| 00:09 | <Hixie> | maybe i should spend more time fixing the spec and less time responding to sam and insulting the tag. |
| 00:10 | <hober> | I must have missed the TAG insult email |
| 00:10 | <Hixie> | just sent it |
| 00:17 | gsnedders | sighs |
| 00:17 | <gsnedders> | Is it silly the sub. docs is longer than xref and TOC/numbering docs put together? |
| 00:42 | <Hixie> | http://www.whatwg.org/issues/data.html now shows labels in browsers that support the html5 fillText() api |
| 00:44 | <takkaria> | damn, I have to go and download one now... |
| 01:16 | gsnedders | realises he's completely screwed up 1.0b1 spec-gen docs by having no external links |
| 01:16 | <gsnedders> | Nevertheless, go get now! |
| 01:17 | <gsnedders> | Also, if there's anyone who should be in the ack but isn't, do email me |
| 01:18 | <gsnedders> | http://hg.gsnedders.com/hgwebdir.cgi/spec-gen/rev/fab6bfa129aa (see the bzip/zip/gz links to download) |
| 01:18 | <gsnedders> | Oh dear. |
| 01:18 | <gsnedders> | The docs say 1.0b1-dev to |
| 01:18 | <gsnedders> | *too |
| 01:19 | <gsnedders> | I really am too tired :P |
| 01:44 | <Hixie> | sam is rich |
| 01:44 | <Hixie> | "please don't dismiss me" he says, after not replying to almost any of the questions i ask him |
| 01:45 | <Hixie> | like, i write an e-mail "here's how you could help us move forward: X. So far you haven't helped us move forward." |
| 01:45 | <Hixie> | and he replies "You say I haven't helped you move forward! Whine whine whine." |
| 01:45 | <takkaria> | I think that thread is a waste of your time, fwiw, and you should probably stop replying |
| 01:45 | <Hixie> | i'm gonna see if he replies to the questions i asked |
| 01:45 | <Hixie> | if he does, we could make progress |
| 01:46 | <Hixie> | if he doesn't, i'll add him to my filter that labels e-mails as being "AAA IMPORTANT/CRITICAL" |
| 01:53 | <takkaria> | Hixie, I assume you saw the posts about getting WF2 integrated? have you plans to do that soon now that we seem to have some consensus on it? |
| 01:54 | <Hixie> | yeah gonna do that after we publish next month |
| 01:55 | <takkaria> | awesome. :) |
| 03:05 | <kangax> | is it possible to modify image opacity (after it was rendered on canvas)? |
| 08:35 | <hsivonen> | apparently, browsers don't treat a bogus internal encoding decl after a BOM as an error in XML: http://upload.wikimedia.org/wikipedia/commons/3/3a/Bahia_Municip_Itapicuru.svg |
| 08:36 | <hsivonen> | nzkoz: you were looking for me? |
| 09:03 | <hsivonen> | I had dinner with friends who write software. |
| 09:04 | <hsivonen> | it seems to me that when people who have had to deal with Namespaces in XML can talk freely, they never have anecdotes about how Namespaces have helped them |
| 09:04 | <hsivonen> | instead, they have negative comments |
| 09:05 | <hsivonen> | OTOH, devil's advocate scenarios where Namespaces could help come from people who don't have to deal with Namespaces as part of their work |
| 09:15 | <othermaciej> | Namespaces are an example of the Fundamental Software Engineering Error |
| 09:15 | <othermaciej> | which is that something too terrible to actually use can be fixed by adding a level of indirection |
| 09:16 | <othermaciej> | sometimes that is true but software engineers try to do it even when it clearly is not |
| 09:23 | <hsivonen> | othermaciej: do you mean that URI-based extensibility is the too terrible thing in this case? |
| 09:23 | <othermaciej> | using URIs as a namespace identifier for tags in a markup language |
| 09:23 | <othermaciej> | is the terrible thing |
| 09:24 | <othermaciej> | if you had to mention the URI on every tag it would be clearly unusable |
| 09:24 | <othermaciej> | but since URIs are *obviously* the one true form of unique identifier, you add a level of indirection instead of rethinking why you are using them |
