| 00:14 | <dolske> | Hixie: why do media elements not get UA-provided controls unless the |controls| attribute is explicitly set? |
| 00:14 | <dolske> | it seems to be a problem people keep stumbling over, and feeds have an annoying habit of stripping it too. |
| 00:16 | <Hixie> | how else would you do it? |
| 00:16 | <Hixie> | basically it was the least bad solution we could find |
| 00:17 | <dolske> | have UA-controls be the default, unless a nocontrols attribute was set? |
| 00:18 | <Hixie> | how would you handle the case where nocontrols is set but js is disabled? |
| 00:19 | <dolske> | have the UA ignore the attribute, and provide controls anyway? |
| 00:19 | <annevk5> | I think it was mostly this way because the original design didn't have controls at all. |
| 00:20 | <Hixie> | dolske: so you would have elements with nocontrols="" have controls? |
| 00:21 | <dolske> | only when JS is disabled, since the page obviously can't provide its own. |
| 00:21 | <Hixie> | seems like the same as a controls="" attribute, except for being a negative attribute (historically a source of confusion) and one that isn't always accurate |
| 00:23 | <dolske> | Doesn't seem any worse than today with no controls attribute set and JS disabled... I'd expect a UA to enable the default controls in that case too. |
| 00:23 | <Hixie> | right but at least the source doesn't lie about what's going on there :-) |
| 00:24 | <Hixie> | it's early days yet, i'd expect feed sanitisers to fix their code in due course |
| 00:24 | <Hixie> | i'm surprised they're letting <video> through at all, that's probably a sign of a serious security bug |
| 00:24 | <dolske> | It seems a little less ambigious when JS is *enabled*, though... with "controls" you can't know for a fact that the page is going to supply its own, maybe something stripped the attribute. It seems unlikely for someone to accidentally add a "nocontrols" attribute. |
| 00:25 | <Hixie> | i don't see why anything would strip the controls="" value specifically but not add a nocontrols="" attribute unless there is already a security bug, as noted |
| 00:27 | gsnedders | wonders how to make php html5lib more conformant without losing perf. |
| 00:27 | <dolske> | I suppose, although there's still author error... |
| 00:27 | gsnedders | wonders about turning the whole tokenizer into a switch($state) statement, hence avoiding function call overhead (which is non-neglible) |
| 00:29 | <annevk5> | the UA should probably have a right click menu with options regardless |
| 00:41 | <Hixie> | yeah the browser should always make the controls available, even if the author said not to -- the attribute is just a way to say that the browser should definitely provide them |
| 00:41 | <Hixie> | as opposed to only if the user wants them |
| 00:41 | <Hixie> | that's another reason i don't really like "nocontrols" |
| 00:41 | <Hixie> | it's more like "no-controls-unless-you-want-them" |
| 04:38 | <zcorpan__> | "Yalin Wang and Jianying Hu have written an interesting paper on table |
| 04:38 | <zcorpan__> | type detection. Here's a link: http://portal.acm.org/citation.cfm?coll=GUIDE&dl=GUIDE&id=511478 They developed an algorithm and tested it on about 14000 tables." - http://markmail.org/message/mmlgkpck35u66qjp |
| 04:47 | <zcorpan> | hmm, not free to read the paper |
| 05:06 | <zcorpan> | http://www.w3.org/mid/1242739380.4531.275.camel@localhost |
| 06:25 | <heycam> | close ACTION-2561 |
| 06:25 | <heycam> | sorry, ww |
| 08:11 | <Philip`> | zcorpan: http://www.research.ibm.com/people/j/jyhu/www2002.pdf |
| 08:14 | <zcorpan> | Philip`: thanks |
| 08:19 | <hsivonen> | whoa. I've failed to reference to email where James Clark calls the BS on XSD in my thesis |
| 08:21 | <hsivonen> | hmm. that sentence needs some parentheses to parse correctly |
| 08:21 | <hsivonen> | I've failed to reference in my thesis the email where James Clark calls the BS on XSD |
| 08:22 | <hsivonen> | this one http://www.imc.org/ietf-xml-use/mail-archive/msg00217.html |
| 08:23 | <hsivonen> | Hixie: were you planning on writing a microdata version of ccREL? |
| 08:24 | <Hixie> | no, as discussed here: http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-May/019668.html |
| 08:24 | <Hixie> | though it has been suggested that being able to link a rel=license (or similar) directly to an <img> or <video> would be beneficial (though without the extra baggage of ccREL, we don't want to be interpreting licenses in microdata or anything) |
| 08:24 | <Hixie> | dunno if that will involve microdata though |
| 08:24 | <Hixie> | i don't really see that it can to be honest |
| 08:25 | <Hixie> | well i guess it could |
| 08:25 | <Hixie> | but it'd be a bit awkward |
| 08:25 | <Hixie> | i dunno |
| 08:25 | <Hixie> | we'll see |
| 08:25 | <hsivonen> | Hixie: it seems to me that attribution-name, attribution-url and license could be properties on an item |
| 08:25 | <hsivonen> | and more-rights |
| 08:26 | <Hixie> | i don't see any reason to make the attribution-name and attribution-url machine readable |
| 08:26 | <Hixie> | such a thing would become intractable quickly and is not really needed by most licenses anyway |
| 08:26 | <Hixie> | (most licenses need the license and/or copyright and/or disclaimer included, not the attribution) |
| 08:27 | <Hixie> | (and i don't see why any of that is of benefit to computers searching through this stuff) |
| 09:05 | <hsivonen> | is it intentional that accesskey allows astral characters? |
| 09:07 | <Hixie> | yes |
| 09:08 | <hsivonen> | ok. thanks |
| 09:08 | <pesla> | np |
| 09:08 | <Hixie> | hsivonen: you might want to warn about it of course :-) |
| 09:08 | <Hixie> | "this seems unlikely to be a useful key" |
| 09:10 | <zcorpan> | what's Document.load? |
| 09:10 | <zcorpan> | seems to be undefined in my browsers |
