| 00:21 | <hober> | Hixie: works in chrome |
| 00:22 | <Hixie> | what do you see exactly? |
| 00:22 | <hober> | when I click on the cats I land on: You have reached this web page by typing "example.com", "example.net", or "example.org" into your web browser.\n\nThese domain names are reserved for use in documentation and are not available for registration. See RFC 2606, Section 3. |
| 00:23 | <Hixie> | sure but is the <canvas> also showing the image map? |
| 00:23 | <Hixie> | i know the <img> supports it |
| 00:23 | <hober> | oh, sorry, didn't notice the <img> |
| 00:24 | <hober> | the <canvas> is the blank area between the ellipses? |
| 00:24 | <Hixie> | yeah |
| 00:24 | <hober> | nothing happens when I click that. |
| 00:24 | <hober> | so yeah, I don't know what chaals was on about there |
| 00:28 | <Hixie> | ok |
| 00:28 | Hixie | goes back to his MUD for now |
| 00:28 | <sicking> | man, i can't believe no-one took the w3c bugzilla bug 10000 before I did |
| 00:29 | <sicking> | it was sitting there wide open for a long time |
| 00:29 | <TabAtkins> | I can't believe that Julian got bug #9001. >_< |
| 00:30 | sicking | doesn't see the significance of #9001 |
| 00:30 | <TabAtkins> | IT'S OVER NINE THOUSAAAAAAAAAAAAAAAAAAAAND |
| 00:31 | <sicking> | riiiiiiight |
| 00:31 | sicking | feels like he is still lacking context |
| 00:31 | <Hixie> | y'all have issues |
| 00:31 | <TabAtkins> | http://www.youtube.com/watch?v=SiMHTK15Pik |
| 00:31 | <TabAtkins> | ^^^ context |
| 00:31 | <TabAtkins> | Well, hm. No, that's not it. That' sjust the source. One sec. |
| 00:32 | <sicking> | urban dictionary is my friend |
| 00:32 | <sicking> | also google |
| 00:32 | <TabAtkins> | That works too. |
| 00:32 | <sicking> | fortunately i've been successful in avoiding Dragon Ball Z |
| 00:33 | <TabAtkins> | over9k doesn't really depend on dbz. It's... a thing all its own. |
| 00:35 | sicking | also notes that 10000 also is over 9000 |
| 00:36 | <sicking> | i know, i know |
| 00:58 | <MikeSmith> | sicking: did anybody tell you yet about the prize for filing bug but 10000? |
| 01:00 | <sicking> | MikeSmith: no?! What do I win? A brand new car? $10000000 from dead Nigerian prince? A years supply of viagra? |
| 01:01 | <MikeSmith> | you won editorship of the WebIDL spec |
| 01:01 | <MikeSmith> | so, congratulations! |
| 01:01 | MikeSmith | pats sicking on the back |
| 01:01 | <sicking> | hahaha |
| 01:01 | sicking | runs |
| 01:01 | <MikeSmith> | heh |
| 01:02 | <MikeSmith> | maybe we should rename that spec Hot Potato |
| 01:03 | <sicking> | Or FileAPI2 |
| 01:05 | <MikeSmith> | or XBL3... |
| 02:17 | <boblet> | re: CSS I was told UAs can’t drop vendor prefixes until the relevant spec goes CR. Given border-radius on Opera even though CSS3 Backgrounds is still Working Draft, that’s wrong, no? |
| 02:18 | <TabAtkins> | Backgrounds went CR for a short time. |
| 02:18 | <TabAtkins> | And then we pulled it back to make some useful changes. |
| 02:19 | <TabAtkins> | It should be back to Last Call any time now, too. |
| 02:19 | <boblet> | TabAtkins: aaah, that explains it. so regarding adding non-prefixed versions to ‘future-proof’ (like for transitions), is that bad? (given the spec could change etc) |
| 02:22 | <TabAtkins> | Yes and no. It's good, and it's bad, and it's good and bad to do the opposite. |
| 02:22 | <boblet> | lol |
| 02:22 | <boblet> | thanks for your definitive answer :) |
| 02:22 | <TabAtkins> | If you put in unprefixed versions, figuring that they won't change, you create some pressure to not change it, since you have deployed content that won't be compatible. |
| 02:22 | <boblet> | (it certainly covered all the bases) |
| 02:23 | <TabAtkins> | But if you don't, then you make it more painful to drop prefixed versions, since you might not update your site to use the unprefixed one. |
| 02:23 | <TabAtkins> | So, shrug? |
| 02:23 | <TabAtkins> | I use unprefixed, but only if I'm using the property in a simple enough manner that I'm confident it won't change. |
| 02:24 | <boblet> | yeah, I personally work on the expectation that most specs won’t change significantly, and if there is a major change it’ll be something I notice and then go fix |
| 02:25 | <erlehmann> | i use prefixes. after all, that is what they are for. |
| 02:25 | <boblet> | there’s no site (that I know of) tracking what browsers are or have been using prefixes, so it’s hard to keep on top of adding unprefixed versions in |
| 02:25 | <erlehmann> | boblet, progressive degradation or how it is called ;) |
| 02:25 | <erlehmann> | make your site work without fancy backgrounds, problem solved |
| 02:26 | <boblet> | erlehmann: graceful degredation ;-) |
| 02:26 | <boblet> | but the unprefixed versions could be seen as progressive enhancement (for the glorious future of unprefixed support) |
| 02:27 | <MikeSmithX> | erlehmann: what if the purpose of your site is two demonstrate fancy backgrounds? |
| 02:28 | <erlehmann> | i would always settle for one fancy background, comrade. |
| 02:44 | <MikeSmith> | you sound like a possible appreciator of the "wabi sabi" aesthetic |
| 02:45 | <MikeSmith> | which, btw and unfortunately, most Japanese websites demonstrate completely the opposite of the wabi-sabi aesthetic |
| 02:45 | <MikeSmith> | with some exceptions |
| 02:45 | <MikeSmith> | like the sites that boblet creates in Japan |
| 02:45 | <MikeSmith> | which are much more wabi and sabi |
| 02:46 | <erlehmann> | wat |
| 02:46 | <boblet> | MikeSmith: i like to think of them as more wasabi |
| 02:47 | <MikeSmith> | boblet: please make em more like yuzu koshou |
| 02:48 | <boblet> | MikeSmith: ooh man! I’m so addicted to that stuff atm. although it prolly should be called yuzu shio huh |
| 02:49 | <MikeSmith> | dunno |
| 02:50 | <MikeSmith> | it's got a lot of pepper in there |
| 02:50 | <MikeSmith> | of some kind |
| 02:50 | <boblet> | the one I’ve got is more salty than peppery |
| 02:50 | <boblet> | regardless it’s awesome |
| 02:55 | MikeSmith | found http://en.wikipedia.org/wiki/Yuzukoshō |
| 02:56 | <MikeSmith> | http://ja.wikipedia.org/wiki/柚子胡椒 |
| 02:57 | <MikeSmith> | pathetic -- dinnet know kanji for 胡椒 til now |
| 02:59 | <MikeSmith> | Hixie: please add a Previous Versions link to source of http://dev.w3.org/html5/spec/ |
| 02:59 | <MikeSmith> | for http://www.w3.org/TR/2010/WD-html5-20100304/ |
| 03:12 | boblet | wants a yuzukosho kitkat! |
| 03:32 | <variable> | I'm looking for more stuff to add to the rationale document - if anyone has any ideas please let me know |
| 03:32 | <variable> | I'm following the mailing list but there isn't that much to go on atm |
| 03:43 | <MikeSmith> | variable: going back through the commit-watchers list and checking commit descriptions might help some |
| 03:43 | <MikeSmith> | http://lists.whatwg.org/pipermail/commit-watchers-whatwg.org/ |
| 03:44 | <MikeSmith> | ...or it might also not help so much |
| 03:45 | <MikeSmith> | most of the commit descriptions don't provide much rationale |
| 03:45 | <variable> | MikeSmith, I'm looking for recent contecious decisions that have some kind of trail on the mailing list |
| 03:45 | <MikeSmith> | just recent ones? |
| 03:46 | <variable> | *contentious |
| 03:46 | <variable> | MikeSmith, older ones are OK as well |
| 03:46 | <variable> | anything that is contentious and needs to be documented ;) |
| 03:46 | <MikeSmith> | for some recent ones, reviewing W3C bugzilla might help |
| 03:47 | <MikeSmith> | http://www.w3.org/Bugs/Public/buglist.cgi?cmdtype=runnamed&namedcmd=HTML%20WG%20issues |
| 03:47 | <MikeSmith> | and for older ones, going back through the whatwg archives and checking Hixie messages |
| 03:48 | <MikeSmith> | to find ones where he indicates that he's made a change to the spec |
| 03:48 | <variable> | MikeSmith, anything specific ? |
| 03:48 | <MikeSmith> | variable: nothing specific |
| 03:48 | <MikeSmith> | http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/ |
| 03:48 | <MikeSmith> | you may find there is some pattern to the messages from Hixie |
| 03:49 | <MikeSmith> | something in the subject line, I dunno |
| 03:49 | <variable> | MikeSmith, I've been following the mailing list - looking for questions where people ask or answer why certain things are the case. But I can't go over everything - hence fishing for more things to add |
| 03:49 | <MikeSmith> | going back through Mark Pilgrim's This Week in HTML5 postings might actually be the best way |
| 03:50 | <variable> | hrm ok |
| 03:50 | <MikeSmith> | and his "What's Next" ones |
| 03:50 | <MikeSmith> | http://blog.whatwg.org/category/whats-next |
| 03:50 | <MikeSmith> | http://blog.whatwg.org/category/weekly-review |
| 03:51 | <MikeSmith> | Mark already did a lot of archaeology on some changes |
| 03:51 | <variable> | ok - I'll read throught them soon |
| 03:51 | <MikeSmith> | and documented them in some of those postings |
| 03:51 | <variable> | I'm trying to update this page: http://wiki.whatwg.org/wiki/index.php?title=Rationale |
| 03:51 | <MikeSmith> | I don't know how much of that might already be reflected in the Rationale page |
| 03:51 | <MikeSmith> | variable: yeah, I know which page you mean |
| 03:52 | <variable> | " I need a ***legitimate*** login and password to continue. " - as if you wanted an invalid username as password ;) |
| 03:53 | <variable> | (from the W3C bugzilla) |
| 03:54 | <variable> | MikeSmith, I added those to my bookmarks - I'll get to them soon - thanks |
| 03:58 | <MikeSmith> | cheers |
| 07:05 | <Hixie> | Lachy: yt? |
| 08:07 | <GPHemsley> | What do we think of Cufón? http://cufon.shoqolate.com/ |
| 08:08 | <Dashiva^2> | krijnh: the logs page seems to be down, but I'm sure you know |
| 08:19 | <krijnh> | Dashiva^2: yeah, my IP changed again |
| 08:51 | <roc> | Philip`: have you run your canvas tests against IE9 dev preview yet? |
| 08:51 | <roc> | or maybe you're the wrong Philip |
| 08:52 | zcorpan_ | has and has emailed the results to Philip` |
| 08:54 | <othermaciej> | how does it do? |
| 08:56 | <hsivonen> | are all the tweets saying things about the performance of IE9 still violating the license? |
| 09:01 | <zcorpan_> | othermaciej: i don't know, waiting for Philip` to put it in the comparison table |
| 09:10 | <jgraham> | Must not scream |
| 09:10 | <jgraham> | Must be calm in the face of ASCII Reference spam |
| 09:10 | <annevk> | that plugin email is pretty daunting too |
| 09:11 | <zcorpan_> | html4 and xml allows > in attribute values. they even allow ]]> in attribute values. |
| 09:11 | <annevk> | Hixie explains a teapot is being rendered and Julian suggests he did not want enhanced there, lol |
| 09:13 | <jgraham> | In the spirit of actually doing useful things, getting Philip`'s canvas tests in the testsuite seems high priority |
| 09:13 | <roc> | yeah |
| 09:13 | <annevk> | when are we getting a useful test framework? |
| 09:14 | <jgraham> | annevk: We are getting a useful test framework? |
| 09:14 | <annevk> | I'd be interested in making tests for a number of things if something is in place |
| 09:14 | <annevk> | jgraham, yeah, so maybe not |
| 09:14 | <jgraham> | I mean I wrote something for general DOM tests |
