| 00:00 | <llrcombs> | basically, replicating what FireFox does for existing videos and non-WebM files |
| 00:00 | <llrcombs> | AGAIN? |
| 00:00 | <llrcombs> | good afternoon, netsplit |
| 00:03 | <llrcombs> | and back again |
| 06:33 | <annevk> | why did undoManager move to Node? |
| 06:33 | <annevk> | I thought the plan was to make Node smaller |
| 06:56 | <annevk> | AryehGregor, would you be okay with renaming the DOM Range component to Selection and either adding a new DOM Range component and moving range bugs there or moving them to DOM Core? |
| 08:16 | <annevk> | Element.replaceElement() |
| 08:16 | <annevk> | Does that seem somewhat appropriate? I think a shorter name is somewhat risky and not in the same style as Node.cloneNode() |
| 08:21 | <zcorpan> | replaceElement sounds like you should pass in an element node |
| 08:22 | <zcorpan> | i dunno, i think rename is fine |
| 08:23 | <annevk> | It is much more than renaming though |
| 08:24 | <annevk> | But yeah, I guess it is clearer |
| 08:29 | <annevk> | zcorpan, you don't think rename() would conflict with custom properties people have out there? |
| 08:30 | <annevk> | zcorpan, that was a problem anyway with some of the forms extensions |
| 08:48 | <asmodai> | hahaha |
| 08:48 | <asmodai> | annevk: http://tweakers.net/nieuws/76475/firefox-vertrouwt-digid-toch-na-verzoek-nederlandse-overheid.html |
| 08:52 | <annevk> | ja zag het |
| 08:53 | <annevk> | mooie overheid hebben we |
| 08:53 | <asmodai> | eh , wth |
| 08:53 | <asmodai> | firefox 7 beta now gives me sec_error_ocsp_unknown_cert for bol.com's https. Worked yesterday. o_O *digs into certificates* |
| 08:56 | <asmodai> | Mmm, Chrome 14 works |
| 09:03 | <annevk> | still one month for Gears of War 3 |
| 09:03 | <annevk> | it better be awesome |
| 09:05 | <asmodai> | That's odd, I don't have the option selected that specifies to treat a certificate as invalid if the OCSP cannot be reached. |
| 09:05 | <asmodai> | Yet it does fail on FF 7, unless I am misunderstanding what is going wrong here. |
| 09:05 | <asmodai> | Wonder if I trip a bug here or not. |
| 09:06 | <asmodai> | annevk: No idea tbh. Just finished DXHR last night, nice story and pretty good game overall. |
| 09:07 | <asmodai> | And yet Aurora doesn't fail. Eh? |
| 09:08 | annevk | is not a big fan of RPGs |
| 09:09 | <asmodai> | annevk: You're more the FPSer? |
| 09:11 | <asmodai> | OK, now I totally don't get it anymore. I temporarily disable OCSP checking. Log into that site, log out again. Make sure I am back on the normal http homepage. Enable OCSP checks again and log in again. No complaints this time. OCSP checks get cached or something? |
| 09:13 | <annevk> | platform and some FPS games I guess |
| 09:14 | <asmodai> | annevk: From what I heard so far, and I deal with a lot of gamedevs, it seems to be quite ok, worthy successor in the series. |
| 09:17 | <annevk> | though nowadays I only play games every month or so |
| 09:17 | <annevk> | and occasionally a couple of times a year a little longer if there's a good game |
| 09:17 | <zcorpan> | annevk: possible. renameElement is fine also |
| 09:17 | <annevk> | oops, that was double :) |
| 09:19 | <asmodai> | annevk: Given I am trying to move into the gamedev industry, I play too much ;) |
| 09:32 | <Ms2ger> | var is_khtml = (navigator.vendor == 'KDE' || ( document.childNodes && !document.all && !navigator.taintEnabled )); |
| 10:01 | <annevk> | so setAttribute can simply overwrite existing xmlns attributes and create xmlns attributes in the null namespace |
| 10:02 | <annevk> | I guess it does not matter that much |
| 10:03 | <annevk> | xmlns attributes should only matter during parsing |
| 10:03 | <annevk> | but the lookup namespace algorithms still use them |
| 10:03 | <annevk> | hmm |
| 11:03 | <hsivonen> | I noticed that the wiki says nofollow is allowed on <link> but the spec says otherwise |
| 11:08 | <karlcow> | annevk: does the Element.replaceElement(), renameElement, etc only change the name? |
| 11:08 | <karlcow> | not a clone. identical copy (two of them after the operation), |
| 11:09 | <karlcow> | a mask? change how it looks but inside still the same. not sure. |
| 11:09 | <karlcow> | is it a mutation? |
| 11:09 | karlcow | wonders if there is a Marvel hero for illustrating it. Hulk? |
| 11:09 | <karlcow> | Element.hulkElement() - green mutation |
| 11:10 | <Ms2ger> | It would create a new element |
| 11:10 | <annevk> | you get more or less what cloneNode does |
| 11:13 | <annevk> | but the details are not yet entirely clear |
| 11:14 | <annevk> | you create a new element, copy attributes and child nodes from the old element and append them to the new element |
| 11:14 | <annevk> | then invoke replaceChild on the old element's parent |
