| 00:01 | <karlcow> | Ice cream - 22 CAD - every flavours at http://iceshop.example.com/ |
| 00:01 | <paul_irish> | the lack of ice cream at that URL is a supreme let-down, karl. :( |
| 00:01 | <karlcow> | ;) |
| 00:02 | <karlcow> | example.com, example.org, example.net are my favorite URI space. You can define anything on them, they will be always persistent as long as they are reserved domains |
| 00:27 | <bga_> | http://demin.ws/norcpu/norcpu.html |
| 00:43 | <Hixie> | http://parsetree.validator.nu/ and http://james.html5.org/parsetree.html are both gone :-( |
| 04:30 | <Sirisian> | I have question. Is the Dom Level 3 Events implemented by any browser? |
| 04:32 | <Sirisian> | oh wait IE9 hmm |
| 04:36 | <karlcow> | http://mir.aculo.us/2011/02/08/visualizing-webkits-hardware-acceleration/ |
| 05:04 | <paul_irish> | karlcow: http://peter.sh/experiments/chromium-command-line-switches/#show-composited-layer-borders |
| 05:05 | <paul_irish> | and --show-paint-rects .. both very interesting |
| 05:52 | <hsivonen> | Hixie: https://developer.mozilla.org/en/Using_the_Mozilla_JavaScript_interface_to_XSL_Transformations |
| 08:21 | jgraham | is scared of the idea that the DOM should be unicode clean |
| 08:21 | <jgraham> | Seems like a big perf. concern for no gain |
| 08:22 | <annevk> | heh, I used to learn from books by David Flanagan and now he's using my work |
| 08:23 | <annevk> | jgraham, the DOM isn't "Unicode-clean" |
| 08:23 | <jgraham> | annevk: I know |
| 08:23 | <jgraham> | But Hixie wants it to be |
| 08:24 | <jgraham> | "That JS uses UTF-16 is a design mistake, but not one that we need to propagate |
| 08:24 | <jgraham> | to the entire platform, nor one that we need to enforce on other languages |
| 08:24 | <jgraham> | should they ever be added to the platform. As such, the DOM should be |
| 08:25 | <othermaciej> | jgraham: you got cut off there |
| 08:25 | <othermaciej> | what does "unicode clean" mean? |
| 08:25 | <annevk> | it means no surrogates |
| 08:26 | <jgraham> | othermaciej: Did I? Should end "UTF-16." |
| 08:26 | <annevk> | basically |
| 08:26 | <jgraham> | What annevk said |
| 08:26 | <othermaciej> | I don't even understand what that proposal means |
| 08:26 | <othermaciej> | jgraham: "UTF-16" doesn't seem like a logical ending to the paragraph you were quoting |
| 08:27 | <jgraham> | othermaciej: See Bug 11298 |
| 08:27 | <othermaciej> | the last bit I saw was "should they ever be added to the platform. As such, the DOM should be" |
| 08:27 | <jgraham> | "Unicode-clean, not UTF-16." |
| 08:28 | <jgraham> | It feels like sanitizing every DOMString is a huge amount of effort for no gain |
| 08:28 | <annevk> | othermaciej, basically Hixie thinks that for most "string" operations you should get characters out of the string (by replacing surrogates with U+FFFD) before doing things with the string |
| 08:29 | <othermaciej> | I see, so other than the minor problems that JS string operations don't work like that and no other DOM operation works like that, neat idea |
| 08:30 | <annevk> | xhr.send() works like that, as you need to convert the data to UTF-8 |
| 08:30 | <annevk> | but that's all I can remember |
| 08:30 | <jgraham> | Right, doing sanitization at the boundaries makes sense |
| 08:31 | <jgraham> | Doing it on every internal operation, not so much |
| 08:50 | <annevk> | jgraham, you should get james.html5.org back up |
| 08:50 | <jgraham> | Yes, I should |
| 08:50 | <annevk> | it seems at least http://james.html5.org/parsetree.html was used |
| 08:50 | <jgraham> | That needs to be updated, so I will do both |
| 08:52 | <annevk> | sweet |
| 09:05 | <hsivonen> | jgraham: I think Hixie is wrong to want the DOM to be Unicode clean |
| 09:05 | <hsivonen> | we should stop him before specs end up with too much damage |
| 09:06 | hsivonen | mumbles about the SVG API that supposedly counted Unicode characters. FAIL! |
| 09:06 | <annevk> | it's sort of similar to the i18n drive to canonicalize everything |
| 09:06 | <hsivonen> | annevk: I don't want to even think of that drive |
| 09:07 | <hsivonen> | I want to get code written today instead of spending the whole day ranting about canonicalization |
| 09:07 | <annevk> | :) |
| 09:07 | <othermaciej> | let's do something more productive, like rename URLs again |
| 09:08 | <othermaciej> | I think it needs to be called a "URU" now |
| 09:08 | <annevk> | Universal Resource Ugh |
| 09:09 | <annevk> | URB Universal Resource Bikeshed |
