00:00:03 a continuation does not mutate itself 00:00:09 yes, why couldn't you do that? 00:00:54 How do you implement re-useable continuations on top of non-reusable coroutines? 00:01:16 maybe we're thinking of different ways of defining continuations in coroutines 00:01:19 i'd explain mine but it's long 00:03:32 wikipedia says that coroutines are "best implemented using continuations" 00:04:10 it says nothing about the possibility of implementing continuations using coroutines, although that does seem intuitively impossible 00:05:21 hm 00:05:23 ah! 00:05:57 i don't see why 00:06:04 One could say that co-routines are more of an imperative thing whereas continuations are more of an impure-functionl thing. 00:06:13 sorry, not totally relevent 00:06:26 true, but co-routines could be said to be functional 00:06:38 test_: so how would you implement continuations with coroutines? 00:06:47 lament, the simple way? :P 00:06:50 the obvious way? 00:06:57 its not easy to cram into an irc message 00:07:05 must be not all that simple then :) 00:07:09 i can think of no obvious way 00:07:15 (of course, i'm convinced it's impossible) 00:07:25 SimonRC, to take a quote from lazy k, the computation part of coroutines is an implementation detail... so if you imagine "yield" as a branching statement, they are pure 00:07:32 In the pure functional world we have things like Control.Monad.Cont.ContT, which is fatal to those with an IQ of less that 150. 00:07:52 test_: erm 00:09:03 that amb macro is waaay beyond me 00:09:13 heh 00:09:56 i must've skipped that chapter in sicp 00:13:10 it seems that coroutines are almost exactly "one-use" continuations 00:13:30 so how would you get "multi-use" out of them is unclear 00:17:48 loops 00:18:41 that amb thing is amazing 00:19:17 -!- test_ has quit (Read error: 104 (Connection reset by peer)). 00:22:28 you can write the specification of the problem instead of how to solve it 00:23:06 bsmntbombdood: that's exactly how Prolog already works 00:23:12 consider looking at it 00:23:38 ok 00:24:27 (as usually, scheme shines at being able to emulate any programming paradigm out there) 00:25:09 i believe Haskell has a LogicT monad somewhere :) 00:25:13 i wonder how the speed is 00:25:44 (based on Oleg (TM) work, i think) 00:27:13 btw i have read a couple of times that _delimited_ continuations are essentially equivalent to arbitrary monads. 00:32:27 fucked-up things nobody understands often tend to be equivalent to one another 00:34:11 (define (an-integer-between lo hi) 00:34:11 (if (>= lo hi) (amb) (amb lo (an-integer-between (+ lo 1) hi)))) 00:34:14 that's right, right? 00:34:38 looks right 00:35:25 shouldn't that be > rather than >= ? 00:35:55 i think he's using Python semantics for lo and hi 00:35:58 it won't work with the a-pythagorean-triple-between proc in sicp 00:38:53 hmm 00:39:25 (a-pythagorean-triple-between 1 10) => (1 1 0) 00:39:44 i think i've been missing something in my life until this if scheme macros can actually archieve stuff like that 00:39:49 prolog has 'between' built-in 00:39:51 ?- between(1, 3, N), write(N), fail. 00:39:55 prints "123" 00:47:38 why does sicp implement amb as an extension to eval rather than just a macro? 01:33:25 -!- oklopol has quit (anthony.freenode.net irc.freenode.net). 01:33:25 -!- sp3tt has quit (anthony.freenode.net irc.freenode.net). 01:33:25 -!- meatmanek has quit (anthony.freenode.net irc.freenode.net). 01:33:35 -!- meatmanek has joined. 01:33:35 -!- sp3tt has joined. 01:33:35 -!- oklopol has joined. 