| 00:02 | <zewt> | <annevk> so that last argument is bogus and will be ignored <- can we apply this to mailing lists |
| 00:02 | <Hixie> | it's applied to the whatwg list fwiw |
| 03:52 | <Hixie> | anyone got windows handy? |
| 04:11 | <roc> | Hixie: yes |
| 04:12 | <Hixie> | can you think of a text box somewhere in the native windows UI (not an app and not IE) that is multiline and has some text in it by default? |
| 04:12 | <roc> | not off the top of my head |
| 05:06 | <hober> | annevk: I'm on the happiana list |
| 05:08 | <hober> | annevk: i had tried to subscribe when it was private, and stpeter added me when it opened up |
| 05:18 | <MikeSmith> | othermaciej: you around? wanted to ask about moving bugzilla bugs to LC components |
| 05:20 | <MikeSmith> | Hixie: I set up LC components for all HTML drafts, and I will need to do a mass-move of them today |
| 05:20 | <MikeSmith> | as we did back in October |
| 05:20 | <MikeSmith> | wow, it's a blooberry |
| 05:20 | <MikeSmith> | blooberry: don't recall seeing you around here for quite a while |
| 05:38 | <Hixie> | MikeSmith: k |
| 05:38 | <Hixie> | blooberry: sup dude |
| 06:10 | <nessy> | MikeSmith: outch! Shame we couldn't avoid that email deluge... |
| 06:25 | <MikeSmith> | nessy: yeah, sorry |
| 06:36 | <MikeSmith> | ok, mail bomb completed |
| 06:37 | <nessy> | uff |
| 06:37 | <MikeSmith> | Hixie: I turned your bugmail off about 30 minutes ago, before I launched the mail bomb |
| 06:37 | <MikeSmith> | but I just now turned it back on |
| 06:39 | <MikeSmith> | othermaciej: all existing bugs are now moved to the LC1 components |
| 06:39 | <MikeSmith> | if there are any problems, I'll be around for maybe 2 more hours |
| 06:40 | <MikeSmith> | then I have to catch a plane and will be offline for 12 hours or so |
| 06:54 | <hsivonen> | the "broadcaster's point of view" of HTML5 is interesting |
| 06:54 | <hsivonen> | treating the video element as TV instead of considering the other stuff around it |
| 06:55 | <hsivonen> | what if the user doesn't have a <video> playing during a tsunami? |
| 06:56 | <hsivonen> | shouldn't the tsunami warning system be a separate daemon that's always in the background? |
| 07:01 | <zcorpan> | like Twitter? |
| 07:06 | <hsivonen> | zcorpan: I was thinking of something running on the OS level and permission to stick stuff on the screen and to play audio |
| 07:07 | hsivonen | wonders why the CSS WG has reviewed a WD from May when sending feedback in August |
| 07:08 | <hsivonen> | "The following references link to Editor's Drafts, not Working Drafts" :-( |
| 07:11 | <zcorpan> | HbbTV-- for using a new doctype (and new MIME type) and wanting HTML entities to work |
| 07:13 | <zcorpan> | maybe i should instead of trying to fight against it suggest that the HbbTV doctype be added to html5 |
| 07:18 | <zcorpan> | although other browsers don't support the mime type anyway |
| 07:32 | <hsivonen> | zcorpan: what's HbbTV? |
| 07:34 | <Hixie> | silly side-effect of the component change is that it de-prioritised everything in the LC components since they've not changed more recently than anything in any other component -_- |
| 07:36 | <shetech> | Hey, gang |
| 07:36 | <shetech> | question about svg |
| 07:37 | <shetech> | hixie: want to take a stab at this? What's the different in and value of embedding svg using <img>, <embed>, <object> or <iframe>? I noticed that the <img> tag doesn't display consistenly across browsers (Chrome sorta barfs on it). Is there a time when one would use one or the other? |
| 07:38 | <shetech> | s/different in/difference in/ |
| 07:38 | <Hixie> | per spec, the only reason you'd not use img is if you want the graphic to be interactive |
| 07:38 | <Hixie> | then you'd use object. |
| 07:39 | <shetech> | Okay, that's pretty much what I wrote. Chrome's handling of the <img> tag for svg is fugly, though! |
| 07:40 | <shetech> | If you're curious, have a look at http://html5.shetech.com/chp_5.html, and look at the "A Circle, using the <img> tag to embed it" under SVG |
| 07:40 | <shetech> | FF does it right, Chrome does not. |
| 07:41 | <shetech> | Unless I'm doing it wrong? |
| 07:41 | <Hixie> | browsers have bugs, news at 11 :-) |
| 07:41 | <shetech> | Heh. No doubt. |
| 07:41 | <shetech> | Dummies wanna know! |
| 07:41 | <shetech> | :D |
| 07:41 | <hsivonen> | shetech: img has different security characteristics and doesn't create a browsing context |
| 07:41 | <shetech> | splain, please? |
| 07:41 | <hsivonen> | shetech: object and iframe have the disadvantage of creating a browsing context |
| 07:42 | <hsivonen> | shetech: inlining doesn't create a browsing context |
| 07:42 | <shetech> | ah |
| 07:42 | <hsivonen> | shetech: I can't remember what <embed> does |
| 07:42 | <hsivonen> | shetech: the main annoyance of a browsing context is that non-targeted links cause navigation within the browsing context instead of navigating the parent |
| 07:42 | <shetech> | yes |
| 07:43 | <shetech> | So for Dummies, maybe I shouldn't mention them as options? |
| 07:43 | <shetech> | Or with heavy disclaimers... |
| 07:44 | <shetech> | Have I mentioned lately that you guys rock? |
| 07:44 | <shetech> | Thanks |
| 07:47 | <shetech> | Okay, on that happy note, g'nite, all |
| 07:47 | <shetech> | Thanks |
| 07:55 | <zcorpan> | hsivonen: "Hybrid Broadcast Broadband TV". a variation of CE-HTML |
| 07:56 | <hsivonen> | zcorpan: It's that something that Opera needs to implement in order to please checklist-oriented embedding customers? |
| 07:56 | <zcorpan> | yes |
| 07:57 | <hsivonen> | my condolences |
| 08:32 | zcorpan | unsubscribes from public-canvas-api |
| 08:36 | <hsivonen> | huh? did the same person file a bug requesting content access to accessibility settings and a bug for making it so that content can't even infer the accessibility settigs? |
| 08:52 | <hsivonen> | I find it a bit odd that there is any question of "figuring out" if accessibility requirements for MathML and SVG stuff belong in MathML and SVG specs or the HTML spec |
| 08:55 | <hsivonen> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=13651 embodies the culture differece over what "conforming" means pretty well |
| 08:56 | <hsivonen> | I expect that bug to become a total bikeshed about the exact number of words and how to apply the requirement no doubt conceived with English in mind to German, Thai, Chinese, etc. |
| 08:59 | <zcorpan> | Documents must not contain more than 1500 words. |
| 09:02 | <zcorpan> | it's annoying that whitespace in javascript depends on unicode version |
| 09:02 | <zcorpan> | would have been nicer with frozen ascii-only whitespace characters |
| 09:10 | <hsivonen> | zcorpan: it seems that TC39 is bad at responding "no way" to pedantic i18n comments |
| 09:11 | <zcorpan> | at least ES doesn't do unicode normalization |
| 09:11 | <zcorpan> | maybe TC39 only cares when performance is affected |
| 09:11 | <hsivonen> | a while ago, I advised another person with generating correct JSON and I was impressed by the correctness of libraries out there |
| 09:12 | <hsivonen> | I expected libraries to be wrong when it case to Unicode paragraph breaks but they weren't |
| 09:12 | <hsivonen> | s/case/came/ |
| 09:13 | <hsivonen> | (maybe my expectations about library authors reading specs are too pessimistic) |
| 09:14 | <hsivonen> | (or all the libraries that were inspected had gone through the "somebody yells at them" phase already) |
| 09:15 | <hsivonen> | oh, and this wasn't just about generating correct JSON but about generating correct JSON that is also correct JS and that doesn't confuse the HTML parser when inlined in <script> |
| 09:16 | <hsivonen> | it's also remarkable how the libraries do a better job at dealing with the oddities of ES than the JSON spec |
| 09:31 | <annevk> | email overflow |
| 09:31 | <annevk> | bah |
| 09:54 | <zcorpan> | Hixie: http://html5.org/tools/web-apps-tracker?from=6359&to=6360 - shouldn't it be "If <var title="">n</var> is greater than *zero*" ? |
| 09:54 | <annevk> | read-world use cases |
| 09:54 | <annevk> | nice |
| 09:55 | <hsivonen> | annevk: do you mean the spec itself is a real-world use case for <var>? |
| 09:57 | <annevk> | was quoting a typo from Hixie |
| 09:59 | <annevk> | zcorpan, no, zero is address later |
| 10:00 | <zcorpan> | annevk: in step 4? |
| 10:02 | <zcorpan> | let's say you have <select></select> and do select[1] = option; |
| 10:03 | <zcorpan> | length is 0, index is 1, n is 1 |
| 10:03 | <zcorpan> | n is not greater than 1, and it's not zero, so per spec it does nothing |
| 10:05 | <zcorpan> | the spec is wrong in another way too |
| 10:06 | <zcorpan> | it says to append n new option elements, but it should append n-1 new option elements and then append value |
| 10:08 | <annevk> | hmm yeah |
| 10:09 | <zcorpan> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1093 - chrome and firefox have empty text nodes in some options |
