| 02:56 | <Lachy> | I've now added the brief descriptions to the element summaries in the HTML 5 Reference http://dev.w3.org/html5/html-author/#the-elements (i.e. paragraphs that say "The foo element represents...") |
| 02:56 | <Lachy> | some of those are identical to what's in the spec, and still need to be updated to be more author friendly |
| 02:57 | <Lachy> | and some, like the a element, are incomplete because the spec splits the description across multiple paragraphs, but my script only extracted the first |
| 02:58 | <Lachy> | I also added the index of elements http://dev.w3.org/html5/html-author/#index-of-elements |
| 03:07 | <Lachy> | I now added the tag requirements (required/optional/empty) to the element summaries |
| 04:11 | <annevk> | html5alt⊙lwe? |
| 04:13 | <annevk> | (from www-archive) |
| 04:43 | annevk | killed wiki.html5.org |
| 04:43 | <dglazkov> | murderer! |
| 04:45 | <annevk> | if you only knew :p |
| 04:45 | <MikeSmith> | dglazkov: about the WSJ article, dude doesn't know his ass from the hole in the ground, so wouldn't worry about it too much |
| 04:46 | <dglazkov> | MikeSmith: :) I know. It still bugs me. |
| 04:47 | dglazkov | mumbles something about manners and civility |
| 04:51 | <Lachy> | annevk, why did you kill it? Was it getting spammed too much? |
| 04:57 | <karlcow> | little agony… |
| 04:58 | <annevk> | it was unused and there was some spam |
| 05:10 | <MikeSmith> | hsivonen: btw, it's been more than one year now since we first contacted the GNU guys about quoting of filenames in GNU error format |
| 05:12 | <Hixie> | i mam now 25% done with this edit |
| 05:12 | <Hixie> | am |
| 05:13 | <MikeSmith> | Hixie: commit to checkpoint what you got done so far? |
| 05:16 | <Hixie> | checkpoint? |
| 05:16 | <Hixie> | the whatwg copy has what i've done so far |
| 05:16 | <Hixie> | i regen it after each subsection |
| 05:16 | <jwalden> | Hixie: did you see the URL I posted here last night directed at you? |
| 05:17 | <Hixie> | which one? |
| 05:17 | <jwalden> | Hixie: http://pastebin.mozilla.org/630108 |
| 05:17 | <jwalden> | another infinite redirect in Google props |
| 05:17 | <Hixie> | what are the steps to get to that point? |
| 05:17 | <Hixie> | i.e. where did you get that link? |
| 05:18 | <jwalden> | Hixie: Google Reader site URL for the feed for http://www.brandonsanderson.com/blog/ |
| 05:18 | <jwalden> | viewing a specific post, clicked on the title of that post and got sent there |
| 05:19 | <Hixie> | so the steps to reproduce are subscribe to that feed, then click the title of a post in google reader? |
| 05:20 | <jwalden> | Hixie: http://feeds.feedburner.com/BrandonSandersonBlog is the feed, select "all items" as the view for it, expand the "AMoL Update...kind of." post, click on the title |
| 05:21 | <Hixie> | weird |
| 05:22 | <jwalden> | I really have no idea how it is I've managed to find two of these redirect-to-self bugs, I'm not making a special effort to use Google stuff weirdly :-) |
| 05:24 | <jwalden> | conveniently for anyone who might have also hit this there's 1) not much in the entry and 2) a Continue Reading link that works :-) |
| 05:24 | <jwalden> | so at least for this one weird case nobody's going to be especially inconvenienced |
| 05:24 | <Hixie> | i wonder if it's because of the encoding declaration errors |
| 05:25 | <jwalden> | I've been watching that feed for months and haven't had problems in the past, so dunno |
| 05:26 | <Hixie> | that particular post has a "..." unicode character |
| 05:26 | <Hixie> | that is badly encoded |
| 05:26 | <jwalden> | ah, yes, http://feedvalidator.org/check.cgi?url=http%3A%2F%2Ffeeds2.feedburner.com%2FBrandonSandersonBlog |
| 05:27 | <jwalden> | that whole site/feed aren't the greatest tech-wise, most posts on the site show up with what seems to be a UTF-8 BOM and flash of white until the rest of the page loads and presumably Firefox decides it's really a different encoding at that point |
| 05:28 | jwalden | hasn't cared enough to really investigate the problem |
| 05:28 | <sayrer> | god that feed is brutal |
| 05:29 | <jwalden> | :-D |
| 05:29 | <Hixie> | well i filed the bug |
| 05:29 | <Hixie> | so thanks |
| 05:30 | <Hixie> | :-) |
| 05:30 | <jwalden> | no problem, always fun to amuse your peeps with obscure bugs :-) |
| 05:30 | jwalden | can't believe he just said "peeps" |
| 05:30 | <Hixie> | :-) |
| 06:07 | <Hixie> | 29% and about to embark on <video> |
| 06:46 | <Hixie> | hsivonen: wow, one missing space character triggered thousands of errors :-) |
| 06:46 | <Hixie> | (the missing space ate the end tag of an inline element that then got AAA'ed into the next eight thousand or so paragraphs, headers, etc) |