01:35:19 @index showIntAtBase 03:26:03 * pikhq is back in Ratpoison. :D 03:29:19 ion ftw 03:32:12 "He plans to release future versions of Ion3 as closed source" 03:33:55 what!?! 03:34:49 Read the Wikipedia page. 03:36:33 http://archlinux.org/pipermail/tur-users/2007-April/004644.html 03:41:43 last time i checked in with tumuo there was an ion4 in the works 03:48:02 anyway, that doesn't make ion any less aweseom 03:53:56 Yes it does. 03:59:20 how? 03:59:35 just because his new code won't be open doesn't make his old code any worse 04:01:09 No, it means that it'll stagnate. 04:01:42 Ratpoison, on the other hand, is still actively maintained (although a bit less actively, since the main author is currently involved in porting it to Common Lisp) 04:02:24 -!- GregorR-L has joined. 04:05:09 i think ion can be extended in lua enough to not need active dev 04:09:06 The dev of Conkeror uses Ratpoison. Since Conkeror rocks, Ratpoison must also rock. 04:09:32 Oh, right. 04:09:51 Shawn Betts writes Conkeror, Ratpoison, *and* StumpWm. 04:09:55 Man, that guy rocks. 04:10:25 O.O 04:10:26 conkeror was much less than i expected 04:10:48 The man has *also* written nethack-el and LiCE (a port of the Elisp engine to Common Lisp). . . 04:11:04 * pikhq bows 04:33:36 -!- immibis has joined. 04:37:13 -!- toBogE has joined. 04:45:21 -!- GreaseMonkey has joined. 04:46:37 -!- immibis has set topic: alin0. 04:46:40 oops 04:47:03 -!- immibis has set topic: Esoteric programming language discussion | FORUM AND WIKI: esolangs.org | CHANNEL LOGS: http://ircbrowse.com/cdates.html?channel=esoteric. 04:47:12 was trying to see if you could /topic a user 04:47:17 sorry 04:50:21 well can you? 04:52:01 no 04:52:10 it sets the topic of the current channel instead 04:52:24 so don't try it unless you don't like the channel you're in 04:52:35 and since you're only on #esoteric... 04:58:18 * pikhq curses 04:58:32 http://www.latimes.com/news/local/la-me-explode27jul27,0,3190584.story?coll=la-home-local Explosion at Scaled kills 2, injures 4 05:00:45 scaled? 05:01:07 immibis: i'm not only on #esoteric 05:01:41 Scaled Composites. 05:01:48 but you're right in that i like this channel 05:01:51 Creator of SpaceShipOne. 05:02:19 why did you dure? 05:02:21 *curse 05:02:41 (curse -> dure... how?) 05:02:48 Magic. 05:03:21 why did you curse? 05:03:30 is that relevant news? 05:03:36 To me, at least. . . 05:03:47 [16:02] ->> oklofok is on :#esoteric 05:03:51 because... you like rockets? 05:03:56 Damned right. 05:04:08 true, [16:02] ->> oklopol is on #haskell #toboge #esoteric-blah @#vjn #osdev #esoteric 05:04:22 immibis: also this one is one multiple channels. 05:04:26 but oklofok isn't 05:04:31 oh yeah it is. 05:04:34 #esoteric-blah? 05:04:43 oerjan: /whois oklofok 05:04:50 oerjan: very-esoteric. 05:05:09 oerjan: he's the only user on that channel 05:05:09 i guess it was created for bot flood not to takeover here 05:05:17 !raw join #esoteric-blah 05:05:45 it often happens channels die out but i don't leave them 05:05:47 in a few months 05:06:51 -!- Sukoshi has joined. 05:07:04 So, I'm guessing a constructor in Java can't throw an exception? 05:09:13 Well, you can. But in this case, I want this constructor to invoke its super's constructor, and the super throws an exception. 05:09:47 But I can't wrap a try/catch around it, because the compiler cries saying it needs the super's constructor to be the first statement in the constructor. 05:09:48 So :| 05:10:20 hmm 05:10:55 ah 05:11:02 Ah? 05:11:14 you can't throw an exception in a constructor and still have the object actually created 05:11:28 Well, I want it to throw another exception. 