| 02:01 | <Hixie> | aboodman: no thoughts at all until now. seems like something that would be interesting as a separate web apps spec. I'd be concerned about doing any crypto anything over HTTP. |
| 06:33 | <zcorpan_> | MikeSmith: your definition of doctype doesn't match the spec |
| 06:33 | <MikeSmith> | zcorpan_: which part doesn't match? |
| 06:33 | <zcorpan_> | MikeSmith: the spec allows whitespace before the '>' |
| 06:34 | <MikeSmith> | ah, OK |
| 06:34 | <zcorpan_> | MikeSmith: can't you copy the spec's definition whole-sale? it's less verbose now |
| 06:35 | <MikeSmith> | zcorpan_: maybe. for now, I've just added "\s+" |
| 06:35 | <zcorpan_> | shouldn't it be \s* |
| 06:35 | <zcorpan_> | ? |
| 06:36 | <MikeSmith> | yeah, \s* |
| 06:41 | <zcorpan_> | MikeSmith: "Non-void elements have an end tag to indicate where they end." -- foreign self-closing elements don't have an end tag |
| 06:41 | <MikeSmith> | zcorpan_: thanks -- OK, will fix that now |
| 06:45 | <zcorpan_> | MikeSmith: "Void elements that are not from the MathML namespace" doesn't make sense since the list of void elements only includes elements from the HTML namespace |
| 06:47 | MikeSmith | looks now |
| 06:52 | <zcorpan_> | MikeSmith: the example on CDATA section doesn't need a CDATA section -- it doesn't contain any < or & |
| 06:53 | <MikeSmith> | zcorpan_: yeah, I know. but I copied that straight from the MathML spec.. that's the example they have in there. Not sure why they're using that as an example |
| 06:53 | <MikeSmith> | maybe I can revise it to add a < or & into it? |
| 06:54 | <zcorpan_> | if it would make sense, otherwise perhaps come up with some other example :) also show an equivalent example that uses entities instead |
| 06:57 | <MikeSmith> | zcorpan_: OK, I've added to my TODO list. I think I might ask the MathML WG to add those to the examples in the MathML spec itself |
| 06:57 | <MikeSmith> | if they don't want to, I guess I'll just invent some |
| 06:58 | <MikeSmith> | but since CDATA sections are only allowed with MathML content, and I don't know anything about MathML, I'd rather they provide the examples |
| 06:59 | zcorpan_ | finds an example on the web that uses <mi><![CDATA[&infinity;]]></mi> which probably isn't what was intended |
| 06:59 | <MikeSmith> | zcorpan_: in the mean time, I did go ahead and check in your other changes |
| 06:59 | <MikeSmith> | e.g, at http://www.w3.org/html/wg/markup-spec/#the-doctype |
| 07:00 | <zcorpan_> | MikeSmith: cool |
| 07:00 | <zcorpan_> | "cdata" is so not searchable |
| 07:01 | <MikeSmith> | thanks, and keep them coming, when you find more |
| 07:01 | <MikeSmith> | zcorpan_: ? |
| 07:01 | <MikeSmith> | searchable in what context |
| 07:01 | <MikeSmith> | ? |
| 07:01 | <zcorpan_> | it means 10 things so if i want to google for usage of cdata sections i get dtd fragments |
| 07:01 | <zcorpan_> | or attribute definitions |
| 07:01 | <zcorpan_> | or cdata elements |
| 07:01 | <zcorpan_> | etc |
| 07:01 | <MikeSmith> | ah |
| 07:02 | <MikeSmith> | you need some semantic web, man |
| 07:03 | <zcorpan_> | or google code search :) |
| 07:03 | <MikeSmith> | yeah, or MAMA! |
| 07:03 | <zcorpan_> | i haven't looked into how to use mama |
| 07:03 | <MikeSmith> | I would really like to have search access to mama |
| 07:04 | <MikeSmith> | blooberry: ↑ |
| 07:04 | blooberry | looks |
| 07:05 | <MikeSmith> | zcorpan_: you a pythonist? what's the best-practice way to remove leading/trailing whitespace from a string in python? |
| 07:05 | <blooberry> | mikesmith: 8-} |
| 07:05 | <MikeSmith> | blooberry: open data, brother. Lead by example! |
| 07:05 | <MikeSmith> | just giving a hard time |
| 07:05 | <blooberry> | mama currently mostly ignores cdata sections. *puts that on the to-do list* |
| 07:06 | <MikeSmith> | .me (re)finds .strip() |
