00:01:25 oh, happy 2008 00:02:12 -!- Traveler2 has joined. 00:02:27 -!- pikhq has quit (Read error: 104 (Connection reset by peer)). 00:03:06 -!- Traveler2 has quit (Client Quit). 00:16:32 Î§Î§Î Î™Î™Î™ 00:16:47 not for another 7 hours 00:16:59 no 00:17:01 it's 00:17 00:17:01 :P 00:17:05 in gmtland 00:17:06 no 00:17:09 it's 17:17 00:17:20 (getchar()+1 || 1) - 1 /* clever trick for a BF-getchar */ 00:22:47 äºŒåƒå…« 00:23:57 hello calamari 00:24:05 hi ehird` 00:25:44 -!- pikhq has joined. 00:26:23 hello pikhq 00:27:08 â€«×ª×©×¡×´×— 00:27:22 damnit haskell 00:28:13 What does this mean? Happy jew year? 00:28:40 hahaha 00:30:06 á³á»á° 00:30:30 hope you guys have all these fonts, otherwise sorry for the line noise :) 00:30:50 Don't have the last one. 00:30:59 all but last 00:31:28 random esoteric idea 00:31:36 there's probably irc-alikes with s-expressions 00:31:39 Ge'ez or Ethiopic has been assigned Unicode 3.0 codepoints between U+1200 and U+137F 00:31:51 but what about irc alikes with stack-based text 00:31:51 like forth 00:31:51 or postscript 00:32:12 Random esoteric idea. Transmit pieces of code to a million monkey with a million typewriter. 00:32:18 Then, magic happens. 00:32:30 unicode is evil 00:32:56 bsmntbombdood: no, unicode is wonderful 00:32:59 i luff unicode 00:34:19 Why is yunicode evil? 00:35:01 Unicode is full of chars you will never use. 00:35:22 3 terabites of it are just asian characters. 00:35:24 -!- slereah_ has changed nick to Slereah. 00:35:42 English is evil! 00:35:57 English is full of words you will never use. 00:36:11 Slereah: I think asian people may disagree. 00:36:31 The phone book is evil! 00:36:43 The phone book is full of people you will never speak with. 00:36:56 bsmntbombdood 00:37:01 what 00:37:07 And let me not even _start_ on wikipedia. 00:37:07 you keep saying this but I don't understand why 00:37:10 My sister was once bitten by a møøse 00:37:39 R6RS is evil, it's full of stuff you'd never want to use. 00:37:41 is that like a moose, except with an Inspector Closeau accent? 00:37:42 oh wait, that's true 00:38:01 Heh. 00:38:01 who the hell wrote R6RS 00:38:07 lol r5rs 00:38:09 *6 00:38:25 faxathisia: various people's arses 00:38:26 Everytime I read this acronym, I can't help but think of a little robot 00:38:37 yeah 00:38:39 R2D2's cousin or something 00:39:29 R2RS was pretty sucky 00:39:29 :P 00:40:06 hebrew don't celebrate new year anyway 00:40:14 not today at least 00:40:42 I'm a jew, and I don't care for the jew calender! 00:40:50 Slereah: are you jewish or are you a jew 00:40:57 hey what a coincedence, i'm not a jew either! 00:40:58 big difference :-) 00:41:09 bsmntbombdood: Slereah *is* a jew 00:41:11 or jewish 00:41:17 Define both terms. 00:41:53 oh i misread 00:42:32 Slereah: Jewish means you adhere to the Abrahamic, one-mega-psychotic God religion 00:42:45 Jew means your heritage is Jewish 00:42:57 I'm an atheist. 00:43:47 http://en.wikipedia.org/wiki/Quinary 00:43:53 someone figure out 2008 in that 00:43:58 :) 00:44:20 I get lost between 100 and 125 00:44:44 according to the calculator I need to encode 401 + 3 00:44:55 31013 00:45:05 err 5 * 401 + 3 00:45:43 1013? 00:46:34 hmph 00:46:37 Hm. Maybe not 00:46:57 Fuck that, I'll just use Mathematica 00:46:58 'hello world' * 'main' channel say 00:47:01 dambumiriw dambumirri rulu vs dambumirri dambumirri rulu 00:47:06 * 'main' channel join 00:47:13 hm. 00:47:31 showIntAtBase 5 intToDigit 2008 "" 00:47:32 31013 00:47:45 Forgot the 3. 00:47:51 pslbnf { primary = /2+2/ => "4" print null }; 00:47:55 (say (channel 'main') 'hello world') -> 'Hello, world!' 'main' channel say 00:52:00 i'm going to write a self-applicable scheme->forth compiler 00:52:02 that is all 00:52:30 what does self-applicable mean\a 00:57:15 faxathisia: it can compile itself 00:57:20 and the result can compile itself again 00:57:20 etc 00:59:28 * oerjan has seen the term self-hosting for this 00:59:36 oerjan: yes 00:59:47 but self-applicable is a better term 01:00:01 self-hosting implies that if you repeat the process multiple times, it'll 'stack up' on self-hosts 01:00:45 -!