| 00:06 | <gsnedders> | TabAtkins_: V8 is one the engine now to have a non-deletable __proto__ and also the only with it on each object. |
| 00:07 | <gsnedders> | Everyone else who has __proto__ has it on Object.prototype as an accessor pair |
| 00:08 | <TabAtkins_> | Weird. |
| 00:11 | <gsnedders> | tl;dr, people surverted what most of TC39 wanted out of conservatism and fear of extra attack surface |
| 00:11 | <gsnedders> | And then we went, "oh, hey, look at the status quo. it appears to have change. soz guys!" |
| 00:12 | <gsnedders> | I say most of TC39, but it wasn't really that clear cut at all. |
| 00:14 | TabAtkins_ | has no real idea what the details around __proto__ mean anyway. |
| 00:26 | FireFly | thinks the <| operator proposal made more sense than __proto__ -- not really a fan of mutable [[Prototype]] |
| 00:38 | <Benvie> | having it as an accessor on the top of the prototype chain means it can be deleted and rendered inert in cases where that security is needed |
| 00:40 | <Benvie> | I've been working on an es6 VM implemented in JS that runs in IE9+ currently and just changed the way it handles __proto__ from matching how V8 does to having it as an accessor. It's also nice in that it removes all sorts of special case checking for __proto__ everywhere and puts it in one spot. |
| 00:46 | <gsnedders> | Benvie: It was always going to be deletable, it was just how magical it was to achieve that without possibly risking leaking a setter function |
| 00:47 | <Benvie> | yeah I meant to say I implemented it as (I thin kthe current concensus says) a magical accessor-as-data property |
| 00:48 | <gsnedders> | Ah, which isn't what V8 does. :) |
| 00:48 | <Benvie> | v8 does the very least protectable of all options, ha |
| 00:48 | <gsnedders> | The current consensus now is an accessor, given everybody has basically bitten the bullet now |
| 00:48 | <gsnedders> | Possibly poison get/set in Object.getPropertyDescriptor |
| 00:48 | <Benvie> | oh that simplifies it slightly from what I have now then |
| 00:49 | <gsnedders> | (Carakan currently poisons set) |
| 00:49 | <Benvie> | the thing I ran into is: Object.create(null) |
| 00:49 | <gsnedders> | (Nobody else poisons anything, which is worse for SES) |
| 00:49 | <Benvie> | is it impossible to set the __proto__ on a newly nullary object? is it ever possible to use __proto__ as getter or setter on one? |
| 00:50 | <gsnedders> | Once you reach a null prototype you can never change it again, by design. |
| 00:50 | <Benvie> | I suppose. The use case I'm thinking of is basically creating a new object graph. The first step in that is to create the equivelent of Object.prototype |
| 00:51 | <Benvie> | with __proto__ as an accessor that's possible, if you can copy the getter/setter over |
| 00:51 | <gsnedders> | (Basically, you either need to poison Object.getPropertyDescriptor or you need a context check in the setter, for the sake of SES being able to lock down all prototypes) |
| 00:51 | <gsnedders> | (I prefer the former, because otherwise you need the context for all objects) |
| 00:52 | <Benvie> | kind of relatedly, implementing the runtime in js itself necessitates the ability to the do Object.create(null), or else insert checks everywhere to try and separate the meta level with the sandbox level |
| 00:53 | <gsnedders> | Anyhow, I need to go and sleep. :) |
| 00:53 | <Benvie> | fortunately I was able to figure out how to replicate the equivelent of Object.create(null) in ie6-8 so I'll be able to introduce es6 to ie6 |
| 00:53 | <gsnedders> | How, out of interest? |
| 00:53 | <gsnedders> | More than welcome to discuss w/e sometime when I'm more awake. :) |
| 00:54 | <Benvie> | basically create a new Object.prototype in an iframe, steal it, strip it bare, use it as a prototype |
| 00:54 | <Benvie> | and then manufacture new empty objects from it |
| 00:54 | <gsnedders> | Oh wow. |
| 00:54 | <Benvie> | it's in es5-shimsham now |
| 00:54 | <gsnedders> | Does that not keep a reference to that whole iframe around though? |
| 00:54 | <gsnedders> | Which seems costly in terms of memory. |
| 00:55 | <Benvie> | I don't believe so |
| 00:55 | <Benvie> | the reference to the iframe itself is nulled immediately |
| 00:55 | <Benvie> | no reference to any object is saved except for the Object.prototype ref |
| 00:55 | <gsnedders> | All depends on how IE relates objects to browsing contexts. |
| 00:55 | <TabAtkins_> | That should still be enough to keep alive the connection to the scripting environment, which should keep alive the iframe, no? |
| 00:55 | <Benvie> | and all properties are immediately deleted from it, so no reference leaves the initialization aside from that |
| 00:55 | <Benvie> | good chance, but |
| 00:55 | <Benvie> | you only need to do it once |
| 00:56 | <gsnedders> | TabAtkins_: I believe it wouldn't in Opera, FWIW. |
| 00:56 | <Benvie> | once you have function Empty(){} Empty.prototype = stolenObjectPrototype |
| 00:56 | <gsnedders> | But I think we're unique in that way. |
| 00:56 | <Benvie> | it is unobservable that it's not prototypeless |
| 00:56 | <Benvie> | but a single step away from the top level |
| 00:57 | <Benvie> | so you can manufacture `new Empty` all day long from that single iframe |
| 00:57 | <TabAtkins_> | Right, you'd only need the one iframe, so it's not too bad. |
| 00:58 | <Benvie> | the value of Object.create(null) is immense for solving certain problems, so I'm glad that it can be counted amongst the list of solved problems |
| 01:00 | <Benvie> | I use it as the basis for hashes in property lists for the es6 vm and I also use it as the basis for each Realm's ObjectPrototype |
| 01:00 | <Benvie> | (http://benvie.github.com/continuum) |
| 01:01 | <Benvie> | seeing __proto__ in ie9 tugs at my heartstrings a little |
| 01:29 | <GPHemsley> | Blast from the past: http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2004-June/000005.html |
| 01:33 | <zewt> | so does fred andrews basically want a big complicated feature to be specced just to support his extension? heh |
| 01:33 | <TabAtkins_> | Argh jeezus, I somehow lost focus-follows-mouse while I was gone at TPAC. >_< |
| 01:33 | <TabAtkins_> | zewt: Yes. |
| 01:34 | <GPHemsley> | Interesting... http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2004-June/000010.html |
| 01:36 | <zewt> | i'll visit anyone who tries to spec ie6 (in the asylum, as they walk in a 5-foot circle chewing away at what used to be fingernails while sporadically singing show tunes at the top of their lungs) |
| 08:11 | <rniwa> | oh no... |
| 08:11 | <rniwa> | is polyglot new long desc? |
| 08:11 | <annevk> | I'm surprised you have the time to be subscribed to that list |
| 08:12 | <annevk> | also, now you made me look at the archives :p |
| 08:13 | <rniwa> | annevk: i subscribe to it as sort of an entertainment and make me appreciate what i'm doing :) |
| 08:13 | <Hixie> | i'm highly amused that the w3c is interested in publishing the spec that is making the IETF have fits |
| 08:14 | <annevk> | heh, when I attended the WebApps and HTML WG meetings last week I had much the same feeling rniwa :) |
| 08:14 | <rniwa> | annevk: LOL. |
| 08:14 | <hsivonen> | must. resist. replying. to. polyglot. thread. |
| 08:15 | <rniwa> | annevk: working on standards was my childhood dream because back then (~1998), w3c seemed to work fairly well |
| 08:15 | <rniwa> | annevk: compared to IETF/IEEE/ISO |