| 07:20 | <hsivonen> | MikeSmith: oh yeah, should I follow up on the GNU stuff? I thought the ball was in their court. |
| 07:20 | <MikeSmith> | hsivonen: yeah, it is |
| 07:21 | <MikeSmith> | actually, I think it's specifically in the hands of RMS |
| 07:21 | <MikeSmith> | I'll e-mail a reminder to Karl Berry |
| 07:30 | <hsivonen> | MikeSmith: thanks |
| 07:40 | <hsivonen> | MikeSmith: what's your preferred user id other than 'mike'? 'sideshowbarker'? |
| 07:42 | <MikeSmith> | hsivonen: yeah, "sideshowbarker" |
| 07:42 | <hsivonen> | k |
| 10:16 | <Hixie> | 34% done |
| 10:16 | <Hixie> | 5% of the spec is video, wow |
| 10:19 | <Lachy> | Hixie, what's 34% done? |
| 10:21 | <Lachy> | gotta go, back soon |
| 10:35 | <Hixie> | lachy: going through doing the author stuff |
| 10:45 | <hsivonen> | http://lists.w3.org/Archives/Member/w3c-wai-cg/2009JanMar/0029.html |
| 10:50 | <Philip`> | https://lists.wisc.edu/read/?forum=html5alt - new list? |
| 10:51 | <Philip`> | (That's a horrid list interface - why can't they just use plain HTML with links, and not squash everything into tiny boxes?) |
| 11:01 | <jgraham> | http://www.w3.org/2009/03/04-cg-minutes.html |
| 11:18 | <jgraham> | http://esw.w3.org/topic/PF/XTech/HTML5/TextAlternativeProposal |
| 11:22 | <beowulf> | i'm just a dumb html author, but why not simply assume an img with alt="" is presentational? |
| 11:24 | <Dashiva> | Isn't that what everyone does? |
| 11:25 | <Philip`> | beowulf: Maybe the idea is to make the cost of incorrectly marking a critical image as presentational greater than the cost of giving it correct alt text, so people will do less of the former and hence more of the latter |
| 11:26 | <Philip`> | But that's probably not the reasoning, and I have no idea what the reasoning is because that wiki page doesn't seem to include any |
| 11:28 | <beowulf> | speaking for the dumb html authors, making things harder isn't a good tactic for success |
| 11:29 | <jgraham> | beowulf: Really? Who knew? |
| 11:29 | <beowulf> | :) |
| 11:30 | <Philip`> | But if it makes valid pages more accessible then it's a net win! |
| 11:30 | <Philip`> | (and don't bother about invalid pages because they're invalid and who cares what they're going to do) |
| 11:31 | <Philip`> | (and people who do care about writing valid pages will clearly go to whatever extreme lengths are necessary to make their pages valid, so we can make it as hard as we want) |
| 11:33 | <jgraham> | There seems to be some idea that having alt as a validity requirement makes authours who would otherwise not have cared about writing alt text suddenly write good, useful alt text |
| 11:33 | <jgraham> | This is just the extension of that argument |
| 11:33 | <beowulf> | Philip`: and making those people that care enough a small isolated island really helps them be advocates for Doing The Right Thing |
| 11:33 | <Philip`> | It seems plausible to believe that telling people to do X (perhaps via a validator's error messages) will result in more people doing X |
| 11:33 | <Philip`> | and the question is how many more people will that be, and how does it balance against the costs |
| 11:34 | <beowulf> | the people that use a validator aren't the people you need to talk to about @alt usage, i believe |
| 11:34 | <beowulf> | (in IRC, what's the channel mode to switch off cynicism?) |
| 11:35 | <jgraham> | Philip`: Sure, if X is "write alt text". But the X we should actually care about is "write accessible pages" |
| 11:35 | <jgraham> | beowulf: In this channel I think you /leave |
| 11:35 | <beowulf> | lol |
| 11:35 | <hsivonen> | I notice that the alt wiki page doesn't say what the alt TF wants rendering clients (browser+AT) to do and doesn't say what they want authoring tools to do |
| 11:36 | <hsivonen> | In my opinion, the right order to deal with things is 1) establish what rendering clients should do (in a way that is compatible with existing content) |
| 11:36 | <hsivonen> | 2) Establish what authoring tools should do considering what rendering clients should do |
| 11:37 | <hsivonen> | 3) Make sure that the validity definition is broad enough not to flag as invalid output from authoring tools that follow the advice for authoring tools |
| 11:42 | <hsivonen> | wow. the wisc list archive UI is indeed fantastically bad |
| 11:42 | <Philip`> | The <time> restriction kind of reminds me of the <img alt="{whatever}"> idea, in that it adds some arbitrary restrictions that make automated generation of markup really painful because you've got worry about all the new edge cases |
