| 03:37 | <mpilgrim> | i've completed my rel=feed research ( c.f. http://krijnhoetmer.nl/irc-logs/whatwg/20090416#l-507 ) |
| 03:38 | <mpilgrim> | i sampled 3 billion web pages from google's latest index |
| 03:38 | <mpilgrim> | weeding out errors like rel='RSS 2.0 feed' and false positives like rel='service.feed', |
| 03:38 | <mpilgrim> | i found exactly 1 page that uses rel='feed' according to specification and to the exclusion of any other autodiscovery mechanism |
| 03:38 | <mpilgrim> | http://seiji.asia/ |
| 03:39 | <mpilgrim> | and they have a visible link on their page that also links to their feed |
| 03:40 | <mpilgrim> | so there would be little harm in removing rel=feed support from the only browser that actually supports it |
| 03:40 | <mpilgrim> | and little harm in removing it from html 5 |
| 07:15 | <zcorpan> | hendry: surely <div class="inlineheader"> <span class="header"> should be <h2>? |
| 07:16 | <zcorpan> | hendry: you could use <footer> in the posts, too |
| 08:43 | <jgraham> | Hixie: http://wiki.dbpedia.org/DBpediaMobile looks like it might be an end-user application that does something with RDF |
| 08:44 | <jgraham> | (although DBpedia seems to be down at the moment since all the other demo pages are not working for me) |
| 08:45 | <Hixie> | interesting |
| 08:45 | <Hixie> | i wonder how it handles types |
| 08:45 | <Hixie> | if you have data in the wrong type |
| 08:45 | <Hixie> | does it notice? |
| 08:45 | <Hixie> | e.g. if it has a date to display but it's not a date |
| 08:45 | <Hixie> | but a string formatted as a date |
| 08:57 | <Hixie> | i am amused that xml, which has a namespacing mechanism, has the */*+xml notation for MIME types, despite not needing it (due to the namespaces), but JSON, which does need it (you could have different types of JSON data that you can only differentiate out of band) only has one MIME typpe. |
| 09:06 | <annevk5> | <div item><img itemprop="image" src="google-logo.png" alt="Google"></div> -- what if you want an itemprop=name to equal Google here? |
| 09:14 | <Philip`> | <div item><meta itemprop=image content=Google><img src=google-logo.png alt=Google></div> |
| 09:14 | <annevk5> | k |
| 09:14 | <Philip`> | (which is ugly) |
| 09:15 | <annevk5> | yeah |
| 09:15 | <Philip`> | (and doesn't work in browsers that move <meta> to <head>) |
| 09:17 | <Hixie> | but which, unlike pretty much every other solution i considered, will actually be understood! |
| 09:18 | <annevk5> | the colon restriction for itemprop does not seem necessary |
| 09:18 | <Hixie> | what colon restriction? |
| 09:18 | <annevk5> | when item has no tokens itemprop cannot contain a colon |
| 09:19 | <Hixie> | sure it can |
| 09:19 | <Hixie> | it just has to be a url |
| 09:19 | <annevk5> | ah ok |
| 09:19 | <annevk5> | just thought of that :) |
| 09:20 | <annevk5> | from what I've seen so far microdata looks nice, but it seems most companies are going with some variant of RDFa... |
| 09:21 | <Philip`> | It seems many are going with RDFa+microformats, i.e. they don't really care what the input syntax is and will just use whatever they can easily convert into their internal data structures |
| 09:21 | <Hixie> | the most recent checkin is what microdata is really about |
| 09:21 | <Hixie> | all the rest is basically noise |
| 09:36 | <jgraham> | Hixie: So you punted on escaping lone braces in the bibtex output? |
| 09:36 | <Hixie> | i was told by users of bibtex that if they were quoted there was no need |
| 09:36 | <Hixie> | so i used quotes |
