| 03:09 | <Hixie> | the level to which DOM2 HTML is underspecified is insane |
| 03:20 | <othermaciej> | is it worse than DOM Core? |
| 03:25 | <Hixie> | to some extent yes |
| 03:25 | <Hixie> | e.g. try to work out what createTHead() does |
| 03:26 | <Hixie> | or what happens when you set table.caption |
| 03:27 | <othermaciej> | no thanks |
| 03:41 | <hober> | Would you consider <legend> to be block-level or inline-level? (Its content model is inline, but what of the element itself?) |
| 04:57 | <a-ja> | anyone here have a clue why an <a> element wouldn't be allowed inside a <details> ? |
| 07:54 | <annevk> | a-ja, probably because <details> is interactive at the moment |
| 07:54 | <annevk> | a-ja, I think it should be allowed though |
| 07:55 | <a-ja> | annevk: seems to take every other strictly inline i've tried |
| 07:56 | <annevk> | I'm not sure how that negates my point |
| 07:56 | <a-ja> | annevk: e.g. img, span, meter |
| 08:01 | <a-ja> | annevk: sure would be handy having hide/show ToC's in a <details>. in any event, if <a>'s not gonna be allowed, spec needs to say so |
| 08:03 | <annevk> | It already does |
| 08:03 | <annevk> | If you look at the definition of <a> you see: "Where strictly inline-level content is allowed, if there are no ancestor interactive elements." |
| 08:03 | <annevk> | And for <details> you can see that it's an interactive element |
| 08:04 | <annevk> | I don't think this restriction makes much sense, but the conformance checker is correct |
| 08:04 | <a-ja> | ah....links are considered interactive, eh? |
| 08:04 | <annevk> | yes |
| 08:14 | <a-ja> | hmm...that'll make <details> useless for things like hiding/showing blog comments/feedback, too since they're likely to contain links :( |
| 08:15 | <annevk> | would you classify those as "details" though? |
| 08:16 | <a-ja> | yeah,,,that might be iffy |
| 08:16 | <annevk> | but the interactive stuff need sto be fixed, I think Hixie is aware of that (it also affects <datalist> for instance), but you might want to re-raise the point on the list |
| 08:17 | <a-ja> | toggling post's full text vs. just a summary would be nice, too |
| 08:17 | <a-ja> | better on whatwg or html-wg, you think? |
| 08:18 | <annevk> | whatwg |
| 08:18 | a-ja | is not a member of either, but has been considering it |
| 08:18 | <annevk> | public-html hasn't adopted the spec yet |
| 08:18 | <annevk> | and you should probably join at least one of them |
| 08:20 | <a-ja> | still not any "no" votes |
| 08:23 | <annevk> | do you have a pointer btw? |
| 08:24 | <a-ja> | to signup info? remember seeing html-wg signup info at hixie's blog, wasn't it? |
| 08:24 | <annevk> | no, to the voting proceedings |
| 08:25 | <a-ja> | http://www.w3.org/2002/09/wbs/40318/htmlbg/results |
| 08:26 | <annevk> | cool, thanks |
| 08:26 | annevk | couldn't find it on the homepage |
| 08:27 | <a-ja> | heh....compact view of the results is non-public |
| 08:29 | <a-ja> | someone in moz #developers gave me the link the other day....never saw it anywhere else |
| 08:30 | <om_afk> | <details> should definitely allow things that are normally not allowed in interactive elements since they are themselves interactive |
| 08:30 | <om_afk> | IMO |
| 08:31 | <annevk> | agreed |
| 08:33 | <othermaciej> | In Mac OS X UI, the control that corresponds to "details" is often used to hide extra controls for a form or dialog |
| 08:33 | <othermaciej> | that are normally not of interest but sometimes are |
| 08:36 | <annevk> | yeah, I think Hixie solved the issues now |
| 08:36 | <annevk> | we can now also safely allow nested links if it weren't for the parsing |
| 08:38 | a-ja | has been playing with <details> equivalent class-swapper thingy for a couple hours...and is frustrated getting it working via keyboard with opera |