- bsmntbombdood has quit (Connection reset by peer). 01:01:22 -!- bsmntbombdood has joined. 01:03:56 ehird: Why don't you finish ISWYM first :P 01:05:35 faxathisia: 'cause it'll be a pain in the neck :P 01:12:11 -!- bsmntbombdood_ has joined. 01:17:15 -!- oerjan has quit ("Good night"). 01:17:35 someone should write a brainfuck->unlambda translator 01:17:47 then, reduced-unlambda->ski 01:17:50 then ski->iota 01:17:55 voila, bf->iota 01:18:13 -!- bsmntbombdood has quit (Connection timed out). 01:18:18 I don't feel very motivated right now. 01:18:20 brainfuck->unlambda translator souynds hardest 01:19:01 ok, bf->something-with-a-path-to-ski 01:19:08 How do you even make some sort of memory in Unlambda? Keep pairing elements? 01:19:55 Presumably, consing, yeah 01:20:00 bf->lambda-calculus might be easier 01:20:12 input should be specified in the prog!input form pre-compilation 01:20:22 and output should be extractable, somehow 01:20:36 so with a wrapper script, we could really have the slowest BF interp ever 01:20:47 Even slower than mine? 01:21:09 Don't be so sure, I'm working on a 1 tape version as we speak! 01:21:12 Probably 01:26:13 C question 01:26:24 does a signal handler call unwind the s tack? 01:33:45 brainfuck->unlambda isn't that hard 01:34:01 first just write a brainfuck->scheme 01:34:22 Then a scheme->unlambda. :p 01:34:24 reduced scheme 01:36:43 brainfuck->scheme should be simple, yeah 01:36:55 if you ensure no imperatives, you can ->unlambda 01:36:58 well, you could do imperative IO 01:37:02 but that won't survive the translation 01:37:09 thus output is not in R-unlambda 01:37:28 R-unlambda? 01:41:35 ehird`: No, the signal is called in the current context - the program continues executing where it was when the signal handler ends. So, no unwinding is done. 01:41:53 excellent 01:42:07 so, assuming i can get myself an invalid page in the right place, my signal handler just needs to realloc 01:42:13 the invalid page bit is the hard part 01:42:23 since even mmap makes no guarantees, unless you pass it the right flag, which can then make it fail 01:55:43 GregorR: any bright ideas? 02:24:14 GregorR: pingify 03:15:46 BF on a one tape Turing machine is quite annoying to do. 03:16:03 I don't have a simple way to load a file on it. 04:50:38 -!- GreaseMonkey has joined. 06:14:14 The program and the input both have to be in the tape to start with. Both are finite, and then you put the infinite brainfuck array after them. How are you setting tape state in the first place? 06:14:56 You'll need a marker to distinguish the program from its input. I suggest '!' of course :) 06:17:09 And then to separate the input from the array, hm. Maybe an EOF stored on tape. Then there are 257 tape symbols in total, which makes sense. 06:19:31 Well, of *course* you suggest !. 06:25:14 dbc : No need for input in the program 06:25:23 It's not just any Turing machine! 06:25:27 It's a choice machine! 06:26:17 I think my version work right now. 06:28:20 The tape looks something like this : http://farm3.static.flickr.com/2143/2153953116_145ff9967d_o.jpg 06:28:50 @ indicates the beginning of the code, $ the current position, LF the end of it. 06:29:12 µ for the beginning of the memory, ¤ or £ for the current position. 06:29:31 -!- faxathisia has left (?). 06:29:48 -!- faxathisia has joined. 06:30:29 Since the machine has to travel most of the memory and code for each instruction, it will probably be a whole lot slower than the 3 tapes version. 06:47:21 So it only handles brainfuck programs that don't take any input? 06:47:33 So, it handles P'' programs? 06:48:01 The Turing machine has an input function. 