| 09:37 | <Hixie> | but really someone should get tex and reverse engineer wtf the spec should say |
| 09:37 | <jgraham> | Oh, I thought we decided that didn't work |
| 09:37 | <Philip`> | I'm almost entirely certain it didn't work |
| 09:38 | <Philip`> | (BibTeX understands it fine, but LaTeX breaks when trying to parse BibTeX's output, I think) |
| 09:38 | <jgraham> | (IIRC you had to define \lonerightbrace or similar) |
| 09:41 | <jgraham> | (Maybe the easiest solution is to ask some of the mathml people for review here? Isn't David Carlisle one of the LaTeX3 people?) |
| 09:42 | <Philip`> | (We're not allowed to ask experts!) |
| 09:42 | <jgraham> | (that was sort of a rhetorical question since I just checked his homepage) |
| 09:42 | <Philip`> | (Experts are dangerous because they might know things) |
| 09:42 | <hsivonen> | Hixie: can one *input* RDF into DBpedia? |
| 09:43 | <Hixie> | jgraham: well if it needs a macro, screw that |
| 09:43 | <Hixie> | jgraham: we've already got much bigger problems with newlines in values and the like, if i'm not mistaken |
| 09:44 | <Hixie> | hsivonen: presumably |
| 09:44 | <hsivonen> | Hixie: how? |
| 09:44 | <Hixie> | no idea |
| 09:45 | <jgraham> | Hixie: Oh the newline thing could be interesting. Is it still the case that property values are taken from unnormalised text content? |
| 09:45 | <annevk5> | wasn't DBpedia created by scraping Wikipedia? |
| 09:45 | <hsivonen> | annevk5: yes. My understanding was that RDF was only the output of DBpedia, not input. |
| 09:45 | <jgraham> | annevk5: I think it is (in theory) frequently recreated by scraping wikipedia |
| 09:45 | <Hixie> | jgraham: currently, at least. i haven't looked at the feedback yet. |
| 09:46 | <annevk5> | hsivonen, mine too |
| 09:46 | <annevk5> | hsivonen, also seems that if you can create it automatically wikipedia could embed the necessary data itself |
| 09:46 | <jgraham> | http://wiki.dbpedia.org/Documentation |
| 09:48 | <Philip`> | http://www4.wiwiss.fu-berlin.de/dbpedia/dev/ontology.htm looks scary |
| 09:55 | <hsivonen> | Philip`: it's a tree, though |
| 09:56 | <hsivonen> | I'm amused that an ontology uses vague semantics like "Adult Actor" instead of clear terms like "Porn Actor" |
| 09:57 | <Philip`> | "The authors of HTML would like us to use <em> (emphasis) instead of <i> and <strong> instead of <b>. However, the frowned-upon <i> and <b> tags are more convenient to type, and have a clearer semantics (does emphasis mean italic or bold?)." (http://chaos.troll.no/~shausman/api-design/api-design.pdf) |
| 09:58 | <Hixie> | wow there's so much wrong with that i don't even know where to start |
| 09:59 | <hsivonen> | Hixie: that being the API design guide? |
| 09:59 | <Hixie> | Philip`'s quote |
| 10:00 | <Philip`> | hsivonen: They might just be reflecting the terminology used by Wikipedia (e.g. http://en.wikipedia.org/wiki/Template:Male_adult_bio) |
| 10:00 | <hsivonen> | Hixie: why? <em> and <strong> are simply longer synonyms for <i> and <b> |
| 10:01 | <Hixie> | no, they're not |
| 10:01 | <Hixie> | and the "authors of HTML" wouldn't like people to use <em> instead of <i> |
| 10:01 | <Hixie> | and <i> isn't frowned upon |
| 10:01 | <Hixie> | and it doesn't have clearer semantics |
| 10:01 | <Hixie> | and it doesn't mean italics or bold |
| 10:02 | <Hixie> | anyway, bed time for me now :-) |
| 10:02 | <Philip`> | That argument seems to be kind of ignoring reality |
| 10:03 | <jgraham> | Philip`: Hixie's argument? |
| 10:04 | Philip` | was just interested in how the quote's use of "semantics" seems to mean the opposite of what it usually means in the context of HTML |
