| 00:58 | <roc> | I wonder if that X3D activity is a response to canvas3d |
| 00:58 | <roc> | ah yes |
| 00:58 | <olliej> | roc: hehehe |
| 00:59 | <roc> | yippee |
| 00:59 | <olliej> | roc: what is x3d? |
| 00:59 | <roc> | it's like the SVG of 3D |
| 00:59 | <roc> | except, I suspect, worse |
| 00:59 | <roc> | I think it evolved from VRML |
| 01:00 | <olliej> | roc: heheh |
| 01:00 | <olliej> | worse the svg? |
| 01:00 | <olliej> | wow |
| 01:00 | olliej | hides |
| 01:06 | <Philip`> | roc: It is indeed evolved from VRML |
| 01:07 | <Philip`> | You can express it in 'classic' VRML syntax as well as in XML syntax |
| 01:07 | <roc> | excellent |
| 01:07 | <roc> | two syntaxes for a format that's not even really deployedyet |
| 01:08 | <Philip`> | There's quite a number of independent implementations |
| 01:08 | <Philip`> | If you're careful, you can even write a scene that works in at least two implementations |
| 01:09 | <Philip`> | (It's also an ISO standard) |
| 01:09 | <Hixie> | aaah, why are we talking about x3d |
| 01:09 | Hixie | tries to get to a safe place |
| 01:10 | <Philip`> | I'm not sure X3D is particularly crazy or anything |
| 01:10 | <Hixie> | but why did it come up! |
| 01:10 | <Hixie> | actually i know nothing about 3d |
| 01:10 | <Philip`> | Because I saw http://www.web3d.org/x3d/wiki/index.php/X3D_and_HTML5 |
| 01:10 | <Hixie> | though people seem to keep asking me about stuff to do with it |
| 01:11 | <Philip`> | (and because somebody emailed me about the X3D-in-canvas3d prototype I wrote ~16 months ago) |
| 01:11 | <Hixie> | jesus, everybody and their dog wants to put their vocabulary in html5 |
| 01:11 | <Philip`> | Hixie: You've added lots of other things into HTML5 you know nothing about :-) |
| 01:12 | <Hixie> | i'm confident in saying that i know more about everything in html5 than i do about 3d anything |
| 01:12 | <Hixie> | but that admittedly doesn't raise the bar very high |
| 01:12 | <Philip`> | 3D is just like 2D, except in a different direction |
| 01:13 | <Hixie> | 2D doesn't have things like lights |
| 01:13 | <Hixie> | and cameras |
| 01:13 | <Philip`> | Depending on your level of abstraction, 3D doesn't have those either :-) |
| 01:14 | <Philip`> | Cameras are just transformation matrices, and lights are just vectors you feed into your shaders, and it's easy when you copy-and-paste the scary maths from somewhere |
| 01:14 | <Hixie> | woosh |
| 01:14 | <Philip`> | I'm sure it's easier to understand than <keygen> :-) |
| 01:15 | <Hixie> | actually <keygen> was easy to understand, it just wasn't documented |
| 01:16 | <Hixie> | i wonder how we should allow people to annotate their HTML documents with a datetime representing the last substantial update |
| 01:16 | <Hixie> | (or if we should not) |
| 01:17 | <Philip`> | Sounds like a job for the microdata hammer |
| 01:17 | <tantek> | one existing solution: they could use hAtom to markup the page as an entry with an "updated" datetime. |
| 01:19 | <Hixie> | i don't want to make html5 hardcode any magical class values; last time i did that i got yelled at so much it had to be removed :-) |
| 01:20 | Philip` | would wonder what a "substantial update" is, but decides instead to go to bed |
| 01:20 | <tantek> | no need to hardcode any magical class values in html5 - simply don't bother to solve the problem in html5 and let people use the existing solution |
| 01:20 | <Hixie> | this is actually part of some other problem i'm solving |
| 01:21 | <Hixie> | (namely adding Atom as one of the formats you can export to in http://www.whatwg.org/specs/web-apps/current-work/#converting-html-to-other-formats ) |
| 01:27 | <roc> | I was quite clearly to say "not really deployed" instead of "not implemented" |
| 01:27 | <roc> | I was quite *careful* |
| 01:27 | <roc> | I wish I knew another language so I could pretend English wasn't my first language |
| 01:27 | <Hixie> | i wonder if the x3d-in-html5 effort is related to the people who asked me if collada could be put in html5 <canvas> recently |
| 01:28 | <Hixie> | if <article pubdate=""> gives the datetime that the article was published, what would be a good attribute name for the datetime that the article was last modified? |
| 01:29 | <Hixie> | (or alternatively, what would be good attribute names for published datetime and updated datetime respectively; it doesn't have to be pubdate) |
| 01:36 | <Hixie> | eh screw it, i'll just have only pubdate |
| 01:36 | <Hixie> | oh hey, updates can already be annoted using <ins>/<del> |
| 01:37 | <Hixie> | so i can just get the date from that |
| 07:05 | <hsivonen> | hober: using the diagram is ok |
| 07:51 | <Hixie> | anything interesting going on in the world of web specs? |
| 07:52 | <othermaciej> | you tell us! |
| 07:54 | <Hixie> | too many other things going on today to get around to spec work sadly |
| 08:03 | <hsivonen> | Hixie: did you add the DC creator stuff just because you could or was it in response to a use case? |