| 09:14 | <jgraham> | it needs more fleshing out but the basic structure is there |
| 09:14 | <jgraham> | I don't plan to port existing testsuites to use it though |
| 09:16 | <annevk> | writing a new dom2html test suite should not be too hard |
| 09:16 | <jgraham> | annevk: http://dvcs.w3.org/hg/html/file/94563fe91cb0/tests/submission/Microsoft/common/domtestcase.js is the test harness |
| 09:16 | <annevk> | though admittedly HTML5 has more complex (and more useful) interfaces |
| 09:17 | <jgraham> | annevk: I plan to add assertions that map to WebIDL keywords |
| 09:18 | <annevk> | why does the W3C have its own BSD license? |
| 09:18 | <annevk> | is that normal? |
| 09:19 | <roc> | for those of you wondering about Ben Lerner (Microsoft) posting on the WHATWG list ... |
| 09:19 | <roc> | he's an intern |
| 09:19 | <roc> | working on MSR's managed-code Web browser |
| 09:19 | roc | talked to him quite a bit at PLDI a couple of weeks ago |
| 09:40 | <a-ja> | annevk: ping |
| 09:40 | <annevk> | yo |
| 09:41 | <a-ja> | annevk: you support http://html5.org/tools/web-apps-tracker ? |
| 09:41 | <annevk> | in a way |
| 09:41 | <a-ja> | something appears to be busted with last few changes |
| 09:41 | <a-ja> | since 5105 |
| 09:41 | <annevk> | yeah, I wonder what is up |
| 09:42 | <annevk> | maybe I should run svn manually to see what comes out |
| 09:46 | <annevk> | I wonder what is different |
| 09:48 | <annevk> | nothing at all it seems :/ |
| 09:48 | <annevk> | oh, maybe it's the caching |
| 09:49 | <zcorpan_> | if nothing is different, it seems appropriate for the diff to say 'no result' :) |
| 09:51 | <annevk> | pretty sure it's the cache now |
| 09:51 | <annevk> | I'll flush it and see what happens |
| 09:51 | <annevk> | though I suppose deleting 30GiB might take a while |
| 09:51 | <zcorpan_> | is the 'next' button screwing things up? |
| 09:52 | <annevk> | no |
| 09:52 | <annevk> | well, I don't think so |
| 09:52 | <a-ja> | you think? |
| 09:52 | <zcorpan_> | i figured maybe someone clicked 'next' and it gave 'no result' and then 'no result' was cached |
| 09:52 | <annevk> | hmm |
| 09:53 | <annevk> | I don't think so |
| 09:53 | <annevk> | otherwise how would you explain http://html5.org/tools/web-apps-tracker?from=5093&to=5095&context=10 |
| 09:54 | <zcorpan_> | ok |
| 09:55 | <annevk> | this is gonna take ages |
| 09:56 | <annevk> | oh great |
| 09:56 | <annevk> | the cache was what made some of it still work |
| 09:56 | <annevk> | lol |
| 09:56 | <annevk> | god dammit |
| 09:57 | a-ja | hides |
| 10:04 | <a-ja> | hm....working now, it seems |
| 10:05 | <annevk> | it seems that I will have to remove the context option as the external diff program caused the whole thing to break down |
| 10:05 | <annevk> | sorry hsivonen |
| 10:05 | <annevk> | though when I looked through the cache it seems nobody was really using that anyway |
| 10:09 | <a-ja> | for those keeping score...ff trunk now has support for figure/figcaption and mark |
| 10:10 | <annevk> | you can keep score here: http://wiki.whatwg.org/wiki/Implementations_in_Web_browsers |
| 10:11 | <a-ja> | haven't checked to see whether special width stuff for figures works right |
| 10:12 | <annevk> | thanks a-ja, all trackers appear to be fully functioning again |
| 10:12 | <annevk> | and we got rid of a feature in the process |
| 10:12 | <annevk> | always a good thing |
| 10:12 | <a-ja> | yeah...looks good from here,too |
| 10:21 | <hsivonen> | Nice IE9 PP3 improves the DOM namespace compliance with HTML and SVG nodes |
| 10:23 | hsivonen | notices "The new rgba() color notation works in quirks, 7, and 8 document modes. In future versions of IE9, this syntax will only apply in the browser's most standard mode." in the relnotes |
| 10:24 | <zcorpan_> | i wonder if ie10 will have yet another mode |
| 10:25 | <annevk> | when quirks mode gets standardized the house of cards will collapse |
| 10:25 | <zcorpan_> | annevk: why? they can just choose to ignore the quirks mode spec |
| 10:26 | <annevk> | from the IE blog |
| 10:26 | <annevk> | "We encourage the other browser vendors to help the W3C finish the HTML5 specification by providing additional tests to the official HTML5 testing task force." |
| 10:26 | <annevk> | zcorpan_, it seems they haven't been able to ignore most other specs |
| 10:27 | <jgraham> | Yeah, we need some more work on the testsuite |
| 10:27 | <zcorpan_> | i wonder if the tests they submit is all tests they have |
| 10:27 | <jgraham> | I have not been impressed with the quality or number of the tests submitted so far |
| 10:28 | <jgraham> | zcorpan_: I doubt it |
| 10:28 | <jgraham> | I assume they vet tests that they submit to make sure they don't fail too many |
| 10:28 | <annevk> | so on the one hand the quality of the tests Microsoft has submitted are quite poor but on the other hand they have actually submitted tests |
| 10:28 | <hsivonen> | it's a pretty big shift in the Web that now the latest preview releases of all top 5 browsers can play YouTube videos without Flash |
| 10:28 | <jgraham> | Yeah |
| 10:29 | <jgraham> | annevk: ^ |
| 10:29 | <jgraham> | annevk: Don't really know how to encourage others to submit though |
| 10:29 | <zcorpan_> | jgraham: lead by example? |
| 10:30 | <jgraham> | zcorpan_: Yeah, Opera need to do better too |
| 10:30 | <annevk> | jgraham, we could submit the html5lib tests and surrounding framework I suppose |
| 10:30 | <annevk> | jgraham, is the i |
| 10:30 | <annevk> | licensing compatible? |
| 10:31 | <jgraham> | annevk: I think so |
| 10:31 | <annevk> | the thing is, each time we add new tests we would then have to update a number of repositories :/ |
| 10:31 | <jgraham> | I am happy for the W3C copy to be the official home for html5lib tests |
| 10:31 | <jgraham> | Assuming that they don't abuse that position |
| 10:32 | <MikeSmith> | roc: thanks -- I had been wondering what he might be working on |
| 10:32 | <MikeSmith> | hey krijnh is back |
| 10:32 | <jgraham> | annevk: I have a better HTML harness for the html5lib tests |
| 10:33 | <jgraham> | also |
| 10:33 | <annevk> | jgraham, that moves it from MIT to BSD |
| 10:33 | <drclue> | OK , I've contributed to the list and such and the only responses I've received is that it is a good idea, and in other venues the only opposition has been that since we have not done it we should not. |
| 10:33 | <annevk> | jgraham, so new tests only contributed to the W3C cannot be moved back? |
| 10:34 | <hsivonen> | https://twitter.com/mhausenblas/status/16912067471 |
| 10:34 | <jgraham> | Hmm, it depends if you think the BSD license is a problem or not I guess |
| 10:34 | <mhausenblas> | yes, hsivonen? :) |
| 10:35 | <drclue> | We should IMHO integrate XSL with HTML5 with the element level attributes XML/XSL and perhaps include an XML attribute |
| 10:35 | <jgraham> | In the short term the next thing that should be submitted is the canvas testsuite |
| 10:36 | <drclue> | Include an XML element I meant |
| 10:37 | <jgraham> | On consideration, I think it would be easiest if Philip` did that, assuming he is willing |
| 10:38 | <drclue> | Be it twitter,fox news , or many other venues , XML is totally out there, Having an XML and XSL attributes simply makes sense , and an XML attribute a natural extension |
| 10:39 | <drclue> | I could certainly fake these things in javascript , but given the volume of traffic , why make what should be in the HTML5 dialect a vendor specific javascript happening? |
| 10:42 | <drclue> | I know that some consider CSS , or other corners of web development not worth their time, but XML/XSL is totally worth the effort |
| 10:43 | <drclue> | Anything in HTML5 that integrates other w3.org specs is worth at least vigorous debate |
| 10:45 | <drclue> | Be it the weather, stocks, ebay,twitter, the daily news , or an appointment calender or hundreds of other situations , ignoring the thought of XML/XSL attributes and possibly an XML element is simply foolish |
| 10:46 | <annevk> | you may have missed it, but we're not big believers in this XML thing |
| 10:46 | <annevk> | it's sort of the reason we started doing HTML5 |
| 10:46 | <drclue> | I know , not believing in XML (HTML4, HTML5 etc) might be an issue |
| 10:47 | <annevk> | having said that, there's XML and XSLT APIs available, so if people want it they can easily use them together |
| 10:48 | <annevk> | compared to DOM manipulation though their usage is negligible |
| 10:49 | <drclue> | Your so full of excuses my friend. There has never been a truly valid recognition of any standard that was not already strongly in ones wallet. I'm championing something that might not be your thought of today but actually respects the standards that would be in your wallet tomarrow. |
| 10:50 | <jgraham> | YOu know I have no idea what you just said |
| 10:51 | <annevk> | it's not about recognition, it's about adoption |
| 10:51 | <drclue> | Why should I ignore a standard that saves copious bandwidth, increases productivity and profit, just because somebody might not at first glance understand it |
| 10:51 | <annevk> | if something is clearly in use patterns emerge and simplifications can be made |
| 10:52 | <annevk> | e.g. lots of sites have a date picker, so we can assume <input type=date> will be useful to them |
| 10:52 | <drclue> | Slavery of black people was a standard , but did not reflect the future , but I guess I should use your claim to justify same |
| 10:52 | <annevk> | jaja and hitler was against smoking, so being against smoking is evil |
| 10:53 | <annevk> | geez |
| 10:53 | <drclue> | input type=date seems fine to me as the input is asked often enough and could benefit from standardization |
| 10:54 | <drclue> | Smoking is a personal decision and not something related to standards , but given such arguments I can see the value in your position (NOT) |
| 10:55 | a-ja | was wondering when Godwin would appear :) |
| 10:55 | <drclue> | Give me your argument against XML/XSL options! |
| 10:57 | <drclue> | I can start the for arguments with the many many many services offering XML outputs and that XSL options allow all them XML outputs to be integrated into any presentation |
| 10:58 | <drclue> | Your against arguments start with ......? |
| 10:58 | <hsivonen> | krijnh: logs aren't responding. maybe DNS out of sync again? |
| 11:05 | <Dashiva^2> | hsivonen: Yeah, he said so earlier |
| 11:05 | <drclue> | As I suspected, either you have no valid argument against the proposal or your argument requires so much encapsulation as to both take to much time to respond and or not even be valid |
| 11:05 | <hsivonen> | Dashiva^2: OK |
| 11:07 | <drclue> | XML is a very common method of interaction and XSL a very free form method of presentation and really should be as a matter of w3.org interoperability be given a fair hearing |
| 11:07 | <drclue> | XSL is but the flip side of CSS |
| 11:07 | <Dashiva^2> | They have been given a fair hearing, look at actual usage |
| 11:07 | <Philip`> | zcorpan_: Ping |
| 11:07 | <zcorpan_> | Philip`: pong |
| 11:08 | <jgraham> | drclue: The valid argument is "very few people use XSLT. Implementors would likely not implement more XSLT features. The functionality you want can be trivially written in javascript. Ergo your proposal is not worth standardising until several of these things change" |