05:11:30 so they don't let you catch the super's exception because you will never need it 05:11:31 ah 05:11:43 * oklofok ponders some more 05:12:02 Because this class is subclassed by other classes, the meaning of the internal exception has different meanings for the subclass's user based on the subclass used. 05:12:22 So the subclass basically throws a domain-specific version of the exception. 05:15:58 immibis: anyway, i'm also on #scheme 05:16:53 Maybe a minor refactor is neccessary hmm. 05:17:07 Now the interface won't be as nice as I wanted it to be, but it'll work. 05:18:26 yeah, well, interfaces are overrated anyway 05:18:55 Okomol Opilovonal <<< i have to compliment me on my choise for name 05:19:25 public MyClass() throws SuperClassException {super("Parameter");} 05:19:51 no you're not 05:19:57 or is it a secret channel? 05:20:17 yes, i see it's secret 05:21:39 wait, no it's not 05:22:06 when i'm on that channel, /whois says you are on it 05:22:11 when i'm not, /whois says you're not 05:22:18 external messages and colors are disallowes 05:22:19 *d 05:22:25 it's always like that 05:22:36 it's not secret once you're on it 05:22:49 but... i don't see a secret flag on #scheme 05:23:46 immibis: that java code doesn't help, she wanted to catch super's constructor's error in the subclass's constructor 05:24:00 ok... 05:25:11 maybe: private MyClass(int parm1) throws Exception {super(parm1);} public static MyClass create() {MyClass c=new MyClass(0); CONSTRUCTOR CODE; return c;} 05:25:15 messy code though 05:25:16 but it works 05:26:29 I think a refactor is better. 05:26:43 Is ``Releases the value stream points to'' a good comment? :P 05:26:59 Because objects *are* technically references. 05:27:52 Dunno if Java coders understand references/pointers in their lexicon though. 05:28:59 probably ``releases the stream" would be better because java does not distinguish much between the reference and the object it refers to 05:29:22 Alright. 05:29:32 Although when I code, I think of Objects as pointers though. 05:29:38 I guess that's just my C background showing. 05:30:13 oklofok: http://www.cecs.csulb.edu/~hill/cecs497/nestreme/howto.html <-- Havas unun bonan tempon. 05:30:40 Unu artikolo ke eksplenas la metodo pri krei unu emulatoro. 05:32:16 Vi kaj via esperanto. :p 05:33:17 (One article that explains the method of creating one emulator. ?) 05:34:28 Yup. 05:35:10 精神の安心為に、種類が必要と思う。 05:35:30 Question marks for everyone! 05:35:39 Not my fault you don't have UTF-8. 05:36:12 oh, i forgot, i actually thought you were totally bewildered about something :D 05:36:39 Maybe the amount of bewilderment you see on IRC will go down if you get UTF-8 support working :P 05:36:44 Sukoshi: i actually stopped making the emulator because sdl was just too hard for me to download :) 05:36:56 i guess it's just an apt-get... 05:37:02 I'm convinced that UTF-8 support sucks in UNIX. 05:37:03 but they never *really* work. 05:37:14 You need libsdl and libsdl-dev. 05:37:29 How hard is that? :P 05:37:51 not very, if i just ask for them to install themselves and they will 05:38:02 So you can't link to them? 05:38:04 if i actually have to know where to put what, it gets quite hard 05:38:16 i've never done *anything* with multiple files 05:38:19 well 05:38:21 guess sometimes 05:38:33 o_O Do you write a lot of code? 05:38:39 all the time 05:38:47 You fit it into one file?! 05:39:03 err yeah, i don't like having many 05:39:04 ah 05:39:07 *sigh* 05:39:07 ..... 05:39:11 well, one file per application 05:39:14 ..... 05:39:16 oklofok: That ain't right. 05:39:18 :P 05:39:20 That just ain't right. 