| 11:14 | <annevk> | open questions are what happens with event handlers, event listeners, and JS properties |
| 11:15 | <annevk> | and maybe what happens with cloning steps; but I think that would require a new concept |
| 11:15 | <Ms2ger> | JS properties should go away, no? |
| 11:16 | <annevk> | I think so, yes |
| 11:17 | <annevk> | Personally I favor the most simple approach |
| 11:17 | <annevk> | copying event listeners could work, but then people might expect event handlers to work too, which won't always be the case... |
| 11:20 | zcorpan | thinks it should be similar to cloneNode |
| 11:23 | <smaug____> | in general event listeners would be broken |
| 11:23 | <smaug____> | since the type of 'this' would change |
| 11:23 | <smaug____> | and 'this' wouldn't suddenly have some needed properties |
| 11:24 | <smaug____> | so, copying event listeners would be very error-prone |
| 11:24 | <zcorpan> | sounds like we have an idea about what it should do |
| 11:24 | <zcorpan> | so here's another case. what if the element has no parent? |
| 11:25 | <zcorpan> | what if it's the root element? |
| 11:25 | <smaug____> | root element should work just fine |
| 11:25 | <zcorpan> | AryehGregor's suggested algorithm wouldn't work for root |
| 11:26 | <smaug____> | cloneAndRename (or whatever to call it) should return the created node, so element without parent should be ok |
| 11:26 | <smaug____> | hmm, what could be a good name. |
| 11:26 | <smaug____> | it isn't a clone, it isn't rename |
| 11:27 | <smaug____> | morph? |
| 11:29 | <Ms2ger> | replaceWithANewElementAndTransferAttributesAndStuff(DOMString localName) |
| 11:29 | <zcorpan> | that's more like it |
| 11:31 | <jgraham> | Can we have another method that moves the element to a random location in the DOM? |
| 11:31 | <jgraham> | We could call it Element.witnessProtection() |
| 11:32 | <jgraham> | *renames and moves |
| 11:32 | <jgraham> | Sigh |
| 11:33 | <annevk> | zcorpan, replaceChild works for root |
| 11:34 | <annevk> | jgraham, all mutation methods already move by default |
| 11:34 | <Ms2ger> | But not to a random location |
| 11:35 | <annevk> | smaug____, I think renameElement(name) is clearer than morphElement(name) |
| 11:35 | <annevk> | smaug____, renaming an element is just a complex operation :) |
| 11:35 | <smaug____> | renameElement isn't clear at all, since it doesn't rename ;) |
| 11:35 | <Ms2ger> | What he said |
| 11:38 | <annevk> | so replaceElement(name)? |
| 11:38 | <annevk> | zcorpan did not like that because he thought it took a node rather than a string |
| 11:40 | <zcorpan> | yeah i think it's too similar to replaceChild |
| 11:40 | <smaug____> | I agree |
| 11:40 | <jgraham> | I think the problem is trying to wrap two logical operations into a single DOM call |
| 11:41 | <zcorpan> | three |
| 11:41 | <jgraham> | three? |
| 11:41 | <jgraham> | Well clone, rename, replace I guess |
| 11:42 | <smaug____> | create ,copy stuff, replace |
| 11:42 | <zcorpan> | 1. make a new element. 2. replace with old element. 3. transfer children and attributes |
| 11:42 | <zcorpan> | maybe that's 4 |
| 11:42 | <jgraham> | It would be nice if cloneNode could just be extended with an extra argument to specify what to replace |
| 11:42 | <zcorpan> | see we can't even agree on the order :P |
| 11:42 | <annevk> | jgraham, how does that help? |
| 11:44 | <jgraham> | annevk: Well if you can get a reference to the subtree that you want to insert then you can use normal DOM operations to insert the new subtree in place of the old one |
| 11:45 | <zcorpan> | renameElement doesn't make a clone |
| 11:45 | <annevk> | it does seem similar to file move |
| 11:45 | <jgraham> | Well I guess it is different for the children |
| 11:45 | <annevk> | move a b |
| 11:45 | <annevk> | a.move(b) |
| 11:45 | <annevk> | but moveElement is all kind of wrong somehow |
| 11:46 | <jgraham> | Well it's not really that similar to a move since nothing moves |
| 11:46 | <jgraham> | That's really the distinguishing feature of a move :) |
| 11:48 | <karlcow> | it looks like the teletransportation in The Fly |
| 11:49 | <karlcow> | Element.jeffGoldblumed |
| 11:50 | <zcorpan> | maybe we should call it dunnoLOL() |
| 11:52 | <karlcow> | hmm Element.lolCat() |
| 11:54 | <annevk> | I still think rename is best |
| 11:55 | <annevk> | it's not actually what it does, but it matches what people expect |
| 11:57 | <jgraham> | substitute? |
| 11:57 | <zcorpan> | renameElement is my vote |
| 11:58 | <Ms2ger> | Mm, I think we should paint the bike shed "substitute" as well |
| 11:58 | <annevk> | substitute is like replace, as confusing |
| 12:08 | <smaug____> | how does rename match what people expect |
| 12:09 | <jgraham> | I think that the differnce between rename and substitute is that it sounds like rename preserves old === new whereas it doesn't sound like substitute does |
