| 01:18 | <jwalden> | hm, is aside the correct element to use to mark up parentheticals, even if those parentheticals should be |display: inline|? |
| 01:22 | <Hixie> | jwalden: no, use Unicode characters U+0028 and U+0029 to mark up parentheticals. |
| 01:46 | <Hixie> | anyone got any opinions on bruce's e-mail regarding whether to use <figure> or <aside> for pullquotes? |
| 01:46 | <Hixie> | i guess <aside> is better |
| 02:18 | <jwalden> | Hixie: thanks; might be worth adding an informative sentence to <aside> mentioning that anti-use case |
| 02:19 | <Hixie> | added a comment, dunno if it'll confuse more than it helps, though |
| 02:21 | jwalden | looks |
| 02:23 | jwalden | waits for http://html5.org/tools/web-apps-tracker (presumably) to pick up the change |
| 07:05 | jwalden | reads the parts of the ffmpeg/chrome thread he hadn't read yet, merely seen as having happened, and is bemused at how quickly things went south when some people chose not to shut up |
| 07:06 | <jwalden> | frankly, I'm amazed it took so long to catch flame :-D |
| 07:28 | <hsivonen> | so predictable: http://twitter.com/jdowdell/status/2085821360 |
| 08:08 | <jgraham> | HTML5 fully speciying something and being silent on whether it is conforming or not seems difficult |
| 08:09 | <jgraham> | Unless HTML5 made no claims about document conformance at all, I suppose |
| 08:09 | <jgraham> | But that seems... unhelpful |
| 09:02 | <annevk42> | "To call out just one change, note that Snow Leopard now defaults to the same display gamma as Windows — 2.2 instead of 1.8." does this mean PNG now works the same everywhere? |
| 09:43 | <annevk42> | <di> ftw |
| 09:46 | <jgraham> | annevk42: I don't think so |
| 09:46 | <jgraham> | (the png thing) |
| 09:48 | <jgraham> | (at least I think that change alone doesn't solve the problem that hsivonen always brings up which is that png colours may not match CSS colours in browsers that have support for PNG colour profiles) |
| 09:48 | <annevk42> | http://hsivonen.iki.fi/png-gamma/ suggests it is about gamma |
| 09:50 | <annevk42> | it seems it was not much of a problem anyway anymore |
| 09:51 | <jgraham> | annevk42: I don't understand this much but there is also the colour space issue |
| 09:51 | <jgraham> | Oh wait, I am thinking about JPEGs although the same applies to PNG afaik |
| 09:54 | <hsivonen> | when browsers do ICC stuff, you can now enjoy the old png-gamma problem with JPEGs, too! |
| 09:55 | <hsivonen> | annevk42: the main old problem was that some old browsers didn't make PNG color "management" opt-in |
| 09:55 | <jgraham> | hsivonen: For JPEGs it makes much more sense since fewer people are using JPEGs as part of their site design |
| 09:56 | <hsivonen> | w00t! finally gamma sanity at Apple with the 2.2 default. |
| 09:56 | <hsivonen> | time to party. at last! |
| 09:56 | hsivonen | reads logs backwards |
| 09:57 | <hsivonen> | existing content and degrading gracefully FTW! |
| 09:57 | <hsivonen> | Mac-created content degrading gracefully on Windows, that is |
| 09:57 | <hsivonen> | and existing Windows-created content working on Mac |
| 10:16 | <Lachy> | hsivonen, PNG Gamma article of yours seems a little out of date. It's mentions browsers like Safari 2.0, IE5 Mac and Opera 7 as if they were still relevant |
| 10:16 | <Lachy> | if you have time one day, it would be nice if you could update it to discuss the current situation |
| 10:17 | <hsivonen> | Lachy: yeah, I should probably update it to say that the article discusses a problem that has now pretty much gone away as a browser interop problem but is now relevant as a photoshop/browser problem |
| 10:17 | hsivonen | has a lot of outstanding feedback on the article :-( |
