| 01:17 | <Hixie> | the more i try to fix this scripting thing the more involved the edit becomes |
| 01:17 | <Hixie> | this is becoming quite the diff |
| 01:17 | <Hixie> | i don't even remember what the original problem was that i was fixing |
| 01:18 | <Hixie> | it's like someone saying "i'd like you to fix the off-white colour on the wall here" and finding that the off-white is caused by rot which is caused by a slow leak which is caused by faulty insulation in the attic and that the slow leak has also weaked the foundations and that you have to basically pull the whole house down and rebuild it |
| 01:22 | <neatnik> | Hixie: and then you discover that the house was built on an ancient Indian burial ground, and the land is tainted with thousands of restless spirits |
| 01:23 | <MikeSmith> | Hixie: not looking forward to reading that diff |
| 01:24 | <kfish> | neatnik, hence the off-white colour on the walls |
| 01:25 | <neatnik> | Interestingly, you never hear about computers being haunted |
| 01:25 | <Hixie> | neatnik: oh we already know about the burial ground |
| 01:25 | <MikeSmith> | Hixie: at least now that you're using Geoffrey's spec-gen thing, the IDs won't get cascade-changed throughout the whole spec |
| 01:25 | <Hixie> | neatnik: this is html, after all |
| 01:25 | <neatnik> | Hixie: lol |
| 01:25 | <Hixie> | MikeSmith: yeah really |
| 01:27 | <MikeSmith> | If we awaken the spirits, maybe they'll bring some peyote with them. |
| 01:27 | <MikeSmith> | that would be an upside |
| 01:27 | <Hixie> | dude i need my mind totally clear to write this stuff |
| 01:28 | <Hixie> | i don't think shooting up would help the spec :-P |
| 01:29 | <MikeSmith> | peyote will clear your mind, as well as your stomach. it will bring ideas straight from your belly and back up through your nostrils |
| 01:30 | <MikeSmith> | good medicine must always have some kind of adverse side effects |
| 01:30 | <MikeSmith> | but I digress |
| 01:30 | <MikeSmith> | ..as usual |
| 01:30 | MikeSmith | goes to drink some tea |
| 01:35 | <MikeSmith> | Hixie: btw, I was doing a bit with dfn.js yesterday |
| 01:35 | <MikeSmith> | that's a sweet little piece of work |
| 01:35 | <MikeSmith> | every spec should use that |
| 01:41 | <Hixie> | hehe |
| 01:41 | <Hixie> | my biggest problem with it is it doesn't support "open in new tab" |
| 01:41 | <Hixie> | what were you doing with it? |
| 01:45 | <MikeSmith> | Hixie: adding it to my document-conformance draft |
| 01:45 | <Hixie> | ah cool |
| 01:45 | <MikeSmith> | works particularly nicely for back-refs for the datatypes stuff |
| 01:46 | <MikeSmith> | btw, I've never been explicitly told we can't use any scripts in W3C TR docs |
| 01:46 | <MikeSmith> | and I'm not going to ask for permission.. |
| 01:48 | <Hixie> | let's hope the script doesn't introduce the risk of an xss |
| 01:48 | <MikeSmith> | ..so I think it would be good have have dfn.js added, e.g., to some webapps specs |
| 01:49 | <Hixie> | hsivonen: i'm shocked that it's possible (modulo your proposal) to parse html without depending on the dom |
| 01:49 | <Hixie> | hsivonen: are you really sure you can't cause the parser to trigger mutations that make no sense? |
| 01:49 | <hsivonen> | Hixie: is there a reason why foster-parented stuff isn't dropped on the floor if the parent of the table is gone? |
| 01:50 | <Hixie> | not really |
| 01:50 | <hsivonen> | Hixie: pretty sure that that's the only show-stopper problem |
| 01:51 | Hixie | is not 100% convinced, but will have to look closer at some point |
| 01:51 | <Hixie> | when writing AAA in particular i found crashes in safari that were due to hyatt not thinking of certain changes you could make to the dom from script |
| 01:52 | <hsivonen> | Hixie: we could make things slightly simpler if foster-parented stuff was dropped if the parent is gone |
| 01:52 | <Hixie> | (so i wouldn't be surprised if such a radical change as you propose reintroduced such errors) |
| 01:52 | <Hixie> | hsivonen: send mail |
| 01:52 | <Hixie> | hsivonen: i don't really care either way |
| 01:54 | <heycam> | MikeSmith, is dfn.js what makes the <dfn>s clickable in the html5 spec? |
| 01:54 | <heycam> | if so, then i want it :) |
| 01:54 | heycam | already borrowed dl.switch from the whatwg style sheet |
| 01:54 | <MikeSmith> | heycam: yeah |
| 01:56 | <MikeSmith> | heycam: I recommend going ahead and adding it to the WebIDL spec |
| 01:56 | <heycam> | there aren't any copyright licensing terms in http://www.whatwg.org/specs/web-apps/current-work/dfn.js |
| 01:57 | <Hixie> | heycam: what license do you want it under? |
| 01:57 | <heycam> | Hixie, anything sufficiently liberal for me to use it |
| 01:57 | <heycam> | i don't mind if attribution is required |
| 01:58 | <Hixie> | public domain ok? |
| 01:58 | <heycam> | sure |
| 01:58 | <heycam> | that's easiest |
| 01:58 | <Hixie> | i hereby grant http://www.whatwg.org/specs/web-apps/current-work/dfn.js into the public domain |
| 01:58 | <heycam> | (though i know some people think public domain is a bad idea) |
| 01:58 | heycam | wonders whether that is sufficient |
| 01:58 | <heycam> | could you add it as comment at the top of dfn.js? |
| 01:58 | <Hixie> | well if you want something more, let me know |
| 01:59 | <Hixie> | done |
| 01:59 | <heycam> | thanks |
| 01:59 | <Hixie> | (neither me nor google will ever sue you over copying that file) |
| 02:00 | <Hixie> | i have to say, the people who are fighting so hard for their copyrights have really made my life noticeably more annoying |
| 02:00 | <Hixie> | because everyone is always asking me for copyright statements and so forth |
| 02:00 | <Hixie> | i wish copyright didn't eixst |
| 02:00 | <Hixie> | exist |
| 02:00 | <hsivonen> | Hixie: were there compat issues with making CDATA sections act like CDATA sections in HTML? |
| 02:00 | <Hixie> | my life would be far better |
| 02:00 | <heycam> | in australia at least explicit copyright statements don't need to be made |
| 02:00 | <Hixie> | hsivonen: yes, pretty sure |
| 02:01 | <Hixie> | heycam: i mean statements saying that copying is ok |
| 02:01 | <heycam> | Hixie, ah |
| 02:01 | <hsivonen> | Hixie: having to check the tree builder state is a problem for tokenizing ahead of tree building |
| 02:01 | <hsivonen> | as in saving the tokens of speculative script, style and img parse |
| 02:02 | <Hixie> | hsivonen: isn't that a problem for things like <script> anyway? or can you move those parts into the tokeniser effectively |
| 02:02 | <hsivonen> | Hixie: document.write is an obvious problem that cannot go away |
| 02:02 | <hsivonen> | CDATA maybe could still made into a non-problem |
| 02:03 | <Hixie> | well if you have proposals, send mail. but i highly doubt we can support "<![CDATA[" as an escaping string in regular text/html |
| 02:03 | <hsivonen> | Hixie: moving bailing out of foreign content into the tokenizer doesn't work |
| 02:03 | <hsivonen> | Hixie: ok |
| 02:03 | <Hixie> | it's probably dangerous enough in svg/mathml as it is |
| 02:03 | <Hixie> | i could be convinced to remove it altogether, but you'd have to get that past the svgwg, which will be interesting |
| 02:04 | <Hixie> | not sure what the status is wrt svgwg and svg-in-text/html |
| 02:04 | <Hixie> | is the ball in our court? |
| 02:04 | <Hixie> | they never replied to my e-mail |
| 02:04 | <Hixie> | but we had that meeting in mandelieu |
| 02:04 | <Hixie> | which went nowhere |
| 02:04 | <hsivonen> | Hixie: they didn't respond to my review either |
| 02:04 | <heycam> | Hixie, we recently just started discussing it again |
| 02:04 | <heycam> | (see public-svg-wg) |
| 02:04 | <Hixie> | ah ok |
| 02:04 | <Hixie> | cool |
| 02:04 | <heycam> | i expect we'll discuss at at the thursday telcon |
| 02:04 | <heycam> | *it at |
| 02:05 | <Hixie> | k |
