00:05:03 -!- kallisti has quit (Read error: Connection reset by peer). 00:10:15 -!- kallisti has joined. 00:10:15 -!- kallisti has quit (Changing host). 00:10:15 -!- kallisti has joined. 00:21:21 The Mittani has horrible hair! 00:24:59 Maccaroni and cheese tastes better than I thought it would 00:29:42 -!- Phantom_Hoover has quit (Read error: Connection reset by peer). 00:32:01 Maybe you should make it so that if you write "`? `slist" and whatever the `*lists are, that it can tell you what that one is for. 00:38:57 -!- Phantom_Hoover has joined. 00:51:40 -!- nooodl has quit (Ping timeout: 245 seconds). 00:52:01 -!- carado has quit (Ping timeout: 246 seconds). 00:57:28 -!- sacje has quit (Ping timeout: 276 seconds). 00:59:05 Sgeo: ...EVE? 00:59:12 talk about random clash of interests for me 00:59:25 -!- ChrisW has quit (Ping timeout: 250 seconds). 00:59:30 Yes. http://www.youtube.com/watch?v=C2fyO4uKkxo 01:00:29 I've never played, and don't know much, but I like reading about it 01:00:32 Especially major scams 01:00:40 I should read the Titans4U thread fully someday 01:03:02 `olist (908) 01:03:04 olist (908): shachaf oerjan Sgeo FireFly 01:04:14 Sgeo, have the redditors been crushed yet then 01:04:28 Phantom_Hoover: I think that battle has come and gone, afaik 01:04:38 Afaiu, the Redditors went into battle knowing they'd lose 01:04:54 ooh list 01:05:04 it's moved on from goons to redditors? 01:05:14 oerjan: Should it be renamed oohlist? 01:05:27 naah 01:07:11 Bike, no, this was the goons declaring war on the redditors 01:07:29 lol. 01:07:42 the game economy shifted around and left them sitting on top of a pile of resources they weren't equipped to defend 01:11:45 -!- sacje has joined. 01:14:33 elliott: The lambdabot dice thing says "shachaf:" even in /msg. 01:30:50 http://designislaw.tumblr.com/post/57473412060/ what in the fuck 01:37:18 Is there a "overlapping rectangle picture format"? I do have idea for making up such a thing, though. 01:38:55 how can people sustain these levels of anger 01:39:32 Phantom_Hoover: chronic physical pain hth 01:39:54 tfh 01:40:03 zzo38: what do you mean by "overlapping rectangle", is it a sort of layers? 01:40:52 lifthrasiir: I mean that the picture is made up by a bunch of rectangles, all of which are black, and everywhere other than such rectangle is white space. 01:41:22 Phantom_Hoover: mostly i was reacting to like the wi-fi thing. 01:41:36 maybe there might be an image format based on quadtrees. 01:41:56 -!- carado has joined. 01:42:03 (but of course individual squares in the quadtree do not overlap, I know) 01:42:16 svg? 01:42:54 Bike, yeah, that's stupid and offensive. 01:44:12 I have made a program in TeX for printing .PBM pictures, but I think a more efficient format could be made up, using the following features: * It isn't too difficult to parse using TeX. * It is monochrome file. * It is ASCII format. * The data does not take up too much space. * It can be used with other programs too. * The output is efficient (such as using rectangles larger than individual pixels). * It can be scaled, but may have a default 01:51:07 -!- oerjan has quit (Quit: leaving). 01:52:26 -!- sacje has quit (Ping timeout: 256 seconds). 01:52:54 -!- sacje has joined. 01:58:07 -!- Bike has quit (Ping timeout: 240 seconds). 02:06:23 -!- Bike has joined. 02:07:03 -!- Bike_ has joined. 02:11:06 -!- Bike has quit (Ping timeout: 268 seconds). 02:11:21 i like how the IBM repair form assumes you live in a city with 8 or fewer letters in its name 02:11:24 i mean doesn't everyone 02:13:14 -!- Bike_ has changed nick to Bike. 02:13:19 @_@ ... really 02:13:19 ... really 02:13:20 also there are two fields for name, one of which is long enough to fit my name, one of which is not 02:13:21 took me five tries to think of one that would fit 02:13:24 "los angeles" 02:13:25 -!- sacje has quit (Ping timeout: 248 seconds). 02:13:32 "san francisco" 02:13:41 um um um "new york" that fits 02:14:32 new york city 02:14:36 I thought of Portland. 02:14:53 "chicago"? 02:16:05 Bike: well, the city's name is New York 02:17:09 Just give the nearest airport code. Surely that's precise enough? 02:17:28 the nearest airport code to me is in another state, probably 02:18:27 -!- sacje has joined. 02:18:29 -!- sacje has quit (Changing host). 02:18:29 -!- sacje has joined. 02:19:05 kmc: keeg mcall from sanfranc 02:19:19 , californ 02:19:43 KMCALL~1 02:21:15 Well, damn, that's probably what actually happened. 02:21:20 does kmc have a middle name? i hope it's "Mc" 02:21:53 Someone decided to store the information as files on an 8.3 filesystem. 02:22:38 keegan #drugz mcallister 02:22:41 including the # 02:22:57 (pronounced "kay octothorpe drugz em cee") 02:25:23 elliott: itym "sulamit drugz" 02:25:28 hth 02:26:19 itym drogues 02:26:34 although idk why kmc's so interested in parachutes 02:34:29 I suppose "Sanfranc" should be sufficient. 02:35:13 what about Saint Francis, Wisconsin 02:35:41 kmc: "15.6" HD (1366 x 768)" 02:35:42 If you can enter the state too then it might help, though. 02:35:53 (The state can be given by a two-letter abbreviation, too, if needed.) 02:36:07 Oh, oh, I know, use the city and state fields to enter geographical coordinates. 02:36:58 could you have numeric filenames in DOS? 02:37:10 "Please mail to 1600 PENN AVE NW, 38-8976N 77-0366W" 02:37:18 by burrito drone 02:37:33 Oh, right, "20500", almost forgot. 02:37:35 Bike: Yes. 02:37:40 well, flying a drone in that particular area might be a bit problematic 02:37:43 elliott: I've met multiple people recently who would say e.g. "hashtag esoteric" for this channel 02:37:58 I think "hashtag" is yet another name for # now 02:38:31 I never pronounced the # in IRC channel names, the scant few times I've had to refer to IRC channels out loud. 02:39:12 it's still "sharp" for me, thankfully 02:39:30 ♯esoteric 02:39:34 ♯drugz 02:39:51 exactly 02:40:05 ♭esoteric 02:40:09 & looks kind of like a flat symbol if you squint and give yourself presbyopia. 02:40:42 -!- copumpkin has quit (Quit: Textual IRC Client: www.textualapp.com). 02:40:55 -!- copumpkin has joined. 02:41:10 -!- zzo38 has quit (Disconnected by services). 02:41:14 -!- zzo38 has joined. 02:41:43 just use a b like everyone else. 02:42:37 & is for server-local channels. 02:43:10 -!- Phantom_Hoover has quit (Read error: Connection reset by peer). 02:43:33 monotone: Yes, that is what it is in IRC. # is network, & is local to server, + is network modeless, ! is network with server-side codes added to the name to prevent takeovers from netsplits. 02:43:49 zzo38: Do you have any document with proposals for improving the IRC protocols? 02:44:01 s/..$/?/ 02:44:10 On servers that support all channel types, you shouldn't use the # type because other types works better. 02:44:34 shachaf: No. 02:45:10 I've never seen any network that supports the latter two types. 02:47:01 The types &!+ are all immune (although for three different reasons). I think I have seen some networks that support those types, but it isn't very common, it appears. 02:47:27 immune... to the irc plague? 02:47:50 No, they are immune to netsplits. 02:49:01 oh. 02:49:42 immune to hugz 02:51:50 -!- sacje has quit (Ping timeout: 240 seconds). 02:52:44 -!- sacje has joined. 02:58:23 -!- carado has quit (Ping timeout: 246 seconds). 03:05:49 `danddreclist 38 03:05:50 danddreclist 38: shachaf nooodl 03:12:33 zzo38: Thanks! 03:12:41 Where can I find it? 03:14:16 http://zzo38computer.org/dnd/recording/level20.tex (the precompiled .dvi is also available) 03:14:30 Maybe that should be noted somewhere. 03:14:36 Like I said, perhaps we should make it so that whatever `*lists are mentioned so you can write "`? `danddreclist" so you know. 03:15:04 `run echo 'http://zzo38computer.org/dnd/recording/level20.tex (the precompiled .dvi is also available)' > wisdom/danddreclist 03:15:07 No output. 03:15:14 `? danddreclist 03:15:16 http://zzo38computer.org/dnd/recording/level20.tex (the precompiled .dvi is also available) 03:15:58 `? slist 03:16:00 slist? ¯\(°_o)/¯ 03:16:00 | 03:16:00 o/`¯º 03:16:21 `run echo 'Update notification for the webcomic Homestuck.' >wisdom/slist 03:16:24 No output. 03:16:37 `run echo 'Update notification for the webcomic Order of the Stick.' >wisdom/olist 03:16:41 No output. 03:16:59 shachaf can write something humorous for `smlist. 03:17:31 how about 'Update notification for the webcomic Super Mega.' hth 03:17:54 Those descriptions are not good, though. 03:18:10 They should have a URL, and they shouldn't say "webcomic". 03:18:35 Yes, probably that would help a bit. 03:19:04 -!- sacje has quit (Ping timeout: 260 seconds). 03:19:42 -!- kallisti has quit (Ping timeout: 256 seconds). 03:20:17 zzo38: Most networks use timestamping for # channels by default to resolve netsplit collisions, so a separate type is superfluous. 03:21:02 -!- sacje has joined. 03:21:43 monotone: Well, it makes ! superfluous anyways, but still only if the network is programmed to use the protocol that does that. Also, it doesn't make + superfluous. 03:27:30 -!- Bike has quit (Ping timeout: 245 seconds). 03:29:53 -!- Bike has joined. 03:31:10 -!- sacje has quit (Ping timeout: 240 seconds). 03:34:10 -!- sacje has joined. 04:04:23 -!- Bike has quit (Ping timeout: 240 seconds). 04:10:29 -!