| 06:46 | <Von_Davidicus> | I got an interesting question today, and I thought I'd ask everyone here, since I', having difficulties answering it: "What does XML + XSLT do that XHTML cannot"? |
| 06:48 | <othermaciej> | if the context is displaying something in a web browser, then the correct answer is "nothing" |
| 06:51 | <Ms2ger> | Turn you insane |
| 06:54 | <Ms2ger> | wycats, anything in particular you wanted to know? |
| 06:58 | <Von_Davidicus> | What other contexts has XSLT been used in, then? |
| 07:00 | <Ms2ger> | Dunno, the Fukushima reactor? |
| 07:06 | <karlcow> | Von_Davidicus: it can be used sometimes for displaying an XML document in another form in the browser: RDF file as XHTML or HTML, XML feed, etc. but this is not used a lot. |
| 07:07 | <karlcow> | The draconian mode of XML has a nice side effect *for me*, which is to check without too much effort if the content is malformed. |
| 07:20 | <Von_Davidicus> | I like that part of it, myself. |
| 07:22 | <hsivonen> | Von_Davidicus: if the context of the question is displaying stuff in the browser, then, as othermaciej said, there's nothing XML+XSLT can display in a browser that XHTML arriving over HTTP cannot. |
| 07:22 | <hsivonen> | Von_Davidicus: so what XML+XSLT allows you to do is to move CPU burn from your server to the user's device |
| 07:23 | <webben> | You can likely get a validator plugin for your browser to tell you whether content is malformed … |
| 07:23 | <hsivonen> | Von_Davidicus: so you get to burn the battery on the user's device |
| 07:24 | <webben> | Although nowadays people are more likely to push JSON to an HTML frontend page that uses JS to convert the JSON to more HTML, rather than XSLT. |
| 07:25 | <hsivonen> | webben: yeah. I'm sure phone users who are low on battery appreciate designs that move processing to the edge of the network |
| 07:26 | <Von_Davidicus> | I see. Hm. |
| 07:26 | <Von_Davidicus> | How good are phones at displaying graphics on webpages? |
| 07:27 | <othermaciej> | there's plenty of ways to burn battery |
| 07:27 | <othermaciej> | complex CSS can be compute-intensive too |
| 07:27 | <othermaciej> | and sending more data means having the radio on longer |
| 07:27 | <othermaciej> | but generally, client-side XSLT tends to be a particularly inefficient way of doing things |
| 07:27 | <tantek> | certainly seems to be more CPU intensive |
| 07:28 | <zcorpan> | and delays rendering, doesn't it? |
| 07:28 | <tantek> | incremental rendering? who needs that? ;) |
| 07:28 | <zcorpan> | normal XHTML can be rendered incrementally |
| 07:28 | <tantek> | (zing) |
| 07:29 | <tantek> | well here's a question about what can x do in a browser that y cannot |
| 07:29 | <tantek> | what can XHTML do in a browser that HTML as well-formed XML cannot? |
| 07:30 | <tantek> | (per http://tantek.com/2010/302/b1/xhtml-dead-long-live-xml-valid-html5 ) |
| 07:30 | <othermaciej> | it can give you a yellow screen of death |
| 07:31 | <tantek> | oh, like http://www.flickr.com/photos/tantek/5126607972/ ? |
| 07:34 | <hsivonen> | It's nice how "yellow screen" has been abstracted so that a pink box counts as yellow screen |
| 07:35 | hsivonen | always has to look up the order of the arguments of strcpy :-( |
| 07:37 | <jgraham> | It allows you to be the subject of browser developer wrath |
| 07:37 | <jgraham> | Although I guess HTML has a number of ways to do that |
| 07:42 | Von_Davidicus | wishes someone, somewhere, would say XSLT is actually *useful* |
| 07:43 | <webben> | Von_Davidicus: Why? |
| 07:43 | <Ms2ger> | jgraham, document.all("foo") |
| 07:43 | <jgraham> | Von_Davidicus: davidc isn't in the channel at the moment, so he is likely to have missed his cue |
| 07:43 | <Von_Davidicus> | Because I spent so much blasted time experimenting with it! |
| 07:43 | <webben> | Von_Davidicus: Did you find it useful? |
| 07:44 | <hsivonen> | Von_Davidicus: if you know XSLT well and need to convert between two XML vocabularies that are not exactly isomorphic but not completely different, either, XSLT can be very useful |
| 07:44 | <Von_Davidicus> | As a client-side processing language for XML files, yes. |
| 07:44 | <webben> | I guess DocBook is an example. |
| 07:44 | <webben> | Von_Davidicus: Well, then you can say it was useful :) |
| 07:44 | <karlcow> | Von_Davidicus: it is useful, but not necessary in the context of live transformation in the browser |
| 07:46 | <hsivonen> | Opera 12 release process reminds me of Firefox 4. I wonder if Opera will move to rapid release after 12... |
| 07:50 | <hsivonen> | Whoa. StatCounter says Silk 1.0 exceeds Opera 11.6 usage in the U.S.! |
| 07:50 | <karlcow> | hsivonen: Opera just synchronized again with the year! |
| 07:50 | <karlcow> | the next one will be 13 in 2013 |
| 07:50 | hsivonen | didn't expect to see Silk on the charts |