| 11:45 | <Philip`> | There should be a counterpart of Postel's law, like "Be liberal in producing metadata; be selective in accepting it from others", like you should mark up all times with <time> because that's easier to working out precisely when there's a use case for marking it up, and the consumer can decide whether to bother with your metadata or not |
| 11:46 | <Philip`> | s/easier to/easier than/ |
| 11:47 | <Philip`> | It avoids the problem of problem of forcing producers (who see the <time> element, and see a time value in their markup) to twist their minds to view everything from the consumer side (to work out whether this a time value that someone's going to copy into the calendar) |
| 11:47 | <Philip`> | s/problem of// |
| 12:01 | <hsivonen> | Philip`: should consumers be required to support e.g. the Mayan calendar if a blogger wants to image (s)he is blogging from the height of the Mayan civilazation? |
| 12:02 | <hsivonen> | *civilization |
| 12:05 | <olliej> | hsivonen: yes, because that would be awesome |
| 12:05 | <olliej> | :D |
| 12:06 | <hsivonen> | olliej: what about the Y2012 problem? :-) |
| 12:06 | <olliej> | hsivonen: what problem? ;D |
| 12:07 | <hsivonen> | olliej: http://www.usatoday.com/tech/science/2007-03-27-maya-2012_n.htm |
| 12:07 | <olliej> | hsivonen: you're not suggesting there will still be people left to complain are you? |
| 12:07 | <olliej> | hsivonen: :D |
| 12:12 | <Philip`> | hsivonen: No, because that would be very rare, whereas it'd be very common for software to take a nice simple ISO8601ish date object and try to serialise it into a <time> but now they've got to add extra checks for year>0 to avoid producing invalid markup |
| 12:14 | <hsivonen> | Philip`: if common libraries are consistent in their serialization of negative seconds since the epoch to earlier than 1 Jan 0001, I suppose that might be OK |
| 12:15 | <hsivonen> | but if common libraries can't handle dates before 1 Jan 0001 consistently, I think it's not worth the implementation and QA effort |
| 12:18 | <Philip`> | $ perl -MDateTime -le'print new DateTime(year=>-100)' |
| 12:18 | <Philip`> | -0100-01-01T00:00:00 |
| 12:18 | <Philip`> | Seems to work fine there |
| 12:18 | <Philip`> | (It's documented that year 0 exists, so year -100 is 101 BC) |
| 12:18 | <hsivonen> | Philip`: so if you take the seconds since epoch from Perl and give them to Java, do you get the same proleptic Gregorian date serialization? |
| 12:36 | <jgraham> | datetime(-100,0,0) |
| 12:36 | <jgraham> | --------------------------------------------------------------------------- |
| 12:36 | <jgraham> | ValueError Traceback (most recent call last) |
| 12:36 | <jgraham> | /home/jgraham/html5lib/trunk/<ipython console> in <module>() |
| 12:36 | <jgraham> | ValueError: year is out of range |
| 12:38 | <jgraham> | (datetime(0,1,1) which has valid month and day values also doesn't work, fwiw) |
| 12:41 | <Philip`> | hsivonen: Java always seems to drop the minus sign when serialising years |
| 12:42 | <hsivonen> | Philip`: Python and Java are enough of reasons for me not to want pre-0001 years |
| 12:42 | <Philip`> | hsivonen: (I can't easily get seconds-since-epoch from Perl since it does some 32-bit maths, but that doesn't seem a real problem since you can use strings for communication between languages) |
| 12:43 | Philip` | doesn't know if there's a better way to serialise dates in Java - he was just use Data.toString, and also the annoyingly complex SimpleDateFormat |
| 12:44 | <Philip`> | *Date.toString |
| 12:44 | <Philip`> | (Having a method that takes a date and returns a string would clearly be too simple for SimpleDateFormat, so it wants an input StringBuffer and a FieldPosition too) |
| 12:46 | <Philip`> | http://joda-time.sourceforge.net/ - looks like someone didn't like Java's date/time library |
| 12:54 | <Philip`> | jgraham: Consuming software in Python would simply check if year > 0 and if so then it would decide to ignore that <time> element |
| 12:55 | <jgraham> | Philip`: Then it would fail testcases, etc. |
| 12:55 | <Philip`> | It already has to check if year <= 9999 |
| 12:56 | <Philip`> | so it's already incompatible with the range of <time> |
| 12:56 | <jgraham> | Ah, well we should reduce it to have the same range |
| 12:57 | <jgraham> | Although practically no one seems to care about <time> for very large years |
| 12:57 | <Philip`> | And if we find a language/library that fails on dates before 1970 and after 2038 then we should reduce <time> to that range too? |
| 12:57 | <Philip`> | jgraham: Maybe practically no one cares, but it would fail test cases |