| 12:10 | <smaug____> | substitute is significantly better than rename |
| 12:13 | <zcorpan> | i like substituteElement better than replaceElement |
| 12:14 | <hsivonen> | when trying to find a bug in the JS of a product that's still being sold, I have a hard time deciding if finding Dreamweaver boilerplate for Netscape 4.x should make me facepalm or think "compatibility FTW" |
| 12:23 | <annevk> | smaug____, they expect the element to be renamed and prolly don't care too much about the side effects in the usual scenario |
| 12:23 | <annevk> | smaug____, such as renaming <strong> to <b> or some such |
| 12:24 | <hsivonen> | ...and code for IE 5.0, too... |
| 12:24 | <smaug____> | the "doesn't copy event listeners" is such a major side effect, that I assume people will care about it |
| 12:25 | <smaug____> | renameElement sounds like the element is just renamed, but all the properties and listeners are kept |
| 12:29 | <annevk> | in the typical case no event listeners are registered though |
| 12:30 | <annevk> | well the scenarios sketched so far |
| 12:30 | <zcorpan> | changeElement()? |
| 12:30 | <Ms2ger> | doSomething() |
| 12:32 | <smaug____> | what is wrong with substituteElement() ? |
| 12:32 | <smaug____> | (or morph, but others don't seem to like it) |
| 12:33 | <smaug____> | (I think substitute is better though) |
| 12:34 | <annevk> | substitute sounds like it takes a node |
| 12:34 | <annevk> | it's also hard to type |
| 12:37 | <zcorpan> | yeah with renameElement it's clear what the expected argument is |
| 12:40 | <zcorpan> | annevk: now spec this out! :) |
| 12:41 | <smaug____> | but renameElement is still wrong, since it doesn't rename the element |
| 12:41 | smaug____ | is being annoying again. Sorry :) |
| 12:43 | <annevk> | I think I will write a WHATWG Weekly instead |
| 12:44 | <jgraham> | I think the fact that rename breaks the implied invariant is worse than the fact that it isn't obvious what arguments substitute takes |
| 12:44 | <jgraham> | Alsom if we extend it to also allow different attributes on the new element then it doesn't make sense to use rename |
| 12:45 | <jgraham> | (possible use case: editor that wants to replace a <span> with an <a href> but preserve children) |
| 12:45 | <annevk> | renameElement followed by setAttribute() |
| 12:46 | <jgraham> | Well sure |
| 12:46 | <jgraham> | But this whole thing is a convenience API |
| 12:47 | <jgraham> | It seems strange to base your decision of what is convenient on your preferred name |
| 12:48 | <annevk> | create/copy/replace is complicated enough as it is, I would not want to overload it more |
| 12:48 | <zcorpan> | the use case also involved restoring selection |
| 12:49 | <jgraham> | I don't think that adding/removing/changing some attribute values is particularly complex compared to the other operations it does |
| 12:50 | <jgraham> | Especially if we get the convenience method for creating elements (I forget the name) that has the syntax whatever(name, {attr_name, attr_value}) |
| 12:50 | <jgraham> | er |
| 12:50 | <jgraham> | whatever(name, {attr_name: attr_value}) |
| 12:50 | <jgraham> | It would be strange not to support the same syntax |
| 12:53 | <zcorpan> | maybe we should solve teh restore selection use case separately |
| 12:54 | <zcorpan> | and wait with convenience apis |
| 12:54 | <zcorpan> | restore selection is the only thing that was impossible today |
| 13:02 | <annevk> | whoa http://upload.wikimedia.org/wikipedia/commons/b/b7/Population_curve.svg |
| 13:02 | <hsivonen> | MikeSmith: I redeployed V.nu with the new code from the repos. Thanks |
| 13:02 | <annevk> | zcorpan, might be worth looking into |
| 13:03 | <hsivonen> | MikeSmith: I'm not ready to express an opinion about <menuitem>, yet, because I haven't discussed it enough with the people who want it |
| 13:27 | <hsivonen> | what's the most current location of the spec that defines createContextualFragment? |
| 13:27 | <hsivonen> | Ms2ger: ^ |
| 13:28 | <hsivonen> | http://html5.org/specs/dom-parsing.html this? |
| 13:29 | <annevk> | yes |
| 13:30 | <hsivonen> | annevk: thanks |
| 13:34 | <annevk> | http://blog.whatwg.org/weekly-rich-text-editing |
| 13:58 | <annevk> | hsivonen, seems to be source order in Opera/WebKit |
| 13:59 | <annevk> | I wish we could turn it into a dictionary though and serialize in alphabetical order |
| 14:09 | <hsivonen> | annevk: yeah, I noticed that Opera and WebKit were less surprising |
| 14:10 | <hsivonen> | annevk: but I think Firefox and IE are enough to make my point that source order can't be required for Web compat |
| 14:10 | <hsivonen> | except for the <embed> element! |
| 14:10 | <annevk> | sure, but can be required for sanity :) |
| 14:11 | <hsivonen> | annevk: depends on what nice things Gecko and Trident get from what they do now |
