script libraries!) could get broken if we change things |
| 19:15 | <hsivonen> | how does GRDDL deal with CURIEs given the relative purity of the XPath data model compared to the DOM? |
| 19:15 | <Lachy> | e.g. given <html xmlns:foo="...">...<foo:bar></foo:bar>... How would you expect the <foo:bar> element to be parsed? |
| 19:17 | <Lachy> | given that we already went through this whole namespace prefix discussion with aria: attributes and the conclusion of that was to go with aria-attr syntax, I don't think we should bother going into it again with xmlns:foo and other prefixed stuff |
| 19:18 | <hsivonen> | Lachy: I don't like it that we have to spend cycles researching if it would break the Web instead of the XHTML2 WG sticking to stuff that doesn't break for sure. |
| 19:19 | <Lachy> | agreed |
| 19:23 | <gsnedders> | danbri: you said you were cc'ing timbl, but you don't appear to have |
| 19:29 | <annevk> | http://fuckyoupenguin.blogspot.com/ lol |
| 19:33 | <Lachy> | haha! :-D |
| 19:34 | <rubys> | hsivonen: yet you have no problem asking danbri to spend cycles doing something that (given your snarky comment above) aren't likely to support |
| 19:37 | <rubys> | I'm also not clear what the importance of localName is... localName is undefined in IE. My experience is that things that aren't supported in IE tend to have a significant reduced chance of being widely adopted. |
| 19:38 | <rubys> | If I were in DanBri's shoes, I would be inclined to think that this is all one big snipe hunt. |
| 19:40 | <rubys> | The net affect of all this: asking of others what you are unwilling to do yourself increases the perception that this effort isn't truly open. |
| 19:41 | <hsivonen> | rubys: as far as I can tell, there has been no dialog between the WHATWG and the RDFa TF--only indication that they are going to REC and we need to deal |
| 19:41 | <danbri> | gsnedders, yeah i forgot the cc:, I fwd'd it to him straight afterwards, cc:'ing Henri |
| 19:42 | <gsnedders> | danbri: Just making sure you hadn't forgotten and not noticed :) |
| 19:42 | <danbri> | thanks :) |
| 19:42 | <hsivonen> | rubys: I'm much more OK with expending effort towards working around SVG legacy things than adjusting parsing to fit new things that were designed with full knowledge of text/html and ignored DOM realities |
| 19:42 | <danbri> | rubys, i prefer to take an optimistic view of apparent conflict, until i get real evidence otherwise |
| 19:44 | <hsivonen> | rubys: I might even be proactive with validating an RDF serialization like I was with ARIA if I can see that supporting the serialization won't be really bad in a context of XML parser & HTML parser connected to a RELAX NG validator |
| 19:45 | <rubys> | hsivonen: would changing the HTML parser to deal with attributes that start with "xmlns:" address your requirement? |
| 19:45 | <danbri> | is validator.w3.org relax-ng powered nowadays? |
| 19:45 | <hsivonen> | danbri: the HTML5 part is |
| 19:45 | <danbri> | cool |
| 19:45 | danbri | likes what little he knows of relaxng |
| 19:47 | <hsivonen> | rubys: mapping those attributes to SAX startPrefixMapping calls would give above-parser consistency, but the idea of adding that parser code for an unproven technology that could easily have avoided this problem offends me |
| 19:48 | <hsivonen> | rubys: that is, I think it would set an awfully bad precedent to let other WGs throw HTML-hostile RECs at us and just keep adjusting HTML5 |
| 19:50 | <hsivonen> | the experience with the current code for dealing with xml:lang and xmlns:* still irks me so much that I'm not feeling at all charitable about the prospect of doing even more contortions for CURIEs |