| 10:04 | <Philip`> | jgraham: Yes |
| 10:04 | <jgraham> | Philip`: Yeah. I tend to agree with hsivonen on this point |
| 10:27 | <gsnedders> | are http://stuff.gsnedders.com/dif1.jpg and http://stuff.gsnedders.com/dif2.jpg right? |
| 10:27 | <gsnedders> | (This is maths) |
| 10:29 | <jgraham> | gsnedders: d/dx(2y dy/dx) = 2 (dy/dy)^2 d^2y/dx^2 |
| 10:29 | <jgraham> | er |
| 10:30 | <gsnedders> | Where is this? |
| 10:30 | <jgraham> | d/dx(2y dy/dx) = 2 (dy/dy)^2 + 2y d^2y/dx^2 |
| 10:30 | gsnedders | realizes his copy is still in the scanner |
| 10:30 | <jgraham> | dif2.jpg |
| 10:31 | <gsnedders> | Where does the ^2 come from? |
| 10:31 | <gsnedders> | (the first one) |
| 10:31 | <Philip`> | "dy/dy"? |
| 10:32 | <gsnedders> | That too |
| 10:32 | Philip` | assumes that should be dy/dx |
| 10:32 | gsnedders | thinks he needs a better maths teacher than jgraham :P |
| 10:32 | gsnedders | could go into school and see if any maths teacher is around |
| 10:32 | <gsnedders> | Apart from the fact that break starts now… |
| 10:32 | <jgraham> | gsnedders: You know I have other things to do right :) |
| 10:33 | <gsnedders> | jgraham: Yes :) |
| 10:33 | <Philip`> | gsnedders: d/dx(ab) = da/dx*b + a*db/dx, and you have a=2y and b=dy/dx so the first bit is d(2y)/dx*dy/dx = 2(dy/dx)^2 |
| 10:33 | <jgraham> | gsnedders: What Philip` said |
| 10:34 | <Philip`> | gsnedders: (You had d(2y)/dx = 2) |
| 10:34 | <gsnedders> | Philip`: Like, what I have on the right hand side of the page? |
| 10:35 | <jgraham> | (also: you can "check dimensions" on this sort of thing e.g. if y was a distance and x was a time then the two terms should both have units of distance^2/time^2) |
| 10:35 | <jgraham> | (but remember that d^2 y has the same dimensions as y, not as y^2) |
| 10:35 | <Philip`> | gsnedders: Yes, except without the error |
| 10:36 | <Philip`> | gsnedders: You're saying "v' = 2", which is misleading because that's dv/dy and you want dv/dx instead |
| 10:37 | <gsnedders> | Philip`: u, not v :P |
| 10:37 | <Philip`> | gsnedders: Same thing :-p |
| 10:37 | <gsnedders> | Philip`: u' = 2 dy/dx |
| 10:37 | <gsnedders> | right? |
| 10:37 | <Philip`> | Yes |
| 10:38 | gsnedders | thought he probably try and write some notes to prove he actually understands this, and blatantly more or less does, just can't actually do it :P |
| 10:39 | <gsnedders> | (So yes, that's my messy handwriting) |
| 10:39 | <Philip`> | You should have done it all in MathML |
| 10:39 | <gsnedders> | I'm not hand-wriing MathML. |
| 10:39 | <gsnedders> | *writing |
| 10:39 | <jgraham> | gsnedders: On the other one, have you written dy/dx = dx/dy*dy/dx? Becuase that seems wrong... |
| 10:40 | <gsnedders> | That's equal to 1. |
| 10:40 | <gsnedders> | WTF? |
| 10:40 | gsnedders | tries to work out what he's on about |
| 10:40 | <Philip`> | That must mean y=x |
| 10:40 | <jgraham> | (maybe you meant dy/dx = dy/dk*dk/dx or something) |
| 10:40 | <Philip`> | It's all a conspiracy! They were the same variable all along |
| 10:41 | <othermaciej_> | I wonder if there are other solutions to that differential equation |
| 10:43 | <gsnedders> | jgraham: Ah, it's because you differentiate a term with respect to y then multiple by dy/dx |
| 10:43 | <gsnedders> | jgraham: That's what I'm on about |
| 10:43 | <gsnedders> | Also, it has reached that point of the morning when the sun shines straight through the south facing window above me blinding me |
| 10:43 | <gsnedders> | Apple blatantly should make brighter laptop screens |
