| 00:05 | <Hixie> | TabAtkins: so i asked around and the conclusion is that using <section><h1> may have some minor effect on SEO, but nothing serious, and we'll adapt the algorithms once we see how people are using it anyway |
| 00:05 | <TabAtkins> | Awesome. Need to file that away in case our idiot SEO company complains. |
| 00:07 | <TabAtkins> | JonathanNeal: Yo, dude, got an answer to the <h1> question. |
| 00:25 | <JonathanNeal> | Reading. |
| 01:11 | <JonathanNeal> | It takes me a long time to read a single paragraph :-D |
| 07:21 | <hsivonen> | Philip`: I distribute the source code for the GPL fonts I use via @font-face. |
| 07:24 | <hsivonen> | little Bobby Tables reminds me of Robert Smith? |
| 07:26 | <JonathanNeal> | I've made good progress on my normalize css @ http://madison.thewikies.com/html5/normalize.html and the fn() for adding support for those html5 elements in ie is pretty small now. |
| 07:48 | <othermaciej> | MikeSmith: what does the NE keyword mean? |
| 07:50 | <othermaciej> | I see, Non-Editorial |
| 08:31 | <othermaciej> | hsivonen: do the bugzilla bugs I filed about encodings seem sensible? |
| 08:41 | <MikeSmith> | othermaciej: I've just been using that NE keyword thing as a way to cause new bug notifications to get sent to the public-html list |
| 08:41 | <othermaciej> | MikeSmith: I see - I thought the bugmail was just automatic |
| 08:43 | <MikeSmith> | it was previously, but I turned it off after we got some spammy comments from the spec-commenting form |
| 08:45 | <othermaciej> | I see |
| 08:45 | <othermaciej> | I guess these were useful bugs for the list to see |
| 08:45 | <JonathanNeal> | The overflow of body is completely different in ie6/7. |
| 08:45 | <othermaciej> | I don't think I'm going to have the time/energy to send more suggested issue closings before tomorrow's telecon |
| 08:46 | <JonathanNeal> | It can't be set visible and behave in the same way as firefox / safari |
| 08:46 | <othermaciej> | so the next round of closings will have to target Sept 3 as the date I guess |
| 08:46 | <othermaciej> | JonathanNeal: quirks or standards mode? |
| 08:46 | <JonathanNeal> | the only solution where all browsers look the same seems to be to either remove the margin on the body as the default, or set overflow to hidden. |
| 08:46 | <JonathanNeal> | standards mode. |
| 08:47 | <othermaciej> | I think this might be too rarefied for my level of CSS understanding |
| 08:48 | <MikeSmith> | othermaciej: btw, I just checked in a version of my markup draft that includes the ARIA attributes -- based the html5+aria schema that hsivonen put together for validator.nu |
| 08:48 | <MikeSmith> | http://dev.w3.org/html5/markup/aria/ |
| 08:50 | <othermaciej> | MikeSmith: cool - I don't find chapter 10 very easy to understand still, but it seems better than the raw RNG |
| 08:51 | <MikeSmith> | othermaciej: yeah, slightly, at least |
| 08:51 | <MikeSmith> | fwiw, the per-element documentation there also shows where the aria attributes integrate in |
| 08:51 | <MikeSmith> | e.g., http://dev.w3.org/html5/markup/aria/address.html |
| 08:53 | <othermaciej> | MikeSmith: I see - still kinda requires following the RelaxNG links to see what is actually allowed though |
| 08:53 | <MikeSmith> | yeah |
| 08:54 | <othermaciej> | I would really like to see a modernized prose / matrix version |
| 08:54 | <JonathanNeal> | othermaciej, it's fine :-) I think this channel doesn't assume anyone is a css cross-browser-compatible guru, although, you never know. |
| 08:59 | <othermaciej> | good night folks |
| 09:02 | <annevk2> | nn |
| 09:06 | <hsivonen> | othermaciej: yeah, the bugs make sense. thanks |
