| 00:05 | <Hixie> | apparently |
| 00:06 | <Hixie> | i just wish someone would tell me wtf the requirements are and wtf the problem is we're solving |
| 00:06 | <Hixie> | all i hear is "we solved this problem years ago, there's been lots of research, rdfa is the answer" |
| 00:06 | <Hixie> | but i don't even know what the question is! |
| 00:07 | <Hixie> | surely a url to the problem description would be a simple thing to provide if it's such a solved problem |
| 00:56 | <nessy> | rdf is about solving the problem of machine-readable interoperable metadata, FAIK |
| 00:57 | <nessy> | rdfa is a particular implementation of rdf through attributes, which I think is stupid |
| 00:57 | <nessy> | but don't quote me on this :) |
| 00:57 | <webben> | nessy: I don't think that's the sort of problem statement Hixie is looking for. |
| 00:57 | <nessy> | I think it should be done through class names |
| 00:57 | <nessy> | machine-readable interoperable metadata is not a problem? |
| 00:58 | <webben> | nessy: "machine-readable interopable metadata" is a vague solution to (set of) user and author problems. |
| 00:58 | <webben> | nessy: Hixie's email is effectively seeking a catalogue of those problems, I think. |
| 00:58 | <webben> | (or at least of examples of said problems) |
| 00:58 | <nessy> | fair enough :) |
| 00:59 | <webben> | nessy: Hence why the discussion has partly revolved around Natural Language Processing as an alternative. |
| 00:59 | <nessy> | http://www.w3.org/TR/xhtml-rdfa-scenarios/ |
| 01:00 | <nessy> | has an example of 9 use cases |
| 01:00 | <nessy> | it's sort-of asking what is the use case for html |
| 01:01 | <nessy> | kinda difficult to destill the essence |
| 01:01 | <nessy> | anyway - i don't really want to get into the semantic web discussion - better to have that discussion with some semantic web enthousiasts |
| 10:42 | <Hixie> | ooh, the first kickoff was this weekend |
| 10:42 | Hixie | signs up to ref at the silicon valley regional again |
| 10:49 | <olliej_> | Hixie: ? |
| 10:49 | <Hixie> | usfirst.org |
| 10:49 | <Hixie> | though the "us" is a lie |
| 10:49 | <Hixie> | it's international |
| 10:49 | <Hixie> | (though still heavily us-skewed currently) |
| 11:08 | <olliej_> | ah |
| 11:08 | <olliej_> | i see |
| 11:32 | <cyclist> | It seems that if I have a <div> inside a <pre>, the "newline removal" applies for the inner div as well. |
| 11:33 | <cyclist> | for example: <pre> \n <div> \n foo \n </div> \n </pre> |
| 11:34 | <cyclist> | If you set backgrounds to the div and the pre blocks then the <div> occupies |
| 11:34 | <cyclist> | two lines. |
| 11:37 | <gpy> | anyone have spotify invites left? |
| 11:55 | <cyclist> | ok this is really weird: |
| 11:56 | <cyclist> | <pre> \n <div> </div> \n </pre> |
| 11:56 | <cyclist> | is exactly the same as: |
| 11:56 | <cyclist> | <pre> \n <div> \n </div> \n </pre> |
| 11:57 | <cyclist> | (everything without any text, just boxes with background) |
| 11:57 | <cyclist> | but this is hilarious: |
| 11:57 | <cyclist> | <pre> <div> </div> </pre> |
| 11:57 | <cyclist> | without newlines actually occupies 3 lines! |
| 11:58 | <cyclist> | is that part of the standard? |
| 13:21 | <hsivonen> | google sure doesn't make it easy to file bugs on their services |
| 13:34 | <annevk> | what problem descriptions for the RDF thread have been tried yet? |
| 13:35 | <annevk> | is "being able to add 'arbitrary' metadata to HTML content so that machines can understand what is going on" accurate? |
| 13:36 | <annevk> | maybe I shouldn't try |
| 13:39 | <hsivonen> | http://intertwingly.net/blog/2009/01/03/iCalendar-Validator#c1231033642 |
| 13:43 | <annevk> | writing good specs is hard |
| 13:43 | <annevk> | and not just because of the politics |
| 13:59 | <olliej_> | annevk: heh |
| 14:13 | <jcranmer> | I'd say this Charles person is doing a lot of explanation of what RDF will solve |
| 14:13 | <jcranmer> | just not why we care |
| 14:14 | <jcranmer> | he also does a lot of "well, the answer is obvious" |
| 14:15 | <jcranmer> | after being asked the question because the answer was, well, not obvious |
| 14:38 | <hsivonen> | Hixie: wiki.whatwg.org has gone 500 |
