| 04:56 | <wilhelm> | ~ |
| 05:57 | <jgraham> | wilhelm: . |
| 05:58 | <wilhelm> | jgraham: I blame the lag. |
| 06:04 | <jgraham> | So would the Jackson Five if they were around today |
| 14:54 | <gsnedders> | It says something that when I see a Java plugin used I'm actively surprised, whereas a decade ago it wouldn't have surprised me in the least (though annoyed me while I waited for the JVM to startup…) |
| 15:02 | bga_ | want jvm in browser by default instead js. If somebody wants js - Rhino |
| 15:03 | <gsnedders> | bga_: I want something where a single company doesn't control the language by virtue of holding patents over it. |
| 15:03 | <bga_> | mm. patents. ok. llvm |
| 15:06 | <gsnedders> | bga_: Too slow, and not designed for non-trusted code-execution. |
| 15:06 | <bga_> | but monopoly is ok. single version. single mind. better than 3-4 companies implements same stuff i.e. just waste haman resources |
| 15:06 | <gsnedders> | (slow in the compile-time sense, not in generated-code performance) |
| 15:07 | <gsnedders> | bga_: Then why have multiple browser engines at all? |
| 15:07 | <gsnedders> | What incentive is there for performance, etc? |
| 15:08 | <bga_> | lets c++ compler maker fights, not we :) |
| 15:08 | <bga_> | and cpu manufacturers |
| 15:09 | <gsnedders> | But that doesn't really affect the performance of canvas that much. |
| 15:11 | <bga_> | but ok. monopoly is ideal. you want stimul to make product better. i see MS make Windows better w/o pressure of concurents. thay just cares about users |
| 15:11 | <bga_> | oracle too |
| 15:12 | <bga_> | that ill called perfectionism |
| 15:13 | <gsnedders> | bga_: MS don't make Windows better for that reason, MS continue to develop Windows because licensing is tied to specific Windows versions. |
| 15:13 | <gsnedders> | To continue to sell site-licenses, they have to invalidate the old licenses. |
| 15:14 | <gsnedders> | The revenue they get from per-unit OEM sales is comparitively minute. |
| 15:15 | <gsnedders> | bga_: A good monopoly example where you have no financial incentive to develop it further is IE/Win, and the abandonment of its development a decade ago. It brings in no revenue to MS, so it's purely a cost-sink. |
| 15:17 | <bga_> | gsnedders for example we have Flash. wa have canvas. Flash api much better and faster than canvas because Adobe is monopoly. single version. but 3 companies enchange canvas speed triple slower. And after 5+ years canvas reach level of flash |
| 15:18 | <gsnedders> | bga_: It's not clear simply having more resources working on a single engine would have addressed that problem any quicker. |
| 15:18 | <bga_> | gsnedders because IE6 was ideal. 3d, vector graphic, FFI(activex), sound, video, smil, etc |
| 15:19 | <gsnedders> | bga_: The IE team didn't, and were fully intending on fixing bugs in the near future after IE6. |
| 15:20 | <gsnedders> | They were intending on fixing places where hasLayout became black-box observable, expecting to continue on the same fairly quick release-cycle IE had had at that point. |
| 15:20 | <gsnedders> | (Fairly quick at least compared with IE today) |
| 17:05 | <annevk> | Mike[tm]: I just filed a bug on appcache and it did not go to public-html automatically but instead was filed under other drafts... |
| 17:05 | <annevk> | Mike[tm]: seems wrong |
| 17:50 | <heycam> | annevk, according to CSSOM, should document.styleSheets include items for <link rel="stylesheet alternate"> even if they don't have a title=""? |
| 17:53 | <annevk> | probably should not include those |
| 17:53 | <annevk> | I think that might not get past the "create a style sheet" steps |
| 17:53 | <annevk> | or should not, anyway, because it's useless |
| 17:55 | <heycam> | annevk, where is "create a style sheet" actually invoked? |
