| 00:04 | <othermaciej> | I don't think any shipping version supports it yet |
| 00:07 | <dglazkov> | sayrer: no |
| 00:08 | <dglazkov> | there's work being done on it, but no concrete impl |
| 00:08 | <dglazkov> | it's a pretty tricky business for a multi-process arch, and jorlow will tell you all about it. |
| 00:09 | <jorlow> | it'll be in eventually |
| 00:09 | <jorlow> | it should be in the dev channel before too long |
| 00:09 | <sayrer> | does it break run to completion in chrome? |
| 00:10 | <sayrer> | nm, probably doesn't matter much |
| 01:04 | <roc> | hey I think I convinced Tom Lord of something |
| 01:05 | othermaciej | wonders whether to subscribe to www-fonts |
| 01:06 | <othermaciej> | hmm, from the archives it looks like a giant flamefest |
| 01:06 | <roc> | sure is |
| 01:21 | <sayrer> | roc, do you mean we'd have to include the entire Theora spec in HTML5? |
| 01:21 | <sayrer> | *snicker |
| 01:21 | <roc> | no |
| 01:22 | <roc> | maybe we would if there was no spec for it, but fortunately there is one |
| 01:23 | <sayrer> | also, this feedback about removing the WAVE requirement, that was removed, didn't really have anything to do with removing the WAVE requirement |
| 01:23 | <sayrer> | just making sure we're all on the same page |
| 01:23 | <roc> | if H.264 was the baseline codec then we might very well have a lot of work to do to figure out which H.264 profile should be the baseline (including undocumented muxing requirements that Apple requires) |
| 01:23 | <sayrer> | also, the baseline codec must meet the requirements for an optimal codec |
| 01:24 | <sayrer> | such as high volume serving and client vendor unanimity |
| 01:24 | <roc> | I wonder who the Wave-hating browser vendor is |
| 01:24 | <roc> | It's not us |
| 01:24 | <roc> | it's not Apple, for once |
| 01:24 | <othermaciej> | I don't think it's Apple |
| 01:24 | <roc> | Chrome doesn't support Wave currently, so I suspect Chrome |
| 01:24 | <sayrer> | the fact that we are guessing is a fucking joke |
| 01:24 | <roc> | yes, I agree |
| 01:24 | <Hixie> | can't argue there |
| 01:24 | <othermaciej> | so far we seem to support the broadest range of codecs |
| 01:25 | <sayrer> | good job! |
| 01:27 | <sayrer> | well, I wish the w3c guys would hurry up with the cvs access |
| 01:27 | <sayrer> | I have a @summary attribute that needs unbanning |
| 01:28 | <Hixie> | yeah, i guess with xhtml2 being end-of-lifed, there is a niche to be filled for specs that are based on idealism rather than pragmatism. :-) |
| 01:29 | <sayrer> | oh, I'll be sure to note that it doesn't work very well |
| 01:30 | <sayrer> | banning it... that would be idealistic |
| 01:30 | <sayrer> | one might even say naive |
| 01:30 | <Hixie> | wait, aren't you planning on removing all authoring conformance criteria? |
| 01:31 | <sayrer> | there are many kinds of authoring conformance criteria in the spec |
| 01:31 | <MikeSmith> | sayrer: I have put in a request for your cvs access and will let you know as soon as I get word from the systems team that it's been set up |
| 01:31 | <sayrer> | I plan on removing most of them |
| 01:31 | <sayrer> | MikeSmith: thanks |
| 01:31 | <sayrer> | didn't really mean to pester, just fed up with the list |
| 01:33 | <sayrer> | Hixie, to pick a less loaded example, I plan to unban @profile if you haven't already |
| 01:33 | <sayrer> | and insert text that says "this attribute doesn't do anything in practice" |
| 01:33 | <sayrer> | aiming to be accurate, but not prescriptive |
| 01:33 | <Hixie> | i would asssume that once you remove the authoring conformance critiera, profile="", summary="", axis="", and gjshewqt3="" would all get unbanned automatically |
| 01:34 | <sayrer> | for authors, that is |
| 01:34 | <sayrer> | Hixie: some things become defined, others are undefined |
| 01:34 | <MikeSmith> | sayrer: all I ask is that you add an additional subdir under /html5 . name it whatever you want (it's cvs, anyway, so we can change names later if we care too, and still preserve the revision history) |
| 01:34 | <sayrer> | yeah, no problem. don't want any confusion |
| 01:37 | <Lachy> | sayrer, are you going to check in and work on your own modified version of HTML5? |
| 01:37 | <Lachy> | in dev.w3.org ? |
| 01:37 | <sayrer> | I am going to check it in and work on it, I do not want to be the owner |
| 01:39 | <MikeSmith> | sayrer: who else do you expect the owner to be? |
| 01:40 | <sayrer> | it will be my proposal initially, obviously. but if I need to exercise control over it, it's probably a failure. |
| 01:45 | <sinuhe> | Would <code></code>, or <kbd></kbd>, or perhaps something else (samp?) be the most appropriate when marking up something like: To view a file type "cat /etc/passwd", ie <p>To view...type <code>cat /etc/passwd</code></p>? |
| 01:46 | <sinuhe> | s/ie/eg/ |
| 01:46 | <MikeSmith> | sayrer: so your plan is to check in the proposal, then leave it up to the chairs and the working group to decide what (if anything) to do with it? |
| 01:47 | <sayrer> | iterate on it a bit |
| 01:47 | <sayrer> | and the end state should be what you ended with |
| 01:47 | <MikeSmith> | I see |
| 01:47 | <sayrer> | cutting is hard to do that way, and I want to cut some stuff |
| 01:47 | <Hixie> | sinuhe: <kbd> |
| 01:47 | <sayrer> | not forever, just for now |
| 01:47 | <sinuhe> | Hixie: Thank you. |
| 01:48 | <T--> | <kbd class="shell"><kbd class="command">cat</kbd> <kbd class="parameter tabbed">/etc/password</kbd></kbd> ;-) |
| 01:48 | <sinuhe> | Nice |
| 01:48 | <sinuhe> | Though would not /etc/passwd be better suited to <code>? |
| 01:49 | <T--> | a user would type or tab-type it? |
| 01:51 | <T--> | Oh. To realise something like the insane tab-idea, HTML 5 has some weird kbd-samp-combination rules. |
| 01:52 | <sinuhe> | Type, but, I was thinking in terms of the phrase "a filename" in the first sentence of the code element description. |
| 01:52 | <Wolfman2000> | ...did I hear things right? There is no more <video> or <audio>? |
| 01:52 | <othermaciej> | Wolfman2000: you didn't hear right |
| 01:52 | <T--> | Sounds like a var? |
| 01:52 | <othermaciej> | <video> and <audio> are still in the spec |
| 01:52 | <othermaciej> | the section saying that a baseline codec was desired, but not stating what it is, has been removed |
| 01:53 | <sinuhe> | s/but,/but/ |
| 01:56 | <sinuhe> | T--: Looking at that again, I think you're right. The code element seems more appropriate in terms of code snippets. Thank you. |
| 01:58 | <Wolfman2000> | othermaciej: thanks |
| 02:05 | <sinuhe> | Perhaps <kbd class="command argument">/etc/passwd</kbd>? |
| 02:07 | <T--> | /etc/passwd is not really command-y. |
| 02:08 | <Lachy> | Hixie, wouldn't a more appropriate place to document the formats, codecs, and standards supported by all browsers be in a wiki that can be easily kept up to date as browsers progress? |
| 02:09 | <Lachy> | judging from the replies, it seems that people are assuming that such a document would itself be a spec of some kind |
| 02:10 | <Lachy> | but that doesn't really seem like a sensible alternaive to me. |
| 02:10 | <sayrer> | hmm, I thought CSS required PNG support |
| 02:10 | <Lachy> | Even if it were just a NOTE, the W3C process would mean it could at best publish infrequent snapshots |
| 02:11 | <heycam> | svg requires png and jpeg |
| 02:13 | <sinuhe> | T--: A shell positional parameter recognized as an argument of a command, though it indeed is also a file, I'm not using command completion (what I assume you mean by tabbing), but giving a description in computer documentation for what a user should type to accomplish a generic task, (specifically, a description of CSH ! history modifiers in Bash and how to use them). I find it confusing knowing when to use samp, code, kbd, and var |
| 02:13 | <sinuhe> | (though the latter seems simplest), and when to avoid them altogether (such as with a password example). |
| 02:19 | <T--> | sinuhe, I'm not authorative in this regard, just reporting my gut instinct. kbd for communicating with the computer, samp for content, the computer reports back and var is orthogonal to both of them, meaning usable whenever one needs something variable or an example. |
| 02:19 | <sinuhe> | T--: OK. Thanks for your input. |
| 02:42 | <roc> | keep digging, Tom |
| 02:43 | <roc> | your hole is only getting deeper |
| 02:47 | <karlcow> | then you throw a rock in the hole to see how deep it is |
| 02:56 | <roc> | huh, Tom went to CMU |
| 04:42 | <Hixie> | Lachy: a wiki seems reasonable, yeah |
| 04:49 | <MikeSmith> | in general for the wiki stuff, it would be nice to ultimately having something that was more programatically read-friendly |
| 04:50 | <MikeSmith> | with a simple, shared/documented API |
| 04:56 | <sayrer> | fucking maciej |
| 05:07 | <MikeSmith> | sayrer: you got a verb in that sentence, but it seems to lack a subject |
| 05:08 | <sayrer> | http://lists.w3.org/Archives/Public/public-html/2009Jul/0093.html |
| 05:08 | <sayrer> | pure patent fud |
| 05:08 | <sayrer> | stay classy |
| 05:10 | <sayrer> | maybe they should assert a patent |
| 05:10 | <sayrer> | like they do in other working groups |
| 05:11 | <MikeSmith> | I'm surprised that nobody other than Doug has mentioned Sun's OMS Video so far in this thread |
| 05:12 | <MikeSmith> | sayrer: you know about OMS video? |
| 05:12 | <sayrer> | no, don't care |
| 05:12 | <sayrer> | not right now anyway |
| 05:12 | <karlcow> | :) |
| 05:13 | <sayrer> | then he emails www-archive |
| 05:13 | <sayrer> | http://lists.w3.org/Archives/Public/www-archive/2009Jul/0019.html |
| 05:13 | <karlcow> | The power to read the language of the environments was the province of the seer in primitive societies; he related the languages which he discerned in the outer world to the inner world, keeping both as a divine secret committed to him. |
| 05:13 | <karlcow> | McLuhan, Marshall, |
| 05:13 | <karlcow> | War and peace in the global village. p.59, McGraw-Hill, 1968. |
| 05:13 | <sayrer> | "rob, I don't think you should describe what I did quite so concretely" |
