00:00:03 exactly 00:00:13 although 1. why the fuck would you and 2. if you have any interest in speed, you do cmpl Function 00:00:13 maybe gradually over time you should tone it down a bit :-) 00:00:22 hehe :D 00:00:35 by the way, does oklotalk have to look like those mindmashers? 00:00:37 i think it'll be much less grandioso once i actually start implementing it... 00:00:44 show me some clean understandable more-than-one-char-var-names oklotalk :P 00:00:46 actually, it can look quite nice 00:01:23 example please :P 00:01:55 i can show you the somewhat public thingie, if you promise not to publish it / say it's not speccy enough, it's just like the haskell factorial website, but with oklotalk, and with explanations 00:02:15 a lot of explanations, but no full formal spec. 00:02:31 that's what always happens when i start writing something. 00:03:38 i won't publish it :-) 00:03:53 Well, I can input a file on the first tape. 00:03:55 by the way, my implementation is called C-oklotalk :P 00:03:59 imaginative i know 00:03:59 Onto that BF interpreter. 00:04:02 I want to see this website 00:04:05 oklofok 00:04:19 the haskell one? 00:04:22 it's great 00:04:25 no the one with oklotalk 00:04:27 but i can't find it 00:04:36 oh... goddammit 00:04:43 :) 00:04:50 you are popular oklofok :-) 00:04:57 why don't you just trust little old #esoteric :P 00:05:40 there is not website for the oklotalk one, which is why it's just semi-public 00:05:48 i've basically jsut sent it to a few ppl 00:05:50 *just 00:05:55 guess i'll upload it.... 00:06:07 but i warn ya, it's *not* that interesting 00:07:25 i find oklotalk extremely interesting 00:07:28 lol, this isn't ready for publishing... 00:07:32 i'm going to implement all my programs in it okay :p 00:07:40 like, my continuation-based webframework 00:07:43 and stuff 00:07:58 basically some scribble i made for the heck of it, and to get some thoughts sorted 00:08:00 http://www.vjn.fi/pb/p551122163.txt 00:08:01 but there it is 00:08:08 it's Oklotalk then 00:08:10 not OkloTalk 00:08:11 or oklotalk 00:08:16 Oklotalk, definitely. 00:08:28 to differentiate between my nick 00:08:40 because, it's a somewhat different origin 00:08:53 very slight somewhat, and i'm not getting into that 00:09:05 wonder what i pasted there.... 00:09:28 ' Because everything is everything in oklotalk,' 00:09:33 hehe 00:09:36 what's factorial+5 00:09:42 umm 00:09:48 that's (factorial (+ 5)) 00:09:51 in lisp notation 00:09:57 i mean, the function factorial, + 5 00:10:06 Factorial + 5 then 00:10:06 * ehird` renames C-oklotalk to C-Oklotalk 00:10:13 Factorial + 5 = ? 00:10:13 wait 00:10:47 that would mean, factorial gets as args the list [#: + 5]... depends on the pattern matches what happens then 00:11:00 most likely they will all fail, and it'll return void 00:11:12 or None 00:11:25 #z to be spesific. 00:11:57 no 00:12:03 functions are first-class in oklotalk right 00:12:06 "guess function" is a bit stupid for a term 00:12:13 sure they are 00:12:17 well 00:12:26 there's just lambdas, but no difference 00:12:38 hmmmm question 00:12:49 please pick something other than + for concat 00:13:00 (x+y)+z = x+(y+z), and x+y = y+x 00:13:04 none of these apply to concatentation 00:13:05 err 00:13:07 the first does 00:13:08 but not the second 00:13:13 oh, the "return" function's explanation has a flaw, which a friend of mine noticed, but anyway, the "return" call is just for show, and it's actually identity there 00:13:17 and, + should be flippable, so. 00:13:36 nope, you can actually specify these rules separately for different types. 00:14:30 and theoretically, by using explicit typing, let the interpreter know it can optimize with the knowledge (x+y)+z == x+(y+z) 00:15:00 yes but x+y is y+z, and this does not apply to concatentation 00:15:04 thus + is a bad name for it 00:15:12 nope 00:15:46 also, x+y for numbers produces a numerical value... does this mean it can't be used for other types unless that too produces a numerical value? 