| 19:50 | <rubys> | I'm sure that there is an equally valid perspective where people can see the HTML and WHATWG as being the ones doing the throwing... is there anyway to get past the "everything is one big pissing context" perspective? |
| 20:00 | <zcorpan> | annevk: http://simon.html5.org/dump/logo3.svg |
| 20:07 | <gsnedders> | Useless code for today: http://pastebin.ca/1311876 |
| 20:07 | gsnedders | stares at the source |
| 20:07 | <gsnedders> | zcorpan: You have to be kidding… |
| 20:10 | <zcorpan> | gsnedders: sorry i don't follow |
| 20:11 | <gsnedders> | zcorpan: What is the point of having that as SVG when it is nothing but paths? Surely it won't scale well. |
| 20:11 | <zcorpan> | what's wrong with paths? |
| 20:12 | <annevk> | zcorpan, that last one is nice |
| 20:12 | <gsnedders> | Surely they won't scale so well as you won't have curved edges properly scaling (and eventually scaling to the point of seeing the straight lines that make up the curve) |
| 20:12 | <zcorpan> | annevk: :) |
| 20:13 | <zcorpan> | gsnedders: paths don't have to be straight (and they aren't) |
| 20:13 | <zcorpan> | try zooming |
| 20:13 | gsnedders | is blatantly a n00b |
| 20:14 | <rubys> | http://www.w3.org/TR/SVG11/images/paths/quad01.svg |
| 20:14 | <zcorpan> | it's 25kb, but maybe rubys can make it smaller and then gzipped it should be acceptably small :) |
| 20:16 | <annevk> | yeah, definitely compared to the several MB home pages you see now and then |
| 20:18 | <zcorpan> | i had to resize the raster image to 800% or so and then patch some lines here and there so it wouldn't be so boxy |
| 20:18 | <zcorpan> | and then livetrace worked a lot better |
| 20:20 | <zcorpan> | also if i checked the 'use strokes' checkbox and tweaked the parameters so that it actually wouldn't output any strokes, the number of paths were radically reduced (but looked about the same) |
| 20:21 | <rubys> | which is 25kb? |
| 20:21 | <zcorpan> | http://simon.html5.org/dump/logo3.svg |
| 20:23 | <rubys> | Looks like I can get that under 10kB.... before compression. |
| 20:23 | <zcorpan> | cool :) |
| 20:24 | <rubys> | I need to debug why the output of my tool is incorrect first. :-) |
| 20:24 | <rubys> | But I trust that the size is is in the right ballpark. |
| 20:24 | <rubys> | and I want to debug my tool anyway |
| 20:26 | <annevk> | http://news.cnet.com/8301-1023_3-10145058-93.html :/ |
| 20:41 | <rubys> | http://intertwingly.net/tmp/logo3.svg |
| 20:44 | <zcorpan> | looks similar enough at 100% zoom :) |
| 20:45 | <rubys> | and yet, a bit more bandwidth friendly. |
| 20:49 | <zcorpan> | it might have been a bit too agressive if one cares about nice looks when zoomed in a bit (or when printed) |
| 20:50 | <zcorpan> | care to share the script you use? |
| 20:51 | <rubys> | http://intertwingly.net/blog/2008/02/01/SVG-Tidy |
| 20:52 | <zcorpan> | thanks |
| 20:52 | <rubys> | it is a bzr repository if you want to check it out |
| 21:09 | <gsnedders> | hmm… This maths has lost me. |
| 21:09 | gsnedders | stabs |
| 21:11 | <Dashiva> | Hm, lastweek post about public-html mail before any replies. |
| 21:12 | <hsivonen> | wher did Larry Masinter get "walled-garden handset operating system makers" from? |
| 21:13 | <Dashiva> | Why, that must be Opera Mini |
| 21:13 | <hsivonen> | oh is iPhone OS itself a walled garden |
| 21:13 | hsivonen | was thinking about the usual mobile walled gardens |
| 21:16 | <zcorpan> | maybe lastweekinhtml5 = Larry Masinter |
| 21:21 | <Lachy> | zcorpan, who is Larry Masinter? |
| 21:22 | <zcorpan> | Lachy: see public-html |
| 21:22 | <zcorpan> | or the blog |
| 21:37 | <annevk> | seems unlikely to me |
| 21:38 | <hsivonen> | last week gives a whole new dimension te W3C werewolf games |
| 21:38 | <hsivonen> | s/te/to/ |
| 21:42 | <gsnedders> | Using integration by parts, find \int x^3\lnx dx |