| 05:13 | <sayrer> | fucking asshole |
| 05:13 | <othermaciej> | sayrer: I'm doing my best to be polite in the face of your nerdrage |
| 05:14 | <sayrer> | fuck you |
| 05:14 | <sayrer> | seriously |
| 05:14 | <othermaciej> | but you don't make it easy |
| 05:14 | <sayrer> | ooh, now would be a good time for a maciej lecture |
| 05:14 | <karlcow> | othermaciej: your initial tone in the mail was not right. Instigating suspicion towards Mozilla. |
| 05:14 | <MikeSmith> | sayrer: you're sounding like Dick Cheney, man |
| 05:14 | <sayrer> | bring it dimbass |
| 05:15 | <MikeSmith> | sayrer: and now you've reached the point of coining new words |
| 05:15 | <sayrer> | yeah, oops :) |
| 05:15 | <karlcow> | and sayrer reacted like a pitbull. Both in opposite corners and peace ;) |
| 05:15 | <sayrer> | I kinda like the new word, though |
| 05:15 | <MikeSmith> | combination of "dumbass" and "dimwit" |
| 05:15 | <MikeSmith> | very classy, sayrer |
| 05:15 | <MikeSmith> | you're setting a great example |
| 05:15 | <sayrer> | THX MikeSmith |
| 05:16 | <othermaciej> | I think it's accurate that Mozilla wants to push Theora via the HTML5 spec notwithstanding the concerns of others, or at least that's the position of all Mozilla representatives who have spoken up |
| 05:16 | <sayrer> | it could be not the fastest bass fish in the stream |
| 05:16 | karlcow | is waiting for mr lastweek post |
| 05:16 | <sayrer> | actually I think it is just the concerns of some people |
| 05:16 | <sayrer> | not "others" |
| 05:19 | <karlcow> | othermaciej: except saying that, you are playing like if you were a noble knight, when in fact, all people *really* involved in html 5 work have business/community interests and all of them are pushing their things. Apple included ;) |
| 05:19 | <othermaciej> | I didn't mean to imply that this is somehow sinister |
| 05:19 | <othermaciej> | sure, Apple has an agenda too |
| 05:19 | <sayrer> | you wrote what you wrote |
| 05:19 | <roc> | has Apple actually gone on the record with "We will not support Theora because of A, B, and C"? |
| 05:19 | <sayrer> | own your words |
| 05:19 | <roc> | All I've heard is "people might not support Theora because of A, B and C" |
| 05:19 | <othermaciej> | it's natural for Mozilla to care more about Mozilla's concerns than about Apple's concerns |
| 05:20 | <othermaciej> | and vice versa |
| 05:20 | <roc> | and Ian saying that unspecified vendors have declined to support it |
| 05:20 | <sayrer> | a fair point, roc |
| 05:20 | <othermaciej> | Apple doesn't support it in any of the products where we have shipped HTML5 <video> |
| 05:20 | <sayrer> | that's not the same thing that ian said, is it? |
| 05:20 | <othermaciej> | we don't have immediate plans to, but I wouldn't say never because things can change |
| 05:20 | <roc> | that's not what I mean |
| 05:20 | <roc> | Ian says vendors have vetoed it |
| 05:21 | <othermaciej> | putting it in the spec would not change our current unwillingness to implement it |
| 05:21 | <roc> | if those vendors don't come forward and say why, it's hard to give their concerns much weight |
| 05:21 | <sayrer> | yeah, I don't really think we need to reverse engineer the Ian Proxy |
| 05:22 | <othermaciej> | I think when I've expanded on Apple's concerns, they haven't been given much weight in any case |
| 05:22 | <sayrer> | is this sort of a rush limbaugh play the victim stance? |
| 05:23 | <sayrer> | sometimes, you state your case, and you lose |
| 05:23 | <sayrer> | that's how it goes |
| 05:23 | karlcow | will not support ogg theora. I'm still trying to implement D-1 with Fortran on a Z-80 |
| 05:23 | <roc> | othermaciej: it's honestly hard to tell which of the concerns you raise are the ones that are actually stopping Apple from supporting Theora |
| 05:23 | <sayrer> | it doesn't mean your position wasn't considered |
| 05:23 | <othermaciej> | sayrer: that is a mature point of view that I wish you would adopt |
| 05:24 | <roc> | othermaciej: since officially, we don't even know that Apple has told Ian they won't support Theora |
| 05:24 | <sayrer> | I know when I lose |
| 05:24 | <roc> | you're blaming Mozilla for failing to read between the lines correctly |
| 05:24 | <othermaciej> | roc: I think Apple has made it's position pretty clear, both now and in the past |
| 05:25 | <othermaciej> | I'm not sure what additional clarifications you are looking for |
| 05:25 | <sayrer> | when reading between the lines completely correctly, and rejecting the conclusion |
| 05:25 | <doublec> | karlcow: is that for an ATJ2085 based MP3 player device? |
| 05:25 | <sayrer> | and, seriously, sending patent FUD to the public-html list? |
| 05:25 | <sayrer> | like... how dare you? |
| 05:26 | <othermaciej> | I know! how shocking! |
| 05:26 | <sayrer> | is apple going to assert a patent? |
| 05:26 | <karlcow> | doublec: hehe ;) was kidding |
| 05:26 | <sayrer> | like in the widgets wg? |
| 05:26 | <sayrer> | it would be good to know now |
| 05:27 | <roc> | othermaciej: have you got a link to an email where Apple acknowledges telling Ian they won't support Theora? |
| 05:27 | <othermaciej> | I don't know of any Apple patents that would read on Ogg Theora |
| 05:27 | <doublec> | karlcow: man I was looking forward to my cheap device playing h.264 and other codecs |
| 05:27 | <karlcow> | ;) |
| 05:27 | <sayrer> | othermaciej: that's an answer to a question I didn't ask |
| 05:27 | <othermaciej> | roc: we've stated on the public list that we're against declaring Ogg Theora the baseline as things currently stand, and we haven't retracted that position; I'm not sure what more you want |
| 05:28 | <othermaciej> | sayrer: I don't see how we could assert a patent we don't have |
| 05:28 | <sayrer> | I think roc's question was pretty plainspoken |
| 05:29 | <sayrer> | othermaciej: that's true, but you didn't say you don't have one |
| 05:29 | <othermaciej> | as far as I know, no one at Apple is aware of patents we hold that might apply to Theora, nor do I think anyone is looking for any |
| 05:29 | <sayrer> | but you have also said you don't know much about it |
| 05:29 | <sayrer> | I think you wrote it today |
| 05:30 | <sayrer> | so I am not sure why this is an interesting argument for you to be having, unless you just want to bloviate. |
| 05:31 | <roc> | othermaciej: You've said why you think it's not a good idea, is that what you mean? |
| 05:32 | <roc> | I don't think I've seen a flat-out "We refuse to support Theora" statement |
| 05:32 | <roc> | I may have missed it in the volume |
| 05:34 | <sayrer> | that would be some good data |
| 05:34 | <othermaciej> | I believe that when the Theora requirement was first introduced we publicly asked for it to be removed, although I would have to study the email archives to be sure of what exactly was and wasn't said |
| 05:35 | <sayrer> | do you refuse to support theora? |
| 05:35 | <othermaciej> | If there are additional statements that you think would have a chance of changing anyone's position instead of just acting as flamebait, let me know |
| 05:35 | <sayrer> | it is a simple question |
| 05:35 | <sayrer> | you don't even have to answer ehre |
| 05:35 | <sayrer> | here |
| 05:36 | <sayrer> | just mail your position to the list |
| 05:36 | <othermaciej> | I already answered: we don't support it, we have no plans to support it, we won't become willing to support it solely because of a spec change, but I don't really have the ability to make a promise for all time no matter what that we won't implement it |
| 05:37 | <roc> | othermaciej: I think it would be honest to state your position clearly |
| 05:37 | <sayrer> | the question can be answered with "yes" or "no" |
| 05:37 | <sayrer> | but, again, I am not asking you to answer here |
| 05:37 | <sayrer> | only to send the answer to the list |
| 05:37 | <othermaciej> | I don't think our position is actually unclear to anyone |
| 05:38 | <sayrer> | I don't know the answer to the question |
| 05:38 | <sayrer> | it seems like it could be "don't want to, but maybe" |
| 05:38 | <sayrer> | or it could be "no" |
| 05:38 | <sayrer> | or it could be "yes" |
| 05:38 | <sayrer> | It is clear that you don't want to actually say anything |
| 05:39 | <sayrer> | so, your position is clear in that respect |
| 05:39 | <roc> | othermaciej: most people have guessed that Apple's feedback caused Ian to drop the codec requirements |
| 05:39 | <roc> | othermaciej: in that sense, your position is clear to everyone |
| 05:40 | <othermaciej> | the codec requirements for video were dropped long ago |
| 05:40 | <roc> | ok, keep them out |
| 05:40 | <sayrer> | I don't get it, are you arguing with roc? |
| 05:40 | <othermaciej> | Apple has always said Theora shouldn't be required |
| 05:40 | <roc> | he is |
| 05:41 | <othermaciej> | I believe our position is pretty much on the record |
| 05:41 | <roc> | othermaciej: let me rephrase the question. Is Apple one of the vendors Ian refers to as vendors who refuse to support Theora? |
| 05:42 | <othermaciej> | roc: what I said above is Apple's position, I don't know whether anyone would describe that as "refuse" or not, but I personally prefer to state it accurately instead of in one-word form |
| 05:42 | <sayrer> | it's a right squirmy worm, this |
| 05:43 | <sayrer> | I did notice that Ian said "browser vendors" today regarding WAVE |
| 05:43 | <sayrer> | but then changed to "browser vendor in question" in a subsequent response |
| 05:43 | <othermaciej> | I'm pretty sure Apple has no problem with WAVE, or at least if anyone at Apple does, they haven't told me |
| 05:43 | <sayrer> | roc asked " Is Apple one of the vendors Ian refers to as vendors who refuse to support Theora?" |
| 05:44 | <roc> | that's a simple question of fact |
| 05:44 | <roc> | not a question about the meaning of "refuse" |
| 05:44 | <sayrer> | my point is that Ian has characterized one browser vendor as plural "browser vendors" |
| 05:45 | <roc> | that might just be a mistake |
| 05:45 | <sayrer> | could be |
| 05:45 | <sayrer> | hard to be optimisitc |
