| 00:29 | <dglazkov> | Hixie: off the top of your head, when did you add transparent content model? I can't seem to find a good way to search commits on the spec |
| 00:36 | <jamesr_> | TabAtkins_, damn you sinped me |
| 00:36 | <jamesr_> | sniped |
| 00:59 | <TabAtkins_> | bwahaha |
| 00:59 | <TabAtkins_> | And I did that *while* I was making a new diagram in SVG. |
| 01:10 | <MikeSmith> | dglazkov: I think more than 5 years ago at least |
| 01:10 | <MikeSmith> | (about the transparent content model) |
| 01:11 | <MikeSmith> | TabAtkins_: yeah, I got it, but it's going to take me a while to add all those |
| 01:11 | <MikeSmith> | will try to do it today |
| 01:12 | <MikeSmith> | I've been working on validator code |
| 01:13 | <MikeSmith> | which work I hope I can finish up on today |
| 01:13 | MikeSmith | sees question from Ms2ger |
| 01:14 | <TabAtkins_> | MikeSmith: Darn, I was hoping adding a new component was super-easy or something. Thanks, though! |
| 01:16 | <MikeSmith> | it is super easy, but that list is like 20 new components or something, so I need to set aside a few minutes rather than just say, doing it during some telcon the way I usually would |
| 01:16 | <MikeSmith> | TabAtkins_: also ideally each of those components should have a default assignee |
| 01:16 | <MikeSmith> | typically, the editor |
| 01:17 | <MikeSmith> | do you want me to try to set default assignees for them? |
| 01:18 | <MikeSmith> | we need an inform bot here? |
| 01:19 | <MikeSmith> | infobot, inform Ms2ger no David Dahl did not ask me about adding a new repo |
| 01:19 | <MikeSmith> | I assume that would be for domcrypt |
| 01:40 | <TabAtkins_> | MikeSmith: I can send you a list of all the default assignees. Does it need to be just one person, or can it be multiple? |
| 01:43 | <MikeSmith> | dunno if it can multiple |
| 01:43 | <MikeSmith> | I think in practice we always make it one person |
| 01:43 | <MikeSmith> | you can see the current default assignees here: |
| 01:43 | <MikeSmith> | https://www.w3.org/Bugs/Public/describecomponents.cgi?product=CSS |
| 01:46 | <TabAtkins_> | Okay, I'll draw up a list tomorrow. |
| 01:47 | <hober> | You can have a list of people who get CCed by default. that's what we have for transforms, transitions, & animations |
| 05:30 | <jamesr> | fffuuuu, is there any way for one data: URL to script an iframe that's loaded as another data: URL ? |
| 05:37 | <jamesr> | woah, anyone want to see some weird shit? |
| 05:38 | <Hixie> | jamesr: webkit has a bug that blocks data: urls from interacting |
| 05:39 | <jamesr> | yeah they appear to always be different origin |
| 05:39 | <jamesr> | sigh |
| 05:39 | <jamesr> | check this shit out: http://webstuff.nfshost.com/tests/outer.html |
| 06:07 | <jamesr> | anyone around who can fix MDN ? |
| 08:10 | <hsivonen_> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=15592 Why hardware tokens are a bad idea, episode 42. |
| 08:11 | <hsivonen_> | Even if the Web platform had a smart card API, how would you stick a smart card into your wifi-only tablet? |
| 08:11 | <hsivonen_> | we know that people just love to buy and lug around USB dongles for their tablets |
| 09:04 | <MikeSmith> | hsivonen_: I guess it's conceivable that tablets could have NFC readers |
| 09:05 | <MikeSmith> | or does "smart card" imply some specific type of reader that requires actually swiping the card? |
| 09:08 | jgraham | slips up on the multiple-meanings-of-NFC banana skin again |
| 09:09 | <MikeSmith> | heh |
| 09:09 | <MikeSmith> | or RFID or whatever |
| 09:10 | <MikeSmith> | it seems some Dell laptops already have such a thing |
| 09:13 | <MikeSmith> | "contactless" cards I guess |
| 09:14 | <Ms2ger> | MikeSmith, yeah, for the crypto stuff |
| 09:14 | <MikeSmith> | OK |
| 09:15 | <MikeSmith> | Ms2ger: I pinged David on #developers but I guess is has been away |
| 09:15 | <MikeSmith> | if you want, I can just go ahead and set it up |
| 09:15 | <MikeSmith> | just "domcrypt" for the name? |
| 09:16 | <MikeSmith> | and what group should I associate it with, if any? |
| 09:16 | <MikeSmith> | for now I can just set it so that david's the only one with write access |
| 09:16 | <MikeSmith> | can add a group or other editors later if needed |
| 09:17 | <MikeSmith> | are contactless cards used in Europe or North America much? |
