| 00:00 | <TabAtkins> | ADD ME OR THE CAT GETS IT. |
| 00:00 | <Hixie> | hmmm |
| 00:00 | <Hixie> | which cat? |
| 00:00 | <TabAtkins> | ALL OF THEM |
| 00:00 | <Hixie> | wait what is this "Forms feedback" e-mail |
| 00:00 | <Hixie> | how dare you suggest i was actually working in april |
| 00:00 | <Hixie> | it's clear from the graph that i haven't worked since october |
| 00:01 | <TabAtkins> | Mail headers don't lie. |
| 00:01 | <Hixie> | oh actually i did do work in april, go figure |
| 00:01 | <Hixie> | that was the last time i did work until about a week ago |
| 00:02 | <Hixie> | i have 40 more e-mails from you |
| 00:02 | <Hixie> | i'm sure one of them will be convincing |
| 00:03 | <TabAtkins> | I have 66 matching 'from:ian⊙hc "tab atkins" label:whatwg' |
| 00:05 | <TabAtkins> | The majority of your responses to me consist of or start with the word "Indeed". |
| 00:05 | <TabAtkins> | So even if I don't trigger changes, I'm apparently right most of the time when I say something. |
| 00:07 | <erlehmann> | haha |
| 00:08 | <micheil> | man.. this is seriously how we roll in #whatwg. |
| 00:10 | <TabAtkins> | JS should gain some ability to define an unclosure. I've seen several places where functions sound like a good idea for something, but the fact that they can close over arbitrary objects makes them too expensive or annoying. |
| 00:10 | <TabAtkins> | Serializing unclosures should be easy, frex. |
| 00:11 | <Hixie> | TabAtkins: Indeed |
| 00:11 | <TabAtkins> | See! ^^^ |
| 00:12 | <erlehmann> | browsers gonna browse |
| 00:16 | <Hixie> | TabAtkins: e-mails to which i reply "Indeed" are my favourite e-mails |
| 00:16 | <Hixie> | TabAtkins: but they don't trigger changes to the spec, so they don't land you in the acks :-P |
| 00:44 | <Hixie> | re earlier discussion, turns out i've received e-mails from a Lobotom Dysmon who didn't sign with the other name |
| 00:45 | <TabAtkins> | I highly doubt they're different people, still. |
| 00:45 | <Hixie> | indeed |
| 00:47 | AryehGregor | wonders if Hixie is referring to the previous discussion, or just said "indeed" again without thinking about it |
| 00:47 | <TabAtkins> | I've got him trained. |
| 00:48 | variable | wonders if AryehGregor is copying from my keyboard ;) --- I was about to say that |
| 00:49 | <Hixie> | i swear it wasn't on purpose! |
| 00:49 | <Hixie> | he just keeps saying things that make sense! |
| 00:49 | <TabAtkins> | Bwahaha |
| 00:52 | <AryehGregor> | It's fun when someone points out some distinctive quirk of someone else's manner of speech, and everyone else is like "Now that you point it out, that's so true." |
| 00:52 | <TabAtkins> | More fun when someone points it out for your own speech, and then everyone suddenly notices. |
| 00:52 | <AryehGregor> | Yes. |
| 00:52 | <AryehGregor> | Like what are quirks of my speech? |
| 00:52 | <AryehGregor> | It's hard to notice your own quirks. |
| 00:52 | <TabAtkins> | Most quirks don't transmit through IRC well. Too much self-editing takes place when writing in text. |
| 00:53 | <variable> | even more fun when someone points out the quirks of someone's speech and then they suddenly overdo or underdo it ;) |
| 00:53 | <AryehGregor> | That's certainly true, people's speech is much more distinctive than their writing. |
| 00:56 | <Hixie> | i spend so much time self-editing on irc that i can't actually speak properly anymore, because i'll try to edit things as i'm saying them and if i've already said the part that i'm trying to edit it all comes out garbled |
| 00:56 | <Hixie> | this is kind of a serious problem |
| 00:56 | <AryehGregor> | TabAtkins' most distinctive quirk of writing is his ludicrous overuse of neologistic abbreviations. |
| 00:56 | <AryehGregor> | Hixie, normal people go back and correct themselves . . . |
| 00:57 | TabAtkins | didn't realize he used neologistic abbreviations in the first place. |
| 00:57 | <AryehGregor> | Although if your speech gets rusty enough, you could get a TTS converter and just type things IRL too. |
| 00:57 | <AryehGregor> | TabAtkins, you use words like "frex" and "abspos" constantly. |
| 00:57 | <AryehGregor> | It's very noticeable (at least to me). |
| 00:57 | <AryehGregor> | I mean, maybe not strictly neologisms. |
| 00:57 | <TabAtkins> | "abs/fix/relpos" is standard terminology around CSS. ^_^ |
| 00:57 | <TabAtkins> | I do use "frex" a lot. |
| 00:57 | <variable> | AryehGregor, report them to the AAAAA (American Association Against Acroynym Abuse) |
| 00:58 | <AryehGregor> | I don't know if you made them up, but you're the one I hear them from first and most. |
| 00:58 | <TabAtkins> | I learned them from others in the CSSWG. |
| 00:58 | <AryehGregor> | [100727 17:59:32] <TabAtkins> Pretty sure there's an esolang named that, which used cat as its interpreter. |
| 00:58 | <AryehGregor> | "esolang" |
| 00:58 | <AryehGregor> | You just make these up all the time. |
| 00:58 | <TabAtkins> | Esolang is a real word. A neologism, sure, but not an abbreviation. |
| 00:58 | <AryehGregor> | Or at least you're *much* more prone to using them than everyone else. |
| 00:58 | <TabAtkins> | I learned it from the internet. |
| 00:59 | <AryehGregor> | It's an abbreviation for "esoteric language". |
| 00:59 | <AryehGregor> | Which is what they're normally called. |
| 00:59 | <TabAtkins> | Yeah, but who wants to type that? |
| 00:59 | <TabAtkins> | Bah. |
| 00:59 | TabAtkins | likes conlangs too. |
| 00:59 | Philip` | has only ever noticed one other person use "frex" |
| 01:00 | <TabAtkins> | I picked that one up from one specific person on an RPG board I frequented. |
| 01:00 | <AryehGregor> | Also stuff like this: <TabAtkins> 100b? That sounds weird. |
| 01:00 | <AryehGregor> | "100b" instead of "100 bits". |
| 01:00 | <TabAtkins> | That's standard terminology! |
| 01:00 | <Philip`> | He was in his 60s so I assumed it was a slightly old fashioned phrase rather than a neologism |
| 01:00 | <AryehGregor> | Yes, but most people don't normally say it anyway. |
| 01:00 | <AryehGregor> | You don't hear people talking about "64b architectures". |
| 01:00 | TabAtkins | actually says "frex" and "abspos" in real speech too. |
| 01:01 | <AryehGregor> | You use way more abbreviations of this sort than anyone else I know. |
| 01:01 | <AryehGregor> | (it's somewhat tricky to quantify what "of this sort" means, since obviously lots of people use many more abbreviations overall) |
| 01:01 | AryehGregor | was confused the first time he saw "frex", but figured it out eventually |
| 01:01 | <TabAtkins> | The only reason I don't say the "b" in real life is because there's not a strong convention for whether it means "b" or "B" when voiced (unlike the higher prefixes, where "B" is generally assumed). |
| 01:01 | <Philip`> | Surely "100b" means "4" |
| 01:02 | <TabAtkins> | Nah, that's 0b100. |
| 01:02 | <AryehGregor> | . . . also, it's exactly the same length as the word "bit" in speech? |
| 01:02 | <AryehGregor> | Not in x86 assembler! |
| 01:02 | <AryehGregor> | At least the dialect I learned. |
| 01:02 | <TabAtkins> | AryehGregor: That doesn't stop people from pronouncing WW2. |
| 01:02 | <AryehGregor> | It stops sane people from pronouncing WW2. |
| 01:02 | <AryehGregor> | And WWW. |
| 01:02 | <AryehGregor> | Well, you can't get around WWW sometimes. |
| 01:02 | <TabAtkins> | wuh-wuh-wuh. |
| 01:02 | <AryehGregor> | See, there you go trying to abbreviate again. |
| 01:02 | <Philip`> | wow-ow |
| 01:02 | <AryehGregor> | I bet you do it in speech too. |
| 01:02 | <TabAtkins> | I do. |
| 01:02 | <AryehGregor> | No, pronouncing "www" is easy, just pretend it's Welsh. |
