| 00:00 | <othermaciej> | although really his discussion mixes up versioning and XHTML2-conflict-avoidance as if they were related when to me they seem clearly not |
| 00:00 | <annevk> | http://lists.w3.org/Archives/Member/member-cdf/2006Feb/0009.html has the original |
| 00:00 | <annevk> | Member-only :/ |
| 00:02 | <Hixie> | http://dev.w3.org/2006/cdf/cdi-framework/#importing |
| 00:02 | <Hixie> | i'm amused that you forgot you put that in there :-) |
| 00:03 | <annevk> | see backlog? |
| 00:03 | <othermaciej> | I was hoping to find the original reference but blech, I'm not gonna cite a member-only link |
| 00:04 | <Hixie> | oh, i see |
| 00:04 | <Hixie> | my bad :-) |
| 00:04 | Hixie | hides |
| 00:04 | <othermaciej> | actually I just realized this isn't directly relevant to the message I am drafting but I'll save it for future reference |
| 00:05 | <annevk> | you could also ask dbaron is you could cite it in full |
| 00:08 | <Lachy> | othermaciej, if you respond again, can you trim the CC list so I'm not in it, unless you're quoting me |
| 00:09 | <othermaciej> | Lachy: sorry for the spam |
| 00:13 | <Lachy> | changing the MIME type as Larry suggests creates other problems too. Surely there's been enough trouble with the lack of support the the existing XHTML MIME type, that introducing a new one isn't in anyone's best interest |
| 00:13 | <othermaciej> | it seems like such an obviously bad idea that it's actually hard work to articulate all the various things wrong with it |
| 00:15 | <Lachy> | that's the same problem with all of Larry's proposals |
| 00:16 | <othermaciej> | well, that's the sort of thing you get when you start proposing and getting emotionally attached to solutions before actually getting informed about the problem |
| 00:17 | <Lachy> | Larry apparently isn't interested in understanding the problem http://lists.w3.org/Archives/Public/public-html/2009Feb/0365.html |
| 00:18 | <Dashiva> | It should've stopped at "The xhtml2 group created the problem, they get to solve it" |
| 00:24 | <dbaron> | annevk, othermaciej: any particular advise on where in the thread I should repost it? |
| 00:25 | <othermaciej> | dbaron: I don't think it needs to be reposted, unless someone seriously proposes changing the XHTML namespace URI in HTML5 |
| 00:26 | <Lachy> | dbaron, you could just post it to www-archive as it could be useful for future reference |
| 00:26 | <othermaciej> | dbaron: but it seems like no one is seriously advocating that, just using it as a pretext to argue for other brands of crazy |
| 00:26 | <dbaron> | Lachy, I was just about to do that |
| 00:26 | <othermaciej> | dbaron: I briefly got confused about it though |
| 00:28 | <Lachy> | I wonder how long it will take for others to use those arguments against chameleon namespace as an argument for why XHTML2 can't change namespace either, under the delusion that XHTML2 is a compatible successor to XHTML1 |
| 00:29 | dbaron | has not exactly been following the thread (especially having been on an actual vacation this weekend) |
| 00:29 | <dbaron> | Lachy, http://lists.w3.org/Archives/Public/www-archive/2009Feb/0065.html |
| 00:29 | <Lachy> | wait... you go on vacations without having to constantly read email? I don't know how that's possible. |
| 00:30 | <dbaron> | No, a vacation means only reading email 20-30 minutes a day. :-P |
| 00:54 | <Hixie> | it really doesn't matter if someone proposes changing the namespace for xhtml5, since that would never get the support of the browser vendors. |
| 00:56 | <othermaciej_> | Hixie: you're just part of the evil browser-vendor/search-engine-provider conspiracy |
| 00:56 | <Hixie> | well, yes |
| 00:57 | <Dashiva> | Don't forget natural-language-processing |
| 00:57 | <othermaciej> | how dare you provide users with information |
| 00:57 | <Hixie> | (if you can have a conspiracy in public) |
| 00:57 | <Dashiva> | #whatwg-secret-treehouse |
| 00:58 | <jcranmer> | I'm not part of said conspiracy! |
| 00:58 | Hixie | finally gets around to reading the namespace thread |
| 00:58 | <Hixie> | good lord people |
| 00:58 | <Hixie> | wtf |
| 00:59 | Dashiva | ponders semantic email |
| 01:01 | <Hixie> | this entire thread is completely crazy |
| 01:01 | <Hixie> | there's no mime type for DOM manipulation |
| 01:01 | <jcranmer> | Dashiva: ew! |
| 01:01 | <Hixie> | there's no DOCTYPE for DOM manipulation |
| 01:01 | <Hixie> | there's no attribte for DOM manipulation |