| 12:58 | <jgraham> | Philip`: Depends ho significant a language/library it is |
| 12:58 | <jgraham> | Are there any popular langauges that are known to fail for a smaller range and for which simple workarounds are not avalaible |
| 13:00 | <Philip`> | http://www.boost.org/doc/libs/1_38_0/doc/html/date_time/gregorian.html says it's limited to 1400-Jan-01 to 9999-Dec-31 |
| 13:02 | <jgraham> | Hmm that's pretty annoying |
| 13:02 | <jgraham> | It does sugest limiting to 4-digit years is sensible |
| 13:04 | <jgraham> | Or rather, maximum 4-digit |
| 13:15 | <Philip`> | http://uk2.php.net/manual/en/datetime.add.php - "interval: The amount to be added. For the date use "P3D", "P3M", "P3Y" or a combination of the three e.g. "P2M5D" (Y = Years, M = Months, D = Days.) MUST BE YEAR MONTH DAY FORMAT "P5Y", "P5M2D", "P5Y4D". For the time use "T3H", "T3M", "T3S" or or a combination of the three e.g. "T5H20M" (H = Hours, M = Minutes, S = Seconds). For dateTime use "P5D2M4YT5H20M". The digit before the letter (NOT P or T) can b |
| 13:16 | <Philip`> | ...What? |
| 13:16 | <Philip`> | Are the people who wrote this API literally insane? |
| 13:35 | Philip` | sees http://fhtr.blogspot.com/2009/03/anatomy-of-canvas-3d-extension.html |
| 13:36 | <Philip`> | "with a 7x7 gaussian blur kernel over a 256x256 Firefox logo (decomposed into a horizontal blur and a vertical blur.) JavaScript took 0.8 seconds to do a single blur. With GLSL, it took 0.4 seconds to do a thousand blurs." - I don't think any amount of clever JITting is likely to beat that |
| 13:38 | <olliej> | Philip`: it's very easy to write a very slow convolution in js, but yes, if you throw pixel manipulation at something that is built from the ground up to do nothing but pixel manipulation, it's going to do well |
| 13:38 | <olliej> | Philip`: i would expect a gpu to easily be at least 100x faster than a C impl |
| 13:39 | <Philip`> | I suppose this is a bit irrelevant in practice because you could just blur the image once and send it as a static file across the network, instead of getting your users to blur it :-) |
| 13:40 | <olliej> | Philip`: unless you're writing photoshop |
| 13:40 | <olliej> | or something |
| 13:40 | <Philip`> | If you're writing Photoshop you probably care about precision, so you wouldn't rely on the GPU to get the maths right |
| 13:40 | <Philip`> | Actually I suppose that's totally untrue |
| 13:41 | <Philip`> | particularly since you'd have real-time previews of effects, where you don't care about precision |
| 13:41 | <Philip`> | and maybe modern GPUs can do proper floating-point maths now |
| 13:41 | <olliej> | Philip`: some can |
| 13:42 | <olliej> | Philip`: people keep saying canvas 3d should expose shaders, i'm like noooooo |
| 13:42 | <Philip`> | (where "proper" means "conforming to IEEEwhatever") |
| 13:42 | <Philip`> | I think Canvas 3D without shaders would be pretty rubbish in practice, because it'd be painful and you couldn't do anything interesting with it |
| 13:43 | <olliej> | Philip`: the variation in a) what the maths actually produces b) what shaders can actually get compiled on given hardware varies dramatically c) running arbitrary untrusted content on the gpu is a really really really bad plan |
| 13:43 | <Philip`> | In the last few years of pre-shader OpenGL you at least had vendor extensions to get similar effects |
| 13:43 | <Philip`> | and without shaders your content will all look like VRML and people will laugh at you |
| 13:44 | <olliej> | people coped without shaders for years and years and years |
| 13:44 | <Philip`> | but I can't imagine it working decently with shaders either, so I've kind of given up on the idea for now |
| 13:44 | <Philip`> | olliej: They did, but that was ages ago and I can't imagine it being visually acceptable now |
| 13:44 | <olliej> | quake3 is still used as a benchmark today |
| 13:45 | <olliej> | halflife/cs was still very popular for years after becoming graphically obsolete (hl/cs was based in quake2's engine iirc) |
| 13:45 | <Philip`> | olliej: (In terms of security, I'd probably be more worried about driver bugs in shader compilers rather than in the actual execution on the GPU) |
| 13:46 | <Philip`> | olliej: (And interoperability is certainly a problem - the very first Canvas 3D example I wrote worked with NVIDIA drivers on Windows but not with NVIDIA drivers on OS X, because I used a clamp() function in a shader) |
| 13:46 | <olliej> | Philip`: while(true); in a shader, on a gpu that supports true conditional branching == hung gpu |
| 13:49 | <Philip`> | olliej: Hmm, good point about infinite loops |
