| 00:35 | <Hixie> | location.reload() always does a GET? |
| 03:42 | <bosky101> | please leave in your comments for http://news.ycombinator.com/item?id=652202 - a library launched by my startup today |
| 03:42 | <bosky101> | and would anyone knwo the channel for talk on the semantic web |
| 03:42 | <jcranmer> | don't we hate that here? |
| 03:43 | <bosky101> | in anycase i'm not sure in which cateogry start.hover.in lies in ...its basically about links using the URI to render behaviour within a webpage |
| 03:44 | <bosky101> | <a href="http://onhover.hover.in/hoverlet/hover.in/relatedyoutube/election">; election videos</a> will therefore render youtube videos when you hover over the link |
| 03:44 | <bosky101> | available events are onhover and onclick |
| 03:45 | <bosky101> | and there are around 20 kinds of return types available now, looking to build a community of it |
| 10:53 | <RdeWilde> | Q: Anyone can tell me where's the right place to ask this question, or has an answer to it? I'd like to know if this header: http://pastebin.ca/1456435 is somehow legal http 1.1 (as in specs). Tried to figure out, but not sure yet.. |
| 11:16 | <archtech> | which part of a tag is considered the "tag body" actually: <foo>bar</baz> |
| 11:16 | <archtech> | it is <foo> or bar |
| 11:16 | <jgraham> | archtech: WWhat are you quoting? |
| 11:16 | <jgraham> | I don't recall hearing the term "tag body" |
| 11:21 | <archtech> | jgraham, kinda myself. I was trying to cauge if the term is intuitive |
| 11:21 | <archtech> | jgraham, it's confusing to talk about "content type of the content, and content type of the attributes" |
| 11:21 | <archtech> | content type of the body and the attributes, now that is kinda more natural |
| 11:25 | <archtech> | gauge I mean |
| 11:26 | jgraham | has never found himself talking about the "content type" of the content or attributes |
| 11:28 | <archtech> | jgraham, well, you know, inline, block, literal... |
| 11:28 | <archtech> | say <script> is a literal |
| 11:28 | <archtech> | an implicit CDATA if you will |
| 11:36 | <jgraham> | Oh content model. |
| 11:36 | jgraham | should have inferred that |
| 12:13 | annevk2 | reads backlog; spelling, startups, and a question about HTTP... |
| 12:16 | <jgraham> | annevk2: See /topic |
| 12:17 | <annevk2> | I lost that long ago |
| 12:17 | <annevk2> | In other news, http://norman.walsh.name/2009/06/10/ajax is interesting considering the Oracle proposals in WebApps |
| 12:25 | <MikeSmith> | annevk2: yeah |
| 12:27 | jgraham | wonders what the Oracle proposals are |
| 12:28 | <gsnedders> | Yay. Adobe have now cancelled my order for not providing proof of eligibility, despite it waiting around on their website the entire time. |
| 12:28 | <gsnedders> | grumblegrumblegrumble… |
| 12:35 | <annevk2> | hmm, I now get Swedish spam |
| 13:12 | gsnedders | would expect them to only cancel an order after the time limit if it wasn't waiting for _them_ to respond |
| 13:21 | gsnedders | is meant to provide non-sensitive photo ID |
| 13:21 | <gsnedders> | The only photo ID I have is a passport which is sesitive… |
| 13:21 | gsnedders | grumbles… |
| 13:22 | beowulf | wonders what a non-sensitive photo id would be? |
| 13:23 | <Lachy> | gsnedders, how do you define what is and isn't a sensitive photo ID? |
| 13:23 | <gsnedders> | Lachy: http://kb2.adobe.com/cps/407/kb407013.html |
| 13:23 | <beowulf> | maybe it's a photo he's not sensitive about? you know, nice haircut, cheeky grin |
| 13:27 | gsnedders | considers taking the fuck-it option and submitting a copy of passport… |
| 13:29 | <hsivonen> | gsnedders: yay for trying to be a paying customer |
| 13:29 | <gsnedders> | hsivonen: And then I'll end up pirating because they refuse to take my order! |
| 15:16 | hsivonen | wonders what kind of UI OOo will have for the ODF 1.2 fork of RDFa |
| 15:16 | <annevk42> | office formats are such a mess |
| 15:17 | <annevk42> | people say HTML is bad, but seriously |
| 15:18 | jgraham | bets on a UI that eposes a feature set that could easilly be replicated by a much simpler underlying technology |
| 15:19 | <hsivonen> | Word has a very CSS-like layout model, but the UI for tweaking the properties is horribly badly exposing the underlying model |
| 15:20 | <annevk42> | From CSS WG meetings I get the impression MS Office uses the same layout engine as MSIE |
| 15:20 | <Rik|work> | I'd like a word processor that would help to do structure first then style, like what good web devs do |
| 15:21 | <annevk42> | You can do that with MS Office |
| 15:21 | <jgraham> | Right but almost no-one actually does |
| 15:21 | <annevk42> | Just takes a bit of learning. Personally I write my documents in plain text or HTML :) |
| 15:22 | <Rik|work> | annevk2: you can, but the UI doesn't encourage you to do so |
| 15:22 | <Lachy> | jgraham, I do |
| 15:22 | <Lachy> | that is, when I use Word, which I don't do so frequently anymore |
| 15:24 | <Lachy> | the problem with Word is, from what I've seen of Word 2007, is that they took a backwards step with regards to encouraging structure-based editing instead of presentational editing |
| 15:27 | <jgraham> | Lachy: You don't really count as the inverse of "almost no-one" |
| 15:28 | <jgraham> | Plus, even hen people do use it, the temptation to use the bold and italics buttons is still rather strong e.g. the ES5 fraft has a bunch of things that could usefully be some sort of inline style but afaict aren't |
| 15:28 | <jgraham> | (it's quite hard to tell becase the document is read-only and I don't know how to change that) |
| 15:29 | <jgraham> | (and it would be useful if they were some sort of inline class and if that got preserved in the HTML export because then it would be easier to hypertextulise the document in a post-hoc fashion) |
| 15:30 | <Lachy> | I never had any temptation to use them for normal editing. I only used them when customising the style of the available structures |
| 15:33 | jgraham | realises that the file system permissions on the document were set to read only |
| 15:37 | <Lachy> | It's interesting that these two articles about the same scientific calculations chose to focus on two different aspects. 1) Mars being ejected from the solar system and 2) Mars colliding with Earth. - and both articles mention the other issue as a minor note. |