| 01:02 | <AryehGregor> | "oooooo". |
| 01:03 | <TabAtkins> | Actually, in speech I pronounced it "dub-dub-dub". |
| 01:03 | <AryehGregor> | Yeah, see? |
| 01:03 | <AryehGregor> | I told you. |
| 01:03 | <TabAtkins> | I pronounce all my acronyms. |
| 01:03 | <TabAtkins> | hetimmle, frex. |
| 01:03 | <AryehGregor> | Even OHRRPGCE? |
| 01:03 | <TabAtkins> | If I knew what that was and had occasion to say it, maybe! |
| 01:04 | <AryehGregor> | http://hamsterrepublic.com/ohrrpgce/index.php/Main_Page.html |
| 01:04 | <AryehGregor> | I started to write a game using that back in . . . 2002 or something, probably? |
| 01:04 | <AryehGregor> | When I was like 14 or whatever. |
| 01:05 | <TabAtkins> | K, so, if I was involved in that community for some reason and had a reason to type the acronym, I'd totally pronounce it, since I hear myself speak what I'm typing. |
| 01:05 | <AryehGregor> | I'm amused that it still exists, since even back that it was seriously dated, what with only supporting VGA in DOS real mode or something. |
| 01:05 | <AryehGregor> | Anyway, it came to mind because the author (a rather eccentric person) said he designed it specifically so that it would be unpronounceable, since pronouncing acronyms is evil. |
| 01:06 | <TabAtkins> | oar-pig-see |
| 01:06 | <AryehGregor> | I think he also specified that adding vowels was cheating. |
| 01:07 | <TabAtkins> | Haters don't get to specify arbitrary rules. |
| 01:07 | <AryehGregor> | I got around it by treating things as fricatives and affricates rather than plosives, particularly the G. So "orp-jseh", with "GC" pronounced as "js" rolled together with no vowel in between. |
| 01:07 | <AryehGregor> | You don't really need to follow any fricative with a vowel, you can roll them together endlessly. |
| 01:07 | <AryehGregor> | It's only the conventions of the English language that permit "strong" but not "strstrstrstrstrstrstrong". |
| 01:07 | <TabAtkins> | Also: adding schwas doesn't count as adding a vowel, since those often get created just by an accent. |
| 01:07 | <AryehGregor> | (we should be thankful for such conventions) |
| 01:08 | <AryehGregor> | Nonsense, schwas are full-fledged vowels in every sense. Many dialects add or drop vowels in various words, that doesn't make them not vowels. |
| 01:08 | <TabAtkins> | Oh, certainly. But it *does* make it legal to add them wherever you want to make a word pronouncable. |
| 01:08 | <AryehGregor> | For example, most people pronounce "police" as "plees", but rednecks pronounce it as "poh-lees", like it's spelled. |
| 01:09 | <AryehGregor> | http://hamsterrepublic.com/ohrrpgce/index.php/Why_did_you_pick_an_acronym_that_is_so_hard_to_pronounce_.html |
| 01:09 | <TabAtkins> | By "rednecks" you mean "southerners in general". There's a difference in accent. |
| 01:09 | <AryehGregor> | Yes, yes, I don't draw fine distinctions. |
| 01:09 | <TabAtkins> | Of course you don't, yankee. |
| 01:10 | <AryehGregor> | I've never been anywhere in the United States outside the East and West Coasts and probably never will be. |
| 01:10 | <Philip`> | I thought in America it was always pronounced "cops" |
| 01:10 | <TabAtkins> | "po-po". |
| 01:10 | <AryehGregor> | Man, now I'm all nostalgic about OHRRPGCE. |
| 01:10 | <AryehGregor> | I'll have to see if Wandering Hamster has progressed any. |
| 01:10 | <AryehGregor> | It should run fine on DOSBox, I'd think. |
| 01:11 | <TabAtkins> | Bringing things back around, I'm extremely amused that my neologisms stick in your eye. |
| 01:12 | <AryehGregor> | That's a bit of an overstatement. |
| 01:12 | <AryehGregor> | Oh, cool, they have .deb's. |
| 01:12 | <AryehGregor> | Actually, maybe it's even available in my package manager. |
| 01:13 | <AryehGregor> | It seems not. |
| 01:14 | <AryehGregor> | This is a grievous bug in Debian that must be rectified. |
| 01:15 | <Philip`> | http://packages.debian.org/search?keywords=dosbox ? |
| 01:16 | <Philip`> | They seem to have some dosbox there |
| 01:16 | <AryehGregor> | I was talking about Wandering Hamster. |
| 01:16 | <AryehGregor> | Which was apparently ported to Linux natively long ago, after being open-sourced. |
| 01:16 | <Philip`> | Ah |
| 01:17 | AryehGregor | reads the old DOS 5, DOS 6.22 system requirements: "Plotscripts and battles run a tad slow on a Pentium 166MHz" |
| 01:19 | <Philip`> | That seems a bit heavy |
| 01:19 | Philip` | doesn't remember having anything faster than 66MHz when still running DOS |
| 01:19 | <variable> | AryehGregor, It reminds me of the map of america with only the states NY, Washington DC, and California given names. The rest are labeled "flyover states" ;) |
| 03:01 | <nessy> | does anyone know why Hixie's original email on timed tracks is not archived in the mail archives, see the thread at http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2010-July/027275.html ? |
| 03:06 | <MikeSmith> | nessy: maybe because Jonas changed the subject line and/or started a new thread? |
| 03:07 | <nessy> | no, he didn't - in my gmail they are all in the same thread |
| 03:08 | <variable> | nessy, gmail is very smart about thread. (it uses the reply-to header) |
| 03:09 | <nessy> | Still, it should at least be in the listing at http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2010-July/027275.html and it isn't |
| 03:09 | <nessy> | I cannot find it anywhere in the archives - pretty difficult to point people to without a record :( |
| 03:09 | <MikeSmith> | wait, which message exactly? |
| 03:10 | <MikeSmith> | I don't find any "Timed tracks for <video>" message that Hixie initiated here either: |
| 03:10 | <MikeSmith> | http://www.mail-archive.com/whatwg⊙lwo/thrd4.html#22396 |
| 03:10 | <MikeSmith> | the thread seems to have been initiated by Carlos Andrés Solís |
| 03:10 | <nessy> | yup, not there either |
| 03:11 | <MikeSmith> | "My proposal of a subtitle format via an XML-like markup (in progress)" |
| 03:11 | <nessy> | it was pretty long - maybe there is a limit on archiving for mail length? |
| 03:11 | <MikeSmith> | I don't think se |
| 03:11 | <MikeSmith> | *so |
| 03:11 | <MikeSmith> | it would be really odd for a message to be missing from two independent archives |
| 03:11 | <nessy> | the first line of that email is "I recently added to the HTML spec a mechanism by which external subtitles |
| 03:11 | <nessy> | and captions can be added to videos in HTML." |
| 03:12 | <MikeSmith> | hmm, yeah |
| 03:12 | <MikeSmith> | weird |
| 03:14 | <nessy> | it was a mime multipart message, but that still shouldn't stop it from being archived |
| 03:14 | <MikeSmith> | maybe you could send a copy of it to www-archive and use that as the reference |
| 03:16 | <MikeSmith> | wait, I don't think I got that message |
| 03:17 | <MikeSmith> | I think the reason it's nat archived is that it never actually made it to the list |
| 03:18 | <MikeSmith> | must be that the reason you got a copy is that Hixie Cc'ed you |
| 03:18 | <MikeSmith> | or Bcc'ed you, maybe |
| 03:20 | <MikeSmith> | AryehGregor: you still around? |
| 03:21 | <MikeSmith> | AryehGregor: wanted to ask if you had worked around the WBS length-limit problem |
| 03:22 | <nessy> | oh dear! |
| 03:22 | <nessy> | maybe it was too long and is still stuck in a moderation queue? |
| 03:23 | <nessy> | is there a moderator around? |
| 03:24 | <nessy> | MikeSmith: it's strange though since a whole lot of people received it and replied |
| 03:25 | <nessy> | looking at original headers, I am not on the cc or bcc |
| 03:27 | <nessy> | ok, I've sent it to www-archive |