| 09:27 | <hsivonen> | whoa! badness in the DOM spec: https://bugzilla.mozilla.org/show_bug.cgi?id=280692 |
| 09:33 | <annevk2> | hsivonen, I thought the idea was to fix the DOM spec for that case |
| 09:36 | <hsivonen> | annevk2: it seems that zcorpan's draft fixed it already. |
| 09:36 | <annevk2> | I wish someone would make that a standards track draft |
| 09:36 | <hsivonen> | we should definitely get more visibility for zcorpan's spec and remove DevMo links to bogus W3C specs |
| 09:37 | hsivonen | is editing DevMo DOM documentation for HTML5 truthiness |
| 10:01 | <hsivonen> | sigh. apparently Firefox 2 had cross-browser compatible getElementsByTagName behavior but was fixed to comply with the W3C spec |
| 10:32 | <jgraham> | <flamewar>Damn people who save code using the tab character for indentation |
| 10:33 | <annevk2> | you mean that as in they took your code, made a change, and now everything uses tabs? |
| 10:34 | <Lachy> | jgraham, using tabs is superior to spaces! |
| 10:34 | <Lachy> | (well, actually, using an appropriate mix of both so that indentation levels aren't adversely affected by different tab widths) |
| 10:35 | <jgraham> | annevk2: No they just used tabs in their file |
| 10:35 | <jgraham> | And now I have to edit it |
| 10:35 | <jgraham> | But changing my code would be even more annoying |
| 10:36 | <Lachy> | jgraham, is that in some Python code? |
| 10:36 | <Lachy> | for that, I agree, spaces are the only sensible option |
| 10:36 | <jgraham> | Lachy: It's javascript |
| 10:39 | <jgraham> | (when I said "changing my code" I actually meant the hypothetical situation where someone did change my code. In this case I could probably have learnt to reconfigure emacs to insert tabs rather than spacews but given that I will never share my changes I just replaced all the tabs) |
| 10:46 | <Philip`> | mpilgrim: The fonts don't strictly follow all the requirements of OpenType, e.g. they don't bother updating the global font bounding box stuff, and sometimes (rarely) they do nasty things like using invalid glyph IDs to disable bits of the font |
| 10:46 | <Philip`> | (I think Apple has some font validator tool which is useful, and finds a few problems) |
| 10:47 | <Philip`> | But they seem to work in all browsers/OSes I've tried, so I've not yet bothered with those minor details |
| 10:52 | <annevk2> | hsivonen, do you know where the Gecko details of handling encodings are? http://mxr.mozilla.org/mozilla-central/source/intl/uconv/src/charsetalias.properties lists aliases, are those all? Where is it specified that us-ascii is treated as windows-1252? |
| 10:52 | <annevk2> | (if anyone else involved in Mozilla/Gecko knows feel free to answer :) ) |
| 10:53 | <annevk2> | hsivonen, also, when you handle the parsing of <meta charset> do you perform any normalization on the value? |
| 10:57 | <virtuelv> | I'm moderately tempted to change /topic for this channel: http://translationparty.com/tp/#2157154 |
| 11:00 | <annevk2> | sweet |
| 11:00 | annevk2 | changes topic to 'WHATWG (HTML5) -- http://www.whatwg.org/ -- Logs: http://krijnhoetmer.nl/irc-logs/ -- For more information about the door, feel the gratitude of logic!' |
| 11:04 | <annevk2> | http://translationparty.com/tp/#2158174 |
| 11:05 | <gsnedders|work> | http://translationparty.com/tp/#2158350 |
| 11:05 | <annevk2> | heh, I found a long one: http://translationparty.com/tp/#2158317 |
| 11:07 | <annevk2> | "Whatever suits you." -> "Wrong suit." |
| 11:09 | <Rik`_> | annevk2: strange, without the dot, "Whatever suits you" -> "Accident litigation" |
| 11:12 | <gsnedders|work> | http://translationparty.com/tp/#2159339 |
| 11:12 | <gsnedders|work> | Peh. You guys are too distracting. |