- sacje has quit (Read error: Connection reset by peer). 04:13:01 -!- Bike_ has joined. 04:46:18 -!- Bike_ has changed nick to Bike. 05:07:30 -!- Vorpal has quit (Quit: ZNC - http://znc.sourceforge.net). 05:09:07 Did you read my Dungeons&Dragons game latest one? 05:12:28 No. 05:12:35 Could you set up some way for me to read it in my browser? 05:12:41 That would make it much easier. 05:13:20 Either serve it as plaintext with text/plain (e.g. with a symlink level20.txt?) or as HTML or as PDF. 05:18:17 Prefix the URL with "view-source:" that might work in some browsers. 05:18:56 Now you can make bets on what would happen next in this game. 05:19:41 I might also add a .htaccess to make it serve as text/plain. 05:22:42 OK I fixed it now. 05:23:15 (Note: The number after "`danddreclist" is the session number.) 05:23:38 shachaf: Now it is text/plain I made the .htaccess to override that file 05:31:36 -!- conehead has quit (Ping timeout: 260 seconds). 05:38:14 zzo38: Thanks! 05:38:21 I kind of feel bad because application/x-tex is more correct. 05:38:29 But my browser is bad. 06:05:20 -!- conehead has joined. 06:14:14 -!- sebbu2 has joined. 06:15:16 http://i.imgur.com/rtFkBme.gif 06:17:19 -!- sebbu has quit (Ping timeout: 264 seconds). 06:18:36 me too, seal. 06:20:13 That was a pretty bad loop 06:20:27 Everyone can identify with rotating seal. 06:29:39 His ratings are off the charts 06:31:43 But he alienates his original fans who are generally cynical and depressed and popularity-averse. 06:32:12 Seal deals with this by rotating. 06:33:09 He owes nothing to you 07:09:39 http://www.youtube.com/watch?v=6G_IU5lK1E8 even in 2029, when we have 32 TB pages, we will use IPv4 07:09:42 and have slow-ass terms 07:12:47 -!- Frooxius has quit (Ping timeout: 264 seconds). 07:20:34 Slow-ass terms without properly aligned text! 07:20:56 Bike: ok but this aesthetic is fantastic. 07:21:14 I don't even know what's going on but still 07:26:59 -!- epicmonkey has joined. 07:28:19 Bike: It actually uses your local IP address. 07:43:50 Why does malloc(sizeof(Macro)) cause a segmentation fault even though usually it doesn't? 07:44:06 -!- Taneb has joined. 07:45:48 sounds like the heap is corrupted by something else before you get to that call 07:46:16 O, yes I suppose that might cause it. 07:47:50 -!- mnoqy has joined. 07:49:57 -!- epicmonkey has quit (Ping timeout: 248 seconds). 08:19:37 -!- oklopol has joined. 08:23:59 okokokokokokokokokokokokokokokokokokokokokokokokokokokokokokoko 08:24:05 okokokokokokokokokokokokokokokokokokokokokokokokokokokokokokoko 08:24:12 okokokokokokokokokokokokokokokokokokokokokokokokokokokokokokoko 08:24:23 oko. 08:32:59 G'day 08:33:07 -!- aloril has quit (Ping timeout: 276 seconds). 08:49:31 -!- aloril has joined. 08:59:26 -!- conehead has quit (Quit: Computer has gone to sleep.). 09:07:19 hi Taneb 09:07:30 did you know most `olist readers have never played d&d?? 09:09:37 -!- epicmonkey has joined. 09:10:33 -!- nooodl has joined. 09:23:04 -!- katla has joined. 09:26:51 I did not 09:28:34 -!- BlueShark has changed nick to BlueShark_. 09:30:39 -!- BlueShark_ has changed nick to BlueShark. 09:31:36 `relcome BlueShark 09:31:39 ​BlueShark: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.) 09:38:27 -!- aloril has quit (Read error: Operation timed out). 09:40:43 -!- desu__ has joined. 09:43:20 -!- katla has quit (Ping timeout: 256 seconds). 09:44:57 -!- desu__ has changed nick to katla. 09:54:29 -!- aloril has joined. 10:02:45 -!- Taneb has quit (Ping timeout: 248 seconds). 10:05:23 -!- CADD has joined. 10:12:28 -!- zzo38 has quit (Remote host closed the connection). 10:34:55 "Thank you for emailing GIGABYTE. We are delighted with your interest in our products. This is a known issue. Please contact Seagate technical support for latest firmware of Samsung HD105SI. (Seagate took over Samsung HDD business)." 10:35:05 Ooh, I might get my SYSTEM to WORK. 10:37:26 Samsung had a HDD business? 10:38:06 I think it was a "brand" only, maybe? I don't know. 10:38:17 http://knowledge.seagate.com/articles/en_US/FAQ/223631en "This patch code is released in order to solve the compatibilty problem between some motherboards (the AMD SB850 chipset and the Intel P67/H67 chipset) and Samsung-brand hard drives, F3 and F3EG models only." 10:38:21 Yay. 10:38:40 (I *did* search for updates before, but couldn't find any; only found this via some direct googling.) 10:40:15 @yow! 10:40:16 There's a little picture of ED MCMAHON doing BAD THINGS to JOAN RIVERS 10:40:16 in a $200,000 MALIBU BEACH HOUSE!! 10:40:33 ?????? 10:40:44 yow indeed 10:41:03 You should remove that picture from the Malibu beach house. 10:41:14 http://en.wikipedia.org/wiki/File:Diagram_of_Hard_Disk_Drive_Manufacturer_Consolidation.svg 10:41:51 Of course I just installed an operating system on a different disk this morning. Now I need to decide whether it'll be faster to try moving and resizing it, or just reinstall. (Probably to reinstall.) 10:42:48 "List of defunct hard disk manufacturers" has an (unintentionally?) hilarious title. 10:42:58 "Yeah, I'm not going to buy one of *those*." 10:50:13 -!- CADD has quit (Quit: Lost terminal). 10:56:47 -!- CADD has joined. 11:01:24 -!- CADD has quit (Client Quit). 11:01:41 -!- CADD has joined. 11:24:10 ``The equivalent of 'tar xf' is 'cpio -i -d -m' (BSD cpio) or 'cpio -i --make-directories --preserve-modification-time --no-absolute-filenames' (GNU cpio).'' 11:24:11 ​/home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: `The: not found 11:27:23 -!- Sgeo has quit (Read error: Connection reset by peer). 11:28:14 --make-directories is -d and --preserve-modification-time is -m 11:28:47 Just use cpio -dim and don't extract weird cpio files 11:29:52 "cpio -bright" 11:30:52 cpio -mid if antonyms confuse you. 11:50:16 -!- MindlessDrone has joined. 12:06:01 -!- yorick has joined. 12:13:00 -!- sebbu2 has changed nick to sebbu. 12:13:30 -!- copumpkin has quit (Ping timeout: 256 seconds). 12:32:26 Deewiant: "cpio -early" "cpio -late" (cf. http://sprunge.us/GNZF for refs). 12:33:24 Eh, that's quite roundabout. 12:33:54 Alternatively, "cpio -extreme" or some-such. 12:35:24 If related words confuse you then you're probably screwed anyway. 12:36:14 I'm easily confused. 12:37:07 You'll probably try to use unzip instead of cpio. 12:37:07 "cpio -idm" -> Intelligent Dance Music -> most antonymical music subgenre -> cpio -something. 12:37:26 (I don't know what the opposite of IDM is.) 12:38:43 Dumb Noise? 12:39:01 Dumb Arythmic Noise 12:39:06 cpio -dan 12:40:01 http://sprunge.us/hDFZ I was sort of not expecting to end up on a permanent announcement mailing list. 12:41:53 Woot 12:42:01 -!- carado has joined. 12:44:14 -!- Taneb has joined. 12:50:13 -!- oerjan has joined. 12:57:14 -!- FireFly has quit (Ping timeout: 264 seconds). 12:57:37 -!- boily has joined. 13:01:35 -!- FireFly has joined. 13:23:57 (I used their bus once in June.) 13:24:23 (In Summer? You are a loon!) 13:27:04 Well, I mean, it was too long a walk. 13:27:27 Fair enough. I'm glad we had this talk. 13:27:49 I see what you did there. (But only just now.) 13:27:51 good fizzie-met-taneb morning! 13:28:08 `relcome fizzie 13:28:11 ​fizzie: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.) 13:28:37 Yay! 13:30:13 I have a poetic flair. I'm glad there's no row. 13:33:00 I thought IDM was dumb arrhythmic noise 13:33:51 Pretty sure that that's Chinese Water Torture 13:35:32 why is Taneb sounding like a markov bot 13:36:14 oerjan: Except one that rhymes a lot. 13:36:14 I've missed half the conversation and decided to reply in rhyme without logreading 13:36:20 fungot: Can you do poetry? 13:36:20 fizzie: black is white. up is down. their mysql backend must've crashed like three times, each time i run it 13:36:31 Apparently the answer is a resounding yes. 13:36:44 ^style 13:36:45 Available: agora alice c64 ct darwin discworld enron europarl ff7 fisher fungot homestuck ic irc* iwcs jargon lovecraft nethack pa qwantz sms speeches ss wp youtube 13:36:58 very avant-garde 13:37:00 Whether that counts as poetry is anyone's guess. 13:37:13 ^style enron 13:37:13 Selected style: enron (subset of the Enron email dataset) 13:37:23 fungot, what's your best limerick? 13:37:24 Taneb: over the next two to the finish line. index 3 of the costs not covered. subject: attached you will find to be next one. 13:38:13 there are too many people on this channel. let's group them according to their similarities! 13:38:15 ^style agora 13:38:16 Selected style: agora (a large selection of Agora rules, both current and historical) 13:38:21 * boily tie-wraps Taneb onto fungot 13:38:22 boily: than the maximum hand size or pending draws for that board of appeals consists of three and fourteen days 13:38:26 =P 13:38:57 * boily duct tapes elliott to oerjan 13:39:09 boily: I've clustered the people here a couple of times. 13:39:10 i'm similar to elliott? 