00:05:01 -!- KingOfKarlsruhe has quit (Remote closed the connection). 00:34:57 -!- tusho has quit. 00:40:07 ff*ky, what a thought 00:40:32 -!- RedDak has quit (Remote closed the connection). 00:50:23 * SimonRC goes to bed 00:59:17 -!- psygnisf_ has joined. 00:59:17 -!- psygnisfive has quit (Read error: 104 (Connection reset by peer)). 01:01:21 -!- psygnisf_ has changed nick to psygnisfive. 01:03:22 -!- lilja has quit ("KVIrc 3.2.0 'Realia'"). 02:37:26 -!- olsner has joined. 02:37:41 -!- CakeProphet has joined. 02:59:42 -!- olsner has quit ("Leaving"). 03:13:26 It seems that right now, compression is about trying to compress as much data as possible into as small a space as possible, and then decompression is achieved just by running the result or something. 03:14:26 It'd be interesting to look at that from the other direction: compress something into a reasonable space using an obvious method, then work really hard to decompress it. 03:14:33 Compression is the act of taking a low-entropy source of data and transforming it into a high-entropy source of data in a reversible manner while reducing it in size. :p 03:15:48 (I keed, I keed) 03:16:23 ok guys i have an idea 03:17:12 what if we created a programming language where the language primitives are things that normally require complicated code in other languages, things that would be large composite entities 03:17:29 but which are incredibly abstract despite 03:18:04 not that this is a good example but, maps are generally composite, formed by defining some map function, but what would it look like to have a language where map was a primitive operation 03:18:04 Imagine putting every other line of "To Kill a Mockingbird" through a low-security hash function of sorts (CRC?), so that there's only one reasonable string that produces that output. 03:18:40 psygnisfive: map as in map (+1) [2,3,5] = [3,4,6]? 03:18:49 sure whatever 03:18:50 psygnisfive: That's a terrible example. 03:18:55 yeah i know ;) 03:19:03 Because functional languages *have* map as a primitive. 03:19:15 You have one primitive: RSA encryption. 03:19:19 well but not really. i mean, in haskell, map is not a primitive, for instance 03:19:24 its built in, but its not primitive 03:19:33 What's the definition of "primitive"? 03:19:37 maybe in APL or some other languages perhaps 03:19:48 primitive as in the building blocks of the rest of the language 03:19:59 Ah. 03:20:03 in old school lisp, for instance, the primitive operations were like 03:20:16 eq, atom, list, cons, car, cdr, and cond 03:20:19 oh and quote 03:20:25 or something along those lines 03:20:36 I guess Haskell's primitives are pattern matching, lambdas, application, and Lots Of Other Stuff. 03:20:45 everything was built out of those primitives 03:21:19 I suggest the brainfuck interpreter as a primitve. :p 03:22:13 :P 03:23:12 I guess Malbolge is famous for being kind of like that. 03:23:21 That's kind of the goal of esoteric programming, really. :-) 03:24:27 what, having a BF interpreter as a primitive operation? :P 03:25:29 No, coming up with a language that uses insanely complex primitives to accomplish something relatively simple. 03:25:49 AHAH! 03:26:01 We do not implement standard arithmetic operators. 03:26:12 *Instead*, we only operate on matrixes. 03:26:24 We implement ONLY standard arithmetic operators, and computation is done via rounding error. 03:26:38 well i was really thinking more like 03:26:43 On matrixes? 