| 15:37 | <Lachy> | http://arstechnica.com/science/news/2009/06/kicking-a-planet-out-of-the-solar-system-physically.ars |
| 15:37 | <Lachy> | http://news.bbc.co.uk/2/hi/science/nature/8093005.stm |
| 15:38 | <Lachy> | what's even more interesting is that both conclusions being so radically different from each other indicate that the scientists really have no clue what's going to happen. |
| 15:39 | <hsivonen> | whee! Windows CE C++ compiler claims "UNICODE" to itself as a reserved word |
| 15:39 | <hsivonen> | or built-in constant |
| 15:40 | <annevk42> | Lachy, they are about the same article and seem to report different aspects of it |
| 15:41 | <Lachy> | annevk42, yeah, that's basically what I said |
| 15:41 | <annevk42> | Lachy, the BBC focuses on the anomaly and Ars on what is more likely to happen |
| 15:41 | <annevk42> | Lachy, no, you said it indicates scientists have no clue what is going to happen which seems like the wrong conclusion |
| 15:41 | <Lachy> | no, read the first sentence of what I wrote |
| 15:44 | <jgraham> | Actually the "no clue" thing is probably more accurate. But not in a negative way |
| 15:45 | <jgraham> | You can run these models and get an answer but they are obviously rather densitive to the initial conditions which you can't know precisely, and the exact dynamics, which you can't know precisely |
| 15:45 | <jgraham> | So the best that you can do is get a grasp on some of the possibilities |
| 15:46 | <hsivonen> | If I want to get proper hosting for a hg repo, what do I need to know about Google Code and BitBucket? |
| 15:46 | <jgraham> | and understand that things are far from being as simple as "all the planets go around in nice perfer ellipses from now until the end of time" |
| 15:46 | <hsivonen> | should I consider something else, too? |
| 15:47 | <jgraham> | hsivonen: AFAIK html5lib has not had any serious problems using Google code. Although the issue with the @ sign in email addresses is annoying. |
| 15:48 | <jgraham> | You get 1GB of space by default (I think) |
| 15:48 | <jgraham> | The main restriction is that you have to choose one of their approved licenses |
| 15:48 | <hsivonen> | oh, right. that |
| 15:48 | <hsivonen> | V.nu is a bit of a mix of licenses |
| 15:49 | <annevk42> | put it in the public domain! |
| 15:49 | <Lachy> | what's the issue with the @ sign in email addresses? |
| 15:49 | <hsivonen> | annevk42: It's not because of me that it is a mix of licenses |
| 15:50 | <hsivonen> | Lachy: the supposedly copy-pasteable snippet is bogus and you need to double-escape @ in your hgrc |
| 15:50 | <jgraham> | hsivonen: bitbucket is 150Mb free and you have to pay for more. Although there might be some open source plans if you contact them |
| 15:51 | <hsivonen> | jgraham: have you used bitbucket? |
| 15:51 | <hsivonen> | does the Google Code license policy extend to dependency jars? |
| 15:51 | <jgraham> | hsivonen: No. I was thinking of moving html5lib just as google announced hg support |
| 15:52 | <hsivonen> | are other projects violating or following rules by checking in differently-licensed 3rd party jars? |
| 15:53 | jgraham | is tempted to clone the html5lib repository on bitbucket anyway |
| 15:55 | <ezyang> | hsivonen: You could cheese it and have your build script download the jars |
| 15:56 | <hsivonen> | ezyang: I still want to check in MIT and BSD code to one repo |
| 15:56 | <ezyang> | Oh, ok, so it's not just JAR problems |
| 15:56 | <hsivonen> | ezyang: also, the C++ translator uses the Mozilla Tri-license |
| 15:57 | <annevk42> | the ToS doesn't say much about having to adhere to license policies |
| 16:00 | gsnedders | needs to get email working on VPS |
| 16:11 | <hsivonen> | is HTML+TIME used on public sites at all? |
| 16:11 | <hsivonen> | is it used on intranets? |
| 16:13 | <Lachy> | is it supported by any browser? |
| 16:14 | <Lachy> | or at least a plugin? |
| 16:28 | <annevk42> | Lachy, IE |
| 16:30 | gsnedders | laughs at Hixie's reply to him last night :) |
| 17:30 | <hober> | http://upload.wikimedia.org/wikipedia/commons/e/e4/Web_development_timeline.png |
| 17:32 | <gsnedders> | That says quite a bit about how client side open web stagnated |
| 17:33 | <hober> | It's weird that the XHTML2 WG is on there, but not the HTML WG (nor WHATWG). Then again, it's odd the XHTML2 WG is on there at all. |
| 17:34 | <gsnedders> | Where's XHTML2 on that? |
| 17:34 | <gsnedders> | Oh, there, at the top |
| 18:59 | <annevk42> | weird that Opera is placed so late |
| 19:46 | gsnedders | ponders whether to get open or closed headphones |
| 19:47 | gsnedders | isn't sure whether he really wants the bulk of closed ones, but could probably do with no sound leakage |
| 19:48 | <jgraham> | gsnedders: For what purpose |
| 19:49 | <gsnedders> | jgraham: For listening to music when I can't practically carry speakers around with me :P |
| 19:50 | <jgraham> | gsnedders: Right. But are you going for "I have to work in an office and I want to cut out background noise" headphones or "I'm a wanker^H audiophile headphones and I want the best possible sound quality" |
| 19:50 | <jgraham> | ? |
| 19:50 | <gsnedders> | jgraham: the latter. |
| 19:50 | <gsnedders> | jgraham: But the former is useful too :P |
| 19:50 | <jgraham> | Oh well I can't help you then. |
| 19:50 | <gsnedders> | Hence my issue. |
| 19:50 | <gsnedders> | jgraham: Are you allowed headphones there? |
| 19:50 | <jgraham> | Yes |
| 19:51 | <gsnedders> | Then I guess I probably should get closed ones :P |
| 19:51 | <gsnedders> | (More for the sound leakage POV than incoming sound) |
| 19:54 | <jacobolus> | gsnedders: get open headphones |
| 19:54 | <jacobolus> | gsnedders: in fact, just go pick out the set of grados that are in your budget :) |
| 19:54 | <gsnedders> | jacobolus: :) |
| 19:54 | <jacobolus> | gsnedders: these are great, I think http://www.gradolabs.com/product_pages/sr60.htm |
| 19:55 | <jacobolus> | definitely for the price |
| 19:58 | <gsnedders> | jacobolus: I had been looking at the Sennheiser HD 595 or 555 |