| 10:44 | <Philip`> | You could turn it around |
| 10:44 | <Philip`> | or use curtains |
| 10:44 | <gsnedders> | But the other side of the desk is against the wall |
| 10:44 | <gsnedders> | Curtains sound like a better idea, but that means bothering to stand up to shut them |
| 10:44 | <othermaciej> | I wish we could make laptop screens as bright as the iphone screen |
| 10:44 | <Philip`> | Put it on a flat surface (e.g. the floor) behind you rather than on the desk |
| 10:45 | <jgraham> | I wish I understood what gsnedders just said to tell him if he is talking nonsense or not |
| 10:54 | <gsnedders> | jgraham: After that comment, I went to find more books of notes… |
| 10:55 | <gsnedders> | I ought to find out if we have to hand them back… |
| 10:56 | gsnedders | notices that according to his revision timetable he has no more time for differentiation |
| 11:25 | <hsivonen> | Yay. My site is an example of a real-world site for browser development: http://khtml-konqueror.blogspot.com/2009/05/kdekdelibskhtml_8374.html |
| 11:26 | <Philip`> | The title of that blog doesn't fill me with great confidence in KHTML |
| 11:28 | <othermaciej> | I'm not sure the remaining KHTML hackers have a lot of confidence either |
| 13:04 | <Philip`> | Hmm, xmlns:* in text/html interacts badly with XMLLiterals in RDFa |
| 13:04 | <Philip`> | because one would expect XMLLiterals to be produced by serialising the DOM to XML |
| 13:04 | <Philip`> | but if it's got attributes called "xmlns:*" in no namespace, it can't be serialised to XML |
| 13:05 | <Philip`> | s/in RDFa/, in RDFa/ |
| 13:48 | <hsivonen> | how annoying that HTML5 parsing breaks Gecko reftests that incidentally rely on Gecko's HTML parsing peculiarities |
| 13:50 | <jgraham> | hsivonen: That seems weird. What fraction of the reftests are broken for that reason? |
| 13:51 | <hsivonen> | jgraham: I don't know yet |
| 13:51 | <jgraham> | OK, well do you know what fraction are broken at all? |
| 13:51 | <jgraham> | (since that would, at least, be an upper bound) |
| 13:52 | <hsivonen> | jgraham: no, I don't |
| 13:52 | <hsivonen> | (I don't know how to get the total # of tests) |
| 14:12 | <gsnedders> | hsivonen: How about the # that are broken? |
| 14:12 | <gsnedders> | (or at least fail) |
| 14:14 | <hsivonen> | gsnedders: 36 reftest failing 34 undiagnosed so far |
| 14:14 | <hsivonen> | *reftests |
| 14:14 | <gsnedders> | At least you can manually check that number in a not totally insane amount of time |
| 14:16 | Philip` | would be much more worried if replacing the parser *didn't* cause a load of tests to fail |
| 14:22 | hsivonen | notes that both Gecko and IE8 don't hoist whitespace from after </body> into body |
| 14:22 | <hsivonen> | but WebKit and Opera do hoist |
| 14:22 | <hsivonen> | I wonder if that's Web-relevant |
| 14:38 | <hsivonen> | I wonder when Google is starting to do stuff like Wonder Wheel in SVG instead of Flash |
| 14:38 | <hsivonen> | s/starting/going to start/ |
| 14:45 | <jgraham> | hsivonen: Maybe 3422 reftests? That is a really bad guess though and just comes from counting lines that don't start # in files called reftest.list using my extremely dubious bash skills |
| 14:52 | <hsivonen> | I guess reftests written solely for Gecko give a hint of how subtle dependencies there'd be if Web authors tested in only one browser |
| 14:52 | <Philip`> | And if Web authors were intentionally trying to highlight bugs in browsers |
| 14:52 | <hsivonen> | Philip`: the failing reftests so far have failed unintentionally |
