| 00:31 | <Hixie> | ok, today's e-mail has been dealt with |
| 00:35 | <othermaciej> | Hixie: I do think that sometimes conformance requirements stated in algorithmic form can be harder to work with, if you are not implementing the algorithm literally |
| 00:35 | <Hixie> | i agree |
| 00:35 | <othermaciej> | but to me that takes a back seat to precision |
| 00:36 | <othermaciej> | for example, for img.width and img.height |
| 00:36 | <othermaciej> | there's probably the following important criteria: |
| 00:36 | <Hixie> | if there are any algorithms that someone can find a non-algorithmic way to describe the same requirements, let me know |
| 00:36 | <othermaciej> | - the height and width should be either 0 or the correct value |
| 00:36 | <othermaciej> | - the height and width should only change from 0 to the final value barring the final load, no intermediate different values |
| 00:37 | <othermaciej> | - height and width should be available no later than the "load" even |
| 00:37 | <othermaciej> | *event |
| 00:37 | <othermaciej> | - height and width should be available at the same time |
| 00:37 | <othermaciej> | I'm not sure all of these can be easily inferred from the algorithm, simple though it is |
| 00:38 | <othermaciej> | for parsing it is likely impossible to express equivalent requirements in declarative rather than as-if imperative form |
| 00:38 | <othermaciej> | I haven't really studied the spec with an eye to this |
| 00:38 | <othermaciej> | I would expect improvements along these lines, and pointing out where the spec overconstrains implementations, will happen in CR |
| 00:39 | <othermaciej> | Larry seems to underestimate how different the internals of different browser engines are, at least when it comes to details |
| 00:42 | <Hixie> | i think your height/width definitions miss some things |
| 00:43 | <Hixie> | e.g. that the values are the rendered dimensions not the actual dimensions |
| 00:43 | <Hixie> | when the image is rendered |
| 00:43 | <othermaciej> | it wasn't intended to be complete |
| 00:43 | <Hixie> | ok |
| 00:43 | <Hixie> | the problem is that making it complete ends up taking three times more room than what we have now |
| 00:43 | <othermaciej> | I was just thinking about some things that I am pretty sure are needed for interoperability |
| 00:43 | <Hixie> | and it's not clear that it's better |
| 00:43 | <othermaciej> | it wasn't suggested to be an alternative |
| 00:43 | <Hixie> | yeah |
| 00:44 | <Hixie> | larry seems to want an alternative though |
| 00:44 | <othermaciej> | it was a thought experiment, to see if the current conformance requirements imply the things that are needed for interop |
| 00:44 | <Hixie> | ah i see |
| 00:44 | <othermaciej> | my impression is that the most important for non-rendered images is that you have width and height by the time load occurs |
| 00:45 | <othermaciej> | I think an image has to be "available" by the present definition no later than the time the "load" event is fired |
| 00:45 | <othermaciej> | I think width and height becoming available at the same time likely matters for interop, even though Larry thought that was an overconstraint |
| 00:46 | <Hixie> | yes, 'load' fires by definition after the image is 'available' |
| 00:46 | <Hixie> | per spec |
| 00:47 | <othermaciej> | he may have a point that the spec overconstrains things needlessly at times, though his specific example does not seem to me a very good one |
| 00:47 | <othermaciej> | plus I do think implementors often point it out when there are overconstraints |
| 00:48 | <othermaciej> | (consider e.g. concerns arising from multiprocess rendering and various serial behavior rules) |
| 00:48 | <othermaciej> | but still, I'd prefer overconstraint to the sever underspecification of HTML4 |
| 00:49 | <othermaciej> | the cost of an unnecessary conformance requirement is less than the cost of a missing one |
| 00:49 | <othermaciej> | if one assumes the spec will continue to be maintained |
| 00:50 | <Hixie> | i'm sure the spec is both over- and under-constrained in parts |
| 00:50 | <Hixie> | since i am sadly not perfect :-) |