| 03:57 | <MikeSmith> | nimbupani: about the "If you were hacking since age 8, it means you were privileged." post, I think it's not necessarily true |
| 03:58 | <MikeSmith> | for the reason that at least one of the comments alludes to |
| 03:58 | <nimbupani> | MikeSmith: which one? |
| 03:58 | <MikeSmith> | which is, a lot of people got started with programming on computers at school, and may not have even had a computer at home at all, or not until much later |
| 03:59 | <nimbupani> | MikeSmith: ah like that. |
| 03:59 | <nimbupani> | MikeSmith: Yes, granted that. But I think there is an advantage to having stable environment around you. |
| 03:59 | <MikeSmith> | sure, true |
| 04:00 | <nimbupani> | I did not have a computer and learnt programming at school on black and white machines. I got my first email ID only in 2000 :/ |
| 04:02 | <nimbupani> | and I think it is funny that countries where there are more gender discrimination in "real life" has such low discrimination specific to technology. |
| 04:02 | <nimbupani> | like, it is a "career" choice not a "way of life". |
| 04:02 | <MikeSmith> | interesting |
| 04:03 | <nimbupani> | there are definitely lot more female programmers in India/China |
| 04:03 | <nimbupani> | than in US. |
| 04:03 | <MikeSmith> | yeah |
| 04:04 | <MikeSmith> | but Japan has a lot of gender discrimination in general, and it seems like it also applies to technology |
| 04:04 | <nimbupani> | ah k. |
| 04:05 | <MikeSmith> | I think not having a computer and learning programming at school on black-and-white machines is more the rule than the exception |
| 04:05 | <MikeSmith> | or at least it used to be |
| 04:05 | <nimbupani> | yeah. that is true. but I was thinking of the larger impact of having parents who encourage you to learn and provide a stable environment. |
| 04:05 | <MikeSmith> | ah yeah |
| 04:06 | <nimbupani> | and sadly that is still a privilege. |
| 04:07 | <nimbupani> | what is interesting that has happened because of this outsourcing in India |
| 04:07 | <nimbupani> | is that a lot of not-really-middle-class Indians have gained knowledge |
| 04:08 | <nimbupani> | and expertise to become the new upcoming middle-class. |
| 04:08 | <nimbupani> | so it is seen as a way out. |
| 04:09 | <nimbupani> | so parents actively encourage kids to learn (sadly mainly boys). |
| 04:09 | <MikeSmith> | I think that's really similar to some other places as well |
| 04:09 | <MikeSmith> | Vietnam, for one |
| 04:17 | <MikeSmith> | though maybe the class boundaries aren't quite as rigid in Vietnam |
| 04:18 | <nimbupani> | how does that work? is it because there are very few people in the so-called "middle class" |
| 04:22 | <MikeSmith> | nimbupani: in part because it's nominally still a communist country |
| 04:23 | <nimbupani> | oh ah I see what you mean. |
| 07:00 | <Cheery> | could I somehow establish P2P connection between HTML5 browsers? |
| 07:01 | <Hixie> | not yet, but see http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2010-July/027129.html |
| 07:14 | <annevk> | nessy, http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2010-July/027386.html ? |
| 08:32 | <nessy> | annevk: how did you find that! that definitely wasn't there before - thanks! |
| 08:36 | <phrearch> | morning |
| 08:37 | <phrearch> | whats the max amount of data thats permitted to be pushed through a websocket to a client? |
| 08:37 | <phrearch> | at once that is |
| 08:37 | <annevk> | there's no limit |
| 08:37 | <franksalim> | phrearch, there is no limit in the protocol as currently specified |
| 08:38 | <franksalim> | annevk, beat me by milliseconds :-) |
| 08:39 | <phrearch> | aha thanks, somehow when i try to push something like http://paste.pocoo.org/show/242549/ |
| 08:39 | <annevk> | it's the IRC game ;p |
| 08:39 | <phrearch> | it stops at command-script |
| 08:39 | <phrearch> | only the part before it is sent |
| 08:39 | <annevk> | servers can have limits |
| 08:40 | <phrearch> | aha yea |
| 08:40 | <phrearch> | http://paste.pocoo.org/show/242550/ |
| 08:40 | <annevk> | I mean, everything will have limits, they're just encouraged to be as non-limiting as possible |
| 08:41 | <phrearch> | im using a twisted websocket server. looks like its capped at MAX_LENGTH = 16384 |
| 08:41 | <annevk> | that seems quite short |
| 08:41 | <phrearch> | yea, and its not closing the connection as expected |
| 08:53 | <phrearch> | aha, it was that limit indeed |
| 08:57 | <hsivonen> | wow. http://www.w3.org/Bugs/Public/show_bug.cgi?id=9894#c4 |
| 08:59 | <hsivonen> | also, I guess the resulution of http://www.w3.org/Bugs/Public/show_bug.cgi?id=10084 hints that there isn't going to be a recusal in the Distributed Extensibility case... |
| 09:21 | <annevk> | hsivonen, fun stuff |
| 10:41 | <hsivonen> | hmm. I wonder what I've been thinking when I wrote U+0000 handling for MARKUP_DECLARATION_HYPHEN |
| 10:43 | <hsivonen> | huh? the spec doesn't even have a state with that name! |
| 10:43 | <hsivonen> | I think Hixie has renamed states... |
| 10:44 | <annevk> | didn't you turn certain states into substates? |
| 10:44 | <annevk> | markup declaration requires consuming several characters or something; I thought you just made more states for that |
| 10:45 | <hsivonen> | annevk: good point. that's what's going on |
| 10:46 | <hsivonen> | anyway, my utterly bogus handling of U+0000 troubles me |
| 10:46 | <hsivonen> | it feels bad to find that I've committed such bogosity |
| 10:59 | <MikeSmith> | nessy: fwiw, I did finally get a copy of that "timed tracks for video" message, and the headers show that although it was sent on the 22nd, it wasn't delivered to the list until today |
| 10:59 | <nessy> | I'm pretty sure somebody went to the mailing list and authorised the email, cause it was finally archived today, too |
| 11:00 | <nessy> | in any case, it's now properly referencable and that's what I was after |
| 11:00 | <nessy> | hsivonen: it's good to find out you're human ;) |
| 11:01 | <annevk> | must have been Hixie's doing then |
| 11:02 | <hsivonen> | I think the two bogus U+0000 cases I found are leftovers from the time when I used U+0000 to signal end of buffer |
| 11:03 | <hsivonen> | we really need test cases that put U+0000 and carriage return in every possible tokenizer state |
| 11:04 | <hsivonen> | U+0000 and carriage return are my least favorite characters |
| 11:21 | <annevk> | Hixie, "with on name" seems like a typo |
| 11:42 | <annevk> | why does twitter only have translate button for search results? |
| 11:42 | <annevk> | often I need to translate myakura |
| 12:06 | <jarib> | whatwg.org down? |
| 12:10 | <Philip`> | Looks it |
| 14:07 | <mpilgrim> | if anyone would like to say something nice about "Dive Into HTML5" and see it printed on the back cover of the printed edition, send your blurb to the address listed here: http://diveintohtml5.org/about.html |
| 14:09 | <Workshiva> | I wonder if mrlastweek would contribute something |
| 14:10 | <Philip`> | "This book is okay but why aren't you just reading the free online one?" |
| 14:12 | <jgraham> | "Better than a box full of bobcats" |
| 14:13 | <Workshiva> | "No risk of drowning" |
| 14:47 | <hsivonen> | http://blogs.adobe.com/conversations/2010/07/adobe_expands_enterprise_softw.html |
| 14:51 | <Lachy> | why does the Polyglot Markup draft purport to have normative requirements, rather than being a non-normative note that reflects the requirements in HTML5? |
| 14:51 | <Lachy> | I've found cases where there are conflicting requirements, so that seems problematic |
| 14:54 | <Workshiva> | I'm pretty sure that question has been asked before, but I don't recall how it was un-answered |