| 01:01 | <jcranmer> | I'm scared of real HTML mail (e.g., someone using marquee) |
| 01:01 | <Hixie> | all you have is a namespace and a tag name |
| 01:01 | <Lachy> | I don't even know what a "DOCTYPE for DOM manipulation" is |
| 01:02 | <othermaciej> | Hixie: I think Larry has a somewhat muddled view of element identity |
| 01:02 | <Hixie> | createElementNS(ns, tag) is ALL we have to do versioning with |
| 01:02 | <jcranmer> | Hixie: onload="document.getElementById('foobar').addChild(document.createTextNode('DOM manipulation'));" ? |
| 01:02 | <Hixie> | any solution that requires more than an ns/tag pair is not going to work |
| 01:03 | <jcranmer> | you're doing DOM manipulation in an attribute |
| 01:03 | <othermaciej> | I'm not really clear on his idea of "different languages sharing the same vocabulary" |
| 01:03 | jcranmer | returns to the thread debating whether or not infinity is a valid mathematical concept |
| 01:04 | <Dashiva> | I think it's like <i> getting redefined, kinda |
| 01:04 | <othermaciej> | from my naiive implementor point of view, a QNAME uniquely defines the operational behavior of the element |
| 01:06 | <othermaciej> | I can see how you could make a different language which has elements from the same set, but (say) a different set of nesting rules so that the valid grammar is different |
| 01:06 | <othermaciej> | but that's really not what is going on with HTML5 and XHTML2 |
| 01:08 | <Hixie> | Lachy: if you want an example of an xhtml2-1 incompatibility, consider <img src=404 alt="XHTML1">XHTML2</img> |
| 01:08 | <Lachy> | oh, yeah, that's another one I forgot about |
| 01:09 | <Hixie> | hey, xhtml2 hasn't got profile="" either |
| 01:09 | <Lachy> | though RB is likely to ignore that one too since, IIRC, he has repeated proposed that we adopt the same approach in XHTML5 |
| 01:09 | <Lachy> | (or maybe it was someone else) |
| 01:09 | <Hixie> | oh i wouldn't worry about arguing with rob |
| 01:09 | <Hixie> | i just meant in general |
| 01:10 | <Lachy> | I'm not going to argue with him. He's at the top of my do-not-respond list for a reason |
| 01:11 | <Hixie> | i don't understand how xhtml2 UAs are supposed to distinguish between <style src=""> and <span src=""> for the purposes of rendering |
| 01:11 | <Hixie> | is src="" just special on <style>? |
| 01:12 | <Hixie> | another example would be <script src="a" type="text/javascript" srctype="text/vbscript"/> |
| 01:13 | <Lachy> | XHTML2 doesn't have a <script> element any more |
| 01:13 | <Lachy> | they renamed it to handler |
| 01:13 | <Hixie> | oh well then even better |
| 01:13 | <Hixie> | should that show as an image or run script? |
| 01:13 | <Hixie> | xhtml2 is very confusing. why does <style> use src="" but <link> use href=""? |
| 01:14 | <Hixie> | man this spec is so much like html4 |
| 01:14 | Hixie | closes the tab |
| 01:15 | <Lachy> | we have similar problems with HTML5 too. Why have <script src=""> for scripts, but <link href=""> for stylesheets? (I know we're stuck with it for historical reasosn, but it's inconsisent) |
| 01:15 | <Dashiva> | Like html4 in all the wrong ways? :) |
| 01:22 | <Hixie> | Lachy: in html5, src and href are not global attributes |
| 01:22 | <Hixie> | Lachy: so there's no conflict or confusion |
| 01:22 | <Hixie> | well, there's confusion. |
| 01:22 | <Hixie> | but not to the same level. |
| 01:23 | <Hixie> | in xhtml2, <link href="data:text/css,link { display: block; }" rel="stylesheet" src="image"/> is a non-clickable image |
| 01:24 | <Hixie> | whereas <style src="data:text/css,style { display: block; }" href="document">link</style> is a clickable text link |
| 01:24 | <Hixie> | which makes no sense to me |
| 01:24 | <othermaciej> | wait, what? |
| 01:24 | <Hixie> | exactly. |
| 01:25 | <Lachy> | heh |
| 02:05 | <john_fallows> | does the concept of origin require a port to be specified, such as http://www.whatwg.org:80 or can a default port be assumed, such as http://www.whatwg.org ? |
| 02:06 | <Hixie> | an origin typically consists of {scheme, host, port} |
| 02:06 | <Hixie> | there's no syntax, though, so I'm not sure exactly what you mean |
| 02:06 | <Hixie> | an origin is just a triple |
| 02:06 | <Hixie> | that is, the "concept of origin" has no syntax |
| 02:24 | <john_fallows> | ic, so what would the canonical syntax be for postMessage event.origin, or the Origin header for cross-site access control, http://www.whatwg.org (without the default port) ? |