| 05:46 | <othermaciej> | If the word "refuse" is very important to you, then I would say by definition 3 of the definition here: <http://dictionary.reference.com/browse/refuse>; you could characterize what I said as "refuse" |
| 05:46 | <roc> | let's forget about what I mean by refuse |
| 05:46 | <othermaciej> | but I would still add that I wouldn't rule out changing conditions would change our mind |
| 05:47 | <roc> | Is Apple one of the vendors *Ian* refers to as vendors who refuse to support Theora? |
| 05:47 | <othermaciej> | just that a codec requirement HTML5 spec is not one of them |
| 05:47 | <othermaciej> | roc: if you want an interpretation of Hixie's statement shouldn't you ask him? |
| 05:47 | <roc> | he won't tell me because it's private feedback |
| 05:47 | <sayrer> | private feedback in the open process |
| 05:47 | <sayrer> | boy howdy |
| 05:48 | <othermaciej> | I don't think I personally used the exact word "refuse" in any private conversation with Ian, I can't be certain that no one else has |
| 05:48 | <sayrer> | I said "garbage" just today! |
| 05:48 | <roc> | othermaciej: if you think Apple's position is clear, can you give Ian permission to answer my question? |
| 05:48 | <sayrer> | almost the same |
| 05:49 | <othermaciej> | I don't think I said anything to him besides the position stated above, and the fact that we think it's technologically infeasible to ship Theora on our mobile devices at a quality level we are happy with due to lack of hardware support |
| 05:49 | <roc> | I assume you read my reply about that |
| 05:49 | <sayrer> | funny story |
| 05:49 | <othermaciej> | I'm fine with Ian answering whether Apple is a vendor he characterized as refusing to support Theora |
| 05:50 | <sayrer> | in the Atom days, a lot of people were worried about J2ME not supporting PUT and DELETE |
| 05:50 | <othermaciej> | I would assume we are but I can't speak for him any more than you can |
| 05:50 | <sayrer> | that turned out to be a pretty foolish thing to value |
| 05:50 | <sayrer> | you don't write tomorrow's standard for today's devices |
| 05:52 | <roc> | othermaciej: I would actually like clarification about what exactly would change Apple's mind |
| 05:52 | <roc> | if "hardware support" means "actual Theora ASICs in the iPhone" then I concede that's going to be difficult |
| 05:53 | <sayrer> | but also, there is some level where a client is non-conforming |
| 05:53 | <sayrer> | HTML5 requires all sorts of things to get the job done |
| 05:54 | <roc> | but I honestly don't see how supporting Theora in software makes any Apple product worse |
| 05:54 | <sayrer> | it seems bogus to draw the line at a video codec and say "that... that is cruel and unusual!" |
| 05:54 | <othermaciej> | that's what Adobe thinks about Flash |
| 05:56 | <roc> | othermaciej: I'd also like clarification on why Vorbis is unacceptable to Apple |
| 05:57 | <sayrer> | it occurs to me that a real answer probably wouldn't come as an IRC reply |
| 05:57 | <sayrer> | just sayin |
| 05:57 | <roc> | it doesn't have to be an IRC reply |
| 05:57 | <othermaciej> | I don't know if Apple has stated a position on Vorbis, and I haven't asked around what Apple thinks about it in official terms |
| 05:58 | <roc> | hmm |
| 05:58 | <sayrer> | roc, but a reply comes, huh |
| 05:58 | <roc> | Ian said on public-html "If anyone can get Apple to implement it, I'd be more than happy to update |
| 05:58 | <roc> | the spec to mention that as the common audio codec." |
| 05:58 | <roc> | I guess it's not clear whether that means he had private feedback vetoing it, or not |
| 05:58 | <roc> | (booo! hisss!) |
| 05:58 | <othermaciej> | as far as I can tell informally Apple doesn't like it |
| 05:59 | <sayrer> | well, private feedback is bad |
| 05:59 | <sayrer> | but a veto is not part of the w3c process |
| 06:01 | <sayrer> | and, if one wanted to get especially pedantic, one could say that Ian conveying private feedback breaks the requirement that a participant must not represent more than one member |
| 06:04 | <sayrer> | this reminds me of those old videos with ari fleischer |
| 06:05 | <sayrer> | (former whitehouse press secretary, for the rest of the world) |
| 06:05 | <roc> | othermaciej: sorry this was so heated. It's not personal. |
| 06:05 | <sayrer> | not in roc's case |
| 06:05 | <roc> | sayrer: we know who Ari Fleischer is |
| 06:06 | <othermaciej> | roc: I know it's not personal in your case; but thanks for saying so, as well |
| 06:06 | <sayrer> | othermaciej: in my case, it is personal. you're an ass. |
| 06:11 | <othermaciej> | sayrer: are you trying to challenge me to a bar brawl or something? |
| 06:12 | <sayrer> | no :) |
| 06:13 | <sayrer> | have you ever been in a bar brawl? I can't say I recommend it. |
| 06:14 | <othermaciej> | actually, yes |
| 06:14 | <othermaciej> | I won, but I also discovered that in a bar fight, everyone is the loser |
| 06:14 | <othermaciej> | even if you win |
| 06:14 | <sayrer> | that's not true |
| 06:14 | <sayrer> | depends what the other guy did |
| 06:15 | <sayrer> | there are just wars, and sometimes it's just a war |
| 06:17 | <othermaciej> | do you believe in the theory of just war? |
| 06:18 | <sayrer> | yes, but I think it is very rare |
| 06:18 | <sayrer> | iow, I find it hard to rule out |
| 06:19 | jwalden | thinks this conversation has gone nowhere, arguably even less than nowhere, over the last two days of IRC argument |
| 06:20 | sayrer | is not so worried about that |
| 06:20 | <sayrer> | lots of information has been gathered |
| 06:20 | <sayrer> | for a data driven effort, that is progress |
| 06:20 | jwalden | wonders who else would agree |
| 06:21 | <sayrer> | well, I could certainly see the argument about lack of progress |
| 06:21 | <sayrer> | do you also think nothing new was learned? |
| 06:22 | <othermaciej> | I think a fair amount of conversation happened between sayrer's swearing and gratuitous personal attacks |
| 06:22 | <othermaciej> | I'm not sure anyone learned anything new |
| 06:22 | <othermaciej> | can't say I did |
| 06:23 | <roc> | I have |
| 06:23 | <sayrer> | othermaciej: it's not clear to me that you have steered clear of personal attacks |
| 06:23 | <sayrer> | let's take "nerdrage" for instance |
| 06:23 | <sayrer> | a funny word, I'll grant you |
| 06:24 | <sayrer> | so I guess you have the high ground on swearing |
| 06:24 | <sayrer> | I will order you a medal from amazon |
| 06:55 | <MikeSmith> | sayrer: yt? |
| 06:55 | <sayrer> | yep |
| 06:58 | <sayrer> | there is a gold medal available for $0.98 |
| 06:58 | <sayrer> | http://www.amazon.com/Century-Novelty-Gold-Medal/dp/B001VSBGKS/ref=pd_sim_hpc_1 |
| 06:59 | <MikeSmith> | sayrer: your cvs access is set up |
| 06:59 | <sayrer> | thanks so much |
| 06:59 | <MikeSmith> | I'll pm you with the details |
| 06:59 | <sayrer> | kk |
| 07:03 | <Hixie> | sayrer: i never said anyone unspecified refused to implement Theora |
| 07:04 | <Hixie> | I said someone or someones unspecified had requested that the requirement for Wave PCM be removed, but that I had changed their mind so that they would implement it anyway |
| 07:04 | <sayrer> | Hixie: I think roc was asking |
| 07:08 | <sayrer> | Hixie: ah, roc has sent you and the list a very specific question via electronic mail |
| 07:08 | <sayrer> | http://lists.w3.org/Archives/Public/public-html/2009Jul/0098.html |
| 07:08 | <Hixie> | k |
| 07:24 | <sayrer> | Hixie, your latest mail seems wrong wrt to SQL |
| 08:37 | <hsivonen> | cool. Prince 7 beta is out |
| 08:45 | gsnedders | notices the time, curses, and wonders whether he'll beat jgraham to work |
| 08:47 | <othermaciej> | are attributes that apply to "all elements" like hidden and draggable supposed to apply only to elements in the HTML namespace or other elements too? |
| 08:48 | <Philip`> | othermaciej: http://www.whatwg.org/specs/web-apps/current-work/multipage/dom.html#global-attributes |
| 08:48 | <Philip`> | "HTML elements" |
| 08:49 | <othermaciej> | all righty |
| 09:31 | <hsivonen> | has Apple always been part of the licensor pool for H.264 or has Apple joined the pool lately? |
| 09:32 | hsivonen | can't remember noticing Apple in there earlier |
| 09:32 | <gsnedders> | I don't remember them there either. I can only remember them being in the container pool. |
| 09:34 | <hsivonen> | I see Nokia in the AAC pool but not in the H.264 pool despite what shepazu said on list |
| 09:34 | <othermaciej> | I don't remember us being in the H.264 pool before either |
| 09:34 | <othermaciej> | (not that I have any specific knowledge about this besides what the site says) |
| 09:35 | <hsivonen> | Apple hasn't always been there: http://web.archive.org/web/20071126123130/http://www.mpegla.com/avc/avc-licensors.cfm |
| 09:38 | <othermaciej> | I'm also not clear on why the MPEG-4 Systems portfolio is no longer offered (I assume that is what would cover containers) |
| 09:54 | <shepazu> | hsivonen: sorry, my mistake, I thought Nokia was an H.264 license holder |
| 09:55 | <othermaciej> | <pendatic>they are indeed a license holder, just not a patent holder |
| 09:55 | <othermaciej> | </pedantic> (he said optimistically) |
| 10:29 | <jgraham> | "I don't think anyone is concerned about risk of additional patent exposure from other HTML5 technologies, to the point that this is a showstopper for implementing" |
| 10:29 | <jgraham> | Didn't Microsoft claim exactly that about <canvas>? |
| 10:29 | <gsnedders> | jgraham: Where is that from? |
| 10:29 | <gsnedders> | jgraham: yes |
| 10:29 | <jgraham> | gsnedders: One of othermaciej's (many) recent posts |
| 10:30 | <jgraham> | to public-html |
| 10:31 | gsnedders | hasn't really been touching public-html, waiting for MikeSmith to get back to me about whether there was any need to do anything being an intern |
| 10:31 | <othermaciej> | I don't recall Microsoft claiming that, though I remember they had some complaints about <canvas> |
| 10:31 | <MikeSmith> | an organizer for an an upcoming event in Hungary has contacted me about possibility of having an HTML5 presentation there. Anybody know of somebody who's been involved in discussions on whatwg or public-html lists who's based in Hungary or nearby? |