| 20:00 | <archtech> | Can someone comment on what is going on with XHTML2? |
| 20:00 | <annevk42> | http://twitter.com/jdowdell/statuses/2120530436 -- isn't this exactly what we're doing? |
| 20:01 | <annevk42> | archtech, this is not the right place for that question, we're not working on it |
| 20:01 | <archtech> | annevk2, all right. I was just wondering if it was abandoned completely in favor of HTML5.. |
| 20:02 | <Hixie> | archtech: no, there is still a working group working on xhtml2 |
| 20:02 | <Hixie> | archtech: i believe they are even planning on releasing an updated draft soon |
| 20:03 | <archtech> | Hixie, I see. |
| 20:03 | gsnedders | thought they were planning on releasing an updated draft before TPAC last year |
| 20:03 | <archtech> | Is HTML5 intending to move towards paragraphs not just for inline content? |
| 20:04 | <archtech> | Like XHTML2 is, apparently. I can't understand that concept. |
| 20:04 | <Hixie> | archtech: you mean like <p> containing <ol>? |
| 20:05 | <archtech> | Hixie, yes, that is one of the examples in the XHTML2 draft it seems |
| 20:05 | <archtech> | However I have yet to see a documentation system that has this kind of concept. |
| 20:05 | <Hixie> | archtech: we actually tried that a few years ago, but we couldn't get it to work in a way compatible with browsers today |
| 20:06 | <archtech> | Hixie, if compatibility wasn't a problem, I wonder if it would work conceptually? |
| 20:06 | <Hixie> | archtech: yeah, seems like a pretty reasonable idea (though it makes WYSIWYG editing much harder to do) |
| 20:09 | <archtech> | Does LaTeX define something like that? |
| 20:09 | <Hixie> | no idea |
| 20:09 | <archtech> | Thanks. |
| 20:09 | <gsnedders> | jacobolus: Peh. Looking at a Grado compared with the Sennheiser, the Senheiser does better at the bass end, but then the Grado has too much treble and the Sennheiser too little :( |
| 20:10 | <gsnedders> | archtech: LaTeX has no concept of a list being part of a paragraph, from memory. |
| 20:10 | <jacobolus> | gsnedders: what do comparison reviews say? |
| 20:11 | <gsnedders> | jacobolus: That there's little to choose between them :P |
| 20:11 | <gsnedders> | (in terms of quality) |
| 20:11 | gsnedders | is tempted by the Sennheiser mainly because it is more comfy |
| 20:11 | <jacobolus> | gsnedders: if you want good bass, you should go for the SR80s, and buy an amp to plug in between your laptop/ipod/whatever and your headphones |
| 20:12 | <gsnedders> | jacobolus: The Sennheiser HD 595 seems to do better with bass with/without amp |
| 20:12 | <gsnedders> | jacobolus: But I don't care that much about bass alone |
| 20:14 | <gsnedders> | Generally I'm still tempted by the Sennheiser |
| 20:14 | <jacobolus> | go for it!' |
| 20:14 | <jacobolus> | I'm sure you'd be plenty happy with them |
| 20:15 | <jacobolus> | those are pretty pricey, no? |
| 20:15 | <gsnedders> | No |
| 20:15 | <archtech> | gsnedders, it seems yes, LaTeX defines lists as their own block elements. |
| 20:15 | <archtech> | I don't know. |
| 20:15 | <gsnedders> | Oh, wait, from having been looking at both the HD 595 and HD 555 I've got mixed up between the two :P |
| 20:15 | <archtech> | Maybe it's a better model to keep it as is, anyway. |
| 20:16 | <jacobolus> | (well, by my super-cheap standards; amazon has $260 marked down to 178) |
| 20:16 | <gsnedders> | archtech: LaTeX is really aimed to get the semantics needed to print correctly, it isn't concerned with actual document semantics, so it makes no sense for LaTeX to support both. |
| 20:16 | <jacobolus> | archtech: LaTeX doesn't care about semantic purity |
| 20:16 | <archtech> | It doesn't, but semantics bear some overlap with presentation. |
| 20:17 | <archtech> | I can't conceptually wrap my mind around a list in a paragraph. |
| 20:17 | <jacobolus> | but inline lists don't need special structure in latex |
| 20:17 | <archtech> | At least the way literary paragraphs are defined. |
| 20:17 | <gsnedders> | The HD 555 I can get quickly (which is good seeming I leave soon). |
| 20:17 | <gsnedders> | And there isn't overly much difference between them and the HD 595 (which cost almost twice as much) |
| 20:17 | <gsnedders> | (£68.41 v. £112.95, which is a lot less than twice) |
| 20:18 | gsnedders | was rounding both numbers down in his head to nearest multiple of ten :P |
| 20:25 | <gsnedders> | jgraham: There again, I guess people at work would prefer people to have closed headphones :P |
| 20:29 | <shelleyp> | I've posted an email about the WAI CG Consensus Resolutions on Text alternatives in HTML 5 to the HTML WG email list. Hopefully we can pull together the threads from both IRC and twitter into one place |
| 20:32 | <archtech> | Is HTML5 going to tweak the definition of inline/block or leave it as is? |
| 20:32 | <archtech> | I see some voices that say this should be purely a styling concept. |
| 20:32 | <archtech> | And not an element concept. |
| 20:34 | <rubys> | hsivonen: ping? |
| 20:38 | <annevk42> | archtech, HTML5 does not mention inline/block-level |
| 20:38 | <archtech> | annevk2, as in? It doesn't define them or it does away with them? |
| 20:39 | <annevk42> | archtech, it groups elements using different terms |
| 20:39 | <annevk42> | archtech, so yeah, it does away with them, effectively |
| 20:39 | <archtech> | annevk2, which section should I check for more info on this? |
| 20:39 | <annevk42> | seems section 3.4 is about this |
| 20:40 | <archtech> | Thanks :) |
| 20:49 | <beowulf> | so, nice weather out... |
| 20:50 | <beowulf> | the heat generated from accessibility discussions makes me try very hard to not use the img element |
| 20:51 | <jgraham> | gsnedders: Given where you are likely to be sitting I can sey with some certianty that there will be a non zero number of people who will not want to hear your music |
| 20:52 | <jgraham> | Especially if you start listening to christian rock |
| 20:52 | gsnedders | was expecting that :P |
| 20:54 | <archtech> | annevk2, I can't find how is the <br> tag classified in this section? |
| 20:54 | <jgraham> | Hixie: I don't see how we get to tell people waht is intuitive |
| 20:55 | <jgraham> | We have to guess what is intuitive to make the most usable language possible |