| 10:21 | <hsivonen> | looks like the story today is Gruber's anti-Google bias eclipsing his anti-Microsoft sentiment |
| 10:23 | <annevk> | that was somewhat clear earlier on |
| 10:25 | <annevk> | http://yaccessibilityblog.com/library/aria-fix-non-standard-images.html :( |
| 10:30 | <zcorpan> | ARIA to the rescue! |
| 10:30 | <zcorpan> | weird that there's a performance difference |
| 10:31 | <annevk> | alright, subscribed to happiana |
| 10:32 | <annevk> | thanks hober |
| 10:46 | <hsivonen> | http://groups.google.com/group/epub-working-group/browse_thread/thread/ea06deb3500ac246?pli=1 |
| 10:47 | <hsivonen> | see also http://code.google.com/p/epub-revision/wiki/Telcon20110803 incl. comments below |
| 11:19 | <annevk> | CORS is now at http://dvcs.w3.org/hg/cors/raw-file/tip/Overview.html |
| 11:19 | <annevk> | yay Mercurial |
| 11:27 | <smaug____> | what is "LC1 HTML5 spec" in W3C? |
| 11:28 | <smaug____> | I mean in W3C bugzilla |
| 11:29 | <annevk> | it means the bug was raised during the Last Call period |
| 11:29 | <Philip`> | The first last call period, in particular |
| 11:29 | <smaug____> | hmm |
| 11:30 | <smaug____> | I haven't cared about last call |
| 11:30 | <smaug____> | since I've filed bugs about problems in whatwg html spec |
| 11:31 | <smaug____> | well, I guess the lc thing is just some bureaucracy I don't need to care about |
| 11:32 | <Philip`> | The spec isn't going to stop improving in response to feedback because of arbitrary deadlines, at least on the WHATWG side |
| 11:32 | <asmodai> | sigh: http://blogs.forbes.com/fredcavazza/2011/07/17/why-opposing-html5-and-flash-is-a-non-sense/ |
| 11:32 | <Philip`> | I guess the main effect it will have is the prioritisation of bugs |
| 11:32 | <Philip`> | (since pre-LC bugs should get fixed by some other arbitrary deadline) |
| 11:32 | <smaug____> | that is a bit strange |
| 11:33 | <annevk> | it's called the W3C |
| 11:33 | <smaug____> | :9 |
| 11:33 | <smaug____> | :) |
| 11:35 | <smaug____> | (not sure I really like whatwg process either, but since I don't know how to make things work better, I try to not complain too much :) ) |
| 11:39 | <zcorpan> | "ok what are you doning" - excellent bug report |
| 12:03 | <heycam> | annevk, I wonder if switching those two SVG accessibility bugs from the LC product/component/whatever will mess with the disposition of comments generation when it comes time to do that |
| 12:03 | <heycam> | (since they won't be found as easily) |
| 12:05 | <annevk> | guess we'll find out |
| 12:05 | <heycam> | :) |
| 12:06 | heycam | sleeps |
| 12:48 | <zcorpan> | foolip: if you want to hack around with the status boxes, please also fix the mispositioned boxes (they have old IDs e.g. "video" instead of "the-video-element") |
| 12:48 | <foolip> | zcorpan, wrong philip? |
| 12:49 | Philip` | certainly doesn't want to do anything about status boxes |
| 12:51 | <Workshiva> | Is this old news? http://www.enisa.europa.eu/media/press-releases/web-security-eu-cyber-security-agency-enisa-flags-security-fixes-for-new-web-standards |
| 12:53 | <zcorpan> | foolip: no. i saw a bug in which you volunteered to fix status box overlap |
| 12:53 | <Philip`> | Workshiva: Yes - http://krijnhoetmer.nl/irc-logs/whatwg/20110802#l-945 |
| 12:54 | <foolip> | zcorpan, oh, perhaps I did volunteer if someone would just point in the right direction |
| 12:54 | <foolip> | can't remember |
| 12:54 | <Workshiva> | Weird, it didn't show up in logs search |
| 12:54 | <Workshiva> | I guess it isn't indexed yet |
| 13:02 | <zcorpan> | foolip: what needs doing is that the IDs need to be updated, and some status boxes need to be removed, but there's no UI to do that (possibly intentionally) |
| 13:03 | <annevk> | ms2ger, I would like to move to full name for references |
| 13:04 | <annevk> | ms2ger, it's sort of the only place where the first name is abbreviated despite many people going by their first name |
| 13:04 | <annevk> | ms2ger, so from now on new entries will have full names and i might slowly migrate older entries |
| 13:05 | <annevk> | I really wish JSON had comments |
| 13:05 | <zcorpan> | "HTML5: Disabling Click-jacking Protection" is interesting |
| 13:09 | <jgraham> | Hmm, thinking "I want a javascript console" and accidentially enabling an ecmascript debugger in a tab with complete.html open seems to be a bad idea |
| 13:56 | <annevk> | zcorpan, html5-diff uses Anolis now |
| 13:57 | <annevk> | zcorpan, I put the Makefile in CVS, you still need to get https://bitbucket.org/ms2ger/specification-data yourself (and Anolis of course) |
| 14:01 | <zcorpan> | annevk: nice! |
| 14:01 | <zcorpan> | will look into it another day |
| 14:12 | <annevk> | kk |
| 14:13 | <zcorpan> | thanks btw |
| 15:18 | <annevk> | smaug____, should prolly say that while aDoneCallback is running you cannot invoke ModificationBatch methods |
| 15:19 | <smaug____> | annevk: why? |
| 15:20 | <annevk> | smaug____, what would happen? |
| 15:20 | <smaug____> | the "batch" would listen for some new modifications |
| 15:21 | <annevk> | I guess that could work |
| 15:21 | <annevk> | also, why not have specialized Modification interfaces? |
| 15:22 | <smaug____> | for simplicity |
| 15:22 | <annevk> | for text / children / attributes |
| 15:22 | <smaug____> | and Modification interface is kind of based on current MutationEvent |
| 15:23 | <annevk> | I noticed that |
| 15:26 | <smaug____> | annevk: can you see something terribly bad in the approach ? |
| 15:26 | <smaug____> | it is quite close to other proposals, of course |
| 15:29 | <annevk> | the only thing I think is a little weird is that ModificationBatch is also the argument passed to the callback |
| 15:29 | <annevk> | what is the 'this' of the callback? |
| 15:29 | <annevk> | and what does .modifications return while not in a callback? |
| 15:30 | <smaug____> | I thought I said that .modifications is empty while not in callback |
| 15:30 | <smaug____> | empty list |
| 15:30 | <annevk> | ah yeah, the comment says so |
| 15:31 | <smaug____> | annevk: the callback needs to get some parameter so that { handleBatch: function(b) {}} can be used |
| 15:31 | <annevk> | can't you pass "Modification[]" as parameter? |
| 15:31 | <annevk> | and not have .modifications |
| 15:31 | <annevk> | I think that would make more sense |
| 15:32 | <smaug____> | I could, but then one would need to track to which batch the modification is related to |
| 15:32 | <smaug____> | I mean if the same callback is used with many ModificationBatches |
| 15:33 | <smaug____> | and the callback itself call some of the methods of the ModificationBatch |
| 15:33 | <annevk> | is there a use for that? |
| 15:33 | <smaug____> | in the callback you may want to call unbatch*() |
| 15:38 | <annevk> | what are the ways to make .modifications longer than 1? |
| 15:38 | <annevk> | is that stuff like innerHTML, doing things from the callback, inserting DocumentFragment? |
| 15:41 | <smaug____> | yeah |
| 15:43 | <smaug____> | and probably dom range modifications could utilize that too |
| 15:43 | <annevk> | but for setAttribute insertBefore etc. not invoked from the callback it would typically be 1? |
| 15:43 | <smaug____> | typically 1 |
| 15:44 | <annevk> | specifying this would require some very careful algorithms to make it nice |
| 15:44 | <annevk> | but seems doable |
| 15:45 | <smaug____> | it shouldn't be too hard |
| 15:46 | <smaug____> | since, now that I think of it, the batching works quite similarly to gecko's DOMSubtreeModified batching |
| 15:47 | <smaug____> | you start a batch and end it, and the outer-most "end" (if there are nested start-ends) fires the callback |
| 15:47 | <smaug____> | but anyway, it possible that Google objects the approach |
| 15:48 | <smaug____> | and if that is the case, I'm not sure how to proceed with mutation events replacements |
| 15:50 | <smaug____> | perhaps someone from Opera could write a new proposal and everyone would be happy with it ;) |
| 15:54 | <smaug____> | but let's see what kinds of comments the proposal gets |
| 15:59 | <annevk> | fwiw (unrelated to the Modification discussion above) https://plus.google.com/112284435661490019880/posts/eivfrHppgLB |
| 16:00 | <annevk> | smaug____, I guess I could write a proposal that splits your Modification interface :p |
| 16:06 | <annevk> | smaug____, I guess innerHTML is actually also one modification |
| 16:06 | <annevk> | of the childList |
| 16:14 | <smaug____> | innerHTML may remove and add several nodes, so it could create several modifications |
| 16:16 | <annevk> | oh I see |
| 16:47 | <hsivonen> | hmm. so G+ has suspended users from Hong Kong who've used their English name |
| 16:47 | <hsivonen> | that sucks |
| 16:47 | <TabAtkins> | Man, our "Common Name" policy sucks. ;_; |
| 16:48 | <smaug____> | Does G+ require using a real name? |