| 14:53 | <hsivonen> | Philip`: i.e. they've tried to test something other than the thing that caused the failure |
| 14:53 | <jgraham> | Philip`: I was under the impression that web authors _were_ trying to highlight bugs in browsers. I mean the state of the web stretches "never attribute to malice what can be adequtely explained by incompetence" to beyond the limit. |
| 14:53 | <jgraham> | ;) |
| 14:54 | <hsivonen> | so far the failures have been stuff like lone </p> being different, whitespace nodes getting pushed around differently and <colgroup> getting inferred differently |
| 15:05 | <hsivonen> | I expect trouble with colgroup inference on sites that use <col> and don't care about testing in IE anymore |
| 16:19 | <zcorpan_> | hsivonen: my copy of minefield-html5 fails http://simon.html5.org/test/html/parsing/form/form-association.html |
| 16:20 | <zcorpan_> | hsivonen: btw i've asked before that <colgroup> parsing (and the content model) be changed to allow lone <col> |
| 16:20 | <zcorpan_> | Philip`: maybe you should point that out on the list re "one if statement" |
| 16:27 | Philip` | gets bored and tries to find out how hard it'd be to write an RDFa-for-HTML5 specification |
| 16:31 | <jgraham> | You must be *really* bored... |
| 16:32 | <Philip`> | Given how many people are implementing it, it'd be kind of nice if there was some interoperability, and currently there isn't much |
| 16:38 | <Philip`> | http://philip.html5.org/docs/rdfa/ - in how many ways is this horridly broken and/or rubbish? |
| 16:42 | <zcorpan_> | Philip`: you should join the xhtml2 wg and rewrite all their specs |
| 16:43 | <Philip`> | zcorpan_: That doesn't seem especially useful for all of them, only for ones that are being actively used :-p |
| 16:43 | <Philip`> | Also, that would be too much work and I'd get bored before then |
| 16:44 | <zcorpan_> | Philip`: you should join the xhtml2 wg and rewrite all their specs that are being actively used until you get bored |
| 16:44 | <Philip`> | The latter condition would only last for a few days, I expect |
| 16:45 | <zcorpan_> | wait, didn't you start doing this because you were bored? |
| 16:45 | <Philip`> | Yes |
| 16:46 | <Philip`> | and soon I'll think of something else to do instead |
| 16:46 | <Philip`> | (That occasion might coincidentally coincide with the release of the TF2 sniper/spy update) |
| 16:50 | <zcorpan_> | Philip`: should "valid URL" be "valid URL that is an absolute URL"? |
| 16:52 | <Philip`> | zcorpan_: Apparently not, as far as I can tell from http://www.w3.org/TR/rdfa-syntax/ |
| 16:52 | <zcorpan_> | ok |
| 16:53 | <Philip`> | It talks about resolving all URIs, and doesn't say they have to be absolute |
| 16:53 | <Philip`> | even for xmlns - "Note that it is generally considered a good idea not to use relative paths in namespace declarations, but since it is possible that an author may ignore this guidance, it is further possible that the URI obtained from a CURIE is relative. However, since all URIs must be resolved relative to [base] before being used to create triples, the use of relative paths should not have any effect on processing." |
| 16:54 | <zcorpan_> | seems silly to allow relative urls imho :) |
| 16:54 | <danbri> | the relative paths in xmlns thing was a huge war in XML land, about 9 nears ago... |
| 16:55 | <Philip`> | zcorpan_: I agree :-) |
| 16:55 | <danbri> | "seems silly to allow relative urls" is a good summary :) |
| 16:55 | <Philip`> | This isn't just about xmlns, though |
| 16:56 | <danbri> | yup |
| 16:56 | <Philip`> | It includes things like <div about="#me"> |
