00:00:03 ais523, well yeah 00:00:10 ais523, it does mean a power of two iirc? 00:00:11 next question: I'm interested to know what you might be working on where you can't assume CHAR_BIT = 8 00:00:13 because it sounds fun 00:00:22 AnMaster: no, IIRC at least one system has CHAR_BIT == 9 00:00:31 with 18-bit ints and 36-bit longs 00:01:19 next question: I'm interested to know what you might be working on where you can't assume CHAR_BIT = 8 <-- actually I can, but I spotted a possible correctness error in cfunge's "print out build info" where it prints out printf(..., sizeof(void*) * CHAR_BIT, sizeof(funge_cell) * CHAR_BIT); 00:01:25 but it turned out if was correct 00:01:48 ais523, what would C do on a ternary system? 00:01:53 "not work"? 00:02:02 not define uint16_t nor uint32_t 00:02:07 and would have to do arithmetic as if in binary 00:02:16 ais523, and CHAR_BIT wouldn't be defined? 00:02:17 which would imply padding and all sorts of similar nasty tricks 00:02:22 it would be CHAR_TRIT? 00:02:34 CHAR_BIT would be the log base 2 of the max value of an unsigned char plus 1, just as in binary 00:02:41 it would be incredibly inefficient, of course 00:02:47 heh 00:02:57 ais523, so basically it would emulate a binary system? 00:03:16 yes 00:03:42 ais523, does ternary intercal do the arithmetics in ternary or in binary? 00:03:50 I mean, internally 00:03:56 AnMaster: the INTERCAL writes it as ternary 00:04:05 internally there's a lot of *3, /3, %3 00:04:07 and loops 00:04:12 ah 00:04:18 ais523, well internally was what I meant 00:04:50 ais523, hm I still think .1i should exist :/ 00:05:05 as well as the whole way up to .32i 00:05:09 at least 00:05:17 AnMaster: extrapolating, that would mean there were no operators and all constants would have the value 0 00:05:24 and be an infinite number of digits long 00:05:32 that makes so little sense I might add it 00:05:43 ais523, hm what about .0i? 00:05:47 and .-1i? 00:06:05 the last one is probably about as silly as turkybomb 00:06:07 AnMaster: I think those actually would be impossible, or at least there would be no way to make them happen 00:06:15 actually, base -2 is a lot saner than base 0 00:06:21 ais523, oh? 00:06:27 what on earth is base -2 00:06:29 please tell me 00:06:43 e.g. 8 = (-2^4) + (-2^3) 00:06:50 so 8 is 11000 in base -2 00:06:59 ais523, is that two-complement? 00:07:01 you'll find all positive and negative integers have a unique representation 00:07:07 it isn't two's complement, it's base -2 00:07:18 just define it with the generic equations for arbitrary bases, and you'll find it works 00:07:33 base 0???? 00:07:35 wait, you don't need any special storage for negative...err positive numbers? 00:07:45 quantumEd: I don't believe that makes sense 00:07:49 ais523, so what is -8 in base -2? 00:07:52 AnMaster: you don't need special storage for negative /or/ positive 00:07:55 -8 is just 1000, of course 00:08:02 because it's (-2)^3 00:08:09 what about 9 and -9 then? 00:08:26 9 would be 11001, -9 would be 1011 00:08:32 * AnMaster can't figure out the storage scheme that makes it work 00:08:38 storage?? 00:08:48 well whatever 00:08:56 ais523, so if there are two leading ones it will be positive? 00:09:20 AnMaster: no, depends on odd/even location 00:09:21 however, shouldn't those be "minus ones" 00:09:25 rather than "ones" 00:09:29 and no, those are 1s 00:09:38 ais523, why not minus ones 00:09:38 you could do the same thing with 0 and -1 as the digits, but 0 and 1 is more conventional 00:09:41 and easier to write 00:09:45 hm true 00:09:53 ais523, is .3i balanced ternary? 00:09:55 * Sgeo is a Stargate: SG-1 addict 00:10:12 AnMaster: as you should know, INTERCAL doesn't put any interpretation on bitstrings/tritstrings usually 00:10:25 it's unbalanced ternary for the purposes of WRITE OUT, READ IN, and # though 00:10:37 which are AFAIR the only places you can tell 00:10:41 ah hm right 00:13:36 ais523, anyway .-2i sounds like a nice idea 00:13:53 -!- coppro has joined. 