| 00:02 | <Dashiva> | I don't think Ian's cat would be a good change controller |
| 00:09 | <AryehGregor> | Why not? |
| 00:10 | <AryehGregor> | Cats are known to be very intelligent. |
| 00:10 | <AryehGregor> | Well, my cats are complete idiots, but maybe I just have losers. |
| 00:10 | <Dashiva> | Cats have short lives, it might die before HTML6 becomes PR |
| 00:10 | <AryehGregor> | Maybe we can cryogenically freeze it. |
| 00:11 | <Philip`> | Write an AI simulation of the cat |
| 00:11 | <AryehGregor> | Should be easy. |
| 00:12 | <Dashiva> | That's a typically "the tools will save us" fallacy, Philip` |
| 00:12 | <Dashiva> | I expected better from you |
| 00:12 | <AryehGregor> | "Stare plaintively at human until petted or fed. Meow if necessary. Run away from thrown objects." |
| 00:13 | <AryehGregor> | We can cryogenically freeze cats right now! |
| 00:13 | <AryehGregor> | Also, it could be hereditary. |
| 00:13 | <Dashiva> | What if it has multiple kittens, and they disagree? |
| 00:14 | <AryehGregor> | This was solved millennia ago by the doctrine of primogeniture. |
| 00:14 | <AryehGregor> | Don't reinvent the wheel. |
| 00:14 | <Dashiva> | Primogeniture is a rathole you don't want to go down |
| 00:15 | <Dashiva> | You have to resolve issues like whether male descendents have priority, and who inherits when the current controller has no children, and how to handle inlaws, etc etc |
| 00:15 | <Dashiva> | Not to mention the potential for abuse by guardians of underage heirs |
| 00:16 | <Philip`> | Put all the kittens into a last-man-standing fight to the death |
| 00:16 | <Philip`> | and then the victor can make uncontested decisions |
| 00:16 | <Dashiva> | That has a high chance of leading to the winner dying of wounds |
| 00:17 | <Philip`> | Wait for all the kittens to have kittens of their own, and then put the parent cats into the deathmatch |
| 00:17 | <Philip`> | and if they all die then repeat with the younger generation |
| 00:17 | <Philip`> | until there is one survivor |
| 00:18 | <Dashiva> | But by the time the younger generation is old enough, you'll have been without a change controller for a long time |
| 00:18 | <Philip`> | Force-feed them with rapid aging drugs |
| 00:19 | <Dashiva> | Also, I don't think fights to the death conforms to local laws in most jurisdictions |
| 00:19 | <Philip`> | Do it on the moon, so it's not in anybody's jurisdiction |
| 00:19 | <Dashiva> | I don't think W3C can afford that |
| 00:20 | <Philip`> | I'm sure TV companies would pay you a lot for live broadcast rights to lunar kitten deathmatches |
| 00:21 | <Dashiva> | If that's the case there should already exist such ventures |
| 00:21 | <Dashiva> | And the W3C could utilize existing infrastructure |
| 00:21 | <Philip`> | I'm sure such ventures do exist - Unreal Tournament demonstrates one of them |
| 00:22 | <Dashiva> | I'm sorry to say, but Unreal Tournament is not a documentary |
| 00:22 | <Philip`> | Oh |
| 00:23 | Philip` | will have to make new business plans, then :-( |
| 01:10 | <AryehGregor> | One annoying thing about reading old spec mail is you aren't sure whether to respond to old messages. |
| 01:10 | <AryehGregor> | "It will probably mean that Mediawiki some time in the future are going to disallow it, as well as Facebook, since using the built in functionality sooner or later will trickle into their code base." |
| 01:11 | <AryehGregor> | Does that suppose that MediaWiki or Facebook actually uses FILTER_VALIDATE_EMAIL? |
| 01:11 | <AryehGregor> | Because in MediaWiki it's been strpos( $email, '@' ) !== false for a long time. |
| 01:11 | <AryehGregor> | Too much effort to validate e-mail addresses, especially when mail agents tend to ignore the RFCs in practice and accept bad addresses. |
| 01:12 | <othermaciej> | probably depends on the current state of the issue |
| 01:12 | <othermaciej> | there were some bugs filed on this topic |
| 01:12 | AryehGregor | responds on a bug instead |
| 01:13 | <othermaciej> | I don't think any of them have been escalated to tracker issues but I am not sure what the current state is |
| 01:15 | <AryehGregor> | www.w3.org/Bugs/Public/show_bug.cgi?id=9392 is WONTFIX. |
| 01:19 | <Dashiva> | I know facebook rejects + in emails |
| 01:19 | <Hixie> | AryehGregor: unless it gets escalated i wouldn't worry about it |
| 01:19 | <AryehGregor> | Guess not. |
| 01:19 | <Hixie> | ("it" being any general issue, not any one in particular) |