| 09:12 | <Hixie> | ok time for bed |
| 09:12 | <Hixie> | nn |
| 09:19 | hsivonen | thought both Gecko and Opera had DOM Load/Save |
| 09:20 | <hsivonen> | from the era when W3C specs were considered presumptively righteous and MS inventions like XHR presumptively dirty |
| 09:23 | <zcorpan> | Hixie: the character separating prefix and property should be the underscore in order to find the definition using google |
| 09:24 | <jgraham> | Bt oh boy would that be ugly |
| 09:24 | <jgraham> | *But |
| 09:24 | <jgraham> | Also: the paper on table classification looks rather interesting |
| 09:26 | <zcorpan> | it would be interesting to look at tables that have nested tables but otherwise has the characteristics of a genuine table |
| 09:42 | <zcorpan> | http://gsnedders.html5.org/outliner/process.py?url=http%3A%2F%2Fwww.w3.org%2F - alt text is not used |
| 09:44 | jgraham | becomes slightly nervous of the conclusions in the paper when it starts talking about <ROWSPAN> and <COLSPAN> tags |
| 09:45 | <jgraham> | Also, it would be interesting to know how biased they are by working with the Java swing XML parser which presumably does not implement robust HTML parsing |
| 09:47 | <Philip`> | Also this was in 2002, when the world was different |
| 09:47 | <zcorpan> | figure 2b has empty tbody elements |
| 09:49 | <jgraham> | Philip`: Maybe someone has doen a followup? Can you search for papers that cite this one? |
| 09:50 | <Philip`> | jgraham: Can't you search for them? :-p |
| 09:50 | <Philip`> | http://portal.acm.org/citation.cfm?coll=GUIDE&dl=GUIDE&id=511478#citedby |
| 09:50 | <Philip`> | http://scholar.google.com/scholar?num=30&hl=en&lr=&safe=off&cites=1801764763188325408 |
| 09:51 | <Philip`> | etc |
| 09:53 | <jgraham> | Philip`: Maybe. I don't know what ACM lets you do without paying but I don know that you can do whatever :) |
| 09:54 | <Philip`> | I don't have any special access now since I'm still at home :-) |
| 09:54 | <jgraham> | Philip`: You could proxy yourself through cam though :) |
| 09:54 | <Philip`> | The page zcorpan linked to has all the information, and only the PDF itself is restricted, as far as I know |
| 09:55 | <Philip`> | jgraham: I could, but I'm too lazy to do that :-p |
| 09:57 | <zcorpan> | jgraham: the google search result points to the pdfs directly |
| 09:57 | <Philip`> | (Also, I haven't figured out how to set up an actual proxy, so I use rdesktop to download the files locally on the remote server, and then copy via sftp) |
| 10:00 | <jgraham> | Philip`: Can't you do something like ssh -D (mor options that I can't remember) |
| 10:03 | jgraham | takes this oppertunity (again) to point out how insane the idea of pay-per-subscription scientific journals are |
| 10:03 | <Philip`> | jgraham: I could, but the last half of your command line is where the problem is |
| 10:04 | <Philip`> | (Also, configuring the browser to use that proxy requires knowledge and effort) |
| 10:04 | <hsivonen> | my hypothesis is that university libraries could run the system more cheaply by paying the salaries of the journal editors than by subscribing to the journals |
| 10:04 | <Philip`> | It does seem a bit silly when you can just Google for the paper's title and download the PDF directly |
| 10:05 | <Philip`> | and I've only seen that method fail with with really ancient (like 1990s) papers |
| 10:05 | <hsivonen> | the whole system seems to be supported by the inertia of universities insisting on giving more academic credit for publishing in one of these behind-the-paywall journals |
| 10:05 | <jgraham> | Philip`: I used the method in http://ubuntu.wordpress.com/2006/12/08/ssh-tunnel-socks-proxy-forwarding-secure-browsing/ |
| 10:06 | <Philip`> | jgraham: Oh, that doesn't sound too complex |
| 10:07 | <hsivonen> | the inertia with Type 3 fonts in LaTeX-originating PDFs is great, too |
| 10:08 | <jgraham> | It seems like the paper above has ~15% of tables being "genuine" tables whereas there is a Google (and others) paper that finds ~1.1% are "relational tables" |
| 10:08 | <jgraham> | (the latter apper is called "WebTables: Exploring the Power of Tables on the Web" if you want to find it) |
| 10:09 | <roc> | hsivonen: the academics are actually changing relatively fast |
| 10:09 | <roc> | see the recent MIT decision for example |
| 10:09 | <hsivonen> | roc: what did MIT decide? |
| 10:09 | <jgraham> | hsivonen: The problem with fonts in LaTeX is that almost no LaTeX users understand how TeX fonts work |
| 10:09 | <roc> | basically the entire faculty is only allowed to publish in journals that allow public access |
| 10:09 | <hsivonen> | roc: cool |
| 10:09 | <jgraham> | hsivonen: e.g. me |
| 10:10 | <hsivonen> | the anecdotes I've heard recently have been about water engineering, marine biology and cancer research |
| 10:10 | <roc> | also you see on a regular basis the entire editorial board of a 'normal' pay-per-view journal resigning and setting up a new open-access journal with a slightly different name |
| 10:11 | <Philip`> | The instructions I got for some paper submission thing was they required a PS version of the paper as well as PDF, because most LaTeX users put the wrong type of fonts in the PDF |
| 10:11 | <hsivonen> | roc: nice |
| 10:12 | <hsivonen> | too bad ACM, Elsevier and Springer still hold the locks to back catalog |
| 10:16 | <jgraham> | roc: I have never seen that. Also making faculty do something is interesting because (in many cases) the faculty don't have to keep publishing much |
| 10:17 | <hsivonen> | (the anecdotes I mentioned above were anecdotes about paywall craziness) |
| 10:19 | <roc> | here's the MIT thing |
| 10:19 | <roc> | http://www.bitsbook.com/2009/03/mit-adopts-an-open-access-policy/ |
| 10:20 | <hsivonen> | thanks |