00:14:15 ais523, oh and for .1i I think mingle would be addition 00:14:16 right? 00:14:47 no, because you're mingling two infinite streams of zeros 00:14:50 into another infinite stream of zeros 00:14:55 select likewise doesn't produce useful results 00:15:02 hm right 00:15:20 what about balanced binary? You have +.5 and -.5 00:15:32 (I'm not sure this makes sense at all) 00:16:06 AnMaster: that's binary shifted one bit 00:16:07 ais523, ^ 00:16:19 because halving values is just the same as moving one digit to the right 00:16:19 ais523, no, half a bit? 00:16:33 base 1 00:16:37 wait, +.5 or -.5? oh, that's like binary where each digit is +1 or -1 00:16:42 which means you can't have finite numbers at all 00:16:54 ais523, well, what about balanced binary then? 00:16:59 is it even possible? 00:17:01 you can't do it 00:17:50 ais523, hm can you do balanced for any even base at all? 00:17:57 yes, all the odd ones 00:18:03 XD 00:18:06 also, I think all bases above 2, but it's unbalanced balanced 00:18:08 ais523, please read again 00:18:16 ais523, hm can you do balanced for any even base at all? yes, all the odd ones 00:18:19 as in, you can do base 4 with -1, 0, 1, 2 00:18:19 all the odd even ones? 00:18:34 ais523, eh? How is 4 odd? 00:18:43 AnMaster, please read again 00:18:48 oh 00:18:54 also, I think all bases above 2, but it's unbalanced balanced 00:18:57 right 00:19:24 ais523, what about fractional bases? 00:19:28 or even irrational ones 00:19:51 AnMaster: the issue there is you either have some numbers unrepresentable if you round down with the number of digits 00:19:58 or more than one representation for numbers, if you round up 00:20:09 but in the second case it's often possible to put a restriction on that makes them unique 00:20:18 ais523, is that for irrational you mean? 00:20:20 or for fractional? 00:20:21 base fibonnacci, for instance, is pretty close to base golden ratio 00:20:26 AnMaster: either, doesn't matter 00:20:45 ais523, hm some numbers are unrepresentable in integer bases too. Like pi 00:20:55 so I'm not sure how this makes a difference? 00:20:57 not really, it just takes an infinite number of digits 00:21:06 I mean, you can't even get to within arbitrary accuracy 00:21:13 ah 00:21:17 e.g. if you have digits 0, 1, 2 in base pi, there's no way to get within 0.01 of 3.1 00:21:42 ais523, you could make a fraction in base pi? 00:21:56 well 00:22:00 AnMaster: it would go on for an infinite number of digits if representing a rational 00:22:02 pi-imals 00:22:04 I guess 00:22:10 just like integral bases go onto an infinite number of digits if representing pi 00:22:13 ais523, well sure. 00:22:15 there'd be more than one way to do it, too 00:22:19 assuming digits 0, 1, 2, 3 00:22:23 oh? 00:22:46 AnMaster: imagine if decimal had digits 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 00:22:49 this bit about there being more than one way is interesting 00:22:51 you could represent 10 either as 1 0 or 10 00:22:59 it's the same concept, just less obvious 00:23:17 oh 00:23:19 right 00:23:46 ais523, less obvious indeed. 00:24:26 ais523, I'm not sure this is a bad thing in an esoteric context however 00:35:44 * ais523 signs petition against the law that would give Peter Mandelson a dictatorship over the UK 00:35:54 although the petition wasn't even targeting the law for that reason 01:08:51 -!- Asztal has left (?). 01:08:59 -!- Asztal has joined. 01:09:20 ais523, who is that 01:09:31 AnMaster: a relatively infamous politicain 01:09:31 and what sort of law would give anyone a dictatorship? 01:09:45 AnMaster: a law that let a specific person amend it with minimal safeguards 01:09:59 it's been done in Nomic so many times that it's easy to recognise in actual law 01:10:17 ais523, is that even allowed by the constitution? 01:10:30 AnMaster: the UK doesn't have a written constitution 01:10:44 and given that it's a new law, it would just be amending all the bits that prevented that happening normally 01:10:49 new laws take precedence over old 01:11:07 (although the UK has a constitution formed by laws and precedences, there's no special restriction on amending it) 01:11:14 heh 01:12:31 ais523, such a suggested law would surely create lot of criticism in newspapers? 