| 11:20 | <jgraham> | I am quite disappointed I can't seem to cause an infinite loop; Google translate has a thing where it sometimes translates proper nouns incorrectly e.g. the word "Swedish" gets translated to "English" in a Swedish->English translation |
| 11:20 | <gsnedders|work> | (Also: it appears giving it lyrics from The Fly js quite effective.) |
| 11:21 | <gsnedders|work> | "Every artist is a cannibal, every poet a thief, all kill their inspiration and sing about the grief." -> "All thieves, poets, artists of the sad song of all meals." |
| 11:21 | <hsivonen> | annevk2: the alias file applies, but there are also other alias-like behaviors that aren't implemented as aliases |
| 11:22 | <hsivonen> | annevk2: the HTML5 parser in Gecko defers to the underlying APIs for name normalization |
| 11:24 | <annevk2> | hsivonen, any chance you can give me pointers to the relevant bits? |
| 11:24 | <annevk2> | hsivonen, then I can create some documentation out of it on the wiki |
| 11:32 | <hsivonen> | annevk2: it seems hard to find the relevant bits :-( |
| 11:32 | <hsivonen> | annevk2: it might be that certain decoders simply magically implement behavior that happens to match a superset |
| 11:32 | <hsivonen> | superset encoding, that is |
| 11:33 | <hsivonen> | annevk2: Simon Montagu or Jungshik Shin might be able to give better pointers |
| 12:12 | <Lachy> | this is the funniest one I've seen, mentioned in our #staff channel yesterday http://www.translationparty.com/#1728383 |
| 12:14 | <hsivonen> | http://future-web.org/ is still just Coming soon |
| 12:16 | <beowulf> | http://www.translationparty.com/ |
| 12:16 | <beowulf> | grr, wrong room |
| 12:18 | <jgraham> | Lachy: Thank you, that helped me find an infinite loop: http://www.translationparty.com/#2168120 |
| 12:26 | <annevk2> | hsivonen, could you pm their email addresses maybe? |
| 12:31 | <jgraham> | http://www.translationparty.com/#2168787 is also fun |
| 12:31 | jgraham | will give up trying to break translationparty now |
| 12:33 | <hsivonen> | annevk2: did I PM the right nick? |
| 16:37 | <paul_irish> | Philip`: can i PM? bug report about the web font optimizer. |
| 16:42 | <annevk2> | hsivonen, seems that Simon was on it already :) |
| 16:42 | <hsivonen> | annevk2: on what? |
| 16:42 | <annevk2> | documenting encoding stuff of Mozilla |
| 16:43 | <hsivonen> | nice |
| 16:43 | <annevk2> | https://wiki.mozilla.org/I18n:Charset_Aliases he'll integrate it in the WHATWG wiki later |
| 16:44 | <hsivonen> | ah. smontagu. not zcorpan. |
| 16:45 | <annevk2> | aah, I see the confusion :) |
| 16:50 | <gsnedders|work> | hsivonen: html5lib parser tests have #document-fragment — in that case, should it be standards or quirks? |
| 16:50 | <gsnedders|work> | (I guess we ought to make it possible to choose either) |
| 16:50 | <hsivonen> | gsnedders|work: Standars, I suppose, but strictly speaking, there needs to be a flag to test quirks |
| 16:50 | <hsivonen> | yay for the single quirk! |
| 16:51 | <gsnedders|work> | hsivonen: yay for acid2! |
| 16:56 | <TabAtkins> | Huh. Google just now changed the url for their search results. Yesterday it was google.com/search. Now it's just google.com, with all the query params stored in the hash. |
| 17:27 | <jgraham> | Random Python question: Anyone got any idea why proc = subprocess.Popen(args, stdout=subprocess.PIPE); stdout, stderr = proc.communicate() would only get the first 64kb of output |
| 17:28 | <jgraham> | ? |
| 17:31 | <jgraham> | (replacing stdout=subprocess.PIPE with a temp. file works so the program is producing the correct output) |
| 17:31 | <jgraham> | Philip`: (sorry to disturb you but you seem like the most likely person to know) |
| 17:33 | <jgraham> | Hmm http://bugs.python.org/issue1256 seems relevant |