| 08:04 | <hsivonen> | Hixie: it violates theoretical purity |
| 08:04 | <Hixie> | just because i could |
| 08:04 | <Hixie> | it does? |
| 08:04 | <Hixie> | i was trying to convey as many of the document's semantics as possible |
| 08:04 | <Hixie> | so that the conversion loses as little as possible |
| 08:04 | <othermaciej> | what theoretical purity? |
| 08:04 | <Hixie> | (it still loses a lot) |
| 08:05 | <hsivonen> | yes. If you build an object model for microdata that contains enough data to output into all the outputs, you need to have an isAddressThatAppliesToDocument flag on items |
| 08:06 | <Hixie> | the RDF output is not about microdata |
| 08:06 | <hsivonen> | oh? |
| 08:06 | <Hixie> | it's a conversion of HTML to RDF |
| 08:06 | <Hixie> | which includes microdata |
| 08:06 | <Hixie> | but is not limited to it |
| 08:07 | <Hixie> | the Atom output i'm doing now barely mentioned microdata at all |
| 08:07 | <Hixie> | mentions, i should say |
| 08:07 | <hsivonen> | hmm. |
| 08:07 | <Hixie> | the vCard, vEvent, and BibTeX outputs are microdata-specific |
| 08:07 | <Hixie> | maybe i should reorganise that section to separate them |
| 08:07 | <hsivonen> | I was thinking of doing SAX-to-model builder |
| 08:08 | <hsivonen> | and a model to XML-represented-as-SAX converter |
| 08:08 | <hsivonen> | that could feed into a RELAX NG validator whose error messages have been munged |
| 08:08 | <Hixie> | sounds useful |
| 08:09 | <hsivonen> | (retaining Locator data, of course) |
| 08:09 | <Hixie> | i can split the RDF output into two, one for microdata and one for documents |
| 08:09 | <Hixie> | (with the output of the latter being a superset of the former's) |
| 08:10 | <hsivonen> | It think rel / <meta name> stuff should be considered part of microdata and be reflected in the JSON output |
| 08:10 | <Hixie> | why? |
| 08:10 | <hsivonen> | because we can :-) |
| 08:10 | <Hixie> | microdata is just items |
| 08:10 | <Hixie> | because it's simpler |
| 08:11 | <Hixie> | simpler > possible |
| 08:11 | <hsivonen> | seriously, though, it makes it possible to use JSON as a document metadata conveyor without having to resort to RDF |
| 08:11 | <Hixie> | (microdata isn't metadata, it's just data) |
| 08:12 | <Hixie> | i don't mind adding the metadata to the JSON stuff |
| 08:12 | <Hixie> | though that affects what gets dragged and dropped |
| 08:12 | <Hixie> | which i'm not sure is what we want |
| 08:16 | <Hixie> | hmm... <h1> or <title> for atom:title? |
| 08:17 | <Hixie> | i'm thinking <h1>, with <hgroup><h2> being atom:subtitle |
| 08:34 | <Philip`> | Hixie: Seems probably unlikely that people from the X3D community would be asking for COLLADA support, since they're completely separate (and partly overlapping) technologies |
| 08:34 | <Hixie> | ah |
| 08:34 | <Hixie> | i wonder why 3d is suddenly so important |
| 08:35 | <Philip`> | (As I understand it, they're sort of competing technologies in certain aspects, but the X3D people now recognise that COLLADA is popular and isn't going to go away so they're trying to live with it rather than fight against it, but it's not what they really want to support) |
| 08:36 | <Philip`> | 3D has always been important :-) |
| 08:36 | <Hixie> | collada, from what i understand, is a disaster |
| 08:36 | <Philip`> | but now the technology is capable of supporting it decently (perhaps) |
| 08:37 | <Hixie> | if i'm not mistaken, collada is about 15 different technologies all jammed into one spec because nobody could agree to remove their proposal |
| 08:37 | <Philip`> | I've used COLLADA (for loading models and animations into a game engine), and it worked quite nicely for what I wanted |
| 08:38 | <Philip`> | It's not intended for connecting arbitrary pairs of producers and consumers, like HTML - it's intended to fit between specific tools in a pipeline, and the tools can pick which subset to support based on what features they want |
| 08:40 | <Philip`> | e.g. you'll export a COLLADA file from 3ds Max, and run it through a custom tool that converts it into some efficient custom geometry format, and maybe have some other tool to extract texture references, and another to get some physics data, and whatever |
| 08:41 | <Philip`> | (and maybe some reusable tools in the middle to optimise geometry or something) |
| 08:43 | <Philip`> | and so you can just write tools specifically for the dialect that's generated at the beginning of the process, and you don't need to worry about any unused features because they're irrelevant until you want to start using them |
| 08:46 | <Hixie> | Philip`: since i've been asked to put it in html, that's the environment in which i was considering its appropriateness |
| 08:52 | <hsivonen> | Hixie: did you actually mean to use a single 'author' or 'editor' field in the HTML version of BibTeX |
| 08:52 | <hsivonen> | Hixie: that's horribly ugly |
| 08:52 | <Hixie> | i meant to use whatever bibtex uses |