| 09:18 | <MikeSmith> | in Japan we use them for trains and buses and also at many stores |
| 09:20 | <jgraham> | Assuming they are the same thing, they are used for the public transport in London and Linköping (amongst others!). Never seen them used in a store though. |
| 09:20 | <Ms2ger> | MikeSmith, http://www.w3.org/2011/11/webcryptography-charter.html I guess, dunno what the status is |
| 09:21 | <MikeSmith> | Ms2ger, thanks |
| 09:21 | <MikeSmith> | ah yeah |
| 09:21 | <MikeSmith> | I need to ask Harry about that group |
| 09:21 | <MikeSmith> | anyway, I can just set it for now so that David's the only one with write access |
| 09:22 | <MikeSmith> | jgraham: do have electronic ID cards and if so are those contactless? |
| 09:22 | <Ms2ger> | Wan-Teh Chang as well, I think |
| 09:23 | <MikeSmith> | OK |
| 09:23 | <jgraham> | MikeSmith: I have some sort of ID card that apparently has some sort of electronic part. Not contactless though and I have never needed to use it electronically. |
| 09:23 | <jgraham> | Indeed I don't know how. |
| 09:23 | <MikeSmith> | ok |
| 09:28 | <MikeSmith> | Ms2ger: ready to go https://dvcs.w3.org/hg/domcrypt/ |
| 09:29 | <Ms2ger> | Great, thanks |
| 10:27 | <hsivonen_> | MikeSmith: so is Apple/Samsung/whatever willing to put the extra hardware to cater to European countries that got enchanted by the smart card lobby? |
| 10:28 | <MikeSmith> | hsivonen: dunno |
| 10:28 | <MikeSmith> | b |
| 10:28 | <MikeSmith> | but if there is money to be made from it in selling more devices, I'd think so |
| 10:30 | <hsivonen> | MikeSmith: back when Finland tried this, we got a couple of crappy computers from Fujitsu with built-in support. that was it. Fortunately, the whole thing flopped. |
| 10:31 | <MikeSmith> | yeah, I wonder how the current trend will play out in the long run |
| 10:34 | <hsivonen> | also, I don't really want any special interest to replay the digital tv coup, where they made a lot of people replace hardware just because |
| 10:34 | <MikeSmith> | yeah, that would be bad |
| 10:34 | <MikeSmith> | but in this case, people seem to like the convenience |
| 10:34 | <hsivonen> | hahahahah |
| 10:34 | <hsivonen> | this smart card stuff has nothing to do with user convenience |
| 10:34 | <MikeSmith> | oh |
| 10:35 | <hsivonen> | it's all about vendors convincing governments that there's a problem that needs solving |
| 10:35 | <MikeSmith> | ah |
| 10:35 | <hsivonen> | when that same problem has never been equally solved pre-Internet and the sky hasn't fallen |
| 10:36 | <hsivonen> | MikeSmith: sorry if "hahahahah" was rude. |
| 10:36 | <hsivonen> | but this is so far removed from the needs of the *users* |
| 10:36 | <MikeSmith> | I guess I don't really know what the difference is between smart cards and the contactless cards and e-wallet features on mobile phones we have in Japan |
| 10:36 | <MikeSmith> | no problem man :) |
| 10:37 | <MikeSmith> | you should laugh more often! |
| 10:37 | <MikeSmith> | I'm glad to provide some mirth |
| 10:38 | <MikeSmith> | anyway, the use case I had in mind is just the use case of being able to pay for stuff easily without needing to carry cash |
| 10:39 | <MikeSmith> | I take it that in Europe, the use case that the vendors are trying to sell governments on is just electronic identification |
| 10:39 | <hsivonen> | what goverments in Europe try to do is to make login to government Web sites more "secure" than login to Google or Paypal |
| 10:40 | <hsivonen> | and more secure than login to banks |
| 10:40 | <MikeSmith> | aha |
| 10:41 | <jgraham> | Hmm, the situation we have is that login to government websites is pretty much equivalent to login to banks |
| 10:41 | <hsivonen> | because, you know, no one has never just phoned a government office and stated their name and id number |
| 10:41 | <hsivonen> | jgraham: in Finland, login to goverment sites is now outsourced to banks, because the smart card thing sucked so badly |
| 10:41 | <jgraham> | Right, taht is mostly the case here |
| 10:42 | <jgraham> | There are various identity providers and banks are one of them |
| 10:42 | <jgraham> | Banks in turn are typically protected by some kind of hardware challenge/response system |
| 10:43 | <hsivonen> | this situation is *much* better than smart cards, because because typing some numbers from a laminated piece of paper is browser, OS and device-neutral |