| 09:09 | <othermaciej> | URR Universal Resource Reference |
| 09:10 | <annevk> | Obviously you are forgetting about the thousands of implementations that handle URBs. |
| 09:10 | <annevk> | Your browser is worthless compared to my CMS. |
| 09:11 | <othermaciej> | well, URB and URR can be unified as Diversified Universal Resource Reference |
| 09:11 | <othermaciej> | as soon as the DURR spec is ready, we can start on converting all software |
| 09:12 | <annevk> | unified as diversified :) |
| 09:12 | <othermaciej> | so what ever happened to public-html-xml? |
| 09:13 | <othermaciej> | did everyone suddenly lose interest? |
| 09:13 | <annevk> | no, the next meeting is next week |
| 09:13 | <annevk> | the idea is for Norm to write up a document |
| 09:14 | <annevk> | it seems he wrote down a summary of some sort: http://norman.walsh.name/2011/02/08/html-xml |
| 09:15 | <annevk> | quite a good read actually and more or less matches what I think |
| 09:16 | <annevk> | not so much near the end though |
| 09:16 | <othermaciej> | he does rant a little about 42-- |
| 09:17 | <annevk> | I wonder when we stop comparing markup languages with programming languages |
| 09:17 | <annevk> | It obviously fails when it comes to error handling... |
| 09:17 | <othermaciej> | who does that? |
| 09:18 | <annevk> | at the end of that post he writes a bit about Perl |
| 09:18 | <annevk> | and compares it with HTML |
| 09:18 | <annevk> | roc once wrote a great email about programming languages and markup |
| 09:19 | <othermaciej> | oh, right |
| 09:19 | <annevk> | I actually have it stored in my inbox, yay! |
| 09:19 | <annevk> | http://lists.w3.org/Archives/Public/www-tag/2008Dec/0132.html |
| 09:19 | <othermaciej> | imagine if, to use a perl module, you had to bind its namespace URI to a prefix |
| 09:20 | <othermaciej> | that's a good email |
| 09:20 | <othermaciej> | I knew from the subject that it must be a reply to Larry |
| 09:32 | <hsivonen> | calling browser vendors "niche implementors" even as a rhetoric device is a pretty much in the center of how the divergence happened in the first place |
| 09:33 | <hsivonen> | because it's effectively saying that the browsable Web is just a niche thing compared to everything else |
| 09:33 | <hsivonen> | so it was OK for everything else to diverge from it |
| 09:33 | <workmad3> | hsivonen: complaints about SOAP? |
| 09:34 | <hsivonen> | workmad3: I think "everything else" in this case includes not only SOAP but also serious technical writing publication workflows and Semantic Web stuff (to the extent RDF/XML is involved) |
| 09:42 | <annevk> | last night I had a thought of checking how much of the revised XML 1.0 (note that the version number stayed the same...) is actually implemented in browser and whether bug reports had been filed |
| 09:42 | <annevk> | my hypothesis was that it was not and that no reports were filed because nobody cares about XML in browsers |
| 09:43 | <annevk> | though I believe I did file a report for Opera which I believe has no duplicates (and no interest from anyone in fixing it) |
| 09:43 | <jgraham> | The whole allowed character range thing is crazy anyway :| (that's what changed, right?) |
| 09:43 | <othermaciej> | xml 1.0 version 5? |
| 09:44 | <annevk> | othermaciej, "fifth edition" |
| 09:45 | <othermaciej> | wait, XML 1.0 5E uses the XML 1.1 character set? |
| 09:45 | <othermaciej> | what the heck was the point of 1.1 then? |
| 09:45 | <annevk> | it obsoleted XML 1.1 actually |
| 09:45 | <annevk> | because it also removed the strict version checking |
| 09:45 | <annevk> | but XML 1.1 is not rescinded still |
| 09:45 | <annevk> | probably because IBM would go mental |
| 09:46 | <othermaciej> | none of this makes any sense |
| 09:46 | <jgraham> | othermaciej: bingo, we have a winner |
| 09:46 | <key> | sup |
| 09:46 | <annevk> | and then XML people like Michael Kay complain on the WHATWG blog about how standards are a contract |
| 09:47 | <annevk> | I do think this XML 1.0 change was for the better myself by the way |
| 09:47 | <annevk> | it's a gentle first step towards XML5 |
| 09:47 | <othermaciej> | I love how they also made it an erratum to XML 1.0 4th edition |
| 09:47 | <othermaciej> | well, it's a better idea than XML 1.1 as a permanent incompatible fork |
| 09:47 | <othermaciej> | would've made more sense to do it in the first place though |
| 09:47 | <key> | what new could xml possibly need? |
| 09:47 | <key> | isn't it bloated enough by now hehe? |