| 08:15 | <rniwa> | but now… all "bureaucrats" seem to have taken over w3c :( |
| 08:16 | <rniwa> | hsivonen: it's a fun game :P |
| 08:17 | <annevk> | WHATWG holds the spirit of the web high at least (openness, permissive licensing, low overhead, etc.) which is why I ended up exclusively there for now |
| 08:17 | <rniwa> | annevk: yeah… i think what's nice about WHATWG is that there aren't that many rules & policies. |
| 08:18 | <rniwa> | annevk: the problem with standards body is that it attracts people who LOVE to create new policies and rules :( |
| 08:18 | <rniwa> | annevk: and very few who that hate policies and rules |
| 08:21 | <annevk> | uhuh; I was minuting this Process Agility meeting (god knows why) and Daniel Glazman was almost screaming how Chairs would be required to attend coordination meetings and be required to send status updates in advance of those meetings and all I could do was write down the irony of what was happening |
| 08:21 | <rniwa> | :( |
| 08:22 | <rniwa> | i always get skeptical when someone stays everyone has to do X. |
| 08:22 | <rniwa> | that's when I start to move away from that person... |
| 08:22 | <rniwa> | or organization for that matter. |
| 08:23 | <annevk> | well yeah, I quit the CSS WG first after he said stuff like that one to many times |
| 08:23 | <Hixie> | Bureaucracy Is Opt-In |
| 08:23 | <annevk> | nobody else there seems to really care |
| 08:23 | <Hixie> | that's the biggest lesson i learnt at the w3c (and now ietf) |
| 08:23 | <annevk> | maybe they're better at ignoring |
| 08:23 | <annevk> | http://www.w3.org/2012/10/31-agile-minutes#item05 are those minutes btw |
| 08:25 | <Hixie> | wow that meeting doesn't look useful |
| 08:25 | <Hixie> | but good to know that it used up the time of people like yourself, dbaron, and hsivonen |
| 08:25 | <Hixie> | i cannot imagine anything that y'all could have been doing that would have helped the world more. :-P |
| 08:26 | <annevk> | heh, we had a lunch discussion later in the week with two people from the AB that felt productive, but it remains to be seen what is actually going to happen |
| 08:27 | <Ms2ger> | And email is a support forum |
| 08:28 | <annevk> | that "Call for Editor" thread had a turn for the better, did not expect that |
| 08:29 | <rniwa> | annevk: i started anything says "objection", "editor", "chair", etc... |
| 08:29 | <rniwa> | ignoring* |
| 08:29 | <rniwa> | anything tha*t says |
| 08:29 | <rniwa> | ugh… i can't type :( |
| 08:30 | <annevk> | prolly bedtime ;) |
| 08:32 | <rniwa> | annevk: i can't… i'm fighting for a great cause of making html collection sane in webkit. |
| 08:32 | <annevk> | is that the NodeList + Array bug? |
| 08:32 | <rniwa> | annevk: oh, no. |
| 08:33 | <annevk> | ah, I guess I'm not following your work then |
| 08:33 | <rniwa> | annevk: we return a static node list when there are multiple elements matching a id on html collection :( |
| 08:33 | <rniwa> | annevk: https://bugs.webkit.org/show_bug.cgi?id=101311 |
| 08:33 | <hsivonen> | It’s so annoying how Launchpad hides the bug filing function as a clue test |
| 08:34 | <annevk> | rniwa: and that behavior is web compatible? |
| 08:34 | <rniwa> | annevk: what behavior is? |
| 08:34 | <annevk> | rniwa: returning a static rather than live list |
| 08:35 | <rniwa> | annevk: i dunno. |
| 08:36 | <Ms2ger> | "Could be slightly more formal?" |
| 08:47 | <Hixie> | i was more formal for him |
| 08:47 | <Hixie> | i hope he enjoys it. |
| 08:49 | <Ms2ger> | I sure am glad that you're opting in to this thread, I cannot imagine anything that you could have been doing that would have helped the world more ;) |
| 08:50 | <Hixie> | i should be sleeping so i can work well tomorrow |
| 08:50 | <Hixie> | but let's be honest |
| 08:50 | <Hixie> | i'm not going to be worknig tomorrow |
| 08:51 | <Ms2ger> | Why not? |
| 08:51 | <Hixie> | elections |
| 08:51 | <Ms2ger> | Oh |
| 08:51 | <Ms2ger> | That's today |
| 08:51 | <Hixie> | "tomorrow" = after i sleep :-) |
| 08:52 | <Ms2ger> | You should get in line now, I hear ;) |
| 08:52 | <Hixie> | if i get in line, i'll get deported |
| 08:53 | <Ms2ger> | You're secretly an illegal alien? |
| 08:53 | <Hixie> | i'm not at all secretely a legal alien |
| 08:53 | <Hixie> | still can't vote! |
| 08:54 | <Ms2ger> | Actually, do you vote anywhere? :) |
| 08:55 | <Hixie> | not recently, though i could vote in switzerland and the UK if I put my mind to it and had any idea what the issues were |
| 08:55 | <Hixie> | anywho, bed time. nn :-) |
| 08:57 | <Ms2ger> | nn |
| 08:58 | <annevk> | oh yes |
| 08:58 | <annevk> | I did my first pull request |
| 08:58 | <annevk> | and succeeded |
| 08:58 | <annevk> | https://github.com/github/github-services/pull/438 |
| 08:58 | <annevk> | <- success kid |
| 09:03 | <MikeSmith> | whoah |
| 09:03 | <MikeSmith> | score |
| 09:03 | <MikeSmith> | pull request against the github sources |
| 09:04 | <MikeSmith> | they should give you a beer mug for that |
| 09:48 | <karlcow> | [03:15] <hsivonen> must. resist. replying. to. polyglot. thread. |
| 09:48 | <karlcow> | do. not. please. :) |
| 09:48 | <karlcow> | hmmm not clear |
| 09:48 | <karlcow> | do. not. reply. resist. |
| 09:49 | <karlcow> | still not clear. :) |
| 09:49 | <karlcow> | baaaaah. punctuation is important |
| 09:49 | <karlcow> | Do not reply. Resist. |
| 09:50 | <hsivonen> | karlcow: I have successfully avoided replying today. |
| 09:50 | <karlcow> | \o/ happiness. |
| 09:50 | <hsivonen> | I’m guilty of only two out of 34 polyglot emails starting with TPAC |
| 09:52 | <jgraham> | The new rule should be that you can only reply to the polyglot thread if you can compose a message that makes sense in two langauges |
| 09:52 | <MikeSmith> | hahaha |
| 09:52 | <MikeSmith> | bravo |
| 09:54 | <asmodai> | What's the best way forward to get HTML5/CSS3 to work with older browsers? Using Modernizr or some JS library that injects the HTML5 elements? |
| 09:54 | <MikeSmith> | jgraham: would someone hypothetically get points for composing multiple messages in that thread that make sense in no language? |
| 09:59 | <karlcow> | automatism and dadaism points? |
| 09:59 | <asmodai> | Mmm, looks like Modernizr does most of what you would want. |
| 10:00 | <MikeSmith> | karlcow: no that would be intentionally not making sense |
| 10:00 | <karlcow> | houla it becomes complicated :) |
| 10:05 | <Ms2ger> | zcorpan_, I think I poked you about a change to Opera's microdata test? |
| 10:06 | <zcorpan_> | Ms2ger: i seem to have forgotten what that was about |
| 10:07 | <Ms2ger> | Anyway, I pushed it :) |
| 10:08 | <zcorpan_> | http://dvcs.w3.org/hg/html/rev/880c38769cff ? |
| 10:11 | <Ms2ger> | Yeah |
| 10:28 | <zcorpan_> | Ms2ger: the test seems to have dup test names |
| 10:28 | <Ms2ger> | Looks that way, yes |
| 10:33 | <zcorpan_> | so opera and firefox both fail the new test but have different length |
| 10:35 | <Ms2ger> | Firefox passes, as of yesterday |
| 10:35 | <jgraham> | MikeSmith: I think people often post things to public-html that make sense in no language |
| 10:35 | <zcorpan_> | Ms2ger: ok |
| 10:35 | <Ms2ger> | jgraham, I thought that was what public-html was for? |
| 10:48 | <Stevef_> | dglazkov: i presume you are not promoting 'raspberry ketones' https://twitter.com/dglazkov/status/265767128945287168 |
| 10:56 | Ms2ger | thinks |