| 14:56 | <hsivonen> | Hixie: the spec annotation script connects to a server that has gone 500, too |
| 14:57 | <hsivonen> | why is http://www.whatwg.org/specs/web-apps/current-work/#attr-meta-http-equiv-default-style in the spec? |
| 15:02 | <Philip`> | hsivonen: That's why I just complain about Google service bugs on IRC |
| 15:03 | <hsivonen> | Philip`: ok let's try that |
| 15:04 | <hsivonen> | http://code.google.com/p/jing-trang/source/browse/ has gone 500 |
| 15:04 | <hsivonen> | (checkout fails with Reference to non-existent node '/.37722223df817abf.519887ccaff96edb.6c29f3f04d5d1dee' in filesystem 'jing-trang') |
| 15:05 | <Philip`> | (http://jing-trang.googlecode.com/svn/trunk/ fails with that too) |
| 17:30 | <rubys> | http://wiki.whatwg.org/wiki/MicrosyntaxDescriptions |
| 17:31 | <rubys> | wiki's down |
| 17:36 | <annevk> | forums as well |
| 17:36 | <annevk> | maybe CPU quota limit has been hit, Hixie? |
| 17:42 | <Philip`> | If the wiki is down, how can you validate a page that uses rel values defined therein? |
| 17:43 | <Dashiva> | It becomes a human-verified conformance criteron :) |
| 17:44 | <Philip`> | How can the human verify it, if they can't access the wiki? :-) |
| 17:44 | <Dashiva> | The human can go and visit the data center |
| 17:46 | <rubys> | The reason why I ask is that apparently I can't build hsivonen's checker without access to that page. |
| 17:52 | <annevk> | maybe get a copy from Google's cache? |
| 17:54 | <rubys> | this is the fourth thing fetched by the build process that doesn't resolve. The second was from google code: http://code.google.com/p/jing-trang/source/browse/ |
| 17:54 | <rubys> | I mean, if you can't trust google to be up, who can you trust? :-) |
| 17:59 | <annevk> | hsivonen was complaining about that earlier as well |
| 17:59 | <Dashiva> | rubys: windows update! |
| 18:00 | <rubys> | Dashiva: doesn't help me much, I'm running Ubuntu :-) |
| 18:01 | <Philip`> | You can only trust your own machine to be up |
| 18:02 | <Philip`> | so what you should do is download all the dependencies to your own computer, and then you'll never have problems accessing them |
| 18:03 | <rubys> | I also tend to prefer fewer dependencies than Henri seems comfortable with. |
| 18:03 | <Dashiva> | Maybe he reckons people won't find his site if google is down :) |
| 18:04 | <annevk> | dependency on wiki.whatwg.org should be fine, but it should not be fatal... |
| 18:05 | <rubys> | Exception in thread "main" java.lang.ExceptionInInitializerError |
| 18:05 | <rubys> | at nu.validator.servlet.VerifierServlet.<clinit>(VerifierServlet.java:85) |
| 18:05 | <rubys> | at nu.validator.servlet.Main.main(Main.java:89) |
| 18:06 | <rubys> | wiki is up |
| 18:14 | <annevk> | not really |
| 18:15 | <annevk> | though sometimes I do get through, seems like the CPU quota thing |
| 19:13 | <hsivonen> | rubys: right now, I'm not comfortable with the wiki deps, either. :-( |
| 19:14 | <hsivonen> | Dashiva: the wiki dependency is an app startup time dependency--the wiki is not consulted later at run time |
| 19:14 | <hsivonen> | I guess the wiki deps need the same treatment as the spec |
| 19:14 | <hsivonen> | the spec is not checked into the svn of the validator itself |
| 19:15 | <rubys> | how does the validator depend on the spec? |
| 19:17 | <rubys> | btw, I have the checker working. I needed one more dependency, but as the dependency had a file name with 2004 in it, I figured I could get that from where I had gotten the checker working last year. |
| 19:20 | <rubys> | it looks like (a) this checker build process checks out all of the files necessary for the cpp generation, so I can get started with what I have got, and (b) unlike what I remember, the checker build process can be used to update an existing build. |
| 19:21 | <hsivonen> | rubys: the spec is mined for UI strings |
| 19:22 | <hsivonen> | rubys: if you run python build/build.py --help, it lists command line options for overriding the wiki URIs |
| 19:22 | <hsivonen> | hmm. I have forgotten to document some options |
| 19:23 | <hsivonen> | --microsyntax=http://wiki.whatwg.org/wiki/MicrosyntaxDescriptions |