| 20:55 | <Hixie> | language design is all about making choices based on what we think is more intuitive |
| 20:55 | <Hixie> | it's like UI design |
| 20:55 | <jgraham> | Sure. That's not the same as telling people what is intuitive |
| 20:55 | <annevk42> | archtech, br is both flow and phrasing |
| 20:55 | <Hixie> | to some level you can test it, but at the end of the day it's a judgement call, even if you base it on some level of testing |
| 20:55 | <Hixie> | how is it not? |
| 20:55 | <annevk42> | archtech, the element definition lists the categories it is in |
| 20:56 | <annevk42> | archtech, the other way around might be added at some point, but that is currently not in the draft |
| 20:56 | <archtech> | isn't it confusing to have each element in multiple categories though? |
| 20:56 | <jgraham> | Hixie: Because "telling people" implies some level of control over the definition of "intuitive". We have no control, we can only infer what is intuitive (e.g. by studying behaviour) and specify from that |
| 20:56 | <annevk42> | HTML4 had a similar concept, called flow |
| 20:57 | <archtech> | I mean, it's useful for classification in a manual, but seems to not be best suited for implementing HTML5 parsers. |
| 20:57 | <archtech> | annevk2, sure, but it was simpler |
| 20:57 | <jgraham> | (the flow of information is in the opposite direction) |
| 20:57 | <annevk42> | seems natural for things to fit in multiple categories |
| 20:57 | <annevk42> | you don't implement HTML5 parsers based on content models :) |
| 20:58 | <annevk42> | you use the section that specifies parser requirements for that |
| 20:58 | <annevk42> | 9.2 iirc |
| 20:58 | <Hixie> | jgraham: my interpretation was that the guy meant that we don't get to decide on what we think is intuitive |
| 20:59 | <jgraham> | Hixie: That wasn't really my interpretation. At least I think your reply is more confusing than helpful |
| 20:59 | <Hixie> | jgraham: ah well. |
| 20:59 | <Hixie> | jgraham: what did you think he was saying then? |
| 21:00 | <jgraham> | Hixie: I thought he was trying to say "as language designers you can't mandate that something ought to be intuitive even if user behavior shows otherwise" |
| 21:01 | <Hixie> | yay, shelley spilt the beans on where the html4all cabal moved to |
| 21:01 | <Hixie> | jgraham: that seems like a truism, how can anyone mandate what is intuitive? |
| 21:03 | <jgraham> | Hixie: They can't obviously. But they could think that they can. Lots of bad UI designers do, for example :) |
| 21:03 | <jgraham> | s/do/think that/ |
| 21:03 | <Hixie> | i don't think they do, i think they just do exactly what we do, which is make assumptions about what is intuitive |
| 21:03 | <jgraham> | s/think that/seem to think that/ |
| 21:03 | <Hixie> | i don't think they they they do either |
| 21:04 | <Hixie> | er |
| 21:04 | <Hixie> | i don't think they think they do either |
| 21:04 | <jgraham> | Well it's not really important. The point is that it is possible to think that you can do the impossible |
| 21:04 | <jgraham> | (you might back it up with arguments like "we can use education to make people understand our model") |
| 21:05 | <Hixie> | education can't affect what's intuitive, by definition :-) |
| 21:06 | <jgraham> | Right but you often hear people say "x is intuitive once you understand principle y" which can be true or can be false. So people think that by teaching y they can make x intuitive |
| 21:07 | <Hixie> | i guess |
| 21:07 | <jgraham> | (but this is a bit of the point I was originally trying to make) |
| 21:07 | <jgraham> | (which is that I think the comment was meant in a different way to the one in which you interpreted it) |
| 21:08 | <Hixie> | that's certainly possible |
| 21:09 | <archtech> | Is this valid according to HTML5: <b>foo<br>bar</b>? |
| 21:09 | <archtech> | This is a bold tag that crosses theboundary of a line, although that line is an implicit tag |
| 21:09 | <archtech> | defined by a break |
| 21:10 | <Hixie> | aww, their archives are member-only http://wilbur.bytowninternet.com/mailman/listinfo/talk_html4all.org |
| 21:10 | <Hixie> | i guess the cabal doesn't believe in the level of transparency we believe in! |
| 21:10 | <Hixie> | what could they possibly have to hide |
| 21:11 | <jgraham> | I guess irony is lost on some people |
| 21:15 | <Hixie> | right, bbiab, lunch |
| 21:17 | <jacobolus> | archtech: why wouldn't it be? |
| 21:19 | <archtech> | jacobolus, the line break defines two implicit line elements, say I have this: Hello, this is <b>some bold text,<br>and we continue on a new line.</b> Done. |
| 21:19 | <archtech> | Now those are two lines |
| 21:19 | <archtech> | <line>Hello, this is <b>some bold text,</line><line>and we continue on a new line.</b> Done.</line> |
| 21:19 | <archtech> | There's no valid way to nest this. |
| 21:20 | <jacobolus> | the same could be said about the '.' character creating two implicit 'sentence' elements |
| 21:20 | <jacobolus> | etc. |
| 21:20 | <archtech> | My point is that a break is a metacharacter. |
| 21:21 | <archtech> | The actual objects are lines, are they not? |
| 21:21 | <jacobolus> | I don't think so, no |
| 21:21 | <archtech> | I see. |
| 21:21 | <jacobolus> | but I don't really know the browser internals that well |
| 21:22 | <jacobolus> | in any case, no, html5 doesn't to my knowledge mandate that "implicit line elements" not included anywhere in the dom be treated in any special fashion |
| 21:34 | <archtech> | thanks |
| 21:38 | <shelleyp> | Hixie, html4all cabal? Isn't that a pretty silly thing to say? |
| 21:39 | <shelleyp> | You know, you might try communicating with people, rather than make snarky asides in the IRC |
| 21:48 | <annevk42> | jacobolus, per CSS you do get multiple line boxes, but you can get that with a single element as well |
| 21:48 | <annevk42> | (e.g. by just having line wrapping; <br> is nothing more than a forced wrap) |
| 21:49 | <annevk42> | (that's a lie to be honest, <br> has a few other tricks up its sleave) |
| 21:49 | <jacobolus> | annevk42: so you can style a particular line (besides the first)? |
| 21:49 | <jacobolus> | I mean, of course at some level lines, glyphs, etc. are objects |