| 09:55 | <hsivonen> | I thought 5th ed. didn't have the IBM thing from 1.1 |
| 09:55 | hsivonen | looks |
| 09:56 | <annevk> | hsivonen, oh, it is incompatible in that sense |
| 09:56 | <hsivonen> | annevk: http://www.w3.org/TR/xml/#NT-S |
| 09:56 | <annevk> | hsivonen, but when you see a 1.1 labeled document now you can process it per "fifth edition" rules |
| 09:56 | <hsivonen> | 5th ed. still has a sane definition of whitespace |
| 09:57 | <hsivonen> | no IBM mainframe gunk |
| 09:57 | <key> | ibm mainframe gunk? |
| 09:57 | <hsivonen> | key: for 1.1, they changed the definition of whitespace to include a character that allegedly was used in text on IBM mainframes |
| 09:58 | <hsivonen> | oh. wait. I'm looking at the worng thing in the spec |
| 09:58 | <hsivonen> | the magic happens in http://www.w3.org/TR/xml11/#sec-line-ends |
| 09:59 | <annevk> | hsivonen, the specific change I meant above with regards to replacing XML 1.1 is http://www.w3.org/TR/xml/#NT-VersionNum |
| 09:59 | <annevk> | contrast with http://www.w3.org/TR/2006/REC-xml-20060816/#NT-VersionNum |
| 09:59 | <othermaciej> | somehow, I feel like mainframes and XML are technologies meant for each other |
| 09:59 | <hsivonen> | annevk: oh, but IBM was wentioned |
| 10:00 | <hsivonen> | and http://www.w3.org/TR/xml11/#sec-line-ends has IBM written all over it |
| 10:00 | <hsivonen> | and that bit didn't make it to 5th ed. thank &Deity;! |
| 10:00 | <othermaciej> | entity not found |
| 10:01 | <annevk> | you are not well formed |
| 10:01 | <annevk> | please leave |
| 10:01 | <hsivonen> | http://lists.xml.org/archives/xml-dev/201007/msg00089.html |
| 10:07 | <jgraham> | hsivonen: The banks moving thier legacy systems sounds quite implausible. |
| 10:07 | <jgraham> | Well maybe banks would. Airlines wouldn't. But then neither would they use newfangled XML |
| 10:07 | <annevk> | always the pessimist |
| 10:07 | <annevk> | :) |
| 10:14 | <annevk> | joy: http://www.ykombinator.com/ |
| 10:14 | <annevk> | oh, and a background: http://wlppr.com/2011/02/09/html-5 |
| 10:17 | <key> | haha |
| 10:29 | <annevk> | cannot believe Norm actually thinks we want to create some barrier to entry |
| 10:30 | <annevk> | he wrote this on twitter: "I already suspect the established vendors are content to have the bar for new entrants set as high as possible." |
| 10:31 | <jgraham> | The bar for new entrants to the browser market is high because browsers are really complicated |
| 10:31 | <annevk> | it's like pretty much what we set out to fight against |
| 10:31 | <jgraham> | But we are bringing the bar down by documenting the right thing to do |
| 10:31 | <jgraham> | But maybe this is the wrong sort of barrier? |
| 10:31 | <annevk> | having to reverse engineer other browsers and that barrier being really high is why we do the things we do |
| 10:32 | <jgraham> | Possibly he doesn't mean browsers though |
| 10:32 | <annevk> | yeah he does |
| 10:32 | <annevk> | this was in the context of HTML |
| 10:32 | <annevk> | and the WHATWG in particular |
| 10:33 | <jgraham> | If you believe in all the stuff about people embedding custom vocabularies in the markup, then he could be talking about barrier to entry of defining the language semantics |
| 10:34 | <jgraham> | And be working on the assumption that special-purpose vocabularies will be processed by special purpose tools |
| 10:34 | <jgraham> | (of course it rather rests on the assumption that the additional data will be exposed via the DOM in just the same way as HTML |
| 10:35 | <jgraham> | rather than via e.g. microdata |
| 10:35 | <hsivonen> | Why is Kendall Clark using a photo of Noah Chomsky as his Twitter avatar? |
| 10:35 | <annevk> | jgraham, not given the context actually |
| 10:35 | <hsivonen> | annevk: yeah, that tweet from Norm is sad |
| 10:35 | <jgraham> | annevk: Pointer to context? |
| 10:35 | <annevk> | I replied btw |
| 10:36 | <hsivonen> | annevk: me, too |
| 10:36 | <hsivonen> | twice only |
| 10:36 | <annevk> | http://twitter.com/ndw/status/35097590135525376 |
| 10:36 | <hsivonen> | I saw you replied thrice |
| 10:36 | <annevk> | I could have said more :) |
| 10:37 | <hsivonen> | s/Noah/Noam/ |
| 10:39 | <jgraham> | Well I guess "existing vendors want to lock others out" is a easy position to take in the absence of evidence |
| 10:39 | <roc> | it's true that the barrier to new browsers is far too high |
| 10:40 | <annevk> | definitely |
| 10:41 | <annevk> | and I suppose more could be done to make it lower, but saying we are content with it being high... |
