element ends up in the DOM properly? [23:18:00.0000] that's it :) I had a h1 before the articles which was hidden so I kinda missed it. thanks a lot! :) [23:23:00.0000] did Mozilla drop their implementation? [23:23:01.0000] I can't remember what happened with that [23:24:00.0000] if anyone is bored and would like to critique: http://paulgendek.com/html5/ html5 blog layout, validates, with some hcard microformats... [23:24:01.0000] going to do single article view with

doesn't work in most browsers yet [00:45:00.0000] is there a javascript fix for that yet? [00:45:01.0000] dunno [00:45:02.0000] thx [00:46:00.0000] if there is, nobody has updated the issue marker at http://www.whatwg.org/specs/web-apps/current-work/#the-legend-element yet [00:46:01.0000] (the fifth icon in those boxes represents js implementations) [00:47:00.0000] What document conformance rules is Optimus based on? [00:55:00.0000] othermaciej, why didn't you suggest closing ISSUE-4? [00:55:01.0000] (html-versioning) [00:56:00.0000] Lachy: I personally think it should be closed with no change to the spec, but there has been continuing debate and Larry intends to make a proposal [00:56:01.0000] Lachy: so I'm leaving that one in the hands of the chairs [00:56:02.0000] ok [00:56:03.0000] Hixie: #refsE163 and #refsRFC2965 are still missing [00:57:00.0000] i just fixed 2965 [00:57:01.0000] E163 is there though [00:57:02.0000] paulgendek: there's a problem with parsing of in all existing browsers, and you can't practically use

with backup styling until it is fixed [00:57:03.0000] paulgendek: I raised the issue with Hixie but he was not inclined to increase the number of elements in HTML to fix it [00:58:00.0000]

's not important enough. just pretend that it's not in html5 and that we added it 3 years from now. [00:58:01.0000] Lachy: I tried to limit my suggested closures to thinks that it's just painfully obvious should be closed, and where there has not really been discussion to the contrary [00:59:00.0000] othermaciej, any idea when the parsing issues will be fixed in webkit? [00:59:01.0000] once all browsers either adopt HTML5 parsing or at least change parsing of not found in a

will be practically usable [01:00:00.0000] on a completely different note, does anyone want to volunteer to take point on the encoding aliases issue? [01:00:01.0000] Lachy: at the very latest, when we adopt HTML5 parsing, but I will propose doing it sooner [01:00:02.0000] Hixie, what's the issue? [01:00:03.0000] lachy: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-July/021207.html [01:00:04.0000] Hixie: I don't see E163 there [01:00:05.0000] e.g. http://www.whatwg.org/specs/web-apps/current-work/#refsE163 doesn't exist [01:00:06.0000] (but is referenced) [01:00:07.0000] Philip`: hm, odd [01:01:00.0000] it's in the source [01:01:01.0000] wonder what broke [01:03:00.0000] Hixie, what specifically needs to be done? Do we need to update the IANA registry? [01:04:00.0000] Lachy: yeah [01:04:01.0000] Lachy: basically i want to remove the "Additional character encoding aliases" table [01:04:02.0000] Lachy: and just have hte IANA registry be all up to date [01:05:00.0000] Philip`: wtf [01:05:01.0000] ok. I don't have time to do it right now, but if no-one else volunteers within a few weeks, I might be able to do it [01:06:00.0000] k [01:06:01.0000] what's the process for updating the IANA registry? [01:06:02.0000] depends on the registry; i haven't looked into what it takes to update that one [01:08:00.0000] given this line: href="#refsE163">[E163] [X121]