| 16:48 | <hsivonen> | what does G+ do about e.g. Russian names if a Russian person officially has a Cyrillic name but uses a Romanization in international contexts? |
| 16:48 | <TabAtkins> | It requires a "Common Name", but the difference between that and a "Real Name" are nebulous and ill-defined. |
| 16:48 | <hsivonen> | smaug____: yes |
| 16:48 | <smaug____> | huh |
| 16:48 | <TabAtkins> | And I think we're really burning goodwill with this silliness. |
| 16:49 | <smaug____> | I think G+ even asked gender when I activated it |
| 16:49 | <hsivonen> | I've so far seen a Romaji Japanese name and a Kanji Japanese name |
| 16:50 | <smaug____> | but there was some way to not answer to that question |
| 16:50 | <hsivonen> | both *so far* unsuspended... |
| 16:50 | <TabAtkins> | smaug____: Yeah, you could say "Other", and also I think we hide Gender by default now? |
| 16:50 | <zewt> | g+ not supporting gapps sort of makes it hard to take seriously |
| 16:50 | <TabAtkins> | I suspect it was just for pronoun use or something. |
| 16:50 | <zewt> | as if i'm going to create a separate google account just for one google product |
| 16:50 | <dglazkov> | good morning, Whatwg! |
| 16:51 | <dglazkov> | it's a lovely component model day for me |
| 16:51 | <TabAtkins> | But still, despite the faults, it's working great for me. |
| 16:51 | <zewt> | it's like the opposite of single sign-in--multiple signins required for the same site |
| 16:51 | <TabAtkins> | Just need a few more people to jump ship from Twitter and I can start winding down over there. |
| 16:52 | <dglazkov> | TabAtkins: certainly one thing that makes + more habitable for some folks are extended rage-scussions on how W3C and WHATWG differ. Can't do those in 140 characters. |
| 16:53 | <TabAtkins> | dglazkov: Indeed! And they're actually pretty good discussions! |
| 16:53 | <TabAtkins> | Especially anything AryehGregor posts. +1 |
| 16:53 | <zewt> | gotta love twitter for training millions of people against writing nontrivial thoughts |
| 16:54 | <dglazkov> | zewt: it's more like |
| 16:54 | <dglazkov> | " |
| 16:54 | <zewt> | " |
| 16:54 | <dglazkov> | discovering your natural ability to say nothing in 140 characters" |
| 16:55 | <zewt> | it's sort of darkly amusing that people will actually defend it, as if not being able to write complex thoughts is a feature |
| 16:55 | <TabAtkins> | It... actually is. |
| 16:56 | <zewt> | sure, it's the great equalizer to help people incapable of complex thoughts mask the fact |
| 16:56 | <TabAtkins> | I enjoy it as one of my many communications channels. |
| 16:56 | <TabAtkins> | zewt: That's indeed one of the benefits. |
| 16:56 | <zewt> | that's not a positive. heh |
| 16:57 | <hsivonen> | zewt: actually, it takes thought and skill to compress thoughts into tweets |
| 16:57 | <TabAtkins> | zewt: Rather than reading too-long blog posts from people who can't think well, you can read pithy and intelligent-sounding tweets! |
| 16:57 | <hsivonen> | zewt: it's not at all clear that the compression is equally easy for different people |
| 16:57 | <zewt> | hsivonen: or, painfully more frequently, turns the internet into SMS-ese D: |
| 16:57 | <TabAtkins> | zewt: You follow the wrong people. |
| 16:58 | <zewt> | i don't follow anyone. heh |
| 16:58 | <TabAtkins> | Then there you go. |
| 16:58 | <TabAtkins> | The good half of Twitter talks in complete sentences and with proper spelling. |
| 16:58 | <TabAtkins> | The bad half is full of people I don't care about. |
| 16:59 | <dglazkov> | ok, the only thing I need on + is IRC. Then we're set. |
| 17:00 | Philip` | always uses proper capitals and hyphens and semicolons when writing SMS messages, which might be why it usually takes him five minutes to write one |
| 17:00 | <TabAtkins> | Philip`: lrn 2 typ fstr |
| 17:00 | <zewt> | i'll forgive sms-ese from people on dumbphones, where typing anything at all is pulling teeth ... but most people aren't, anymore, so |
| 17:00 | TabAtkins | had to think for a bit to compress that. |
| 17:00 | <dglazkov> | nothing emulates shouting past each other IRL better than IRC |
| 17:01 | <zewt> | heh, and it's probably *harder* to type that way on a smartphone, since autocorrect will flip out |
| 17:01 | <TabAtkins> | So yeah, zewt, you're making an incorrect generalization. A significant portion of Twitter (which includes pretty much everyone in tech, as far as I can tell) talks like real people on Twitter. |
| 17:01 | <gsnedders> | There's no way I could go back to a phone without a QWERTY keyboard now. So much quicker to type properly on. |
| 17:01 | <TabAtkins> | gsnedders: hard or soft keyboard? |
| 17:02 | <gsnedders> | TabAtkins: Either. |
| 17:03 | <TabAtkins> | kk. Just wondering if you had a preference. I used to require hard keyboards due to my gigantic thumbs, until the Nexus S finally had a large enough screen for me to use. |
| 17:03 | <TabAtkins> | (My thumbs are the size of a baby's head.) |
| 17:03 | <zewt> | i used a G1 initially, and couldn't use the virtual keyboard at all, but eventually switched to nS and crash coursed for a day to learn it |
| 17:04 | <Philip`> | 1Maybe I wi2l1 write with run-length encoding and s2e1 if the compre2s1ion helps me type faster3. |
| 17:05 | <Philip`> | 1H2m1, maybe not |
| 17:05 | <hsivonen> | Philip`: I WONTFIXed an HTML WG bug about that |
| 17:05 | <jgraham> | No doubt G+ will eventually get a "<140 characters" stream. It is obviously trying to merge the feature sets of twitter and facebook |
| 17:05 | <TabAtkins> | Hah, the <3br> one? |
| 17:05 | <hsivonen> | TabAtkins: yeah |
| 17:06 | <zewt> | http://redmine.ruby-lang.org/issues/5054 D: |
| 17:06 | <TabAtkins> | wut |
| 17:06 | <zewt> | i can't even tell if that's a troll |
| 17:07 | <hsivonen> | zewt: Python wins! |
| 17:07 | <TabAtkins> | "I'd prefer bad code to look ugly." <-- yes |
| 17:07 | <zewt> | well, it's Ruby, that's a given |
| 17:08 | <Philip`> | Should have labelled end statements |
| 17:08 | <Philip`> | just like labelled break and labelled goto |
| 17:08 | <TabAtkins> | Dammit, Philip`, I was *just* typing that. |
| 17:09 | <zewt> | <div><div><div><///div> |
| 17:09 | <TabAtkins> | /wrist |
| 17:09 | <TabAtkins> | Haha at comment #20 |
| 17:09 | <Philip`> | TabAtkins: lrn 2 typ fstr |
| 17:10 | <TabAtkins> | Philip`: Touche. |
| 17:11 | <zewt> | end⑤ |
| 17:12 | <TabAtkins> | Oh jeez, he was "inspired by Lisp's cdddr". Those are some of the worst functions in the language. |
| 17:12 | <TabAtkins> | Replaced in a ridiculously more readably fashion by NTH and NTH-CDR. |
| 17:13 | Philip` | tries to work out what that's an abbreviation for |
| 17:13 | <Philip`> | cod udder? |
| 17:13 | <TabAtkins> | "contents of decrement register" |
| 17:13 | <Philip`> | Oh, of course |
| 17:13 | <TabAtkins> | Originally the address and decrement registers were used to store Lisp conses (2-tuples), thus the names CAR and CDR. |
| 17:14 | <TabAtkins> | cdddr is just repeated cdr, equivalent in this case to "throw away the first three elements and gimme what's left" |
| 17:15 | <TabAtkins> | And you can mix a and d there. Most impls support up to three. |
| 17:15 | <TabAtkins> | Though I always forget which order they're applied in, which is why I dont' use them. |
| 17:15 | <Philip`> | Stick to palindromes so the order doesn't matter |
| 17:16 | <TabAtkins> | Sensible idea. |
| 17:16 | <Philip`> | though you may have to reorganise your data structures in order to access them in that manner |
| 17:16 | <Philip`> | but that's a minor price to pay |
| 17:16 | <TabAtkins> | While you're at it, you can haiku-optimize the code too. |
| 17:22 | Philip` | wonders if any languages other than Perl have a poetry chapter in their primary reference book |
| 17:29 | <Hixie> | the html spec has poems in it that were specifically written for the spec :-) |
| 17:29 | <jgraham> | Hixie: Being like perl is not a positive quality! |
| 17:30 | <annevk> | http://blog.chromium.org/2011/08/connecting-web-apps-with-web-intents.html is pretty interesting |
| 17:30 | <annevk> | wonder why new bug was filed yet |
| 17:33 | <jgraham> | It would be nice if someone at Google or Mozilla would post to the list about that |
| 17:36 | <annevk> | TabAtkins, is WebKit going to remove background-position-x/y? |
| 17:36 | <TabAtkins> | annevk: I have no idea. Probably not. |
| 17:36 | <annevk> | TabAtkins, so lets add it |
| 17:36 | <annevk> | IE has it too |
| 17:36 | <annevk> | I think we're planning on adding it some day |
| 17:36 | <TabAtkins> | "We" being Opera? |
| 17:37 | <annevk> | yeah |
| 17:37 | <TabAtkins> | Dammit, people. |
| 17:37 | <TabAtkins> | Well, get Elika and Brad to put it into B&B4, then. |
| 17:37 | <nimbu> | who has ops in #css? |
| 17:37 | <annevk> | I should say damnit to you |
| 17:37 | <TabAtkins> | nimbu: I dunno. |
