| 01:52 | <Hixie> | othermaciej: any news on http://www.w3.org/mid/4ACB684A.9020505⊙wo ? |
| 06:57 | <hsivonen> | someone has said something to me but my IRC client no longer remembers |
| 06:58 | <annevk42> | Hixie, <span title="dom-input-value-default">value</span> s/value</default</ |
| 06:59 | <Hixie> | thanks fixed |
| 07:00 | <hsivonen> | I'm surprised that Larry seems unaware of the reason for the Windows-1252 default. |
| 07:18 | <othermaciej> | Hixie: Jeanne agreed to refile those as bugs |
| 07:19 | <Hixie> | k thx |
| 08:40 | <ginger> | M65FLXFAILZAOVV3MGBZKTGQDLOPNXSP6ORA4VCGOV5S6YEIMHJK- |
| 08:40 | <ginger> | MH3GP3PAP4MTWCLHNEB62EZON43CNIHWOBXMSATQKC6VMLXTC6QJ- |
| 08:40 | <ginger> | 4JCRXRKCLNMSNRTDZTYX25DU5BXTRDLE5BBPIWSPIVSCGOAQY |
| 08:40 | <ginger> | ups |
| 08:40 | <ginger> | sorry |
| 08:42 | <ginger> | exit |
| 08:43 | <pesla> | lol |
| 08:45 | <othermaciej> | hsivonen: do you know where in a Firefox .app bundle I can find the default encoding used for HTML in this localization? or is the table in the source somewhere? |
| 08:55 | <hsivonen> | othermaciej: in the local jar (en-US.jar in the en-US case) in file locale/LOCALENAME/global-platform/OS/intl.properties |
| 08:55 | <hsivonen> | othermaciej: pref intl.charset.default |
| 08:56 | <hsivonen> | where LOCALENAME is a code like en-US |
| 08:56 | <hsivonen> | and OS is mac, unix or win |
| 08:56 | <hsivonen> | s/local jar/locale jar |
| 08:56 | <othermaciej> | is that in the source code or in the binary? |
| 08:57 | <othermaciej> | ok I found it |
| 08:57 | <hsivonen> | http://mxr.mozilla.org/l10n-mozilla1.9.1/find?text=&kind=text&string=intl.properties |
| 08:58 | <othermaciej> | we should probably copy the Firefox behavior for Safari/WebKit |
| 08:58 | <hsivonen> | Korean sets it to EUC-KR, for example: http://mxr.mozilla.org/l10n-mozilla1.9.1/source/ko/toolkit/chrome/global/intl.properties |
| 08:58 | <othermaciej> | right now we use WinLatin1 as the default for all locales |
| 08:59 | <hsivonen> | othermaciej: interesting! I guess it's good enough, then. Do you know if users routinely change it e.g. in Japan or Russia? |
| 09:00 | <othermaciej> | hsivonen: I think it's not good enough for at least some locales |
| 09:00 | <hsivonen> | ok |
| 09:00 | <othermaciej> | hsivonen: Apple just doesn't sell as much to those locales so it hasn't been a priority |
| 09:00 | <othermaciej> | I know that in Russia, defaulting to Latin1 fails |
| 09:00 | <othermaciej> | I believe it's bad in China and Japan too |
| 09:00 | <othermaciej> | and probably Korea |
| 09:00 | <othermaciej> | other than that, I don't know |
| 09:01 | <othermaciej> | I have not heard of problem reports from Israel, Greece, Eastern/Central Europe, etc where I believe we have non-trivial presence |
| 09:02 | <othermaciej> | I just wanted to verify that some of the countries in Hixie's map of the "Western World" don't get Latin-1 in Firefox |
| 09:03 | <hsivonen> | othermaciej: which country doesn't? |
| 09:03 | <othermaciej> | Poland, Greece, Israel are the three I checked |
| 09:04 | <othermaciej> | based on that I would expect that Czech, Hungary, the baltic states, slovakia, and croatia do not either |
| 09:04 | <hsivonen> | where did Hixie define any of those countries as Western? (I see Firefox defines Greece as Western) |
| 09:04 | <othermaciej> | he linked to this map of "the Western world": http://upload.wikimedia.org/wikipedia/commons/6/6f/Westerncultures_map.png |
| 09:05 | <othermaciej> | (not in the spec, in email discussion) |
| 09:05 | <hsivonen> | OK. that clearly doesn't properly match Windows-1252 |
| 09:06 | <othermaciej> | I suggested that instead of "Western demographics", the spec should refer to "locales that use the Latin script", perhaps with explicit reference to windows-1252 completely or almost completely covering their alphabets |
| 09:06 | <othermaciej> | but it might be that Windows-1252 is also the default in some locales for historic reasons only; I did not check |