| 08:44 | <annevk> | A future version of Opera will have addressed that... |
| 08:44 | <annevk> | (by supporting tabindex for HTMLElement) |
| 08:47 | <a-ja> | adding tabindex to a link child of legend let me get there, but won't fire "onactivate" for some reason, unless i add a href....but that make the page scroll |
| 08:50 | <a-ja> | annevk: in Kestrel, per chance? |
| 10:32 | <zcorpan_> | re <a> in <details>: could be reasonable to disallow <a> in <details><legend> |
| 10:33 | <zcorpan_> | since it's really the legend that is interactive |
| 10:51 | <mpt> | precisely |
| 11:07 | <annevk> | note that that argument doesn't apply to <datalist> |
| 12:15 | <Lachy> | zcorpan_, this article might be useful for you to post in that forum thread http://lachy.id.au/log/2005/05/script-comments |
| 13:37 | <Philip`> | Is anybody testing HTML5 backward-compatibility with UAs that don't understand <script>/<style> and still need the content-hiding comments? (I would have thought it's very unlikely for anyone making an HTML5 page to expect it to work decently in ancient browsers, and if it's going to break badly anyway then there's no need for the spec to explain the script-hiding syntax) |
| 13:38 | <othermaciej> | how ancient are we talking about |
| 13:38 | <othermaciej> | Netscape 1? |
| 13:39 | <Lachy> | the browser on the Motorolla v3 phone! |
| 13:40 | <Lachy> | but that one can be ignored, since it would be mostly used in walled gardnes anyway |
| 13:40 | <Lachy> | *gardens |
| 13:41 | <Philip`> | The only age-related references I've seen are saying that HTML 3.2 added script/style and would correctly not display the contents, but I've got no idea how that matches implementations from that time period |
| 13:46 | <Lachy> | Netscape 2.0 added support for script, so any browser released after relatively soon afterwards would have added support too |
| 13:47 | <Philip`> | Wow, not even Google works in Netscape 3 |
| 13:47 | <Philip`> | (It has JavaScript error boxes coming up) |
| 13:47 | <Philip`> | I wonder if that suggests it's old enough to be worth abandoning... |
| 13:48 | <Lachy> | anything before IE6 is worth abandoning |
| 13:50 | <Lachy> | heh, "What's the point of having different browsers if they are all required to work the same way ... ?" -- http://lists.w3.org/Archives/Public/www-html/2007Apr/0211.html |
| 13:56 | <zcorpan_> | Lachy: it's a forum, anyone can post there :) |
| 13:56 | <Lachy> | zcorpan_, I'm not registered |
| 13:57 | <Lachy> | I'm also avoiding it because I don't like forums |
| 13:58 | <zcorpan_> | ok |
| 13:59 | Lachy | responds once more to www-html (though I'm getting tired of the baseless arguments) |
| 13:59 | <Philip`> | Aha: Netscape Gold 2.02 and Navigator 3.04Gold shows the contents of <style>s, but not <script>s. Navigator 4.08 shows neither. |
| 14:00 | <Lachy> | I surprised you have all of those browsers installed |
| 14:00 | <Philip`> | I didn't, until ten minutes ago :-) |
| 14:00 | <Philip`> | http://sillydog.org/narchive/full123.php etc |
| 14:03 | <Philip`> | Hmm, Netscape 2 takes quite a while to load the HTML5 spec, and then it gets stuck trying to download an image and doesn't draw the rest of the page :-( |
| 14:04 | <Lachy> | can you put up a screenshot of google, or something in NN2? |
| 14:05 | <Lachy> | or the Acid2 test :-) |
| 14:08 | <zcorpan_> | wouldn't <noscript><img src="tracker"/></noscript> work even in xhtml5, given that images that are display:none are not fetched? |
| 14:09 | <Lachy> | zcorpan_, are you sure they're not fetched? |
| 14:09 | <Lachy> | isn't that up to the browser? |
| 14:10 | <Philip`> | (Hmph, Gmail makes Netscape crash...) |
| 14:10 | <Lachy> | lol |