| 01:21 | <estellevw> | If the client has in memory maxed out browser cache, is there a spec as to how the browser / user agent should handle having no more local storage? |
| 01:22 | <Hixie> | estellevw: for appcache, http cache, local storage, cookies, something else? |
| 01:22 | <estellevw> | app cache |
| 01:23 | <Hixie> | no, nothing really defines that |
| 01:23 | <Hixie> | it's basically up to the ua |
| 01:23 | <estellevw> | i don't really see a way of controlling it, other than from the developer using dates... does the user have a way to set preferences. |
| 01:23 | <Hixie> | it falls under the "hardware limitations" clause |
| 01:23 | <estellevw> | ok |
| 01:24 | <estellevw> | should we be defining a path... oldest out first... or warning to user that their storage is full? |
| 01:24 | <estellevw> | or other? |
| 03:07 | <TabAtkins__> | AryehGregor: As far as I know, basically no one uses PHP's email validation filter, because it's buggy as all hell from what I understand. |
| 03:09 | <boblet> | re: email validation, I’d wager almost none of the scripts for that took multibyte domains into account |
| 03:09 | <boblet> | (random aside since I have one ;-) |
| 07:15 | <annevk> | hmm, Gecko has computed style for elements not in the DOM |
| 07:15 | <annevk> | how does that even work? |
| 07:17 | <annevk> | see e.g. http://dump.testsuite.org/2009/getcomputedstyle/liveness/002.htm |
| 07:17 | <annevk> | it's very weird |
| 07:17 | annevk | goes to download a nightly |
| 07:20 | <annevk> | same deal |
| 07:20 | <annevk> | CSS is matched against elements not even in the DOM, isn't that somewhat weird? |
| 07:47 | <annevk> | in other news, Tyler seems to continue suggesting forking CORS |
| 07:47 | <annevk> | http://lists.w3.org/Archives/Public/public-webapps/2010AprJun/0162.html |
| 07:47 | <annevk> | *sigh* |
| 07:49 | <Hixie> | i wish the hybi chairs would clear up once and for all whether amateur programmers are part of the target audience or not |
| 07:49 | <Hixie> | all we're doing is arguing around in circles because there are two completely different opinions about who the protocol should be aimed for |
| 07:50 | <annevk> | it seems Mike is sort of arguing for both sides, assuming SSL is too complex for amateur programmers |
| 07:50 | <Hixie> | i think he's saying what i said, namely that we should assume experts deploy using tls, but still allow a non-tls variant |
| 07:51 | <annevk> | doing everything over SSL and saving a roundtrip is the first novel suggestion in months on that list I think |
| 07:51 | <Hixie> | yeah |
| 07:51 | <MikeSmith> | a header named "U" could certainly not be accused of being verbose.. nor misleading, I guess.. since whatever meaning it's supposed to have can't be discerned from the name at all |
| 07:51 | <Hixie> | wish i understood enough about TLS to know if it made any sense :-) |
| 07:51 | <MikeSmith> | annevk: I think I'll post to suggest that the header should be named "X" |
| 07:52 | <annevk> | Hixie, ah yeah, I guess you can have two separate handshakes, one for non-tls usage and for tls usage |
| 07:52 | <annevk> | MikeSmith, it's more that he wants to replace current access control headers with something new |
| 07:53 | <MikeSmith> | yeah |
| 07:53 | <annevk> | MikeSmith, I think he's missing that this only solves the response side, the request side also has a whitelist |
| 07:53 | <MikeSmith> | I just particularly liked the suggestion for the alternativ eheader name |
| 07:53 | <annevk> | maybe I should point that out |
| 07:56 | <MikeSmith> | It seems like a lot of time might be saved if somebody from the Chrome team were to publicly say whether they'd have any plans at all to actually implement UMP |
| 07:56 | <MikeSmith> | or even not publicly |
| 07:59 | <annevk> | TabAtkins, maybe you can figure out what the Google Chromium team is interested in? |
| 07:59 | <annevk> | TabAtkins_, i.e. CORS or UMP |
| 07:59 | <annevk> | or Hixie, or anyone |
| 07:59 | <annevk> | it doesn't even matter so much I suppose |
| 08:00 | <annevk> | if CORS is not going to be removed it makes no sense to add UMP |
| 08:17 | <othermaciej> | ah, CORS vs UMP, too bad it's only about #10 on the list of total messes the I feel obligated to clean up |
| 08:18 | <othermaciej> | annevk: the Chrome security people I've talked to don't seem interested in removing CORS or in adding a separate UMP implementation, but no one has given an authoritative statement |
| 08:18 | <othermaciej> | MikeSmith: have any other browser vendors besides Apple made a public statement on the matter? |
