| 00:02 | <annevk> | I sort of dislike subdomains but I guess it's the easiest |
| 00:04 | <Hixie> | subdirectories would be much more difficult to deal with |
| 00:04 | <jcranmer> | obviously, you should replace all subdirectories with subdomains |
| 00:05 | <Hixie> | also subdomains have the advantage of being great in autocomplete in browsers :-) |
| 00:05 | <Hixie> | jcranmer: i mean, when you have multiple people involved |
| 00:05 | <annevk> | whatwg.org/c versus html.spec.whatwg.org |
| 00:05 | <annevk> | i guess we could still have shortcuts |
| 00:05 | <Hixie> | whatwg.org/c would still work |
| 00:05 | <Hixie> | but remember, that's why we added c.whatwg.org |
| 00:06 | <Hixie> | if i do this btw i'm merging complete.html and HTML together and calling the whole thing HTML again |
| 00:06 | <annevk> | yes |
| 00:06 | <annevk> | want |
| 00:07 | <annevk> | so html.spec.whatwg.org dom.spec.whatwg.org |
| 00:07 | <annevk> | and spec.whatwg.org gives an overview? |
| 00:08 | <Hixie> | hadn't thought of spec.whatwg.org but sure |
| 00:08 | <Hixie> | we can have that show the same as platform.html5.org |
| 00:08 | <annevk> | yeah or a redirect to the wiki |
| 00:08 | <Hixie> | yeah |
| 00:09 | <Hixie> | btw in other news i think i've done the merging of all the dom core stuff |
| 00:09 | <Hixie> | annevk: do you have a dreamhost account? |
| 00:10 | <annevk> | yes |
| 00:11 | <Hixie> | do you want dom.spec.whatwg.org in your account? |
| 00:11 | <Hixie> | i wonder if we can do that |
| 00:14 | <Hixie> | hmm |
| 00:18 | <Hixie> | i'll ask them |
| 00:22 | <annevk> | you can by setting the nameservers accordingly |
| 00:22 | <annevk> | hmm actually, dunno |
| 00:23 | <annevk> | because you'd have the same nameservers :) |
| 00:27 | <Hixie> | yeah |
| 00:28 | <Hixie> | i sent a support req |
| 00:34 | <annevk> | cool about DOM Core integration |
| 00:34 | <annevk> | will review |
| 01:04 | <annevk> | so SVG wants on* |
| 01:04 | <annevk> | so infrastructure should probably be in DOM Core |
| 01:04 | <annevk> | problem: SVG needs "evt" exposed |
| 01:04 | <annevk> | also, it should be in DOM Core because browsers implement it on Element, not HTMLElement |
| 01:13 | <Hixie> | are we having all the on* handlers on every element? |
| 01:13 | <Hixie> | or only the SVG ones on SVGElement and HTML ones on HTMLElement? |
| 01:13 | <Hixie> | also which are going on Window? |
| 01:14 | <annevk> | all on Element |
| 01:14 | <annevk> | with the "evt" compat integrated |
| 01:14 | <annevk> | I guess on Document and Window too |
| 01:15 | <Hixie> | that's a _lot_ of redundant event handler attributes |
| 01:15 | <annevk> | I was planning on defining the infrastructure |
| 01:15 | <annevk> | isn't that the situation we have now? |
| 01:15 | <annevk> | anyway, I was planning on defining the infrastructure and let the "expose this list" to other specs |
| 01:18 | <Hixie> | we don't have that many events currently |
| 01:18 | <Hixie> | i guess it depends on how many SVG would be adding |
| 01:18 | <Hixie> | i'm used to SVG adding six bazillion of everything |
| 01:32 | <annevk> | whoa http://www.w3.org/Bugs/Public/show_bug.cgi?id=12417 |
| 01:32 | <annevk> | Hixie, SVG overlaps a lot with what we have |
| 01:32 | <annevk> | Hixie, in the meeting this morning they identified about 4/5 events |
| 01:33 | <annevk> | Hixie, that would need to be added |
| 01:33 | <annevk> | Hixie, I think sort of prefer this stays in HTML though given the dependency on script context, and returnValue and such |
| 01:34 | <annevk> | but I think bz wants them to move from HTMLElement to Element |
| 01:35 | <Hixie> | ok, glad it's not much |
| 01:36 | <Hixie> | i still think we should just merge dom core and html, personally |
| 01:36 | <Hixie> | but... :-) |
| 01:36 | <Hixie> | i mean, it's not like SVG can avoid depending on Window |
| 01:37 | <annevk> | hmm yeah |
| 01:37 | <annevk> | only Progress Events solely has a dependency on DOM Core |
| 01:37 | <annevk> | and not HTML |
| 01:38 | <annevk> | anyway, I rather not go there for now |
| 01:38 | <annevk> | slow change works better with the rest of the world |
| 05:21 | <annevk> | what is the use case of DocumentFragment? so you can have several element siblings instead of always a root? |
| 05:24 | <nimbu> | annevk: ut? |
| 05:26 | <annevk> | yeah |
| 05:26 | <nimbu> | k pmm |
| 05:53 | <annevk> | added a link to platform.html5.org on http://html5.org/ |
| 06:04 | <annevk> | where is ms2ger? |
| 06:04 | <annevk> | hmm |
| 06:04 | <annevk> | http://html5.org/temp/dom-mutation-methods.txt |
| 06:04 | <annevk> | i think those steps are sufficient |
| 06:06 | <annevk> | but I'd like some review before I go change the spec |
| 06:34 | <boblet> | wassup with the blog roll nav example in #the-aside-element? I would have thought a blog roll would prolly not be major navigation… |
| 06:35 | <boblet> | just a case of “major nav for the author of this page”? |
| 07:07 | <hsivonen> | I wonder what possessed the Google News team to use dc.date.issued instead of dc.issued or dcterms.issued |
| 08:05 | <hsivonen> | "SVG would probably have remained a dead technology if Google hadn't started carrying its banner with Chrome" |
| 08:05 | <hsivonen> | eh? |
| 08:07 | <heycam> | s/Google/MS/ and s/Chrome/IE/? :) |
| 08:13 | <hsivonen> | it's remarkable how Chrome has managed to create a reality distortion field like this |
| 08:13 | <heycam> | who wrote that quote? just some random person? |
| 08:14 | <hsivonen> | heycam: Kurt Cagle |
| 08:14 | <heycam> | huh. |
| 08:14 | <hsivonen> | https://plus.google.com/114141433688365651943/posts/aGDpEJX7jj8 |
| 08:14 | <heycam> | he used to be big into SVG, so I am surprised he would have that "reality-challenged" view |
| 08:14 | <heycam> | (being close to things a number of years ago, that is) |
| 10:27 | <boblet> | http://www.htmlvalidator.com/htmlval/whycseisbetter.html lol |
| 10:28 | <boblet> | also, ppl pay for validators? :o |
| 10:28 | <boblet> | hsivonen: ^^^ |
| 10:32 | <espadrine`> | boblet: "All HTML documents should have titles and "Untitled" is not a good title"... |
| 10:32 | <espadrine`> | They are checking for the word "Untitled" in <title>! |
| 10:32 | <boblet> | espadrine`: validator+ ? ;) that’s more a linting feature |
| 10:33 | <boblet> | (it is good advice tho) |