| 14:56 | <Lachy> | yeah, it probably has. But I have been too busy to pay much attention to that draft till now. |
| 15:05 | <hsivonen> | Lachy: feel free to write your rationale in http://www.w3.org/Bugs/Public/show_bug.cgi?id=9969 |
| 15:06 | hsivonen | sighs at http://www.w3.org/Bugs/Public/show_bug.cgi?id=10167 |
| 15:13 | <karlcow> | hsivonen: to be honest, I do not understand the requirements for issue 10167 in the prose. |
| 15:14 | karlcow | wishes we had more annotated specs, but resources, etc yada yada. :) |
| 15:15 | <Lachy> | oh, that whole section was wrong in so many ways, as I just noted in my e-mail I just sent. |
| 15:16 | <Lachy> | it lists non-conforming attributes, wrongly claims case-sensitivity of values and states an incorrect requirement for lowercase values |
| 15:18 | <karlcow> | Lachy: thanks, I thought so. |
| 15:58 | <AryehGregor> | MikeSmith, I eventually posted it by just reducing the length, although it's a pretty silly bug. |
| 16:29 | <AryehGregor> | For anyone who hasn't seen this yet on TDWTF: view source on http://www.korea-dpr.com/ and look for "<strong><strong>". |
| 16:29 | <AryehGregor> | Fascinating. |
| 16:29 | <AryehGregor> | Based on validator.nu output, they do seem to be properly nested. |
| 16:32 | <hsivonen> | AryehGregor: I don't see the text "OCN Articles" on the page, but Firefox's "Find" function sees it |
| 16:32 | <AryehGregor> | I see it in Chrome. |
| 16:32 | <hsivonen> | AryehGregor: I see it in Midori, too |
| 16:32 | <AryehGregor> | It's in the bottom right. |
| 16:33 | <hsivonen> | AryehGregor: apparently the deep nesting hits the DoS-by-author-incomptence prevention code in the HTML5 parser in Gecko |
| 16:33 | <hsivonen> | but it's not clear to me why the text isn't visible |
| 16:33 | <AryehGregor> | So did I inadvertently find a site that will break with the HTML5 parser? |
| 16:34 | <hsivonen> | AryehGregor: yes |
| 16:34 | <AryehGregor> | Hurrah, I thought I was just spamming an amusingly stupidly-written site. |
| 16:34 | <AryehGregor> | The DoS part is for real, Web Inspector barely works. |
| 16:34 | <AryehGregor> | Maxes out CPU and becomes unresponsive. |
| 16:36 | AryehGregor | counts around a nesting depth of ~883 |
| 16:40 | <hsivonen> | AryehGregor: filed as https://bugzilla.mozilla.org/show_bug.cgi?id=582620 |
| 17:04 | <AryehGregor> | Mozilla people: Direct2D doesn't appear to work for me when I fiddle with the appropriate about:config settings and restart the browser, on Windows 7 with a GeForce 6200 SE and the latest drivers from NVIDIA. It supports DirectX 9.0c. IE9PP3 plays all the IE demos at good frame rates, but Firefox 4.0b2 doesn't, D2D doesn't seem to be used at all. Should I report a bug? |
| 17:10 | <hsivonen> | AryehGregor: I don't know the details about Gecko's D2D support, but on surface that looks like something that should be filed as a bug |
| 17:32 | <Rik`> | AryehGregor: it seems your card is too old for that |
| 17:32 | <AryehGregor> | Really? I thought it was supposed to work with any card supporting DirectX 9. |
| 17:33 | <AryehGregor> | I didn't see clear requirements listed anywhere. |
| 17:33 | <AryehGregor> | The instructions I got are from some forum post written by I don't know who. |
| 17:33 | <Rik`> | ok, now I'm told "it might work" :) |
| 17:34 | <AryehGregor> | It does work in IE9PP3, so if it doesn't work in Firefox, that should be fixed for parity. |
| 17:34 | <AryehGregor> | But I don't know if it's complete/mature enough to be worth filing bugs yet. |
| 17:34 | <AryehGregor> | (I found a bug in 4.0b1, fullscreen video running super-slowly on nouveau, but that was fixed in b2, so I have hope) |
| 17:38 | <Rik`> | AryehGregor: http://www.basschouten.com/blog1.php/2010/03/02/presenting-direct2d-hardware-acceleratio ? |
| 17:38 | <AryehGregor> | "a high-end DirectX 9 graphics card" |
| 17:38 | <AryehGregor> | 6200 SE is definitely what I'd call "low-end". :) |
| 17:39 | <AryehGregor> | But if IE9PP3 can do it, surely Firefox 4 should be able to, at least by final release. |
| 17:39 | <AryehGregor> | Too bad my only decent graphics cards are in Linux or XP machines. Oh well. |
| 17:51 | <AryehGregor> | TabAtkins, when pasting data URLs into e-mails, you have to make sure they don't line-wrap at inconvenient places. Newlines are eaten completely when you paste into the URL bar, not converted to spaces. (at least for me) |
| 17:52 | <AryehGregor> | So it has to be manually fixed or you get things like <divstyle= instead of <div style=. |
| 17:53 | <Workshiva> | Aren't spaces encoded? |
| 17:54 | <AryehGregor> | No, that would be terrible for readability. |
| 17:55 | <jgraham> | data urls are supposed to be readble? |
| 17:57 | <Workshiva> | They could encode as + |
| 18:32 | <TabAtkin1_> | AryehGregor: Ah, sorry about that. |
| 18:32 | <AryehGregor> | I don't know why newlines are eaten instead of being converted to spaces. |
| 18:33 | <TabAtkin1_> | Workshiva: When I'm writing a test document in my address bar, I don't hand-encode, and neither does the address bar. I often try to encode before posting it into a bug or something, but that requires an explicit "javascript:alert(encodeURIComponent(''));" call |
| 18:34 | <Philip`> | Opera converts them to spaces, I believe |
| 18:35 | <Workshiva> | It does, and it's terrible |
| 18:35 | <Workshiva> | It breaks URLs in any document or email with hard line wrapping |
| 18:35 | <TabAtkin1_> | In general, though, I consider data: url readability not very important, since I can just paste it into the address bar and view-source it if I want. |
| 18:36 | <AryehGregor> | I prefer to be able to read it directly. |
| 20:06 | AryehGregor | wonders if YouTube's fake HTML5 fullscreen gets accelerated in Firefox 4 |
| 20:16 | <jgraham> | /me guesses not if it is fake |
| 20:21 | <AryehGregor> | It sets it to something like position: absolute; top: 0; bottom: 0; left: 0; right: 0. In principle that could be accelerated just as well as fullscreen, if nothing overlays it, no? |
| 20:21 | <AryehGregor> | Particularly if the user hits F11. |
| 20:26 | <jgraham> | I love wikipeda |
| 20:26 | <jgraham> | "Other than the situation of egg donation, the identity of a child's mother is seldom in doubt" |
| 20:26 | <jgraham> | and then it has a citation! |
| 20:26 | <Workshiva> | There are no axioms, only references |
| 20:27 | <jgraham> | Who needs to read up on why the identity of mothers is generally secure? |
| 20:27 | <jgraham> | I am imaining some extremely basement-closeted geeks here |
| 20:28 | <jgraham> | *imagining |
| 20:28 | <jgraham> | Or maybe it is the same people who depend on the reference in HTML5 to ell them what ascii is |
| 20:30 | <jgraham> | The number of people whose assumed father is not their biological father seems to be surprisingly high |
| 20:30 | <Workshiva> | It's only surprising that you find it surprising |
| 20:33 | <jgraham> | Well umm... I can't really disagree with that, can I |
| 20:33 | <jgraham> | But they suggest 1 in 30 |
| 20:33 | <jgraham> | Which is like one kid from your class at school |
| 20:33 | <jgraham> | OTOH it probably varies a lot by socio-economic status |
| 20:33 | <Workshiva> | Probably |
| 20:34 | <jgraham> | e.g. if you are the prince regent of england, it is like 1 in 2 of your children |
| 20:34 | <jgraham> | </scandal> |
| 20:34 | <jgraham> | or probably </libel> |
| 20:34 | <Workshiva> | Yeah, nice knowing you, hope you enjoy bankrupcy |
| 20:37 | jgraham | should note that that is almost certianly not true :) |