| 07:07 | <blooberry> | mikesmith: I'm working on it. 8-} I need to do a current re-crawl of the URLs MAMA looked at last year first and improve the search perf. |
| 07:07 | <MikeSmith> | blooberry: you mostly still doing the work on it yourself, or you got some assistance now? |
| 07:08 | <blooberry> | it's all me still. 8-/ |
| 07:08 | <zcorpan_> | StoreRaw("<ow:math><![CDATA[" & Replace(pText, "]]>", "]]>") & "]]></ow:math>") -- that surely won't do what's intended |
| 07:08 | <zcorpan_> | as it happens i can't find a single valid use of cdata in mathml |
| 07:08 | <blooberry> | I'm adding new features though to the analysis script right now. Have any other requests? |
| 07:09 | <MikeSmith> | blooberry: nothing I can think of now |
| 07:09 | <zcorpan_> | blooberry: is it possible to do regexp search on the source markup? |
| 07:10 | <blooberry> | zcorpan_: no. It stores a greatly deflated version of the page right now. |
| 07:11 | <zcorpan_> | that's what Philip` usually does (http://philip.html5.org/data.html) and is basically what i would like to do most often |
| 07:11 | <blooberry> | The "manual" version of the analysis script can do regexp searching of all content though, but you are probably looking for regexp searching on a batch of pages, yes? |
| 07:11 | <zcorpan_> | yeah |
| 07:11 | <zcorpan_> | doesn't need to be a million pages |
| 07:12 | <blooberry> | yeah, I've thought of doing something like that for a much smaller subset |
| 07:12 | <zcorpan_> | consider it a feature request :) |
| 07:12 | <blooberry> | philip` is covering that for some values of smaller sets though already |
| 07:12 | <blooberry> | 8-} |
| 07:12 | <zcorpan_> | yeah but i hate to bother him all the time |
| 07:13 | <zcorpan_> | and his batch of pages is out of date :P |
| 07:13 | <blooberry> | so is mine, at the moment. 8-} |
| 07:14 | <blooberry> | (how "out of date" is his set?) |
| 07:14 | <zcorpan_> | at least a year i think |
| 07:14 | <zcorpan_> | but could be wrong |
| 07:14 | <blooberry> | about what mine is too. |
| 07:15 | <zcorpan_> | <replacetoken><![CDATA[<math>]]></replacetoken> |
| 07:15 | <zcorpan_> | <replacevalue><![CDATA[<math xmlns="http://www.w3.org/1998/Math/MathML">]]></replacevalue>; |
| 07:15 | <blooberry> | I'd like to start a re-crawl by the new year, but it may not happen that soon. Should be january at the latest though. |
| 07:16 | <zcorpan_> | cool |
| 07:18 | <blooberry> | so, I'm contemplating having MAMA store some stats about replaced object dimensions, but I'm not sure I can come up with a solution that addresses interesting problems and is also a good storage compromise |
| 07:19 | <MikeSmith> | zcorpan_: are those replacetoken examples good examples? |
| 07:20 | <blooberry> | and if I factor in resizing via CSS or (worse) Javascript, then the problem possibly becomes a silly proposition |
| 07:20 | <zcorpan_> | MikeSmith: no, they're just interesting |
| 07:20 | <MikeSmith> | OK |
| 07:21 | <zcorpan_> | MikeSmith: the code assumes that mathml markup thrown at it won't have the namespace declaration |
| 07:21 | <MikeSmith> | OK |
| 07:22 | <zcorpan_> | MikeSmith: i recommend replacing the example with svg <script><![CDATA[... when svg is back in the spec |
| 07:23 | <MikeSmith> | zcorpan_: yeah, definitely |
| 07:25 | <zcorpan_> | MikeSmith: a:-webkit-any-link translates to a:link, a:visited |
| 07:26 | <MikeSmith> | zcorpan_: ah, OK |
| 07:26 | MikeSmith | fixes now |
| 07:37 | <zcorpan_> | MikeSmith: the <input> section would be more useful if split up to its type='' states |
| 07:39 | <MikeSmith> | zcorpan_: yeah, I think so. I guess I'll need to have the build script do that. It's too much of a mess to read through the way it is now. |
| 07:40 | <zcorpan_> | it looks like the schema treats them as different elements basically |
| 07:40 | <BenMillard> | the different states of <input> are more like different elements in the way they are used |