| 00:50 | <othermaciej> | anyway, even if I disagree with Larry on many things, I think his substantive commentary using a concrete example is much more helpful than title trolling |
| 00:50 | <Hixie> | it's certainly a step up |
| 00:50 | <annevk4> | Hixie, I mentioned earlier over IRC that "valid image" and still being downloaded are somewhat at odds with each other |
| 00:50 | <Hixie> | i can actually address his concerns when he mentions them |
| 00:50 | <othermaciej> | as was Sam's identification of a specific issue bothering him |
| 00:51 | <Hixie> | annevk4: i removed "valid image" from the spec a few minutes ago |
| 00:51 | <othermaciej> | since it turns out that issue is already slated to be fixe |
| 00:51 | <annevk4> | Hixie, I suppose that helps :) |
| 00:51 | <Hixie> | the content-sniffing thing? |
| 00:51 | <Hixie> | adam said he's doing some work on it next week |
| 00:51 | <Hixie> | iirc |
| 00:52 | <Hixie> | and then we can submit it for rfc or something |
| 00:52 | <Hixie> | i don't fully understand how the ietf works |
| 00:52 | <Hixie> | did you know that once you submit an rfc, you can no longer fix it? |
| 00:52 | <Hixie> | it's amazing there are only a few thousand rfcs so far, given that |
| 00:52 | <Hixie> | i guess errors just go unfixed for years |
| 00:53 | <othermaciej> | so there's no provision for errata, you have to make a new RFC? |
| 00:54 | <othermaciej> | that seems kind of wacky, it would be like locking down a W3C spec at FPWD |
| 00:55 | <othermaciej> | and yeah it is amazing how rarely there are updates to fix errors |
| 00:55 | <annevk4> | I saw you can report errata against an RFC in some fashion (not sure if it's official) |
| 00:56 | <Hixie> | i believe they consider RFC publication more like CR |
| 00:59 | <annevk4> | e.g. with URIs it took several RFCs to become a STD |
| 00:59 | <othermaciej> | yeah but it seems like even the "Proposed Standard" state is one where one would expect further changes are needed |
| 00:59 | <othermaciej> | there's nothing wrong per se with using a new RFC number for that |
| 00:59 | <othermaciej> | just surprising it hasn't happened more |
| 01:01 | <othermaciej> | interesting, RFC2068 was a Proposed Standard but RFC2616 which obsoleted it isn't yet (a decade after coming out) |
| 01:10 | <Hixie> | annevk4: any idea what happens when you set xhr.responseXML.body.onload ? |
| 01:10 | <Hixie> | (re http://www.w3.org/Bugs/Public/show_bug.cgi?id=6515 ) |
| 01:13 | <annevk4> | it seems that by the time you can access that it has already fired |
| 01:13 | <annevk4> | but no, I haven't tested |
| 01:15 | <Hixie> | fired? i would have presumed that no events ever fire on those documents |
| 01:15 | <Hixie> | by the ua, anyway |
| 01:15 | <Hixie> | i shall poke around |
| 01:18 | <annevk4> | you once made a testcase with form submission in such a document and expected submit to fire |
| 01:18 | <annevk4> | and it did |
| 01:18 | <annevk4> | iirc |
| 01:19 | <Hixie> | ah |
| 01:20 | <annevk4> | http://twitter.com/heycam/status/1918008204 hehe |
| 01:20 | <Hixie> | i don't understand why this is an issue for html but not svg |
| 01:21 | <Hixie> | did the svgwg get feedback along the lines of "you should have a subtitle that says you're a browser spec"? |
| 01:21 | <heycam> | SVG Tiny 1.2: A specification that defines an XML-based vector graphics language, associated APIs, and other APIs for non-graphics related things that we needed but weren't specified elsewhere |
| 01:21 | <heycam> | nope |
| 01:21 | <annevk4> | Hixie, it did get feedback that it did too much |
| 01:22 | <heycam> | but not that it needed to call that out in the title |
| 01:22 | <heycam> | (and in the end we did rip a lot of stuff out) |
| 01:22 | <annevk4> | right |
| 01:23 | <annevk4> | but don't tell people they're using the title as an argument for something else because they don't like that |
| 01:27 | <annevk4> | hmm yeah, SVG 1.2 didn't call out any of those things in title/subtitle/abstract |
| 01:28 | <annevk4> | but I don't really have a problem with calling out page load processing in the subtitle/abstract |