| 20:37 | <AryehGregor> | The guy who writes overcomingbias.com argues that we should have routine paternity testing so fathers aren't forced to support kids that aren't theirs. |
| 20:38 | <jgraham> | I have a feeling that too much information can be bad for you |
| 20:38 | <AryehGregor> | Probably. |
| 20:39 | <Workshiva> | I'm not sure I agree with the problem |
| 20:39 | <Workshiva> | If you aren't asking for a test yourself, that means you were probably expecting a child anyway |
| 20:39 | <AryehGregor> | What do you mean? |
| 20:40 | <jgraham> | I guess it is more of an issue in cases where the parents are seperated and there is a dispute |
| 20:40 | <Workshiva> | Yeah |
| 20:42 | <AryehGregor> | You can seriously get a 500 GB drive for $40 now? Geez. |
| 20:42 | <AryehGregor> | It's Hitachi, though. Not touching it, my last Hitachi died on me out of nowhere. |
| 20:43 | <Workshiva> | I wouldn't buy a 500 GB anyway, not efficient use of the drive wells |
| 20:45 | <AryehGregor> | What would you buy? |
| 20:45 | <AryehGregor> | 1 TB? |
| 20:47 | <AryehGregor> | WTF, you can get 1 TB for $70. |
| 20:47 | <AryehGregor> | Geez. |
| 20:51 | <jgraham> | AryehGregor: I see you buy hardware about as often as I do |
| 20:51 | <jgraham> | i.e. almost never |
| 20:52 | <AryehGregor> | Yeah, why should I bother buying hardware often? |
| 20:52 | <AryehGregor> | I'm just annoyed at my server backups and things getting close to 500 GB. |
| 20:52 | <jgraham> | I don't know |
| 20:52 | <jgraham> | some people seem to enjoy it |
| 20:52 | <AryehGregor> | Don't want to worry about it, so may as well get a couple of 1 TB drives. |
| 20:53 | <AryehGregor> | I can just use the first 500 GB for now, to get 1 TB in RAID 10, which will be more space and also faster. Then if I want, I can replace the 500 GB with 1 TB later to get 2 TB total. |
| 20:53 | <Workshiva> | AryehGregor: Yeah, 1 TB is the sweet spot for price now, as far as I know |
| 20:53 | <AryehGregor> | Last time I bought, it was closer to 500 GB. |
| 20:53 | Rik` | just bought a 120Gb SSD for 350$ |
| 20:54 | <AryehGregor> | My drive has been powered on for 6241 hours, which means I probably got them a year ago or somewhat less. |
| 20:54 | <AryehGregor> | Rik`, I got an X25-M. It was great for my gaming machine, but basically useless for web browsing. |
| 20:54 | <AryehGregor> | (is great, I should say) |
| 20:54 | <AryehGregor> | You're just not I/O-bound often enough with normal desktop use. |
| 20:54 | <AryehGregor> | Also good for booting, but I almost never reboot my desktop. |
| 20:54 | <Rik`> | well, opening a lot of browsers, closing them |
| 20:55 | <AryehGregor> | This is why we have page caches. |
| 20:55 | <AryehGregor> | It's usually just a few times a day that you start programs. |
| 20:55 | <Workshiva> | For web browsers, shouldn't a ram-backed fs work even better? |
| 20:55 | <AryehGregor> | So not worth the money yet IMO. |
| 20:55 | <AryehGregor> | But it really helps out for games with level load times. |
| 20:55 | <AryehGregor> | Workshiva, yes, except for the part where you crash. |
| 20:56 | <Rik`> | yeah, I should also upgrade my RAM, but that's for another time |
| 20:56 | <Workshiva> | Why would you crash? |
| 20:58 | <AryehGregor> | Workshiva, *I* wouldn't, but my system has been known to do it with nonzero frequency. |
| 20:59 | <Workshiva> | In that case, keep only the cache stuff in memory :) |
| 21:00 | <AryehGregor> | I am reasonably certain that browsers do, in fact, attempt to do that. |
| 21:00 | <Workshiva> | Yeah, but they also have disk caches |
| 21:00 | <Workshiva> | Not all of them let you disable disk cache, I believe |
| 21:00 | <AryehGregor> | Which are on disk because they're a) meant to persist across reboot and/or b) too large to easily fit in memory. |
| 21:01 | <Workshiva> | b) is not so important anymore with 64-bit systems and ample ram |
| 21:01 | <AryehGregor> | But even more ample disk space. On a typical computer, browsers should really be using at least 10 GB of disk space for cache. |
| 21:01 | <AryehGregor> | Currently they're way too stingy. |
| 21:02 | <Workshiva> | Even if you don't cache video? |
| 21:02 | <AryehGregor> | http://www.stevesouders.com/blog/2010/04/26/call-to-improve-browser-caching/ |
| 21:02 | <AryehGregor> | "55% of people surveyed have a cache that’s over 90% full." |
| 21:03 | <Workshiva> | Isn't that natural if you don't ever delete from the cache unless it's full? |
| 21:04 | <AryehGregor> | Okay, that's true. Better data: "40-60% of unique users hit the site with an empty cache at least once per day." |
| 21:05 | <AryehGregor> | There's no conceivable reason for this except over-aggressive cache clearing. |
| 21:06 | <Workshiva> | Yeah, but unless you stop caching video video will just eat up every new MB, like said in the comments |
| 21:06 | <AryehGregor> | Does anyone actually put video in the same cache as everything else? |
| 21:07 | <AryehGregor> | That sounds like a crazy bad idea. |
| 21:07 | <AryehGregor> | AFAIK, at least Firefox has a separate media cache. |
| 21:07 | <AryehGregor> | (which I don't think persists across browser restart) |
| 21:07 | <AryehGregor> | (since usually you don't watch the same video every day) |
| 21:09 | <Rik`> | only < 100MB of cache for the piece of software you use most feels really small |
| 21:11 | <Philip`> | Do the caches have O(1) access costs? |
| 21:11 | <Philip`> | i.e. does caching some data you'll never use have any detrimental effect except on disk space? |
| 21:11 | <AryehGregor> | I'm guessing O(log(N)) at worst. |
| 21:12 | <AryehGregor> | Which is probably negligible at the sizes we're talking about. |
| 21:12 | <AryehGregor> | Compared to the cost of extra cache misses. |
| 21:13 | <Hixie> | good god |
| 21:13 | <Hixie> | now not only is julian arguing about the ascii reference |
| 21:13 | <Hixie> | but he even thinks that the reference he is arguing about is one NO ONE WILL FOLLOW |
| 21:13 | <Rik`> | the other part of the post about improving eviction algorithms is also important |
| 21:13 | <Hixie> | this is like the epitome of pointlessness |
| 21:13 | Philip` | tried to follow the Bezier reference a few times but found it impossible :-( |
| 21:14 | <Rik`> | like fonts, javascripts and CSS are much more important to rendering than images |
| 21:14 | <Hixie> | Philip`: yeah i wish we could get better references for some of these things |
| 21:14 | <Philip`> | It'd be more useful to reference the Wikipedia entry |
| 21:15 | <AryehGregor> | Philip`, I thought that too. |
| 21:16 | <AryehGregor> | We don't *need* a normative reference for ASCII. |
| 21:16 | <jgraham> | It turns out that people get upset if you reference things that are useful but not offical-looking |
| 21:16 | <Hixie> | Philip`: yeah, i've considered doing that |
| 21:17 | <jgraham> | Wikipedia is practically the definition of not-official-looking |
| 21:17 | <jgraham> | In fact I should edit it to say that... |
| 21:21 | <jgraham> | Hmm, someone is sending email to the WHATWG with a trademark sign in their name |
| 21:21 | <jgraham> | Didn't MikeSmith patent that |
| 21:21 | <jgraham> | ("A method for adding distinguishing characters to a name such that it would be recogniased as unique") |
| 21:22 | <jgraham> | wtf |
| 21:22 | <jgraham> | How badly did I mangle "recognised" |
| 21:28 | <Philip`> | jgraham: The Levenshtein distance is 1, so it doesn't seem badly mangled at all |
| 21:29 | <jgraham> | Philip`: I was actually timing how long it would take someone to say that |
| 21:30 | <jgraham> | annevk: Hey, what's "Hear the Wind Sing"? and more to the point why haven't I got it |