| 08:53 | <hsivonen> | Hixie: comparend to having multiple and munging to and from .bib's " AND " |
| 08:53 | <Hixie> | i was originally going to do that but it means hard-coding them in the conversion |
| 08:53 | <hsivonen> | Hixie: I suggest special casing those two, because .bib's intra-field format makes no sense for DRY HTML |
| 08:53 | <Hixie> | which means hard-coding a non-localised string |
| 08:54 | <hsivonen> | Hixie: consider it a reserved word in .bib |
| 08:54 | <hsivonen> | Hixie: letting 'and' leak into HTML content in whatever language is worse, much worse |
| 08:54 | <Hixie> | i guess |
| 08:54 | Hixie | isn't a fan of bibtex |
| 08:55 | <Hixie> | i wish there was a better vocabulary to use here |
| 08:55 | <Hixie> | but the alternatives are all ten times worse |
| 08:55 | <hsivonen> | particularly when foo and bar and baz doesn't make DRY sense even in English |
| 08:55 | <Hixie> | Refer for example has the worst vocabulary terms ever |
| 08:55 | <Hixie> | and most of the others have much more domain-specific usage |
| 08:55 | <Hixie> | or no usage to speak of |
| 09:00 | <hsivonen> | the microdata representation sucks big time for format like "Doe, J.", but that's because munging of personal names sucks |
| 09:00 | <hsivonen> | and people should just put "John Doe" in their bibliographies |
| 09:02 | <Philip`> | Hixie: Yeah, it does seem pretty inappropriate as a general standard model/animation format |
| 09:03 | <Philip`> | The whole being-written-in-XML thing makes it inappropriate for content delivery anyway |
| 09:04 | <Hixie> | hsivonen: thanks for the bug |
| 09:04 | <Hixie> | collada is xml? |
| 09:04 | <Philip`> | Uh... Yes |
| 09:04 | <Philip`> | I guess you didn't look at it in any detail :-p |
| 09:08 | <Philip`> | (It's not terrible in terms of space/time efficiency, since the main issue is just that it stores numbers as space-separate strings of decimals rather than as 32-bit floats, but it's (intentionally) not great) |
| 09:08 | <Philip`> | s/e/ed/ |
| 09:09 | <Hixie> | i said i knew nothing! |
| 09:09 | <Hixie> | wow i had no idea collada was xml |
| 09:09 | <Hixie> | isn't that, like, dumb |
| 09:09 | <Hixie> | surely 3d modelling is one of those things that you'd want a binary format for |
| 09:10 | <Philip`> | There are lots of binary format, but they're almost universally awful |
| 09:10 | <Philip`> | e.g. the only widely-supported one is 3DS, which comes from the days of DOS and is limited to 64K vertices and supports almost no features |
| 09:10 | <Philip`> | and there are some newer proprietary binary ones |
| 09:12 | <Philip`> | The aim was to make something extensible and relatively easy to understand, and not to make something that's as efficient as possible (because everyone will have different rendering engines and want their own custom formats for that), and XML was a better way to achieve that than a new binary format |
| 09:13 | <Hixie> | fair enough |
| 09:15 | <othermaciej> | 3D canvas will probably get implementation traction |
| 09:15 | <othermaciej> | although I don't quite get why Mozilla decided to take it to Khronos instead of a Web standards group |
| 09:16 | <Philip`> | Possibly because they need involvement from hardware/driver developers in order to make it work securly, and all those people are in Khronos |
| 09:16 | <Philip`> | s//e/ |
| 09:39 | <Hixie> | ok bed time nn |
| 09:39 | <annevk42> | g'night |
| 10:09 | <zcorpan_> | if you remove the "green", you still have "left border", so even ignoring markup the requirements are met |
| 10:11 | <zcorpan_> | i guess i would just be contributing noise if i pointed that out on the list |
| 10:12 | <zcorpan_> | though i wonder if john would be satisfied if the border was changed from green to black |
| 10:14 | <zcorpan_> | oh maciej already said so |
| 10:14 | <annevk42> | I doubt borders are exposed to e.g. screen readers though |
| 10:15 | <zcorpan_> | annevk42: the requirements didn't say anything about that |
| 10:16 | <annevk42> | Private email suggests that's not a concern :/ |
| 10:16 | <Philip`> | The quote from WCAG makes it sound like you don't have to actually make it work in e.g. screen readers, you can just convey the information with markup (and presumably class="example" is markup) and it's somebody else's problem if they can't actually access that information through their UA |
| 10:18 | <zcorpan_> | "WCAG 2 - 1.4.1 Use of Color: Color is not used as the only visual means of conveying information, indicating an action, prompting a response, or distinguishing a visual element. (Level A);" - this is met, there is still a border as a visual means |
| 10:18 | <othermaciej> | the WCAG rules lawyering is completely off point |
| 10:18 | <annevk42> | That was my point, yes. Nobody has replied to that yet as far as I can tell. |
| 10:18 | <annevk42> | There are quite a lot of specifications that use class="example" in this way... |
| 10:18 | <hsivonen> | Philip`: I think mpilgrim had a blog post about that kind of accessibility enablement |