| 10:18 | <Lachy> | yeah, I always had trouble getting photoshop to export PNGs in ways that played nicely with browsers |
| 10:19 | <hsivonen> | the great thing about OS X is that old releases fade into irrelevance much faster than old Windows releases |
| 10:19 | <Lachy> | I had much more luck with Fireworks, which somehow seems to get it right |
| 10:19 | <hsivonen> | so pre-Tiger Safari doesn't matter anymore |
| 10:20 | <Lachy> | so, is the best way to get PNGs with consistent colour to omit the colour profile information, and rely on current browsers matching the CSS colours? |
| 10:21 | <hsivonen> | Lachy: yeah. with unlabeled color for everything, CSS, bitmaps and plug-ins match |
| 10:21 | <Lachy> | good |
| 10:22 | <Lachy> | that would have been useful to know when I was doing web development a couple of years ago. |
| 10:22 | <hsivonen> | Lachy: you should have googled for my article! |
| 10:23 | <Lachy> | I did read it a long time ago, but didn't really understand it too well |
| 10:23 | <Lachy> | I just knew if I exported PNGs with Fireworks, they were right, and if I used Photoshop, they'd be wrong in some browser/platform combinations |
| 10:24 | <beowulf> | you can use software to remove the gamma black from pngs, that's what i do anyway |
| 10:25 | <Lachy> | I still don't really understand the concept of colour profiles too well. They seem like a hack to make poor quality hardware give good colour, instead of getting hardware vendors to produce better quality colour |
| 10:25 | <hsivonen> | most of the feedback I haven't addressed yet revolves around endorsing various removal tools |
| 10:26 | <Lachy> | the pngcrush command the article mentions seems good enough |
| 10:26 | <Lachy> | although, I haven't used it |
| 10:26 | <jgraham> | Lachy: define "good colour" |
| 10:27 | <beowulf> | i use pngthing, but i can't imagine it's any better worse than something else that does the same thing, considering |
| 10:27 | <Lachy> | jgraham, good colour = colour that looks like what it's supposed to look like |
| 10:29 | <jgraham> | Lachy: Define "what it's supposed to look like" |
| 10:31 | <sprasanna> | im trying to use "string".is_utf8? in my code but it gives false all the time... i dunno why it happens |
| 10:35 | <jgraham> | sprasanna: You probably would need to give more context. But at a guess you want a more ruby-related channel |
| 11:47 | <MikeSmith> | hsivonen: question about v.nu backend behavior: As far as I can see now, there's not currently a way to overwrite contents of the local-entities subdir with freshly downloaded copies, right? |
| 11:49 | <hsivonen> | MikeSmith: do you mean 3rd party ones? |
| 11:50 | <MikeSmith> | hsivonen: no, I mean for the case where the contents of the upstream resource might have changed |
| 11:51 | <hsivonen> | MikeSmith: but resources that aren't in the whattf repo? |
| 11:51 | <hsivonen> | MikeSmith: I'm also curious which upstream resources have changed |
| 11:52 | <MikeSmith> | hsivonen: stepping back, I'm thinking about it in the context of trying figure out the best way to handle running v.nu in "local only" .. so the upstream resources are the language-subtag-registry file, MicrosyntaxDescriptions file, Validator.nu_alt_advice file |
| 11:53 | <MikeSmith> | what I want to add is some way to have the build download copies of those and store them locally |
| 11:54 | <hsivonen> | ah. there's no code for that download action right now, but build.py should have parameters that can take file URL once downloaded |
| 11:54 | <MikeSmith> | hsivonen: so what I was thinking is, it'd be easiest to just use your existing entity-map.txt and downloadLocalEntities() mechanism |
| 11:55 | <MikeSmith> | but - |