| 14:11 | <hsivonen> | which is why I CCed bz |
| 14:11 | <hsivonen> | I believe bz knows *why* CSS-mapped legacy presentational hint attributes sort differently in Gecko |
| 14:12 | <hsivonen> | my hypothesis is that the ordering in IE simply depends on the implementation details of a hashtable somewhere in IE internals |
| 14:18 | <hsivonen> | does ubuntu have downgrade manager now or what's the thing with down-pointing arrow in https://launchpadlibrarian.net/78619733/Launcher%2520washout.png |
| 14:19 | <zcorpan> | downloads maybe? |
| 14:25 | <zcorpan> | hsivonen: does the attribute order for <font> have web compat implications? |
| 14:29 | <hsivonen> | zcorpan: I'm not aware of any |
| 14:30 | <hsivonen> | zcorpan: since browsers have different orders, I'm guessing it doesn't have Web compat implications |
| 14:39 | <annevk> | can ranges ever be outside a document? |
| 14:47 | <annevk> | smaug____, could you comment on http://www.w3.org/Bugs/Public/show_bug.cgi?id=12205#c21 maybe? |
| 14:47 | <annevk> | smaug____, how likely is it that Gecko would implement it? |
| 14:47 | <annevk> | I can probably convince someone at Opera to do it if Gecko has done it |
| 14:51 | <smaug____> | looking |
| 15:07 | <hsivonen> | btw, now that Zakim supports SIP, has anyone figured out what the least troublesome way of getting SIP access is? |
| 15:07 | <hsivonen> | the more I read about SIP, the less I appreciate it, though |
| 15:07 | <hsivonen> | and the more it seems that Skype deserves to win |
| 15:26 | <AryehGregor> | annevk, sounds good to me. |
| 15:27 | <AryehGregor> | TabAtkins, DOM operations never trigger repaints. Repaints only happen when the event loop spins. Except in Opera, which is broken. |
| 15:27 | <annevk> | "broken" |
| 15:28 | <jgraham> | Hmm, why is that broken? |
| 15:28 | <AryehGregor> | Because it will display transient states to the user. |
| 15:28 | <AryehGregor> | Also, it's different from every other browser, which is broken more or less by default. |
| 15:28 | <AryehGregor> | I saw someone from Opera say it caused site compat issues, IIRC. |
| 15:29 | <jgraham> | Displaying transient states to the user is normal |
| 15:29 | <jgraham> | Think progressive loading |
| 15:29 | <timeless> | asmodai: i'm fairly certain ocsp checks are cached |
| 15:29 | <asmodai> | timeless: That would explain then |
| 15:29 | <AryehGregor> | Depends on your definition of "transient". |
| 15:29 | <jgraham> | It can cause slowness if you don't carefully control how often you repaint ofc |
| 15:29 | <annevk> | it causes uglyness indeed mostly because other browsers have a silly blocking implementation |
| 15:30 | <annevk> | and authors authored against the silly implementations |
| 15:30 | <AryehGregor> | In Opera's implementation, there's no way to have the browser not display a half-mutated DOM. It will display some snapshot at random depending on when it decides to repaint. |
| 15:31 | <jgraham> | Roughly the same as every browser and loading off the network really |
| 15:31 | <AryehGregor> | As opposed to other browsers, which will never display anything unless the event loop spins, so a bunch of consecutive operations will always be run atomically from the user's perspective. |
| 15:31 | <AryehGregor> | Loading off the network is a special case. |
| 15:31 | <jgraham> | But there's no intrinsic reason it has to be |
| 15:31 | <jgraham> | If you have a very slow running script, displaying something is better than displaying nothing |
| 15:32 | <AryehGregor> | Yes, if it's slow-running enough, maybe. |
| 15:32 | <AryehGregor> | Particularly if you try to be intelligent about which point to display. |
| 15:32 | <jgraham> | That is often the case on lower end devices |
| 15:35 | <zcorpan> | hmm. setting textarea.value = '\r\n' it gets normalized to '\n' in webkit and gecko |
| 15:35 | <zcorpan> | but spec doesn't say so afaict |
| 15:39 | <jgraham> | There is a bug somewhere about normalizing textarea on suibmission |
| 15:39 | <jgraham> | I'm not sure if it covers the DOM as well |
| 16:34 | <zcorpan> | yeah it seems to me the DOM should be normalized to \n and submission normalized to \r\n |
| 16:39 | <annevk> | hmm event handlers |
| 16:45 | <dglazkov> | good morning, Whatwg! |
| 16:45 | <zcorpan> | afternoon dglazkov |
| 16:49 | <Ms2ger> | Night |
| 16:53 | <Ms2ger> | Interesting subject: http://lists.w3.org/Archives/Member/tag/2011Aug/0015.html |
| 17:12 | <annevk> | http://www.w3.org/TR/DOM-Level-3-Core/introduction.html#ID-Conformance suggests DOM3 Events can be split |
| 18:11 | <zcorpan> | matjas: you forgot to RT :) |
| 18:12 | <matjas> | no I didn’t! *whistles* |
| 18:12 | <zcorpan> | :) |
| 18:13 | <annevk> | WHATWG forums looks different |
| 18:14 | <zcorpan> | since when? |