| 19:24 | <hsivonen> | actually, it's worse than just not documenting stuff. the build script can't override the alt advice URI |
| 19:25 | <hsivonen> | given how easy it is for remote resources to break the validator.nu build script, I'm not very optimistic about the idea of UAs fetching namespace URIs at runtime to load OWL |
| 19:32 | <annevk> | hsivonen, didn't DTDs tell us that wasn't going to work, ever? |
| 19:33 | <hsivonen> | annevk: I suppose |
| 19:34 | annevk | remembers some outcry over the Netscape RSS DTD going dead |
| 19:34 | <rubys> | this stuff should be cached, and tags with ETAGs |
| 19:34 | <rubys> | s/tags/tagged/ |
| 19:35 | <hsivonen> | rubys: do you mean the cvsdude repo should seed the cache? |
| 19:35 | <annevk> | can't the validator SVN carry a copy of the page that's updated from time to time and used when the wiki page is not available? |
| 19:36 | <hsivonen> | annevk: yes |
| 19:36 | <rubys> | and furthermore, running instances should update their own private cache when they can successfully fetch the wiki page |
| 19:37 | <hsivonen> | rubys: fwiw, the main reason why jars aren't checked in is that I don't want to deal with all the license compliance stuff that would come with distributing the binaries |
| 19:37 | rubys | nods |
| 20:29 | <Dashiva> | And as we're talking about how last week has gone downhill, it plunges into the swamp.... |
| 20:29 | <gsnedders> | Yeah, it really isn't very good any more. |
| 20:29 | <gsnedders> | Damned Mr Last Week. |
| 20:30 | <Philip`> | Dashiva: I think you should set up your own Last Week-like blog |
| 20:30 | <gsnedders> | And call it Next Week in HTML5! |
| 20:31 | <gsnedders> | I mean, if you're a teamster, you should be able to know what's gonna happen! |
| 20:31 | <Philip`> | Backdate all the blog posts, and nobody will ever realise that you can't actually read the future |
| 20:32 | <Dashiva> | Last week on last week |
| 21:56 | annevk | thought the bit about chaals was funny in the latest lastweek |
| 21:59 | <annevk> | http://robertbrook.com/html-5-validation-and-openid-validation/ |
| 22:00 | <Philip`> | People really like using trailing slashes in HTML5 |
| 22:00 | <annevk> | http://adactio.com/journal/1540 mentions a bunch of HTML5 articles I missed |
| 22:01 | <annevk> | e.g. http://forabeautifulweb.com/blog/about/more_on_developing_naming_conventions_microformats_and_html5/ and http://jontangerine.com/log/2008/03/preparing-for-html5-with-semantic-class-names and http://boblet.tumblr.com/post/60552152/html5 |
| 22:03 | <annevk> | Philip`, he uses both, no? |
| 22:03 | annevk | thinks part of his stuff is leftover from WordPress |
| 23:33 | <BenMillard> | wait a minute...when did classitis become a good thing? |
| 23:35 | <BenMillard> | hmm, <p><pre><code>...</code></pre></p> |
| 23:35 | gsnedders | throws parse error on the </p> |
| 23:36 | annevk | returns <p></p><pre><code>...</code></pre><p></p> |
| 23:36 | <gsnedders> | Yeah, that's what I got too in my non-draconian mode. |
| 23:36 | Hixie | brings his CPU back up at dreamhost |
| 23:37 | <Hixie> | i'd tried reducing it over the holidays |
| 23:37 | <Hixie> | but apparently that was a mistake |
| 23:37 | <gsnedders> | That obviously didn't entirely work. |
| 23:37 | <BenMillard> | all the focus outlines are turned off: http://forabeautifulweb.com/demo/2008/11/15/index.html |
| 23:37 | <BenMillard> | that's rather sad |
| 23:38 | <BenMillard> | funny how things which actually help users, like focus outlines, are being turned off when things that don't help anyone, like classitis, are becoming vogue |
| 23:38 | <BenMillard> | although I wholeheartedly approve of his intention style :P |
| 23:39 | <annevk> | Hixie, does it take some time to go into effect? |
| 23:39 | annevk | still gets 500 on the wiki |
| 23:39 | <BenMillard> | s/intention/indention/ |
| 23:39 | <Hixie> | oh the problem is memory not cpu |
| 23:40 | <Hixie> | there |
| 23:40 | <Hixie> | should work now |
| 23:43 | <annevk> | http://forums.whatwg.org/viewtopic.php?t=4037 |
| 23:52 | <Hixie> | if anyone wants to reply, onerror is actually in html5 already |
| 23:53 | <Hixie> | the callstack sounds like something to ask the ecmascript guys though |