13:39:25 ^style lovecraft 13:39:25 Selected style: lovecraft (H. P. Lovecraft's writings) 13:39:33 fungot, how do you feel? 13:39:35 Roujo: with bowers of vines and sweet blossoms, and seats and benches of marble and went down the steep slope that they might find men to whom he flung the remaining coil of fnord and after the crash there was nothing left standing in the street, swarthy, slender, and sinister with latent horror. hieroglyphics had covered the walls and windows shifted. nahum did not send her to the county asylum, but let her wander about the ho 13:39:58 that does not sound very healthy 13:40:15 oerjan: nah. I just wanted to see what happens. you know. for pure scientific reasons. yeah. science. *whistles* 13:41:03 boily: For example, here are some #esoteric people projected to two dimensions: http://xn--nxa.zem.fi/~fis/esomapnz.png 13:41:18 woooooah... 13:41:28 my god, it's full of stars... 13:41:42 fizzie: what are the axies? 13:42:02 fizzie: who is CADD? 13:42:11 CADD: First two principal components from a... I think something 34-dimensional. 13:42:12 boily: you can /whois me.. 13:42:24 boily: who be you? 13:42:47 fizzie: those are expectation-maximisation ellipses, yes? 13:43:19 fizzie: that tells me a lot, lol 13:43:34 boily: They're maximum-likelihood ellipses for 2D gaussians. You don't need EM to fit single ones. 13:43:42 * oerjan smells a lithuanian 13:43:48 oh right. my machine learning stuff is getting rusty. 13:43:51 CADD: http://xn--nxa.zem.fi/~fis/esomap-comp.png here's the pre-projection variant, I think it will help. 13:43:51 oerjan: exactly 13:44:11 fizzie: excellent 13:44:29 fizzie: very interesting 13:44:48 It's all very simple textual features. (The word classes are cheats.) 13:45:19 * boily is all contemplative and swoonish 13:45:20 http://xn--nxa.zem.fi/~fis/esoconf.png and that's how well they work in a classifier. 13:45:43 oerjan: good catch, most people guess greek 13:45:46 (Of course it was a p. easy task, with just those ten folks and giant blobs of text.) 13:46:24 can you easily recreate your figures with current members? 13:46:33 -!- metasepia has joined. 13:46:50 CADD: yeah i thought greek for a second but something was off 13:46:52 boily: No, but I have a much better set of figures that also aren't easily recreatable. 13:47:00 boily: There's a SVM classifier and everything. 13:47:10 boily: (Sadly, I don't remember where I put them.) 13:47:46 you really went full force on that. I should get back on track with ML; those were fun times. 13:48:09 oerjan: well with your last name im guessing you are somewhere in scandinavia 13:49:29 boily: ML is great! you should join us in #machinelearning 13:49:37 boily: I'll tell you the story of the cheated word class features in these plots, it's sort of related. 13:49:41 boily: Basically, anything that'd actually try some parsing would've been terribly slow and/or not much better than guessing, so I assigned each word a word class based on its wordnet entry; if the word had entries in several categories, I picked the one with the highest frequency in wordnet's tagged texts. 13:49:46 boily: So every instance of, say, "die", has been considered a verb. 13:50:19 I can't seem to find the newer plots. Possibly because I have a habit of naming files things like "blort4.png". (Cf. http://users.ics.aalto.fi/htkallas/blort4.png ) 13:51:19 fizzie: if I ever get back to academia, I'll name a figure “blort4” in your honour, and have a crappitude somewhere in the article. 13:51:52 I don't quite recall what's the proper name for the metric called "crappitude" in that graph. 13:52:21 I think it's the average of within-class accuracy. 13:52:32 (Distinct from accuracy if the classes are of non-uniform size.) 13:52:52 Arguably it's a misleading name, because you want to have a high crappitude. 13:53:30 you should create a graph of fizziliciousness 13:53:50 -!- Bike has quit (Ping timeout: 240 seconds). 13:54:10 http://users.ics.aalto.fi/htkallas/alise_vs_vorpal.png I have no idea whatsoever what's happening here, but it's certainly something #esoteric-related. 13:54:12 is it just a coincidence that fizzie and boily are both properties held by some liquids? 13:54:35 quintopia: As far as you know. 13:54:38 quintopia, so's elliott 13:54:49 And HackEgo 13:54:58 elliott is a liquid? 13:55:07 Some liquids are elliott 13:55:21 As in, "That's some very elliott coffee" 13:55:21 * boily stirs elliott with a spoon 13:55:41 * boily thwacks Taneb with a not quite ripe banana 13:55:49 -!- Bike has joined. 13:55:54 fizzie: so, you can recognize us just by the words we use or...? 13:57:42 quintopia: With the features used above, only for a sufficiently (read: very) large sample. Otherwise there's too much noise in the statistics. (Since they're things like word length distributions and the like.) 13:58:13 Actually, it doesn't really look at any words, except for the counts of particular things like classes of pronouns. 13:58:29 fizzie: that also assumes that your language patterns dont drift either 13:58:44 fizzie: i wonder what an optimal classifer of us would look like, one that actually used words as features 13:58:56 CADD: i'm norwegian 13:59:11 oerjan: cool cool 14:00:04 "Adversarial Stylometry: Circumventing Authorship Recognition to Preserve Privacy and Anonymity" -- https://psal.cs.drexel.edu/index.php/Main_Page -- was recently featured in /.. 14:00:25 (You can use that to avoid being recognized on #esoteric.) 14:00:37 (Possibly combining that with changing your nickname would be advisable, however.) 14:01:41 (and perhaps with connecting from a different network than usual) 14:02:08 (And possibly taking a bath first.) 14:09:59 -!- jsvine has joined. 14:30:09 -!- augur has quit (Remote host closed the connection). 14:38:28 > not (3 < 5) 14:38:30 False 14:39:40 cat /dev/urandom >> lambdabot 14:40:20 I once did `cat /dev/urandom` while connected via PuTTY 14:40:26 PuTTY didn't really like it 14:40:27 =P 14:45:54 fancy 15:13:28 ~eval not (3 < 5) 15:13:30 Error (1): 15:13:32 ~eval not (3 < 5) 15:13:34 False 15:15:01 -!- Taneb has quit (Quit: Leaving). 15:15:42 -!- Taneb has joined. 15:18:33 -!- augur has joined. 15:18:57 ~eval (35 > 2) 15:18:57 Error (1): :1:3: lexical error at character '\SYN' 15:20:43 go and syn no more 15:21:00 ^^ 15:22:02 lol 15:22:15 -!- KingOfKarlsruhe has joined. 15:23:33 Roujo: what are you trying to do exactly, you vile elliotty bot abuser? 15:23:55 I'm experimenting 15:24:01 Some bots strip control codes 15:24:07 Looks like this bot doesn't =P 15:36:10 -!- oerjan has quit (Quit: leaving). 15:43:28 I am making a boot floppy, in 2013. 15:49:52 wat 15:49:54 (The Samsung drive flash utility only works from DOS, and does not work in the computer that can't see the drives, so I had to use an old desktop from the closet, and it's too stupid to boot from USB and I don't want to waste a CD-R.) 15:49:54 Nice =)< 15:50:01 ^^ 15:50:07 Also the flash utility was too big to fit on a floppy (1.7M). 15:50:25 Fortunately the computer in question was new enough to support USB storage emulation, even if it doesn't boot from one. 15:54:40 -!- Taneb has quit (Ping timeout: 246 seconds). 15:59:44 -!- CADD has quit (Ping timeout: 260 seconds). 16:01:15 hi everybody 16:01:41 -!- CADD has joined. 16:02:16 -!- augur has quit (Ping timeout: 264 seconds). 16:02:32 kmc: Heya 16:11:52 fizzie: awesome 16:12:34 reflashing hard drives is always a party 16:12:54 I liked it when Samsung released a fix for a critical data loss issue but didn't bother bumping the version number, so you couldn't tell if your drives had it already 16:21:04 The reflash made the drive appear and work everywhere else, except that Windows 8 setup utility says Windows cannot be installed on it because "the system may not support booting from it", or some-such. 16:21:23 Same applies to the SSD, but not the two disks reserved for Linux use. 16:24:33 o_O 16:24:41 is there anything unusual about these drives 16:27:32 -!- jsvine has quit (Quit: Leaving.). 16:39:52 Ok, apparently none of the shader code I ever wrote compiles with Mesa ): 16:41:23 Ok, got it to work by making all integer constants into float constants 16:44:21 -!- iamfishhead has joined. 16:45:11 Which means just putting a whole lot of decimal points all over the place 16:47:59 -!- Frooxius has joined. 16:53:33 -!- jsvine has joined. 17:03:13 isn't it great how every renderer implements a slightly different shader language 17:03:47 "yes" 17:07:21 #define "yes" maybe 17:07:30 Bret Victor - The Future of Programming http://vimeo.com/71278954 17:08:34 `? yes 17:08:35 yes? ¯\(°_o)/¯ 17:08:35 | 17:08:36 º¯`\o 17:09:03 I should really switch to a monospaced font >_> 17:12:34 as long as you're not using papyrus, I guess everything's fine. 17:12:47 (or maybe papyrus is the secret to understanding fungot?) 17:12:48 boily: i talked with the policemen, seemed at first as perplexed as armitage and his companions ride on fnord or in fnord chariots. carter and his guide climbed up an alley that was all they ever did was clutch and fly and fnord that was not of 17:13:31 There we go 17:16:34 Is there a monospace version of comic sans? 17:18:06 fungot: don't you write code in comic sans? 17:18:07 olsner: i soon found that it slipped from sight as quickly as i had never before been an extravagant dreamer. many of these took lodgings in the battered houses that had once known the songs of iranon. but though the dreaded martenses were gone, whilst most of the fine scroll-and-urn overmantels and fnord cupboard fnord were gone, pabodie, sherman, ropes, sherman, and the 17:18:37 !style c64 17:18:39 hmm, I guess that counts as avoiding the subject 17:20:09 ^style c64 17:20:09 Selected style: c64 (C64 programming material) 17:20:21 fungot: pabodie, sherman. sherman. sherman. sherman. 