| 18:34 | <gsnedders|work> | Does an iframe set to a data URI not fire load in Fx 3.6+html5? |
| 18:35 | <gsnedders|work> | It does, just not in the case I'm trying. WTF? |
| 20:40 | <Hixie> | anyone know what "Assistive technologies must treat an article like a document in that the article must be processed like an application." means? |
| 20:43 | <othermaciej> | context? |
| 20:43 | <othermaciej> | (the text itself I find baffling) |
| 20:44 | <Hixie> | http://www.w3.org/WAI/PF/aria/#article |
| 20:45 | <othermaciej> | I would assume "like a document" means "same as the document role" |
| 20:46 | <othermaciej> | however, "processed like an application" seems to directly contradict that |
| 20:46 | <othermaciej> | because the "document" role seems to be defined mainly be contrasting its behavior with the "application" role |
| 20:47 | Hixie | fails to figure out what role="" <title> should have by default |
| 20:48 | <othermaciej> | hsivonen suggested that it should have no role and not allow a role |
| 20:49 | <othermaciej> | (along with other special elements like <script> and <meta> that are really part of the document's content presentation) |
| 20:50 | <othermaciej> | in VoiceOver, the title is read when reading the whole window, in the order in which it appears in the chrome, and not read at all when browsing the content area, which would in fact correspond to no role |
| 20:50 | <Hixie> | wait, things without roles aren't read out? |
| 20:50 | <Hixie> | that can't be right |
| 20:50 | <othermaciej> | title doesn't fail to be read out because it has no role |
| 20:50 | <Hixie> | what happens if head,title{display:block} ? |
| 20:50 | <othermaciej> | it fails to be read out because it's not rendered |
| 20:52 | <othermaciej> | if title was visible, Safari would read it |
| 20:52 | <othermaciej> | is there a role appropriate to normal prose content? |
| 20:53 | <othermaciej> | I guess no role means normal prose content, for things that are rendered |
| 20:53 | <Hixie> | seem that way |
| 20:54 | <othermaciej> | I could see the argument for giving title a role of heading instead, but the case where it is made visible seems pretty unusual |
| 21:01 | <Hixie> | this aria thing is going to be a lot of work, sigh |
| 21:01 | <Hixie> | time for lunch |
| 22:25 | gsnedders | sighs at the suckiness that is British trains |
| 22:28 | <Hixie> | i wonder what role="" i should use for <input type=file> |
| 22:29 | Hixie | wonders how to make radio buttons work in aria |
| 22:30 | <gsnedders> | Ergh. £130 on three train tickets. |
| 22:32 | <gsnedders> | (Inevitably, being Britain, the trains _will_ be late.) |
| 22:35 | gsnedders | wonders how much more he'll spend on trains in Sept |
| 22:37 | <TabAtkins> | Hixie: Hrm, I don't see anything appropriate, at least in http://www.w3.org/TR/wai-aria/ |
| 22:38 | <TabAtkins> | For files, at least. |
| 22:38 | <gsnedders> | How the hell does it take 7 hours to get from Coventry to Leuchars? |
| 22:38 | <othermaciej_> | I think <input type="file"> can't be accurately represented by any existing ARIA role |
| 22:38 | <gsnedders> | (I <3 British trains) |
| 22:38 | <othermaciej_> | but you also can't fake a file input with anything else, by design |
| 22:39 | <TabAtkins> | Not sure how to represent <input type=radio> properly, as there's no requirement to have a "radiogroup" element. |
| 22:41 | <TabAtkins> | gnedders: wtf? Travelling across Britain north-south seems about twice as far as the distance between Houston and Dallas, which can be *driven* in just over 3 hours. |
| 22:41 | <TabAtkins> | And that 'seems' is with the misrepresentative Mercator projection I'm looking at. |