| 21:50 | <annevk42> | there are proposals |
| 21:50 | <jacobolus> | but are the opposite sides of a <br> actually treated as special, "implicit objects"? |
| 21:50 | <Philip`> | hsivonen: I've used BitBucket to a trivial extent, and it worked trivially and I haven't seen any problems |
| 21:51 | <annevk42> | jacobolus, no, <br> in that sense is just a forced linebreak |
| 21:56 | <Hixie> | shelleyp: i was just using the term that html4all participants themselves used for themselves: http://html4all.org/pipermail/list_html4all.org/2007-August/000004.html |
| 21:57 | <Hixie> | shelleyp: and i have been talking directly to everyone involved; i would involve myself in more discussions but since e.g. html4all is a closed group, it is not possible for me to do so |
| 21:59 | <shelleyp> | Hixie, the context of your use of the term belies your later justification |
| 22:00 | <annevk42> | html4all has two forums I think shelleyp |
| 22:00 | <annevk42> | one public and one private |
| 22:00 | <annevk42> | the former is no longer in use and is the one Hixie is pointing to |
| 22:00 | <Hixie> | i really was just using the term because that's how the group of people involved self-identifies |
| 22:00 | <shelleyp> | OK, and since the html4all group does not have controlling interest over any aspect of any W3C document, what's the point annevk42? |
| 22:00 | <Philip`> | hsivonen: (I suppose there's the question of what you mean by "proper hosting", seeing as Hg isn't reliant on a central repository like SVN and so the requirements can be looser) |
| 22:00 | <Hixie> | if you have a problem with the way the people you are associating with self-identify, i urge you to take it up with them |
| 22:01 | <annevk42> | shelleyp, point? |
| 22:01 | annevk42 | was just trying to explain something |
| 22:01 | <shelleyp> | annevk42, Ah, OK, thanks for the explanation. |
| 22:02 | <Philip`> | hsivonen: (e.g. maybe the goal is that people can write "hg clone some-url-which-won't-change-ever", in which case BitBucket is nice because it can apparently do CNAMEs and you could move the same domain to a different host later) |
| 22:03 | <shelleyp> | Hixie, why you're concerned about this one group, I have no idea. There's plenty of people to communicate with on the HTML WG. |
| 22:04 | <Hixie> | i'm not in the least bit concerned about the html4all group |
| 22:04 | <shelleyp> | But communication has to be more than telling people that no, their idea won't work, and here's what you're going to do instead |
| 22:04 | <Hixie> | i agree entirely |
| 22:05 | <shelleyp> | Actually, I don't want to come across as pugnacious, but I'm not sure that you do agree with me |
| 22:05 | <Hixie> | i agree entirely that ignoring feedback and simply asserting positions is not a good way to participate in any group |
| 22:06 | <shelleyp> | Hixie, one could say that what you just described is quite similar to many of your recent communications with folks |
| 22:06 | <Hixie> | indeed many do, i'm sure |
| 22:06 | <Hixie> | mr last week for instance |
| 22:07 | <Hixie> | though it's interesting that he hides behind a screen of anonymity to do so |
| 22:07 | <shelleyp> | I'm not talking about an anonymous individual. |
| 22:08 | <Hixie> | he is merely one example among many, indeed |
| 22:08 | <Hixie> | maybe even dozens :-) |
| 22:08 | <shelleyp> | No, he's not. The folks I'm talking about are those who participate in the groups, and identify themselves by name |
| 22:09 | <Hixie> | i am talking about all the people who claim that i ignore them despite my writing e-mails literally hundreds of lines long explaining my positions based on detailed research and reasoning |
| 22:10 | <Hixie> | mostly, if not exclusively, people who have found themselves not agreeing with my reasoning, sadly |
| 22:10 | <shelleyp> | Hixie, take a break, have a beer, come back and read the line you just wrote. You just about summed up the entire problem. |
| 22:10 | <Hixie> | i don't drink and i'm on a break right now, but thanks for the suggestion :-) |
| 22:10 | <jacobolus> | shelleyp: but quite often what happens, that I've observed, is that people re-write the same emails, completely ignoring Hixie's responses, and then pretend that he's ignoring them |
| 22:11 | <jacobolus> | it's frustrating even to just watch from the sidelines |
| 22:11 | <Hixie> | the sad truth of the matter is that on many topics, there are people with a variety of opinions, and i could never make the spec agree with them all |
| 22:11 | <Hixie> | so merely disagreeing with any one group doesn't mean i'm ignoring them, it merely means that they were not the group that was most convincing |
| 22:12 | <shelleyp> | Hixie, no one says you should. But when you're the only author of the spec, your being in a position of power means you have to go the extra mile |
| 22:12 | <Hixie> | it's not even that i'm just doing what i want -- there are a variety of things in the spec where the group that "won" is actually the group i disagree with |
| 22:12 | <Hixie> | (many things, in fact) |
| 22:12 | <shelleyp> | jacoblus, then that's incorrect -- people should specifically address Ian's comments in response |
| 22:12 | <hober> | He's already going the extra distance-unit-much-longer-than-a-mile |
| 22:14 | <hober> | I mean, seriously, exchanges of the form 1. initial feedback, 2. hixie replies with what's changed, rationale, etc., 3. "new" feedback which doesn't take anything in (2) into account and is in fact indistinguishable from (1) happen all too frequently |
| 22:15 | <hober> | Most people would stop replying with (2) at all, because all the work going into it doesn't seem to have any effect. Clearly Hixie isn't normal, to continue to engage like he does. |
| 22:15 | <shelleyp> | hober, Hixie has worked his butt off, and I appreciate his effort. But Ian also has a habit of discounting other's expertise, which is disrespectful |
| 22:16 | <shelleyp> | I only checked in because I was not happy to see the term "cabal", especially related to a hot topic like accessibility |
| 22:16 | <shelleyp> | I'm sure this group would rather get back to tech stuff |