| 09:07 | <othermaciej> | I suspect there are also countries not on Hixie's map that would probably use 1252 (probably Albania for instance) |
| 09:07 | <hsivonen> | othermaciej: that would still be wrong for Polish, Czech, etc. |
| 09:08 | <othermaciej> | hsivonen: yes, just saying "Latin script" would be wrong, but saying "locales that primarily use the Latin script, and where Windows-1252 completely or almost completely covers their alphabet" would be right |
| 09:08 | <hsivonen> | othermaciej: I'm not yet sure if that covers Estonia the right way |
| 09:08 | <othermaciej> | the Estonian language is not covered by 1252 |
| 09:09 | <othermaciej> | whether 1252 gets used anyway, I don't know |
| 09:09 | <hsivonen> | othermaciej: but do browsers still traditionally ship with a Windows-1252 default to Estonia? |
| 09:09 | <annevk2> | othermaciej, such a definition is very close to circular |
| 09:09 | <othermaciej> | annevk2: it's not circular at all |
| 09:10 | <annevk2> | i guess not, but still |
| 09:10 | <othermaciej> | it assumes a text encoding based on the alphabet of the locale |
| 09:10 | <hsivonen> | I still think it would be more useful to define the defaults for locales where you don't get away with shipping with Windows-1252 |
| 09:10 | <othermaciej> | which is in fact how that decision is actually made |
| 09:10 | <othermaciej> | hsivonen: there's definitely some locales where shipping Windows-1252 is a big problem |
| 09:11 | <othermaciej> | hsivonen: I don't think any current browser has a custom default encoding for only those locales |
| 09:11 | <hsivonen> | othermaciej: FWIW, in the 1980s, Apple (rightly) decided to ship MacRoman to Finland even though MacRoman doesn't cover the whole alphabet according to orthographers in ivory towers |
| 09:13 | <othermaciej> | hsivonen: I'm not saying it would be wrong to change the default encoding for only a small number of locales; I'm just not sure anyone actually would do that |
| 09:24 | <Hixie> | hsivonen and i were talking about this earlier, and i think the preferable solution would be a table that lists the places where 1252 isn't the default, and lists what the default should be there |
| 09:24 | <Hixie> | and then says "1252 for everything else" |
| 09:25 | jgraham | thinks this has generated a lot of discussion for a non-normative sentence that could be cut from the spec without significant harm |
| 09:25 | <jgraham> | (unless you do actually spec the default in other places but I guess that will cause even more discussion) |
| 09:27 | <annevk2> | Hixie, that sounds good, although it still sucks |
| 09:27 | <annevk2> | one day I'll find a better solution |
| 09:28 | <Hixie> | HTML sucks, but one day we'll find a better solution. :-) |
| 09:28 | <jgraham> | Listing it based on places rather than language sounds wrong. |
| 09:28 | <Hixie> | i don't mind what the key is |
| 09:28 | <othermaciej> | Hixie: I like that idea, if browsers are actually willing to follow such a list |
| 09:28 | <annevk2> | having pages render correctly based on the local of your browser is just a tad too wrong |
| 09:29 | <annevk2> | though apparently so far the only workable solution |
| 09:29 | <jgraham> | Yeah the whole idea of a locale-specific default is clearly broken. But there is nothing we can do |
| 09:29 | <othermaciej> | annevk2: another possibility is to try to guess the language, or do it based on country code TLDs, but I suspect that would be more broken too |
| 09:29 | <hsivonen> | annevk2: Opera Mini doesn't have charset defaulting UI, and Opera Mini works in Norway and Russia. How does it deal? |
| 09:29 | <jgraham> | othermaciej: The spec already allows for that of course |
| 09:30 | <annevk2> | hsivonen, we don't? I've no idea |
| 09:30 | <othermaciej> | The spec does let you autodetect encoding, I don't think it lets you guess based on the TLD |
| 09:30 | annevk2 | is fixing XHR bugs |
| 09:30 | <annevk2> | after that I'm going to strongly consider working on CSSOM again |