| 10:20 | <jgraham> | roc: Oh it doesn't actually require that they publishin an open access journal. It just requires that they *also* publish the information in the (open access) MIT repository |
| 10:21 | <jgraham> | Which is not that different to putting the article on arXiv |
| 10:21 | hsivonen | wonders why MIT ops to disallow selling the articles for profit |
| 10:21 | <roc> | yes but the journal has to *allow* open access |
| 10:21 | <hsivonen> | if they are already available on the Web, why bother disallowing business model innovation that could spread the ideas even further |
| 10:22 | <roc> | most traditional journals don't |
| 10:22 | <roc> | many of them have been turning a blind eye to personal Web publication |
| 10:23 | <jgraham> | roc: Where does it say that? Of course the journal has to allow you to also put the article in another place but since they were all allowing that for arXiv anyway it doesn't seem that helpful |
| 10:23 | <jgraham> | (maybe arXiv isn't so prevelant in other fields so there it might be a big change) |
| 10:23 | <roc> | most traditional journals, at least in CS-related fields that I'm familiar with, technically required you to sign over copyright, at which point you lose the right to publish your article on your own Web site |
| 10:23 | <hsivonen> | so the policy is a great improvement, but I don't understand MIT's fascination with the *-nc flavors of CC licensing and now adopting a policy that doesn't allow the Provost to apply Free licenses to the papers |
| 10:24 | <roc> | it's true that people were just ignoring that |
| 10:25 | <roc> | but the MIT policy is flat-out incompatible with copyright assignment to the journal, unless the journal explicitly allows the open-access MIT publication |
| 10:25 | <jgraham> | roc: IIRC the blackwells publishing copyright form that I had to sign used to say that you were signing all the rights to blackwells and that you should try to disseminate the paper using arXiv |
| 10:25 | <jgraham> | I never understood how the cluses were compatible |
| 10:25 | <jgraham> | *clauses |
| 10:25 | <roc> | me neither |
| 10:26 | <hsivonen> | It's insane that strictly speaking one needs to ask a journal company for permission to publish a compilation of articles as a PhD thesis |
| 10:26 | <jgraham> | I think recently they might have changed it so that you retained your own copyright |
| 10:27 | <roc> | could be, I've been out of the loop |
| 10:27 | <roc> | anyway, licensing and other details aside, there's a clear desire by MIT faculty and various other academic groups to make their work open access |
| 10:50 | <zcorpan> | hsivonen: when elements get cloned, lots of unhelpful messages are generated |
| 10:51 | <zcorpan> | hsivonen: consider <div><div><div><ul><b><li></ul>x</div>x</div>x</div> |
| 10:52 | <zcorpan> | hsivonen: all "End tag div seen but there were unclosed elements." messages are just adding noise |
| 10:54 | <zcorpan> | hsivonen: or <ul><b><li></ul>x<p><p><p> |
| 10:56 | <zcorpan> | hsivonen: http://www.chasejobs.co.uk/index.php?option=com_catsone has such markup |
| 10:59 | zcorpan | filed a bug |
| 10:59 | <gsnedders> | zcorpan: The spec doesn't say to use alt text. Blame Hixie |
| 11:00 | <zcorpan> | gsnedders: does the spec say to use textContent? |
| 11:00 | <jgraham> | Doesn't the spec just say to use textContent |
| 11:00 | <jgraham> | Oh, you already asked that |
| 11:01 | <gsnedders> | zcorpan: No, it just says to use the element. Doesn't say how. |
| 11:01 | <gsnedders> | zcorpan: And textContent seems a sane soltion |
| 11:01 | <zcorpan> | gsnedders: well then it's your bug to not show the image or its text equivalent |
| 11:02 | <gsnedders> | zcorpan: Not showing it seems perfectly reasonable though |
| 11:02 | <zcorpan> | i disagree :) |
| 11:04 | <zcorpan> | it seems reasonable to be able to use <h1><img src=... alt="Foo"></h1> and expect the image or "Foo" to show up in the outline |
| 11:04 | <jgraham> | zcorpan: It seems entirely unreasonable for an image to show up in an outline :) |
| 11:05 | <jgraham> | (But quite reasonable for "Foo" to show up) |
| 11:05 | <gsnedders> | zcorpan: Get a better textContent spec'd :P |
| 11:06 | <zcorpan> | 1. Replace all descendant "img" elements with a new text node that has the .data set to the element's alt attribute. 2. Use textContent |
| 11:07 | <jgraham> | zcorpan: "Replace" is reeally the wrong word. You don't want to mutate the document :) |
| 11:08 | <zcorpan> | jgraham: for gsnedders' purpose, mutating the document seems ok, but he can implement it in another way that is equivalent :) |
| 11:09 | <gsnedders> | In Anolis I think I do actually copy the nodes already for TOC |
| 11:10 | <hsivonen> | hrm. the simples microdata to XML mapping I can think of would use OPMLish text-in-attributes antipattern |
| 11:11 | <hsivonen> | because XML element content can't distinguish between empty string and absent string |
| 11:14 | gsnedders | laughs at Hixie getting a reference to Avenue Q in yet another email |
| 11:19 | <zcorpan> | jgraham: without mutation: The textContentWithAlt attribute, on getting, must return a concatenation of the data of all the descendant Text nodes, as well as the value of the alt attribute of all the descendant img elements, of the context node, in tree order. |
| 11:20 | <zcorpan> | where attributes are said to be in no namespace and elements in the xhtml namespace |
| 11:29 | <zcorpan> | hmm, document.body.onload doesn't seem to be the same as window.onload in gecko and webkit |
| 11:32 | <zcorpan> | Hixie: the live dom viewer seems to give js error in ie8 |
| 11:35 | <zcorpan> | window.onload and document.body.onload seems to be the same in ie |
| 11:50 | <Philip`> | gsnedders: I vote in favour of cross-document cross-reference support in Anolis |
