2008-09-01 [17:11:00.0000] "First of all, it's not like this code needs to be maintained *once*, it needs maintenance every time the set of void elements change. I would call this a stupid design of the target language." [17:11:01.0000] we did learn about particle and anti particles in undergrad [17:11:02.0000] I'm tempted to reassign the "stupid design" tag to whoever decided it was a good idea to output HTML with XSLT :) [17:12:00.0000] can't recall if the hawking radiation was part of the discussion though [17:12:01.0000] anyway, bbl [17:14:00.0000] jgraham, what is it about particles and antiparticles that is a lie? [17:15:00.0000] Is it just the fact that the antiparticle is also a particle, just with opposite charge (or something like that)? [17:16:00.0000] Dashiva, yeah, I don't really get why we need to fix HTML5 while they really want to fix IE... [17:17:00.0000] jgraham, I'm redesigning the test file table to see if our views on that making it better are true or not [17:18:00.0000] BenMillard, did you get that table I emailed you? Have you had a chance to investigate it yet? [17:19:00.0000] National Enquirer: Dave Hyatt spotted in Memphis supermarket [17:20:00.0000] Lachy: it's not that it's a lie exactly ... the truth is a bunch of mathematics. Then to explain it to people, the physicists have to say something which is an English approximation to the mathematics. And the approximation is necessarily rough. [17:20:01.0000] Lachy, checking inbox now [17:20:02.0000] Lachy, oh you mean the search result table? Yeah, I got it. [17:20:03.0000] Dashiva, :p [17:22:00.0000] Lachy, I filed in my "Research" folder for now. [17:25:00.0000] Lachy, summary="Post query results" stood out as being text which would be better in or just omitted, since there's a preceeding heading which says "Displaying 1 to 50 of 854 reports" [17:27:00.0000] how do the various header algorithms deal with the dual-row items, each grouped by its own tbody? [17:28:00.0000] Lachy, some header cells aren't using , but if they were then parts of it would work [17:28:01.0000] Lachy: Nothing. I meant as an explaination for Hawking radiation [17:29:00.0000] in a way, it seemed much like a demo table I presented once before with a similar dual-tr structure. http://lachy.id.au/dev/2007/table-headers.html [17:30:00.0000] Lachy, that table you just linked to is badly designed in several respects, imho. [17:30:01.0000] it uses 2 levels of column headers but they don't make sense stacked up like that [17:32:00.0000] the header cells and each row of data can be flatted to 1 row each, making it into a regular table that works without with plain [17:32:01.0000] Title Last Name First Name Middle Name Date of Birth Mailing Address City Zip Code [17:33:00.0000] erm, "Title; Last Name; First Name; Middle Name; Date of Birth; Mailing Address; City; Zip Code" [17:34:00.0000] an interesting thing in the query table, which the "smart colspan" step handles, is what I call "sectional headers" [17:35:00.0000] roc: Perhaps for physicists to explain it to themselves, they have to say something which is a mathematical approximation to the reality, and that approximation is necessarily rough [17:35:01.0000] these are cells which colspan the entire width of the table, creating implicit sections but where the main headers at the start of the table must still penetrate [17:35:02.0000] Lachy, for example (if it were using ): 29 Aug 2008 [17:35:03.0000] that's not how it works [17:37:00.0000] Lachy, much like the table you linked to, the search query table is packing information into cells where it isn't logical to put it [17:38:00.0000] Lachy, refreshing to see it using

for pagination links, though :) It follows a common design pattern for their order and delinks current location, which is all cool. [17:40:00.0000] Lachy, it uses to avoid lots of presentational markup. neat [17:41:00.0000] Lachy, it contains unlabelled checkboxes (sad and much more common than labelled checkboxes for things like this, from what I've seen) [17:42:00.0000] Lachy, the alt text is sensible and useful on the whole. :) [17:42:01.0000] Lachy, :( [17:49:00.0000] BenMillard, that table I linked to was actually designed by the designer of a project I was working on, cause it needed to have all that information fit into a narrow space that wouldn't handle having it all in a single row [17:52:00.0000] I didn't notice the target=_blank. Since I have support for that disabled in my browser, it never really bothers me in practice, unless the site actually depends on the new window opening for some functionality [17:55:00.0000] Lachy, I set mine the same way. :) [17:56:00.0000] Needing