| 01:38 | <Hixie> | i was considering coming up with a boatload of alternative subtitles |
| 01:38 | <Hixie> | and just cycling through them using a script |
| 01:38 | <Hixie> | so every time you load the spec it gives you a new one |
| 01:39 | <othermaciej> | that would be lols |
| 01:45 | <annevk4> | not exactly a new concept: http://krijnhoetmer.nl/irc-logs/ :) |
| 01:55 | annevk4 | takes another stab at sleeping |
| 04:55 | <Hixie> | oooh tough one |
| 04:55 | <Hixie> | marquee has some custom event handler attributes |
| 04:55 | <Hixie> | do i add them to the general set? |
| 04:55 | <Hixie> | or do i make marquee an exception? |
| 04:56 | <Hixie> | you know, i think the thing i find annoying about a lot of the people who keep complaining about html5 in various ways is that they never give advice for cases like this |
| 04:56 | <Hixie> | so it feels like they're complaining and complaining but not ever being constructive in areas that aren't their pet peeve |
| 05:01 | <sayrer> | http://www.rfc-editor.org/errata.php |
| 05:01 | <sayrer> | http://www.rfc-editor.org/errata.php |
| 05:01 | <sayrer> | fwiw |
| 05:01 | <sayrer> | most errata lists I've encountered seem to be informally |
| 05:02 | <sayrer> | maintained |
| 05:02 | <sayrer> | like, by people who actually need to consult them |
| 05:08 | <Hixie> | doesn't seem to have any serious errata for any of the rfcs i looked at |
| 05:34 | <Hixie> | i feel like leif is trying to make a point but i have no idea what that point is |
| 05:35 | <Hixie> | i wish he'd just come out and say it instead of quoting me and implying his point |
| 05:49 | <sayrer> | Hixie, there is a provision for errata, that's all |
| 05:49 | <sayrer> | and, you can sort of just make them |
| 05:49 | <sayrer> | like the HTTP errata that were maintained for years |
| 05:49 | <Hixie> | indeed |
| 05:49 | <Hixie> | errata are a poor way of maintaining specs, though, as I learnt with CSS |
| 05:50 | <sayrer> | sure, when the http-bis thing started |
| 05:50 | <sayrer> | figuring out what to do with the errata was a problem |
| 05:51 | <sayrer> | dang, larry found a section of the spec I would like to delete |
| 05:51 | <sayrer> | I guess the little google summaries might count |
| 05:58 | <othermaciej> | the Outline section? |
| 05:58 | <othermaciej> | I am not really sure what that is for |
| 05:58 | <othermaciej> | what kind of client produces or displays an outline? |
| 05:59 | <Hixie> | accessibility tools, amongst others |
| 05:59 | <Hixie> | but that section is present not so much for tools, so much as to make sure that it is well-defined what the heading is for each part of the document |
| 05:59 | <othermaciej> | VoiceOver doesn't make an outline and if it did, I'm not sure it would necessarily want to use that algorithm to compute it |
| 06:00 | <Hixie> | one of the questions that kept coming up with html4 was "what exactly is my document structure" |
| 06:00 | <othermaciej> | perhaps "what the heading is for each part of the document" has more immediate relevance |
| 06:04 | <sayrer> | well-defined sounds pretty circular |
| 06:05 | <sayrer> | Hixie, interesting use of passive voice. who was asking the question? |
| 06:13 | <Hixie> | many people |
| 06:13 | <Hixie> | i didn't take names |
| 06:14 | <othermaciej> | really in practical terms this section is only defining semantics, not conformance requirements for typical UAs |
| 06:14 | <othermaciej> | but doing it in algorithm terms |
| 06:14 | <Hixie> | yes |
| 06:14 | <Hixie> | i'm happy to phrase it in non-algorithmic terms if anyone can come up with a sane way of describing it that way |
| 06:14 | <Hixie> | i couldn't work out how to do it |
| 06:17 | <othermaciej> | I don't find the definition of this particular piece of semantics very important |
| 06:17 | <othermaciej> | I can see how it might be hard to define in purely declarative terms without potentially missing cases |
| 06:23 | <hsivonen> | pasting some URLs for to have a record in the log: |
| 06:24 | <hsivonen> | http://twitter.com/jdowdell/status/1608188445 |
| 06:24 | <hsivonen> | http://twitter.com/jdowdell/status/1352144527 |