| 17:38 | <TabAtkins> | nimbu: Why? |
| 17:38 | <nimbu> | some idiot just spammed |
| 17:38 | <annevk> | cannot argue against features without removing them first |
| 17:38 | <TabAtkins> | nimbu: Oh jeezus. |
| 17:38 | <nimbu> | yeah wanna kick/ban but cant without ops :/ |
| 17:42 | <nimbu> | really nobody has ops :||| |
| 17:43 | <TabAtkins> | Could someone go to slide 59 of http://www.slideshare.net/jaffathecake/optimising-where-it-hurts-jake-archibald and tell me which browsers are diverging? I cant' tell the blue/greens apart. |
| 17:44 | <nimbu> | the biggest spike is ff3 |
| 17:44 | <nimbu> | the one less than that is IE8 |
| 17:44 | <nimbu> | 3rd is saf 3 |
| 17:44 | <TabAtkins> | Ok, cool. Thanks! |
| 17:44 | <AryehGregor> | MikeSmith, thanks for changing my Bugzilla address (although you're not here and probably won't see this). |
| 17:44 | AryehGregor | wonders why he does that |
| 17:45 | <Philip`> | TabAtkins: Optimising where it hurts Jake Archibald sounds like a surprisingly specific optimisation strategy |
| 17:46 | <TabAtkins> | And yet, effective! |
| 17:46 | <TabAtkins> | It's good to know that scope depth doesn't affect modern browsers. |
| 17:47 | <TabAtkins> | Damn you, unlabelled axises! I have no idea what these numbers are or why I should care. |
| 17:47 | <nimbu> | axes |
| 17:47 | <nimbu> | FTFY |
| 17:48 | <TabAtkins> | NO |
| 17:48 | <AryehGregor> | I hate Bugzilla. Does it really have no way to do mass changes without making me see "Inbox (339)"? |
| 17:49 | <AryehGregor> | Oh, and that's only the ones Gmail decided were important. |
| 17:49 | <AryehGregor> | in:inbox from:w3.org actually matches 411 conversations. |
| 17:49 | <AryehGregor> | Sigh. |
| 17:49 | <jgraham> | AryehGregor: Well probably one can access the db directly somehow. I imagine that to be… foolhardy |
| 17:50 | <AryehGregor> | I was suggesting maybe there could be a feature that was actually intended to let you do that. |
| 17:50 | <AryehGregor> | It seems like a kind of obvious need. |
| 17:50 | <AryehGregor> | Like, you know, so people could receive one e-mail for all bugs they're subscribed to? Would that be so hard? |
| 17:50 | <jgraham> | Also, I must remember the 1s of driving === 1 day of charger on standby quote |
| 17:50 | <AryehGregor> | . . . what? |
| 17:50 | <jgraham> | AryehGregor: See the presentation that TabAtkins couldn't read |
| 17:51 | <Philip`> | There should be a "make these changes without emailing people" slider, and each user should have a "actually email me about changes even if someone says not to" slider, and a user only gets emailed if their slider is higher than the changer's slider |
| 17:51 | <Philip`> | so people who really want to get emails can opt to do so |
| 17:51 | <jgraham> | Or couldn't colour-distinguish, rather |
| 17:51 | <AryehGregor> | Philip`, and people should be allowed to choose arbitrary real numbers for the slider, right, without bounds? |
| 17:51 | <Philip`> | and people who really really want to get them, even if the changer thinks people who merely really want to get them shouldn't get them, can get them |
| 17:52 | <zewt> | what if both people select INF? |
| 17:52 | <Philip`> | That's not a real number |
| 17:54 | Philip` | wonders how to implement UI for an unbounded slider |
| 17:54 | <AryehGregor> | Philip`, map it bijectively to (0, 1), and have it zoom in as you push the slider toward either edge. |
| 17:55 | <AryehGregor> | All we really need is a dense totally ordered set without endpoints here, it doesn't have to be R. |
| 17:55 | <zewt> | of course, experience with "importance" flags on email tells us that some people will set all of their changes, no matter how trivial, to the highest value available |
| 17:55 | <AryehGregor> | Of course, any two dense totally ordered sets without endpoints are isomorphic if they're of the same cardinality. |
| 17:55 | <AryehGregor> | zewt, that's why there is no highest value! |
| 17:56 | <AryehGregor> | It will depend on how long people are willing to type digits, or let the slider scroll. |
| 17:56 | <zewt> | unless you're willing to send 500-megabyte numbers over the wire, there's always a highest value |
| 17:56 | <AryehGregor> | Well, yeah, eventually you're going to hit the server's configured POST size limit. |
| 17:57 | <zewt> | and more to the point (as far as this has a point), the fact that some people will always use a really high value regardless of importance would break the system for everyone else |
| 17:58 | <zewt> | i suppose that suggests needing a weighting value: if people rate the stuff you're doing as "unimportant" and you're marking it "important", your importance values are reduced |
| 17:58 | <Philip`> | I suppose a boringer but maybe more practical solution would be to send a single email on bulk changes, which contains a link you can click that causes the bug tracker to send you an individual email for every bug that was changed, for people who want an email archive of every bug change |
| 17:59 | <AryehGregor> | Or how about we just don't care about people who want an e-mail archive of every bug change? |
| 17:59 | <AryehGregor> | I mean, why would they? |
| 17:59 | <zewt> | ocd |
| 17:59 | <Philip`> | The first implementation could make that link point to a 404, then you only bother implementing the multiple-mail feature if enough people complain about it |
| 17:59 | <AryehGregor> | Okay, hope I received nothing important from w3.org between 1:00 AM and 1:36 AM, because I'm archiving it all without reading it. |
| 18:01 | Philip` | doesn't know why people would want all that mail but is assuming some people must be objecting to changes that would stop that mail, else surely someone would have implement a bulk-change-with-no-mail feature already |
| 18:02 | <AryehGregor> | That seems optimistic. |
| 18:21 | <AryehGregor> | Fun fact: when I want to find out what day of the week something is, I now open Google Calendar in Chrome instead of clicking the GNOME clock in the upper-right corner of my desktop, because it's faster to load Google Calendar. |
| 18:21 | <AryehGregor> | Seriously. |
| 18:22 | <AryehGregor> | If Unity actually takes UI responsiveness seriously, I'm going to switch in 10.10 without looking back. |
| 18:25 | gsnedders | just responded to GNOME3 and Unity by finally moving to a tiling WM |
| 18:25 | <AryehGregor> | Heh. |
| 18:26 | <AryehGregor> | Unity seems okay so far, although I haven't used it much. |
| 18:26 | <AryehGregor> | (I didn't upgrade my desktop to 10.10 because I plan to ditch it soon in favor of my shiny new laptop.) |
| 18:26 | <zewt> | "unity" is a good name for a feature when you want it to be confused with 25 other products |
| 18:26 | <AryehGregor> | (My desktop has an NVIDIA card anyway, so no shiny stuff for me unless I go the proprietary driver route.) |
| 18:27 | <zewt> | (the first thing that "unity" brings to mind for me is VMware's, which is the *worst* thing to bring a comparison to if you're talking about responsiveness) |
| 18:27 | <Philip`> | (Why not go the proprietary driver route?) |
| 18:28 | <AryehGregor> | Because it caused instability and broke stuff when I last used it. |
| 18:28 | <Philip`> | Oh, okay |
| 18:28 | <AryehGregor> | It didn't actually crash the system for me anytime after boot completed, to be fair. |
| 18:28 | <AryehGregor> | But it did cause a lot of crashes on boot. |
| 18:29 | gsnedders | is wondering whether to care about accelerated 3D on his new system |
| 18:29 | <gsnedders> | I mean, I care in Windows for games. But under Linux? Not so much… |
| 18:29 | <AryehGregor> | It looks like Linux is moving to actually use accelerated 3D where available. |
| 18:29 | <AryehGregor> | For video and such. |
| 18:29 | <AryehGregor> | So I'd get it, even if it's just an integrated Intel chip. |
| 18:30 | <AryehGregor> | My new laptop has an Intel card that works well enough for stuff like Unity. |
| 18:30 | <AryehGregor> | I don't plan to try it with games, though. |
| 18:31 | Philip` | has an oldish laptop with Intel graphics as his main platform for some 3D game development |
| 18:31 | <Philip`> | (although admittedly not hugely demanding 3D) |
| 18:32 | <Philip`> | (but it's usable, which is good) |
| 18:32 | jgraham | never worked out how tiling wms were supposed to work when you had windows that really wanted to be certain sizes |
| 18:33 | <Philip`> | ((although sometimes buggy OpenGL code can cause the drivers to get confused so the scrollbars on all my other applications change colour and then everything else gradually change colour until I reboot)) |
| 18:33 | <jgraham> | (after tring Xmonad for a bit some years ago) |
| 18:33 | <gsnedders> | jgraham: Depends if they're floating windows or not |
| 18:33 | <Philip`> | ((which is unideal) |
| 18:33 | <AryehGregor> | Argh, Gmail doesn't believe me when I want to make a filter for "[editing]". It thinks I mean "editing". |