| 10:43 | <jgraham> | (not sure if I am using the right technical terms here) |
| 10:43 | <hsivonen> | jgraham: over here, the "hardware" that the user possesses is a piece of paper with some plastic coating |
| 10:43 | <hsivonen> | jgraham: works great except with gullible people and phishing |
| 10:43 | <jgraham> | (but the system is device neutral to the extent that the bank website works on different devices/browsers) |
| 10:44 | <jgraham> | (and one option for government login requires java. There is also a mobile phone based alternative) |
| 10:44 | <hsivonen> | but then, smart cards are also vulnerable to gullible users |
| 10:45 | <hsivonen> | in fact, using smart cards requires more leaps of faith from the user |
| 10:46 | <jgraham> | hsivonen: Here you get some hardware for which you have a PIN. The bank site presents a number, you enter that number into the device and then enter the corresponding number into the bank website. |
| 10:47 | <hsivonen> | jgraham: I guess that protects gullible people from blatant phishing better than what we have |
| 10:47 | <hsivonen> | jgraham: but probably doesn't help people who are excessively gullible |
| 10:47 | <hsivonen> | compared to the piece of paper that is |
| 10:48 | <hsivonen> | with the piece of paper, the bank is supposed to ask for one printed number at a time |
| 10:49 | <hsivonen> | but since the user has all the numbers available, a particularly audacious phisher can ask the user to type *all* of them at once |
| 10:49 | <hsivonen> | of course, the user has to be super-gullible to fall for that one |
| 10:49 | <hsivonen> | I don't know if people who fall for that one can be helped in terms of computer security |
| 10:58 | <MikeSmith> | hsivonen: I'm wondering how the validator deals with "xmlns:" in text/html |
| 10:59 | <MikeSmith> | I see the parser does not expose "xmlns:*" as attributes to the rest of the SAX pipeline |
| 10:59 | <MikeSmith> | as far as I can see |
| 11:00 | <MikeSmith> | and doesn't support the sax namespace-prefixes feature |
| 11:01 | <MikeSmith> | so I'm wondering how the validator service is able to see report xmlnis:* attributes |
| 11:03 | <hsivonen> | MikeSmith: IIRC, the parser reports them, then doesn't show them to the pipeline |
| 11:04 | <MikeSmith> | ah, OK |
| 11:05 | <MikeSmith> | hmm, but jing sees them |
| 11:05 | <hsivonen> | oh. |
| 11:05 | <hsivonen> | then I misremember |
| 11:06 | <MikeSmith> | well, I mean I can see that the error message seems to be a kind that jing emits |
| 11:06 | <MikeSmith> | "Attribute xmlns:foo not allowed here" |
| 11:07 | <MikeSmith> | oh |
| 11:07 | <MikeSmith> | no |
| 11:07 | <MikeSmith> | that is emitted by the parser |
| 11:07 | <MikeSmith> | my bad |
| 11:08 | <MikeSmith> | that's in the treebuilder code |
| 11:08 | <hsivonen> | MikeSmith: yes. private void checkAttributes(HtmlAttributes attributes, @NsUri String ns) |
| 11:08 | <MikeSmith> | yep yep |
| 11:08 | <MikeSmith> | OK, that is good news |
| 11:08 | <MikeSmith> | or good for me to get confirmation of that |
| 11:08 | <hsivonen> | MikeSmith: fun with RDFa, eh? |
| 11:08 | <MikeSmith> | you guessed it man |
| 11:09 | <MikeSmith> | xmlns is deprecated in RDFa 1.1 |
| 11:09 | <hsivonen> | MikeSmith: I suggest not supporting deprecated parts |
| 11:09 | <MikeSmith> | however, the spec currently says, "Conformance checkers must accept attribute names that have a case insensitive prefix matching "xmlns:" as conforming. Conformance checkers should generate warnings noting that the use of xmlns: is deprecated." |
| 11:09 | <MikeSmith> | yeah |
| 11:09 | <MikeSmith> | agreed |
| 11:12 | <MikeSmith> | if anything that part of the spec should be changed to say, "Conformance checkers may report xmlns attributes as errors" |
| 11:12 | <MikeSmith> | anyway, it's also not clear whether xmlns is an error in RDFa Lite or not |
| 11:12 | <MikeSmith> | not clear to me anyway |
| 11:21 | <MikeSmith> | oh, looks like gecko implementation of Network Status API just landed |
| 11:21 | <MikeSmith> | https://bugzilla.mozilla.org/show_bug.cgi?id=677166 |
| 11:23 | <Ms2ger> | And some SMS stuff too |
| 11:23 | <MikeSmith> | navigator.mozConnection.bandwidth |
| 11:23 | <MikeSmith> | SMS stuff? |
| 11:27 | <Ms2ger> | https://bugzilla.mozilla.org/show_bug.cgi?id=674725 |
| 11:31 | hsivonen | wishes this stuff wasn't moz-prefixed |