[01:08:01.0000] Hixie, hsivonen: does either of you think it is useful to continue tracking this as an issue: http://www.w3.org/html/wg/tracker/issues/17 [01:09:00.0000] why does this not match two things: while ($substedline =~ m/\G.*href="#refs([^"]+)"/osg) { [01:10:00.0000] Looks like it will take a few e-mails to this list http://www.iana.org/assignments/charset-info [01:10:01.0000] othermaciej: depends on what MQ says now. I'll have to check [01:11:00.0000] http://tools.ietf.org/html/rfc2978 [01:11:01.0000] would it be inappropriate spamming if I suggested a bunch more issue closures? [01:11:02.0000] I am doing one per message to associate them in the issue tracker [01:11:03.0000] othermaciej, doesn't the issue tracker allow one message to be associated with multiple issues? [01:12:00.0000] Lachy: I don't know - I'm counting on the automated detection of ISSUE-nn in the subject [01:12:01.0000] othermaciej: we fixed that years ago [01:12:02.0000] othermaciej: as far as I can tell, the MQ WS issue still stands [01:12:03.0000] looks like it does [01:13:00.0000] othermaciej: (U+000B isn't a whitespace character anymore in HTML5) [01:13:01.0000] is it fixed in HTML5? [01:13:02.0000] Hixie: oh [01:13:03.0000] see http://www.w3.org/html/wg/tracker/issues/1 and http://www.w3.org/html/wg/tracker/issues/4 -- your last mail appears in both [01:13:04.0000] Hixie: nice [01:13:05.0000] I wonder how I forgot [01:13:06.0000] i'd forgotten too [01:13:07.0000] i just checked :-) [01:13:08.0000] had to look up wtf vertical tab was first [01:14:00.0000] Lachy: is it searching the text? [01:14:01.0000] othermaciej: so yeah, the issue has been addressed. No point in keeping it open. [01:14:02.0000] othermaciej, yes [01:14:03.0000] Lachy: do you think it would be better to gang them up? [01:14:04.0000] yes [01:15:00.0000] Hixie: Because the .* is greedy and will match the whole first href, I expect [01:16:00.0000] Hixie: You probably want .*? [01:16:01.0000] ohhh! [01:16:02.0000] yes [01:16:03.0000] thanks [01:18:00.0000] /me would just use "while ($substedline =~ /href="#refs([^"]+)"/g)" which works the same and has fewer funny characters [01:19:00.0000] there's an implied \G? [01:19:01.0000] i always had problems using /g [01:19:02.0000] Hixie: is there anything to do in response to this: http://www.w3.org/html/wg/tracker/issues/26 [01:20:00.0000] (without the \G, i mean) [01:20:01.0000] (it appears to be suggestions for accessibility of the spec text itself) [01:20:02.0000] for one of those things we're waiting for gsnedders to add features to anolis [01:21:00.0000] i can't read the bottom two paragraphs, they're too "wall of text" for me [01:21:01.0000] i guess i have a disability, and whoever wrote that doesn't know how to write accessible comments :-) [01:21:02.0000] what's still pending about issue 37? [01:21:03.0000] #3 explains why hex/rgb colors are supposed to be bad [01:21:04.0000] also, wasn't issue 54 addressed long ago? [01:22:00.0000] Philip`: ok fixed, thanks for catching that [01:22:01.0000] issue 60 is not realistically actionable [01:23:00.0000] #4 in issue-26 looks too vague to be actionable [01:23:01.0000] issue 60 should have been closed at the telcon a month ago. It was kept open for silly reasons [01:23:02.0000] I guess I won't suggest closing it if there are anolis changes planned which will help [01:23:03.0000] hsivonen: I don't understand why 37 is open, indeed [01:24:00.0000] also I thought 54 was addressed [01:24:01.0000] Is anyone but Larry expressing active interest in keeping issue 4 open? [01:24:02.0000] and 60 should obviously be closed in light of the news about the XHTML2 WG [01:24:03.0000] issue 54 is pending the registration of the about: URI scheme [01:25:00.0000] I haven't heard from Joseph in a while, so I don't know what he's doing about that RFC [01:25:01.0000] hsivonen: no idea - I would hope the chairs would make some assessment of who has objections [01:26:00.0000] Hixie: It's similar to an implied \G, except it starts where the previous execution of the current regexp on the current string left off, whereas \G starts where the previous regexp on the current string left off (even if it's a different regexp) [01:27:00.0000] Hixie: (so \G is useful when chaining multiple regexps together, and /g works fine if you're just looping over a single one) [01:27:01.0000] Philip`: ah, i see. yeah, i usually use /g with multiple regexps, that's why i had trouble without \G. [01:27:02.0000] Philip`: thanks. [01:29:00.0000] do these remain issues of controversy (it's not clear to me from the issue tracker item what the point is): [01:29:01.0000] for 27, i'm fine with the WHATWG registry, but others are probably not [01:29:02.0000] Hixie: (Also the /o is redundant unless you've got interpolated variables in the regexp, and the /s is redundant since you're only processing a single line) [01:30:00.0000] 28 is fixed by abarth's mimesniff