| 10:59 | <annevk> | can someone review http://annevankesteren.nl/2012/11/copyright maybe? |
| 11:00 | <Ms2ger> | Sure, if you tell me what WebIDL thinks about http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1895 |
| 11:00 | <SimonSapin> | annevk: s/clausule/clause/ ? |
| 11:01 | <annevk> | thanks |
| 11:01 | <annevk> | Ms2ger: heh |
| 11:01 | <Ms2ger> | lgtm |
| 11:04 | <SimonSapin> | annevk: I personally agree with you on this, but what kind of review do you want? |
| 11:06 | <annevk> | SimonSapin: whether it makes sense, that's all :) |
| 11:06 | <annevk> | sometimes I write stuff and nobody knows what's going on |
| 11:07 | <Ms2ger> | ... a party? |
| 11:07 | <jgraham> | annevk: The last sentence is a bit weird |
| 11:08 | <jgraham> | It seems like mixed metaphors in English |
| 11:08 | <jgraham> | and a parody of them at that |
| 11:08 | <annevk> | jgraham: ah yeah, but that was meant to be weird |
| 11:08 | <annevk> | but maybe it didn't turn out the way I intended? |
| 11:08 | <annevk> | maybe I should leave out the marbles thing |
| 11:09 | <jgraham> | Well the phrases it recalls are "picked up my toys and went home" and "left for greener pastures" |
| 11:09 | <jgraham> | But I don't think you want to mix them like that |
| 11:09 | <SimonSapin> | annevk: I think your post makes sense, but I already had some context |
| 11:09 | <jgraham> | (I think it makes sense too) |
| 11:09 | <jgraham> | (as a whole) |
| 11:09 | <SimonSapin> | annevk: also, I read in some ML archive that you disagree with the Invited Expert Agreement (o |
| 11:10 | <SimonSapin> | (ok so far) but were ok to work at W3C through a Member company |
| 11:10 | <SimonSapin> | is that the case? |
| 11:11 | <annevk> | I'm not sure that's the case |
| 11:11 | <annevk> | it depends on Member's company legal department and their reading of the Agreement |
| 11:12 | <jgraham> | The agreement seems decidedly confused on this point |
| 11:12 | <zcorpan_> | Ms2ger: thanks for the test |
| 11:12 | <annevk> | the Member Agreement is different fwiw, but the W3C and I disagree on how |
| 11:13 | <jgraham> | " Member acknowledges that all such jointly owned inventions, software or other copyrightable materials, or materials owned by Member made available by Member for Consortium activities, will be made available to the general public pursuant to the then-current W3C Software Notice and License" |
| 11:13 | <jgraham> | Seems like the W3C agreement should see everything published under the software license |
| 11:14 | <annevk> | Ms2ger: that post was pretty literal :) |
| 11:14 | <jgraham> | (except when the AC and director both agree otherwise) |
| 11:14 | <Ms2ger> | :) |
| 11:34 | <annevk> | jgraham: guess it depends on what the actual Member Agreement says that was signed by the Member and W3C, which is prolly private to the Member and W3C |
| 11:35 | <jgraham> | Yeah, would be weird to have a different agreement on the website to that which was actually signed |
| 11:35 | <jgraham> | But I guess weird is par for the course |
| 11:36 | <annevk> | the one on the website says [DRAFT] for a reason |
| 11:42 | <SimonSapin> | annevk: I don’t know about other members, but I can tell you that the agreement signed by Kozea is the same as on the website, except for [DRAFT] and the signatures |
| 11:42 | <SimonSapin> | [DRAFT] is there only because it’s not really an agreement until it is signed |
| 11:42 | <SimonSapin> | at least that’s what we were told |
| 11:44 | <hsivonen> | annevk: “This in part is why left for greener style sheets.” missing “I”? |
| 11:45 | <SimonSapin> | annevk: what about "a greener logo" ? |
| 11:46 | <annevk> | SimonSapin: the joke is usually about style sheets |
| 11:46 | <annevk> | hsivonen: ta |
| 11:46 | <annevk> | SimonSapin: cool, so then the Document License thing is indeed dubious |
| 11:46 | <annevk> | someone else told me about that as well |
| 11:46 | <annevk> | oh well, gotta go |
| 12:45 | <karlcow> | "I'm responding privately (cc'ed to www-archive)" |
| 12:45 | <jgraham> | hahaha |
| 12:45 | <karlcow> | :) |
| 12:58 | <zcorpan> | hmm, not sure i'm happy with the banning of some encodings not being normative anymore, but i guess it'll fix itself when the spec is updated to use the Encodings spec |
| 13:22 | <MikeSmith> | hmm weird |
| 13:22 | <MikeSmith> | question about form submission with file upload |
| 13:25 | <MikeSmith> | if I'm sending multipart/form-data request with a file upload, should it matter whether the file-upload part is the first part in the request or the last part? |
| 13:26 | <MikeSmith> | because in the case of the vaidator.nu backend at least it does seem to matter |
| 13:27 | <MikeSmith> | in that if the file-upload part is the first part in the request, it seems to ignore all the other form-data parts that follow |
| 13:28 | <MikeSmith> | whereas if I send the request with the other form-data parts first and he form-data part for the file upload part last, it doesn't ignore the other form-data parts the precede it |
| 13:54 | <zcorpan> | MikeSmith: is this about what the browser sends over the wire or how the server exposes what it got over the wire? |
| 13:55 | <MikeSmith> | zcorpan: I was asking about it in the context of what the browser is supposed to send |
| 13:55 | <MikeSmith> | but in the mean time I think I found the cause of the problem in the validator code |
| 13:56 | <MikeSmith> | which is, it appears that when it's looping over the form-data parts, as soon as it sees a filename part, it processes that and then breaks out of the loop |
| 13:56 | <MikeSmith> | intentionally |
| 13:57 | <MikeSmith> | why I dunno |
| 13:58 | <zcorpan> | hmm, the multipart/form-data part of the spec references rfc2388 to get a byte stream of the data |
| 13:59 | <MikeSmith> | yeah I followed that reference |
| 13:59 | <MikeSmith> | which in turn references RFC1867 |
| 14:00 | <zcorpan> | and rfc2047 |
| 14:00 | <zcorpan> | wonder if i should file a spec bug to rewrite the requirements in the html spec |
| 14:00 | <MikeSmith> | hmm but it doesn't seem to actually be normatively referencing those |
| 14:01 | <MikeSmith> | yeah maybe |
| 14:01 | <MikeSmith> | rfc2388 doesn't seem like such a great thing to reference |
| 14:01 | <MikeSmith> | there's hardly anything in it |
| 14:02 | zcorpan | notices there are errata |
| 14:02 | <MikeSmith> | man there are virtually not conformance requirements stated in that RFC |
| 14:06 | <zcorpan> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=19879 |
| 14:07 | <MikeSmith> | good |
| 14:12 | <MikeSmith> | so for validator.w3.org I'm using virtually the same markup for the input form as Henri has for the validator.nu input form, and only a slightly modified version of the JavaScript from the validator.nu UI |
| 14:12 | <MikeSmith> | and both have the file-input part of the form first, in document order, before all the other parts of the form where you can choose other options |
| 14:12 | <MikeSmith> | w |
| 14:13 | <MikeSmith> | which presets and which parser etc. |
| 14:14 | <MikeSmith> | but when I submit that form using the stock validator.nu stuff and sniff the request, I see it's somehow always putting the file-input part last |
| 14:14 | <MikeSmith> | different from the document order |
| 14:15 | <MikeSmith> | but when I submit it with the modified W3C version it always sends it in document order, so the file-upload part is first |