| 02:05 | <Hixie> | i have to say, for all the problems that it introduces, the whatwg model definitely does make things faster |
| 02:06 | <Hixie> | svg-in-text/html was resolved in about a month once i started working on it, from start to finish (notwithstanding the svgwg's feedback, of course) |
| 02:06 | <Hixie> | including responding to all feedback |
| 02:08 | <Hixie> | (all of which is ironic given our timetable vs the w3c's timetables, heh) |
| 02:11 | <kfish> | Hixie, I was wondering, what tools do you use to summarize discussions in email? or do you just cut and copy? |
| 02:11 | <Hixie> | pine |
| 02:11 | <Hixie> | it has a multi-message reply facility |
| 02:11 | <heycam> | do you use an external editor? or the builtin one? |
| 02:12 | <Hixie> | i use pine's editor generally, but if i need to do anything special i escape out to emacs |
| 02:12 | <Hixie> | i often fix up formatting in dozens of e-mails at once using emacs |
| 02:12 | <Hixie> | then do the actual replying in pine |
| 02:12 | <heycam> | i suppose at least the pine editor's key bindings are similar to emacs |
| 02:12 | <Hixie> | pine has a slightly better mail reflowing mechanism |
| 02:12 | <heycam> | but apart from that, /me shudders :) |
| 02:13 | <Hixie> | heh |
| 02:13 | <Hixie> | i'd use emacs if i could work out how to configure it to reflow blocks of indented mail sanely |
| 02:13 | <kfish> | cool, thanks |
| 02:13 | <Hixie> | but my current configuration isn't good at distinguishing where the indent level changes |
| 02:13 | <heycam> | M-x reflow-mails-as-hixie-likes-them |
| 02:13 | <Hixie> | and ends up screwing everything up into one paragraph |
| 02:13 | <Hixie> | which is suboptimal |
| 02:14 | <heycam> | my workflow is to use mutt with vim as the editor, and i highlight the stuff to reflow and do :!par |
| 02:14 | <heycam> | i just haven't figured out how to make par put two spaces between sentences instead of one |
| 02:14 | <Hixie> | ctrl-j in pine just figures out what the current paragraph is, reflows it, and moves the cursor to below the paragraph |
| 02:15 | <heycam> | that seems nice |
| 02:15 | <Hixie> | so i can reflow an entire mail paragraph-by-paragraph just by hitting ctrl-j..j..j..j..j..j..j... |
| 02:15 | <Hixie> | which is awesome |
| 02:15 | <Hixie> | it screws up with mail from emacs gnu mail users (they indent using " >" instead of "> " which pine doesn't know what to do with) |
| 02:16 | <Hixie> | and it screws up with mail from gmail (pine can't parse the html from gmail and ends up not indenting anything) |
| 02:16 | <Hixie> | so for those two cases i jump to emacs, do a quick regexp search-replace, and go back |
| 02:16 | <Hixie> | all under screen(1) over ssh |
| 02:17 | <heycam> | i should use screen i suppose |
| 02:17 | <heycam> | i don't find it too annoying to quit mutt and restart it later though |
| 02:17 | <Hixie> | (my mail client configuration has basically not changed since 1996, except for adding screen about 6 years ago) |
| 02:18 | <heycam> | heh |
| 02:18 | <Hixie> | well the great thing with screen is that i can be half-way through an e-mail and i can just close my laptop, move elsewhere, reopen it, and the cursor will still be at the same point in the e-mail |
| 02:18 | <Hixie> | which is awesome |
| 02:18 | <Hixie> | if a cat sits on my keyboard, i can just pull out another laptop and continue and it's like nothing happened, without having to move the cat |
| 02:18 | <heycam> | lol |
| 02:18 | <Hixie> | (i did that once and the cat was like, wtf) |
| 02:19 | <Hixie> | ("you're supposed to pet me. nobody else has been able to ignore me when i sit on the keyboard. who are you, freak?") |
| 02:37 | <heycam> | nice, http://dev.w3.org/2006/webapi/WebIDL/ with clickable <dfn>s now |
| 02:38 | <heycam> | i like how google thinks that www.w3.org/TR/WebIDL/ is written by a mr (or mrs) "C Changes" -- http://www.google.com/search?q=web+idl |
| 02:40 | <Hixie> | yes, you should be receiving a letter any day now announcing that we've filed to have your name changed |
| 02:40 | <Hixie> | sorry for the delay, but we have to do it by postal mail for legal reasons so it takes a while |
| 02:40 | <heycam> | such is google's influence on the world |
| 02:41 | <Hixie> | well it's just that your parents made a mistake, we're just correcting it now |
| 02:41 | <heycam> | :) |
| 02:42 | <heycam> | google actually has crawled the web and generated an immense genealogy tree and determined that my great^5-grandfather was illegitimate and was given the wrong name |
| 02:43 | heycam | marvels at what computers can do these days! |
| 02:43 | <Hixie> | :-) |
| 02:44 | <Philip`> | Looking at e.g. http://www.google.com/search?q=site:www.ietf.org+rfc it seems to be wrong about as often as it's right |
| 02:44 | <Hixie> | that it gets any right at all is pretty amazing |
| 02:44 | <Hixie> | i wonder how it works |
| 02:45 | <Hixie> | i guess i could look |
| 02:46 | <Philip`> | It seems kind of silly to display that information at all, when there's so little confidence that it's right |
| 02:46 | <heycam> | i wonder whether it looks at <address> |
| 02:46 | <heycam> | (if one exists) |
| 02:46 | <Philip`> | I've seen Google sometimes say how many replies there were to forum posts in its results, which is a much more useful thing since it seems to get it right and lets you ignore unanswered queries |
| 02:47 | <Hixie> | well well well |
| 02:48 | <Hixie> | designMode = 'on' doesn't stop other frames from calling your scripts |
| 02:53 | <Hixie> | woah |
| 02:53 | <Hixie> | designMode = 'on' doesn't stop ANY scripts! |
| 02:54 | <Hixie> | where did i get the idea that it did from? |
| 03:15 | MikeSmith | thanks heycam for tip about par(1) |
| 05:50 | <zcorpan> | Hixie: opera supports <![CDATA[ in text/html, and we haven't got many bugs about it |
| 06:01 | <MikeSmith> | CDATA sections in text/html would be great to have |
| 06:01 | <MikeSmith> | would make life easier for author/developers who want to include HTML/markup examples in their docs |
| 06:01 | <MikeSmith> | e.g., spec editors |
| 06:05 | <zcorpan> | MikeSmith: you could use <xmp> |
| 06:07 | <MikeSmith> | zcorpan: <xmp> supported across browsers? IE? |
| 06:07 | <zcorpan> | yes |
| 06:08 | <MikeSmith> | I guess I remember hearing that now, but don't remember why it's not part of HTML5, then |
| 06:10 | <zcorpan> | http://lists.w3.org/Archives/Public/public-html/2007Dec/0120.html |
| 06:14 | <MikeSmith> | zcorpan: thanks |
| 06:14 | MikeSmith | sees not reply from Hixie there, wonders which of his issue folders <xmp> might fall in |
| 06:19 | <zcorpan> | http://www.google.com/search?q=site%3Ahttp%3A%2F%2Fcanvex.lazyilluminati.com%2Fmisc%2Fcgi%2Fissues.cgi+xmp |
| 06:19 | MikeSmith | finds http://krijnhoetmer.nl/irc-logs/html-wg/20070822#l-301 |
| 06:19 | <MikeSmith> | zcorpan: thanks |
| 06:21 | <zcorpan> | oh right, Philip` started with an html reference document |
| 06:22 | <zcorpan> | MikeSmith: perhaps you can steal some stuff from it |
| 06:22 | <MikeSmith> | zcorpan: what're your own current thoughts on making <xmp> conforming? good idea? bad idea? |
| 06:22 | <MikeSmith> | zcorpan: URL for Philip` s reference doc? |
| 06:22 | <zcorpan> | http://canvex.lazyilluminati.com/misc/ref/ref.html |
| 06:23 | <zcorpan> | MikeSmith: i don't mind it being conforming but i don't feel strongly about it |
| 06:23 | <MikeSmith> | OK |
| 06:23 | <zcorpan> | it's pretty useless in xhtml though |
| 06:23 | <MikeSmith> | yeah |
| 06:25 | <MikeSmith> | zcorpan: btw, splitting up input into multiple sections is still a high priority for me for the document-conformance draft.. I got sidetracked a bit messing around with the Datatypes section. among other things, borrowed Hixie's dfn.js and put it to use for backrefs for the datatypes |