17:20:21 boily: if not impossible. if the msb of the video matrix and character dot-data area to act as the current status of that register, switch in the mob-mob collision bit for each sprite can be used 17:24:48 ^style 17:24:48 Available: agora alice c64* ct darwin discworld enron europarl ff7 fisher fungot homestuck ic irc iwcs jargon lovecraft nethack pa qwantz sms speeches ss wp youtube 17:25:03 ^style nethack 17:25:03 Selected style: nethack (NetHack 3.4.3 data.base, rumors.tru, rumors.fal) 17:25:14 fungot: GO! =P 17:25:14 Roujo: looking glass, mirror: " if you don't need any help from the silver mountain, while he was born from the cave; the ferocious fear the only good djinni is a long-drawn-out affair. fresh whole tripe calls for a good talisman if handled right. 17:25:19 ... 17:25:20 " 17:25:21 there 17:26:59 I haven't played Djinni yet; I'm still waiting for 0.13 to be out of trunk. 17:33:10 error: mismatched types: expected `u32` but found `uint` (expected uint but found u32) 17:37:25 boily: advice: dont 17:38:34 but, but... all I wanted is to die in new creative ways... won't djinni even offer me that simple pleasure? 17:39:03 kmc: Turns out that creating partitions on the disk and then restarting the installer made it consider the disk kosher. 17:39:35 I had put an empty GPT disklabel on it, and I guess that was maybe too confusing? 17:41:02 -!- epicmonkey has quit (Ping timeout: 240 seconds). 17:45:47 17:21:04 The reflash made the drive appear and work everywhere else, except that Windows 8 setup utility says Windows cannot be installed on it because "the system may not support booting from it", or some-such. 17:45:59 I think this means "it's GPT but I'm not sure you have EFI" 17:46:10 Windows only supports booting GPT from EFI 17:47:48 I had already set the bios to UEFI-only boot. 17:48:00 But I guess it could've been just confuzzled. 17:51:48 -!- aloril has quit (Ping timeout: 256 seconds). 17:52:34 fizzie: "set the bios to UEFI-only" 17:55:08 Yes? I didn't want to boot in legacy mode by accident. 17:56:22 elliott: is this the thing where uefi is a replacement for bios 17:56:46 yes 17:57:20 I was holding off trying to think of a good analogy to describe my reaction to that statement but failed 17:57:22 dammit. I have sacramental bread stuck to my lips. 17:57:27 ow. 17:57:43 boily: Is it still sacramental bread, or is it already Jesus? 17:57:46 dare I ask why 17:58:26 fizzie: no jesus yet. 17:58:32 elliott: you may dare. 18:03:22 -!- gsteinbrenner has joined. 18:03:46 `relcome gsteinbrenner 18:03:49 ​gsteinbrenner: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.) 18:04:09 hi all 18:04:12 boily: I have no dares 18:04:15 this channel seems promising 18:04:23 hi 18:05:16 -!- aloril has joined. 18:05:36 fizzie: huh, I would think that to make a disk kosher you would have to cover it in salt and bury it for two weeks 18:05:44 elliott: well, the truth is that a colleague has «retailles d'hosties» on his desk (you can get them for very cheap at grocery stores). these things are evil, and become like glue if you let them on your lips for too long. 18:05:47 D 18:05:48 :D 18:05:57 Yeah, those =P 18:06:05 They don't taste like anything though 18:06:08 Why even have them 18:06:37 they were there, and I was hungry. 18:07:02 kmc: read that as computer disk 18:07:04 can recommend the experience 18:07:15 it was a joke on what fizzie said 18:07:19 oh 18:07:22 Roujo: sure, I could have had a banana, but our new batch's morphogenic field is already too weak, and I don't like pears. 18:07:23 I didn't actually look at what fizzie said though 18:07:26 so that's okay 18:07:27 good job elliott 18:07:30 team player 18:07:36 elliott: he was talking about some disk problem before, I think 18:08:28 boily: Est-ce qu'on parle des mêmes bananes? o_0 18:08:36 Oh 18:08:44 Joke de banane molle 18:08:46 Got it 18:08:47 Nevermind 18:08:55 `relcome soft banana 18:08:57 ​soft: banana: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.) 18:09:09 `cat dog 18:09:11 cat: dog: No such file or directory 18:09:15 -!- Taneb has joined. 18:09:16 `cat /dev/urandom 18:09:18 ​knCvkjxڡqc1}K*hOgiӸ@4۸j0r˭S9"Ѿ!\EZX 18:09:25 That's it? 18:09:26 Good 18:09:51 `? ngevd 18:09:52 ​_E%ntC\ \ ?+)i/}G_<",b 18:09:59 Oh god 18:10:24 `relcome Cthulhu 18:10:26 ​Cthulhu: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.) 18:10:29 Oh 18:10:32 So it's not permanent 18:10:34 That's nice 18:11:49 nice, sounds like we have some thunder rolling in 18:11:55 Nice, nice 18:13:14 I love `? ngevd 18:14:04 one of the many high points of `? 18:18:03 ~metar CYUL 18:18:04 CYUL 061800Z 21010KT 30SM FEW040 BKN240 24/12 A3007 RMK CU1CI5 SLP182 DENSITY ALT 1000FT 18:18:21 olsner: where were you, again? 18:18:35 -!- zzo38 has joined. 18:18:45 Why is that output so hard to parse? =P 18:18:57 Also, does `? ngevd always put out garbage? 18:18:57 `? ngevd 18:18:59 ​ȃyuݴl*yqyhqfj#qL'>L|9\2)_O 18:19:02 Huh 18:19:06 `? ? 18:19:08 ​? is wisdom 18:19:21 In a way, I guess =P 18:19:42 `? !echo `? 18:19:43 Roujo: METARs are for airplane pilots, with a (not quite, in fact far from it) standard format. 18:19:43 ​!echo `?? ¯\(°_o)/¯ 18:19:44 | 18:19:44 º¯`\o 18:19:56 !echo Hats 18:19:57 ~metar ESSL 18:19:58 Hats 18:19:58 ESSL 061750Z 16006KT 9999NDV SHRA SCT025CB 20/19 Q1011 18:20:08 oh, SHRA! lucky you! 18:20:10 boily: Oh, cool =) 18:21:33 what does 9999NDV mean? 18:21:41 first time I see that. 18:22:21 “No Direction Variable”. 18:22:50 another source cites “No Direction Visibility”. 18:23:02 Ah, standards 18:25:06 olsner: 9999 means “don't even think about being bothered about ground visibility” and NDV is “and if you're thinking of it, direction doesn't matter”. hth. 18:29:48 boily: thanks, I'm not so worried about the visibility now 18:30:21 I like how humidity seems to be recorded in dewpoint temperature rather than e.g. humidity 18:30:52 right now, we're at 47.0%. 18:32:31 you're at 94.0%. I think it's pouring outside. 18:33:41 it's not actually raining now, just ominously rainylooking 18:34:20 at which distance are you from ESSL? 18:34:34 20 minutes or so 18:34:41 shachaf: hichaf 18:35:10 olsner: by car? 18:36:10 by foot! 18:36:44 so you're probably in northern linköping. how much do you weigh? 18:37:19 hi kmc 18:37:29 yay 18:37:30 -!- Bike has quit (Ping timeout: 245 seconds). 18:37:32 how's it going? 18:38:31 my sleeping is back to being kind of messed up :'( 18:38:43 i was waking up at ~7 for a while, it was nice 18:39:38 I've been waking up around 07:30 but failing to actually get out of bed until like 08:10 18:39:39 -!- Bike has joined. 18:39:42 olsner: I'll put you on Norrköpingsvägen in my file. 18:40:00 that kind of schedule would be fine with me 18:40:16 -!- desu__ has joined. 18:40:23 `relcome desu__ 18:40:25 ​desu__: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.) 18:40:31 -!- sacje has joined. 18:40:54 `relcome sacje 18:40:56 ​sacje: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.) 18:41:16 why thank you Roujo 18:41:35 You'relcome ^^ 18:42:02 Roujo: Have you considered the many advantages of a seven-letter nick? 18:42:11 I think you'd be much better off with one. 18:42:36 -!- katla has quit (Ping timeout: 260 seconds). 18:42:46 Roujo: don't listen to that unorthodox heretic. 5 is the way to godliness. 18:43:25 shachaf: Meh. I'm aither named Roujo or Skynet, really. 18:43:31 shachaf: still going on about that? 18:43:34 Or Gilnaur/Feyllin in games. 18:43:39 But never on IRC 18:43:40 =P 18:44:20 mnoqy: you can keep your five-letter nick 18:45:25 -!- desu__ has changed nick to katla. 18:45:31 -!- Bike has quit (Ping timeout: 264 seconds). 18:45:51 shachaf: So, advantages, you say? Do tell. 18:47:27 -!- Bike has joined. 18:51:24 -!- epicmonkey has joined. 18:58:31 -!- KingOfKarlsruhe has quit (Quit: ChatZilla 0.9.90.1 [Firefox 22.0/2013062200]). 19:02:53 -!- Bike_ has joined. 19:03:50 -!- Bike has quit (Ping timeout: 240 seconds). 19:06:56 hmm, so according to this speaker voice they were going to reveal "the identity of the 12th doctor" 19:07:30 -!- Bike_ has quit (Ping timeout: 245 seconds). 19:09:13 -!- Bike has joined. 19:09:16 -!- CADD has quit (Quit: Lost terminal). 19:10:23 "This partition is formatted with the ntfs." Oh debian-installer, you so funny. 19:15:36 I should participate in the IOCCC but I have no ideas :( 19:16:18 I'm perusing the crawl sources through gitk, and I stumbled upon a commit about “nooodl-y vaults (#7353)”. 19:16:19 Gregor: maybe you should participate in icfpc 19:16:23 is that from our own nooodl? 19:16:30 should i do that? i have 4 hours to decide 19:16:34 boily: yes 19:16:39 nooodl: :D 19:16:49 i'm in CREDITS.txt, too! 19:16:49 icfp always seems so boring to me 19:17:05 all for making some crappy vaults 19:17:07 i dont unerstand how people can bear playing it 19:17:09 shachaf: when is it? 19:17:25 icfpc? 19:17:26 olsner: In 52 hours or so. 19:17:28 boily: the elliotts are also me 19:17:32 (I am nooodl) 19:17:38 Registration deadline is 48 hours before. 19:18:00 can you enter without registering? 