| 21:42 | gsnedders | can't get the right answer for this :\ |
| 21:48 | <Dashiva> | werewolf games? |
| 21:53 | <roc> | a party game that's ultra lame yet inexplicably popular in tech circles |
| 21:54 | <Dashiva> | Oh, that silly ripoff of mafia |
| 21:55 | <hsivonen> | I'm seeing the YSoD at http://realtech.burningbird.net/and-all#comments |
| 22:00 | <zcorpan> | http://simon.html5.org/dump/daddy.svg now has the proper font |
| 22:01 | <Lachy> | zcorpan, is it intentional that "SVG SUCKS" doesn't fit within the speech bubble? |
| 22:02 | <zcorpan> | Lachy: it does if the linked font is used |
| 22:03 | <Lachy> | I tried 9.6. I'll get the latest internal build and see if that works |
| 22:03 | <roc> | works in Firefox trunk! |
| 22:04 | <Lachy> | works in opera 10 too |
| 22:04 | <hsivonen> | and WebKit |
| 22:04 | <hsivonen> | interop! |
| 22:04 | <Lachy> | will it work in Firefox 3.1? |
| 22:04 | <roc> | yeah |
| 22:05 | <Lachy> | ok. When is 3.1 expected to be released? |
| 22:05 | <annevk> | guess I have the wrong Opera 10 |
| 22:05 | <zcorpan> | webkit seems to wait with rendering the text until the font is loaded, which is seems nice |
| 22:06 | <zcorpan> | s/is // |
| 22:06 | <roc> | I dunno, a couple of months |
| 22:06 | <Lachy> | annevk, I used build 6196 Mac |
| 22:06 | <annevk> | s/seems // |
| 22:06 | <roc> | sometimes you want to wait to show the text, sometimes you don't |
| 22:06 | <annevk> | works for my menu though, hmm |
| 22:06 | <Lachy> | yeah, Opera had a flash of fallback font |
| 22:07 | <zcorpan> | firefox too |
| 22:07 | <Lachy> | which looked like Helvetica, or other similar sans-serif font. Could you pick a better fallback font? |
| 22:07 | <zcorpan> | Lachy: which one? |
| 22:07 | <annevk> | guess my Firefox 3.2a1pre is somehow out of date then as @font-face does not work |
| 22:07 | Lachy | checks Font Book |
| 22:08 | annevk | doesn't care about fallback |
| 22:08 | annevk | rather has browsers fix their bugs |
| 22:08 | <zcorpan> | annevk: will you file a bug? |
| 22:08 | <roc> | annevk: you on Linux? |
| 22:08 | <annevk> | yeah |
| 22:09 | <annevk> | and I might file a bug ;) |
| 22:09 | <roc> | ok, that build might not have it |
| 22:09 | <annevk> | k |
| 22:10 | <roc> | a later build should work |
| 22:10 | zcorpan | looks at "The heavier “Impact” sans serif stack" at http://www.sitepoint.com/article/eight-definitive-font-stacks/2/ |
| 22:10 | <annevk> | guess I'll update then |
| 22:11 | <Lachy> | annevk, is the font your using under a free licence? |
| 22:12 | <annevk> | it's a free font, haven't checked license details |
| 22:12 | Lachy | finds http://www.1001fonts.com/font_details.html?font_id=2379 - says freeware |
| 22:15 | <Lachy> | damn, IE8 doesn't do SVG :-( |
| 22:16 | <hsivonen> | conditional freeware with redistribution prohibited |
| 22:18 | <Lachy> | yeah, the licence isn't particularly good, though it's not clear whether using it as a web font is considered to be a redistribution |
| 22:19 | annevk | doesn't care |
| 22:20 | <annevk> | it's not like I abide the law when it comes digital rights anyway |
| 22:21 | <Lachy> | yeah, I know. most people don't |
| 22:29 | <zcorpan> | Lachy: provided some fallback fonts |
| 22:30 | <zcorpan> | also used text-anchor="middle" so it won't misalign |
| 22:32 | <Philip`> | nils-dagsson-moskopp⊙dn says "(No, please don't point out that I'm using Wordpress as well and vulnerable to similar things.)", so I'll be careful not to point out anything like http://blog.dieweltistgarnichtso.net/?s=%ef%bf%bf |
| 22:32 | <Lachy> | zcorpan, Impact is a good choice |
| 22:51 | <hsivonen> | http://www.w3.org/2009/01/14-xhtml-minutes.html |