draft [01:30:01.0000] Hixie: (Also you should get rid of all the pesky whitespace and variable names, they're just wasting bytes :-( ) [01:30:02.0000] othermaciej: Anolis changes for #4 in issue-26? [01:30:03.0000] Philip`: yeah i just always include /o and /s [01:30:04.0000] othermaciej: Or one of the others? [01:30:05.0000] gsnedders: Hixie mentioned that something in http://www.w3.org/html/wg/tracker/issues/26 is waiting on anolis changes - I do not know what he had in mind [01:30:06.0000] ah, yeah [01:30:07.0000] Philip`: /s actually isn't redundant here, one of the operations that makes $line into $substedline can introduce more lines [01:30:08.0000] Hixie: does the HTML5 spec completely defer to mimesniff now? [01:31:00.0000] othermaciej: yeah [01:31:01.0000] othermaciej: where appropriate, anyway [01:31:02.0000] /me wonders where in the multi-level feedback queue of his to-do list that is [01:32:00.0000] oh, there [01:32:01.0000] /me shrugs, and heads off to work [01:32:02.0000] Hixie: Oh, okay [01:32:03.0000] Hixie: I'll try and work out the encoding change today though [01:32:04.0000] Hixie: Clearly you shouldn't use such confusing variable names like "$substedline", and should try to put absolutely everything into $_ [01:33:00.0000] gsnedders: i don't mind the encoding change at all [01:33:01.0000] Hixie: I do. You specify US-ASCII and get UTF-8. [01:33:02.0000] Philip`: this is nested inside several loops, one of which uses $_ :-) [01:33:03.0000] gsnedders: k :-) [01:34:00.0000] Hixie: If you specify US-ASCII, you should get US-ASCII. [01:34:01.0000] I spoke to jgraham about it briefly yesterday, I suspect it is in the PMS code the bug :P [01:34:02.0000] Hixie, what do you expect calling mutator methods like setHours() on a Date object returned from HTMLTimeElement.time and friends to do? [01:34:03.0000] anything? [01:35:00.0000] gsnedders: can i specify cross-spec cross refs and get those too? :-) [01:35:01.0000] /me stabs Hixie [01:36:00.0000] gsnedders, great, now who will edit the spec? [01:36:01.0000] manu, apparently [01:36:02.0000] Lachy: I didn't say I killed him, or harmed him that much. [01:36:03.0000] heycam: i expect the Date object to be a real Date object with no relationship to the underlying attribute [01:36:04.0000] ok, so a new one gets returned every time you access it? [01:36:05.0000] yeah [01:36:06.0000] ok you'll need to say that i think [01:36:07.0000] Lachy: Just a bit of knife play :P [01:37:00.0000] heycam: can you add a comment in the comment box to that effect and hit the button? it'll file a bug for you [01:37:01.0000] sure [01:37:02.0000] /me hopes gsnedders used a stab-proof knife [01:39:00.0000] He just used one of those plasic toy knives where the blade goes in [01:40:00.0000] Hixie, i'm still not sure of the usefulness of having separate date/time/timezone attributes [01:40:01.0000] it might be better to have one Date and then booleans to state whether certain parts were present [01:40:02.0000] Philip`, are they anything like those useless plastic, cut-proof scissors often given to children so they can't cut themselves, or paper? [01:41:00.0000] heycam: it seems that whatever you do, you're going to have to end up creating these Date objects for yourself, typically [01:41:01.0000] heycam: having just these dates seems as useful or more than one date and booleans [01:42:00.0000] Lachy: They're like http://www.foxnews.com/story/0,2933,526590,00.html [01:44:00.0000] the claim about them being inneffective as a weapon seems a little inaccurate [01:44:01.0000] Hixie, I suppose I'd like to see use cases to convince me. maybe i'll send mail about it. [01:44:02.0000] they may not be effective for stabbing due to the blunt end, but they look to be perfectly effective for slitting someone's throat [01:45:00.0000] btw are s meant to represent an instance in time or an instance in time and an associated time zone? [01:45:01.0000] they seem unlikely to reduce the risk of injuries [01:45:02.0000] most accidental knife injuries while cooking are from the edge, not the point [01:46:00.0000] heycam: heycam: a date, a time, or an instance in time with time zone [01:46:01.0000] othermaciej: It seems it's designed to reduce injuries when e.g. a burglar breaks into your house and you disturb him and he grabs the nearest knife and stabs you, not when cooking [01:46:02.0000] Hixie, ok. currently the spec says only "The time element represents a precise date and/or a time in the proleptic Gregorian calendar." [01:47:00.0000] Philip`: s/injuries/accidental injuries/ [01:47:01.0000] it might prevent deliberate stabbings, but it would also prevent self-defense with a knife [01:47:02.0000] (Presumably it's no good for premeditated attacks, because you've have to be pretty stupid to intentionally take a stab-proof knife to stab someone with) [01:47:03.0000] then again, the UK doesn't believe in self-defense any more [01:48:00.0000] bbiab [01:48:01.0000] it may not be effecitve for stabbing, but slicing attacks are still as effective as before [01:49:00.0000] unless they make it blunt on the edge as well [01:49:01.0000] Were slicing attacks still effective before? [01:49:02.0000] I imagine they're much harder to do [01:49:03.0000] and much harder to cause fatal injuries [01:49:04.0000] if you are any good at knife fighting you can mess up an attacker really bad, probably fatally [01:49:05.0000] just grab someone from behind, hold the knife to their throat and slice [01:49:06.0000] or if you are strong enough to overpower someone, you can cut their throat or the like [01:50:00.0000] /me notes that discussing murdering techniques is a new low for this channel :-) [01:50:01.0000] just slashing at the arms of someone coming for you, they will quickly reach a point where they are likely to bleed to death [01:50:02.0000] (forearms) [01:52:00.0000] Hmm, seems more difficult and situational than just shoving the knife into someone's chest and then running off before anyone notices you [01:53:00.0000] (Mostly this is in the context of e.g. http://www.telegraph.co.uk/news/newstopics/politics/2078201/Gordon-Brown-urges-tougher-punishments-for-teenage-knife-crime.html and not trained knife fighters) [02:28:00.0000] I guess the main use case for the stab-proof knife is "violent hasband flies into a rage and attacks wife" [02:28:01.0000] *husband [02:29:00.0000] Are you calling myself and Hixie husband and wife? :\ [02:30:00.0000] I'd like to have stab-proof knives just to mitigate the consequences of dropping a knife [02:30:01.0000] /me shan't enquire which recesses of gsnedders psyche caused him to infer that [02:30:02.0000] I wonder if the stab-proof knives will be delayed by 20 due to a patent [02:31:00.0000] *20 years [02:32:00.0000] jgraham: Well, I seemingly often attack Hixie in a rage… [02:32:01.0000] /me thinks something is a bit backwards though, seeming he's the one wearing a girls T-shirt [02:33:00.0000] well, this is an interesting conversation to walk into... [02:35:00.0000] it's inappropriate for use cases to assign gender to attacker and victim [02:35:01.0000] GPHemsley: See /topic :) [02:36:00.0000] you people really should have a logic detector at the entrance [02:36:01.0000] you can't expect me to remember to voluntarily relinquish it every time I come in here [02:36:02.0000] /me also thinks non-violent discourse without stabbing would be good on #whatwg [02:38:00.0000] aargh. I just realized that if nodes have been moved from a document to another during the parse, the HTML5 parser in Gecko notifies the subsequent child insertion on the wrong document [02:38:01.0000] fixing that in a performant way seems like a can of worms [02:40:00.0000] I guess I need to ship a pointer to the parser's doc to each tree operation [02:40:01.0000] sigh. [02:45:00.0000] hey. Theora support has migrated from Chrome dev channel to beta channel [02:45:01.0000] nice [03:26:00.0000] I wonder how hard it would be to generate static status markers from the WHATWG dynamic status marker data [03:27:00.0000] hsivonen: Not hard but possibly not worthwhile either. [03:28:00.0000] I can't imagine that it would make people happy [03:29:00.0000] status markers are a bit of a sideshow [03:29:01.0000] issues need to actually be resolved in some final way - it's only the total lack of closure on anything in the HTML WG that makes people play games with publishing WD and adding status markers and such [03:31:00.0000] Indeed. I would rather we spent time resolving issues than shuffling little red boxes around [03:31:01.0000] jgraham: it seems to me that autogenerating the markers into the W3C snapshots would address the issue Manu's draft is said to address [03:31:02.0000] I agree that actully resolving issues is better [03:32:00.0000] hmm. looks like Wikipedia now tells a bogus story about video codec debate as far as dates of events go [04:43:00.0000] http://www.exalead.com/search/web/results/?q=html5 - wow, a site that actually uses SVG for real [04:44:00.0000] (in the Languages/Countries pie charts) [04:45:00.0000] Philip`: http://www.opera.com/company/jobs/ ;-) [04:49:00.0000] Oh, looks like they didn't write the SVG themselves, they just used http://raphaeljs.com/ [05:18:00.0000] hey people! could you take a look at http://bearditch.com/stuff/debate.html and give your 2cents about the markup. Would you structure it differently? It's for a general tabbed content section of a site. If you have the time or mood ofc. [05:24:00.0000] http://twitter.com/mattur [05:26:00.0000] hsivonen: followed... [05:28:00.0000] StefanShipor: looks pretty good! [05:30:00.0000] paulgendek, StefanShipor: You may find http://gsnedders.html5.org/outliner/ useful in assessing whether you are using sectioning elements in the right way [05:31:00.0000] cool, thanks! [05:32:00.0000] paulgendek: In your case I think you have too many sibling