| 06:24 | <hsivonen> | http://twitter.com/jdowdell/status/1344855975 |
| 06:24 | <hsivonen> | http://twitter.com/jdowdell/status/1464529893 |
| 06:24 | <hsivonen> | (end of flood) |
| 06:26 | othermaciej | is not sure he gets the theme there |
| 06:26 | <hsivonen> | seems like jd considers the variability of HTML runtimes a problem, so I guess HTML5 should err on the side off well-defined behavior |
| 06:38 | <othermaciej> | oh you're quoting the Adobe trollblogger? |
| 06:44 | <hsivonen> | othermaciej: I'm quoting the Adobe Flash blogger who seems to blog a lot about HTML these days |
| 06:49 | <MikeSmith> | johnny one-note |
| 06:49 | <othermaciej> | hsivonen: after reading some of his recent posts, I stand by my prior assertion |
| 06:49 | <othermaciej> | MikeSmith: do you have administrative access to the w3c CVS server? |
| 06:51 | <MikeSmith> | othermaciej: yeah |
| 06:51 | <MikeSmith> | sorta |
| 06:52 | <MikeSmith> | I have access to the actual filesystem where the cvs ,v files are |
| 06:56 | <othermaciej> | yeah, he's basically Adobe's Asa Dotzler |
| 07:01 | <MikeSmith> | another possible competitor in this race - |
| 07:01 | <MikeSmith> | http://twitter.com/mattmay/status/1884053829 |
| 07:02 | <MikeSmith> | but I think he'll need to try harder than that if he really wants to win it |
| 07:02 | <othermaciej> | it makes Adobe look sad and desparate to try to fight against HTML having more features |
| 07:04 | <Hixie> | i don't think matt is fighting html having more features in that twitter |
| 07:05 | <Hixie> | since leaving the w3c would make progress on html5 far quicker and easier |
| 07:05 | <Hixie> | if anything, he's fighting _for_ html having more features |
| 07:06 | <othermaciej> | I wonder why John likes the catchphrase "browser brands" so much |
| 07:06 | <othermaciej> | why does he say that instead of "browsers"? is that an Adobe thing? |
| 07:09 | <hsivonen> | othermaciej: the impression I get is that he wants to portray browsers as different chrome designs |
| 07:09 | <othermaciej> | but wouldn't that portray the engines as essentially interchangeable? |
| 07:09 | <othermaciej> | which is contrary to his point? |
| 07:09 | <hsivonen> | othermaciej: I suppose |
| 07:10 | <hsivonen> | othermaciej: although I think the point is that you pick your favorite toolbar and run Flash in the space below it |
| 07:10 | <othermaciej> | ah |
| 07:14 | <othermaciej_> | "browser brand" is not a very common phrase outside his blog |
| 07:14 | <othermaciej_> | but yeah I can see how he might want to take the "browsers are just Flash loaders" position |
| 07:31 | <MikeSmith> | hsivonen: when you have a few minutes, I could use a bit of help figuring out how to write a test for valid surrogate pairs in accesskey values |
| 07:36 | <zcorpan_> | Hixie: yt? |
| 07:40 | <Hixie> | zcorpan_: am now |
| 07:54 | <Dashiva> | http://lists.w3.org/Archives/Public/public-html/2009May/0375.html seems to be a great example of how to abuse the design principles |
| 08:01 | <Hixie> | it's more an example of not understanding the principles |
| 08:01 | <Hixie> | it might be worth just removing the names of the principles and just numbering them |
| 08:03 | <othermaciej> | tempting! but would make it hard to cite them |
| 08:04 | <Hixie> | that's the point :-) |
| 08:04 | <Hixie> | it would also mean that they couldn't be mischaracterised based on their names |
| 08:05 | <Hixie> | but seriously, the biggest problem with the design principles is that people have tried to apply them like blind rules |
| 08:05 | <Hixie> | instead of taking them as what they are -- principles of design |
| 08:07 | <Dashiva> | Could give them code names. Easier to remember without having names to infer meaning from |
| 08:08 | <othermaciej> | people also object to the fact that I said they are just guidelines and rules of thumb not absolute laws |
| 08:08 | <othermaciej> | the one that bugs me the most abuse-wise is "Pave the Cowpaths" |
| 08:08 | <othermaciej> | where opponents of it keep claiming it is being misused |
| 08:08 | <othermaciej> | in a way that, as far as I can tell, no one in favor of the principle actually does |