| 10:31 | <othermaciej> | Apple has a known patent on <canvas> which we've never enforced and which we disclosed and didn't exclude as part of the HTML5 FPWD patent disclosure requirement |
| 10:32 | <othermaciej> | I think Microsoft complained about the difficulty and expense of doing patent research on <canvas> for patent disclosure purposes, presumably a non-issue now |
| 10:33 | <MikeSmith> | gsnedders: for now, you are OK to continue to participate in public-html discussions as an individual |
| 10:33 | <gsnedders> | MikeSmith: k |
| 10:33 | <MikeSmith> | we would need chaals to add you as Opera rep, and I think he may have just left for a vacation until the 13th |
| 10:34 | <Philip`> | jgraham: I think Chris Wilson mentioned the issue of canvas patents, before Apple disclosed it in the patent non-exclusion period |
| 10:35 | <Philip`> | but I think I only remember it being mentioned somewhat tangentially, not as a reason why they weren't implementing canvas |
| 10:35 | <jgraham> | I had got the impression that it was not the apple patent in particular that they were concerned about |
| 10:35 | Philip` | hadn't got that impression |
| 10:35 | <jgraham> | But I don't really remember the details; it was discussed a bit at the 2007 TPAC iirc |
| 10:36 | <Philip`> | Oh |
| 10:36 | Philip` | just knows what was said on lists |
| 10:37 | <othermaciej> | I believe Apple disclosed the fact that we had a patent even before the non-exclusion period |
| 10:37 | <Philip`> | http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-March/010129.html |
| 10:39 | <othermaciej> | personally I'm surprised that Apple's patent has held up, but fortunately it's now not a problem for anyone |
| 10:39 | <Philip`> | The patent seems to be basically a dump of Apple's canvas documentation |
| 10:40 | <Philip`> | and then repeated in lawyerish |
| 10:40 | <gsnedders> | Well of course, what do you expect a patent to be? |
| 10:40 | <Hixie> | i don't understand why this codec issue is so popular |
| 10:40 | <Hixie> | i've received more private feedback abotu this than almost anything else |
| 10:41 | <Hixie> | put together |
| 10:41 | <gsnedders> | Because it is the free web v. the evil mega-corps! |
| 10:41 | <othermaciej> | it seems to generate a lot of public feedback too |
| 10:42 | <Philip`> | Hixie: Because the list of involved parties is large enough that everyone can find someone they hate |
| 10:42 | <jgraham> | Hixie: Because people wwant usable video on the web and are concerned that the current spec will impede that? |
| 10:43 | <jwalden> | "almost" anything else? |
| 10:43 | <jwalden> | I'm guessing only rdfa could possibly surpass |
| 10:43 | <jwalden> | but I don't see why such feedback would have been private |
| 10:44 | <othermaciej> | I'm kinda curious who is sending Hixie private codec feedback |
| 10:44 | <othermaciej> | I gave him some when asked but I believe I have also made all the same points in public |
| 10:44 | <roc> | people don't want to have to pay money to produce or consume Web content |
| 10:45 | <jwalden> | or maybe the more reasonable interpretation is s/any/every/ |
| 10:45 | <gsnedders> | roc: I don't think anybody contends that |
| 10:45 | <Hixie> | othermaciej: random people from all over the web |
| 10:45 | <othermaciej> | there are a surprising number of H.264 advocates / Ogg haters in threads about this topic on, e.g., reddit |
| 10:45 | <roc> | gsnedders: then no-one should be surprised |
| 10:46 | <othermaciej> | (I don't think any of them are Apple shills) |
| 10:46 | <roc> | othermaciej: a lot of them are x264 developers and their fans |
| 10:46 | <jwalden> | I want the bikeshed to not be splotchy |
| 10:46 | <jwalden> | (oversimplification, to be sure, but a vague attempt at humor) |
| 10:46 | <jwalden> | even if only self-humor :-) |
| 10:46 | <jwalden> | and I suppose s/splotch/block/ to use the technical term |
| 10:47 | <Philip`> | The discussion is great because you can complain about Apple, Microsoft, Google, Mozilla, W3C, Adobe, and Nokia - everyone can find someone to argue against |
| 10:48 | <roc> | there seems to be a large community of codec geeks who hang out in doom9 and other places who tend to focus on the technology, don't care about licensing issues, and hence tend to be pro-H.264 anti-Theora |
| 10:48 | <jgraham> | As long as no one can complain about Opera :) |
| 10:48 | <Philip`> | (Seemingly nobody thinks Opera has enough influence to be worth caring about) |
| 10:48 | <jgraham> | :( |
| 10:48 | <roc> | jgraham: the only thing worse than being talked about is not being talked about |
| 10:49 | <jgraham> | Maybe I should try to convince someone here to make our policy really controversial |
| 10:49 | <jgraham> | We could send out teams of ninjas to hunt down pepople who author videos in non-RF formats |
| 10:49 | jwalden | is tangentially sad that the person named singer does not work for the company named opera |
| 10:50 | <roc> | jgraham: announce support for Dirac |
| 10:50 | <Philip`> | jgraham: Use the Opera Mini/Turbo proxies to dynamically transcode content from H.264 into Theora |
| 10:50 | <othermaciej> | roc: that would be more confusing than controversial |
| 10:50 | <othermaciej> | Philip`: for delivering content to phones, the other way around might make more sense... |
| 10:51 | <takkaria> | Opera need a Therora to Dirac transcoder |
| 10:51 | <takkaria> | it's the only way to preserve free content on the Web |
| 10:51 | <Hixie> | oh that's just what we need, a lossy codec to lossy codec convertor |
| 10:52 | <Philip`> | takkaria: Just use ffmpeg or mencoder, they should support that already |
| 10:52 | <othermaciej> | I'm starting to think motion jpeg might be a good idea |
| 10:52 | <Philip`> | (if compiled with suitable flags and libraries) |
| 10:53 | <othermaciej> | it's so stupid that everyone can hate it |
| 10:53 | <Hixie> | othermaciej: these long flamewars really cause people to lose good judgement :-P |
| 10:53 | <Philip`> | MJPEG might be a good way to distract from the issue of finding a solution that anyone can seriously use |
| 10:53 | <roc> | and remarkably, it has actual proponents |
| 10:54 | <othermaciej> | well, a baseline should be a least common denominator |
| 10:54 | <othermaciej> | hard to think of anything lesser than MJPEG |
| 10:54 | <takkaria> | HD video in MJPEG would be amusing |
| 10:55 | <gsnedders> | Anything prior to MPEG-1, of those related codecs, has fixed resolutions, right? |
| 10:55 | hsivonen | has a camera that outputs HD MJPEG (720p@24fps) |
| 10:55 | <Hixie> | othermaciej: one of the e-mails i received suggested i make animated GIFs the baseline |
| 10:56 | <othermaciej> | Hixie: I stand corrected |
| 10:56 | <gsnedders> | hsivonen: And it manages to have high enough IO to cope? That's actually somewhat impressive. |
| 10:56 | <Hixie> | :-) |
| 10:56 | <hsivonen> | Hixie: btw, another area where Theora is more righteous than H.264 is the profile stuff |
| 10:57 | <hsivonen> | Hixie: Theora doesn't do profiles. H.264 has profiles times levels |
| 10:57 | <hsivonen> | Hixie: and we don't like profiles |
| 10:57 | <Hixie> | i personally don't care which of 264 or theora is better |
| 10:57 | <hsivonen> | gsnedders: I think I have a "Class 6" SDHD card |
| 10:57 | <roc> | on my blog, an H.264 advocate posted a link to an H.264 vs Theora comparison |
| 10:57 | <hsivonen> | (or whatever the abbreviation is today) |
| 10:57 | <Hixie> | it's academic unless they could both be implemented, which they can't |
| 10:57 | <roc> | his H.264 file didn't play in Quicktime |
| 10:58 | jgraham | doesn't understand "could both be implemented" |
| 10:58 | <roc> | Hixie: I have disputed that, but I need to take a break from all this for a few days |
| 10:58 | <gsnedders> | They could be implemented :P |
| 10:59 | <Hixie> | roc: you could implement 264? |
| 10:59 | <roc> | hopefully in the meantime someone will reply to all my emails on the subject |
| 10:59 | <gsnedders> | Esp. with H.264 the legally is questionable, but it could be implemented. |
| 10:59 | <roc> | sorry, I see what you mean |
| 10:59 | <Philip`> | What would happen if you took MJPEG but 'rotated' the input so you were compressing x-time planes instead of x-y planes? That would make it benefit from temporal coherency, which is better than standard MJPEG |
| 11:00 | <roc> | then you don't benefit from Y-locailty |
| 11:00 | <roc> | fail |
| 11:00 | <Philip`> | Oh :-( |
| 11:00 | <hsivonen> | Wikipedia should push XiphQT to Safari users. are they doing that already? |
| 11:01 | <roc> | that's kind of impractical |
| 11:01 | <roc> | it would be more practical to create a virus that installs XiphQT on Mac worldwide |
| 11:03 | <hsivonen> | nessy: who is working on the Ogg ActiveX control you mentioned had been mentioned in the open video conference? |
| 11:04 | <Philip`> | Okay, so if you consider the video to be a 3D x-y-time cube, you could quantise it into a series of nested cubes of decreasing size, and then flatten the surface of each cube into a plane, and compress that series of planes as MJPEG |
| 11:04 | <Philip`> | Then you'll benefit from locality in all three dimensions |
| 11:04 | <Philip`> | I can't imagine why nobody has done this already |
| 11:04 | <othermaciej> | I doubt binary plugins will be effective at resolving a format war, the last time the industry tried that, Flash won |
| 11:05 | <hsivonen> | well, yeah, an implementation in ActionScript would be better |
| 11:05 | <othermaciej> | Philip`: it's desirable for codecs to be able to decode frames sequentially, for smooth playback |
| 11:05 | <othermaciej> | Philip`: your scheme would require decoding N frames at a time for possibly large N |
| 11:05 | <othermaciej> | there's a Java implementation of Theora |
| 11:06 | <othermaciej> | I think Wikipedia uses it |
| 11:06 | <roc> | they do |
| 11:06 | <hsivonen> | othermaciej: I use it, too, on my site. It sucks when you put two videos on one page |