| 11:51 | <Philip`> | (For http://philip.html5.org/docs/rdfa/ I hacked it so <span title="html5:resolve-a-url"> would go to the right place) |
| 11:51 | <Philip`> | ((i.e. http://whatwg.org/html5/#resolve-a-url, and then the multipage magic sends you to the correct page)) |
| 11:51 | <Philip`> | (but I guess that's not a very general solution) |
| 11:52 | <Philip`> | (and it doesn't work for #linkTypes because the references get lowercased :-( ) |
| 12:03 | <gsnedders> | Philip`: I'm lazy, OK? |
| 12:05 | <gsnedders> | Philip`: I'll get around to it, OK? :P |
| 12:05 | <gsnedders> | Philip`: The two Anolis things I want to get done before I start at Opera are merging in biblio, and xdoc xref |
| 12:08 | <hsivonen> | gsnedders: will biblio use BibTeX microdata? |
| 12:08 | <jgraham> | gsnedders: I want microdata support in anolis |
| 12:08 | <jgraham> | Not for any particular reason, just because |
| 12:08 | <gsnedders> | hsivonen: No, it uses refer, because I concluded BibTeX was impossible to sanely parse at all |
| 12:08 | <hsivonen> | gsnedders: but you can now store bib data in .html with sane parsing! |
| 12:09 | <gsnedders> | hsivonen: When I wrote it you couldn't :P |
| 12:09 | hsivonen | mumbles something about off-the-shelf libraries being available for Java |
| 12:10 | <hsivonen> | gsnedders: if you had used Java, you could have used a .bib parser already there |
| 12:10 | gsnedders | mumbles something about off-the-shelf libraries being unbelievably buggy in Python |
| 12:10 | <gsnedders> | I could've tried Jython, but I'm heavily reliant upon lxml, which being a CPython extension won't work |
| 12:10 | <gsnedders> | hsivonen: I don't know Java, which is another problem. |
| 12:13 | <zcorpan> | gsnedders: web dom core has xdoc xrefs to html5 |
| 12:14 | gsnedders | goes back to working on solving diophantine equations |
| 12:15 | <gsnedders> | (well, linear Diophantine equations) |
| 12:49 | <zcorpan> | hsivonen: itemprop="http://example.org/foo/#"; |
| 12:49 | <hsivonen> | zcorpan: don't do that :-) |
| 13:04 | <annevk5> | zcorpan, how does <frameset onload> work when there's multiple framesets? |
| 13:04 | <annevk5> | come to think of it, that question applies to <body onload> too |
| 13:06 | <hsivonen> | what boxes did I miss? http://hsivonen.iki.fi/web-stack/ |
| 13:07 | <hsivonen> | I notice I don't have Atom or RSS there, but they aren't involved when content is loaded into a browsing context |
| 13:10 | <annevk5> | browsers do implement the sniffing rules so they sort of are |
| 13:11 | <gsnedders> | Yeah, but nothing else really does |
| 13:11 | <gsnedders> | Most feed readers ignore content-type |
| 13:13 | <zcorpan> | annevk5: the body or frameset element (or window object) that gets the attribute set last overrides |
| 13:15 | <annevk5> | also when the element is not inserted yet? |
| 13:15 | annevk5 | wonders if this is defined |
| 13:16 | <zcorpan> | annevk5: yep. the spec just says "... and with corresponding content attributes and DOM attributes exposed on the body element" |
| 13:16 | <zcorpan> | (which should say "body and frameset elements") |
| 13:18 | <zcorpan> | when the attribute is set, it overrides the element's ownerDocument's window.onload |
| 13:18 | <annevk5> | seems weird for non inserted elements to have an effect |
| 13:19 | zcorpan | points at topic |
| 13:19 | <annevk5> | meh |
| 13:25 | <jgraham> | hsivonen: It is not quite obvious why HTML, SVG, etc. are one side or the other of DOM |
| 13:26 | <jgraham> | (they should sit inside DOM in some sense) |
| 13:27 | <Philip`> | Does anyone have cross-document cross-references that aren't referring to HTML5? |
| 13:28 | <jgraham> | Philip`: You, surely? |
| 13:29 | <Philip`> | jgraham: I don't have any cross-references like that |
| 13:29 | <Philip`> | I just have links |
| 13:29 | <Philip`> | which, uh, I suppose are kind of cross-references |
| 13:30 | Philip` | presumes there's some good reason to have special syntax rather than just using <a href> everywhere |
| 13:31 | <annevk5> | yeah, from XHR2 to CORS |
| 13:31 | <annevk5> | HTML5 will have those too |
| 13:31 | <annevk5> | well, might |
| 13:50 | <hsivonen> | jgraham: they sit on top of the DOM |
| 13:52 | hsivonen | considers putting Unicode under everything but bitmaps and video |
| 13:55 | <jgraham> | hsivonen: In what sense? |
| 13:55 | <jgraham> | (It depends what you mean by HTML, SVG, etc.) |
| 13:56 | <hsivonen> | jgraham: in the sense HTML5 is a DOM language |
| 13:56 | <annevk42> | Philip`, the reason is not having to type <a href> everywhere |
| 13:56 | <hsivonen> | jgraham: and MathML, SVG and ARIA are in implementation practice, too |
| 13:57 | <Philip`> | annevk42: Typing <a href> doesn't seem like a great burden |
| 13:57 | <jgraham> | hsivonen: Sure. I think it is a bit confusing though, especially if you don't already have that wordview. In particular if you view HTML as a stream of bytes |
| 13:57 | <Philip`> | annevk42: but I guess typing the content of the href does |
| 13:58 | <Philip`> | jgraham: I thought the diagram was meant to represent how things are, rather than how some people misunderstand them to be |
| 14:00 | <jgraham> | Philip`: If the goal is to educate those people then doing so in a way that causes confusion is not the best approach |
| 14:01 | <jgraham> | hsivonen: Also, events are generally not exposed directly to the user |
| 14:02 | <hsivonen> | jgraham: what should I have between Events and the User? |
| 14:02 | hsivonen | considers being politically incorrect and using labels Ogg and Theora |
| 14:02 | <jgraham> | hsivonen: Javascript :) |
| 14:02 | <hsivonen> | perhaps adding WAV and PCM |
| 14:03 | <jgraham> | (Events seems to live within js/dom) |
| 14:03 | <hsivonen> | jgraham: having JS between user and events doesn't seem right |