| 09:25 | <othermaciej> | or why URIs that are not meant to be dereferenced should start with http: and have a hostname |
| 09:38 | <hsivonen> | Hixie: if Google Translate isn't observing <code> now, why would it observe some other "do not translate" marker? |
| 09:41 | <hsivonen> | I wonder if Web authors would bother to annotate their stuff for machine translation |
| 09:42 | <Philip`> | If sometimes there are <code>s it ought to translate, it could just default to not translating and have some popup UI when you move the mouse over that text to offer to translate it |
| 09:46 | <Hixie> | hsivonen: good question |
| 09:47 | <Hixie> | hsivonen: though it wouldn't help with the wikipedia example, since that's not marked up right |
| 09:53 | hsivonen | expects a new round of Distributed Extensibility around ITS |
| 10:15 | <hsivonen> | In case anyone is wondering about Validator.nu weirdness, the DNS server that Validator.nu use for resolving addresses for outgoing connections is being really slow to respond today |
| 10:15 | <Hixie> | is it being attacked? |
| 10:15 | <Hixie> | i hear there are attacks going on now |
| 10:15 | <hsivonen> | I don't know. |
| 10:19 | <virtuelv> | are the mozilla devs still trapped in Whistler? |
| 10:19 | <virtuelv> | https://bugzilla.mozilla.org/show_bug.cgi?id=448604 |
| 10:21 | <MikeSmith> | hsivonen: I just updated my local validator.nu and now getting "Exception in thread "main" java.lang.NoClassDefFoundError: org/mortbay/jetty/Connector" error |
| 10:22 | <hsivonen> | MikeSmith: did you run build.py with target 'all' or 'dldeps'? |
| 10:22 | <MikeSmith> | with "run" |
| 10:22 | <hsivonen> | MikeSmith: try dldeps first and then run again |
| 10:22 | <MikeSmith> | k |
| 10:23 | <MikeSmith> | OK, I see it's downloading the new dependencies now |
| 10:27 | <MikeSmith> | hsivonen: btw, the dldeps can sometimes be a PITA because certain downloads often fail with "Connection reset by peer" messages, and the download doesn't retry, so I have to go back and retry it manually |
| 10:27 | <MikeSmith> | happening now with the http://download.icu-project.org/files/icu4j/4.0/icu4j-4_0.jar download |
| 10:28 | <hsivonen> | MikeSmith: yeah, I'm experiencing problems with DNS right now. It has worked until now, so there hasn't been a need to make it retry before... |
| 10:28 | <MikeSmith> | OK |
| 10:28 | <hsivonen> | but yeah, I should probably make it retry |
| 10:33 | <Hixie> | MikeSmith: the subversion link is so that people can use a svn client to get the complete log, blame, diffs, etc (it's not supposed to be accessed from a browser) |
| 10:34 | <Hixie> | MikeSmith: would it make sense to expose it as a non-hyperlinked url, maybe? |
| 10:34 | <Hixie> | MikeSmith: (other changes look fine) |
| 10:35 | <MikeSmith> | great |
| 10:35 | <MikeSmith> | Hixie: yeah, I think a non-hyperlinked "svn checkout http://svn.whatwg.org/webapps/"; would be good |
| 10:35 | <Hixie> | cool, will do that then |
| 10:35 | <MikeSmith> | thanks |
| 10:38 | <Philip`> | 'svn blame' isn't very useful, since it blames Hixie for everything |
| 10:38 | <gDashiva> | You mean it isn't all his fault? |
| 10:38 | <Hixie> | i use it a lot to track which version number a line was last edited in |
| 10:39 | <Philip`> | gDashiva: It is, but we know that already |
| 10:39 | <hsivonen> | Philip`: I deployed a new XML serializer. Feel free to try to break it. |
| 10:40 | <Philip`> | hsivonen: I might have a look when I have fewer urgent things to work on :-) |
| 10:41 | <Hixie> | MikeSmith: "choice of means" kind of sounds kooky to me so i'm changing that paragraph |
| 10:42 | <Hixie> | now it just reads: <p>There are various ways to follow the change history for the specification:</p> |
| 10:42 | <MikeSmith> | Hixie: yeah, sounded funny to me too :) I just couldn't think of better wording.. |
| 10:42 | <MikeSmith> | your revision sounds great to me |