to cramming too much data into too little width is a fairly common problem...I often come across it in my day job as a website developer. [17:56:01.0000] my favoured solution is to cut the data down so only the most useful stuff is shown, then it usually fits [17:57:00.0000] (sometimes linking each entry to a page where all the rest of the data is available) [17:57:01.0000] yeah, that's what I try to do. But when I'm not in control of the content, and I just have to implement what the client wants, there's little I can do [17:57:02.0000] Lachy, that's also an experience I share. :( [17:58:00.0000] right, time for dinner...probably return in ~1 hour [18:00:00.0000] hmm, it seems that of all the media players I have, trying to find one that supports playing MKV that contains 2 1080p VC-1 video tracks and 5 AC3 audio tracks, and has UI for selecting which of those tracks to play, is a lot harder than I thought it would be :-( [18:00:01.0000] Even VLC? [18:00:02.0000] VLC doesn't support VC-1 [18:00:03.0000] yet [18:01:00.0000] QuickTime might work, but with the Perian plugin, it takes forever to finish loading MKV videos, and while it's loading, it's having trouble playing the HD [18:04:00.0000] I suppose I should just remux them into separate files, one for each video track. Though it's a full quality blu-ray rip and I wanted to have it all self contained in one. [18:05:00.0000] Zoom player seems to work on windows, but it has some bugs [18:06:00.0000] I haven't had much luck with MKV in general. It's probably user error, but I tried it out a couple times about 6 months apart with 3 different media players and an assortment of MKV files. My playback was always choppy and the audio was out of sync. [18:07:00.0000] You could buy a Blu-Ray player and the movie disc instead :-) [18:08:00.0000] Philip`, you're assuming I actually have the disc :-) [18:08:01.0000] Indeed. :) [18:08:02.0000] Lachy: That's why I said "buy" :-p [18:09:00.0000] oh, ok. I thought you just said to buy the player [18:09:01.0000] I used the word "and", to indicate that I meant both :-) [18:10:00.0000] I'll get a blu-ray player from Australia when a) I get confirmation that the manufacturer is legally required to sell in region free (or at least provide RPC-1 firmware on request), and b) blu-ray drives are available for Mac, and c) I can afford to replace my current Macs [18:11:00.0000] So in the meantime, you deserve to get everything for free? :-) [18:11:01.0000] I'm fairly sure (a) is true, because of the Australian trade practices act and a nice court case that ruled it illegal for DVDs [18:13:00.0000] no, I pay $20 a month for my usenet subscription and even more for my internet connection. It's just that my money isn't going to the movie studios until they start showing that they want to treat their customers fairly [18:13:01.0000] I would buy more movies legally from iTunes instead, but they're really low quality compared to what I can get elsewhere [18:14:00.0000] and there's limited selection [18:15:00.0000] The quality is generally good enough for me on Itunes, and so if I can rent it there I do, but most of their new releases are purchase only, which then sends me to usenet, and they lose out on my rental. [18:15:01.0000] annevk: First comment is moderated [18:15:02.0000] annevk: Approved [18:16:00.0000] GregHouston, I refuse to rent from iTunes because it's not possible to strip the DRM from those, whereas its easy to do from purchases [18:16:01.0000] Why is it unfair that they are offering their product (which you clearly want to consume) in various media (DVD, Blu-Ray, iTunes, etc) with various quality and limitations and price? [18:17:00.0000] Philip`, the DRM is unacceptable, especially the region encoding if it's done in the firmware instead of just software. [18:17:01.0000] are people here using alt.binaries? [18:17:02.0000] :-) [18:17:03.0000] jcranmer, yes [18:17:04.0000] Lachy: I can see that. It's not much of an issue for me, because I watch everything on the same computer, don't own a tv, and most movies I will never watch a second time. [22:32:00.0000] I've sent an e-mail requesting peer review of a follow-up to my last headers+id reply. Off to sleep now. I'll check the logs and my inbox when I return, just in case. :) [23:28:00.0000] http://lists.w3.org/Archives/Public/www-archive/2008Sep/0000.html is interesting [23:29:00.0000] if you look at it more closely: [23:30:00.0000] Sent Recvd Ratio Subjects Sent/Subj Author [23:30:01.0000] 136 109 0.80 89 1.53 Ian Hickson [23:30:02.0000] 66 57 0.86 30 2.20 Julian Reschke [23:30:03.0000] 53 62 1.17 27 1.96 Henri Sivonen [23:30:04.0000] 56 36 