| 00:29 | <MikeSmith> | anybody else from #whatwg at Web Directions South this week? |
| 00:35 | <Hixie> | weinig: here |
| 00:36 | <weinig> | Hixie: hey, have Adam Barth or Collin Jackson (or others) talked to you about specing sending the Origin header for non-XMLHttpRequest requests? |
| 00:37 | <weinig> | ie. https://bugzilla.mozilla.org/show_bug.cgi?id=446344 and https://bugs.webkit.org/show_bug.cgi?id=20792 |
| 00:47 | <Hixie> | it's a topic that has come up in conversation |
| 00:50 | <Hixie> | i agree with maciej's comment on the webkit bug |
| 00:50 | <Hixie> | i'm happy to spec it in html5 for form submission and the like, but http seems like a place to define the header, if not the value |
| 00:52 | weinig | nods |
| 03:16 | <gsnedders> | Hixie: Nobody is working on a new version of HTTP, though: 2616bis is just revising the spec, and can't make any normative changes to it. |
| 03:16 | <gsnedders> | Hixie: Which is kinda problematic if that's the place to define the header. |
| 03:17 | <gsnedders> | (and yeah, I really am up at 3:19am) |
| 03:32 | <gsnedders> | This draft of the personal statement is absolutely shite. |
| 04:10 | Lachy | finally met MikeSmith in person at Web Directions |
| 04:11 | <gsnedders> | So MikeSmith is indeed a person? |
| 04:12 | <Lachy> | gsnedders, either that, or a really life-like robot. |
| 04:12 | <gsnedders> | I still suspect the latter. |
| 04:12 | <Lachy> | considering his name is a trademark, yeah, me too :-) |
| 04:13 | <Lachy> | he's probably made by some company in Japan |
| 04:13 | <gsnedders> | Mikeā¢ Smith Robots, Inc.? |
| 04:14 | <Lachy> | heh |
| 04:14 | <Lachy> | gotta go, cya |
| 04:14 | <gsnedders> | Have fun |
| 04:15 | <gsnedders> | But he left before I could say even that :'( |
| 04:54 | <gsnedders> | I HATE THIS. |
| 04:54 | gsnedders | sighs |
| 05:10 | <gsnedders> | Philip`: ping |
| 05:11 | <gsnedders> | (though I expect I'll be at school when you pong me) |
| 06:15 | <gsnedders> | I really dislike writing a personal statement. |
| 06:25 | <gsnedders> | The first and last sentences are the hardest to write, certainly. |
| 07:17 | <hsivonen> | yay. the server handling my email got electricity back |
| 07:27 | <krijnh> | hsivonen: do you have an article online somewhere about "[10:53] <hsivonen> CSS positioning is much worse than tables" ? :) |
| 07:37 | <hsivonen> | krijnh: I don't, but dbaron had a presentation touching on the matter at XTech 2006. |
| 07:37 | hsivonen | looks up the presentation paper |
| 07:37 | <hsivonen> | (though that was about floats, iirc) |
| 07:37 | <krijnh> | Ah, yeah, I think I've read that |
| 07:38 | <zcorpan_> | hasn't dbaron whined on his blog about the matter too? |
| 07:38 | <krijnh> | Yep |
| 07:38 | <krijnh> | Re the One True Layout thing |
| 07:38 | <zcorpan_> | http://dbaron.org/log/2005-12#e20051228a |
| 07:39 | <krijnh> | And now the anti-positioning thing :) |
| 07:40 | <hsivonen> | the paper is http://xtech06.usefulinc.com/schedule/paper/146 but the paper doesn't really contain the point I imagine to recall from the presentation |
| 07:40 | <hsivonen> | krijnh: anyway, my own experience with positioning is that people who go with positioning in order to avoid tables end up having their boxes overlap on narrow viewports |
| 07:41 | <hsivonen> | krijnh: with tables at least the formatter takes care of cells not overlapping |
| 07:42 | <krijnh> | Our government advises http://www.webrichtlijnen.nl/english/manual/development/production/page-structure/body/#r-pd-6-2 - which would almost require position: absolute in all of the cases |
| 07:44 | <krijnh> | (If a main navigation which should be on top of the page visually is considered less important and put below the main body) |
| 07:45 | <krijnh> | Dunno if that would count as using positioning in order to avoid tables |
| 08:11 | <Hixie> | heycam: any news on the issue of changing the way we do getters and setters to be just exposed properties? |
| 08:23 | <heycam> | hi Hixie, no i haven't fixed that up yet |
| 08:23 | <heycam> | (that's the issue garrett raised, yes?) |
| 08:24 | <heycam> | i haven't had any time for webidl lately, unfortunately, with conferences and other work taking up my time |
| 08:24 | <heycam> | perhaps soon... |
| 08:24 | heycam | heads home and will be back online in an hour |
| 08:32 | <Philip`> | gsnedders: Pong |
| 09:08 | <Hixie> | am i missing something? |
| 09:08 | <Hixie> | headers/id IS in the spec |
| 09:11 | <zcorpan_> | Hixie: it can't point to <td> |
| 09:11 | <Hixie> | that doesn't seem to be what hsivonen was talking about in his last e-mail |
| 09:11 | <annevk2> | (maybe he was confused with Laura, who wasn't very precise) |
| 09:11 | <Hixie> | headers="" can't point to a <td> because we flipped the "cells that are both data and headers should be <td>" to "should be <th>" to improve accessibility |
| 09:13 | <Hixie> | headers="" isn't allowed on <th> because it hadn't been shown that it was necessary, though maybe data since then has changed that, i haven't checked |
| 09:13 | <Hixie> | (i'm doing studies to see what the Web can tell us on the matter as we speak) |
| 09:16 | <annevk2> | btw, did you see sicking's message that the Web Workers demos spell data as message? |
| 09:16 | <Hixie> | yeah |
| 09:16 | <Hixie> | someone should remind me of that when i work on that part of the spec, if i forget :-) |
| 09:16 | <Hixie> | i wish aa and sicking would make their mind up about workers though |
| 09:19 | Hixie | wonders where to put the <dfn> for min/max/step given that those attributes are different in each case they apply |
| 09:20 | <Hixie> | where's mike |
| 09:20 | <annevk2> | in Australia |
| 09:21 | annevk2 | is reminded of TR/ publication |
| 09:24 | <hsivonen> | whoa. I should have checked the status of headers in spec before replying. |
| 09:25 | <hsivonen> | what is the argument about then? chains? |
| 09:25 | <Hixie> | they want headers="" on <th> so they can point to other <th>s (and maybe <td>s?) |
| 09:25 | <Hixie> | and they might well be right |
| 09:25 | <Hixie> | i haven't actually gone through their feedback yet |
| 09:26 | <Hixie> | there are 70 e-mails on the subject waiting in my folder |
| 09:26 | <Hixie> | 70! |
| 09:26 | <annevk2> | people are using html5lib: http://simonwillison.net/2008/Sep/24/htmlwhitelist/ |
| 09:26 | annevk2 | is surprised |
| 09:27 | <Hixie> | apparently though they think that somehow their pet issue, which affects a tiny, tiny fraction of the web (so few pages that when i do studies of this subject matter i have to use larger-than-usual samples from google's index), deserves some sort of fast-track treatment |
| 09:28 | <Hixie> | i haven't worked out on what grounds, given that what they're asking for is already implemented (that's one of the reasons they use to argue for it) |
| 09:29 | <Hixie> | and given the utter disdain i got from them when i fast-tracked their last pet issue (img alt), i am not at all inclined to put them ahead of the people who actually politely work on far more important and more pressing issues |