| 10:18 | <othermaciej> | the way examples are indicated is indeed inadequate, but for reasons that have nothing to do with John's complaint |
| 10:19 | <zcorpan_> | "WCAG 1 - 2.1 Ensure that all information conveyed with color is also available without color, for example from context or markup. [Priority 1];" - there is no information conveyed with color. information is conveyed with a border |
| 10:19 | <annevk42> | E.g. http://www.w3.org/TR/css3-namespace/ |
| 10:19 | <othermaciej> | also it is lame of him to complain about something I myself noticed and proposed to change |
| 10:19 | <zcorpan_> | "and Section 508 § 1194.22 c) Web pages shall be designed so that all information conveyed with color is also available without color, for example from context or markup." - same here |
| 10:19 | <othermaciej> | and to give some huge lecture on how bad it is |
| 10:19 | <zcorpan_> | so WCAG2 is met, and WCAG1 and section 508 are not appliciable |
| 10:20 | <othermaciej> | annevk42: the way examples render there looks much better to me |
| 10:20 | <othermaciej> | like the way they say "Example" |
| 10:20 | <annevk42> | othermaciej, that's a matter of style, not substance, in this case |
| 10:20 | <othermaciej> | annevk42: I would say rather that it's a matter of style, not markup |
| 10:20 | <othermaciej> | but in this case the style affects the substance |
| 10:21 | <othermaciej> | b/c I think in retrospect it was unclear to people that the examples are just examples, and separate from the statements of principle |
| 10:21 | <annevk42> | fair enough |
| 10:22 | <zcorpan_> | othermaciej: in some sections the examples are clearly marked as such, e.g. "2.2.1. Examples" |
| 10:23 | <zcorpan_> | the first two batches of examples |
| 10:24 | <othermaciej> | zcorpan_: good point! maybe using that style in most places would be sufficient |
| 10:24 | <othermaciej> | also, sadly, some of the design principles are still missing examples |
| 10:25 | <annevk42> | I think you rewrote the first two principles to be more lengthy and then the group decided to publish |
| 10:25 | <othermaciej> | also I wonder if perhaps "Evolution Not Revolution" could be retitled "Incremental Improvement" and thus made to encapsulate Hixie's proposed "Baby Steps" principle |
| 10:25 | <othermaciej> | yeah |
| 10:25 | <othermaciej> | this document could use some polish |
| 10:25 | <othermaciej> | the thing is, certain parties may turn any revision process into a rathole |
| 10:25 | <othermaciej> | I re-read the votes on the invidual principles |
| 10:25 | <annevk42> | s/may/will/ I think |
| 10:26 | <othermaciej> | and found that of all the objections, nearly all came from a few people who in turn objected to more than half the principles |
| 10:26 | <othermaciej> | so basically we have a set of Design Principles that most of the group agrees on, and then a handful of people who object to almost all of them |
| 10:26 | <othermaciej> | it's hard to see how to increase consensus in that light |
| 10:26 | <othermaciej> | you could water down the documet, but then the many people who largely agree with it could object |
| 10:29 | <gsnedders> | Hixie wanted visible (symbolic, IIRC) marks for things like examples, created by Anolis. I haven't done that yet. |
| 10:29 | <annevk42> | ah, so John Foliot can blame you! |
| 10:30 | <Philip`> | Only if the design principles document uses Anolis |
| 10:30 | <gsnedders> | Oh, we're talking about a11y of the design principles doc now? |
| 10:31 | gsnedders | is way behind on the ml |
| 10:31 | <Philip`> | I hope we are, otherwise I'm totally lost |
| 10:32 | <annevk42> | I thought it was about both |
| 10:33 | Philip` | shrugs |
| 10:35 | <zcorpan_> | hmm, the headings are blue |
| 10:36 | <zcorpan_> | wonder if that's also a violation of wcag? |
| 11:01 | <hsivonen> | whoa! role=math really made it into ARIA last call |
| 11:01 | <hsivonen> | I'd love to see two interoperable implementations of that one |
| 11:01 | <annevk42> | funny |
| 11:02 | <annevk42> | I was just reading Member-only ARIA discussions as I encountered something ARIA-related in my inbox |
| 11:02 | <om_sleep> | what does role=math mean? |
| 11:02 | <zcorpan_> | i wonder why ARIA discussions are still happening in Member-only lists |
| 11:02 | <annevk42> | Not too optimistic about my comments getting addressed adequately |
| 11:02 | <hsivonen> | om_sleep: it means that the the thing represents math and its text equivalent is MathML or TeX |
| 11:03 | <om_sleep> | I see |
| 11:03 | <om_sleep> | so you'd put it on an image that's an equation? |
| 11:03 | <om_sleep> | or something like that |
| 11:03 | <hsivonen> | which makes me ask: TeX with which implicitly defined macros? |
| 11:03 | <zcorpan_> | i also still wonder why MathML can't stand on its own and needs role="math" |
| 11:03 | <om_sleep> | anyway |
| 11:03 | <om_sleep> | bedtime |