| 16:56 | <Philip`> | which I suppose you do want to allow relative URLs for |
| 16:56 | <danbri> | but re xml, this did blow up. http://lists.w3.org/Archives/Public/xml-uri/ http://lists.w3.org/Archives/Public/xml-uri/2000May/ |
| 16:56 | <danbri> | (...1700 messages in two months) |
| 16:57 | <Philip`> | I guess it's impossible to require anything of xmlns values, but requiring CURIEs to be absolute (after prefix processing) seems sane |
| 16:57 | <danbri> | yup |
| 16:58 | <Philip`> | danbri: Only 1700? |
| 16:58 | <danbri> | plus a few offlist i'm sure :) |
| 16:58 | <Philip`> | public-html had 1773 in *one* month :-) |
| 16:58 | <danbri> | i consider rdfa with curies to be about as fragile as Javascript ... but without the built-in "aw crap it doesn't work, what did I screw up?" corrective that comes with separating .js usage from .js declarations |
| 16:59 | <danbri> | ...so i think we need some utilities that replace that part of the lifecycle. "aww crap it doesn't work in google/yahoo" might operate the same, if their indexes updated in real time or there were better checkers... |
| 17:00 | <Philip`> | Or just an HTML validator? |
| 17:00 | <Philip`> | Hmm, someone needs to invent a real-time search engine indexer update protocol |
| 17:00 | <danbri> | there are so many ways of making technically valid stuff that's still not v useful |
| 17:01 | <danbri> | but yeah, integration with normal validation ideally |
| 17:01 | <Philip`> | Maybe you could set up an <event-source> server that the indexers connect to, so you can immediately send them all your updates |
| 17:01 | <danbri> | http://www.sitemaps.org/ is sorta related, but i guess those sitemaps are just periodically polled |
| 17:02 | <Philip`> | What kinds of problems would still be technically valid? |
| 17:02 | <Philip`> | Typos in namespace URIs is the obvious one, I suppose |
| 17:02 | <danbri> | using an unfashionable property or old style URI |
| 17:02 | <danbri> | yeah |
| 17:02 | <Philip`> | (I'm assuming the validator would be clever enough to detect references to undeclared prefixes etc) |
| 17:03 | <danbri> | eg. if someone validates using http://purl.org/dc/elements/1.0/Creator I'd like an easy way for them to learn that http://purl.org/dc/elements/1.0/creator is more widely used, and http://purl.org/dc/elements/1.1/creator even more widely used |
| 17:03 | <danbri> | (similar things hold for microformat idioms, since different constructs become fashionable over time, or are consumed by different apps) |
| 17:05 | <Philip`> | (Quite a few of the errors in http://philip.html5.org/data/rdf-namespace-status.txt are just plain typos - people find these things unusually tricky) |
| 17:07 | <danbri> | rdf-namespace-status.txt is cool |
| 17:07 | <danbri> | i had a script somewhere that tried to extract from dns/whois the time left for each popular domain |
| 17:08 | <danbri> | was quite fiddly as i recall, maybe there's a better library out there to try somewhere |
| 17:08 | <Philip`> | whois+grep? :-) |
| 17:09 | <Philip`> | Hmm, I guess they're not a very consistent format |
| 17:10 | <Philip`> | xmlns.com appears to have its whois information in YAML |
| 17:10 | <Philip`> | (via Gandi) |
| 17:11 | <danbri> | whois+grep --- yeah exactly what i was doing initially |
| 17:11 | <danbri> | then i think i found some perl api |
| 17:11 | <danbri> | yaml, cool hadn't seen that |
| 17:12 | <danbri> | so one issue is figuring out from the domain "xxx.yyy.zzz.aaa.bbb.ccc" which one to look up |
| 17:12 | <Philip`> | Hmm, good point |