| 08:20 | <annevk> | I could make a statement, but we haven't implemented yet so I'm not sure whether it's worth it |
| 08:22 | <othermaciej> | it seems relevant to me, since people specifically asked for statements of interest in implementing UMP or CORS |
| 08:22 | <othermaciej> | I'm not sure if there was any clear position taken by Mozilla |
| 08:25 | <MikeSmith> | othermaciej: none that I remember |
| 08:25 | <othermaciej> | or Microsoft for that matter |
| 08:26 | <othermaciej> | why were you specifically interested in the Chrome team's opinion? |
| 08:26 | <othermaciej> | fwiw neither Mark Miller nor Tyler Close are on the Chrome team, and Adam Barth is (as a contractor) |
| 08:27 | <annevk> | well, they implement it |
| 08:27 | <annevk> | and would need to implement UMP |
| 08:28 | <MikeSmith> | othermaciej: because they work at the same company |
| 08:28 | <othermaciej> | I don't know of them having any special inside track with the Chrome team, but I'm not the expert |
| 08:28 | <othermaciej> | certainly no one has proposed adding UMP to WebKit |
| 08:37 | <annevk> | emailed asking for implementor interest in a separate thread |
| 08:38 | <annevk> | I pointed out your reply in that email othermaciej and added Opera's position |
| 08:38 | <othermaciej> | gah, the hybi list generates stupid amounts of mail |
| 08:56 | <annevk> | othermaciej, if it was new information it might be more acceptable... |
| 08:56 | <MikeSmith> | hsivonen: any clue on what might be the problem here: |
| 08:56 | <MikeSmith> | http://validator.w3.org/check?uri=http%3A%2F%2Fwww.trigofacile.com%2Fjardins%2Flucullus%2Fgrec%2Fdemocrite%2Ffragments.htm&charset=%28detect+automatically%29&doctype=Inline&ss=1&group=0&user-agent=W3C_Validator%2F1.767#line-123 |
| 08:57 | <MikeSmith> | HTML5 facet of W3C validator is reporting "End of file seen and there were open elements", but validator.nu says its valid |
| 08:57 | <othermaciej> | annevk: I'm supposed to be a co-editor for the requirements document but it's almost impossible to keep up with email enough to know what has consensus |
| 08:57 | <annevk> | MikeSmith, some kind of zero byte issue? |
| 08:58 | <othermaciej> | annevk: I think the hybi chairs need to raise some trac issues on clear permathreads so we can stop discussing those points endlessly |
| 08:58 | <annevk> | some kind of chair guidance would be good yes |
| 08:58 | <MikeSmith> | annevk: zero byte? it seems to actually be reading in the whole file OK |
| 08:59 | <annevk> | maybe the W3C has smaller buffers or some such? |
| 08:59 | <hsivonen> | MikeSmith: that sure is odd |
| 09:00 | <hsivonen> | MikeSmith: does it work with W3C-hosted v.nu with the same parser revision? |
| 09:00 | <MikeSmith> | annevk: if so, that's not the error I'd expect to see, I guess |
| 09:00 | <hsivonen> | MikeSmith: i.e. without the perl front-end but otherwise the same |
| 09:01 | <MikeSmith> | hsivonen: yeah - http://www.w3.org/html/check?doc=http%3A%2F%2Fwww.trigofacile.com%2Fjardins%2Flucullus%2Fgrec%2Fdemocrite%2Ffragments.htm |
| 09:01 | <MikeSmith> | though that's a different instance and a different machine even |
| 09:02 | <MikeSmith> | must be due to some monkeying that the perl layer is doing |
| 09:03 | <hsivonen> | MikeSmith: which part is responsible for showing the source? |
| 09:03 | <hsivonen> | MikeSmith: that is, does the full source showing prove anything about the full source reaching the back end? |
| 09:03 | <MikeSmith> | hsivonen: hmm, yeah, it may not prove that |
| 09:04 | <annevk> | http://www.w3.org/2009/12/vnu/style.css uses padding:none which should be padding:0 or some such |
| 09:04 | <MikeSmith> | I don't know much at all about how the W3C valdiator works |
| 09:04 | <hsivonen> | ok. I'm going to claim the front end cuts the content |
| 09:04 | <MikeSmith> | annevk: that one is my fault, thanks |
| 09:04 | <annevk> | it also has various -pre-wrap stuff that can prolly become just pre-wrap |
| 09:04 | <JonathanNeal> | and 0px is te same as 0 |
| 09:04 | <JonathanNeal> | Those jerks! |
| 09:04 | <MikeSmith> | heh |
| 09:04 | <MikeSmith> | "jerk", singular |
| 09:04 | <MikeSmith> | =me |
| 09:05 | <MikeSmith> | I'm not a CSS Ninja |
| 09:05 | <JonathanNeal> | MikeSmith, doing something right in a way that could be done another way that others might prefer? |
| 09:05 | <MikeSmith> | I'm more of a CSS cro-magnon |
| 09:05 | <annevk> | CSS maximalist |
| 09:06 | <JonathanNeal> | I'm not wearing my "I Less-Than-Three MikeSmith" shirt tomorrow. |