| 06:26 | <MikeSmith> | e.g, http://www.w3.org/html/wg/markup-spec/#common.data.integer |
| 06:27 | <zcorpan> | MikeSmith: ok |
| 06:29 | <zcorpan> | MikeSmith: the entity table should omit entities without trailing ; |
| 06:31 | <MikeSmith> | zcorpan: added to my TODO list -- which is, so far, all stuff from you |
| 06:31 | <MikeSmith> | http://www.w3.org/html/wg/markup-spec/TODO |
| 06:31 | MikeSmith | needs to add a few more things to that... |
| 06:34 | <zcorpan> | MikeSmith: put it on the todo list that you have to add a few more things to it |
| 06:34 | <MikeSmith> | heh |
| 06:34 | <MikeSmith> | smartass |
| 06:59 | zcorpan | uploads http://simon.html5.org/dump/titles-with-line-breaks-all.txt |
| 06:59 | <zcorpan> | some have just a trailing line break |
| 07:45 | <MikeSmith> | trying to recall what the rationale is for allowing years with more than 4 digits |
| 07:51 | <Lachy> | MikeSmith, to avoid the imminent Y10k bug |
| 07:52 | <MikeSmith> | ah |
| 07:54 | <Lachy> | we could easily restrict it to 4 digit years now and then, if HTML is still around in 8000 years, let the future generations solve the problem by permitting 5 digit years |
| 07:55 | <roc> | I fear what HTML will have grown into in 8000 years |
| 07:55 | <jwalden> | but what if I want to talk about dates in the future in H. G. Wells's "The Time Machine"? |
| 07:55 | <jwalden> | (I kid, mostly) |
| 07:56 | <Lachy> | jwalden, you don't need the time element for that |
| 07:57 | <MikeSmith> | OK, also wondering what the rationale is for allowing month to be 00-12, instead of just 01-12 |
| 07:57 | <Lachy> | that's got to be a mistake |
| 07:58 | <MikeSmith> | Lachy: I'd hope so, but I don't think it is |
| 07:58 | <MikeSmith> | http://www.whatwg.org/specs/web-apps/current-work/multipage/infrastructure.html#months |
| 07:58 | <MikeSmith> | "Two digits, representing the month month, in the range 0 ≤ month ≤ 12" |
| 07:58 | <Lachy> | looks like a mistake to me |
| 07:59 | <Lachy> | it also says in the parsing rules "If month is not a number in the range 1 ≤ month ≤ 12, then fail." |
| 07:59 | <MikeSmith> | ah |
| 07:59 | <MikeSmith> | OK |
| 07:59 | <MikeSmith> | thanks |
| 07:59 | <Lachy> | Hixie, ^ |
| 07:59 | <MikeSmith> | definitely must be a mistake, then |
| 08:00 | MikeSmith | breathes a sign of relief.. then takes a bong rip |
| 08:01 | <MikeSmith> | somehow the effects of the bong rip are significantly more calming than a simple sign of relief |
| 08:15 | <Lachy> | Hixie, same error the the day in "Two digits, representing day, in the range 0 ≤ day ≤ maxday where maxday is the number of days in the month month and year year" |
| 08:43 | <Hixie> | Lachy: send mail or file a bug (or hope i remember) :-) |
| 08:47 | <Lachy> | ok |
| 09:06 | <Lachy> | I just booked my hotel in London for my New Years holiday :-) |
| 09:14 | jgraham | fails to believe that Lachy found somewhere cheap and nice |
| 09:18 | <Philip`> | http://www.crummy.com/2008/12/09/0 |
| 09:20 | <jgraham> | Philip`: Interesting |
| 09:21 | <annevk5> | maybe we should tell him html5lib is not as fast |
| 09:25 | <jgraham> | Maybe we should encourage takkaria to produce python bindings for hubbub and call it chtml5lib :) |
| 09:27 | <Lachy> | jgraham, I found a place that cost me £317 for 4 nights in Sussex Gardens |
| 09:28 | <Lachy> | arr Mon 29th, dep Fri 2nd |
| 09:43 | <annevk5> | just published: http://www.devarticles.com/c/a/HTML/Upgrading-from-HTML-to-XHTML/ |
| 09:43 | <annevk5> | seems like a copy from some other site though |
| 09:48 | <roc> | I can't bear it |
| 10:41 | BenMillard | cringes at the "upgrading" article. |
| 10:41 | BenMillard | is reminded of things he used to say about 3 years ago. :( |
| 11:13 | <gsnedders> | heyhey |
| 11:16 | <BenMillard> | hey G |
| 11:17 | <gsnedders> | how's you? |
| 11:18 | <BenMillard> | oh I'm alright |
| 11:19 | <BenMillard> | been catching up on the logs |
| 11:21 | <BenMillard> | Philip`, I think ref.html would look a bit cleaner without the grey border-bottom. |
| 11:22 | gsnedders | is on train going back home |
| 11:22 | gsnedders | left his jacket on the train from Cambridge to Peterborough and thus is cold |
| 11:23 | <BenMillard> | Philip`, using <strong> to highlight the relevant part of samples is good. It seems correct for <html> and <base> but incorrect for both <title> examples. |
| 11:24 | <zcorpan> | BenMillard: i don't think Philip`is maintaining ref.html :) |
| 11:25 | <gsnedders> | zcorpan: Do you know off-hand how Document.gEBID should work on XML multi-namespace documents? |
| 11:25 | <zcorpan> | gsnedders: same as in html documents |
| 11:26 | <gsnedders> | zcorpan: Only match @id on an element in the HTML namespace, or @id on an element in any namespace? |
| 11:26 | <BenMillard> | zcorpan, oh right. I thought it was the early days of a side project/demo type thing. |
| 11:27 | <zcorpan> | gsnedders: web dom core just talks about "ID" |
| 11:27 | <gsnedders> | BenMillard: A lot of Philip`'s things are early demos but with little intention of being maintained :) |
| 11:27 | <zcorpan> | gsnedders: what "ID" means depends on if you support xml:id, xhtml5, svg, etc |
| 11:27 | <gsnedders> | zcorpan: Ergh. Yay. :\ |
| 11:28 | <zcorpan> | (until someone sane comes along and defines XML ID 5) |
| 11:30 | <gsnedders> | Train wifi sucks. |
| 11:30 | <gsnedders> | Email won't send |
| 11:36 | <gsnedders> | wow! |
| 11:37 | <gsnedders> | 12KB/s! |
| 11:38 | gsnedders | wonders if it'd be quicker to use the web on his phone |
| 11:40 | <Lachy> | zcorpan, what improvements would a hypothetical XML ID 5 spec include? |
| 11:44 | <gsnedders> | The wifi was quick for a minute (like 75KB/s quick), but now back to < 10 |
| 11:48 | <annevk5> | zcorpan, it sort of makes sense if you define DOM ID and then have the respective specs define how they map to that |
| 11:55 | <zcorpan> | Lachy: it would make 'id' an ID attribute on any element and drop DTD ID and xml:id |
| 11:56 | <zcorpan> | annevk5: ok |
| 12:24 | <MikeSmith> | zcorpan: build for my draft doc now strips out semicolon-less entity names from the giganto table - |
| 12:24 | <MikeSmith> | http://www.w3.org/html/wg/markup-spec/#named |
| 12:31 | <zcorpan> | MikeSmith: nice |
| 12:38 | <annevk5> | might be nice to have a link there to skip over the table given it's so huge |
| 12:52 | <MikeSmith> | annevk5: OK, added |
| 12:52 | <MikeSmith> | it says Skip past table to next section (References) |
| 12:59 | <annevk5> | i should probably not have gotten involved in that link relationship debacle |
| 12:59 | <annevk5> | I don't at all see evidence for all the complexity they are adding |
| 13:00 | <annevk5> | evidence that justifies the complexity* |
| 13:03 | <MikeSmith> | annevk5: extensibility? ... |
| 13:03 | <MikeSmith> | that oftens seems to be the rationale for complexity |
| 13:03 | <zcorpan> | html5 link rels can be extended through the wiki |
| 13:04 | <annevk5> | MikeSmith, where is evidence that we need URIs for extensibility? |
| 13:04 | <annevk5> | MikeSmith, it's not like there are millions of relationships going around currently |
| 13:05 | <MikeSmith> | annevk5: true |
| 13:05 | <MikeSmith> | I'm not defending the rationale.. just speculating for what it might be |
| 13:05 | <annevk5> | they are just making a silly feature we should probably remove (Link HTTP header) even more complex and making HTML and Atom more complex in the process |
| 13:05 | <annevk5> | i mean, wtf |
| 13:06 | <annevk5> | MikeSmith, link looks nice btw |
| 13:32 | <Philip`> | BenMillard: Yeah, that grey border looks pretty weird |
| 13:32 | <Philip`> | BenMillard: For <title>, I think the point was to highlight the "contrast[...] with the top-level headers", hence highlighting the headers as well as the titles |