| 14:03 | <hsivonen> | should I put mouse/keyboard there? |
| 14:03 | <hsivonen> | USB? |
| 14:03 | <hsivonen> | gotta go |
| 14:04 | <jgraham> | Oh, I see. I wasn't thinking of the flow of information going that way |
| 14:04 | <jgraham> | Since it goes the opposite way everywhere else |
| 14:04 | <jgraham> | (this is the problem with the "stack" concept) |
| 14:05 | <Philip`> | Human Interface Devices? |
| 14:05 | <annevk42> | it's much more a graph than stack |
| 14:05 | Philip` | realises that he doesn't know how to parse XML in Perl |
| 14:06 | <Philip`> | (except using XML::Simple, which is no good for parsing XHTML) |
| 14:07 | <Philip`> | Hmm, I wonder if XML::LibXML is sane |
| 14:10 | <Philip`> | Hmm, it seems to hang when parsing |
| 14:11 | <Philip`> | Oh, great, it's trying to download a zillion XHTML module DTDs |
| 14:11 | <gsnedders> | EPIC FAIL. |
| 14:12 | <jgraham> | Philip`: Surely you can turn that off? |
| 14:12 | <Philip`> | jgraham: I can once I know it's going to be doing that |
| 14:12 | Philip` | apologies to the w3.org server hosting people for accidentally contributing to the DDOS against them |
| 14:13 | <Philip`> | *apologises |
| 14:39 | <mpilgrim> | oh look, more false dichotomies from rubys: http://lists.w3.org/Archives/Public/public-html/2009May/0161.html |
| 14:40 | mpilgrim | wonders if rubys ever read the mailing list archives before he was appointed from on high |
| 14:40 | gsnedders | waits for mpilgrim to make another appearance on MLW's infamous blog |
| 14:41 | <mpilgrim> | the title of the spec was settled a long time ago |
| 14:43 | mpilgrim | is inexplicably reminded of the classic "bungee boss" dilbert strip |
| 14:44 | <mpilgrim> | "Hi-I'm-Your-New-Boss-Let's-Change-Everything-Before-I-Get-Reassigned-Oops-Too-Late-Goodbye" |
| 14:44 | <mpilgrim> | sproing |
| 14:44 | <mpilgrim> | He was like a mentor to me |
| 14:44 | <Dashiva> | Do you expect he'll be reassigned? |
| 14:45 | <mpilgrim> | no one seems to know why or how he was assigned in the first place |
| 14:45 | <mpilgrim> | so i couldn't say |
| 14:45 | <Philip`> | mpilgrim: You should volunteer to be chair |
| 14:46 | <Philip`> | I can't imagine any ways in which that could fail to work |
| 14:46 | <mpilgrim> | lately the working group seems to have been infested with bungee trolls |
| 14:48 | <gsnedders> | Hmm… according to this textbook, -2\frac{dy}{dx}-1 = \frac{2(x+y)}{x} - 1 |
| 14:48 | <mpilgrim> | "Hi-I-Care-Deeply-About-The-Future-Of-The-Web-Let's-Change-Everything-Before-I-Get-Exasperated-Oops-Too-Late-Goodbyte" |
| 14:48 | <gsnedders> | How is the first term not negative? |
| 14:48 | <gsnedders> | (on the RHS) |
| 14:49 | <mpilgrim> | "We specifically need people to put forward suggestions for @profile and @summary. If no such proposals are produced, such issues will be summarily closed." <-- noted for future reference |
| 14:50 | <Philip`> | gsnedders: How do you go from dx to x? |
| 14:50 | <gsnedders> | Philip`: I'm not entirely sure about that either |
| 14:50 | <Dashiva> | Is that an identity or an equation to solve? |
| 14:50 | mpilgrim | isn't sure rubys understands the biggest sticking point with the spec license |
| 14:50 | <gsnedders> | Identity |
| 14:51 | <Dashiva> | Then I'm with Philip`, it seems to be missing something |
| 14:54 | jgraham | joins Philip` and Dashiva |
| 14:55 | <Philip`> | Fortunately nobody seems have cared about the bit in my document that says "This is blatant copyright infringement." |
| 14:56 | <gsnedders> | Or using the W3C stylesheet |
| 14:56 | jgraham | wonders what mpilgrim thinks the biggest sticking point is |
| 14:56 | <Philip`> | gsnedders: That too |
| 15:01 | <gsnedders> | http://secret.gsnedders.com/school/id.jpg |
| 15:01 | <gsnedders> | (This is blatant copyright infringement.) |
| 15:02 | <Philip`> | gsnedders: The question makes much more sense when we can see that dy/dx = -(x+y)/x |
| 15:03 | <gsnedders> | Oh, so they substitute for it |
| 15:03 | <gsnedders> | That makes sense |
| 15:03 | <Philip`> | (by rearranging the expression on the first line) |
| 15:05 | jgraham | rhetorically wonders if it is worth pointing out to Shelley that she doesn't get to define who "participated" in HTML 5 |
| 15:06 | Dashiva | rhetorically avoids answering |
| 15:06 | <mpilgrim> | jgraham: they don't own it. it isn't theirs. |
| 15:07 | jgraham | doesn't know who is supposed to be owning anything |
| 15:07 | <mpilgrim> | traditionally, choosing a license is reserved for the copyright owner |
| 15:07 | <jgraham> | Oh, I follow now |
| 15:07 | <Philip`> | Doesn't the W3C traditionally own the copyright on the specs it publishes? |
| 15:08 | <gsnedders> | Philip`: The W3C host sites do, not W3C |
| 15:08 | <Philip`> | "Copyright ©1997-1999 W3C® (MIT, INRIA, Keio), All Rights Reserved.", "Copyright © 2009 W3C® (MIT, ERCIM, Keio), All Rights Reserved.", etc |
| 15:10 | <mpilgrim> | Philip`: traditionally, yes |
| 15:19 | <gsnedders> | http://secret.gsnedders.com/school/dif2.jpg |
| 15:19 | <gsnedders> | is that right? |
| 15:23 | <Philip`> | gsnedders: Looks about right to me |
| 15:23 | <Philip`> | though you could have cancelled out the 2s much earlier, to save a lot of writing :-) |
| 15:24 | <gsnedders> | That is true. |
| 16:02 | <mpilgrim> | i love how rubys seems to think that there is a stable of potential spec editors just chomping at the bit to be given a chance to publish |
| 16:02 | <mpilgrim> | we've been desperately trying to find editors for years |
| 16:03 | <mpilgrim> | nobody ever steps up and sticks with it long enough to matter |
| 16:04 | mpilgrim | notes, somewhat hypocritically, that he is not volunteering either |
| 16:05 | <jgraham> | mpilgrim: You seem to be assuming naivity which is not necesssarily true |