0.64 13 4.31 Boris Zbarsky [23:30:05.0000] 39 36 0.92 14 2.79 Philip TAYLOR [23:30:06.0000] 45 26 0.58 14 3.21 Leif Halvard Silli [23:30:07.0000] (these numbers are sorted on the total sent+recvd, not just sent) [23:31:00.0000] note the Subjects line :-) [23:50:00.0000] there, sent a reply with my version of the stats :-) [00:20:00.0000] Hixie: interesting development in ratios since http://junkyard.damowmow.com/291 [00:21:00.0000] people ignoring me more, you mean? :-) [00:23:00.0000] Hixie: no, the people whose ratio is > 1.0 [00:26:00.0000] heh [01:21:00.0000] So YouTube conflates accessibility captions and translation subtitles under a "CC" menu [01:22:00.0000] stats for all time http://junkyard.damowmow.com/342 [01:23:00.0000] if anyone has the complete whatwg archives in mbx format i can do the same for whatwg [01:24:00.0000] I thought lists.whatwg.org had the complete archives. what's missing? [01:24:01.0000] the mbx format part [01:24:02.0000] afaik [01:24:03.0000] ah [01:28:00.0000] hmm. RB has sent more email than I have... [01:29:00.0000] oh hey [01:29:01.0000] the text files in the whatwg archives ARE mbx format [01:45:00.0000] http://junkyard.damowmow.com/343 [01:45:01.0000] http://junkyard.damowmow.com/342 is public-html from the start [01:45:02.0000] http://junkyard.damowmow.com/343 is whatwg from the start [01:46:00.0000] whatwg has had over 700 people contribute ideas, wow [01:46:01.0000] hmm. I've sent a lot of email :-( [01:46:02.0000] Hey, I made the whatwg list twice :-) [01:47:00.0000] hsivonen: for a while, you had the #1 spot in terms of outstanding e-mails on my list [01:47:01.0000] i think you're #2 now [01:47:02.0000] Hixie: who's #1 now? [01:47:03.0000] Hixie, if you combine my @lachy.id.au and @iinet.net.au emails into one, how do I score? [01:49:00.0000] let's see [01:50:00.0000] hmm. Juan R's Sent/Subj... [01:51:00.0000] a high sent/subj means one of three things as far as i can tell: [01:51:01.0000] 1. you aren't explaining yourself well [01:52:00.0000] 2. you're repeating yourself [01:52:01.0000] 3. i'm responding to each e-mail individually instead of doing one mass e-mail [01:52:02.0000] (#3 probably only applies to me :-P ) [01:53:00.0000] ok reload 343 [01:55:00.0000] so in about a third of the time, public-html received about the same amount of mail [01:55:01.0000] that's insane [01:57:00.0000] Hixie, both lists have been high volume recently. I haven't been able to keep up with them. [01:57:01.0000] the whatwg list got to about 500 unread mails in my folder, mostly thanks to that RDF thread :-( [01:57:02.0000] yeah i tried to damp down the rdfa discussion when it got too verbose [01:58:00.0000] I suppose I should try to read them one day, but I mostly ignored it since I don't like RDF [01:59:00.0000] well the issue isn't really rdf [01:59:01.0000] Hixie: what's the issue in your view? [02:00:00.0000] the issue is how do we address the needs of the communities that RDFa is addressing, and whether doing so is appropriate given our overall goals [02:00:01.0000] i need to look at some of manu's and ben's e-mails more closely to determine the answer to both [02:00:02.0000] but that will have to wait until after wf2 [02:01:00.0000] (underlying both of those points is "what are the needs of the communities that RDFa is addressing") [02:05:00.0000] to the extent the needs include compatibility with the RDF data model, it's about RDF [02:08:00.0000] /me suspects the RDF data model is too hard for mass adoption [02:17:00.0000] hsivonen: i don't think that's actually been listed as a need [02:20:00.0000] AFAICS, everything beyond rel=license is totally unnecessary verbosity when it comes to expressing copyright info. [02:22:00.0000] i'm not even convinced rel=license is needed [02:22:01.0000] it's only needed if the copyright info is going to actually be used by computers [02:22:02.0000] and i'm not convinced that it is [02:27:00.0000] Hixie: one of the reasons why I chose Flickr was that Flickr is the place people go to when they want to search for CC-licensed photos [02:27:01.0000] Hixie: they don't go to Google image search [02:28:00.0000] there is value in being able to search by license [02:28:01.0000] hsivonen: Would it work well in an open system? [02:29:00.0000] It doesn't work all that well on flickr [02:29:01.0000] jgraham: it would not work if people can make licensing claims cross-Origin about other people's works [02:29:02.0000] jgraham: but it could work it the licensing info is near the data object [02:29:03.0000] preferably inside it [02:30:00.0000] jgraham: what sucks on Flickr is that people don't know what they are doing when the flip the