| 18:33 | <gsnedders> | jgraham: Try awesome, it's pretty good out of the box. |
| 18:33 | <gsnedders> | Philip`: hah, awesome |
| 18:33 | <AryehGregor> | Philip`, that sounds like Linux to me! |
| 18:33 | <zewt> | google is starting to ignore +keywords in search :( |
| 18:33 | <zewt> | i've always used a lot of +to force it to stop fuzzing searches to death, and now that's starting to not work |
| 18:34 | <jgraham> | gsnedders: I doubt I would like it more than compiz grid which basically allows me to tile in the simplistic way that my brain can handle and just dump things in random places when tiling is inconvenient |
| 18:34 | <zewt> | building python from source like it's 1996 |
| 18:35 | <jgraham> | zewt: Can't imagine why that was never a hit single |
| 18:35 | <Philip`> | I happened to be searching for some three-character postcode today, and Google was occasionally (depending on search terms) quite insistent that I wanted to find the sixth Harry Potter book |
| 18:36 | <zewt> | google needs a "knows how to search" slider, +so +i +don't +have +to +search +like +this |
| 18:36 | <jgraham> | Philip`: That is evidence that Google knows best, I would say |
| 18:37 | <gsnedders> | jgraham: Well, I have moved to it from grid, which has worked for me. |
| 18:37 | <jgraham> | I mean buckinghamshire isn't that exciting really |
| 18:38 | <Philip`> | It's got some hills, which are exciting, if you're into hills |
| 18:40 | <jgraham> | Those hills even have a wikipedia page: http://en.wikipedia.org/wiki/Category:Hills_of_Buckinghamshire |
| 18:41 | <jgraham> | Clearly the are Notable |
| 18:44 | <gsnedders> | jgraham: Though each hill in Cambridge has an individual Wikipedia page! |
| 18:47 | <jgraham> | Being the only hill in Cambridge is, by definition, quite notable |
| 19:02 | <annevk> | "With the host of problems this country is currently facing, the fact that our president is devoting time to the human process of aging is an affront to Americans everywhere," http://www.theonion.com/articles/obama-turns-50-despite-republican-opposition,21061/ |
| 19:03 | <AryehGregor> | Yeah, that sounds like a typical Onion story. |
| 19:04 | <hober> | annevk: you might want to follow along with https://bugs.webkit.org/show_bug.cgi?id=65535 [CSSOM] |
| 19:06 | <annevk> | I should really make a proposal |
| 19:06 | <annevk> | :( |
| 19:06 | <annevk> | That is, make the CSSOM stuff somewhat more concrete for various properties |
| 19:06 | <hober> | annevk: please do! :) we're all ears |
| 19:07 | <oal> | I'm working on a website where I'm including a website in an iframe, however the iframe tries to forward the user to that website, and navigating away from my site. How'd I handle that? |
| 19:07 | <annevk> | I know, I talked to dino the other week; although he was more like, do it now |
| 19:07 | <hober> | annevk: hahahhaa, yeah :) |
| 19:08 | <AryehGregor> | annevk, PLEASE standardize serialization of CSS values, argh. |
| 19:08 | <AryehGregor> | It's been a huge headache in my editing implementation. |
| 19:08 | <AryehGregor> | And spec. |
| 19:08 | <AryehGregor> | I wish I could just reference some algorithm saying "parse and serialize this CSS value" and be sure that it would come out in some normalized form so I can use string comparison. |
| 19:09 | <AryehGregor> | oal, iframes are allowed to navigate the parent page, so in principle there's not much you can do. Sites are allowed to stop themselves from being framed if they don't want to be. |
| 19:10 | <annevk> | serialization is somewhat done |
| 19:10 | <annevk> | but is going to be moved from CSSOM to individual CSS modules |
| 19:11 | <oal> | AryehGregor: Oh, ok. Thanks :( |
| 19:11 | <AryehGregor> | oal, <iframe sandbox> might theoretically give you this ability in some of the newest browsers, but I don't know if it would do exactly what you want. |
| 19:13 | <oal> | AryehGregor: Thanks, will try |
| 19:27 | <Ms2ger> | Boo, MS |
| 19:30 | <gsnedders> | paul_irish: I was guessing that because he already has those patches in the pull queue somebody else had to do something :) |
| 19:35 | <annevk> | Ms2ger, oh fun |
| 19:35 | <Ms2ger> | They always are |
| 19:35 | <Ms2ger> | And plh |
| 19:37 | <AryehGregor> | "Boo, MS" why? |
| 19:38 | <paul_irish> | gsnedders: yeah i know. :) no worries. rick will stay on 'em for the release cycle |
| 19:44 | <Ms2ger> | W3C politics |
| 19:44 | <AryehGregor> | Oh, DOM Core? |
| 19:44 | <AryehGregor> | To be honest I also think that it's confusing to name specs things like "HTML", "DOM Core", or "DOM Range". |
| 19:44 | <AryehGregor> | In the long term it makes sense given lack of versioning, but it's confusing for the interim period. |
| 19:45 | <AryehGregor> | Also, the W3C likes to version stuff, so you need a version number at least when you make it into an obsolete snapshot that everyone will ignore. |
| 19:45 | <AryehGregor> | The complaint about scope is unreasonable, of course. |
| 19:45 | <AryehGregor> | We'll see how it goes. |
| 19:46 | <Ms2ger> | Want to reply? I'm afraid I'd offend people if I ded :) |
| 19:46 | <Ms2ger> | did* |
| 19:46 | <AryehGregor> | Aren't we all? |
| 19:46 | <annevk> | I replied |
| 19:47 | <AryehGregor> | I was going to leave it to -- yeah, Anne. |
| 19:47 | AryehGregor | doesn't see a reply |
| 19:47 | <annevk> | wait for it |
| 19:47 | <annevk> | I might be okay with adding a 4 |
| 19:47 | <AryehGregor> | It only makes sense to add a 4 for the W3C draft. |
| 19:47 | <Ms2ger> | So, if "DOM Core" is too confusing for Microsoftees... |
| 19:48 | <Ms2ger> | How about "DOM"? |
| 19:48 | <AryehGregor> | Googling "DOM Core" still produces useless stuff as the top results. |
| 19:48 | <annevk> | yeah, it should be DOM at some point |
| 19:48 | <AryehGregor> | "Web DOM Core" used to work fine. |
| 19:48 | <AryehGregor> | "DOM" is a bad name, as is "HTML". |
| 19:48 | <AryehGregor> | Too short and non-distinctive. |
| 19:48 | <AryehGregor> | Maybe if you called it "DOM Standard", "HTML Standard", something like that. |
| 19:49 | <annevk> | sure |
| 19:49 | <Ms2ger> | Technically, sure |
| 19:49 | <Ms2ger> | Politically? I dunno |
| 19:49 | <AryehGregor> | I mean from, say, the WHATWG's perspective. |
| 19:49 | <AryehGregor> | For the W3C, just call it DOM 4 Core. |
| 19:50 | <Ms2ger> | "Document Object Model Level Four Core" |
| 19:50 | <Ms2ger> | (tm)(c) |
| 19:50 | <annevk> | o_O |
| 19:50 | <Ms2ger> | Or Core-XML? |
| 19:53 | <jamesr> | if you have to add a number make it 5 |
| 19:54 | <jamesr> | 5 stuff is way cooler than 4 these days |
| 19:54 | <annevk> | CSS3, DOM4, HTML5 |
| 19:54 | <annevk> | makes perfect sense |
| 19:54 | <jamesr> | SVG2 |
| 19:54 | <annevk> | there you go |
| 19:54 | <jamesr> | MathML6? |
| 19:54 | <Ms2ger> | XHTL1 |
| 19:54 | <annevk> | XML1 |
| 19:54 | <jamesr> | or MathML1 |
| 19:54 | <annevk> | teehee |
| 19:54 | Ms2ger | kicks his "M" key |
| 19:55 | <jamesr> | XML1 SVG2 CSS3 DOM4 HTML5, maybe we can get ECMA to make ECMA-262 6th ed |
| 19:55 | <scor> | can someone explain to me how I can force HTML snippets inside microdata values? it seems all microdata values can only be plain text |
| 19:56 | <annevk> | you cannot have HTML there |
| 19:56 | <scor> | s/can/will |
| 19:56 | <scor> | the HTML is stripped out |
| 19:57 | <scor> | annevk: any idea why such restriction? |
| 19:57 | <scor> | annevk: there can be HTML in the page, but the microdata parsing will strip it out (that's how I understand it) |
| 19:58 | <jgraham> | Well I guess you can if the vocabulary allows it. I mean it will be plain text but you can always specify that the text should be fed into an HTML parser |
| 19:58 | <annevk> | no compelling use case I think scor |
| 19:58 | <jgraham> | Just generic tools won't do that |
| 19:58 | <jgraham> | HTML in microdata seems like a really bad idea |
| 19:59 | <scor> | jgraham: exactly, that's extra processing not specified in the microdata spec |
| 19:59 | <scor> | hum, why is it a bad idea jgraham |
| 19:59 | <scor> | ? |
| 19:59 | <jgraham> | It only allows a very limited range of things, but hugely increases complexity |
| 19:59 | <scor> | annevk: what do you think of these use cases http://www.w3.org/Bugs/Public/show_bug.cgi?id=13468#c7 |
| 20:00 | <scor> | jgraham: complexity in the md parsing algo? or? |
| 20:00 | <annevk> | scor, seems you're already talking to the right people :) |
| 20:00 | <jgraham> | Complexity in the data model, yeah |
| 20:01 | annevk | isn't really sure annotated data is going to work out |
| 20:01 | <jgraham> | Having people implement HTML parsers just to understand microdata is a burden |
| 20:01 | <scor> | annevk: I only wish there were more than one person I was talking too |
| 20:01 | <jgraham> | Having to design APIs that can retrun text-or-markup is hard |