| 10:35 | <hsivonen> | boblet: I'm quite happy with my business model that doesn't involve worrying about collecting my income from end users |
| 10:36 | <boblet> | hsivonen: I was meaning more why would you pay for a validator when validator.nu is so awesum |
| 10:37 | <espadrine`> | it *does* find more errors than their CSE HTML Validator Lite v9.02 |
| 10:39 | <hsivonen> | I wonder what accessibility and SEO checking involves in practice |
| 10:39 | <Ms2ger> | longdesc and meta keywords? |
| 10:40 | <jgraham> | hsivonen: "Missing meta description tag used by some search engines" |
| 10:40 | <hsivonen> | jgraham: is that an actual quote from that product? |
| 10:41 | <jgraham> | hsivonen: Yes |
| 10:42 | <jgraham> | See the page boblet linked |
| 10:43 | <boblet> | espadrine`: also that’s only validation errors, not other stuff (linting, SEO etc) |
| 10:43 | <hsivonen> | jgraham: oh |
| 10:43 | <boblet> | W3 validator finds a bunch using HTML5 mode too, unsurprisingly |
| 10:43 | <hsivonen> | boblet: the product seems to be a syntax-highlighting text editor and validators in one product |
| 10:44 | <hsivonen> | boblet: which is useful |
| 10:44 | <hsivonen> | boblet: Validator.nu (alone) doesn't give you an editor to fix the errors in |
| 10:44 | <boblet> | yeah I’d like validator.nu locally to operate on a project level |
| 10:44 | <hsivonen> | hendry has integrated Validator.nu into vi, though |
| 10:44 | <boblet> | hsivonen: but they’re putting the validation aspect front and center in their branding huh |
| 10:45 | <boblet> | but that would mean learning vi :p |
| 10:46 | <hsivonen> | boblet: well, the validation features seem to be the bulk of the product |
| 10:46 | <hsivonen> | boblet: so it would be silly to advertise it as an editor that, oh by the way, has a validator |
| 10:47 | <espadrine`> | hsivonen: Is vi integration offline? |
| 10:47 | <boblet> | hey adactio, nice article on IDs and ARIA roles from Jan that I somehow missed. linked it up |
| 10:47 | <espadrine`> | I assumed it used internet |
| 10:47 | <hsivonen> | espadrine`: if you run Validator.nu on localhost, yes |
| 10:47 | <adactio> | boblet: merci |
| 10:48 | <espadrine`> | That could be interesting! |
| 10:48 | <espadrine`> | One day, there will be a port dedicated to html validation. |
| 11:23 | <Ms2ger> | [07:12] <annevk> where is ms2ger? |
| 11:23 | <Ms2ger> | At 7 AM? What were you thinking? :) |
| 12:25 | <matjas> | any feedback on http://www.w3.org/Bugs/Public/show_bug.cgi?id=13118? (“Consider firing the `input` event for contenteditable areas”) /cc Hixie |
| 12:33 | <matjas> | seems like a no-brainer to me, I must be missing something |
| 12:47 | <Ms2ger> | All you're missing is that there's a couple of hundred bugs that need to be addressed before yours |
| 13:25 | <hsivonen> | foolip's Java looks Pythonic. it's shockingly compact as far as Java code goes |
| 13:26 | <foolip> | hsivonen, I'm not sure if that is a compliment or not :) |
| 13:27 | <hsivonen> | foolip: it's a compliment |
| 13:27 | <foolip> | found any horrible bugs yet? |
| 13:27 | <hsivonen> | foolip: not yet |
| 13:28 | <foolip> | great, hope it works out |
| 13:28 | <foolip> | I was hacking on collecting itemValue yesterday, with that vocabulary validation should be in reach |
| 13:28 | <foolip> | but I'm not sure how to structure |
| 13:28 | <foolip> | it |
| 13:29 | <foolip> | did you have ideas about transforming it to something and using existing schema languages? |
| 13:30 | <hsivonen> | foolip: I've considered mapping Microdata to XML in order to use RELAX NG on it |
| 13:30 | <foolip> | hsivonen, could you do that while still warning at the appropriate source location? |
| 13:30 | <jgraham> | Needs more AbstractSingletonProxyFactoryBean |
| 13:30 | <hsivonen> | foolip: so far, I'm unsure if that would help or hurt |
| 13:30 | <hsivonen> | foolip: the source location could be preserved |
| 13:31 | <hsivonen> | foolip: but I'm not sure what would happen with property names that aren't NCNames |
| 13:31 | <foolip> | I was thinking about just representing items and properties with the Element class, mirroring the DOM API |
| 13:31 | <hsivonen> | so far, I've learned that itemref doesn't mean what I thought it meant |
| 13:31 | <foolip> | ok, what did you think it did? |
| 13:32 | <foolip> | the main question is of course how to represent all of schema.org |
| 13:32 | <hsivonen> | I thought its referent had to be an item so that it gave a dislocated item value to an item-valued property |
| 13:32 | <foolip> | ah, the item* prefix is misleading sometimes |
| 13:33 | <foolip> | just like itemValue is not the value of an item, but of a property |
| 13:34 | <foolip> | there's bound to be things that need code to validate, something like http://schema.rdfs.org/all.json won't suffice for all vocabularies |
| 13:35 | <hsivonen> | as I read the spec more, itemref is totally different from what I thought it was |
| 13:35 | <hsivonen> | I wonder why I thought it was something other than what it is |
| 13:36 | <foolip> | without itemref, microdata would be totally trivial |
| 13:36 | <foolip> | I wonder if Google search will actually implement it properly |
| 13:37 | <foolip> | it seems rather unlikely, really |
| 13:38 | <hsivonen> | what use case lead to the current design of itemref? |
| 13:39 | <foolip> | let me dig up a real-world example |
| 13:39 | <hsivonen> | the spec itself mentions having items as columns of a table while the properties are in the cells |
| 13:39 | <hsivonen> | but that's not really a use case |
| 13:39 | <foolip> | http://www.2gc.co.uk/a2gc-people |
| 13:40 | <hsivonen> | foolip: hmm. ok |
| 13:40 | <foolip> | in a nutshell: the information is interleaved throughout the page |
| 13:41 | <foolip> | I'm not sure we're not going to regret itemref down the line, but so far it's not too bad |
| 13:42 | <cygri> | at least itemref degrades somewhat gracefully |
| 13:42 | <foolip> | if leaving out half of the information is graceful, then yes :) |
| 13:43 | <foolip> | 50% is infinitely better than 0% |
| 13:43 | <foolip> | (as if 1%) |
| 13:43 | <foolip> | s/if/is/ |
| 14:07 | <hsivonen> | foolip: given how much complexity itemref adds to the language, it seems that the it wouldn't be that big of a deal to make properties forward-compatible with changing from string-valued to item-valued |
| 14:08 | <hsivonen> | although that's in the department of theoretical problems until we get far enough that someone has that problem |