| 14:15 | <MikeSmith> | and I can't see anything in my modifications that would be causing it to behave any differently with respect to that |
| 14:16 | <MikeSmith> | anyway clearly I need to ask Henri what's up |
| 14:16 | <MikeSmith> | this is what caused the problem that dude reported on the help⊙lwo list yesterday |
| 15:02 | <zcorpan> | MikeSmith: http://validator.nu/script.js maybeMoveDocumentRowDown() |
| 15:02 | <MikeSmith> | ah |
| 15:02 | <MikeSmith> | will take a closer look at that |
| 15:02 | <annevk> | fwiw, the form submission thing has been discussed in the past |
| 15:02 | <annevk> | someone needs to reverse engineer it and write down the new spec |
| 15:03 | <annevk> | or write a bunch of tests so Hixie can write the spec |
| 15:05 | <annevk> | zcorpan: maybe Microsoft didn't pass that test? o_O |
| 15:05 | <annevk> | re http://dvcs.w3.org/hg/webapps/rev/d0ec1879951a |
| 15:06 | <zcorpan> | i thought they usually submit tests that they pass |
| 15:16 | <MikeSmith> | zcorpan: yeah maybeMoveDocumentRowDown() is what I'm missing |
| 15:17 | <MikeSmith> | well, it's still there in the modified version |
| 15:17 | <MikeSmith> | but it's not having the intended effect because I made multiple tables where Henri's source has the whole form in one table |
| 15:18 | <MikeSmith> | the file-upload part is in one table and the other parts in another |
| 15:18 | <zcorpan> | i can see that changing the document tree can break the script |
| 15:18 | <MikeSmith> | yeah |
| 15:18 | <MikeSmith> | I should of looked at it more carefully |
| 15:19 | <MikeSmith> | anyway thanks I know I can see how to fix it now |
| 15:24 | <zcorpan> | np |
| 15:41 | <Hixie> | hsivonen: the reason i mentioned the plagiarism angle is that i've started fielding e-mails from people who insult people at the whatwg on the grounds that we're stealing the w3c's work |
| 15:43 | <asmodai> | >_< |
| 15:44 | <asmodai> | I love how people have such a distorted view of the world. |
| 15:44 | <annevk> | Now the W3C claimed HTML for itself again, it seems a lot of people forgot how that thing started out... |
| 15:49 | <wilhelm> | The consortium consumes n million USD per year. They need a significant number of paying members to bring in that much money. As usual, just follow the money trail to understand the political priorities. |
| 15:50 | <wilhelm> | (n is a surprisingly high number.) |
| 15:55 | <Stevef_> | most of the money goes to fill MikeSmith's Gin can |
| 15:56 | <annevk> | so I'm writing a few blog posts to cover TPAC, for public interest and to hopefully give Mozilla the impression it was worth the expensive hotel |
| 15:56 | <annevk> | I'm wondering, should I just post them as I finish them or one a day? how does this work? |
| 15:56 | <annevk> | Stevef_: weekly bottle of Shōchū you mean |
| 15:57 | <wilhelm> | That simplifies things. We get Mike sacked, and W3Cs money addiction will cause less damage. |
| 15:58 | <annevk> | if you get Mike sacked, the W3C's ability to do good will greatly diminish |
| 15:59 | <wilhelm> | (c: |
| 15:59 | <karlcow> | wilhelm: n in fact is not high |
| 16:00 | <karlcow> | A couple of years ago it was ~10 times LESS than Mozilla |
| 16:00 | <karlcow> | annevk: I would do one a day |
| 16:01 | <karlcow> | to give time for readers to do other things in between. |
| 16:01 | <annevk> | I'll guess I put my draft online somewhere for this channel |
| 16:02 | <foolip> | annevk, I want one post every morning on the bus, like I got today ;) |
| 16:03 | <annevk> | heh okay, I'll try to make it so |
| 16:06 | <annevk> | sweet, my GitHub fix is already live, tweets no longer start with "Git: " |
| 16:07 | <annevk> | I guess for the @WHATWG account we could remove the em dash |
| 16:09 | <wilhelm> | karlcow: I strongly disagree - and a comparision to Mozilla is irrelevant. The money addiction leads to poor decisions to avoid offending paying, but irrelevant members. And the ROI per dollar spent is shockingly low compared to the WHATWG. |
| 16:10 | <GPHemsley> | annevk++ for removing "Git:" from tweets |
| 16:12 | GPHemsley | might be eligible to reply to polyglot, by jgraham's standards. |
| 16:12 | <GPHemsley> | But I wonder... are those rules applied before or after the reply? |
| 16:12 | <GPHemsley> | Also, what is polyglot? ;) |
| 16:13 | <annevk> | history will judge Polyglot Markup as an interesting exercise to keep the XML pipe dream alive just a little longer, but overall a gigantic waste of everyone's time |
| 16:14 | <wilhelm> | karlcow: Don't get me wrong here. I'm a W3C chair because I believe we need the consortium for a number of reasons. But I was shocked when I saw the budget numbers last week. I would spend those funds differently. And a big chunk of it - not at all, to maintain political independence. |
| 16:15 | <annevk> | then you don't quite get how the W3C operates |
| 16:15 | <wilhelm> | I do now. (c; |
| 16:15 | karlcow | wonders what you call money addiction. |
| 16:16 | <annevk> | they try to find certain activities to wiggle themselves into and then get those people to join the W3C so they get more money, forgetting they then also have to spend more money, and for some reason they now have to resort to asking money beyond their Members |
| 16:17 | <karlcow> | WHATWG budget is difficult to evaluate, you would have to calculate the time paid by companies. The time right now we are spending chatting for example. |
| 16:17 | <annevk> | Hixie: fwiw, I also started to call bullshit on "WHATWG uses a dictator model", "WHATWG is just browsers", "WHATWG is innovation whereas W3C is stability", and similar such themes |
| 16:18 | <annevk> | Hixie: no longer really interested in letting it slide, it seems |
| 16:18 | <annevk> | karlcow: you're draining my savings account, stop chatting! |
| 16:18 | <karlcow> | ;) |
| 16:19 | <SimonSapin> | karlcow: if you count people’s time in WHATWG budget, it’s the same with W3C |
| 16:20 | <karlcow> | SimonSapin: for WG participants, yes. |
| 16:20 | <karlcow> | and it's why the cost of participation is not the membership for most companies, but the cost of contributing. :/ which is unfortunately not easy for all engineers, people of good will who would like to participate. |
| 16:21 | <wilhelm> | karlcow: If you need n million dollars per year to survive, you may do things you otherwise wouldn't have done. There's a continuum stretching from "not offending people unneccessarily" to prostitution. |
| 16:22 | <GPHemsley> | To all of those in the United States: Don't forget to vote today! For all of those in other places: Go see if there's something you can vote for today! |
| 16:22 | <karlcow> | GPHemsley: USA spam |
| 16:23 | karlcow | tempted to bring in a bot announcing elections of all countries in the world |
| 16:23 | <karlcow> | wilhelm: such as cutting the staff? |
| 16:23 | <annevk> | is anyone following the IETF meeting this week? |
| 16:23 | <wilhelm> | When annevk says "Although Jeff Jaffe personally agrees with licensing specifications as the WHATWG does.., his defence for not doing it at the W3C is because there is no consensus among the W3C Members to do so", alarm bells should go off. |
| 16:24 | <karlcow> | wilhelm: why? |
| 16:24 | <karlcow> | (trying to understand what bothers you in the previous sentence) |
| 16:25 | <SimonSapin> | Aren’t there both USA resident who can’t vote there and expats who can? |