| 07:50 | <karlcow> | :p |
| 07:57 | <hsivonen> | From the outside at least, Opera 12 looks like Firefox 4 in the sense that it has gotten stuck into over-long beta. Presumably there's something holding back release but it means that all the good stuff that's done isn't reaching users. |
| 07:57 | <hsivonen> | the sort of situation rapid release is meant to fix |
| 08:32 | Von_Davidicus | dances. |
| 08:32 | <odinho> | o/-< o\-< |
| 08:33 | <Von_Davidicus> | My assignment is complete. |
| 08:34 | <Von_Davidicus> | I think I mentioned earlier that I was doing an assignment for an HTML class to build a series of webpages--and instead of doing them in HTML, I've done them in XML + XSLT. |
| 08:41 | <Von_Davidicus> | Anyways, the assignment is done--hopefully I get my point across. And for your enjoyment, the mess starts here: http://www.mrinitialman.com/Coding/D2Runes/Words/ancients_pledge.xml and http://www.mrinitialman.com/Coding/D2Runes/Runes/runeEl.xml |
| 08:48 | <zcorpan> | Von_Davidicus: http://annevankesteren.nl/2004/01/xml-versus-xhtml |
| 08:50 | <charlvn> | yeah i don't know how this is in any way new, we worked past this like 8 years ago |
| 08:52 | <charlvn> | afaik the conclusion was that server-side xslt is cool but client-side is a fail, unless anything changed since then? |
| 08:52 | <Von_Davidicus> | This isn't really meant to be new or edgy--but my HTML teacher has thus far refused to believe that I've done this before (I *have* to take an introductory course in HTML at college, dems da rules), so hopefully something overcomplicated like this will get him to reconsider. |
| 08:54 | <Von_Davidicus> | I -have- done a webpage in HTML5; a professional one: http://sheepmaster.zzl.org/ Some of the differences between HTML 4.01 and HTML 5 forced me to quite being so lazy with my JavaScript. :D |
| 08:54 | <charlvn> | to be honest i don't entirely "get" how this is relevant to the whatwg in 2012 but i guess it's cool as an academic exercise |
| 08:54 | <jgraham> | Von_Davidicus: It might make them believe that you need to be retaught the basics |
| 08:55 | <charlvn> | yeah exactly |
| 09:00 | <Von_Davidicus> | I thought it was fun. |
| 09:05 | <MikeSmith> | so apparently because of a problem with the way we have the W3C bugzilla instance set up, when you have several bugzilla tabs open, it's possible that a comment you submit for one bug in on tab ends up wrongly getting added to a bug open in another tab |
| 09:05 | <othermaciej> | that sounds extremely bogus |
| 09:06 | <annevk> | how is that possible? |
| 09:06 | <annevk> | named windows? |
| 09:07 | <odinho> | Ahhh. Just like some bank websites. I love that. |
| 09:07 | <odinho> | They keep state using cookies with an ID I think, and keep state server side, -- so when you're suddenly in another place it messes up things pretty badly. |
| 09:07 | <Ms2ger> | Sounds like a browser bug to my neighborhood bugzilla developer |
| 09:08 | <odinho> | (I was not speaking of bugzilla btw, -- but evil ways of doing web services, which has frustrated me to no end before :P) |
| 09:08 | <Ms2ger> | "when you submit the comment, the bug id is fetched from the form, not from some other location" |
| 09:09 | <MikeSmith> | othermaciej, annevk - report came from a user who had observed the problem |
| 09:11 | <MikeSmith> | Ms2ger: ok |
| 09:11 | <MikeSmith> | I will let the systems team know |
| 09:12 | <MikeSmith> | I think they had been thinking it had something to do with their mirroring setup |
| 09:27 | <MikeSmith> | foolip: systems team is doing a DB migration right now |
| 09:27 | <foolip> | MikeSmith, thanks, that explain it |
| 09:27 | <foolip> | s |
| 09:27 | <MikeSmith> | it'll probably be a couple more hours at least before they're done |
| 09:28 | <MikeSmith> | I'll try to remember to ping you once they are |
| 09:28 | <foolip> | MikeSmith, thanks! |
| 09:31 | <Philip`> | Von_Davidicus: From your page: xsi:schemaLocation="http://www.mrinitialman.com/Coding/D2Runes/RuneWords/ ../DataFiles/Runewords.xsd" |
| 09:31 | <Philip`> | Von_Davidicus: Surely that's broken (and invalid?) because of the space? |
| 09:32 | <Von_Davidicus> | Oh, thanks for the heads-up. I forgot to edit my files after I changed some directory names. Beg pardon. |
| 09:34 | <Philip`> | Seems a bit odd having all this unwieldy machinery for validation and then letting through errors like that :-p |
| 09:39 | <Von_Davidicus> | Which validator did you use, that I may recheck my coding? |
| 09:40 | <Philip`> | I used my eyes |
| 09:44 | <Von_Davidicus> | Hmmm.... *Tries to remember how to set up the schema location.* |
| 09:45 | jgraham | wonders which schema format Philip`'s eyes use |
| 09:47 | <charlvn> | eye-based validation... i wonder if there is a spec for that :) |