| 20:01 | <annevk> | jgraham, you need an HTML parser to get Microdata |
| 20:01 | <annevk> | jgraham, and APIs can already change from string to object for other reasons |
| 20:01 | <jgraham> | annevk: Not if you consume e.g. JSON |
| 20:02 | <scor> | foolip: you were also involved in that thread re HTML snippets in microdata |
| 20:02 | <annevk> | jgraham, oh right, but you could expose the markup in an objecty way |
| 20:02 | <jgraham> | You infect everyone with the need to process HTML, not just the edges of the system |
| 20:03 | <scor> | jgraham: but you also prevent those who need that HTML from using it |
| 20:03 | <jgraham> | I'm also not sure that being able to say H<sub>2</sub>O is really that useful |
| 20:03 | <scor> | jgraham: H<sub>2</sub>O and H2O are completely different things |
| 20:04 | <jgraham> | What if you want to express the chemical structure rather than just the elements? |
| 20:04 | <jgraham> | HTML isn't any use as a graph language |
| 20:05 | <jgraham> | So "html in microdata" quickly turns into "arbitary markup in microdata" |
| 20:05 | <Ms2ger> | There's a Unicode character for <sub>2</sub>, so that's fine |
| 20:05 | <scor> | Ms2ger: how do I find it out? |
| 20:05 | <annevk> | jgraham, true true, strings just want to become markup typically :) |
| 20:05 | <jgraham> | But at that point I think you might as well just make it opaque to generic processors and allow specific vocabs to define processing |
| 20:06 | <Ms2ger> | www.google.com/search?q=subscript+2+unicode |
| 20:06 | <Ms2ger> | ₂ |
| 20:06 | <jgraham> | I mean it's not like a *generic* processor can do much that is useful with H<sub>2</sub>O as nodes compared to as a literal string |
| 20:07 | <jgraham> | It only really helps when the use caseis "format for display" |
| 20:07 | <jgraham> | (so there we go, I just conclusively proved that HTML is a presentational langauge :) ) |
| 20:08 | Ms2ger | thwaps jgraham |
| 20:08 | <Ms2ger> | Lang-gauge? |
| 20:08 | <scor> | Ms2ger: ok. so I suppose I would have to write this HTML snippet using unicode instead of HTML then, but how user friendly is that? (in comparison to the HTML version) |
| 20:09 | <jgraham> | Ms2ger: See the google webstats. That's like the most misspelt word in HTML :) |
| 20:09 | <jgraham> | scor: Or make up some vocabulary-specific processing |
| 20:11 | <jgraham> | I just can't imagine many UAs that display raw microdata as strings to end users, so the fact that H<sub>2</sub>O doesn't make much sense as a string doesn't matter |
| 20:12 | <jgraham> | The other problem with generic HTML processing is that H<video src="myvideo"><font size=2>2</font>O isn't much use as a chemical formula |
| 20:12 | <jgraham> | so you still need vocab-specific rules |
| 20:12 | <Ms2ger> | I hear that kind of formula is used in astrophysics, though |
| 20:13 | <jgraham> | Are you insane? It's way too precise. |
| 20:13 | <gkellogg> | I don't see why rules for expressing markup should be vocab-specific. As much as I hate @itemvaltype, this is a case where it seems necessary. |
| 20:13 | <scor> | jgraham: so how would these vocab-specific rules work wrt to the md parsing? when would these rules be applied? |
| 20:14 | <jgraham> | scor: In generic UAs everything would be parsed as strings. The angle brackets would just be angle brackets, not anything magic |
| 20:14 | <Ms2ger> | Good point |
| 20:15 | <jgraham> | Although astrophysiscis do love a good <video> to go with a theory |
| 20:15 | <jgraham> | Or at least an <img> containing an "artist's impression" |
| 20:15 | <Ms2ger> | "So all the planets in the solar system kinda follow a pattern, except that we lost one planet" |
| 20:17 | Ms2ger | hopes he summarized the Titius–Bode law correctly |
| 20:18 | annevk | suggests DOM4 |
| 20:22 | <jgraham> | scor: The other reason to dislike markup in microdata is that it gives the whole feature way more surface area for security problems e.g. if someone sets the value of an item to be <script src=evilscript.js></script> you want to be rather sure that you don't execute it when you don't mean to |
| 20:22 | <scor> | jgraham: absolutely, the consumer has to strip out insecure tags |
| 20:23 | <Hixie> | i hope you mean strip out anything not known to be secure |
| 20:23 | <scor> | I'm not saying HTML snippet should be the default, I'm just wondering why there is not option for allowing it |
| 20:23 | <Hixie> | rather than stripping out anything known to be insecure |
| 20:24 | <scor> | yes |
| 20:24 | <Hixie> | just checking :-) |
| 20:24 | <scor> | :) |
| 20:25 | <jgraham> | scor: Which is still more risky than just not allowing markup, and doesn't address the use cases where "risky" things make sense as item values |
| 20:27 | <scor> | jgraham: not allowing markup makes sense as a default |
| 20:29 | <Hixie> | scor: the main reason microdata doesn't have a way to use the DOM Elements as part of the value (as opposed to just the text) is that none of hte use cases it was designed for needed it |
| 20:29 | <Hixie> | (though there are certainly additional reasons to avoid it as jgraham has pointed out) |
| 20:30 | <scor> | Hixie: when was the list of use cases for designing microdata closed? |
| 20:30 | <Hixie> | it wasn't closed |
| 20:30 | <Hixie> | if there are new use cases to consider, the design could be augmented accordingly |
| 20:31 | <Hixie> | it depends on whether the use cases are deemed important enough |
| 20:31 | <scor> | but I guess the use cases provided in the tracker are not important enough |
| 20:31 | <dabaR> | can a header contain sections? |
| 20:32 | <Hixie> | scor: off-hand, i've no idea |
| 20:32 | <scor> | Hixie: so what do you require then? more people asking for these use cases? |
| 20:32 | <Hixie> | i don't think i've ever seen a use case for which it makes sense for the element tree to be part of the value, so far |
| 20:33 | <Hixie> | but it's quite possible that some have been sent to the whatwg list or on bugzilla that i haven't seen yet |
| 20:33 | <Hixie> | the volume of feedback is not important, only the quality matters |
| 20:33 | <Hixie> | dabaR: do you mean a <header> or a <h1> header? |
| 20:34 | <Hixie> | dabaR: if you mean <header>, then yes. Search the spec for "Little Green Guys With Guns" for an example where a <header> contains a <nav> section. |
| 20:35 | <dabaR> | thank you |
| 20:35 | <Hixie> | np |
| 20:37 | <dabaR> | On here: http://www.w3.org/TR/html5/ ? |
| 20:37 | <Hixie> | http://whatwg.org/html |
| 20:38 | <dabaR> | Thanks |
| 20:38 | <dabaR> | Nope |
| 20:38 | <dabaR> | No little green guys with guns |
| 20:38 | <dabaR> | In fact, none of those words appear :) |
| 20:39 | <Hixie> | oh that's the multipage copy |
| 20:39 | <Hixie> | switch to the one-page copy |
| 20:39 | <annevk> | if you load whatwg.org/c they will, but beware of crashing your browser |
| 20:39 | <Hixie> | there's a link at the top |
| 20:39 | <dabaR> | Odd it is there |
| 20:40 | <dabaR> | Oh thats why |
| 20:40 | <dabaR> | Well that's nav inside header |
| 20:40 | <dabaR> | But also section is OK? |
| 20:41 | <Hixie> | <nav>, <Section>, <Article>, and <aside> are all sections |
| 20:42 | <dabaR> | Oh OK |
| 20:42 | <Hixie> | (<section> is just the more generic one) |
| 20:42 | <dabaR> | so <section> can go inside <header> |
| 20:43 | <dabaR> | You know how on government sites there is often a more web-presence-wide header and footer? |
| 20:43 | <dabaR> | Containing links to other sites they have, and also a search bar for all of their sites kinda thingy? |
| 20:44 | <dabaR> | I guess that is a <nav>. Is that a good guess? |
| 20:44 | <Hixie> | yeah, that's basically that <nav> example |
| 20:44 | <dabaR> | OK. Thank you. |
| 20:46 | <dabaR> | But a nav gets its own heading? |
| 20:46 | <dabaR> | Not in this example |
| 20:46 | <dabaR> | A <nav> can have an <h*> |
| 20:50 | karlcow | read the discussion with scor and had difficulties to figure out the benefit of markup. |
| 20:51 | <karlcow> | it seems like alt="" and title="" attributes |
| 20:53 | <scor> | karlcow: ok, let microdata have the same limitations as @alt and @title then |
| 20:54 | <scor> | karlcow: did you check the use cases in the tracker too? |
| 20:54 | <karlcow> | scor: link? |
| 20:54 | <scor> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=13468#c7 |
| 20:54 | <karlcow> | ah yes what I was talking about |
| 20:56 | <karlcow> | years after years of using html, I have grown a feeling that attributes were not the right place for rich markup. |
| 20:56 | <karlcow> | attributes most of the type should be more "operational" than "informational". |
| 20:57 | <karlcow> | (just a feeling) |
| 20:57 | <karlcow> | scor in the bug tracker could you put a full example of code? |
| 20:58 | <jgraham> | karlcow: One could of course design it to not have markup in attributes if one wanted that extension to the data model |