| 11:56 | <MikeSmith> | I would want to make it possible for a user to make changes to the remote sources of those, and then re-run v.nu and incorporate those remote changes |
| 11:57 | <MikeSmith> | but I think if I use the existing entity-map.txt and downloadLocalEntities() mechanism as-is, then I can't do that |
| 11:57 | <hsivonen> | ok. perhaps there should be a build.py action for redownloading |
| 11:57 | <MikeSmith> | OK, I can add that |
| 11:57 | <hsivonen> | since downloadLocalEntities() never redownloads unless you delete the local files |
| 11:58 | <MikeSmith> | hsivonen: Yep. I'm just wondering if it'd be preferable to you for me to add that redownload mechanism, or just to add separate, new handling for the language-subtag-registry, MicrosyntaxDescriptions, Validator.nu_alt_advice stuff |
| 11:59 | <MikeSmith> | I mean, the upstream DTDs don't change, so I suppose it's a bit of a waste to re-download them |
| 12:00 | <hsivonen> | MikeSmith: it might be best to commit microsyntax and alt advice into svn and handle languag tags like 3rd-party dtds |
| 12:32 | <jgraham> | public-html totally needs the /topic from here to be appeneded to the end of every message |
| 12:33 | <jgraham> | It ould make it so much easier to follow |
| 12:47 | <hsivonen> | I don't buy it that accessibility were a topic that weren't tractable to empirical research methods |
| 12:50 | <jgraham> | hsivonen: Shell'y argument appeared to me to be "we should keep features that ere designed for accessiblity independent of their actual usefulness because their presence makes disabled people feel better even if they provide no tangible benefit" |
| 12:51 | <jgraham> | Which seems like a pretty poor way to design a language (I would prefer to provide features that have actual benefits) not to mention rather patronising |
| 12:55 | <Lachy> | hey, in WebStorage, it says "Multiple separate objects implementing the Storage interface can all be associated with the same list of key/value pairs simultaneously" -- http://dev.w3.org/html5/webstorage/#the-storage-interface |
| 12:56 | <Lachy> | Is the only way that can happen when there are two pages open from the same site simultaneously, either in separate tabs or frames? |
| 14:30 | jgraham | fails to find the microsoft ES5 testsuite |
| 15:55 | <zcorpan_> | Hixie: http://www.w3.org/Bugs/Public/show_bug.cgi?id=6880 |
| 15:57 | jgraham | has never got accessed denied on W3C bugzilla before |
| 15:59 | annevk42 | now wonders what that bug is about |
| 16:02 | <zcorpan_> | oh i didn't know it had limited access |
| 16:02 | <Lachy> | hmm, damn, I had no luck finding a bugmail archive that would have a copy of the report. I guess not all bugs go to a list like public-html-bugzilla |
| 16:03 | <zcorpan_> | it's an aria bug that's pending input from the html wg |
| 16:03 | <zcorpan_> | wonder how the html wg can provide input if html wg participants can't access the bug! |
| 16:03 | <jgraham> | Oh well good luck on getting input if no one in the WG can see the bug :) |
| 16:03 | <jgraham> | Oh you just said that :( |
| 16:05 | <annevk42> | WAI should really fix their transparency issues :/ |
| 16:06 | <Lachy> | oh, where did you find the bug mentioned? |
| 16:08 | <jgraham> | Isn't zcorpan in the wg or something? |
| 16:08 | <zcorpan_> | i'm subscribed to aria-ua-impl |
| 16:14 | <remysharp> | Hi - I'm looking for how to markup a slider in HTML5 (type="range") - but going by the documentation, I can't see anywhere where it says to use "range" as the type: http://www.whatwg.org/specs/web-apps/current-work/multipage/forms.html#range-state |
| 16:14 | <remysharp> | I was just wondering if there were docs/specs I had missed and you might be able to point me in the right direction |
