00:00:22 what about it? 00:01:44 hm who can we hack in greenland 00:02:45 -!- augur has quit (Ping timeout: 252 seconds). 00:10:41 -!- sirdancealot7 has joined. 00:11:36 -!- sirdancealot7 has quit (Read error: Connection reset by peer). 00:22:00 -!- impomatic has quit (Quit: http://corewar.co.uk). 00:29:32 -!- sirdancealot7 has joined. 00:33:45 I have added specification of ICO and of a few of the ETC/8000???? 00:39:19 -!- c00kiemon5ter has joined. 00:53:48 -!- itrekkie has quit (Ping timeout: 245 seconds). 01:02:24 -!- Phantom_Hoover has quit (Remote host closed the connection). 01:10:12 -!- oerjan has quit (Quit: leaving). 01:22:35 -!- augur has joined. 02:09:58 -!- TeruFSX has joined. 02:11:53 When I google for newspeak nof, I do not expect Google to change nof to of without telling me explicitly 02:13:01 stick it inside quotes 02:15:09 I did, that works, but I do expect at least a "Showing results for of; Search instead for nof" 02:27:22 I don't think I like Actors 02:27:35 is that the Scheme thing or what 02:27:37 Or at least, I don't like having a single mailbox for each actor 02:27:43 Bike, most well known in Erlang 02:27:59 oh, that. 02:29:25 -!- tromp has quit (Ping timeout: 260 seconds). 02:31:28 -!- Bike_ has joined. 02:31:34 So is there an Erlangtalk here? 02:31:47 -!- Bike has quit (Disconnected by services). 02:31:52 -!- Bike_ has changed nick to Bike. 02:32:17 moar bikes! 02:34:11 -!- zzo38 has quit (Remote host closed the connection). 02:40:14 bah the old newspeak forums are dead and I can't see what I posted 02:40:21 Other than that I posted two threads 02:41:03 I searched for newspeak sgeo 02:41:09 "Did you mean: newspeak good " 02:41:36 -!- evincar has joined. 02:42:35 google likes you! 02:45:54 -!- Arc_Koen has quit (Quit: Arc_Koen). 02:47:16 kmc, you've done Newspeak stuff? 02:47:28 At least, there's a Keegan on BitBucket who has 02:48:21 I went to Google today and was under NDA but they didn't tell us anything that isn't public. :( 02:48:49 lol "SqueakNOS is back! - 2006-05-16 - gera 02:48:49 " 02:49:27 not me 02:49:54 I did run a forkbomb on one of the the Chromebooks they had on display though. 02:50:00 well played 02:50:17 hmm, how does that work? Is there a way to access a console? 02:50:34 Oh, or some Javascript that crashes the brows... wait, hmm 02:50:41 There is an underpowered console which runs in the browser. 02:50:54 you can forkbomb IE6 from the address bar 02:50:57 And has access to some native things via Chrome. 02:51:01 So yes, JS. 02:51:14 about: 02:51:16 or something like that 02:51:32 heh 02:51:41 does navigating to a javascript: url not work? 02:51:45 that might work too 02:52:08 this is a dim memory from approximately a billion years ago 02:52:45 Also the keyboard is shitty. 02:53:12 Capslock is replaced with a super key with a search icon on it. 02:53:29 And there's no menu, home/end/pgup/pgdn/etc. 02:53:43 Not even by way of a function key. 02:54:14 damn it intel why do you have 38573645 different instructions named "shuffle" 02:54:24 Then again I'm one of the 5 people left in the world who still uses a menu key. 02:54:41 evincar: I use it by mistake sometimes, and intentionally a bit less often 02:55:21 keyboard keys ftw 02:55:24 don't take my keys! 02:55:30 I have it bound to WM stuff, so I don't really use it. 02:55:39 i really hate how dell or whoever rearranged the six pack 02:55:44 why would you do that? 02:55:54 i can understand removing that whole chunk of keyboard, but not just destandardizing it 02:56:18 One rationale I've heard from a Logitech guy is that "Insert" confuses people. 02:56:26 So they use a double-high Delete. 02:56:28 fuck people. 