| 21:30 | Philip` | had to look up the term on Wikipedia :-( |
| 21:32 | jgraham | wonders where one finds can find a copy |
| 21:32 | <jgraham> | Hmm, that was a whole extra word |
| 21:36 | <annevk> | http://www.schneier.com/blog/archives/2010/07/dnssec_root_key.html weird stuff |
| 21:36 | <annevk> | jgraham, I bought one via a store that sells via amazon.com |
| 21:37 | <annevk> | jgraham, that and Pinball, 1973 are his first two books of which the English editions are only published via some Japanese library? publisher |
| 21:37 | <annevk> | jgraham, together with the wild sheep chase they form the trilogy of the Rat |
| 21:39 | <jgraham> | annevk: Ah. Amazon. Taht huld have been obvious, really |
| 21:39 | <jgraham> | *That should |
| 21:40 | jgraham | would like to blame the silly lag for his even-worse than usual typing |
| 21:40 | <annevk> | I really quite like that first novel |
| 21:40 | <annevk> | he goes into this fictional writer a whole lot more and overall it reads pretty good |
| 21:43 | <jgraham> | It's unsurprisingly rather expensive |
| 21:51 | <[Tyrant]> | I'm having a bit of trouble grasping some of the semantics of html5 |
| 21:52 | <TabAtkins> | shoot |
| 21:52 | <[Tyrant]> | I keep looking at <article>'s needed to be put in a list |
| 21:52 | <[Tyrant]> | im so used to lists |
| 21:53 | <TabAtkins> | And? |
| 21:53 | <[Tyrant]> | so would Twitter's posts be a list of <article> |
| 21:53 | <[Tyrant]> | and the feed is a <section> |
| 21:54 | <[Tyrant]> | <section><ol><li><article> |
| 21:54 | <TabAtkins> | The individual posts would be <article>s, yeah (they're independent, appropriate to be linked directly to, etc.). It would be fine to put them in a list. |
| 21:54 | <TabAtkins> | And then, sure, the entire segment that shows the feed is a section of the page, so <section> is fine there. |
| 21:55 | <[Tyrant]> | Ah, well thanks! Feel a bit better now. :) |
| 21:57 | <Hixie> | in case people are wondering why whatwg.org is slow, I present a sample line from the logs: GET /specs/web-apps/current-work/#dom-textarea/#text/#flow-content |
| 21:57 | <TabAtkins> | Wut? |
| 21:59 | <Hixie> | runaway buggy bot |
| 22:00 | <hsivonen> | Hixie: can it be banned by IP? |
| 22:01 | <annevk> | TabAtkins, furthermore, unlike what WebKit tolerates, "red" and such are not even valid values for <input type=color> |
| 22:02 | <TabAtkins> | I thought so! But I wasn't quite certain enough to say so. |
| 22:02 | <TabAtkins> | I'm pretty sure we just hand the value to our color parser. |
| 22:02 | <Hixie> | hsivonen: already done |
| 22:02 | <Hixie> | i wonder why this doesn't work?: |
| 22:02 | <Hixie> | <FilesMatch "#"> Order Deny,Allow Deny from All |
| 22:02 | <Hixie> | </FilesMatch> |
| 22:02 | <AryehGregor> | annevk, they can't validly be submitted to the server, but the user can enter them. |
| 22:03 | <Hixie> | er, with different line breaking |
| 22:03 | <AryehGregor> | Does WebKit actually send them? |
| 22:04 | <annevk> | AryehGregor, it matches :valid |
| 22:04 | <annevk> | AryehGregor, but then :valid and :invalid don't really make sense for that control if you implement proper UI |
| 22:05 | <AryehGregor> | No, they don't. |
| 22:05 | <AryehGregor> | Hmm, is WebM also an audio format? |
| 22:06 | <Philip`> | Are the PNGs in http://www.w3.org/TR/SVG/painting.html#FillProperties broken, or is it just me? |
| 22:06 | <TabAtkins> | Broken. |
| 22:06 | <TabAtkins> | I looked at that 2 weeks ago... |
| 22:08 | <annevk> | AryehGregor, container format, yes |
| 22:18 | <jgraham> | Hixie: Don't become a UI designer :) |
| 22:21 | <Hixie> | i have no plans to :-) |
| 22:22 | <Hixie> | (what don't you like about my proposal?) |
| 22:23 | <jgraham> | The dropdown... |
| 22:25 | <Hixie> | what would be better? |
| 22:25 | <jgraham> | Not sure. That's why I'm not a UI designer either :) |
| 22:25 | <Hixie> | or what's wrong with it? |
| 22:25 | <Hixie> | i've seen apple use that kind of UI, e.g. in Keynote in their Inspector panels |
| 22:25 | <Hixie> | but I'm no expert |
| 22:26 | <jgraham> | It adds a lot of complexity to something that is conceputally simple. It makes an existing UI harder than it already is |
| 22:26 | <jgraham> | Well I'm not either, so I could be totally wrong |
| 22:26 | <Hixie> | it's hard not to add complexity when adding functionality |
| 22:26 | <AryehGregor> | What proposal is this? |
| 22:26 | <jgraham> | Yes, I think that's what they pay the big bucks for |
| 22:27 | <jgraham> | UI designers good enough to add functionality but reduce complexity |
| 22:27 | <Philip`> | Showing the same list of many options to the user on all sites using <input type=file> seems to ignore that some options are going to be used far more frequently than others on some sites, and that the browser has no way to know that |
| 22:27 | <Philip`> | so it can't optimise the UI for the particular use when there's no way to distinguish it from other uses |
| 22:28 | <AryehGregor> | Philip`, easy, just have the browser upload statistics to a central server and apply a heuristic based on past users of that page. |
| 22:28 | <Hixie> | that would be awesome |
| 22:28 | <Philip`> | AryehGregor: That's a privacy violation |
| 22:28 | <AryehGregor> | You could anonymize the data if you want. |
| 22:28 | <jgraham> | The word you're looking for is awful :) |
| 22:28 | <Hixie> | why? |
| 22:29 | <AryehGregor> | You could also claim to anonymize it but actually define "anonymize" to mean something counterintuitive and figure it's ethical because no one who cares believed you in the first place. |
| 22:29 | <Philip`> | It'd be awful for people writing pages, because they'll write it and test it and see it works okay, but then tomorrow it suddenly starts acting differently |
| 22:29 | <jgraham> | Rearranging the UI based on use frequency has a certian Office 97 (or whichever version it was) feel to it |
| 22:29 | <AryehGregor> | That wasn't a serious suggestion, by the way. It was a joke. |
| 22:29 | <jgraham> | AryehGregor: That was what I supposed |
| 22:30 | <jgraham> | Then I paniced because I wasn't sure if other people were taking you seriously... |
| 22:31 | <jgraham> | Anyway, there is something to be said for always having the controls in the same place |
| 22:31 | <Philip`> | We could fix all web compatibility problems by just letting browser users edit the markup of the site they're visiting and submit it to a central server so the same fix is applied to all subsequent users |
| 22:32 | <jgraham> | Philip`: That sounds like site-patching |
| 22:32 | <TabAtkins> | wiki-wide web? |
| 22:32 | <jgraham> | Except with user-submitted site patches |
| 22:32 | <jgraham> | Rather than hallvord doing the whole web by himself |
| 22:32 | <jgraham> | (not actually true) |
| 22:32 | <TabAtkins> | s/<body>(.*)/COCKSCOCKSCOCKS/ |
| 22:32 | <AryehGregor> | Philip`, I wonder if anyone here has familiarity with any software that could be used for that purpose. |
| 22:33 | <jgraham> | I think TabAtkins identified a major flaw :) |
| 22:33 | <jgraham> | You really don't want to push out site patches without review |
| 22:33 | <AryehGregor> | Are you discussing this seriously? |
| 22:34 | <jgraham> | Well it actually isn't that far from what Opera already do |
| 22:34 | <Philip`> | AryehGregor: If you anonymise the data for the file inputs, how do you prevent abuses, like some Flickr competitor sending a zillion fake messages to the server to tell it to favour totally useless upload options and harm the user experience? |
| 22:34 | <jgraham> | Except without the "members of the public" bit |