| 09:31 | Hixie | receives another e-mail on the subject (from hsivonen) -- 71 e-mails... |
| 09:31 | <hsivonen> | oops. I cited the whatwg URL on public-html |
| 09:32 | <Hixie> | annevk2: dewitt works with me, i was probably the one who brought it to his attention |
| 09:32 | <BenMillard> | Hixie, it's fair to describe data table header association as being one of my pet issues. :) |
| 09:33 | <BenMillard> | Hixie, although I'm well aware it's not a pressing issue in the context of HTML5, so I'm fine with it having a longer cycle of review |
| 09:33 | <Hixie> | having pet issues is fine, it's the crazies who want it prioritised above all else that get to me |
| 09:33 | <Hixie> | i mean i have plenty of pet issues, e.g. Web Socket, but I don't bring them to the front of the queue |
| 09:33 | <Hixie> | I mean, Web Socket probably had the longest delay of anything in the spec so far |
| 09:34 | Hixie | gets off his soapbox and goes back to working out wtf <input type=datetime min/max/step> do |
| 09:36 | <BenMillard> | Hixie, yeah a couple of website jobs and another pet project are ahead of any HTML5 in my personal queue atm |
| 09:42 | <zcorpan_> | hsivonen: i always cite the whatwg url on public-html :) |
| 09:42 | <zcorpan_> | if i am to cite the spec |
| 09:42 | Hixie | too |
| 09:42 | <Hixie> | the whatwg spec has far more features |
| 09:42 | <Hixie> | like click-on-dfn |
| 09:42 | <Hixie> | which i use all the time now |
| 09:43 | <annevk2> | that's seriously useful |
| 09:43 | <hsivonen> | I always read the whatwg spec, but as a courtesy, I try to remember to substitute the W3C url before hitting send |
| 09:44 | <Philip`> | You can click on dfns? I never knew that :-/ |
| 09:45 | <Hixie> | it's very recently added |
| 09:45 | <Hixie> | a week or so |
| 09:45 | <Philip`> | Ah |
| 09:45 | <Hixie> | i use it all the time now though |
| 09:45 | <Hixie> | it's really useful in the wf2 section |
| 09:45 | <Hixie> | or for finding out which elements are flow-level, etc |
| 09:45 | <Philip`> | I guess it doesn't work so well in the multipage version |
| 09:46 | <Hixie> | indeed not |
| 09:46 | <BenMillard> | I expect clickable navigation things to use <a href>. |
| 09:46 | <BenMillard> | <dfn><a href>...</a></dfn> perhaps |
| 09:47 | <Philip`> | (and the single-page version has awful performance and freezes for tens of seconds after loading, presumably while it's computing all these cross-references) |
| 09:47 | <Hixie> | actually no |
| 09:47 | <Hixie> | the freezing is caused by the status boxes being spawned |
| 09:47 | <Hixie> | the dfn links are calculated asynchronously in small batches |
| 09:48 | <Hixie> | and shouldn't cause slowness |
| 09:48 | <Hixie> | i should rewrite the status thing to be async too |
| 09:49 | <BenMillard> | <dfn><a href> could happen while the spec is being generated? that would reduce the amount of work every UA must do when it visits |
| 09:50 | <BenMillard> | and it could support the multi-page version quite easily |
| 09:51 | <hsivonen> | was the change that allowed headers not marked as conformance-checker-sensitive? |
| 09:51 | <annevk2> | BenMillard, that doesn't give the same feature at all, did you try it? |
| 09:51 | <hsivonen> | I'm trying to track how I managed to miss it |
| 09:51 | <Hixie> | BenMillard: yeah, i was going to ask gsnedders to do something along these lines when we were transferred over to his script |
| 09:51 | <Hixie> | hsivonen: not sure |
| 09:51 | <Hixie> | hsivonen: if so, i apologise |
| 09:51 | <hsivonen> | Hixie: np. I just wanted to check my script wasn't broken |
| 09:52 | <Hixie> | i was looking at my table headers implementation earlier |
| 09:52 | <Hixie> | and noticed i wrote it in june 2007 or so |
| 09:52 | <Hixie> | and realised in great sadness that that meant i had over a year of updates to apply |
| 09:52 | <Hixie> | and decided it wasn't worth it |
| 09:53 | <Philip`> | Hixie: I don't remember it being so slow when it was just doing status boxes; now it spends about 70 seconds at 100% CPU load with the browser (Opera 9.5) being almost entirely unresponsive |
| 09:53 | <Hixie> | but anyway it does i mean i narrowed it down and it happened sometime in the last year and a half! |
| 09:53 | <Hixie> | hope that helps! |
| 09:53 | <Hixie> | Philip`: woah |
| 09:53 | <Hixie> | Philip`: sounds like a bug |
| 09:53 | <Hixie> | Philip`: i don't notice anything more than a second or so of slowness in webkit |
| 09:54 | <Philip`> | In Firefox 2 it's more like 5 seconds |
| 09:54 | <Philip`> | so I suppose something's interacting badly with Opera |
| 09:55 | <Hixie> | and firefox 2 is slow on html5 |
| 09:55 | <BenMillard> | annevk2, I was guessing based on Hixie's description of it...trying Fx2 on the HTML5 single-page spec now |
| 09:55 | <BenMillard> | Hixie, yeah it's been running 4 minutes for me...now |
| 09:55 | <Hixie> | woah |
| 09:55 | <Hixie> | y'all should start using webkit, apparently |
| 09:55 | <Hixie> | i didn't realise html5 was so slow in other browsers |
| 09:55 | <Hixie> | that's insane |
| 09:56 | <Hixie> | it's really snappy in webkit builds |
| 09:56 | <Philip`> | Nobody makes WebKit work nicely on Linux |
| 09:56 | <Hixie> | browsers on linux suck |
| 09:56 | <BenMillard> | Hixie, I've dragged a window across Firefox 2 and it's not repainting and there's no "A script is stuck" message box, so I guess it's still doing stuff and hasn't died |
| 09:56 | <Hixie> | (maybe chrome for linux will help that) |
| 09:56 | <Philip`> | and even if they did, I'd still use Opera, because changing would be far too traumatic |
| 09:56 | <Hixie> | heh |
| 09:56 | <Philip`> | Hixie: I'd prefer a browser that has features, i.e. not Chrome |
| 09:57 | <Hixie> | ah well can't help you there |
| 09:57 | <roc> | Fx3 (and trunk) takes a while to load the HTML5 page for me, but it doesn't lock up |
| 09:57 | <Hixie> | Philip`: (like what features?) |
| 09:58 | <BenMillard> | roc, the titlebar of Firefox 2 now says (Not Responding) |
| 09:58 | <BenMillard> | roc, after nearly 7 minutes |
| 09:58 | <roc> | try 3 |
| 09:58 | <BenMillard> | roc, I can't stand the UI of 3, sorry |
| 09:58 | <Hixie> | what os? |
| 09:58 | <BenMillard> | Hixie, XP SP3 |
| 09:59 | <zcorpan_> | BenMillard: isnt't here a firefox 2 skin for firefox 3? |
| 09:59 | <Hixie> | yeah i was going to say, try a skin :-) |
| 09:59 | <BenMillard> | zcorpan_ & Hixie, I expected there to be one but couldn't find it when I looked the other month |
| 10:00 | <BenMillard> | I currently use Winestripe and that over-engineered tabs extension to make it sort of fit in with XP |
| 10:00 | <Hixie> | https://addons.mozilla.org/en-US/firefox/addon/6898 |
| 10:01 | <Philip`> | Hixie: (Ad blocking, mouse gestures, feed reading, the ability to quickly disable all stylesheets, etc) |