| 02:24 | <Hixie> | oh that's the serialisation of the orogin |
| 02:24 | <Hixie> | origin, even |
| 02:25 | <Hixie> | http://www.whatwg.org/specs/web-apps/current-work/#unicode-serialization-of-an-origin |
| 02:25 | <Hixie> | http://www.whatwg.org/specs/web-apps/current-work/#ascii-serialization-of-an-origin |
| 02:25 | <Hixie> | (depending on where you are using it) |
| 02:26 | <Hixie> | and then to parse it back into an origin triple, see http://www.whatwg.org/specs/web-apps/current-work/#origin |
| 02:26 | <john_fallows> | thanks Hixie, that's the section i was looking for :-) |
| 02:27 | <Hixie> | which you use depends on what you're doing |
| 02:28 | <Hixie> | (postMessage() and Origin: use different ones, in particular) |
| 02:37 | <john_fallows> | the HTML 5 postMessage section seems to reference Unicode / ASCII serialization as the value of event.origin, and cross-site access control refers back to ASCII serialization in HTML 5 |
| 02:37 | <john_fallows> | http://www.w3.org/TR/access-control/#origin-ascii |
| 02:38 | <john_fallows> | how do postMessage and Origin header serializations differ? |
| 02:38 | <john_fallows> | sorry, how do postMessage event.origin and Origin header serializations differ? |
| 02:39 | <Hixie> | postMessage is unicode, i think, and origin header is ascii, i hope |
| 02:40 | <john_fallows> | ok, now i understand - thanks |
| 02:41 | <Hixie> | was the spec not clear on this? |
| 02:41 | <Hixie> | i should fix it if not |
| 02:42 | <john_fallows> | no that's fine - i had missed the serialization section, and then just wanted to clarify your comment regarding a difference between postMessage and cross-site, since I was asking specifically about the need to include a default port when you mentioned there was a difference |
| 02:42 | <john_fallows> | it's all very clear now though - thanks for the clarification. |
| 02:43 | <Hixie> | make sure to follow links in the spec |
| 02:43 | <Hixie> | they're there for a reason :-) |
| 02:44 | <Hixie> | that is, there should be a link from the postMessage() bit that talks about origin to the relevant origin sections -- if there isn't, of course, please do tell me! |
| 02:44 | <john_fallows> | ok, i'll certainly let you know if I discover a problem like that |
| 02:46 | <franksalim> | There is no link in the WebSocket handshake section which mentions the origin value "encoded as US-ASCII" |
| 02:47 | <franksalim> | by the way :-) |
| 02:51 | <Hixie> | really? crap. |
| 02:51 | Hixie | looks |
| 02:51 | <Hixie> | franksalim: the "origin" value at that point is a string, not a triple |
| 02:52 | <Hixie> | the serialisation happens in http://www.whatwg.org/specs/web-apps/current-work/#dom-websocket |
| 02:52 | <Hixie> | step 4 |
| 02:52 | <Hixie> | i guess i could make that clearer |
| 02:55 | <franksalim> | Hixie: i see now. i think what confused me was that the serialization and the conversion to lowercase were split into two steps |
| 02:56 | <Hixie> | hm, why do i lowercase it |
| 03:09 | <Hixie> | franksalim: i updated the spec, is it clearer now? |
| 03:12 | <franksalim> | Hixie, yes it is, and thank you for clarifying |
| 03:12 | <Hixie> | Lachy: i finally worked out what was causing the multipage copy to have weird problems with the update detection stuff |
| 03:12 | <Hixie> | franksalim: my pleasure! |
| 03:12 | <Hixie> | Lachy: it should be better now |
| 03:29 | Hixie | reluctantly joins the versioning debate in attempt to reduce the amount of e-mail he gets |
| 07:29 | <zcorpan> | wow <head><isindex> is really weird in opera |
| 07:37 | <annevk> | not too weird given how <isindex> works elsewhere |
| 07:40 | <zcorpan> | annevk: the weird part is that we get two <head>s |
| 07:44 | <annevk> | oh, hadn't noticed that :D |
| 07:44 | <zcorpan> | our html parser creates a siamese twin! |
| 07:44 | <annevk> | we should have a easter egg that whenever there's three <head>s we start playing that scene from Monty Python |
| 07:57 | <annevk> | so Hixie really did stop the discussion or did I just stop receiving e-mail at that point? |
| 07:57 | <Hixie> | which discussion |
| 07:58 | <annevk> | the one with lots of things I don't care about: media types, namespaces, and versioning |
| 07:59 | <Hixie> | ah |
| 07:59 | <Hixie> | i guess so |
| 08:54 | <annevk> | Hixie, form submission should also use character encoding mapping rules |
| 08:54 | <annevk> | Hixie, e.g. iso88591 is windows1252 for the purposes of form submission |