| 14:08 | <hsivonen> | (and then it's too late) |
| 14:08 | <foolip> | hsivonen, I'm not sure what you mean |
| 14:08 | <foolip> | what's string-valued vs item-valued? |
| 14:08 | <hsivonen> | foolip: what TabAtkins sent email to the list about |
| 14:08 | <hsivonen> | foolip: properties have either strings or items as their values |
| 14:09 | <foolip> | right |
| 14:09 | <foolip> | is this an old email I've forgotten? |
| 14:09 | <foolip> | the one I replied to perhaps |
| 14:09 | <hsivonen> | foolip: so suppose you have a list of tracks of an album and each track is a string that is the tracks name |
| 14:09 | <hsivonen> | foolip: then someone who isn't the author of the page writes consumption code that expects string values |
| 14:10 | <hsivonen> | foolip: then publishers decide that tracks are now items |
| 14:10 | <hsivonen> | foolip: the consuming code breaks |
| 14:10 | <foolip> | yeah, I agree that it is a problem |
| 14:10 | <foolip> | but I'm not thrilled about having a dual interpretation of properties where some consumers end up having to consider *both* at once |
| 14:11 | <hsivonen> | it could be address by having the consumer say to his/her microdata tooling "get property foo as string" |
| 14:11 | <foolip> | for example, because people only tested with consumers that consider it a string, and didn't notice the item representation was broken |
| 14:11 | <hsivonen> | and the tooling to have a rule about what to do if property foo is an item |
| 14:11 | <hsivonen> | foolip: good point |
| 14:11 | <hsivonen> | maybe I'm just a Complicator here |
| 14:12 | <foolip> | and if you also have the reverse, you have to look at both and apply heuristics |
| 14:12 | <foolip> | but without seeing what the consuming code would look like, it's hard to make guesses at solutions |
| 14:13 | <hsivonen> | foolip: the solution could be an itemname that's a privileged property |
| 14:13 | <foolip> | itemNAME, a new attribute? |
| 14:14 | <hsivonen> | foolip: so all vocabularies would have to use <span itemname>Whatever</span> for their most nameish or titleish property |
| 14:14 | <hsivonen> | instead of <span itemprop=title>Whatever</span> |
| 14:15 | <foolip> | solving it by making authors cater to the problem sounds like it'll only work 1% of the type |
| 14:15 | <foolip> | it broke because the author was careless in upgrading from string to item without checking what existing consumers do... |
| 14:15 | <hsivonen> | well, the solution requires vocabulary designers to cooperate |
| 14:16 | <hsivonen> | not authors per se |
| 14:16 | <foolip> | the problem is bound to happen, but I've failed to come up with an opinion on the matter :) |
| 14:16 | hsivonen | gets some gloves http://thedailywtf.com/Articles/The_Complicator_0x27_s_Gloves.aspx |
| 14:17 | <hsivonen> | though I don't know what the gloves would be in this case |
| 14:17 | <foolip> | oh, is The Complicator a meme of some sort? |
| 14:18 | foolip | replies to bug 850 |
| 14:18 | <hsivonen> | foolip: yes |
| 15:46 | <matjas> | Ms2ger: I can wait :) was just wondering if this is something that has a chance to be specced, since only WebKit has this behavior atm |
| 15:46 | <Ms2ger> | If webkit has it and Gecko is in favour, sounds like it does |
| 16:21 | <foolip> | hsivonen, do you want the changes based on your feedback as a new complete patch, or as a delta on top of the old one? |
| 16:53 | <dglazkov> | good morning, Whatwg |
| 16:53 | <jcranmer> | wake up and smell the ashes |
| 16:56 | <hober> | morning dglazkov |
| 16:56 | <hsivonen> | foolip: either way works |
| 16:56 | <foolip> | ok |
| 16:57 | <Ms2ger> | Evening |
| 16:57 | <Michael> | Hello :) |
| 16:58 | <dglazkov> | you are a friendly bunch I see |
| 18:08 | <AryehGregor> | . . . why does http://platform.html5.org/ link to the W3C version of HTML? |
| 18:08 | <Ms2ger> | Because MikeSmith works for the W3C? |
| 18:09 | <Ms2ger> | Not saying it's a conspiracy, just saying |
| 18:09 | <AryehGregor> | Phooey. |
| 18:16 | <annevk> | do we need to define that "an /x/ element" means "an element with local name /x/ and namespace ?!"? |
| 18:17 | <Ms2ger> | Where? |
| 18:20 | <annevk> | in DOM Core |
| 18:20 | <AryehGregor> | annevk, that would be nice, actually. |
| 18:20 | <annevk> | we talk about "html element" sometimes but that is never really mapped to local name |
| 18:20 | <annevk> | I guess you would still need to mention the namespace |
| 18:20 | <AryehGregor> | Currently I just make /x/ an xref to the relevant element in the HTML spec. |
| 18:20 | <Ms2ger> | Mm |
| 18:21 | <annevk> | an /x/ element in namespace /y/ |
| 18:21 | <AryehGregor> | But that only works if I have a specific name, and it's not a variable. |
| 18:21 | <annevk> | also that does not work for DOM Core as long as we do not want to depend on HTML |
| 18:21 | <AryehGregor> | Otherwise I'm forced to say "an <span>HTML element</span> with <span ...>local name</span> <var>tag name</var>" or such. |
| 18:22 | <AryehGregor> | Where I've defined "HTML element" to mean "Element in the HTML namespace". Dunno if that duplicates definitions elsewhere. |
| 18:33 | <AryehGregor> | Okay, my latest Google+ post is going to cause a massive drama-fest. |
| 18:33 | <AryehGregor> | But I hope Jeni et al. appreciate the points I make and it makes some kind of difference. |
| 18:33 | <AryehGregor> | If the peanut gallery wants to freak out, I'm happy to ignore them. |
| 18:33 | <AryehGregor> | https://plus.google.com/u/0/105458233028934590147/posts/h7nsT7wuNmX |
| 18:33 | AryehGregor | can't figure out if you can link directly to posts |
| 18:34 | AryehGregor | finds an id for the post of z135hbfqbruhcpvbj04chplqevjesd4xigk#1311874590176000, but somehow doubts that's meant to be used for anchors |
| 18:36 | <Ms2ger> | "I know you're all perfectly reasonable people who are committed to improving the web" |
| 18:36 | <Ms2ger> | My sarcasm detector is failing me |
| 18:36 | <AryehGregor> | I wasn't being sarcastic at all. |
| 18:37 | <AryehGregor> | I think they're doing a bad job at improving the web, in certain key respects, but they're trying. |
| 18:37 | <Ms2ger> | Maybe it's just my cynicism that made me doubt that :) |