| 16:05 | <Philip`> | I don't think he thinks there is, which is why he includes the condition that if there aren't editors for a document then the group will drop that document, rather than saying the group will block on issues that don't have the support of an editor |
| 16:07 | <mpilgrim> | jgraham: well, like the rest of us, i sincerely hope that he DOES have a stable of potential spec editors |
| 16:08 | <mpilgrim> | though i think it's a bit cruel to make the call for editors without acknowledging how much work it will be |
| 16:08 | <jgraham> | mpilgrim: I interpreted his email more as "if you want to push an idea that you can't get someone else to spec you can spec it yourself (or stop complaining)" |
| 16:08 | <mpilgrim> | otherwise we'll just end up with more orphaned drafts |
| 16:09 | mpilgrim | rereads it |
| 16:10 | <mpilgrim> | i think we're both right |
| 16:10 | <mpilgrim> | it's "put up or shut up" |
| 16:10 | <gsnedders> | That's how I read it. |
| 16:11 | <mpilgrim> | but it's not as if we've been overwhelmed with a plethora of competently edited specs that we're ignoring |
| 16:12 | <annevk42> | I wish we had more editors for Web specs |
| 16:13 | <mpilgrim> | perhaps the spec fairy will visit us |
| 16:14 | <jgraham> | Is the spec fairy the one who gives you money wwhen you give her a spec? |
| 16:14 | <jgraham> | Because that is another issue... |
| 16:14 | <mpilgrim> | hmm, good point, the analogy doesn't work |
| 16:15 | <mpilgrim> | anyway, if rubys can get people to put in the work and produce a half-decent spec, more power to him |
| 16:15 | <jgraham> | (actually I think it sort of works for one of the problems which is that there just aren't a lot of people willing to pay for decent web specs, despite the huge value that they have) |
| 16:16 | <jgraham> | (it is a tragedy of the commons type thing where it is in everyone's interest for someone else to do it) |
| 16:16 | <jgraham> | (so Hixie is one of the very few people paid to edit specs full time) |
| 16:19 | <mpilgrim> | actually, most w3c groups are composed entirely of employees of member companies |
| 16:19 | <mpilgrim> | who are presumably paid to collaborate/discuss/edit their respective specs |
| 16:19 | <gsnedders> | Just not FT |
| 16:20 | <jgraham> | In many cases not just not FT but not enough of the time to make a good result |
| 16:20 | <mpilgrim> | not sure about that either |
| 16:20 | <mpilgrim> | at ibm there was a whole department of people who did standards work full-time |
| 16:21 | <mpilgrim> | but i agree with the fundamental point, that spec editing is hard and few people have the time + energy + resources + expertise to do it |
| 16:22 | <Philip`> | + motivation |
| 16:22 | <jgraham> | If Hixie an't being paid to work full time on HTML 5 it would have died da long time ago |
| 16:23 | <Philip`> | (I guess there are people who could edit specs but don't care enough about it to do so) |
| 16:25 | <Philip`> | jgraham: Sounds like an unsustainable model for spec development |
| 16:26 | <jgraham> | Philip`: Right, but what else do you suggest? |
| 16:26 | <Philip`> | jgraham: No idea |
| 16:27 | <jgraham> | Unless you can make enough money from the spec itself to hire editors. But that doesn't seem likely |
| 16:28 | <Philip`> | It seems like companies don't mind paying their own employees to write their own specs |
| 16:28 | <Philip`> | e.g. OOXML |
| 16:29 | <Philip`> | but it seems much harder to get them to contribute similarly to open collaborative processes |
| 16:30 | <mpilgrim> | adsense for specs |
| 16:30 | <mpilgrim> | you heard it here first |
| 16:30 | <Philip`> | You could say that anybody writing a web page using your spec MUST provide a link to the spec |
| 16:31 | <Philip`> | Then you'll get massive PageRank and can sell links |
| 16:31 | <mpilgrim> | presumably that would work on all specs except rel=nofollow |
| 16:31 | <mpilgrim> | ;) |
| 16:32 | <Philip`> | The problem is the editor would need up-front funding to develop the spec, and would only get any return once it's finished and widely used |
| 16:36 | <mpilgrim> | spec on spec |
| 16:37 | <Philip`> | There's always the ISO model of charging people to download the spec |
| 16:38 | <jgraham> | Philip`: I may not have mentioned that due to it being evil and wrong |
| 16:38 | <Philip`> | Also you could hide some patented technologies in the spec, and then sue Microsoft for a billion dollars a few years later |
| 16:39 | <mpilgrim> | submarine specs |
| 16:39 | <Philip`> | (See e.g. MP3) |
| 16:39 | <mpilgrim> | pretty sure we have rules against that |
| 16:39 | <annevk42> | patent policy doesn't protect against that if the spec writer is just a person |
| 16:40 | <annevk42> | at least, I believe it doesn't |
| 16:40 | <mpilgrim> | as opposed to spec writers who are superhuman? |
| 16:40 | <mpilgrim> | i'm confused |
| 16:40 | <annevk42> | spec writers employed by legal entities who could be counter sued |
| 16:40 | <mpilgrim> | no, i mean the w3c has rules against that |
| 16:40 | <Philip`> | Patent policy doesn't help if the patent is owned by somebody outside the W3C |
| 16:41 | <gsnedders> | s/W3C/WG/ |
| 16:41 | <Philip`> | so the editor could get his/her friend to patent the technology and then sue Microsoft |
| 16:41 | <mpilgrim> | pretty sure we have rules against that too, but whatever |
| 16:42 | <Philip`> | There may be rules, but you can break the rules |
| 16:42 | <mpilgrim> | oh dear god: http://lists.w3.org/Archives/Public/public-html/2009May/0170.html |
| 16:42 | <Philip`> | If the punishment is getting thrown out of the W3C, but you get a billion dollars, it's pretty good overall |
| 16:43 | <mpilgrim> | it's really amazing that the w3c ever publishes anything at all |
| 16:45 | <Philip`> | Most WGs aren't as objectionable as the HTML WG |