| 14:10 | <Lachy> | can anyone else access hixie.ch? I'm trying to get to the live dom viewer |
| 14:11 | <Philip`> | That and whatwg.org seem to be down |
| 14:14 | <Philip`> | http://canvex.lazyilluminati.com/misc/nn2/ |
| 14:14 | <Philip`> | (Google is at least still generally usable, if you ignore the script error on every page) |
| 14:17 | <Philip`> | In any case, I think I feel justified in not doing <style><!--... on my sites because anybody using Netscape 3 is already going to be used to having a hard time on the web |
| 14:24 | <Philip`> | (I have to go back to Netscape Navigator 1.22 before there's any point in doing <script><!--...//--></script>) |
| 14:26 | <Lachy> | what about past versions of IE? |
| 14:26 | <Lachy> | or TimBL's WWW browser |
| 14:27 | <Philip`> | (In relation to earlier discussions about parsing bogus colour values, NN1 parses "black" as #b0ac00, which disturbs my site's colour scheme) |
| 14:29 | <Lachy> | really? I thought it would have support for colour keywords |
| 14:29 | <Lachy> | I wonder when X11 colours were added to NN then |
| 14:30 | <Philip`> | NN2 handles black correctly |
| 14:31 | <Lachy> | try some more http://en.wikipedia.org/wiki/X11_color_names |
| 14:31 | <Philip`> | http://browsers.evolt.org/?worldwideweb/NeXT - maybe that would work if I had a NeXT computer, but sadly I don't :-( |
| 14:32 | <Philip`> | http://browsers.evolt.org/?ie/32bit - aha, handy archive |
| 14:34 | <Lachy> | see which colours NN2 uses for Gray, Green Maroon and Purple, which clash between X11 and CSS |
| 14:38 | <othermaciej> | Jukka's message seems poorly thought out |
| 14:38 | <othermaciej> | what's the point of having more than one browser if they all act completely different? |
| 14:38 | <Lachy> | did you read my response? |
| 14:39 | <zcorpan_> | i need to figure out exactly which color names are supported by html parsers, in order to put forward a proposal |
| 14:39 | <Lachy> | a proposal for what? |
| 14:39 | <othermaciej> | www-html seems much more insane than public-html |
| 14:39 | <zcorpan_> | for changing the html5 parsing algorithm |
| 14:39 | <othermaciej> | glad there's other people to cover it |
| 14:40 | <Lachy> | I thought all the supported colour names were known, you just need to determine the algorithm to use for unknown names |
| 14:40 | <Philip`> | Lachy: NN1 handles none of those colours; NN2 handles them all (and uses the HTML-like values rather than X11-like) |
| 14:40 | <zcorpan_> | don't we need to spec what the set of color names are to be known? |
| 14:41 | <Lachy> | CSS defines those, HTML can just reference that |
| 14:41 | <zcorpan_> | css3-color? |
| 14:42 | <Lachy> | yes |
| 14:42 | <zcorpan_> | could work |
| 14:43 | <Lachy> | oh, Tina responded to me on www-html, wanting to know which parts of HTML4 can't be implemented |
| 14:43 | <Lachy> | Now, where do I start? |
| 14:43 | <othermaciej> | <object> |
| 14:43 | <Philip`> | ...and the Wikipedia list is precisely the same set as is listed in NN2's .exe |
| 14:43 | <othermaciej> | is it "can't be implemented" or "can't be implemented interoperably" |
| 14:44 | <zcorpan_> | it surely contradicts itself in various places, doesn't it? |
| 14:44 | <wilhelm> | Lachy: Error handling. |
| 14:44 | <othermaciej> | also are we looking for self-contradiction in the spec, or contradiction to expectations of deployed content? |
| 14:44 | <Lachy> | othermaciej, both |
| 14:44 | <othermaciej> | tag minimization is one obvious thing in the latter category |
| 14:44 | <zcorpan_> | </foo closing style and script |
| 14:45 | <Philip`> | (...and precisely the same list of colour names in NN4 too) |
| 14:45 | <Lachy> | we should record this stuff in the whatwg wiki too! |
| 14:45 | <othermaciej> | you can't implement <foo/> as SGML calls for |