02:56:38 people can learn 02:56:47 But still put Ins above Del on the same row as the F keys. 02:56:56 i've never used sysrq but you don't see me complaining! 02:56:57 ;) 02:57:04 Causing Pause/Break and Scroll Lock to get out of whack. 02:57:13 -!- Bike_ has joined. 02:57:14 It's this terrible chain reaction. 02:57:34 Yeah that. SysRq. 02:58:29 -!- TeruFSX has quit (Read error: Connection reset by peer). 03:00:17 -!- Bike has quit (Ping timeout: 248 seconds). 03:01:55 -!- evincar has quit (Ping timeout: 244 seconds). 03:03:23 -!- augur has quit (Remote host closed the connection). 03:03:57 -!- augur has joined. 03:04:37 -!- augur has quit (Read error: Connection reset by peer). 03:04:48 -!- augur has joined. 03:05:43 -!- augur has quit (Read error: Connection reset by peer). 03:05:53 -!- augur has joined. 03:07:19 -!- Bike_ has changed nick to Bike. 03:10:25 -!- augur has quit (Ping timeout: 248 seconds). 03:11:49 -!- augur has joined. 03:19:59 kmc: The good one is pshufb. 03:20:21 Or was it vpshufb? 03:22:27 i don't know what the difference is 03:22:37 sometimes the mnemonics differ between intel and AT&T :( 03:22:46 do they really 03:22:58 -!- augur has quit (Remote host closed the connection). 03:23:12 think so 03:23:16 Yes. 03:23:22 Isn't v the one that takes three operands? 03:23:29 New in AVX. 03:23:30 -!- augur has joined. 03:23:38 (I knew once.) 03:24:25 https://en.wikipedia.org/wiki/VEX_prefix 03:24:26 I think it's that. 03:24:39 -!- augur has quit (Read error: Connection reset by peer). 03:24:50 -!- augur has joined. 03:29:11 oh 03:29:15 ok 03:29:49 i love x86 instructions 03:31:22 kmc: I hear the "cool byte shuffling instructions" are Knuth's MOR and MXOR. 03:33:58 yes 03:34:00 those are nic 03:34:01 e 04:00:32 -!- TeruFSX has joined. 04:02:04 -!- augur has quit (Remote host closed the connection). 04:02:39 -!- augur has joined. 04:05:09 kmc: Oh, I missed it. 04:05:17 * shachaf doesn't track megaseconds. 04:05:21 -!- augur has quit (Read error: Connection reset by peer). 04:05:45 -!- augur has joined. 04:06:19 -!- augur has quit (Remote host closed the connection). 04:06:53 -!- augur has joined. 04:07:50 -!- WeThePeople has quit (Quit: Leaving). 04:07:52 At least I noticed kmc over 9000 day! 04:11:11 -!- augur has quit (Ping timeout: 256 seconds). 04:14:29 ? 04:15:58 kmc, you are over 9000 units of some dimension 04:30:42 «We're grateful for Greg's many contributions to Sequoia over the last 12 years. His wikipedic knowledge, quick wit, and uncanny ability to connect seemingly unrelated ideas made him a joy to work with.» 04:31:04 Is that a sort of backhanded compliment? 04:31:24 Wikipedic knowledge sounds like a bad thing. 04:31:41 it is :( 04:35:45 that's definitely an insult 04:35:51 i don't know how they got away with that 04:36:11 what's the difference between MOVAPS and MOVDQA 04:36:40 kmc: 3 04:36:40 how many jokes are there about intel's mnemonics not being very mnemonic 04:37:50 kmc: http://stackoverflow.com/questions/6678073/difference-between-movdqa-and-movaps-x86-instructions 04:38:04 i wonder if loongson has official han character mnemonics 04:38:08 probably not but that would be amazing 04:38:20 thanxchaf 04:38:30 'On some (but not all) micro-architectures, there are timing differences due to "domain crossing penalties"' 04:38:37 ok now you're just making shit up 04:38:54 is that term especially made-up-sounding...? 04:39:09 no it's just more complexity than i want to think about 04:39:19 welcome to assembly? 04:39:22 MOVAPS two cycles slower due to transfer at Oxford Circus 04:40:18 What are you doing with x86 instructions? 