2012-10-08T00:15:33 *** mceier has quit IRC (Quit: leaving) 2012-10-08T01:31:21 *** grc has joined #aichallenge 2012-10-08T01:53:55 *** mpstyler has joined #aichallenge 2012-10-08T02:00:33 *** coeus has joined #aichallenge 2012-10-08T02:27:01 *** mceier has joined #aichallenge 2012-10-08T02:28:33 *** pairofdice has joined #aichallenge 2012-10-08T02:49:46 *** mpstyler_ has joined #aichallenge 2012-10-08T02:51:29 *** mpstyler has quit IRC (Ping timeout: 245 seconds) 2012-10-08T02:51:39 *** mpstyler_ is now known as mpstyler 2012-10-08T03:08:27 *** mlp has joined #aichallenge 2012-10-08T03:11:46 *** coeus has quit IRC (Ping timeout: 240 seconds) 2012-10-08T04:01:57 *** mviel has joined #aichallenge 2012-10-08T04:28:30 *** replore has joined #aichallenge 2012-10-08T04:44:45 *** sigh has joined #aichallenge 2012-10-08T05:23:32 *** replore has quit IRC (Remote host closed the connection) 2012-10-08T05:33:22 *** replore has joined #aichallenge 2012-10-08T05:38:47 *** replore has quit IRC (Remote host closed the connection) 2012-10-08T05:58:11 *** bqf has quit IRC (Ping timeout: 256 seconds) 2012-10-08T05:59:01 *** bqf has joined #aichallenge 2012-10-08T06:05:35 *** mleise has joined #aichallenge 2012-10-08T06:10:57 *** bqf has quit IRC (Changing host) 2012-10-08T06:10:57 *** bqf has joined #aichallenge 2012-10-08T07:27:01 *** epicmonkey has joined #aichallenge 2012-10-08T07:29:27 *** grc has quit IRC (Quit: Leaving) 2012-10-08T07:30:18 *** replore has joined #aichallenge 2012-10-08T08:59:17 *** Scooper has joined #aichallenge 2012-10-08T09:04:18 *** mpstyler has quit IRC (Read error: Connection timed out) 2012-10-08T09:11:47 *** InsaneMalkavian has joined #aichallenge 2012-10-08T09:18:49 *** replore has quit IRC (Remote host closed the connection) 2012-10-08T09:23:28 *** replore_ has joined #aichallenge 2012-10-08T09:25:07 *** sigh has quit IRC (Remote host closed the connection) 2012-10-08T09:33:55 *** mpstyler has joined #aichallenge 2012-10-08T09:40:30 *** replore_ has quit IRC (Remote host closed the connection) 2012-10-08T09:51:20 *** mpstyler has quit IRC (Read error: Connection timed out) 2012-10-08T10:07:36 *** thestinger has joined #aichallenge 2012-10-08T10:28:52 *** mpstyler has joined #aichallenge 2012-10-08T10:31:28 *** UncleVasya has joined #aichallenge 2012-10-08T10:41:01 mleise, there is a funny spam message in moderator queue. ;) 2012-10-08T10:47:25 *** mceier has quit IRC (Quit: leaving) 2012-10-08T10:51:13 *** replore has joined #aichallenge 2012-10-08T10:55:26 *** replore has quit IRC (Ping timeout: 240 seconds) 2012-10-08T10:57:48 what where? 2012-10-08T10:58:18 *** monstaokki has joined #aichallenge 2012-10-08T10:58:26 *** monstaokki has left #aichallenge 2012-10-08T10:58:38 i can't read cyrillic 2012-10-08T10:59:02 oh an online drug store 2012-10-08T11:01:47 mleise: about lisp 2012-10-08T11:02:12 now the list is empty 2012-10-08T11:02:27 but I wasn't fast enough to read anything about lisp 2012-10-08T11:07:53 *** mpstyler has quit IRC (Read error: Connection timed out) 2012-10-08T11:08:31 mleise: http://i25.fastpic.ru/big/2012/1008/67/2a97cdac67e9a0781a2c7f0f7c4a9467.png 2012-10-08T11:11:17 After you make user account... Operating system Windows XP or better. Thank you. :) 2012-10-08T11:12:31 the new programming language /better than LISP language because contain the lisp source and C+ scripts/ 2012-10-08T11:12:33 I'm sold 2012-10-08T11:12:37 I found it odd because I neve thought lisp is a good theme for spam messages :) 2012-10-08T11:13:14 that's right. and even if i was a lisp programmer I'd hardly understand what they want to sell 2012-10-08T11:13:50 *** replore_ has joined #aichallenge 2012-10-08T11:14:06 besides, no lisp programmer would ever click on such spam links 