| 11:32 | <hsivonen> | that SMS landing sure had lots of changesets |
| 11:32 | Ms2ger | wishes this stuff was being specced |
| 11:32 | <Ms2ger> | hsivonen, even better, it didn't build |
| 11:33 | <hsivonen> | Ms2ger: IIRC, there's a draft for the network status API |
| 11:33 | <hsivonen> | Ms2ger: is there not a spec for SMS? |
| 11:33 | <Ms2ger> | On wikimo |
| 11:34 | <MikeSmith> | the implementation is different from network status API draft |
| 11:34 | <MikeSmith> | for good reasons |
| 11:34 | <MikeSmith> | hopefully the spec will get revised |
| 11:34 | <Ms2ger> | AIUI, the plan is to try to establish a significant legacy through b2g, and then spec our implementation |
| 11:44 | <smaug____> | Ms2ger: is wikimo not a spec but draft in some random website which has html or html5 in the domain name is a draft? |
| 11:44 | <smaug____> | though, perhaps you did mean that wikimo draft is a draft spec |
| 11:51 | <Ms2ger> | No, what I meant is that wikimo is owned by one vendor, while html5.org or whatwg.org isn't |
| 11:56 | <hsivonen> | html5.org and whatwg.org are both owned by employees of vendors |
| 11:57 | <Ms2ger> | Sure |
| 11:57 | <Ms2ger> | But wikimo is a Mozilla asset, and whatwg.org isn't a Google asset |
| 11:57 | <hsivonen> | ok |
| 12:13 | <MikeSmith> | hsivonen: sent you some patch review requests |
| 12:15 | <hsivonen> | MikeSmith: I'll take a look |
| 12:15 | <MikeSmith> | thanks |
| 12:15 | <MikeSmith> | sorry it's all pretty boring stuff |
| 13:57 | <MikeSmith> | hsivonen: resent the review request but with the actual patch attached this time |
| 13:57 | <MikeSmith> | sorry about that |
| 14:07 | <hsivonen> | MikeSmith: I replied |
| 14:07 | <MikeSmith> | thanks |
| 15:18 | <AryehGregor> | hober, are the transform specs being actively edited? I filed a bunch of bugs, some of them very simple, and while some have gotten responses, none have resulted in changes yet. |
| 15:21 | <AryehGregor> | TabAtkins_ or jamesr_ or some other Chrome person: could you run the test-case in this comment and see if it actually works for you in Chrome? It still produces incorrect results for me in dev, but someone claims it works in canary: https://bugs.webkit.org/show_bug.cgi?id=68551#c0 |
| 15:22 | <AryehGregor> | (I'm on Linux, so I don't have canary or WebKit nightlies to test in) |
| 15:22 | <AryehGregor> | (but the patches landed over a month ago, they should be in dev, right?) |
| 15:26 | <beverloo> | AryehGregor, I don't know about the context, but I'm getting an empty alert in a build from 2 minutes ago |
| 15:26 | <AryehGregor> | beverloo, thanks, that's the correct (fixed) behavior. |
| 15:26 | <AryehGregor> | I guess dev just hasn't gotten the updates yet or something. |
| 15:27 | <beverloo> | I think this is the change which is relevant? http://trac.webkit.org/changeset/103953 (cssText change for border-width) |
| 15:28 | <beverloo> | the fix probably isn't on the branch in that case |
| 15:31 | <AryehGregor> | Lame, Wikipedia blackout implemented by JS that runs onload or something. |
| 15:31 | <AryehGregor> | Why couldn't they reconfigure the Squids instead? |
| 15:31 | <AryehGregor> | beverloo, but it will migrate there soon enough, I guess? |
| 15:32 | <beverloo> | yeah, should be in Chrome 18 |
| 15:32 | <AryehGregor> | Thanks. |
| 15:34 | <jgraham> | AryehGregor: It seems to have been a design goal to make it easy to get around. Dunno why |
| 15:34 | <AryehGregor> | Laziness? |
| 15:35 | <zewt> | because actually taking sites offline in "protest" is stupid? heh |
| 15:35 | <jgraham> | I doubt it. I mean it seems almost easier to do it right |
| 15:36 | <zewt> | if a site i need is offline because of some "protest", there's zero chance i'm going to go "damn you SOPA" and 100% chance i'm going to get angry at the people who *actually* took the site offline |
| 15:39 | <wilhelm> | You appear to be special, in that case. Following the discussions in local media, ~everyone that comment on the issue seem to be positive to the action. The issue has been widely covered today, thanks to the blackout. |
| 15:40 | <wilhelm> | I have no numbers – just a few more anecdotes. (c: |
| 15:41 | <jgraham> | The coverage on the BBC seems to be quite anti-blackout |
| 15:41 | <wilhelm> | Interesting. |
| 15:42 | <jgraham> | Not people's opinions but the reporters |