06:48:07 Ahhh. 06:48:33 choice machine being this : 06:48:35 "For some purposes we might use machines (choice machines or c-machines) whose motion is only partially determined by the configuration (hence the use of the word “possible” in §1). When such a machine reaches one of these ambiguous configurations, it cannot go on until some arbitrary choice has been made by an external operator. " 06:50:08 That's right, I think you said the 3 tapes version was slow. Why was that? 06:50:24 Well, because it is. 06:50:44 I tried some programs, some never finished 06:51:09 The mandelbrot.b ran for 10 hours, outputing two lines before I turned it off. 06:51:22 Does it average more than ten Turing instructions per brainfuck instruction? 06:51:50 Or did it take a long time to do each Turing machine instruction... 06:52:05 See for yourself! 06:52:07 http://paste-it.net/5500/ 06:52:10 Okay. 06:52:43 The number of instructions isn't fix. But there's less than 5 states per instructions read. 06:53:00 I think it's mostly the fault of the Love Machine 9000. 06:53:47 Well, actually, there's exactly 5 states for the brackets. 06:53:57 But other than that, it's usually 2. 06:54:30 One state to do the instruction, one to go back on the code and increment it. 06:57:08 Ah. Increment it, meaning move the code head right? 06:57:18 Yes. 06:57:32 That's the state g in the code. 07:05:39 I should put the unicode-like version online, too. 07:05:46 Errr, ASCII. 07:26:02 -!- calamari has quit ("Leaving"). 07:59:59 -!- clog has quit (ended). 08:00:00 -!- clog has joined. 08:59:18 -!- GreaseMonkey has quit ("Welcome to 2008. I will be your tour guide."). 10:12:38 http://www.codu.org/wiki/?title=Hydra // card game we invented at the New Years party 10:45:09 -!- slereah_ has joined. 10:57:53 -!- Slereah has quit (Read error: 110 (Connection timed out)). 12:09:24 -!- oerjan has joined. 13:24:13 -!- Hiato has joined. 13:24:25 Wassabi all :) 13:24:28 -!- RedDak has joined. 13:29:41 just chili'ng out... 13:29:58 cool, same here, after last night ;) 13:31:09 oh yeah, happy new year 13:31:17 yip, you too :) 13:31:46 PS: 8tt.org is back so Full 0 now sits back at http://www.full0.8tt.org 13:34:55 -!- Hiato has quit ("Leaving."). 13:50:29 -!- jix has joined. 14:21:05 "the new years party" 14:21:13 GregorR likes to pretend he's sociable 14:21:44 GregorR: it's ok, we understand you here 14:48:06 hey, he didn't say it was an offline party... 14:49:00 hm actually GregorR has to be sociable. he has to show off his hats after all... 14:50:39 Well, he could send photos on the interwebs! 15:11:11 -!- jix has quit (Read error: 145 (Connection timed out)). 15:12:05 -!- jix has joined. 15:25:34 -!- biribau has joined. 15:26:18 hi, prolog basics questions have site here? 15:33:39 -!- oerjan has quit ("Dinner"). 15:39:07 -!- biribau has left (?). 15:47:35 -!- Sgeo has joined. 15:48:59 Hi all 15:52:15 Anyone awake? 15:52:33 * Sgeo threatens to not work on PSOX if no one talks 16:00:47 Hi. 16:01:06 -!- slereah_ has changed nick to Slereah. 16:02:05 hi 16:28:12 wow, er, that's impressive.. i didn't know you could patternmatch lambdas 16:28:12 oklopol: you'd like that! 16:28:16 what? 16:28:29 why :\ 16:36:22 ((?:[^\x00]|\x00{2}(?:.|\Z))*)(\x00.*)? is the most important regex in PSOX? 16:37:38 Can that regex be made more efficient? 17:07:02 -!- Hiato has joined. 17:07:46 Hi Hiato 17:07:58 Wassup Sgeo :) 17:08:09 ((?:[^\x00]|\x00{2}(?:.|\Z))*)(\x00.*)? 17:08:25 say what? 17:08:46 That's the most important regex in PSOX 17:09:12 lol, regexp's are not my strong point :P (especially in Ruby) 17:09:18 Why, what does it do? 17:09:47 breaks a line apart into the outputted text with 0x00 0x00 etc. and the command part with 0x00 17:09:55 aha 17:10:03 oh yeah huh, ^\xoo 17:10:04 :) 17:10:26 so, anything news, perhaps a version I can start meesing with? 17:10:30 *messing 17:11:21 *new 17:11:38 meh, I hat tyops :P 17:12:23 x:xs can be used as a pattern matcher because a list can be represented as just one cons. a+b cannot pattern match because there's an infinite amount of conses for addition, 6=6+0=5+1=... now, if one were to introduce the concept of virtual cons cells into a language, and make a cons be a set of rules of dependence, so that an addition cell knows it's (result - b, result - a). 