| 17:56 | <heycam> | annevk, for link elements that is? I can see there's a section for <?xml-ss?> and Link headers... |
| 17:56 | <annevk> | in HTML I think |
| 17:57 | <annevk> | though I'm not sure if Hixie is doing that properly |
| 17:57 | <heycam> | mm |
| 17:58 | <heycam> | an IE demo fails in Firefox because it doesn't expose all the title-less alternate stylesheet <link>s in document.styleSheets, so I wanted to point to some spec justification for that |
| 17:59 | <annevk> | yeah I heard about that the other day |
| 18:00 | <annevk> | style sheet loading, creating of StyleSheet objects, etc. is all very sketchy :( |
| 18:00 | <heycam> | mm |
| 18:00 | <heycam> | :) |
| 18:02 | <annevk> | are you eating something? |
| 18:02 | <annevk> | mm |
| 18:05 | <heycam> | annevk, no, just sleepy after flight so I have a limited set of irc responses to use :) |
| 18:06 | <annevk> | mm |
| 18:29 | <kennyluck> | Is it a known issue that validator.nu throws more "parse error" than it should be for character references? |
| 18:41 | <annevk> | kennyluck: you can check its bugs no? |
| 18:43 | <kennyluck> | annevk, it will be easier if someone says "yes" to my question and then I'll skip this. ;) |
| 18:44 | <annevk> | I don't know the answer :) |
| 18:45 | <annevk> | ew http://wiki.whatwg.org/wiki/StringEncoding |
| 18:47 | <zewt> | only particularly "ew" thing is "detectEncoding", which is being dropped |
| 18:48 | <zewt> | rest is rough but straightforward |
| 19:33 | <annevk> | other feedback: |
| 19:33 | <annevk> | appcache does not mean "works offline" |
| 19:33 | <annevk> | even if it did, users don't get it |
| 19:34 | <annevk> | they search google for URLs so a) they can't find the apps when offline and b) they don't know they work offline |
| 19:34 | <zewt> | then why is it in a section titled "offline web applications"? heh |
| 19:34 | <annevk> | ironyfreezone |
| 19:36 | <Philip`> | Is it more like "works when the internet connection temporarily drops because you're going through a tunnel but you don't need to notice"? |
| 19:38 | <annevk> | the thing you want is e.g. angry birds offline |
| 19:38 | <annevk> | but users won't remember that URL |
| 19:38 | <Philip`> | Sounds like we need an offline Google |
| 19:39 | <Philip`> | Shouldn't be too bad if it just searches through every page you've already visited |
| 19:39 | <zewt> | since when do users remember any URLs, that's not the user's job |
| 19:44 | <annevk> | I guess you need to get to the point where the application is "installed" so you can launch it from where you launch other apps |
| 19:45 | <zewt> | chrome can do that (bookmark webpages to the start menu like an app), but it's really obscure |
| 19:51 | <zewt> | "create application shortcuts" |
| 21:40 | <annevk> | http://tantek.com/2011/238/b1/many-ways-slice-url-name-pieces is pretty awesome |
| 22:09 | <annevk> | http://lists.w3.org/Archives/Public/public-webapps/2009JulSep/1358.html |
| 22:17 | <annevk> | http://dev.w3.org/2006/webapi/WebIDL/#es-attributes |
| 23:00 | <annevk> | we should organize a WHATCAMP; a place for developers and standards people to meet |
| 23:00 | <annevk> | (not my idea) |
| 23:01 | <annevk> | I get the feeling there's a lot of subtle things we can change that would improve the web for everyone |
| 23:01 | <annevk> | having met with some developers in the past week |
| 23:08 | <smaug____> | annevk: hopefully somewhere in Europe |
| 23:08 | <smaug____> | but yes, sounds like a good idea |
| 23:09 | <zcorpan> | Örebro maybe? :-) |
| 23:09 | <smaug____> | Opera could arrange it... Oslo perhaps? I haven't been there :) |
| 23:12 | <smaug____> | might be easier to arrange it in Brussels close to Fosdem, like couple of days before Fosdem |
| 23:12 | <annevk> | that does not sound too bad |