| 09:50 | <Von_Davidicus> | Nope, that space is supposed to be there. But for some reason, it's not validating. I'll have to figure out why. |
| 09:50 | <webben> | the space is supposed to be there? |
| 09:50 | <webben> | seems very odd |
| 09:56 | <MikeSmith> | Von_Davidicus: maybe you need to make it tab instead |
| 09:56 | <Von_Davidicus> | Yeah, it separates the namespace from the filename of the schema. |
| 09:56 | <MikeSmith> | or maybe it's a non-breaking space |
| 10:00 | <charlvn> | acording to my understanding it's any whitespace character http://www.w3.org/TR/xmlschema-0/#schemaLocation |
| 10:01 | <zcorpan> | each(im.getAttribute('srcset').split(','), function (def) { |
| 10:01 | <zcorpan> | https://github.com/davidmarkclements/Respondu/blob/master/R.js |
| 10:03 | <Philip`> | Von_Davidicus: Hmm, I guess I'm confused - isn't http://www.w3.org/TR/xmlschema-1/#xsi.schemaLocation defining that schemaLocation is a single URL (anyURI)? |
| 10:04 | <Von_Davidicus> | Not in any of the examples I've seen. |
| 10:06 | <webben> | Von_Davidicus: Oh i c. |
| 10:07 | <hsivonen> | so many responsive image emails |
| 10:07 | <hsivonen> | every time I feel I have something to say, I feel I have to catch up with the threads first |
| 10:07 | <hsivonen> | and then there's more to catch up with |
| 10:08 | Von_Davidicus | has even checked to make sure the schema URIs are correct |
| 10:10 | <charlvn> | Philip`: it's not entirely clear to me either but from my understanding it could be a list http://www.w3.org/TR/xmlschema-1/#variety |
| 10:11 | <Von_Davidicus> | Well, my files follow every example I've seen, so I've no idea what's going on. :( |
| 10:12 | <charlvn> | the example on http://www.w3.org/TR/xmlschema-0/#schemaLocation also shows this |
| 10:14 | <Von_Davidicus> | Oh, I was clicking a box I shouldn't have: "Check As Complete Schema"--the validator was treating my XML file as a schema, which it isn't. *Puts on dunce cap* |
| 10:14 | <Von_Davidicus> | With that box unchecked, things check out fine. :) |
| 10:15 | <Von_Davidicus> | And another validator doesn't do custom schemas. So, all is well. |
| 10:20 | <Von_Davidicus> | I have a question about the HTML5 DOM--something I found a pain in HTML 4.01 |
| 10:20 | <Von_Davidicus> | In HTML5, if you leave out the <tbody> tags, does that element still exist in the DOM? |
| 10:20 | <webben> | yes. |
| 10:21 | <Von_Davidicus> | What about XHTML5? |
| 10:21 | <webben> | the XML serialization has no implied tags |
| 10:21 | <Von_Davidicus> | Okay. |
| 10:22 | <Von_Davidicus> | Boy, were tables a trick to script for. It took me ages to figure out what was happening. >< |
| 10:24 | <Von_Davidicus> | Ladies and Gents, I am headed out. :) |
| 11:03 | <odinho> | hsivonen: Yes. I participated heavily in the beginning, but felt that it didn't really help, because the ones replying didn't really read and understand it anyway... |
| 12:20 | <david_carlisle> | jgraham: Just for you: XSLT is actually *useful* |
| 12:21 | Ms2ger | sniggers |
| 12:25 | jgraham | just came acrss an old copy of "Dr Dobbs Journal" from like 2000 or so containing a Microsoft advert featuring someone from Ask Jeeves explaining how XML in visual studio was helping them scale to meet business needs |
| 12:25 | <jgraham> | It was hillarious |
| 12:26 | <jgraham> | No justification of why XML solved any problems |
| 12:26 | <jgraham> | Just the magic letters X M and L were assumed to be a selling point |
| 12:27 | <david_carlisle> | jgraham: well that's always the way. json or html(5) or whatever or sticking "java" at the front of "javascript", people like to have buzzwords |
| 12:30 | <jgraham> | Indeed. I look forward to the future when people laugh at things that talk about "mobile-first responsive HTML5" or whatever today's buzzword bingo entries are |
| 12:31 | <gsnedders> | Web 2.0-enabled HTML5 mobile interface featuring WebGL! |
| 12:33 | <Philip`> | We should fight against the buzzwordification of pragmatic technologies by giving them names that cannot possibly be used with a straight face |
| 12:34 | <Philip`> | Rename HTML5 to something like "SynergML" and then nobody can use it unironically |
| 12:34 | <jgraham> | Like AJAX? |
| 12:34 | <Philip`> | "ECMAScript" was a good one |
| 12:35 | <gsnedders> | W3CML |
| 12:35 | <gsnedders> | WHATML |
| 12:35 | <jgraham> | Really should have gone with the task force in that case |
| 12:39 | <odinho> | I said "responsive" "mobile-first" and a few others at a party, annevk and richt laughed at it IIRC. So given the right people we're already there! |
| 12:40 | <jgraham> | Wait, what? |
| 12:40 | <jgraham> | I don't understand |
| 12:40 | <jgraham> | "party"? |
| 12:43 | <odinho> | jgraham: -> 14:32 < jgraham> Indeed. I look forward to the future when people laugh at things that talk about "mobile-first responsive HTML5" or whatever today's buzzword bingo entries are |