| 13:32 | <Philip`> | (Those examples came from the spec, I think) |
| 13:33 | <Philip`> | Then I got bored and nobody seemed to care a lot, so I gave up with it :-) |
| 13:36 | <MikeSmith> | OK, just for kicks: in the named-character-references table, I made all the names link to themselves |
| 13:36 | <MikeSmith> | e.g., http://www.w3.org/html/wg/markup-spec/#charref-clubsuit |
| 13:36 | <MikeSmith> | not sure how useful it is |
| 14:24 | <annevk5> | http://q42.nl/q42-lego-logo anyone up for doing the WHATWG logo in lego? |
| 14:28 | <Philip`> | annevk5: It's circular, and circles are hard in Lego :-( |
| 14:30 | <Philip`> | Also it's green and white, which are fairly rare colours in Lego |
| 14:33 | <annevk5> | I guess you don't read Dutch :) |
| 14:34 | <annevk5> | The idea is that you design it on your computer (Sjoerd wrote some software to calculate which bricks were needed for the rounding) using some software from lego.com and then you can order the result |
| 14:38 | <zcorpan> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=6296 |
| 14:43 | <Philip`> | I emailed the online Lego store some years ago, asking for enough bricks to build a replica of the Earth |
| 14:44 | <Philip`> | (Their web-based ordering form was limited to 9,999 units, which was far too many orders of magnitude below what I needed) |
| 14:44 | <Philip`> | s/replica/life-sized replica/ |
| 14:45 | <Philip`> | They were very nice and wrote back to me, though I can't quite remember what they said |
| 14:45 | <Lachy> | LOL |
| 14:47 | <MikeSmith> | OK, in the spirit of taking things to their logical extreme, I have a named-character table now with a 3rd column that shows the glyphs (or replacment char if you font doesn't have the glyph for whatever code point) |
| 14:47 | <Lachy> | http://www.theregister.co.uk/2008/12/04/toy_taliban/ |
| 14:47 | <MikeSmith> | e.g., http://www.w3.org/html/wg/markup-spec/#charref-expectation |
| 14:48 | <MikeSmith> | that's a great name |
| 14:48 | <MikeSmith> | euitable for use in Emotion ML |
| 14:48 | <Philip`> | Surely that should be http://www.w3.org/html/wg/markup-spec/#charref-ℰ |
| 14:49 | <Philip`> | Hmph, my IRC font doesn't like that character |
| 14:50 | <Lachy> | MikeSmith, for the authoring guide, I intend to use images of the glyphs instead of relying on people's fonts |
| 14:50 | <MikeSmith> | Lachy: yeah, that'll be better |
| 14:50 | <Lachy> | although the current chart I have for that doesn't use images yet |
| 14:50 | <Lachy> | I need to find a set of images with an appropriate copyright licence that I can use |
| 14:51 | <annevk5> | better to encourage people to fix issues than include workarounds, imo :) |
| 14:51 | <Lachy> | annevk5, there are other technical reasons to use images, even if people have the appropriate fonts |
| 14:51 | <zcorpan> | MikeSmith: s/Glyph/Character/ ? |
| 14:51 | <Lachy> | like ensuring that each glyph fits into the design properly, and doesn't overlap |
| 14:52 | <MikeSmith> | Lachy: zvon.org has some licenseable images |
| 14:52 | <MikeSmith> | mabye |
| 14:52 | zcorpan | wouldn't want 2000 images to be loaded |
| 14:52 | <zcorpan> | how about using a web font that includes all the glyphs? |
| 14:52 | <Lachy> | annevk5, e.g. see fflig in http://dev.w3.org/html5/html-author/charref |
| 14:53 | <Lachy> | zcorpan, that might work |
| 14:53 | <MikeSmith> | I think those character refs should just be published as a separate doc |
| 14:53 | <Philip`> | Why would anyone want to use a character reference when their own computer can't display the character? |
| 14:54 | <Philip`> | It seems better for authors to be discouraged from using such characters at all |
| 14:54 | <Lachy> | Philip`, I guess some people find them easier to input, or for use with non-Unicode encodings |
| 14:55 | <Lachy> | Philip`, for most cases, I agree, but there are some legitimate cases where' they're useful, such as for non-breaking spaces |
| 14:57 | <MikeSmith> | I don't think it's helpful at all to encourage authors to use named character references |
| 14:57 | <MikeSmith> | but we are kind of stuck with them now |
| 14:57 | <MikeSmith> | thanks to the MathML folks |
| 14:57 | <Philip`> | Lachy: I was meaning the characters themselves, not just the references to them |
| 14:57 | <MikeSmith> | giving names to every character under the sun |
| 14:58 | <Philip`> | (In particular, the characters which would cause problems when displayed in a character reference table because the author doesn't have glyphs for them) |
| 14:59 | <Lachy> | Philip`, there are some specialised use cases where such characters are useful for people, but for which your average user may not have all the characters available. |
| 14:59 | <Lachy> | e.g. a lot of those math related characters aren't available in default installs, but mathematicians could easily install them if needed |
| 15:10 | <MikeSmith> | Lachy: so what do you think about publish that table as a standalone separate doc? |
| 15:11 | <Lachy> | MikeSmith, I was intending to just make it a separate page within the authoring guide |
| 15:11 | <MikeSmith> | or better yet, asking the MathML group to do it so other can just reference it |
| 15:12 | <Lachy> | well, actually, when it's finished, I'm hoping it will be more like an app that let's people search interactively |
| 15:12 | <MikeSmith> | Lachy: what you mention above makes me think even more strongly that most users don't need that info |
| 15:12 | <MikeSmith> | mostly it's a just for math |
| 15:12 | <Lachy> | why? There are other sections that will get their own pages too |
| 15:13 | <zcorpan> | maybe hide the mathml entities behind a checkbox |
| 15:13 | <MikeSmith> | Lachy: that list frankly doesn't seem to me like something worth putting a whole lot of time into.. e.g., not worth making in a searchable app thing |
| 15:13 | <Lachy> | zcorpan, there will be various filters that let people show and hide characters based on all sorts of categories |
| 15:13 | <MikeSmith> | we already got the searchable thing at zvon.org |
| 15:14 | <MikeSmith> | http://zvon.org/other/charSearch/PHP/search.php |
| 15:14 | <MikeSmith> | not sure what the benefit in reinventing that wheel would be |
| 15:14 | <Philip`> | It seems much more useful to have a general-purpose Unicode character searching tool, which might happen to tell you the HTML entity or maybe some other tool does a simple number->name lookup, rather than making a clever specialised tool just for HTML entities |
| 15:15 | <MikeSmith> | http://zvon.org/other/charSearch/PHP/search.php?request=8496&searchType=2 |
| 15:15 | <MikeSmith> | Philip`: what you describe seems like just what the zvon tool does already |
| 15:18 | <MikeSmith> | anyway, the more I think about it, the more it seems to me that properly the MathML WG should be the ones to produce a normative list of those things |
| 15:18 | <MikeSmith> | I mean, the only reason we need most of those is because of the requirement to allow MathML in text/html, right? |
| 15:19 | <zcorpan> | right |
| 15:20 | <Lachy> | yeah, but it makes sense for the authoring guide to at least provide some information about them |
| 15:21 | <MikeSmith> | Lachy: yeah, but the authoring guide could just reference whatever document the MathML WG produced |
| 15:23 | <Lachy> | MikeSmith, the MathML specs are notoriously hard to find information about the charcter references in |
| 15:24 | <MikeSmith> | Lachy: i guess they could solve that problem by publishing a standalone doc with the list themselves |
| 15:24 | <MikeSmith> | maybe nobody's suggested yet that they do it |
| 15:24 | <Lachy> | but if they do produce a user friendly document, then I could. But until that happens, I see no reason to drop it from the guide |
| 15:25 | <MikeSmith> | OK, I'll shut up about it.. in the mean time, I guess I'll suggest it to them and see where we get |