| 11:08 | <Philip`> | zcorpan_: Could you send the IE9 results as an attachment or something? I don't know how to easily fix the linewrapping in the email |
| 11:09 | <drclue> | The actual usage of XSL has not been held up by it's value but rather it's implementation and standard. I could most easily site the same excuses for / against the deployment of the video element |
| 11:10 | <drclue> | Can you cite an excuse against the XSL that has any more validity than one might have against the video element? |
| 11:11 | <Dashiva^2> | People code non-standard and non-cross-browser-compatible features without restraint if they actually want them |
| 11:11 | <Dashiva^2> | Video, you may have noticed, was rather popular before <video> was added |
| 11:12 | <drclue> | Again , can you cite a reason against the support of the w3.org standard XSL that would not apply to the video element? |
| 11:13 | <jgraham> | drclue: (a clue: the words "w3.org standard" don't magically win arguments here) |
| 11:13 | <hsivonen> | drclue: lolcats with XSL is harder than with video |
| 11:13 | <hsivonen> | drclue: most W3C standards are not implemented in browsers |
| 11:13 | <drclue> | So your argument would be like those that used bgcolor="" as opposed to those that used the CSS equivalent |
| 11:14 | <drclue> | XSL is implemented in all major and most minor browsers |
| 11:14 | <Dashiva^2> | So there's no problem |
| 11:14 | <Dashiva^2> | The feature already exists |
| 11:14 | <drclue> | The exploitation of the feature is not implemented , but the proposed standard would make it more mainstream |
| 11:15 | <annevk> | if it was desired it would be mainstream, through workarounds |
| 11:15 | <annevk> | see video |
| 11:15 | <annevk> | see <input type=date> |
| 11:15 | <hsivonen> | drclue: if the usage of the current JS APIs for XSLT become ridiculously mainstream, then, maybe, there's a reason to provide a declatative way |
| 11:16 | <annevk> | see what went before Web Sockets |
| 11:16 | <Philip`> | Someone ought to make an XSLT Compact Syntax |
| 11:17 | <drclue> | Server load would be greatly reduced , developer coding would be greatly reduced. To say that because one has not added something to the spec is a reason not to add it is so lame as to cause me to choke on the laughter |
| 11:17 | <zcorpan_> | drclue: http://wiki.whatwg.org/wiki/FAQ#Is_there_a_process_for_adding_new_features_to_a_specification.3F |
| 11:18 | <drclue> | Yippie skippy , good try , not playing here |
| 11:18 | <Dashiva^2> | I get the feeling drclue isn't actually reading what anyone else is writing |
| 11:18 | <drclue> | I read lots of things , but perhaps not always what you want me to read or in the case of reading same not always agreeing |
| 11:18 | <jgraham> | It's like having mookid back |
| 11:19 | <Dashiva^2> | I thought they were the same person? |
| 11:19 | <jgraham> | Oh well that's why then |
| 11:19 | <drclue> | So in lue of a position you write what you do. Obviously you have no validity to your position |
| 11:20 | <drclue> | One should argue their position , not play junior highschool kid games |
| 11:21 | <annevk> | we did plenty to explain our position |
| 11:21 | <drclue> | Plese cite your valid arguments |
| 11:21 | <annevk> | please read what we wrote |
| 11:21 | <MikeSmith> | I miss mookid |
| 11:22 | annevk | -> lunch |
| 11:22 | <drclue> | I've been reading what you have all written and it sounds like a bunch of kids at lunch |
| 11:22 | <MikeSmith> | drclue: I strongly recommend you look into content negotation |
| 11:22 | <zcorpan_> | lol |
| 11:22 | <AryehGregor> | annevk, when quirks mode gets standardized, I'm betting IE9 will support it in standards mode (ironically), but keep some or all of the old modes for IE-specific code paths that still aren't updated. |
| 11:23 | <Philip`> | zcorpan_: http://philip.html5.org/tests/canvas/suite/tests/results.html - thanks |
| 11:23 | <drclue> | So you think that citing "content negotiation" is supposed to scare me? |
| 11:23 | <Philip`> | roc: ^ |
| 11:23 | <roc> | thanks |
| 11:23 | <AryehGregor> | drclue, do you think repeated claims that we're all ignoring you and not making any arguments, rather than actually addressing the arguments, is going to persuade us? |
| 11:23 | Philip` | also fixed a few test bugs but didn't update the results so they don't quite match any more |
| 11:24 | <zcorpan_> | AryehGregor: how could they implement quirks mode in standards mode? that doesn't make any sense |
| 11:24 | <AryehGregor> | You can't make browser implementers implement what you want just by putting it in a spec. If you can't convince them of your argument, there's no point in proceeding. |
| 11:24 | <roc> | people on my blog keep annoying me by claiming Mozilla is evil for not implementing every W3C REC |
| 11:24 | <drclue> | I don't think that closed minds ignoring years of w3.org efforts are easily persuaded as those ignoring such efforts need to take responsibility for their actions which just may be impossible for them to do |
| 11:25 | hsivonen | points to http://spreadsheets.google.com/ccc?key=0AtnDcoh7FXfAdEFuWVlVVDZTVkFWdnRqaWFGMzNYM3c&hl=en |
| 11:25 | <AryehGregor> | zcorpan_, I mean that someone will spec the behavior other browsers use for choosing quirks mode, and will spec quirks mode itself, but IE will still let you explicitly opt out of that and rely on blacklists and other heuristics. |
| 11:26 | <AryehGregor> | drclue, then I guess you're out of luck. Try making your own web browser, I guess. |
| 11:26 | <AryehGregor> | Or get a law passed. |
| 11:26 | <AryehGregor> | Not much else for you to do about it. |
| 11:26 | <drclue> | @AryehGregor: There indeed is not much to actually implement as all the code is already in the browsers , even old IE |
| 11:27 | <jgraham> | roc: You don't implement Web Services Choreography Description Language?! Clear evil. |
| 11:27 | <drclue> | The only fight here is should we leverage what is in all browsers to the benefit and advancement of the w3.org standards or simply talk shit for the sake of doing same |
| 11:27 | <roc> | We've actually been removing support for W3C RECs. that makes us extra evil |
| 11:28 | <AryehGregor> | Which ones? |
| 11:28 | <jgraham> | (I guess that one is actually CR) |
| 11:28 | <roc> | we had some support for SOAP, WSDL and XForms |
| 11:28 | <roc> | all gone |
| 11:28 | <drclue> | The spec I suggest simply leverages code already implemented in ALL browsers |
| 11:29 | <roc> | (I don't recall if they ever actually shipped in a browser) |
| 11:29 | <MikeSmith> | ^- |
| 11:29 | <MikeSmith> | oops |
| 11:29 | <MikeSmith> | fat fingered |
| 11:29 | <roc> | I think we've removed some XLink and XPointer stuff too but I don't remember the details |
| 11:30 | <roc> | not sure if they're RECs |
| 11:30 | <hsivonen> | I *think* SOAP shipped in Netscape circa 2002 or 2003 |
| 11:30 | <drclue> | Even IE6 (Gasp if I can even call that a browser) includes the support as does every piece of software we would actually call a browser |
| 11:31 | <drclue> | All I'm asking is to map a couple of att4ributes to the existing support in those browsers |
| 11:32 | <hsivonen> | drclue: you need to spec, implement and test all sorts of things starting with how this feature would integrate with the event loop and the progression of the parse of the document having those attributes |
| 11:32 | <drclue> | Of course there will always be those that would argue against light at sunrise, but asking to leverage what is already there is simply a no brainer |
| 11:33 | <drclue> | @hsivonen are you trying to make me laugh? |
| 11:33 | <drclue> | You are either trying to make me laugh or are trying to snow me |
| 11:34 | <drclue> | I've been at the keyboard probably longer than you have been alive |
| 11:34 | <hsivonen> | Chrome thinks software.hixie.ch is written in Irish Gaelic and offers to translate |
| 11:34 | <hsivonen> | source says lang="en" |
| 11:35 | <drclue> | I'm always bothered by spec that thinks in lang-en |
| 11:36 | <hsivonen> | and it offers to translate to Icelandic in a Finnish Chrome running on English Windows |
| 11:36 | <hsivonen> | FAIL |
| 11:36 | <drclue> | All machines , all languages and that is just a base line |
| 11:37 | <AryehGregor> | hsivonen, I found that feature annoyingly unreliable. I turned it off. |
| 11:37 | <drclue> | Obviously those developing something are apt to look at things through colored lenses |
| 11:37 | <AryehGregor> | It would offer to translate all-ASCII Apache directory listings with English file names (I forget from what). |
| 11:38 | <drclue> | ASCII is biased too as some languages could not even fit in such a constrained space |
| 11:39 | <drclue> | It's a hard biscuit to think in broad terms , but it is required to do so as otherwise ones opinions are stupid |
| 11:39 | <drclue> | BRB |
| 11:41 | <hsivonen> | curiously, Google decides what UI language to install for Chrome from the IP address of the downloader |
| 11:42 | <hsivonen> | as opposed to making the installer stub sniff for the UI language of the OS, for example |
| 11:42 | <AryehGregor> | You'd think that the latter would be more reliable . . . do they have data to the contrary? |
| 11:43 | <AryehGregor> | Maybe it's less reliable in some places, but more reliable in others. |
| 11:44 | <Philip`> | Maybe the application UI languages aren't a subset of the OS UI languages |
| 11:44 | <Philip`> | i.e. the OS doesn't provide enough information to select one of the application languages |
| 11:45 | <hsivonen> | Philip`: that's possible but Finnish is a in the set of possible UI languages for Windows, Mac OS X and the top Linux distros |
| 11:45 | <drclue> | There are a lot of things in spec and implementation that bother me , like CSS trying to do display:table etc , but in the end I would rather allow a spec to over reach than to under reach |
| 11:48 | <drclue> | CSS was always to me style while HTML was structure but it is inevitable that any particular spec is going to reach as far as it can, which as far as I've seen has had more benefits than detractions as much as such detractions might at time disturb me |
| 11:51 | <drclue> | My basic hope is despite these excesses that each spec will in the course of things try to advance other specs in a mutual sort of way so that the entire spec ecosystem can grow together |
| 11:52 | <roc> | admirable restraint, everyone |
| 11:54 | <drclue> | Of course I personally like all the pluses that XML/XSL has to offer an over strained system where the USA struggles to be even number 15 in the world. I try to advance the integration of this spec if for nothing else that it would allow number 15 in the world to function better than the current number one |
| 11:58 | <drclue> | @roc we were restrained? Oh ya , I did not deploy my fusion weapon , shit , I guess your right :) |
| 12:12 | <annevk> | "I've been at the keyboard probably longer than you have been alive" -- I'm waiting for claims of axiomatic proofs now. This stuff is great :-) |