| 22:16 | <shelleyp> | But I will remember what was said about people not responding to Ian's comments, and make sure that I do |
| 22:18 | <sicking_> | shelleyp, what i feel happens a lot is that people join and say "Do X. You should listen to me because i'm an expert". What would be much more useful is people saying "here is the argument for why you should do X" |
| 22:21 | <annevk42> | I did not know: http://www.olsenhome.com/gif/ |
| 22:22 | <annevk42> | I'm not sure what to do with this knowledge though, pronouncing it as "gift" seems more understandable... |
| 22:22 | <shelleyp> | sicking: I've not see that behavior. Can you point out where someone said that, specifically? |
| 22:23 | <Philip`> | annevk42: That's just prescriptivist pronunciation - you should ignore it and say it the sensible way with a hard 'g' |
| 22:24 | <sicking> | shelleyp, every time someone tries to argue from authority. If you have expertise you shouldn't have to point it out i'd think. It should be obvious from your well layed out arguments |
| 22:29 | <jruderman> | the simple solution is to switch from GIF to PNG ... except that PNG can also be pronounced in two ways |
| 22:30 | <annevk42> | yeah, they should've called it PING |
| 22:30 | <hober> | PNG gets pronounced like "ping" (and GIF like "jiff") |
| 22:31 | <hober> | What's the other way to say PNG? "pong"? |
| 22:31 | <annevk42> | s/gets/is supposed to be/ |
| 22:31 | <gavin_> | pee en gee |
| 22:31 | <annevk42> | that's how I call it usually |
| 22:32 | <Philip`> | You could pronounce it like "snog" but with a p |
| 22:34 | <gsnedders> | Philip`: Instead of the s? I can't get it out as one syllable then :P |
| 22:34 | <Hixie> | "spog"? |
| 22:34 | <Hixie> | "snpg"? |
| 22:36 | <Philip`> | gsnedders: Yes |
| 22:37 | <Philip`> | gsnedders: It's not any harder than plenty of syllables in other languages :-p |
| 22:50 | <shelleyp> | sicking: your argument makes no sense. I, for one, am more likely to listen to someone with proven experience and expertise in the field, then someone who doesn't |
| 22:51 | <sicking> | shelleyp, right, but to most people on the list, someone saying that they have expertise is not a "proven expert" |
| 22:52 | <sicking> | shelleyp, since they haven't spent time in that field |
| 22:52 | <shelleyp> | sicking: again, I've not seen anyone imply an expertise they don't have -- not in any of the debates I've followed |
| 22:53 | <sicking> | shelleyp, the word "expertiese" have been brought up several times. Though it could be that most of the time in the form of "people on this list isn't listening to experts" |
| 22:53 | <sicking> | shelleyp, i don't want to get into a meta discussion here though, that's just a time waster |
| 22:54 | <shelleyp> | sicking: I agree. If you find specific examples, twitter or email me with them. |
| 22:54 | <sicking> | shelleyp, next time i see it i'll email it to you |
| 22:55 | <shelleyp> | sicking: thanks. |
| 22:58 | <sicking> | shelleyp, i'll note though, that the page you linked to today, from WAI CG, contains very little rationale |
| 22:58 | <sicking> | shelleyp, i happen to agree with most of it, but that's beside the point |
| 23:00 | <sicking> | shelleyp, either they think that the rationale is obvious, which it IMHO it isn't, or they are arguing that we should listen to them because they have expertise. Though I'm hoping it's because of some other reason that I can't think of |
| 23:02 | <arun_> | I have a rough editor's draft of the File API spec, which I'd welcome discussion on. There are a few issues with it still (Callbacks *and* ErrorCallbacks have to be handled; editorial nits; sample code nits) but I'd love to get feedback from the channel: http://dev.w3.org/2006/webapi/FileUpload/publish/FileAPI.xhtml |
| 23:03 | <annevk42> | .xhtml? o_O :p |
| 23:03 | <arun_> | annevk42, it was just convenient to edit. |
| 23:03 | <arun_> | I'll change that |
| 23:03 | <annevk42> | I was joking, no worries |
| 23:04 | <arun_> | In particular, I'd like to hear about FileDialog. I think it's a good idea, but I've had nits to the effect that honing <input type="file" ... > behavior is better. |
| 23:05 | <annevk42> | "This interface describes a single file in a FileList," I think something else for this would work better since we might not always want to have them in lists |
| 23:05 | <sicking> | yeah, i had some comments like that too |
| 23:05 | <Hixie> | to repeat the comments i made privately: you'll want to make better use of rfc2119, and i recommend splitting File into two, FileData with everything but the name, and File with the name, where File inherits from FileData |
| 23:05 | <sicking> | it also assumes that lists come from the FileDialog |
| 23:06 | <annevk42> | would FileData be like Blob? |
| 23:06 | <sicking> | that was my question too |
| 23:07 | <sicking> | i'm not sure if Blob will have a mimetype though? |
| 23:08 | <Hixie> | File doesn't have a mime type either right now does it? |
| 23:08 | <arun_> | Hixie, agree on rfc2119. I'm amenable to splitting File into two. You'd want a slice() on FileData at some point, I take it, for "blobs"? |
| 23:08 | <annevk42> | I'd rather tie it to HTMLInputElement and DataTransfer than to FileDialog to be honest |
| 23:08 | <Hixie> | i agree that whatever File inherits from shouldn't have a type |
| 23:08 | <sicking> | Hixie, hmm.. good point, i think they should |
| 23:09 | <sicking> | Hixie, or rather, in the firefox implementation they do |
| 23:09 | <Hixie> | arun_: yeah that would be one of the ways this would get extended |
| 23:09 | <Hixie> | arun_: similarly binary data coming from a WebSocket would be a FileData (or Binary or Blob or whatever you prefer), not a File |
| 23:09 | <sicking> | Hixie, oh, no, they don't |
| 23:09 | <annevk42> | spawning file picker dialogs to users seems open to abuse and we already have somewhat sandboxed way for picking files we can reuse |
| 23:09 | <arun_> | OK; you also suggested that obliging developers to handle both the data callback and the error callback was a stylistic choice. No strong opinions either way? |
| 23:10 | <sicking> | Hixie, ah, now i remember, we figured it was mostly useful for piecing together a data: uri, but since we expose a data: uri we didn't feel that mimetype was that important |