04:40:29 AES-NI 04:40:57 This isn't for mosh, is it? 04:42:06 no 04:42:18 What is it for? 04:42:54 Oh, those things you were working on? 04:48:49 oerjan: Don't you ever get tired of logreading? 04:50:27 -!- augur has joined. 04:52:01 monqy: you should draw a self portrait of oerjan 04:52:58 i remember when i drew self portraits 04:53:09 they were so good 04:53:17 -!- augur_ has joined. 04:53:22 imo some of your finest work 04:53:37 -!- augur has quit (Read error: Connection reset by peer). 05:00:52 -!- monqy has quit (Quit: hello). 05:22:35 "Are you proud when you explain this [primitives] to a non-Smalltalker?" 05:22:47 (On a presentation about Newspeak) 05:22:53 Referring to Bike's objection 05:25:59 Security vulnerability in the pure-Haskell implementation of TLS! 05:26:01 How unexpected. 05:26:12 gosh i'm like some kind of smalltalkkin 05:26:28 shachaf: No wai 05:29:38 shachaf: haha 05:29:40 what's the vuln? 05:31:58 -!- zzo38 has joined. 05:32:43 https://github.com/vincenthz/hs-tls/issues/29 05:33:58 :( 05:34:51 I'm not sure why Haskell people seem to insist on implementing everything themselves. 05:35:12 Implementing crypto by yourself – what could possibly go wrong? 05:35:55 shachaf: To me "mathematically correct", maybe. 05:35:58 software is automatically better if written in haskell 05:35:59 s/me/be/ 05:36:02 that said, a lot of language communities do this 05:36:11 With cryptography? 05:36:16 yeah 05:36:21 I don't see Ruby people implementing their own cryptography things so much. 05:36:24 you can find pure python implementations of everything 05:36:26 Maybe I'm just out of touch. 05:36:32 it's a question of whether people actually use it or whether they're toys 05:36:33 kmc: Sure, but are they *standard*? 05:36:58 and there is a certain appeal to having your security critical primitives written in a high level memory safe language with some level of static assurance 05:37:12 is hs-tls standard? what do you mean by that 05:37:14 is it in the Platform? 05:37:32 I mean the things that all the web frameworks etc. seem to use. 05:37:57 I don't think there's much of a high-level cryptography API that's made over bindings to OpenSSL, say. 05:38:54 hs-tls is used by all the web frameorks? 05:39:35 not clear your web framework should even be doing SSL 05:39:35 I don't mean hs-tls specifically, but most of the cryptography libraries that I've come across. 05:39:45 terminating SSL at a frontend reverse proxy / app container is common 05:40:26 to be fair OpenSSL is written in an unsafe language and is also shitty 05:40:34 i mean, somebody has to try to do better 05:40:39 True. 05:40:39 but they are going to get burnt repeatedly 05:40:43 especially if they are not experts 05:40:46 and SSL is fucking complicated 05:40:49 i'm so glad Mosh doesn't use DTLS 05:40:52 this came up again today 05:44:29 -!- TeruFSX has quit (Ping timeout: 255 seconds). 05:45:30 To some, deism means that God created the universe but does not affect it now. When I said that I don't believe the Jews are God's chosen people and all that stuff is just mythology, I did not mean that God does nothing since. God is beyond comprehension and now you are trying to put it into something you understand, even though it cannot understand, so that is an error. 05:45:50 zzo38: Would you say that everything exists? 05:45:55 shachaf: Yes. 05:46:03 What Wikipedia says about deism is different. 05:46:13 And I think that God is a hypothesis with near-zero likelihood. 05:46:58 People don't even agree what "God" is, regardless if you think so or not. (This applies to me too, of course.) 