| 08:11 | <Hixie> | people always do that |
| 08:11 | <Hixie> | it's the same as the way people quote my "handling-people" file and say "look! he's doing this and that!" when they are losing an argument |
| 08:11 | <hsivonen> | looks like removing the head of the pin generates more heat than deciding how many angels can dance on it |
| 08:11 | <hsivonen> | (profile) |
| 08:11 | <othermaciej> | how dare you use the 2 + 2 = 4 principle to claim that 1 + 1 is an odd number! |
| 08:11 | <othermaciej> | that's an abuse of the principle! |
| 08:11 | <othermaciej> | "... wait, what?" |
| 08:12 | <hsivonen> | Hixie: does one lose an argument if one catches someone else than you practicing "handling people"? |
| 08:13 | <Hixie> | there's never any benefit to pointing it out, since people would always just deny it |
| 08:13 | hsivonen | looked up the sabotage manual from Gutenberg the other day |
| 08:13 | <Hixie> | that file is just useful as a way to know what to look for |
| 08:13 | <Hixie> | since once you spot someone doing something from that list, you can then apply appropriate countermeasures |
| 08:13 | <Hixie> | instead of falling into their trap |
| 08:15 | <zcorpan_> | Hixie: we're implementing spellcheck, so if you're going to simplify the algorithm, we'd appreciate if you did it soonish :) |
| 08:16 | <Hixie> | noted |
| 08:18 | <othermaciej> | crap, I forgot that IE means I can't use CSS generated content |
| 08:19 | <zcorpan_> | othermaciej: ie8 supports it |
| 08:19 | <othermaciej> | would it be ok for W3C documents to be IE8-and-up? |
| 08:19 | <zcorpan_> | sure |
| 08:19 | <zcorpan_> | i'm sure the html5 spec doesn't work correctly in ie :) |
| 08:21 | <zcorpan_> | othermaciej: although you should expect people to complain because they use ie7 and a screen reader and now you're against blind people |
| 08:21 | <othermaciej> | well |
| 08:22 | <othermaciej> | I could preface every example with the words "Example:" by hand |
| 08:22 | <othermaciej> | but that seems annoying |
| 08:22 | <othermaciej> | was trying to decide whether to do that instead of a :before rule |
| 08:23 | <zcorpan_> | othermaciej: the html5 spec uses generated content to insert "Note:" and "Warning:" |
| 08:24 | <Hixie> | the spec gen script will eventually do this for us |
| 08:24 | <Hixie> | once gsnedders is out of school or whatever |
| 08:25 | <zcorpan_> | when you're in school you have plenty of time |
| 08:25 | <othermaciej> | would anyone like to request quick fixes to the Design Principles or point out errors, while I am editing? |
| 08:25 | <Hixie> | baby steps! |
| 08:25 | <Hixie> | baby steps! |
| 08:26 | <zcorpan_> | othermaciej: i sent an email before |
| 08:26 | <othermaciej> | Hixie: oh that's right |
| 08:27 | <zcorpan_> | othermaciej: s/irrelevant/hidden/ |
| 08:27 | <othermaciej> | Hixie: I was thinking of changing Evolution not Revolution to "Incremental Improvement" so it can include the spirit of "baby steps" without causing a fight over the name "baby steps" |
| 08:27 | <othermaciej> | zcorpan_: thanks! |
| 08:27 | <othermaciej> | Hixie: I think that's too substantive a change to sneak in with minor edits though |
| 08:27 | <Hixie> | i don't care what you call it, but the key point is not adding so many new features that implementors get lost behind and never catch up |
| 08:28 | <othermaciej> | Hixie: I think you might already be ahead of the curve on that... |
| 08:28 | <Hixie> | yeah |
| 08:28 | <othermaciej> | although there aren't many things I would call for removing at the moment |
| 08:28 | <Hixie> | hence how important it is! |
| 08:28 | <Hixie> | it was on my feedback on that survey n years ago |
| 08:28 | <Hixie> | on the list of things i'd like to see before we publish |
| 08:28 | <Hixie> | if that helps :-) |
| 08:30 | <othermaciej> | I guess I can include it as part of addressing the old feedback |
| 08:31 | zcorpan_ | ponders about an example to Solve Real Problems |
| 08:34 | <Hixie> | Ollie: hi :-) |
| 08:35 | <othermaciej> | would anyone care to help me proofread http://dev.w3.org/cvsweb/~checkout~/html5/html-design-principles/Overview.html?rev=1.24 before I note it on the list? |