| 11:03 | <hsivonen> | om_sleep: you basically dump fooTeX stuff in alt |
| 11:03 | <om_sleep> | zcorpan_: presumably the idea here is that the UA doesn't implement MathML but the assistive technology does |
| 11:04 | <zcorpan_> | hsivonen: i asked that to be defined before |
| 11:04 | <om_sleep> | (!) |
| 11:04 | <hsivonen> | I wonder how top on the list of e.g. JAWS developers a TeX subsystem is |
| 11:04 | <zcorpan_> | om_sleep: even in that case, MathML can stand on its own -- it has a namespace and a tag name to identify itself as "math" |
| 11:05 | <hsivonen> | zcorpan_: oh yeah, the first SHOULD makes no sense |
| 11:05 | hsivonen | starts drafting feedback |
| 11:09 | <zcorpan_> | "In order to be perceivable, images SHOULD also be labeled by text that describes the math formula as it should be spoken, using the aria-describedby attribute." |
| 11:09 | <zcorpan_> | wait, if the author should write how the formula should be spoken, what does role="math" help? |
| 11:10 | <zcorpan_> | oh well |
| 11:11 | <zcorpan_> | interesting that there are are both SHOULDs and shoulds |
| 11:11 | <webben> | Hmm. Yep that's a pretty confusing bit of spec. |
| 11:13 | <zcorpan_> | when role="math" was new, i asked "how would it be implemented?" |
| 11:13 | <zcorpan_> | evidently, that's not a great concern |
| 11:18 | <hsivonen> | email sent |
| 11:30 | <zcorpan_> | hsivonen: where did you send it? |
| 11:31 | <zcorpan_> | found it |
| 12:13 | <hsivonen> | am I understanding correctly that aria-label exists because title attribute is badly supported by AT? |
| 12:15 | <hsivonen> | or is the purpose to suppress the tooltip? |
| 12:20 | <hsivonen> | hmm. the conformance of aria-checked is interesting |
| 12:20 | <hsivonen> | it must be present but may have the value "undefined" |
| 12:24 | <zcorpan_> | hsivonen: my impression is that it exists because it's hard or annoying to position <h2> elements off-screen with css |
| 12:24 | <hsivonen> | zcorpan_: ah. |
| 12:25 | <hsivonen> | zcorpan_: do you know why aria-required is global but aria-expanded isn't? |
| 12:25 | <hsivonen> | the former seems less universally applicable than the latter |
| 12:25 | <zcorpan_> | hsivonen: no |
| 12:25 | <zcorpan_> | hsivonen: (...and that Lynx users can be ignored) |
| 12:26 | <hsivonen> | If Lynx users wish to become unignored, Lynx should catch up with a lot of Web tech that isn't even tied to bitmap rendering |
| 12:27 | <hsivonen> | someone should take Gecko or WebKit and replace the bitmap renderer with a character grid renderer |
| 12:27 | <annevk42> | I wonder that given the time it takes to get ARIA deployed and given that IE is updating as well if it would not have been more productive to argue for the sane solutions from the start... |
| 12:28 | <annevk42> | Of course, with hindsight things are easier. |
| 12:29 | <hsivonen> | well, assuming that IE8 doesn't support role=math, feel free to argue for a saner solution for that one |
| 12:30 | <hsivonen> | the obvious solution is MathML-in-text/html, but for that to work, IE and WebKit need MathML support |
| 12:30 | <annevk42> | you've already done that no? |
| 12:31 | <hsivonen> | annevk42: I effectively argued for the reading of the math as words being used as alt |
| 12:31 | <hsivonen> | the major use case is math on wikipedia |
| 12:31 | <annevk42> | another problem with this silo approach to comments is that the WG can back each other up while people commenting are all alone |
| 12:32 | <hsivonen> | I guess they should have the resources to feed their math to AsTeR in order to generate the words |
| 12:32 | <hsivonen> | at least in English |
| 12:32 | <zcorpan_> | is alt text for math images covered in html5? |
| 12:32 | <annevk42> | e.g. following the progress on each other's comments is much harder, etc. |
| 12:33 | <annevk42> | zcorpan_, yeah, use MathML :) |
| 12:35 | <hsivonen> | I'm updating v.nu's ARIA support |
| 12:36 | <hsivonen> | it seems that PF hoisted nearly everything except aria-expanded into global states/props, and I don't understand the logic behind that move at all |
| 12:37 | <hsivonen> | also, I don't like the design pattern for aria-checked, but I guess it's too late to complain about that one |
| 12:37 | <annevk42> | hmm Hixie complicated things so much with the storage mutex and task queues |
| 12:37 | annevk42 | grmbls |
| 12:41 | <hsivonen> | http://twitter.com/myakura/statuses/1867322694 |
| 12:43 | <annevk42> | while certainly practical, do we really want to alienate yet another bunch of people? :) |
| 12:50 | <Dashiva> | Bikeshed a new name instead? |
| 12:50 | <hsivonen> | Maybe the ODF proponents shouldn't have bragged about how short their spec is: http://blogs.msdn.com/dmahugh/archive/2009/05/09/1-2-1.aspx |
| 12:54 | <annevk42> | "Text: The specific conversion is implementation-defined" |
| 12:54 | <annevk42> | yay |
| 12:55 | <hsivonen> | they should have defined .doc5, .xls5 and .ppt5 instead of these ODF and OOXML things |