| 14:45 | <zcorpan_> | that ua's must not assume a default encoding |
| 14:45 | <othermaciej> | Lachy: this isn't quite the kind of thing you asked for, but it amuses me to no end that the definition of "conforming implementation" is self-contradictiory |
| 14:45 | <othermaciej> | in HTML4 |
| 14:46 | <Lachy> | I'll mention that too |
| 14:46 | <zcorpan_> | perhaps it implies that being conforming is optional for conformance? :) |
| 14:47 | <othermaciej> | http://www.w3.org/TR/html401/conform.html |
| 14:47 | <othermaciej> | "A conforming user agent for HTML 4 is one that observes the mandatory conditions ("must") set forth in this specification, including the following points:" |
| 14:47 | <othermaciej> | and then it lists one should-level, one must-level and one recommend-level requirement |
| 14:48 | <othermaciej> | but you can also find whole chapters where the word "must" does not appear |
| 14:49 | <Philip`> | Whoops, my default browser is now set to Netscape 1 |
| 14:49 | <Philip`> | At least the later versions asked me if I wanted to change it |
| 14:49 | <othermaciej> | this section for instance contains no UA conformance requirements: http://www.w3.org/TR/html401/struct/lists.html |
| 14:52 | <Philip`> | Lachy: Unfortunately I can't test old versions of IE - "Internet Explorer 3.0 cannot be installed on a system that has Internet Explorer 4.0 installed" |
| 14:53 | <Philip`> | although I may actually have a spare Windows 3.1 virtual machine somewhere, which may run it... |
| 14:53 | <Lachy> | cool |
| 14:55 | <zcorpan_> | Philip`: does it need to be installed? isn't there a standalone version that just runs? |
| 15:01 | <Philip`> | (Hmm, installing network drivers in Win3.11 is too much of a pain so I think I'll give up) |
| 15:02 | zcorpan_ | has or had mosaic installed... most of the web didn't work at all |
| 15:02 | <Philip`> | zcorpan_: Ah, hadn't tried that, but the ones at http://browsers.evolt.org/?ie/32bit/standalone seem to work |
| 15:03 | <zcorpan_> | ...which wasn't because of the markup used but something else (probably http something) |
| 15:04 | <Philip`> | NN1 seems to have HTTP problems - it's not telling the server what hostname it's connecting to, or something, so it acts the same as if you typed in the IP directly (which rarely works) |
| 15:04 | <Philip`> | How peculiar - the HTML5 spec in IE3 has a textarea in the top left corner |
| 15:04 | <zcorpan_> | perhaps it sees the <textarea> in the last comment |
| 15:05 | <zcorpan_> | closes comments at > or something |
| 15:08 | <Philip`> | (Oh, there's no standalone IE2 there, and the non-standalone one fails to install with "Internet Explorer requires Windows 95.") |
| 15:09 | <Philip`> | (IE3 has precisely the same list of colour names as NN2) |
| 15:12 | <Lachy> | does <applet> also suffer from interop problems? |
| 15:13 | <othermaciej> | <applet> has a pretty serious flaw in Safari |
| 15:13 | <othermaciej> | at times, it is known to load a Java applet |
| 15:14 | <zcorpan_> | "If your argument was correct, then everyone would learn about mark-up simply by reading other people's mark-up, which is obviously not what happens." |
| 15:14 | <zcorpan_> | wow |
| 15:14 | <othermaciej> | that's the only way I ever learned... |
| 15:14 | <zcorpan_> | it might not apply to *everyone*, but surely most |
| 15:15 | <othermaciej> | even people who learn from books copy the examples in those books |
| 15:15 | <Lachy> | zcorpan_, who wrote that? |
| 15:15 | <zcorpan_> | Lachy: mark http://www.w3.org/mid/640dd5060704300529u70a16269xec412a16436f3e2b⊙mgc |
| 15:16 | <Lachy> | he is so out of touch |
| 15:17 | <Lachy> | even I learned by view source initially, though I started reading specs when that was too difficult to understnad |