| 22:42 | <othermaciej> | HTML radio button groups are implicit and don't have an element, which I guess doesn't quite match the ARIA "radio" role semantics |
| 22:42 | <gsnedders> | TabAtkins: British trains suck. |
| 22:42 | <gsnedders> | TabAtkins: Though they suck less than American trains, as far as I can tell. |
| 22:42 | <TabAtkins> | Our trains do suck, I must admit. |
| 22:42 | <gsnedders> | TabAtkins: The East Coast Main Line is decent, though, taking around 4.5 hours to get from London to Edinburgh |
| 22:43 | <gsnedders> | TabAtkins: As well as the Channel Tunnel Train Link, from London to the Channel Tunnel… but that's the only high speed line in Britain :\ |
| 22:43 | <gsnedders> | TabAtkins: (Also: having a driving license helps at times.) |
| 22:43 | <annodomini> | In many cases, the American trains simply don't exist. |
| 22:44 | <gsnedders> | Well, Dr Beaching tried to make that true in Britain |
| 22:44 | <TabAtkins> | Hmm. That still blows. London to Edinburgh is 400 miles. Houston to Dallas is 230 miles. It takes about 3.5 hours to drive between houston and dallas. It shoudl take about 6 to drive between london and edinburgh. |
| 22:44 | <gsnedders> | *Beeching |
| 22:44 | <TabAtkins> | trains *should* go faster than cars. |
| 22:44 | <gsnedders> | "More than 4,000 miles of railway and 3,000 stations closed in the decade following the report, a reduction of 25 per cent of route miles and 50 per cent of stations. To this day in railway circles and among older people, particularly in parts of the country that suffered most from cuts, Beeching's name is still synonymous with mass closure of railways and loss of many local services." |
| 22:44 | <TabAtkins> | Wait. My brain just went crzy. |
| 22:44 | <TabAtkins> | 4.5 hours is fine. |
| 22:45 | <gsnedders> | It's not, at all. |
| 22:45 | <gsnedders> | This is Europe. It should have an average speed of at least 200 km/h on such a major route. |
| 22:45 | <TabAtkins> | Ah. Looks like it's average... about 90 mph. 150 km/h or so? |
| 22:47 | <gsnedders> | It's 632km (393 miles), with top speed of 200km/h (125mph). |
| 22:48 | <gsnedders> | It's also completely over capacity. |
| 22:51 | <gsnedders> | (Bear in mind that most other countries in Europe have at least 300 km/h high speed lines) |
| 22:52 | <TabAtkins> | Man, that sucks. |
| 22:53 | <JonathanNeal> | Hey TabAtkins, how are you doin'? |
| 22:53 | <JonathanNeal> | Don't let it get'cha down! :-D |
| 22:54 | <TabAtkins> | Doing good, JonathanNeal. |
| 22:54 | <TabAtkins> | Rewriting part of an intranet CMS because it's sucky and hard to navigate now that we have a lot of items in it. |
| 22:55 | <JonathanNeal> | Oh that's no fun, just rewriting the frontend or the backend or both? |
| 22:56 | <TabAtkins> | Just the front-end. That back-end is just fine. A touch crufty, but not enough to worry about. |
| 22:56 | <TabAtkins> | And it is fun! I get to stretch my information architect muscles. ^_^ |
| 22:57 | <TabAtkins> | Also, now, fixing a bug that I just discovered in my permissions manager. |
| 22:57 | <TabAtkins> | Damn namespaces. |
| 23:07 | <JonathanNeal> | TabAtkins, most of the chatty folks here seem to be devs of one kind or another, that include you? |
| 23:08 | <JonathanNeal> | of w3 or whatwg, I mean. |
| 23:08 | <TabAtkins> | Nah, I'm just webmaster for my company, and general webdev on the side. |
| 23:10 | <TabAtkins> | I don't do any work on browsers or anything like that. |
| 23:10 | <TabAtkins> | But I do like participating in the evolution of the languages that I use. |
| 23:17 | <annevk2> | hsivonen, http://bugzilla.validator.nu/ is broken |
| 23:28 | <nessy> | midnight there .... |
| 23:29 | <annevk2> | same timezone then :) |