defaults [02:31:00.0000] hsivonen: Inside might work technically but I think user-invisible metadata outside the object would almost never be right [02:32:00.0000] jgraham: also, what sucks is that people take photos of people, CC doesn't deal with model release, and the photographer can't unilaterally concoct model release [02:32:01.0000] hsivonen: People not knowing waht they are doing would be even worse without a consistent UI that explained everything, right? [02:32:02.0000] jgraham: could be [02:32:03.0000] I don't think Flickr's UI explains everything, though [02:32:04.0000] hsivonen: AFAIK Model release laws vary wildly by country [02:34:00.0000] Flickr says "You can choose to use a Creative Commons license to allow more liberal use and sharing of your photos or video while still maintaining reasonable copyright protection." [02:34:01.0000] one could argue that "reasonable copyright protection" is not a good UI term [02:36:00.0000] hsivonen: Yeah, the explaination is worse than I remembered [02:36:01.0000] Does flickr default to a CC licence or ordinary copyright? [02:36:02.0000] Lachy: defaults to ARR [02:36:03.0000] ok [02:36:04.0000] Lachy: A CC default would be disaterous [02:37:00.0000] Lachy: Lots of photog. types are seriously paranoid about any copying [02:37:01.0000] jgraham, yeah. But that means that people are somehow being coerced into changing to a CC licence, without actually understanding the concequenses of their actions [02:38:00.0000] e.g. http://duncandavidson.com/2008/04/the-copyright-conspiracy.html [02:38:01.0000] Lachy: CC and FSF have very different marketing [02:39:00.0000] Lachy: In what sense coerced? They are given several options but none are presented as "better" [02:39:01.0000] Hixie, did you add fora⊙an to annevk⊙oc? [02:39:02.0000] (for the WHATWG list) [02:39:03.0000] annevk: looks like it [02:42:00.0000] hsivonen: google has cc search that is directly based on the cc metadata; what is your opinion of it? [02:42:01.0000] Hixie: too well hidden [02:42:02.0000] Hixie: 3095, Anne: 904, Henri: 590 [02:42:03.0000] I should e-mail less :) [02:43:00.0000] whoa, Matthew and Jim Ley make it to the top five [02:43:01.0000] annevk: the script automatically merges names that are the same [02:43:02.0000] jgraham, coerced is probably the wrong word. I probably should have said tempted or something, because there are people with CC-licenced stuff on flickr who still complain when their stuff gets reused [02:43:03.0000] Hixie: also, not available in image search [02:43:04.0000] annevk: note that the script does things based on time active, so unless jim posts again, he'll stay high [02:44:00.0000] hsivonen: well there's no way with rel=license to label an image [02:44:01.0000] hsivonen: so if that's your use case, rel=license won't help google [02:44:02.0000] Hixie: yeah. that's a problem. [02:44:03.0000] that copyright conspiracy article is interesting. But in which countries does any form of copyright registration apply? Is that just a US thing, or more widespread? [02:44:04.0000] /me shrugs [02:44:05.0000] if you say so [02:44:06.0000] :-) [02:44:07.0000] Hixie: I want a way to have a license field in EXIF/JFIF/PNG [02:44:08.0000] Hixie: and then I want tool support for it [02:45:00.0000] hsivonen: ah well that's not my problem, luckily [02:45:01.0000] Hixie: It annoys me the CC isn't standardizing embedded license metadata for EXIF/JFIF/PNG/Ogg/MP4 [02:45:02.0000] and PDF in a simple way [02:46:00.0000] Lachy: it's a U.S. thing [02:47:00.0000] You can register copyright in the UK [02:48:00.0000] Dunno how it compares to the US [02:48:01.0000] ok, that's what I thought. But I find it annoying that whenever I read anything related to copyright, it's usually very much US-centric [02:48:02.0000] Lachy: you may safely omit "related to copyright" [02:48:03.0000] the US system gets exported everywhere else eventually [02:49:00.0000] so I basically have no idea what the copyright situation is in Australia, although I do know that we also have excessively long copyright terms [02:49:01.0000] say hello to WIPO [02:49:02.0000] roc: It would be great if the U.S. system had been exported without the Victor Hugo input [02:49:03.0000] I think it's life-of-author+70 years, which is absolutely absurd [02:49:04.0000] Lachy: that stuff came from Europe [02:49:05.0000] who's Victor Hugo? [02:50:00.0000] the disaster is taking the worst parts of French and American copyright things and mixing them [02:50:01.0000] I wish we could go back to 14 year copyright terms [02:51:00.0000] Lachy: a famous writer who was so annoyed at Americans pirating his books that he wrote what lead to the