05:39:41 Good luck having someone else read your code :P 05:39:46 Or making your code even somewhat modular. 05:39:51 * pikhq really, really hates having locale set to en_US.UTF-8 and not having Unicode work at all 05:40:08 UTF-8 works like a beaut on me system har. 05:40:08 oklofok: It's not all that hard. . . 05:40:09 it's pretty modular, i don't see what that has to do with the fact i don't have multiple files 05:40:27 Sukoshi: I've got USE="unicode", a Unicode local, and yet my terminal hates Unicode. 05:40:31 Well, it's considered a Very Bad Thing (TM). 05:40:36 i guess it's not, i just never learned how to link other than one .cpp 05:40:39 or .c 05:40:47 gcc -c -o foo.o foo.c 05:40:51 gcc -c -o bar.o bar.c 05:40:51 except with an ide, but i don't have one in linux 05:40:54 oh 05:40:57 gcc bar.o foo.o -o foobar 05:41:11 are the sdl-things .libs? 05:41:20 If you need to link anthing else, one adds it to the last command line. . . 05:41:27 i know how to use those in visual c++, but i'm guessing gcc does it a bit differently 05:41:33 okay 05:41:39 No, they're in /lib. 05:41:43 then i don't see how i would not succeed 05:41:43 Attach the flag -LSDL IIRC. 05:41:45 huh? 05:41:48 .lib 05:41:53 Rather -Lsdl 05:41:57 library files 05:42:08 No, -lSDL 05:42:14 oklofok: That's not how UNIX works. 05:42:43 i see. 05:42:54 The libraries are under /lib or /usr/lib. . . To link to them, you just do -lLIBRARY_NAME. 05:43:09 And then, the dynamic linker magically takes care of the rest. 05:43:19 C libraries are compiled to object files (.o and .so) and these object files are kept in a directory which is maintained by LD_CONFIG. You give gcc an -l flag and it includes the corresponding lib as neccessary. 05:43:43 oklogok: the -lFILENAME flag to ld (or gcc) will link the library called libFILENAME.so or libFILENAME.a which is somewhere on the search path, into your program. 05:43:53 *oklofok 05:44:07 that is on unix 05:44:19 on windows...not sure 05:44:51 apt-get isn't the way to get libsdl 05:45:08 apt-get install sdl-dev 05:45:15 . . . I think. 05:45:28 nope 05:45:43 apt-cache search sdl 05:45:53 aptitude is better. 05:45:57 True. 05:46:00 Since it resolves package removes in a nicer way. 05:46:06 aptitude search sdl 05:46:09 * pikhq uses Gentoo, anyways 05:46:15 Slackware :P 05:46:26 Another good distro. 05:46:54 i have no idea what to download there 05:47:04 that's a lot of files. 05:47:05 .... Read the package descriptions ... 05:49:03 how do i get the rest of a description? 05:49:30 "Simple DirectMedia Layer development fil" might be it 05:49:52 That'd be it. 05:52:33 "unable to fetch some archives" 05:52:40 guess that's okay 05:52:40 ...? 05:52:45 Uh. . . 05:52:51 aptitude update 05:53:00 When was the last time you did something like that? 05:53:34 that'd be a long time ago. 05:53:38 Ah. 05:53:55 aptitude update&&aptitude upgrade; 05:53:55 "Some index files failed to download, they have been ignored, or old ones used instead." 05:54:02 this, again, is from update. 05:54:02 ... 05:54:11 What did you do? 05:54:53 oh error with connection or something 05:55:39 Fetched 2B in 6s (0B/s) 05:56:48 Do them seperately. ``aptitude update'' ``aptitude upgrade'' 05:57:27 Sukoshi: "aptitude update&&aptitude upgrade" will only continue to aptitude upgrade if the aptitude update returns 0. 05:57:47 Oh. Hmph then. 05:58:02 I don't like BT's weird tracker request thinger. 05:58:20 actually i did "apt-get update", aptitude *did* work 05:58:38 i have no idea what aptitude is... but i guess it's like apt-get? 05:58:44 Bit better. 