| 08:55 | jgraham | discovers his email wasn't working properly yesterday |
| 08:55 | <annevk> | Hixie, do you want a bug on that? |
| 08:56 | <zcorpan> | hsivonen: <isindex> seems to add name="isindex" to all elements in the tree (except the <form>) in the gecko html5 build |
| 08:58 | <zcorpan> | hsivonen: also affects other documents that are parsed afterwards |
| 09:01 | <Hixie> | annevk: yes please |
| 09:02 | zcorpan | exposes the isindex bug at v.nu |
| 09:03 | <zcorpan> | hsivonen: (sorry i just wanted to see if it worked, didn't mean to bump priority) |
| 09:04 | <zcorpan> | hmm found another bug |
| 09:05 | <zcorpan> | guess i should file them |
| 09:07 | <Hixie> | annevk: i disagree with that bug actually. The browser should use ISO 8859-1 if that's what the author set. |
| 09:08 | <zcorpan> | Hixie: so what should be done with curly quotes? |
| 09:08 | <annevk> | that would break pages Hixie |
| 09:09 | <annevk> | (it's what Opera at the moment) |
| 09:09 | <Hixie> | opera bug#? |
| 09:09 | <Hixie> | zcorpan: what do you mean? |
| 09:09 | <annevk> | CORE-10326 |
| 09:10 | <zcorpan> | Hixie: author says <meta charset=iso-8859-1>, user submits U+201C |
| 09:11 | <Hixie> | annevk: that site isn't coming up for me |
| 09:12 | <annevk> | Hixie, given that all other browsers do the mapping, we're going to follow them to avoid future issues |
| 09:13 | <Hixie> | zcorpan: 4.10.15.3:6.1 |
| 09:14 | <Hixie> | annevk: is there an actual site affected by this that's still up? |
| 09:14 | <zcorpan> | Hixie: ? |
| 09:14 | <Hixie> | zcorpan: "For each character in the entry's name and value that cannot be expressed using the selected character encoding, replace the character by a string consisting of a U+0026 AMPERSAND character (&), one of more characters in the range U+0030 DIGIT ZERO (0) to U+0039 DIGIT NINE (9) representing the Unicode codepoint of the character in base ten, and finally a U+003B SEMICOLON character (;)." |
| 09:14 | <annevk> | Hixie, the most recent dupe was 2008-09 complaining about not being able to use Spanish characters |
| 09:15 | <annevk> | Hixie, other dupes are complaining about not being able to use the euro sign for login/password; it definitively seems like something that needs to be fixed |
| 09:16 | <Hixie> | annevk: i believe that it happens, i'm just looking for an actual working site i can test |
| 09:16 | <Hixie> | the most recent dup has no url |
| 09:19 | <annevk> | can't help you with that I'm afraid |
| 09:19 | <Hixie> | well if there's no actual affected site... |
| 09:20 | <annevk> | I didn't say that ;) |
| 09:21 | <Hixie> | how have you determined that this is the actual underlying problem in these cases if you have no site to check this against? :-) |
| 09:21 | <zcorpan> | "If the token has an attribute called "action", set the action attribute on the resulting form element to the value of the "action" attribute of the token." - html5lib doesn't seem to do this |
| 09:24 | <zcorpan> | (isindex) |
| 09:25 | <jgraham> | zcorpan: File a bug or send email or something |
| 09:25 | <jgraham> | Or provide a patch, I guess ;) |
| 09:25 | <zcorpan> | doesn't support prompt either |
| 09:26 | <annevk> | has HTML5 been updated to include these attributes? |
| 09:26 | <annevk> | I seem to recall there were more |
| 09:28 | <annevk> | are there any two browsers that implement <isindex> in the same way that do not share a rendering engine? |
| 09:30 | <zcorpan> | annevk: not that i'm aware of :) |
| 09:31 | <zcorpan> | annevk: i don't get any other attributes do anything in ie |
| 09:32 | <annevk> | enctype? |
| 09:32 | <annevk> | method target? |
| 09:32 | <zcorpan> | nope |
| 09:32 | <zcorpan> | they end up on the <input> |
| 09:38 | <zcorpan> | jgraham: http://code.google.com/p/html5lib/issues/detail?id=87 |
| 09:42 | <annevk> | I think the same goes for <keygen>, that element is a mess as well |
| 09:42 | <jgraham> | zcorpan: Thanks |
| 09:43 | jgraham | notes that the versioning thread is pure undistilled crazy |
| 09:43 | <jgraham> | s/un// |
| 09:44 | <annevk> | so much material for standards suck and we're not doing anything with it... |
| 09:45 | <Hixie> | keygen is for april |
| 09:45 | <Hixie> | i think i have enough info on it now |
| 09:45 | <annevk> | cool |
| 09:51 | <zcorpan> | "It will do so for all feeds that contain ten or more entries all with the same id." - http://intertwingly.net/blog/2009/02/15/White-House-Feed-Now-Declared-Invalid |