| 10:45 | <Hixie> | i also changed your <dd><ul><li> construct to just a list of <dd>s, since <dl>s can have multiple <dd>s per <dt>s |
| 10:46 | <MikeSmith> | yeah, that's cleaner |
| 10:47 | <Hixie> | hey there's no link to the issues list either |
| 10:47 | <Hixie> | should we add taht? |
| 10:47 | <Hixie> | i guess i forgot to add it when i added it to the whatwg copy when daniel asked |
| 10:49 | <MikeSmith> | Hixie: yeah, seems like that would definitely be good to have too |
| 10:51 | <Hixie> | ok here's what i have so far: http://www.whatwg.org/specs/web-apps/current-work/.w3c/Overview.html |
| 10:52 | MikeSmith | looks now |
| 10:52 | Hixie | isn't sure he likes the text of the "HTML 5 bug/issue-tracking service" link but doesn't have a better suggestion |
| 10:52 | <Hixie> | too many capitals, numbers, and types of punctuation in short successon |
| 10:52 | <Hixie> | succession |
| 10:54 | <MikeSmith> | Hixie: yeah, that "HTML 5 bug/issue-tracking service" wording definitely klunky |
| 10:54 | <MikeSmith> | anyway, revised SOTD overall looks great |
| 10:54 | <Hixie> | how about just "our public bug tracker"? |
| 10:54 | <Hixie> | or database |
| 10:55 | <MikeSmith> | "public bug database" sound best, i think |
| 10:56 | <MikeSmith> | "submit them to our public bug database" |
| 10:56 | <Hixie> | ok: http://www.whatwg.org/specs/web-apps/current-work/.w3c/Overview.html |
| 10:56 | <Hixie> | oh you think s/using/to/? i can do that too if you want |
| 10:56 | <MikeSmith> | beautiful |
| 10:57 | <MikeSmith> | nah, "using" is fine |
| 10:57 | <Hixie> | oh i should probably update the link to the bug list to not be the list that i use but hte list that includes all the bugs i am hiding from myself too! |
| 10:57 | <Hixie> | e.g. the ones i reassign to you :-) |
| 10:57 | <MikeSmith> | heh, yeah |
| 10:58 | <hsivonen> | hmm. the resolf.conf on the validator.nu machine is interesting |
| 11:00 | <MikeSmith> | Hixie: http://www.w3.org/Bugs/Public/buglist.cgi?quicksearch=ALL+product:HTML+-status:RESOLVED+-status:CLOSED |
| 11:00 | <MikeSmith> | I think |
| 11:00 | <MikeSmith> | for the "bug database" link |
| 11:00 | <MikeSmith> | hmm, though I see that picks up the authoring-guide also |
| 11:01 | <MikeSmith> | anyway, some form that quicksearch feature would seem best |
| 11:03 | <Hixie> | http://www.w3.org/Bugs/Public/buglist.cgi?component=Spec%20bugs&component=Spec%20proposals&product=HTML%20WG&resolution=NEEDSINFO&resolution=LATER&resolution=REMIND&resolution=---&order=bugs.resolution%2Cbugs.priority%2C%20bugs.bug_severity |
| 11:04 | <Hixie> | checked in |
| 11:39 | <Lachy> | wow, so Dean Edridge might be becoming an editor? http://lists.w3.org/Archives/Public/www-archive/2008Aug/0002.html |
| 11:39 | <Lachy> | it'll be interest to see how well he manages |
| 11:53 | <hsivonen> | hmm. Jigsaw is pretty seriously vintage Java... |
| 11:56 | <hsivonen> | Namespaces (java packages) don't solve the problem of the contents of the namespace being different in 2000 and 2008 |
| 12:01 | <MikeSmith> | hsivonen: I have another question about the http://svn.versiondude.net/whattf/syntax/trunk/relaxng HTML5 schema |
| 12:03 | <MikeSmith> | does it actually capture the content-model constraints around phrasing prose/phrase/flow content? |
| 12:03 | <hsivonen> | MikeSmith: it should |
| 12:03 | <hsivonen> | subject to bugs, of course |
| 12:03 | <hsivonen> | MikeSmith: however, exclusions are handled in Schematron |
| 12:04 | <hsivonen> | and it doesn't capture the new transparent <a> thing yet |
| 12:04 | <MikeSmith> | OK. maybe I need to look at the assertions. I haven't much yet. |
| 12:04 | <hsivonen> | MikeSmith: do you have a test case that misvalidates? |
| 12:04 | <MikeSmith> | for a specific for example, where is the constraint that a <p> can't have a <ul> as a child? |
| 12:05 | hsivonen | looks |