| 16:17 | <annevk42> | in http://www.whatwg.org/specs/web-apps/current-work/#attr-input-type it defines how type attribute values map to states |
| 16:17 | <remysharp> | cheers. |
| 16:18 | <remysharp> | from a developers point of view - don't you think it might be good to have either a link to that page or a practical example implementation? |
| 16:18 | <remysharp> | I would have found that link sooner or later, |
| 16:18 | <remysharp> | but I'm thinking newbies to html. |
| 16:18 | <annevk42> | that link is present from your section |
| 16:18 | <annevk42> | type is a link |
| 16:18 | <remysharp> | ah - from the 'type' |
| 16:19 | <annevk42> | I think at some point Hixie will add more examples and fluff to make it more understandable |
| 16:19 | <remysharp> | okay, not sure that's obvious, but at least it's there - |
| 16:19 | <remysharp> | and good to hear that more examples will be added - definitely worth it for learning from. |
| 16:19 | <remysharp> | cheers. |
| 16:20 | <annevk42> | you can play with range in Opera/Safari fwiw |
| 16:20 | <remysharp> | I know ;-) |
| 16:36 | <a-ja> | anyone else noticing changes in <br /> in Opera 10 beta ? |
| 16:38 | <a-ja> | as in not starting a new line |
| 16:39 | <jgraham> | a-ja: Testcase? |
| 16:39 | <a-ja> | checking....think it's within an inline-block |
| 16:41 | <a-ja> | NSFW: http://stlrope.info/grue2/ |
| 16:44 | <jgraham> | a-ja: Is that a testcase? |
| 16:45 | <a-ja> | not a reduced one....it's an example |
| 16:46 | <jgraham> | OK |
| 16:46 | annevk42 | finds http://www.openscreenproject.org/ |
| 16:46 | <a-ja> | jgraham: witihin the vevent div's |
| 17:07 | gsnedders | needs to get an SSD for /… IO slows me down way too much. |
| 17:09 | <jgraham> | gsnedders: More than IRC? |
| 17:09 | <gsnedders> | jgraham: Yes, I run more than just IRC :P |
| 17:10 | <gsnedders> | (This is when starting everything up after booting up, which is a side-effect of having so much open) |
| 17:10 | <jgraham> | gsnedders: That's not what I meant :p |
| 17:11 | <gsnedders> | jgraham: Hey, I don't look at the sort of websites zcorpan does at work! |
| 17:17 | <Philip`> | gsnedders: Why care about disk IO when you can just get lots of RAM and everything should be magically cached? |
| 17:17 | <gsnedders> | Philip`: It still has to get from IO to RAM in the first place, and that is my problem :P |
| 17:18 | <Philip`> | gsnedders: That should happen about twice a year when you reboot your machine |
| 17:18 | <gsnedders> | Philip`: The problem is I do it closer to 52 times a year. |
| 17:18 | <Philip`> | gsnedders: Then stop doing that :-p |
| 17:29 | <jgraham> | gsnedders: Or stop worring about optimizing something that ony happens once a week |
| 17:29 | <jgraham> | *worrying |
| 17:40 | gsnedders | doesn't think Opera is meant to do that (i.e., crash) |
| 17:47 | <jgraham> | http://abcdefu.wordpress.com/2009/06/09/introducing-yip-an-unified-notification-system-for-the-web/ |
| 17:47 | <jgraham> | Seems prople want a standardised notification API |
| 17:47 | <shelleyp> | jgraham, it would be helpful that if you have a comment about something I write at the HTML WG mailing list, you respond to it there, instead of here |
| 17:47 | <shelleyp> | That way we have a discussion, rather than asides that pass each other in the internet night |
| 17:49 | <jgraham> | shelleyp: Sure I was planning to, but not till this evening |
| 17:49 | <shelleyp> | jgraham: cool, thanks |
| 17:49 | <jgraham> | I have other things to do fortunately / sadly (delete as required) |
| 17:49 | <jgraham> | :) |
| 17:50 | <shelleyp> | I understand, we're all overbooked |
| 17:51 | <jgraham> | shelleyp: FWIW the comment I was planning on making is basically "I don't think that solutions that make people feel good but don't actually work should be an acceptable design goal" |