01:12:36 -!- puzzlet has quit (Remote closed the connection). 01:12:41 -!- puzzlet has joined. 01:12:43 and other media 01:12:55 AnMaster: except that it's part of an anti-copyright-infringement law 01:13:03 and I suspect the newspapers want the law itself to pass 01:13:09 ... 01:13:32 for some reason politicians seem unaware of the possibility that you don't have to accept or reject an entire law, you can amend bits 01:13:56 ais523, what about EU level? Surely it would somehow be stopped there? 01:14:11 possibly, if they notice 01:14:15 it might be worth telling them 01:14:29 ais523, yes exactly. Go do that? 01:15:33 ais523, also, might be worth emigrating somewhere. Noway or Switzerland maybe? 01:15:38 (neither are in EU iirc) 01:15:42 AnMaster: I'm not planning to emigrate 01:16:02 and I think that given how mostly unpopular Mandelson is, if he started passing arbitrary laws the government would just get overthrown by force 01:16:27 ais523, how did he managed to become elected then 01:16:34 if he is so unpopular 01:17:02 he wasn't 01:17:05 he was appointed to the house of Lords 01:17:21 ais523, what does that mean. I forgot how UK politics work 01:17:31 there's two houses of parliament 01:17:36 the commons, which is completely elected 01:17:37 right 01:17:41 and the lords, which is completely unelected 01:17:45 um 01:17:47 traditionally, it was hereditary 01:17:51 and nowdays? 01:17:53 as in, if your parents were lords, so were you 01:17:55 nowadays, it's a mess 01:18:04 meaning? 01:18:08 nowadays you pay 01:18:18 there was a huge scandal recently where Lords memberships were allegedly paid for 01:18:30 that's no democracy... 01:18:44 also, it's generally accepted that the government can get anyone into the Lords they want to 01:18:49 if they need them to have political power somehow 01:18:54 AnMaster: yes, and not intended to happen either 01:18:56 the police got involved 01:19:14 ais523, I mean the house of lords in general is undemocratic 01:19:36 yes, but it's strangely often saner than the commons 01:19:39 having such a thing should disqualify UK as being counted as a democracy 01:19:55 at least hereditary lordships tended not to have an agenda 01:20:02 and so actually made decent decisions as a result 01:20:05 ais523, 01:20:41 anyway, the lords is powerless in theory because laws are passed only if both houses agree, or if the commons makes the same decision three times it overrides the lords whatever they say (remarkably recent rule, that one) 01:21:39 heh, technically he's Baron Mandelson now 01:21:44 I think the whole hereditary thing got reversed 01:21:52 in that you get a title when you become a Lord, rather than the other way round 01:23:14 "Democracy is the worst form of government except for all those others that have been tried." strangely enough http://en.wikiquote.org/wiki/Winston_Churchill does *not* list it 01:24:05 ais523, does that include people like famous artists and such that become "sir"? 01:24:17 no, that's knighthood 01:24:25 it's an honour given by the Queen, for the purpose of honouring people 01:24:36 ah. so it is not related? 01:24:59 yep, unrelated 01:25:32 in the traditional Swedish system they were related (iirc), thus the confusion 01:25:40 hmm... ehird would probably be highly amused to know that Lord Mandelson was the Baron of Hartlepool 01:25:45 I mean, Hartlepool! 01:25:49 Hartlepool? 01:25:52 never heard of it 01:26:13 would have been funnier if it was that small place where ehird lives though 01:26:26 oh no 01:26:31 I doubt there even is a Baron of Hexham 01:27:08 ais523, is Hartlepool a small place? 01:27:10 wow, there was as well (isn't one currently) 01:27:16 AnMaster: it's large enough that people have heard of it 01:27:19 but small enough that they can't remember wh 01:27:21 *why 01:27:26 baron wasn't a very high kind of noble 01:27:30 baron was line manager of nobles 01:27:59 ais523, what about the new place where ehird moved though? 