2012-10-08T11:14:29 right, lisp programmers can't click on link 2012-10-08T11:14:38 because they use a wget daemon that emails web pages to themselves 2012-10-08T11:26:13 *** mcstar has joined #aichallenge 2012-10-08T11:26:57 *** aarossig has quit IRC (Changing host) 2012-10-08T11:26:57 *** aarossig has joined #aichallenge 2012-10-08T11:27:01 emacs has a browser fyi 2012-10-08T11:27:55 no operating system is complete without a browser these days 2012-10-08T11:33:39 these russians really know how to throw a programming language together 2012-10-08T11:33:49 remember that last one they made? 2012-10-08T11:34:06 the one that became a success 2012-10-08T11:40:38 mcstar: turns out you actually have to _think_ to write a sudoku solver 2012-10-08T11:40:49 nowaaaay! 2012-10-08T11:41:13 yeees. I thought I'd just strike out the wrong options and end up with a solution 2012-10-08T11:41:22 mleise: i only went as far as thinking about the representation 2012-10-08T11:41:40 mcstar: and that is? 2012-10-08T11:42:08 well, 2d array, 9x9 obviously, of 9 element doubly linked of numbers form 1-9 2012-10-08T11:42:26 and i can eliminate numbers from that as the contraints go 2012-10-08T11:42:29 that->those 2012-10-08T11:42:35 Mine is now: for each digit (1-9), for each aligment (row, column, block), store a bitmask with one bit for each position where that digit can be 2012-10-08T11:42:57 but then i thought, maybe a 2^n representation would be better, to easily sum rows/columns/blocks, to check for correctness 2012-10-08T11:43:04 hmm, yes your's is way easier to update 2012-10-08T11:43:11 *** mlp has quit IRC (Quit: Page closed) 2012-10-08T11:43:22 if something changes I have to compute coordinates and touch 3 lists 2012-10-08T11:44:02 I haven't looked at other people's solutions 2012-10-08T11:44:08 me neither 2012-10-08T11:44:18 except that quick glance of your pastes 2012-10-08T11:44:20 I guess quite a few of the short ones are brute force recursive iterations 2012-10-08T11:44:36 obfuscated code isn't readable anyway ;) 2012-10-08T11:44:51 i hope i can use randomness in the solutino 2012-10-08T11:45:04 randomness? O.o 2012-10-08T11:45:07 yeah 2012-10-08T11:45:13 some random algorithm 2012-10-08T11:45:15 it is a deterministic problem though 2012-10-08T11:45:28 thats how i solved antimatroid's problem too :) 2012-10-08T11:45:55 i was too lazy to solve a certan subproblem, and i decided i just randomly flip something, and check for a criterion 2012-10-08T11:46:00 so you would randomly try a digit in a slot and see how far you get? 2012-10-08T11:46:03 turned out to be fast and conscise 2012-10-08T11:46:23 that sounds like an evolutionary algorithm or so 2012-10-08T11:46:34 mleise: idk, really, havent ironed out the details 2012-10-08T11:46:44 and i dont think ill be thinking about it 2012-10-08T11:46:58 when ill get tired with physics, ill take a crack at it 2012-10-08T11:49:33 mleise: http://sprunge.us/dNgF 2012-10-08T11:50:29 O_o 2012-10-08T11:50:41 the outer while just repeats the problem 2012-10-08T11:50:43 not haskell? 2012-10-08T11:50:55 the inner while tests for correctness, i.e when the problem is solved 2012-10-08T11:51:26 and the bottom just does a manipulation, that can be done in several ways, and the random number decides what way 2012-10-08T11:51:32 which way 2012-10-08T11:51:39 lol 2012-10-08T11:51:42 UncleVasya: he likes cpp :( 2012-10-08T11:51:45 what? 2012-10-08T11:51:49 sry 2012-10-08T11:51:52 mleise: ^^ 2012-10-08T11:52:09 i pressed tab on u 2012-10-08T11:52:14 instead of m 2012-10-08T11:52:26 my unconscious mind want to talk to vasya 2012-10-08T11:52:29 wants* 2012-10-08T11:53:31 antimatroid should switch to D. I don't know why I couldn't convince him yet 2012-10-08T11:54:04 lol 2012-10-08T11:54:10 he should have switched to haskell 2012-10-08T11:54:18 i dont know why i couldnt convince him yet 2012-10-08T11:54:24 no he seems to like the C syntax 2012-10-08T11:54:33 but he is a mathematician! 