05:47:15 pikhq: imo there is a very high likelihood for my existence 05:47:21 near 1 in fact 05:47:49 shachaf: However, although I may say that everything exists, I now say, what is exist? Physical existence? Universal existence? Mathematical existence? It is different! 05:48:44 thinking that questions about god are stymied by people not agreeing on what "god" means is called "ignosticism", for reference 05:49:12 Yes, I know that is called "ignosticism"; I did see that on Wikipedia. 05:49:28 So I do sometimes call some of my ideas ignostic, too. 05:51:42 I do study philosophy of much. 05:51:58 `addquote I do study philosophy of much. 05:52:06 933) I do study philosophy of much. 05:52:28 kmc: lens now has a function which pays a 1000x constant factor for the common case to avoid a space leak in the unusual case. 05:52:33 :( 05:52:34 I feel kind of bad about that. 05:52:49 But I don't know what the right answer is. 05:52:53 Space leaks are really annoying. 05:53:05 shachaf: Can you perhaps fix it, or make up a separate function for the unusual case? 05:53:06 You can profile a slow thing, but a space leak will just freeze your computer. 05:57:29 I am not a humanist religion. Humanity is just one of the life on the Earth, the Earth is just one of the planets in the solar system, and so on. The absolute reality is the universe as a whole; we divide it to understand it in different ways. 05:57:53 zzo38: You should fix our issue with the function! 05:57:55 The function is lastOf. 05:57:59 It gets the rightmost element of a tree. 05:58:40 I may look, but I don't even know exactly how to fix it necessarily. 05:59:24 https://github.com/ekmett/lens/issues/245 06:03:17 Here's an idea for a FUSE filesystem: Just overlay over another filesystem, presenting all targets from Makefiles as pseudofiles that will be created on-demand. 06:03:25 Smalltalkers are utterly utterly insane http://vimeo.com/50530082 06:06:27 this is silly. 06:06:43 I do not like the way FUSE works; I would prefer it to continue running while active rather than terminate, and to be accessible in a subdirectory of the process directory (in /proc/) instead of the other one. 06:06:48 http://netjam.org/quoth/ 06:07:03 Search that page for Play 4 times 06:07:03 (And you could create a link if you wanted it in a separate directory) 06:07:21 zzo38: IIRC FUSE filesystems don't terminate, they just daemonize. 06:30:36 -!- FreeFull has quit. 06:38:35 -!- aloril has quit (Ping timeout: 255 seconds). 06:51:56 -!- aloril has joined. 07:04:04 -!- Bike has quit (Quit: dead). 07:07:12 -!- ogrom has joined. 07:11:04 -!- sebbu2 has joined. 07:11:22 -!- sebbu2 has quit (Changing host). 07:11:22 -!- sebbu2 has joined. 07:14:51 -!- sebbu has quit (Ping timeout: 260 seconds). 07:24:48 shachaf: it's so lame that system calls have to use memory to transfer data 07:24:55 i want to read 16 bytes from a file directly into %xmm0! 07:26:03 Making an asynchronous interface for that sounds like fun. 07:28:14 The OS is going to read it into the page cache anyway unless you do something weird, so you might a well just mmap it and copy it from memory or something. 07:28:30 I guess you want the disk controller to copy data straight to a register too? 07:28:35 You should patent DRA before someone else does. 07:28:36 what if it's a terminal 07:28:37 heh 07:28:54 Oh, a terminal. 07:29:20 terminals: best type of file? 07:29:51 > compare "terminal" "disk" 07:29:53 GT 07:29:56 Looks like it. 07:33:48 sleep, 'night all 07:43:17 kmc: movdqantfs 07:57:10 Fiora: Do you like the guessing game "x86 instruction mnemonic or CAPTCHA?"? 07:57:32 XD 07:57:58 I think my favorite insane one is still phminposuw 07:58:41 Pft. That one's almost pronounceable. 