| 15:23 | Philip` | wonders how different the web would be today if the early browsers didn't have a "view source" menu item |
| 15:24 | <Lachy> | people would have been forced to use editors to write HTML, just like very few people hand code RTF, PDF, Work docs, etc. |
| 15:24 | <Lachy> | s/Work/Word/ |
| 15:27 | <Lachy> | here's the draft of my reply to Tina http://lachy.id.au/temp/html4 |
| 15:29 | <othermaciej> | SHORTTAG syntax is kind of cool |
| 15:29 | <othermaciej> | too bad it can't be used |
| 15:37 | <Lachy> | I made a few more additions and modifications, and I think I'm ready to send it. I'm sure there would be more, but I couldn't be bothered looking. |
| 15:40 | <othermaciej> | I think "do not assume a default character set" is one worth listing |
| 15:40 | <othermaciej> | (don't remember the precise wording of the spec) |
| 15:40 | <Lachy> | where does it say that in the sepc? |
| 15:40 | <Lachy> | found it |
| 15:40 | <Lachy> | it says "Therefore, user agents must not assume any default value for the "charset" parameter." |
| 15:40 | <Lachy> | what should I say about it? |
| 15:42 | <othermaciej> | user agents have to assume the default encoding is Windows Latin 1 |
| 15:42 | <othermaciej> | (windows-1252 officially) |
| 15:43 | <othermaciej> | in fact, browsers have to treat latin1 as windows-1252 always, but I don't think HTML is to blame for that |
| 15:43 | <Philip`> | "<table border> is actually shorthand for <table frame="border">" - shouldn't that be border="border"? |
| 15:44 | <Lachy> | no |
| 15:44 | <Lachy> | in HTML5, yes |
| 15:44 | <Lachy> | in SGML, no |
| 15:44 | <Philip`> | Oh, okay, HTML4 has funny rules for that |
| 15:45 | <Lachy> | it's actually the attribute name that can be omitted, not the value as most people think |
| 15:45 | <Lachy> | but UAs only implemented it for attrs that had the same name and value |
| 15:46 | <wilhelm> | othermaciej: In some countries, UAs have to use other encodings as default. |
| 15:46 | <Lachy> | othermaciej, I added the default charset thing |
| 15:47 | <Lachy> | wilhelm, yes, I wrote "usually forced to default to win 1252" |
| 15:47 | <othermaciej> | wilhelm: well, the per-country default encoding thing is even worse - we don't actually do that in Safari because we don't want your web browsing to be affected by language setting |
| 15:47 | <wilhelm> | Most Japanese websites don't bother defining what charset they use. And it usually isn't windows-1252. |
| 15:47 | <met_> | what exactly mean "concur" in the W3C wbs? feel something like no but without canceling the approval? |
| 15:47 | <Lachy> | it means you agree with whatever the majority say |
| 15:48 | <wilhelm> | othermaciej: It is terrible, I agree. But you can't ship on Japanese devices without it. |
| 15:48 | <met_> | what is difference pro abstaining? |
| 15:49 | <othermaciej> | wilhelm: I might consider it if it was for localized hardware sold only in the given region |
| 15:49 | <Lachy> | met_, see the logs of #html-wg from a few hours ago for an explanation |
| 15:50 | <Lachy> | we just need to get japanese markets to use UTF-8 or UTF-16 for everything. |
| 15:50 | <wilhelm> | othermaciej: Yes. That is what we (Opera) do on the Wii and on phones for Japanese customers. |
| 15:51 | <Lachy> | the major problem is that OSs ship with different default charsets for different languages, instead of Unicode |
| 15:51 | <met_> | Lachy thx |
| 15:52 | <Lachy> | what does a japanese user do when they visit a site that needs to default to win 1252? |
| 15:52 | <wilhelm> | Lachy: Getting everyone to use Unicode would be ideal. But just getting sites to define which charset they use would solve most of these problems. |