| 13:49 | <Philip`> | GLSL says "Non-terminating loops are allowed. The consequences of very long or non-terminating loops are platform dependent." which doesn't help much |
| 13:50 | <olliej> | Philip`: if you have good drivers they will kill the shader (eventually) |
| 13:50 | <olliej> | iirc that will take out the app however |
| 13:50 | <olliej> | maybe i'm wrong |
| 13:50 | <olliej> | it's been a long time since i did anything really gpu related |
| 13:51 | <Philip`> | olliej: I'd imagine Canvas 3D uses would include games (which would compete with non-browser-based games), adverts (like shiny rotatable 3D iPhones for you to examine), and I can't really think of much else |
| 13:51 | <Philip`> | and both those cases seem to need modern decent-quality graphics (i.e. shaders) |
| 13:51 | <jgraham> | Philip`: They wwouldn't really have to compete with non-rowser-based games |
| 13:52 | <jgraham> | People don't really expect to play GTA in their browser |
| 13:52 | <jgraham> | They expect to be able to play tetris |
| 13:52 | <olliej> | Philip`: i think a canvas capable of doing sprite/layers etc would be a good intermediate step |
| 13:52 | <olliej> | but then |
| 13:53 | <olliej> | you could basically use css :D |
| 13:53 | <Philip`> | jgraham: They still expect to be able to play a decent-quality Tetris, e.g. if it's 2D then it should have pretty sprites and everything |
| 13:53 | <jgraham> | Philip`: Sure. But the level of graphics you need to make an acceptable tetris are lower than those needed for a 3D shooter |
| 13:53 | <Philip`> | jgraham: so if they play a game that is 3D, they should have similar expectations of quality, and they don't want it to look like http://www.dmp.cz/rt3d/onsi/diplomka/images/bitmanagement_neruda.jpg |
| 13:54 | <Philip`> | and the only sensible way to get decent lighting and shadows and materials is by using shaders |
| 13:57 | <Philip`> | (and the methods used in e.g. Quake are not sensible, because they require hours of preprocessing and large texture maps and static geometry stuff, since they're just cheating) |
| 13:57 | <Philip`> | s/stuff/and stuff/ |
| 13:57 | <olliej> | Philip`: hours of preprocessing == decade or two ago |
| 13:58 | <olliej> | Philip`: that said, yes, there are many reasons i haven't done any canvas-3d like thing in wk |
| 13:58 | <Philip`> | olliej: Okay, so maybe it's more like seconds now :-) |
| 13:58 | <Philip`> | olliej: but it's still annoyingly limited and painful to write code that way, particularly when it's just an artifact of the API and you know the hardware is capable of so much more |
| 13:59 | <olliej> | Philip`: doom3 recomputes the level BSP for every frame -- that used to be an half and hour or something back when quake 3 was released |
| 13:59 | <olliej> | Philip`: ah ha |
| 13:59 | <olliej> | Philip`: you *don't* |
| 14:00 | <olliej> | websites don't have minimum requirements in the same way shop purchased software does |
| 14:00 | <olliej> | eg. you have a much more varied set of platforms |
| 14:00 | <Philip`> | Sure they do |
| 14:00 | <Philip`> | "You must install the Flash plugin" |
| 14:00 | <Philip`> | "This site requires 1024x768 resolution" |
| 14:00 | <Philip`> | etc |
| 14:01 | <olliej> | Philip`: yes, and that just means people complain about your site not working |
| 14:01 | <Philip`> | Canvex has minimum requirements that exclude a majority of all people |
| 14:01 | <Philip`> | and nobody has complained to me about it not working in IE :-) |
| 14:02 | <Philip`> | Even integrated graphics and mobile phones can do GLSL now, so in five years it's not really going to be an issue |
| 14:03 | <Philip`> | (Also I think the Mozilla people wanted to use Mesa to implement shaders in software when it's not supported by hardware, though I don't know if that's possible or practical) |
| 14:05 | Philip` | wonders if Doom 3's BSPing has the same output as Q3's, or if it does it much more loosely and relies more on the graphics card to cope with unnecessary polygons |
| 14:07 | Philip` | remembers seeing some discussion of lots of complex terrain LOD algorithms, and then someone suggesting just throwing all the polygons at the GPU and letting it cope with them all itself, which actually worked surprisingly well in their example |
| 14:14 | <hsivonen> | hrm. jgraham's test case at http://news.ycombinator.com/item?id=500781 seems to be gone / unresponsive |
| 14:17 | <jgraham> | hsivonen: Yeah, the server seems to be down |
| 14:24 | <hsivonen> | I wonder if it is significant if scripts run immediately upon </script> or if timeouts/intervals have an opportunity to run between </script> being tokenized and the script starting running |