| 12:43 | <odinho> | jgraham: Or is it the party that's hard to understand? :P It's was Lachy's birthday party. |
| 12:47 | <Ms2ger> | odinho, shh, jgraham wasn't invited |
| 12:48 | <odinho> | oh f... |
| 12:49 | <gsnedders> | I guess that's what he gets for not having FB. |
| 12:49 | <odinho> | I didn't either, but I was in the same office as Lachy, so it was OK. I guess I'll be missing out now though... |
| 13:51 | <zcorpan> | // Time in seconds within which the media is seekable. |
| 13:51 | <zcorpan> | var seekableEnd = audio.seekable.end(); |
| 13:51 | <zcorpan> | http://html5doctor.com/html5-audio-the-state-of-play/ |
| 13:52 | <zcorpan> | maybe we should give up and make the argument optional |
| 14:07 | <zewt> | seems wrong that a 5. isn't a "valid floating point number" |
| 14:37 | <foolip> | MikeSmith, still migrating? |
| 14:38 | <MikeSmith> | foolip: ah! |
| 14:38 | <MikeSmith> | no, should be done now |
| 14:38 | <foolip> | ok, I'll try filing that issue again |
| 14:38 | <MikeSmith> | yup |
| 14:39 | <foolip> | MikeSmith, it works now! |
| 14:39 | <MikeSmith> | super |
| 14:43 | <david_carlisle> | Philip`: (schemaLocation is a space separated list of URI (that's one of it's problems) the other one (in general) is that every other URI points at an XSD schema, and in this forum the remaining problem is that the remaining half of the URI are namespaces. Apart from those three minor issues. It's brilliant. |
| 14:44 | <david_carlisle> | its |
| 14:57 | <Philip`> | david_carlisle: Am I missing some part of the spec where schemaLocation is defined as a space separated list, instead of as a single URI, other than examples and notes? |
| 14:57 | <Philip`> | (Not that it actually matters; I'm just curious about to what extent I'm misunderstanding stuff) |
| 15:01 | <Philip`> | Oh, I think see something I didn't see |
| 15:03 | <Philip`> | namely that http://www.w3.org/TR/xmlschema-1/#xsi.schemaLocation says "{variety} list" |
| 15:04 | <Philip`> | so it's less confusing than I thought |
| 15:04 | <Philip`> | so never mind |
| 16:00 | <gsnedders> | I wonder what a reimpl of the jQuery API targetting just nightly builds of browsers would look like. |
| 16:00 | <gsnedders> | (wrt the shrinking libraries thread) |
| 16:22 | <dglazkov> | good morning, WhatWg! |
| 16:27 | <Ms2ger> | Good, dglazkov |
| 16:46 | <MikeSmith> | david_carlisle: you around? |
| 16:46 | <MikeSmith> | I can push the mathml3 validator updates any time |
| 16:47 | <MikeSmith> | but I would like to have help with testing it after I do |
| 16:47 | <MikeSmith> | to make sure I get everything pushed |
| 16:47 | <MikeSmith> | the instance I pointed you to before was code straight from my workspace |
| 16:48 | <MikeSmith> | and experience tells me that sometimes I don't manage to get everything pushed that's needed |
| 16:50 | <david_carlisle> | MikeSmith: yes |
| 16:51 | <MikeSmith> | OK |
| 16:51 | <MikeSmith> | I will give it try now |
| 16:51 | <david_carlisle> | ooh OK I'll stay here for a bit then:-) |
| 16:51 | <MikeSmith> | I don't have the most efficient system for deploying |
| 16:51 | <MikeSmith> | there are 5 validator hosts and I need to ssh into each an manually do updates |
| 17:09 | <Ms2ger> | MikeSmith, dvcs.w3.org appears down |
| 17:09 | <MikeSmith> | Ms2ger: thanks, checking ow |
| 17:09 | <MikeSmith> | *now |
| 17:09 | <Ms2ger> | Ta |
| 17:10 | <MikeSmith> | "Restarting web server: apache2 ... waiting ............." |
| 17:11 | <Ms2ger> | Reported back |
| 17:11 | <MikeSmith> | workign now |
| 17:11 | <MikeSmith> | yeah |
| 17:11 | <MikeSmith> | if it becomes not available again please ping me |
| 17:11 | <Ms2ger> | Will do, thanks |
| 17:23 | <Ms2ger> | Philip`, failed https://www.w3.org/Bugs/Public/show_bug.cgi?id=17141 for you |
| 17:47 | <jgraham> | Doesn't allowing min-* and max-* in srcset significantly complicate the processing |
| 17:47 | <jgraham> | ? |
| 17:48 | <jgraham> | For example at the moment it isn't possible to have gaps |
| 17:49 | <jgraham> | I guess you could use the rule that the src image is used if there's no other candidate |
| 17:49 | <TabAtkins> | jgraham: Yeah, it does. I suspect it's worthwhile, though. |
| 17:51 | <jgraham> | I'm not sure it's not worthwhile, it just takes a conceptually simple model and makes it less simple :) |
| 17:52 | <jgraham> | s/sure/saying/ |
| 18:19 | <david_carlisle> | im back:-) |
| 18:31 | <TabAtkins> | david_carlisle: Yo, I had a question for you a few weeks back. |
| 18:32 | <TabAtkins> | Does MathML have a good definition of "bounding box" of its internal elements? Content: CSS Images 4 has an element() function that lets you use an element as an image. All it needs to be able to do so is the bounding box of the element, so it knows what size the image is. |