19:18:03 No. 19:18:25 * boily ties ellioerjan to nooodl with hemp rope 19:18:30 "To participate in the contest you need to pre-register with EasyChair. The pre-registration will close 72 hours prior to the contest, so register now." 19:18:58 "You need to pre-register for the contest, 48 hours prior (extended from 72h) to the contest at the latest." 19:19:02 difficultchair 19:19:11 "this year's contest will involve an element of program synthesis" 19:19:52 im kind of interested but i know ill give up as soon as i see the actual problem set :/ 19:21:20 why are they using what appears to be a system for submitting papers? 19:21:59 Because they're part of ICFP. 19:22:05 And probably just inherited the system from that. 19:23:55 I don't think this system was used before? I don't know. 19:26:43 can someone paste me the challenge once its released 19:26:54 or maybe itll be an annoyig one where you have to submit to test your code 19:27:00 Should I enter the ICFPC? I'd probably suck at it 19:27:10 you should enter if you're going to seriously try 19:27:47 It'd be a solo project by a bored teenager against teams of professionals possibly 19:28:13 professional icfpc competitor 19:28:26 :P 19:30:25 -!- MindlessDrone has quit (Quit: MindlessDrone). 19:30:30 What challenge is that? Does it involve submitting a paper? 19:30:33 “Fix a Kirke crash by deporkalating transiting hogs correctly.” :D 19:31:04 zzo38: No, it involves writing a programs. 19:31:39 zzo38: Examples: https://en.wikipedia.org/wiki/ICFP_Programming_Contest#Past_tasks 19:32:01 If the program is written in WEB then you can submit a papers too, though? 19:32:24 If you want. 19:32:39 What programming languages do they use? And it says "what appears to be a system for submitting papers" so is it a system for submitting papers or not? 19:32:41 Once the contest is finished, a few teams will be asked to submit two page document describing their solutions to be considered for the judges prize. 19:33:00 There's a system which was designed for submitting papers but instead it's used for registering your team for the contest. 19:36:09 shachaf: do you have something like a team to register yourself with? 19:36:10 -!- katla has quit (Quit: Leaving). 19:36:21 You need a team? 19:36:33 I've just registered as an individual 19:36:42 It's teams or individuals 19:36:54 olsner: Me? No. 19:37:04 I think I will probably not do the contest. 19:37:50 O, so you are using a system designed for something else. It seems also like how registration for Z-Comp uses a system designed for a MUD (to enter Z-Comp, you register an account on ifMUD, and then go into my apartment, and then go northwest, and then fill up the forms in that room; but first you have to calculate the SHA-1 hash of your entry before filling up the forms). 19:38:40 zzo38: where is your apartment? 19:38:53 boily: The apartment number is 11011 19:38:54 -!- NihilistDandy has quit (Quit: NihilistDandy). 19:39:04 -!- conehead has joined. 19:39:53 Free [] almost represents S-expressions 19:40:01 zzo38: I mean, IRL, unless if 11011 is already irly, then it's fine. I was subtly trying to subvertly extract your approximate coördinates. 19:40:12 boily: It has no location IRL. 19:40:30 (If you want my approximate coordinates the best thing you can do is by the IP address.) 19:40:54 zzo38: What about your (approximate?) body weigh? 19:41:06 I didn't figure it out recently and I don't want to tell to you. 19:42:47 zzo38: Delta, BC? 19:43:00 boily: Yes, that's close enough. 19:46:01 Do you know my ifMUD apartment has three floors, and the top one is rotating? 19:48:58 zzo38: you are tempting me with your mud talk. 19:54:07 The temple is so bright, that I install a darkness generator, so that it is a bit less bright. 19:57:17 as I understand it, ifMUD is like minecraft? 19:57:39 I don't think so. 19:58:13 You look (you can connect as a guest if you do not want to registration) 19:58:23 It is ifmud.org port 4000 20:03:10 -!- Bike has quit (Ping timeout: 246 seconds). 20:06:05 -!- NihilistDandy has joined. 20:07:07 -!- Bike_ has joined. 20:09:11 zzo38: how do I reach your apartment? 20:09:38 -!- Bike_ has changed nick to Bike. 20:09:47 -!- Taneb has quit (Quit: Leaving). 20:09:57 boily: The closest place you can teleport to is #20080. (Type '@teleport #20080') 20:10:01 And then type 11011 to enter. 20:11:10 slime? bletch! 20:13:56 Do you like this I wrote some messages in side of there 20:14:15 it sounds interesting. 20:14:59 -!- gsteinbrenner has quit (Quit: gsteinbrenner). 20:18:20 -!- pqh has joined. 20:21:23 that was fun! 20:21:53 If you write 'finger zzo38' then you can see my plan contains a quine. 20:22:51 The @print command in JotaCode is for concatenating strings; it doesn't actually print anything. 20:28:03 ~ping 20:28:03 Pong! 20:28:51 JotaCode feels very esoteric... 20:29:23 -!- pqh_ has joined. 20:29:25 -!- pqh has quit (Ping timeout: 246 seconds). 20:29:26 -!- pqh_ has changed nick to pqh. 20:29:43 -!- pqh has quit (Client Quit). 20:34:19 -!- lifthrasiir has quit (Ping timeout: 268 seconds). 20:36:57 how reliable are those sound-of-keyboard attacks? I wonder if you could use that to make a wireless keyboard by just disconnecting your keyboard and connecting a microphone 20:37:27 anyone here in boston? 20:37:38 careful of pianists who drum their keyboards! 20:38:42 quintopia: I think kmc is in boston. otherwise, the closest person is jsvine in NYC. 20:39:00 kmc moved to caliland 20:39:34 that's correct 20:39:56 edwardk is in Boston and copumpkin is in connecticut 20:39:59 but neither is in this channel :< 20:40:27 kmc: looking at your nëw coördinates, the closest member is Fiora. 20:40:37 Do any of you ever go in to Victoria, BC sometimes? 20:40:43 -!- sacje has quit (Ping timeout: 264 seconds). 20:40:51 there's nobody else in SF? 20:40:52 boily: you must not have coördinates from shachaf or Gracenotes then 20:40:53 i'm pretty sure zzo38 goes there most of the time 20:40:55 copumpkin used to be here, did he leave or is he just intermittent? 20:41:07 (or douglass_ or iamfishhead) 20:41:34 @ask edwardk hi wanna be friends in boston 20:41:34 Consider it noted. 20:41:45 also wasn't somebody here a MIT student? Vorpal? 20:41:48 am I completely imagining that? 20:42:21 kmc: indeed. I don't have Gracenotes yet, and shachaf is the current Unchallenged Champion of Not Wanting to be Coördinated. 20:42:31 -!- lifthrasiir has joined. 20:43:20 -!- sacje has joined. 20:43:28 zzo38: it's a few hours north of me, i gues 20:44:44 Bike: can I get your approximate coördinates and body weigh? as a bonus, you'll be the first entry in my file! 20:45:07 `? megalun 20:45:09 megalun is a chain of a million SCSI devices. FreeFull weighs 482 of them. 20:45:34 45 north, 122 west, 200 lbs. 20:45:44 zzo38: I would like to go to Vancouver sometime and maybe I would go to Victoria as well 20:46:08 i don't have a passport so i gues canading would be hard 20:46:14 you should fix that 20:46:18 Bike: thanks! 20:46:55 Canading is plenty worth it. you'll get to meet nice people, eat nice and healthy Canadian food, and enjoy the local scenery and mooses! 20:47:07 we get moose here anyway. 20:48:09 http://nedroid.com/2011/05/passport/ also 20:48:16 but our mooses are the most authenticest there is. 20:49:33 http://nedroid.com/2013/05/oh-canada/ boily. 20:50:41 YOU LIE! 20:51:09 we... uhm... we don't have any fascination whatsoever to funnels. yes. we don't. right. no funnels. 20:51:46 Roujo: you don't have any funnels on you, do you? in fact, you don't even know what a funnel is? 20:51:51 * boily *wink* *wink* 20:52:12 sm;oth 20:52:46 ~duck sm;oth 20:52:46 --- No relevant information 20:53:02 boily: look, funnel! 20:53:23 * boily twitches underneath his "poker" face 20:56:37 -!- oerjan has joined. 21:01:21 -!- epicmonkey has quit (Ping timeout: 264 seconds). 21:03:29 -!- boily has quit (Quit: FUNNELS!). 21:03:34 -!- metasepia has quit (Remote host closed the connection). 21:03:47 -!- neena has quit (Quit: leaving). 21:08:55 -!- Bike has quit (Ping timeout: 264 seconds). 21:08:58 somebody want to add fundeps to Rust 21:09:01 that would really make my day 21:10:41 didn't know Rust had something fundeps could be applied to 21:11:09 hmm, should I learn Rust? is that a good idea or anything? 21:11:09 -!- Bike has joined. 21:11:38 rust is the wave of the future, imo, 21:13:53 olsner: it kind of accidentally sort of has multi-parameter typeclasses 21:14:53 it has single-parameter "traits" which are like typeclasses, but you can also parametrize a trait on other type variables (like parametrizing a struct or a function) 21:15:02 so you could say that Foo implements trait Bar 21:15:13 which is like a 3-param typeclass except that the Foo location is still special in some ways 21:15:16 kinda gross imo 21:15:21 I guess Java interfaces + generics are like this too? 21:16:01 like, public class Node> 21:16:33 olsner: Rust is a cool and unique language; I suggest learning some stuff about it 21:16:53 actually writing and maintaining a library or something could be painful because the language still changes a huge amount every week 21:17:24 -!- KingOfKarlsruhe has joined. 21:17:30 -!- Bike has quit (Ping timeout: 245 seconds). 21:18:41 -!- Bike has joined. 