| 15:39 | <Lachy> | note to self: never mention on public-html that there is a potential issue about what to name something. |
| 15:40 | Lachy | tries to stop the bikeshedding and ask people to focus on the real content of the HTML 5 Reference |
| 15:40 | rubys | likes blue. Or is it red these days? :-) |
| 15:41 | <Lachy> | I like blue too |
| 15:42 | <Lachy> | I just find it strange that of the 5 responses to my email about the HTML5 reference, all of them have focussed on what to call polyglot documents, rather than providing any useful feedback about what I had written in the other sections |
| 15:43 | <svl_> | "strange"? You must be new here. |
| 15:43 | <svl_> | *g* |
| 15:46 | <jgraham> | Lachy: Call them Bob |
| 15:46 | <jgraham> | Bonus points if you can make it sound like Blackadder |
| 15:55 | <Lachy> | jgraham, I never watched Blackadder, so I'm not sure what you mean |
| 15:58 | <jgraham> | Lachy: Rowan Atkinson who plays blackadder has a speech impediment that makes it difficult for him to pronounce b sounds. So in one episode they introduced a character called "bob" which he found very difficult to say. As a result... well I don't knoww how to explain it but the way he says it is memorable |
| 15:59 | <jgraham> | It kind of builds up gently then suddenly pops out |
| 16:10 | beowulf | replied to lachy's tweet with 'good work' but now considers asking for it in pink |
| 16:11 | <gsnedders> | pink? I'd rather have it in green, personally. |
| 16:11 | <gsnedders> | But more important is what to make the roof of. |
| 16:12 | gsnedders | is sure he should remember who svl is |
| 16:14 | <hsivonen> | hmm. all of Steven Pemberton's posts to public-html pertain to issues arising from using colonified attribute names in text/html: http://www.w3.org/Search/Mail/Public/search?keywords=&hdr-1-name=from&hdr-1-query=steven.pemberton%40cwi.nl&index-grp=Public__FULL&index-type=t&type-index=public-html |
| 16:17 | <virtuelv_> | hsivonen: does validator.nu offer a json interface? |
| 16:17 | <hsivonen> | virtuelv_: yes. (with callbacks and without) |
| 16:17 | <hsivonen> | virtuelv_: &out=json&callback=foo |
| 16:17 | <virtuelv_> | thanks |
| 17:02 | <Lachy> | woah, I think I need to take a break from editing the HTML5 Reference. I didn't realise how late it had gotten here. Although, I've managed to do a fair bit on it today. |
| 17:05 | gsnedders | wonders whether he can send comments without getting bikesheded |
| 17:12 | <Lachy> | gsnedders, just avoid issues about the name of polyglot documents and you'll be fine |
| 17:12 | <gsnedders> | Wanna bet? |
| 17:12 | <gsnedders> | :) |
| 17:13 | rubys | will take that bet too |
| 17:13 | <gsnedders> | rubys: Which way? |
| 17:13 | <rubys> | same side as you |
| 17:13 | <gsnedders> | Ah, another realist. |
| 17:13 | <gsnedders> | :P |
| 17:14 | Lachy | bets $2.50 on the bikeshed being painted green |
| 17:16 | hsivonen | wants papayawhip |
| 18:36 | <gsnedders> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cdiv%3E%3C%2Fdiv%3E%0A%3Cscript%3E%0Adocument.getElementsByTagName(%22div%22)%5B0%5D.innerHTML%20%3D%20%22%3Cbase%20href%3Dhttp%3A%2F%2Fexample.com%3E%3Cp%3E%3Ca%20href%3D%2F%3Efoo%22%3B%0A%3C%2Fscript%3E |
| 18:36 | <gsnedders> | Where does the link link? |
| 18:37 | <gsnedders> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cdiv%3E%3C%2Fdiv%3E%0A%3Cscript%3E%0Awindow.onload%20%3D%20function()%20{%0Adocument.getElementsByTagName(%22div%22)[0].innerHTML%20%3D%20%22%3Cbase%20href%3Dhttp%3A%2F%2Fexample.com%3E%3Cp%3E%3Ca%20href%3D%2F%3Efoo%22%3B%0A}%0A%3C%2Fscript%3E |
| 18:37 | <gsnedders> | That's needed for compat. with some browsers |
| 19:08 | <hdh> | in chrome, both to http://example.com/ |
| 19:09 | <hdh> | firefox trunk, http://software.hixie.ch/ |
| 19:09 | <gsnedders> | Got IE? |
| 19:09 | <gsnedders> | In HTML 5, http://example.com/ |
| 19:10 | Hixie | wonders what he did that causes long urls to screw up in terminal now |
| 19:11 | <gsnedders> | Hixie: n00b |
| 19:11 | <hdh> | IE6 http://example.com/ |
| 19:12 | <Hixie> | score, the spec wins again |
| 19:13 | <gsnedders> | IE7: http://software.hixie.ch/ |
| 19:13 | <Hixie> | d'oh |
| 19:13 | <Hixie> | is there a /saved url for these tests? |
| 19:13 | <gsnedders> | 25 is a more complex example but doesn't seem to work in IE7 |
| 19:14 | <gsnedders> | But otherwise no |
| 19:15 | gsnedders | is back to reverse-engineering base URL behaviour |
| 19:16 | <gsnedders> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0A%3Cp%3Efoo%0A%3Cbase%20href%3D%22http%3A%2F%2Fexample.com%22%3E%0A%3Cp%3Ebar |
| 19:16 | <gsnedders> | That moves base to head in both Saf and FF |