creation of the Berne Convention [03:16:00.0000] Hixie: I think searching images by license is a more common use case than searching text by license, since images have different remixability characteristics [03:16:01.0000] no disagreement there [03:17:00.0000] so in that sense, standardizing and implementing a license field for EXIF should be a higher priority than rel license [03:17:01.0000] agreed [03:17:02.0000] EXIF isn't my problem though [03:17:03.0000] :-) [03:24:00.0000] wow, Microsoft has just patented the page up/down keys http://news.zdnet.com/2424-9595_22-218626.html [03:25:00.0000] i try and avoid reading about patents [03:25:01.0000] companies become three times more liable to damages in the US if it can be demonstrated that they (or their employees) had knowledge of an infringed patent [03:26:00.0000] that's actually changing [03:26:01.0000] so it's better to be ignorant [03:26:02.0000] roc: oh thank god [03:26:03.0000] roc: for the better? [03:26:04.0000] depends on your point of view, but yeah [03:26:05.0000] good to hear [03:29:00.0000] I only read about patents that get mentioned on places like /., and are so absurd that there's obviously plenty of prior art [03:30:00.0000] like the Eolas patent? [03:30:01.0000] yeah [03:30:02.0000] oh, the Patent Reform Act 2007 is stalled :-( [03:30:03.0000] lachy: being absurd and having plenty of prior art doesn't seem to matter [03:30:04.0000] roc: :-( [03:31:00.0000] for some reason, people like the concept of Trial by Jury despite cases like Eolas [03:31:01.0000] and people like spec design by consensus despite cases like XLink [03:32:00.0000] some people even like XLink [03:33:00.0000] /me notes that the WHATWG is closer to arguing things to a professional judge and a panel of semi-pro lay judges than to Trial by Jury [03:33:01.0000] yup [03:34:00.0000] hsivonen, what about public-html? [03:34:01.0000] Lachy: They're not patenting the keys at all - they're patenting a method of scrolling (which jumps in a one-page increment and preserves the vertical offset of your view of your page, regardless of zoom, etc), which might be triggered by those keys [03:34:02.0000] (using the term professional to mean literally "paid", as opposed to "competent") [03:34:03.0000] Lachy: no comment [03:34:04.0000] Philip`, yeah, I realised that. But I just summarised it poorly [03:35:00.0000] (using semi-pro to refer to the Finnish trial system--not the proness of the WHATWG panelists) [03:36:00.0000] Lachy: ZDNet summarised it equally poorly/inflammatorily :-) [03:36:01.0000] Philip`, in any case, that method of using pg up/dn has been used for decades [03:37:00.0000] /me can't easily remember any programs that scroll like that [03:41:00.0000] Philip`, web browsers, word processors, text editors, etc. [03:41:01.0000] they all scroll like that [03:41:02.0000] Linux man pages even [03:41:03.0000] Web browsers and text editor don't usually split the document into pages [03:42:00.0000] Word processors scroll by roughly the viewport height (which depends on your zoom setting), not by a single page [03:43:00.0000] in the case of web browsers, since a web page is a continuous medium, a "page" could be considered equivalent to the content of the viewport [03:43:01.0000] no matter how high it is [03:43:02.0000] If you're looking at just the top of a page, and press page-down, you jump to the bottom of page, whereas the patent is talking about jumping to the top of the next page [03:44:00.0000] Philip`: @media projection [03:44:01.0000] Philip`, Print Preview [03:44:02.0000] Lachy: Uh, you can't just consider things to be things that they aren't, and say it's prior art :-p [03:45:00.0000] Philip`, the patent is too broad [03:45:01.0000] /me tries print preview in Opera and Firefox, and in both cases it scrolls by the zoom-dependent viewport height and not by a single page [03:45:02.0000] Philip`: and virtually any other slideshow tool [03:46:00.0000] Any PDF viewer whose zoom is set to "Fit to window" or similar [03:46:01.0000] man [03:46:02.0000] Lachy: the patent is a software patent, it's a given that it's not a good idea. but lots of companies have software patents, microsoft is no exception. [03:46:03.0000] virtuelv: apple's pdf viewer doesn't do what this patent says, i just checked [03:48:00.0000] Hixie: evince does, if set to "Best fit" [03:48:01.0000] Acrobat reader has a similar means [03:49:00.0000] even Notepad scrolls like that (though, I'm aware its an MS program, it's been around for years, and to patent the process now is just absurd [03:49:01.0000] virtuelv: That only works because the viewport height happens to be the height of a single