| 22:51 | <hsivonen> | "markbirbeck: we could produce a document that describes the rendering behavior of the various html document types (?)" |
| 22:52 | <hsivonen> | "Steven: The HTML 5 approach to events is not scalable to the future, for example." |
| 22:53 | <annevk> | scaled pretty well for ten years |
| 22:54 | <annevk> | assuming he refers to onX |
| 22:55 | <annevk> | the namespace crowd always makes assumptions about growth that are never really backed by any research, afaict |
| 22:55 | <annevk> | just assumptions |
| 22:56 | <hsivonen> | annevk: London Gazette sure had NS growth :-) |
| 23:01 | <tantek> | ironically it is not lack of namespaces that doesn't scale, but rather the *use* of namespaces themselves as evidenced by how quickly documents get messy (hard to read and view source / copy-paste) with added namespaces. |
| 23:01 | <tantek> | theoretically, namespaces scale. in practice, they don't. |
| 23:04 | <Hixie> | i'm sure there are technical solutions to namespaces that could scale |
| 23:04 | <Hixie> | java's for example works reasonably well |
| 23:04 | <Hixie> | as i understand it |
| 23:05 | <roc> | that works because of the "import" statements at the top of the file which mean you never have to type them in the body |
| 23:06 | <annevk> | hsivonen, the other thing I find funny about that site is that they wrote a custom DTD to pass validation |
| 23:06 | <roc> | also, Java IDEs have an "organize imports" command which basically auto-generates the correct list of import statements for you and sticks it in your source files |
| 23:06 | <hsivonen> | also, Eclipse autogenerates them and maintains them, so you never have to type them anywhere |
| 23:07 | <jcranmer> | even with packages, I think name collisions are relatively rare |
| 23:07 | <jcranmer> | with the exception of some classes like "Pair" (how many people do you think reinvent this wheel?) |
| 23:07 | <annevk> | from the design doc of that site: "A <base> element needed to be added giving the identifier URL for the notice; this acts as the initial object during RDFa parsing. Relative URLs in the rest of the page needed to be updated to account for the change in base URL." sounds scary |
| 23:08 | <Hixie> | roc: yeah it's quite possible that java's mechanism isn't appropriate for html |
| 23:08 | <Hixie> | my point is just that while certainly the existing mechanisms for namespaces in web content have proved suboptimal, that doesn't mean that there isn't a better solution. |
| 23:08 | <Hixie> | now i don't know what it is |
| 23:08 | <tantek> | Hixie, also, namespaces for code and namespaces for data behave very very differently in practice, in usage, in maintenance etc. |
| 23:08 | <annevk> | "(In RDFa, the href attribute can hold a CURIE or a URL; we used CURIEs when referring to terms in known vocabularies, and URLs for more general purpose URLs because they are recognisable to non-RDFa processors.)" |
| 23:08 | <Hixie> | indeed |
| 23:08 | <Hixie> | (to tantek) |
| 23:09 | <jcranmer> | I suppose importing would be horrendous if you had a high collision rate |
| 23:09 | tantek | is always referring to namespaces for data when he criticizes namespaces in general. |
| 23:09 | <roc> | some languages deal with collisions fairly well by writing something like "import A from B as C" |
| 23:10 | <roc> | Java doesn't |
| 23:10 | <tantek> | Perhaps I need to say data:namespaces vs. code:namespaces ;) |
| 23:10 | gsnedders | probably ought to learn Java |
| 23:10 | <jcranmer> | the only collision I've ever had to deal with is java.util.List and java.awt.List |
| 23:11 | <jcranmer> | import java.awt.* and import java.util.* and List don't mix :-) |
| 23:11 | <roc> | one important difference is that Java has a compilation step that processes source text in a known environment and emits code with fully qualified names |
