| 00:08 | <Philip`> | Hmm, createRadialGradient is pretty much entirely non-interoperable, which looks like fun... |
| 00:13 | <Philip`> | (Oh, actually, Firefox and Safari are pretty close for at least the portion of the image which they both draw, though Opera is totally crazy. (Maybe Opera is following the spec and the spec is crazy, but I haven't tried reading it yet...)) |
| 00:16 | <Dashiva> | radial gradients are crazy to begin with, if you ask me :) |
| 00:19 | <Philip`> | I wouldn't disagree with that :-) |
| 00:20 | <Philip`> | I should have stopped after I reached linear gradients, because everything after that is hard :-( |
| 00:52 | <Dashiva> | Hmm... acid3... |
| 00:57 | <Philip`> | Aha, the spec is indeed crazy, though less so than the three incompatible implementations |
| 00:58 | Philip` | tries to guess what should happen instead |
| 01:03 | <Dashiva> | I can't find anything limiting 'element' in CSS, so :first-child and related selectors would include <script>, <style>, etc, right? |
| 01:23 | <gavin> | Is Matthew Raymond on IRC? |
| 01:27 | <zcorpan> | Dashiva: yes, any element that is in the dom tree is selectable |
| 01:27 | <zcorpan> | Dashiva: try * { display:block; margin:1em; border:solid; } |
| 01:28 | <Dashiva> | So that tricky selector seems to be checking for whether *:first-child selects an implicit head element |
| 01:28 | <zcorpan> | Dashiva: yes |
| 02:05 | <Hixie> | can someone explain the paragraph with the highlighted text in http://groups.google.com/group/alt.html/browse_thread/thread/811af37c5b2c2a96/fda33e528fee9ad6?q=whatwg&rnum=1#fda33e528 ? |
| 02:08 | <zcorpan> | Hixie: seems he observes that you can work with namespaces in the dom in browsers, and hence the text/html serialization should have a declarative namespacing mechanism |
| 02:09 | <Hixie> | is that what he's saying? |
| 02:10 | <Hixie> | i couldn't really tell |
| 02:12 | <zcorpan> | can't say for sure but that's how i read it |
| 02:16 | <zcorpan> | it could also be that he thinks that the dom that comes from html parsing doesn't know about namespaces |
| 02:17 | <Hixie> | ah |
| 02:27 | <othermaciej> | his message does seem confused |
| 02:30 | <Philip`> | (Ooh, excellent, I get totally different behaviour in Firefox 2 vs Firefox 3) |
| 02:30 | <Philip`> | (and I wouldn't be surprised if it changed on different OSes...) |
| 02:31 | <zcorpan> | Philip`: different results for what? |
| 02:32 | <Philip`> | Radial gradients in <canvas> |
| 02:32 | <zcorpan> | ok |
| 02:33 | <Dashiva> | Hixie: What's the status of the acid3 thing? Just a draft, work in progress, finished? |
| 02:33 | <Philip`> | (FF doesn't really do anything interesting with gradients, so I assume the changes are in Cairo somewhere - and I get a third behaviour when doing radial gradients with the version of Cairo I've got installed here) |
| 02:34 | <Hixie> | Dashiva: extremely early work |
| 02:34 | <Hixie> | there are no tests in it yet, to start with |
| 02:34 | <Hixie> | :-) |
| 02:35 | <Dashiva> | Well, it reveals two opera bugs already |
| 02:36 | <Hixie> | cool |
| 02:37 | <Dashiva> | But I gather I shouldn't use it for bug reference just yet |
| 02:39 | <Hixie> | well feel free to narrow down the bugs to file them |
| 02:40 | <Dashiva> | They were both already reported, to my disappointment :) |
| 02:47 | <Hixie> | :-) |
| 03:49 | <Hixie> | can anyone think of a way to programatically cause a bubbling event to fire without actually using dispatchEvent? |
| 03:50 | <Hixie> | HTMLInputElement.click() maybe |
| 03:51 | <othermaciej> | that should work |
| 03:51 | <othermaciej> | submitting a form would be another way |
| 03:52 | <othermaciej> | also DOMFocusIn / DOMFocusOut on focus changes |
| 03:54 | <Hixie> | i best avoid changing the focus or submitting a form during the test, i think :-) |
| 03:55 | <othermaciej> | well, you can cancel the form submission, perhaps in some way unlikely to fail; the click() method does seem simpler in any case |
| 03:55 | <Hixie> | yeah |
| 03:55 | <Hixie> | might do some form stuff |