| 09:30 | <annevk2> | seems more high-pri than Web Encoding |
| 09:30 | <annevk2> | s |
| 09:31 | <hsivonen> | I have no idea how charsets work in Opera Mini, but the inputs have to be page, heuristics, server domain, client localization and/or client address, because there's no UI |
| 09:32 | <jgraham> | othermaciej: I don't see why "other algorithms to the data stream" couldn't include looking at the URL |
| 09:32 | <hsivonen> | Interestingly, Nokia felt they should include the charset menu in MicroB |
| 09:33 | <annevk2> | mkay, I'm somewhat intrigued now, if I run across someone from what team I'll ask |
| 09:33 | hsivonen | isn't curious enough to install a Russian localization of Mini for testing if the UI language is an input |
| 09:34 | <jgraham> | annevk2: You could go up a couple of floors and you would run into dozens of them ;) |
| 09:34 | <annevk2> | that'd require effort :) |
| 09:36 | <jgraham> | othermaciej: (ALthough I guess data stream could be interpreted as the document rather than the document+headers but it's hard to tell because that section variously talks about "resource", "file" and "data stream") |
| 09:36 | <othermaciej> | jgraham: if it turns out to be a workable solution I'm sure no one will let the spec stop them |
| 09:37 | <jgraham> | othermaciej: Indeed, so we may as well make it explicitly allowed |
| 09:38 | jgraham | will file a bug |
| 10:02 | <annevk2> | I guess XHR should get domintro boxes as well |
| 10:02 | <annevk2> | but maybe during CR so I don't have to fix bugs all the time in two places |
| 10:02 | annevk2 | adds a note |
| 10:04 | <gsnedders> | domintro? |
| 10:05 | <gsnedders> | Just introduction to the DOM in general, or? |
| 10:08 | <annevk2> | gsnedders, class=domintro |
| 10:08 | <annevk2> | see html5 |
| 11:34 | <annevk2> | funny thing: UTF-8 doubles size of requests so use a single-byte encoding and percent-escape it |
| 11:35 | <Dashiva> | ... what? |
| 11:35 | <Dashiva> | Percent-encoding can potentially triple the size, though |
| 11:41 | <hsivonen> | annevk2: huh? |
| 11:41 | <annevk2> | Dashiva, the triple sizing is the joke |
| 11:41 | <annevk2> | for reference: http://lists.w3.org/Archives/Public/public-webapps/2009OctDec/0136.html |
| 12:17 | <hsivonen> | Out of personal interest, I started a spreadsheet where I try to score W3C TR page publications from 0 to 4 depending on how many of Gecko, WebKit, Presto and Trident make a serious effort to support a given spec |
| 12:18 | <hsivonen> | Three findings: 1) Lots of zeros. 2) I'm ignorant about the implementation status of a lot of stuff. 3) I'm ignorant about the way the SVG WG organizes its specs. |
| 12:20 | <annevk2> | is it an online spreadsheet so others can help out? |
| 12:20 | <hsivonen> | annevk2: not yet. |
| 12:21 | <hsivonen> | annevk2: should I put it on Google Docs or elsewhere? |
| 12:21 | <annevk2> | wiki or google works I suppose |
| 12:21 | <hsivonen> | OK. |
| 12:21 | <othermaciej> | hsivonen: how would you could "serious effort"? |
| 12:22 | <othermaciej> | and what counts as supporting? Does Gecko "support" RDF for instance? |
| 12:22 | <hsivonen> | othermaciej: I marked Gecko as supporting RDF/XML and RDF concepts |
| 12:22 | <hsivonen> | othermaciej: and I marked all four supporting ARIA |
| 12:22 | <Hixie> | i thought we dropped our rdf support |
| 12:22 | <othermaciej> | does Gecko expose its support for RDF/XML to the Web somehow? |
| 12:22 | <hsivonen> | othermaciej: so maybe the bar should be any support at all |
| 12:22 | <hsivonen> | othermaciej: as opposed to "serious" |
| 12:22 | <hsivonen> | othermaciej: I have no idea |
| 12:22 | <othermaciej> | there's a lot of TR publications, wow |
| 12:23 | <hsivonen> | othermaciej: if it doesn't, it shouldn't count towards to score |
| 12:23 | <hsivonen> | othermaciej: see my finding #2 above |
| 12:24 | <annevk2> | you should prolly also have something like spec-obsolete-in-favor-of-something-else |
| 12:24 | <annevk2> | e.g. for HTML4, CSS1, etc. |