00:15:58 what exactly is the use of having those identities always hold? 00:17:12 operator input and return types should not be constant 00:17:16 but properties of application should be 00:17:24 why? 00:17:28 (Overtalk, an Oklotalk successor? oh fuck no) 00:17:32 :)) 00:17:41 oklofok: it means you can reason about programs better 00:17:47 and is simply cleaner 00:18:07 nah, oklotalk is too dynamic for that to be true 00:18:18 and too conciseness-oriented 00:18:26 still 00:18:27 it's nicer 00:18:35 what could i use then? 00:18:40 (also, Oklotalk should not be actor-based. actor model == function model) 00:18:44 (for the mostpart.) 00:18:52 (see: why scheme was created, to test that hypothesis) 00:18:59 oklofok: depends what you have unused 00:19:03 .. 00:19:03 nothing, really 00:19:03 ~ 00:19:08 .. would work well 00:19:13 that's for ranges 00:19:25 although i'd use ¨ if it was more standard. 00:19:43 errr why should it not be actor based? 00:19:59 because mostly actors are equiv. to functions 00:20:03 and? 00:20:09 well, it makes it pointless :) 00:20:23 by the way, is 3 spaces the standard Oklotalk indentation width? i like that, it's crazy :P 00:20:26 well, everything is equivalent to everything. 00:20:55 err i actually use 3 there :) 00:20:57 errrrr 00:20:58 i mean, 4 00:22:39 do you? damn 00:22:42 3 would be cool 00:23:00 hmmmm, there's some obsolete strictness information there... 00:23:31 can you make oklotalk standard 3-spaces? pleeeez? :P 00:23:39 it says ´(...) != (...) when it comes to strictness, but that would make no sense, since ´ and ` never do but okenity changes 00:23:45 yeah, i can :P 00:23:59 :D 00:25:11 {|...} i think is the dynamically scoped lazily evaluated list... hard to keep up, really ;) 00:25:43 i change the scope types quite often, since there's always something wrong with them... 00:25:47 does oklotalk allow multiple return values? 00:25:54 i'm pondering the best way to implement return values 00:25:54 you can return a list 00:26:04 by the way a while ago you posted a parsing spec, where's that? 00:26:18 F={1 2} returns 1 and 2 00:26:28 ummm, that's actually somewhat lost. 00:26:41 if you recall the date, i can find it though. 00:26:56 ooooh 00:27:26 http://www.vjn.fi/mb/?p=oklotalkspec operators are wrong, and i'm not sure about the rest either. 00:27:45 what is that interface written in? xD 00:28:05 "Nesting basically means making a list of tokens one single token with information about how it was nested." wait, what :) 00:28:09 Comments can be handled in a separate phase before the actual parsing (even before tokenizing) by just replacing every comment with a space (' '). nuh uh what about strings :-) 00:28:30 wtf have i been smoking. 00:29:01 heh, i actually realized that later myself 00:29:14 ¤ 00:29:14 ¨ I CAN'T TYPE THESE :P 00:29:19 the one online isn't the newest, just almost 00:29:23 heh, perhaps :) 00:29:29 nor ´ 00:29:37 ´ is kinda crucial :) 00:29:53 i might make an emacs mode that lets me type that stuff :P 00:30:01 oklotalk-mode.el! 00:30:03 each time you need to use a function as first-class, you need ´ 00:30:15 hahahah imagine trying to syntax highlight 00:30:17 well, just when you actually create a lambda on the fly 00:30:21 :D 00:30:26 hmm, i guess what i could do is write a program in oklotalk which hooks into the parser 00:30:33 -!- puzzlet has quit (Remote closed the connection). 00:30:34 and returns an s-expression parse tree 00:30:35 -!- puzzlet_ has joined. 00:30:40 hehe 00:30:42 so that i can simply call it and use emacs lisp's parser 00:30:48 now, can you let me code for a while? :P 00:30:58 invalid syntax ofc would be a problem but i would just make it highlight them as 'error' regions or something :P 00:31:04 -!- helios_24c3 has quit ("Leaving"). 