| 07:12 | <Hixie> | &&= isn't valid in JS? |
| 07:24 | <othermaciej> | I don't believe so |
| 07:24 | <othermaciej> | it's definitely not in C, and I don't think JS extends the set of update operators |
| 07:25 | <Hixie> | hm |
| 07:25 | <Hixie> | so how do you do a &&= b without evaluating a twice? |
| 07:25 | <Hixie> | var x = a; x = x && b? |
| 07:27 | <othermaciej> | yes |
| 08:24 | <met_> | 2 spams http://blog.whatwg.org/charmod-norm-checking#comment-3616 http://blog.whatwg.org/feed-autodiscovery#comment-3609 |
| 08:24 | <Hixie> | do you have a blog account? |
| 08:25 | <Hixie> | we should just make you a moderator or whatever it is that has access to get rid of spam :-) |
| 08:25 | <Hixie> | Lachy should be able to hook you up |
| 08:26 | <met_> | don't have account |
| 08:26 | met_ | is just registering |
| 08:33 | <Lachy> | hey met_, I can set you up with admin rights if you like |
| 08:35 | met_ | ok |
| 08:35 | <met_> | login is met |
| 08:36 | <Lachy> | I know, I got the email when you registered |
| 08:36 | <met_> | 8-) |
| 08:36 | <Lachy> | I upgraded you to Editor. That should give you sufficient rights to moderate comments, I believe |
| 08:36 | <Lachy> | if not, let me know |
| 08:37 | <othermaciej> | Lachy: obviously the blog needs a non-WHATWG Editor |
| 08:37 | <Lachy> | lol |
| 08:37 | <Lachy> | if there are any nomineeds, I'll consider them for the role also |
| 08:37 | <met_> | see 'Delete comment' in admin, fine |
| 08:37 | <Lachy> | *nominees |
| 08:38 | <Lachy> | can you access the moderation queue? |
| 08:38 | <Lachy> | in Manage > Awaiting Moderation |
| 08:40 | <met_> | yes I see, it's empty now |
| 08:40 | <Lachy> | yep |
| 10:14 | <hsivonen> | hmm. Unicode 5.0.0 still isn't online. the ETA is in May. is it just me or wasn't the ETA supposed to be in March? |
| 10:51 | <krijnh> | http://xhtml.se/2007/04/20/im-not-arguing-im-just-making-fun-of-them/#comment-1358 - haha |
| 10:52 | <krijnh> | rel="nofollow" removed jarvklo ;) |
| 11:06 | <hsivonen> | yeah, the channel should be considered as public as the mailing list |
| 11:06 | <hsivonen> | since it is |
| 11:06 | <krijnh> | Yeah |
| 13:00 | hsivonen | learns something new about CSS namespace rules and class selectors |
| 13:00 | hsivonen | considers the interaction counterintuitive |
| 13:01 | <Lachy> | hsivonen, what did you learn? |
| 13:02 | <hsivonen> | Lachy: I learned that declaring the default namespace for element selectors breaks class-only selectors for elements from another namespace |
| 13:03 | <zcorpan> | ditto * |
| 13:04 | <zcorpan> | you need *|* (and *|*.foo) for any namespace if you have declared a default namespace |
| 13:04 | <hsivonen> | zcorpan: so it appears |
| 13:04 | <Lachy> | oh, right. That is counter-intuitive |
| 13:04 | <hsivonen> | I'm disappointed |
| 13:05 | <Lachy> | well, in a way it makes sense. By declaring the default namespace, you're basically saying you only want to select elements in that namespace unless you explicitly state otherwise |
| 13:08 | <hsivonen> | bah. now that I fixed my SVG class selectors, they broke in Prince |
| 13:09 | <hsivonen> | I guess I'll just get rid of the namespace declarations and rely on DWIM |
| 13:10 | <zcorpan> | yeah, in most cases namspace declarations in css is not necessary |
| 13:10 | <zcorpan> | (even if you used mixed namespaces in the document) |
| 13:11 | <hsivonen> | the XML API user in me feel that not declaring namespaces in wrong |
| 13:12 | <hsivonen> | *is wrong |
| 13:13 | <zcorpan> | if two elements from different namepsaces get styled but you only wanted one of them, you can always qualify the selector with a parent or descendant combinator, which might not be as elegant but this situation is very rare anyway |
| 13:14 | <Lachy> | doesn't the pragmatic and sensible user in you make you feel that declaring namespaces is pointless? |
| 13:14 | <hsivonen> | Lachy: yes |
| 13:15 | <hsivonen> | the moron and the asshole fighting in me |
| 13:15 | <hsivonen> | ok. I'm going back to namespaceless CSS |
| 13:15 | <Lachy> | lol |