| 17:51 | <jgraham> | (irrespective of whether @summary actually works) |
| 17:51 | <shelleyp> | jgraham: I don't want to get too far into here, I think the discussion should happen in the mailing list but... |
| 17:52 | <shelleyp> | My point was that measurements of success aren't always based on empirical observations |
| 17:52 | jgraham | doesn't really understand how you can have a non-empirical measurement |
| 17:53 | <jgraham> | Purely on semantic grounds |
| 17:54 | <annevk2> | that YIP is basically what showNotification() in HTML5 was |
| 17:54 | <jgraham> | (I can see logically ho you could say something like "success shouldn't be defined on empirical grounds" meaning "we won't measure anything before declaring a success or not") |
| 17:55 | gsnedders | grumbles about Adobe's ordering. They're saying they will cancel the order if they don't receive proof of academic identification tonight. It's was submitted over a week ago. |
| 17:55 | <jgraham> | annevk2: Whatever happened to showNotification? |
| 17:56 | <jgraham> | gsnedders: Didn't you just email / web form it? |
| 17:56 | <gsnedders> | jgraham: I did. |
| 17:56 | <gsnedders> | jgraham: They claim to have not got it! |
| 17:56 | <jgraham> | gsnedders: Oh. WFM |
| 17:56 | <gsnedders> | WFM in the past too. |
| 17:56 | <jgraham> | gsnedders: Resubmit it? |
| 17:56 | <gsnedders> | I am. Oh well. They get two open cases for the same order. Their problem. |
| 17:57 | <shelleyp> | jgraham: Interesting viewpoint, looking forward to reading more in your email response later tonight |
| 17:57 | <annevk2> | jgraham, only Google was interested and wanted full HTML support |
| 17:57 | <jgraham> | Oh |
| 17:57 | <jgraham> | Well obviously that's not quite true |
| 17:58 | <a-ja> | jgraham: opera 10 beta doesn't seem to be handling br:before{content:"\A";} |
| 17:59 | <gsnedders> | "You can expect a response within 1 business day." — bullshit. :\ |
| 18:00 | <jgraham> | a-ja: Oh interesting. I will check if there is a bug already reported (but not right now because I'm going home, so if someone else wants to do it, feel free :) ) |
| 18:01 | <a-ja> | pretty sure it worked in prior weekly.....wil back out to that and verify, in a bit |
| 18:07 | <annevk2> | a-ja, maybe we finally fixed the bug that requires you to add white-space:pre-line for that to work? |
| 18:11 | <a-ja> | annevk2: will give that a try......wasn't aware that pre-line 's required |
| 18:12 | <annevk2> | or pre |
| 18:12 | <annevk2> | \A is just a newline character |
| 18:15 | <a-ja> | there's some white-space:nowrap stuff in same <p>. will mess with that a bit and see what happens |
| 18:34 | gsnedders | grumbles at PHP some more |
| 20:01 | <takkaria> | gsnedders: are you at Opera too this summer? |
| 20:03 | <jgraham> | takkaria: Yeah you are both interning :) |
| 20:03 | <jgraham> | takkaria: When do you start? |
| 20:04 | <takkaria> | I'm there on the 22nd |
| 20:04 | <takkaria> | until the 18th September |
| 20:04 | <jgraham> | OK, nice |
| 20:04 | <jgraham> | gsnedders starts the week after iirc |
| 20:05 | <takkaria> | do you know if there are any plans on what we'll be working on? |
| 20:05 | <jgraham> | takkaria: Nothing specific. You are QAing, right? |
| 20:06 | <takkaria> | I'm not sure, I specified that I'd be happy to code or QA |
| 20:06 | <jgraham> | (I mean I don't know anything speific, not that you won't be woking on anything specific) |
| 20:06 | <takkaria> | yeah, not working on anything specific seems like a pretty easy job :) |
| 20:07 | <jgraham> | OK. Well if you are QAing I can think of at least three or four things that you could work on and that's just things that I happen to be interested in :) |