01:28:03 * AnMaster forgot the name of it 01:28:04 Aubrey Geoffrey Frederick Rippon, Baron Rippon of Hexham 01:28:09 AnMaster: it's allegedly a very small village 01:28:16 ais523, yeah 01:28:25 ais523, so there should be a duke of it 01:28:28 just because 01:28:51 (isn't duke one of the higher ones?) 01:28:59 I think so, not sure 01:30:02 duke is the highest you can get without being a king or prince or something 01:30:24 princes tend to collect duchies 01:31:53 by the way. what exactly is the difference between an emperor and a king? 01:31:59 apart from the name 01:34:17 emperors are supposed to be above kings 01:34:30 emperor is a sort of king-of-kings type arrangement 01:34:39 Pthing, so you could have several kings under a single emperor? 01:34:45 yes, plausibly 01:35:02 the emperor may be those kings, however 01:35:08 Pthing, I thought both were "head of state" technically though? 01:35:13 Pthing, hah! 01:35:37 HEY GUYS would you mind if I asked whether I could ask a question regarding permission to inquire about whether I may ask a question? 01:35:41 then you have the situation of like the british empire 01:35:58 where victoria became empress when she became Empress of India 01:36:05 in addition to being Queen of the United Kingdom 01:36:15 Gregor, I did that before 01:36:15 duh 01:36:24 Gregor, I think it was the same number of iterations 01:36:33 * oklofok has changed the topic to: don't ask to ask. http://tunes.org/~nef/logs/esoteric/?C=M;O=D 01:36:33 * adam_d__ is now known as adam_d 01:36:33 * FireFly (n=firefly@1-1-3-36a.tul.sth.bostream.se) has joined #esoteric 01:36:33 oklofok, may I ask if I can ask to ask? 01:36:38 Gregor, there ^ 01:37:09 I have four, you have three. 01:37:12 SO I WIN. 01:37:55 Gregor, I was faster. So I win 01:38:17 Your girlfriend told me that, but she elided the "I win" part. 01:38:33 (See what I did there hyuk hyuk) 01:38:38 Gregor, you mean she just said "so"? 01:38:59 I guess she was talking about dynamic linking then. 01:39:19 WELL PLAYED SIR 01:41:33 -!- Pthing has quit (Remote closed the connection). 01:42:15 Gregor, I don't play. I'm always deadly serious. 01:42:49 Funny, your girlfriend also said that. (OK, this one doesn't work as well but gimme a break, you're not giving me much material) 01:45:17 Gregor, nor is she. Due to not existing. 01:53:00 night → 02:17:45 -!- oerjan has joined. 02:22:25 -!- coppro has quit (Remote closed the connection). 02:23:47 -!- coppro has joined. 02:29:54 -!- Asztal has quit (Read error: 110 (Connection timed out)). 02:49:11 -!- cal153 has quit. 03:09:40 -!- puzzlet_ has joined. 03:21:56 -!- puzzlet has quit (Read error: 110 (Connection timed out)). 03:31:49 -!- coppro has quit (Remote closed the connection). 03:32:23 -!- coppro has joined. 03:53:07 -!- oerjan has quit ("leaving"). 04:02:37 -!- quantumEd has quit ("Leaving"). 04:03:43 -!- GreaseMonkey has joined. 04:05:18 -!- coppro has quit (Read error: 54 (Connection reset by peer)). 04:07:23 -!- coppro has joined. 04:15:59 so 04:16:01 sup guys 04:16:04 hows it goin 04:22:35 fine 04:23:10 i dont know you. :| 04:23:38 augur distrusts everyone he doesn't know. 04:23:49 unless hes hot and has a big cock. 04:24:36 So long as you've got your priorities straight. Or, y'know, not so straight. 04:26:05 so i learned a bit today about dependent types 04:26:16 Goodie. 04:26:19 and ive decided that types should be first class, just for fun 04:26:42 So, prototypes. 04:27:43 no what 04:27:48 'cuz I wurve prototypes. 04:28:08 -!- Sgeo_ has joined. 04:28:27 Prototypes, the brief: Every object has a prototype. That prototype is another object. To look up a method, you first check if the object itself has it, then if it doesn't, check its prototype (and its prototype, and its prototype, etc etc) 04:28:37 i know what prototypes are 04:28:40 Oh 04:28:44 but i mean real first class types 04:28:46 -!