| 20:58 | <jgraham> | But I think the concept is bad even then |
| 20:59 | karlcow | wonders if jgraham meant "to have markup" |
| 21:00 | <karlcow> | I remember all the discussion around description element in RSS feed. |
| 21:00 | <scor> | karlcow: you mean a full microdata example? |
| 21:01 | <scor> | which I would expect to produce HTML? |
| 21:02 | <jgraham> | karlcow: I mean you could define a mechanism where, say, the child nodes of the element with @itemval represented the content |
| 21:04 | <jgraham> | But even in the presence of such a mechanism, allowing markup in values would be more harm than good |
| 21:04 | <karlcow> | scor: yes |
| 21:05 | <karlcow> | jgraham: without counting for what a microdata means inside a microdata |
| 21:06 | <jgraham> | That is an example of the additional complexity that allowing markup in values introduces, yes |
| 21:21 | <scor> | karlcow: http://www.w3.org/Bugs/Public/show_bug.cgi?id=13468#c12 |
| 21:24 | <scor> | karlcow: what was the conclusion re. "description element in RSS feed" |
| 21:25 | <karlcow> | scor: for RSS, dirty hacks, including escaping markup. |
| 21:25 | <karlcow> | Atom introduced a content type |
| 21:26 | <karlcow> | <content type="xhtml">… |
| 21:30 | <Ms2ger> | The bugs line at http://www.whatwg.org/issues/data.html?period=1 does look interesting |
| 22:37 | <jarek> | Hello |
| 22:37 | <jarek> | is there any performance difference between element.setAttribute('blah1', 'blah2') and element.dataset.blah1 = 'blah2'? |
| 22:39 | <jarek> | I'm currently using custom attributes all over the place and I'm wondering whether switching to HTML5 datasets would make sense |
| 22:39 | <Philip`> | Seems very unlikely that there would be any noticeable difference |
| 22:39 | <Philip`> | since dataset is just an API to get/set content attributes |
| 22:40 | <Philip`> | so it's doing the same amount of DOM manipulation either way |
| 22:43 | <jarek> | perhaps setting an attribute directly on node (e.g. element.blah1 = 'blah2') could be any faster? |
| 22:46 | <Hixie> | depends what you're trying to do |
| 22:46 | <The_8472> | jarek, you shouldn't set custom attributes directly on dom nodes anyway. it's bad practice for various reasons. |
| 22:46 | <Hixie> | are you doing this in a tight loop? |
| 22:46 | <Hixie> | if you're not, the performance aspects of this are probably irrelevant anyway |
| 22:47 | <AryehGregor> | The_8472, why not? That's what data-* is meant for. |
| 22:48 | <The_8472> | AryehGregor, i mean this: <jarek> perhaps setting an attribute directly on node (e.g. element.blah1 = 'blah2') |
| 22:48 | <jarek> | The_8472: I'm using custom XML namespace inside XHTML5 document, so custom attributes should be fully valid in my case |
| 22:48 | <The_8472> | attributes may have been the wrong word |
| 22:48 | <The_8472> | object properties? |
| 22:50 | <Hixie> | the spec uses the terms IDL attributes and content attributes |
| 22:51 | <Hixie> | though in your case you're talking abotu object properties, since they're not IDL attributes either |
| 22:51 | <Hixie> | they might conflict with IDL attributes in the future though, so it's not good practice :-) |
| 22:51 | <jarek_> | oops, I got disconnected |
| 22:52 | <jarek_> | yeah, I already learned that extending core objects is considered to be a bad practice: http://perfectionkills.com/whats-wrong-with-extending-the-dom/ |
| 22:52 | <jarek_> | but I want to do this anyway :) |
| 22:52 | <The_8472> | good good |
| 22:53 | <jarek_> | btw, why it's possible to use element.dataset on non-html elements? |
| 22:54 | <jarek_> | while it's not possible to use e.g. element.style |
| 22:54 | <Hixie> | i thought it was the other way around |
| 22:55 | <jarek_> | let me double check... |
| 22:55 | <jarek_> | yum, |
| 22:55 | <jarek_> | element.style always returns null if element is from custom XML namespace |
| 22:56 | <jarek_> | s/yum/yup |
| 22:56 | <Hixie> | jarek_: and dataset works? |
| 22:56 | <jarek_> | element.dataset works fine no matter whether element is from custom namespace or not |
| 22:56 | <Hixie> | weird |
| 22:57 | <Hixie> | sounds like a bug |
| 22:57 | <jarek_> | no, it's a feature |
| 22:57 | <jarek_> | please don't fix it |
| 22:57 | <The_8472> | it's a bug |
| 22:57 | <Hixie> | what if another namespace defines attributes that start with data-? |
| 22:57 | <heycam> | move it to Element and into DOM Core |
| 22:58 | <heycam> | if we're going to have global id and class attributes, we should do the same for data- |
| 22:58 | <Hixie> | i still don't think we should have global id and class :-) |
| 22:58 | <The_8472> | id is namespaced in xml documents, isn't it? |
| 23:00 | <heycam> | Hixie, if dataset doesn't move to Element, would you be able to stick it in a separate interface? beacuse I'm pretty sure it would be useful to use on SVG elements too, and in that case we'd want to have `SVGElement implements DataSetThingos` |
| 23:00 | <Hixie> | i doubt i'll win the id/class battle |
| 23:00 | <Hixie> | in which case i'll let anne take data-* to dom core |
| 23:01 | <heycam> | ok |
| 23:01 | <Hixie> | i'm not sure how anne is planning on defining content attributes though |
| 23:01 | heycam | ducks out for a bit |
| 23:03 | <AryehGregor> | Man, Scalia writes awesome opinions. http://www.supremecourt.gov/opinions/10pdf/08-1448.pdf |
| 23:04 | <AryehGregor> | Cogent and extremely sarcastic. |
| 23:04 | <AryehGregor> | "One study, for example, found that children who had just finished |
| 23:04 | <AryehGregor> | playing violent video games were more likely to fill in the blank letter in “explo_e” with a “d” (so that it reads “explode”) than with an “r” (“explore”). App. 496, 506 (internal quotation marks omitted). The prevention of this phenomenon, which might have been anticipated with common sense, is not a compelling state interest." |
| 23:04 | <Hixie> | hah |
| 23:07 | <Hixie> | i also like "Since California has declined to restrict those |
| 23:07 | <Hixie> | other media, e.g., Saturday morning cartoons, its video-game regulation is wildly underinclusive, raising serious doubts about whether |
| 23:07 | <Hixie> | the State is pursuing the interest it invokes or is instead disfavoring |
| 23:07 | <Hixie> | a particular speaker or viewpoint" |
| 23:07 | <Hixie> | man this is full of win |
| 23:08 | <AryehGregor> | Scalia's decisions are always like that. |
| 23:08 | <AryehGregor> | Roberts is good too. |
| 23:10 | <Hixie> | " There is no contention that any of the |
| 23:10 | <Hixie> | virtual characters depicted in the imaginative videos at issue here are |
| 23:10 | <Hixie> | criminally liabl" |
| 23:10 | <Hixie> | that's good to know |
| 23:13 | <Hixie> | lol they cite grimm |
| 23:14 | <Hixie> | by page number no less |
| 23:15 | <The_8472> | writing their opinions probably provides them some fun after dealing with aneurysm-inducing lawyer reasoning. |
| 23:15 | <AryehGregor> | I think Alito's concurrence makes sense too. |
| 23:15 | <AryehGregor> | But then, I think court interpretations of the First Amendment can be kind of extreme sometimes. |
| 23:15 | AryehGregor | gets to the dissent |
| 23:16 | The_8472 | points at corporations having 1st amendment rights in the US. |
| 23:17 | <Hixie> | page 11 has a sentence that just says "Who knows?" |
| 23:17 | <AryehGregor> | Why shouldn't they have them? They're basically just collections of people. Your right to free speech shouldn't be affected by who you're being paid by. |
| 23:18 | <Hixie> | AryehGregor: if we could imprison the collection of people, that might have more weight |
| 23:18 | <AryehGregor> | You can imprison individual people who individually did things wrong. Corporations don't have volition, they only have assets. They're controlled entirely by individual people. |
| 23:18 | <The_8472> | http://www.nytimes.com/2010/01/22/us/politics/22scotus.html |
| 23:18 | <The_8472> | therefore corporations shouldn't have the same rights as people. |
| 23:19 | <AryehGregor> | They don't. The people they employ do. |
| 23:19 | <The_8472> | see... that's how it should be, not how it is. |
| 23:19 | <AryehGregor> | Or rather, the rights that have are corollaries of the rights of their employees. |
| 23:20 | <AryehGregor> | People who run ads or lobby Congress on behalf of a corporation aren't any less entitled to do so than people who do so on behalf of individuals or themselves. |
| 23:20 | <AryehGregor> | Just because you were told to do it by a manager in exchange for a paycheck from a corporate bank account doesn't mean you have any less right to do it. |
| 23:21 | AryehGregor | always gets bored of reading court decisions by the time he gets to the dissents |
| 23:21 | <Hixie> | i don't mind them having rights if they have responsibilities to go with them |