| 07:40 | <BenMillard> | oh, zcorpan_ types faster than me :P |
| 07:41 | <MikeSmith> | IMHO, they should in fact have been different elements from the beginning. it's bad language design the way it is now |
| 07:41 | <MikeSmith> | the complexities of documenting it expose that |
| 07:42 | <zcorpan_> | html has lots of bad langauge design :) |
| 07:42 | <MikeSmith> | true |
| 07:42 | <MikeSmith> | but <input> sticks out like sore thumb, in its particular badness |
| 07:43 | <zcorpan_> | for your use case yes :) |
| 07:43 | <MikeSmith> | zcorpan_: added to TODO list |
| 07:44 | <zcorpan_> | if you're writing html you're probably more concerned about <img alt> not taking markup |
| 07:44 | <zcorpan_> | it doesn't bother me so much (when writing markup) that <input type=image> and <input type=file> aren't different elements |
| 07:46 | <zcorpan_> | i wonder why map { display:inline } is in the style sheet |
| 07:48 | <MikeSmith> | zcorpan_: moz stylsheet has: |
| 07:48 | <MikeSmith> | html, div, map, dt, isindex, form { |
| 07:48 | <MikeSmith> | display: block; |
| 07:49 | <zcorpan_> | so they had it as block and then changed it to inline? or is it different in quirks mode? |
| 07:50 | <MikeSmith> | zcorpan_: dunno, webkit has it as inline, moz has a block |
| 07:50 | <MikeSmith> | as far as I can see |
| 07:51 | <zcorpan_> | oh. in moz it's inline in quirks |
| 07:52 | <zcorpan_> | in webkit it's always inline. still, i wonder why webkit has the declaration -- it's inline by default |
| 07:53 | <zcorpan_> | MikeSmith: the spec has class="bad example" or some such in a different style than normal examples |
| 07:53 | <MikeSmith> | zcorpan_: OK, will take a look |
| 07:54 | <zcorpan_> | MikeSmith: should say that noscript is only allowed in HTML |
| 07:55 | <MikeSmith> | zcorpan_: you mean as opposed to MathML or SVG? |
| 07:56 | <zcorpan_> | MikeSmith: as opposed to XHTML |
| 07:56 | <MikeSmith> | ah |
| 07:57 | <zcorpan_> | hmm, webkit's default style for <q> doesn't handle nested <q>s well |
| 07:57 | <zcorpan_> | in fact i think webkit doesn't support open-quote etc at all |
| 07:58 | <BenMillard> | zcorpan_, I've seen sites which set quotes on <q> via q:before and q:after, which might be due to that? |
| 07:59 | <BenMillard> | (the content property via those selectors, not the quotes property) |
| 08:01 | <MikeSmith> | zcorpan_: I've set pre.bad to "color: red" for now |
| 08:05 | <MikeSmith> | zcorpan_: fixed the -webkit-any-link thing too |
| 08:05 | <MikeSmith> | http://www.w3.org/html/wg/markup-spec/#a-display |
| 08:09 | <MikeSmith> | zcorpan_: I'll try to get input split up by some time next week |
| 08:12 | <MikeSmith> | for consistency, maybe it make sense to split up button and command as well |
| 08:13 | <MikeSmith> | hmm, or maybe not command, but button at least |
| 10:38 | <zcorpan_> | BenMillard: you'd still have q:before { content:open-quote } |
| 11:16 | <Philip`> | zcorpan_: It isn't any bother for me to run regexp searches, but if you're looking for legitimate <![CDATA[ uses then I probably couldn't help much since I have almost no XML pages |
| 11:23 | <Philip`> | zcorpan_: Also, if you just want to run regexps yourself, you could probably easily download a load of pages (e.g. get the list from dmoz.org, pick a random sample, pass to curl, it's only about 25KB per page so a million fits in 25GB of disk easily), and then run grep and be very patient :-) |
| 11:28 | <MikeSmith> | zcorpan_: I added build support for generating per-element "Additional constraints" subsections. So far, the only one is for stating the "The noscript element must not be used in a document in the XML syntax." constraint. |
| 11:28 | <MikeSmith> | http://www.w3.org/html/wg/markup-spec/#noscript-constraints |
| 12:01 | <gsnedders> | Anyone know of any definition of BibTeX? |
| 13:35 | <gsnedders> | "two fields must be separated by a coma" — nice to know |