17:12:25 * Sgeo is working on the psox.types module 17:12:45 this way you could pattern match like the haskell n-1 pattern, but in a more sophisticated fashion. 17:13:11 well, haskell prolly has a much more sophisticated reason for that to work, but i don't know it, and i prefer my own ideas. 17:14:01 ?? 17:14:22 :), I generally prefer my own ideas too, over others that is. Some call it stubbornness ;) Oh, and I agree with Sgeo: ?? 17:15:26 these would be virtual cells of course, so when doing a pattern match for 5 on (n+3), 5 would be converted to an addition cell as (result-b, 3), where b=3 = (2,3), so n=2 17:15:50 IF the conversion into an addition cons cell should fail, the pattern match simply would fail. 17:16:11 this would be correct at least in this case, since only numbers can be subtracted from 17:16:19 I'm still not quite with you... but I guess it's not aimed at me :) (Otherwise ti would be allot longer) 17:16:25 Is this anything to do with PSOX? If so, please be warned that I don't have a drop of understanding of what you're talking about 17:16:51 but, you could also have the possibility to specify domains for which the conversion always works 17:17:13 is this part of the Graphica spec? 17:17:35 oh, right, in case in the final pattern, there were still cells that are defined in terms of other cells, the pattern itself would be incorrect, for example if you tried to pattern matchon (a+b) 17:17:37 *match on 17:17:53 ... 17:18:08 although, that could just take a random cons cell that represents 5 as an addition. 17:18:41 sometimes, all that matters might be that you separate the object into *some* other objects that, when consed, result into the original object. 17:18:56 not in the case of numbers and addition, prolly, but with lists, for example. 17:19:08 Sgeo: is he talking to us? 17:19:20 :) 17:19:30 this is very esolang-related, and not that high-level 17:19:33 so... yes? 17:19:37 anyway, i'll continue 17:19:40 :P 17:19:50 sure, don't let me stop you 17:19:58 If it's PSOX related, you are hereby REQUIRED to translate it into something that makes sense.. 17:20:01 just /ignore me for a while in case not interested, this is just may way to clear my thoughts, if okay. 17:20:22 Sure, proceed 17:20:38 so, for instance quicksort's pivoting, pmatch the list into lpart++[pivot]++rpart 17:20:41 I will interject when (rather if) I have something to say 17:20:47 because all we want is *some* pivot 17:21:15 (++ lpart (++ [pivot] rpart)) 17:21:36 if i didn't have this headache, i'd manually pattern match that 17:21:46 might be an interesting concept 17:22:04 but once again it's so clever yet simple it must already be in use. 17:22:08 er.. yes .. amm.. no .. whichever is more appropraite :) 17:22:18 well, not clever, but... fun? 17:23:07 sorry, was lost in my thoughts, so didn't read what you said between my messages 17:23:22 you didn't *get* that? in that case i'll happily explain 17:23:26 sure, no problem, and believe me, it wasn't important 17:23:28 i just assumed you don't care :-) 17:23:49 lol, I care, but don't *concentrate* 17:23:53 hehe 17:23:53 :) 17:24:01 do you know what pattern matching is? 17:24:14 amm, well yes (I think) 17:24:40 as in "The cat ate the mouse" pattern would be "at' twice? 17:24:47 or not 17:24:48 .... 17:24:51 nevermind 17:24:55 amm.. No 17:24:59 I don't 17:26:11 hehe 17:26:13 sooooo 17:26:20 like, if you have the list [1,2,3] 17:26:26 yip 17:26:33 that's basically (cons 1 (cons 2 (cons 3 nil))) 17:26:38 in lispy fashion 17:26:40 do you get that? 17:26:41 cons? 17:26:49 well, other than the cons, yes 17:26:51 construct, i guess 17:26:55 oh, ok 17:26:56 sure 17:27:08 it means (cons head-of-list tail-of-list) 17:27:14 where tail-of-list is usually the rest of the list 17:27:15 aha 17:27:19 cool 17:27:22 proceed 17:27:32 (with caution ;) ) 17:27:45 now, let's define a function that takes the second element of a list and returns that *2 17:28:03 ok sure 17:28:17 you could do it with L -> (tail tail List) * 2 17:28:26 where tail does (cons a b) -> b 17:28:29 got it? 17:28:33 1 sec 17:28:39 sdfapojoaifjg 17:28:42 sorry. 