07:59:04 packed horizontal minimum position unsigned word 07:59:37 You should get MOR and MXOR into x86. 07:59:41 take a vector of 8 uint16_ts, return {minimum value, position of minimum value} 08:00:10 single-cycle (throughput at least) 8-way minimum~ it's fun, the name is just ridiculously long XD 08:00:14 ... MOR? 08:00:15 PUNPCKLQDQ and such make for funky names, too, due to how they dropped the 'a' from 'pack'. 08:00:34 pun-pack-liq-diq 08:00:39 If only they'd gone all out with vowel-dropping and made it PNPCKLQDQ. 08:00:47 pun-pack-liq-bow 08:00:48 It's not "pack" when it's PCK. 08:00:57 Also "liq-diq" sounds dirty. 08:01:11 .... oh geez it does >_< 08:01:59 MOR is great. 08:02:15 what's MOR? 08:02:17 It does matrix multiplication of two 64-bit registers treated as 8x8 matrices. 08:02:21 .... @_@ 08:02:23 what architecture is that? 08:02:24 Using AND and OR instead of + and * 08:02:27 MMIX. 08:02:35 wait, so, each entry in the matrix is 1-bit? 08:02:39 Yes. 08:02:44 wow I think that could actually be used in some FEC code I saw a while ago 08:02:51 It's pretty nice. 08:02:56 that's a crazy instruction, was it fast? 08:03:07 MMIX has never been built, so who knows? 08:03:55 oh... I guess I was confusing it with something else 08:04:09 ohhhh. it's Knuth's thing 08:04:13 Yes. 08:04:40 * Fiora tries to think of odd simd instructions she's seen 08:05:00 I think NEON has a simd leading-zero count. which isn't really that weird I guess 08:05:08 MOR/MXOR give you pshufb, among other things. 08:05:18 Except instead of shuffling you can or/xor bytes. 08:05:41 (So instead of a byte index like 0,1,2,3, you give a mask like 1,2,4,8 for the byte you want.) 08:05:48 oh apparently there's an FPGA implementation of MMIX o_O 08:06:02 There is? 08:06:15 I didn't hear about it. 08:06:22 Maybe MMIX is the future of computing. 08:06:27 http://en.wikipedia.org/wiki/MMIX#Hardware_implementations 08:06:51 FPGA isn't really hardware, is it? 08:06:54 Oh, it doesn't say that. 08:08:22 -!- carado has joined. 08:09:57 Maybe I should write an MMIX emulator to learn about writing emulators. 08:14:01 -!- evincar has joined. 08:21:25 GCC even can compile into MMIX. (LLVM cannot) 08:21:42 wait, really? XD 08:23:24 Oh, that's exciting. 08:24:25 -!- epicmonkey has joined. 08:24:26 Is there a Linux port? 08:24:37 What's arguably more funky is that it's part of mainline GCC distribution, and not just someone's patchset. 08:25:39 -!- ais523 has quit. 08:26:53 http://gcc.gnu.org/onlinedocs/gcc/MMIX-Options.html and so on. 08:29:31 Apparently there is a Linux port. 08:35:56 I think MMIX is good but I don't like the rN register and think it should be replaced. 08:36:34 Remind me what that register is? 08:37:23 -!- epicmonkey has quit (Ping timeout: 255 seconds). 08:37:51 Read-only register specifying the serial number. 08:38:47 Oh. 08:39:10 Yes, it should be replaced with a register that specifies the National Identity Number of the current user. 08:41:36 No, it should probably be replaced with something more useful, such as a hypervisor register. 08:42:28 i love registers 08:42:30 they are so easy 08:48:25 -!- evincar has quit (Quit: leaving). 08:49:53 Have I mentioned that De Bruijn sequences are the best thing yet? 08:49:57 (They are.) 08:50:20 those are the cool things you can use for stuff like leading zero count, right? 08:50:23 and population count 08:50:33 Isn't it trailing 0 count? 08:50:38 you can do both, I think 08:50:39 ? 08:50:41 Oh. 08:51:06 That's pretty neat if true. 08:51:19 Anyway I first heard of them in a talk about magic. 08:51:37 As in card magic. 08:52:28 The presentation I saw was, roughly: 08:52:42 Giving 5 people a deck of cards, telling one of them to cut it randomly. 