| 12:05 | <MikeSmith> | hsivonen: <p><ul><li>foo</li></ul></p> doesn't misvalidate, but validator.nu doesn't actually seem to get to the point of validating it |
| 12:06 | <MikeSmith> | because it seems that the parser fixes it before it gets to the validation stage |
| 12:06 | <hsivonen> | MikeSmith: in block.rnc, p.inner is defined to be ( common.inner.phrase ) |
| 12:07 | <MikeSmith> | hsivonen: right, and common.inner.phrase = text & common.elem.phrase* |
| 12:07 | <MikeSmith> | and common.elem.phrase = common.elem.embedded |
| 12:07 | <hsivonen> | MikeSmith: yeah, in that case, stuff happens according to the parsing algorithm before it reaches the schema layer |
| 12:07 | <MikeSmith> | and common.elem.embedded = notAllowed |
| 12:07 | <MikeSmith> | hsivonen: right, that's what I meant for that particular case |
| 12:07 | <MikeSmith> | the </p> gets implied |
| 12:08 | <MikeSmith> | before the <ul> |
| 12:08 | <hsivonen> | that has nothing to do with the schema |
| 12:08 | <MikeSmith> | right, I understand that |
| 12:09 | <MikeSmith> | it seems like with a conformant HTML5 parser, there are many such cases |
| 12:09 | <MikeSmith> | with one consequence being that the error messages aren't going to be very helpful |
| 12:09 | <hsivonen> | but for XHTML5, the restriction is that common.elem.prose |= ul.elem does not end up augmenting common.inner.phrase |
| 12:10 | <MikeSmith> | OK |
| 12:10 | <hsivonen> | MikeSmith: there's a pending feature request to get warnings on implied tags |
| 12:10 | <MikeSmith> | ah |
| 12:10 | <MikeSmith> | that would be great to have |
| 12:11 | <MikeSmith> | ideally I think a user should see a message saying, e.g., "the <p> element cannot contain a <ul> as a child" |
| 12:11 | <MikeSmith> | or whatever |
| 12:11 | <MikeSmith> | that makes it very explicit |
| 12:12 | <hsivonen> | MikeSmith: the thing is, that omitting </p> is a legitimate way to end the <p> |
| 12:12 | <MikeSmith> | hmm, yeah, I realize that now |
| 12:13 | <MikeSmith> | god, all this stuff must make building a conformance checker a major PITA |
| 12:13 | <MikeSmith> | :) |
| 12:15 | <hsivonen> | right now, the PITA is that Jigsaw doesn't print informative diagnostics when stuff fails :-) |
| 12:16 | <MikeSmith> | I saw you had mentioned Jigsaw but I'm clueless so far about what you need it for |
| 12:16 | <MikeSmith> | what problem does it potentially solve for you? |
| 12:16 | <hsivonen> | MikeSmith: getting the W3C run an instance of Validator.nu under their preferred container |
| 12:17 | <MikeSmith> | ah |
| 12:17 | <MikeSmith> | that would definitely be really nice to have |
| 12:20 | <MikeSmith> | hsivonen: getting back to the HTML5 schema, am I confused, or is it the case that if you expand the content-model references out, common.inner.phrase just amounts to text & notAllowed |
| 12:20 | <hsivonen> | hmm. interesting. when the servlet-relative path is "/", Jigsaw gives it as null |
| 12:21 | <hsivonen> | MikeSmith: each phrase-level element definition augments that stub definition |
| 12:21 | <MikeSmith> | OK |
| 12:23 | <MikeSmith> | hsivonen: I see now... I just need to quit being lazy and to actually read the schema |
| 12:34 | <hsivonen> | do all browsers default to submitting the form to base uri if the action attribute on the form is missing? |
| 12:34 | <MikeSmith> | hsivonen: do you know of any tools that are able to generate a flattened version of an rng/rnc schema with the combine=choice definitions for a pattern actually combined into a single definition? |
| 12:35 | <hsivonen> | MikeSmith: I'm not aware of such a tool, but here's a guess |
| 12:35 | <hsivonen> | you might get that result if |
| 12:35 | <hsivonen> | you run Trang to convert the schema to RELAX NG XML syntax |
| 12:36 | <hsivonen> | and then run Kohsuke Kawaguchi's schema converter to convert the schema from RELAX NG to RELAX NG |
| 12:37 | <hsivonen> | but that's just a guess |