| 10:01 | <BenMillard> | Hixie, I found that one before. it doesn't fit in with XP...look at the tabs! :P |
| 10:02 | <Philip`> | Hixie: (where "etc" means I can't think of any more things, though I'd probably notice if they were missing :-) ) |
| 10:02 | <Hixie> | BenMillard: i've no idea what xp looks like, the only xp i have access to is in a VM and has all the fancy UI disabled :-) |
| 10:02 | <Hixie> | Philip`: yeah i like access over style sheets, and chrome is missing that. the other features, can't say i like them or use them :-) |
| 10:05 | <BenMillard> | Hixie, I've used (and adored) the "Royale" theme that ships as the default with XP Media Centre Edition for a few years: http://projectcerbera.com/ui/windows/royale/royale.png (1,6000x1,200) |
| 10:05 | <BenMillard> | *1,600 |
| 10:05 | Philip` | can't survive without the left-click-right-click mouse gesture for 'back' |
| 10:05 | <Hixie> | BenMillard: i don't know how you can stand xp :-) |
| 10:06 | <BenMillard> | Hixie, I live in hope that one day things will get better. :) |
| 10:06 | <Philip`> | BenMillard: Needs more padding in Explorer :-p |
| 10:06 | <Hixie> | BenMillard: let me know how that turns out :-) |
| 10:08 | <Philip`> | Sadly Vista makes the Windows Classic theme (i.e. the one that's like Win95) unusable, because it has crazy toolbars that look awful if you're not using the Aero themes instead :-( |
| 10:10 | <annevk2> | zcorpan_, the security check is still there for adoptNode() |
| 10:13 | <BenMillard> | Hixie, after 21 minutes and 30 seconds, I think it's safe to conclude that Firefox 2.0.0.17 on XP has crashed on the single-page HTML5 spec (although multipage worked fine beforehand) |
| 10:14 | <Hixie> | weird |
| 10:16 | <BenMillard> | sent error report...but the "More information" link doesn't work even though the error message that pops up shows the URL and that URL does work...I guess Hixie's right about XP :P |
| 10:17 | <Philip`> | Browser bugs are meant to be exposed by HTML5 test suites, not by the spec document itself :-) |
| 10:19 | <zcorpan_> | annevk2: thanks |
| 10:20 | <Dashiva> | Philip`: You could combine the features, though. Make the spec text only appear properly in the correct rendering |
| 10:20 | <BenMillard> | Dashiva, that's basically what I experienced just now |
| 10:20 | <annevk2> | aah, "access control check" is already written in such a way that HTML5 can easily use it for <img> |
| 10:21 | <annevk2> | and the CSS Fonts Module for @font-face, etc. |
| 10:21 | annevk2 | is pleased |
| 10:31 | <Hixie> | http://www.w3.org/mid/55687cf80809250152o796d6b74n57ce73e01eccf110āmgc is incredible |
| 10:32 | <annevk2> | he only ever flames on public-html |
| 10:32 | <Hixie> | i just love the fact that he says he can't find any way for headers="" to be useful without chaining |
| 10:35 | <Hixie> | BenMillard: i just loaded html5 in a VMWare XP box with Firefox2, and it didn't even stall, loaded in a few seconds. |
| 10:36 | <Hixie> | this is on a quite old mac book pro, with the VMWare instance having at most one CPU and not much memory. |
| 10:37 | <nessy> | Hixie: you mentioned your talk being published through YouTube - is that up yet? |
| 10:39 | <Hixie> | seems not |
| 10:40 | <annevk2> | will it be on googletechtalks? |
| 10:40 | <BenMillard> | Hixie, meh. Steve is wrong, specifically because irregular data tables can use <th> for one layer of header cells and associate the relevant <td>s with them using headers+id (which I imagine is the use case you had for adding that ability) |
| 10:41 | <BenMillard> | along with using headers+id as a "patch" for largely regular cases which include an irregularity which Smart Headers, scope and other features aren't able to fill |
| 10:41 | <Hixie> | annevk2: i assume so |
| 10:42 | <Hixie> | BenMillard: there are bazillions of examples where headers="" can be used without chaining, absolutely |
| 10:42 | <annevk2> | hey abarth |
| 10:42 | <abarth> | hi annevk2 ! |
| 10:43 | <abarth> | i read that you've been having fun registering HTTP headers :) |
| 10:43 | <annevk2> | they should move it to a wiki :) |
| 10:44 | <annevk2> | "expert review" is a silly concept anyway, it's in a spec that gets review from experts already |
| 10:44 | <abarth> | i think you're an expert, you should review it |
| 10:44 | <Hixie> | BenMillard: (many of them -- at least two thirds of tables that use headers="" today according to a study i did last year -- can be done with scope="" too, but still) |
| 10:44 | <annevk2> | hehe |
| 10:45 | <Hixie> | annevk2: more review is always good :-) |
| 10:45 | <annevk2> | not if it doesn't happen |
| 10:45 | <Hixie> | then it's not more review :-) |
| 10:45 | <annevk2> | and if it's the only way to get some header names on a page |
| 10:45 | <annevk2> | Hixie, :p |
| 10:45 | <hsivonen> | is there only one resident expert for a given registry at IANA? |
| 10:46 | <annevk2> | I think it's just whoever signs up to the mailing list |
| 10:46 | <hsivonen> | oh. I thought there were appointed guardians of IANA registries |
| 10:49 | jgraham | complains that the Tokyo subway map is hard to understand |
| 10:53 | <annevk2> | test |
| 10:54 | <abarth> | test succeeds |
| 10:54 | <hsivonen> | hmm. It seems that I have noticed the headers/scope spec change, implemented it, and forgotten about it |
| 10:54 | <hsivonen> | Hixie: so probably not a problem with the spec annotations... |
| 10:54 | <Hixie> | ok i have some stats on tables |
| 10:55 | <Hixie> | out of about 8 billion pages, guess how many had a <th> element with a headers="" attribute |
| 10:55 | <hsivonen> | 200? |
| 10:56 | <annevk2> | 10000? |
| 10:56 | <annevk2> | (billion is 1E9 here right?) |
| 10:56 | <Hixie> | and then guess how many of those pages had no cells with duplicate IDs, no headers="" pointing to non-cells, no headers="" pointing to cells in other tables, no headers="" pointing to themselves, and no cycles. |
| 10:57 | <annevk2> | 2% |
| 10:57 | <Hixie> | (yes, on the order of 8e9 pages) |
| 10:57 | <hsivonen> | 5? |
| 10:57 | jgraham | thought html4 didn't really encourage headers on th |
| 10:58 | <Hixie> | i found 350,000 pages with <th headers=""> (<0.005%), and about 150,000 had none of the above errors (less than half) |
| 10:58 | annevk2 | is less and less confusing short scale and long scale |
| 10:59 | <annevk2> | that's a whole lot better than longdesc, but that's not saying much |
| 10:59 | <hsivonen> | looks like I was too pessimistic ;-) |
| 10:59 | annevk2 | too |
| 11:01 | <hsivonen> | Hixie: aren't the errors trivially recoverable, though, except for the case where there are duplicate ids across tables and the other-table-id comes in document order first, so that smart recovery couldn't use gEBI |
| 11:02 | <Hixie> | they're recoverable, sure |
| 11:02 | <Hixie> | but that doesn't mean the tables will be accessible |
| 11:02 | <Hixie> | which is the whole point here |
| 11:02 | <hsivonen> | Hixie: but are the 150000 tables accessible? |