| 12:55 | <Dashiva> | <insert joke about ODF5> |
| 12:55 | Dashiva | shakes fist at hsivonen |
| 12:58 | <annevk42> | I guess the conversion is impl specific because it can depend on locale and such as well |
| 12:59 | <annevk42> | Sort of glad I'm not involved with that though admittedly we have similar issues with e.g. character encoding |
| 13:04 | <annevk42> | also check the comments on that MS blog post btw |
| 13:09 | <annevk42> | hmm: http://www.schneier.com/blog/archives/2009/05/on_the_anonymit.html |
| 13:30 | <Dashiva> | "So that everyone can be sure about which parts are which, the normative parts use specific phrases like “shall” and “shall not” to clearly label the things the standard actually requires you to do." |
| 13:30 | <Dashiva> | Reminds me of that specs vs blogs post :) |
| 13:32 | <annevk42> | if I'm to believe that person it sounds like ODF 1.1 has been written in HTML4 style |
| 13:38 | <annevk42> | That last email from Laura is interesting. Rather than saying what she thinks she selectively highlights parts from past surveys. (I.e. not mentioning that there was 5 to 1 agreement for the principle but rather that people disagreed and as well as strongly disagreed with the principle.) |
| 13:38 | <annevk42> | I seem to recall this happening more often, but I could be wrong |
| 13:40 | <annevk42> | And the last bit seems simply misleading as Larry was talking about the name of the draft which was a separate question from the one she highlights. |
| 13:57 | <Philip`> | Hmm, people are saying things related to me on Twitter but I've got no way to respond |
| 13:57 | Philip` | guesses he should just send email or something |
| 13:57 | <annevk42> | they are? |
| 13:58 | <Dashiva> | Philip`: Post it anywhere else and get someone to RT it ;) |
| 13:59 | <zcorpan_> | hsivonen: "Actually, I’ve noticed that on this page the social bookmark links create this error: “WHITESPACE in QUERY” Er, what?" - http://www.tomleadbetter.co.uk/blog/entry/html5/ |
| 14:01 | <zcorpan_> | that page also has <input type="url" id="url" name="url" value="http://" /> which is not valid |
| 14:01 | <annevk42> | ah, burningbird + samruby |
| 14:01 | <annevk42> | seems we're following similar twitter queries :) |
| 14:03 | <Philip`> | I'm not really following any, I just randomly look at random people's Twitter pages sometimes :-) |
| 14:03 | <zcorpan_> | hsivonen: and "</l↩</ul>" which could have slightly clearer messages maybe |
| 14:04 | <Philip`> | I assume proper following would require a Twitter account, which is far too advanced for me when I haven't even got a blog yet |
| 14:04 | <Philip`> | (I haven't even got a proper homepage yet, so I'm ~15 years behind the times) |
| 14:06 | annevk42 | hits the next issue in understanding task queues |
| 14:07 | <zcorpan_> | "The contact form also has several ARIA attributes, and if you look at it in Opera, you’ll see it has a nifty built-in validation." |
| 14:07 | <hsivonen> | what kind of validations does Opera have for ARIA? |
| 14:07 | <hsivonen> | or is that about WF2 validation? |
| 14:07 | <zcorpan_> | hsivonen: none, it's type="url" etc |
| 14:07 | <zcorpan_> | yep |
| 14:57 | <annevk42> | sigh |
| 14:57 | <annevk42> | event loops differ between workers and browsing contexts |
| 14:58 | <hsivonen> | how? |
| 14:58 | <annevk42> | in browsing contexts they're always associated with a Document |
| 14:58 | <annevk42> | so when a Document gets garbage collected all the associated events get too (I think) |
| 14:59 | <annevk42> | this means that XHR cannot tie into existing event loops or has to tie into different ones depending on whether XMLHttpRequest Document exists |
| 14:59 | <annevk42> | neither is nice |
| 15:00 | <hsivonen> | I have tried not to look at GC interaction of native and JS objects |
| 15:00 | <hsivonen> | I'm assuming magic happens while I'm not looking |
| 15:00 | <annevk42> | it would probably be much easier if Hixie sorted all this out though one problem with that is he'd be the only spec writer to understand it |
| 15:05 | <Philip`> | Do implementors understand it? |
| 15:05 | <hsivonen> | mpilgrim: fwiw, I read Sam's process post as "put money where your mouth is (by producing camera-ready text) or stop complaining" |
| 15:07 | <hsivonen> | mpilgrim: and the license thing matters, because camera-ready text only works if it doesn't come with copyright strings attached |
| 15:07 | <annevk42> | Philip`, I think so, given the recent WebKit comments on Workers |
| 15:08 | <annevk42> | Philip`, but I'm not a 100% sure |
| 15:08 | <annevk42> | hmm, midori crashes all the time and the current chromium-browser cannot load any page whatsoever -- FAIL |
| 15:30 | <Philip`> | http://linuxcentre.net/adobe-has-issued-a-dmca-removal-request-for-rtmpdump/ - hooray for DRM |
| 15:31 | <Philip`> | (rtmpdump is a tool for saving Flash streaming video to disk) |
| 15:32 | <Philip`> | I guess I'll have to delete the version off my disk, and stop downloading TV programmes that are broadcast for free anyway |