| 16:25 | <wilhelm> | karlcow: Jeff is taking a position he personally believes is wrong to avoid losing the stream of money. |
| 16:25 | <wilhelm> | That is harmful. |
| 16:26 | <wilhelm> | And yes, if cutting staff is what's needed to maintain political independence, that is what one should do. It'll suck, of course, but finding new jobs should be easy in our industry these days. |
| 16:26 | <karlcow> | Jeff has a personal opinion. We all do. Then we are hired for position/role. That's life reality. You can as a personal position resign if you think it is too difficult to be aligned with your own beliefs or you may decide to try to convince to change things. |
| 16:26 | <karlcow> | Jeff is an employee. |
| 16:28 | <karlcow> | I'm pretty sure most of the people here working for companies Opera, Google, Apple, do not agree 100% with their companies. Still they didn't leave, they are trying to accomodate with it (personal reasons, etc.) and work for what they like. |
| 16:28 | <wilhelm> | I'm still a W3C chair. I'm ringing alarm bells, not jumping ship. |
| 16:29 | <annevk> | I haven't quite jumped ship either. I'm probably running for the TAG. |
| 16:30 | <gsnedders> | You can run for TAG without being a member of any WG? Heh. |
| 16:30 | <annevk> | (The way that'll sort itself out with licensing is that I will sign the Agreement, but not join any WG.) |
| 16:31 | <wilhelm> | annevk: I'd vote for you if I could. (c: |
| 16:32 | <wilhelm> | I don't have 1950 EUR available to buy a vote, unfortunately. |
| 16:32 | <karlcow> | http://www.w3.org/2005/10/Process-20051014/organization.html#tag-participation |
| 16:32 | <karlcow> | wilhelm: you do not need |
| 16:33 | <SimonSapin> | wilhelm: who votes? |
| 16:35 | <wilhelm> | karlcow: Can random invited experts vote? (I quit Opera a year ago, and am running my own business now. My company is not a member.) |
| 16:36 | <karlcow> | I think just the W3C AC vote. |
| 16:37 | <SimonSapin> | wilhelm: so your company has no "conflict of interest" that prevents you from being invited expert? |
| 16:39 | <wilhelm> | SimonSapin: No, I own and run the company. I can do whatever I want. (c: |
| 16:40 | <karlcow> | Maybe there could be a proposal for sharing the vote in between. Some seats elected by AC, some seats elected by WG participants. |
| 16:40 | <karlcow> | foodforthought |
| 16:41 | <SimonSapin> | wilhelm: sure, but when I was invited expert w3c said that could be only for 6 months since my employer is doing w3c-related stuff like implementing specs |
| 16:43 | <wilhelm> | SimonSapin: Oh. Uh, I haven't checked. I'm just doing work for free here. If W3C has a rule that causes me to be kicked out several months ago, that's their loss. (c: |
| 16:44 | <SimonSapin> | Sure. I’m just curious as to how things work. I certainly don’t want to get anyone kicked out. |
| 16:45 | <karlcow> | SimonSapin: because of patents and responsibilities of employes with regards to their employers :/ You gonna hate legal framework around IP. ☹ |
| 16:45 | <annevk> | SimonSapin: it's a weird model imo; productive people get to pay to play and large corporations hardly do shit other than be conservative when it comes to changing things around |
| 16:46 | <wilhelm> | I have no idea how any of this works. I just want to get the work done. |
| 16:46 | <annevk> | karlcow: sounds like bullshit |
| 16:46 | <annevk> | karlcow: you don't need pay-to-play for his employer to sign of on patents |
| 16:46 | <SimonSapin> | we don’t even have any patent |
| 16:46 | <karlcow> | why bullshit? |
| 16:47 | <annevk> | karlcow: because the reason they want his company to join is for the money |
| 16:47 | <karlcow> | so much hate each time. I don't get that. |
| 16:48 | <karlcow> | ah, understood what you meant. |
| 16:48 | <wilhelm> | Money addiction leads to poor decisions. |
| 16:49 | <Stevef_> | annevk: i have been involved at W3C for years no money has changed hands |
| 16:49 | <karlcow> | yes because it's a source of money and if everyone participates without paying the system collapses. And we are back to initial situation. How do we create a system which is financially viable. |
| 16:49 | <manu-db> | TabAtkins, TabAtkins_: ping |
| 16:49 | <karlcow> | wilhelm: It is not money addiction. You need money to run a system. |
| 16:50 | <annevk> | Stevef_: sure |
| 16:50 | <karlcow> | an organization, infrastructure may decide to run the system differently, but usually it takes time to do so. |
| 16:51 | <wilhelm> | karlcow: Sure, but needing n million dollars is absurd. And leads to poor decisions. |
| 16:51 | <karlcow> | ISO is another system. financed by states, and I'm not sure I would like to see that for W3C. honestly. It might be another way. |
| 16:51 | <karlcow> | Mozilla found Search Revenues as a system for injecting the money. |
| 16:52 | <karlcow> | If we think of Not For Profit organizations |
| 16:52 | <karlcow> | Wikipedia has a system of donations, not sure if they run with something else. |
| 16:52 | <wilhelm> | I'm sure we could do a good job with a bootstrapped W3C for $1M per year. That's a fraction of the current cost. |
| 16:52 | <karlcow> | Openstreetmap → donations too |
| 16:52 | <karlcow> | wilhelm: It's also nice on paper. |
| 16:53 | <karlcow> | Reality is a different thing. |
| 16:53 | <annevk> | You need a server, some services, and an IP framework, a lot of the other things can be scrapped |
| 16:53 | <annevk> | (to put it simply) |
| 16:54 | <karlcow> | that's a start. |
| 16:54 | <karlcow> | Who manage the server? |
| 16:54 | <wilhelm> | annevk: Some administrative and techinical staff may be useful. |
| 16:54 | <karlcow> | Who pays the bandwidth bill? |
| 16:54 | <annevk> | wilhelm: yeah, I meant that to be included |
| 16:54 | <wilhelm> | $1M gets you about five people. You can get far with that. |
| 16:54 | <karlcow> | huh? |
| 16:54 | <karlcow> | in which country? |
| 16:55 | <karlcow> | 5 persons + taxes + bills + offices + server + bandwidth + health insurance + etc. |
| 16:56 | <wilhelm> | In the most expensive country on the planet. Average Norwegian employees cost approximately 1M NOK per year, everything included, which translates to 175k USD. |
| 16:59 | <wilhelm> | Maintaining that would require membership fees from about 12 organizations, not 384(!!). |
| 17:00 | <karlcow> | ok so a system with no staff a bit like IETF |
| 17:00 | <wilhelm> | Limited, not zero staff. |
| 17:00 | <karlcow> | (a part the maintenance as you said) |
| 17:02 | <karlcow> | How do you dodge the patent attacks? |
| 17:02 | <karlcow> | special budget when it comes around? |
| 17:03 | <Ms2ger> | Kill software patents ;) |
| 17:04 | <karlcow> | hehe Ms2ger and rainbows and ponies |
| 17:04 | <Ms2ger> | Yes! |
| 17:04 | karlcow | would add copyrights too ;) |
| 17:05 | <wilhelm> | The IP policies at W3C are useful. They should be maintained. 384 members buying into that model is great too. But we won't have to resort to prostitution in times of desparation just to bring in money from said member organizations. |
| 17:05 | <karlcow> | 70 years after the death of the author is just insane |
| 17:06 | karlcow | wonders when wilhelm will make a kickstarter project :p |
| 17:07 | <wilhelm> | I won't. I propose reducing expenses, not increasing income. |
| 17:11 | <karlcow> | This will mean cutting the staff basically. |