| 11:06 | <hsivonen> | othermaciej: like here http://hsivonen.iki.fi/makasiinit/ |
| 11:06 | <roc> | it's not too bad, for a Java implementation of a video codec |
| 11:06 | <othermaciej> | hsivonen: using Actionscript would improve that to one |
| 11:07 | <Philip`> | othermaciej: To decode frame 1, you would only need the outermost cube; to decode frame 2, you would need the next smaller cube plus the edges of the previous one; etc; so you could actually do it sequentially |
| 11:07 | <othermaciej> | you know, the more I read the thread title "Nothing is really hidden", the more I want to take it as a sublime philosophical point |
| 11:07 | <hsivonen> | I don't mean playing them at the same time. Just weird breakage. |
| 11:07 | <hsivonen> | at least Flash works with multiple ads on one page. |
| 11:07 | <othermaciej> | Philip`: oh, nested cubes, well that seems like it would be pretty terrible |
| 11:07 | <othermaciej> | I was assuming cube slices |
| 11:07 | <hsivonen> | (the videos on that page were encoded 3 years ago) |
| 11:08 | <hsivonen> | (I should re-encode them with Thusnelda) |
| 11:08 | <Philip`> | othermaciej: Hmph, maybe I should leave codec design to "experts" then :-( |
| 11:09 | <othermaciej> | well if you're going to do something more complicated than vanilla motion jpeg, it may as well be good |
| 11:10 | <Philip`> | What if I consider overcomplication to be inherently good? |
| 11:10 | <takkaria> | othermaciej: you should take it as a philosophical point, since it's pretty philosophical in nature, and it ends up like most philosophy, with people redefining other people's words in order to be right |
| 11:11 | Philip` | notes that his contributions to this discussion are probably not especially productive |
| 11:12 | <othermaciej> | Philip`: so far your contributions have been more useful than those of some people who think they are contributing seriously |
| 11:13 | <othermaciej> | takkaria: I've given up on actually reading the mails in that thread, so I wouldn't know |
| 11:14 | <hsivonen> | so Mozilla, Apple and Opera are now saying PCM WAVE isn't a problem... |
| 11:16 | <othermaciej> | by process of elimination, it was probably Google (the Chromium bit of it) that complained, but I buy Hixie's argument that their complaint was irrelevant if they agreed to implement it anyway |
| 11:16 | <Philip`> | Anonymous feedback isn't so anonymous when you know it comes from one of a finite group and you can ask all the others in the group if it was them |
| 11:18 | <othermaciej> | I feel bad for duplicating Hixie and doublec's replies |
| 11:19 | <doublec> | repitition is important :) |
| 11:21 | Philip` | has a revelation |
| 11:21 | <Philip`> | This MJPEG-over-surfaces-of-cubes-in-x-y-t thing clearly ties into the notion of the Time Cube |
| 11:22 | Hixie | gets more private feedback from people trying in good faith to be helpful but only suggesting things that were attempted years ago |
| 11:22 | <Hixie> | i've been trying to go to bed for the past 3 hours, but there's always more mail! |
| 11:23 | <jwalden> | othermaciej: if you're going to take a high ground, don't snark at someone behind his back |
| 11:23 | <jwalden> | notwithstanding that I basically agree |
| 11:23 | <takkaria> | doublec: if repetition is important, why doesnt't HTML5 support repetition in forms? hmm? |
| 11:23 | <takkaria> | see, your logic is inconsistent |
| 11:24 | <doublec> | hehe |
| 11:24 | <hsivonen> | no edits to http://en.wikipedia.org/wiki/Ogg_controversy yet in the last 3 days |
| 11:24 | <Hixie> | holy crap, we've surged in subscribers |
| 11:24 | <Hixie> | 1068! |
| 11:25 | <Hixie> | i should post controversial things about theora every 6 months |
| 11:25 | <hsivonen> | this is all just a way for Hixie to attract more slashdotters to subscribe |
| 11:25 | <Hixie> | public-html hasn't changed, still 341 |
| 11:26 | <Philip`> | That's unsurprising given the relative effort needed to join each list, and the relative benefits |
| 11:26 | <othermaciej> | jwalden: what did I say that you thought was snarking about someone behind his back? the thing about the "nothing is really hidden" thread? |
| 11:26 | <Philip`> | (Joining the HTML WG doesn't even give you the benefit of being able to post to the list, since everyone has that anyway) |
| 11:26 | <gsnedders> | takkaria: lunch now? |
| 11:26 | <takkaria> | gsnedders: aye |
| 11:27 | <jwalden> | othermaciej: "Philip`: so far your contributions have been more useful than those of some people who think they are contributing seriously" |
| 11:27 | <othermaciej> | (though I guess if that was snarking, it would be about Shelley) |
| 11:27 | <nessy> | hsivonen: the work has not really been started yet on the ActiveX control, but the guy who is maintaining oggcodecs is interested |
| 11:27 | takkaria | snarks (scare quotes implies) |
| 11:27 | <hsivonen> | nessy: ok. any plan on whether it's going to be VLC minus everything else or a new thing? |
| 11:28 | <othermaciej> | jwalden: if you took that to be about a specific person rather than about the many people who posted useless things, then that person probably earned some snark |
| 11:28 | <hsivonen> | nessy: and whether it will expose an HTML5-like API for easy JS shims? |
| 11:28 | <othermaciej> | jwalden: but I didn't mean it that way |
| 11:28 | <jwalden> | othermaciej: undoubtedly (to the first), guess I misinterpreted then (to the second) |
| 11:29 | <Hixie> | othermaciej: would i be right to assume that apple wouldn't ship dirac by default either at this point? |
| 11:29 | <jwalden> | it probably doesn't help that I read all of the last two days of scrollback at once, so I got the full force of all the argument in one sitting :-) |
| 11:29 | <Hixie> | (it seems the same reasons you've stated for not doing theora apply to dirac, but it's always worth asking...) |
| 11:29 | <othermaciej> | Hixie: I don't think Dirac is currently in a state to be a viable choice for Web video |
| 11:30 | <othermaciej> | but I think all the concerns about Theora would apply |
| 11:30 | <othermaciej> | jwalden: that must have been a hair-raising experience |
| 11:31 | <hsivonen> | BBC had a partner that did Dirac in hardware (albeit not in the mobile form factor), right? |
| 11:31 | <hsivonen> | maybe that was without interframe compression? |
| 11:32 | <nessy> | hsivonen: it will be built on oggcodecs but be native otherwise |
| 11:32 | <jwalden> | othermaciej: heh |
| 11:32 | <Philip`> | IRC clients should have a rewind / fast-forward mechanism, that plays back the conversation at 2x normal speed, so you still get the real-time aspects rather than a sudden onrush of dozens of pages of chat log |
| 11:32 | <nessy> | hsivonen: oggcodecs is a directshow filter for anything ogg |
| 11:32 | <othermaciej> | Hixie: Dirac uses novel techniques, so it might be less likely to infringe the nest of existing patents, but on the other hand, there's less US-based big company exposure to it so far |
| 11:32 | <hsivonen> | nessy: ok. |
| 11:32 | <nessy> | hsivonen: and yes, it should play nicely with HTML5 - that's the whole idea |
| 11:33 | <hsivonen> | nessy: excellent |
| 11:33 | <othermaciej> | Hixie: but I'm not sure it's worth worrying about because it doesn't even have a stable spec yet and really isn't ready for prime time by anyone's estimation |
| 11:33 | <othermaciej> | lazyweb side note: does anyone know of grants or disclaimers of On2's patents on Theora that aren't limited to code derivatives of the reference implementation? |
| 11:33 | <othermaciej> | (I've been trying to research what patent grants or non-enforcement promises there are) |
| 11:33 | <nessy> | othermaciej: but Dirac has gone through a standardisation process at SMPTE and even Apple did not object to it being standardised there |
| 11:34 | <nessy> | I agree thought that it's not quite the choice for Web video (yet) |
| 11:34 | <hsivonen> | nessy: wasn't the SMPTE thing without interframe compression? |
| 11:34 | <jgraham> | Philip`: Maybe the people who black out MP's expenses reports could be employed to remove all the bits that are likely to raise ire in the IRC logs too |
| 11:34 | <othermaciej> | jgraham: but then MLW would be out of a job |
| 11:34 | <nessy> | hsivonen: it was some subset of Dirac, which is now called VC-2, yes, but I don't know the details |
| 11:35 | <Hixie> | ok no e-mail for 5 seconds, time to go to bed |
| 11:35 | <Hixie> | nn |
| 11:35 | <hsivonen> | nn |
| 11:35 | <jgraham> | gn |
| 11:36 | <Philip`> | jgraham: That's easy - just add <style>body { background: black }</style> on the logs page |
| 11:36 | <nessy> | excellent - I am sick of the email avalanche, too :) |
| 11:36 | <jgraham> | Nah, you have to let something through to give a sense of being transparent |
| 11:38 | <othermaciej> | the only licenses/disclaimers I found are these: <https://trac.xiph.org/browser/branches/vp32/vp32/VP32_license.txt>; <https://svn.xiph.org/trunk/theora/LICENSE>; |
| 11:38 | <othermaciej> | from googling it seems I am not the first to wonder if independent implementations wouldn't be covered |
| 11:42 | <jwalden> | yay, untrusted connection :-\ |
| 11:43 | hsivonen | likes how Chrome handles those |
| 11:43 | Philip` | just clicks "Accept" in Opera automatically |
| 11:44 | <othermaciej> | amazingly I didn't even notice they had a bogus cert |
| 11:44 | <othermaciej> | which goes to show you how useful the warning is |
| 11:46 | <jwalden> | it's not bogus, just issued by an untrusted authority |
| 11:46 | <Philip`> | It's more of a problem because in many cases (e.g. this one, viewing a public file) you really don't care about security at all |
| 11:47 | <othermaciej> | I shouldn't have used the word "bogus" so loosely |
| 11:47 | <jwalden> | I make the effort to add temporary exceptions when I encounter things like this |
| 11:47 | <othermaciej> | in a sense you could argue all certs are bogus, given the extremely weak verification done by CAs |
| 11:47 | <jwalden> | and complain when I can, e.g. http://bugs.freedesktop.org/show_bug.cgi?id=20250 |