03:26:47 array programming languages have array operations as primitives, right 03:26:56 pikhq: sure. 03:27:00 Yes. >:-) 03:27:04 >:D 03:27:06 so that if youre coding something that does a lot of array operations, its intuitive to code it in an array programming language 03:27:21 I'm sure the IEEE standard floaty things are a rich computational environment... 03:28:01 but what is the extreme of that? what would it look like if we decided that an array programming language is too low level or something 03:28:23 what if Language X is to an array programming language what an array programming language is to an assembly language 03:28:31 what would Language X look like 03:29:48 Problem: nobody speaks of an 'array programming language'. 03:29:59 No more than someone speaks of an 'if programming language'. 03:30:24 Either it has arrays as primitives or they're trivial to implement. 03:30:39 ... Or it's called Malbolge. 03:32:00 uh 03:32:13 pikhq: http://en.wikipedia.org/wiki/Array_programming 03:33:13 Ah. 03:33:22 thanks for playing. 03:33:22 :P 03:33:25 I was, obviously, thinking of something completely different. 03:33:38 Sorry; I tend to think of those as vector operations. 03:33:41 YOU SIR ARE A DISGRACE 03:33:41 :P 03:33:43 Not array operations. 03:33:57 omg omg omg hey so i make my own tee-shirts 03:34:04 lets come up with eso teeshirts 03:34:06 :o 03:35:16 Amusingly, GNU C, courtesy of its attempts to support SSE and such well, is getting a bastardised set of array operations. :p 03:35:33 s/array/vector/ 03:35:40 Damn you... 03:35:43 also, on a semi-related note, data-oriented programming is sexy 03:36:26 On an unrelated note, EBIL!!!-oriented programmig is sexy... 03:36:32 In a BDSM sort of way. 03:36:41 EVIL? 03:36:43 .. 03:36:44 EBIL* 03:37:04 -!- ihope has quit (Read error: 104 (Connection reset by peer)). 03:38:00 Yes. 03:38:17 You know, ones that require n-dimensional thought. 03:38:55 no i dont know 03:41:16 DIMENSIFUCK. 03:41:19 :D 03:43:18 ? 03:43:30 -!- ihope has joined. 03:43:32 Firefox got a migraine. 03:43:46 whats firefox have to do with this? 03:45:36 I don't know, but soon after the backwards typing it froze up and I had to restart it. 03:45:54 but why does that affect your irc??? 03:46:03 Um, no reason. 03:46:07 * ihope whistles innocently 03:46:11 Chatzilla? 03:46:15 You bastard! 03:46:18 Use XULrunner! 03:46:21 oh. thats lame. 03:46:29 There is no browser, there is only XUL. 03:47:22 Using sed | nc | sed would be fun. 03:48:03 Brilliant. 03:48:22 so cmon! 03:48:31 i want to make #esoteric teeshirts 03:50:02 Stick some of the more odd Hello, World examples on there. 03:52:30 noo it has to be interesting not just a shitty tee-shirt 03:52:43 most of the tee-shirts im designing for the linguistics club at school are horrible puns 03:52:53 like 03:53:06 Might I recommend a Fugue program? 03:53:25 3-inch-white thumbtacks (the flat disc-like ones), 7 of them, each with one of the 7 deadly sins written on it 03:54:05 get it? 03:54:08 they're sin tacks.. 03:54:10 syntax 03:54:10 :D 03:54:45 im doing another one thats a log cabin 03:54:49 "tree structure" 03:54:51 :B 03:55:29 How would you feel about sticking Var'aq on a shirt? 03:55:37 whats it look like 03:56:43 Klingon. 03:56:49 No, seriously. 03:57:00 why would we put that on a tee-shirt 03:57:02 that would be silly 03:57:23 It's a programming language. 03:57:29 All the keywords are in Klingon. 03:57:32 yes so? 03:57:44 It's twice-esoteric. 03:57:51 but its not FUNNY 03:57:59 Oh, you want *funny*? 