| 15:42 | <jgraham> | I mean it isn't saying "OMG this is bad" or anything |
| 15:43 | <jgraham> | But it is written in such a way that it sounds like the protest is in favour of piracy |
| 15:43 | <jgraham> | That could just be my bias reading it of course |
| 15:43 | <jgraham> | It doesn't help that Jimmy Wales rambled in his interview whereas the MPAA goon was well-drilled and had clear soundbites |
| 15:44 | <jgraham> | http://www.bbc.co.uk/news/technology-16604990 |
| 15:45 | <AryehGregor> | jgraham, doing it right means involving sysadmins, in Wikipedia's case, to reconfigure the Squids (or Varnish or whatever they use these days). Doing it wrong means any old sysop adding something to [[MediaWiki:Common.js]]. |
| 15:46 | <AryehGregor> | Where "sysadmin" means "paid ops person with root access, of which there are like five", while "sysop" means "any 16-year-old who spent a few hundred hours scaring away well-intentioned newcomers by posting scary anti-vandalism warnings on their talk pages". |
| 15:46 | <jgraham> | OTOH Richard Murdoch came out in favour of SOPA which might just be enough to get everyone to realise it is a bad idea |
| 15:46 | <AryehGregor> | (of which there are like 2,000) |
| 15:47 | <jgraham> | (since he is universially reviled, at least in the UK) |
| 15:47 | <wilhelm> | The biggest radio station here attempted to find someone who was in favour of SOPA for a debate. They couldn't find anyone. They did ask the local equivalent of Encyclopædia Britannica, who were offended by the question. The result was that the encyclopedia considered shutting down too, but ended up doing what no.wikipedia.org did – put up a big, black banner. |
| 15:47 | <jgraham> | heh |
| 15:47 | <Ms2ger> | Nice |
| 15:48 | <AryehGregor> | jgraham, because the UK is entirely liberal, by American standards, like basically all of Europe on most issues. American conservatives don't revile him at all. |
| 15:48 | <AryehGregor> | American liberals do, though. |
| 15:48 | <jgraham> | AryehGregor: American conservatives probably don't read the BBC though |
| 15:48 | <AryehGregor> | No, it's a liberal rag as far as they're concerned. |
| 15:48 | <AryehGregor> | (I'm an American conservative, incidentally) |
| 15:49 | <Ms2ger> | Well, everything except for Fox is a liberal rag, no? |
| 15:49 | <AryehGregor> | More or less, yes. Also the Wall Street Journal is probably okay. |
| 15:49 | AryehGregor | discovers that the 4x4 identity matrix with the last 1 equal to k is the same as scale3d(1/k, 1/k, 1/k) |
| 15:49 | <AryehGregor> | This perspective stuff is kind of interesting. |
| 15:50 | <AryehGregor> | Also, setting the last 1 to a 0 makes everything completely vanish. |
| 15:50 | <Ms2ger> | The last one equal to k? |
| 15:50 | <Ms2ger> | diag(1,1,1,k)? |
| 15:50 | <AryehGregor> | [1 0 0 0; 0 1 0 0; 0 0 1 0; 0 0 0 k] |
| 15:50 | <AryehGregor> | Yeah. |
| 15:50 | Ms2ger | is lazy |
| 15:50 | <AryehGregor> | Which implies that you can always have that entry be either 1 or 0 by adjusting the other entries appropriately. |
| 15:50 | <AryehGregor> | I guess. |
| 15:50 | <wilhelm> | American TV is wonderfully bizarre. Even more so than British tabloids. |
| 15:51 | <jgraham> | (BTW thw Mozilla SOPA page is very good) |
| 15:52 | <AryehGregor> | Actually, if what Philip` said is correct, then diag(1, 1, 1, 0) plasters everything out to a sphere at infinity, squishing every ray coming from the origin to its endpoint at infinity. |
| 15:52 | <AryehGregor> | Philip`, so what's the graphical significance of [0, 0, 0, 0]? |
| 15:52 | <Ms2ger> | jgraham, duh? ;) |
| 15:52 | <jgraham> | (not least because it explicitly says "we think piracy is bad, but still think this legislation is bad") |
| 15:53 | <Ms2ger> | We've got complaints already of people switching to other browsers because we support piracy |
| 15:53 | <wilhelm> | Fascinating |
| 15:55 | <jgraham> | Ms2ger: Haha |
| 15:56 | <Ms2ger> | The fact that we still have users who use that treat must mean they're just as unhappy with other browsers |
| 15:56 | <Ms2ger> | If they still end up with us.. |
| 15:59 | <Philip`> | jgraham: Richard Murdoch, the deceased comedy actor? |
| 15:59 | <bga> | lol http://stopsoap.com/ |
| 15:59 | <jgraham> | Philip`: Did I say that? |