| 09:06 | <MikeSmith> | JonathanNeal: I have perfected this method that I call "brute force trial-and-error" |
| 09:08 | <Hixie> | i just read an article that suggested the following as a strategy "Find a problem, find the simplest solution, punt the rest" |
| 09:08 | <Hixie> | i think that pretty much summarises a lot of how we've been approaching problems in html5 and related specs |
| 09:10 | <JonathanNeal> | I just want abbr! |
| 09:10 | <JonathanNeal> | Dangit ... I mean |
| 09:10 | <JonathanNeal> | alt! |
| 09:12 | <JonathanNeal> | the<wbr>rapist, what an infinitely useful element. |
| 09:12 | <JonathanNeal> | And by alt, I mean for microdata, Mr. Hixie. |
| 09:14 | <MikeSmith> | Hixie: no, that's all wrong. The correct strategy is to start with a vague "framework" that can be used to solve all kinds of problems -- including problems we didn't even know we had -- and then to introduce a magical "policy" mechanism to control it all, then people show up and ask to pile on all kinds of additional features and we do that to make them happy |
| 09:14 | <MikeSmith> | that's the way it's supposed to work |
| 09:17 | <wirepair> | wow you guys are bitter |
| 09:17 | <wirepair> | ;> |
| 09:22 | <Hixie> | wirepair might have us figured out :-P |
| 09:22 | <wirepair> | hehe |
| 09:39 | <MikeSmith> | hsivonen: I tested with localhost 8888 on the W3C validator host and can't reproduce that error |
| 09:39 | <MikeSmith> | so it must be some problem in the perl layer |
| 09:40 | <hsivonen> | MikeSmith: ok. Good. :-) |
| 09:46 | <ment> | hmm this is maybe not the best channel to ask this question but it's somehow vaguely releated - are there any guidelines for writing css3 rule matcher for html*/xhtml* or any kind of previous work in this field? |
| 10:25 | <othermaciej> | I think I just killed w3c bugzilla |
| 10:26 | <annevk> | finally made IRIs work on my blog |
| 10:26 | <Yudai> | |
| 10:26 | <Yudai> | oops |
| 10:26 | <Yudai> | MikeSmith: do you have any idea about "validator/entity-map" ? |
| 10:27 | <Yudai> | build.py all fails on loading the file |
| 10:29 | <othermaciej> | MikeSmith: I just discovered a downside to Cc-ing public-html on all bug keyword changes.... |
| 10:29 | <othermaciej> | MikeSmith: any chance we could turn that off temporarily? |
| 10:30 | <annevk> | he's in some kind of meeting currently I believe |
| 10:30 | <othermaciej> | I just spammed public-html with hundreds of emails accidentally :-( |
| 10:31 | <annevk> | oh, so those twenty-odd were just the beginning |
| 10:31 | <othermaciej> | yeah |
| 10:31 | <othermaciej> | there's gonna be, like, 250 |
| 10:31 | <othermaciej> | only noticed what was happening partway |
| 10:31 | <annevk> | I wonder what our company mail server will do |
| 10:36 | <roc> | I have nightmares about triggering something like this |
| 10:36 | <roc> | one time one of my kids found an open email window and invited a couple of dozen of my colleagues to GMail |
| 10:36 | <roc> | that was embarrassing |
| 10:37 | <othermaciej> | well - people will get to learn a lot about some of the old bugs we had hanging around :-/ |
| 10:37 | <othermaciej> | it could be worse, I could have accidentally dropped prototype hardware in a seedy bar |
| 10:39 | <Dashiva> | You're more anonymous that way, though |
| 10:39 | <othermaciej> | looking at these bugs in a mail folder, some of them seem really ridiculous |
| 10:41 | <annevk> | othermaciej, ah, so it wasn't you? :p |
| 10:42 | <othermaciej> | annevk: honestly - I have no idea if it's even real |
| 10:42 | <othermaciej> | that kind of stuff is on a need-to-know basis |
| 10:42 | <roc> | what are you talking about? |
| 10:42 | <othermaciej> | god damn did I ever mailbomb the list |
| 10:43 | <annevk> | engadget claimed to have found an next-gen iphone |
| 10:43 | <annevk> | s/an/a doh |
| 10:45 | <roc> | hehe |
| 10:46 | <Dashiva> | othermaciej: lists, even. public-html-bugzilla as well :) |
| 10:47 | <othermaciej> | Dashiva: anyone on public-html-bugzilla has volunteered to be spammed by bugmail, I think |
| 10:49 | <Dashiva> | Well, there's bugmail and then there's 250 messages adding NoReply |
| 10:50 | <Lachy> | Hixie, re this about mediawiki from a few days ago, I thought you knew I'd given AryehGregor access and basically made him responsible for upgrading the wiki. http://krijnhoetmer.nl/irc-logs/whatwg/20100416#l-1070 |
| 10:50 | <Dashiva> | It's too bad bugzilla can't generate one mail for each multichange |