| 18:32 | <TabAtkins> | For HTML and SVG this is pretty simple, but I'm not familiar enough with MathML's rendering model to know if it has such a concept. |
| 18:40 | <david_carlisle> | TabAtkins: Probably good enough see perhaps http://www.w3.org/Math/draft-spec/mathml.html#chapter3_presm.mpadded the wording in MathML is a bit wooly as mathml doesn't enforce a css style box model, so eg mathematica or tex or whatever can render it, but you don't care about that in a mathml in html context |
| 18:42 | <TabAtkins> | david_carlisle: Okay, so you do refer to "the bounding box of its content", but that term isnt' linked. |
| 18:44 | <david_carlisle> | TabAtkins: "linking" bit modern for us old timers you know. You should be grateful it's not fixed width plain text ascii:-) |
| 18:45 | <TabAtkins> | Tell that to TimBL twenty years ago. ^_^ |
| 18:46 | <david_carlisle> | TabAtkins: Probably the nearest is the pictures in the next section http://www.w3.org/Math/draft-spec/image/mpadded-resize.png but basically mathml assumes every construct _has_ a bounding box but doesn't always specify what that is as there is quite a bit of leeway in how constructs are laid out. But for element() taht should be OK shouldn't it? |
| 18:47 | <TabAtkins> | Not really. :/ It's not really acceptable if browsers define bounding boxes differently for the same markup. |
| 18:49 | <david_carlisle> | TabAtkins: maybe I misunderstood, but don't you just need to access the bounding box of the markup as rendered in the system applying the function. The bounding boxes will differ on different systems anyway for font reasons if nothing else |
| 18:51 | <david_carlisle> | even something as simple (switching to tex notation) as P_1 will have different bounding boxes on different systems even using the same fonts as different systems will have different ideas about whether to kern the 1 under the P |
| 18:52 | <TabAtkins> | It's fine for them to end up rendering differently, as long as there's a strict definition of what should actually be rendered. |
| 18:52 | <david_carlisle> | TabAtkins: I don't understand. |
| 18:53 | <TabAtkins> | So, like, for SVG the bounding box is "the smallest axis-aligned rectangle that includes the strokes of all the content within the element". |
| 18:54 | <TabAtkins> | For HTML it's the element's border box, modified by border-iamge. |
| 18:59 | <david_carlisle> | http://www.w3.org/Math/draft-spec/mathml.html#appendixd_dt-blackbox and http://www.w3.org/Math/draft-spec/mathml.html#appendixd_dt-blackbox and #appendixd_dt-boundingbox plus the previously mentioned mpadding description are what you have. That seems as definite as the svg definition you gave. If css does need something more proscriptive here we'd certainly be open tp a clarifying note on... |
| 18:59 | <david_carlisle> | ...mathml/css interaction but hopefully there is enough in the spec |
| 19:02 | <TabAtkins> | The "black box" might work. I'm not sure if it's equivalent to SVG's definition, as the definition of how far a stroke extends is rigorous in SVG (it's part of the geometry used for, e.g. pointer events). |
| 19:02 | <david_carlisle> | Although I guess you're not going to like the paragraph immediately before section 3.2 |
| 19:02 | <TabAtkins> | Is the black box what is used for matching :hover, for example? |
| 19:04 | <david_carlisle> | TabAtkins: The only thing MathML says about CSS (so :hover) is that if you have a system supporting mathml and css they may interact. http://www.w3.org/Math/draft-spec/mathml.html#chapter6_world-int-style |
| 19:04 | <TabAtkins> | Yay. |
| 19:04 | <TabAtkins> | ^_^ |
| 19:08 | <jgraham> | TabAtkins: It's not that surprising that the interaction of CSS and MathML is poorly defined |
| 19:09 | <TabAtkins> | jgraham: Oh, definitely. It's just problematic for me. |
| 19:09 | <jgraham> | I mean, the interaction of CSS with itself isn't always that well defined ;) |
| 19:09 | <jgraham> | TabAtkins: The solution is probably to fix MathML :) |
| 19:10 | <TabAtkins> | Yes, here the "right" solution would be for MathML to explicitly say what the bounding box of its elements are, so that other CSS things like :hover and element() could just hook into that. |
| 19:10 | <jgraham> | (I mean it would be surprising if it was the exception to the rule that no spec is good enough) |
| 19:13 | <david_carlisle> | TabAtkins: I'm not sure that it's possible to say what the bounding box is without first fixing a particular layout model such as css box model. Clearly when MathML is implemented natively in a browser it is rendered using a (possibly extended) version of the css layout rules so you can ask the question but what would an "exact" definition of teh bounding box of P_1 be in a purely mathml context? |