| 17:12 | <karlcow> | and "hide" the management cost in relevant organizations/participants. |
| 17:12 | <Ms2ger> | How many employees does the W3C have nowadays? |
| 17:12 | <karlcow> | http://www.w3.org/People/ |
| 17:13 | <karlcow> | 72 |
| 17:13 | <karlcow> | http://www.w3.org/2011/05/w3cteam |
| 17:13 | <Ms2ger> | That sounds like a lot |
| 17:14 | <Ms2ger> | Never heard of most of them, too |
| 17:15 | <karlcow> | Ms2ger: it depends on how much you consider a lot. |
| 17:15 | <Ms2ger> | I consider 72 a lot |
| 17:15 | Ms2ger | ducks |
| 17:15 | <karlcow> | heh |
| 17:15 | annevk | was about to say, what he considers a lot should be clear from the context |
| 17:16 | <karlcow> | it depends on the number of groups. So I guess in wilhelm plans, some staff disappears because we abandon some activities. |
| 17:17 | <karlcow> | Mozilla is now how many? 600 |
| 17:17 | <karlcow> | Opera around 700 |
| 17:17 | <wilhelm> | karlcow: "A lot" is a relative term. If one could maintain a staff of 72 without making stupid decisions to keep all member organizations, I have no objections to the number. |
| 17:18 | <karlcow> | I don't know how many people there are at wikipedia. |
| 17:18 | <Ms2ger> | At Mozilla, I actually see results from those 600 ;) |
| 17:18 | <karlcow> | http://wikimediafoundation.org/wiki/Staff_and_contractors |
| 17:18 | <wilhelm> | Mozilla has not yet made stupid decisions to make more money. Opera has, unfortunately. |
| 17:19 | <karlcow> | wikipedia seems to be around ~100 |
| 17:19 | <Ms2ger> | wilhelm, we try :) |
| 17:20 | <karlcow> | and recruiting http://wikimediafoundation.org/wiki/Job_openings |
| 17:20 | <wilhelm> | Wikimedia has a business model that doesn't put them at odds with their principles. If they can get away with 100 employees - good for them. |
| 17:20 | <wilhelm> | W3C obviously can't. Not in its current shape. |
| 17:20 | <wilhelm> | Ms2ger: Good. (c: |
| 17:20 | <karlcow> | Mozilla is in a difficult position too :/ |
| 17:21 | <Ms2ger> | Mozilla does a lot of silly things, but afaik not for the money ;) |
| 17:22 | karlcow | has to run for lunch. will come back later. |
| 17:24 | <kennyluck> | Yeah, as a community member, I do think MozCamp is probably a silly idea. |
| 17:26 | <wilhelm> | That said, 72 people is a lot. |
| 17:27 | <JonathanNeal> | Hello |
| 17:39 | GPHemsley | doesn't take karlcow's temptation to be a threat. |
| 17:59 | <miketaylr> | back |
| 17:59 | <Ms2ger> | Hi |
| 18:15 | <Hixie> | annevk: christ, running for the TAG? |
| 18:15 | <Hixie> | i told you TPAC was a bad idea |
| 18:26 | <jgraham> | I dunno, it would be nice if we got a critical mass of web people onto the TAG |
| 18:27 | <jgraham> | so it wasn't all RDF and architecture astronautics |
| 18:30 | karlcow | has been caught with judgment of annevk in http://annevankesteren.nl/2012/11/copyright but then I checked and found it was correct. I didn't know |
| 18:36 | <jgraham> | (everone should ask hsivonen to join the TAG so that his "how many times have I been asked to join the TAG" counter overflows) |
| 18:37 | <karlcow> | the issue is not about joining the TAG, but more about the expectations the person has by doing so, and the will to compromise pushing these goals. |
| 18:38 | <karlcow> | make a compromise |
| 18:38 | karlcow | is checking the English definition |
| 19:07 | <Hixie> | jgraham: you seem to assume the tag is relevant |
| 19:07 | <Hixie> | jgraham: as far as i can tell there is no more benefit to getting "web people" onto the TAG than there is getting them onto an NFL team |
| 19:16 | <jgraham> | It isn't terribly relevant today, but that doesn't mean that something like the TAG couldn't be useful if it was doing something different |
| 19:16 | <Hixie> | so you're saying that maybe a group of people different than the TAG, doing something different than the TAG, might be useful? |
| 19:17 | <Hixie> | ...why don't we just form that group of people to do that useful thing, then, instead of adding the task of "convince the tag" to do that thing |
| 19:19 | <jgraham> | Well there are certian advantages to revolution-by-democracy rather than revolution-by-setting-up-your-own-thing-and-ignoring-the-old-thing |
| 19:20 | <karlcow> | NFL team… is IMHO more dangerous. |
| 19:24 | <Hixie> | jgraham: what is the thing the TAG could be doing that would be useful? |
| 19:26 | <karlcow> | The sense of useful depends on what the community defines it to be useful. |
| 19:34 | <Hixie> | hmm, no roc |
| 19:34 | <Hixie> | so i'm looking at this canvas-in-workers idea |
| 19:35 | <Hixie> | the main thing i'm wondering is, how do we get the data from the thread onto the display? |
| 19:35 | <Hixie> | i can think of three main approaches: |
| 19:35 | <Hixie> | 1. getImageData(), postMessage(), putImageData() |
| 19:35 | <Hixie> | this could be implemented somewhat efficiently (lazily) so long as the author keeps his hands off |
| 19:36 | <Hixie> | 2. something similar to getImageData() but that is officially async and hands-off, only useful for this purpose |
| 19:36 | <Hixie> | 3. a way to have a <canvas> in the DOM, but _hand the context over to the worker_ |
| 19:36 | <Hixie> | so the drawing always happens async straight from the worker |
| 19:36 | <Hixie> | any opinions? |
| 19:42 | <Hixie> | i'm thinking 3, with canvas.getContext('worker') returning a Transferable object |
| 19:43 | <Hixie> | once you have a worker context like this, the methods on canvas like toDataURL() are blocked |
| 19:43 | <Hixie> | the object itself just has a getContext() method |
| 19:43 | <Hixie> | and once it has a context gotten, it can no longer be transferred |
| 19:43 | <Hixie> | hmmm |
| 19:56 | <annevk> | Hixie: to support Alex Russell |
| 19:59 | <hober> | how many open seats are there? |
| 20:00 | <hober> | 4? |
| 20:08 | <Hixie> | hmmmm |
| 20:08 | <Hixie> | how do we get the equivalent of HTMLImageElement into a worker |
| 20:08 | <Hixie> | XR? |
| 20:08 | <Hixie> | XHR even? |
| 20:08 | <Hixie> | have it return an opaque Image object |
| 20:08 | <Hixie> | ? |
| 20:08 | <yuhong> | <annevk> history will judge Polyglot Markup as an interesting exercise to keep the XML pipe dream alive just a little longer, but overall a gigantic waste of everyone's time |
| 20:09 | <yuhong> | Thanks god IE8 is finally beginning to die. |
| 20:09 | <yuhong> | Another ~10 years old feature IE8 fails to support is DOM level 2. |
| 20:09 | <yuhong> | Google Docs will drop IE8 support soon. |
| 20:11 | <yuhong> | jQuery is planning to maintain two parallel codebases, 1.9 and 2.0, with 1.9 supporting IE8 and 2.0 not. |
| 20:13 | <yuhong> | 37signals's Basecamp too. |
| 20:34 | <Hixie> | annevk: how do you feel about making XMLHttpRequest have an 'image' responseType that returns an HTMLImageElement in the main thread and an Image object in a worker thread? |
| 20:34 | <Ms2ger> | That sounds somewhat fishy to me |
| 20:35 | <Hixie> | i need a way to get an opaque non-dom image object in workers for canvas workers |
| 20:35 | <Ms2ger> | Can we put that Image API on the main thread too? |
| 20:35 | <Hixie> | API? |
| 20:36 | <Hixie> | canvas is already on the main thread |
| 20:36 | <Ms2ger> | Interface, whatever you want to call it:) |
| 20:36 | <Hixie> | what use would it have? |
| 20:37 | <Ms2ger> | Making it easier to share code between workers/non-workers |
| 20:37 | <Ms2ger> | And a simpler API to learn |