| 11:48 | <Philip`> | Opera also complains "An encryption level that is not secure was requested by the server" |
| 11:51 | <othermaciej> | I found a thread in the theora mailing list archives about this from 2004 |
| 11:52 | <othermaciej> | there didn't seem to be a clear conclusion on whether the patent disclaimer applies to independent implementations |
| 11:52 | <doublec> | yeah I'm reading that too |
| 11:52 | <doublec> | maikmerten is on irc a lot and can probably comment more |
| 11:55 | <othermaciej> | from the thread it doesn't seem like there are any relevant patent licenses or disclaimers |
| 11:55 | <othermaciej> | does maikemerten work for or represent On2? |
| 11:56 | <doublec> | No, he was involved in the thread which is quite old so thought he'd be able to comment on how relevant it is. |
| 11:56 | <doublec> | eg. if another thread came up clearing things up later |
| 11:56 | <doublec> | an On2 person has frequently posted in the reddit threads about theora I notice |
| 11:57 | <othermaciej> | where on IRC can he be found? |
| 11:58 | <othermaciej> | I am a bit hesitant to dredge this up with him because of the risk of accidentally instigating an additional flamewar |
| 11:58 | <doublec> | he hangs out here, as well as in #annodex, #theora |
| 11:58 | <doublec> | he's a reasonable guy - very easy to discuss things with |
| 11:58 | <othermaciej> | I'll stick to here then |
| 11:59 | <doublec> | not a raving zealot anyway :) |
| 11:59 | <doublec> | not that there are any of those in codec discussions... |
| 12:01 | <othermaciej> | if independent implementations of Theora were disallowed or non-RF I assume that would be a problem |
| 12:01 | <hsivonen> | ffmpeg has an independent impl, right? |
| 12:01 | <othermaciej> | it would be nice if On2 made a clearer statement |
| 12:02 | <othermaciej> | dunno |
| 12:02 | <doublec> | so apple has licensed vp3, which is similar to theora |
| 12:03 | <doublec> | since theora was based on it |
| 12:03 | <doublec> | does that mean apple are covered anyway by the patents that on2 have on theora? |
| 12:04 | <hsivonen> | has Apple shipped VP3, too? |
| 12:04 | <doublec> | oh, my mistake, it's the 'third party components' quicktime page I'm ready |
| 12:04 | <doublec> | reading |
| 12:04 | <hsivonen> | ah |
| 12:04 | <othermaciej> | I think the people who said Apple has licensed VP3 were arguing based on faulty evidence (the codec download page) |
| 12:05 | <doublec> | yes, correct |
| 12:05 | <othermaciej> | I don't know if Apple has in fact licensed it, but I would guess no, b/c Apple started phasing out proprietary codecs with the advent of MPEG-4 |
| 12:05 | <othermaciej> | and VP3 appeared around the same time (in fact I believe it was a rejected candidate for MPEG-4 Part 2 video, the pre-H.264 MPEG-4 video codec) |
| 12:07 | <othermaciej> | if independent implementations are not allowed or must pay, then I would find that more of a problem from the spec POV than for Apple specifically (since the reference implementation is BSD-licensed) |
| 12:11 | <Lachy> | wikipedia says that On2 released the codec under an open source licence and disclaimed all rights to their patent, citing this licence http://trac.xiph.org/browser/branches/vp32/vp32/VP32_license.txt |
| 12:11 | <doublec> | "VP3 allows for the highest quality video, at the lowest possible data rates while still retaining the video's full frame-rate. The result is a clear 320x240 video that can be expanded to a full-screen 640x480 resolution. The power of this codec has been validated by the fact that it has been licensed by both RealNetworks and Apple for their internet video players." |
| 12:11 | <othermaciej> | that patent doesn't actually disclaim all rights (it is one of the two I linked) |
| 12:11 | <doublec> | from random internet news source of course ;) |
| 12:12 | <Lachy> | then I assume what wikipedia says is poorly written |
| 12:12 | <othermaciej> | s/that patent/that license/ |
| 12:12 | <othermaciej> | the original VP3.2 license actually has a field-of-use restriction on the patent in addition to limiting it to derivative works |
| 12:13 | <othermaciej> | the later patent disclaimer (not license) seems limited to derivative works, but doesn't have the field-of-use restriction |
| 12:15 | <othermaciej> | it's not even exactly a disclaimer, just a promise not to assert the patent |
| 12:15 | <othermaciej> | legalese is fun |
| 12:16 | <Lachy> | othermaciej, where's the patent disclaimer you're referring to? |
| 12:16 | <othermaciej> | Lachy: https://svn.xiph.org/trunk/theora/LICENSE |
| 12:17 | <othermaciej> | (despite the name of that file, it appears to contain a promise not to assert the patent, not a patent license) |
| 12:17 | <othermaciej> | (I have no idea if that makes any legal difference) |
| 12:20 | <Lachy> | that legalese is certainly confusing to me too |
| 12:21 | <doublec> | according to On2's CTO at the time: "On2's VP3 codec was licensed by Apple Computer for inclusion in their 3rd party auto-update component server, starting with QuickTime 5." |
| 12:21 | <doublec> | Not sure what 'licensed by..." actually means in regards to the 3rd party auto update component server |
| 12:22 | <othermaciej> | me neither |
| 12:22 | <othermaciej> | like I said, I think this is more of a concern from a standards process perspective than an Apple-specific perspective |
| 12:23 | <othermaciej> | if independent implementations were not allowed and RF, that would subvert the RF requirement and make it problematic to satisfy the usual CR exit conditions |
| 12:24 | <doublec> | yes, clarification from Xiph on that is really needed |
| 12:24 | <Lachy> | I wonder when On2's patent expires? |
| 12:25 | <othermaciej> | I'm not sure Xiph can clarify, unless they have some additional documents from On2 |
| 12:25 | <othermaciej> | Lachy: if we knew what the patents are it would be easy to look it up |
| 12:26 | <othermaciej> | however I can't find any record of specific patent numbers |
| 12:26 | <Lachy> | yeah, I was just looking for a patent number |
| 12:27 | <Lachy> | othermaciej, any idea when h.264 patents will all expire? The patent list on MPEG-LA only listed the expiration dates for the few that already expired. |
| 12:27 | <hsivonen> | https://twitter.com/sjespers/status/2452546402 |
| 12:27 | <Lachy> | LOL |
| 12:28 | <othermaciej> | Lachy: the spec was finalized in 2003, so the upper bound would be 2024 I think; but I haven't looked up all the patents |
| 12:29 | <Lachy> | damn, that's a long time. |
| 12:29 | <othermaciej> | there are a number of old patents that read on it, so the expiration will be very staggered |
| 12:30 | <othermaciej> | my understanding is that at least some members of the MPEG-LA pool would be willing to give a royalty-free license for some kind of baseline profile but there are some holdouts |
| 12:30 | <othermaciej> | I don't know who the holdouts are, when their particular patents expire, or if they continue to be stubborn |
| 12:30 | <othermaciej> | I might be able to learn those things though |
| 12:32 | <Lachy> | I wonder if it's possible to create a baseline codec that would exclude anything covered by the patentds owned by those holding out |
| 12:34 | <othermaciej> | no idea |
| 12:34 | <Lachy> | though, it might be a challenge even if it were possible, since we would need to get video editing software that can output content that complies with the more limited baseline |
| 12:38 | <hsivonen> | Lachy: even if MPEG-LA made H.264 Baseline RF, it would be a challenge to make video encoder software really clamp to it and to make decoders not implement anything beyond |
| 12:39 | <Lachy> | hsivonen, that's basically what I just said. |
| 12:39 | <hsivonen> | Lachy: not quite :-) |
| 12:39 | <othermaciej> | hsivonen was talking about the existing H.264 Baseline profile |
| 12:39 | <hsivonen> | Lachy: H.264 Baseline profile is already defined |
| 12:40 | <othermaciej> | not a hypothetical new profile that tries to dodge patents of stubborn patent holders |
| 12:40 | <hsivonen> | Lachy: and it's still a challenge |
| 12:40 | <Lachy> | hsivonen, where I said "baseline", I was referring to a hypothetical RF baseline |
| 12:40 | <othermaciej> | I would guess x264 can properly handle target profiles |
| 12:41 | <hsivonen> | othermaciej: if it does, it requires a lot of effort on the part of its user to set the right knobs to the right positions |
| 12:42 | <Lachy> | maybe we should just start fighting to have all the h.264 patents invalidated, at least for those who won't agree to RF licencing |
| 12:42 | <Lachy> | it would take a lot of effort though |
| 12:45 | <jwalden> | speak of the devil :-) |
| 12:45 | <jwalden> | perhaps not coincidentally, I bet |
| 12:47 | <Lachy> | jwalden, what are you referring to? |
| 12:47 | <othermaciej> | Lachy: I think he's referring to maikmerten joining the channel |
| 12:47 | <jwalden> | ^ |
| 12:48 | <Lachy> | oh |
| 12:49 | <othermaciej> | hsivonen: it looks like someone would have to undergo the painful experience of making an x264 profile definition in XML and after that it would be easy (whether for an official H.264 profile or some arbitrary other set of features) |
| 13:44 | <maikmerten> | othermaciej_, oh, hello :) |
| 13:44 | <maikmerten> | so... what devil was talked about? |
| 13:46 | <Lachy> | maikmerten, they were discussing On2's VP3 patent licensing issues in relation to implementing Theora |
| 13:47 | <maikmerten> | ah |
| 13:47 | <Philip`> | and whether it covers independent reimplementations |
| 13:48 | <Lachy> | it's not clear whether it only covers derivatives of the Xiph implementation, or any implementation |
| 13:49 | <Lachy> | maikmerten, discussion started around here http://krijnhoetmer.nl/irc-logs/whatwg/20090703#l-728 |
| 13:49 | <Lachy> | or maybe slightly before that |
| 13:50 | <maikmerten> | from what I know it should cover independent reimplementaton as the deal with On2 was done in that spirit. One can see current Theora as a non-VP3 related implementation, given it has been rewritten from scratch |
| 13:50 | <maikmerten> | Lachy, thanks |