| 15:25 | <Lachy> | but I still don't see why you want to keep this information so separate anyway. What problem are you trying to solve by removing them from the authoring guide? |
| 15:26 | <MikeSmith> | Lachy: the problem of having a giganto table in there with information that only a small percentage of users will probably have any use for |
| 15:27 | <MikeSmith> | .. or the problem of it appearing to encourage authors to use gazillions of named character references in their docs |
| 15:27 | <hallvors> | Dr. MikeSmith sets out to cure featuritis :) |
| 15:27 | <Philip`> | It's the same problem that would be solved by removing the complete works of Shakespeare from the authoring guide, if the works were in there in the first place - they're a waste of space for most of the audience, and make it harder to download or read or print out or anything |
| 15:28 | <MikeSmith> | hallvors: I learned about the problems of featuritis from one of my previous jobs :) |
| 15:28 | <Lachy> | it won't encourage them, because I intend to have a section in there that discusses the use of unicode encodings for their documents and explain why they should use the real characters instead of the references |
| 15:28 | <MikeSmith> | Lachy: they'll skip right past your words and go to the pretty pictures |
| 15:28 | <MikeSmith> | I would, at least |
| 15:29 | <Philip`> | Lachy: The probability of a reader reading the giganto table (and getting the wrong impression from it) seems much higher than the chance of them reading one particular paragraph that tells them to not use the table |
| 15:29 | <MikeSmith> | I know Rob Tyner of the MC5 probably would skip the words and go for the pictures also |
| 15:29 | MikeSmith | raises his Goblet of Rock to the MC5 |
| 15:29 | <Lachy> | I don't know who Rob Tyner is |
| 15:29 | <Lachy> | or what MC5 is |
| 15:29 | <MikeSmith> | Lachy: educate yourself, brother |
| 15:29 | <Philip`> | It's the precursor to MD5 |
| 15:29 | <MikeSmith> | heh |
| 15:30 | <MikeSmith> | Philip`: don't blaspheme |
| 15:30 | <Philip`> | Rob Tyner is a world-renowed cryptologist |
| 15:30 | <Lachy> | :-D |
| 15:30 | <Philip`> | *renowned |
| 15:32 | <MikeSmith> | renow is something you do when you when you forget something, then remember it. but just without the "k" |
| 15:33 | <MikeSmith> | "renow" is an abbreviation for "reknow" in that way.. as "whatcha" is an abbreviation for "what are you" |
| 15:33 | zcorpan | discovers that safari can play and pause animated gifs in <video>, and resizes the video to something when loading in the ahem ttf font |
| 15:33 | <Philip`> | How can you remember something unless you've already membered it? |
| 15:34 | MikeSmith | writes that one down |
| 15:34 | <zcorpan> | safari can also play a .bmp in <audio> |
| 15:34 | <MikeSmith> | "member" is a good word in a lot of ways |
| 15:35 | <Philip`> | zcorpan: Can you pass a .txt into <audio> and get text-to-speech from it? |
| 15:35 | <zcorpan> | Philip`: no |
| 15:36 | <Philip`> | zcorpan: That's a shame |
| 15:36 | <hallvors> | zcorpan: we should implement that :-) |
| 15:36 | <Lachy> | Philip`, you could if there were a codec available that did that |
| 15:37 | <zcorpan> | hallvors: seems better to just load the .txt in a browsing context and hit v |
| 15:39 | <Philip`> | zcorpan: That doesn't work for e.g. interactive games which want to speak to you |
| 15:41 | <zcorpan> | Philip`: true. but we have X+V for that, i think |
| 15:41 | MikeSmith | finds http://www.w3.org/TR/MathML2/byalpha.html |
| 15:45 | <jgraham> | fwiw authouring math without named characters would be even more nightmarish than it is with them |
| 15:46 | <MikeSmith> | jgraham: very glad I don't have to do it, then |
| 15:46 | <MikeSmith> | jgraham: are there not tools that allow you to select them from a menu? |
| 15:47 | <MikeSmith> | I mean, you can set up even Emacs to do stuff like that |
| 15:47 | <Philip`> | When you're writing a page full of equations, you really don't want to be selecting each character from a menu |
| 15:47 | <Philip`> | Also you can't even read the characters, because they all look the same when they're in a little font on the screen |
| 15:48 | <Lachy> | editors should just allow users to insert characters with keyboard shortcuts |
| 15:48 | <jgraham> | Lachy: The problem is remembering the keyboard shortcut for integral is harder tahn remembering ∫ or whatever |
| 15:49 | <Dashiva> | Use a postprocessor that lets you write entities and replaces them with characters? |
| 15:49 | <jgraham> | Dashiva: That counts as "even harder" |
| 15:50 | <jgraham> | (since in that case you could just use a preprocessor that lets you write something that looks like TeX and outputs mathml |
| 15:50 | <jgraham> | ) |
| 15:50 | <Lachy> | jgraham, the keyboard shortcut should be something that brings up a special input, where you start typing part of it's name, and it offers you a selection of matching characters to insert |
| 15:50 | <zcorpan> | wouldn't a sane author write something like TeX anyway? |
| 15:50 | <Lachy> | like an autocomplete system |
| 15:51 | <jgraham> | Lachy: Well I guess that might work but there are rather a lot of editors that are good at ascii input and rather few that implement the functionaility you describe |
| 15:51 | <jgraham> | So it suffers from being a "tools will save us" argument |
| 15:51 | <Philip`> | It's impossible to visually distinguish ∲foo from ∳foo when you're editing your document source, whereas it's easy to distinguish ∲foo from ∳foo |
| 15:52 | <jgraham> | zcorpan: Yes. |
| 15:53 | <MikeSmith> | and now I come across http://www.w3.org/TR/xml-entity-names/ |
| 15:53 | <jgraham> | (but that is not always possible particularly if you care what the markup looks like in the end) |
| 15:53 | <MikeSmith> | and now wondering why the above doc has not progressed beyond WD |
| 15:54 | <MikeSmith> | http://www.w3.org/TR/2007/WD-xml-entity-names-20071214/byalpha.html |
| 15:55 | <MikeSmith> | good enough for me |
| 15:56 | <zcorpan> | MikeSmith: doesn't contain AMP; |
| 15:57 | <MikeSmith> | zcorpan: so where did AMP; come from? |
| 15:57 | <MikeSmith> | something that's supported in browsers but not spec'ed? |
| 15:58 | <zcorpan> | MikeSmith: it's in the html5 spec |
| 15:59 | <MikeSmith> | zcorpan: but it's not something we want authors to actually use, right? |
| 15:59 | <MikeSmith> | just something that browsers need to support? |
| 15:59 | <zcorpan> | MikeSmith: it's conforming html5 |
| 16:00 | <MikeSmith> | zcorpan: so maybe it shouldn't be |
| 16:00 | <zcorpan> | MikeSmith: fine by me. convince Hixie :) |
| 16:00 | <MikeSmith> | as far a document-conformance goes |
| 16:01 | zcorpan | learns about &debug=1 in validator.w3.org |
| 16:02 | <Philip`> | I thought http://svn.whatwg.org/webapps/entities-unicode.inc was generated from some XML data provided as part of xml-entity-names |
| 16:02 | <Philip`> | (which includes AMP;) |
| 16:03 | <Philip`> | (but I may be wrong) |
| 16:05 | <MikeSmith> | Philip`: would be interesting to find out |
| 16:05 | <MikeSmith> | in the mean time, i'm dis-membering that table from my draft at least |
| 16:07 | <annevk5> | yeah, the XML guys included & |
| 16:07 | <MikeSmith> | annevk5: it doesn't appear to the in the version above |
| 16:07 | <MikeSmith> | is it maybe in an editor's draft? |
| 16:08 | <MikeSmith> | I seem to re-member now somebody posting a URL to one of their EDs a while back |
| 16:09 | <annevk5> | it's part of http://www.w3.org/2003/entities/2007xml/unicode.xml |
| 16:09 | <annevk5> | linked from http://www.w3.org/TR/xml-entity-names/#source |
| 16:10 | <annevk5> | your version is at least 6 months old anyway |
| 16:13 | <MikeSmith> | annevk5: OK |
| 16:28 | zcorpan | has now filed 1 bug and 2 enh req on validator.w3.org |