| 16:11 | Dashiva | is catching up on public-html mail from the two last weeks... |
| 16:11 | <Dashiva> | Can anyone tell me, is there a happy ending? |
| 16:12 | <annevk42> | anything going wrong? |
| 16:13 | <annevk42> | hehe, I like how I found http://www.w3.org/TR/qaframe-spec/ again |
| 16:14 | <annevk42> | thanks for writing that karlcow! |
| 16:14 | <Dashiva> | Well, so far there seems to be a lot of heat about rdfa |
| 16:16 | <Philip`> | Someone tried to start writing a concrete document on how to use RDFa in HTML5 |
| 16:16 | <Philip`> | Don't know if it's a good idea or not |
| 16:17 | <gsnedders> | w00t! |
| 16:17 | <gsnedders> | Maths exam over! |
| 16:17 | <Dashiva> | There was one line, about how the RDFa community would gladly ensure incorporating RDFa was as painless as possible |
| 16:18 | <Dashiva> | And I was wondering if that was the same RDFa community insisting on namespaces and CURIEs |
| 16:19 | <Philip`> | Dashiva: I think some people in that community have expressed a willingness to consider alternate prefix-binding mechanisms rather than xmlns:* |
| 16:19 | <Philip`> | which would avoid the text/html-specific pain |
| 16:21 | <Dashiva> | That sounds good |
| 16:21 | <Philip`> | I don't know whether they want to incompatibly change RDFa-in-XHTML to use that same mechanism, though |
| 16:22 | <Philip`> | or change it and leave the old deprecated mechanism in there too, so that there's three ways to specify prefixes |
| 16:22 | <Philip`> | or do something else |
| 16:23 | <Dashiva> | Mhm |
| 17:35 | <gsnedders> | Time to do more work on html5lib-php, me thinks |
| 17:35 | <gsnedders> | (Like working on the 23 failing tokenizer tests) |
| 17:49 | gsnedders | fixes 10 |
| 17:50 | <annevk4> | the phone discussion just now was interesting |
| 17:50 | <annevk4> | apparently Larry thinks that the design principles we drafted do not reflect how HTML5 is written though he did not elaborate much on why when asked |
| 17:51 | <gsnedders> | http://www.w3.org/2009/05/21-html-wg-minutes.html |
| 17:51 | <smedero> | well he did refer to table @summary and head @profile as his examples |
| 17:51 | <smedero> | i think i kinda got what he meant... but it felt like he was dancing about something else he really wanted to say |
| 17:52 | <smedero> | he did note he needed some time to formulate his comments... |
| 17:53 | <annevk4> | still not really clear to me how @profile is related to caving cowpaths, but fair enough |
| 17:55 | <annevk4> | I think the problem with the HTML WG meeting is that a lot of time people start thinking while the telcon is happening rather than planning ahead of time what they want to say whenever the next telcon is |
| 17:56 | <annevk4> | This causes a lot of lost time for the other participants if someone is semi-brainstorming while everyone is dialed in |
| 17:56 | <smedero> | indeed. |
| 18:02 | <gsnedders> | So I think all the tokenizer test cases html5lib-php fails are either disputable or the parse errors thrown in 9.2.4 Tokenization (and not subsections) |
| 18:03 | <Philip`> | Why are they disputable? |
| 18:03 | <gsnedders> | Philip`: The spec doesn't say what order parse errors have to be thrown in |
| 18:04 | <gsnedders> | (Or more to the point, for null bytes and non-characters, when they have to be thrown) |
| 18:05 | <gsnedders> | html5lib-php throws all those parse errors first, then does the actual tokenizing |
| 18:07 | <Philip`> | hsivonen: "Thora" typo in diagram |
| 18:07 | <Philip`> | gsnedders: The test cases should have ignoreErrorOrder:true (or something like that) in those cases |
| 18:14 | <Dashiva> | "Microdata will only work in HTML5/XHTML5. XHTML 1.1 and yes, 2.0 will be around for years, decades." |
| 18:14 | <Dashiva> | So we're still on the "If it doesn't validate, it doesn't work" page? |
| 18:15 | <gsnedders> | Dashiva: Yes |
| 18:20 | <Dashiva> | "Reversed domains are as ugly as CURIEs (but at least CURIEs resolve to something useful [...]" |
| 18:21 | <Dashiva> | Just avoiding the redundant http:// makes them a lot prettier in my book :) |
| 18:21 | <gsnedders> | s/resolve/might resolve/ |
| 18:22 | <Dashiva> | "Are you thinking that someone will be using different namespaces but the same prefix? Come on -- do you really think that will happen?" |
| 18:23 | <Dashiva> | Well, if prefixes never collide, why do we bother with namespaces? Why not just use the globally unique prefixes? |
| 18:25 | <hober> | Dashiva: indeed. |
| 18:30 | <annevk4> | Dashiva, can you please provide pointers with these quotes? |
| 18:39 | <Dashiva> | Sure |
| 18:39 | <Dashiva> | Any particular one you want right now? |
| 18:40 | <Dashiva> | @annevk |
| 18:40 | <annevk4> | 1 and 3 but 2 might be fun as well for completeness |
| 18:48 | <gsnedders> | http://code.google.com/p/html5lib/source/detail?r=8f225971195ed9f2e6a8e28abaa01dc35100decb |
| 18:48 | <gsnedders> | w00t fun |