03:58:13 it has to be interesting and humorous 03:58:18 witty 03:58:19 ironic 03:58:20 something 03:58:41 "Esoteric languages: because the levels of hell deserve it." 03:59:03 *snore* 04:01:45 "Brainfuck? I hardly even know you!"] 04:02:22 also lame 04:03:27 im thinking a haskell one that mocks IO as being a useless complication. :T 04:03:58 Tempting. 04:04:49 I wish Python had callCC. 04:05:20 And handled it nicely, as there are many ways to handle it... 04:05:24 BTW, I'd just like to announce: 04:05:34 Sweetness; PEBBLE is in DMOZ. 04:12:56 OH GOD 04:12:59 THATS GENIUS 04:13:00 Hrm. How many people here have websites in Dmoz? 04:13:06 psygnisfive: What? 04:13:23 a tree made up of bits 04:13:28 a binary tree! 04:13:36 LMAO 04:14:02 who wants one? :O 04:14:24 * ihope frowns, as a binary tree is not made up of bits 04:14:37 ihope, you're lame. 04:14:59 * ihope frowns, as he's lame 04:15:06 Wait, what? 04:15:13 its a pun 04:15:18 God dammit; now I want to actually work more on my PEBBLE game engine. 04:17:41 * ihope frowns, as it's a pun 04:17:55 I mean, a and b = not (not a or not b). 04:18:04 i love silly puns 04:18:36 So take the complement of each regular expression, alternate them, and take the complement of that, and see whether it's empty or not. 04:21:24 And the valid characters in hostnames are letters, numbers, hyphen and dot. Yay. 04:21:53 wait, whats the complement of a regular expression?? 04:26:59 The regular expression that matches everything that one doesn't. 04:27:50 The complement of /red/ is /|[^r].*|r[^e].*|re[^d].*|red..*/, I believe. 04:28:07 And taking complements is really annoying. :-) 04:28:33 Better to just have a complement operator, maybe. 04:28:54 i think thats wrong. :P 04:29:04 would it have to be like 04:29:57 blah: /.+red|red.+|.*[^r]ed.*| .... 04:30:10 or something 04:30:24 why isnt there some way of grouping 04:30:43 Isn't there? 04:30:46 i dont know. 04:30:48 anyway 04:30:53 What's wrong with mine? 04:30:58 its wrong. :P 04:31:07 * ihope hits psygnisfive 04:31:33 It's right. 04:31:38 maybe. :D 04:31:53 no you're right, it is 04:40:59 I propose a new feature for Perl regexps. ^{{: regexp :}} 04:41:04 Complement of regexp. 04:41:13 (please say they don't already use that) 04:41:20 (Please?) 04:41:28 oh. 04:41:29 i know 04:41:48 dont even bother 04:41:57 just have a function that returns an object that behaves exactly like a regexp 04:42:03 same methods, etc 04:42:11 only it behaves like the complement 04:42:39 or if regexp's arent objects just data 04:42:54 then have a wrapper 04:43:01 i mean figure 04:43:07 if this is how you test a regexp: 04:43:21 test( expression, string ) 04:43:32 My goal here is to see if two regexes overlap, and to see if one is a subset of another. 04:43:34 then this: test( complement_of_expression, string ) 04:43:36 is the same as 04:43:42 !test( expression, string ) 04:43:42 ihope: I hate you. 04:43:43 ] 04:43:48 And all regexp fans. 04:43:49 :p 04:44:09 so your test function could just look at the first arg to see if its a complement or not 04:44:18 There is no alternative. Wait, there is: wildcards. 04:45:53 Yes, there is an alternative. 04:46:03 if. 04:46:08 BWAHAHAAH! 04:46:11 what? 04:46:32 "if"? 04:46:47 With enough if statements, you can do just the same thing as the regexp. . . Of course, by the time that you have done so, you will be stark raving mad. 04:47:44 Do this with if statements: ihope/.* 04:47:57 I guess "if it begins with ihope/". 04:48:28 * ihope begins conjuring up an appropriate regex to hit pikhq with 04:49:17 -!- Corun has quit ("This computer has gone to sleep"). 04:52:45 I don't want to be stark raving mad, however. 