| 16:00 | <Philip`> | 15:48 < jgraham> OTOH Richard Murdoch came out in favour of SOPA which might just be enough to get everyone to realise it is a bad idea |
| 16:00 | <jgraham> | See, I hate him so much I can't even bear to type his name |
| 16:00 | <jgraham> | Either that or I am an idiot |
| 16:00 | <Philip`> | That makes it hard to tell which of the Murdochs you hate :-( |
| 16:00 | <Philip`> | or maybe there's no need to make that distinction |
| 16:01 | <Philip`> | AryehGregor: It's a light source at infinity with intensity 0 (so its direction is irrelevant) |
| 16:02 | <AryehGregor> | Philip`, how does it have intensity zero? |
| 16:11 | <Philip`> | I think (not based on much expertise) that w=0 vectors occur most commonly in lighting equations, where you compute stuff like I * (N dot L) where N is a (unit) normal vector and L is the (usually unit) light direction and I is the light intensity |
| 16:12 | <Philip`> | and the only time you'd get a vector like [0, 0, 0, 0] is if you tried folding the intensity into L in advance (when I = 0) |
| 16:13 | <Philip`> | That's probably a rubbish answer; it's probably better to say that graphics almost always uses unit vectors, and non-unit vectors have no significance, and [0, 0, 0, 0] probably results from a bug like scaling your world by 0 |
| 16:18 | <AryehGregor> | Philip`, interesting. |
| 16:19 | Philip` | has only ever used 4-vectors in graphics when applying transforms to 3D models, since you compute vertex positions as [x, y, z, 1]*T and vertex normals as [x, y, z, 0]*T, and all the vectors were stored as 3-vectors anyway and the w was just implicit in the computation |
| 16:20 | <AryehGregor> | What do you mean by vertex normals? |
| 16:21 | <Philip`> | The normal vector of the surface at a vertex |
| 16:23 | <AryehGregor> | Which surface? |
| 16:25 | <Philip`> | The surface of the mesh that the code is rendering |
| 16:25 | <bga> | http://en.wikipedia.org/wiki/Surface_normal |
| 16:27 | AryehGregor | knows what a surface normal is, in the sense of something normal to the tangents, where the tangents are the derivatives of the parametrization, but isn't sure how [x, y, z, 0]*T is relevant |
| 16:27 | <Philip`> | (It's not really a mathematical surface, e.g. if you have a cube then there are three coincident vertexes at each corner with different normals (one per face of the cube)) |
| 16:27 | <AryehGregor> | Well, probably I don't need to know for what I'm doing right now. |
| 16:27 | <AryehGregor> | Oh, interesting. |
| 16:27 | <AryehGregor> | But each face is a proper surface with boundary, right? |
| 16:27 | <AryehGregor> | So it's really the union of six surfaces. |
| 16:29 | <Philip`> | T is a translation+rotation matrix representing the object-space-to-world-space transform of the model, and you need to apply T to the normals for use in subsequent lighting calculations, but the translation component of T is irrelevant so you treat the normals as being vectors [x,y,z,0] in the matrix multiplication |
| 16:29 | <Philip`> | (whereas translation matters for positions, so you treat them as [x,y,z,1]) |
| 16:30 | <Philip`> | (Actually there's probably some inverses and transposes in there but they cancel out so they're not important) |
| 16:30 | <AryehGregor> | Oh, I see. |
| 16:30 | <AryehGregor> | Interesting. |
| 16:30 | <AryehGregor> | Amazing what you can do with linear algebra. |
| 16:31 | <AryehGregor> | Of course, a large fraction of pure math is just devoted to reducing stuff to linear algebra. |
| 16:31 | <AryehGregor> | (see, e.g.: multivariable calculus, homology) |
| 16:31 | <AryehGregor> | Applied math even more so. |
| 16:32 | <Philip`> | (In code you just do stuff like http://trac.wildfiregames.com/browser/ps/trunk/binaries/data/mods/public/shaders/model_common.vp lines 14-20, where instancingTransform is the object-to-world-space transform matrix, and "DP3 a, b, c" means a = b.x*c.x + b.y*c.y + b.z*c.z, and "DP4 ..." is the 4-vector version of that) |
| 16:33 | <Philip`> | (although if you're sane you probably don't use assembly for shaders) |
| 16:37 | <AryehGregor> | Fun question: should transform: scaleZ(0) have any visible effect on rendering? |
| 16:37 | <AryehGregor> | In Gecko it seems to make everything vanish, which doesn't make a lot of sense to me.. |
| 16:37 | <AryehGregor> | me. |
| 16:37 | <TabAtkins_> | It shouldn't, no. |
| 16:37 | <TabAtkins_> | Because they're already zero in the z dimension? |