| 18:48 | <gsnedders> | I <3 languages with shit Unicode support. |
| 18:48 | <annevk4> | PHP6! |
| 18:50 | <gsnedders> | Um, have we been talking about that for as long as DNF yet? :P |
| 18:50 | gsnedders | can remember when there was talk of getting an alpha out by the end of 2005 |
| 18:50 | <gsnedders> | (As of today, we are still pre-alpha) |
| 18:52 | <Dashiva> | Hmm, can I easily get a link from the whatwg archives from a message-id? |
| 18:52 | <annevk4> | no :( |
| 18:53 | <annevk4> | maybe we should write a script that does enables that |
| 18:53 | <Dashiva> | First one: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-May/019751.html |
| 18:54 | <Hixie> | Dashiva: if you have the body of the e-mail, searching for that on google works |
| 18:55 | <gsnedders> | 10 test failures… |
| 18:56 | <Dashiva> | Second: http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-May/019749.html |
| 18:58 | <Dashiva> | Third: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-May/019752.html |
| 18:59 | <Dashiva> | annevk4: Would also be nice if the "by date" view segmented the list by day |
| 19:03 | <annevk4> | Dashiva, hmm, I might just be compelled enough to try to do something |
| 19:03 | <annevk4> | been a while since I wrote Python |
| 19:14 | <Madness> | A quick question about html5 and links around stuff: Suppose I have a link with overflowing content. Should the clickable area be limited o the a element or extended to the overflow too? |
| 19:16 | <Hixie> | Madness: the click on any text node or box that is a descendant of the element would bubble to the element and then cause the link to be followed |
| 19:17 | <Madness> | Thank you. |
| 19:22 | <Hixie> | there's a (!@$%*! registration fee for the tpac meeting? |
| 19:22 | <Hixie> | are you kidding me? |
| 19:23 | <Hixie> | we have enough trouble getting people to attend the htmlwg meeting in the first place |
| 19:24 | <annevk42> | how much? |
| 19:25 | <Hixie> | $50 |
| 19:25 | <annevk42> | per day or for the whole event? |
| 19:25 | <Hixie> | event |
| 19:25 | <Hixie> | er |
| 19:25 | <Hixie> | day! |
| 19:25 | <Hixie> | jesus! |
| 19:25 | <Hixie> | that's even worse! |
| 19:26 | <annevk42> | so I guess it's more than food costs then |
| 19:26 | <Hixie> | i can't justify $350 for a meeting i don't even want to go to in the first place! |
| 19:26 | <Hixie> | that's insane! |
| 19:26 | <annevk42> | you paid more last year |
| 19:26 | <annevk42> | well, your employer |
| 19:27 | <Hixie> | i only paid travel last year |
| 19:27 | <Hixie> | and hotel |
| 19:27 | <annevk42> | and food |
| 19:27 | <Hixie> | i have to pay for food even if there's n meeting |
| 19:27 | <smedero> | is that info on a members only list? |
| 19:27 | <Hixie> | http://www.w3.org/2009/11/TPAC/overview.html#Registration |
| 19:27 | <annevk42> | I usually pay a lot more for food when there's a meeting :) |
| 19:28 | <smedero> | lovely. following links around the w3.org site, I ended up here: http://www.w3.org/2002/09/TPOverview.html |
| 19:28 | <smedero> | heh |
| 19:28 | <annevk42> | "The registration fee will increase to 75 USD per person, per day after a date approximately 4-6 weeks prior to the meeting." |
| 19:29 | <smedero> | yowsers |
| 19:29 | <annevk42> | fortunately the dollar ain't worth shit |
| 19:29 | <annevk42> | :p |
| 19:29 | <Hixie> | yeah well i'm paid by the $ so... |
| 19:29 | <Hixie> | also, google already paid the insane 5-digit membership fee this year |
| 19:30 | <Hixie> | so charging again for the meeting is crazy |
| 19:35 | annevk42 | investigates cron jobs for the alternate mail index |
| 20:08 | <hsivonen> | is the HTML WG having a meeting at TPAC? |
| 20:12 | <smedero> | I thought so... |
| 20:12 | <smedero> | Didn't Chris Wilson send something around? |
| 20:13 | <gsnedders> | He was asking about interest for a meeting, IIRC |
| 20:13 | gsnedders | isn't really sure, as it's highly unlikely he'll be there, and thus doesn't care |
| 20:37 | <gsnedders> | What file do I add tests to for the tokenizer? |
| 20:38 | <gsnedders> | What's the diff between test1, test2, test3 and test4? |
| 20:39 | <Philip`> | One of them is (originally) automatically generated, and full of weird short inputs, so you probably shouldn't use that one |
| 20:39 | <Philip`> | There's no clear difference between the others, as far as I'm aware |
| 20:40 | <Philip`> | so pick one that has some existing tests that are vaguely similar to the one you're adding |
| 20:40 | <Philip`> | or pick one at random |
| 20:42 | <Philip`> | or think up a new system that makes sense, and rewrite all the tests into a different layout |
| 20:47 | <annevk4> | Philip`, any chance you can put <meta name=robots content=index,nofollow> on your survey pages? |
| 20:47 | <annevk4> | Philip`, I'm getting reqeusts from spammers if they can put spam links on your page |
| 20:48 | <annevk4> | Philip`, e.g. http://philip.html5.org/data/charsets.html |
| 20:49 | <annevk4> | Philip`, it's not too serious fwiw |