| 20:07 | <takkaria> | fair enough |
| 20:07 | <takkaria> | what do you do, out of interest? |
| 20:09 | <jgraham> | QA. I've been working with the JS engine a bit and on general document-related bugs (as opposed to layout-related bugs) |
| 20:10 | <takkaria> | ah, cool |
| 20:12 | <gsnedders> | takkaria: 28th |
| 20:12 | <gsnedders> | takkaria: I'm doing QA |
| 20:12 | <gsnedders> | takkaria: What office you at? |
| 20:15 | <Hixie> | i am SO glad i gave xhr to anne :-) |
| 20:16 | <jgraham> | gsnedders: I am pretty sure he is in Linkoping |
| 20:18 | <annevk42> | Hixie, I recall I just took it; I'm also not glad with it :p |
| 20:18 | <annevk42> | sync/async is a bitch to define properly |
| 20:19 | <annevk42> | and now I'm looking at XHR2 it gets even worse because then it's sync same-origin/sync cross-origin/... |
| 20:20 | <jgraham> | annevk2: Just define that sync cross-origin requests always fail :) |
| 20:20 | <gsnedders> | jgraham: I'm pretty sure he is too |
| 20:20 | <gsnedders> | takkaria: So I arrive after you and leave before you :P |
| 20:20 | <annevk42> | jgraham, compelling, but it's "needed" for Workers |
| 20:21 | <jgraham> | annevk42: Oh. |
| 20:23 | <annevk42> | seamless sandboxed iframes are neat |
| 20:23 | <annevk42> | especially the origin feature |
| 20:46 | <slightlyoff> | hey Hixie |
| 20:46 | <slightlyoff> | is there an SVN repo w/ acid3 in it somewheres? |
| 20:49 | <slightlyoff> | n/m, dug it out of webkit sources... |
| 20:55 | <annevk42> | Hixie, are we getting <iframe document> at some point before LC? |
| 20:58 | <rubys> | hsivonen: ping? |
| 20:59 | <rubys> | java.lang.ClassCastException: org.xml.sax.InputSource cannot be cast to nu.validator.xml.TypedInputSource |
| 21:05 | <Hixie> | annevk42: depends on implementors. I don't want to add it until we have experience with sandbox. |
| 22:36 | <jgraham> | Hmm I thought people had grasped the difference between UA requiements and author requirements, but I guess not |
| 22:36 | <sayrer> | jgraham: I think people grasp them |
| 22:37 | <sayrer> | they just don't think the concept is useful |
| 22:38 | <sayrer> | it all depends on how effective you think prescriptive author requirements will be |
| 22:39 | <sayrer> | if there were any advantage to complying with that sort of thing, websites would do it |
| 22:39 | <jgraham> | sayrer: I think you may be projecting your view onto other people :) |
| 22:40 | <sayrer> | well maybe, but I also want to know why they are useful |
| 22:40 | <sayrer> | smart people seem to think they are |
| 22:40 | <sayrer> | so I am willing to entertain the notion that I am wrong |
| 22:41 | <jgraham> | sayrer: Generally they are useful as QA tools. If you stay within conformance you generally won't hit sharp edges that are likely to get you burnt |
| 22:41 | <sayrer> | there is a lint aspect to them |
| 22:41 | <sayrer> | I see that |
| 22:41 | <sayrer> | but that is different than a conformance requirement |
| 22:41 | <jgraham> | There is also the question of simplifying the language |
| 22:41 | <sayrer> | hmm? |
| 22:42 | <sayrer> | <font> is much simpler than css |
| 22:42 | <sayrer> | for authors |
| 22:42 | <jgraham> | By restricting the set of all working ways of achieving an effect we can focus resources on those that have the best properties for the whole ecosystem |
| 22:42 | <sayrer> | ah, but we can't |
| 22:43 | <sayrer> | since the spec gives quite precise requirements for everything, whether they are conformant for authors or not |
| 22:43 | <sayrer> | that is something I fully support |