| 08:35 | <Ollie> | Hixie: Hi there :) |
| 08:35 | <Hixie> | othermaciej: "The working group is working on a new version of HTML not yet published under TR" is no longer accurate |
| 08:36 | <Dashiva> | gsnedders: For extra credit, show this to the teacher: http://2.media.tumblr.com/9pDtThz04nxemo71EGVHHWQgo1_500.gif |
| 08:36 | <hsivonen> | othermaciej: I'll read |
| 08:37 | <hsivonen> | othermaciej: it's no longer the *First* public WD |
| 08:37 | <hsivonen> | othermaciej: what Hixie said |
| 08:38 | <othermaciej> | fixed what Hixie said, changed front matter to say "an Editor's Draft" |
| 08:38 | <hsivonen> | othermaciej: it's probably unnecessary to put the vote tally up there |
| 08:39 | <othermaciej> | where is there a vote tally? |
| 08:39 | <Hixie> | in the SotD |
| 08:39 | <hsivonen> | othermaciej: yes. the next paragraph after that is odd, too |
| 08:39 | <othermaciej> | should I just remove that part? |
| 08:40 | <othermaciej> | I think I will |
| 08:40 | <othermaciej> | removed both those paragraphs |
| 08:40 | <hsivonen> | thanks |
| 08:42 | <Hixie> | othermaciej: i don't think "Evolution Not Revolution" really conveys the "do things slowly" approach i think is important |
| 08:42 | <othermaciej> | Hixie: I don't think it does either! |
| 08:42 | <othermaciej> | I would like to edit it to do so in addition to changing the title |
| 08:42 | <Hixie> | k |
| 08:42 | <othermaciej> | changing the name rather |
| 08:42 | <othermaciej> | I think one principle can cover both concepts |
| 08:42 | <othermaciej> | don't boil the ocean, and don't try to do too much at one go |
| 08:42 | <othermaciej> | if I may speak in shorthand |
| 08:49 | <hsivonen> | othermaciej: looks good. However, I notice that you didn't bother to annotate each priciple by saying if it applies to conforming language or supported language |
| 08:49 | <othermaciej> | I guess that's another thing I should fix as part of the next pass |
| 09:41 | <MikeSmith> | use of Unicode surrogate pairs is restricted to content encoded in UTF-16, right? |
| 09:42 | <Hixie> | in theory |
| 09:45 | <hsivonen> | MikeSmith: scripts could introduce lone surrogates to pages that had any encoding |
| 15:10 | <annevk42> | so I follow a link to sitepoint and say yes to answering some survey |
| 15:10 | <annevk42> | first question "Based on today's visit, how would you rate your site experience overall?" ... |
| 15:11 | <annevk42> | second "Which of the following best describes the primary purpose of your visit?" doesn't list "following a link" as one of the options... |
| 15:11 | <annevk42> | third "Were you able to complete the purpose of your visit today?" only gives yes/no |
| 21:15 | <gsnedders> | Oh well, I guess we should change the interface then :D |
| 21:17 | <jgraham> | Dashiva: good point |
| 21:18 | <hsivonen> | cgriego: not deployed. I'll deploy it in 12 hours or so |
| 21:19 | <gsnedders> | jgraham: LOL |
| 21:19 | <gsnedders> | jgraham: I do like how you're using my implementation of the outlining algorithm as an example and not your own |
| 21:45 | <ezyang> | I think the doctype syntax is also ambiguous |
| 21:46 | <ezyang> | Since if I only specify systemId, it's not distinguishable from publicId |
| 21:49 | <ezyang> | Anyway, pushed the DOM -> test format converter. |
| 21:49 | <ezyang> | Probably the most dirty piece of code I've written in a long while. |
| 21:59 | <jgraham> | gsnedders: Yours has the advantage of being relatively well tested and well commented |
| 22:00 | <gsnedders> | jgraham: It's almost entirely untested, it's just people rely upon it! |
| 22:00 | <gsnedders> | (Those idiots) |
| 22:02 | <ezyang> | Developers want shiny things NAO. |
| 22:03 | <jgraham> | gsnedders: It is known to work in some situations |
| 22:03 | <jgraham> | In other news we should write some tests for this stuff |
| 22:04 | <gsnedders> | jgraham: "news"? :D |
| 22:05 | <jgraham> | News of the form "MPs claim everything they can on expenses" i.e. everyone already knew, it has been explicit for months and yet it is still "news" |