| 09:51 | <zcorpan> | so 9 duplicate ids is ok but 10 is not? |
| 09:51 | <annevk> | it's not like feed readers do anything with duplicate IDs anyway |
| 09:52 | <jgraham> | zcorpan: Yeah that seems odd |
| 09:52 | <annevk> | e.g. Google Reader has an interesting bug I found out about |
| 09:52 | <annevk> | I have feed 1 (RSS) and 2 (Atom) |
| 09:52 | <annevk> | I redirect 1 to 2 (via 301) |
| 09:53 | <annevk> | Google Reader treats them as separate feeds and always has entries marked as unread in both despite them using the same Atom ID |
| 09:54 | <annevk> | there's probably a feed reader or two that handles this stuff correctly, but whether it is worth all the trouble of making Atom IDs required... |
| 09:56 | <Philip`> | zcorpan: I think 2..9 is a warning (because it's probably incorrect but the checker can't tell if you're actually using it legitimately), and 10 is an error because you're almost certainly using it incorrectly |
| 09:57 | <jgraham> | Is the use case for repeated atom:ids in a single feed an entry that gets udated without deleting the previous atom:entry |
| 09:58 | <jgraham> | Philip`: But it is not actually invalid per spec is it? So it seems wrong to make a validator signal that it is invalid even though it almost certianly is |
| 10:06 | <Philip`> | jgraham: A useful validator seems more valuable than a validator that's pedantically correct according to the spec in weird edge cases that will never occur in practice |
| 10:12 | <annevk> | Philip`, if you make <canvas> testcases again maybe include tests for drawImage(document.createElement("canvas"), ...) together with globalCompositeOperation does the right thing |
| 10:13 | <annevk> | ok, that sentence does not run quite right, but you get the idea |
| 10:13 | <Philip`> | annevk: You mean an empty not-yet-drawn to canvas, in particular? |
| 10:14 | <annevk> | Philip`, yes, in theory such a canvas is transparent black and has a certain width/height |
| 10:14 | Philip` | adds it to his list of things to maybe write tests for eventually |
| 10:51 | Hixie | reluctantly posts on sam's blog |
| 10:52 | <Lachy> | Hixie, the new defintion for <small> is better |
| 10:52 | <Hixie> | good |
| 11:36 | <hendry> | what are the reasons again for not using iso8601? http://www.whatwg.org/specs/web-apps/current-work/multipage/infrastructure.html#valid-date-string # someone i think has told me this before, but I forgot :) |
| 11:38 | <Philip`> | hendry: Do you mean using the whole of ISO 8601, or a subset of it? |
| 11:38 | <hendry> | Philip`: either or |
| 11:39 | <hendry> | Philip`: i guess html5 define a subset as the whole thing is too complex? |
| 11:39 | Philip` | (based on about zero knowledge) would guess that the whole of ISO 8601 is too complex a thing to require, and what's in the spec maybe is a compatible subset |
| 11:40 | <hendry> | ok, sounds familiar |
| 11:40 | <zcorpan> | iirc iso8601 allows crazy things like comments |
| 11:40 | <Philip`> | Also I'd guess (based on never having read it or anything so I'm probably wrong) ISO 8601 doesn't define how to parse invalid strings |
| 11:46 | <gsnedders> | zcorpan: No, it doesn't |
| 11:46 | <gsnedders> | hendry: Backwards compatible |
| 11:46 | <gsnedders> | hendry: Stuff already uses whitespace instead of "T" for example |
| 11:46 | <gsnedders> | Also, ISO8601 allows three different forms for the date |
| 11:47 | <gsnedders> | (Which is stupid because nobody cares about any apart from YYYYMMDD) |
| 11:47 | <hendry> | thanks guys! i need to write a little blog entry next |
| 11:50 | <Philip`> | gsnedders: You mean nobody except for everybody who cares about dd/mm/yyyy and mm/dd/yyyy etc? |
| 11:54 | <Lachy> | Philip`, people who care about using dd/mm/yyyy and mm/dd/yyyy can be ignored because those formats are annoyingly ambiguous |
| 11:57 | <jgraham> | Lachy: It is easier just to ignore people who use mm/dd/yyyy since that is crazy. dd/mm/yyyy is only marginally worse than yyyymmdd |
| 11:58 | Lachy | will use yyyy-dd-mm from now on just to confuse everyone |
| 12:00 | <jgraham> | mmyyyydd ftw |
| 12:00 | <Philip`> | I like ymydymyd |
| 12:00 | <jcranmer> | mydymyd is better |
| 12:00 | <jcranmer> | er, mydymydy |
| 12:01 | <jcranmer> | 02102079! |
| 12:01 | <Philip`> | How about mydyrmydy, where 'r' stands for a random digit? |
| 12:01 | <jcranmer> | f! |
| 12:12 | hendry | blogs http://dabase.com/blog/HTML5_is8601/ |
| 12:14 | <zcorpan> | hendry: gsnedders says it doesn't allow comments |