| 13:51 | <Lachy> | maikmerten, it would be nice to get some official clarification on the issue, preferably from On2 themselves |
| 13:51 | <maikmerten> | I'll forward this |
| 13:52 | <Lachy> | thanks |
| 13:52 | <hsivonen> | maikmerten: is it a known problem that Cortado acts weird if there are two instances on one page in IE? |
| 13:52 | <hsivonen> | maikmerten: also, what's the difference between the Wikipedia and Xiph Cortado code bases? |
| 13:53 | <maikmerten> | hsivonen, no, didn't hear about that issue before |
| 13:53 | <maikmerten> | hsivonen, development moved to Xiph.org Git |
| 13:53 | <hsivonen> | maikmerten: ok. |
| 13:53 | <maikmerten> | hsivonen, we're trying to set up a proper project page |
| 13:54 | <maikmerten> | hsivonen, currently the Xiph.org version features some decoder fixes over the Wikimedia one |
| 13:54 | <doublec> | if it's a problem caused by two applets onthe page it sounds more like a jvm issue |
| 13:54 | <hsivonen> | doublec: could be |
| 13:55 | <hsivonen> | I'll try to test it some more |
| 13:57 | <doublec> | I've used cortado a bit and it's worked well for me - but only one per page. I'll try two as well. |
| 14:00 | <maikmerten> | mkay, it seems like any spec-compliant Theora product is derived from the original VP3 codebase simply because the spec was written from scratch with the VP3 source-code as input |
| 14:01 | <maikmerten> | there were zero pages of specification when the VP3/Theora deal was done |
| 14:01 | <hsivonen> | hmm. an implementation of the spec isn't a derivate work of the spec necessarily, is it? |
| 14:02 | <Philip`> | I thought the write-implementation-from-spec-written-from-implementation approach was often used as a way to *avoid* being considered a derivative |
| 14:02 | <Philip`> | like in http://en.wikipedia.org/wiki/Clean_room_design |
| 14:02 | <gsnedders> | Philip`: yes |
| 14:02 | <hsivonen> | Philip`: though that's cleaning copyright derivation |
| 14:04 | <maikmerten> | I think nobody here is lawyer enough to settle that question. The only reliable source would be the actual agreement. |
| 14:04 | <maikmerten> | I'll forward this request |
| 14:06 | <Philip`> | hsivonen: Is there any other form of derivation that would be relevant? |
| 14:06 | <hsivonen> | Philip`: the kind of derivation that the On2 promise means |
| 14:08 | <Philip`> | hsivonen: Could that kind be anything other than copyright derivation? |
| 14:08 | <Philip`> | I suppose I don't actually care about the answer, and probably nobody knows enough to answer it anyway |
| 14:09 | <Philip`> | s/nobody/nobody here/ |
| 14:11 | <gsnedders> | Huh, #location has irrelevant tests linked |
| 14:12 | <hsivonen> | Philip`: I think it's better if I don't try to guess what someone else's legalese means |
| 14:58 | <Philip`> | Hmm... I have a set of Unicode codepoints. Someone can construct a string of those codepoints, and then convert the string to NFC. How I can tell what set of codepoints the output string can contain? |
| 14:59 | <hsivonen> | font subsetting is hard |
| 15:00 | <Philip`> | It'd be easier if OS X wasn't silly enough to convert strings to NFC before trying to render them |
| 15:00 | <jgraham> | I thought OSX converted to NFD |
| 15:00 | <Philip`> | Not when rendering text, as far as I'm aware |
| 15:01 | <Philip`> | If you write HTML with "i{combining grave}" then it'll try to find the glyph corresponding to "{i grave}" |
| 15:02 | <Philip`> | (I already have code that finds a superset of the necessary codepoints, but want something stricter) |
| 15:05 | <gsnedders> | OS X uses NFD for HFS+, but otherwise uses NFC. |
| 15:07 | <hsivonen> | gsnedders: well, almost NFD for HFS+, IIRC, it's a special normalization that isn't quite exactly any real normalization form |
| 15:07 | <hsivonen> | I think the differences are in the Korean block |
| 15:07 | <gsnedders> | How can you find out if something is in an array? |
| 15:07 | <gsnedders> | (in JS) |
| 15:07 | <hsivonen> | Hangul isn't decomposed or something of that nature |
| 15:07 | <hsivonen> | I forget |
| 15:09 | <Philip`> | Argh |
| 15:09 | <Philip`> | NFC(U+1e60 U+0323) = NFC(U+1e62 U+0307) = U+1e69 |
| 15:09 | <Philip`> | s/1e62/1e63/ |
| 15:09 | <Philip`> | s/1e60/1e61/ |
| 15:09 | <Philip`> | Uh, something like that |
| 15:11 | <Philip`> | What I currently do is convert everything to NFD, and then find all codepoints that decompose into a subset of those NFDs, but that is both inefficient and incorrect |
| 15:13 | <Philip`> | (Incorrect because UnicodeData.txt doesn't always give maximal decompositions, and inefficient because e.g. {i grave} will get decomposed into a combining grave even though there's no need to include that character) |
| 17:46 | <sayrer> | ls |
| 17:46 | <sayrer> | oops |
| 17:48 | <sayrer> | ls |
| 17:48 | <sayrer> | gah |
| 17:48 | <sayrer> | sorry about the noise |
| 17:55 | <beowulf> | whatwg: ls: command not found |
| 19:47 | <Philip`> | http://tech.slashdot.org/comments.pl?sid=1291113&cid=28572213 - "The [HTML5] spec is quite readable" - really? |
| 19:52 | <Hixie> | Philip`: actually i get feedback regularly along those lines |
| 19:53 | <Hixie> | Philip`: though frankly with a spec the size of html5, there are bits that are easy and bits that are hard |
| 19:53 | <Hixie> | Philip`: maybe they're only reading the easy parts :-) |
| 19:57 | <Philip`> | http://tech.slashdot.org/comments.pl?sid=1291113&cid=28573027 - "Suddenly, I don't resent HTML5 anymore." - wow, somebody actually changed their mind in a discussion |
| 20:00 | <Hixie> | on /. no less |
| 20:01 | <Hixie> | :-) |
| 20:24 | <Lachy> | I've had people tell me that the spec is unreadable. |
| 20:24 | <Lachy> | I was either told that or read a comment to that effect a few days ago, but not sure where exactly |
| 20:25 | <Lachy> | IMHO, the only problem I have with the spec is that being so long, it takes forever to read |
| 20:30 | <Hixie> | i've certainly had people complain it's unreadable also |
| 20:33 | <Lachy> | Hixie, what's the point of section 1.4 History? I never noticed that before. Why was it added? |
| 20:34 | <Hixie> | someone asked for it |
| 20:34 | <Lachy> | ok, it just seems that a description of how the work started and who held the copyright for it at particular points in time isn't particularly relevant content for the spec itself |
| 20:35 | <Hixie> | *shrug* |
| 20:35 | <Hixie> | it was easier to add it than not add it, iirc |
| 20:35 | <Lachy> | well, I don't have any serious objection to it being there. I just thought it was odd |
| 20:42 | <Lachy> | Hixie, now that XHTML2 is dead, the relationship to XHTML 2.0 should be considered for removal |
| 20:42 | <Lachy> | leaving just the relationship to XForms parts of that section |
| 20:43 | <Hixie> | good point |
| 20:43 | <Lachy> | I will send mail, if you don't want to do it now |
| 20:43 | <Hixie> | i'll do it now |
| 20:43 | <Lachy> | ok |
| 20:50 | Philip` | discovers that Apple's web site lets you reserve a slot for technical support at one of their stores, with nothing more than a name and an (unverified) email address |
| 20:50 | <Philip`> | That's absurdly vulnerable to DOS attacks |
| 20:51 | <Philip`> | but it's user-friendly, and I guess people must not abuse it too much |
| 20:51 | <Lachy> | yeah, but I think it only lets you book one appointment per day with the same name and email address |
| 20:51 | <Philip`> | That's sure to stop a determined attacker who can't think up more than one fake name or fake email address in a day |
| 20:51 | <Lachy> | exactly! |
| 20:52 | <Lachy> | actually, I was thinking that it would be nice to be able to reserve numerous consecutive slots so that the staff were free enough to spend extra time with you |
| 20:54 | <Hixie> | Philip`: the twitter form on www.whatwg.org is also vulnerable to DOS attacks, but in practice DOS attacks aren't interesting to a lot of people |
| 20:54 | <Hixie> | so such features survive with occasionaly abuse |
| 20:54 | <Lachy> | I'm only aware of one organised attempt at DOSing Apple's Genius bar |
| 20:55 | <Lachy> | but I'm not sure how effective it was |
| 20:55 | <Lachy> | http://www.defectivebydesign.org/apple-challenge |
| 20:57 | <Philip`> | Hixie: That's not a DOS attack, because there's no limit to Twitterings |
| 20:58 | <Philip`> | whereas there's a very limited number of 20-minute time slots in a day, and if they're all reserved then people can't get support for their stupid buggy defective expensive equipment |
| 20:58 | <Lachy> | Hixie, would it be possible for you to remove some of those scripts running on the whatwg copy of the spec and replace them with equivalent server side processing that inserts the necessary status markers |
| 20:58 | <Lachy> | reloading the spec has become a really time consuming operation now |
| 21:01 | Hixie | stabs himself trying to open a DVD |
| 21:01 | <Lachy> | at least toc.js and status.js seem like they might work with a significant portion of their function moved server side |
| 21:01 | <Philip`> | Hmm, their "Thanks for making a reservation" email provides a handy "Get directions to the store" link which is a 404 |
| 21:01 | <Hixie> | i've seen HOUSES with less elaborate security systems than DVD boxes do these days |
| 21:01 | <Hixie> | it's ridiculous |
| 21:01 | <Philip`> | (because it doesn't include the country code in the URL, and so it'll only work in the US) |
| 21:01 | <Hixie> | can we please go to download-only media already? |
| 21:02 | <Lachy> | Hixie, is that a new form of DRM the MPAA are adding to DVDs now to prevent people even opening the case? :-) |
| 21:02 | <Philip`> | Lachy: More like ARM |
| 21:02 | <Hixie> | Lachy: if you use safari it's pretty quick actually |
| 21:02 | <Lachy> | ARM? |
| 21:02 | <Philip`> | Lachy: (Analogue) |
| 21:02 | <Hixie> | Lachy: but yes, i've been waiting for your colleague gsnedders to implement a variety of server-side versions of these features for a while now |
| 21:03 | <gsnedders> | heh. |
| 21:03 | <gsnedders> | How much of status.js do we want server side? |