| 10:52 | <Lachy> | othermaciej, what does the NoReply tag mean? |
| 10:52 | <othermaciej> | Dashiva: that would have been much better in this case |
| 10:52 | <othermaciej> | Lachy: per our decision policy, it means we're going to assume we are not getting a reply to this bug, and it is now in a terminal state |
| 10:52 | <othermaciej> | Lachy: these are the bugs I moved to VERIFIED a month ago |
| 10:53 | <Dashiva> | Timeout while waiting for CLOSED |
| 10:53 | <othermaciej> | Lachy: that didn't get reopened, closed, or moved to the tracker |
| 10:53 | <Lachy> | ah, ok. So they have to move from VERIFIED FIXED to CLOSED? |
| 10:55 | <MikeSmith> | othermaciej: sorry, was afk |
| 10:55 | <MikeSmith> | I can try to turn it off now |
| 10:56 | <MikeSmith> | but I think it might be too late |
| 10:56 | <othermaciej> | MikeSmith: I stopped bugzilla at half the bugs |
| 10:56 | <MikeSmith> | ok |
| 10:57 | <othermaciej> | Lachy: technically, the policy says they stay in VERIFIED with the NoReply keyword, but in the next rev of the decision policy I'm proposing CLOSED with the NoReply keyword |
| 10:57 | <Dashiva> | Luckily it looks like nobody posted anything during the bug storm |
| 10:57 | <Lachy> | but for the 3 bugs I filed, I can now move them to closed myself because they are fixed? |
| 10:57 | <Dashiva> | That's what you were supposed to do these preceding weeks :) |
| 10:58 | <Lachy> | yeah, but for the last few weeks, I haven't been paying too much attention to HTMLWG stuff |
| 10:58 | <Dashiva> | But yes, go ahead. You can remove the NoReply keyword when you set it to CLOSED. |
| 11:00 | <MikeSmith> | othermaciej: you still want me to disable e-mail notifications for now? |
| 11:00 | <othermaciej> | MikeSmith: that would be nice, yes |
| 11:00 | <MikeSmith> | ok |
| 11:01 | <othermaciej> | Lachy: yes, you can move stuff to CLOSED and remove NoReply (and optionally add Disagree) |
| 11:01 | <othermaciej> | MikeSmith: I still have 235 more bugs to mass-modify |
| 11:01 | <MikeSmith> | hai |
| 11:01 | <Lachy> | does Disagree mean that I don't agree with the fix? |
| 11:03 | <othermaciej> | Lachy: it means you disagree with the resolution, but you don't care to escalate, reopen, or raise a formal objection |
| 11:04 | <Lachy> | ok. I don't disagree with them anyway, but good to know. |
| 11:05 | <MikeSmith> | othermaciej: OK, I just disabled all bugzilla mail for the public-html⊙wo user |
| 11:06 | <othermaciej> | why do people keep bringing up the idea of tunneling SPDY over WebSocket? |
| 11:07 | <othermaciej> | o_O |
| 11:16 | <Lachy> | MikeSmith, thank you! I've always hated having public-html spammed with bug mail |
| 11:19 | <othermaciej> | MikeSmith: let me know when the coast is clear please |
| 11:26 | Lachy | realises how even more seriously annoying it is to be spammed with 500 duplicate bug mails each on both public-html and public-html-bugzilla, on the same day my mail server seems to be responding rather slowly :-( |
| 12:25 | <zcorpan> | Hixie: should the spec say to fail the connection if the server sends a binary frame with a length that would cause a buffer overflow (say we store the length in a 32bit int) |
| 16:47 | <TabAtkins> | annevk: I was going to ask around today. I'll give you a headsup on the thread later. |
| 17:43 | <franksalim> | othermaciej, re:SPDY over WebSocket, is that a rhetorical question? |
| 17:54 | <AryehGregor> | The real problem with the bug spam isn't the volume, it's that now I don't know which bugmail is actually meaningful, what with Gmail's conversation grouping . . . |
| 18:06 | <Philip`> | AryehGregor: Open your mail folder in an IMAP client, sort messages by date, select the ones in the offending time period, mark as read |
| 18:06 | <AryehGregor> | That sounds like more effort than just hitting ] a few hundred times. |
| 18:06 | <Dashiva> | Does it really |
| 18:06 | <AryehGregor> | (assuming Gmail doesn't decide that that should take ten seconds on a regular basis, as it is sometimes wont to do) |
| 18:07 | <AryehGregor> | Well, I've never used an IMAP client in my life. |
| 18:07 | <AryehGregor> | I used AOL up to 2005, and Gmail since then. |
| 18:07 | <jgraham> | Umm |
| 18:08 | <jgraham> | Wow |
| 18:08 | <gsnedders> | AryehGregor: Are you a masochist or something? |
| 18:08 | <Philip`> | It's fairly trivial to configure Thunderbird to access your Gmail account |
| 18:08 | <AryehGregor> | gsnedders, nope. Webmail is convenient, why should I use a real mail client? |
| 18:08 | <AryehGregor> | I'm not clear on the benefits. |