| 12:35 | <drclue> | @annevk: You have already had plenty of axiomatic proofs that XSL is worthless |
| 12:45 | <Dashiva^2> | No, those proofs are mostly empirical |
| 12:45 | <drclue> | Well, each would seem about as worthy an argument against XSL |
| 12:48 | <drclue> | There are many instances where I don't care for` a particular aspect of a w3.org standard , but none where I would simply because it was not my favorite argue against it in the face of it's benefits |
| 12:50 | <annevk> | AryehGregor, that seems pretty hard to do, but maybe to some extent |
| 12:50 | <Dashiva^2> | Those benefits are irrelevant until people show they care about them enough to use them |
| 12:52 | <drclue> | Well, I for one care about them enough to use them, and in the course of learning them consulted with scores of others who had learned them, just as when I learned CSS I lerned from many others in the face of at the time many nay sayers |
| 12:53 | <drclue> | There are many religions that object to vaccination , but I would be happy to disagree with them too |
| 12:54 | <drclue> | One could get a room full of folks to tell me that blacks should but shine shoes and I would disagree with them as well |
| 12:55 | <drclue> | Now if you would actually like to discuss the merits of the propositions as opposed to taking a bar stool poll , I'm more than up for that |
| 12:58 | <Dashiva^2> | Scores of people is not very much on internet scale |
| 12:59 | <drclue> | Well , are we talking the might of stagnation or the right of advancement? |
| 13:00 | <Dashiva^2> | We're talking about priorities, doing work on features that people actually want |
| 13:01 | <drclue> | Did people really WANT display:table vs element TABLE? |
| 13:01 | <Philip`> | We're talking about spending time developing features that are likely to give the most benefit to the most people, based on evidence of how widespread are people demonstrating a need for that feature |
| 13:02 | <Philip`> | Oh, like what Dashiva^2 said |
| 13:03 | <annevk> | display:table is sunk cost and anyway this is not the CSS WG |
| 13:03 | <drclue> | Well, I'm talking about the minimal investment to exploit functionality already built into every browser, to the benefit of all be they server side developers , client side developer , those paying for bandwidth etc. |
| 13:03 | <Philip`> | drclue: Lots of people clearly wanted table-like layouts, and there were clear disadvantages to using <table> for layout (unmaintainable markup, can't adapt to widely varied display devices, etc), so there was a problem that needed to be solved |
| 13:04 | <Philip`> | (display:table may or may not be an adequate solution to that problem) |
| 13:04 | <Dashiva^2> | The only benefit seems to be convenience |
| 13:05 | <Dashiva^2> | The feature set remains unchanged, and they are features people clearly do not care much about |
| 13:06 | <drclue> | @Philip I'm not arguing against the CSS attributes , as obviously you think they have value, but there was not a huge public demand for that feature , but rather a thought be the spec folks that it had value. Here I'm just saying that what I suggest has extreme value despite a public outcry for it's implementation. Specifications are often an act of consderation , not mob rule |
| 13:07 | <drclue> | If we had put everything to a bar stool vote , blacks would still be picking cotton |
| 13:08 | <Philip`> | drclue: The intended approach for HTML5 has not been to look for demand expressed as public outcry, but to do research into what people do in practice |
| 13:09 | <drclue> | XSL saves server load , saves bandwidth , affords portability , scales between the ignorant and the savvy |
| 13:09 | <Dashiva^2> | drclue: You are the one who wants to force other people to follow your commands |
| 13:09 | <Dashiva^2> | Adjust your analogies accordingly |
| 13:09 | <Peter`> | .. or are likely to be practicing when the features become available |
| 13:09 | <Philip`> | which gives a more accurate indication of what problems people experience and are having to work around in more inefficient ways and are worth solving better |
| 13:10 | <drclue> | Well , maybe in practice one utilized selective questioning where they only aske the opinions of those on the bar stool next to them |
| 13:11 | <drclue> | I'm not forcing anyone to use XSL , but rather trying to make it available in a context beyond vendor specific javascript and http header mdness |
| 13:13 | <drclue> | XSL while supported in every browser, has been forced to contend with an implementation madness that has thwarted its value and for which I hope to provide a reasonable avenue of normalcy |
| 13:13 | <Dashiva^2> | Authors have coped with all kinds of madness. They would not be stopped if they actually wanted XSL. |
| 13:15 | <hsivonen> | annevk: IE9 PP3 bloated their Accept header. |
| 13:16 | <annevk> | I noticed that |
| 13:16 | <drclue> | I would not put an effort behind XSL if I did not believe it to have a very significant merit. XSL is an amazing specification both server side and client side , but alas the various implementations have made it almost impossible for the authors to deploy it. It was not simply a matter of author choice , but having to mess with apache add-type and a maze of mumbo jumbo to get browsers to properly deploy. |
| 13:17 | <annevk> | hsivonen, guess that might increase the amount of complaints a little, but I think we have a pretty strong case |
| 13:17 | <drclue> | @annevk , the only case you;ve made is to say you have one |
| 13:19 | <drclue> | I can say that XSL saves server load , saves bandwidth , reduces developer load and cite all day use cases for it that leverage existing sources all over the net |
| 13:19 | <drclue> | By simply adding an XSL and XSL attribute , I can immediately leverage many thousands of data sources |
| 13:20 | <annevk> | you can implement your proposal in a few lines of JavaScript as we have said several times so the arguments about load and saving bandwidth are bullshit |
| 13:20 | <drclue> | Fox,CNN,Twitter,Flicker, NOAA, and almost any other place of note |
| 13:21 | <drclue> | We could implement <video> in a few lines of javascript leveraging Flash , if all you have is that kind of argument |
| 13:22 | <annevk> | yes we could, but we do not want Flash |
| 13:22 | <annevk> | so we made <video> |
| 13:22 | <drclue> | Of course we would both stand up on each of our points citing the go do it in javascript thing as a bogus argument |
| 13:22 | <annevk> | and you still need JavaScript for <video> most of the time |
| 13:22 | <hsivonen> | IE blog says: "Note this video uses the HTML5 video tag (with the H.264 codec) if your browser supports it, and falls back to other methods otherwise. It’s a good example of same markup in action." I see "Install Microsoft Silverlight" but no video. |
| 13:22 | <Peter`> | hsivonen: in Firefox? |
| 13:23 | <drclue> | But of course you need far less javascript , making things more standardized and robust. Same holds for XSL |
| 13:23 | <annevk> | hsivonen, yeah, "same markup" is such nonsense, but I lost the will to argue it long ago |
| 13:23 | <hsivonen> | Peter`: in Minefield with Flash Player installed |
| 13:23 | <Peter`> | hsivonen: ah, then they're only supplying Silverlight fallbacks... makes sense (not) |
| 13:24 | <drclue> | I hate Flash too, but I really dislike arguments that hide behind such similar thinking as well |
| 13:24 | <annevk> | drclue, it's pretty clear people are willing to use JavaScript; if they want XSLT on the client in the way you suggest they can do that too, using some JavaScript; if they don't maybe they're not interested in XSLT |
| 13:24 | <Peter`> | drclue, you could start implementing the "jQuery for XSLT" in Javascript and become famous! |
| 13:24 | <drclue> | I could use javascript to inflict flash video too , so lets ditch the lame arguments |
| 13:25 | <drclue> | XSL has true legitimate value , can you at least admit that , or do you have no honesty in this discussion at all? |
| 13:26 | <annevk> | lol |
| 13:26 | <MikeSmith> | http://flashlab.com/html5/ is fun |
| 13:27 | <drclue> | LOL , I can too LOL at the idea that a quality w3.org specification can be given the short sheet by a wannbe specification that I also support |
| 13:29 | <drclue> | Why is it that one specification should disrespect another specification when the goal should be interoperability betyween all quality specs? |
| 13:29 | <Dashiva^2> | I think I'm going to stop feeding drclue now |
| 13:29 | <drclue> | HTML is after all like SVG , just more XML |
| 13:29 | <jgraham> | hsivonen: YT? |
| 13:30 | <jgraham> | I was thinking of doing something to fix the html5lib tokenizer tests with unpaired surrogates |
| 13:30 | <jgraham> | But I want to consult on what |
| 13:30 | <drclue> | XSL is a transformation syntax that allows XML to become any other thing |
| 13:30 | <gsnedders> | drclue: How can HTML after XML when it predates XML? |
| 13:31 | <jgraham> | hsivonen: Do those tests actually work with your test harness? They are a problem in the python case because the built-in json module (possibly erronously) believes that JSON forbids unpaired surrogates |
| 13:32 | <hsivonen> | jgraham: the surrogate tests work with the Java harness, IIRC, but not with the tokenizer |
| 13:32 | <drclue> | Because HTML while originating earlier was realized to be in the end XML which is the basis of a number of of things like SVG and XSL and many others. Sometimes it takes a bit to understand that something that came first is really a derivative of something more basic |
| 13:33 | <hsivonen> | jgraham: the Java tokenizer assumes its input is valid UTF-16 |
| 13:33 | <jgraham> | hsivonen: Ah. OK |
| 13:33 | <MikeSmith> | "HTML5 compared to Flash is like a hat compared to an umbrella. They can be used together, but a hat won't replace an umbrella." |
| 13:34 | <jgraham> | hsivonen: So what do you do with those tests? |
| 13:34 | <hsivonen> | jgraham: fail them |
| 13:34 | <drclue> | If we were to go strictly by timelines and such multicol would still be in HTML and not trying to live in CSS |
| 13:36 | <MikeSmith> | HTML *is* a hat - in fact, it's a whole bunch of hats: http://item.rakuten.co.jp/icefield/10008496/ and http://item.rakuten.co.jp/icefield/10001179/ and http://item.rakuten.co.jp/icefield/10001141/ and http://item.rakuten.co.jp/icefield/10001180/ |
| 13:36 | <drclue> | Although multicol in HTML was only ever a Netscape thing |
| 13:36 | <hsivonen> | MikeSmith: where's the hat quote from? |
| 13:37 | <annevk> | hsivonen, flashlab.com |
| 13:37 | <MikeSmith> | hsivonen: http://flashlab.com/html5/ |
| 13:37 | <annevk> | but Flash is a Marvel character, not an umbrella |
| 13:37 | <jgraham> | hsivonen: So if I were to move them into a seperate file and do something like double escape them (and add an extra flag to indicate that they were double escaped), would you be able to skip that file without too many problems? |
| 13:37 | <drclue> | Anyway , my point is that despite your favorite approach whatever that might be , having ready access to to XSL transforms , be they for XML to HTML or XML to SVG or whatever , ther eis a major value here and almost any majoe site you could think of provides data to drive it |