page [03:49:02.0000] Lachy: Notepad doesn't have pages [03:50:00.0000] Philip`, define what a page is? [03:51:00.0000] Lachy: A page is like a rectangular piece of paper, but on a computer [03:51:01.0000] if you're only talking about the on-screen representation of an A4 page, then ok. But that's irrelevant, since the concept of keeping the cursor in place on screen and shifting it down by the viewport has been around forever [03:52:00.0000] The point is not to shift it down by the viewport - the point is to shift it down by a page, regardless of the viewport [03:52:01.0000] the whole point of this patent as i understand it is that it isn't about shifting by the viewport but shifting by the page height [03:52:02.0000] and i think virtuelv is right that acrobat does that, i remember it driving me crazy [03:54:00.0000] either way, patents are just Evil [03:54:01.0000] patents as originally defined make some sense [03:55:00.0000] Hixie: yes, but the problem is that the patent system does not scale [03:55:01.0000] if you come up with an idea for a physical product, it makes sense to give you some protection to allow you to sell it before some big multinational can use your idea and destroy your margins [03:55:02.0000] and the threshold for getting a patent is ridiculously low these days [03:55:03.0000] but i agree that at least for software it makes no sense [03:56:00.0000] like that guy getting a patent for a special case of multiply-linked lists [03:56:01.0000] because the economics just aren't the same for software [03:56:02.0000] patents on methods of construction or on algorithms, etc, also make no sense imho [03:58:00.0000] a friend of mine with a PhD in control systems theory says most of the patents she encounters are also complete bullshit [03:58:01.0000] and obvious to anyone skilled in the trade [03:58:02.0000] What about e.g. the MP3 algorithm? That presumably took substantial effort to develop, and seemed far in advance of any alternatives at that time, and the patent was specific enough that it didn't prevent development of similar competitors like Vorbis [03:58:03.0000] (mostly for industrial processes, AIUI) [03:59:00.0000] Philip`: discovering the general theory of relativity also took substantial time and effort, and Einstein's theories were specific enough [03:59:01.0000] should he have been granted a patent on it? [04:00:00.0000] algorithms are discovered, not invented [04:00:01.0000] algorithms are invented [04:00:02.0000] and if the mp3 patent was awarded, then a patent on the algorithm for predicting time dilation should be too [04:01:00.0000] and probably would be [04:02:00.0000] Philip`: i could see an argument for that, if it was limited to a very small time frame (low single digit years), and was limited to people actually using the technology to make money [04:02:01.0000] Philip`: it shouldn't prevent me from writing an implementations and sharing it with people online [04:02:02.0000] Hixie, wouldn't that just be a mathematical equation, which can't be patented? [04:03:00.0000] Lachy: all software is "a mathematical equation" in the same sense [04:04:00.0000] virtuelv: There's only one theory of relativity, and even if it was phrased differently you'd get the same results, so that theory just needed to be discovered; but there are millions of ways you can encode audio, each slightly different, and the problem is choosing the right set of tradeoffs, so that's much more like invention than like discovery of a pre-existing thing [04:04:01.0000] there's not just one theory of relativity [04:04:02.0000] the whole point of how science works is that there are many theories on everything [04:04:03.0000] we test them to see which one is closest to reality each time [04:04:04.0000] and then keep improvign it [04:05:00.0000] /me tries to work out how to phrase it better [04:06:00.0000] coming up with an algorithm to predict how the universe will act is the same thing as writing an algorithm to compress music [04:06:01.0000] it just does something different [04:07:00.0000] no, I think there's a difference between science and engineering [04:08:00.0000] I think I'm trying to say something like how Newton's theory of gravity and Einstein's theory of gravity are substantially different, and there aren't a million theories in between that they could have just as easily decided on but that might not have worked quite so well in practice [04:09:00.0000] roc: what's the difference? [04:09:01.0000] Einstein didn't do some experiments to decide whether E=mc^2 or E=mc^2.01 or E=0.01+mc^2 - there's just one way that makes sense [04:10:00.0000] there's "just one that makes sense" in the