| 18:08 | <Philip`> | Apparently webmail is inconvenient in this particular situation |
| 18:09 | <AryehGregor> | Well, one particular design choice of Gmail (which I mostly like a lot) is inconvenient in this particular situation. |
| 18:09 | <AryehGregor> | (I mean, I mostly like this design choice a lot, threaded viewing) |
| 18:10 | <AryehGregor> | Things like this seem to come up rarely. |
| 18:10 | <Philip`> | It's good to have an alternative for the rare inconvenient situations |
| 18:10 | <Philip`> | and continue to use webmail when it's good enough |
| 18:11 | <Dashiva> | The nice thing about IMAP is that you can have both |
| 18:11 | <AryehGregor> | Well, I suppose I technically do. There's just never been a situation so inconvenient that I felt bothered to set it up. |
| 18:11 | <AryehGregor> | Maybe I should set up a mail client to at least download all my mail, though. |
| 18:11 | <jgraham> | AryehGregor: You know you can use a non-web client and a web client together on the smae mailbox, right? |
| 18:11 | <AryehGregor> | Yes, yes, I know. |
| 18:11 | <AryehGregor> | I've never seen the need. |
| 18:58 | <Hixie> | AryehGregor: do you know how to embed an external image on a wikimedia wiki? |
| 18:59 | <AryehGregor> | Hixie, you don't. You can't include any external content for privacy reasons. |
| 18:59 | <AryehGregor> | IP sniffing and such. |
| 18:59 | <Hixie> | Lachy_: ah, indeed, nevermind then, just my poor memory :-) |
| 19:00 | <AryehGregor> | MediaWiki supports it (although kludgily), it's just disabled by default and on Wikimedia. |
| 19:00 | <Hixie> | AryehGregor: hm |
| 19:00 | <AryehGregor> | Did you mean Wikimedia or MediaWiki? |
| 19:00 | <Hixie> | oh i meant mediawiki |
| 19:00 | <AryehGregor> | Oh, right. |
| 19:00 | <Hixie> | didn't know there was a difference |
| 19:01 | <AryehGregor> | Different story then. |
| 19:01 | <Hixie> | ah, gotta go.brb |
| 19:01 | <Hixie> | wanted to add them to our subtitles page |
| 19:01 | <AryehGregor> | The names were chosen before Wikimedia had enough money to consider thinking seriously about stuff like branding. |
| 19:01 | <AryehGregor> | So, well, they're terrible. |
| 19:01 | <AryehGregor> | Wikimedia is the charitable foundation that runs Wikipedia, MediaWiki is the name of the software. |
| 19:02 | <AryehGregor> | Anyway, the way MW supports external images last I checked is that if you enable the feature, then bare links that look like image URLs are turned into <img> instead of <a>, basically. |
| 19:02 | <Philip`> | Is there a Mikiwedia too? |
| 19:02 | <AryehGregor> | Which doesn't allow specifying any attributes or anything. |
| 19:02 | <AryehGregor> | Philip`, apparently. http://www.o--o.info/ |
| 19:03 | <daedb> | Can't you just whitelist img elements? |
| 19:03 | <AryehGregor> | Well, yes, if you hack the code. The element whitelist isn't configurable as it stands. |
| 19:03 | <AryehGregor> | (I'm not sure why not.) |
| 19:04 | <daedb> | oh, I thought I had read something a while back about that, but I must be misremembering then :) |
| 19:05 | <AryehGregor> | I could make it configurable, I suppose. |
| 19:05 | <AryehGregor> | The WHATWG wiki is already running off a development version, I could upgrade to a more recent one. |
| 19:06 | <AryehGregor> | (hopefully header processing hasn't changed again, or hsivonen will get annoyed again) |
| 19:09 | <othermaciej> | franksalim: it just seems like the opposite of the layering that makes sense |
| 19:12 | <Hixie> | back briefly, sorry for running off |
| 19:13 | <Hixie> | AryehGregor: is there something we can do to make it possibel for me to embed images from around the web on http://wiki.whatwg.org/wiki/Timed_tracks ? |
| 19:13 | <AryehGregor> | Hixie, I'm writing a patch right now to just allow whitelisting <img>. That seems simplest. |
| 19:13 | <AryehGregor> | When I've committed that, I can update to latest trunk, fix any fallout, and then whitelist <img>. |
| 19:13 | <AryehGregor> | If someone reverts my commit, well, I'll deal with that. :) |
| 19:14 | <Hixie> | sweet |
| 19:14 | <Hixie> | so in the meantime i just stick <img src="" alt=""> in the markup? |
| 19:17 | <AryehGregor> | I guess so. |
| 19:17 | <Hixie> | cool, thanks man, you rock |
| 19:19 | <AryehGregor> | This is a pretty horrendous interface. I can imagine there might be objections to exposing it as an API we have to maintain. |
| 19:19 | <AryehGregor> | Hmm. |
| 19:20 | <AryehGregor> | Maybe a more targeted patch would be a better idea. |