| 11:02 | <Hixie> | http://junkyard.damowmow.com/344 has a sample of the pages that the script found <th headers=""> on with no obvious problems |
| 11:03 | <Hixie> | so you can find out :-) |
| 11:03 | Hixie | starts looking at them |
| 11:03 | <hsivonen> | Hixie: and the does headers/id after recovery make the remaining 200000 worse? |
| 11:04 | <Hixie> | no idea |
| 11:04 | <Hixie> | http://webopac.gosford.nsw.gov.au/libero/WebopacOpenURL.cls?DATA=GCL&ACTION=DISPLAY&RSN=4363 |
| 11:05 | <annevk2> | headers="" was not needed on http://articles.moneycentral.msn.com/Investing/Extra/FiveStrongDividendPayingStocks.aspx |
| 11:05 | <Hixie> | on the above one, headers="" is redundant in the first table and points to a header called "Field Name" in the second |
| 11:05 | <Hixie> | which is about as pointless a header as you can get |
| 11:06 | <Hixie> | http://articles.moneycentral.msn.com/Investing/Extra/PovertyNowComesWithAColorTV.aspx misuses <th> |
| 11:06 | <Philip`> | 0.005% <th headers> is far more than I found :-( |
| 11:06 | <Philip`> | (I see it on one out of 130K pages) |
| 11:07 | <Hixie> | based on the sample urls, they're all on very "deep" pages |
| 11:07 | <jgraham> | http://www.maastricht.nl/maastricht/show/id=93271/notextonly=42282 almost certainly doesn't need @headers |
| 11:07 | <Hixie> | so likely not those you'd find in your sample source |
| 11:08 | <Hixie> | http://library.woollahra.nsw.gov.au/libero/WebOpac.cls?VERSION=2&ACTION=DISPLAY&RSN=74818&DATA=DWO&TOKEN=Y8uUcWNA859284&Z=1&SET=13 apparently has changed since my script looked at it |
| 11:08 | <Hixie> | since i can't find headers="" on it |
| 11:08 | <Philip`> | (I see http://www.skydebanegaard.dk/ which seems to be referring to a cell in a different table) |
| 11:09 | <Philip`> | (though I'm not sure why it's necessarily a problem for headers to point to a different table) |
| 11:09 | <Hixie> | because it's not accessible to users who don't have TAs |
| 11:10 | <Hixie> | http://webopac.slub-dresden.de/libero/WebOpac.cls?LANG=DE&RSN=14195675&ACTION=DISPLAY&TOKEN=y7QBaahcVB3613 should just lose its <thead> altogether and not use headers="" at all |
| 11:11 | <Hixie> | http://www.hildesheim.de/hildesheim/buergerservice/rathaus/soziales.shtml points to <th>s that are empty! |
| 11:12 | <BenMillard> | dotjay, hi you've joined us during an accessibility research session :) |
| 11:12 | <dotjay> | BenMillard: Marvellous. :) |
| 11:13 | <Hixie> | i don't understand the language on http://www.kiesbeter.nl/patientenorganisaties/detail/tno-preventie-en-zorg/ to determine if it's useful or not |
| 11:13 | <BenMillard> | dotjay, http://krijnhoetmer.nl/irc-logs/whatwg/20080925#l-343 onwards |
| 11:13 | <dotjay> | BenMillard: Thanks. Will have a read in a moment. |
| 11:14 | <BenMillard> | Hixie, looks like a profile for a company |
| 11:15 | <Hixie> | yeah i mean i don't understand the words in the headers to work out if they're useful or not |
| 11:15 | <annevk2> | Hixie, the kiesbeter.nl duplicates a lot |
| 11:15 | <Hixie> | (in the way that "field name" is blatently not useful) |
| 11:15 | <annevk2> | "Naam" means "Name" |
| 11:16 | <BenMillard> | Hixie, seems to be a regular layout which could use a blank top-left cell |
| 11:16 | <annevk2> | not sure why the table has the header "TNO, Preventie en Zorg" as that just duplicates information |
| 11:16 | <hsivonen> | Hixie: it may not be useful, but it's a case of chained headers :-) |
| 11:16 | <annevk2> | BenMillard, actually, the first row could be removed and it would be useful, but maybe the CMS software doesn't support that |
| 11:16 | <Hixie> | http://www4.mdr.de/www3/sport/ergebnisse/inhalt_e_del_m_08.shtml is the first one i've seen that'd arguably useful, but even there i'd argue that the user doesn't need to know that "EisbƤren Berlin" is a "Mannschaft" since those reading the subject matter will almost certainly know that anyway |
| 11:16 | <BenMillard> | annevk2, yeah that's a good point. I guess it doesn't need column headers at all |
| 11:17 | <Hixie> | hsivonen: hm? |
| 11:17 | <annevk2> | BenMillard, though even if it was kept, headers="" would not be needed there, and in fact it uses scope="row" too! |
| 11:18 | <annevk2> | (which with a smart enough algorithm also wouldn't be needed) |
| 11:18 | <hsivonen> | Hixie: Omschrijving is a heading and Naam is a heading for Omschrijving |
| 11:18 | <BenMillard> | annevk2, yeah it's a regular layout so it only needs plain <th> and a simple-as-mud algorithm |
| 11:18 | <Hixie> | http://articles.moneycentral.msn.com/Investing/Extra/FiveStrongDividendPayingStocks.aspx makes the same mistake as the other moneycentral page |
| 11:18 | <BenMillard> | annevk2, what does "gezondheidsonderzoek" and the other entries in that cell mean? Is their use of lowercase significant or just lazy? |
| 11:18 | <Hixie> | hsivonen: sure, but we shouldn't add the feature if it doesn't actually help when used, so does it help? |
| 11:19 | <annevk2> | sicking, https://bugs.webkit.org/show_bug.cgi?id=20879 (also interesting for aboodman) |
| 11:19 | <hsivonen> | Hixie: I guess you'd have to ask an AT user whether it helps in that particular case |
| 11:20 | <annevk2> | BenMillard, it's a list of things they do, "gezondsheidsonderzoek" means "health research" or some such |
| 11:20 | <annevk2> | (without the s after gezond) |
| 11:20 | <Hixie> | http://www.afpa.fr/formations/derniere-minute/places-disponibles-en-formation.html?tri=date&domain=M has the same redundancy as the mdr.de one -- the heading is obvious given the contents of the cells, so doesn't have to be repeated each time |
| 11:21 | <BenMillard> | annevk2, cool I can guess what other languages say from context. :) |
| 11:21 | <Hixie> | same with http://sport.ard.de/sp/komponente/ergebnisse/tabellen/tennis.jsp?liga=WTA%7CTier+III%7CBali&runde=Damen%7CEinzel%7C1.+Runde |
| 11:22 | <Hixie> | http://neuerwerbungen.slub-dresden.de/cgi-bin/getrsn.pl?rsn=14223077 is the same as the earlier one on the same domain |
| 11:22 | <Hixie> | http://library.bankstown.nsw.gov.au/libero/WebopacOpenURL.cls?DATA=DBK&ACTION=DISPLAY&RSN=709279 too |
| 11:22 | <annevk2> | so I guess most add the complexity because browsers and screen readers are silly with respect to table headers |
| 11:23 | <annevk2> | not because it's actually needed by the tables |
| 11:23 | <BenMillard> | Hixie, I noticed tables tend to have consistent authoring throughout a website, but there are exceptions. |
| 11:23 | <Hixie> | yeah |
| 11:23 | <Hixie> | what does lang=nl Praktijknaam mean? |
| 11:24 | <BenMillard> | annevk2, there's also a sense among content providers (not HTML authors) I've worked with that a table isn't really a table unless it has a complete set of column headers, even if they are useless. |
| 11:24 | <Hixie> | looks like the thead in http://www.kiesbeter.nl/eerstelijnszorg/Dietist/detail/thuiszorg_het_friese_land_4/ is redundant, assuming "Praktijknaam" means "name" or some sort |