2012-10-08T11:54:38 he likes haskell 2012-10-08T11:54:42 he just doesnt know it 2012-10-08T11:54:59 he's just not open minded enough 2012-10-08T11:55:05 mleise: seriously, how would he benefit from D? 2012-10-08T11:55:24 because it wants to be a better C++ ? 2012-10-08T11:55:36 built-in arrays and associative-arrays 2012-10-08T11:55:39 how would he benefit from it? 2012-10-08T11:55:40 garbage collection 2012-10-08T11:55:50 gc is an improvement 2012-10-08T11:55:53 no shitty zero-terminated strings and unicode 2012-10-08T11:56:15 if you are a programmer for a living, those might matter 2012-10-08T11:56:16 mleise: indexing by code point is wrong and doesn't help with doing anything 2012-10-08T11:56:24 if it doesn't index by grapheme, it's not a useful feature 2012-10-08T11:56:26 meta programming up to the point where you can literally load a script file at compile time 2012-10-08T11:56:39 mleise: not interesting for him 2012-10-08T11:57:05 he could benefit from python and haskell, imho 2012-10-08T11:57:15 mleise: C++ is useful because you can generate fast code that doesn't depend on gc 2012-10-08T11:57:30 that's the baseline for a C++ replacement 2012-10-08T11:57:34 if it doesn't do that, it doesn't fit the same niche 2012-10-08T11:57:40 D doesnt depend on it, does it? 2012-10-08T11:57:47 all the features do 2012-10-08T11:57:55 thestinger: Ok there is no grapheme support, but how many languages offer that? 2012-10-08T11:58:06 mleise: why does that matter? 2012-10-08T11:58:12 you say it's better than what we already use 2012-10-08T11:58:32 indexing by code point is helpful in exactly 1 way - to implement a unicode library 2012-10-08T11:58:38 it's never what you want in your actual code 2012-10-08T11:58:48 delete a character - you want to delete a grapheme 2012-10-08T11:58:57 display a grid of characters - it's a grid of graphemes 2012-10-08T11:59:05 many languages suffer from bad unicode support, I think D is an improvement over what C offers 2012-10-08T11:59:17 C offers indexing by char32_t, it's not helpful 2012-10-08T12:00:08 Go has graphemes in the stdlib, as does perl 2012-10-08T12:00:37 D doesn't depend on the GC, but much of Phobos and actually dynamic arrays and string allocations depend on the GC, mcstar 2012-10-08T12:00:49 ok 2012-10-08T12:01:00 i can live with that 2012-10-08T12:01:09 i dont use D XD 2012-10-08T12:02:07 Also the data types are standardized: int is 32-bit, long is 64-bit, and the unsigned types are just a prefixed 'u' 2012-10-08T12:02:09 also, none of you took a big leap from your usual languages 2012-10-08T12:02:20 nope :p 2012-10-08T12:02:34 leaving Delphi was the biggest leap 2012-10-08T12:02:34 mleise: no one stops you from using uint32_t in C 2012-10-08T12:02:44 or the same things in go/rust/haskell 2012-10-08T12:02:59 thestinger: sure, but I have to include int.h or something, right? 2012-10-08T12:03:15 mleise: and? 2012-10-08T12:03:28 and it looks ugly 2012-10-08T12:03:28 you usually include stdio.h too 2012-10-08T12:03:45 mleise: a smaller language and more stuff in libraries is a good thing 2012-10-08T12:03:56 how often do you need a fixed width integer? 