| 22:10 | <ezyang> | Excellent! 369 failures, but at least it doesn't fatally error out. |
| 22:11 | <gsnedders> | ezyang: BTW, do you have exams? When? |
| 22:11 | <ezyang> | It looks like there are some stray newlines |
| 22:11 | <ezyang> | gsnedders: I'm done with exams ;-) |
| 22:11 | <gsnedders> | ezyang: re your latest tweet, @id :) |
| 22:11 | <gsnedders> | ezyang: Ah, I still have one to go. |
| 22:11 | <ezyang> | Ok. So they totally got rid of it? |
| 22:11 | <ezyang> | Heh |
| 22:11 | <gsnedders> | yeah |
| 22:11 | <ezyang> | Good luck mate |
| 22:12 | <gsnedders> | Like, computing exam. The one I haven't done any work for all year. The one I got 77/80 in the coursework. The one I barely need to pass to get a A. |
| 22:12 | <ezyang> | Oh hoho |
| 22:12 | <gsnedders> | I had physics exam… I was thirty minutes late, because I thought it was on Thursday, until I got a phone call from school! |
| 22:13 | <gsnedders> | s/…/today…/ |
| 22:13 | <ezyang> | You're scott free. Play video games or work on html5lib or something |
| 22:13 | <ezyang> | Eek. At least they called. |
| 22:13 | <gsnedders> | ezyang: Hixie's answer would be work on Anolis :) |
| 22:13 | <ezyang> | I'm still not quite sure what anolis is. |
| 22:13 | <gsnedders> | An HTML document post-processor :P |
| 22:14 | <gsnedders> | (Sorry, that's what I call it because I couldn't really work out what to call it.) |
| 22:14 | <ezyang> | Ohh, like XHTML Compiler? |
| 22:14 | <gsnedders> | What? |
| 22:14 | <ezyang> | It like, takes HTML files and turns them into more awesome HTML files |
| 22:15 | <gsnedders> | What it does is take HTML files and turn them into more awesome HTML files with things like a table of contents and cross-references. |
| 22:15 | <ezyang> | Yeah. I've done this before. |
| 22:15 | <ezyang> | That's neat. What's the language? |
| 22:15 | <gsnedders> | Python. |
| 22:15 | <gsnedders> | It's kinda slow. |
| 22:16 | <gsnedders> | Well, for the biggest user of it it is. |
| 22:16 | <ezyang> | Slow == ? |
| 22:16 | <gsnedders> | But that's Hixie, and HTML 5 is a kinda big file. :) |
| 22:16 | <ezyang> | Aha. |
| 22:16 | <ezyang> | Yeah, HTML 5 is big. |
| 22:16 | <gsnedders> | It's been a while since I've run it against HTML 5 myself. I can't remember how long it takes. |
| 22:17 | <gsnedders> | I'll probably someday get around to rewriting it in some other language, possibly Java (seeming that has an HTML 5 parser for it, and is quick, once the VM is running). |
| 22:18 | <ezyang> | Ummm... I'm not convinced you actually want to rewrite it in Java. |
| 22:18 | <Hixie> | "once the VM is running" isn't helpful since the VM would have to be started each time in most scenarios :-) |
| 22:19 | <ezyang> | Anyway, I presume you're loading the document entirely into memory and then doing transforms? |
| 22:20 | <jgraham> | gsnedders: Just wait for the python VM nd/or Hixie's computer to get faster |
| 22:20 | <Hixie> | actually i use your computer :-P |
| 22:20 | <gsnedders> | jgraham: Hixie doesn't own the computer. :P |
| 22:20 | <jgraham> | Oh |
| 22:21 | <gsnedders> | Hixie: But on a webservice, like v.nu, it's fine, because the start-up time doesn't matter. |
| 22:21 | <gsnedders> | ezyang: Yes. |
| 22:21 | <Hixie> | if jgraham can set it up that way :-) |
| 22:21 | <jgraham> | Well wait for my sebserver to get faster which still should happen rather faster than HTML 5 grows |
| 22:21 | <jwalden> | ezyang: hey, I think I remember seeing your name in occasional skims of r-h-t over the past year, are you living in random? (I was, graduated last year) |
| 22:21 | jgraham | is filled with fear at the hought of trying to run java on the server |
| 22:22 | <jgraham> | I mean I only get about 1 in 2 java applications to run on my own machine |
| 22:23 | <ezyang> | jwalden: Yeah! |
| 22:23 | <jwalden> | ezyang: clam, or elsewhere? |
| 22:23 | <jwalden> | I was on clam first, third, fourth years, pecker third |
| 22:24 | <jgraham> | In no way is MIT a weird cult. At all. Honest. |
| 22:24 | <gsnedders> | Hixie: I would likely do so :) |
| 22:27 | Dashiva | bakes jgraham on a slow roast |
| 22:27 | jgraham | is not sure what he did o deserve being roasted |