| 11:24 | <annevk2> | it means "name of the practice" |
| 11:25 | <annevk2> | which is "Thuiszorg het Friese Land" in that table |
| 11:26 | <Hixie> | oh |
| 11:26 | <annevk2> | this was also the case in the first kiesbeter.nl example |
| 11:26 | <Hixie> | then that's just wrong |
| 11:26 | <Hixie> | shouldn't be a <th> at all |
| 11:26 | <annevk2> | yes, both are actually completely wrong |
| 11:26 | <BenMillard> | I'd say they shouldn't even be tables, as the type of data in each cell is pretty clear from the data itself |
| 11:27 | <Hixie> | http://www.maastricht.nl/maastricht/show/id=93271/textonly=42282 is wrong too |
| 11:27 | <annevk2> | I'm guessing the CMS works in a way that the first row is mandatory and always headers |
| 11:27 | <BenMillard> | like an address is obviously an address as it follows the typographic conventions of an address and has place names and area codes in it |
| 11:28 | <Hixie> | and http://www.sestao.eu/agenda/agenda.asp?sestao_idioma=EU&nivelMenu=6&actualMenu=14&dia_seleccionado=24/5/2008&p=12 is a calendar, with the months pointing at the year |
| 11:28 | <Hixie> | which seems unnecessary, but is probably the closest to a valid use case in the set |
| 11:29 | <BenMillard> | summary="Egutegia" |
| 11:29 | <Hixie> | my conclusion is that the data doesn't support the assertion that headers="" is indispensable on <th> |
| 11:29 | <Hixie> | maybe, in some cases, it's marginally useful |
| 11:29 | <annevk2> | Hixie, maastricht.nl doesn't use headers="" or <td>; I think that page works fine probably if screen readers just read from LTR TTB |
| 11:29 | <Hixie> | but that's as much as this data suggests |
| 11:29 | <annevk2> | I meant, or <th> |
| 11:29 | <Hixie> | annevk2: be sure to look at the textonly version |
| 11:30 | <annevk2> | hmm |
| 11:30 | <Hixie> | anyway, that's enough time wasted on this for now, back to forms... |
| 11:31 | <annevk2> | oh yeah, when I use curl I get indeed something that's wrong |
| 11:31 | <annevk2> | may I ever fail at my current job, I can always start fixing random sites on the Web |
| 11:32 | hsivonen | wonders if fixing random sites pays rent/mortgage |
| 11:34 | <annevk2> | ok, not so random then :) |
| 11:37 | <annevk2> | oh shit, lunch |
| 11:42 | <BenMillard> | hsivonen, fixing sites (usually CSS layout problems) is part of my day job and pays Ā£25 per hour but work comes in at irregular intervals. |
| 11:42 | <BenMillard> | auditing government websites against web standards and writing a report pays the same and is a bit more regular |
| 11:43 | <BenMillard> | I know people who charge incredibly larger amounts for doing either type of work, though |
| 11:43 | <hsivonen> | BenMillard: not random, though, I presume. Do people with broken sites find you or do you find them? |
| 11:44 | <BenMillard> | hsivonen, it's a bit of both. |
| 11:44 | <hsivonen> | it's encouraging that there's business in fixing broken stuff on the Web |
| 11:44 | <BenMillard> | yeah, I think so :) |
| 11:45 | <BenMillard> | my boss (http://www.sdesign1.com) sometimes approaches places, other times people we've worked with before come back to us for more, sometimes other agencies subcontract us |
| 11:46 | <BenMillard> | there's definitely work to be had doing that kind of thing, but the site commissioner needs to care about it in the first place |
| 12:04 | <zcorpan_> | hsivonen: the <ol><li></li> ...rest of page... case isn't the html parser though, is it? or maybe you could detect that in the parser |
| 12:05 | <hsivonen> | zcorpan_: I was thinking of making the fatal check in the tree builder |
| 12:05 | <zcorpan_> | hsivonen: btw, could the bugmail include full URLs? |
| 12:05 | <zcorpan_> | hsivonen: ok |
| 12:05 | <hsivonen> | zcorpan_: I'll try to find the configuration parameter |
| 12:07 | <hsivonen> | zcorpan_: I have now set the urlbase bugzilla parameter |
| 12:07 | <zcorpan_> | hsivonen: thanks |
| 12:11 | <zcorpan_> | Hixie: perhaps <dfn>s should have tabindex=0 |
| 12:12 | <Hixie> | on the long term, as i was telling BenMillard, my plan is to do all this statically with the spec gen tool |
| 12:13 | <Hixie> | at which point, we should consider this stuff |
| 12:13 | <Hixie> | until then, i'm happy to just wait |
| 12:14 | <annevk2> | by static you mean generating an awful lot of additional markup to be downloaded? |
| 12:14 | <Hixie> | yes |
| 12:14 | <Hixie> | it'll be offset by the removal of the data templates section :-P |
| 12:15 | <annevk2> | i'll make sure to measure that |
| 12:15 | <Hixie> | :-P |
| 12:15 | <Hixie> | anyway, the bonus is it'll then work in the static copy too |
| 12:15 | <Hixie> | multipage even |
| 12:19 | <annevk2> | it's too bad Content-Location failed |
| 12:20 | <annevk2> | I was thinking of setting up an HTML5 archive service that would allow you to get a given SVN version |
| 12:20 | <annevk2> | would've been a lot easier with Content-Location |
| 12:29 | <Philip`> | Like using something like Trac, where you can view any revision of the repository? |
| 12:31 | <annevk2> | not all resources are in the repository |
| 12:38 | <BenMillard> | Hixie, another bonus is links are a more familiar UI for navigating between and within documents on the web. |
| 12:39 | <Hixie> | well, links are what i use to do the actual navigating |
| 12:39 | <Hixie> | but there has to be some way to show the links |
| 12:41 | <zcorpan_> | Hixie: what's a "value URL"? |
| 12:41 | <Hixie> | should be valid |
| 12:41 | <Hixie> | will fix |
| 12:44 | <BenMillard> | Hixie, I'm looking at the single-page version in Opera and see how it works now |
| 12:49 | <Hixie> | ok bed time |
| 12:49 | <Hixie> | nn |
| 12:50 | <BenMillard> | Hixie & annevk2, so <a href title="Referenced by ..."><dfn>...</dfn></a> could work directly when there's only 1 reference, but for multiple references <a onclick title="View References"><dfn>...</dfn></a> would show the popup list of links? |
| 12:50 | <BenMillard> | (or something to that effect) |
| 12:51 | <zcorpan_> | BenMillard: what's wrong with <dfn onclick>? |
| 12:52 | <annevk2> | I don't think <a href="..."><dfn> makes sense, I'd rather have <dfn onclick> consistently |
| 12:53 | annevk2 | grmbls at the access control cache stuff |
| 12:53 | <BenMillard> | zcorpan_, if it's "something to that effect" then I'd say it's fine |
| 12:54 | <zcorpan_> | i'd like to be able to open the popup with the keyboard |
| 12:57 | <annevk2> | krijnh, https links don't work and the link in http://krijnhoetmer.nl/irc-logs/html-wg/20080816 is also wrong |
| 13:00 | <BenMillard> | zcorpan_, maybe <a onclick> should be focusable like <a href> is? would doing that for <dfn onclick> also make sense? |
| 13:01 | <annevk2> | yes |
| 13:01 | <zcorpan_> | BenMillard: not sure all onclicks should be focussable by default but you could set tabIndex = '0' |
| 13:02 | <zcorpan_> | e.g. i might want to set onclick on a container and do something with the target without wanting the container to be focusable |