| 13:38 | <hsivonen> | MikeSmith: why is the last one Checkers instead of VR? |
| 13:38 | <hsivonen> | jgraham: moving them to a separate file would be great. I'd just skip the file |
| 13:39 | <drclue> | It matters not if it is twitter , fox , cnn , flicker, many infarstructure and database tools etc , to ignore XSL simply because it aint your favorite is not the best expression of wisdom |
| 13:39 | <MikeSmith> | hsivonen: because I didn't have time to build a VR demo.. my Flash skillz are limited to Checkers-level stuff |
| 13:39 | <jgraham> | abarth: I hope that works for you too |
| 13:42 | <hsivonen> | MikeSmith: are those Flash-based images supposed to be clickable to get actual demos? they aren't clickable for me |
| 13:44 | jgraham | doesn't have an umbrella, but has a coat with an integrated hat to replace it |
| 13:44 | <hsivonen> | oh. the Flash stuff doesn't work with click events in my browser. I guess flashlab.com needs a "support matrix" to use jd's terminology |
| 13:44 | <jgraham> | Dunno if that says more about me or the mataphor though |
| 13:44 | <MikeSmith> | hsivonen: dunno.. I didn't actually try the demos. I just like the words |
| 13:46 | <hsivonen> | oh. now I see why the links don't behave right |
| 13:46 | <hsivonen> | the whole page is Flash |
| 13:47 | <jgraham> | hsivonen: So you want me to move all the tests with invalid unicode characters? |
| 13:47 | <hsivonen> | jgraham: only the surrogate tests and the tests that test U+0000 next to CR |
| 13:47 | <jgraham> | hsivonen: OK |
| 13:48 | <drclue> | Flash needs but the nuclear explosion that is supposed to precede it |
| 13:49 | <MikeSmith> | hsivonen: did you open a bug for the "bug when emitting tokenizer start tags" bug? |
| 13:49 | <hsivonen> | MikeSmith: nope |
| 13:50 | <MikeSmith> | oh |
| 13:50 | <MikeSmith> | you waiting for dude to open it? |
| 13:50 | <MikeSmith> | or no bug is needed? |
| 13:50 | <MikeSmith> | or ? |
| 13:50 | <hsivonen> | MikeSmith: I figured he can file a bug and if not, the email probably makes it to Hixie's mailboxes |
| 13:50 | <MikeSmith> | ok |
| 13:55 | <drclue> | Really XSL is just the flip side of CSS if one thinks about it. |
| 13:57 | <jgraham> | hsivonen: It seems that the only test that is problematic for the json module is the one with input "\uD800" |
| 13:57 | <jgraham> | (suggests there should be more cases covered) |
| 13:57 | <jgraham> | Do you have a list of the other ones do you fail for this reason? |
| 14:00 | <drclue> | with XSL and but perhaps an onclick handler that called but one setAttribute one could have a news reader for every RSS feed on the planet |
| 14:01 | <drclue> | With XSL one could have but an inline onclick handler to show month , week , and day views of a calender |
| 14:02 | <hsivonen> | jgraham: Invalid Unicode character U+D800 and the test that tests U+0000 and carriage return next to each other |
| 14:03 | <drclue> | With XSL , forum programs could output perhaps 10% of what they have to do now and every forum could be totally unique even though driven by a simpler server side program |
| 14:04 | <jgraham> | hsivonen: Where is the latter? for some reason I can't find it |
| 14:04 | <drclue> | Every tabular listing from a database could share but one XSL file no matter the table involved |
| 14:05 | <jgraham> | ah, seen it |
| 14:06 | <drclue> | Widgets like the weather could be deployed with zero javascript or for that matter any scripting at all |
| 14:09 | <drclue> | Charts and graphs could be displayed against existing real time data , again without ever writing a line of code. That same data could be words,graphs or infographics and never require any scripts |
| 14:11 | <drclue> | changes to database browses often would require nothing but the change in SQL statement and no changes to code as of course with XSL there would be no code required |
| 14:14 | <MikeSmith> | the channel seems to have sensed some disturbance in the Force, resulting in krijnh site being busted |
| 14:14 | <MikeSmith> | I wonder if we can imagine the disturbance back into the cornfield |
| 14:14 | MikeSmith | concentrates |
| 14:14 | <MikeSmith> | or make it into a jack-in-the-box |
| 14:15 | <drclue> | More corn grows as a result of the concentration , but the field does not go away :) |
| 14:16 | <Dashiva^2> | MikeSmith: Maybe you can imagine the DNS updating to the new IP faster instead ;) |
| 14:17 | <MikeSmith> | ommm ommm ommm ommm ommm |
| 14:17 | <drclue> | I code a lot of javascript, and have done so since the 90's , but I do know that XSL is in many cases a much better alternative with the only issue having been the weird things one has to go through to access it in the various browsers these days. That is why I am fighting hard to have a clean way to use it implemented in HTML5 |
| 14:18 | <MikeSmith> | rama rama hare hare |
| 14:19 | <drclue> | Emm Rama, did not they do everything in threes? |
| 14:21 | <MikeSmith> | nam myo ho renge kyo |
| 14:22 | <drclue> | I had a date once that chanted like that. It was amazing , after four hours of heavy sweating they were still able to chant , although it sounded much more like a whimper at that point |
| 14:24 | <drclue> | Of course the next day the IRS revoked their 501.c3 |
| 14:27 | <drclue> | Yup, I would much rather that HTML5 provide a decent hook to XSL than have to continue a voodo of server side and client side code to make XSL work with those complications used as an excuse for not fixing XSL access in HTML5 |
| 14:30 | <drclue> | The only reason XSL sucks has in reality nothing to do with XSL itself, however all them things that currently make XSL hard to deploy is why XSL is not popular. If XSL were accessible in a consistent way people would not understand how they lived without it |
| 14:31 | <daedb> | ...and after four hours, you're still able to sit on irc and chant about the magical XSL unicorns. Amazing! |
| 14:32 | <drclue> | You bet your arse , I can chant about it, because XSL is amazing |
| 14:32 | <zcorpan_> | drclue: don't you have anything better to do today? |
| 14:32 | <Philip`> | I've used XSLT about once, and one of the most annoying things is that it didn't have a way to serialise XML :-( |
| 14:33 | <Philip`> | so I had to write (by which I mean copy-and-paste from a blog post from 2002) a partial serialiser in XSLT |
| 14:33 | <drclue> | Exactly what did you want to do that you could not figure out how to do in but "1" time using it? |
| 14:34 | <drclue> | Dis you also give up on CSS after "1" time using it? |
| 14:34 | <drclue> | How did you manage to learn HTML after giving up on it after using it but "1" time? |
| 14:35 | <Philip`> | I was generating HTML documentation of a file format based on a RelaxNG grammar, and wanted to include fragments of the grammar inside <pre> blocks in the output |
| 14:35 | <Philip`> | I didn't give up, I checked various documentation and mailing lists and it appeared that the only solution was to manually write an XML serialiser |
| 14:36 | <drclue> | I use XSL to ease the development of documentation, and it works great for me. I'm able to embed HTML , XML , whatever I like |
| 14:37 | <drclue> | The real problem has been the odd ways in which XSL was deployed in the browsers. It made it really hard for folks to sort out the documentation |
| 14:38 | <Philip`> | I like having a pattern-matching language that works cleanly over XML, but I'd like it more if there was extensive library support and if the syntax was much less verbose |
| 14:38 | <drclue> | One had to screw around with content-type headers that varied by browser and browser version and do so much screwing around before one could actually get to just doing XSL |
| 14:39 | <drclue> | The verbosity of the syntax was indeed annoying at first , but once I got the hang of it , the verbosity was no longer an issue |
| 14:40 | <Peter`> | You must realize that most web developers will have given up before getting the hang of it at all, because of that |
| 14:40 | <drclue> | Well, that would be the excuse to give up everything you know about web development |
| 14:41 | <drclue> | HTML sucks , CSS sucks , javascript sucks , XSL sucks , at l;east until you know them |
| 14:42 | <drclue> | How many times have you written the code to show a tabular list of something? |
| 14:43 | <MikeSmith> | the disturbance in the Force seems to be persisting |
| 14:43 | <drclue> | If you had to type the words a little longer but only had to do it once , might not it be worth it? |
| 14:43 | <MikeSmith> | I wonder if disturbances in the Force ever sleep, and if so, at what time |
| 14:43 | <Peter`> | Just don't switch over to the dark side, MikeSmith |
| 14:43 | <drclue> | @MikeSmith well , when are you going to sleep? |
| 14:44 | <MikeSmith> | Philip`: real men always write their own XML serializers |
| 14:44 | <Philip`> | MikeSmith: Real men are bozos |
| 14:45 | <drclue> | Both my C++ and PHP tools output dual XML and JSONized XML which on the client side always end up being run through XSL anyways |
| 14:45 | <MikeSmith> | o |
| 14:46 | <drclue> | I used to do it via other means , but I actually learned that XSL was more than worth the trip |
| 14:46 | <MikeSmith> | I sense that the disturbance in the Force may have tried to reach out to me a little bit ago, to try to draw me into further communion |
| 14:47 | <MikeSmith> | I feel a kind of "fascination with the abomination" |
| 14:47 | <drclue> | Mind you , at first I hated XSL , cussed and screamed and stomped my feet , but like every other language I've learned , I made it through to the other side and appreciated the worth of the effort |
| 14:49 | MikeSmith | wonders if anybody may perhaps have a clue what timezone the DITF might be in, and if it possibly might be close to bedtime |
| 14:49 | <drclue> | If XSL were made uniformly accessible I tend to believe folks would find a liking for it, as it is really a cool spec |
| 14:50 | <Peter`> | MikeSmith: Unfortunately in minus seven, so it might be a while |
| 14:51 | <MikeSmith> | Peter`: ah |
| 14:51 | <MikeSmith> | understood, thanks |
| 14:52 | <MikeSmith> | so I think I might deal with it by medicating and coming back again after a while |
| 14:52 | MikeSmith | looks around for his prescription from Dr. Kush |
| 14:53 | <drclue> | Writes MikeSmith a prescription for medical marijuana |
| 14:58 | <drclue> | Hers a really simple XSL I used for documenting some XML commands to an Asterisk PBX middle ware I'm writing using HTML5 and WebSockets http://asterclick.drclue.net/Assets/XML/API.xml |
| 15:23 | <drclue> | It is interesting that each time I lobby for the support of XSL which aims to support the good programming practice of having a separation in data ,code and presentation , that real smart folks come out of the woodwork to say I should do that presentation work in javascript. Who would have thunk it. |
| 15:24 | <drclue> | If to lobby for the support of w3.org standards is a DITF , then listen to me breath in your helmet |