| 22:44 | <jgraham> | sayrer: <font> doesn't scale well as a skill and tends to lead to pages that are hard to repurpose beyond the original author's intentions (e.g. to a user using a different type of UA or a screen scraper or...) |
| 22:44 | <sayrer> | I'm not sure that's true |
| 22:44 | <sayrer> | given the example at hand |
| 22:44 | <sayrer> | <font color="blue"> |
| 22:45 | <jgraham> | sayrer: There is far from enough context in that example o deduce anything |
| 22:45 | <sayrer> | well, I agree that it's possible you could do something more reusable in some situations |
| 22:45 | <jgraham> | If you could show me that on real pages people only use <font> when they "should" have used <span>+CSS I would be more convinced |
| 22:46 | <jgraham> | and very surprised, frankly |
| 22:46 | <sayrer> | but a class name + a css file might also buy you an extra file and CSS |
| 22:46 | <sayrer> | when all the author wants is a blue word |
| 22:47 | <sayrer> | jgraham: anything syndicated generally goes for font |
| 22:47 | <gsnedders> | Oh man. That did not just happen. In a race on GTA 4, my car broke down while in the lead coming out of the final corner. |
| 22:48 | <jgraham> | sayrer: I guess the times that "all the auhor wants is a blue word" with no context for the blueness are rather few |
| 22:48 | <jgraham> | Compared to the cases where <font> is used |
| 22:49 | <sayrer> | that seems quite difficult to quantify |
| 22:49 | <jgraham> | sayrer: Sadly lots of useful things are hard to quantify in practice |
| 22:50 | <sayrer> | sure, we don't go making MUST requirements about them :) |
| 22:50 | <jgraham> | It is at least clear that "leading" web designers voluntarily forgo the use of <font>, suggesting that the advntages to not using it are non-theoretical |
| 22:51 | <jgraham> | (and, I should add, they can back up their decision with cognent reasoning about mantainability and so on) |
| 22:56 | jgraham | is going to try sleeping now |
| 22:56 | <sayrer> | jgraham: they also write "xhtml" |
| 22:56 | <sayrer> | so I think your point is questionable |
| 22:56 | <sayrer> | night |
| 22:57 | <Dashiva> | And use floats to make table layouts |
| 23:03 | <gsnedders> | And use floats for tabular data, because tables are bad. |
| 23:13 | <sayrer> | snicker |
| 23:13 | <sayrer> | google.com |
| 23:13 | <sayrer> | perhaps the most closely measured page in the history of the world |
| 23:13 | <sayrer> | contains a font element |
| 23:14 | <ezyang> | jaja |
| 23:14 | <sayrer> | somebody better tell them |
| 23:14 | <sayrer> | about their suboptimal markup |
| 23:15 | <ezyang> | It's probably really portable! |
| 23:34 | <hober> | I don't think I understand Shelley's latest email. |
| 23:35 | <sayrer> | hober, why not? |
| 23:35 | <sayrer> | it seems to contain a good point |
| 23:35 | <hober> | which was? |
| 23:36 | <sayrer> | you don't want to edit a font element out of a document in order to produce "conformant" markup |
| 23:37 | <hober> | Who is 'you' in this case? (I think I would want to do that...) |
| 23:37 | <sayrer> | hober, if you agree that a <span style=""> is no better |
| 23:38 | <sayrer> | then you would want to move the CSS to a style element |
| 23:38 | <sayrer> | that is madness |
| 23:38 | <sayrer> | since you would be changing the "lexical scope" that a given piece of markup depends on |
| 23:38 | <hober> | Well, I'd like to know why the text is blue, so I could write <strong>...</strong>, or <span class="some-semantic-missing-from-the-default-set-of-elements">...</span> |
| 23:39 | <sayrer> | hober, well, I don't agree that the semantics of the blue text are always worth writing down, but where do you put the styling information if it's clear that the element should be <strong> |
| 23:40 | <hober> | In an external stylesheet, linked to with <link> |