| 23:13 | <annevk> | I think we should have it in the bay area too, but experimenting closer to home works better for me :) |
| 23:14 | smaug____ | is all for trying to make technology world less US centric |
| 23:15 | <smaug____> | annevk: perhaps Opera and Mozilla Europe could arrange something together. |
| 23:16 | <zcorpan> | what do you guys think about http://www.w3.org/Bugs/Public/show_bug.cgi?id=14364 ? |
| 23:16 | <annevk> | yeah |
| 23:17 | <annevk> | zcorpan: there's a lot of interest here |
| 23:17 | <annevk> | zcorpan: pretty much everyone dealing with appcache also wants dynamic; but it's kind of dinstict |
| 23:18 | <annevk> | zcorpan: appcache is atomic, this is somewhat separate from that |
| 23:18 | <annevk> | wycats had these notes: |
| 23:18 | <annevk> | window.localResources.add(url, callback) => window.localResources.remove(url) => window.localResources.update(url, callback) |
| 23:18 | <annevk> | window.localResource.on("update", callback) |
| 23:19 | <annevk> | I think sicking thinks this should be solved using Indexed DB |
| 23:19 | <annevk> | download first and then store it in the database |
| 23:19 | <annevk> | but that seems less convenient as you cannot keep using URLs at that point |
| 23:20 | <annevk> | zcorpan: does that help? |
| 23:21 | <zcorpan> | so sicking thinks appcache shouldn't be used at all? |
| 23:22 | <annevk> | app cache is what is required to make your app run offline |
| 23:22 | <annevk> | resources on top of that such as media or large amounts of email is typically stored in different ways |
| 23:23 | <annevk> | e.g. for email attachments it seems sensible to store them in Indexed DB |
| 23:23 | <annevk> | but there is probably room between the two for cached URLs |
| 23:24 | <annevk> | tantek: hey |
| 23:24 | <annevk> | tantek: so I'm just using macports |
| 23:24 | <tantek> | greetings |
| 23:24 | <tantek> | oh ok |
| 23:26 | <annevk> | tantek: updated the spec just now |
| 23:26 | <annevk> | hg clone, hg pull, hg update, hg commit, hg push |
| 23:27 | <annevk> | it's "fairly" trivial |
| 23:27 | <tantek> | ah ok - sounds not too different than git |
| 23:30 | <smaug____> | obviously tantek should do some gecko development to get used to hg |
| 23:30 | <annevk> | he does work for Mozilla... :p |
| 23:30 | <smaug____> | I know :) |
| 23:32 | <annevk> | btw tantek, see the logs for "WHATCAMP"; standards/developers/browsers |
| 23:34 | <tantek> | ooh that sounds like a good idea |
| 23:34 | <tantek> | annevk - I took a look at the draft, we want the OWFa statement to autoupdate as well |
| 23:34 | <tantek> | s/4 November 2011/[DATE: 01 Jan 1901] |
| 23:34 | <tantek> | smaug :) |
| 23:35 | <zcorpan> | ok who will be the one to save <hgroup>? |
| 23:35 | <tantek> | zcorpan - I'm considering it |
| 23:35 | <annevk> | tantek: oh I didn't realize that |
| 23:35 | <tantek> | annevk - that way it indicates that we've agreed to OWFa on the very latest draft |
| 23:36 | <annevk> | also fixed |
| 23:36 | <tantek> | I like <hgroup> semantically, and am unsure about the usefulness of its impact on the outline algorithm |
| 23:36 | <tantek> | nice |
| 23:37 | <zcorpan> | tantek: cool |
| 23:37 | <tantek> | I'm considering writing up a change proposal just to keep the semantics part, and drop the impact on the outline algorithm. |
| 23:37 | <tantek> | I'm wondering if there are enough people that would care however. |
| 23:37 | <tantek> | Because if I'm the only one that cares then it doesn't make sense to save hgroup. |
| 23:38 | <zcorpan> | wait what? impacting the outline algorithm is the semantic <hgroup> has! |
| 23:38 | <annevk> | tantek: I want to use it for my specs at least |
| 23:39 | <annevk> | tantek: afaik people want to keep it and I doubt it will be dropped across both specs... |
| 23:39 | <zcorpan> | nobody has written a no-change CP |