| 10:42 | <annevk> | competition is what keeps this thing alive |
| 10:43 | <annevk> | otherwise you just get another IE6 disaster |
| 10:43 | <annevk> | or NN4 |
| 10:44 | <roc> | I can imagine alternative platforms to the Web that are simpler and therefore more amenable to new entrants |
| 10:44 | <roc> | I just can't see how to get there from here |
| 11:50 | <annevk> | oh heh, the images are meant to load on http://www.fixtheweb.net/ |
| 11:59 | <AryehGregor> | What? |
| 12:00 | <annevk> | I clinked on that link and since images were taking over 10 seconds to load for me I wondered whether the site was making a point instead |
| 12:01 | <annevk> | but apparently it was just a little slow |
| 12:04 | <annevk> | seems mattur is doing some history project |
| 12:14 | <matjas> | TabAtkins: What’s up with your feed? Looks like it’s down, or is it just me? |
| 12:16 | <annevk> | whoa |
| 12:16 | <annevk> | http://lists.w3.org/Archives/Public/public-html/2011Feb/0188.html is the like the weakest objection ever |
| 12:16 | <annevk> | and also does not cite new arguments |
| 12:36 | <annevk> | can't believe mattur only has 100 followers |
| 12:37 | <annevk> | maybe the irony is hard to follow if you're not closely involved |
| 14:53 | <annevk> | http://www.w3.org/QA/2011/02/last_year_i_published_several.html look at that, HTML5 is generating money for the W3C! |
| 14:54 | <annevk> | http://www.w3.org/2011/01/w3c2011 -- W3C 2011: Select Priorities and Milestones |
| 14:55 | <annevk> | "W3C expects to advance HTML5 to Last Call in May 2011, including accessibility support for important new features." I wonder whether they argued long over that line |
| 14:58 | <annevk> | "Community Groups" thing seems somewhat interesting |
| 14:59 | <annevk> | though I guess another way of viewing that is postponing changing the W3C Process |
| 14:59 | <jgraham> | I like the fact that HTML5 gets a LC date but CSS2.1 doesn't get a dtae for Rec. |
| 15:00 | <jgraham> | Also, it says the testsuite is important but doesn't really say what they will do about making it good |
| 15:01 | <gsnedders> | jgraham: Date for REC depends upon a lot more, though |
| 15:01 | <gsnedders> | For LC only the WG needs to have consensus, a lot more needed for REC |
| 15:04 | <jgraham> | gsnedders: Given the makup of the two groups I would imagine it should be easier to get CSS2.1 to Rec. than HTML to agree on anything |
| 15:04 | <jgraham> | More pointless too, obviously |
| 15:04 | <gsnedders> | jgraham: Dunno how easy AC vote is going to be |
| 15:04 | <jgraham> | (well having a testsuite is good) |
| 15:05 | <jgraham> | (but it is good for its own sake not for having an implementation report) |
| 16:16 | <TabAtkins> | matjas: It works for me now. Try again? |
| 16:17 | <matjas> | TabAtkins: yep, works now |
| 18:43 | <karlcow> | MikeSmith: http://kyoto-client.org/ |
| 18:44 | <MikeSmith> | Kyoto Cabinet? |
| 18:44 | <MikeSmith> | not Tokyo Cabinet |
| 18:47 | <MikeSmith> | hmm, I guess it's the same developer and he end-of-lifed Tokyo a while back |
| 19:04 | <AryehGregor> | Okay, turns out WebKit's createContextualFragment() doesn't work too differently from Firefox, it was range.selectNode() that was buggy. |
| 19:05 | <hsivonen> | sigh. https://bugzilla.mozilla.org/show_bug.cgi?id=264871 |
| 19:28 | <MikeSmith> | [[ |
| 19:28 | <MikeSmith> | I now |
| 19:28 | <MikeSmith> | frequently see ISO-8859-1 content mis-identified as cyrillic (windows-1251) |
| 19:28 | <MikeSmith> | which didn't happen before |
| 19:28 | <MikeSmith> | ]] |
| 19:29 | <MikeSmith> | hsivonen: ↑ |
| 19:29 | <MikeSmith> | that don't sound so good |
| 19:30 | <hsivonen> | MikeSmith: context? |
| 19:31 | <MikeSmith> | hsivonen: most recent comment posted to https://bugzilla.mozilla.org/show_bug.cgi?id=264871 |
| 19:31 | <hsivonen> | MikeSmith: well, that's a good reason not to turn it on for everyone! |
| 19:32 | <MikeSmith> | I'd say |
| 19:34 | <MikeSmith> | hsivonen: btw, I got a short validator patch for review if you have a few minutes |
| 19:34 | <MikeSmith> | https://gist.github.com/819101#file_check_meta_refresh.diff |
| 19:35 | <MikeSmith> | and/or I can send you be e-mail |
| 19:35 | <MikeSmith> | this is for implementing a check for a change Hixie made recently |
| 19:35 | <MikeSmith> | https://gist.github.com/819101#file_check_meta_refresh.diff |
| 19:35 | <MikeSmith> | oops |
| 19:35 | <MikeSmith> | http://html5.org/tools/web-apps-tracker?from=5839&to=5840 |