04:52:54 And it was a gigantic fucking joke. 04:53:06 Yet oddly true. Yay, Turing. 04:53:26 what? 04:54:53 It's midnight; I don't have to make sense. 04:56:11 Your regex is almost ready. 04:56:20 And I ain't touching it. 04:58:30 /Gr((er?)*e(n(R(er?)*en)*R(er?)*e(a(p(er?)*e|n(R(er?)*en)*R(er?)*ea|a)p(er?)*e(n(R(er?)*en)*R(er?)*er|r)|r)|a(p(er?)*e|n(R(er?)*en)*R(er?)*ea|a)p(er?)*e(n(R(er?)*en)*R(er?)*er|r)|r))?/ 05:03:42 'Trivial'. 05:03:48 pœnis! 05:03:52 If I were in a looney bin. 05:07:17 -!- pikhq has quit ("leaving"). 06:23:59 -!- Sgeo has quit ("Ex-Chat"). 06:28:58 we should design a language where the primitive operations are operations that inherently are parallel 07:20:26 -!- dbc has quit (Client Quit). 07:23:40 -!- dbc has joined. 07:59:59 -!- clog has quit (ended). 08:00:00 -!- clog has joined. 08:00:46 idea for complement-regexps: forget expressing the complement using another regexp 08:00:58 just build the automata for the positive expression 08:01:08 and flip the accept/reject states 08:01:38 that way the positive and the negative are clearly related 08:02:00 and you wouldnt need a special method of figuring out the complement automata 08:02:14 er, the complement expression 08:15:46 /|[^r].*|r[^e].*|re[^d].*|red..*/ cannot be a complement of /red/, because (using the usual "matches anywhere" semantics, not the "endpoints are tied to the ends of the string" ones) both match "redx". For the latter type of matching it might work. 08:19:07 As for Perl, usually you just use !/.../ to complement regexps, or maybe the zero-width negative lookahead (?!...) thing if it's inside a larger regex. 08:21:36 yes, obviously what he meant was that /^$|^[^r].*$|^r[^e].*$|^re[^d].*$|^red..*$/ is the complement to /^red$/ 08:21:51 using the matches-anywhere semantics 08:22:17 If you just flip the accept/reject states, your complemented regexps end up having an undefined "which part of the string matched" property. Not that the "which part matched" result given by an explicitly complemented expression would be very useful, either. 08:22:49 well i wasnt assuming that thered be some response other than booleans 08:23:06 so there'd be no relevance to "which part matched" 08:23:46 obviously any non-zero pattern will have atleast one part of a string matching its complement 08:24:02 given that its complement includes the zero string, and thus theres atleast one zero-string match in any string 08:24:33 so i dont think the "which part matches" is relevant, since most complements will always match the zero string on all inputs 08:24:59 the idea i think was the match whole strings, from beginning to end 08:25:02 and then complement that 08:25:11 not to match a substring 08:25:16 Yes, it certainly sounds a lot more useful. 08:25:27 night. 08:25:33 Morning. 09:08:42 -!- RedDak has joined. 09:11:56 -!- olsner has joined. 09:16:35 BAD: the top of the stack after y isn't equal to what 1y pushes 09:16:36 hrrm 09:16:55 Deewiant, while I know mycology doesn't do Funge-108, I still don't see how that error can happen 09:25:38 " This means any isolated argument can be a null string, but no two consecutive arguments may be null strings - a rather contrived scenario, null string arguments being rare in themselves. 09:25:38 The first string is the name of the Funge source program being run." 09:25:44 that is from Funge-98 Deewiant ^ 09:25:55 which means you can be sure to not have two of them in a row 09:26:00 the spec forbids it 09:26:05 bbiab 09:41:50 AnMaster: yeah, but still: foo 0 0 0 <- the 0gnirts "foo" followed by an empty string, or the 0gnirts "foo" followed by the end of the environment variables 09:45:02 -!