00:31:11 invalid syntax, what's that? 00:31:43 i've been considering an ide that actually runs the code in the background to check for errors... 00:31:58 -!- jix has quit ("CommandQ"). 00:32:11 oklotalk is definitely not user-oriented 00:32:28 which is kinda lame since it's not really efficient either :) 00:33:09 i'm mainly going for as much theoretical freedoms as possible, conciseness, and as much madness i can fit in. 00:33:19 now, i need to get back to graphica 00:33:21 oklofok: invalid syntax will exist in oklotalk as much as you try and make that not happen 00:33:23 it just will exist 00:33:34 really? 00:33:35 also, what if i wrote a program to delete the hd 00:33:36 :P 00:33:53 oklofok: and yeah, because 'everything makes sense' can not be true without making your language useless 00:33:56 safety is actually easy to do in oklotalk 00:34:26 everything doesn't make sense semantically, but every string will definately a correct oklotalk program 00:34:27 *definitely be a 00:35:09 DSA:P{}AWO)(_)(U)(~*(HE(AJHWEC }}LÅ’Í”ÒÎ{:!@"E:OP$U)($)($)(@£_$(_+ 00:35:15 that's valid syntax? 00:35:18 sure 00:35:23 i can parse that for ya 00:35:28 well, there'll be something that's invalid 00:35:48 oklofok: actually, the nested braces implicitness, please remove that... it means it's impossible to do matching-brace-highlighting 00:35:55 also, typos could cause weird semantic errors 00:36:13 sorry, i wont remove that, one of my favorites ;) 00:36:35 also, what would incorrect parens do without it! 00:36:42 cause an error 00:36:51 also, what if i forget to type a brace and it ends up thinking it's in the wrong place? 00:37:00 and what about matching brace highlighting in all good editors? srsly :< 00:37:10 heh, sorry, i ruin my language as i please :) 00:37:38 answer my questions! :P 00:38:08 well, i don't really care about those :) 00:38:17 not my prob 00:39:19 anyhow, just do code~[\"("´'")"|"{"´'"}"|"["´'"]"] 00:39:36 and you'll know whether it has the correct digits 00:40:17 or something close to that, i don't remember which operator does string matching :P 00:40:47 at least make it runtime-disableable :P 00:40:53 by the way, oklotalk file ext is .ot right? 00:41:01 prolly 00:41:17 i'll make a strict.ot which disables some of the crazier things :-) 00:42:19 the parser is (supposedly...) fully accessible at runtime, and somewhat parametrizable, you can just include a parser at the beginning of your program and start coding "on top of" that 00:42:57 there are multiple syntaxes for oklotalk, mainly differentiating on tokenization 00:43:01 would you reccomend a strict.ot for use or is it anti-Oklotalkian? :-) 00:43:45 if strict is an oklotalk program parsing your code as you write it, i'd say definitely oklotalkian 00:43:57 no 00:44:05 it is an emacs mode? 00:44:12 strict.ot is a file that you include in your oklotalk programs, like perl's "use strict" 00:44:16 ah 00:44:19 it does things like disabling implicit parens 00:44:27 just to make the language a bit saner 00:44:44 sure, the built-ins should provide such parser modes anyway. 00:46:04 but actually using it, is that oklotalkian or do real oklotalkers just deal :P 00:46:43 real oklotalking uses the K philosophy, it's originally designed to be write-only. 00:47:03 the parenthesis thing is partly because of conciseness 00:47:40 -!- cherez has joined. 00:47:59 -!- cherez has left (?). 00:49:21 so real oklotalking = no strict.ot :P 00:49:49 i'd say so 00:50:07 the paren precedence doesn't *always* make things look bad 00:50:27 especially if you add some more complex whitespace enhanced parsing 00:50:49 like out 4 + somfunc(otherfunc(soidf(oier(value 00:51:15 it makes auto-indentation basically impossible 00:51:52 perhaps, but say what you will, it stays, and i gotta code graphica now :P 00:53:58 oklofok: aww come on :P 00:54:01 how about 00:54:06 making the implicitness optional 00:54:10 and turned off by default, but stil lthere 00:54:47 by the way, using pattern matching in oklotalk, how do i match on multiple params? 