| 15:34 | <jgraham> | "In practice, no one puts ">" in attribute values." -- that sounds like a foolishly testable assertion |
| 15:35 | <Philip`> | I'm not sure what his point was about > being invalid in rel="...", because surely > would be equally invalid (since it's not part of the expected datatype) |
| 15:35 | <jgraham> | I don't think he is making any sense in general |
| 15:36 | <jgraham> | It's not like we could change parsers to make foo="abc>" break even if it was invalid |
| 15:36 | <Philip`> | I don't think he's not making sense, I think he's just wrong |
| 15:37 | <Philip`> | Probably he's assuming that parsers only have to parse valid documents |
| 15:37 | <jgraham> | Well I guess the need to parse actual documents is such a fundamental part of my world view that people who don't share it slip from "wrong" to "not making sense" |
| 15:39 | <drclue> | Hopefully though if a parser wants to blow up on a badly formed document , it will at least leave a decent complaint :) |
| 15:41 | jgraham | wishes hg log interacted with the console like git log |
| 15:43 | <Philip`> | How does git log interact with the console? |
| 15:43 | <Philip`> | (Is it something that can't be emulated with "hg log|less"?) |
| 15:43 | <jgraham> | no, it is just like hg log | less |
| 15:44 | <jgraham> | but without having to remember to type |less |
| 15:44 | <jgraham> | and it works sensibly when writing to files and so on, as you would expect |
| 15:46 | <Philip`> | I remember to type |less so frequently that I often write "svn commit|less" which causes the nano UI to be displayed via less, which causes quite a mess |
| 15:49 | <zcorpan_> | Philip`: did you get my second email with the attachment? |
| 15:50 | <Philip`> | zcorpan_: I did |
| 15:50 | <Philip`> | zcorpan_: Did you get my link to the results page that incorporated your results? |
| 15:50 | <Philip`> | (on IRC) |
| 15:51 | <zcorpan_> | hmm no |
| 15:52 | <Philip`> | 11:28 < Philip`> zcorpan_: http://philip.html5.org/tests/canvas/suite/tests/results.html - thanks |
| 15:54 | <Philip`> | (IE9 seems to do quite well, particularly given that it hasn't got globalCompositeOperation yet) |
| 15:56 | <zcorpan_> | yeah |
| 15:58 | <Philip`> | Peculiar that it parses rgb(255 0 0) as red |
| 15:58 | <Philip`> | but rejects all the other invalid values |
| 15:59 | <jgraham> | Hmm, I think lots of those unicode tokenizer tests are wrong. The spec says to convert surrogates to U+FFFD |
| 16:00 | <jgraham> | I wonder when that changed... |
| 16:13 | <zcorpan_> | ie9 supports ' |
| 16:23 | <zcorpan_> | http://simon.html5.org/sandbox/html/w3c-home-in-html5 renders correctly in ie9 (but the <code>s are tiny) |
| 16:24 | <zcorpan_> | seems they don't support drawImage(video) |
| 16:30 | Philip` | has no tests for drawImage(video), sadly |
| 16:31 | <zcorpan_> | annevk: have you noticed that hovering the logo on your blog in opera is ... weird? |
| 16:32 | <Philip`> | Is http://www.w3.org/TR/CSS/fonts.html#font-shorthand more correct than http://www.w3.org/TR/css3-fonts/#font for defining syntax? |
| 16:34 | <Mike^ses> | jgraham: http://html5.org/tools/web-apps-tracker?from=3870&to=3871 |
| 16:42 | <zcorpan_> | hmm, ie9 parses the "<b>" as text in <svg><title><b> |
| 16:43 | <zcorpan_> | but they break out correctly for <svg><b> |
| 16:43 | <zcorpan_> | <svg><style><b> is like title |
| 16:43 | <zcorpan_> | but <svg><script><b> inserts a B element |
| 16:54 | <taptapir> | hey Philip` |
| 16:54 | <Philip`> | Hello |
| 16:55 | <taptapir> | so what you are suggesting is that, everytime an image is loaded they should set a flag or somethign to true, and when all of the image flags are set to true the last one should start drawing and other thigns... |
| 16:56 | <Philip`> | If your aim is to wait until they've all downloaded, and you want to download them all in parallel, then you'll need to do something like that |
| 16:56 | <taptapir> | yep. it's a webgl game so I need my assets (images and audio) loaded before anythign happens |
| 16:57 | <Philip`> | You could keep an array of all the images and check the img.complete property to see if they've been downloaded yet |
| 16:57 | <taptapir> | oh that would be more elegant fo sure |
| 16:57 | <Philip`> | though that's probably slow if you've got hundreds of files being loaded |
| 16:57 | <taptapir> | or what about hidden <img> fileds and placing an onLoad event to the body tag of the page and doing everything after that? |
| 16:58 | <Philip`> | You could just increment a counter and see if it's reached number_of_images |
| 16:58 | <Philip`> | or have a timer that polls a few times a second to see if everything's loaded yet, etc |
| 16:59 | <taptapir> | why would the "img.complete" method be slower? |
| 16:59 | <Philip`> | Yeah, you could use <body onload> too if you know all the images before the page has first finished loading |
| 17:00 | <Philip`> | If you have 1000 images, and each time an image loads you loop through all 1000 images to check whether they're all loaded yet, you'll be doing 10^6 checks |
| 17:01 | <Philip`> | (though if you have 1000 images you should stick them in a texture atlas or something so that you don't have 1000 images :-) ) |
| 17:01 | <zcorpan_> | waiting for <body onload> seems like the cleanest solution |
| 17:02 | <taptapir> | Philip`: probably :) |
| 17:02 | <taptapir> | zcorpan_: can hidden img tags slow down the browser |
| 17:03 | <zcorpan_> | taptapir: new Image() also delays the load event |
| 17:03 | <taptapir> | zcorpan_: because in that method I could end up with 1000 hidden img tags |
| 17:03 | <zcorpan_> | don't need to be in the document |
| 17:03 | <Philip`> | I like doing it more like http://canvex.lazyilluminati.com/83/play.xhtml where it doesn't force you to wait for everything to download first :-) |
| 17:04 | <Philip`> | (That just polls for img.complete every frame) |
| 17:04 | <taptapir> | Philip`: are textures loaded dynamically every frame? |
| 17:05 | <Philip`> | taptapir: Not quite sure what you mean |
| 17:08 | <taptapir> | zcorpan_: can you elaborate it a little. I'm too new to js and dom :/ what I know is that I can use an image from the dom by querying its id. how can I query it if its not in the document ? |
| 17:08 | <taptapir> | I would understand if you 2 would choose to ignore me hehe :) |
| 17:09 | <taptapir> | Philip`: you said it polls for im.complete every frame and I thought you load the textures when they are needed |
| 17:09 | <taptapir> | thats what I meant by dynamic |
| 17:09 | <zcorpan_> | taptapir: you don't query, you create the image from script |
| 17:09 | <zcorpan_> | taptapir: <script> var x = new Image(); x.src = 'foo'; </script> <body onload="alert(x.complete)"> |
| 17:10 | <taptapir> | ohh that clarifies a lot of things. somehow I thought body onload would only work for html data |
| 17:10 | <Philip`> | taptapir: Ah - it starts loading all of the textures immediately when you start the game, not dynamically based on visibility, and then just renders each one as the 'loading' image until the real image has been downloaded |
| 17:11 | <taptapir> | ok, it was real fast so I didn't noticed the loading images :) |
| 17:11 | <taptapir> | well that's another way. not bad actually. like a webpage :) |
| 17:13 | <zcorpan_> | taptapir: are you making your game multiplayer with websockets? :) |
| 17:13 | <taptapir> | zcorpan_: not this one. it's a single player adventure game in the spirit of Machinarium that I originally wrote in C (not complete) |
| 17:14 | <zcorpan_> | ok |
| 17:14 | <taptapir> | but I have lots of ideas about websockets and multiplayer games :) |
| 17:14 | <zcorpan_> | awesome :) |
| 17:14 | <Peter`> | Porting OpenTTD to canvas/websocket would be awesome |
| 17:15 | <Peter`> | but highly unlikely |
| 17:15 | <taptapir> | somtehing like Transformice. it's a flash game where a shaman mouse leads other mice to cheese :) there is physics involved etc.. |
| 17:15 | <Peter`> | (fyi: open source version of Transport Tycoon Deluxe) |
| 17:15 | <taptapir> | Peter I think it's likely, did you see freeciv |
| 17:16 | <Peter`> | Oh it certainly can be done |
| 17:16 | <Peter`> | but it would take an insane amount of time |
| 17:17 | <taptapir> | hmm. if it uses opengl I think it should be pretty straight forward to port it to webgl |
| 17:17 | <taptapir> | input etc. would take time though |
| 17:19 | <Philip`> | http://philip.html5.org/demos/canvas/spritepick/example.html - that's like half of TTD already |
| 17:19 | <taptapir> | I should get the hang of this async. business first. Coming from C, I'm lost! |
| 17:19 | <Philip`> | It just needs the buildings to be arranged a bit better, and for gameplay to be added |
| 17:19 | <taptapir> | Philip`: wow nice sprites. |
| 17:20 | <taptapir> | oww you stole them :p |
| 17:20 | <Philip`> | Of course :-) |
| 17:22 | <Philip`> | Peter`: Would it be more complex than a straightforward (but tedious) line-by-line C++->JS port? |
| 17:23 | <Philip`> | I guess the biggest pain is when games use lots of giant external libraries (for graphics or GUI or physics etc), because you wouldn't want to port all of them too |
| 17:23 | <taptapir> | so from what I understand, async means: every function is run from top to down but the VM doesn't wait for them to complete their job? |
| 17:23 | <taptapir> | and everything in JS works that way? |
| 17:23 | <Philip`> | but otherwise it might not be too hard |
| 17:24 | <Philip`> | taptapir: Every function call always runs to completion |
| 17:25 | <Philip`> | The asynchronous bit is like in 'img.onload = function () { do something }' - that's just defining a function, not executing it, and then storing it somewhere (in the image) and carrying on |
| 17:25 | <Philip`> | and some time after that script has finished, the browser will decide to call the function you stored in onload |
| 17:26 | <Philip`> | JS itself doesn't do anything asynchronous - it just registers callbacks with the browser environment |
| 17:26 | <Philip`> | effectively |
| 17:26 | <taptapir> | hmm, I started some javascript books yesterday but none of them seems to mention the async nature of it. any good books/sources you can recommend? |
| 17:26 | Philip` | doesn't know of any |
| 17:27 | <Philip`> | Are you familiar with function pointers and callbacks in C? |
| 17:27 | <taptapir> | yep |
| 17:28 | <Philip`> | It's basically the same thing, except instead of using a function pointer like "&function_that_does_stuff" you can define it inline like "function() { do stuff }" |
| 17:30 | <taptapir> | finally! thanks I get it a little bit better now |
| 17:31 | <Philip`> | (You can also do "function do_stuff() { do stuff }; ... img.onload = do_stuff;" which is equivalent to "img.onload = function() { do stuff }") |
| 17:31 | <taptapir> | yeah the first one is more C like :) |
| 17:31 | <Philip`> | (except in the latter case the function can refer to variables in its container scope) |
| 17:31 | <taptapir> | inline callbacks, that made sense. |
| 17:32 | <Philip`> | JS books should talk about closures, which are the relevant thing here |
| 17:32 | <taptapir> | checking... |
| 17:33 | <taptapir> | yes it's mentioned, I'm not there yet though. I'll pay special attention to that chapter |