17:28:47 lol, wait, thinking 17:28:50 L -> (tail head List) * 2 17:28:59 think about the correct one, though :) the latter 17:29:00 ? which one 17:29:02 --- 17:29:04 FUCK. 17:29:06 ok, wait some more 17:29:08 :) 17:29:12 OKAY, that's incorrect too, sorry 17:29:14 headache. 17:29:17 meh 17:29:20 L -> (head tail List) * 2 17:29:36 means, take L as param 17:29:42 ooh 17:29:42 then return what's after -> 17:29:43 ok 17:29:52 wait 17:29:59 tail list of head is 2,3 17:30:02 not just 2 17:30:02 could explain my code if i invent the notation on the fly... 17:30:18 L = 1,2,3 17:30:21 yip 17:30:24 so tail L = 2,3 17:30:28 yes 17:30:30 so head tail L = 2 17:30:35 oh, I see 17:30:39 for a more imperative version 17:30:40 not tail of head 17:30:44 but head of tail 17:30:45 roger 17:30:47 :) 17:30:50 gotcha 17:30:53 im with you 17:31:27 L -> ListTail = tail List; SecondHead = tail ListTail; return SecondHead*2 17:31:33 although you already got it. 17:31:39 i just wanted to write that for some reason 17:31:45 sure 17:31:46 :) 17:32:22 okay, no the idea of pattern matching is, instead of explicitly taking the *head* of a list, we just take the list in as if it were already spliced into a head and a tail. 17:32:24 so 17:33:11 (: 1 (: 2 (: 3 Nil))), ":" is cons here, it's just a shorter word so i don't have to type cons, although this explanation is longer than all the conses i've ever written, so it is pretty useless. 17:33:34 ok 17:33:37 go on 17:33:40 so, we define the function as (: A (: B L)) -> 2*B 17:33:51 so, instead of just putting the param into a variable 17:33:59 we have a pattern we match for the parameter 17:34:10 aha 17:34:14 and inside the pattern, there are variables, that hold the parts of the parameter. 17:34:20 L being significant of the rest of the list? 17:34:32 this works because most data can be, and is often represented by a tree., 17:34:43 which you prolly know? 17:34:49 trees 17:34:51 yip 17:35:00 and somewhat Alpha Bet Pruning 17:35:08 *Alphabeta 17:35:11 and minimaxing 17:35:18 BFStmt=+|-|>|<|[BFStmt]|,|. 17:35:26 no, sorry 17:35:27 hmmm 17:35:31 alphabeta? 17:35:31 what the hell is that 17:35:38 what is what? 17:35:38 never seen it like that 17:35:45 BFStmt=+|-|>|<|[BFStmt]|,|. 17:35:47 oh 17:35:51 you claim it to be alphabeta 17:35:52 just a random notation. 17:35:57 oh, ok 17:36:05 | was jsut my way to separate possibilites for a branch 17:36:18 actually a pretty standard way 17:36:23 i don't know alphabeta... 17:36:25 oh, ok 17:36:36 well, its pruning, like minimaxing 17:36:37 anyway, L was the rest of the rest of the list. 17:36:40 except different 17:36:42 :) 17:36:49 cool 17:37:29 (: A (: B L)) -> 2*B can be written A:B:L->2*B if we notate : as infix and right grouping, right 17:37:31 ? 17:37:41 1:2:3:Nil is the list. 17:37:56 yip 17:37:59 get it 17:38:02 so... we pattern match as 1->A, 2->B, 3:Nil->L 17:38:30 actually, the cons representation is much clearer when you pattern match, just felt it'd look bad if i wrote it on irc. 17:38:34 so, actually L = 1->A, 2->B, 3:Nil 17:38:37 in other words 17:38:44 nope! 17:39:02 L would just be (: 3 Nil) 17:39:09 wa? 17:39:13 because A and B take the first two elements 17:39:18 L only gets the last. 17:39:25 oh, I see 17:39:37 just that the 3:NIL->L got me 17:39:39 yip 17:39:41 3:Nil 17:39:45 is the last one 17:39:48 roger 17:39:51 back with you 17:40:20 in (: A (: B L)) -> 2*B, (: 1[place of A] (: 2[Place of B] (: 3 Nil)[place of C])) 17:40:32 uhh that was ugly. 17:40:44 but 17:40:46 yueah 17:40:48 I get it 17:41:04 (: 3 Nil)[place of C] surely place of L 17:41:12