| 19:36 | <MikeSmith> | to disallow the URL part of meta refresh values from starting with an apostrophe or double-quote |
| 19:36 | <MikeSmith> | and to also to report an error for any leading/trailing whitespace in the URL part |
| 19:37 | <MikeSmith> | because the IRI checker itself allows leading/trailing whitespace now |
| 19:43 | <AryehGregor> | Hixie, createContextualFragment() tests: <http://aryeh.name/spec/dom-parsing-and-serialization/test/createContextualFragment.html>; Any further ideas? (Current spec, for reference: <http://html5.org/specs/dom-parsing.html#dom-range-createcontextualfragment>;) |
| 19:43 | <AryehGregor> | The spec doesn't cover what happens if the Range's start isn't an Element. |
| 19:43 | Hixie | looks |
| 19:44 | <AryehGregor> | The spec is pretty short, so not much to test (given that I'm not testing the HTML parser proper beyond basic sanity checks). |
| 19:44 | <hsivonen> | MikeSmith: sorry, about to go to a telecon and not on a dev machine, so I eed to look at you patch later |
| 19:44 | <AryehGregor> | (which Opera fails, but I've confirmed that's a legitimate createContextualFragment() failure, it's inserting <body>s and stuff) |
| 19:44 | <MikeSmith> | hsivonen: no problem |
| 19:44 | <Hixie> | we'll have to test the fragment part of the parser at some point, but maybe not as part of this |
| 19:45 | <MikeSmith> | hsivonen: will e-mail it to you |
| 19:45 | <AryehGregor> | (also not testing XML, since I was told we aren't sure what we want to happen for XML, although the spec does specify behavior) |
| 19:45 | <hsivonen> | MikeSmith: thanks |
| 19:46 | <Hixie> | AryehGregor: test to make sure the end point of the range has no effect |
| 19:46 | <AryehGregor> | Oh, right. |
| 19:46 | <AryehGregor> | I thought of that but forgot to test it. |
| 19:47 | <Hixie> | AryehGregor: other than that, looks pretty reasonable to me. |
| 19:47 | <AryehGregor> | I mean, I remembered to test it but then later forgot again before I did it. |
| 19:47 | <Hixie> | AryehGregor: i'm not familiar enough with that api to really know where the bugs are, so i don't know what else to test :-) |
| 19:55 | <AryehGregor> | Okay, now I'll go back to writing DOM Range tests. |
| 19:56 | <AryehGregor> | That way I should get more familiar with browser bugs and not try to do silly things like use selectNode() anymore. |
| 19:56 | <Hixie> | heh |
| 19:57 | <AryehGregor> | (all the WebKit tests for createContextualRange() were failing, turns out range.selectNode(document.documentElement) actually puts the start in the Document object instead of the root element . . .) |
| 19:57 | <AryehGregor> | Actually, there's no spec for that method. |
| 19:57 | <AryehGregor> | So I'll write that first. |
| 20:02 | <Hixie> | can anyone find any ARIA on http://developer.yahoo.com/yui/examples/carousel/carousel-ariaplugin_source.html http://developer.yahoo.com/yui/examples/tabview/tabview-ariaplugin_clean.html or http://developer.yahoo.com/yui/examples/menu/menuwaiaria_source.html ? |
| 20:10 | <Peter`> | Hixie, carouselariaplugin.js seems to do that. |
| 20:29 | <Hixie> | Peter`: any idea how to turn it on? |
| 20:32 | <Peter`> | it's sniffing |
| 20:32 | <Peter`> | for Gecko >= 1.9 or IE >= 8 |
| 20:34 | <Peter`> | seems to work in a firefox beta here |
| 20:40 | <Hixie> | ah, it avoids chrome |
| 20:40 | <Hixie> | man that's lame |
| 20:43 | <Hixie> | anyone here know anything about STUN/ICE/SDP? |
| 20:47 | <MikeSmith> | http://twitter.com/#!/tmpvar/status/34356630250328064 "xml namespaces suck. In related news, only 1 test left in level2" |
| 20:47 | <MikeSmith> | from implementor of jsdom for Node |
| 20:47 | <key> | if i have a main menu, structured with a UL, and wrapped with a semantic NAV, this makes sense to me. but then how should i handle the sub menu which shows for the main menu section selected? should my main menu include a nested UL for each section? or should server side scripting only nest a UL for the current section? |
| 20:48 | <key> | http://www.pastie.org/private/z4e91xoa8dx8twu4etl2ua |
| 20:48 | <key> | version 2 |
| 20:51 | <deane> | Hi Mike |
| 20:53 | <MikeSmith> | hey deane |
| 20:54 | <key> | http://www.pastie.org/private/z4e91xoa8dx8twu4etl2ua <-- this is my question, version 2 (include all submenus whether they're active or not) or version 3 (have server only include submenu that is active)? |
| 20:54 | <deane> | sorry, haven't been online in ages |