| 23:41 | <hober> | (ideally the selector for applying the blue style is simply "strong", too) |
| 23:41 | <sayrer> | so you introduce a dependency like that via a wysiwyg editor? |
| 23:42 | <hober> | I use Emacs to write markup. :) |
| 23:43 | <sayrer> | hober: I take it you forfeit? |
| 23:43 | <sayrer> | ;) |
| 23:44 | <sayrer> | hober, but even with Emacs |
| 23:44 | <sayrer> | are you aware of any editor that would perform that transformation? |
| 23:44 | <hober> | I think the question of how to build a WYSIWYG web page editor is a separate one from what we were talking about |
| 23:44 | <sayrer> | hober, not according to the spec |
| 23:44 | <sayrer> | but I am willing to engage on Emacs editing modes |
| 23:44 | <sayrer> | are there Emacs editing modes that perform the transformation you described? |
| 23:45 | <hober> | What do you mean by transformation? |
| 23:45 | <hober> | the author types in < s t r o n g > ... |
| 23:45 | <sayrer> | <font color="blue"> --> <strong> + css file |
| 23:45 | <sayrer> | the editor doesn't always create the pages it edits |
| 23:46 | <hober> | The author doesn't always create the page he/she is editing, sure. |
| 23:46 | <sayrer> | so, if I copy and thenpaste a section with <font> |
| 23:46 | <sayrer> | should the editor paste in a font? |
| 23:46 | <sayrer> | a font element, I mean |
| 23:46 | <hober> | I mean, this is generally true of any document editing. How does an author change foo to bar in a document of type baz? Well, he/she finds all the foos and changes them to bar.. |
| 23:47 | <sayrer> | authoring tools are subject to the same requirements that authors are |
| 23:47 | <hober> | If you copy source from somewhere and paste it in somewhere, I would be really annoyed if it wasn't the same source... |
| 23:47 | <sayrer> | yes, I agree |
| 23:48 | <sayrer> | so it will be quite impossible for authoring tools to meet the current conformance requirements |
| 23:50 | <hober> | so you've got markup in an intermediate state, of unknown conformance. I don't see why it would be "quite impossible" to have a "export to conformant html5" button |
| 23:51 | <hober> | with an annoying "some formatting may be lost" confirmation dialog |
| 23:51 | <sayrer> | so it would require data loss? |
| 23:52 | <sayrer> | lame |
| 23:52 | <hober> | user writes unkown attribute or element (which are non-conformant), what does an editor replace them with? |
| 23:52 | <sayrer> | nothing, obviously |
| 23:53 | <hober> | <unknown> could become <div class="com.example.unknown"> |
| 23:53 | <sayrer> | but it might suggest something |
| 23:53 | <hober> | unknown attributes though, get lost I think |
| 23:53 | <sayrer> | a far more likely example might be <dvi></div> |
| 23:53 | <hober> | data-com-example-unknown? |
| 23:53 | <Hixie> | hober: you could use data-com.example.attr="foo" |
| 23:53 | <Hixie> | yeah |
| 23:53 | <Hixie> | what you said |
| 23:54 | <sayrer> | but if we veer into wysiwyg, this is irrelevant |
| 23:54 | <sayrer> | it looks how it looks |
| 23:54 | <hober> | ... and when you click that button, and OK that dialog, the result is conformant HTML5, right? |
| 23:55 | <hober> | I don't see why it wouldn't be, or rather why it can't be. |
| 23:55 | <sayrer> | two things |
| 23:55 | <sayrer> | I am not aware of any editor that would do that |
| 23:55 | <sayrer> | it might be undesirable |
| 23:56 | <sayrer> | the HTML5 standard is an astonishingly bad place to speculate on Editors Of The Future |
| 23:58 | <sayrer> | it is sort of antithetical to the "don't mess with headers you don't understand" ethos of email |
| 23:59 | <sayrer> | and more complicated given the external dependencies that a "proper" file would have |