08:52:50 Then have each one take the top card and pass the deck to the next one. 08:52:54 having 08:53:17 Then doing all sorts of mind-reading things and saying "I'm getting a strong sense of RED... Stand up if you have a red card". 08:53:28 Then naming all the cards. 08:53:53 This is the only good "mathematical" magic thing I've ever seen. 08:55:23 I think my favorite magic trick is the one where the magician plays a simultaneous chess game with extremely skilled people 08:55:30 and beats/ties half 08:55:49 by copying the moves of one half of the players against the other half, to play them against each other 08:55:51 I saw that presented as a puzzle, not a magic trick. 08:56:04 really? I think I remember hearing it done by a magician or something 08:56:08 maybe I'm wrong 08:56:18 Magicians will do anything these days! 08:56:25 I first came across it in a computer game. 08:56:44 I once wrote a program that used the same idea to spy on people's conversations on omegle.com 08:56:52 The trouble is, most people's conversations are very boring. 08:57:36 (I assume other people have done the same thing there. Who knows how many relays your messages are going through! That's the trouble with anonymity.) 08:59:47 -!- fizzie has quit (Excess Flood). 08:59:54 -!- fizzie has joined. 09:00:28 Fiora: Anyway isn't that card magic thing great? 09:00:53 it works by putting the cards in a de bruijin sequnece or something? 09:01:12 Yes. 09:01:15 ahhh 09:01:21 You have red/black as bits, so you make a base-2 de bruijn sequence. 09:01:34 If you have 6 people, that's enough for 64 cards. 09:01:45 You can encode the suit in two bits and the value in 4 bits. 09:01:59 He came up with a system such that he doesn't have to memorize the whole sequence, too, he can just work it out. 09:02:12 It's pretty neat how the more people you have, the easier it gets. 09:02:17 huh, cool :o 09:02:20 (With 5 people you can only do 32 cards.) 09:03:26 @brain Are you thinking what I'm thinking? 09:03:27 Troz! 09:03:30 Hrm. 09:03:31 @brain Are you thinking what I'm thinking? 09:03:31 I think so, Commander Brain from Outer Space! But do we have time to grease the rockets? 09:04:20 There are so many neat applications of De Bruijn sequences. 09:04:54 If you have a keypad lock that doesn't require an "end" key, you can come up with the shortest sequence that will try every combination. 09:05:19 * shachaf tries to remember some other neat applications. 09:10:54 I will look it up in Wikipedia so that I can understand it better. 09:11:41 zzo38: Perhaps you should move to California. 09:11:47 That way you would understand it well. 09:11:51 Is certainly a good idea, I think. 09:12:32 Moving to California? 09:13:00 No, I mean De Bruijn 09:13:07 I like it 09:13:17 De Bruijn invented everything good. 09:14:10 No, not everything; I just like mathematics, though. 09:14:58 He invented De Bruijn indices, too. 09:15:05 And other good things. 09:15:55 -!- carado has quit (Remote host closed the connection). 09:18:05 -!- carado has joined. 09:27:16 -!- oerjan has joined. 09:28:01 shachaf: sometimes. 09:28:31 Do you search for your nick or read the whole thing? 09:29:12 both. 09:29:29 @ty both 09:29:31 Applicative f => (a -> f b) -> (a, a) -> f (b, b) 09:29:50 unless it's too damn long or i'm too tired or the goblins in the cupboard are too noisy. 09:29:52 Any opinions on debru ijn sequences? 09:31:27 "something about how they're less repetitive than you". 09:34:14 Ooh, a "de burn-ij" reply. 09:35:44 DARN FINNS STEALING OUR PUNS 09:35:55 oerjan: Are yoerjan calling me repetitive? 