| 18:39 | <Philip`> | AryehGregor: Given that Google thinks that strings like 105458233028934590147 and h7nsT7wuNmX and aapbdbdomjkkjkaonfhkkikfgjllcleb etc are good things to use in URLs in place of human-readable identifiers, I don't see why they wouldn't think z135hbfqbruhcpvbj04chplqevjesd4xigk#1311874590176000 was perfectly acceptable too |
| 18:39 | <AryehGregor> | Philip`, to start with, it has a # in it, so it wouldn't register as a valid URL in lots of contexts. |
| 18:41 | <Ms2ger> | FWIW, I have vague plans to move my specs to the W3C, but it's low enough priority that it may never happen |
| 18:41 | <Philip`> | Wouldn't that just get replaced with a %34 or whatever? |
| 18:42 | <Philip`> | Oh, %23 |
| 18:42 | Philip` | was only off by one (twice) |
| 18:49 | <brucel> | Who's up for anwering a stoopid microdata question (prompted by TabAtkins' blogpost)? |
| 18:50 | <annevk> | AdobeGuest network is so slow |
| 18:50 | <annevk> | just ask brucel |
| 18:50 | <Ms2ger> | Maybe if you pushed an Adobe employee into a lake... |
| 18:50 | <annevk> | more chance than with asking to ask |
| 18:51 | <brucel> | OK - itemid: according to spec, "The itemid attribute must not be specified on elements that do not have both an itemscope attribute and an itemtype attribute specified, and must not be specified on elements with an itemscope attribute whose itemtype attribute specifies a vocabulary that does not support global identifiers for items, as defined by that vocabulary's specification.? |
| 18:52 | <brucel> | how do I know if a vocab "supports" global indetifiers? |
| 18:53 | <brucel> | Tabh's blogpost http://www.xanthir.com/blog/b4570 leads me to assume that it means: if a vocab has something that's primary-key like, eg ISBN or social security number, then you can use that |
| 18:53 | <cygri> | brucel: the documentation of the vocabulary should explain whether it supports itemid |
| 18:54 | <cygri> | if the documentation doesn't say anything about itemid, then don't use it |
| 18:54 | <brucel> | but "supports" could mean that it conforms to some technical schema full of OWLs and SPARQLs and things that upset my hippocampus |
| 18:55 | <brucel> | microdata spec tells me that certain vocabs support global identifiers, but then don't use itemid in the examples, so am not any wiser |
| 18:56 | <cygri> | brucel, look at the vevent, vcard and so on vocabs that are part of the microdata spec |
| 18:56 | <cygri> | they explain where you can use itemid |
| 18:56 | <cygri> | OWL and SPARQL has nothing to do with this really |
| 18:57 | <annevk> | euh |
| 18:58 | <annevk> | wtf is going on here: http://lists.w3.org/Archives/Public/www-archive/2011Jul/0025.html |
| 18:58 | <annevk> | hint: dates |
| 18:59 | <brucel> | cygri have been reading here http://www.whatwg.org/specs/web-apps/current-work/multipage/microdata.html#mdvocabs but none of the examples employ itemid |
| 19:00 | <cygri> | brucel, hm, how to explain this |
| 19:00 | <cygri> | itemid usually wouldn't be something like the social security number |
| 19:01 | <cygri> | it would just be a URL that, according to the publisher of the HTML page, is a globally unique identifier for that item |
| 19:02 | <cygri> | so, if i were to add microdata to my own homepage, i'd use itemid="http: //richard.cyganiak.de/#cygri"; or something like that, and not my social security number |
| 19:03 | <brucel> | cygri thanks - and what's the benefit of your doing that |
| 19:04 | <Ms2ger> | lolurls |
| 19:05 | <Ms2ger> | Having taken ten years to confirm that we implemented your feedback... |
| 19:05 | <cygri> | well, then this url can be used in an itemprop value to refer to me |
| 19:05 | <Ms2ger> | That's worse than Hixie |
| 19:05 | <annevk> | Ms2ger, did you look at http://html5.org/temp/dom-mutation-methods.txt ? |
| 19:05 | <Ms2ger> | " This draft is expected to be updated or made obsolete within three months of its publication (3 October 2002)." |
| 19:06 | <Ms2ger> | Looked at it, yes |
| 19:06 | <Ms2ger> | Read, no |
| 19:07 | <cygri> | brucel, a better example might be an event. you can give it a unique identifier with itemid, and then someone could use that id to state that they're attending the event or whatever, assuming you have a vocabulary that makes use of itemids in that way |
| 19:07 | <annevk> | SVG WG is going to discuss HTML in SVG soonish btw |
| 19:07 | <annevk> | if people have thoughts |
| 19:07 | <cygri> | brucel, you *could* still use things like urn:isbn:123456 for itemids |
| 19:09 | <brucel> | ok, thanks cygri |
| 19:09 | <cygri> | yw brucel |
| 19:10 | <Philip`> | Ms2ger: At least they implemented the feedback promptly, whereas Hixie can take 5 years to read a message and then replies asking for clarification on some of the points |
| 19:10 | <brucel> | feels a bit ... edgecasey.. but I think I understand enough to know that I don't need to use it if I'm marking up people, isbns, events |
| 19:11 | <Ms2ger> | Philip`, please elaborate. |
| 19:12 | <Philip`> | Ms2ger: I'll elaborate in 2016 if that's alright |
| 19:12 | <AryehGregor> | Philip`, be fair, he just said that his oldest piece of feedback was from 2007. |
| 19:12 | <Ms2ger> | Sure |
| 19:12 | <AryehGregor> | That's only four years. |
| 19:13 | <AryehGregor> | (BTW, Hixie, it is a real problem that you don't respond to feedback for months or years, it massively discourages feedback) |
| 19:13 | <AryehGregor> | (although you edit so much that maybe you don't have any option) |
| 19:13 | AryehGregor | plan to never edit so many specs that he can't respond to feedback promptly |
| 19:14 | <dglazkov> | Hixie has become a one-person bureaucracy!!! :P |
| 19:14 | <Philip`> | http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2011-June/032026.html was 6.5 years after some comments, or 5 if you're feeling charitable |
| 19:16 | <Philip`> | (...and asks non-rhetorical questions in a few cases) |
| 19:16 | <Ms2ger> | annevk, looks good |
| 19:17 | <annevk> | cool, now I need to figure out how to rewrite it nicely |
| 19:26 | <dglazkov> | I am genuinely surprised by negative reaction to AryehGregor's spec announcement. |
| 19:27 | <AryehGregor> | dglazkov, then you aren't familiar with W3C politics. |
| 19:27 | <jamesr> | negative reaction where? |
| 19:27 | <dglazkov> | I do avoid them at all costs. |
| 19:27 | <Ms2ger> | All over the W3C, I guess |
| 19:34 | <annevk> | inner substeps yay |
| 19:35 | <annevk> | at least these either continue or fail |