| 23:21 | <The_8472> | just because you work for a company does not mean they are entitled to represent your political opinion |
| 23:22 | <AryehGregor> | Of course they aren't. But they can pay you to exercise your right of political speech on their behalf. That's all corporations can ever do -- they can't talk or write anything *themselves*. They're legal fictions. Any speech is necessarily speech by some person. |
| 23:22 | <wilhelm> | The_8472: They usually don't. They represent the political opinion of the owners, if any. |
| 23:23 | <AryehGregor> | Hixie, constitutional rights in America are generally viewed as unalienable, not as being granted in exchange for fulfilling responsibilities. |
| 23:23 | <AryehGregor> | Convicted murderers spending a life sentence in prison still have the same right to free speech as anyone. |
| 23:24 | <Hixie> | AryehGregor: i'm not saying they're directly linked, just that i would be happy with one if the other existed. |
| 23:25 | <AryehGregor> | Hixie, well, corporations basically only have assets. When it comes to the disposition of those assets, they have similar responsibilities to individuals, often much greater responsibilities (depending on the type of corporation). That is, they can be held fiscally liable for breaking the law. But if it's anything non-monetary, you can't really take actions against a corporation, you have to take action against the individuals responsible. |
| 23:25 | <Hixie> | AryehGregor: e.g. if someone were to kill all the fish in a sea due to reckless negligence while doing something in that sea, they could be expected to suffer consequences, not be immediately invited to continue doing that thing in the same sea and others, as well as spending large sums of money influencing political decisions that control those consequences |
| 23:26 | The_8472 | points at the wallstreet. diffusion of responsibility to the max. they rake in millions and billions by gambling with other people's money (and losing) and get away with a slap on the wrist (to individuals) and practically no regulation at all (towards the companies) |
| 23:26 | <The_8472> | they cause far more damage than your lowly criminals that may get many-year sentences... |
| 23:26 | <Hixie> | AryehGregor: i'd be fine with it if the individuals were held responsible too |
| 23:27 | <AryehGregor> | Hixie, that's orthogonal to the question of corporations. "Corporation" doesn't mean "big business", it means an entity that can legally act as a person in some respects. Small charities and whatever are also often incorporated. |
| 23:27 | <Hixie> | AryehGregor: unfortunately, in the US (and many other places), corporations act as responsibility shields |
| 23:27 | <wilhelm> | Indeed. “because the corporation is legally considered the "person," individual shareholders are not legally responsible for the corporation's debts and damages beyond their investment in the corporation” |
| 23:27 | <Hixie> | AryehGregor: for both the matter of responsibility and the matter of free speech, the problem is proportionally bigger as the company gets bigger. |
| 23:28 | <AryehGregor> | It's not the corporation that's shielding anything, it's the money and influence. Corporations are designed to be a good vehicle for concentrating arbitrarily large amounts of money, so it just so happens that any organization of note is going to be incorporated. |
| 23:28 | <AryehGregor> | That's not the fault of the incorporation. |
| 23:28 | <The_8472> | Hixie, it's not just corporations. political parties or bureaucratic machineries are exactly the same. it's all about diffusing responsibility so much that in the end nobody is responsible at all. or that just some symbolic heads will roll even after biggest fuckups that you can imagine |
| 23:28 | <AryehGregor> | If incorporation didn't exist, you'd see the same thing with large private businesses. |
| 23:29 | <Hixie> | The_8472: yeah, responsibility diffusion really is the problem. the rights get concentrated, but the responsibilities diffused. |
| 23:29 | <wilhelm> | The shielding of individual shareholders is a wonderful tool to encourage the establishment of new businesses, and awfully scary when said businesses grow bigger and more powerful than small nation-states. |
| 23:29 | <AryehGregor> | It's simpler than that, it's just that large businesses are powerful enough to have bargaining power with the government. |
| 23:30 | <AryehGregor> | In many cases they can credibly threaten the government. |
| 23:30 | <AryehGregor> | For instance, a big business could threaten to leave a particular state if it's taxed too heavily. |
| 23:30 | <wilhelm> | Or threaten its employees, customers or the general populace. |
| 23:30 | <AryehGregor> | That could cause thousands of people to lose their jobs, which would create a big backlash by the general public against the politicians in power. |
| 23:31 | <The_8472> | shareholders are another issue when they get dividents. they're syphoning off money from other people's work, even when the company has reached the point of being self-sustaining. so instead of accumulating money within the company to reinvest it gets drained, thus potentially slowing innovation just to fill a few private people's pockets. |
| 23:31 | <The_8472> | i know that shares are an important tool to get investments... but in some cases it's just counterproductive |
| 23:31 | <The_8472> | the value of the shares themselves should be sufficient |
| 23:31 | <AryehGregor> | The_8472, well, yes, that's how capitalism works. In a capitalist society, it's very easy to start businesses, because investors have the promise of exorbitant returns. |
| 23:32 | <AryehGregor> | Most large companies don't pay dividends, though. |
| 23:32 | <AryehGregor> | And if they do, it's only a very small fraction of profit, like one or two percent. |
| 23:32 | <AryehGregor> | So it's not a big deal in the scheme of things. |
| 23:32 | <wilhelm> | The growth of the value of the company is usually sufficient. |
| 23:32 | <The_8472> | unless someone buys them up, changes the rules and syphons off money, seen that all too often |
| 23:33 | <AryehGregor> | Right, and dividends slow down the company's growth sometimes. |
| 23:33 | <AryehGregor> | The_8472, hostile takeovers of that sort are very rare. A publicly-traded company almost always has far higher market capitalization than assets. Usually when you have that sort of systematic buy-out, it's an attempt to merge one company into a larger one or such. |
| 23:34 | <AryehGregor> | But it's hard to pull off, because trying it causes the stock price to shoot up. |
| 23:34 | <The_8472> | <AryehGregor> It's simpler than that, it's just that large businesses are powerful enough to have bargaining power with the government. <- that's one aspect, but not the only one. just look at all the externalties that companies cause and that the taxpayer has to wipe up after them |
| 23:34 | <AryehGregor> | Anyway, if you do buy a lot of stock in the company, it's rarely in your interest to try grabbing the assets and running. You want to hold onto the stock and let it go up in price, then sell it. That almost always provides better returns. |
| 23:35 | <AryehGregor> | Except if the company's stock price has really crashed, then it's sometimes worthwhile to buy it just for the sake of liquidating it. |
| 23:35 | <The_8472> | oh, you don't grab the assets. you slowly bleed money from them, then sell it again while it's still somewhat performing |
| 23:35 | <AryehGregor> | The_8472, practically any economic activity causes externalities, both positive and negative. Big businesses can bring a lot of positive economic effects, like greatly reduced costs through economy of scale. |
| 23:36 | <The_8472> | i'm talking about negative externalties, obviously. |
| 23:36 | <AryehGregor> | They can devote a much larger amount of their budget to R&D and innovation, etc. |
| 23:36 | <AryehGregor> | Well, yes, but you can't fairly complain about the negative externalities unless you have evidence that they outweigh the positive externalities. |
| 23:36 | <The_8472> | which sometimes exceed even the profit that the companies themselves make (look at nuclear fuel processing and waste storage cycles) |
| 23:37 | <The_8472> | it's basically cheap power for the economy in exchange for the govt funding the whole infrastructure and waste management for millions of years to come. |
| 23:39 | <wilhelm> | AryehGregor: Sure you can. I spend all day searching for an complaining about the few negative issues in otherweise good software. I don't see why the management of this planet's productive resources should be exempt from any criticism. Quite the contrary. |
| 23:43 | <The_8472> | you could go even further... we're putting some of those negative externalties on those who don't see anything of the positive ones (i.e. everything that isn't human) |