| 20:37 | <Hixie> | when i say "opaque non-dom image object" i mean opaque |
| 20:37 | <Hixie> | as in, has no members |
| 20:37 | <Hixie> | interface Image { }; |
| 20:37 | <Hixie> | so it would be a strict subset of HTMLImageElement |
| 20:38 | <Hixie> | not much point exposing it on the main thread... |
| 20:44 | <Hixie> | roc: opinions on http://junkyard.damowmow.com/514 is welcome |
| 20:44 | <Hixie> | anyone else, too |
| 20:44 | <Hixie> | it's a strawman for canvas in workers |
| 20:44 | <Hixie> | 2d and gl, though i have only looked at the 2d part |
| 20:45 | <Hixie> | afk, bbiab |
| 20:50 | <roc> | Hixie: sounds pretty good |
| 20:51 | <roc> | a convenient way to pass binary image data across to the worker might be interesting |
| 20:52 | <roc> | I suppose you can kinda do it by creating a Blob on the main thread and using XHR with the Blob URI on the worker? |
| 20:52 | <roc> | but that requires the awful revocation stuff |
| 20:52 | <roc> | anyway, it looks like a good start |
| 21:33 | <annevk> | Hixie: happy with that in general, file a bug |
| 21:33 | <annevk> | Hixie: had been thinking about it |
| 21:36 | <zcorpan> | roc: or paint the image on a canvas and send the ImageData. but that's also not very convenient |
| 21:36 | <zcorpan> | convenient would be to postMessage(<img>) and have it come out as an Image on the other end. but that wouldn't make sense for cross-document messaging |
| 21:37 | <zcorpan> | Hixie: since "Image" already exists as a constructor in window, maybe we should call it something else |
| 21:37 | <zcorpan> | Hixie: e.g. WorkerImage |
| 21:54 | <MikeSmith> | "URI/IRI/URL thread among IETF/W3C/WHATWG" discussion at IETF f2f is starting in a few minutes |
| 21:54 | <MikeSmith> | http://www.ietf.org/proceedings/85/agenda/agenda-85-iri |
| 21:55 | <MikeSmith> | http://tools.ietf.org/wg/iri/minutes |
| 21:58 | <hober> | MikeSmith: is there an IRC channel into which things get scribed / where we can follow along? |
| 21:58 | <MikeSmith> | hober: no, XMPP |
| 21:59 | <MikeSmith> | XMPP chat |
| 21:59 | <MikeSmith> | which I think you can do via Adium |
| 21:59 | <hober> | do you have a link to the xmpp conference room? |
| 22:00 | <MikeSmith> | xmpp:iri⊙jio?join |
| 22:00 | smaug____ | assumes annevk will attend the meeting |
| 22:01 | hober | actually managed to join |
| 22:01 | <MikeSmith> | oh iChat also |
| 22:02 | <hober> | yeah, i'm using ichat :) |
| 22:22 | <MikeSmith> | http://www.meetecho.com/ietf85/iri works too I guess |
| 22:22 | <MikeSmith> | browser-based |
| 22:24 | <MikeSmith> | Roy at the mic |
| 22:25 | <MikeSmith> | "will not resolve the anti-social behavior within the WHATWG" |
| 22:26 | <hober> | wat |
| 22:26 | <hober> | i don't understand the relationship between the minutes and whatever's actually happening in that room |
| 22:27 | <wilhelm> | Did the scribe die? |
| 22:29 | <MikeSmith> | hahaha "negotiating with terrorists" |
| 22:29 | <MikeSmith> | way to sling the rhetoric man |
| 22:29 | <MikeSmith> | "I have no problem ignoring their work in the future" |
| 22:29 | <MikeSmith> | where "their work" is the URL standard |
| 22:30 | <hober> | wow, sounds like a wonderfully productive meeting... |
| 22:30 | <MikeSmith> | "You don't have critical mass for doing work. You're done." |
| 22:30 | <MikeSmith> | heh |
| 22:31 | <gsnedders> | Critical mass of random server people, where most of the weirdness is handled by either clients or CGI scripts? |
| 22:31 | <MikeSmith> | hober, wilhelm : yeah in my experience the minutes/scribing are not nearly as detailed as most W3C meetings are |
| 22:31 | <MikeSmith> | it's more like the minutes are notes |
| 22:31 | <MikeSmith> | scratchpad |
| 22:32 | <MikeSmith> | "URXes" "HURLs" |
| 22:32 | <hober> | how very open of them |
| 22:32 | <MikeSmith> | "get them to call it something else" [other than URLs] |
| 22:32 | <MikeSmith> | bravo Larry |
| 22:32 | <MikeSmith> | Larry is now pointing out that they are called URLs |
| 22:33 | <MikeSmith> | Larry replying to whoever it was who made the "get them to call it something else" comment |
| 22:33 | <MikeSmith> | Larry, "I don't see any point in fighting over the name." |
| 22:35 | <MikeSmith> | "Let the WHATWG go for it. They'll break the Web but not the Internet." |
| 22:35 | <wilhelm> | Did he just say "browsers are not the web"? |
| 22:35 | <MikeSmith> | yeah |
| 22:35 | <MikeSmith> | no the speaker didn't say it |
| 22:35 | <MikeSmith> | Roy did |
| 22:36 | <MikeSmith> | he said Roy said it and he agreed with it |
| 22:36 | <wilhelm> | Fascinating. |
| 22:37 | <MikeSmith> | whoo boy now they're talking about LEIRIs |
| 22:39 | <MikeSmith> | funny how they keep using the term "the outside world" to refer to well, the actual world and the actual people in the world |
| 22:40 | <MikeSmith> | we need to get Steven Soderbergh to direct the screen version of the multi-year drama |
| 22:40 | <MikeSmith> | someday |
| 22:40 | <MikeSmith> | *this multi-year drama |
| 22:41 | <MikeSmith> | this is like word-for-word the discussion we had at the BOF in Hiroshima before they even chartered the IRIbis WG |
| 22:41 | <MikeSmith> | groundhog day |
| 22:41 | <wilhelm> | Interesting to observe Larry as a voice of reason here. |
| 22:41 | <MikeSmith> | wilhelm: indeed yeah he is |
| 22:42 | <MikeSmith> | the scribing here is near worthless |
| 22:42 | <MikeSmith> | I don't know if it's just Martin not bothering to even try or what |
| 22:42 | <MikeSmith> | but he's not even indicating who the current speaker is |
| 22:43 | <MikeSmith> | the keep scribing "Mic:" to indicate that that somebody is talking at the mic |
| 22:43 | <MikeSmith> | instead of actually noting who the person is |
| 22:43 | <MikeSmith> | really fucking helpful |
| 22:43 | <wilhelm> | (c: |
| 22:45 | <MikeSmith> | god almighty we need some metaphor that's way beyond "going off into the weeds" |
| 22:45 | <MikeSmith> | we are way way past the weeds here |
| 22:45 | <nessy> | wow, you got me curious! |
| 22:46 | <MikeSmith> | this is like a Virginia Wolfe novel |
| 22:46 | <hober> | going off the cliffs of insanity? |
| 22:47 | <wilhelm> | MikeSmith: http://upload.wikimedia.org/wikipedia/commons/1/19/Train_wreck_at_Montparnasse_1895.jpg |
| 22:49 | <MikeSmith> | "ok other speaker you rambled on again about some stuff you talked about in almost the same exact words 4 years ago but you had your say (again) and now it's my turn to (again) further pile my stream of consciousness on top of what you said" |
| 22:50 | <MikeSmith> | I am beginning to suspect this is some highly refined art form |
| 22:50 | <MikeSmith> | like Noh drama |
| 22:52 | <Hixie> | zcorpan: I was thinking of using Image precisely because it already exists :-) |
| 22:52 | <MikeSmith> | performance repeated each couple years or whatever with the performers just making slight variations in the way they hold their folding fans |
| 22:53 | <MikeSmith> | oh somebody just dropped the phrase "doesn't give a rat's ass" |
| 22:54 | <MikeSmith> | guess which performer that is |
| 22:54 | <Hixie> | when i was talking to IETF folk about during URL at the IETF, one of the things I said was that if we did it, we had to have a guarantee that the tone of discussion would be productive and respectful |
| 22:54 | <Hixie> | I CAN'T IMAGINE WHY I THOUGHT THAT WAS IMPORTANT |
| 22:55 | <Hixie> | not like there's any reason for me to fear it might not be! |