| 17:12 | <danbri> | foo.co.uk you look up foo.co.uk |
| 17:12 | <danbri> | foo.co.com you look up co.com ... |
| 17:12 | <Philip`> | http://publicsuffix.org/ seems too much like hard work |
| 17:12 | <danbri> | yeah that's the site i was just thinking of |
| 17:13 | <danbri> | i guess you can crawl down and look 'em all up |
| 17:13 | <Philip`> | What did you want that data for? |
| 17:13 | <danbri> | i'd like to make a "health check" site for namespaces, ... so people have some sense of QA before they commit to using one |
| 17:13 | <Philip`> | I guess most people would leave it until the last minute before renewing their registration, and it's hard to tell the difference between that and an abandoned domain |
| 17:14 | <danbri> | 'cos it's all well and good being decentralised, but not if it's going to leave holes in the data in 10 years time cos someone's random experiment accidentally got worldwide use |
| 17:14 | <danbri> | last minute - yeah, also i'm told some registrars (eg. in Germany) don't let you register lots of years |
| 17:14 | <Philip`> | You could solve that problem by putting the RDF schema in a data: URL and use that as the namespace, instead of http:// |
| 17:15 | <danbri> | so for eg dbpedia.org seemed to be perillously close to expiration, but the owner/admin had it on some rolling auto-debit arrangement |
| 17:15 | <danbri> | heh, that'll really help with our consiseness problem :) |
| 17:16 | <Philip`> | Then you could define that namespace URI as an entity in an external DTD |
| 17:16 | <Philip`> | and import that into your document and use the entity reference |
| 17:16 | <Philip`> | which would avoid verbosity |
| 17:17 | <Philip`> | Then anyone publishing a vocabulary can put a DTD containing their schema data: URI onto their server |
| 17:17 | <Philip`> | The advantage is that you can move it onto a different host without changing the namespace that your data uses |
| 17:18 | <danbri> | heh |
| 17:18 | <Philip`> | The disadvantage is that it's clearly a stupid idea |
| 17:18 | <danbri> | how could the schema reference the properties defined within it? |
| 17:18 | <danbri> | btw http://prefetch.net/code/domain-check looks handy |
| 17:19 | <Philip`> | I guess it could still reference them with "#" |
| 17:19 | <Philip`> | If it wants to reference anything in another schema, it can just import that schema's DTD |
| 17:20 | <danbri> | it's mad but worth writing up! |
| 17:21 | <Philip`> | Also it ties in nicely to the discussion on the WHATWG list about having browsers download external DTDs |
| 17:21 | danbri | gotta run (eats...) |
| 18:14 | <Philip`> | Why does http://validator.nu/?doc=http://philip.html5.org/docs/rdfa/ not like my <ins>es and <del>s? |
| 18:34 | <Philip`> | http://www.robweir.com/blog/2009/05/battle-for-odf-interoperability.html - "Standards are written documents -- text -- and as such they require interpretation. [...] When interpreting a document using an purposive view, you look at the purpose, or intent, of a document in its full context, and interpret the text harmoniously with that intent." - sounds like an excuse for writing vague specs |
| 20:04 | <gsnedders> | Philip`: A valid CURIE should probably have to have the prefix it uses defined |
| 20:04 | <Philip`> | gsnedders: My thing already requires that |
| 20:05 | <Philip`> | at least in the sense that it's got a big red box saying it should require that |
| 20:05 | <Philip`> | unless I missed something |
| 20:05 | <Philip`> | which is quite possible |
| 20:05 | <gsnedders> | Oh, no. |