| 19:35 | AryehGregor | pokes hober with a pointy stick in the direction of innerText |
| 19:35 | <AryehGregor> | Pretty soon I'm going to give up on getting a response from WebKit here. |
| 19:42 | <gsnedders> | Ms2ger: All those specs are WebApps WG not HTML WG, so nothing for the HTML testsuite |
| 19:42 | <Ms2ger> | appcache? |
| 19:42 | <gsnedders> | Oh, that isn't |
| 19:42 | gsnedders | can't read |
| 19:43 | <Ms2ger> | And webapps doesn't have a test list afaik |
| 19:47 | <jgraham> | WebSQL is dead isn't it? |
| 19:56 | <MacTed> | jgraham - spec appears so, judging by http://www.w3.org/TR/webdatabase/ ... |
| 19:56 | <MacTed> | unfortunate, and surprising that they say all implementors used same backend (SQLite), as we implemented bridges to ODBC amd XMLA -- http://wikis.openlinksw.com/dataspace/owiki/wiki/UdaWikiWeb/InstallConfigHTML5SQLBridges -- and thus backends may be Oracle, DB2, Informix, MSSQL, MySQL, PostgreSQL, etc. |
| 19:58 | <Ms2ger> | Very fortunate |
| 19:58 | <jgraham> | Tests should be ported to testharness.js too |
| 19:58 | <Ms2ger> | Yeah |
| 19:58 | <MacTed> | Ms2ger - why fortunate? |
| 19:58 | <Ms2ger> | Because that would have led to a sqlite monoculture |
| 19:59 | <Ms2ger> | And as nice as it might be, it would have had all the problems of a monoculture |
| 19:59 | <MacTed> | Ms2ger - did you miss the part where we implemented ODBC and XMLA binding? as in, monoculture no more? |
| 19:59 | <annevk> | will we avoid the monoculture with IndexedDB or is everyone going to use the open source work from Google for Chrome? |
| 20:01 | <Ms2ger> | We have an independent implementation AFAIK |
| 20:01 | <annevk> | on top of SQLite |
| 20:01 | <jamesr> | currently there are sqlite backends for indexdb and leveldb |
| 20:01 | <annevk> | teehee |
| 20:01 | <jamesr> | so that's 2 |
| 20:01 | <annevk> | what is leveldb? |
| 20:01 | <jamesr> | http://code.google.com/p/leveldb/ |
| 20:01 | <jamesr> | storage engine |
| 20:02 | <annevk> | but it uses sqlite? |
| 20:02 | <jamesr> | no |
| 20:02 | <jamesr> | it's a storage engine |
| 20:02 | <annevk> | i got thrown of by sqlite backend above |
| 20:03 | <MacTed> | I don't think you mean "backend" the same as I do... |
| 20:03 | <jamesr> | i'm just talking about indexdb |
| 20:03 | <MacTed> | ... |
| 20:04 | <MacTed> | "there are sqlite backends for indexdb and leveldb" |
| 20:04 | <MacTed> | please rephrase? |
| 20:05 | <MacTed> | I would understand that to mean "sqlite is (or can be) the storage engine used by indexdb and leveldb" |
| 20:05 | <jamesr> | that is true |
| 20:07 | <MacTed> | so... indexdb and leveldb are 2 masks/wrappers over 1 body/engine (sqlite) |
| 20:07 | <MacTed> | how does that avoid monoculture? |
| 20:08 | <MacTed> | reading further in http://www.w3.org/TR/webdatabase/, this line seems like it should have been in the read box above it -- "The Web Applications Working Group continues work on two other storage-related specifications: Web Storage and Indexed Database API. " |
| 20:08 | <jamesr> | wait |
| 20:08 | <jamesr> | sorry i misread |
| 20:08 | <jamesr> | sqlite is _not_ the storage engine used by leveldb |
| 20:09 | <annevk> | i think what you meant is that indexeddb has two potential backends: sqlite and leveldb |
| 20:09 | <jamesr> | right |
| 20:09 | <Philip`> | I think the main problem is the API being a non-standardised monoculture (which WebSQL's is, since it incorporates SQLite's SQL syntax and semantics which aren't defined clearly enough for independent implementation) |
| 20:10 | <jamesr> | websql had only one possible backend, sqlite |
| 20:10 | <annevk> | if we had kept the SQL parsing separate it would have been better |
| 20:10 | <annevk> | but nobody ever defined that |
| 20:11 | <Philip`> | It's okay if everyone uses the same implementation, as long as the API is standardised and lets them easily use a different implementation if they ever had any reasons to do so |
| 20:11 | <dglazkov> | https://plus.google.com/103035368214666982008/posts/43eHRWqsMEP |
| 20:11 | <Ms2ger> | dglazkov, will I regret following that link? |
| 20:11 | <dglazkov> | define regret |
| 20:11 | <Ms2ger> | I.e., is it goatse or W3C process discussion? |
| 20:11 | <jamesr> | Ms2ger: is there a difference? |
| 20:12 | <MacTed> | I think I get you... |
| 20:12 | <MacTed> | OK, so we'll have to go back to the labs and implement IndexedDB-to-ODBC. I think that's a can-do. |
| 20:12 | <jamesr> | Philip`: it is a little dangerous if everyone uses the same impl since in that situation it's very difficult to avoid creating strange dependencies |
| 20:12 | <MacTed> | (now looking at http://www.w3.org/TR/IndexedDB/ and http://www.w3.org/TR/webstorage/ ) |
| 20:13 | <Philip`> | MacTed: Does your ODBC/etc bridge support precisely the same SQL syntax/semantics as SQLite? |
| 20:15 | <Philip`> | If it's just the same WebSQL method calls but different SQL then it's not really compatible with WebSQL (people will write web pages depending on the quirks of SQLite so they won't work if you try to swap it out for a different SQL implementation) |
| 20:18 | <gsnedders> | Does importScripts work cross-origin? |
| 20:20 | <jgraham> | Anybody got any idea what to do when you get ImportError: /home/jgraham/lib/python2.6/lib-dynload/unicodedata.so: undefined symbol: _PyUnicodeUCS4_T\ |
| 20:20 | <jgraham> | oNumeric |
| 20:20 | <gsnedders> | "Attempt to fetch each resource identified by the resulting absolute URLs, from the entry script's origin, with the synchronous flag set." — what about scripts not from the same origin? Silently ignore them. |
| 20:20 | <jgraham> | trying to run something under mod_wsgi |
| 20:20 | <gsnedders> | jgraham: Looks like you have something compiled for UCS4 and something compiled for UCS2 |
| 20:20 | <jgraham> | That seems to work fine under normal python |
| 20:21 | <jgraham> | With the same executable and the same python path |
| 20:21 | <jgraham> | gsnedders: Well I worked that much out |
| 20:21 | <jgraham> | I just can't work out *what* it might be |
| 20:22 | <gsnedders> | jgraham: mod_wsgi? |
| 20:22 | <jgraham> | Well that is the only thing I haven't compiled by hand |
| 20:22 | <jgraham> | But it doesn't *say* this should be necessary |
| 20:22 | <gsnedders> | Oh, the "from the entry script's origin" is an argument to "fetch". Well that's certainly unclear. |
| 20:26 | <AryehGregor> | Can I disable section numbering in anolis? |