| 12:14 | <zcorpan> | so i was probably thinking of something else, like email addresses |
| 12:17 | <annevk> | or media types, wtf |
| 12:17 | <jgraham> | http://hydracen.com/dx/iso8601.htm |
| 12:19 | <annevk> | btw, http://forum.krijnhoetmer.nl/irc-logs/html-wg/20081202#l-100 |
| 12:19 | <annevk> | was the original discussion for that issue |
| 12:19 | <krijnh> | forum ? |
| 12:19 | <krijnh> | What the |
| 12:20 | annevk | copied the URL from Google |
| 12:20 | <zcorpan> | http://foobar.krijnhoetmer.nl/irc-logs/html-wg/20081202#l-100 |
| 12:20 | <hendry> | annevk: thanks, i'll add that |
| 12:20 | <svl> | exploit spammers try forum.*.tld and in time that filters through to google |
| 12:20 | <svl> | rel="canonical" to the rescue - or something... :) |
| 12:21 | <krijnh> | Nah, redirects |
| 12:21 | <krijnh> | One moment.. :) |
| 12:21 | <krijnh> | F5 ? |
| 12:21 | <krijnh> | (Can't test it from my within my LAN) |
| 12:22 | <svl> | wfm |
| 12:23 | <krijnh> | foobarshizzle. gets redirected? |
| 12:23 | krijnh | wonders why he didn't do that already.. |
| 12:23 | <annevk> | ja |
| 12:24 | <zcorpan> | hsivonen: "<table>test" results in a text node for each character |
| 12:39 | <zcorpan> | hmm? <ruby> is one of the escape-foreign-lands tags? |
| 12:46 | <annevk> | i wonder how much would break if we make that list empty |
| 12:47 | <annevk> | i.e. never escape |
| 12:49 | <hsivonen> | zcorpan: yeah, foster parenting text sucks |
| 12:49 | <hsivonen> | I just gave my HTML5 lecture |
| 12:49 | <hsivonen> | time to see if the recording is any good |
| 12:51 | <jgraham> | hsivonen: How did it go? |
| 12:51 | <hsivonen> | jgraham: I mostly didn't get mixed up in my words and I got a few questions |
| 12:52 | <annevk> | any interesting questions? |
| 12:52 | <hsivonen> | the questions were about when can HTML5 features be used, are there JS implementations, relationship of new elements to microformats and RDFa, extensibility |
| 12:53 | <annevk> | fun :) |
| 12:53 | <zcorpan> | hsivonen: also see earlier in the logs; i've broken v.nu (nothing will validate until you restart the service) |
| 12:54 | hsivonen | looks |
| 12:54 | <hsivonen> | good thing my demos were all playback and didn't rely on live services working |
| 12:54 | <zcorpan> | :) |
| 12:54 | <jgraham> | :) |
| 12:55 | <hsivonen> | http://html5.validator.nu/?doc=http%3A%2F%2Fwww.whatwg.org WFM |
| 12:55 | <zcorpan> | hsivonen: http://validator.nu/?doc=http%3A%2F%2Fwww.whatwg.org |
| 12:56 | <hsivonen> | zcorpan: now, that's weird |
| 12:56 | <zcorpan> | hsivonen: i only validated <isindex> on v.nu, not h.v.nu |
| 12:56 | <zcorpan> | hsivonen: they're different instances, right? |
| 12:56 | <hsivonen> | zcorpan: aah. yeah. I've fixed this bug in svn. |
| 12:56 | <hsivonen> | zcorpan: yes, different instances |
| 12:57 | <zcorpan> | http://bugzilla.validator.nu/show_bug.cgi?id=446 |
| 12:58 | <hsivonen> | what happened was that the singleton that represents empty attributes accidentally got attributes added to it |
| 12:58 | <zcorpan> | why did that affect <body>? |
| 13:05 | <hsivonen> | zcorpan: if there was a body without attributes, the empty attributes singleton got used |
| 13:05 | <zcorpan> | interesting |
| 13:05 | <hsivonen> | restarted with the same bogus code. uploading a fix now |
| 13:05 | <annevk> | no questions about why we didn't use XForms and such? |
| 13:05 | <hsivonen> | annevk: no |
| 13:06 | <hsivonen> | annevk: the exercise for the course relies on server-side XForms--not client side |
| 13:06 | <hsivonen> | (with Orbeon) |
| 13:06 | <annevk> | ok |
| 13:07 | <hsivonen> | zcorpan: thanks for finding the bug. I thought it only affected the buffered mode--not the streaming mode |
| 13:09 | <zcorpan> | welcome :) |
| 13:16 | <hsivonen> | now, to distribute the recording of my lecture in a proper HTML5 <video> dogfood fashion, I should encode it as Theora... |
| 13:16 | hsivonen | wants a Thusnelda build of XiphQT or ffmpeg2theora |
| 13:21 | jgraham | wonders if there is a term for a single, unmached codepoint that is supposed to form a surrogate pair |
| 13:22 | <hsivonen> | jgraham: lone surrogate? |
| 13:22 | <jgraham> | I guess. Or simply "surrogate" |
| 13:22 | <annevk> | surrogate character? |
| 13:23 | <annevk> | nm |
| 13:23 | <annevk> | http://blogs.msdn.com/michkap/archive/2005/07/27/444101.aspx |
| 13:23 | <annevk> | :) |
| 13:24 | <annevk> | it's a surrogate code point |