| 20:05 | <gsnedders> | I just can't read. |
| 20:05 | <Philip`> | Hmm, maybe I should list "ability to read" in the prerequisites for reading the document |
| 20:15 | <jgraham> | Speaking of the public suffix list, I should hunt down gerv |
| 20:19 | <Philip`> | Hunt him down and kill him? |
| 20:23 | <jgraham> | Perhaps |
| 20:38 | <Traveler0> | can anyone help me with hymns requiem? |
| 20:40 | <Traveler0> | currently using 1.8.7 and itunes 8.0.2.2 and ever time i try and decrypt a file it just says "failed" |
| 20:44 | jgraham | wonders how Traveler0 ended up in this channel |
| 20:45 | <jgraham> | Traveler0: In general this seems like an unlikely place to get help for that particular problem. The main Requium user here isn't around at the moment |
| 20:47 | <annevk5> | the logs have a pagerank that is too high |
| 20:47 | <annevk5> | at least for certain topics :) |
| 20:48 | <jgraham> | annevk5: I assumed that but I couldn't reverse engineer the right search terms :) |
| 20:52 | <Traveler0> | jgraham: yes i found this from google |
| 20:53 | <Traveler0> | when does he usually chat? |
| 20:57 | <jgraham> | Traveler0: He's not really a hardcore user or anything. He jsut mentions it sometimes :) And I don't really know when he will be around, afaik he is away at the moment |
| 21:01 | <Traveler0> | from what the logs say Lachy is the one talking about it |
| 21:46 | <gsnedders> | Traveler0: He's a bit of a loner really, talking to nobody but himself about it :P |
| 21:49 | <Traveler0> | gsnedders, well i'd really like to have a few words with him |
| 21:50 | <Traveler0> | do i have is correct nick? |
| 21:53 | <gsnedders> | Traveler0: Yes, but I doubt he can really help |
| 21:54 | <gsnedders> | Whenever I send a txt I get odd sounds from the speakers here |
| 21:55 | <Traveler0> | gsnedders i get crazy feed back to for some reason |
| 22:15 | <roc> | hsivonen: I'd guess that reftests actually would under-estimate problems found on the real Web |
| 22:15 | <roc> | since most of the people who write reftests are quite clueful |
| 22:16 | <roc> | if I do say so myself |
| 22:16 | <roc> | you can easily get the total number of reftests ... just run them, it's in the window title |
| 22:16 | <Traveler0> | what is a reftests ? |
| 22:17 | <roc> | then again, in the context of reftests, a "problem" means "does not render pixel-accurately the same as the reference page" |
| 22:17 | <roc> | which is more strict than really matters on the Web |
| 22:17 | <roc> | http://weblogs.mozillazine.org/roc/archives/2008/12/reftests.html |
| 22:36 | <Hixie> | i love when people say "why can't we use prefixes? people understand css!" |
| 22:36 | <Hixie> | leaving quite alone the problem with ignoring evidence (people fail to understand prefixes, so that's more relevant than css), the sad fact is people don't understand css either |
| 22:36 | <Hixie> | most stylesheets just do .foo { properties... } |
| 22:37 | <Hixie> | with no more compelx selectors than that |
| 22:40 | <zcorpan> | Hixie: that's still indirection, though. using style="" is without indirection |
| 22:41 | <inimino> | I think the main problem is that prefixes look meaningful, but aren't supposed to be |
| 22:42 | <annevk5> | interesting point |
| 22:51 | <Traveler0> | Hixie, do you use requiem? |
| 23:15 | <zcorpan> | http://www.w3.org/TR/html-design-principles/#degrade-gracefully s/irrelevant/hidden/ |
| 23:20 | <annevk5> | it's not like someone is actually maintaining that document |
| 23:53 | <Hixie> | http://www.w3.org/mid/7B28DCC4-BFAB-4BC3-9C96-7B37BF196349⊙gc |