| 20:26 | <AryehGregor> | I don't want people reporting things by section numbers, since they don't appear in the source. |
| 20:27 | <Ms2ger> | .secno { display: none; } |
| 20:27 | <AryehGregor> | Blech. |
| 20:27 | <Philip`> | Shouldn't you try to make things as easy as possible for reviewers, to maximise the amount of feedback you get? |
| 20:27 | <AryehGregor> | They can give the section name. |
| 20:27 | <gsnedders> | AryehGregor: Anolis 2 made it easier. :P |
| 20:27 | <Philip`> | (even if that means you have to cross-check an old HTML version to match up numbers) |
| 20:27 | <Ms2ger> | It probably wouldn't be hard to hack that in |
| 20:27 | <AryehGregor> | Giving the section name is about as easy for them and saves me effort. |
| 20:28 | <AryehGregor> | I should also change all my ol's to ul's. |
| 20:33 | <MacTed> | Philip` - partially true. anyone who writes to the quirks of SQLite wouldn't be able to move to our bridge, but anyone who didn't want SQLite could start (and stay) with our bridge, whether they then used ODBC standard (recommended) or DBMS-specific dialect |
| 20:33 | <Hixie> | AryehGregor: given how much feedback we get, discouraging feedback isn't a real problem :-P |
| 20:34 | <MacTed> | Philip` - and if they wrote to ODBC spec, and then wanted to use SQLite ... the bridge still works, with their chosen backend. |
| 20:34 | <AryehGregor> | Hixie, I'd love to get more feedback myself. |
| 20:35 | <Ms2ger> | Hixie, how much feedback *you* get, you mean ;) |
| 20:41 | <jamesr> | lawl |
| 20:41 | <jamesr> | now charles pritchard is quoting IRC chatlogs on email on public-canvas-api |
| 20:41 | <jamesr> | the circle has completed |
| 20:46 | <othermaciej> | did his email get quoted on IRC first? |
| 20:46 | <jamesr> | not directly |
| 20:46 | <annevk> | Ms2ger, how does this read: "If node is a DocumentFragment node, inserts its children (preserving tree order), before child or at the end of parent if child is null." |
| 20:46 | <jamesr> | but the IRC conversation was about public-canvas-api email threads |
| 20:46 | <jamesr> | and how many of them were kind of useless |
| 20:46 | <annevk> | Ms2ger, child is the new refChild |
| 20:47 | <Ms2ger> | s/inserts/insert/? |
| 20:47 | <annevk> | if that's it, great |
| 20:47 | <Ms2ger> | I think it's fine |
| 20:48 | <AryehGregor> | Okay, is there any git hosting site that will serve text/html from my repo without $$$? I'd prefer to use something like github or gitorious instead of gitweb.cgi here. |
| 20:49 | <annevk> | alright |
| 20:49 | <annevk> | completed insert node |
| 20:49 | <annevk> | now replace node |
| 20:49 | <annevk> | then we are pretty close to being able to spec any kind of mutation events |
| 20:50 | <jgraham> | AryehGregor: Will github not do that? |
| 20:50 | <Ms2ger> | Then, LC |
| 20:50 | <The_8472> | AryehGregor... get a virtual server and do it yourself? |
| 20:50 | <AryehGregor> | jgraham, AFAICT you need to buy some kind of better account. |
| 20:50 | <The_8472> | cloud is overrated |
| 20:50 | <AryehGregor> | Maybe I should. |
| 20:50 | <AryehGregor> | The_8472, I have a dedicated server that I'm already using. The problem is, gitweb stinks compared to sites like github. |
| 20:50 | <AryehGregor> | Has way fewer features, is uglier, etc. |
| 20:51 | <The_8472> | mhm... and no better software available? |
| 20:51 | <AryehGregor> | It has syntax highlighting too. |
| 20:51 | <AryehGregor> | Nothing nearly as slick as the professional stuff that I'm aware of. |
| 20:51 | <jgraham> | Oh, I assumed all those people with github hosted pages were using the free service |
| 20:52 | <AryehGregor> | Hmm, does it have syntax highlighting? |
| 20:52 | <AryehGregor> | You'd assume so. |
| 20:52 | <annevk> | Ms2ger, apparently we're going to spec on* |
| 20:52 | <AryehGregor> | But I don't see it. |
| 20:53 | <annevk> | Ms2ger, and DOM Range should prolly move in once AryehGregor fixed all its bugs :) |
| 20:53 | <Ms2ger> | Sounds good |
| 20:53 | <annevk> | DOM Range does a lot of mutation stuff so having that tightly coupled would be good |
| 20:53 | <jgraham> | AryehGregor: "All plans come with [...] Wikis, Issues, Downloads, Pages & more" |
| 20:54 | <jgraham> | Which sounds like text/html |
| 20:54 | <Philip`> | AryehGregor: Use GitHub plus cron on your own server to clone and publish the HTML? |
| 20:54 | <AryehGregor> | Philip`, hmm. Interesting thought. |
| 20:54 | <Ms2ger> | Ugh, 22-step algorithms |
| 20:54 | <AryehGregor> | jgraham, I mean serving raw text/html of files in the repo. |
| 20:54 | <AryehGregor> | Ms2ger, only 22? |
| 20:54 | <Ms2ger> | That's the first I noticed |
| 20:55 | <AryehGregor> | "Delete the contents" is 32 steps. |
| 20:55 | <AryehGregor> | Plus lots of substeps, naturally. |
| 20:55 | <AryehGregor> | insertParagraph is also 32. |
| 20:55 | <AryehGregor> | Did I ever mention execCommand is complicated? |
| 20:55 | <Ms2ger> | Those aren't in DOM Range :) |
| 20:57 | <annevk> | back in 30-45min |
| 20:57 | <AryehGregor> | Yeah, but the 22-step algorithm in DOM Range was written by me too. :) |
| 20:58 | <Ms2ger> | AryehGregor, is commonAncestorContainer hard or did you just not get around to it? |
| 20:58 | <AryehGregor> | Ms2ger, I never got around to it. |
| 20:59 | <AryehGregor> | I'm pretty sure it's trivial. |
| 20:59 | <Ms2ger> | OK |
| 20:59 | <AryehGregor> | var ret = range.startContainer; while (ret != range.endContainer && !isAncestor(ret, range.endContainer)) ret = ret.parentNode; return ret; |
| 20:59 | <AryehGregor> | Something like that. |
| 21:14 | <jgraham> | Victory! |
| 21:14 | <jgraham> | http://test-review.hoppipolla.co.uk/changeset/6e56634565c9/ |
| 21:14 | <jgraham> | Or in general http://test-review.hoppipolla.co.uk/ |
| 21:14 | <jgraham> | The UI kinda sucks and there are a few non-trivial problems with the software |
| 21:14 | <jgraham> | But it is something |
| 21:15 | <jgraham> | I don't think itt's setup so that people can comment at the moment |
| 21:18 | <Ms2ger> | 502 Bad Gateway |
| 21:18 | <Ms2ger> | It rocks! |
| 21:19 | <jgraham> | Try again? |
| 21:19 | <AryehGregor> | Hixie, there's no longer any plan to merge the editing spec into HTML, right? |
| 21:19 | <Ms2ger> | So, how about you test it by reviewing http://test-review.hoppipolla.co.uk/changeset/abb5f54243bf/? |
| 21:20 | <jgraham> | Ms2ger: I see what you did there |