| 19:38 | <AryehGregor> | Okay, feature added: http://www.mediawiki.org/wiki/Special:Code/MediaWiki/65286 |
| 19:38 | <AryehGregor> | Now let's see if updating breaks anything exciting. |
| 19:42 | <AryehGregor> | Hixie, seems to be working. |
| 19:43 | <AryehGregor> | (I'm not sure why you don't upload those, though. They look like they're likely to bitrot as-is.) |
| 19:57 | AryehGregor | ponders configurably whitelisting some new-to-HTML5 tags too |
| 20:39 | <Hixie> | AryehGregor: thanks |
| 20:39 | <Hixie> | AryehGregor: not uploading them because the lifetime of this page is going to be about 3 days |
| 21:03 | <Hixie> | well |
| 21:03 | <Hixie> | so far i haven't found any subtitles that do anything fancy at _all_ |
| 21:03 | <Hixie> | not even italics! |
| 21:03 | <Hixie> | i must look harder |
| 21:07 | <AryehGregor> | I've noticed that in computer games, subtitles also rarely do anything fancy. Most games don't even use italics or bold. |
| 21:07 | <AryehGregor> | Some actually even resort to *asterisks* for emphasis. |
| 21:08 | <Hixie> | heh |
| 21:08 | <Hixie> | amusingly irssi made your asterisks into bold :-P |
| 21:09 | <Philip`> | http://philip.html5.org/misc/portal-caption.jpg has colours and italics |
| 21:09 | <Philip`> | but only per-line styles, I think |
| 21:10 | AryehGregor | almost said "If subtitles are intended for blind people, it's not surprising", but remembered just in time that that makes absolutely no sense |
| 21:11 | <AryehGregor> | Philip`, it styles them according to speaker, I think, no? Or something like that. The pink italics is GLaDOS. |
| 21:11 | <AryehGregor> | Interesting approach. |
| 21:11 | <AryehGregor> | That way different speakers have distinct appearances, without having to tell you their name. |
| 21:12 | <Philip`> | AryehGregor: Well, there's only one speaking character in the game |
| 21:12 | <Philip`> | so it doesn't make much difference |
| 21:12 | <AryehGregor> | Not true! |
| 21:12 | <AryehGregor> | The robots speak. |
| 21:12 | <AryehGregor> | "I'm sorry." |
| 21:12 | <Philip`> | Oh, I suppose that's true |
| 21:12 | <AryehGregor> | . . . I guess that's all, though. |
| 21:13 | <Philip`> | They're more like sound effects than real speech |
| 21:13 | <AryehGregor> | How so? |
| 21:14 | <Philip`> | You could replace them with something like a whirring effect and it would make no real difference |
| 21:14 | daedb | has at least one file with subtitles that is in two languages (simultaneously), use different colours, and has at least one part with a different font... |
| 21:15 | <Philip`> | http://philip.html5.org/misc/subtitles-3.xml uses colours to differentiate speakers within a single line occasionally |
| 21:23 | <KaOSoFt> | Good morning. |
| 21:25 | <KaOSoFt> | Did you people know of this? http://apirocks.com/html5/html5.html |
| 21:28 | <AryehGregor> | KaOSoFt, to begin with, slide 3 is grossly wrong. |
| 21:28 | <AryehGregor> | HTML5 started in 2004, for practical purpose.s |
| 21:28 | <AryehGregor> | s. |
| 21:29 | <AryehGregor> | "Tableless Web Design" and "AJAX" also don't belong on the same list as actual standards. |
| 21:29 | <AryehGregor> | Some form of HTML predated 1991, too, AFAIK. |
| 21:30 | <AryehGregor> | I'm pretty sure not all the relations listed in slide 18 are new to HTML5, although I could be wrong. |
| 21:30 | <AryehGregor> | Slide 21 lists some things as unimplemented that are implemented, at least by Opera. |
| 21:31 | <AryehGregor> | The CSS stuff has nothing to do with HTML5, even if you construe it pretty broadly, unless by "HTML5" you mean "any vaguely recent development in web technology". |
| 21:31 | <AryehGregor> | That's just at a glance. |
| 21:35 | <jgraham> | (if this is the page I think it is, the fineprint somewhere says that it is only really about Chrome. Which is a pretty crappy thing to bury in the fineprint of a presentation nominally about a cross-vendor standard with multiple implementations) |
| 21:39 | <jgraham> | (where "only really about" means "the demos - which are basically the whole point of the presentation - have only been written to work in Chrome even when trivial modifications would allow them o work in more browsers) |
| 21:39 | <jgraham> | s/)/")/ |
| 21:44 | <TabAtkins> | Slide 19 uses @item - this was changed to @itemscope after the usability testing. |
| 21:57 | <AryehGregor> | Thought that looked vaguely wrong. |