| 18:14 | <annevk> | since green? |
| 18:15 | <annevk> | zcorpan, cannot read news without being logged in? |
| 18:16 | <zcorpan> | huh, that should be possible |
| 18:20 | <zcorpan> | fixed |
| 18:20 | <zcorpan> | thanks |
| 18:30 | <zcorpan> | annevk: it has had this design for like a year |
| 18:31 | <annevk> | evidently I'm more of an IRC guy |
| 18:31 | <annevk> | made a RT from @whatwg |
| 18:34 | <zcorpan> | thanks |
| 18:37 | <annevk> | so <textarea> submission basically makes text/plain decoding lose |
| 18:37 | <annevk> | or is text/plain submission \n separated? |
| 18:38 | <zcorpan> | does anyone support text/plain submission? |
| 18:38 | <zcorpan> | what's the use case for text/plain again? |
| 18:40 | <annevk> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1126 |
| 18:40 | <annevk> | debugging I guess? |
| 18:41 | <annevk> | does not seem to work in WebKit? |
| 18:41 | <annevk> | well Chrome |
| 18:43 | <zcorpan> | seems to me like it is a dubious use case and it changes security assumptions about form submission |
| 18:44 | <annevk> | what assumptions? |
| 18:44 | <annevk> | this is not a new feature fwiw |
| 18:45 | <zcorpan> | that browsers can only submit in two specific formats |
| 18:51 | <annevk> | seems like a false assumption |
| 18:54 | <zcorpan> | evidently. but we should be careful about what changing what can be sent from the browser. c.f. websockets |
| 19:02 | <timeless> | Ms2ger: grumble, your link was members only |
| 19:04 | <Ms2ger> | Grumble at the tag, please |
| 19:04 | <timeless> | of note, that email links to http://www.w3.org/2010/html-xml/snapshot/ |
| 19:04 | <timeless> | which i believe is less confidential |
| 19:06 | <timeless> | do they take editorial bug reports? |
| 19:06 | <timeless> | > Where HTML goes to great lengths to defined how |
| 19:06 | <timeless> | s/defined/define/ |
| 19:07 | <Philip`> | "It seems that the world at large is unlikely to adopt polyglot markup as the standard way to encode all HTML documents" |
| 19:07 | <Philip`> | They seem good at understatement |
| 19:08 | <Ms2ger> | I hear they're all Brits |
| 19:09 | <timeless> | Philip`: yeah well |
| 19:10 | <timeless> | they're wearing diplomatic hats |
| 19:10 | <timeless> | a coworker asked me if i spoke diplomat today |
| 19:10 | <timeless> | i looked at his problem and threw up my hands :( |
| 19:10 | <timeless> | if anyone here happens to have a diplomat hat handy and 10mins, please let me know, i'd be glad to share our problem :) |
| 19:10 | <timeless> | > makes all of the XML power available to any HTML document. |
| 19:11 | <timeless> | what's this `XML power` of which they speak? :) |
| 19:11 | <Philip`> | "How can XML be made more forgiving of errors?" - I think the problem with XML is not that it's unforgiving of errors, it's that there are lots of obscure edge cases that are considered errors |
| 19:11 | <zewt-> | isn't not being forgiving of errors sort of the point of xml? heh |
| 19:12 | timeless | thought that it was the point of xml |
| 19:13 | <Philip`> | Draconically requiring matched tags and quoted attributes is probably okay; requiring that text doesn't contain U+FFFE or "]]>", and that comments don't contain "--", and that xmlns:* does not have the value "http://www.w3.org/XML/1998/namespace";, is not okay and trips people up |
| 19:13 | <timeless> | > the most successful approach is likely to involve parsing the XML with an XML parser and the HTML with an HTML parser<span class="revision-deleted"><span>, combining the resulting DOMs through some other process.</span></span></p> |
| 19:13 | timeless | likes how they deleted the period from the end of that sentence |
| 19:14 | <jcranmer> | http://tools.ietf.org/html/rfc6366.txt |
| 19:14 | <AryehGregor> | Philip`, I'd say that draconian enforcement of quoted attributes is fine, but matched tags is already dodgy. It means well-formedness isn't local. It's common to want to stitch together a bunch of elements with the start and end tags added in separate places. |
| 19:15 | <AryehGregor> | But it's very rare you want to output an attribute other than all at once, so it's relatively easy to vet for well-formedness. |
| 19:15 | <timeless> | > Also, conforming trees that have tr elements as children of table elements will be replaced with semantically equivalent but tree-wise different construct where the tr elements gain a tbody parent which is a child of the table. |
| 19:15 | <AryehGregor> | MediaWiki has known cases where its parser misnests tags in output, which are very hard to fix, but attribute quoting is just fine. |
| 19:15 | <timeless> | is it me, or is `construct` in the above misconjugated? |
| 19:15 | <AryehGregor> | s/construct/constructs/ |
| 19:18 | <annevk> | timeless, prolly best to email public-html-xml |