| 20:58 | <MikeSmith> | deane: well, while you were away, somebody unfortunately summoned the Dark Lord Inglip and I reckon it's just a matter of time before he awakens to his full power and the world is destroyed |
| 20:59 | <MikeSmith> | in the mean time, there's an SVG f2f in Auckland soon |
| 20:59 | <MikeSmith> | Doug Schepers in attendance |
| 20:59 | <MikeSmith> | among others |
| 21:00 | <MikeSmith> | so good chance to meet some people while they're there |
| 21:02 | <deane> | MikeSmith: oh no, doesn't sound good mate :) |
| 21:03 | <Philip`> | http://html5.validator.nu/?doc=http://dvcs.w3.org/hg/html/raw-file/tip/tests/submission/AryehGregor/base64.html |
| 21:03 | <key> | is your name really mike smith? or a pseudonym |
| 21:03 | <Philip`> | I guess it'd be nice if it made a mention that the apparent problem is in the HTTP header |
| 21:06 | <AryehGregor> | Philip`, on the test list? |
| 21:08 | <Philip`> | ? |
| 21:09 | <Philip`> | I just mean it'd be nice if the validator's error message said where the problem was |
| 21:09 | <AryehGregor> | Oh. |
| 21:09 | <AryehGregor> | "it", not "you". |
| 21:09 | <Philip`> | Now that I think about it, I suppose it'd also be nice if the dvcs server didn't trigger that error |
| 21:10 | <AryehGregor> | I'm not even sure what the error means. |
| 21:11 | <Philip`> | It means it saw Content-Type: text/html; charset="UTF-8" |
| 21:12 | <AryehGregor> | And what did it want? |
| 21:15 | <Philip`> | charset=UTF-8 |
| 21:15 | <Philip`> | (I assume) |
| 21:16 | <AryehGregor> | So no quotes? |
| 21:16 | <key> | i'd use quotes |
| 21:16 | <key> | i always do |
| 21:17 | <TabAtkins> | Yes. |
| 21:23 | <key> | what's with people not using 6 characters for hex colors? like the #444 ? |
| 21:24 | <AryehGregor> | It's shorter. |
| 21:24 | <AryehGregor> | Equivalent to #444444. So why not use it, if you don't care much for precision? |
| 21:25 | <key> | who said i don't care much for precision? i just said i always use quotes. <shrug> |
| 21:25 | <AryehGregor> | That part was in reply to "what's with people not using 6 characters for hex colors? like the #444 ?" |
| 21:25 | <key> | oh |
| 21:25 | <key> | so why not just #4 ? |
| 21:25 | <AryehGregor> | With three characters you have less precision, but often it doesn't matter. |
| 21:25 | <AryehGregor> | That's not legal. |
| 21:25 | <AryehGregor> | It has to be three or six. |
| 21:25 | <key> | when did this happen? 3 being valid |
| 21:26 | <AryehGregor> | It was always true, as far as I know. |
| 21:26 | <key> | interesting |
| 21:26 | <hsivonen> | can someone explain to me how the failure mode seen in https://bug631751.bugzilla.mozilla.org/attachment.cgi?id=509978 in Chrome arises? |
| 21:26 | <AryehGregor> | Back to the early days of CSS. |
| 21:26 | <key> | i was in web dev a lot back in the 90s, never saw it that way |
| 21:26 | <hsivonen> | (the page is BOMless unlabeled UTF-16) |
| 21:26 | <TabAtkins> | Yes, CSS has had both 3 and 6 hexit colors since the beginning. Most applicatinos that accept hex colors understand the 3-hexit version, too. |
| 21:27 | <key> | what's hexit mean? |
| 21:27 | <hsivonen> | the failure mode in Firefox 4 and Safari 5 at least makes sense to me |
| 21:28 | <hsivonen> | is a heuristic encoding detector in Opera always enabled and not user-togglable? |
| 21:28 | <TabAtkins> | A hexit is a hexadecimal digit. |
| 21:28 | <TabAtkins> | Similar to how "bit" is a binary digit. |
| 21:29 | <hsivonen> | can someone with IE tell me how https://bug631751.bugzilla.mozilla.org/attachment.cgi?id=509978 works in IE? does IE detect UTF-16 or does it discard U+0000 ahead of tokenization? |
| 21:29 | <TabAtkins> | And "trit" for trinary digit. Not sure if there are any more common words for digits in particular bases. |
| 21:29 | hsivonen | is without IE at this time of day |
| 21:30 | <hsivonen> | unlabeled BOMless UTF-16 is evil and makes me sad |
| 21:31 | <TabAtkins> | s/unlabeled BOMless // |
| 21:31 | <TabAtkins> | s/sad/HULK ANGRY/ |
| 21:32 | <hsivonen> | TabAtkins: I was trying to be polite |
| 21:32 | <key> | whoa, i never knew bit was short for binary digit! |
| 21:32 | <key> | in all this time, damn |
| 21:32 | <key> | thanks TabAtkins! |
| 21:33 | <key> | i wonder why 7 bit bytes never took off |
| 21:34 | <key> | what was it about 8 that made it the clear winner? |
| 21:35 | <hsivonen> | key: smallest power of 2 that's enough for characters in American text |
| 21:36 | <webr3> | "power of 2" might be the overriding factor |