| 22:06 | <Hixie> | AryehGregor: I tried to put "<style scoped>img { max-width: 350px; }</style>" into a wiki page, but that didn't work... is there some way to set styles on a per-page basis? |
| 22:06 | <Hixie> | Philip`: thanks for that image, that's great |
| 22:12 | <daedb> | Hixie: Use the class that exists on the body element for that page? |
| 22:12 | <Hixie> | how do i do that? |
| 22:13 | <daedb> | In the global stylesheet, put in .page-pagename img { etc } |
| 22:14 | <Hixie> | sorry i'm a complete wikimedia n00b |
| 22:14 | <Hixie> | oh i can edit the global css? |
| 22:14 | <daedb> | http://wiki.whatwg.org/wiki/MediaWiki:Monobook.css |
| 22:14 | <daedb> | You can add styles there ^^ |
| 22:14 | <Hixie> | nice |
| 22:14 | <daedb> | And every page has the pagename in a class on the body ;) |
| 22:15 | <Hixie> | yeah, just found it |
| 22:15 | <Hixie> | you rock, thanks |
| 22:15 | <Hixie> | sweet, it worked |
| 22:16 | <daedb> | Every user account has its own css file in Mediawiki too, which is really fun to play with :) |
| 22:17 | <Hixie> | cool |
| 22:19 | <TabAtkins> | +1 for <video>! Switching youtube to <video> makes the audio actually work, instead of sounding like Max Headroom using the flash player. |
| 22:19 | <TabAtkins> | (on my goobuntu box) |
| 22:19 | <daedb> | http://wiki.whatwg.org/wiki/MediaWiki:Common.css is the really really global one btw. Monobook.css is for the monobook skin only. |
| 22:21 | <Hixie> | daedb: k |
| 22:55 | <Hixie> | i've put a bunch of examples in http://wiki.whatwg.org/wiki/Timed_tracks |
| 22:57 | <Hixie> | i've yet to find a single example that needs anything more than plain text and line feeds, the ability to have multiple distinct voices, some vague positioning, karaoke, and ruby |
| 23:00 | <TabAtkins> | It appears that removing most of the styling decisions from people would be a great improvement, if these examples are at all representative (and I'm betting they are...). |
| 23:00 | Philip` | notes that the Portal one isn't really timed tracks, since it's interactive rather than playing back pre-prepared sequences |
| 23:08 | <Philip`> | Hixie: http://philip.html5.org/misc/eva-captions.jpg is one with seemingly more important positioning requirements |
| 23:09 | <Hixie> | Philip`: wow, that's quite the trainwreck :-) |
| 23:11 | <Philip`> | Hixie: Why so? :-) |
| 23:12 | <Philip`> | It's just translating the text that's on the video, in a layout that roughly matches the video, plus the audio subtitles |
| 23:12 | <Hixie> | multiple fonts, three different kinds of text alignment, including one block of text with two different alignments... i'd have to see the video to tell how bad it really was in context |
| 23:13 | <Hixie> | but i've added it to the pile and changed the requirements section to match :-) |
| 23:14 | <TabAtkins> | Hixie: All but the bottom text is meant to more-or-less copy the alignment of the hand-written text on the picture, since it's a translation of that. |
| 23:14 | <Hixie> | oh i see what it's doing, but that doesn't make it any less of a trainwreck :-) |
| 23:15 | <Hixie> | a lot of these are pretty bad |
| 23:15 | <Hixie> | like the green subtitles in one of the examples |
| 23:15 | <Hixie> | ew |
| 23:15 | <TabAtkins> | Yeah, that's horrifying. |
| 23:15 | <TabAtkins> | I think it's outlined in red, too? |
| 23:15 | <Hixie> | or the second one in the "ruby" examples, though that one might have been ok before compression artefacts had their go at it |
| 23:15 | <Philip`> | Hixie: The video is http://www.youtube.com/watch?v=xG9KluukpJI at 2:43 |
| 23:15 | <Philip`> | but that version has totally different text |
| 23:16 | <Hixie> | heh |
| 23:16 | <Hixie> | ok the start of that video is pretty insane too |
| 23:17 | <Philip`> | (The version I've got has the English translation of on-screen text baked into the video, plus an English subtitle track which translates a Japanese audio track, plus an English audio track which uses a different translation) |
| 23:17 | <Hixie> | between the subtitles, the subtitle credit, the main credit, and the youtube ad, i feel assaulted |
| 23:18 | <Hixie> | ooh, that video's 2:43 is interesting for another reason |
| 23:18 | <Hixie> | it adds another subtitle above the first one |
| 23:18 | <Hixie> | i wonder if that's an example of multiple voices or of positioning or something else |
| 23:18 | <Philip`> | (...using DivX video and Vorbis audio in an OGM container) |
| 23:20 | <Hixie> | actually i guess that should happen automatically just by having two subtitles with overlapping segments |