s [05:32:01.0000] jgraham: but why do the untitled sections appear? http://gsnedders.html5.org/outliner/process.py?url=http%3A%2F%2Fbearditch.com%2Fstuff%2Fdebate.html [05:32:02.0000] I waas expecting only one since it doesn't have a heading [05:34:00.0000] StefanShipor: You have a with no heading, a

with no heading at the top and two

s with no heading further down [05:34:01.0000]

should have within? [05:34:02.0000] (FWIW I think that the approach to sections with no heading in the spec needs to change) [05:35:00.0000] paulgendek: Theoretically, yes [05:36:00.0000] i stared at my dom for a few hours and couldn't think of any other ways of ordering elements, etc. it seems like the most practical for a real world solution, as a base theme, etc [05:36:01.0000] paulgendek: e.g. you miught say [05:37:00.0000] paulgendek: In yoour specific case, I think you want to structure the page so that the main page header is outside of the

s [05:37:01.0000] that doesn't make sense, because from what i read, you would put the h-tags in the

, followed by the

[05:38:00.0000] Like

My blog

My post

[05:38:01.0000] i think body's header should be reserved for accessibility links, in a magazine/blog layout [05:39:00.0000] there will be lots of navs on sites for navigation through posts and such. I hope they don't leave the heading inside nav mandatory [05:39:01.0000] I don't know what you mean by accessibility links really (do you mean "skip to content" type links?) but that seems like a strange belief [05:39:02.0000] cause if you fave "older" and "newer" it's already pretty self-explanatory [05:40:00.0000] have* [05:40:01.0000] it's nowhere near 2022, we still need to be practical here... [05:40:02.0000] in 2009/2010, i still need skip to content links [05:41:00.0000] may be a want, idk [05:41:01.0000] paulgendek: Maybe, I'm not sure what the state of the art in AT is in this regard. But I see no readon why that precludes having a header for the body [05:42:00.0000] because in modern web design, where we have containers, centered, with borders, padding, maybe even nested with special corners, etc [05:42:01.0000] the body header should be reserved for top-most content [05:43:00.0000] What is the title of your blog if not topmost content? [05:43:01.0000] like an admin bar that is position absolute, or leaderboard ads that are outside of the centered/bordered container, or the accessibility links that are hidden for screen [05:45:00.0000] In any case, at the very least the articles on the page should be nested below the main page title otherwise your markup has the wrong structure (which will show up as an incorrect outline) [05:45:01.0000] we can agree that title of the blog should be an h1, which will take its order correctly just by being an h1... let it be in section#header's header with the site slogan, logo, and primary nav... better for designing real website header sections [05:46:00.0000] paulgendek: That's not how headers work in HTML5 [05:46:01.0000] paulgendek: Why should one section's header apply to one of its siblings? [05:46:02.0000] Any header can be a