00:55:16 F={A B -> A*B}; 4 f 2 -> 8 00:55:52 A = { 0 n -> n+1; m 0 -> '(m-1 1) m n -> '(m-1 '(m n-1)) } 00:55:54 correct? 00:56:21 you can spesifically match for a certain arity with F={#1 A B -> A+B;#2 A B -> A*B}; f 2 3 -> 5; 2 f 3 -> 6 00:56:24 hmm 00:56:37 ackermann? 00:56:43 Yes. 00:56:46 Using the definition: 00:56:46 A(0, n) = n + 1 00:56:46 A(m+1, 0) = A(m, 1) 00:56:46 A(m+1, n+1) = A(m, A(m+1, n)) 00:56:48 you need a ; there 00:56:55 apart from that 00:57:04 then, almost :) 00:57:09 you need to have vars uppercase 00:57:19 why 00:57:23 and m-1 1 needs a comma or something. 00:57:42 because lowercase variables are considered funcokens and are used as operators :) 00:57:49 youch 00:57:50 ok :P 00:57:52 A = { 0 N -> N+1; M 0 -> '(M-1, 1); M N -> '(M-1 '(M, N-1)) } 00:57:54 m-1 == (m (- 1)) 00:58:05 M-1 == (- M 1) 00:58:26 how do i mutate a variable in OT? 00:58:31 that's a bit weird, and oklotalk will definitely need a bit special ide anyway... 00:58:48 well, = is the only thing you can actually change a var with 00:58:55 i mean, the contents of a variable 00:59:11 everything else is sent to the object like [oper params] 00:59:16 in a list like that 00:59:19 also, emacs can handle oklotalk, mostly 00:59:26 since it can do turing-complete things for language modes 00:59:36 oklofok: how do i do IO? 00:59:42 out and in 00:59:43 specifically, there is a concept of ports or streams right? 00:59:50 and printing to them is generic? 00:59:58 so a socket has the same interface, etc? 01:00:02 as a file, or stdout 01:00:13 yes, although there is not *byte* in oklotalk, really, well, there is, but you don't see it. 01:00:18 *is no 01:00:30 i mean, any stream will have to handle any object. 01:00:43 uh 01:00:46 that's barely possible 01:00:47 so... no, there isn't a stream in the sense you mean, proly. 01:00:49 *prolly 01:00:52 :) 01:00:54 can you wrap that around a low-level byte based interface? :P 01:01:03 we need to interact with the real world with our programs here 01:01:10 hehe, outn will call str() for the object prolly 01:01:19 well, byte OR character 01:01:23 oklofok: it is unicode, right? 01:01:30 strings are unicode, files are read as unicode, etc 01:01:35 with of course bytes and bytestrings 01:01:36 i haven't actually thought about that, but i think it is. 01:02:34 yaaaay 01:02:43 what about loading files 01:03:16 also, i have the tentatively-named ehirdtalk to implement too :P 01:03:55 :D 01:03:56 fact<-{<3->1=>*$-1} is factorial, iirc 01:04:03 more clearly: 01:04:03 you can add the sanity i lack :) 01:04:11 ah, i recall that 01:04:26 stream / code flow oriented 01:04:35 fact <- { ((_ < 3) -> 1) => (_ * ($ (_ - 1))) } 01:04:37 at least those were the things you explained back when i first saw that 01:04:43 and naw, not really 01:04:50 but the evaluation is weird 01:04:51 just a guess 01:05:03 hmm that should be: 01:05:12 fact<-{<3=>1->*$-1} 01:05:13 i think 01:05:20 well the pattern matching is crazy 01:05:38 as you can see, since -> or => or whatever 'if' is matches on a tuple on the left side 01:05:47 but it doesn't do it by matching on the tuple (which would mean it'd already be evaluated) 01:05:52 it matches on the call to the tuple constructor 01:06:03 that fact is wrong 01:06:08 errr 01:06:11 or am i wrong... 01:06:18 err, you are right 01:06:23 you have <3 in both 01:06:26 fact<-{<2->1=>*$-1} is factorial, iirc 01:06:34 hmm 01:06:53 what implies the implicit _ after => there? 