| 13:03 | <BenMillard> | zcorpan_, how would a non-pointer user operate that feature? |
| 13:04 | <zcorpan_> | BenMillard: e.g. by activating links in the container |
| 13:04 | <krijnh> | annevk2: somebody smarter than me should fix the regex making those links |
| 13:04 | <annevk2> | krijnh, https should work by s? or something no? |
| 13:04 | <BenMillard> | zcorpan_, I thought <a href> was allowed more exotic content now, so you could wrap some (or all) of cases in <a href> ? |
| 13:04 | <annevk2> | or [s], hmm |
| 13:05 | annevk2 | sighs |
| 13:05 | <krijnh> | The issue is bigger though |
| 13:05 | <zcorpan_> | BenMillard: you can't nest links |
| 13:05 | <zcorpan_> | BenMillard: consider a menu with expandable items |
| 13:05 | <zcorpan_> | BenMillard: without script you fall back to server side expansion |
| 13:05 | <annevk2> | maybe it should be an app-engine service |
| 13:06 | <BenMillard> | zcorpan_, oh, I see |
| 13:06 | <annevk2> | pass a string of text and it will linkify stuff |
| 13:06 | <krijnh> | annevk2: https links work now |
| 13:06 | <annevk2> | sweet |
| 13:06 | <krijnh> | Testing http://krijnhoetmer.nl/irc-logs/ testing |
| 13:06 | <krijnh> | Testing http://krijnhoetmer.nl/irc-logs/'; testing |
| 13:06 | <krijnh> | Hmm |
| 13:07 | <annevk2> | it shouldn't consider - to be special, except when it's followed by a space or EOL |
| 13:07 | <BenMillard> | zcorpan_, but if that expansion is useful to pointer users, it would also be useful to non-pointer users? so the onclick should still be focusable so they can activate it to expand/collapse that section, or just Tab past it to the next section of navigation? |
| 13:08 | <annevk2> | annevk2, well, and ., etc. :p |
| 13:08 | <zcorpan_> | BenMillard: it is available to non-pointer users |
| 13:08 | <krijnh> | Yah, teh win \o/ |
| 13:08 | <zcorpan_> | BenMillard: activating a link with the keyboard sends a click event |
| 13:09 | <zcorpan_> | BenMillard: you wouldn't want to focus the *container* |
| 13:09 | <annevk2> | krijnh, maybe you should open source the log converter script so we can suggest improvements more easily |
| 13:09 | <krijnh> | First suggestion would be: "Rewrite this, you fool!" |
| 13:09 | <krijnh> | :) |
| 13:09 | <krijnh> | (from scratch, that is) |
| 13:10 | <krijnh> | Testin' http://krijnhoetmer.nl/irc-logs/. |
| 13:11 | <Philip`> | krijnh: Put the live source code on a wiki, so we can insert backdoor security vulnerabili-- uh, I mean bug fixes |
| 13:11 | <krijnh> | Enough spam for today |
| 13:11 | <annevk2> | Testin' http://krijnhoetmer.nl/irc-logs/, |
| 13:11 | <krijnh> | FU! :P |
| 13:11 | <annevk2> | win win win |
| 13:11 | <krijnh> | And fixed |
| 13:12 | <krijnh> | Although http://foobar.com/?p=this,sucks http://foobar.com/?p=this,sucks&this=aswell probably breaks now |
| 13:12 | <krijnh> | Ow wow |
| 13:12 | <annevk2> | you should patent this and sell the algorithm, it's better than Opera Mail and my IRC client at this point |
| 13:13 | <annevk2> | Testing http://krijnhoetmer.nl/irc-logs/"; |
| 13:13 | <krijnh> | Opera management: when reading this, you can hire m e;) |
| 13:13 | <krijnh> | me even |
| 13:13 | Philip` | 's IRC client never makes mistakes in detecting URLs |
| 13:13 | <Philip`> | (since it doesn't try detecting them at all) |
| 13:14 | <BenMillard> | all those seem to work in Opera's IRC client (which is what I use) |
| 13:14 | <Philip`> | (Click-and-drag to select and then copy-and-paste works perfectly well) |
| 13:14 | <zcorpan_> | BenMillard: not the apostrophe test |
| 13:15 | <zcorpan_> | Testing http://krijnhoetmer.nl/irc-logs/) |
| 13:15 | <BenMillard> | oh yeah, it was stuck to the end of the / so I didn't see it :P |
| 13:15 | <zcorpan_> | Testing http://krijnhoetmer.nl/irc-logs/)x |
| 13:15 | <BenMillard> | there are URLs which end with (foo), such as the disambiguation at the end of of some Wikipedia URLs |
| 13:15 | <zcorpan_> | hey maybe it'll work for live dom viewer urls by now |
| 13:16 | <krijnh> | http://disambiguation.com/wiki(foo) |
| 13:16 | <krijnh> | Bleh |
| 13:16 | <BenMillard> | Opera IRC gets that wrong |
| 13:16 | <annevk2> | krijnh too :/ |
| 13:16 | <krijnh> | Yeah |
| 13:16 | <annevk2> | I guess you need to do () matching and such :/ |
| 13:16 | <BenMillard> | if there's a ( in the URL, I guess you should include one ) |
| 13:17 | <Philip`> | It should work in most cases if you only match balanced parentheses, which is provably impossible in regular expressions but fortunately real regular expressions aren't regular |
| 13:17 | <krijnh> | Foo sucks (as seen on http://foo.com/) |
| 13:17 | <krijnh> | I think Philip` is the only one using ( and ) in this channel :) |
| 13:18 | <krijnh> | Oh well |
| 13:18 | <krijnh> | This is Good Enough(tm) |
| 13:18 | <krijnh> | You can continue your silly chattery stuff :) |
| 13:18 | <BenMillard> | it's certainly an invaluable service you provide |
| 13:18 | <BenMillard> | archiving our silly chattery stuff |
| 13:18 | <BenMillard> | the world thanks you! |
| 13:18 | <krijnh> | Nothing compared to what you guys are doing :) |
| 13:19 | <krijnh> | </slimeball> |
| 13:19 | <BenMillard> | :D |
| 13:19 | <krijnh> | I should ask some Mozilla funding for these logs ;) |
| 13:20 | <annevk2> | a server with a bit more speed wouldn't hurt |
| 13:20 | <krijnh> | I think it's my internet connection, not my server |
| 13:25 | <annevk2> | I find http://juicystudio.com/wcag/tables/nestedheaders.html highly confusing |
| 13:26 | <annevk2> | oh well, back to access control |
| 13:27 | Philip` | doesn't see why it's confusing |
| 13:27 | <krijnh> | If you want more feedback on that, rename it to accessibility control |
| 13:30 | <annevk2> | Philip`, maybe I'm bad with tables :) |
| 13:41 | <zcorpan_> | the summary in that table seems useful for anyone |
| 13:42 | <annevk2> | hmm, you side tracked me again; looks like a <caption> to me |
| 13:42 | annevk2 | back to caches |
| 14:09 | <jgraham> | That email from Gez was pretty useful, it actually identified a specific problem which could be good to consider |
| 14:10 | <jgraham> | since the solution in this case is pretty obvious and might well generalise |
| 14:14 | <annevk2> | i was thinking the same |
| 14:14 | annevk2 | made some progress meanwhile |
| 14:23 | <zcorpan_> | jgraham: i thought the smart headers algorithm did the right thing in that case |
| 14:24 | <zcorpan_> | at least iirc it was a case ben and i were discussing when he was looking in to this first |
| 14:24 | <hsivonen> | zcorpan_: the example was contrived, so there's insufficient data regarding what the real-world Right Thing is here |
| 14:25 | hsivonen | expects Ben to have better data |
| 14:25 | <zcorpan_> | hsivonen: yes we had actual tables with the same thing except for columns iirc |
| 14:28 | <zcorpan_> | or perhaps it was a colspan case we were discussing |