| 19:19 | <timeless> | annevk: i wonder what kind of headache that will give me |
| 19:19 | <annevk> | from this channel hsivonen and I are involved |
| 19:20 | <timeless> | how many advils has it cost you so far? :) |
| 19:20 | <annevk> | I think in particular if hsivonen had not been involved that document could have turned out less realistic |
| 19:22 | <hober> | annevk: indeed |
| 19:23 | <timeless> | that doesn't at all surprise me |
| 19:23 | <timeless> | or perhaps better, that matches my expectations |
| 19:25 | <timeless> | heh |
| 19:25 | timeless | is amused by <script type="application/xml"> <![CDATA[ your xml cdata here ? ]]> </script> |
| 19:26 | <Ms2ger> | >.< |
| 19:26 | <Ms2ger> | Are these the guys who invented RSS? |
| 19:26 | <timeless> | ATOM is indirectly referenced repeatedly |
| 19:27 | <timeless> | their only *actual* references are to MTOM and XOP |
| 19:27 | <timeless> | which is amusing |
| 19:27 | <timeless> | since they should formally reference at least DOM (since they reference it in the body) |
| 19:28 | <timeless> | > Working out all of the details to assure that the necessary error correction produced expected results in all cases might be tedious |
| 19:28 | <timeless> | do they mean `produced` or `produces`? |
| 19:31 | <Philip`> | Or 'produce'? |
| 19:32 | <Philip`> | Tenses are hard |
| 19:33 | <zcorpan> | it should be producez |
| 19:33 | <Philip`> | The more fundamental problem with that sentence is that they consider trying to come up with a widely accepted definition of "expected" to be "tedious" instead of "an unending exercise in bikeshedding" |
| 19:34 | <timeless> | Philip`: i only deal in grammatical errors when sending editorial comments |
| 19:34 | <timeless> | technical details like " |
| 19:34 | <timeless> | utter futility" |
| 19:34 | <timeless> | are outside my scope :) |
| 19:35 | <Philip`> | (Better to come up with something roughly agreeable just for the most obvious cases, and let the other "all cases" be whatever the algorithm happens to produce) |
| 19:36 | <Philip`> | You might want to suggest they use a spelling checker too ("gauranteed", "concatentation", etc) |
| 19:37 | <timeless> | you're kidding |
| 19:37 | timeless | sighs |
| 19:37 | <zewt> | heh |
| 19:37 | Philip` | wonders if a concatentation is when you join two incantations together |
| 19:37 | <timeless> | "simulataneously" |
| 19:37 | <timeless> | you're not kidding |
| 19:37 | <zewt> | am i the only person who assumes that people who say "2" or "u're" are doing that to cover the fact that they can't keep to/two/too and your/you're straight |
| 19:38 | <Philip`> | zewt: Yes |
| 19:38 | <Ms2ger> | u're? Don't think I've seen that before |
| 19:38 | <Philip`> | I assume they're just too lazy to type properly |
| 19:38 | <timeless> | Philip`: `yes he's the only person`? |
| 19:38 | <zewt> | i saw it recently, i was dimly impressed |
| 19:38 | <Philip`> | (or grew up in a peer group where that has become the expected form of communication) |
| 19:38 | timeless | just watched a clock turn backwards by a minute |
| 19:38 | timeless | wonders why |
| 19:39 | <zewt> | lsd? |
| 19:39 | <Philip`> | ntp? |
| 19:40 | <zewt> | by "turn" i pictured an analog clock |
| 19:40 | <Philip`> | Analogue clocks could be driven by a synchronised digital time source |
| 19:40 | <timeless> | zewt: have you seen android's clock? |
| 19:40 | <timeless> | it's the old fashioned clock radio style |
| 19:40 | <timeless> | (flip digits) |
| 19:40 | <zewt> | i remember my highschool had analog wall clocks that were synchronized to the office, and once in a while you'd see them change as someone adjusted them |
| 19:41 | <timeless> | 1?#>!@$ |
| 19:41 | <timeless> | wow |
| 19:41 | <zewt> | not on my android (nexus s stock) |
| 19:41 | <Philip`> | I remember my school had a clock that took about 10 seconds for the seconds hand to count from 0 to 30, and then about 50 seconds to drag itself all the way back up to 60 |
| 19:41 | <timeless> | http://www.google.com/imgres?imgurl=http://appstorehq-production.s3.amazonaws.com/senseanalogclockwidget4x2-android-239034.320x480.1275100026.63102.jpg&imgrefurl=http://www.appstorehq.com/senseanalogclockwidget4x2-android-239034/app&usg=__7btBcdHqoaaikl31ObYlzbLzOB0=&h=480&w=320&sz=35&hl=en&start=0&zoom=1&tbnid=RjpJhSBnt3aZrM:&tbnh=140&tbnw=95&ei=JIFeTu2HGILK0AHhxIjmAg&itbs=1&iact=hc&vpx=1216&vpy=320&dur=1109&hovh=275&hovw=183&tx=133&ty=112&pa |
| 19:41 | <zewt> | i'm never a fan of digital things that try to pretend they're analog things--digital oscilloscopes are really bad like that |
| 19:42 | <Philip`> | which was disturbing when that clock was used for timing exams |
| 19:42 | <timeless> | Philip`: wow |
| 19:42 | <zewt> | timeless: that's pretty gross, but i havn't seen it--ui looks 1.xish though |