| 12:24 | <othermaciej> | sXBL scores a 0 *and* is obsolete in favor of something else |
| 12:25 | <annevk2> | maybe we could just color those light grey |
| 12:26 | <annevk2> | e.g. colors for draft maturity and numbers for impl status per vendor |
| 12:27 | <othermaciej> | I'm definitely interested in seeing this data, and possibly in helping to complete it |
| 12:28 | <smaug> | hsivonen: so are you looking at *all* the TRs? |
| 12:30 | <Hixie> | annevk2: so i have some feedback regarding legacy encodings and stuff like that... should I tell them you're dealing with that? |
| 12:31 | <hsivonen> | smaug: all except Notes and WDs officially classified as abandoned |
| 12:31 | <jgraham> | At least one Note has multiple implementations (XPath) |
| 12:32 | <smaug> | There are of course TRs which aren't for browsers |
| 12:34 | <hsivonen> | smaug: those would simply score 0 as far as implementations in browser engines go |
| 12:34 | <smaug> | right |
| 12:34 | hsivonen | has trouble getting Google Docs accept an HTML file as a spreadsheet |
| 12:35 | <Hixie> | ok bed time now |
| 12:35 | <Hixie> | nn |
| 12:35 | <smaug> | hsivonen: do you count extensions as "make a serious effort to support"? |
| 12:36 | <hsivonen> | smaug: I don't |
| 12:36 | <hsivonen> | hmm. Do Google Docs spreadsheets not support hyperlinks in table cells? |
| 12:37 | <annevk2> | Hixie, not sure, if I'm going to do CSSOM again also dealing with encodings would be distracting |
| 12:37 | annevk2 | needs a real annevk2 |
| 12:37 | <smaug> | Btw, gecko does still support some rdf |
| 12:37 | <smaug> | at least with templates |
| 12:38 | <hsivonen> | smaug: is it Web-exposed? I tentatively scored RDF concepts and RDF/XML as being implemented in Gecko |
| 12:38 | <hsivonen> | I'll try to battle Google Docs later. |
| 12:38 | <smaug> | yes, you can use templates in content pages |
| 12:38 | <hsivonen> | clearly, Google Docs still has a long way to go to compete with Excel and OOo Calc |
| 12:38 | <hsivonen> | smaug: OK. thanks |
| 13:09 | <hsivonen> | OK. I finally managed to get Google Docs ingest the spreadsheet after I used .xls format... |
| 13:09 | <hsivonen> | http://spreadsheets.google.com/ccc?key=0AtnDcoh7FXfAdEFuWVlVVDZTVkFWdnRqaWFGMzNYM3c&hl=en |
| 13:09 | <hsivonen> | now back to work |
| 13:10 | <smaug> | hsivonen: I think Opera has some VoiceXML support (at least Opera used to support XHTML+Voice) |
| 13:11 | <smaug> | ah, you marked it ? |
| 13:11 | <hsivonen> | I think I now made it world-writable |
| 13:13 | <smaug> | hsivonen: should I ? if there is a patch waiting for review to implement a spec? |
| 13:13 | <smaug> | patch which might even land |
| 13:14 | <hsivonen> | smaug: go ahead. Please use the notes column to mention it's about to land |
| 13:17 | <daedb> | CSS question: Are styles with a :lang() selector really supposed to be applied to descendants in a different language? |
| 13:19 | <gsnedders> | Well sure, normal inheritance rules apply. |
| 13:19 | <gsnedders> | How you specify what element to set properties on has no effect on how they are inherited. |
| 13:19 | <gsnedders> | daedb: ^^ |
| 13:20 | <hsivonen> | I hate it when I break Hixie's Live DOM Viewer |
| 13:21 | <daedb> | gsnedders: That seems counter-intuitive to me, but alright :) |
| 13:24 | <Dashiva> | daedb: If you want to limit it, use :lang() to target the descendants with specified language |
| 13:29 | <daedb> | Dashiva: It's not terribly important to me. I was just playing around with it and was surprised with some of the results. |
| 13:37 | <othermaciej> | hsivonen: interesting that there's more 4s than 3s |
| 13:38 | <othermaciej> | doesn't IE support Selectors API? |
| 13:41 | <gsnedders> | othermaciej: IE8 does |
| 13:41 | <othermaciej> | gsnedders: yeah, I fixed it already |
| 13:43 | hsivonen | wonders if someone else broke Hixie's Live DOM Viewer |
| 13:43 | <othermaciej> | hsivonen: I updated a few things - should I re-sort by score? |