- Sgeo has quit (Read error: 104 (Connection reset by peer)). 04:28:46 I figured you didn't 'cuz you said "no what" :P 04:28:55 Just not typish enough for you? 04:29:00 no 04:29:05 i mean like a real type system 04:29:09 Right. 04:29:18 i eman, your junk could be prototypes, i dont care, thats details 04:29:23 i just like the idea of stuff like 04:29:25 oh, say 04:29:40 functions from types to functions over that type 04:30:02 Fair enough. 04:30:08 like in system f 04:30:49 \\a -> \x :: a -> x 04:31:00 which takes a type and returns an identity function for that type 04:31:08 or dependent types 04:31:09 Yesh. 04:31:09 like 04:31:22 \n -> List n 04:31:33 which takes a number and returns the type of lists of length >= n 04:31:42 or just type-to-type functions 04:32:01 its cool 04:32:04 dependent types especially 04:32:39 augur: I've wanted to make a language with first-class types for a while 04:32:40 i had this idea of constructing a programming language that would be able to catch all sorts of errors by doing a sort of back-calculation of what might go wrong 04:33:20 Though I wasn't looking at a functional approach 04:33:28 e.g. it would look at every single use of a function and say, ok, suppose this spits out an error, lets trace back where all the values could come from that are related to that error 04:33:50 coppro: when you think about it, Gregor is kind of right; prototypes are first class types 04:33:56 so are Class objects in Ruby and Smalltalk 04:34:07 but its hard to define new Class objects derived from old ones 04:34:23 There are lots of ways of making types first-class that aren't prototypes, and even have a prayer of static analysis :P 04:34:26 I just like prototypes :) 04:34:28 Prototypes in what sense? 04:34:32 and its hard to define a prototype that is "a thing that is not this other thing" 04:34:58 unless ofcourse your prototype logic has some interesting fundamentals 04:35:03 coppro: Prototype-based object systems. 04:35:28 like a propositional aspect where you can construct a new object and make assertions about it such as opposite-of 04:35:43 or not-equal-to 04:36:01 which sort of derives onto everything else regardless of whether or not they were explicitly prototyped off this object 04:36:26 hmm.. I don't know of this concept 04:36:29 link? 04:36:34 or better, language? 04:36:46 which concept? 04:36:49 prototype OO? 04:36:51 yeah 04:36:53 Self, JavaScript 04:36:56 oh, right 04:36:57 javascript does it poorly tho 04:36:58 urgh 04:37:10 think of it like this: object literals + copy 04:37:10 now that you give me an example, I know what you mean 04:37:35 speaking of 04:37:44 Yeah, JavaScript is a suckfest :) 04:37:48 But I love it. 04:37:51 ive been trying to think of ways to construct a prototype logic 04:40:15 Plof takes prototypes around another twist. 04:40:19 But that's beside the point. 04:43:35 gregor, wanna work on a prototypal logic? 04:43:57 Perhaps. 04:50:14 i want to try to construct a form of logic that employs prototypes rather than the normal set of inference tools 04:50:17 dunno how it'd work but 04:53:42 Inference over prototypes? You could probably dance in that direction with something involving abstract interpretation, but inference is usually most successful in languages with limited state (that is, functional languages), whereas prototypes are basically pointless without state, so hmmmm. 04:54:03 Not that it's unresolvable, 'ts just an uphill battle. 04:54:48 no no by inference i mean like 04:55:08 logics usually let you do various sorts of calculations and stuff 04:55:14 of interesting sorts 04:55:31 and prototypes can potentially let you do different sorts of calculations 04:56:01 an example is stuff like typical vs. atypical behaviors 04:56:33 unless otherwise specified, you can infer that any given dog will bark, have fur, four legs, etc. 04:56:35 Heh, totally different direction. 