- dbc has quit (Remote closed the connection). 09:45:22 -!- dbc has joined. 09:47:50 Deewiant, no that is easy 09:48:07 3 in a row == null + end of args 09:48:25 4 in a row == end of args, end of env 09:48:41 Deewiant, anyway 108 will have lengths in it 09:58:18 AnMaster: foo 0 0 0 0 <- "foo", end of args, end of env or "foo", null, end of args 09:58:52 foo 0 0 0 0 0 <- "foo", end of args, null envvar, end of env or "foo", null, end of args, end of env 09:59:22 errr...? 09:59:36 foo 0 0 0 0 09:59:45 that can't be "foo", null, end of args 09:59:58 because that would be "foo" 0 0 0 10:00:23 Deewiant, and null env isn't valid afaik 10:00:32 because the format is NAME=VALUE 10:00:44 null env variable just isn't valid 10:01:06 AnMaster: args are terminated by double null 10:01:14 correct 10:01:16 "foo", null is two zeroes, end of args makes it four 10:01:16 so are env 10:01:21 no 10:01:24 env is terminated by a single null 10:01:58 no.... 10:01:59 AnMaster: and the format is NAME=VALUE followed by null, so if there's no name or value it could theoretically be null 10:01:59 a series of strings, each terminated by a null, the series terminated by an additional null, containing the environment variables. (env) 10:02:08 yes, exactly... 10:02:11 that will be double null at the end 10:02:11 "an additional null", not two 10:02:21 "a series of sequences of characters (strings), each terminated by a null, the series terminated by an additional double null, " 10:02:22 hm 10:02:27 that must be a typo? 10:02:30 I do not care 10:02:33 that's what it says 10:02:55 in any case, it's either a) damn hard to parse it always correctly or b) impossible 10:03:14 so unless you can write Befunge code that handles all cases correctly, I'm not going to bother 10:03:24 heh 10:03:43 Deewiant, it is weekend now 10:03:48 yes, quite 10:03:51 so working on fixing ccbi and mycology? :) 10:04:03 no, but I'm working on working on it ;-) 10:04:08 hah 10:04:09 what's the k consensus 10:04:15 everything in mycology is correct except for the 3k4? 10:04:21 no 10:04:27 Deewiant, you use k in some other places 10:04:33 you need to check them 10:04:44 s/everything/every place where k is checked/ 10:04:47 ? 10:04:58 BAD: 0y pushes wrong stack stack size 10:05:04 that one must use k 10:05:09 yes yes, whatever 10:05:22 Deewiant, also I suspect BAD: "a b" takes more than 5 ticks 10:05:23 uses k 10:05:25 but all the k tests that currently show GOOD for CCBI, except for 3k4, are correct 10:05:28 or?! 10:05:39 Deewiant, um that makes no sense? 10:05:59 Deewiant, with CCBI style k everything says GOOD in my version of mycology (from last week) 10:06:11 with Funge-98 style k: 10:06:12 BAD: 3k4 leaves more than 3 fours on stack 10:06:13 AnMaster: !!!!! get what I'm saying, will ya 10:06:15 GOOD: 0k^ doesn't execute ^ 10:06:15 GOOD: 1k[ turns left from k 10:06:15 [UNDEF: k executing space] 10:06:15 GOOD: 2k# jumps twice from k 10:06:17 GOOD: ak47k$ leaves 3 fours on stack 10:06:19 BAD: "a b" takes more than 5 ticks 10:06:20 GOOD: 3k4 leaves 3 fours on stack 10:06:22 BAD: 0y pushes wrong stack stack size 10:06:22 the question is 10:06:25 are all of those correct 10:06:28 with the exception of the last 10:06:30 AnMaster: ?? 10:06:37 Deewiant, um.... good question 10:06:41 ak47k$ 10:06:46 that one should be wrong I guess? 