| 14:28 | <jgraham> | zcorpan_: Maybe it got lost in the great smart foo rewrite that I did |
| 14:29 | <zcorpan_> | jgraham: clearly we need regression tests :) |
| 14:29 | <jgraham> | zcorpan_: Does it make sense for columns also? |
| 14:29 | <jgraham> | zcorpan_: I had regression tests at one point but I don't know what I did with them |
| 14:29 | jgraham | was sure he checked them in to svn |
| 14:30 | <zcorpan_> | jgraham: yeah in the tables we looked at it was columns |
| 14:31 | <zcorpan_> | but this was so long ago i don't remember the details |
| 14:32 | <jgraham> | zcorpan_: Should be fixed now |
| 14:33 | <zcorpan_> | jgraham: cool |
| 14:38 | <zcorpan_> | hmm actually |
| 14:38 | <zcorpan_> | it doesn't work with http://projectcerbera.com/web/study/2007/tables/clark2006/06-gui/minimal |
| 14:41 | <zcorpan_> | jgraham: btw, would it make sense to preserve dir="rtl"? |
| 14:41 | <zcorpan_> | jgraham: in the inspector |
| 14:45 | <zcorpan_> | maybe it was this case i was thinking of http://projectcerbera.com/web/study/2007/tables/clark2006/11-controls/minimal |
| 14:46 | <zcorpan_> | where headers of the same width would be skipped after seeing a TD in between |
| 14:48 | <zcorpan_> | http://projectcerbera.com/web/study/2007/tables/clark2006/19-movies/minimal is another case where it fails |
| 14:49 | <zcorpan_> | dunno about http://projectcerbera.com/web/study/2007/tables/clark2006/23-watches/minimal |
| 14:49 | <zcorpan_> | http://projectcerbera.com/web/study/2007/tables/finance/01-wisc-rev/minimal another example of the case i was probably thinking of |
| 14:50 | <zcorpan_> | and http://projectcerbera.com/web/study/2007/tables/finance/02-wisc-alloc/minimal |
| 14:51 | <zcorpan_> | and http://projectcerbera.com/web/study/2007/tables/thatcher/01-water/minimal |
| 14:52 | <zcorpan_> | http://projectcerbera.com/web/study/2007/tables/thatcher/02-monitor/minimal might be a case of a "conceptual header data cell" |
| 14:54 | <annevk2> | I'm a bit uncomfortable with the statement that intranet usage should be given as much weight as Web usage... They're quite different after all |
| 14:58 | <hsivonen> | annevk2: they indeed are quite different. However, are Google's apps more like intranet apps or more like the Web? |
| 14:59 | <hsivonen> | I'd expect Web technology to be for the Web and intranets be an incidental spillover--not a design target. |
| 15:01 | <hsivonen> | in fact, so far it seems that bad design happens when intranets become the design target rather than the spillover |
| 15:09 | <annevk2> | hsivonen, lots of intranets are a "design target" |
| 15:10 | <annevk2> | hsivonen, IE serves two different markets, the intranet and the Web, one of the issues the IE Team is having |
| 15:12 | <smedero> | Don |
| 15:12 | <smedero> | erm |
| 15:12 | <smedero> | (itchy trigger finger) |
| 15:12 | <hsivonen> | annevk2: isn't it exactly a problem that IE has two markets to serve? |
| 15:13 | <hsivonen> | annevk2: or more to the point: that their intranet market limits agility in the Web market |
| 15:16 | <hsivonen> | interestingly, http://www.webaim.org/discussion/mail_message.php?id=10610 mentions axis |
| 15:17 | <annevk2> | hsivonen, yes, hence me being uncomfortable with it :) |
| 15:45 | <annevk2> | reviews appreciated: http://dev.w3.org/2006/waf/access-control/ |
| 15:48 | <Philip`> | annevk2: s/JSRONRequest/JSONRequest/ |
| 15:58 | <hsivonen> | annevk2: is that version of the spec in Minefield nightlies? |
| 16:02 | <annevk2> | hsivonen, sicking claims having it implemented, not sure it landed |
| 16:03 | <hsivonen> | annevk2: is now a sensible time to sync my server side impl, in your opinion? |
| 16:05 | <annevk2> | the author API should be stable (and I hope I'm right this time) |
| 16:06 | <hsivonen> | ok. I'll sync. |
| 17:10 | <BenMillard> | Great Scot! My e-mail client just caught up with this thread: http://lists.w3.org/Archives/Public/public-html/2008Sep/thread.html#msg362 |
| 17:10 | BenMillard | is glad hsivonen wrote up that response for the list. |
| 17:15 | <BenMillard> | LOL! http://www.paciellogroup.com/blog/misc/complexdatatable.html |
| 17:17 | <annevk2> | where you see LOL, I see more pain for implementors |
| 17:22 | <BenMillard> | annevk2, I'm guessing any implementor would LOL at that table and therefore it won't get implemented. |
| 17:23 | <annevk2> | :p |
| 17:26 | <BenMillard> | like I said, I live in hope :) |
| 17:27 | <annevk2> | I'm not sure what the status of aria-labelledby is |
| 17:27 | <annevk2> | I do know people will misspell that attribute all the time |
| 17:28 | <BenMillard> | ...and use it to associate things which are right next to each other |
| 17:29 | <annevk2> | yeah, sometimes I suspect these WAI groups of making things intentionally complicated |
| 17:30 | <annevk2> | for instance, when ARIA was still full of namespaces |
| 17:30 | <annevk2> | (and ARIA itself is highly problematic too, of course, in terms of ease of usage etc., though some of that is mitigated by libraries) |
| 17:32 | <BenMillard> | annevk2, I guess they haven't taken the red pill yet? (http://annevankesteren.nl/2007/04/html-red-pill) |
| 17:41 | <annevk2> | hmm, #html-wg discussing ownage of rel= values |
| 17:41 | <annevk2> | apparently the XHTML2 WG claims ownership |
| 17:41 | <annevk2> | fortunately none of it really impacts anything |
| 17:47 | <smedero> | annevk2: that would be this email in case you missed it: http://lists.w3.org/Archives/Public/www-tag/2008Sep/0037.html |
| 17:49 | <annevk2> | hah |
| 18:23 | <BenMillard> | I just checked the logs and saw the discussion about Smart Headers with spanned cells and suchlike |
| 18:29 | <BenMillard> | Gez Lemon wrote: 'The table inspector reports "Budgeted", "Actual", and "Forecasted" as headers for the data in the columns "TCO", "ROI", and "NPV". That's not correct. The automatic association has assumed the headers run to the end of the row.' http://juicystudio.com/wcag/tables/nestedheaders.html |
| 18:30 | <BenMillard> | I think that is indeed a case zcorpan and I talked about way back last year |
| 18:31 | <BenMillard> | it was his idea to apply the same cleverness to rowspan that we had found useful for colspan, but based on observing real tables |
| 18:31 | <BenMillard> | s/but/both/ |
| 21:06 | <gsnedders> | Hixie: If you want me to do something, email me :P |
| 21:32 | <Hixie> | gsnedders: i don't yet |
| 22:33 | <Hixie> | 1e1 is scientific notation, 0.1e3 is engineering notation, what do we call "10" again? |
| 22:37 | <gsnedders> | Hixie: Sanity. |
| 22:37 | <Dashiva> | Hixie: decimal notation? |
| 22:38 | <Hixie> | Dashiva: aren't the other two also decimal notation? |
| 22:39 | <Dashiva> | You could call it 'ordinary decimal notation' to be explicit |
| 22:39 | <webben> | Hixie: You could express it in a regex to be 100% specific maybe |
| 22:40 | <Hixie> | webben: oh it'll be expressed in detail, i just need a section title |
| 22:40 | <Dashiva> | decimal-digits-with-possible-minus-sign notation |