| 16:38 | <AryehGregor> | It's a funny case, because normally zero-determinant transforms make everything vanish, but here should be an exception. |
| 16:39 | <niloy> | I have a question regarding window.getComputedStyle(), if I ask for "border-radius", its returning empty string, is this okay? |
| 16:39 | <AryehGregor> | Or rather, anything should be an exception if it squashes everything to a plane that's not perpendicular to the x-y plane. |
| 16:39 | <TabAtkins_> | I'm wondering what effect scaleZ() can have at all, actually. |
| 16:39 | <TabAtkins_> | niloy: Code sample? Sounds like your'e doing it wrong. |
| 16:39 | <AryehGregor> | TabAtkins_, well, you can intersperse it with other transforms. perspective() is an obvious one, but also try rotate3d(). |
| 16:39 | <TabAtkins_> | AryehGregor: Ah, yeah. |
| 16:40 | <niloy> | TabAtkins_, http://jsfiddle.net/LEZA6/ |
| 16:42 | <TabAtkins_> | Oh, right. border-radius is a shorthand, so it doesn't have a computed style. |
| 16:42 | <TabAtkins_> | You need to ask for each individual corner. |
| 16:42 | <TabAtkins_> | (This is a dumb rule.) |
| 16:42 | <niloy> | ohkay |
| 16:42 | <niloy> | ^_^ |
| 16:43 | <TabAtkins_> | Also, any particular reason you're using .getPropertyValue('foo-bar') instead of just .fooBar? |
| 16:44 | <niloy> | no particular reason, I tried everything, nothing worked |
| 16:45 | <hober> | AryehGregor: yes, they're being actively edited. but none of the editors are working *solely* on the specs, so edits happen in a somewhat unpredictably sporadic/batchy way. each issue tracked in bugzilla will (if neccessary) eventually result in a spec change, which will be noted in bugzilla at bug-resolving time. |
| 16:45 | <AryehGregor> | hober, okay, just checking that someone will get around to it eventually. Do you have any ideas on what the timeframe might be like? Weeks, months, years? |
| 16:48 | <AryehGregor> | TabAtkins_, okay, now what should happen if you have two things overlaid on top of one another and then you do scaleZ(0)? Which color do you get? The topmost? |
| 16:48 | <TabAtkins_> | I... think? |
| 16:48 | <hober> | AryehGregor: I'll try to ensure things happen quickly but no, timeframe is probably somewhat dependent on each particular issue's complexity etc. |
| 16:49 | <AryehGregor> | hober, okay. |
| 16:49 | <AryehGregor> | TabAtkins_, so the behavior of scaleZ() should be dependent on orientation, then? Reflecting through the x-y plane doesn't commute with scaleZ()? |
| 16:50 | <TabAtkins_> | AryehGregor: I'm talking out my ass. |
| 16:50 | <TabAtkins_> | Possibly Gecko's behavior is correct, and normal elements have epsilon depth in the z axis. |
| 16:50 | <AryehGregor> | That means <div style="transform:matrix3d(1,0,0,0, 0,1,0,0, 0,0,-1,0, 0,0,0,1)"><div style="transform:scaleZ(0)"> is different from <div style="transform:matrix3d(1,0,0,0, 0,1,0,0, 0,0,-1,0, 0,0,0,1) transform:scaleZ(0)">. |
| 16:50 | <AryehGregor> | Er. |
| 16:50 | <AryehGregor> | That means <div style="transform:matrix3d(1,0,0,0, 0,1,0,0, 0,0,-1,0, 0,0,0,1)"><div style="transform:scaleZ(0)"> is different from <div style="transform:matrix3d(1,0,0,0, 0,1,0,0, 0,0,-1,0, 0,0,0,1) scaleZ(0)">. |
| 16:51 | <TabAtkins_> | I can't interpret matrices. |
| 16:51 | <AryehGregor> | That matrix reflects everything through the x-y plane. |
| 16:51 | <AryehGregor> | I.e., negates the z-coordinate. |
| 16:51 | <AryehGregor> | Actually, it's just scaleZ(-1). |
| 16:52 | <AryehGregor> | The thing is, implementations probably premultiply all the matrices here for nested elements. |
| 16:52 | <AryehGregor> | So it's not intuitive what element is "on top". |
| 16:52 | <AryehGregor> | Actually, maybe it's not a problem. |
| 16:52 | <AryehGregor> | Just follow the regular z-index. |
| 16:53 | <AryehGregor> | Or document order if z-index is the same. |
| 16:53 | <AryehGregor> | This is related to transform-style:preserve-3d, methinks. |
| 16:54 | <AryehGregor> | Really, the question is no different from translating one div on top of another within the X-Y plane. |
| 16:54 | <AryehGregor> | It amounts to the same transformation matrix in the end. |
| 16:54 | <AryehGregor> | So it's not a problem, I guess. |
| 16:54 | <TabAtkins_> | Glad you satisfied yourself. ^_^ |
| 16:55 | <jgraham> | We need a LaTeX mode for irc |