| 13:43 | <hsivonen> | othermaciej: sure if you can figure out how to resort. I couldn't. |
| 13:46 | <othermaciej> | hsivonen: hmm, how serious should an attempt at "support" be? I noticed you didn't score anyone for HTML5 |
| 13:46 | <hsivonen> | HTML5 is a tough one |
| 13:46 | <othermaciej> | obviously no one has an implementation that you could claim to be complete yet |
| 13:47 | <hsivonen> | Maybe it should be marked as supported, since the point of the table is to see which specs have part-of-the-platform kind of browser vendor buy-in |
| 13:47 | <othermaciej> | but everyone has a partial implementation, in some cases including quite complicated new parts of the spec, and everyone has an intent to do more |
| 13:48 | <hsivonen> | Whoa! WebKit has MathML support now? |
| 13:48 | <othermaciej> | partial |
| 13:48 | <hsivonen> | cool |
| 13:48 | <othermaciej> | with more in progress |
| 13:48 | <othermaciej> | I think everyone supports 'Syntax of CSS rules in HTML's "style" attribute' |
| 13:49 | <jgraham> | othermaciej: nice |
| 13:49 | <jgraham> | (about the MathML) |
| 13:51 | <othermaciej> | how does one determine if browsers support "Modularization of XHTML 1.0"? |
| 13:51 | <othermaciej> | are the implementation requirements any different than regular XHTML 1.0? |
| 13:51 | <hsivonen> | othermaciej: AFAICT, browsers that don't support DTD loading can't support it |
| 13:52 | <hsivonen> | when there are multiple versions of a spec, I've marked the latest not supported if only the subset already covered by a previous version is supported |
| 13:52 | <hsivonen> | so no browser supports XHTML 1.1 |
| 13:52 | <hsivonen> | since the delta from 1.0 to 1.1 is ruby and no browser supports ruby in the right mime type |
| 13:53 | <hsivonen> | (unless Google's contribution to WebKit landed already) |
| 13:54 | <othermaciej> | I sucessfully sorted |
| 13:54 | <othermaciej> | Google has landed some Ruby stuff, though I don't know if it works sufficiently per XHTML 1.1 requirements |
| 13:55 | <othermaciej> | it's sad that "1" is the second-largest score category after "4" |
| 13:56 | <othermaciej> | though maybe Opera's next major release will greatly increase the quantity of "3" |
| 14:01 | <othermaciej> | I'm tempted to flip some of the HTML5 columns to 1 but I'll leave it to hsivonen's judgment |
| 14:04 | <othermaciej> | it looks like Gecko+WebKit is the most popular 2-combo, and the 1s are pretty evenly distributed among non-Trident engines |
| 14:26 | <boblet> | wonders if <h> might be worth considering for the ongoing <figure> title saga. If there are no repurposable elements that might be a nice one to add… |
| 14:48 | <hsivonen> | i18n FTW. Someone's email client broke Mark Davis' and Martin Dürst's names in email and the errors propagate down the thread. |
| 15:00 | <annevk2> | thanks othermaciej for chiming in |
| 15:00 | <othermaciej> | annevk2: I figured since Mark cited a specific scenario I should actually read it - I kind of tuned this discussion out last time around |
| 15:36 | <TabAtkins> | This is heartening information: http://4734020732036341599-a-1802744773732722657-s-sites.googlegroups.com/site/macchiato/main/updated-unicode-growth/growth_of_unicode_on_the_web.png?attachauth=ANoY7cqYWCRCtFSL4QvZl_L1Bza9-sscXBYmqJ3juoLgiogEdU53GeXZ6kVdMxir7WxjpmA1n0T0yUuu1mRKtw4bTgBXFO9-LalamXut92p1_QkEJrEqjYcyfcatfMF57R9AkxLiXEnXSljJdqihzNtp0KJDYfU6RR3Idxwc5oNgsMKqIJSm6GYFaiyj4dMyps_gby7qLjZhyIQ |
| 15:36 | <TabAtkins> | bLql0rUWFelspOgAlb2n0VcrsvqQso6LR-XLA6gPEwHXVqJTieitMguV0EPw0&attredirects=0 |
| 15:38 | <jcranmer> | UTF-8 is only ~38%? |
| 15:38 | <TabAtkins> | Apparently. The heartening part is the upward climb. |
| 15:39 | <hsivonen> | though there's still non-Unicode Chinese growth |
| 15:39 | <jcranmer> | how is it measured? |
| 15:39 | <TabAtkins> | jcranmer: Not sure. It came from i18n thread. |
| 15:39 | <hsivonen> | jcranmer: I'd expect it to be percentages from the Google index as detected by Google's detector |