| 19:14 | <TabAtkins> | david_carlisle: I disagree. For example, SVG doesn't use the CSS layout model, but it has a precise definition of the bounding box of any element. |
| 19:14 | <TabAtkins> | You can state the definition in terms of whatever layout model you use, as long it's an answer that will be implementable compatibly across browsers. |
| 19:15 | <TabAtkins> | So that, for example, you dont' end up with one browser doing a tight box around glyphs, but another including the spacing on either side. |
| 19:15 | <othermaciej> | TabAtkins: is there anything I should do to ensure that my comment about Media Queries mostly not defining anything gets addressed by the WG? |
| 19:15 | <othermaciej> | TabAtkins: I heard that there is a bug tracker but the CSS WG Chair will get mad at people who use it |
| 19:15 | <TabAtkins> | othermaciej: >_< >_< >_< I yelled at Ms2ger about saying that. |
| 19:15 | <othermaciej> | TabAtkins: is it actually ok to use it? |
| 19:16 | <TabAtkins> | File bugs freely. We don't want *conversation* on the bug tracker, is all - those should happen on the list. |
| 19:16 | <othermaciej> | since no one responded to my email there seems to be little risk of a conversation |
| 19:16 | <david_carlisle> | TabAtkins: yes but it defines a precise layout as that's it's purpose. <msub> in MathM< means a subscript and if you define teh bounding box of P_1 in terms of teh extent of black pixels then TeX is doomed as it doesn't know which pixels are black. If you define it in terms of font metrics it's possible but may be not what you want either. |
| 19:17 | <TabAtkins> | othermaciej: Yeah, it's fine. If discussion is necessary there's the email thread already started, or someone can start a new one. But go ahead and file a bug. |
| 19:18 | <TabAtkins> | david_carlisle: Out of curiosity, why doesn't TeX know that? |
| 19:18 | <jgraham> | david_carlisle: It seems clear that, at the very least, MathML-on-the-web should have a precisely defined layout model |
| 19:18 | <jgraham> | I doubt anyone cares very much about precise layout compatibility with non-web systems |
| 19:18 | <TabAtkins> | But yes, what jgraham says. The limitations of things that aren't browsers shouldn't hold back defining useful things for browsers. |
| 19:18 | <david_carlisle> | TabAtkins: TeX knows nothing about glyph shapes, each character is just 4 integers extracted from the font metrics (height depth width and italic correction) |
| 19:20 | <david_carlisle> | jgraham: as I said above if we restrict to mathml-in-acss-based-rendering model there is definitely more that could be said, and if it neess saying we could drum up a WG not eto say it |
| 19:20 | <jgraham> | Please don't make a Note :) |
| 19:20 | <TabAtkins> | david_carlisle: The rendering engine for TeX certainly knows about black pixels, though. But it's irrelevant anyway, as TeX has no :hover or element() or similar. |
| 19:20 | <jgraham> | The most pointless of all publication types |
| 19:20 | <TabAtkins> | david_carlisle: Yes, please don't make a Note. It needs to be normative, and ideally contained within the main spec. |
| 19:21 | <david_carlisle> | TabAtkins: there isn't _a_ rendering engine for TeX. And even if yiou restrict to a particular one the point is that you can't access it from within TeX. |
| 19:22 | <TabAtkins> | david_carlisle: Exactly, so it doesn't matter for TeX what the definition is. ^_^ |
| 19:22 | <TabAtkins> | (I wasn't implying a single rendering engine for TeX. It's of course rendered by many engines, just as web content is.) |
| 19:23 | <david_carlisle> | TabAtkins: well I think formally we are not chartered to do another rec-track thingy at present, but that's just bureaucracy. If the words get written it could presumably be made a normative document by some combination of math/css/htmlWG |
| 19:24 | <TabAtkins> | You need a charter for errata? |
| 19:24 | <TabAtkins> | (This qualifies as MathML3 errata, I would think.) |
| 19:25 | <david_carlisle> | TabAtkins: No but errata need an error and this is feature creep:-) |
| 19:25 | <TabAtkins> | I disagree. The draft currently defines that MathML can interact with CSS, but it does so incompletely. |
| 19:26 | <TabAtkins> | If you can't even reasonably respond to :hover, that's an error. ^_^ |
| 19:27 | <david_carlisle> | TabAtkins: If a document specifies something and does it wrong, that's an error. If a document explictly says that in order to support multiple rendering systems some aspects of rendering are left unspecified, then it is not an _error_ that they are unspecified. Nor is it surprising that users of one particular system decide to specify it for their case. |
| 19:37 | <jgraham> | It doesn't really matter what the Process is as long as the right text ends up in a document that implementors can use |
| 19:37 | <david_carlisle> | jgraham: yep |