| 21:36 | <key> | interesting to think how the origin of 8 bit byte is so arbitrary, then |
| 21:37 | <key> | why would you need to have a byte based on power of 2? eg, why not a 7 bit byte, 9, etc |
| 21:40 | <webr3> | because computer memory is built on binary logic, so requires powers of 2? |
| 21:41 | webr3 | apologies, was watching issue-129 cp1 get sent in to a black hole |
| 21:51 | <AryehGregor> | Why does Opera always fail so many tests compared to Gecko and WebKit? http://aryeh.name/spec/dom-range/test/Range-selectNode.html |
| 21:51 | <AryehGregor> | This time it passes more tests than IE9 . . . one more test. |
| 21:53 | <jgraham> | AryehGregor: Because you are testing something we aren't very good at, possibly |
| 21:53 | <jgraham> | Or maybe we tried to copy IE rather than conform to any spec |
| 21:53 | <AryehGregor> | This is Ranges, IE didn't have them before IE9 AFAIK. |
| 21:54 | <hsivonen> | http://lists.w3.org/Archives/Public/www-rdf-comments/1998OctDec/0014.html is that a different David Singer from Apple's? |
| 21:54 | <jgraham> | Oh, well presumably we just didn't have good QA of this feature then |
| 21:54 | <AryehGregor> | Maybe you just have fewer compatibility reports than WebKit when copying Gecko stuff, so you wind up with a less correct implementation? |
| 21:54 | <jgraham> | and it hasn't caused many site-compat issues |
| 21:54 | <AryehGregor> | This is what specs and tests are for. :) |
| 21:55 | <AryehGregor> | WebKit and Gecko presumably get compat reports faster than Opera does. |
| 21:56 | <jgraham> | afiak all the DOMRange stuff predates me woking at Opera, so I'm not in the best position to comment on it |
| 21:57 | <hsivonen> | it's a shame that http://www.w3.org/Member/Meeting/98JanAC/xml-req.html is confidential |
| 21:57 | AryehGregor | can now read it since he's a member via Google and not just a fake Invited Expert, yay |
| 21:58 | <AryehGregor> | hsivonen, can it be declassified somehow? |
| 21:58 | <hsivonen> | AryehGregor: I don't know |
| 21:58 | <hsivonen> | AryehGregor: maybe there's a Process for that. maybe not |
| 22:02 | <MikeSmith> | there is not process for that, unfortunately |
| 22:02 | <MikeSmith> | no afaik at least |
| 22:02 | <MikeSmith> | that's the general downside of making anything Secret to begin with |
| 22:02 | <MikeSmith> | as with governments or whatever else |
| 22:04 | <jgraham> | Governments are actually often quite reasonable at declassifying stuff after a while |
| 22:04 | <Philip`> | webr3: Power-of-two isn't a requirement - there were early computers that had e.g. 36-bit words, which could be split into e.g. 6-bit or 9-bit bytes |
| 22:06 | <jgraham> | hsivonen: It does seem to be a different David Singer, in case you didn't already find that out |
| 22:07 | jgraham | can't even find where you asked to see if someone answered |
| 22:07 | <jgraham> | Philip`: Power-of-two not being a requirement doesn't mean its not an advantage |
| 22:10 | <MikeSmith> | jgraham: I wouldn't say that the US government at least is quite reasonable about it |
| 22:10 | <annevk> | hsivonen, if you remind me tomorrow I will finally email what I found out about our heuristics |
| 22:10 | <annevk> | hsivonen, it's all in my inbox I just never went through it and posted a reply to your question |
| 22:10 | <annevk> | (though it's not a complete answer I'm afraid) |
| 22:12 | <hsivonen> | jgraham: ok. thanks |
| 22:12 | <hsivonen> | annevk: ok, I'll ping you tomorrow |
| 22:13 | <jgraham> | http://en.wikipedia.org/wiki/System/360 claims that the IBM 7030 had variable sized bytes |
| 22:14 | <annevk> | Hixie, btw, did I read correctly you want to give up on <device>? |
| 22:14 | <jgraham> | MikeSmith: With the Freedom of Information act and declassification of old documents it may be easier to get interesting historical information out of the UK government than W3C (as a non-Member). But I could be quite wrong |
| 22:14 | <annevk> | Hixie, sounds reasonable to me; wondering if you will make the change shortly? |
| 22:17 | <jgraham> | annevk: Where? |
| 22:18 | <annevk> | jgraham, IRC logs I guess |
| 22:20 | <Hixie> | annevk: i'm waiting to find someone who can write a spec for the STUN/ICE/SDP glue before working on that spec |
| 22:27 | <karlcow> | jgraham: how old must be the documents? |
| 22:28 | <karlcow> | (before being declassified) |
| 22:29 | <jgraham> | karlcow: No idea. Maybe rather old |