| 12:37 | <hsivonen> | then you could run Trang againg to compact syntax to make the result human-readable :-) |
| 12:37 | <hsivonen> | MikeSmith: Trang preserves the structure of the schema |
| 12:38 | <MikeSmith> | yeah, tried trang .. doesn't do it, unfortunately -- or fortunately, depending on how you look at it. trang faithfully preserves the RNC structure in RNG output in such a way that is seems like it's actually round-trippable |
| 12:38 | <hsivonen> | MikeSmith: Kohsuke Kawaguchi's converter builds an abstact model and reserializes it without preserving structure |
| 12:38 | <hsivonen> | but IIRC, him tool doesn't read compact syntax |
| 12:39 | <MikeSmith> | I tried Dave Tolpin's incelim and it doesn't combine them either |
| 12:39 | <hsivonen> | hence, the need to use Trang, too |
| 12:39 | <MikeSmith> | hsivonen: OK |
| 12:39 | <MikeSmith> | will try Kohsuke's tool |
| 12:40 | MikeSmith | apologizes again for not actually reading carefullywhat hsivonen wrote above |
| 12:40 | <MikeSmith> | I'll shut up now :) |
| 12:40 | <MikeSmith> | for a while at least |
| 12:46 | <hsivonen> | w00t. I got Validator.nu to run inside Jigsaw. (without file upload support, without gzip support and without non-ASCII input support) |
| 12:49 | <MikeSmith> | hsivonen: congats |
| 12:49 | <hsivonen> | MikeSmith: thanks. now I need to document what I did. :-) |
| 16:18 | <MikeSmith> | hsivonen: any clues on getting Kohsuke's rngconv working with the HTML datatype library? |
| 16:19 | <MikeSmith> | hsivonen: I'm trying to run a conversion, but I'm getting "http://whattf.org/datatype-draft"; is not a recognized data type vocabulary" |
| 16:29 | <tantek> | hsivonen, othermaciej just saw your dialog re: namespaces earlier (last night) http://krijnhoetmer.nl/irc-logs/whatwg/20080801#l-154 |
| 16:30 | <tantek> | feel free to add a new section (or sections), like "implementation experience" and/or "fundamental software engineering error" to http://microformats.org/wiki/namespaces-considered-harmful |
| 16:34 | <hsivonen> | MikeSmith: no clue. do you have the library in classpath? |
| 16:35 | <hsivonen> | tantek: I recently started a wiki page, too: http://wiki.whatwg.org/wiki/Namespace_confusion |
| 16:35 | <hsivonen> | not much there yet |
| 16:36 | <tantek> | still, a good collection |
| 16:36 | <tantek> | feel free to link to your page also from http://microformats.org/wiki/namespaces-considered-harmful |
| 16:37 | <hsivonen> | tantek: ok. I will. (gotta run now, though) |
| 16:44 | <MikeSmith> | hsivonen: yeah, I got the dist/html5-datatypes.jar subdir of my http://svn.versiondude.net/whattf/syntax/trunk/relaxng/datatype/java working directory |
| 16:44 | <MikeSmith> | it's just that one jar file, right? |
| 19:08 | hober | wonders if getting involved with this thread was a good idea after all |
| 19:35 | <takkaria> | it wasn't |
| 19:36 | <takkaria> | neither is the way trackback/pingback have been brought up at all |
| 19:42 | <hober> | maybe I should set a cron job to email http://xkcd.com/386/ to me every morning... |
| 19:45 | <takkaria> | the debate on extensibility is fundamentally a religious one, I don't see how either side will ever buckle |
| 19:46 | <takkaria> | I have this suspicion that HTML5 will never become a W3C recommendation as a result of this and other permadiscussions |
| 21:29 | <takkaria> | libxml2's APIs suck a little bit |
| 22:13 | <Lachy> | "And authors want to add metadata. Instead of forcing it into containers that haven't been designed for it (@title, @data-*), let them do it properly." -- http://lists.w3.org/Archives/Public/public-html/2008Aug/0023.html |
| 22:13 | <Lachy> | I don't get what other way would be considered the proper way to embed metadata, beyond the mechanisms designed for adding metadata?! |
| 22:15 | <Lachy> | if, as Julian claims, title and data-* weren't designed for adding some type of metadata, then I must be missing something. |