| 19:16 | <gsnedders> | Doesn't in HTML 5 |
| 19:19 | <Hixie> | the <base> handling in the spec is based on some research of what pages need in practice, iirc |
| 19:19 | <Hixie> | i wanted to simplify what the browsers do now since they're not all perfectly compatible |
| 19:45 | gsnedders | goes insane |
| 19:45 | <gsnedders> | I have this one bit of hair that will not stay tied back today. |
| 19:46 | <Philip`> | Cut it off |
| 20:16 | gsnedders | wonders what Mellblomenator references |
| 20:16 | <gsnedders> | I mean, that _has_ to be a joke. |
| 20:21 | <gsnedders> | Anyone able to give me a hand with writing a security considerations section? |
| 20:21 | <Philip`> | gsnedders: Seems quite likely that it's referring to the same person as http://www.mellblom.us/blog/ |
| 20:21 | <Philip`> | particularly since the photo on that site indicates the person works for Google |
| 20:22 | <Philip`> | and the latest blog post talks about the Mellblom Browser, mere days before the similar text was inserted into HTML5 |
| 20:33 | <Hixie> | damn, you're good |
| 20:51 | rubys | wonders who Hixie is referring to... |
| 20:54 | <Hixie> | philip |
| 20:57 | Philip` | used his incredible skills to type Mellblom into Live Search and find that page |
| 20:58 | <Hixie> | and you actually found it? |
| 20:58 | <Hixie> | impressive |
| 20:59 | <gsnedders> | Hixie: How do you keep spam off whatwg? |
| 20:59 | <Philip`> | I used my supersight abilities to look through plastic and liquid crystals to see the first search result |
| 20:59 | <gsnedders> | http://lastweekinhtml5.blogspot.com/2009/03/hair-in-gait.html |
| 20:59 | <gsnedders> | oh dear… |
| 21:00 | <Hixie> | gsnedders: off spam where? |
| 21:00 | <gsnedders> | Hixie: I mean the whatwg⊙wo list |
| 21:00 | <Hixie> | only subscribers can post |
| 21:01 | <gsnedders> | That's all? |
| 21:03 | <Hixie> | yup |
| 21:04 | <roc> | hang on, are there multiple Philip Taylors? |
| 21:04 | <Hixie> | yes |
| 21:04 | <gsnedders> | roc: Yes |
| 21:04 | <Hixie> | two |
| 21:04 | <Hixie> | both of which have used at least two different e-mail addresses |
| 21:04 | <roc> | wow, that explains a lot! |
| 21:04 | <Hixie> | possibly three |
| 21:04 | <Hixie> | lol |
| 21:04 | <gsnedders> | Three |
| 21:07 | <Philip`> | I'm pretty sure I've used at least three addresses |
| 21:11 | <gsnedders> | roc: One has "Philip TAYLOR" in the From header and isn't on IRC; the other has "Philip Taylor" in the From header and is Philip` on Freenode and Philip on irc.w3.org |
| 21:11 | <roc> | thanks |
| 21:14 | <Philip`> | I'm the one who's on IRC |
| 21:15 | <Hixie> | they are otherwise undistinguishable |
| 21:15 | Hixie | ducks |
| 21:16 | <Hixie> | s/u/i/ |
| 21:16 | Philip` | hopes that substitution was meant to apply to the last-but-one line |
| 21:17 | <Hixie> | yes |
| 21:17 | <Hixie> | you can't apply substitutions to what i'm doing, silly |
| 21:17 | Hixie | ducks again |
| 21:21 | <Hixie> | btw roc, re your mail, things out of the document shouldn't delay the event |
| 21:21 | <Hixie> | i'll make it clearer at some point |
| 21:22 | <roc> | images do |
| 21:22 | <roc> | boris says this matters for existing content |
| 21:22 | <roc> | image preloading stuff |
| 21:22 | <Hixie> | really? |
| 21:22 | <Hixie> | wow |
| 21:22 | <Hixie> | good to know |
| 21:22 | <Hixie> | well then i guess it's up to you |
| 21:23 | <Hixie> | i can go either way |
| 21:23 | <Hixie> | per spec right now technically they do even out of their doc, though i should make it clearer either way |
| 21:29 | <roc> | I think video should be consistent with images |
| 21:30 | <Hixie> | right-o |
| 21:33 | gsnedders | waves his maths working at #whatwg |
| 21:33 | <gsnedders> | Anyone able to see the mistake? |
| 21:34 | <Philip`> | gsnedders: I think you calculated x wrongly |
| 21:34 | <gsnedders> | Philip`: Nah, that's given. |
| 21:35 | <Philip`> | gsnedders: Oh, maybe you got y wrong then |
| 21:35 | <gsnedders> | Philip`: There is no y. |
| 21:35 | <gsnedders> | This is question five, BTW. |
| 21:36 | <Philip`> | Hmm, beats me |
| 21:36 | <gsnedders> | It would work if I got ln 2/3 + ln 2/3 and not ln 2/3 + ln 3/2 |
| 21:57 | <gsnedders> | "A solid is formed by rotating the curve $y = e^{-2x}$ between $x = 0$ and $x = 1$ through $360\circ$ about the $x$-axis. Calculate the volume of the solid that is formed." |
| 21:57 | <gsnedders> | Why would I ever want to be able to do this? |
| 21:57 | <gsnedders> | :P |
| 22:00 | <roc> | I've got one of those at home |
| 22:00 | <roc> | makes a great toilet plunger |
| 22:01 | <gsnedders> | And Boris realizes I'm being dumb on /. |