| 22:30 | <Philip`> | jgraham: It appears to be similar to how modern architectures let you read a variable number of units from the address of the highest unit, which in this case is 1/2/4/8 bytes but in the 7080 is 1-8 bits |
| 22:31 | <karlcow> | in France it seems in between 50 years and 100 years but also with a right to destroy the document before that (if we can trust wikipedia) |
| 22:33 | <Philip`> | (Addresses nowadays are 32/64 bits pointing at bytes; on the 7080 it seems to be 18-bit addresses pointing at 64-bit words, plus an extra 6 address bits to point to individual bits) |
| 22:38 | <karlcow> | http://googlecode.blogspot.com/2011/02/optimizing-your-site-for-tv-is-now.html |
| 22:38 | <Philip`> | s/7080/7030/g |
| 22:39 | <jgraham> | karlcow: 30 years seems to be somewhat normal though it has gine down to 20 |
| 22:39 | <jgraham> | And of course you can opt to keep things secret |
| 22:40 | <karlcow> | yes |
| 22:40 | <karlcow> | one I would love to see public is the minutes of the 1998 workshop |
| 22:40 | <AryehGregor> | karlcow, I think in the US, you're allowed to request any information that doesn't fall under the exemptions, no matter how new. |
| 22:40 | <karlcow> | http://www.w3.org/MarkUp/future/htmlcall.html |
| 22:40 | <AryehGregor> | Of course, lots of stuff falls under the exemptions. |
| 22:42 | <karlcow> | http://www.w3.org/MarkUp/future/ |
| 22:49 | <karlcow> | hsivonen: david singer is a different one from IBM |
| 22:49 | <karlcow> | http://www.w3.org/MarkUp/future/papers/singer/im-164149.htm |
| 22:50 | <karlcow> | # Some Requirements on the future of HTML |
| 22:52 | <jgraham> | The interesting thing is seeing people identify all the right issues and then make all the wrong choices for addressing them |
| 22:56 | <AryehGregor> | Well, it's easy to identify errors in hindsight. |
| 22:56 | <othermaciej_> | I love how that list of future requirements includes |
| 22:57 | <othermaciej_> | both "XML compliance" and "Toleration of bad markup" |
| 22:58 | <MikeSmith> | Strange Case of Dr Jekyll and Mr Hyde |
| 23:02 | <othermaciej> | and then he concludes that "Toleration of bad markup" is the requirement to be rejected |
| 23:03 | <deane> | jgraham: I agree |
| 23:35 | <karlcow> | the change proposal of hixie is huge |
| 23:42 | <AryehGregor> | Hixie, should we be aiming to get DOM Range accepted as a W3C spec someplace? |
| 23:42 | <Hixie> | *shrug* |
| 23:42 | <Hixie> | so long as the browsers are on board, i don't care personally who publishes it |
| 23:43 | <AryehGregor> | Particularly since IE9 is implementing the features, and quite possibly won't look at a spec on someone's Bitbucket account. |
| 23:43 | <annevk> | yeah, we should publish at WebApps |
| 23:43 | <annevk> | Ms2ger is a member of that group now on behalf of the Mozilla Foundation |
| 23:43 | <AryehGregor> | Oh, neat. |
| 23:44 | <annevk> | so he can get it published if it is difficult for you to join |
| 23:44 | <annevk> | as I guess he's still co-editor? |
| 23:45 | <deane> | It should be published at the WHATWG site, not at the W3C |
| 23:46 | <annevk> | oh, per http://html5.org/specs/dom-range.html he's the only editor |
| 23:47 | <annevk> | deane, publishing a spec on a couple of places should not be a big deal I think |
| 23:50 | <deane> | annevk, yeah, I guess |
| 23:50 | <deane> | Hixie, There's no such thing as "The HTML syntax" or "The XHTML syntax". HTML and XHTML are variants, formats, or languages (I prefer 'variants'). We have syntaxes *for* HTML, and we have syntaxes *for* XHTML, but those syntaxes themselves are not HTML or XHTML. This is important as it has caused many problems and confusions (and arguments :) ). HTML has many 'legal' syntaxes (for better or worse), on the other hand, XHTML has just a few 'legal' |
| 23:50 | <deane> | syntaxes. |
| 23:51 | <deane> | Hixie, It could/should read: "The HTML variant" or "The XHTML variant". |
| 23:54 | <deane> | Hixie, HTML and XHTML are distinguished by parsers, *not* syntaxes. We need to be spot on with our language here. Are you able to change it please. Thanks buddy. |
| 23:54 | <key> | RFC -> http://www.pastie.org/private/z4e91xoa8dx8twu4etl2ua |
| 23:55 | <Dashiva> | The not-actually-SGML variant |
| 23:56 | <AryehGregor> | annevk, I'm sort of a de facto editor of DOM Range now. I assume I can join the Web Apps WG by asking Google to wave its magic wand, so that should be no trouble. |
| 23:57 | <annevk> | AryehGregor, you should put your name on it |
| 23:59 | <AryehGregor> | Probably. |