| 21:20 | <Ms2ger> | :) |
| 21:20 | <Ms2ger> | Want me to link you to the spec? :) |
| 21:21 | <jgraham> | If you like |
| 21:22 | <Ms2ger> | http://www.whatwg.org/specs/web-apps/current-work/multipage/webappapis.html#handler-window-onload |
| 21:24 | <AryehGregor> | What's the Mac equivalent of Shift-Enter? |
| 21:27 | <jgraham> | Evidently the allow_anon_comments setting doesn't do much :) |
| 21:28 | <dglazkov> | option-enter |
| 21:28 | <AryehGregor> | dglazkov, thanks. |
| 21:29 | <dglazkov> | it actually varies from editor to editor :( |
| 21:29 | <AryehGregor> | :( |
| 21:30 | <dglazkov> | for key in ['shift','option','command']: |
| 21:31 | <AryehGregor> | Hmm, should backspacing immediately after a link unlink? |
| 21:31 | <AryehGregor> | I think yes. |
| 21:31 | <AryehGregor> | IE9 and Word 2007 do it, but other browsers and OO don't. |
| 21:32 | <AryehGregor> | I'll change the spec to match IE and Word, it seems more useful. |
| 21:33 | Ms2ger | touches DOM Range for once |
| 21:33 | <dglazkov> | BOOM |
| 21:34 | <david_carlisle> | Ms2ger: on the xml serialialisation bug, might be interested in helping spec that , but no time until end of august, just wanted to get comments in before LC deadline |
| 21:34 | Ms2ger | whacks dglazkov |
| 21:34 | <dglazkov> | :D |
| 21:35 | <Ms2ger> | david_carlisle, Shelly might throw a tantrum over it, but innerHTML probably will be moved out of HTML soon |
| 21:35 | <Ms2ger> | And there's no LC in sight for my spec :) |
| 21:35 | <AryehGregor> | Into where, DOMPS? |
| 21:35 | <david_carlisle> | Ms2ger: well 'twas me put in a buzilla request asking for that... |
| 21:35 | <Ms2ger> | Yeah |
| 21:36 | <david_carlisle> | so it would apply to element not htmlelement and mathml would benefit |
| 21:37 | <foolip_> | what does innerHTML do outside of HTML? |
| 21:38 | <Ms2ger> | Same as it does inside |
| 21:38 | <david_carlisle> | foolip if you have a html span with mathml in it then innerhtml on the span has to, by spec and current implementations serialise the math |
| 21:38 | <foolip_> | parse it as HTML and insert? |
| 21:38 | <david_carlisle> | but you can't access the seialisation by doing innerhtml on math |
| 21:38 | <foolip_> | ah, getting innerHTML, not setting it |
| 21:38 | <david_carlisle> | well both |
| 21:38 | <foolip_> | setting doesn't magically parse as XML or anything, right? |
| 21:39 | <Ms2ger> | No |
| 21:39 | <foolip_> | maybe I should just read the spec when I'm bored :) |
| 21:39 | <foolip_> | ignore me |
| 21:39 | <david_carlisle> | parsing of mathml is fully specified in html(5) parse algorithm |
| 21:39 | <gsnedders> | It parses it as XML in XHTML, at lesat. |
| 21:39 | <Ms2ger> | Like using <svg> in HTML doesn't magically parse as XML ;) |
| 21:39 | <gsnedders> | So it'd make sense to parse as XML in XML. |
| 21:39 | <foolip_> | right, silly me |
| 21:39 | <david_carlisle> | gsnedders: in application/xml yes not in text/html |
| 21:40 | <gsnedders> | david_carlisle: Well, that's not XHTML or XML. |
| 21:40 | <david_carlisle> | gsnedders: sorry I'm lost, back up:-) |
| 21:41 | <Ms2ger> | text/html \not\in \{XHTML, XML\} |
| 21:41 | <david_carlisle> | gsnedders: worry about xml mime types later, immediate concern is mathml in html served as text/html |
| 21:41 | <Ms2ger> | And apologies for the LaTeX |
| 21:41 | <gsnedders> | david_carlisle: innerHTML parses as XML in XHTML. I made no comment about HTML. |
| 21:41 | <gsnedders> | Ms2ger: Can't you hand-write MathML? |
| 21:41 | <Ms2ger> | Only with innerHTML |
| 21:42 | <david_carlisle> | Ms2ger:http://www.w3.org/Bugs/Public/show_bug.cgi?id=11204 |
| 21:42 | <david_carlisle> | LaTeX, what's that? |
| 21:42 | <Ms2ger> | "Firefox doesn't know how to open this address, because the protocol (ms2ger) isn't associated with any program." |
| 21:43 | <david_carlisle> | sorry firefox (well chatzilla) thought it would be clever and stick your name at front, probably i caught the tab key or something:-) |
| 21:43 | <Ms2ger> | :) |
| 21:44 | <annevk> | sounds like a bug in Firefox |
| 21:45 | <Ms2ger> | annevk, of course not, there are no bugs in Firefox |
| 21:54 | <Michael> | ^^ |
| 22:34 | AryehGregor | seriously does not understand why stuff like http://www.w3.org/Bugs/Public/show_bug.cgi?id=13431 has to mention anything about accessibility at all |
| 22:40 | <TabAtkins_> | AryehGregor: That's explained. People in the a11y TF came up with it, so they decided to phrase it in terms of a11y. |
| 22:41 | <AryehGregor> | Yes, but I don't understand how the explanation makes sense given that it really manifestly has nothing at all to do with a11y. |
| 22:41 | <AryehGregor> | But it's good feedback in this case anyway. |
| 22:46 | <annevk> | it's just missing "e.g. " |
| 22:48 | <annevk> | it's good feedback, but to report a small typo they could have done with a somewhat simpler bug report |
| 22:49 | <AryehGregor> | Yeah, but they're used to getting their bug reports closed NEEDSINFO with demands for explanation and use cases, so I don't blame them. |
| 23:01 | <Hixie> | AryehGregor: define "plan" |
| 23:01 | <Hixie> | AryehGregor: i don't think long-term that we should have execCommand specced in a different spec than HTMLDocument |
| 23:02 | <Hixie> | AryehGregor: then again on the long term I don't think we should have HTMLDocument specced in a different spec than Document either |
| 23:02 | <AryehGregor> | Hixie, you don't think we should have more than one web spec in the long term. :) |
| 23:02 | <AryehGregor> | (except maybe orthogonal stuff like HTML vs. SVG) |
| 23:02 | <Hixie> | i expect i'll merge the editing stuff into HTML in the next year or two |
| 23:02 | <AryehGregor> | Even if I'm still actively editing it? |
| 23:03 | <Hixie> | if you're actively editing it, i'm happy to delay the merge |
| 23:03 | <annevk> | AryehGregor, that is not completely true; Hixie wants multiple specs, just not that small ;) |
| 23:04 | <AryehGregor> | Given that we're going to have multiple people editing significant things indefinitely, don't you think it makes more sense to have separate specs that reference each other? You're going to have somewhat different stylistic conventions and so on, I don't think it makes sense to actually merge everything. |
| 23:04 | <AryehGregor> | The editing stuff is pretty orthogonal to the rest of HTML. |