| 23:11 | <Hixie> | arun: that's just an api design decision, no right answer really. personally i design callback-style apis such that they call either the success or the failure callbacks, never both, but it depends on the way you'd expect them to be used. |
| 23:11 | <annevk42> | while we brainstorm, it would be sort of nice to just like with XMLHttpRequest you can get a Document out of it and maybe even HTMLImageElement / ImageData / JSON |
| 23:12 | <Hixie> | getting an HTMLImageElement would be nice, but long term what would be even nicer is being able to get a short (not data:) URL out of it |
| 23:12 | <arun_> | annevk42, I agree it is open to abuse. I had an ednote about conditions where implementations would want to define when to raise a FileError with a SECURITY_ERR, and FileDialog abuse seemed like a prime choice. |
| 23:12 | <sicking> | annevk42, since it's an async API, the UA can show a unintrusive warning if it suspects misuse. A'la GeoLocation |
| 23:12 | <annevk42> | sicking, but why have two ways to do things? |
| 23:13 | <Hixie> | e.g. .getAsUrl() and it returns "magic-file-or-whatever://origin.example.com/0"; where that URL works only for browsing contexts in the same origin |
| 23:13 | <arun_> | Basically, major web applications like Flickr and GMail use the Flash uploader. Sure, we could address some of this by changing how user agents handle input type="file", but a programmatic FileDialog seemed useful. |
| 23:13 | <sicking> | annevk42, because a lot of sites are working around the uglyness of the current <input type=file> |
| 23:13 | <Hixie> | that would let us automatically take a FileData object and plug it into anything like <video> or <img> or whatever |
| 23:13 | <sicking> | annevk42, gmail being an obvious example |
| 23:13 | <annevk42> | I'd argue we should solve that with CSS, not ECMAScript |
| 23:14 | <annevk42> | the reason Gmail/Flickr use Flash for uploading is mostly multiple files (<input type=file multiple>) and I suppose async upload (using XMLHttpRequest) |
| 23:14 | <arun_> | annevk42, it's not just ugliness, it's also functionality, but HTML5 does allow input type="file" to have multiple selects, etc. I'm in two minds about this. |
| 23:14 | <sicking> | annevk42, no-one has been able to crack that nut for many many years now |
| 23:14 | <annevk42> | did we even try? |
| 23:15 | <annevk42> | for many years progress on CSS has been so-so |
| 23:15 | <sicking> | styling UI elements have been attempted many times i think |
| 23:15 | <sicking> | webkit is currently attempting scrollbars |
| 23:15 | <sicking> | there has been suggestions of various pseudo-elements |
| 23:15 | <arun_> | Hixie, wouldn't that mean defining yet another URI scheme? |
| 23:16 | <sicking> | Hixie, an alternative would be to have streams |
| 23:16 | <Hixie> | arun_: yes |
| 23:16 | <Hixie> | sicking: well i imagine we'd need streams to use the same feature |
| 23:17 | <arun_> | Hixie, that doesn't seem desirable. I'd like that use case better articulated. |
| 23:17 | <sicking> | Hixie, i agree it's something we should look into, not sure if it needs to be part of this spec though. At least not initial version |
| 23:17 | <Hixie> | sicking: unless we want to add an API to everything that accepts a url :-) (e.g. background-image in CSS, list-style-image... SVG...) |
| 23:17 | <Hixie> | sicking: absolutely, not for this version |
| 23:17 | <annevk42> | Hixie, how can the URL be shorter than with data URLs? |
| 23:17 | <Hixie> | annevk42: it wouldn't include the data |
| 23:18 | <Hixie> | annevk42: just a locally minted pointer to the data |
| 23:18 | <sicking> | Hixie, agreed, we might need both |
| 23:18 | <Hixie> | annevk42: this would not be the kind of URL that ever goes over the wire |
| 23:18 | <annevk42> | oh I see |
| 23:18 | <arun_> | Hixie, hmmm... now I get that better |
| 23:18 | <sicking> | annevk42, magical elves |
| 23:18 | <annevk42> | fancy |
| 23:18 | <sicking> | annevk42, with pixie dust |
| 23:18 | <sicking> | it's pretty awesome |
| 23:18 | <annevk42> | hixie dust |
| 23:18 | <Hixie> | anyway i wouldn't expect this for this version |
| 23:19 | <Hixie> | but i think it's a better solution long term than getting HTMLImageElements out of FileData objects |
| 23:19 | <sicking> | yeah |
| 23:19 | <sicking> | i think we can stick to data-uris for now |
| 23:19 | <arun_> | OK, this is good stuff. I'll leave FileDialog in for now, but clean up File a bit more, maybe break it into two (File, FileData or something) |
| 23:19 | <sicking> | otherwise we'll have to define things like what File.magicalHTMLImageProducer().src returns and such |
| 23:20 | <annevk42> | i'm still opposed to FileDialog fwiw |
| 23:20 | <annevk42> | I'd rather wait and see if <input type=file multiple> + XHR can do the trick |
| 23:20 | <arun_> | annevk42, so's Hixie. But I'm unmoved for now ;-) |
| 23:20 | <sicking> | annevk42, XHR?? |
| 23:20 | <Hixie> | i'm not a fan of FileDialog either, i think leaving the security to the last minute is a bad way of doing things |
| 23:20 | <sicking> | annevk42, you mean CSS? |
| 23:20 | <Hixie> | i don't mind something new, but security needs to be first and foremost in mind |
| 23:21 | <sicking> | Hixie, how is it different from geolocation? |
| 23:21 | <Hixie> | such that the entire design precludes abuse |
| 23:21 | <annevk42> | sicking, people articulated the desire for async upload / fragment upload |
| 23:21 | <arun_> | Hixie, I'm not sure we're quite leaving it to the last minute; we're discussing it out of the starter gate :) |
| 23:21 | <Hixie> | sicking: geolocation doesn't pop up a modal dialog |
| 23:21 | <sicking> | Hixie, neither would this need to |
| 23:21 | <arun_> | Hixie, even here, we're NOT suggesting a modal dialog |
| 23:21 | <arun_> | Hixie, in fact I mention a non-modal one |
| 23:21 | <Hixie> | arun_: what i mean is that you have the API designed before the security solution |
| 23:21 | <sicking> | Hixie, no, we talked about this a lot |
| 23:22 | <sicking> | Hixie, not sure if all possible implementation strategies ended up in the draft, but we have several in mind |