| 13:15 | <hsivonen> | I guess I should file a bug about Prince, though |
| 13:30 | <hsivonen> | well, typical. I tried to do the right thing. I learned that it didn't work like I expected. I learned that it didn't help any. and I found a bug. |
| 20:37 | <hsivonen> | Hixie: do you have stats on the usage proportions of all HTML elements as opposed to just top 10? |
| 20:37 | <hsivonen> | specifically, am I right when I think that <dfn> is used very rarely? |
| 20:42 | <Hixie> | hsivonen: ask me again when i get to work (a few hours from now) |
| 20:42 | <Hixie> | (i have stats for more than the top 10) |
| 20:52 | <hsivonen> | Hixie: ok |
| 20:55 | <hsivonen> | btw, there's a thread about the legitimacy of html5 on www-html |
| 20:56 | <hsivonen> | start: http://lists.w3.org/Archives/Public/www-html/2007Apr/0053.html |
| 20:57 | <hsivonen> | the main feedback point is that Tina Holmboe does not like the semantic retrofitting of <i> |
| 20:58 | <hsivonen> | also, Lynx developers have not researched <script/> behavior thoroughly |
| 20:58 | <hsivonen> | Hixie: do you read www-html? |
| 20:58 | <Hixie> | yes |
| 20:59 | <hsivonen> | ok. no need to mention this then |
| 21:00 | <Hixie> | i saw the thread. i figured i wouldn't have any credibility from tina's point of view, so i didn't bother to reply |
| 21:01 | <Dashiva> | Poor Hixie, his life's work demolished by the mere flick of Tina's send button |
| 21:03 | <gavin_> | Tina Holmboe has some interesting opinions :) |
| 21:05 | <hsivonen> | I bothered to reply, because back when she was on ciwah, she was in the clueful minority |
| 21:17 | <zcorpan> | " No, I meant HTML 4.1. Starting with the 4.01 specification would make perfect sense; take out deprecated elements, clean it up, remove even more presentational markup ... Then release it. After that we can see." perfect plan for a high-quality and useful spec, why didn't we think of that? |
| 21:32 | <Hixie> | i tried that but when i did the "clean it up" stage it fell apart |
| 21:32 | <Hixie> | </sarcasm> |
| 21:35 | <zcorpan> | a first step could be to remove everything that isn't normative... but then the only thing left would be "UAs MUST NOT assume a default encoding" |
| 21:35 | <zcorpan> | :) |
| 21:36 | <Dashiva> | we should take the IRC approach |
| 21:36 | <Dashiva> | Treat everything as a sequence of octets, and let the clients decide on a per-case basis |
| 21:36 | <hsivonen> | Dashiva: not interoperable |
| 21:36 | <Dashiva> | octets are about as interoperable as it gets |
| 21:37 | zcorpan | thinks the xhtml2 wg would like that proposal |
| 21:37 | <hsivonen> | Dashiva: irssi is so good at guessing that it is hard for people who aren't experts to figure out what they are talking to clients with worse heuristics |
| 21:37 | <hsivonen> | Dashiva: so irssi users can talk to each other and non-irssi users complain |
| 21:38 | <hsivonen> | and the complaints look ok to the irssi users :-) |
| 21:38 | <Dashiva> | See, mIRC got recode support for utf-8 last year, so it works out somehow |
| 22:01 | <Hixie> | i'm thinking placeholder="" as an attribute to hide things |
| 22:01 | <Hixie> | <section placeholder>, <output placeholder>, that kind of thing |
| 22:07 | <Hixie> | man, the arguments in www-html are very polar |
| 22:08 | <Hixie> | i'm extremely amused by the way that people in the ivory tower say that i'm pro-presentationalism at the expense of semantics, and people in the trenches say i'm pro-semanticism at the expense of their ability to do what users want |
| 22:08 | <bewest> | yeah, there seems to be some idea that every possible semantic should have it's own tag |
| 22:09 | <bewest> | and that users somehow want to use it |
| 22:09 | <hsivonen> | "I think you and I, and I believe TBL, have a higher view in that it is to help people to communicate, but I don't think that that is what the WHATWG's constituency wants." |
| 22:09 | <othermaciej> | Hixie: Murray is in the ivory tower and he thinks you're not going far enough with presentationalism |