| 19:37 | <jgraham> | Prefereably in a Rec. or something with similar properties |
| 19:40 | <david_carlisle> | TabAtkins: Out of interest what's the bounding box of <span>hello</span> ? that is can you point me to an exact wording somewhere in html spec (or any spec) that I could see what you want me to say for the bounding box of <mtext>hello</mtext> (which should be the same) |
| 19:44 | <tantek> | david_carlisle - that would be inline layout as defined in CSS2.1 |
| 19:45 | <tantek> | (re: the span) |
| 19:45 | <tantek> | and yes, one possible answer is for MathML to simply phrase everything in terms of defined CSS layout terms |
| 19:49 | <david_carlisle> | tantek: I'm looking at that(and Ive read it before:-) but where for example do I find if the width of the box is based on rendered pixels or on the font metrics so including the left sidebearing of h and right of o for example? (and similarly for height) as that seemed to be level of detail TabAtkins was asking for MathML |
| 19:50 | <tantek> | david_carlisle I believe that is defined where stated (CSS 2.1 inline layout) and used by various things such as painting the background, drawing borders etc. |
| 19:50 | <tantek> | (when such spans have backgrounds and borders) |
| 19:50 | <tantek> | f |
| 19:50 | <tantek> | r |
| 19:50 | <tantek> | om |
| 19:51 | <tantek> | from what I recall, that was quite a bit of work in the last few years of CSS 2.1 |
| 19:51 | <tantek> | figuring out inline layout details to get interop background/border rendering |
| 19:53 | <david_carlisle> | "inline layout" doesn't occur in the toc http://www.w3.org/TR/CSS21/cover.html#minitoc I could read the whole spec, or do you mean somewhere else? (clear;y there is a lot of detail about inline and layout but I don't see that but I'm not sure if I'm looking at the right spec:-) |
| 19:54 | <david_carlisle> | Ah found it.. CSS 2.1 says: The 'height' property does not apply. The height of the content area should be based on the font, but this specification does not specify how. |
| 19:55 | <david_carlisle> | hmmmmmm |
| 19:56 | <david_carlisle> | and goes on to say If more than one font is used (this could happen when glyphs are found in different fonts), the height of the content area is not defined by this specification. |
| 19:57 | <david_carlisle> | got to go, back later |
| 20:57 | <dglazkov> | MikeSmith: it's dead again. |
| 21:02 | <jgraham> | {insert Monty Python reference here} |
| 21:03 | <Ms2ger> | I'm not dead yet |
| 21:04 | <dglazkov> | http://downforme.org/is-dvcs.w3.org-down-today-for-everyone |
| 21:04 | <jgraham> | Ms2ger: YOu are an AI not a parrot |
| 21:04 | <jgraham> | Although it seems you are an AI that thinks it is a parrot |
| 21:05 | <Ms2ger> | Excuse me. |
| 21:05 | <Ms2ger> | I would like a fish license. |
| 22:09 | <splend> | Hi! Where can I download the specification for offline reading? |
| 22:16 | <edwardbc> | splend: http://developers.whatwg.org/offline.html |
| 22:18 | <splend> | edwardbc: Oh, I'm sorry I wasn't clear. I want to download the current HTML specification as file(s), so I can read it without a connection to Internet. |
| 22:20 | <edwardbc> | oh I see, can't help you with that (maybe print that to PDF?), but the whatwg site actually uses the offline manifest, so after the first load you should be able to open that same page without internet access |
| 22:20 | <gsnedders> | Or just save the file using your browser of choice. |
| 22:22 | <edwardbc> | indeed, heh |
| 22:26 | <splend> | gsnedders: Thanks for the suggestion, but I think the whole specification as a single HTML file is too much form my low-spec computer to handle easily. |
| 22:27 | <splend> | edwardbc: Printing to a PDF seems like a good suggestion – I think it will do fine. Thank you! |
| 22:30 | <edwardbc> | no problem, you can build a local copy as well: https://github.com/benschwarz/developers.whatwg.org |
| 22:30 | <edwardbc> | (until now I get that you meant the whole spec) |
| 22:32 | <splend> | Either is fine, whole or Web Dev Edition. |
| 22:32 | <splend> | I'll see if I can manage to built it on my own. Thank you! |
| 22:32 | <edwardbc> | :) |
| 23:15 | <MikeSmith> | dglazkov: should be working now |
| 23:16 | <dglazkov> | MikeSmith: \o/ |
| 23:16 | <MikeSmith> | I think the cause may be that the swap partition on that machine needs to be bigger |
| 23:20 | <benschwarz> | splend left, but if people want to try build the developer spec I can help them. |
| 23:20 | <benschwarz> | also, it uses application cache, so they can already read it offline |
| 23:33 | <ojan> | annevk, tantek: i'm about to file a bunch of bugs on chromium's fullscreen implementation and i want to double check that i understand the current state of the spec correctly |
| 23:33 | <tantek> | ok |