| 22:56 | <MikeSmith> | pretty much all of the people in this discussion are respectful at least. except for one |
| 22:56 | <MikeSmith> | if not productive |
| 22:57 | <Hixie> | given that you've quoted multiple people already... |
| 22:57 | <MikeSmith> | Hixie: they have an empty chair in the room that they're referring to as "the WHATWG ghost" |
| 22:58 | <MikeSmith> | can't make this stuff up |
| 22:58 | <Hixie> | MikeSmith: and you say they're all being respectful? :-) |
| 22:58 | <MikeSmith> | heh |
| 22:58 | <MikeSmith> | at least that's what I think they are referring to |
| 22:58 | Hixie | ponders roc's thing about how to transfer images across postMessage() |
| 22:59 | <Hixie> | having HTMLImageElement transfer into an Image, and transfer back into an HTMLImageElement, is one option |
| 22:59 | <MikeSmith> | it's possible that they have collectively ascended to some other level of abstraction with the "the WHATWG ghost" metaphor that's of too great refinement for me to appreciate |
| 23:00 | <hober> | MikeSmith: that's probably it |
| 23:01 | <Hixie> | if i have this object in the worker that's like an HTMLImageElement except for not being a DOM node, should I just call it HTMLImageElement? Or Image? or something else and we'll eventually put it on the main thread even though it's redundant? |
| 23:05 | <MikeSmith> | Hixie: would seem odd to call it HTMLImageElement |
| 23:05 | <MikeSmith> | or misleading |
| 23:05 | <Hixie> | yeah |
| 23:05 | <Hixie> | but if we ever put the DOM into workers, would be weird to not have it called that... |
| 23:06 | <MikeSmith> | oh |
| 23:08 | <MikeSmith> | yay finally Pete Resnick talking now |
| 23:09 | <MikeSmith> | Hixie: mentioning that he talked along with others with you and Anne |
| 23:10 | <MikeSmith> | I like Pete |
| 23:10 | <annevk> | so sorry for not attending the IRI thing |
| 23:10 | <annevk> | I was busy drinking wine |
| 23:10 | <MikeSmith> | heh |
| 23:10 | <annevk> | back tomorrow |
| 23:15 | <MikeSmith> | "I'm looking for some reason not to close down this working group." |
| 23:15 | <MikeSmith> | "Does anyone want to explain to me why the working group should not be shut down?" |
| 23:42 | <Hixie> | heycam: yt? |
| 23:43 | <heycam> | hi Hixie |
| 23:43 | <Hixie> | hey hey |
| 23:43 | <Hixie> | hope you had a good vacation |
| 23:43 | <heycam> | yes thanks :) |
| 23:43 | <Hixie> | question about "implements", what happens if you have an interface that has an "attribute DOMString foo", and it implements another interface with "readonly attribute DOMString foo"? |
| 23:43 | <Hixie> | is that just non-conforming? |
| 23:44 | <heycam> | yeah I think the spec either says that's non-conforming, or that which one "wins" is undefined |
| 23:44 | <Hixie> | k |
| 23:44 | <heycam> | what do you want to do? |
| 23:44 | <Hixie> | dunno. my general problem is that i want to represent images in both the main thread and workers |
| 23:44 | <Hixie> | and ideally want to make the objects postMessage()able from one to the other |
| 23:45 | <Hixie> | i'm scared of making an HTMLImageElement (or just Image) object in the worker that is a subset of HTMLImageElement in the main thread because people might rely on it not being a Node and then if we ever add the DOM to workers we'll be in a world of hurt |
| 23:45 | <Hixie> | but then i don't want to introduce a separate object that isn't HTMLImageElement because it's bound to get into the main thread and then it's Yet Another Way to represent images there |
| 23:45 | <heycam> | ok so you don't want some properties not being on the worker-version of the object, in case they pop up later |
| 23:46 | <Hixie> | well my question above was about making HTMLImageElement implement Image, and having just Image in the worker |
| 23:46 | <Hixie> | and having Image have a readonly "src" |
| 23:46 | <heycam> | oh, what's Image |
| 23:46 | <heycam> | that used to be just a named constructor |
| 23:46 | <Hixie> | some theoretical new object for workers that is a subset (non-Node) of HTMLImageElement, named to coincide with the constructor Image() that creates an HTMLImageElement |
| 23:47 | <Hixie> | the naming being intentional to indicate that they're hte "same object" in principle |
| 23:47 | <Hixie> | and to make it so we can never bring Image to the main thread |
| 23:47 | <Hixie> | maybe i should just bite the bullet and introduce the Yet Another Way To Represent Bitmap Images |
| 23:47 | <heycam> | ok so Image would now be an interface, its contructor would return an HTMLImageElement in non-workers |
| 23:48 | <Hixie> | yeah |
| 23:48 | <Hixie> | (well, it'd be an interface in workers, and not in main thread) |
| 23:48 | <Hixie> | not present |
| 23:48 | <heycam> | do you think it would really be an issue if the worker version of the object some time in the future gained the reaminder of the HTMLImageElement properites? |
| 23:48 | <Hixie> | i don't know, but it sure seems plausible |
| 23:49 | <heycam> | btw what do you get from passing images to workers? |
| 23:49 | <heycam> | that isn't just what you get from passing a CanvasPixelArray of the image? |
| 23:50 | <Hixie> | well this is in the context of http://junkyard.damowmow.com/514 - adding canvas to workers |
| 23:50 | <Hixie> | and canvas needs images, really |
| 23:50 | <Hixie> | now i imagine we'll get images in a number of ways, e.g. from Blobs fetched from IndexDB (assuming we add a toImage() method to Blob), from XHR (responseType='image') |
| 23:51 | <Hixie> | but it makes sense that we'd also get images from the main thread |
| 23:51 | <heycam> | I see |
| 23:51 | <heycam> | I guess it is kind of weird to have the DOM element be the canonical representation of images |
| 23:51 | <Hixie> | welcome to the web, but yeah |
| 23:51 | <heycam> | :) |
| 23:53 | <Hixie> | current ways to represent binary (bitmap) image data in the main thread that I can think of: HTMLCanvasElement, HTMLImageElement, ImageData, Blob, a data: URL |
| 23:53 | <Hixie> | I suppose I could add a new object that can be constructed from any of those |
| 23:53 | <Hixie> | and just be done with it |
| 23:54 | <heycam> | well what about ImageData |
| 23:54 | <heycam> | or does that not include some information that the others have |
| 23:54 | <Hixie> | ImageData's problem is that it exposes every pixel and so it is trivial for the author to monkey with the data, which immediately breaks every optimisation you can think of |
| 23:54 | <Hixie> | if you don't allow that, you can do everything async or even off the main thread |
| 23:55 | <Hixie> | so we really don't want the primary representation to be ImageData |
| 23:55 | <Hixie> | or at least, it seems like a bad idea to me |
| 23:57 | <Hixie> | yeah ok i think i should just bite the bullet and make a new object for everywhere |
| 23:58 | <Hixie> | with a factory method that can take any of the above and asynchronously returns an object that has the data loaded |
| 23:58 | <Hixie> | BitmapImage maybe? |
| 23:59 | <zewt> | in principle you should be able to fast path until someone changes bits, but it probably does make sense to have an "async-optimized" container for images, in the same way Blob is that for generic data (vs. ArrayBuffer) |
| 23:59 | <Hixie> | zewt: yeah, and i expect people will do that (lazy ImageData), but the problem is that it's really easy to run into the slow path by mistake |