| 23:05 | <annevk> | I think we should try to converge on stylistic behavior when possible |
| 23:05 | <AryehGregor> | I don't think we should bother, unless it's likely to confuse anyone. |
| 23:05 | <annevk> | Makes it easier for people to read specs |
| 23:05 | <AryehGregor> | Otherwise we're reinventing pubrules. |
| 23:05 | <annevk> | style guide is way different from pubrules |
| 23:05 | <AryehGregor> | Different people will have different styles, it's not something that we should waste time negotiating about if different editors disagree. |
| 23:06 | <gsnedders> | annevk: At what level? |
| 23:06 | <annevk> | pubrules is about boilerplate |
| 23:06 | <gsnedders> | Someone using en-gb and someone using en-us is mostly irrelevant, for example. |
| 23:06 | <annevk> | sure |
| 23:06 | <Hixie> | AryehGregor: well as an example, i intend to work with ms2ger so that the common text in dom core and html actually is merged in from one common source |
| 23:07 | <Hixie> | annevk: ok dreamhost said they can do this co-hosting thing |
| 23:07 | <AryehGregor> | I mean yeah, let's all use the same stylesheets and same basic approach and not use conflicting terminology or anything, but if I want to xref everything as direct links and not have a references section, or want to xref more or fewer common terms than other people, or want to use <ul> instead of <ol> because otherwise people will report bugs using ever-changing step numbers that aren't in the source, I don't think we should worry about it. |
| 23:07 | <annevk> | sweet, so I just claim a subdomain and it works? |
| 23:07 | <annevk> | i guess that would be pretty bad |
| 23:07 | <AryehGregor> | Hixie, why isn't it just in DOM Core and referenced from HTML? |
| 23:08 | <annevk> | not everything is needed for DOM Core |
| 23:08 | <Hixie> | annevk: we have to both agree |
| 23:08 | <Hixie> | annevk: and then tell them |
| 23:08 | <Hixie> | annevk: so what do you want, dom.spec.whatwg.org? |
| 23:08 | <Hixie> | AryehGregor: because that makes HTML a huge pain to read |
| 23:08 | <AryehGregor> | Hixie, how so? |
| 23:09 | <Hixie> | AryehGregor: i don't have the details at hand, but basically html right now says "a, b, c, d" and dom core says "b, d" and each of these is like one sentence long |
| 23:09 | <annevk> | Hixie, that makes the most sense I think |
| 23:09 | <Hixie> | AryehGregor: it takes just as much to point the reader to the other spec as to just say it |
| 23:10 | <Hixie> | AryehGregor: plus it gets confusing having a and c in one spec yet have to read another for b and d when they are all so similar concepts |
| 23:10 | <AryehGregor> | Yeah, for minor things like that duplication makes sense. |
| 23:10 | <Hixie> | none of this is anything but definitions and so on |
| 23:10 | <AryehGregor> | But if there's a huge chunk that's orthogonal to the rest of the spec, I don't think it makes sense to try having it be part of the same spec. |
| 23:11 | <Hixie> | i took out all those bits yesterday |
| 23:11 | <Hixie> | (execCommand() isn't orthogonal though) |
| 23:11 | <AryehGregor> | Unless the same person winds up editing it. Obviously if you wind up editing the editing stuff, then you'll want to merge it into your spec. |
| 23:12 | <AryehGregor> | How is it not orthogonal? |
| 23:12 | <Hixie> | it's HTML's editing model |
| 23:13 | <Hixie> | it's less orthogonal than, say, PeerConnection, which i also think belongs in HTML |
| 23:14 | gsnedders | thinks anything that has only one-way references should be a separate spec |
| 23:15 | <Hixie> | annevk: what's your dreamhost id? |
| 23:15 | <Hixie> | annevk: or dreamhost e-mail address |
| 23:15 | <Hixie> | annevk: i'll cc you |
| 23:23 | <annevk> | annevankesteren⊙gc |
| 23:24 | <annevk> | yay for dglazkov |
| 23:24 | <dglazkov> | I win! |
| 23:24 | <dglazkov> | what did I win? |
| 23:24 | <annevk> | writing the spec |
| 23:25 | <dglazkov> | that doesn't seem like a win |
| 23:26 | <dglazkov> | spec-writing for me is like squeezing water out of a stone. |
| 23:28 | <hober> | AryehGregor: yes, sorry, will get back to you today or tomorrow on that |
| 23:29 | <AryehGregor> | hober, thanks. |
| 23:33 | <Hixie> | foolip_: if things are marked with class=impl incorrectly, you can file that as a regular bug in bugzilla |
| 23:33 | <foolip_> | Hixie, ok, so class=impl is what it is |
| 23:33 | <foolip_> | will do |
| 23:33 | <annevk> | dglazkov, haha |
| 23:34 | <annevk> | dglazkov, without a spec it's kind of hard to see how Shadow DOM differs from XBL |
| 23:35 | <annevk> | and how it works :) |
| 23:35 | <hober> | AryehGregor: which list did that email go to? |
| 23:35 | <dglazkov> | annevk: writing this up now, btw |
| 23:35 | <AryehGregor> | hober, public-html, since I wanted feedback from Microsoft. |
| 23:36 | <annevk> | I have a vague idea and wonder how closely it needs to be defined together with DOM Core given how it affects that, but I'd like to know a little more |
| 23:39 | <annevk> | oh yes |
| 23:39 | <annevk> | done: http://dvcs.w3.org/hg/domcore/raw-file/tip/Overview.html#concept-node-insert |
| 23:40 | <annevk> | Ms2ger is sleeping already of course, but he can look tomorrow |
| 23:42 | <Hixie> | foolip_: so this algorithm was left in because it's used to define the validator requirements |
| 23:42 | <Hixie> | foolip_: any suggestions on getting around that? |
| 23:43 | <foolip_> | foolip_, well, I just implemented the validator bits and I wasn't looking at the developer version, so I don't think it's a practical problem |
| 23:43 | <foolip_> | you could try to express the validity constraints in prose instead, as it's rather obscure right now |
| 23:44 | <foolip_> | oops, talking to myself :) Hixie ^ |
| 23:44 | <Hixie> | yeah, dunno |
| 23:45 | <foolip_> | the best I could come up with for an error message for "microdata error" was "The itemref attribute contained redundant references", which is neither completely true nor clear |
| 23:46 | <Hixie> | yeah coming up with good prose was hard, which is why the spec says what it does |
| 23:46 | <Hixie> | i can just hide these requirements from the developers version i guess |
| 23:46 | <foolip_> | that would solve the problem at hand :) |
| 23:58 | <benschwarz> | oh hai you two |
| 23:58 | <Hixie> | foolip_, benschwarz: fixed |
| 23:59 | <foolip_> | benschwarz, quick, make it work :P |
| 23:59 | <benschwarz> | I just saw. |