| 16:36 | <zcorpan> | hsivonen: i guess you may find this interesting http://www.w3.org/Bugs/Public/show_bug.cgi?id=6296 |
| 16:41 | <annevk5> | at some point we should just obsolete HTML 1-4 and require them to be parsed like HTML5 |
| 16:43 | <zcorpan> | annevk5: html5 allows validators to use legacy validation for legacy doctypes |
| 16:43 | <MikeSmith> | zcorpan: is the lack of AMP; QUOT; etc. the only difference you notice between the list in that MathML WG WD and the list that's in the HTML5 spec? |
| 16:44 | <MikeSmith> | I notice the unicode.xml file has all of those in a "html5-uppercase" set |
| 16:44 | <zcorpan> | MikeSmith: it was the only thing i looked for, so yes :) |
| 16:44 | <MikeSmith> | OK |
| 17:23 | <MikeSmith> | fyi, David Carlisle just updated the "XML Entity definitions for Characters" draft |
| 17:24 | <MikeSmith> | http://www.w3.org/2003/entities/2007doc/overview.html#sets |
| 17:24 | <MikeSmith> | http://www.w3.org/2003/entities/2007doc/html5-uppercase.html |
| 17:25 | <MikeSmith> | http://www.w3.org/2003/entities/2007doc/html5-uppercase.html |
| 17:25 | <MikeSmith> | http://www.w3.org/2003/entities/2007doc/byalpha.html |
| 17:26 | <MikeSmith> | the MathML guys rock |
| 18:07 | <hsivonen> | Hixie: I now see what you might have meant when you asked what about <script> in speculative tokenization |
| 18:08 | <hsivonen> | so yeah, speculating past <svg> or <math> is a problem |
| 19:01 | <yecril71> | I would like to remind that the whole problem with duplicated identifiers |
| 19:01 | <yecril71> | has been caused by using identifier references in table attributes. |
| 19:01 | <yecril71> | Replacing an identifier with a class name is not the way to go, |
| 19:02 | <yecril71> | unless the algorithm for resolving such references changes as well. |
| 19:02 | <yecril71> | But it looks like a misuse. |
| 19:03 | <yecril71> | Classes are not �Look, it�s there!� things. |
| 19:03 | <yecril71> | quot, amp and apos entities are all in the XML specification. |
| 19:04 | <yecril71> | There is no need for MathML to reproduce them. |
| 19:05 | <yecril71> | The logo WHATWG logo is a nologo. |
| 19:05 | <yecril71> | Instead of building it using LEGO, one should rather make a better one. |
| 20:15 | <avandenhoven> | Quick question. |
| 20:16 | <avandenhoven> | Does anyone know if there is a companion to the <noscript> tag. |
| 20:16 | Philip` | wonders why the Microsoft Connect site is so unbearably slow |
| 20:16 | <Philip`> | avandenhoven: Uh, like <script>? |
| 20:16 | <avandenhoven> | I don't mean the <script> tag per se. I mean tag that will render its TEXT contents if scripting is enabled. |
| 20:17 | <avandenhoven> | Often times we do all sorts of things to avoid the flash of unstyled (or wrongly styled) text that one sees if you're doing one thing when you have script and another when you dont. |
| 20:18 | <avandenhoven> | often we'll write a script class name to the html or body tag and hang style off that class. |
| 20:18 | <avandenhoven> | but there is that instant in which you have no script stles. |
| 20:18 | <avandenhoven> | er styles. |
| 20:18 | <avandenhoven> | it occured to me the other day that you can do <noscript><link rel="stylesheet"..." /></noscript> and that should work |
| 20:19 | <avandenhoven> | but why not also have <yesscript><link rel="stylesheet" href="myscriptedstyle.css" /></yesscript> |
| 20:19 | <Philip`> | What are the problems with <script>document.write('text contents')</script> or <script>document.write('<style>#foo{display:inherit !important}<\/style>')</script><div id=foo style=display:none>text contents</div> or something along those lines? |
| 20:20 | <avandenhoven> | Does that not totally creep people out? |
| 20:20 | <Philip`> | It doesn't creep me out :-) |
| 20:21 | <avandenhoven> | it should. |
| 20:21 | <Philip`> | (You could do it with DOM manipulation if you're opposed to document.write) |
| 20:21 | <avandenhoven> | I realize that there are JS ways to write content. |
| 20:21 | <avandenhoven> | its inherently obtrusive. |
| 20:21 | <yecril71> | I would suggest hiding the content using CSS and revealing it from script. |
| 20:22 | <avandenhoven> | Again, I realize that there are ways to do this now. I'm suggesting that an improvement would be to find a way to do it without resorting to JavaScript. |
| 20:22 | <yecril71> | By clearing the class attribute. |
| 20:23 | <yecril71> | Why exactly would that be an improvement? |
| 20:23 | <avandenhoven> | Easier to read. |
| 20:23 | <Philip`> | avandenhoven: <noscript>/<yesscript> seem inherently obtrusive too, since they're explicitly encoding script-related issues into the markup |
| 20:23 | <yecril71> | div[class="ifscript"] is easy to read. |
| 20:24 | <avandenhoven> | using <yesscript> (aside from its stupid name) is obtrusive but it also means that you don't need to translate the thing you want to include into JavaScript. |
| 20:25 | <yecril71> | But you do not, do you? |
| 20:25 | <avandenhoven> | Converting HTML or CSS into JavaScript is fine if its a one off, "I'm hacking my own site" kinda thinkg |
| 20:25 | <yecril71> | There is no need to include it into the script code. |
| 20:25 | <avandenhoven> | but if you're building something with a CMS, then you have to do something different. |
| 20:26 | <avandenhoven> | An arbitrary block of HTML or CSS is not guaranteed to make sense if you wrap it in a document.write('...'); |
| 20:26 | <yecril71> | Just say DIV.ifscript { DISPLAY:NONE } and you are done. |
| 20:27 | <yecril71> | You should never ever document.write, yep. |
| 20:27 | <Philip`> | avandenhoven: I think I'd have to agree that document.write('all the content') is not the best idea if you have a lot of content |
| 20:28 | <avandenhoven> | Another problem with relying on JavaScript to change things is timing. |
| 20:28 | <yecril71> | That means? |
| 20:28 | <avandenhoven> | I can't type fast enough and think coherently... give me a sec |
| 20:28 | <avandenhoven> | :) |
| 20:29 | <yecril71> | Well, this may be called a problem with timing, yes :-) |
| 20:29 | <avandenhoven> | Except in a handcoded, static site, you are going to want to do do things together. |
| 20:29 | <avandenhoven> | usually bootstrapping all your ajaxy widgets and what have you comes on DOM load. |
| 20:30 | <yecril71> | Bootstrapping? |
| 20:30 | <avandenhoven> | Adding JS to markup to make to do pretty things. |
| 20:30 | <yecril71> | This is not bootstrapping. |
| 20:30 | <avandenhoven> | that's what I call it. Sorry. |
| 20:30 | <yecril71> | Bootstrapping is when you want to compile Scheme. |
| 20:31 | <avandenhoven> | I'm pretty sure I never want to compile scheme :) |
| 20:31 | <yecril71> | Because you usually use Scheme to compile Scheme, you are pretty much stuck :-) |
| 20:31 | <avandenhoven> | anyways. if you do things "all at once" you're going to want to tend to put add your "isScripted" or what ever class onDomLoad with the rest of your JavaScript. |
| 20:32 | <avandenhoven> | Now you've suddenly changed what the user sees mid way through. |
| 20:32 | <yecril71> | I think the elements should be .ifscript at the beginning, and the script should clear that afterwards. |
| 20:33 | <avandenhoven> | Either way, the JS people are going to see the unscripted view until dom load. |
| 20:33 | <yecril71> | You can say onload="show()" to get them ASAP. |
| 20:34 | <avandenhoven> | onLoad happens after onDomReady. |
| 20:34 | <yecril71> | You mean window.onLoad or element.onLoad? |
| 20:34 | <Philip`> | avandenhoven: Isn't it easy enough to simply modify your script so the single line of code that sets the isScripted class will be executed immediately when the script is loaded? |
| 20:35 | <yecril71> | The script should clear .ifscript, not set it. |