| 19:43 | <timeless> | zewt: it's really disturbing when you switch back to the home screen |
| 19:43 | <timeless> | because it doesn't update while it isn't visible |
| 19:43 | <timeless> | so when you switch back, it immediately does Flips to get to <now> |
| 19:43 | <timeless> | and you blink and think "did i just see it do something?" |
| 19:43 | <zewt> | a pretty terrible ui--why would you waste half of the very limited space on your home screen for something that's already shown in the notification bar? |
| 19:44 | <AryehGregor> | Because lots of people use their phones as clocks, and want to be able to quickly pull them out and glance at the time. |
| 19:45 | <zewt> | no, when you do that you're looking at the lock screen, not the home screen |
| 19:45 | <zewt> | and the lock screen does have a bigger clock (not that big) |
| 19:45 | <timeless> | zewt: i'm not defending it |
| 19:45 | <AryehGregor> | My home screen doesn't have a clock on it . . . are we talking about Android here? |
| 19:45 | <timeless> | i'm merely noting that it was what i saw when i encountered droids |
| 19:45 | <timeless> | it drove me nuts |
| 19:46 | <zewt> | timeless: i've just never seen it before--are you sure that isn't some vendor hack or third party widget? |
| 19:46 | timeless | shrugs |
| 19:46 | <AryehGregor> | My stock Nexus One doesn't have anything like that. |
| 19:46 | <timeless> | i'm not an android owner |
| 19:46 | <timeless> | unless you count my n900 |
| 19:46 | <timeless> | or some other non hardware droid systems |
| 19:46 | <timeless> | (and presumably you don't..) |
| 19:47 | <timeless> | fwiw, i do like having a large clock (digital please) on a lock screen |
| 19:47 | <timeless> | the n8 does that pretty well (although it's disturbing at first since they use an Outline font because the screen is OLED) |
| 19:47 | <zewt> | http://images.google.com/search?hl=en&gbv=2&tbm=isch&q=usb+oscilloscope these things are all the most absurd uis |
| 19:47 | <timeless> | http://www.google.com/imgres?imgurl=http://4.bp.blogspot.com/_PVz_hhcxiro/TUoC022Sj0I/AAAAAAAAAT0/RYO1jL-VYec/s1600/lockscreen.png&imgrefurl=http://www.nabiltewolde.com/2011/02/symbian3-ui-improvements.html&usg=__u80a9d2JbT3Lqy4UTfpJITPSchE=&h=640&w=360&sz=7&hl=en&start=0&zoom=1&tbnid=UF_W4KVixJzmQM:&tbnh=175&tbnw=103&ei=hYJeTt3mMOHa0QGcrOHRDg&itbs=1&iact=rc&dur=637&page=1&ndsp=31&ved=1t:429,r:16,s:0&tx=57&ty=73 |
| 19:54 | <timeless> | Philip`: ooh, you missed "arbirary " :) |
| 19:55 | <Philip`> | timeless: I wasn't trying to be comprehensive :-) |
| 19:56 | <timeless> | Philip` / Ms2ger / annevk : http://pastebin.mozilla.org/1316571 |
| 19:56 | Philip` | didn't even try to read the whole thing, because it seemed uninteresting |
| 19:56 | timeless | needs to get into the habit of feeding all documents through MS Word |
| 19:56 | <Ms2ger> | Don't want anything to do with it :) |
| 19:58 | <Philip`> | (In particular it just seemed to be XML people trying to figure out what HTML5 people have been saying about HTML/XML integration for the past five years) |
| 20:00 | <annevk> | timeless, okay if I email that to the list? |
| 20:01 | <timeless> | annevk: i was going to subscribe and email it |
| 20:02 | <timeless> | step one is figuring out how to subscribe, step 2 is getting the pastebin into my mail client |
| 20:02 | <Ms2ger> | You're a brave man |
| 20:03 | <timeless> | the word you sought was "foolish" |
| 20:03 | <annevk> | timeless, public-html-xml-request with subject subscribe should do |
| 20:04 | <timeless> | annevk: yeah, but that's easier to do if you're willing to use your desktop to do it |
| 20:04 | <annevk> | heh, gonna be fun meeting you again I can tell :) |
| 20:04 | <timeless> | i don't like giving my personal credentials to my corporate desktop computers |
| 20:04 | <timeless> | oh right, i need to file an expense claim for the tpac registration (it already hit my credit card statement) |
| 20:04 | <timeless> | and i need to book travel |
| 20:05 | <timeless> | (oddly, i trust my corporate phone w/ my personal credentials) |
| 20:05 | <annevk> | can Mozilla sponsor you Ms2ger? |
| 20:06 | <Ms2ger> | Wasn't planning to go |
| 20:06 | <timeless> | grr |
| 20:06 | <timeless> | annevk: can you poke gmail for me? |
| 20:06 | <timeless> | i'm using opera mobile on my n900 |
| 20:06 | <timeless> | and when i try to do stuff w/ m.gmail.com (?) i get XML parsing failed |
| 20:06 | <timeless> | XML parsing failed: syntax error |
| 20:06 | <timeless> | Reparse document as html |
| 20:07 | <timeless> | which would be fine if it were one document |
| 20:07 | <timeless> | but it' |
| 20:07 | <timeless> | s each link for a session |
| 20:07 | <timeless> | so i have to click that extra intermediate page each time i do anything :( |