| 17:34 | <Philip`> | I think they're a slightly strange concept when coming from C, but pretty useful once you understand them properly :-) |
| 17:35 | <taptapir> | got it! thanks Philip`and zcorpan_ |
| 19:16 | <taptapir> | in JS, functions are objects right? |
| 19:16 | <taptapir> | howcome "document.body.insertAdjacentElement(document.createElement("div").innerHTML = "ljkghljkg");" is illegal? |
| 19:18 | <TabAtkins> | The return value of 'document.createElement("div").innerHTML = "ljkghljkg"' isn't a valid arg for insertAdjacentElement(). |
| 19:18 | <TabAtkins> | This isn't jQuery, where everything returns a reasonable value for chaining. ^_^ |
| 19:27 | <taptapir> | arghh, right sry |
| 19:30 | <TabAtkins> | Was your "functions are objects" question unrelated? |
| 19:31 | <TabAtkins> | Because yes, they are. And most objects are functions, so you can get a constructor. |
| 19:32 | <ayapo> | woah, just checked out the site hosting the irc logs. The yellow highlight function is epic awesome. |
| 19:34 | <taptapir> | TabAtkins: i thought the createElement().innetHTML part was the illegal one so I was assuming that createElement was not an object. |
| 19:34 | <TabAtkins> | Ah, that's unrelated. createElement() returns a Node object. |
| 19:37 | <taptapir> | what's a node object? html tags? |
| 19:37 | <TabAtkins> | Yeah, it's a class that represents an element in the DOM. Also other things, like text or comments. |
| 19:46 | <paul_irish> | boblet proposed using css transitions and transforms without the prefix (in addition to the prefixed variants) as the spec isnt likely to change and end users are unlikely to update their code when the spec hits CR... |
| 19:46 | <paul_irish> | i can totally see that POV but i've been holding off on going unprefixed for un-cr things so far.. |
| 19:47 | <paul_irish> | anyone here take a side on that? |
| 19:47 | <TabAtkins> | As long as you're willing to go back and fix your code, there's no problem. |
| 19:47 | <hsivonen> | yes, there is |
| 19:47 | <hsivonen> | implementors don't know if your site is maintained by a person who is willing to adjust |
| 19:47 | <paul_irish> | i can handle either case, i just want to make the right recommendation to other developers. |
| 19:47 | TabAtkins | wants to write a mixin spec for CSS which would solve the prefix problem without the problems of removing them entirely. |
| 19:48 | <paul_irish> | who are likely a lot more lazy |
| 19:48 | <hsivonen> | (I use unprefixed stuff myself, though.) |
| 19:48 | <TabAtkins> | hsivonen: But going with or without prefixes both cause problems for implementors. |
| 19:48 | <taptapir> | let's say I create an image in JS and point it to a PNG file in the server with _.src property, when does the browser actually starts to download it? also when the image.onload event occurs, does it mean that the download is finished or that the image is in the memory? or do I make any kind of sense :) |
| 19:48 | <hsivonen> | TabAtkins: indeed, I use unprefixed versions partly in protest against long-term prefixing |
| 19:49 | <paul_irish> | webkit animation is one i'd certainly not leave unprefixed. that seems to be a bit more contentious, yes? |
| 19:49 | <hsivonen> | yeah |
| 19:49 | <TabAtkins> | Animation's complex enough that I wouldn't unprefix it. Too much change of *something* changing underneath you. |
| 19:49 | <paul_irish> | For sure. thx guys |
| 19:50 | <TabAtkins> | And yeah, the convo's died down somewhat, but I'm agitating for some changes to animations. |
| 19:50 | <Philip`> | taptapir: It could start to download in the background as soon as you set .src, though I don't know whether browsers will wait for the current script to return first |
| 19:51 | <Philip`> | taptapir: onload will be called some time (probably immediately) after the image is downloaded and decoded and found to not have errors (else it'll call onerror instead, I think), and you can safely use the image for drawing in canvases etc after that |
| 19:52 | <taptapir> | Philip`: so in this phase image is still not read to the memory right? |
| 19:53 | <Philip`> | taptapir: Depends on what you mean by being in memory |
| 19:54 | <Philip`> | taptapir: Browsers are free to do whatever optimisations they want in terms of caching data in memory, lazy decoding, etc, so you can never really know what they're doing - you just know whether or not it's safe to perform certain actions on images (like drawing them) |
| 19:55 | <taptapir> | I mean, is the image data transferred to a buffer in the memory? or is it simply presented by something like file pointers in C? |
| 19:56 | <TabAtkins> | Again, you don't know, because browsers are allowed to do what they want. All you know is that certain actions are allowed at specific points. |
| 19:56 | <Philip`> | It's been downloaded from the network (so you can use it immediately), but everything else is just implementation details that you shouldn't need to know about |
| 19:57 | <AryehGregor> | annevk, the entire IE compatibility list thing can be fairly described as "hard to do", but IE has to do it anyway, since they have to deal with IE-only sites and IE-only code paths, so if they just implemented standards without compatibility they'd break. (See: IE7.) |
| 19:57 | <taptapir> | Philip`: you know I want to download my game assets before the game starts, what I wonder is that if I do that will it fill up the memory? |
| 19:57 | <taptapir> | lets say for 1000 images |
| 19:58 | <AryehGregor> | Everyone else tends to get the same code paths in most cases, due to "develop your site for standards-compliant browsers, then fix for IE". So they can converge on a common behavior without causing more bugs than they prevent, usually. |
| 19:58 | <AryehGregor> | IE9's goal is explicitly to work with the same exact code everyone else gets, so hopefully IE10 won't need an IE9 compatibility mode. |
| 20:00 | <Philip`> | taptapir: That's an implementation detail :-) |
| 20:01 | <Philip`> | taptapir: It seems plausible that some implementations would store all loaded images as decoded RGBA in RAM, so you don't want to load huge amounts of image data and just hope the browser will do something clever with it |
| 20:02 | <taptapir> | Philip`: I better load each levels' assets when the time to use comes just to be safe then. |
| 20:02 | <Philip`> | taptapir: Depends on whether you're talking about megabytes or hundreds of megabytes or gigabytes of decoded image data |
| 20:03 | <Philip`> | taptapir: (If you're using them with WebGL then you have to worry about VRAM too, which is much more limited than RAM+swap) |
| 20:03 | <taptapir> | exactly |
| 20:03 | <taptapir> | I'M talking about couple hundreds of MB image data |
| 20:03 | <taptapir> | and about 100 MB of audio |
| 20:05 | <taptapir> | hmm wait, I'm not sure I can use an Image() object directly with webgl |
| 20:05 | <taptapir> | maybe there is a loadtexture function like in opengl to read it to VRAM |
| 20:05 | <airox> | hi guys |
| 20:06 | <airox> | Would watchPosition in the Geolocation API still report positions when the browsers (safari on iphone for example) is not active ? |
| 20:07 | <Philip`> | The web environment is not designed to provide much control of resource usage (since it's meant to provide portability instead, which means taking resource control away from authors and leaving it to browser developers), so if you want to push the boundaries then you'll just have to do lots of experimentation and make browser-specific assumptions |
| 20:07 | <AryehGregor> | Or wait till everyone supports NaCl. |
| 20:07 | <Philip`> | which is an irritating concept when you're used to having complete control over your environment |
| 20:08 | <Philip`> | You can pass an Image to glTexImage2D or something like that, I think |
| 20:08 | <taptapir> | Philip`: yeah I'm already irritated but the concept of playing hardware accelerated games in browser is just too good to pass |
| 20:09 | <taptapir> | Philip`: yes I checked it. I need to explicitly pass it to a webgl texture so I don't have to worry about browser filling the VRAM. and I can live with a couple hundreds MB of RAM usage |
| 20:12 | <taptapir> | I think I'll start by writing a decent game resource manager in javascript before anything else. |
| 20:24 | <taptapir> | is Image() object inherited form a regular node object? if so why there is an Image() object when it can also be created with .createElement("img")? |
| 20:44 | <AryehGregor> | What format do HTML5 tests need to be in? It occurs to me that I wrote some tests for my Mozilla patch to enable maxlength on textarea, and could probably submit those. |
| 20:46 | <Philip`> | I don't think any format has really been agreed yet |
| 20:46 | <AryehGregor> | Well, there must be some way to run all the tests and figure out how many passed, no? |
| 20:47 | <Philip`> | There ought to be |
| 20:48 | <Philip`> | Actually, there probably shouldn't be a way to run *all* the tests |
| 20:48 | <Philip`> | because nobody will really want ten thousand results all at once, they'll just focus on a single feature |
| 20:49 | <TabAtkins> | AryehGregor: That's the part that I'm working on. |
| 20:49 | <AryehGregor> | Ah. |
| 20:50 | <AryehGregor> | When there is a standard test format, hopefully all the browsers can write tests for new features using that format and wrap it in their own format for internal use. |
| 20:50 | <AryehGregor> | That would be cool. |
| 20:50 | <TabAtkins> | Exactly. |
| 20:50 | Philip` | will just submit his own idiosyncratic format for now |
| 20:50 | <AryehGregor> | Since it seems to me that WebKit and Gecko must be writing tons of HTML5 tests for internal use, but they aren't submitting them like Microsoft is. |
| 20:50 | <AryehGregor> | I mean, both of them have an "all new features must have tests" policy . . . |
| 20:51 | <TabAtkins> | Yeah, I know our format is very ideosyncratic. |
| 20:51 | <AryehGregor> | Everyone's is. |
| 20:56 | <othermaciej> | AryehGregor: WebKit definitely has tests for HTML5 features |
| 20:56 | <othermaciej> | I think there would be a few steps required to make them into appropriate tests for a standards test suite: |
| 20:57 | <othermaciej> | 1) ensure that everything tested is associated with specific conformance requirements in the spec |
| 20:57 | <othermaciej> | 2) annotate the tests with the relevant conformance requirements |
| 20:57 | <othermaciej> | 3) convert to whatever format the w3c wants them in |
| 20:57 | <AryehGregor> | Come to think of it, my maxlength tests won't work at all cross-browser, because they synthesize keyboard input. Oops. |
| 20:58 | <othermaciej> | we have a lot of tests that are script-driven, and would totally work cross-browser |
| 20:58 | <othermaciej> | we have some that are based on dumps of the render tree structure which would need substantial reworking |
| 20:58 | <othermaciej> | I don't believe Apple, Google, or any of the other big players in WebKit-land have anyone currently tasked with converting WebKit regression tests for test suite use |
| 21:00 | <Rik`> | I think Opera contributes a lot of test, no ? |