05:58:48 It's a more full-featured version of apt-get. 06:00:07 -!- oerjan has quit ("leaving"). 06:00:58 How did BT even catch on, I can't understand. It's a mishmash of technologies. 06:01:46 BitTorrent, you mean? 06:01:50 No idea. 06:01:54 Magic, no doubt. 06:03:07 Yep. BitTorrent. 06:03:15 ...is there an alternative? 06:03:35 Distributed IP over Avian Carrier. 06:03:53 It wouldn't be hard to write something cleaner than the BT standard. 06:05:17 i think the reason people use it is that not everyone cares if its standard is ugly 06:05:26 ./ knows 06:05:29 And that there's nothing better, sadly. 06:05:45 is there actually something wrong with torrents? 06:05:49 Writing a BT tracker must be a *royal* pain. 06:06:15 The standard is heavily baroque. Part of it is CGI, part of it is a peer-wire protocol, another is a BT specific bencode protocol. 06:06:40 unicode is evil 06:06:54 No, Unicode is most holy. 06:06:59 * Sukoshi pats the missionary. 06:07:14 Of course, I come from Tcl, where Unicode is par for the course. 06:07:44 The only complaint I've seen articulated correctly against Unicode is Matz's objection. 06:07:45 if you must use unicode, you should at least have the decency to use UTF-32 06:07:55 (seriously: everything is a string, and every string is UTF-8) 06:10:03 varied width encodings-- 06:12:38 incr {ASCII compatible character sets containing more writing systems than any other} 06:17:08 now that i have SDL, i wonder if i should also have sdl.h somewhere 06:17:49 I think it's SDL.h, but I'm not sure. 06:19:48 be it either one of those, i can't find it 06:36:51 bsmntbombdood: what's wrong with them? 06:37:09 just don't like them 06:37:45 would you rather all words in a language be the same length, too? :) 06:39:02 pikhq: i'm learning tcl, and is that legal, incrementing a string? 06:39:07 probably not 06:39:30 don't you need to increment a _variable_? 06:40:21 That *is* incrementing a variable. 06:40:37 You type [incr var], not [incr $var]. 06:40:46 isn't it pretty inefficient having strings only? 06:41:19 oklofok: The interpreter stores different representations of the string. . . 06:41:50 If you're only using a variable as a number, you only are dealing with a C int. . . If you're only using a variable as a list, then you're only dealing with a list. 06:42:11 It gets a different representation when you try to use it differently. 06:43:22 immibis: To dereference that variable, BTW, one would type either ${{ASCII compatible character sets containing more writing systems than any other} or [set {ASCII compatible character sets containing more writing systems than any other}] 06:43:32 s/{{/{/ 06:44:12 so variable names can contain spaces? 06:44:15 Yeah. 06:44:23 Have you read the Tcl man page? 06:44:31 but who would want a variable called "ASCII compatible character sets containing more writing systems than any other" 06:44:31 no 06:44:40 No idea, but it's legal. 06:44:43 Read it. 06:44:47 All 11 rules. 06:44:48 probably just "applicable_charsets" would do 06:44:53 no man here 06:44:55 except for me 06:44:59 no man(1) here 06:45:20 http://www.tcl.tk/man/tcl8.4/TclCmd/Tcl.htm 06:45:45 That's the full syntax and semantics of Tcl. . . Memorize it, and your Tcling will be better off. 06:50:47 thats a clever website name, tcl.tk 06:51:09 Yeah. ;) 06:52:03 i assume it has info on tcl/tk as well? 06:52:39 * immibis would go to the website if he actually wanted to know that. 06:52:53 Yeah, it's the offical Tcl/Tk website. 06:52:56 but immibis doesn't have tcl on this computer 06:53:14 * I don't 06:55:05 Download a Tclkit. 