09:36:35 Has Norwegian invented puns yet? 09:38:03 of course not, we stole them from the brits, who stole them from the french, who stole them from the romans who stole them from the greeks who stole them from the phoenicians who stole them from the babylonians who got them from ancient aliens and what was i saying about being repetitive again. 09:38:40 (you don't steal from ancient aliens, look how that went for the gomorrans) 09:40:02 ^rot13 stolenpuns 09:40:02 fgbyrachaf 09:40:16 the ancient aliens didn't invent them either, they just got it from time travelers (together with the time travel) 09:42:55 `addquote I searched for newspeak sgeo "Did you mean: newspeak good " 09:43:00 934) I searched for newspeak sgeo "Did you mean: newspeak good " 09:47:16 damn it intel why do you have 38573645 different instructions named "shuffle" <-- are they in random order too? 09:47:47 and they're formatted differently too! 09:47:53 934 quotes, isn't that a bit too many quotes? 09:47:57 shufps, pshufd~ 09:48:27 fizzie: not if they're all good 09:49:03 I bet kmc is working on a super-secret weblog post that we all can't know about. 09:49:15 That's why he's being so secretive. 09:49:43 oerjan: Are they all good? 09:50:17 i dunno but when they do that five quote thing they frequently don't find anything to delete 09:50:19 -!- epicmonkey has joined. 09:50:51 -!- ais523 has joined. 09:52:04 I think 934 quote is not too much, is OK to have many quotations, I think. 09:52:43 -!- zzo38 has quit (Remote host closed the connection). 09:53:35 fungot: Is it okay to have many quotations? 09:53:36 fizzie: ( the native sparc variant, that is and it seems to me that it's lf not crlf 09:53:52 fungot: Not helpful. 09:53:54 fizzie: i feel bad now. that feels weird. riastradh, are you putting the poor parentheses all apart and away from their comrades? 09:54:11 fizzie: I don't see why CPU architectures can't have favoured line endings… 09:54:28 ais523: Not helpful re the amount of quotes, I mean. 09:54:37 ah right 09:55:10 (Also I didn't know fungot even had feelings.) 09:55:10 fizzie: for instance clc-intercal exposed a bug in my code 09:55:32 fungot fungot fungot 09:55:33 shachaf: if we have ( ( rec ( f n d c) ( ( a) it determines what code one uses to access structure members. a functor is a class 09:55:37 help 09:55:49 ais523: Do you know all about subtyping? 09:56:08 shachaf: slightly 09:56:22 mostly in the context of Javaish OO, because I teach Java 09:56:23 I feel like subtyping is way more complicated than not-subtyping. 09:56:30 and am going up to the lecture room in a couple of minutes 09:56:44 And also mutability + subtyping sounds really complicated. 09:56:59 I don't normally teach on Thursdays this term, but I'm covering for someone who hadn't been hired at the time 09:57:00 kmc: Oh, I missed it. <-- wat. 09:57:06 ais523: Did you know Java has covariant arrays?! 09:57:11 Mutable arrays, I mean. 09:57:14 Pretty crazy, huh. 09:57:14 shachaf: yes 09:57:24 oerjan: ? 09:57:39 shachaf: * shachaf doesn't track megaseconds. 09:57:42 shachaf: and the fact that they only have the one sort of variance catches people out 09:57:57 They only have one sort of variance? 09:58:11 oerjan: Oh, it was megasecond 1,359. 09:58:11 mutable arrays are supposed to be invariant. 09:58:16 `run date +%s 09:58:17 1359021496 09:58:43 what's significant about that... 09:58:43 kmc used a secret type of message not visible to glogbot to say it, though. 10:00:06 Every megasecond is significant. 10:00:21 There's only so many of them left, after all. 10:00:47 `frink megasecond in days 10:00:56 2194560000 m s^2 (unknown unit type) 10:01:01 ...wat. 10:01:10 `frink Ms in days 10:01:17