| 21:02 | <othermaciej> | I believe they do in general, I am not sure how many have been contributed for HTML5 in particular |
| 21:08 | <annevk> | zcorpan, yeah, haven't filed a bug yet |
| 22:34 | <MikeSmith> | Hixie: can you please add http://www.w3.org/TR/2010/WD-html5-20100304/ to the set of Previous Version links in the spec |
| 22:34 | <Rik`> | anyone know a very simple module to work with WebSockets ? (php, python, nodejs, whatever) |
| 22:37 | <Hixie> | MikeSmith: yeah, hold on, i have to check in what i'm doing at the moment first |
| 22:37 | <MikeSmith> | k |
| 22:37 | <AryehGregor> | Why does the "Negative effects" part of the change proposal thing even exist? Half the proposals don't seem to even use it, and in the ones that do it's usually not very convincing. |
| 22:40 | <MikeSmith> | Rik`: dunno how they might compare |
| 22:41 | <MikeSmith> | I guess the python one that the chrome dev team made was the first one |
| 22:41 | <MikeSmith> | but might not be the simplest to start with |
| 22:41 | <TabAtkins> | AryehGregor: Shrug. Yeah, someone writing a change proposal usually isn't the best person to write up negative effects against their proposal. |
| 22:42 | <MikeSmith> | Rik`: http://jwebsocket.org/ looks quite active at least |
| 22:43 | <Hixie> | Rik`: write your own :-) |
| 22:43 | <Rik`> | Hixie: well, I might end up doing that but I wanted a simple one to start and then see where I can go from there |
| 22:44 | <MikeSmith> | http://twitter.com/_romuald/status/16914358097 "Web standards my ass : http://blog.chromium.org/2010/06/websocket-protocol-updated.html #html5" |
| 22:44 | <Rik`> | MikeSmith: thanks but, hum, not Java :) |
| 22:44 | <MikeSmith> | Rik`: yeah, I figured that would probably not end up at the top of your list :) |
| 22:45 | <Hixie> | Rik`: well you can play with my perl demo: http://damowmow.com/playground/demos/websocket |
| 22:47 | <Rik`> | Hixie: I meant something to abstract the protocol for me so that I can just implement my application |
| 22:51 | <Rik`> | http://github.com/miksago/node-websocket-server might be what I'm looking for |
| 22:54 | <MikeSmith> | Rik`: that's micheil I guess |
| 22:54 | <MikeSmith> | he's around here on #whatwg sometimes |
| 22:58 | <Hixie> | Rik`: there's not much to abstract, but ok :-) |
| 23:01 | <othermaciej> | hello all |
| 23:01 | <MikeSmith> | hei |
| 23:04 | <Rik`> | MikeSmith: well, next time you see him, tell him congrats, works like a charm |
| 23:06 | <MikeSmith> | Rik`: cool |
| 23:07 | <MikeSmith> | http://www.reuters.com/article/idUSLDE65N14720100624 |
| 23:07 | <MikeSmith> | includes "<^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Graphic on Nokia N-series losing to iPhone: here ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^>" |
| 23:08 | <MikeSmith> | good technique |
| 23:08 | <MikeSmith> | much easier than taking the trouble to actually create a graphic that illustrates whatever point you are trying to make |
| 23:08 | <MikeSmith> | to be fair the "here" part is a link to an actual graphic, but let's pretend for a moment that's not really there |
| 23:09 | <paul_irish> | TabAtkins: i dont see this announced but looks like flexible box is going into ie9 http://paulirish.com/i/2520.png |
| 23:09 | <Hixie> | MikeSmith: ok, i'm at a stopping point -- what was the change you wanted? |
| 23:09 | <MikeSmith> | to add http://www.w3.org/TR/2010/WD-html5-20100304/ to the set of Previous Version links |
| 23:09 | <Hixie> | ah ok |
| 23:09 | <Rik`> | paul_irish: wow, seriously ? |
| 23:09 | <TabAtkins> | paul_irish: Interesting. That would explain my AlexMog has been bugging about flexbox so much. |
| 23:10 | <TabAtkins> | That image gonna stick around for a bit, paul_irish? |
| 23:10 | <Hixie> | man, i wish they'd implement stuff that's stable like DOM Events rather than stuff that's still in development |
| 23:10 | <annevk> | it's already June 25 here; did we fail to publish? |
| 23:11 | <paul_irish> | TabAtkins: yup. |
| 23:11 | <Rik`> | Hixie: haven't they done that with the first preview ? |
| 23:11 | <TabAtkins> | K, just asked AlexMog about it. |
| 23:11 | <MikeSmith> | yeah, Alex, implement DOM Events! forget about that CSS crap |
| 23:11 | <MikeSmith> | annevk: I failed to publish |
| 23:11 | <Hixie> | Rik`: there's tons of stuff they could implement that's more stable than unfinished css3 or html5 stuff |
| 23:12 | <MikeSmith> | we will publish on June 25 and pretend that it's June 24 |
| 23:12 | <Hixie> | Rik`: do they really have a bug-free implementation of dom events yet? |
| 23:12 | <TabAtkins> | Especially Flexbox, which is definitely changing from the current draft. |
| 23:12 | <annevk> | MikeSmith, kk |
| 23:12 | <TabAtkins> | Not sure *how much* it'll change now, but change it will. |
| 23:12 | <annevk> | it's prefixed so that seems cool |
| 23:12 | <Hixie> | MikeSmith: k checked it in |
| 23:12 | <annevk> | having an experimental impl is nice |
| 23:12 | <Hixie> | MikeSmith: anything else? |
| 23:12 | <TabAtkins> | Yeah, that's good at least. |
| 23:12 | <MikeSmith> | Hixie: thanks |
| 23:12 | <annevk> | what do you mean "at least"? |
| 23:12 | <MikeSmith> | Hixie: that should be it I believe |
| 23:13 | <Hixie> | MikeSmith: cool |
| 23:13 | annevk | doesn't see the problem |
| 23:13 | <Rik`> | Hixie: don't know about the bug free |
| 23:13 | <annevk> | anyway, naptime |
| 23:13 | <TabAtkins> | I didn't really mean anything. Habit. |
| 23:13 | <Rik`> | Hixie: it seems to me they've implemented pretty much all the stable stuff used on the web |
| 23:14 | <TabAtkins> | http://www.tehrantimes.com/index_View.asp?code=221957 |
| 23:14 | <MikeSmith> | browser projects need to find more engineers who really, really prefer to implement stable boring old stuff rather than fun and interesting new stuff |
| 23:14 | <MikeSmith> | that'd solve the problem |
| 23:14 | <paul_irish> | annevk: can i ask you about elem.style.VendorProps in pm? |
| 23:16 | <Hixie> | Rik`: it's the stable stuff not used on the web that they're holding back |
| 23:17 | <Hixie> | Rik`: my concern is that they have a habit of implementing a great breadth of standards with a very shallow depth of quality |
| 23:17 | <Hixie> | Rik`: and we end up having to clean up the mess of all the bugs they introduce into the platform |
| 23:18 | <annevk> | paul_irish, email annevk⊙oc |
| 23:18 | <annevk> | paul_irish, I'm going to bed |
| 23:18 | <annevk> | nn |
| 23:18 | <Hixie> | nn |
| 23:22 | <AryehGregor> | Hixie, are there major bugs in IE8's CSS2.1 support? |
| 23:23 | <Rik`> | yeah, bugs are bad. but if there were more W3C tests… |
| 23:23 | <AryehGregor> | Has anyone reported big interoperability problems with IE9's stuff, either? I imagine a major goal would be "get all the big canvas/video/etc. demos working flawlessly". |
| 23:23 | <TabAtkins> | Well, it passes the CSS2.1 test suite better than anyone else, iirc. |
| 23:23 | <AryehGregor> | It looks to me like Microsoft has been doing a good job of standards support since IE8. |
| 23:23 | <AryehGregor> | Arguably since IE7, if you account for the fact that that was crippled by not throwing out IE6's engine and writing a CSS engine from scratch. |
| 23:23 | <Rik`> | AryehGregor: http://jhop.me/ie8-bugs |
| 23:24 | <Hixie> | AryehGregor: honestly it's been so long since i've actually done real testing, i've no idea |
| 23:24 | <AryehGregor> | Rik`, well, there are probably a lot more than 68 CSS2.1 bugs filed in either Mozilla's or WebKit's tracker. |
| 23:25 | <AryehGregor> | In practice, it seems to be just fine. |
| 23:25 | <AryehGregor> | Also, "Only partial implementation of CSS3 writing-mode property" <-- that doesn't look like CSS2.1 to me. |
| 23:27 | <Hixie> | back when i was testing as a full-time job, most "lists of bugs" for browsers for which i knew the filed bug counts were off by a factor of 100 or more |
| 23:27 | <Hixie> | if that ratio is still true today, then 68 bugs is equivalent to about 7000 bugs if you did real thorough testing |
| 23:28 | <AryehGregor> | And the probably hundreds or thousands of CSS bugs filed in other UAs' bug trackers is how many? :) |
| 23:28 | <Hixie> | hundreds or thousands |
| 23:29 | <AryehGregor> | I'm pretty sure I've seen bzbarsky or someone say IE8 has the best CSS2.1 implementation. |
| 23:29 | <Hixie> | well i guess we'll see when it comes time for me to write acid4 :-P |
| 23:29 | <Hixie> | since acid4 is likely to be a CSS test suite |
| 23:29 | <TabAtkins> | And I've seen tables listing CSS2.1 test suite failures across browsers produced by arronei. Consider the source, of course, but still. ^_^ |
| 23:29 | <Hixie> | how are they doing with acid3 at the moment? |
| 23:30 | <AryehGregor> | 83 or something. |
| 23:30 | <AryehGregor> | They got a lot less hostile to it with PP3. |
| 23:31 | <AryehGregor> | Previously they played it down a lot, but now they have it displayed on their landing page and everything. |
| 23:31 | <AryehGregor> | I wouldn't be at all surprised if they shoot for 100 on the final release. |
| 23:31 | <Hixie> | their rendering still has some way to go |
| 23:31 | <Hixie> | but it's getting there |
| 23:31 | <Hixie> | (based on the screenshot on wikipedia) |
| 23:32 | <Hixie> | wow they do @font-face with ttf fonts? |
| 23:32 | <TabAtkins> | Yup, that made it in. |
| 23:32 | <Rik`> | ttf, otf, woff |
| 23:33 | <Hixie> | that's quite a nice |
| 23:33 | <daedb> | IE9 PP3 has the funniest Acid3 bug though, it shows fail if png files are associated with a non-default program :) |
| 23:33 | <Hixie> | change |
| 23:33 | <Hixie> | daedb: hah |
| 23:33 | <daedb> | They noted it in the release notes and everything. |
| 23:33 | <Hixie> | do they show the right favicon? |
| 23:34 | <Hixie> | (which is to say, no favicon or a cat, but not a red square) |
| 23:34 | <daedb> | There's no adress bar, so can't see favicons... |
| 23:34 | <Hixie> | ah |
| 23:39 | <Hixie> | anyone got any opinions on how WebSRT cues should parse if they don't have a valid timestamp? |
| 23:40 | <TabAtkins> | Drop on the floor? |
| 23:41 | <Hixie> | the whole cue up to the next double blank line? |
| 23:41 | <Hixie> | that's kinda what i figured should happen |
| 23:41 | <Hixie> | other options are to just ignore the one line and try reading a timestamp from the next one |
| 23:41 | <Hixie> | or to drop the whole file from there on |
| 23:42 | <Hixie> | or to assume a timestamp of 0 |
| 23:42 | <Hixie> | or of equal to the last timestamp seen |
| 23:43 | <TabAtkins> | I don't think anything that assumes a timestamp would be good. |
| 23:43 | <TabAtkins> | Dropping the whole file is no good either, if the rest is properly formatted. |
| 23:43 | <TabAtkins> | Ignoring just that line and trying the next sounds potentially cool. |
| 23:44 | <TabAtkins> | So, one of those two. Either skip the line, or skip everything until the next double blank line. |
| 23:45 | <Hixie> | i guess skip the block is safest |
| 23:47 | <variable> | is there any undefined or illegal code here: http://codepad.org/Ghv5oeHJ ?? |
| 23:48 | <AryehGregor> | #c++? |
| 23:48 | <variable> | woops |
| 23:48 | <variable> | wrong channel sorry |