| 23:33 | <ojan> | annevk, tantek: for now, we've punted on the nested requestFullScreen situation by saying the requestFullScreen throws if you already have an element fullscreened |
| 23:33 | <ojan> | annevk, tantek: is that correct? |
| 23:34 | <tantek> | no, you should be able to do nested fullscreens |
| 23:34 | <tantek> | but I suppose an implementation could treat that as incremental implementation |
| 23:34 | <tantek> | with Mozilla we're implementing nested fullscreens |
| 23:34 | <ojan> | tantek: i guess i don't understand these lines from the spec then... |
| 23:34 | <ojan> | tantek: The context object's node document fullscreen element stack is not empty and its top element is not an ancestor of the context object. |
| 23:34 | <ojan> | A descendant browsing context's document has a non-empty fullscreen element stack. |
| 23:35 | <ojan> | tantek: from the 1st step of requestFullscreen |
| 23:35 | <tantek> | ojan, annevk will be faster at answering specific spec wording questions |
| 23:35 | <ojan> | tantek: saying to error |
| 23:35 | <tantek> | and I think it's sleeptime in his TZ right now |
| 23:36 | <tantek> | let me check - http://wiki.whatwg.org/wiki/Irc-people |
| 23:36 | <tantek> | hmm - no TZs for annevk |
| 23:36 | <tantek> | ojan - add yourself: http://wiki.whatwg.org/wiki/Irc-people |
| 23:36 | <ojan> | tantek: k. no worries. just curious, is mozilla implementing a stack of fullscreens? e.g. if you hit escape it removes you form the top-level fullscreened element only? |
| 23:36 | <tantek> | no escape pops all the fullscreens off |
| 23:36 | <tantek> | the only situation when it doesn't is when there are intermingled dialogs |
| 23:36 | <tantek> | which we don't implement yet |
| 23:36 | <tantek> | but the model is supposed to be: |
| 23:37 | <tantek> | e |
| 23:37 | <tantek> | s |
| 23:37 | <tantek> | cap |
| 23:37 | <tantek> | e |
| 23:37 | <tantek> | escape pops all the fullscreens off the stack until the next dialog element or there are no more fullscreens |
| 23:37 | <ojan> | tantek: hmm...apparenltly i don't have access to the whatwg wiki and need an administrator to create me an account? |
| 23:38 | <ojan> | tantek: interesting...so is there a way to just pop off the top element? |
| 23:38 | <tantek> | ojan - from API yes |
| 23:38 | <ojan> | tantek: ok...you'd have to provide UI for it in your page |
| 23:38 | <tantek> | however if the user hits the escape key, they're all supposed to pop off until the next dialog element |
| 23:38 | <tantek> | right |
| 23:38 | <tantek> | that's what sites do today |
| 23:39 | <tantek> | provide a fullscreen button |
| 23:39 | <tantek> | an |
| 23:39 | <tantek> | d |
| 23:39 | <tantek> | and an exit fullscreen button |
| 23:39 | <tantek> | (sorry for the extra lines, not sure what Colloquy is doing on my side) |
| 23:39 | <tantek> | for each level |
| 23:39 | <ojan> | tantek: that's true...but escape, in practice would only escape the top of the stack...although the reason is mostly accidental |
| 23:40 | <ojan> | tantek: the case i'm thinking of is having a fullscreened slideshow with a youtube video in it that you then fullscreen |
| 23:40 | <ojan> | tantek: i suppose if you did the slide show and youtube in flash, hitting escape would exit both |
| 23:40 | <tantek> | right |
| 23:40 | <ojan> | it's surprising how complicated fullscreen behavior is. |
| 23:41 | <tantek> | ojan - many new things on the web are apparently complicated |
| 23:41 | <tantek> | until you're used to their intricacies |
| 23:41 | <tantek> | a |
| 23:41 | <tantek> | n |
| 23:41 | <tantek> | and then you take them for granted |
| 23:41 | <ojan> | i suppose that's true of most computer science endeavors...once you really dig into it, all the problems are hard |
| 23:41 | <ojan> | or at least...harder than they look |
| 23:41 | <ojan> | tantek: anyways...thanks... that helps inform my bug-filing |
| 23:42 | <tantek> | ojan - no problem |
| 23:42 | <tantek> | the escape / fullscreen / dialog interaction is non-trivial |
| 23:43 | <tantek> | and it took a whole lunch in-person between myself, hober, Hixie, Tab to work it out. with printouts of use-cases and everything |
| 23:43 | <tantek> | I've also walked folks through the scenarios here on the Mozilla side in our security reviews. |
| 23:50 | <tantek> | ojan - for the wiki you need to ask Hixie for an account |
| 23:54 | <TabAtkins> | Or get annevk to do it for you, I think. |
| 23:55 | <TabAtkins> | tantek: Any idea why sometimes you end up sending multiple single-letter lines in irc? |
| 23:55 | <TabAtkins> | Oh, I see that you notice that already. |
| 23:55 | <TabAtkins> | Never mind. |
| 23:57 | <tantek> | sorry about that TabAtkins |
| 23:57 | <tantek> | I'm trying to track down the problem |
| 23:58 | <TabAtkins> | No problem. As long as you're aware of it I'm okay. |