| 20:07 | <annevk> | works here |
| 20:07 | <annevk> | Ms2ger, too bad |
| 20:08 | <timeless> | this is https://mail.google.com/mail/u/0/x/ .... |
| 20:08 | timeless | wonders how to check the opera version |
| 20:08 | <timeless> | it doesn't *always* happen |
| 20:08 | <timeless> | but it does seem to always happen for a given session |
| 20:09 | timeless | goes to see how this list writes its message summaries |
| 20:10 | timeless | doesn't detect any patterns |
| 20:10 | timeless | waits for the subscribe bot |
| 20:16 | timeless | finds all sorts of unread mail :( |
| 20:19 | <timeless> | annevk: ok, almost sent |
| 20:21 | <timeless> | Ms2ger: http://lists.w3.org/Archives/Public/public-html-xml/2011Aug/0048.html |
| 20:22 | <Ms2ger> | Hah, telcon |
| 20:22 | <timeless> | hurricane trumps |
| 20:22 | <timeless> | :) |
| 20:24 | <timeless> | annevk: ok, my email appears in the archives |
| 20:41 | <timeless> | Ms2ger: for unknown reasons, i only now saw your email about `you were unhappy about the ordering` :( |
| 20:43 | <timeless> | oh odd |
| 20:43 | timeless | got offlist mail about something |
| 20:43 | timeless | wonders why someone expected that to work |
| 20:45 | <zewt> | why would anyone expect offlist mail to not work? |
| 20:58 | <TabAtkins> | AryehGregor: Did I say "repaint"? If so, I meant "relayout", which happens in the middle of script execution all the time. |
| 21:02 | <AryehGregor> | TabAtkins, you said "Even when JS worked, you'd get significant and unavoidable flashes of unstyled content (multiple, because each swap would almost certainly trigger relayout)." You did say relayout, but I'm not clear how you'd get FOUCs without repaints. |
| 21:02 | <TabAtkins> | Oh, whoops, indeed. Bad on me. You'd get a lot of slowdown, rather. |
| 21:22 | <timeless> | TabAtkins: http://www.guardian.co.uk/artanddesign/2010/nov/21/sargy-mann-blind-painter-tim-adams is amusing fwiw |
| 21:23 | <timeless> | s/amusing/educational+informative/ |
| 21:23 | <timeless> | you have to watch the video |
| 21:23 | <timeless> | but my conclusions from it is that one generally will need an `assistant` who can answer questions like `what color is this point` |
| 21:24 | <timeless> | i haven't actually seen ATs that do that, oddly |
| 21:37 | <timeless> | it also seems relatively clear that one is unlikely to generally want a list of all arcs which compose a forest |
| 21:37 | <timeless> | since most of them will be overlapped by other things and are thus irrelevant to you or anyone who would be at your perspective |
| 21:38 | <timeless> | anyway |
| 21:39 | <timeless> | it seems that in the end, if someone is going to design something for people who need accessibility support, they will need to ensure that the widgets they provide aren't ambiguous and have sufficient detail to indicate what would happen if they were poked |
| 21:39 | <timeless> | a button labeled "button" would thus fail, especially if it were next to 4 others with the same label.. |
| 21:40 | <timeless> | whereas buttons labeled "floodfill", "stamp", "line", "erase" wouldn't have such a problem, and one wouldn't necessarily need to care where they are so long as one could discover they exist.. |
| 21:41 | timeless | has a feeling that a lot of this is already addressed by ARIA's ability to list areas which are subject to change, and identify changes |
| 21:41 | timeless | heads out for the day |
| 23:06 | <Moo^_^> | does it taint the canvas if I use images loaded from the same domain, different port? (repeated from #html5) |
| 23:26 | <jacobolus> | do any desktop browsers fire events for multitouch trackpad gestures? |
| 23:32 | <jacobolus> | desktop safari doesn't seem to fire 'touchstart' or 'gesturestart', but I guess I can get both X and Y movements out of a 'mousewheel' event |
| 23:32 | <jacobolus> | any way to get pinches or rotations? |
| 23:33 | <jacobolus> | do other browsers support trackpad multitouch? |
| 23:38 | <roc> | Moo^_^: yes I think so |
| 23:39 | <Moo^_^> | roc: do you know if there exist anything similar to CORS but for images? |
| 23:39 | Moo^_^ | has not read the spec |
| 23:40 | <jacobolus> | Moo^_^: what's the use case? |
| 23:40 | <Moo^_^> | jacobolus: I want to use images from another domain, but not tainting the canvas |
| 23:40 | <Moo^_^> | need to mash up images from two different sources |
| 23:41 | <roc> | there is CORS for images now |
| 23:41 | <roc> | see the "crossorigin" attribute |
| 23:41 | <roc> | only implemented in latest versions of Chrome and Firefox AFAIK |
| 23:41 | <zewt> | yeah, CORS already does that, for webgl |
| 23:44 | <Philip`> | Moo^_^: Tainting is based on origins, which are scheme+host+port |
| 23:45 | <Philip`> | so changing port is the same as changing host |
| 23:45 | <Moo^_^> | roc: zewt: Philip`: Thanks a lot! |