06:55:34 * immibis thinks he's already gone over the "limit" on his "unlimited download" connection though. 06:55:50 which is crazy, really 06:57:13 Very. 06:58:06 -!- toBogE has quit (Read error: 104 (Connection reset by peer)). 06:58:26 you can ask greasemonkey about that, he appears to have the same isp, i have to go now. 06:58:29 bye 06:58:30 -!- immibis has quit ("Give a man a fish and he will eat for a day. Teach him how to fish, and he will sit in a boat and drink beer all day"). 07:26:00 Does TCL have live objects? 07:26:50 "live"? 07:26:57 Like Smalltalk and Lisp. 07:27:08 . . . That doesn't help me. 07:27:12 ... :P 07:27:18 Fully introspective at runtime? 07:27:25 On-the-fly editable? 07:27:33 Oh. 07:27:42 Duh. 07:27:58 That's something it's had from day one. 07:28:27 pikhq: can you fuck around with the execution stack? 07:28:34 That was my next question. 07:28:37 is it accessible as a first-class object of some kind? 07:28:46 Sadly, no. 07:28:50 boo 07:29:03 One can go and pull variables from things up the stack, though. . . 07:29:08 Are blocks/methods first class objects? 07:29:22 Not yet. 07:29:33 ``Not yet'' ? 07:29:35 oh, really? 07:29:44 Planned for Tcl 9. . . 07:29:45 those { blah } things aren't first-claass? 07:29:50 Oh. 07:29:51 just syntax? 07:29:56 Those aren't blocks, those are *strings*. 07:30:09 I misinterpreted you. 07:30:12 but if you can execute strings... 07:30:18 that basically makes them blocks too, doesn't it 07:30:27 since everything's a string anyway... 07:30:28 Yeah. 07:30:35 Those are first-class. 07:30:37 My bad. 07:30:41 okay, so tcl does have blocks :) 07:30:44 Now, if only Tcl had Smalltalk style syntax ... :P 07:31:04 Io is a very nice language, by the way. 07:31:05 can you specify function signatures in tcl? 07:31:33 Could you do me a favor and use terminology I can find via Wikipedia? 07:31:47 Wikipedia's CS articles are woefully bad :P 07:31:54 at least, can you specify how many arguments the function has 07:31:58 Sure. 07:32:10 can you do that for a string? :) 07:32:17 proc foo {arg1 arg2 arg3} {# code here} 07:32:55 what exactly does proc do? 07:33:08 Defines a function. 07:33:23 To call that function: foo arg1 arg2 arg3 07:33:45 proc itself is a builtin function? 07:33:50 Yeah. 07:33:53 what does it do? 07:34:06 does it create a variable called foo and put the function in it? 07:34:10 No. . . 07:34:22 Functions, unfortunately, aren't first class. 07:34:25 :( 07:34:26 :( 07:38:47 *That* is what's planned for Tcl 9. 07:38:47 like ruby! 07:38:47 Sadly, functions can only be passed by reference ATM. 07:38:47 What a lame shortcoming :| 07:38:47 (or, of course, you can do some radical language modification, adding anonymous functions) 07:38:47 pass-by-reference/pointer stuff is so ugly 07:38:47 newMethod := method(x, x + 2) 07:38:47 That was Io. 07:38:47 -!- GreaseMonkey has quit (Connection timed out). 07:38:47 -!- GreaseMonkey_ has joined. 07:38:47 Io is nice but some design decisions seemed silly 07:38:47 proc newMethod x {return [+ $x 2]} 07:38:47 The only decision I think was silly was the decision to nix the ST syntax. 07:38:47 -!- GreaseMonkey_ has changed nick to GreaseMonkey. 07:38:47 Sukoshi: actually yes, thats one of mine 07:38:47 newMethod := [ :x | x + 2] is nicer, IMO. 07:38:47 Io's method calling is much less pretty than smalltalk-style would be 07:38:47 and that 07:38:47 lament: Wanna fork Io to use ST style? 07:38:47 well, why? 07:38:47 ST is there, no? 07:38:47 But most ST implementations ... erm... suck? :P 07:38:47 I suppose the difference is Io is not supposed to be inside a virtual computer 07:38:47