| 22:40 | <webben> | "Long decimal notation"? |
| 22:40 | <Hixie> | yeah see that becomes a bit verbose |
| 22:40 | <Hixie> | "10" is shorter than both 1e1 and 0.1e3 :-P |
| 22:41 | <Dashiva> | Wikipedia article on scientific notation called it 'ordinary' when comparing |
| 22:41 | <Hixie> | hm ok |
| 22:41 | <Hixie> | thanks |
| 22:42 | <webben> | http://www.google.com/search?hl=en&q=%22ordinary+decimal+notation%22&btnG=Search looks promising |
| 22:43 | <Hixie> | hmm... |
| 22:43 | <Hixie> | looks like i can just do away with the ordinary form |
| 22:43 | <Hixie> | and support the scientific form |
| 22:43 | <Hixie> | since i only use the ordinary form for new features in html5... |
| 23:00 | <Hixie> | anyone know of a way i could forward a usb port from one machine to another over the network? |
| 23:01 | <Hixie> | i.e. plug in a device in one machine and have it appear on another machine as a local usb device? |
| 23:11 | <gavin> | "usb over ip" seems to return some promising results |
| 23:12 | <Hixie> | oooh, good call |
| 23:18 | <Hixie> | no good mac solutions it seems |
| 23:18 | <Hixie> | oh well |
| 23:20 | <phroggy> | hi! How do you guys feel about Apple's <input type="search">? |
| 23:21 | <phroggy> | I see tons of sites using JS/CSS hacks to implement similar functionality; it seems like standardizing it should be helpful, and Apple already has a working implementation. |
| 23:27 | <webben> | phroggy: What exactly would you require of a user agent to conform? |
| 23:28 | <webben> | phroggy: and what problem does it solve? |
| 23:28 | <smedero> | a while back lachlan & mpt also (briefly) made the case for <input type="search"> here: http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2007-May/011367.html |
| 23:29 | <phroggy> | here's the details as I understand them: https://bugzilla.mozilla.org/show_bug.cgi?id=456229 |
| 23:30 | othermaciej | likes <input type="search"> |
| 23:31 | <othermaciej> | and also placeholder |
| 23:31 | <phroggy> | I've heard that the "placeholder" option is being added to all (relevant) input types (not just search); I'm not sure if that's an official HTML 5 thing or just something Mozilla is doing. |
| 23:31 | <othermaciej> | which is separately useful even for non-search input types |
| 23:31 | <phroggy> | agreed |
| 23:31 | <othermaciej> | WebKit has placeholder on all text-like input types |
| 23:32 | <othermaciej> | some of the <input type="search> stuff duplicates more general mechanisms in WF2 |
| 23:32 | <webben> | the autocomplete stuff? |
| 23:32 | <othermaciej> | I would like to deprecate them when we add the relevant WF2 stuff but not sure we can ever remove |
| 23:32 | <othermaciej> | well incremental |
| 23:32 | <othermaciej> | and the event that results |
| 23:32 | <phroggy> | that makes sense. |
| 23:32 | <othermaciej> | would be oninput |
| 23:32 | <othermaciej> | in WF2 |
| 23:32 | <webben> | ah okay |
| 23:33 | <phroggy> | the main issue is, right now everybody is reinventing the wheel, and some implementations are better than others. |
| 23:34 | <othermaciej> | Mozilla sounds like they are just copying what WebKit has |
| 23:34 | <othermaciej> | so I don't think there is reinvention exactly |
| 23:34 | <phroggy> | if there were a standard way to do it in HTML, that everybody could use, then they'd just use that, and either use JS to fake compatibility with old browsers (same as they're doing now), or let old browsers have reduced functionality if that's easier. |
| 23:35 | <phroggy> | othermaciej: that Mozilla bug was something I recently submitted, when I couldn't find an existing one. |
| 23:35 | <phroggy> | I attempted to describe Apple's implementation, but that doesn't mean Mozilla is actually doing it. |
| 23:35 | <phroggy> | I think they did say placeholder will be supported soon though. |
| 23:36 | <webben> | phroggy: Do you think devs are after the Safari functionality or the Safari look-and-feel? |
| 23:36 | <othermaciej> | phroggy: I see |
| 23:36 | <othermaciej> | I think it would probably be better to support <input type="search"> with the relevant WF2 replacement features |
| 23:37 | <webben> | phroggy: If they were offered IE-native, Fx-native, Op-native look-and-feel would they take it? Or would they continue to try to craft out the Safari look? |
| 23:37 | <phroggy> | well, one of the things Google mentioned was that they feel Safari's rounded corners would be inappropriate on Windows, so they'd like to come up with a different appearance |
| 23:38 | <webben> | http://www.w3.org/TR/css3-ui/ and http://www.w3.org/TR/css3-background/ are two other drafts that seem relevant here. |
| 23:38 | <webben> | phroggy: Oh I suspect browser makers would make the interface look native; I'm more thinking about what site designers are trying to do. |
| 23:39 | <phroggy> | oh, I gotcha |
| 23:40 | <phroggy> | no, I think web designers aren't too picky about the appearance, although obviously I'd expect Apple to continue making the search box on www.apple.com look the way it does today |
| 23:41 | <webben> | hmm, less certain of either :) |
| 23:41 | <webben> | I suspect the search box on www.apple.com will track Apple UI changes generally. |
| 23:42 | <phroggy> | personally I love the little magnifying glass icon - it's a bit weird that it's the "results" option that makes it appear, but having a tiny little icon like that, that I don't have to make myself, the appearance of which could be platform-specific, would be great. |
| 23:42 | <phroggy> | heh, well yeah, that's what I meant ;-) |
| 23:43 | <roc> | we already have code to style search boxes with platform-specific look |
| 23:43 | <roc> | we use it in XUL |
| 23:43 | <webben> | I'd tend to agree that being able to distinguish search inputs (I wouldn't want to be more specific than that in terms of conformance) and being able to jump to search inputs (e.g. with a quick key) would be good. |
| 23:44 | <phroggy> | yeah, Firefox uses it for things like the download window |
| 23:44 | <webben> | seems likely to make the web a lil more usable :) |
| 23:45 | <webben> | although it's possible jumping might be better helped with a container for the form as a whole (e.g. in case of search instructions). |
| 23:48 | <phroggy> | I recently implemented something where I wanted the functionality to basically work like searching in iTunes - it filters the results with each keypress. I had to experiment with using onkeypress, onkeydown, onkeyup, see what worked and what didn't |
| 23:51 | <phroggy> | I wound up using setTimeout with a 50ms delay, to make sure the value of the textbox would have changed by the time my function was called... |
| 23:52 | <phroggy> | and added code to that function to copy the value to a variable, after checking to see if it was different, to avoid repeatedly searching for the same string |
| 23:53 | <phroggy> | anyway, it just felt like I was reinventing the wheel, and probably poorly. |