21:25:30 Rust actually looks at what has been happening in the programming language world 21:25:33 Unlike Go 21:26:16 yep 21:26:37 they aren't really in the same niche either 21:27:50 Go uses garbage collection for everything, which is fine for many applications, but we already have Java, C#, Haskell, OCaml, Python, Ruby, JavaScript, etc. 21:28:15 Rust is a replacement for C++, which almost nobody else is seriously attempting 21:28:43 What is the replacement for C? 21:29:20 or you know I should say "Rust is aiming to become a viable replacement for C++ years from now" 21:30:09 I see Go as Java, simplified even further, with syntax weird enough that the "C is h4rdc0re" crowd will accept it 21:30:23 it does have some things going (ha!) for it, though 21:33:05 @MVPoliceBlotter "Man wearing Google Glass breaks window after walking into it while watching YouTube on El Camino Real near Calderon Ave." 21:33:05 Unknown command, try @list 21:34:14 -!- Bike has quit (Ping timeout: 240 seconds). 21:35:34 Do you know if they are making a replacement of C as well or only of C++? There are some variants of C which are mostly C, such as Cyclone, Objective-C, and Black-C, but do they do it differently? 21:36:01 kmc: hi 21:36:16 hi 21:36:45 where are you 21:37:36 desk 7115, 2 Harrison St., San Francisco, CA 94110, United States of America 21:37:44 I did see the macro system in Rust seems like it does a lot of things. 21:38:28 whoa. what a coincidence. that's where i launched the ICBM 21:38:40 :O 21:39:54 :o 21:39:56 I'm going to get kicked from ##c++ for telling them to use Rust won't i 21:40:14 telling ##c++ to use rust sounds like a good idea 21:40:25 possibly even better than being in ##c++ to begin with 21:40:29 kmc: oh no you're like those people who come into #haskell :'( 21:40:32 yep 21:40:39 I do actually answer C++ qusetions though 21:40:43 maybe correctly 21:42:09 -!- FreeFull has quit (Ping timeout: 268 seconds). 21:43:34 -!- FreeFull has joined. 21:46:49 what are the standard workarounds for lack of fundeps anyway 21:47:46 -!- Bike has joined. 21:48:14 ok, does anyone know a good linux distro for an idiot with a netbook. because my ubuntu is really not cutting it 21:48:16 lots of type annotations hth 21:48:30 Bike: ubuntu with not gnome 21:48:35 xubuntu or whatever 21:48:37 that's what i'm using now :( 21:48:39 kubuntu 21:48:40 ok 21:48:43 :P 21:48:43 well whats not cutting it 21:49:00 well the main problem, i don't even know if it's ubuntu really 21:49:14 Bike: check for a Canonical Genuine Ubuntu sticker 21:49:23 anyway, soon enough C++ will be The Haskell Done Right and all rusts and haskells will be irrelevant 21:49:31 but my wireless cuts out at random intervals and i can't figure out a way to recover that doesn't involve restarting my computer (which is irritating) 21:49:46 and i can't for the life of me figure out what the fuck is happening, there's a lot of closed issues in linux-wireless and ugh 21:49:51 kmc: I doubt it. 21:50:04 Bike: Does rmmod-modprobe work as a workaround instead of restarting? 21:50:08 nope 21:50:26 what's yer kernel version? 21:50:27 the reason i'm irritated at ubuntu is, in concert with this, the last upgrade whatever made it so that it waits /two whole minutes/ every boot to get a nonexistent network configuration 21:50:35 might be you can get a newer kernel from a PPA or something 21:50:38 Does dmesg say anything good? 21:50:39 i just edited an /etc to fix that particular issue but i'm not liking where this is going 21:50:52 shachaf: i can paste the dmesg if you want 21:50:57 kmc: um how do i check (see: idiot) 21:51:03 uname -r 21:51:13 i'm no expert on debugging wireless things 21:51:14 3.2.0-48-generic 21:51:50 i used some backported driver i don't understand to make it fail every few minutes instead of every few seconds, but bluh 21:52:51 imo give up and weep hth 21:52:55 been doing that. 21:53:09 not the giving up part apparently 21:53:24 the dmesg is like Failed to wakeup in 500us // Failed to stop TX DMA! // DMA failed to stop in 10 ms AR_CR=0xffffffff AR_DIAG_SW=0xffffffff DMADBG_7=0xffffffff and then keeps on going like that forever 21:53:57 i'm reasonably sure this is far beyond my ken and just want it to stop :( 21:54:00 https://bbs.archlinux.org/viewtopic.php?id=137643 21:54:03 "There were lots of fixes to ath9k recently. So I suggest you try compat-wireless or the latest kernel, which currently is 3.3-rc7" 21:54:10 maybe that's what your backported driver is, though? 21:54:15 well i can try getting compat-wireless again. 21:54:16 but maybe try the latest kernel anyway 21:54:48 That particular post is from over a year ago. 21:55:08 yeah. i know. i've googled around and gotten lots of stuff like that. 21:55:25 I'm sure. 21:55:29 So did you try a new kernel yet? 21:55:35 lemme see. 21:55:43 give up and buy http://www.amazon.com/Edimax-EW-7811Un-Wireless-Adapter-Wizard/dp/B003MTTJOY/ref=sr_1_1?ie=UTF8&qid=1375826132&sr=8-1&keywords=usb+wifi+adapter 21:55:51 Or maybe did you try the nohwcrypt=1 thing? 21:55:55 except the driver for that might also suck 21:56:08 I do have one like this working very well ,but not with in-tree drivers 21:56:18 could dig up the exact part number and the driver tarball I used, if desired 21:56:32 i have a usb wireless adapter you're free to take if you stop by 21:56:35 i think i tried the nohwcrypt at some point... yes it's still there. 21:57:03 now excuse me while i try to figure out how to get a new kernel through apt 21:57:04 well, maybe that's what's causing the problems!! 21:57:24 Bike: look for a PPA 21:57:38 there's one called kernel-ppa 21:57:45 maybe that's a good one 21:59:12 i've been thinking of moving away from ubuntu anyway, though. i don't use gnome (or other default things like NetworkManager) and all the wiki instructions assume i do, and stuff like that. 21:59:17 my system is really frankenstein and shitty. 22:00:43 i'm glad you don't use gnome and NetworkManager 22:01:01 -!- bicyclidine has joined. 22:01:10 hi bicyclidine 22:01:19 have you considered running your irc client on a server somewhere 22:01:36 Well, here we go, I could just use my phone for everything! V_V 22:01:38 You use irssi anyway, it seems. 22:01:38 -!- bicyclidine has quit (Client Quit). 22:01:51 shachaf: idiot, lazy, etc. 22:02:21 kmc: when unity came out i was still using gnome, and i was all like huh, it's for netbooks? maybe this is for me! and then i tried it and all the 3d gobledegook made it unusable on my integrated graphics card :D 22:02:22 * boily ties ellioerjan to nooodl with hemp rope <-- this is starting to get kinky hth 22:02:39 Bike: That's why I'm telling you to do it! 22:03:09 because i'm a stupid idiot? 22:03:27 Bike: Because you're lazy, so you need encouragement. 22:03:50 oh. well, i'd need a server first. i don't think i'd really use a linode or whatever. 22:05:36 i kind of like just disconnecting when i'm not online instead of making a computer spin when i don't need it. it only gets annoying when i disconnect for stupidly having to restart, and uch. 22:06:30 imo trick someone into letting you use their server 22:06:39 I tried that but elliott is also lazy. 22:06:49 -!- augur has joined. 22:07:12 have you considered all the other possibilities 22:07:32 like the free ec2 thing kmc talks about?? 22:07:56 I think what shachaf is trying to say is that he's got a server that he'd like you to use. ;) 22:10:03 apparently you need a credit card for AWS even if you're only using free stuff. so that's cool 22:10:20 yes 22:10:54 maybe you can trick other people (not me though!!) 22:11:23 like kmc or monotone or Fiora 22:11:25 i don't even know 22:11:40 so i can buy things from amazon without a credit card, but can't not buy things, i guess. 22:13:08 I wonder if you can use a prepaid debit card 22:13:08 -!- augur has quit (Read error: Connection reset by peer). 22:13:24 i buy stuff on amazon with a debit card. 22:13:29 I wonder if you can use someone else's credit card. 22:13:37 oh, you can sign up for AWS with a debit card 22:13:39 I did that 22:13:43 I don't have a credit card 22:13:50 -!- augur has joined. 22:13:56 it's funny how people get really agitated when I say that I don't have a credit card 22:14:05 I don't have one either...? 22:14:06 and start listing all the advantages as though surely I'm just unaware of them 22:14:09 but my debit card is a visa so it's the same thing 22:14:17 well, it's not 22:14:27 they're made of PLASTIC, kmc. you're missing out 22:14:36 you can finally stop bartering with ducks and gooses 22:14:44 I don't have much that a goose wants, anyway 22:14:50 some bread I guess 22:14:55 okay okay, unlike a credit card, the APR is 0% :P 22:16:06 I have a credit card, and the APR is 0%. 22:16:14 -!- Bike_ has joined. 22:16:18 I mostly use my credit card for online transactions. My bank lets me generate one-off numbers, which is handy for beating back my paranoia. 22:16:23 well i gues i can do that. 22:16:31 ooh, one-off credit card numbers? that's kind of cool 22:16:34 (It's been configurated to pay 100% of the thing every month, and it has no monthly fixed payment.) 22:16:45 I mean god knows they've already gotten hundreds of dollars out of that card 22:16:51 -!- Bike has quit (Disconnected by services). 22:16:56 -!- Bike_ has changed nick to Bike. 22:17:04 maybe at school i'll get a bit of server space to use 22:17:24 I would host Bike's IRC except I don't trust myself to not abuse my root powers 22:17:25 Only noticeable issue: they keep sending letters every now and then about how it'd be much more "flexible" if I changed the minimum-payment amount to something that's less than 100%. 