| 13:25 | <Philip`> | jgraham: There aren't any codepoints that are supposed to form a surrogate pair |
| 13:26 | <Philip`> | There's just codepoints that happen to be numerically equal to some of the values used by UTF-16 to encode surrogates, but those codepoints are not supposed to be anything at all :-) |
| 13:27 | Philip` | suggests using the term "U+D800..U+DFFF" |
| 13:27 | <annevk> | that's a range |
| 13:27 | Philip` | suggests using the term "a codepoint in the range U+D800..U+DFFF" |
| 13:29 | <jgraham> | Philip`: The Unicode FAQ uses the term surrogate code point |
| 13:29 | <hsivonen> | at least when encoded as UTF-16 they are code *units* |
| 13:32 | <Philip`> | jgraham: Sure, but "a codepoint in the range U+D800..U+DFFF" might be clearer/obviouser depending on the context and audience |
| 13:32 | Philip` | can imagine people interpreting "surrogate code point" as meaning "a code point that has to be represented using surrogates" or something like that |
| 13:36 | <hsivonen> | eww. my microphone sucked |
| 13:37 | <hsivonen> | my words are comprehensible, but it sounds like I'm in some kind of muffled container |
| 13:38 | <zcorpan> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/9 has interesting results in browsers - in particular it makes ie8 freeze |
| 13:40 | <zcorpan> | (maybe because the parser tries to insert stuff into something that has been garbage collected) |
| 13:46 | <zcorpan> | i think firefox is correct for that test |
| 13:46 | <zcorpan> | (the second script isn't executed because the script is not inserted to the document) |
| 13:52 | <zcorpan> | Hixie: i uploaded http://software.hixie.ch/utilities/js/live-dom-viewer/saved/10 but your server is saying 404 not found |
| 13:54 | <zcorpan> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cbody%20onload%3D%22w(1) - executes per html5 and in moz but not in ie/opera/webkit |
| 14:13 | <aroben> | hi ap! |
| 14:13 | <aroben> | hi everyone! |
| 14:13 | <ap> | hi aroben! |
| 14:28 | <hsivonen> | grr. Keynote's QuickTime export insists on writing to the boot volume even when told to save to another volume |
| 14:35 | <zcorpan> | hsivonen: sometimes in the live dom viewer with the gecko html5 build, the doctype disappears in the dom view |
| 14:36 | <hsivonen> | zcorpan: I'll have to look into that. thanks |
| 14:36 | <hsivonen> | is there a way to debug XHR serialization without an echoing script? |
| 14:38 | <hsivonen> | "Or, if this fails because the Document cannot be serialized act as if data is null." |
| 14:38 | <hsivonen> | I want to see when that happens |
| 14:38 | hsivonen | starts live HTTP headers |
| 14:40 | <hsivonen> | annevk: the spec should be clearer on when a documunt cannot be serialized |
| 14:40 | <hsivonen> | annevk: it seems that "xmlns:dublincore" local name in no namespace serializes in Gecko |
| 14:44 | <annevk> | the idea is that the spec uses the same serialization tricks as HTML5 for XML |
| 14:44 | <annevk> | unless the content is HTML in which case it uses the HTML serialization tricks |
| 14:49 | <zcorpan> | "</head><!>" puts the comment after head |
| 14:50 | <zcorpan> | but "</body><!>" doesn't put the comment after body |
| 14:50 | <zcorpan> | why this assymetry? |
| 14:50 | <zcorpan> | Hixie? |
| 15:21 | <zcorpan> | hsivonen: http://validator.nu/?doc=http%3A%2F%2Fwww.damowmow.com%2Fplayground%2Fdemos%2Fflash%2F001.html reproducably gives an internal error |
| 15:21 | <zcorpan> | hsivonen: it also has two warnings about the hyphens in <![endif]--> |
| 15:25 | <zcorpan> | hsivonen: it seems "<param>" with the html4 schema reproduces the internal error |
| 15:35 | <rubys> | jgraham: ping? |
| 15:35 | <jgraham> | rubys: Hej |
| 15:35 | <rubys> | I don't have a moderation queue... I just say I do to scare off spammers. |
| 15:36 | <jgraham> | rubys: Oh, well I got some message about it. Was there some button I was supposed to click? Maybe I didn't look too hard |
| 15:36 | <rubys> | "submit" |
| 15:37 | <jgraham> | Oh right. I obviously got confused by the UI :) |
| 15:37 | <jgraham> | Nevermind :) |
| 15:37 | Philip` | wonders (while we're on the subject of rubys's site) whether rubys saw the issue on http://intertwingly.net/stats/internalsearches.html |
| 15:37 | <rubys> | Cool! Nope. Will take a look at that. |
| 15:39 | <rubys> | I have a log of 3238 spam attempts in the past 24 hours (much lower than normal). None contain jgraham - I scanned in a case insensitive manner |