as long as you get the structual elements right [05:47:00.0000] or for the sake of having a clean outline, and the elements where they need to be in the dom for design, should we then hide the h1 in body header, and use an h2 for the logo in the actual header section of the design? [05:47:01.0000] paulgendek: Why is it a section, even? [05:47:02.0000] because it's a page layout section [05:47:03.0000] Note that
is not a sectioning element [05:47:04.0000] "header, content, sidebar, footer" [05:47:05.0000] about.validator.nu, etc. going down now [05:48:00.0000] exactly, section is sectioning, which is why i apply the id to sections, for building the page structure [05:48:01.0000] paulgendek: "section" is _not_ a page layout section. It is a semantic section. [05:48:02.0000] gsnedders|work: the spec says that it's flow AND structural [05:48:03.0000] Doing
My blog instead of
My blog is both unnecessary and wrong [05:49:00.0000] Unless you further nest all other content in that outer section [05:49:01.0000] headers are only for sections and articles [05:49:02.0000] the body is a section. [05:49:03.0000] body's header is reserved for top content imo [05:50:00.0000] at least, in blog layout [05:50:01.0000] maybe for a single page site or basic app [05:50:02.0000] sure [05:50:03.0000] paulgendek: I don't see how to reach your opinion from the spec or from logic [05:50:04.0000] but not for magazine/blog layout that can be fairly complex [05:50:05.0000] http://tools.ietf.org/html/bcp19#section-2.4 how can HTML be a character set? [05:50:06.0000] But I have toi go now... [05:50:07.0000] jgraham: i'll continue to work on it, and we'll see how it goes, thanks [05:52:00.0000] paulgendek: If you absolutely don't want the body's header to be that, which I think is what you do, then don't you want the whole page in that section? [05:53:00.0000] i see it as the body has a header and a footer, and sections... and all of the structural sections are grouped in a div#container to remain ambiguous... body's header for top most meta-ish content, body footer for debug/version info, javascripts, etc [05:54:00.0000] the #container can then constrain the width and be centered of all the structural sections of the layout [05:54:01.0000] What is wrong with using header for the header? [05:55:00.0000] i think there is confusion between
as the top part of a section, and #header which is the container for the structural header of a design [05:56:00.0000] gsnedders|work: because any css done to the body header for a complex design's header section would nest a lot of selectors based off of "header", which would inherit down to all other headers, you would need to use body > header:first-child, which is much more to type, and not really much more semantic, body is more like a canvas for sections imo [05:57:00.0000] paulgendek: Per the spec, body is a section. [05:57:01.0000] i'm a front-end developer and i bring this up from tons of real-world application in common practice [05:57:02.0000] paulgendek: And yes, you do end up with complex selectors in CSS doing almost anything [05:58:00.0000] hmm, all the charset registry lists are offline? [05:58:01.0000] it's much easier for css and javascript to target #header than body > header:first-child [05:58:02.0000] btw, if I'm right and we cannot use the Unicode mapping algorithm for charsets we prolly need to register more aliases [05:58:03.0000] then give the body > header:first-child an id of header [05:58:04.0000] annevk42: http://www.iana.org/assignments/character-sets wfm [05:58:05.0000] but then you'll have
which is then pointless [05:59:00.0000] might as well go back to it being a div [05:59:01.0000] at least it's 3 less chars on open/close [05:59:02.0000] paulgendek: @id has no semantic value anyway, id="foobar" is just as meaningful. It's just a way to select the element. [05:59:03.0000] gsnedders|work, that's not a mailing list [05:59:04.0000] and the container with #header is a structural section of a design, not necessarily the header of body [05:59:05.0000] "12:52 < jgraham> [...]
My blog is both unnecessary and wrong" - sounds like an accurate description of many blogs [05:59:06.0000] annevk42: It is a charset registry list :D [06:00:00.0000] sure, but not the one I'm after [06:00:01.0000] /me wants ietf-charsets⊙io archives [06:00:02.0000] annevk42: I don't think I ever managed to subscribe to that when I tried [06:01:00.0000] it is supposed to be based here: http://mail.apps.ietf.org/ietf/charsets/threads.html [06:01:01.0000] That's long gone [06:01:02.0000] doh [06:01:03.0000] where's Julian when you need him [06:02:00.0000] Not on IRC? :) [06:04:00.0000] paulgendek: can't you select every header but the body header on the basis that they're all children of another element? [06:05:00.0000] annevk42, http://lists.w3.org/Archives/Public/ietf-charsets/ [06:05:01.0000] Lachy_: That's older and more dead :P [06:05:02.0000] I noticed taht [06:05:03.0000] that [06:05:04.0000] beowulf: Only with :not that really isn't well supported [06:06:00.0000] yeah, that's not the archive I was looking for [06:06:01.0000] I wonder if we should email iana⊙io [06:06:02.0000] header {foo:bar}; body header {foo:none} # no? [06:06:03.0000] I guess [06:06:04.0000] beowulf: That would work, but you then have to override everything [06:07:00.0000] gsnedders|work: c'est la vie [06:07:01.0000] beowulf: Only CSS 3 Selectors can be relied upon, body > header and :not(body) > header [06:07:02.0000] Once [06:08:00.0000] If you want to style a specific element, why would you not simply use id? [06:08:01.0000] Philip`: iurro [06:09:00.0000] Philip`: Because @id is _so_ yesterday, and XPointer is _so_ the future. [06:09:01.0000] like i said, maybe in 2015 or whatever, but right now i have to cover all bases, which means older browsers and screen readers, and search crawlers [06:09:02.0000] i don't have bare headers in my documents, so it doesn't apply to me :) [06:09:03.0000] the top header is usually nested in a
or some such [06:09:04.0000] paulgendek: Why doesn't @id work now? [06:10:00.0000] paulgendek: If you want to get one specific element, it seems the most sensible way [06:10:01.0000] header for body, section, article, aside, nav, sure... but that doesnt mean the same as the structural header of the design [06:11:00.0000] paulgendek: but that has nothing to do with css, surely? [06:11:01.0000] paulgendek: How does that effect selecting elements? [06:11:02.0000] it affects css, with the > and :first-child selectors not being supported in ie6 [06:11:03.0000] paulgendek: Well how would you select a specific p element for example? [06:12:00.0000] nested selectors [06:12:01.0000] section#primary article header p [06:12:02.0000] And when you can't do that? [06:12:03.0000] (Because of other siblings, or whatever) [06:13:00.0000] paulgendek: doesn't that get all the p's in a header? how do you get just one specific p? [06:13:01.0000] if there is a reason to target that specific p versus the others, then the article would need an id to specify why it's different [06:13:02.0000] section#primary article.different-view header p [06:13:03.0000] How does that not apply to the case of getting one specific header? [06:13:04.0000] What is so different between the two cases? [06:13:05.0000] seems like going backwards though [06:13:06.0000] paulgendek: that would still get all the p's in the header, no? [06:13:07.0000]
[06:14:00.0000] just looks wrong [06:14:01.0000] Is that more backwards than
? [06:14:02.0000]
[06:14:03.0000]
? [06:14:04.0000]
[06:14:05.0000]
[06:14:06.0000] beowulf: ok, header#banner, sure, but that still is not the header of a layout where you would place user meta, logo, site name/slogan, primary menu, search, etc [06:15:00.0000] Well what header is it? [06:15:01.0000] I'm sure you can come up with a sensible id for it., [06:15:02.0000] the body's header, which i said i'd prefer to reserve for accessibility links, banners, admin bars, etc [06:15:03.0000] paulgendek: when i do html like that i create a div called banner, and place header in it with all that other stuff as siblings [06:15:04.0000] div with id banner [06:15:05.0000] the "header" that contains a design's header elements are a section, imo [06:16:00.0000] selecting the header is then simply #banner header [06:16:01.0000] why a div? it's structural [06:16:02.0000] div/span seems to be block/inline containers for design [06:17:00.0000] is that not what we're talking about? [06:18:00.0000] perhaps i'm being stupid [06:19:00.0000] /me thinks that trying to describe what design you're talking about in words is never going to be as effective as showing us it graphically [06:20:00.0000] i have to run to work [06:20:01.0000] but here's the design in question [06:20:02.0000] http://paulgendek.com/html5 [06:21:00.0000] We could avoid a lot of debate if the new elements were named , , etc, instead of
,
, etc, because that would avoid the problem of people having different intepretations of what the names mean [06:21:01.0000] well i don't agree that
is the same a section#header [06:21:02.0000] but we will continue this some other time ;) [06:21:03.0000] It isn't. [06:22:00.0000] Ooh, there's the "WAHTWG" image again [06:22:01.0000] Philip`: That would be too easy to confuse. How about naming them after animals instead? [06:22:02.0000] /me wonders where it originally came from [06:23:00.0000] Dashiva: Hmm, that could work [06:24:00.0000] Dashiva: Then we could add easy-to-remember conventions to distinguish different classes of elements, e.g. inline elements can be named after insects while block elements are named after mammals, so you never have to wonder "hmm, is

Heading

Foo

Bar