| 15:52 | <othermaciej> | yeah, relying on the default encoding is bad |
| 15:53 | <wilhelm> | Yes. It breaks Norwegian sites in Japan, and Japanese sites in Norway. |
| 15:53 | <Lachy> | wilhelm, we've got more chance of fixing the situation by getting editors to use Unicode as default encoding for everyting, than by getting authors to declare charsets |
| 15:55 | <Lachy> | any more suggestions for my list of HTML4 issues before I send? |
| 15:55 | <wilhelm> | That is possible. |
| 15:56 | <othermaciej> | Lachy: in Asia (Japan especially), enthusiasm for unicode is not that widespread |
| 15:56 | <Lachy> | why not? |
| 15:57 | <othermaciej> | they like how compact encodings like shift-JIS are for their language |
| 15:57 | <othermaciej> | same reason Americans don't tend to use UTF-16 |
| 15:57 | <MikeSmith> | the Japanese developer community is not fond of Unicode. Many Japanese mobile devices don't have proper Unicode support (or even fonts for non-ASCII Latin characters). Several Japanese desktop mail clients don't have Unicode support, and many (or most) e-mail clients on mobile devices in Japan don't either |
| 15:58 | <Lachy> | yeah, full support for unicode is a problem everywhere |
| 18:02 | <othermaciej> | good morning |
| 18:06 | <gsnedders> | is it possible to have a third state with a checkbox (like the line in OS X?) |
| 18:10 | <Dashiva> | Not as of currently, no |
| 18:15 | <gsnedders> | so I haven't forgotten some input type value |
| 18:49 | <gsnedders> | I take it the SOLIDUS is allowed in all empty elements? |
| 19:32 | <deltab> | gsnedders: I take it you mean inside the start tag? |
| 19:33 | <deltab> | void elements only |
| 19:33 | <deltab> | i.e. base, link, meta, hr, br, img, embed, param, area, col, input |
| 19:33 | <gsnedders> | yes |
| 20:31 | <bewest> | "the w3c should define HTML, and browser manufacturers should be willing to accept that definition (or to reject it, at their own risk: this is a free world), but it would be a great boost for standards were the "W3C HTML 5" logo to be as applicable to /browsers/ as it will be to web pages." |
| 21:21 | <gsnedders> | mm… looking through the HTML 4.01 Trans DTD, there are more EMPTY elements than those allowed to have a SOLIDUS character in their start tag in the parsing section of WA1.0 |
| 21:29 | othermaciej | is tempted to +1 Hixie's email about +1 messages, but restrains his impish side |
| 21:32 | <bewest> | othermaciej: your buddy dave hyatt beat you to it |
| 21:33 | <othermaciej> | I guess he's more of an ass than I am :-) |
| 21:34 | <bewest> | I'm not sure how much +1 matters on the public-html list either |
| 21:34 | <bewest> | it appears the chairs intend to use web surveys |
| 21:34 | <bewest> | othermaciej: you have some nice replies on that list, btw |
| 21:35 | <bewest> | imo one of the voices of sanity |
| 21:37 | <Philip`> | Hixie's email was only explaining why +1 is not useful in discussions about technical issues, where there can be objectively wrong answers - it wasn't arguing that +1 is not useful in discussions about whether +1 is not useful in discussions, and hence it seems quite logical to +1 it |
| 21:37 | <jgraham> | +1 |
| 21:38 | jgraham | has wanted to type that in response to /almost anything/ for weeks. He realises this is deeply illogical |
| 21:39 | <gsnedders> | +1 |
| 21:39 | <gsnedders> | I restrained replying to Anne's (I think) call for an end of +1s on public-html, on grounds that he was citing the number of messages as a reason, making it further illogical |
| 21:40 | <othermaciej> | -1 |
| 21:50 | <Hixie> | if anyone's bored and wants to argue for a while, this one's hot: http://www.w3.org/Bugs/Public/show_bug.cgi?id=1500 |
| 21:56 | <othermaciej> | the very URL is discouraging |
| 23:29 | <Hixie> | annevk? |