| 16:49 | <mpilgrim> | most standards aren't as important as HTML 5 either |
| 16:49 | <mpilgrim> | and most of the ones that are, aren't standardized in public |
| 16:51 | <jgraham> | There seems to be an awful lot of "over my dead body" around publication issues in the HTMLWG |
| 16:51 | <jgraham> | which doesn't bode well |
| 16:52 | <beowulf> | didn't the guy who raised the objection to the design principals quit the wg? |
| 16:55 | <Philip`> | Philip Taylor? |
| 16:55 | <Philip`> | If it was him, then yes |
| 16:56 | <hober> | hsivonen: mind if I use (with attribution) your web stack diagram in a blog post of mine? |
| 16:58 | <annevk4> | jgraham, there's not too much "over my dead body" considering the size of the group |
| 16:59 | <annevk4> | jgraham, I don't think the people who disagree strongly with certain decisions has grown since the start |
| 17:00 | <annevk4> | I listened to one of the TED talks yesterday where the speaker suggested that if nobody disagreed with what you're doing you're doing it wrong. |
| 17:01 | <Philip`> | I'm not sure I'd agree with that statement |
| 17:03 | <annevk4> | http://www.ted.com/index.php/talks/lang/eng/seth_godin_on_the_tribes_we_lead.html |
| 17:05 | <annevk4> | It does resonate with me that if you try to change the status quo people will disagree. |
| 17:05 | <annevk4> | And it's pretty clear we try to change the status quo |
| 17:05 | <gsnedders> | I thought most of what we did was standardizing the status quo |
| 17:06 | <Philip`> | Lack of disagreement might just mean that nobody knows what you're doing |
| 17:06 | <beowulf> | or it might mean you're being given a lot of slack in the hope something might happen |
| 17:06 | <Philip`> | And when you reach an extreme where everybody disagrees with you, you're almost certainly wrong |
| 17:07 | <annevk4> | I don't really see any of that applying to HTML5 |
| 17:07 | <Philip`> | so it's not a particularly helpful guideline, since it doesn't tell you whether you're experiencing too much disagreement or too little |
| 17:10 | <annevk4> | Philip`, it seems pretty evident that as far as the standards community is concerned people are aware of HTML5 |
| 17:10 | <annevk4> | Philip`, that said, I'm not sure it should be used as guideline |
| 17:11 | <gsnedders> | http://rossching.com/little-bribes/ is really impressive |
| 17:47 | <jgraham> | annevk4: If you look at people who have been actively involved, there seems to be quite a lot of people making objections and not being amenable to discussion or compromise |
| 17:48 | <annevk4> | My perception is different. Maybe it's worth gathering data? Then again, not sure what we could usefully do with the outcome. |
| 17:49 | <archtech> | Have you guys seen the ECMA5 presentation: http://www.youtube.com/watch?v=Kq4FpMe6cRs |
| 17:49 | <jgraham> | archtech: Not yet but I plan to watch it |
| 17:49 | <archtech> | I like the juxtaposition of HTML5 and ECMAScript5 :) The version where things finally are well done ;) |
| 17:49 | <jgraham> | annevk4: I expect I am being too negative |
| 17:50 | <jgraham> | It wouldn't be the first time |
| 17:50 | <jgraham> | It will be interesting to see what happens when we try to transition to LC |
| 17:52 | gsnedders | thinks we need DOM5 and CSS5 then Web5 will be complete! |
| 17:52 | <Philip`> | Hmm, rdfQuery throws an exception if you have <link rel=bogus> |
| 17:52 | <Philip`> | which isn't very forward compatible |
| 17:52 | <Philip`> | gsnedders: Don't forget HTTP5 |
| 17:52 | <gsnedders> | And TCP/IP5 |
| 17:52 | <Philip`> | Also we need User5 |
| 17:53 | <gsnedders> | And Earth5 |
| 17:53 | <annevk4> | and URL5 etc. |
| 17:56 | <beowulf> | and IE5. No, wait... |
| 18:53 | <archtech> | In the presentation we learn arguments object is removed, but nothing is there to replace it. |
| 18:53 | <archtech> | We have no default params, so weird. |
| 18:57 | <inimino> | archtech: the arguments object isn't removed |
| 18:58 | <archtech> | inimino, I misunderstood, can you please tell me what is their strategy? |
| 18:58 | <inimino> | I haven't seen the presentation so I'm not sure what was being referred to |
| 18:59 | <inimino> | but have no fear, 'arguments' is alive and well in ES5 |
| 19:00 | <archtech> | inimino, to clarify myself, what the presentation seems to say is, it will be in ES5 no-strict mode, but any call to arguments, arguments.callee, arguments.caller will throw immediately in strict mode |
| 19:00 | <inimino> | I'll check |
| 19:01 | <archtech> | javascript |
| 19:01 | <archtech> | ops |
| 19:05 | <inimino> | archtech: arguments.caller and .callee will now throw |
| 19:06 | <archtech> | Arguments will not? |
| 19:06 | <inimino> | no |
| 19:06 | <archtech> | I see, thanks. |
| 19:07 | <archtech> | Now if only they agreed on private/protected flags for properties as well ;) |
| 19:27 | <Hixie> | jgraham: if i wasn't being paid to work on html5, my social life would be far less interesting, but html5 would still be happening |
| 19:30 | <Philip`> | So Google is effectively paying for your social life? |
| 19:30 | <Philip`> | Sounds like a good trick you've got worked out there |
| 19:30 | <Hixie> | it's a pretty awesome trick |
| 19:33 | annevk42 | more or less catched up with email and is wondering about XMLHttpRequest fetch again |
| 20:20 | <jgraham> | Hixie: I would prefer it if we didn't have to do that experiment :) |
| 20:23 | <mpilgrim> | oh look, john foliot is going on again about "experts" |
| 20:25 | <mpilgrim> | the WCAG WG was chartered on february 5, 2004: http://www.w3.org/2004/04/wcag-charter |
| 20:26 | <mpilgrim> | on december 11, 2008, they published "Techniques for Web Content Accessibility Guidelines 2.0": http://www.w3.org/TR/2008/NOTE-WCAG20-TECHS-20081211/ |
| 20:27 | <mpilgrim> | it includes an entire section on "Using alt attributes on img elements": http://www.w3.org/TR/WCAG-TECHS/H37.html |