| 23:22 | <Hixie> | arun_: what flash does is a modal dialog. If it's supposed to be something else, then that's better, but the spec needs to make that crystal clear, not just leave it up to hte UA. |
| 23:22 | <Hixie> | arun_: and if it's not what flash does, then it's not clear it addresses the same use case :-) |
| 23:23 | <arun_> | Hixie, ahh, there are two things here. 1. The actual File Picker and 2. The user interaction for permission. But yeah, that must be made clearer. |
| 23:23 | <sicking> | Hixie, what I want to do is to leave it up to the UA to decide what to do, but give a few non-normative possible strategies that the implementation can use |
| 23:24 | <annevk42> | modal or non-modal it's a new API for authors, new UI for users (sounds more complex already) and more complexity for implementors |
| 23:24 | <annevk42> | for something that's already possible |
| 23:25 | <arun_> | annevk42, good points -- I think really explaining the use case will get us on the way to resolution. |
| 23:25 | <annevk42> | the main use case I'm hearing is styling |
| 23:25 | <annevk42> | i'm not at all convinced that's a good reason |
| 23:26 | <annevk42> | there's been no attempt whatsoever by any browser at letting authors style the file input control |
| 23:26 | <arun_> | annevk42, we could say that the *existing* input type ="file" mechanism in browsers today is ugly and causes developers of big web apps to use Flash. FileDialog is one solution, and doesn't need to be bound to using an element in your application. |
| 23:26 | <annevk42> | people use Flash because they can allow users to pick multiple files |
| 23:26 | <annevk42> | that's far more a reason than styling |
| 23:26 | <sicking> | annevk42, It's a complaint I hear a lot, that people think that <input type=file> is ugly and wanting to use something else. |
| 23:26 | <arun_> | But multiple selections (as spec'd in HTML5) is a big reason. |
| 23:27 | <sicking> | annevk42, i don't think it's the whole truth that people use flash just for multiple |
| 23:27 | <arun_> | annevk42, yes -- multiple selection is a big one, and if that's the prime use case, I agree that it can be solved within input type = "file" ... but I'm not sure it is |
| 23:28 | <annevk42> | sicking, yes, but address the styling problem where it should be addressed |
| 23:28 | <sicking> | annevk42, first of all I think that many people aren't able to select multiple files in a filepicker anyway (requires keyboard shortcuts on windows) |
| 23:28 | <annevk42> | arun_, my suggestion is to first fix multiple selections and async upload for files and then see if it's good enough |
| 23:28 | <annevk42> | sicking, no, you can just drag around with your mouse iirc |
| 23:29 | <sicking> | annevk42, ah, yeah, if files have neighboring names |
| 23:29 | <annevk42> | for photo sites that seems likely |
| 23:29 | <sicking> | annevk42, true |
| 23:30 | <sicking> | annevk42, for gmail, not so sure |
| 23:30 | <annevk42> | depends on what people attach primarily |
| 23:30 | <annevk42> | hixie would be in a better place to answer that one I guess :) |
| 23:30 | <sicking> | annevk42, and again, i'm not convinced people generally know about that |
| 23:30 | <sicking> | annevk42, we can of course ask google :) |
| 23:31 | <annevk42> | but you're talking about users, i was suggesting why web developers want it |
| 23:31 | <sicking> | i mean we can ask google why they use flash |
| 23:31 | <sicking> | i.e. what made them not consider HTML good enough |
| 23:31 | <annevk42> | when i log in to gmail it doesn't use flash |
| 23:34 | <Lachy> | wow, so html4all moved to a whole new secret list now? That explains why I've had no mails from the old list for a few months |
| 23:34 | <annevk42> | but yeah, there should definitely be substantive evidence that we really need to introduce another API for opening a file dialog |
| 23:34 | <annevk42> | imo |
| 23:35 | <sicking> | arun_, it occurred to me that we should specify the .files property on HTMLInputElement as well though |
| 23:35 | <sicking> | arun_, if you aren't already |
| 23:36 | <Hixie> | i'll do that in html5 |
| 23:36 | <Hixie> | i've been waiting for this spec |
| 23:36 | <Hixie> | i have a whole folder of issues to deal with that are blocked on not having this spec |
| 23:38 | <sicking> | Hixie, ah, cool |
| 23:40 | <ojan> | annevk42: I see a flash uploader in gmail. |
| 23:41 | <annevk42> | I don't :) |
| 23:42 | <sicking> | annevk42, do you have flash installed and enabled? They detect if you don't and then just use a normal <input type=file> |
| 23:42 | <annevk42> | yes, when I go to youtube.com I can see videos of weird people |
| 23:43 | <sicking> | weird people = flash. Check! |
| 23:43 | <sicking> | annevk42, note that you have to click on 'attach' for any of this to happen |
| 23:43 | <annevk42> | I know |
| 23:43 | <annevk42> | I checked the source |
| 23:43 | <annevk42> | <input id=":jz" name="f_fvu266070" class="LIODqc" size="50" tabindex="2" type="file"> |
| 23:43 | <sicking> | annevk42, and that there won't be any weird people indicating flash on gmail |
| 23:43 | <sicking> | annevk42, dunno, appears they're failing to detect flash for you then |
| 23:44 | <sicking> | annevk42, WFM |
| 23:44 | <arun_> | annevk42, for me too. It seems to invoke Flash pretty consistently. |
| 23:44 | <arun_> | annevk42, and the experience when i turn off Flash is dramatically different. |
| 23:45 | <arun_> | annevk42, though I should say, I'm a weird person. |
| 23:45 | <annevk42> | to test this I'm using Firefox 3.0 for Ubuntu which is the only browser where I have Flash |
| 23:45 | <arun_> | annevk42, OK, I'll bet there's a "Flash version doesn't do foo on Ubuntu" issue which causes you to always default to the HTML input element for file selection. |
| 23:47 | <annevk42> | anyway, whether Gmail does or does not use Flash does not really sway my opinion one way or another :) |
| 23:49 | <arun_> | annevk42, it shouldn't. i just want a good sense for *why*. Sure, there's multiple selection and style, but I'm interested in other features. |
| 23:50 | <arun_> | So this has been great. I'm going to tweak it and report back :) |