01:07:00 the * binary operator 01:07:10 ah, not everything can be any-ary 01:07:25 should've known 01:07:25 any left-out dangling params are _ 01:07:32 i know 01:07:55 i forgot how i did param names :( 01:09:44 hmm 01:10:00 an actually clear ehirdtalk program: 01:10:02 fib <- { 0 -> 0 => 1 -> 1 => > 1 -> $(-1) + $(-2) } 01:10:03 err 01:10:11 with newlines 01:10:12 fib <- { 01:10:15 0 -> 0 => 01:10:20 1 -> 1 => 01:10:27 > 1 -> $(-1) + $(-2) 01:10:29 } 01:10:42 hmmmm 01:10:46 i don't buy that 01:10:54 - can't be unary? 01:11:11 _ is unary negation 01:11:14 ah 01:11:20 or you could just do 0- 01:11:22 then i do buy that 01:11:37 yeah 01:11:55 though i am going to remove some of that stuff 01:12:00 because fact<-{*1..} should work 01:12:21 maybe ? 01:12:35 fact<-{?<2->1=>?*$?-1} 01:12:43 you could do like K 01:12:48 what does k do 01:13:13 \ is a primitive that does map when preprended to an oper 01:13:17 / does reduce 01:13:20 io think 01:13:22 *i think 01:13:28 it's 3 years since i looked at taht 01:14:13 ooh i know 01:14:15 {} matches on tuples 01:16:47 oklofok: gimme a call/cc in oklotalk 01:18:00 that is: (call/cc f) is (f k), where k is a function with a saved call stack from the point of call/cc, which on call with an argument restores the call stack and returns its argument as the return value of call/cc 01:18:09 also, k is callable multiple times 01:18:16 hmm... @@ may be the current continuation... i recall i had to make a new one because ' was clearly unoverloadable 01:19:52 oklofok: would cps-transformation of oklotalk be possible 01:19:54 you can do `'@#vars to get the current namespace, but without a state list, you can't make a goto without the explicit current continuation @@ 01:20:25 i'm going to make it possible if it's no yet 01:20:28 *it's not yet 01:20:36 everything must be possible 01:20:59 I think I have a BF interpreter now. 01:21:08 cool 01:21:12 oklofok: just make the call-stack accessable 01:21:14 in what language? 01:21:18 But writing the + and - function in full is awfully boring. 01:21:22 Turing machine. 01:21:36 specifically, you can access the current call stack, replace the current callstack, and unwind with a return value to a certain point 01:21:42 then, you can implement continuations in about 3 lines 01:21:59 so basically add a global callstack object that you can assign to, and give it an 'unwind' operator 01:21:59 bf interpreter in what Slereah? 01:22:05 ooh sorry 01:22:07 I didn't notice 01:22:17 cool :D 01:22:24 What does it look like"? 01:22:41 ehird`: yeah, if you can access the exact point where you are at the code 01:22:48 the code pointer to be exact 01:22:59 So far, http://paste-it.net/5460/raw/ 01:23:05 dont need to do that oklofok 01:23:17 i don't? 01:23:18 why 01:23:21 The + isn't complete, there's no - yet, and it's not wrapping so far. 01:23:34 It also isn't in ASCII, but well. 01:23:58 (It's also on a Turing machine with 3 tapes, because I don't want to do a 1 tape version now) 01:24:34 ah, indeed, the call stack would hold the points of return anyway :) 01:24:35 silly me 01:31:31 so, any callcc oklofok ? :) 01:31:45 right, f@@ 01:32:08 (f current_continuation) 01:32:13 (call/cc f) 01:32:41 what's the implicit paramater name in oklotalk 01:32:43 you call-with-current-continuation by calling with the current continuatino 01:32:49 ? 01:32:51 _ 01:33:02 call/cc = {_@@} 01:33:13 almost 01:33:14 feel free to give a more concise definition :P 01:33:33 CallCC = {`_ @@} 01:33:39 oh 01:33:51 actually, that will use the continuation in CallCC :) 01:33:57 i need to ponder about this a bit. 01:33:58 hehe 01:34:25 oh, actually doesn't matter, i think 01:34:33 oklofok: oklotalk booleans 01:34:35 what are they 01:34:38 since callcc will return what was returned anyway 01:34:45 you usually use the atoms #t and #f 01:35:16 hey, schemeish 01:35:17 i like it 01:35:28 errr actually §t and §f... it seems i've added another set of atoms for typing and then mixed them up :o 01:35:59 #<...> should be reserved for stuff you don't actually need to see, but you *can* 01:36:00 ffffffff do i really have to type that char 01:36:06 :) 01:36:09