same way that there is "just one" WebSocket spec definition that makes sense [04:10:01.0000] lol, instead of volunteering to do something, TAYLOR starts debating Hixie's list, typical [04:11:00.0000] well my list had the desired effect [04:11:01.0000] someone (i won't say who in case they don't actually want to do it!) contacted me and volunteered to work on something maybe :-) [04:11:02.0000] The effect of demonstrating that people are more willing to complain about editors than to help as editors? :-) [04:12:00.0000] no, getting someone to volunteer [04:12:01.0000] at least if they're complaining about the list they aren't complaining about the spec, though [04:13:00.0000] happy mailman mailing list membership reminder day, btw [04:14:00.0000] haha pt(w)'s e-mail is funny [04:14:01.0000] let's see... i've written four specs like html5... he has written none... but hell, maybe he does know better after all [04:17:00.0000] I like how he challenged some of the requirements, without actually making a case against them [04:17:01.0000] my favourite has to be the last one: [04:17:02.0000] "a pragmatic attitude that is willing to put the needs of the users, authors, and implementors (in that order) far ahead of technical purity" [04:17:03.0000] ...which he challenges [04:17:04.0000] ironically, he partially challenged the "the ability to defend a decision" one without actually defending his :-) [04:18:00.0000] well it makes sense that he wouldn't defend a decision to not defend decisions :-P [04:18:01.0000] that's not ironic that's just plain sensible :-P [04:18:02.0000] You should have added "The editor must not drown kittens" and see if anyone objected [04:19:00.0000] i didn't expect people to object! [04:19:01.0000] Well, someone fulfilling those requirements would probably produce yet another spec he'd be unhappy with. Your axioms do result in a very specific type of spec. [04:20:00.0000] Maybe his view is the editor is just meant to edit, and decisions are made by WG consensus and not by the editor, and therefore the editor does not need to defend decisions [04:22:00.0000] in engineering we're optimizing a very complex utility function and we make a lot of trade-offs [04:22:01.0000] in science the utility function is much simpler, primarily "does this theory fit the data" [04:22:02.0000] and you make very few tradeoffs [04:22:03.0000] wilhelm: that's what i led with. "a spec of the quality of html5" or some such [04:23:00.0000] Philip`: that doesn't get you a spec like html5. :-) [04:23:01.0000] roc: i'll grant you that, but i'm not sure that changes the argument at the point of "is it patentable" [04:23:02.0000] Hixie: Many people may consider that to be a good thing :-) [04:24:00.0000] Philip`: probably :-) [04:25:00.0000] /me guesses consensus works best on specs which not many people care about, which is either boring specs that they're not interested in or new specs that they haven't heard of [04:25:01.0000] i don't know of any spec where consensus has worked well [04:25:02.0000] it's hard to find such a spec, I tried [04:26:00.0000] i think the most i've found is specs where consensus didn't work worse than one poor editor would have [04:26:01.0000] and those were all very small specs, where there simply weren't that many design decisions to screw up [04:26:02.0000] Are there any specs were only one person was involved in development? That kind of spec would surely be the result of the unanimous consensus of all who were involved [04:26:03.0000] I think the idea is that in science, you're so constrained by observations that there's little creativity and so little """invention""" [04:26:04.0000] s/were/where/ [04:27:00.0000] but don't ask me to defend the theory of patents [04:27:01.0000] roc: i think that relativity is very creative. and quantum even more so. [04:27:02.0000] roc: but yeah. [04:27:03.0000] Philip`: specs are really just software written in prose, and the same applies to both, really [04:27:04.0000] Philip`: you can't write good software in a committee [04:28:00.0000] actually I think you're wrong there. The observations force you to adopt those theories, or theories equivalent to those theories [04:28:01.0000] roc: the sheer number of competing GUTs right now suggests otherwise [04:29:00.0000] they're all "relativity + QM + extensions" [04:29:01.0000] roc: science works the other way around. You come up with many theories, then you make predictions using them, and only the ones that predict reality survive. [04:29:02.0000] roc: but you still have to come up with theories, and that's creative. [04:29:03.0000] and the fact that the extensions have so many degrees of