| 15:41 | <rubys> | 7 contain 'graham', but all are clearly spam. |
| 15:42 | <rubys> | jgraham: at this point, I don't believe I have the comment; please consider resubmitting it. |
| 15:44 | <jgraham> | rubys: At this point I don't believe that *I* have the comment either ;) I will think about rewriting it. |
| 15:47 | <rubys> | in the log, I see attempts to post at 12:24Z and 13:25Z, but both match Shelley's IP and browser; I don't see any POST request to that entry in between. |
| 15:50 | <rubys> | which is weird, as you don't see the message about moderation until a POST request is issued |
| 15:51 | <jgraham> | You are looking at Yesterday, right? |
| 15:51 | <jgraham> | s/Y/y/ |
| 15:52 | <rubys> | oh, I was looking at today. I don't keep logs of aborted comments for more than 24 hours |
| 15:52 | <gsnedders> | Lachy (if you see this): you want YYYYMMDD without the separators to confuse people |
| 16:16 | <Philip`> | gsnedders: I suggest an alternate way of dealing with the world, in which confusing people is not a goal |
| 16:16 | <gsnedders> | Philip`: That's boring. |
| 18:41 | <annevk> | haha, what is RB smoking? I want too! |
| 19:11 | <annevk> | and it continues |
| 19:13 | <annevk> | on a related note, it seems the people on www-tag only half read e-mails or so |
| 19:14 | <annevk> | no wonder they are so frequently confused |
| 19:20 | <annevk> | (hi there, mr last week!) |
| 19:23 | gsnedders | wonders who Mr Last Week is, again |
| 19:24 | <peter_12> | The HTML 5 spec doesn't give guidance on the intuitive choice between strong and em tags. Is there any way to choose? |
| 19:26 | <annevk> | peter_12, doesn't it state that one is for marking up importance and one for emphasis and gives examples and stuff? |
| 19:27 | <peter_12> | annevk: it gives examples in both but it doesn't make it clear |
| 19:27 | <gsnedders> | It does |
| 19:27 | <annevk> | peter_12, interesting, could you explain why you think it is not clear? |
| 19:28 | <peter_12> | annevk: the explanations hinge on the difference between the phrases "stress emphasis" and "strong importance" |
| 19:28 | <peter_12> | those English phrases are not intuitively different. |
| 19:29 | <peter_12> | I'd say they could be synonymous. |
| 19:29 | <annevk> | you don't think emphasis and importance are different things? |
| 19:29 | <peter_12> | no |
| 19:29 | <annevk> | then again, I'm not a native speaker, so what do I know |
| 19:30 | <peter_12> | also using the word "emphasis" in the em description is a circular definition |
| 19:30 | <peter_12> | the same for using "strong" in the strong tag description |
| 19:30 | <annevk> | element names do not mean anything so I'm not sure it is |
| 19:31 | <peter_12> | In English we emphasis something in our speech because it is important. That blurs those two phrases together immediately. |
| 19:32 | <annevk> | e-mail public-html-comments⊙wo or whatwg⊙wo or so? |
| 19:35 | <gsnedders> | The difference is in the nuances. They are different words. |
| 19:36 | <gsnedders> | karlcow had some good diagram of the overlap somewhere on w3.org/2007 |
| 19:36 | <peter_12> | gsnedders: nuance is not a good spec quality |
| 19:36 | <peter_12> | people misinterpret nuance all the time |
| 19:37 | <gsnedders> | Emphasizing something is giving special attention to it; importance is a statement that is crucial |
| 19:42 | <peter_12> | gsnedders: that is not a clear distinction to me |
| 19:42 | <peter_12> | you want to give special attention to something that is crucial |
| 19:43 | <peter_12> | annevk: sent to whatwg⊙lwo |
| 19:45 | <annevk> | cheers |
| 22:25 | yecril71 | is reading forms.html#application/x-www-form-urlencoded-encoding-algorithm |
| 22:25 | <yecril71> | Is it correct that extraneous characters should be sent as &XXXX;? |
| 22:25 | <yecril71> | MSIE7 sends them as &#XXXX;, i.e. NCRs. |
| 22:28 | yecril71 | �s computer has just run out of virtual memory |
| 22:44 | yecril71 | pings from http://msdn.microsoft.com/en-us/library/ms533033(VS.85).aspx#ctl00_rs1_WikiContent_ctl00_Container |
| 22:46 | <annevk> | seems that Kristof just pointed out that issue yecril71 |
| 23:10 | Hixie | adds a "+" to his regexp on the .../saved script so that /saved/10 works... |
| 23:11 | <Hixie> | i wonder if the rdfa people will actually reply to http://lists.w3.org/Archives/Public/public-rdfa/2009Feb/0098.html |
| 23:45 | yecril71 | �s name is Kristof, actually. |
| 23:45 | yecril71 | has a "cri" in it |