| 13:47 | <gsnedders> | BibTeX is ewww |
| 13:56 | <Philip`> | gsnedders: If you could design something better, and could do it 23 years ago and could get everyone to use it, that would be greatly appreciated |
| 13:56 | <gsnedders> | :) |
| 13:56 | <gsnedders> | I mean, what's wrong with RDF!? |
| 16:34 | <gsnedders> | Hmm, all the introduction sections should really be numbered section 0 within their depth |
| 18:05 | <annevk3> | http://twitter.com/maercsratS/statuses/1040686309 :) |
| 18:10 | <Philip`> | That looks like the result of excellent marketing - they're excited about HTML5 because of two features which are not actually in HTML5 and never have been |
| 19:39 | <Lachy> | Philip`, the web workers stuff is understandable since the spec was also edited by Hixie and was discussed in the whatwg and/or html wg lists |
| 19:40 | <Lachy> | but I'm surprised about how anyone could possibly think geolocation is part of html5 |
| 19:42 | gsnedders | wonders WTF is up with Anolis |
| 19:54 | <gsnedders> | html5lib bug |
| 19:58 | <gsnedders> | Hixie: you around? |
| 21:23 | <Hixie> | Philip`: web workers is in html5 according to the web workers spec |
| 21:43 | <gsnedders> | Hixie: Can you not have to have .copyright on each p, but just have a div instead around them? |
| 22:22 | <Hixie> | gsnedders: why? |
| 22:46 | <Lachy> | http://lists.w3.org/Archives/Public/public-webapps/2008OctDec/0424.html I'm surprised the svg wg wants me to remove the authoring requirements from selectors api |
| 23:18 | <gsnedders> | Hixie: So I don't have to repeat @class='copyright' |
| 23:18 | <Hixie> | ? |
| 23:18 | <gsnedders> | Hixie: I'm lazy. |
| 23:19 | <Lachy> | gsnedders, what is it that you're doing with the copyright stuff which is made more difficult? |
| 23:20 | <gsnedders> | Lachy: Nothing. I'm just lazy. |
| 23:20 | <Lachy> | what? Then how are you affected by it at all? |
| 23:21 | <gsnedders> | Lachy: Because I'm too lazy to repeat @class='copyright' :P |
| 23:21 | <gsnedders> | I want <div class="copyright"><p>foo<p>bar</div> instead of <p class=copyright>foo<p class=copyright>bar |
| 23:21 | <Lachy> | is this something to do with anolis? |
| 23:21 | <gsnedders> | Nope |
| 23:22 | <Lachy> | you're not being clear at all. |
| 23:22 | <gsnedders> | I'm too lazy to repeat the class attribute. That is all. |
| 23:22 | gsnedders | doesn't see how this is unclear |
| 23:24 | <Lachy> | because I don't know what you are doing that makes you have to repeat it? All I can tell is that this has something to do with the whatwg spec's header file |
| 23:25 | <gsnedders> | I have the MIT license which has four paragraphs. |
| 23:26 | <gsnedders> | I, like a number of other people here, use the whatwg stylesheet |
| 23:27 | <Lachy> | oh, so your problem is in fact the stylesheet using p.copyright instead of just .copyright |
| 23:27 | <gsnedders> | yeah |
| 23:28 | <Lachy> | well, why didn't you just say so?! |
| 23:28 | gsnedders | thought he did |
| 23:28 | gsnedders | is half alseep |
| 23:29 | <Lachy> | no, you asked Hixie to change the markup in the spec, which doesn't directly address your problem, although it would have indirectly made him update the stylesheet. |
| 23:30 | <gsnedders> | Lachy: Don't expect sense out of me at this time! |
| 23:31 | <Lachy> | it's ok. I'm' used to people failing to explain their problems and use cases properly :-) |
| 23:33 | <Hixie> | gsnedders: why do you have two copyrights? |
| 23:33 | <gsnedders> | Hixie: For the same reason you do, the copyright info is more than one paragraph. |
| 23:34 | <gsnedders> | Hixie: As I said, I have the full MIT license |
| 23:34 | <gsnedders> | four paragraphs off the top of my head |
| 23:34 | <Hixie> | don't include the whole license on the page :-P |
| 23:34 | <Hixie> | stick it in an appendix if you must |
| 23:34 | <Hixie> | or just refer to it |