| 22:01 | <gsnedders> | roc: So what's its volume? |
| 22:01 | <gsnedders> | (Or rather: why would I care?) |
| 22:02 | <roc> | so you know how much rubber to buy to make it |
| 22:02 | <gsnedders> | But then surely I care about the circumference at the bottom and not the volume? |
| 22:03 | <roc> | I give up |
| 22:04 | gsnedders | concludes that it has no use and ignores the question |
| 22:10 | <Hixie> | why is knowing the volume not useful? |
| 22:10 | <Hixie> | for the case roc gave? |
| 22:10 | <Philip`> | gsnedders: Isn't it just something like integral(pi*y^2, dx)? |
| 22:10 | <gsnedders> | Philip`: Well, there's some equation, yeah |
| 22:11 | <Philip`> | gsnedders: What I mean is "Isn't it exactly integral(pi*y^2, dx)? (except if that's wrong then it's just because I've not spent much time thinking about it, not because I'm stupid, honest)" |
| 22:12 | <gsnedders> | Philip`: The answer to which is "I don't know because I can't be bothered looking it up or thinking about it." |
| 22:13 | <Hixie> | is RT retweet? |
| 22:13 | <Lachy> | Hixie, I think it is. |
| 22:13 | <gsnedders> | Hixie: Yeah |
| 22:13 | <Lachy> | today's the first time I've seen that too |
| 22:13 | <gsnedders> | Sorry Hixie, forgive me :P |
| 22:14 | <Philip`> | gsnedders: You don't need to think about it, it's just obvious that the volume is the sum of lots of discs, and sums are integrals, and discs have volume pi*r^2*dx, and that's about it :-) |
| 22:15 | <Philip`> | and the mere existence of a question is enough reason to attempt to solve it |
| 22:18 | <gsnedders> | http://flickr.com/photos/gsnedders/3331908722/ |
| 22:18 | <gsnedders> | (MLW take note) |
| 22:25 | <Hixie> | i sure hope this RT thing doesn't take off |
| 22:25 | Hixie | reads the latest AB minutes and raises several eyebrows |
| 22:26 | <Dashiva> | RT? |
| 22:26 | <Lachy> | AB? |
| 22:26 | <Lachy> | Dashiva, re-tweet |
| 22:27 | <Dashiva> | Is that the twitter-IRC interface thing? |
| 22:27 | <Lachy> | no |
| 22:28 | <deltab> | http://bloggingbits.com/the-art-and-science-of-retweeting-for-twitteraholics/ |
| 22:28 | <deltab> | reposting, in other words |
| 22:28 | <Lachy> | Dashiva, see http://twitter.com/gsnedders/statuses/1285533654 and http://twitter.com/migrosch/statuses/1283809127 |
| 22:29 | <hsivonen> | Hixie: URL? |
| 22:33 | <Dashiva> | Wow. Retweeting sounds like dupe heaven |
| 22:37 | <deltab> | surely it'd be better to give a link and your own thoughts or at least rephrasing |
| 22:37 | <rubys> | Hixie: ping? |
| 22:47 | <Hixie> | rubys: pong |
| 22:48 | <rubys> | My read is that the discussion on the proposed Excerpt License is simply going to rediscover the original use cases that we already sent, after a lot of email. |
| 22:49 | <rubys> | It occurs to me that if you happened to agree with the last two paragraphs of this: <http://lists.w3.org/Archives/Public/public-html/2009Mar/0144.html>; and were to say so, it might (just might) shorten the process a bit. |
| 22:49 | <Hixie> | that would certainly follow past patterns of public-html e-mail |
| 22:49 | <rubys> | Of course, if you disagree, feel free to say that also |
| 22:50 | <Hixie> | seems reasonable to me, which is why i didn't post again to the list after saying my bit about license proliferation :-) |
| 22:51 | <rubys> | I'm just suggesting that you do consider posting again. You're call. I agree with you about license proliferation, but given the brokenness of not having a dialog, it unfortunately could be seen as us raising the goal posts. |
| 22:52 | <Hixie> | rubys: i disagree with your earlier point (though i don't care enough to say so on the list), in that i think that the change you talk about happened in 2004 |
| 22:52 | <Hixie> | and the w3c license is really a change back to the old ways |
| 22:52 | <Hixie> | er |
| 22:52 | <Hixie> | <Hixie> rubys: i disagree with your earlier point (though i don't care enough to say so on the list), in that i think that the change you talk about happened in 2004 |
| 22:52 | <Hixie> | <Hixie> and the w3c license is really a change back to the old ways |
| 22:53 | <Hixie> | but if you want me to say +1 then i'm happy to do so |
| 22:53 | <Hixie> | it might set up a bad precedent, though :-) |
| 22:53 | <Hixie> | (public-html had a bad problem with people +1'ing each other historically) |
| 22:53 | <rubys> | your call |
| 23:01 | <Hixie> | rubys: sent |
| 23:01 | <Hixie> | i really wish dragging the w3c into modernity wasn't so painful a process |
| 23:04 | <gsnedders> | Well, you should be less demanding, and you should just be a post-modernist. |
| 23:58 | <Hixie> | woot, done <canvas>. |