| 22:35 | <AryehGregor> | Philip`, by having the sites pay you an annual fee to change their database entry to match the appearance they want. |
| 22:35 | <Philip`> | Maybe you'd have to give authors a way to access the central server and override the default heuristic behaviour, so they can fix problems on their own site |
| 22:35 | <AryehGregor> | You didn't think of the "paying" part. |
| 22:35 | <Philip`> | but you'd need a way to verify the authors really own that pae |
| 22:36 | <AryehGregor> | Well, that's easy. Whoever pays you the most is presumptively the legitimate owner. |
| 22:36 | <Philip`> | so you could require the authors to put some special magic code in their markup to identify themselves |
| 22:36 | <TabAtkins> | AryehGregor: You're right, by the way. I just invented and used "lexenv" in conversation. |
| 22:36 | <Philip`> | Perhaps they could add a marker like 'directory' to their <input> markups to do that |
| 22:36 | <AryehGregor> | TabAtkins, of course I'm right. |
| 22:37 | <jgraham> | TabAtkins: Do you mostly shorten words to increase the ratio of "x" to other letters in your speech? |
| 22:37 | <TabAtkins> | jgraham: That's a secondary goal, yes. |
| 22:38 | <TabAtkins> | I hope to throw off frequency analysis in my naive cryptographic experiments. |
| 22:38 | <hsivonen> | crowdsourcing some hallvors activities for Gecko-based browsers has been seriously suggested |
| 22:38 | <hsivonen> | and you guys seem to be joking about it... |
| 22:39 | <TabAtkins> | hsivonen: Only because a large part of the submissions will be of the COCKSCOCKSCOCKS variety. |
| 22:39 | <jgraham> | hsivonen: That seems... scary |
| 22:39 | <hsivonen> | jgraham: indeed |
| 22:40 | <hsivonen> | (FWIW, I think we should avoid site patching, since Gecko-based browsers seems to be doing quite well without it) |
| 22:40 | <jgraham> | I would imagine that the resources you would need to review the submissions would be close to the resources that you would need to just write them yourself |
| 22:40 | <hsivonen> | jgraham: yeah |
| 22:40 | <AryehGregor> | WebKit had to do a site patch for MediaWiki. |
| 22:40 | <jgraham> | (I would prefer that Opera could avoid site-patching too, but we have much less clout with authors) |
| 22:40 | <AryehGregor> | More of a software-specific patch. |
| 22:41 | <AryehGregor> | Wasn't an issue of not having clout with authors there, more like not being able to get all sixty bajillion installations updated. |
| 22:41 | <Philip`> | $ GET http://www.w3.org/TR/SVG/images/painting/fillrule-nonzero.png|hexdump -C|head -1 |
| 22:41 | <Philip`> | 00000000 89 50 4e 47 0d 0d 0a 1a 0d 0a 00 00 00 0d 49 48 |.PNG..........IH| |
| 22:41 | <Philip`> | It's nice of PNG to include some line-ending characters in its header |
| 22:41 | <Philip`> | to detect exactly the error that seems to have occurred here |
| 22:42 | <jgraham> | AryehGregor: Yoah, distributed installations are a problem, as the IE team found out with jQuery |
| 22:43 | <AryehGregor> | Philip`, haha, ick. |
| 22:43 | <AryehGregor> | (kind of odd juxtaposition of interjections) |
| 22:43 | <Philip`> | Would be much better if it automatically error-corrected in that case |
| 22:43 | <Philip`> | since at least the lf->crlf transformation is reversible |
| 22:43 | <hsivonen> | AryehGregor: what was the patch tha WebKit applied to mediawiki? |
| 22:44 | AryehGregor | searches |
| 22:44 | <AryehGregor> | Why is it that it takes orders of magnitude less time for Google to search the entire web than it takes for them to search my mailbox? |
| 22:45 | <TabAtkins> | More resources thrown at the problem. |
| 22:45 | <TabAtkins> | *Vastly* more resources. |
| 22:45 | <jgraham> | Mwahhahhaha? |
| 22:45 | <AryehGregor> | And these can't be, you know, reused to solve similar problems? |
| 22:45 | <TabAtkins> | That was implied. |
| 22:45 | <TabAtkins> | AryehGregor: Simply put, no. |
| 22:46 | <AryehGregor> | hsivonen, https://bugs.webkit.org/show_bug.cgi?id=28350 |
| 22:46 | <hsivonen> | AryehGregor: thanks |
| 22:46 | <AryehGregor> | TabAtkins, now you're being all superior and mysterious just because you work at Google. |
| 22:46 | <hsivonen> | AryehGregor: you just have to put your email into gmail |
| 22:46 | <jgraham> | AryehGregor: Presumably that would only work if you had spare resources in the large pool |
| 22:46 | <AryehGregor> | hsivonen, see also https://bugs.webkit.org/show_bug.cgi?id=29792 |
| 22:47 | <TabAtkins> | No, because I (a) don't know much more than what I'm saying, and (b) can't talk about it anyway. |
| 22:47 | <AryehGregor> | TabAtkins, yes, as I said. |
| 22:47 | <jgraham> | In which case why not just reallocate them to the smaller pool |
| 22:47 | <jgraham> | (unless you are suggeting the pools could be merged, but it's not clear that the problem domains are really sufficiently similar) |
| 22:48 | <AryehGregor> | It seems somewhat pathetic anyway. I mean, it often takes ten or twenty seconds to do a search. Fulltext search on a few gigs of data should be faster than that on bottom-end commodity hardware with nothing in memory. |
| 22:49 | <AryehGregor> | jgraham, I'm pretty sure Google has a giant pool of servers that they reallocate on the fly to whatever's convenient. There are papers published about it. |
| 22:49 | <jgraham> | But "resources" don't have to be physical servers |
| 22:49 | <TabAtkins> | A large percentage of those have been allocated to running the moon base now, though. |
| 22:50 | <jgraham> | It can be the fraction of the time allocated to one task vs the other |
| 22:50 | <AryehGregor> | I'm certainly pretty sure that it shouldn't have to take fifteen seconds to archive the current e-mail and go to the next one. |
| 22:50 | <AryehGregor> | Which sometimes it does anyway. |
| 22:50 | <AryehGregor> | (Although not that often anymore, thankfully.) |
| 22:51 | <jgraham> | Yeah, I agree that sounds slow |
| 22:51 | <Philip`> | AryehGregor: There's no competitive need for Gmail to be extremely fast at searching email, since you put up with its current speed, whereas nobody would accept web searches being so slow |
| 22:52 | <AryehGregor> | Philip`, there's no competitive need for Gmail to do much of anything, because it's clearly superior to all other web mail services. |
| 22:52 | <AryehGregor> | Whereas search does actually compete to some degree with Bing and Yahoo!, if only because those are the default search engines in a nontrivial number of cases. |
| 22:52 | <jgraham> | If you can make money off an email service it seems like a good market to be in because the switching costs are unusually high |
| 22:53 | <jgraham> | Unlike search where they are basically zero |
| 22:53 | <AryehGregor> | Not really. Gmail supports forwarding to other addresses indefinitely. |
| 22:53 | <Philip`> | Also your mailbox is searched extremely rarely, so the data has extremely low utilisation, and it would be wasteful to use high-speed data stores when most of the time it's not going to be used at all |
| 22:53 | <Philip`> | whereas the web is searched quite often |
| 22:54 | <AryehGregor> | That's true. |
| 22:54 | <jgraham> | AryehGregor: I bet a lot of gmail users don't know that and wouldn't be able to set it up if they did |
| 22:54 | <AryehGregor> | Working set for web searches is probably in memory practically all the time, working set for Gmail searches must typically be on disk . . . |
| 22:55 | <jgraham> | There are still people who use @operamail.com addresses even though it offers 3MB of storage |
| 22:55 | <Philip`> | You can afford to make things much faster when the cost is amortised across a billion users than when it's for one user |