freedom is in fact very disturbing to many/most scientists [04:29:04.0000] roc: it's a bit like having a test suite, and then writing a bunch of programs, and then throwing the ones away that don't pass the test suite. [04:30:00.0000] What do you do when you find there was a bug in the test suite? [04:30:01.0000] ok, but in engineering you get to choose the test suite and the results you want from each tests [04:30:02.0000] subject to rather vague and broad requirements [04:30:03.0000] in science, you don't get to choose the results [04:31:00.0000] and you're not supposed to be choosy about the tests either [04:31:01.0000] roc: the problem is mostly that the predictions are things we can't yet test (e.g. LHC should help with that), but once we get to the stage where practical physics is testing the theoretical stuff again, they'll go away. [04:31:02.0000] Philip`: science by definition has a bug-free test suite. :-P [04:31:03.0000] roc: granted [04:32:00.0000] roc: but i still don't think that difference is, or should be, relevant for the purposes of the patent system [04:34:00.0000] concensus definitely screws things up, especially when that concensus overrules the editor's decision [04:35:00.0000] More than when the editor's decision overrules the consensus? [04:35:01.0000] just look at the selectors api naming crap. People were happy with the names I chose. Then I got overruled, and people were disappointed [04:35:02.0000] the naming thing just made different people sad [04:35:03.0000] naming decisions are a pain and really should never be something that people have to vote on [04:36:00.0000] If people were happy, how was there consensus to overrule that decision? [04:36:01.0000] Philip`, there should be no concensus. In fact, in a really large group, concensus is almost impossible [04:36:02.0000] Lachy: Perhaps the solution is to avoid really large groups, not to avoid consensus [04:36:03.0000] there wasn't concensus in the case of the selector naming at any point iirc, it was just a bunch of narrowly-won votes that kept swinging the results back and forth [04:37:00.0000] Philip`, because it was decided by a vote, with each member ranking their preferences in order. The winner wasn't the most preferred by anyone, it just happened to have the highest average vote [04:37:01.0000] (there definitely needs to be some oversight of the editor, though) [04:38:00.0000] (i just think it needs to be oversight of the form "keep us happy enough or leave", not "keep us very happy or do what we say") [04:38:01.0000] (because that both gives more motivation to the editor, and more responsibility, and thus more care, from the oversight) [04:39:00.0000] I prefer "if you're not happy, you leave!" :-) [04:39:01.0000] well an editor with no oversight has nothing to keep him in check, and that's dangerous [04:40:00.0000] because it is supremely easy to go down a path where you think you're doing the right thing and you're ignoring people saying that you're not [04:40:01.0000] /me thinks someone should fork HTML 5, to promote some competitive spirit between the editors of each fork [04:40:02.0000] even when they're right [04:40:03.0000] that way, the editor is always happy and the only people left in teh group are those who agree [04:40:04.0000] Philip`: i'm definitely up for that [04:40:05.0000] Philip`: i totally agree that that would be a good thing (seriously) [04:40:06.0000] Lachy: that's how xhtml2 came to be, btw [04:40:07.0000] Hixie, I know :-) [04:41:00.0000] so i think it's not a good thing :-) [04:42:00.0000] Hixie, forking the spec wouldn't be good if the two diverged significantly and different implementations followed different specs [04:42:01.0000] they wouldn't [04:43:00.0000] at least, my copy of it would always follow what the UAs did [04:43:01.0000] what wouldn't? The specs significantly diverging, or the implementers following different specs/ [04:43:02.0000] ? [04:43:03.0000] the implementers following different specs in a way that resulted in lack of interop [04:44:00.0000] I suppose, since they generally try to follow each other [04:45:00.0000] so if the spec was forked, and both editors tried to equally follow what the implementers did, then by the end, we'd have 2 specs saying roughly the same thing in 2 different ways [04:45:01.0000] because of human error, the specs would say two subtly but very different things, and we'd be in hell [04:46:00.0000] i don't think you'd see the browser vendors following different copies of the spec [04:47:00.0000] i also don't think you'd find two people who'd edit the spec competently enough for the two to both be credible for long [04:47:01.0000]