| 21:03 | <Philip`> | Someone should set up a fund for implementing these features |
| 21:04 | <Philip`> | Start at $1, increase by $1 every day, and eventually somebody will crack and implement the feature for you |
| 21:04 | <Lachy> | if only gsnedders were here. I would start beating him till he stopped being lazy. I guess I will have to let jgraham and zcorpan have the honours of that while gsnedders is in sweden. |
| 21:04 | <gsnedders> | Lachy: Sadly, for your sake, zcorpan is on holiday. |
| 21:04 | <Lachy> | ah, that explains why he isn't responding to his bugmail! |
| 21:05 | <Lachy> | at least jgraham is there |
| 21:05 | <gsnedders> | His computer's turned on :P |
| 21:05 | <gsnedders> | He's away for the weekend. |
| 21:05 | <Hixie> | status.js should be mostly client-side |
| 21:05 | <Hixie> | toc.js should be server-side |
| 21:05 | <Hixie> | and the css sheet should change to be server-side too |
| 21:05 | <Hixie> | the one that inserts Note: and stuff |
| 21:05 | <gsnedders> | toc.js certainly should be server-side |
| 21:05 | <gsnedders> | Yeah, that too |
| 21:06 | <Lachy> | status.js could probably be sped up by moving part of it to the server side, such as inserting the actual boxes, and leave just parts of it as JS that can be done a little quicker |
| 21:06 | <Hixie> | how could you do it server-side? |
| 21:06 | <Hixie> | you'd have to generate it each time the page was served |
| 21:06 | <Hixie> | and that would be even slower than what we do now, given the relative speeds of html5 parsers |
| 21:07 | <Philip`> | HTML5 parser speed isn't important - just parse once per revision, then cache it as XML |
| 21:07 | <Lachy> | you check in so frequently anyway, and the statuses change so infrequently, that a relatively small delay between changes to the statuses and when they appear probably won't matter too much |
| 21:08 | <Lachy> | that way, they could just be updated once at checkin time when the spec in generated |
| 21:08 | <gsnedders> | Lachy: I tried reproducing the crashing bug, BTW, and failed. |
| 21:08 | <Philip`> | It'd be nice if definition cross-reeferences were computed on the server, so that they could still work in the multipage version |
| 21:08 | <Philip`> | s/ee/e/ |
| 21:08 | <Lachy> | gsnedders, there's a bug in BTS about the spec crashing Opera |
| 21:08 | <gsnedders> | Lachy: I saw. |
| 21:08 | <Lachy> | I filed it, but had difficulty reproducing it too |
| 21:08 | <gsnedders> | Lachy: I saw. |
| 21:09 | <Hixie> | Lachy: i think that it would be very confusing if, say, we had a wg meeting where we were going through and adding annotations, and then we had to reload, and the annotations were gone. |
| 21:09 | <Lachy> | ok, fair enough |
| 21:09 | <Hixie> | oh yeah if we could do dfn.js server-side, that would rock too |
| 21:09 | <gsnedders> | Lachy: That's why I'm telling you :P |
| 21:09 | <gsnedders> | Hixie: Yeah, you asked for that. |
| 21:09 | <gsnedders> | Hixie: I took a brief look at that when I got bored once. |
| 21:10 | <Lachy> | I wonder if there's a way it could be optimised in some way |
| 21:10 | <gsnedders> | Hixie: Problem is getting section titles. |
| 21:10 | <gsnedders> | Hixie: Doing this without massively changing Anolis and without large perf. regressions is hard. |
| 21:10 | <Hixie> | also i'd like to create a static version of the author (no .impl) version |
| 21:10 | <gsnedders> | Hixie: Just no .impl? I guess that's quite easy. |
| 21:11 | <Lachy> | Hixie, why is the current stylesheet approach ineffective? |
| 21:11 | <Hixie> | gsnedders: what i'd really like is a script that takes html on stdin, removes all nodes that match certain filters (like .impl), and outputs the result on stdout |
| 21:11 | <gsnedders> | Hixie: CSS selectors as filters? |
| 21:11 | <Hixie> | gsnedders: that would be ideal but is far beyond what would be necessary |
| 21:12 | <Lachy> | gsnedders, how many anolis feature requests from me do you still have in your queue? |
| 21:12 | <Hixie> | Lachy: it's not optimal to be shipping 2MB down for 1MB's worth of rendering, or whatever it is |
| 21:12 | <gsnedders> | Lachy: Dunno. I may have lost metadata as to who things are from. |
| 21:12 | <Hixie> | Lachy: and it means you have to have JS |
| 21:12 | <Hixie> | well anyway, i'm happy to do change all these things (except making status things disappear as noted above) |
| 21:13 | <gsnedders> | Lachy: biblio, and inclusion (which I was basically going to reply to and say no) |
| 21:13 | <Hixie> | it should be noted that in safari, this is still pretty fast and snappy |
| 21:13 | <Hixie> | so if you're seeing perf issues, file browser bugs |
| 21:13 | <Lachy> | damn, that inclusion thing is somthing I really want |
| 21:13 | <Lachy> | didn't we already get biblio implemented? |
| 21:13 | <gsnedders> | Lachy: Yes, I just need to merge this. |
| 21:13 | <Lachy> | ok |
| 21:13 | <Hixie> | (stick &profile=1 into the url to see timings btw, which might help track down what is slow) |
| 21:14 | <Hixie> | i'm now going afk to see that movie whose dvd i was bitching about earlier. |
| 21:14 | <Hixie> | bbl |
| 21:14 | Lachy | wonders what DVD it is |
| 21:14 | <gsnedders> | Hixie: I can do xpath or limited CSS |
| 21:14 | Lachy | guess it's a Pixar film |
| 21:15 | <Lachy> | Hixie, load: 14065ms; toc.js: 56ms; styler.js: 2ms; updater.js: 19ms; dfn.js: 15068ms to do 22280 links; status.js: 27187ms |
| 21:19 | <gsnedders> | OK, so vague plan: |
| 21:19 | <Lachy> | gsnedders, without that inclusion feature, how should I manage combining the auto-generated sections with the manually written sections? |
| 21:19 | <gsnedders> | This weekend, try and move WordPress's WXR/RSS importer over to using an XML parser instead of regex. |
| 21:19 | <gsnedders> | Weekend after, merge in biblio to Anolis. |
| 21:19 | <gsnedders> | Weekend after, take an attempt at dfn. |
| 21:20 | <gsnedders> | Lachy: The problem with that is execution order is undefined in Anolis. |
| 21:20 | <gsnedders> | Lachy: So I could do includes after TOC, which would make your doc fail. |
| 21:21 | <Lachy> | ok, so I would have to write a separate pre-processor from which I then pipe the output into anolis to get the final version |
| 21:21 | <gsnedders> | Lachy: yeah |
| 21:21 | <Philip`> | gsnedders: You could just define the execution order |
| 21:21 | <Lachy> | ok |
| 21:22 | Philip` | notes that if there's lots of tools and piping, it really would be a good idea to support XML import/export in all of them, for huge efficiency gains |
| 21:22 | <gsnedders> | Philip`: Indeed. |
| 21:22 | <Lachy> | Philip`, what is special about XML import/export? |
| 21:23 | <Philip`> | Lachy: It's a zillion times faster than HTML5 import/export |
| 21:23 | <gsnedders> | Lachy: it's just XML parsing/serializer is cheaper in Python, by a long way. |
| 21:23 | <Lachy> | is that because html5lib is just really slow? |
| 21:23 | <gsnedders> | Lachy: Yes |
| 21:23 | Philip` | should make the spec splitter accept XML, and have Anolis output XML, and then it'd save tens of CPU seconds on each spec change |
| 21:23 | <Lachy> | and because Python's built in XML parser is actually written in C or C++ or something? |
| 21:23 | <gsnedders> | Lachy: C |
| 21:24 | <Lachy> | ok, that's fair enough. So html5lib doesn't just suck, it just can't compete with C for speed |
| 21:24 | <Philip`> | Hmm, actually, I think the spec splitter already uses lxml's HTML mode for input, which is close enough to XML speed |
| 21:24 | <Lachy> | has anyone ported html5lib to C yet? |
| 21:25 | <gsnedders> | Philip`: Hixie uses lxml's HTML parser for Anolis (and html5lib serializeR) |
| 21:25 | <Philip`> | python $SPLITTERDIR/spec-splitter.py --html5lib-serialiser index multipage |
| 21:25 | <ezyang> | Lachy: No ;-) |
| 21:25 | <Lachy> | I didn't realise it was possible to force anolis to use lxml's parser. Doesn't it default to html5lib though? |
| 21:26 | <Philip`> | gsnedders: Looks like I do the same |
| 21:26 | <gsnedders> | Lachy: Yes |
| 21:26 | Philip` | defaults to lxml |
| 21:26 | <Lachy> | ok. I'm not too concerned about speed though. I don't generate nearly as often as Hixie, and my spec is smaller |
| 21:26 | <Philip`> | The HTML5 serialiser is still a bit slow, though |
| 21:26 | <Philip`> | (but I think the lxml serialiser had correctness problems) |
| 21:27 | <gsnedders> | Philip`: Yes |
| 21:34 | <takkaria> | evening |
| 21:35 | <gsnedders> | takkaria: hi, how are you? |
| 21:35 | <takkaria> | I'm good |
| 21:35 | <takkaria> | I went for a ride to IKEA in the end |
| 21:35 | <takkaria> | quite a nice view from just near there |
| 21:35 | <gsnedders> | Ah |
| 21:36 | <Philip`> | "Long ago in days of yore / It all began with a god named Thor / There were Vikings and boats / And some plans for a furniture store" |
| 21:39 | <Lachy> | The Vikings used to purchase their Viking ships in kit form from IKEA. |
| 21:40 | <takkaria> | it makes sense |
| 21:40 | <gsnedders> | takkaria: Where's http://www.flickr.com/photos/takkaria/3669356238/? |
| 21:45 | <takkaria> | gsnedders: at the side of the canal just near Hemköp |
| 21:47 | <gsnedders> | takkaria: ah |
| 23:09 | <Hixie> | Lachy: Primer |
| 23:35 | <roc> | it's easy to make something a derived work of something else |
| 23:41 | <Viper550> | Wait, What working group |
| 23:42 | <Philip`> | Viper550: I don't know what working group |
| 23:56 | <ndim> | From "10.2 Parsing XHTML documents": |
| 23:56 | <ndim> | Certain algorithms in this specification spoon-feed the parser characters one string at a time. In such cases, the XML parser must act as it would have if faced with a single string consisting of the concatenation of all those characters. |
| 23:56 | <ndim> | Does "spoon-feeding characters one string at a time" make any sense? |
| 23:57 | <Viper550> | so for say a site like this http://mnsi.net/~playdium (I designed it though), would HTML5 provide any advantages? |
| 23:57 | <ndim> | Spoon-feeding strings one character at a time, I find an obvious remark on XML parsers. |
| 23:57 | <ndim> | But not the other way round |