04:56:35 because the prototypical dog does 04:56:41 Yesh. 04:56:53 and having three legs doesnt make something not a dog, it makes it an atypical dog 04:57:28 im just thinking of how to capture facts about human cognition and the problems of traditional necessary-and-sufficient-conditions definitions of things 04:58:21 Well, presently I'm thinking of how to sleep. 04:58:26 I believe the first step is to go to bed. 04:58:29 So baheeeeeeee 04:58:38 :P 04:58:40 night 05:41:22 -!- GreaseMonkey has quit (Client Quit). 06:14:45 -!- puzzlet has joined. 06:15:24 -!- puzzlet_ has quit (Remote closed the connection). 06:41:35 -!- Sgeo_ has quit (Read error: 110 (Connection timed out)). 06:59:20 -!- cal153 has joined. 07:59:59 -!- clog has quit (ended). 08:00:00 -!- clog has joined. 09:09:01 -!- kar8nga has joined. 10:00:10 -!- kar8nga has quit (Read error: 54 (Connection reset by peer)). 10:06:42 -!- puzzlet has quit (Read error: 60 (Operation timed out)). 10:07:50 -!- BeholdMyGlory has joined. 10:08:38 -!- puzzlet has joined. 10:26:39 -!- ais523 has quit (Read error: 54 (Connection reset by peer)). 10:33:21 Prototypes, the brief: Every object has a prototype. That prototype is another object. To look up a method, you first check if the object itself has it, then if it doesn't, check its prototype (and its prototype, and its prototype, etc etc) <-- what about the top prototype? 10:36:56 Gregor, when did you begin spelling "yes" as "Yesh" instead? 10:46:42 there's obviously an object that's the prototype of the first ones 10:47:01 which is its own prototype, in case the designers wished to make it follow the same pattern. 10:47:22 i mean not that i know anything about prototyping 10:50:40 augur: in a language of mine i had a sort of prototyping using "Cat is Dog but ...", most things were just data, because it was used for making maps for a game, so it was nice to be able to give a few properties things usually have and just say but if one is slightly different 10:50:48 dunno if that's at all close to what you meant 10:57:19 really i just wanted to mention my but 11:12:33 oklofok, the problem with an object being it's own prototype is that you can get an infinite loop when looking up a method that turns out to not exist. 11:12:39 like, if you made a typo 11:13:00 unless you do some sort of cycle detection 11:13:43 I'm not sure how bad the added overhead for cycle detection would be, but calling a method would probably be a very common operation. 11:14:49 oklofok, prototyping for game world objects isn't uncommon. Seen that several times before in MUDs and similar 11:35:29 -!- rodgort has quit (Client Quit). 11:35:40 -!- rodgort has joined. 11:42:16 -!- kar8nga has joined. 11:44:54 AnMaster: yeah it'd be really hard to special case that one object 11:45:12 oklofok, that's cheating! 11:45:49 oklofok, because then it wouldn't have itself as prototype *in practise* 11:46:25 practice* 11:46:50 also cycle detection can be done with O(1) overhead in non-threaded code, just take some number that represents "now", and tag all the things you check for methods with it, cycle = now-tagged prototype 11:47:23 we just need one tag field for each object 11:47:32 oklofok, you would need to untag afterwards, to prepare for next lookup? 11:47:36 no 11:47:40 oklofok, how comes? 11:47:52 that's why "number representing now", and not "boolean" 11:48:17 (boolean would still be O(1), i just knew you'd wine about this) 11:48:25 oklofok, well okay, that would break if system clock was corrected by something like ntp 11:48:51 yes, if you're a fucking retard and actually use current time 11:48:53 well I guess you could use a incrementing counter 11:48:57 yeah 11:49:03 you could not be a retard 11:49:08 sorry, a bit edgy today 11:49:10 :P 11:49:15 eating time! -> 11:49:15 still, it isn't O(1) in threaded code right? 11:49:34 well umm, you'd probably have a lock when checking for methods anyway. 11:49:37 threaded here meaning "multitasking" not "forth threading" 11:49:40 oklofok, why? 11:49:41 yeah multi 11:50:02 oklofok, just make the GC lock, you could have concurrent lookups 11:50:17 you can use multi-read/single-write lock 11:50:32