| 20:27 | <mpilgrim> | which lists a grand total of 2 examples |
| 20:27 | <mpilgrim> | only 1 of which contains actual markup |
| 20:28 | <mpilgrim> | HTML 5 has 14 sections devoted to using alt attributes on img elements: |
| 20:28 | <mpilgrim> | 1. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-0.html#a-link-or-button-containing-nothing-but-the-image |
| 20:28 | <mpilgrim> | 2. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-0.html#a-phrase-or-paragraph-with-an-alternative-graphical-representation:-charts,-diagrams,-graphs,-maps,-illustrations |
| 20:28 | <mpilgrim> | 3. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-0.html#a-short-phrase-or-label-with-an-alternative-graphical-representation:-icons,-logos |
| 20:28 | <mpilgrim> | 4. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-0.html#text-that-has-been-rendered-to-a-graphic-for-typographical-effect |
| 20:28 | <mpilgrim> | 5. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-0.html#a-graphical-representation-of-some-of-the-surrounding-text |
| 20:28 | <mpilgrim> | 6. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-0.html#a-purely-decorative-image-that-doesn't-add-any-information |
| 20:28 | <mpilgrim> | 7. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-0.html#a-group-of-images-that-form-a-single-larger-picture-with-no-links |
| 20:28 | <mpilgrim> | 8. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-0.html#a-group-of-images-that-form-a-single-larger-picture-with-links |
| 20:28 | <mpilgrim> | 9. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-0.html#a-key-part-of-the-content |
| 20:28 | <mpilgrim> | 10. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-0.html#an-image-not-intended-for-the-user |
| 20:29 | <mpilgrim> | 11. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-0.html#an-image-in-an-e-mail-or-private-document-intended-for-a-specific-person-who-is-known-to-be-able-to-view-images |
| 20:29 | <mpilgrim> | 12. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-0.html#general-guidelines |
| 20:29 | <mpilgrim> | 13. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-0.html#guidance-for-markup-generators |
| 20:29 | <mpilgrim> | 14. http://www.whatwg.org/specs/web-apps/current-work/multipage/embedded-content-0.html#guidance-for-conformance-checkers |
| 20:29 | <annevk42> | So reading WCAG-TECHS and looking at http://www.whatwg.org/specs/web-apps/current-work/images/parsing-model-overview.png I wonder why John thinks the alternative text is too long |
| 20:30 | <annevk42> | as WCAG-TECHS clearly says to include the words from the image in the alternative text |
| 20:30 | krijnh | wonders if mpilgrim would also paste all this if this channel wasn't logged :) |
| 20:30 | <Philip`> | On the other hand, the WCAG section on alt is readable, whereas the HTML 5 version is a stupidly large and complex amount of content that will just get totally ignored and/or misunderstood |
| 20:30 | <Philip`> | See e.g. http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-May/019631.html |
| 20:30 | <mpilgrim> | with a total of 32 examples |
| 20:31 | <mpilgrim> | "experts" are valuable if-and-only-if they share their expertise |
| 20:31 | <mpilgrim> | the accessibility "experts" -- who were given ALL THE TIME IN THE WORLD to share their expertise -- failed to do so |
| 20:32 | <mpilgrim> | so they don't get to complain that somebody came in later to fill the GAPING VOID OF KNOWLEDGE that they left |
| 20:37 | <Philip`> | http://www.web3d.org/x3d/wiki/index.php/X3D_and_HTML5 - "The X3D Working Group will participate in the HTML Working Group for the purpose of best integrating X3D with HTML." |
| 20:59 | annevk42 | wonders how http://www.w3.org/mid/000601c9d984$a79d63b0$f6d82b10$@org will work out |
| 21:13 | jgraham | wonders if turning all bibliography properties into URIs won't destroy usability |
| 21:14 | <jgraham> | s/all/some/ |
| 21:14 | <Philip`> | Abbreviate the URIs with a "biblio:" prefix and it'll be perfectly usable |
| 21:14 | <jgraham> | :p |
| 21:15 | jgraham | remembers that he tried to use zotero once but it seemed so complex that he didn't bother to learn to use it properly |
| 21:18 | <Philip`> | http://www.librocket.com/ - "HTML/CSS Game Interface Middleware" - "libRocket uses the time-tested open standards XHTML1.0 and CSS2.0 (while borrowing features from HTML5 and CSS3)" |
| 21:18 | Philip` | can't find any references to what features were borrowed from HTML5 |
| 21:29 | <tantek> | mpilgrim, so what you're really saying is that "experts" are valuable if-and-only-if they make their expertise accessible to others. |
| 21:50 | <mpilgrim> | tantek: haha. but seriously, i still firmly believe what i wrote in 2004: http://diveintomark.org/archives/2004/08/16/specs |
| 21:50 | <mpilgrim> | "experts" are morons who write about their mistakes |
| 21:51 | <mpilgrim> | there isn't anything intrinsic to being an expert |
| 21:51 | <mpilgrim> | you're only as "expert" as what you share |
| 21:53 | <tantek> | heh - yes, classic essay |
| 21:54 | <tantek> | mpilgrim - and sharing in email is not particularly accessible IMHO - so if expertise isn't shared on a web page with a permalink (email archives notwithstanding because they don't search/index well), then it's not sufficiently shared. |
| 21:55 | <tantek> | mpilgrim - the IE5Mac/Tasman team pretty much followed the "Angels" description in your essay when developing/shipping IE5Mac |
| 21:55 | <tantek> | (modulo some interpretation of "thoroughly" - not sure how much is enough there, right?) |
| 22:03 | <mpilgrim> | IE5/Mac totally rocked |
| 22:03 | <mpilgrim> | for its time |
| 22:24 | <annevk42> | o_O http://code.google.com/p/chromium/issues/detail?id=988#c9 |
| 22:31 | <gsnedders> | :o |