| 23:11 | <jcranmer> | yep |
| 23:11 | <annevk> | (from http://assets.expectnation.com/15/event/3/SemWebbing%20the%20London%20Gazette%20Paper%201.pdf ) |
| 23:11 | jcranmer | heads to supper |
| 23:11 | <roc> | so the data that gets exchanged is all fully qualified names |
| 23:12 | <roc> | HTML doesn't have this step; data transmitted exactly as authored |
| 23:12 | <hsivonen> | "We don't actually use DOM level 2 in XHTML 1.x at this time. However, |
| 23:12 | <hsivonen> | there is good text in there so I have referenced that as well. |
| 23:12 | <hsivonen> | " |
| 23:12 | <hsivonen> | http://lists.w3.org/Archives/Public/public-xhtml2/2009Jan/0041.html |
| 23:12 | <roc> | that's a key strength and weakness of HTML IMHO |
| 23:12 | <Philip`> | gsnedders: Java is a boring language, so unless you're being forced to write code in Java it'd probably be more fun to learn something else :-) |
| 23:13 | <gsnedders> | Philip`: I'm gonna be forced to learn it an uni, so I may as well get a slight clue |
| 23:13 | <jcranmer> | most interpreted languages will be compiled as the first step of being interpreted (although at a very high level, generally) |
| 23:13 | <Philip`> | gsnedders: If you learn it in advance, then it'll just be more boring when you're taught it again because you'll already know everything |
| 23:13 | <annevk> | for some reason data and code gets confused more often, e.g. people saying XML works because C compilers work |
| 23:13 | <gsnedders> | Philip`: That is true. |
| 23:14 | <jcranmer> | I prefer Java over C++ |
| 23:14 | <Philip`> | jcranmer: In terms of a productive language to write things in, or in terms of being fun to use? |
| 23:14 | <jcranmer> | probably productivity, but I'm a practical guy |
| 23:15 | <hsivonen> | Philip`: do you find C++ fun? |
| 23:15 | <roc> | I find getting things done fun |
| 23:15 | <jcranmer> | most "fun" languages tend to just be annoying to me |
| 23:16 | gsnedders | ought to re-write Anolis in a language that tends to be quicker than Python |
| 23:16 | <annevk> | gsnedders, it's quick enough, it just needs a nice serializer and new features :) |
| 23:17 | <gsnedders> | annevk: It's not that quick on the spec. |
| 23:17 | <gsnedders> | annevk: And the serializer is html5lib's fault :P |
| 23:17 | <annevk> | there are many specs that are not insanely huge |
| 23:17 | <Philip`> | hsivonen: Sometimes, when it lets you write a nice elegant efficient clever easy-to-read interface or implementation; whereas in Java you just write a load of classes and methods, which is really dull |
| 23:18 | <annevk> | gsnedders, well, html5lib has a serializer that is good for html5lib, but you need a spectree serializer, that's special |
| 23:18 | <gsnedders> | Why does it need to be different? |
| 23:19 | <Philip`> | (Of course the Java code is likely to be much more maintainable, but that's not the point) |
| 23:19 | <gsnedders> | annevk: and 40s for html5 is too slow :P |
| 23:21 | <annevk> | i suppose you could give html5lib a prettyprint serializer fla |
| 23:21 | <annevk> | g |
| 23:22 | <gsnedders> | 1.7s to do string substitutions, 1.2s to add cross-references |
| 23:22 | <gsnedders> | Too slow. |
| 23:31 | annevk | notes that Origin is currently named XXX-Origin |
| 23:34 | Philip` | assumes XXX-Origin is meant to be used to identify adult content |
| 23:40 | <zcorpan> | annevk: for completeness :) http://simon.html5.org/dump/find.svg |
| 23:46 | <annevk> | the "Adding Semantics" section of that gazette doc is interesting |
| 23:46 | <annevk> | points out a lot of (known) issues |
| 23:47 | <annevk> | and "solutions", e.g. trying to automate generation of RDFa data |