22:17:28 It's also handy for when stored numbers get compromised. As has happened to me before. 22:18:05 elliott: one time at a party a friend and i swapped ircs for the hell of it 22:18:32 also i'm actually unsure of what i would use a credit card for 22:19:20 Running up massive quantities of debt? 22:19:28 i guess if i wanted to pay for something before i had a paycheck? 22:20:20 Well, depending on the card you can get little niceties like cash back on certain transactions or whatever. 22:22:57 -!- Bike_ has joined. 22:24:35 -!- Bike has quit (Ping timeout: 245 seconds). 22:25:32 -!- Bike_ has quit (Client Quit). 22:25:50 -!- Bike has joined. 22:26:51 -!- Nisstyre has quit (Quit: Leaving). 22:29:20 so, this is cool. if i call a number on the landline the internet cuts out, but if i call on my cellphone - which, out here, uses the wi-fi - internet's fine. 22:29:24 is it possible i'm in robot hell? 22:30:08 Bike: did you get those dsl thingies 22:30:57 no 22:31:08 do you really not have filters on your phone sockets? 22:31:15 those come with the modem I think 22:31:49 i think i'm just o bad with computers.. so bad. in twenty years i'm going to be like the CS professor who doesn't know what version control is 22:32:40 Bike: imo get the filters 22:32:47 otherwise you might get internet cancer 22:32:50 i'm moving out this week. 22:32:52 (well you get that anyway) 22:32:54 Oh. 22:32:56 To CA? 22:33:07 Northern CA. The part not in CA. 22:33:08 yeah, the filters are important because diesel internet makes a lot of toxic fumes 22:33:45 Bike: "who needs landlines anyway" -- norwegian telenor 22:33:46 ooh where in northern ca? 22:33:48 Bike: Oh, no, Medford, OR? :-( 22:33:56 slightly north of medford 22:34:13 i'm not actually moving to california. i am moving across the state i am already in, which is not california. 22:34:31 I thought you were in WA. 22:34:44 Anyway, don't move to Medford. 22:36:03 can do. 22:36:52 How slightly north? 22:36:54 i used to live in Medford, MA 22:37:53 north of oregon. 22:38:30 are you moving to canada 22:38:53 There's still one state left in between Oregon and Canada. 22:39:34 Bikes of Canada 22:40:08 which is where Bike lives now, and also where i lived for almost a decade 22:40:51 you got me. i'm moving to iqaluit 22:41:48 Say hi to a polar bear for me. 22:52:18 moving to ᐃᖃᓗᐃᑦ 22:53:07 very scenic 22:53:29 kmc: does Rust have any OpenGL bindings (should I just google this) 22:54:02 "Due to Newfoundland's high debt load, arising from World War I and construction of the Newfoundland railroad, and decreasing revenue, due to the collapse of fish prices, the Newfoundland legislature voted itself temporarily out of existence in 1933 in exchange for loan guarantees." 22:54:07 elliott: yes we're using them for Servo 22:54:19 kmc: oh! should have thought of that 22:54:24 interesting 22:54:40 do you use the fancy programmable pipeline with shaders and all that? 22:54:47 https://github.com/mozilla-servo/rust-opengles https://github.com/bjz/glfw-rs 22:54:49 uh i don't think so 22:54:58 -!- nooodl has quit (Ping timeout: 256 seconds). 22:54:58 you kind of have to with OpenGL ES? 22:55:12 i dunno man 22:55:17 well I mean, do you use the deprecated glBegin type stuff 22:55:24 I would assume not since servo is meant to be all GPU-y isn't it 22:55:24 ddon't think that exists in ES 22:55:30 right 22:55:31 What is "Calcutta (Taxi, Taxi, Taxi)"? I have a cover of this which is rewritten for 2A03+MMC5 soundchips, by Archj. 22:55:41 our drawing is behind several layers of abstraction though 22:56:12 this binding looks somewhat incomplete 22:56:24 ``patches welcome'' 22:56:25 ​/home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: `patches: not found 22:56:26 or maybe OpenGL ES is just tiny?? 22:56:56 kmc: do you have to use the... servo fork of the rust compiler... for this 22:57:02 (why does servo have its own rust) 22:57:08 it's not a fork just an old version 22:57:25 cool servo uses hubbub? 22:57:26 because the Rust language changes every day and we would like not to spend all day every day porting Servo to today's version of Rust 22:57:32 I started working on Haskell bindings to hubbub 22:57:33 instead we spent a few days every month on that 22:57:37 that's what I'm doing right now actually 22:57:38 gave up when I had to modify one of its dependencies to be thread-safe 22:57:46 well really what I'm doing is slamming my head into a brick wall due to lack of fundeps, but in theory 22:57:47 (their interning library or whatever) 22:58:00 what's hubbub 22:58:04 Bike: boring 22:58:07 (HTML5 parser library) 22:58:07 gotcha 22:59:04 incredible worldwide gangster computer god secret containment policy 22:59:11 http://24.media.tumblr.com/b7c651eba14e9a34af8a5eeab58c040f/tumblr_mr4rbolZZt1s5f5n9o1_500.jpg hey good news, i found a worse version of that xkcd comic. 23:02:54 great 23:03:15 shachaf: maybe I should give up on real polymorphism and use macros 23:04:00 kmc: imo implement polymorphism using macros 23:04:14 well yah 23:04:22 What do you need fundeps for, anyway? 23:04:44 fun 23:04:56 no but actually https://github.com/mozilla/servo/blob/master/src/components/util/tree.rs#L88 23:05:30 there should be a 1-to-1 relationship between N and NR but sometime in the past month the compiler realized that it has no reason to assume there is one 23:05:36 and I have no way to tell it that there is one 23:05:56 it upsets me how rust does this privilegign a singel object thing 23:06:02 like its "existentials" too 23:06:04 *privileging *single 23:06:04 yep 23:12:07 http://sourceware.org/bugzilla/show_bug.cgi?id=13271 drepper is fucking amazing 23:12:54 i'm not sure general existentials make that much sense in rust......... 23:13:30 does rust include what you're closing over in the type of funptrs 23:14:52 a little 23:15:39 you can write &fn: Clone(x: int) -> char 23:15:56 I don't understand that syntax :/ 23:16:07 the ': Clone' bit says that all the free variables need to implement the Clone trait, i.e. be copyable 23:16:16 I see 23:16:35 you know the thing ski always talks about with -># ? 23:16:38 do you know the "a -> b = exists s. (# s, (# s, a #) -># b #)" thing 23:16:40 haha 23:16:41 yeah 23:16:42 way ahead of you shachaf 23:16:50 people brought that up in #rust actually 23:16:50 I was wondering if Rust had anything corresponding to -># I think 23:16:54 yes 23:17:01 extern "Rust" fn(x: int) -> char 23:17:06 ew (but okay) 23:17:13 is guaranteed to have no free variables and is represented by a raw code pointer 23:17:18 (but not with the C calling convention) 23:17:29 it uses the \rainbow{Rust calling convention!} 23:17:42 how are closures represented with this anyhow 23:17:56 elliott: you mean way behind hth 23:18:05 I don't know exactly how closures are represented in Rust 23:18:11 way ahind of you shalliott 23:18:51 you can have @fn (managed box with free variables in it) and &fn (borrowed pointer to somebody else's box of free variables) 23:19:00 ok next pointless question, do things that are usually uh stack-allocated (like fixnums) implement Copy 23:19:11 if you call a function and use a lambda expression for one of the args, then that'll stack-allocate free variables and pass an &fn, I think 23:19:17 in particular I'm broadly interested in a languages where you actually construct closures with that existential construction 23:19:27 i like how awodey always calls his category C and an object in that category C 23:19:49 and where the notion of references is advanced enough to allow denoting read/write access to variables in the context as elements of that tuple 23:19:59 and I guess there's also ~fn, where the function is uniquely owned and so are all its free variables 23:20:08 elliott: any chance you could walk me through that type signature 23:20:09 that's what the task-spawning function takes 23:20:49 (sort of a pet vapourware project of mine -- a functional language with things like linear types and this -># thing etc. that has full memory control (and no automatic hidden allocation or GC, etc.) but still has all the goodies like higher-order functions, without requiring turning everything into a pointer or closure etc.) 23:20:59 (I think it overlaps heavily with Rust, but I started thinking about it before I knew of Rust) 23:21:11 yeah that sounds similar 23:21:38 Bike: the one with a lot of #s? 23:22:01 yeah 23:22:04 Bike: (# a, b #) is an unboxed tuple of a and b, i.e. (# a, b #) -> c literally just takes "two arguments", there's no pointers allocated etc. in the construction of the unboxed tuple 23:22:16 right 23:22:24 Bike: (exists s. ...) is ... with a hidden type "s" that the user of the value doesn't know what it is -- the person who constructs the value chooses it 23:22:34 (doesn't exist with that syntax in Haskell but you can construct the same thing) 23:22:40 is s the free variables? 