| 15:40 | <jcranmer> | I expect that it may be based on what pages claim |
| 15:40 | <TabAtkins> | hsivonen: Yeah, a bit. I dunno quite why, though - utf8 is more efficient on most chinese pages. |
| 15:40 | <hsivonen> | TabAtkins: it's probably more about legacy software or legacy mindset than about efficiency |
| 15:40 | <TabAtkins> | True. |
| 15:43 | <annevk2> | and gov endorsed maybe, in case of gb23.. |
| 15:43 | <zcorpan_> | the number of pages using legacy encodings might still be growing (just that utf-8 is growing faster) |
| 15:44 | <jgraham> | Given that the number of webpages is probably growing exponentially and nothing seems to be declining fater than linearly that would appear to be the case, yes |
| 15:44 | <jgraham> | *faster\ |
| 15:45 | <hsivonen> | Is the govt still actively endorsing GB18030? |
| 16:13 | <annevk2> | hmm interesting, shouldn't data:... in data URLs be case-insensitive i.e. data:,x and datA:,x are the same? |
| 16:14 | <gsnedders|work> | Yeah |
| 16:14 | <annevk2> | in Firefox it seems case-sensitive |
| 16:15 | <gsnedders|work> | Scheme and host should be case-insensitive, every other part of the URI should be case-sensitive |
| 16:15 | <gsnedders|work> | (ASCII case-insensitive, that is) |
| 16:20 | <TabAtkins> | omg I'm bored with the frameset discussion. |
| 16:22 | <Dashiva> | I don't get the security aspect |
| 16:22 | <Dashiva> | Anyone can just right-click and choose 'open frame', then bookmark that |
| 16:22 | <TabAtkins> | I suspect the OP is completely irrational about that point. |
| 16:22 | <hendry> | has someone crafted a good HTML5 audio test? Something that tests looping, fallback to different formats and media that can perhaps test stereo? |
| 16:23 | <zcorpan_> | mmm stereo |
| 16:27 | <TabAtkins> | Dashiva: I hadn't thought of that, though, so I've gone ahead and posted it to the list. Maybe we can get him to shut up about breaking bookmarking. |
| 16:29 | <TabAtkins> | It's really not hard to do properly on the server-side, if you want to. Just pass an expiring secret around, and redirect to the front-page if the secret is out of date. |
| 16:30 | hendry | discovers http://en.wikipedia.org/wiki/Reference_tone |
| 17:16 | <annevk2> | CSSOM is a lot of work; sigh |
| 17:16 | <annevk2> | for every bit you write you can immediately think of like ten questions you have to answer so you can rewrite it |
| 17:17 | <annevk2> | exponential pain :p |
| 17:24 | <annevk2> | class task: define "supported style sheet language" so that CSS is included and XSLT is out without naming either |
| 17:25 | <TabAtkins> | "not xml-based" |
| 17:25 | <annevk2> | fallback: forget about architecture and just deal with it explicitly |
| 17:25 | <annevk2> | I should add that it is about the fundamental difference |
| 17:26 | <annevk2> | CSS could be xml-based and still work pretty much the same way |
| 17:26 | <Dashiva> | "Using a non-distributed or non-extensible language" |
| 17:26 | <annevk2> | lol, also an F for you :p |
| 17:26 | <Dashiva> | Then you just define language to be CSS |
| 17:26 | <TabAtkins> | Since you've only got two stylesheet languages, and I don't *think* it's likely that another will emerge in the near future, I'd go with the fallback. Specific language can be easily amended in the future if more appropriate languages are created, but a seemingly-general classification (that's really just meant to differentiate two things) is harder to work around conceptually. |
| 17:27 | <TabAtkins> | ...if it turns out, upon the creation of the new language, that the classification used doesn't capture the correct details. |
| 17:43 | <TabAtkins> | Man, db abstractions are complicated. |
| 17:43 | TabAtkins | is going through legacy code to update to current abstractions, and should probably add another feature to eliminate even more direct SQL strings. |
| 17:44 | <othermaciej> | annevk2: would XSL-FO styling count, if anyone implemented it? |