| 20:35 | <avandenhoven> | clear or set, has no real difference, except in the selectors. |
| 20:35 | <yecril71> | No real difference but easier to follow if everyone sees the same picture though. |
| 20:36 | <avandenhoven> | yes you could put a script immediately after the opening body tag. But that still doesn't elminate the need for |
| 20:36 | <avandenhoven> | <script type="text/javascript">document.write("<li class=\"item0\"><a href=\"#\" onClick=\"window.print()\" onKeyPress=\"window.print()\" />Print Map & Results</a></li>");</script> |
| 20:36 | <yecril71> | What whas that for? |
| 20:36 | <yecril71> | You cannot have "</" in a script block. |
| 20:36 | <avandenhoven> | it makes a link that fires the print dialog. |
| 20:37 | <avandenhoven> | not necessary per se, but my clients like it. |
| 20:37 | <yecril71> | Why not put it directly into HTML? |
| 20:37 | <avandenhoven> | why pollute the DOM with elements that are not valuable unless you have JavaScript. |
| 20:38 | <yecril71> | What is in the DOM is pretty irrelevant where there is no script. |
| 20:38 | <yecril71> | There could be no DOM whatsoever in such an environment. |
| 20:39 | <avandenhoven> | Hmm. |
| 20:39 | <avandenhoven> | maybe I should rethink the logic of some of the code that has been handed down through the aeons. |
| 20:40 | <yecril71> | And clean things up for Christmas. |
| 20:40 | <avandenhoven> | you have no idea what you're asking. |
| 20:41 | <avandenhoven> | We do an integrated online banking/corporate website platform for canadian Credit unions. |
| 20:41 | <yecril71> | Augian stables? |
| 20:41 | <avandenhoven> | I should be so lucky. |
| 20:41 | <yecril71> | Need help? |
| 20:42 | <avandenhoven> | try adding a slave master who doesn't want to change anything becuase its stable. |
| 20:42 | <gsnedders> | yecril71: But layout and (from that) rendering is done from the DOM, so it is still relevant when there is no script |
| 20:42 | <yecril71> | But DISPLAY:NONE has no effect on rendering. |
| 20:43 | <gsnedders> | if CSS isn't enabled, then that's true |
| 20:44 | <avandenhoven> | He means that if you display none the elements you don't want to show when javascript is disabled, you can insert what ever you want. |
| 20:44 | <avandenhoven> | to be really general though (the thing you need when you have complex systems like this) is a class on the body or HTML element and a class on the individual elements. lets you do: |
| 20:45 | <avandenhoven> | .noScript .scriptOnly {display:none;} |
| 20:45 | <yecril71> | If CSS is not enabled, I would just DISPLAY:NONE a warning message: no script? not for you, pal! |
| 20:46 | <yecril71> | This may cause minor problems to users with no CSS, yes. |
| 20:46 | <avandenhoven> | ick. No CSS and JavaScript? |
| 20:47 | <avandenhoven> | you're on your own buddy. |
| 20:47 | <avandenhoven> | that is CSS disabled and JavaScript enabled. |
| 20:47 | <yecril71> | It seems like a very unexpected setup to me. |
| 20:48 | <yecril71> | Do you think it should be supported? |
| 20:48 | <avandenhoven> | yecril71: as it turns out, my team (we have 30+ java developers and 9 web developers, I'm with the web dev team) is looking for a web developer in the next little while. |
| 20:48 | <avandenhoven> | CSS disabled and JavaScript enabled? I don't think you can do anything too reasonable. |
| 20:49 | <avandenhoven> | Except make sure your markup works well with out CSS and have your JS bail at the first opportunity. |
| 20:50 | <yecril71> | Do you think JS-CSS is a valid target for your application? |
| 20:51 | <avandenhoven> | you mean using JS to write the css. I remember looking at that way back in the day and it struck me as being sub optimal. |
| 20:51 | <yecril71> | I mean JS enabled, CSS disabled. |
| 20:52 | <avandenhoven> | oh that. |
| 20:52 | <avandenhoven> | I don't think so. |
| 20:52 | <yecril71> | And what about -JS-CSS? |
| 20:53 | <avandenhoven> | Sure. |
| 20:53 | <avandenhoven> | We always make sure that you can use our sites with lynx. I find it makes it more likely that screen readers will not choke. |
| 20:54 | <avandenhoven> | We don't worry about it too much but we do think about it. |
| 20:54 | <avandenhoven> | Its not pretty but it is functional. |
| 20:54 | <yecril71> | I thought Lynx should support display:none at least. |
| 20:54 | <yecril71> | Doesn�t it? |
| 20:54 | <krijnh> | It doesn't |
| 20:55 | <yecril71> | Ouch. |
| 20:55 | <yecril71> | Then it seems you have to use irrelevant instead. |
| 20:55 | <yecril71> | (which Lynx does not support either, but it is an HTML5 chat after all) |
| 20:56 | <krijnh> | avandenhoven: I totally get your problem, hard to solve :) |
| 20:57 | <yecril71> | Or just say "Please ignore the following message, it is not for you" |
| 20:57 | <krijnh> | We get thought unobtrusive scripting is the way to go, even though that creates a suboptimal experience for like 95% of your users :) |
| 20:59 | <yecril71> | OTOH, short of document.write, you could use a JSON representation of the node to be inserted |
| 20:59 | <yecril71> | and have a library function insert it for you. |
| 21:00 | <yecril71> | Because even if yesscript were defined, Lynx would probably not support it anyway. |
| 21:00 | <yecril71> | So the solution would not solve anything. |
| 21:01 | <Philip`> | krijnh: That's not a hard problem to solve - you just have to ignore what everyone is telling you to do :-) |
| 21:01 | <yecril71> | It would be easier to get Lynx to honour DISPLAY:NONE. |
| 21:02 | <Philip`> | yecril71: That would involve Lynx getting a complete CSS implementation, which is not trivial |
| 21:02 | <krijnh> | Philip`: well, document.write and inline event handlers don't solve it all either :) |
| 21:03 | <yecril71> | A CSS implementation limited to DISPLAY:NONE is not a complete implementation |
| 21:03 | <yecril71> | and that is the only thing that is needed |
| 21:04 | <krijnh> | That would totally suck |
| 21:04 | <krijnh> | ul ul { display: none; } ul li:hover ul { display: block; } |
| 21:04 | <krijnh> | Go Lynx :) |
| 21:05 | <yecril71> | That sux anyway |
| 21:05 | <krijnh> | I knew it |
| 21:05 | <yecril71> | Lynx or not lynx |
| 21:07 | <Philip`> | yecril71: It would still need all the selectors (and hence a tree model of the document) and the cascading rules, and media types and so on, even if it then ignores all but one of the properties |
| 21:08 | <yecril71> | Doesn�t W3C have a reference implementation for that? |
| 21:08 | <krijnh> | yecril71: btw, Opera makes that less sucky (and afaik the only browser that does :hover states when keyboard navigating) |
| 21:09 | <avandenhoven> | Sorry, i had to get back to "real" work for a bit. :) |
| 21:09 | <Philip`> | yecril71: Not that I'm aware of, and reference implementations probably aren't the kind of thing you'd want to ship in a browser anyway |
| 21:10 | <avandenhoven> | Adding something like "yesscript" isn't going to work retroactively. That's fine. |
| 21:10 | <avandenhoven> | I wonder if |
| 21:10 | <avandenhoven> | <script type="text/html"> would be the way to go. |
| 21:10 | <krijnh> | That does work already |
| 21:10 | <avandenhoven> | really? |
| 21:11 | <avandenhoven> | you're saying that if I put <script type="text/html"><h1>You have script enabled</h1></script> I'll actually have what I think I should have? |
| 21:11 | <krijnh> | Didn't JohnResig write something about that? |
| 21:12 | <krijnh> | http://ejohn.org/blog/javascript-micro-templating/ |
| 21:14 | <avandenhoven> | Ah that's different but very cool. |
| 21:14 | <Philip`> | avandenhoven: All browsers should parse the content of that <script> as a single text node, and then (if they support scripting) do nothing with it because it's an unknown type, or else (if they don't support scripting) do nothing with it |