23:22:44 also there's ~once fn which is only called once, so after it's been called you get the owned free variables back???? tbh dont understand this one 23:22:44 right, it binds s 23:22:50 (a -># b) doesn't exist, but it's basically a raw function pointer from a to b 23:22:58 ahem FunPtr a b 23:23:02 like, say the syntax is (\# x -> ...) :: (a -># b) 23:23:09 in ..., you can't use any variables in the context 23:23:09 er is it FunPtr (a -> b) w/e 23:23:10 ok i was just missing what s was, i think, thanks 23:23:13 because you can't in C 23:23:23 so let's say you wanted to do 23:23:30 foo :: Int -> (Int -> Int) 23:24:09 foo = (# (), (\# ((), m) -> (# m, (\# (m', n) -> m' + n ) #) ) #) 23:24:17 this is hideously ugly but hopefully you can see what's going on 23:24:23 inside the inner \# you couldn't reference m 23:24:27 you can to explicitly close over it 23:24:30 the corresponding C structure is 23:24:43 struct closure { void *data; B code(void *data, A param); }, basically 23:24:47 Yes it is sensible to me. 23:24:55 where you do closure->code(closure->data, param) to call it 23:24:58 that is really fucking ugly 23:25:01 just you get fancy type magic here 23:25:16 but yeah that makes sense, i'm used to closure = raw ptr + some vector 23:25:23 Bike: well then what you do is you introduce (\x -> ...) as sugar for (# closure, \# (# closure, x #) -> ... #) 23:25:28 where closure is all the local variables you use in ... 23:26:00 right, right 23:26:18 in kernel magic idea i should really work on i was thinking of having $vau work like that 23:26:37 also has the convenience that the "raw pointer" is obviou 23:26:37 s 23:27:10 i think you could do lambda lifting pretty nicely that way 23:29:04 that is, uh, you make (\x -> ...) just (\x -># ...) when possible, obviously. 23:35:07 -!- Sgeo has joined. 23:36:44 -!- KingOfKarlsruhe has quit (Quit: ChatZilla 0.9.90.1 [Firefox 22.0/20130618035212]). 23:37:27 kmc: Bike: in particular, I think it would be cool to have a non-existential version of this 23:37:32 where the "s" is actually part of the type 23:37:41 elliott: oh re graphix libraries, look at https://github.com/mozilla/servo/tree/master/src#supporting-libraries 23:37:43 yeah that could be nice 23:38:01 like (Int ->{a :: Int, b :rw: Int} String) 23:38:13 so that closes over two things, read-only Int a and read-write Int b 23:38:20 and at runtime it'll look like e.g. 23:38:22 so if you had a function that took a thing-that-closes-over-1-Int you could vector pass the number and the ptr in some situations, or something 23:38:37 struct this_closure { int a; int *b; string (*foo)(int a, int *b, int param); } 23:38:39 elliott: I think azure calls skia and skia calls opengles??? 23:38:41 don't know why 23:38:49 kmc: will look after finishing rambling 23:38:59 anyway the idea then that higher-order functions would be written *polymorphic* to this closure part 23:39:17 so that things don't necessarily have to get bundled up in a closure to be passed higher-order at all 23:39:19 mostly i got irrationally annoyed once i realized after far too long that "returning a closure" allocated memory same as actual alloc calls, but looks exactly like returning a bare function 23:39:24 and you can have control structures that you know get all inlined out and stuff 23:39:28 (skipping even the need for a funptr) 23:39:46 this sort of stuff is similar to C++ lambdas I think with the kind of control you get 23:39:49 "eventually just optimize everything into IMUL cals" 23:40:18 basically my hidden motivation here is that I want something that looks like Haskell but you can also e.g. write a memory manager to run on the bare metal with 23:40:30 (a strongly-typed memory manager) 23:40:33 that's the opposite of hidden. 23:40:48 well, I just told you. 23:40:52 so now it's no longer hidden 23:40:55 no i mean, it was obvious before. 23:41:11 btw are we getting "mad hintz" of @ here 23:42:20 I would use this kind of stuff in @ yes, but actually it's a bit less vapourware than @ 23:42:21 the other thing that would be nice is if you could use this as a jit, which moves me into full no idea what i'm talking about territory 23:42:33 in that I can imagine trying to implement it if I got a good enough grip on the design and the design looks vaguely grippable 23:42:50 what is a hidiom 23:43:11 Bike: though the language I'd want for @ would be basically fully dependently-typed. this thing is only kind of dependently-typed. 23:43:35 elliott: er isn't everything you've said so far independent of dependent types? 23:43:43 yes 23:43:55 then i mean... why say that. 23:44:01 well you asked about mad hintz of @. 23:44:08 the sorta dependent types come in when describing memory layout, in this language. 23:44:18 "though the language i'd want for @ would have full complex arithmetic. this thing doesn't have any complex arithmetic" 23:44:22 haha 23:44:27 wait, memory layout? how? 23:44:39 Bike: well uh 23:45:01 man i'm so bad at PLT, i still don't get how dependent types are different from like, integral range types 23:45:06 I should start reporting bugs against Pharo 23:45:08 Bike: do you know the kind of bullshit structure layouts you get in OSes and networking 23:45:11 and when i say "PLT" i mean "anything in this channel". 23:45:12 -!- yorick has quit (Remote host closed the connection). 23:45:15 elliott: kinda. 23:45:29 There are just so many bugs :( 23:45:41 like addrinfo? 23:45:55 Bike: ok well take a look at the second and third pictures here http://wiki.osdev.org/Global_Descriptor_Table#Structure 23:46:03 note how base and limit are like 23:46:08 sprawled across multiple words 23:46:12 in random order 23:46:34 oh. yeah. fuck those 23:46:36 and this is awful to deal with because you have to do the bitpacking yourself etc. 23:46:47 so I'd like a way to first specify precise bit alignments etc. of structures 23:47:00 and then say "there's a base field made by bitpacking these other precisely-sized smaller-than-word fields together" 23:47:05 and then just use that everywhere. 23:47:21 what does that have to do with dependent types 23:47:24 and then I also want to be able to e.g. represent a TCP packet directly as a structure within the language, with the correct in-memory layout 23:47:28 Bike: I'm gettting theeeeeeeeeeere 23:47:51 so uh you know how ADTs are generally represented with a tag word 23:47:51 like 23:47:55 data Maybe a = Nothing | Just a 23:47:59 you store a word for whether it's Nothing or Just 23:48:00 yeah 23:48:08 ok i didn't know it was a whole word though? 23:48:12 well, details 23:48:18 actually GHC does it a bit different It hink but w/e 23:48:27 and that kind of thing also comes up explicitly in a lot of protocoly type stuff 23:48:35 like "if this flag is this then it has this layout otherwise it has another"? 23:48:40 oh. yeah, i see. 23:48:43 so I'd like to be able to say (syntax made up on the spot) 23:48:57 struct Maybe A { enum { Nothing, Just } tag; switch (tag) { Nothing: ; Just: A value; } } 23:49:17 and then the sorta-dependent-typing is the rest of the structure type depending on the value of the tag. 23:49:44 that's... uh. 23:49:44 then you could do things like switch on tag, and if it's Just you can access .value safety, but if not you can't access .value (without explicitly marking it as an unsafe access, perhaps) 23:49:58 the point is that this generalises to other stuff 23:50:04 where it's not as simple as just one tag value 23:50:12 (obviously you can build sugar for the common ADT-y case on top of this) 23:50:15 yeah i'm imagining some weird ass "switch"es here 23:50:23 the point is that the existence of a tag word in memory is explicitly represented in the language 23:50:35 because the idea is not to hide anything about the memory layout 23:50:58 hm, well, doesn't that possibly inhibit optimizations that work by eliding the tag word when possible? 23:51:34 I'd say it more enables you to make optimisations by not having a tag word in the first place 23:51:40 or packing the tag word as bits in another part etc. 23:51:49 and e.g. avoiding casing on the tag for data shared between all variants of a struct 23:51:58 more broadly, optimisations that change in-memory layout aren't appropriate for something like an OS 23:52:02 (most of the time) 23:52:08 granted. 23:52:28 (though I think this language would also apply nicely to game development and similar things, with the GC-less nature) 23:52:48 i was thinking of i dunno, \x -> let fn = ... in case x of Nothing -> foo; Just _ -> fn x where it can obviously be seen that fn will always get a Just 23:53:10 right, I guses 23:53:20 but yeah i can see why you wouldn't want that in an OS 23:53:28 though an example that isn't a bad idea would help (since what is fn Nothing going to be? you could just pass the value inside the Just) 23:53:45 off the top of my head and i'm bad at programming, sorry~ 23:53:50 but also I don't think compilers generally do this kind of optimisation in practice, so losing it doesn't seem like such a big deal 23:56:32 kmc: Rust does something similar with forcing you to case to use things safely, right? 23:56:33 -!- iamfishhead has quit (Ping timeout: 264 seconds). 23:58:13 I didn't read this whole convo 23:58:15 so, what? 23:58:45 uh I don't know. isn't there a thing where you can't access fields of an object in Rust before you check a tag or something 23:58:51 I don't know how to describe what I am thinking of 23:58:59 that sounds like pattern matching 23:59:05 and you are required to check all cases 23:59:57 -!- doesthiswork has joined.