| 17:44 | <othermaciej> | annevk2: XSLT isn't really a stylesheet language in the same sense as CSS, it's a transform language, so there's no runtime style information resulting |
| 17:45 | <othermaciej> | annevk2: anyway, what part of CSSOM needs to reference the concept of "supported style sheet language"? |
| 18:11 | <annevk42> | othermaciej, XSL-FO would count I suppose |
| 18:11 | <annevk42> | othermaciej, it's the part on style sheets and the DOM; in particular whether or not a style sheet is to be created for certain Link headers and ProcessingInstruction nodes |
| 18:12 | <annevk42> | HTML and SVG presumably need to reference such a concept as well for <style>, <link>, etc. so maybe I should look there |
| 18:12 | <othermaciej> | annevk42: I see |
| 18:12 | <othermaciej> | annevk42: I think you should just limit it to CSS |
| 18:12 | <othermaciej> | there's no actually existing stylesheet language where it would be useful to expose a DOM stylesheet object, other than CSS |
| 18:13 | <annevk42> | Hixie just uses "specified styling language" without a link to a definition |
| 18:13 | <othermaciej> | or you could say MUST include CSS stylesheets, MAY include other recognized stylesheet languages that produce runtime style information rather than modifying the content |
| 18:14 | <annevk42> | ta |
| 18:14 | <TabAtkins> | annevk42: Actually, it looks like Hixie offloads the definition to CSSOM. ^_^ |
| 18:14 | <annevk42> | guess I'll fiddle with it a bit tomorrow |
| 18:15 | <annevk42> | TabAtkins, not that one |
| 18:15 | <TabAtkins> | Not 4.2.8? |
| 18:15 | <annevk42> | yes, but that term is not said to be defined in the CSSOM |
| 18:16 | <TabAtkins> | Okay, true, but it does defer the idea of what styling information is to CSSOM. |
| 18:16 | <annevk42> | just some of it |
| 18:17 | <annevk42> | seems hixie also defines some other concepts around style sheets that do assume CSS |
| 18:17 | <TabAtkins> | Those may be spec bugs. |
| 18:18 | annevk42 | wonders whether SVG elements can have style sheets blocking scripts |
| 18:18 | <TabAtkins> | HTML5 should merely assume that CSS is a commonly supported styling language, not that it is the "correct" language. |
| 18:19 | <annevk42> | it's inside an e.g. actually |
| 18:19 | <TabAtkins> | Oh, right. Those are just examples, then. |
| 18:20 | <TabAtkins> | Or rather, that one line (since there's only one) is. |
| 20:13 | <remysharp> | I've a question about the example code in the spec |
| 20:13 | <remysharp> | http://www.whatwg.org/specs/web-apps/current-work/multipage/sections.html#the-section-element |
| 20:14 | <remysharp> | the code directly above that point - the online - offline example |
| 20:14 | <remysharp> | I'm sure there's lots more like it, but why is it using inline JavaScript rather than DOM scripting? |
| 21:45 | <annevk42> | othermaciej, CaretPosition wfm too |
| 21:45 | <annevk42> | othermaciej, I think only WebKit implemented it and only recently |
| 21:45 | <annevk42> | othermaciej, so that should certainly be possible |
| 21:46 | <othermaciej> | annevk42: yeah it's a recent addition |
| 21:47 | <annevk42> | othermaciej, so containingNode is one of Element or Text? |
| 21:48 | <annevk42> | prolly don't need offsetKind |
| 21:49 | <othermaciej> | annevk42: well, you could identify form controls by their tag name |
| 21:50 | <othermaciej> | annevk42: but if the pointer is between two adjacent images for instance, you could get an Element that's not a form control (not sure what actually happened in our implementation) |
| 21:50 | <annevk42> | and maybe change containingNode to ownerNode (that is used in e.g. StyleSheet anyway) |
| 21:51 | <othermaciej> | I'm not sure ownerNode is the right terminology for the node containing a position - it doesn't "own" the position, that's just where the position was found |
| 21:51 | <annevk42> | hmm true |
| 21:52 | <annevk42> | just container would be somewhat compatible with DOM Range |
| 21:54 | <annevk42> | anyway, this can certainly work |
| 21:54 | <annevk42> | lets see what the Google guy thinks |