| 16:55 | <TabAtkins_> | Slightly difficult in ascii. |
| 16:56 | <jgraham> | Well the output would have to be non-ascii of course |
| 16:56 | <jgraham> | Unless you just read the input :) |
| 16:56 | <AryehGregor> | jgraham, so I can write a program that sends your IRC client into an infinite loop? |
| 16:56 | <AryehGregor> | Or just a LaTeX subset like Wikipedia? |
| 16:56 | <jgraham> | AryehGregor: Pretend I added the word subset |
| 16:56 | <jgraham> | Yeah |
| 16:57 | <jgraham> | It could have been worse, I might hace said MathML |
| 16:57 | <AryehGregor> | :) |
| 17:20 | <david_carlisle> | jgraham: you are being watched |
| 17:22 | <jgraham> | david_carlisle: You don't scare me :p You might pretend to hand-author MathML, but I know you really have a LaTexX subset to MathML convertor written in XSLT :p |
| 17:23 | <david_carlisle> | jgraham: well, since you ask.... |
| 17:23 | <david_carlisle> | jgraham: I do have an xml parser written in TeX so of course you can use mathml syntax in your tex files if you want save all that old \ and {} nonsense.... |
| 17:26 | <jgraham> | Do you ever write in languages other people like? Or are all your "shell scripts" in Cobol, just because? |
| 17:26 | <jgraham> | ;) |
| 17:27 | <david_carlisle> | jgraham: shell scripts are in bash and sed and perl but In my heart I know they should all be xproc really |
| 17:28 | <jgraham> | Heh |
| 17:58 | <dglazkov> | good morning, Whatwg! |
| 17:59 | <Ms2ger> | Good night, dglazkov |
| 17:59 | <dglazkov> | good night, Ms2ger! http://www.youtube.com/watch?v=w-7BT2CFYNU |
| 18:02 | <Ms2ger> | Oh, a giraffe... That makes sense, I guess |
| 18:02 | <dglazkov> | :D |
| 18:27 | <dglazkov> | (figurative) penny for your thoughts: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2012-January/034410.html |
| 18:30 | <Ms2ger> | Don't do that |
| 18:31 | <Ms2ger> | My penny please? :) |
| 18:37 | <dglazkov> | here: http://upload.wikimedia.org/wikipedia/commons/thumb/3/3f/1990-issue_US_Penny_obverse_2.jpg/220px-1990-issue_US_Penny_obverse_2.jpg |
| 18:39 | <Ms2ger> | Thanks |
| 18:41 | <Ms2ger> | Opera seems to run <script for="window" event="bar"> |
| 18:44 | <dglazkov> | Ms2ger: it's in the spec, right? |
| 18:44 | <Ms2ger> | Only if event == onload or onload() |
| 18:44 | <dglazkov> | ah |
| 18:44 | <dglazkov> | bar looks pretty close to onload |
| 18:45 | <Ms2ger> | Yeah, keys so close together and stuff |
| 18:45 | <dglazkov> | usually when at bar, you load on |
| 19:12 | <Ms2ger> | "[webmaster] Why I am deleting Firefox, in favor of Opera" "Because you are boycotting SOPA for one day, it tells me that you are in favor of piracy and copyright theft, and against policing same. Therefore I am permanently boycotting Mozilla, and all other organizations who are in favor of piracy and theft. Switching out browsers is easy, switching e-mail addresses is harder, but by the end of the week I will no longer be using any products or web |
| 19:12 | <Ms2ger> | sites assocaited with people who are for piracy." |
| 19:12 | <Ms2ger> | Anyone at Opera want to make a statement? :) |
| 19:14 | <divya> | hahaha |
| 19:32 | <Velmont> | Heh... Ohwell. |
| 21:33 | <dglazkov> | is dreamhost protesting SOPA too? Or is whatwg.org just down? |
| 21:42 | <jgraham> | Ms2ger: Now that I know being against piracy is such a winning strategy I will ensure that we add heavy armaments to Howcome's boat and send him off to maintain law on the high seas |
| 21:44 | <jgraham> | dglazkov: Never ascribe to political protest what can be adequetly explained by incompetence |
| 21:44 | <dglazkov> | :) |
| 22:57 | <Hixie> | dglazkov: neither as far as i can tell |
| 22:58 | <dglazkov> | Hixie: it's back up |
| 22:58 | <dglazkov> | it was a quantum fluctuation. Heisenberg applies. |
| 23:04 | <dglazkov> | Hixie: did you see that parser thread? What's the best way to do this? |
| 23:06 | <Hixie> | saw it, haven't studied it yet |
| 23:07 | <dglazkov> | ok. I am available for lunches, breakfasts, and beverages as well. |
| 23:08 | <dglazkov> | Hixie: also volunteer to do the dirty work (excludes murder and washing dishes). |
| 23:08 | <jamesr_> | dglazkov, afternoon tea? |
| 23:08 | <jamesr_> | dglazkov, bar mitzvahs? |
| 23:09 | <dglazkov> | you bet. My rendition of hava nagila will leave you breathless. |
| 23:10 | <dglazkov> | or at least speechless. |