| 22:55 | AryehGregor | is using 4248 MB of e-mail storage :/ |
| 22:56 | <jgraham> | Or 25MB if you pay money |
| 22:56 | <AryehGregor> | lol. |
| 22:56 | <jgraham> | (but I don't think we are accepting money any more) |
| 22:57 | <jgraham> | Oh my |
| 22:57 | <jgraham> | If you try to register an account the captcha is an animated gif |
| 22:57 | <jgraham> | "Please enter the letters moving up and down in the centre of the image " |
| 22:57 | <jgraham> | I have never seen that before |
| 22:57 | <Philip`> | For good reason |
| 22:58 | <jgraham> | Indeed |
| 22:59 | <jgraham> | "If you can't see the picture, please click here" - reloads the page with a similar Captcha but in Comic Sans |
| 22:59 | <TabAtkins> | hahahahaha |
| 22:59 | <AryehGregor> | Shouldn't you at least use an obscure font? |
| 22:59 | <AryehGregor> | When I set up captchas for my site, I grabbed some moderately-legible fonts off some free font site. |
| 22:59 | <jgraham> | (actually I don't think it is Comic Sans, but it looks quite like it) |
| 23:00 | <jgraham> | Anyway, I take it there were no a11y people involved with that design... |
| 23:02 | <jgraham> | (the page also has a "Buy Opera" link on it, which helps date it somewhat) |
| 23:03 | <jgraham> | Oh wow |
| 23:04 | <jgraham> | The whole page is document.write(unescape(long_escaped_string)) |
| 23:05 | <AryehGregor> | Impressive. |
| 23:06 | <AryehGregor> | Maybe as an extra deterrent for poorly-written spam bots. |
| 23:07 | <jgraham> | Yeah, I guess that is a logical explaination |
| 23:07 | <jgraham> | I was thinking "one too many Norwegian winters" |
| 23:08 | <AryehGregor> | jgraham, I guess irssi doesn't have built-in spellcheck? |
| 23:08 | <TabAtkins> | Nope. |
| 23:09 | <jgraham> | AryehGregor: How can you tell? :) |
| 23:09 | <AryehGregor> | I dislike XChat overall, but spellcheck is convenient. |
| 23:09 | <jgraham> | Yeah, I would like spellcheck |
| 23:10 | <jgraham> | But irissi is awesome just because I can use it over ssh |
| 23:10 | <AryehGregor> | Yeah, that would be nice to have. |
| 23:10 | <jgraham> | Seems to be a plugin for it |
| 23:11 | <hober> | This is when I'm supposed to put in a plug for ERC |
| 23:11 | <jgraham> | Although the UI for the spllcheck seems dreadful |
| 23:12 | <hober> | (which gets things like spellcheck for free. thanks, Emacs!) |
| 23:15 | <AryehGregor> | UI in IRC clients is generally horrible. |
| 23:15 | <KaOSoFt> | Hey, mIRC works just fine. |
| 23:16 | <KaOSoFt> | I mean, it's not like you need smilies and Live Messenger stuff. |
| 23:16 | <AryehGregor> | Like everything gets flooded with join/part messages, and if you want to turn them off, you turn them off completely. |
| 23:16 | <AryehGregor> | I said UI, not features. |
| 23:16 | <AryehGregor> | Features are also lousy, but that's a whole separate issue. |
| 23:24 | <AryehGregor> | Is it just me, or does http://dev.w3.org/2006/webapi/WebTiming/ fail to help the use-case of cross-browser performance comparisons by not being precise enough? It doesn't say exactly what constitutes when a "navigation begins", for example, so UAs might adopt significantly different approaches. |
| 23:25 | <AryehGregor> | Like one choosing "as soon as the user clicks a link/hits 'Go'/etc.", and another saying "as soon as the HTTP request is on the wire". |
| 23:26 | <AryehGregor> | I guess the latter would really be "fetch", not "navigate". |
| 23:26 | AryehGregor | is still dubious |
| 23:26 | AryehGregor | could send feedback, but already has enough on his plate . . . |
| 23:28 | <AryehGregor> | I guess it's precise enough. |
| 23:28 | <dandaman> | if i have an inline style of background: and have that tag also have a class with a style of background |
| 23:28 | <dandaman> | the inline style will overtake the class background |
| 23:28 | <dandaman> | right? |
| 23:28 | <AryehGregor> | Unless the class rule has !important specified, yes. |
| 23:28 | <dandaman> | but what if i remove the background with javascript |
| 23:29 | <dandaman> | by going tag.style.background=''; |
| 23:29 | <dandaman> | will that make the background the default white |
| 23:29 | <dandaman> | or the class's specified background? |
| 23:29 | <AryehGregor> | Then you're setting the element's style, so it still overrides the rule in the stylesheet, I assume. |
| 23:29 | <AryehGregor> | Although I don't know offhand. Try it and see. |
| 23:29 | <dandaman> | ok |
| 23:30 | <dandaman> | do you know of a way to remove the inline style with javascript |
| 23:30 | <AryehGregor> | Oh, I didn't read far enough. The timing spec seems quite precise. Will be nice to have, I guess. |
| 23:30 | <dandaman> | to make the class's stylesheet take over? |
| 23:30 | <Hixie> | should we just obsolete <address>? people never understand what it's for |
| 23:31 | <AryehGregor> | Hixie, yes. |
| 23:31 | <AryehGregor> | Also <kbd> and <samp>, too obscure. Those can be covered by expanding <code> a bit, or resurrecting <tt> in the fashion of <i> and <b>. |
| 23:31 | <AryehGregor> | I filed a bug about it. |
| 23:31 | <Hixie> | <kbd> and <samp> are fine, no need to remove them |
| 23:32 | <AryehGregor> | Someone is pressuring me to whitelist <address> in MediaWiki at this very moment, as it happens, so this is good timing. |
| 23:32 | AryehGregor | is dubious. |
| 23:32 | <Hixie> | the problem is that there are lots and lots of pages that use <address> correctly |
| 23:32 | <Hixie> | automatically generated man pages and the like |
| 23:33 | <AryehGregor> | Does anyone consume it, though? |
| 23:34 | <hober> | IIRC <address>+hCard is used for "canonical hCards" |
| 23:34 | <hober> | or "representative hCards" |
| 23:34 | hober | looks up the term |
| 23:34 | <Hixie> | AryehGregor: does anyone consume <em>? |
| 23:35 | <AryehGregor> | Well, it's rendered in italics, so in that sense yes. I don't know if anyone distinguishes it from <i>, practically speaking. |
| 23:35 | <AryehGregor> | If no one does, that seems like it calls into question the point of distinguishing them. But probably too late to change that now. |
| 23:36 | <dandaman> | AryehGregor: background=' ' made it go to the class's specified background in case you were wondering |
| 23:36 | <AryehGregor> | Okay, if you say so. |
| 23:36 | <Hixie> | AryehGregor: <address> is also rendered in italics, so in that sense also yes, then. :-) |
| 23:36 | <AryehGregor> | Sounds like a hack. |
| 23:37 | <AryehGregor> | Touché. |
| 23:38 | <AryehGregor> | Meh, I don't really care. These things are cool ideas, but all basically useless in practice as far as I can tell. They're used incorrectly such a large percentage of the time that they can't typically be assigned useful non-presentational meaning anyway. |
| 23:38 | <Hixie> | is <address> used more incorrectly than <blockquote>? |
| 23:38 | <AryehGregor> | Well, it's sure used a lot less in absolute terms. |
| 23:38 | <Hixie> | my concern isn't so much that it's used incorrectly, so much as that people don't use it because they think it's for something else |
| 23:39 | <AryehGregor> | Whether the usage is more often incorrect percentage-wise, I dunno. |
| 23:39 | <AryehGregor> | It should be called <author>. |
| 23:39 | <Hixie> | yeah |
| 23:39 | <AryehGregor> | Then everyone would use it. |
| 23:39 | <Hixie> | indeed |
| 23:39 | <AryehGregor> | I mean, at least the people who want to use trendy semantic markup. |
| 23:39 | <Hixie> | eh, i guess i'll punt this to the future and we can look at it again |
| 23:41 | <AryehGregor> | Shrug. Fine by me. |