| 21:15 | <Philip`> | so you'd never see that text, unless you have some extra scripts that extract the contents and insert them as HTML somewhere else in the document |
| 21:15 | <krijnh> | Doh, that's different indeed :) |
| 21:15 | <avandenhoven> | I was thinking along the lines of: http://gist.github.com/34491 |
| 21:16 | <avandenhoven> | I was hoping the style there would be unnecessary. |
| 21:16 | <avandenhoven> | Either way, when you have JS disabled, you see both messages. |
| 21:17 | <krijnh> | document.write('<body class=js-enabled>'); works good enough for styling now |
| 21:17 | <avandenhoven> | I would never write out a body tag like that. |
| 21:17 | <krijnh> | Why not? |
| 21:17 | <avandenhoven> | i'm pretty sure the source wouldn't validate. |
| 21:18 | <krijnh> | Why not? |
| 21:18 | <avandenhoven> | because I'm pretty sure validators don't run JS before validating |
| 21:18 | <krijnh> | They don't need to |
| 21:18 | <krijnh> | The body tag isn't mandatory |
| 21:18 | <krijnh> | http://html5.validator.nu/?doc=http://www.youngprofessionaloftheyear.nl/ (don't mind the iframe) |
| 21:19 | <yecril71> | The problem to solve is what to do with no CSS. |
| 21:21 | <yecril71> | Now that Philip` busted my attempt to enhance Lynx, I think @irrelevant is the way to go. |
| 21:21 | <avandenhoven> | either way, I require that a file that opens a tag also closes it (for instance if you're using SSI or a some CMS system... we literally have dozens of includes in any given generated page). |
| 21:21 | <Philip`> | yecril71: (I think it's called 'hidden' now, not 'irrelevant') |
| 21:21 | <yecril71> | Thanks, memorized. |
| 21:22 | <avandenhoven> | I'll look at the irrelevant/hidden attributes. |
| 21:22 | <avandenhoven> | Thanks for the useful discussion. I |
| 21:22 | <krijnh> | avandenhoven: HTML doesn't require it though.. |
| 21:22 | <Philip`> | <body><script>document.body.className = 'js'<script> might look nicer than optionally writing <body> |
| 21:23 | <Philip`> | s//\// |
| 21:23 | <krijnh> | So much nicer :) |
| 21:23 | <avandenhoven> | krijnh: No, but if I want to keep my code quality up, I do (I have 30 develoeprs who aren't HTML saavy). |
| 21:23 | <avandenhoven> | Philip: yeah that's what I would have to do. |
| 21:23 | <yecril71> | In spite of them, you can closeTag("body") at the end. |
| 21:24 | <yecril71> | So that the code markup can balance. |
| 21:24 | <yecril71> | But body opened and not closed is all right as well. |
| 21:24 | <avandenhoven> | I'm of the kindergarden school of web development. |
| 21:24 | <avandenhoven> | If you open it close it. |
| 21:24 | <krijnh> | Agreed, especially if you work with others |
| 21:25 | <avandenhoven> | I know HTML doesn't require it but I find that it avoids trouble. |
| 21:25 | <yecril71> | Like </br >? :-) |
| 21:25 | <avandenhoven> | no <br /> |
| 21:25 | <avandenhoven> | that's closed. |
| 21:25 | <avandenhoven> | <img /> |
| 21:25 | <avandenhoven> | that too. |
| 21:25 | <yecril71> | It is XHTML, not HTML. |
| 21:25 | <krijnh> | *sigh* |
| 21:25 | <krijnh> | It is HTML5 |
| 21:25 | <Philip`> | It is Real World HTML |
| 21:25 | <yecril71> | In HTML, it is <BR >> |
| 21:26 | <avandenhoven> | that lack of closing for empty tags is a hold over of SGML that was, IMHO, a mistake. |
| 21:26 | <yecril71> | It is a manifestation of Occam razor. |
| 21:27 | <avandenhoven> | Until you start making tag soup. |
| 21:27 | <avandenhoven> | which is what HTML5 in current browsers really is. |
| 21:27 | <Dashiva> | It's only tag soup if you want it to be |
| 21:28 | <yecril71> | <br /> in HTML4 is tag soup |
| 21:28 | <avandenhoven> | what is the model for <body> <section><p></body> if you aren't an HTML5 browser? |
| 21:28 | <yecril71> | <body ><section /><p /></body > in IE |
| 21:28 | <avandenhoven> | yecril71: exactly and its absolutely clear for a use what it is. |
| 21:29 | <yecril71> | I am not sure about Firefox? |
| 21:30 | <avandenhoven> | I would love to continue this but I'll waste a whole day if I don't sign off. |
| 21:30 | <avandenhoven> | I'll talk to you guys later! |
| 21:30 | <yecril71> | Are the positions in your team public? |
| 21:31 | <Philip`> | Only a whole day? I've wasted over a year here! |
| 21:35 | <yecril71> | script type="text/bourne-shell" |
| 21:35 | <yecril71> | cat <<EOF |
| 21:36 | <yecril71> | Hey, what have I done? |
| 21:36 | <yecril71> | All right, I am back. |
| 21:36 | <yecril71> | \/script. |
| 21:37 | yecril71 | does not know how to insert a leading slash. |
| 21:37 | <Dashiva> | /press ctrl while pressing enter |
| 21:38 | <Dashiva> | or use /say /whatever with leading slash |
| 21:38 | <jcranmer> | / in irssi |
| 21:38 | <gsnedders> | yecril71: <br /> isn't tag soup, it just means something else :P |
| 21:38 | <gsnedders> | double / |
| 21:38 | <jcranmer> | (as in `/ ', not `/') |
| 21:38 | <gsnedders> | i.e., ///foo |
| 21:39 | <gsnedders> | *// |
| 21:39 | gsnedders | has screwed up |
| 21:58 | yecril71 | finds out it is hard to DT an adjective |
| 21:58 | yecril71 | decides to LI DFN instead |
| 22:56 | Philip` | downloads the new IE8 build |
| 22:57 | <gsnedders> | jgraham: I wasn't asked any sort of order of magnitude question — I know several maths applicants were asked how many grains of sand there are in the world, though. |
| 22:58 | <Philip`> | gsnedders: Oh, how did the interviews go? |
| 22:59 | <gsnedders> | Philip`: The first (physicsy one) went really badly; the second went fine |
| 23:00 | <Philip`> | gsnedders: Ah |
| 23:01 | <gsnedders> | Basically, all the questions I was asked in the first were the things that I'm meant to know for AH physics but don't actually know |
| 23:03 | <Philip`> | Sounds like the new IE8 build fixes the ARIA attribute properties |
| 23:03 | <gsnedders> | I did work it out mostly, albeit slowly and only just not making bad mistakes with integration |
| 23:05 | <Philip`> | Interviews don't really feel like the right environment for carefully solving equations |
| 23:05 | <gsnedders> | No, they really don't |
| 23:05 | <roc> | what new IE8 build? |
| 23:06 | <Philip`> | roc: I think it's only in the Tech Beta program at the moment |
| 23:06 | <roc> | how often do they release new builds? |
| 23:07 | <Philip`> | They're basically the same as the public betas |
| 23:07 | <Philip`> | (This is the first change since Beta 2, as far as I'm aware) |
| 23:07 | <roc> | you mean this is the first build they've released that wasn't a public beta? |
| 23:08 | <gsnedders> | They've been distributing a newer build with Win7 pre-beta apparently |
| 23:08 | <gsnedders> | It's probably just the same build as that |
| 23:08 | <Philip`> | roc: It's the first one via the Tech Beta thing |
| 23:09 | <Philip`> | (which is free to join if you can convince them you're not going to spam their bug database with garbage, or if you convince an existing member to give you an access code, etc, and if you agree to all the relevant terms and conditions) |
| 23:12 | <Philip`> | gsnedders: It's dated to the 10th of December (at least in the readme file), so it may not be the same |
| 23:12 | <Philip`> | (Oddly, the readme file is a PDF named Readme.exe) |
| 23:19 | <Philip`> | (When I said I was downloading the IE8 build, I didn't realise it was going to go at 15KB/sec, or that it would take half an hour to update Windows and uninstall IE8b2 and still not get around to installing the new build...) |
| 23:39 | <Philip`> | Aha, there it goes |
| 23:44 | <Philip`> | It's a lot less slow at rendering the spec than before |
| 23:45 | <Philip`> | It still takes half a second to work out where to go when you click on a fragment link, and maybe a quarter of a second to draw the background highlight when hovering items in the table of contents, but it's quite bearable |