| 23:41 | <tantek> | zcorpan - the use-case of having it impact the outline algorithm which then would affect things like tables of contents - is not described as it is used in practice. |
| 23:41 | <tantek> | e.g. in a book with a title with subhead, or a chapter title with subhead, the text of both is grouped and shown together, rather than hidden |
| 23:41 | <tantek> | rather than the subhead being hidden that is |
| 23:42 | <tantek> | zcorpan - if you'd like to work together on a change proposal - I'm for it |
| 23:42 | <tantek> | feel free to get started and I'll contribute |
| 23:42 | <zcorpan> | the subhead doesn't need to be hidden in the ToC, i think |
| 23:43 | <zcorpan> | the <hgroup> is the heading, and it just gets its *rank* from the highest ranked h* child |
| 23:43 | <zcorpan> | what text to use in the ToC is not defined |
| 23:44 | tantek | is working on a few change proposals already (enhanced time, keep/introduce data, etc.) |
| 23:45 | <tantek> | zcorpan that wasn't my understanding - but I can re-read the spec. |
| 23:46 | <zcorpan> | hmm. seems i was wrong. |
| 23:46 | <zcorpan> | "For the purposes of document summaries, outlines, and the like, the text of hgroup elements is defined to be the text of the highest ranked h1–h6 element descendant of the hgroup element, if there are any such elements, and the first such element if there are multiple elements with that rank. If there are no such elements, then the text of the hgroup element is the empty string." |
| 23:46 | <zcorpan> | i think that should be considered a UI issue and not restricted like that |
| 23:46 | <tantek> | zcorpan - what you described above (not quoted) is the effect I want because that's what I've seen happen in outlines/ToCs |
| 23:46 | <tantek> | zcorpan - I would be happy with that too. |
| 23:50 | <zcorpan> | tantek: if the effect hgroup has was removed from the outline algo altogether, then <hgroup><h1></h1><h2></h2></hgroup> would create two *sections*, which is not wanted |
| 23:50 | <tantek> | zcorpan - aha I see what you mean |
| 23:51 | <tantek> | In practice, semantic subheads are joined with their heading using some sort of combination character. |
| 23:51 | <tantek> | e.g. |
| 23:51 | <tantek> | http://en.wikipedia.org/wiki/The_Diamond_Age |
| 23:51 | <zcorpan> | i don't have time right now to write a CP (have to sleep), but i'd suggest no-change CP except loosen the paragraph quoted above |
| 23:53 | <zcorpan> | (rationale: what text to use in ToC etc is a UI issue, and the spec usually doesn't restrict UI choices) |
| 23:54 | <tantek> | effectively has <hgroup><h1>The Diamond Age</h1> <h2>Or, A Young Lady's Illustrated Primer</h1></hgroup> |
| 23:54 | <annevk> | back later |
| 23:54 | <tantek> | but when the h1 and h2 are combined, they're combined with a ":" |
| 23:54 | <tantek> | The Diamond Age: Or, A Young Lady's Illustrated Primer |
| 23:54 | <tantek> | ttyl annevk |
| 23:55 | <zcorpan> | yep |
| 23:55 | <tantek> | thus it would be particularly useful if <hgroup> allowed for specifying how to join the headings when making a single string |
| 23:55 | <zcorpan> | that smells like overengineering to me :) |
| 23:55 | <tantek> | I don't have a specific suggestion except perhaps a combiner attribute that takes a string to insert between the different heading children |
| 23:55 | <tantek> | perhaps |
| 23:55 | <tantek> | we can mention it as considered and rejected too. |
| 23:56 | zcorpan | would keep it simple |
| 23:56 | <zcorpan> | if you write a CP, i'm happy to review it tomorrow |
| 23:57 | <zcorpan> | or maybe i can write one tomorrow if you don't |
| 23:57 | <tantek> | zcorpan - tomorrow is fine - go ahead and start one |
| 23:57 | <tantek> | I would but I'm already working on two |
| 23:57 | <zcorpan> | k |
| 23:57 | <zcorpan> | g'night |