| 22:29 | <Dashiva> | You mentioned cult after clams had been brought up |
| 22:29 | gsnedders | had people at TPAC trying to bully him into applying to places like MIT |
| 22:30 | <ezyang> | gsnedders: We've touched on this topic before, but do you know where you're going? |
| 22:31 | <gsnedders> | ezyang: No. Well, Opera for this summer. |
| 22:31 | <ezyang> | Nice! |
| 22:31 | <ezyang> | So you get to do the runaround this winter :-P |
| 22:31 | <jgraham> | Can you roast clams? |
| 22:32 | <gsnedders> | ezyang: Seeming you're from the US you might have a clue: when would I have to do SAT by? |
| 22:33 | <ezyang> | Erm, why do you want to take the SAT? |
| 22:33 | <ezyang> | As a non-US person, you can't get away with, say, IBs? |
| 22:33 | jgraham | knows you have to take some special exams to go to grad school in the US |
| 22:34 | <jwalden> | MIT was awesome, crazy being in a place where so many other people are just as weird as you are :-D |
| 22:34 | <gsnedders> | ezyang: At places like MIT, no. |
| 22:35 | <jwalden> | last I heard IBs were never really in style, and even APs are going out these days |
| 22:35 | <ezyang> | I don't know the answer to your question offhand, since I did my SATs really early. |
| 22:35 | <ezyang> | collegeboard.com can probably tell you when the last time for seniors to take it is. You'll want to get your scores in before the application date. |
| 22:36 | <ezyang> | Also, if you don't know about it already, College Confidential is a really awesome resource. I'm a crufty college student turning Sophomore, so I've basically forgotten everything about the process. |
| 22:37 | <jgraham> | gsnedders: January |
| 22:38 | <gsnedders> | jgraham: Says where? |
| 22:38 | <jgraham> | The internet is a wonderful thing |
| 22:38 | <jgraham> | http://www.mitadmissions.org/topics/apply/international_applicants_helpful_tips/index.shtml#testingrequirements |
| 22:41 | <hendry> | evening, anyone have any thoughts on Website backups to share? http://kitenet.net/~joey/blog/entry/website_backup_RFC/ |
| 22:41 | <ezyang> | Awesome! 84 passes. That's totally commit worthy. :-D |
| 22:42 | <ezyang> | (285 failures, btw) |
| 22:42 | <ezyang> | gsnedders: Shall I commit? |
| 22:42 | <gsnedders> | ezyang: Yup |
| 22:43 | <ezyang> | Oh, wait, it's really chatty. Let me remove those var_dumps |
| 22:46 | <ezyang> | Pushed. gsnedders, have fun! |
| 22:46 | <gsnedders> | Anyhow, time for me to go sleep. |
| 22:46 | <ezyang> | (me, I'm going outside for a walk) |
| 22:46 | <ezyang> | Ahaha |
| 22:46 | <ezyang> | Good night! |
| 22:46 | <gsnedders> | Actually, more realistically, eat a few biscuits, read Anna Karenina for a couple of hours, then sleep. |
| 23:08 | <Hixie> | #$%^*@$^@ |
| 23:09 | Hixie | finally admits to himself that he needs to reverse engineer <marquee> |
| 23:09 | <Hixie> | i was really hoping to be able to define html5 without doing this |
| 23:11 | <ezyang> | What needs to be reverse engineered about it? |
| 23:12 | <Hixie> | how does loop=4 work? |
| 23:12 | <Hixie> | if you change the behavior, does the looping restart? |
| 23:13 | <ezyang> | Aha |
| 23:13 | <ezyang> | That doesn't sound too difficult, albeit slightly time consuming |
| 23:15 | <Hixie> | it's not difficult, no :-) |
| 23:25 | <Hixie> | great, all the browsers act weirdly and differently from each other |
| 23:25 | <Dashiva> | And oddly enough I suspect that's not an ironic "great", but actually a positive thing? |
| 23:26 | <ezyang> | Hixie: In cases like this, is documenting existing behavior or deciding on a standard the first priority? |
| 23:31 | <Hixie> | ezyang: what's the difference? |
| 23:31 | <Hixie> | Dashiva: it was ironic, though i guess it is actually good too, gives me more flexibility :-) |
| 23:45 | <jwalden> | browsers don't actually create anything new, they just copy like mad all the time :-) |
| 23:45 | <jwalden> | so writing a browser standard is just more copying :-) |
| 23:49 | <othermaciej> | we were all kinda trying to reverse engineer IE's <marquee> |
| 23:49 | <othermaciej> | if you can reverse engineer more accurately then win |