10:06:49 4 and 7 cancel each other out 10:06:53 er 10:06:53 ah 10:06:54 a and 7 10:07:01 so it does 11 pushes and 8pops 10:07:04 instead of 10 and 7 10:07:15 that's why I added the 3k4 test originally :-P 10:07:35 GOOD: 2k# jumps twice from k 10:07:41 that... 10:07:42 um 10:07:44 not sure :P 10:07:45 what's cfunge's opinion on that 10:07:58 Deewiant, cfunge for some reason gets a GOOD from that 10:08:01 I guess Mike would have complained if it were wrong 10:08:03 but I'm not sure that is correct 10:08:20 and actually yeah, I think he reasoned about it in his mail to me 10:08:20 and said it was correct 10:08:20 GOOD: 3k< takes two ticks 10:08:20 BAD: "a b" takes more than 5 ticks 10:08:20 aha 10:08:23 that explains the tick issue 10:08:33 you get desynced before there I bet 10:09:11 Deewiant, because I get the right value when having CCBI k 10:09:15 so I'd blame that on CCBI 10:09:17 err 10:09:19 on mycology 10:15:34 Deewiant, right? 10:16:08 well, you could blame it on the tides of the moon if you wanted to :-P 10:16:41 Deewiant, well I debugged the code in question and it only happens if I change the k 10:16:47 so I assume it must be related to k right? 10:16:58 of course, goes without saying? 10:17:18 could be a hidden bug before, but I don't think so 10:17:34 due to it checking k just before 10:19:25 Deewiant, you have a bug on line 91 too 10:19:31 >"celfer y :DAB"82*k,@ 10:19:53 on line 112: 10:19:55 a"sehsup y1 tahw ot lauqe t'nsi y retfa kcats eht fo pot eht :DAB"88*k,@ 10:20:09 yeah, and the reason those are generally used is because there's no room for >:#,_... 10:20:21 Deewiant, you just need to adjust them by one 10:20:31 on line 113 you got another one: 10:20:32 yeah, so add "1-" there which takes two cells 10:20:33 >"mialc y1 :DAB"6a*1-k, 10:20:36 not easy 10:20:42 wait...? 10:20:45 hm 10:21:03 @,k+2f"BAD: 1y reflects"a< 10:21:13 there is two sapces on line 113 10:21:16 that you can use 10:21:23 well that's an addition anyway 10:21:26 so it doesn't matter 10:21:29 just change 2 to 1 10:21:45 Deewiant, well I just point them out, you need to fix every k really :P 10:21:56 yeah yeah, thanks for the info 10:22:05 there's bound to be a few incorrect outputs left with the BADs anyway :-P 10:22:10 not like I'm going to test them all 10:22:12 sure 10:28:31 Will continue to produce textual output, so strings must work correctly where concurrency is concerned: "a b" should take 5 ticks, 'a should take 1. 10:28:34 BAD: "a b" takes more than 5 ticks 10:28:40 Deewiant, um a question... 10:28:50 how many spaces are you talking about? 10:29:01 they differ between the two texts 10:29:02 ? 10:29:06 yes... 10:29:12 is it intentional? 10:29:20 because the first says that the code assumes that one space works correctly 10:29:27 ah right 10:29:28 true 10:29:28 whereas the second is testing that two spaces work correctly 10:31:34 Deewiant, the two space could actually be a bug in cfunge, but then why did it work with the different k? 10:31:46 I don't know!! 10:35:57 -!- tusho has joined. 11:10:05 Ooh, there's another feature I can add to botte. 11:10:20 An #esoteric link log. You can just add a link to it in IRC with a simple botte command. 11:10:36 botte? 11:10:53 AnMaster: ESO's (ostensibly, but really I'm the one doing it) bot for #esoteric 11:11:05 ah 11:11:15 tusho, you forgot to join #eso btw 11:11:18 and where the heck is ais 11:11:25 ais comes at 3pm 11:11:28 its 11 11:11:38 tusho, well he wasn't here yesterday was he? 11:11:40 anyway, botte (in the future will:) - does logs with awesome search - has useful interwebby stuff like a google/wikipedia lookup - and now this link log thing 11:12:01 AnMaster: He was doing stuff maybe? 11:13:37 oh wow, ellio.tt is available XD 11:13:41