| 22:10 | <Hixie> | it's not clear to me what murray wants or where he stands |
| 22:10 | <Hixie> | i think he might be in the tower's subbasement |
| 22:10 | <Hixie> | or maybe he's on a tall ladder in the trenches |
| 22:10 | <Hixie> | or something |
| 22:10 | <hsivonen> | :-) |
| 22:11 | <othermaciej> | I think he built another tower on top of the tower and wrapped around to underneath the trenches |
| 22:11 | <Hixie> | hah |
| 22:11 | <othermaciej> | (damn spherical geometries) |
| 22:11 | <Hixie> | that's either a tall tower or our metaphor is in a universe with a very small word size |
| 22:11 | <hsivonen> | I got the impression that he is even more pro reflowable presentation atoms than I am |
| 22:14 | <othermaciej> | he wants an <indent> element |
| 22:14 | <othermaciej> | and an unspecified number of other weakly semantic constructs |
| 22:15 | <othermaciej> | we can always get Dan Connolly to tell him whether his elements are good or not |
| 22:15 | <hsivonen> | othermaciej: I can sympathise with <indent>, but the thread seems like a rathole. and for compatibility, <indent> has to be called <blockquote> :-) |
| 22:15 | <Hixie> | i didn't get what <indent> was supposed to do |
| 22:16 | <hsivonen> | Hixie: my understanding was that it would behave like <blockquote> on the visual media but the element name would not suggest quoting |
| 22:17 | <Hixie> | so like a <div> but with a margin-left: 40px in the UA style sheet? |
| 22:17 | <jgraham> | The problem being it doesn't work in any existing browser so no one would use it anytime soon |
| 22:17 | <Hixie> | that's certainly one problem |
| 22:17 | <hsivonen> | Hixie: yeah. something that you could map to indent/outdent buttons in Word/OOo |
| 22:18 | <Hixie> | i see |
| 22:18 | <hsivonen> | jgraham: see has to be called <blockquote> above |
| 22:18 | <Hixie> | well, it seems clear that that would be a step backwards in terms of the markup cleanup... |
| 22:18 | <jgraham> | hsivonen: Indeed |
| 22:19 | <hsivonen> | Hixie: it could be a step forward in admitting reality, though :-/ |
| 22:19 | <Hixie> | hsivonen: not really, authors rarely want 40px these days |
| 22:19 | <Hixie> | making <style> allowed in <div>s is the way to do this |
| 22:20 | <Hixie> | <div> <style> div { margin: 0 2em; border: ...; background: ...; } </style> ... </div> |
| 22:20 | <othermaciej> | <indent> seems like a poor name for an element to add extra horizontal margin on both sides |
| 22:20 | <hsivonen> | Hixie: that doesn't work with copy-paste to a document with another style sheet in a way that the indent depth adapts to the style of the target doc |
| 22:21 | <Hixie> | hsivonen: ironically, that's the opposite of the requirement i keep hearing that copy-paste should preserve the styles exactly |
| 22:21 | <Hixie> | so |
| 22:21 | <othermaciej> | only difference between <indent> and <div class="indented"> is default style |
| 22:21 | <Hixie> | go figure |
| 22:21 | <hsivonen> | Hixie: and you've got a verbose way of saying style='' there ;-) |
| 22:21 | <Hixie> | style="" isn't media-independent, nor does it support alternate stylesheets |
| 22:22 | <hsivonen> | Hixie: I want an editor with generic indent and italics but no copying of computed style |
| 22:22 | <othermaciej> | hsivonen: by that do you mean an authoring tool or a spec editor? |
| 22:23 | <othermaciej> | (because people are getting awfully specific in their requirements for HTML spec editors) |
| 22:23 | <hsivonen> | glazou says I have an orphan disease |
| 22:23 | <hsivonen> | othermaciej: authoring tool |
| 22:23 | <Hixie> | othermaciej: hah |
| 22:24 | <othermaciej> | I want an editor who will add <canvas> and <video>, but not XML data islands or namespaces in HTML |
| 22:24 | <Hixie> | hsivonen: i think editors can already do whatever they need with the current state of things, especially if it's not got to interoperate with other editors |
| 22:24 | <Hixie> | anyway |
| 22:25 | <othermaciej> | please everyone fill out this checklist of desired features before you apply to be editor |
| 22:25 | <Hixie> | gotta go to work |