2012-10-08T12:04:14 how often do you need at least x bits? 2012-10-08T12:04:25 often, and you don't need any includes for that 2012-10-08T12:04:31 it doesn't complicate the language. there aren't more types than in C 2012-10-08T12:04:49 long is at least 32-bit 2012-10-08T12:04:55 long long is at least 64-bit 2012-10-08T12:05:14 great, we were trying to determine, from what language would antimatroid benefit, and look where we got 2012-10-08T12:05:18 shame on you both 2012-10-08T12:05:19 and when you want to store a struct in a file, you are fucked 2012-10-08T12:05:31 mleise: sure and many languages have modules 2012-10-08T12:05:37 making a language that sucks less than C or C++ isn't hard 2012-10-08T12:05:38 antimatroid probably wants to serialize integers ^^ 2012-10-08T12:05:45 making a language that fits the same niche and sucks less is apparently very hard 2012-10-08T12:06:20 niche is probably not a term you can use with C++ 2012-10-08T12:06:35 fool me once, shame on you, fool me twice.. shame on.. you cant get fooled again! 2012-10-08T12:07:11 There are probably few areas where C++ is not used professionally by someone. 2012-10-08T12:07:50 you dont teach children c++ 2012-10-08T12:07:58 and why would they use D over another language? what production software is written in D? 2012-10-08T12:08:08 (pls god, make that statement true!) 2012-10-08T12:08:29 probably dmd is written in D 2012-10-08T12:08:48 mcstar: no, it couldn't be tacked onto other compilers written in C++ then 2012-10-08T12:09:02 it could 2012-10-08T12:09:06 llvm bootstraps itself 2012-10-08T12:09:16 ghc bootstraps itself 2012-10-08T12:09:21 rust bootstraps itself 2012-10-08T12:10:11 That must be magic. All I've heard is that the communication between front end and back end is easier when the front-end is in C++ too 2012-10-08T12:10:54 * mcstar doesnt have a good boot 2012-10-08T12:11:15 i think sbcl can builds itself too 2012-10-08T12:11:16 also does C++ still lack foreach with C++11 ? 2012-10-08T12:11:38 at least, i see a lot of CL compilation, when i build it XD 2012-10-08T12:11:47 for_each? 2012-10-08T12:11:55 there is for_each 2012-10-08T12:11:58 isn't that a boost macro? 2012-10-08T12:12:01 but that takes a funtion, a lambda 2012-10-08T12:12:03 for example 2012-10-08T12:12:13 not boost 2012-10-08T12:12:19 it came from there probably 2012-10-08T12:12:23 like most of the good stuff 2012-10-08T12:12:26 ah, you mean a functional style "apply this function to each" ? 2012-10-08T12:12:31 yeah 2012-10-08T12:12:38 but the new for is better 2012-10-08T12:13:02 for(auto& class: container){ 2012-10-08T12:13:08 for exmaple 2012-10-08T12:13:32 *** thestinger has quit IRC (Ping timeout: 246 seconds) 2012-10-08T12:13:58 strcat lost his battle :D 2012-10-08T12:14:10 *** thestinger has joined #aichallenge 2012-10-08T12:14:21 hey, thestinger. long time no see 2012-10-08T12:14:29 wifi sucks 2012-10-08T12:14:50 make sure noone else is using a channel close to yours 2012-10-08T12:15:04 can't, there are like 60 wifi networks 2012-10-08T12:15:14 O.O 2012-10-08T12:15:16 none of the others are 5GHz N though 2012-10-08T12:15:49 I wonder if 5GHz would even be allowed in Germany 2012-10-08T12:16:21 probably, it's just a specific part of that range 2012-10-08T12:16:31 they might have a different part allocated 2012-10-08T12:16:51 like you can't use 2.4GHz channel 12 and 13 in north america 2012-10-08T12:16:55 shops sell antennas for it, so it must be legit 2012-10-08T12:21:48 *** FireFly has quit IRC (Excess Flood) 2012-10-08T12:23:33 *** mceier has joined #aichallenge 2012-10-08T12:24:54 great 2012-10-08T12:25:16 i walked my mom through changing her thunderbird/xp passwords 2012-10-08T12:25:26 ove the phone 2012-10-08T12:26:19 *** Simas_J has joined #aichallenge 2012-10-08T12:27:10 thestinger: what type of house do you live in? 2012-10-08T12:27:16 *** FireFly has joined #aichallenge 2012-10-08T12:27:25 mcstar: it has a roof 2012-10-08T12:27:32 hehe 2012-10-08T12:27:38 mcstar: 3 floor detached house 2012-10-08T12:27:38 does it have a garden? 2012-10-08T12:27:59 well there's a garden in the front behind the yard and there's a backyard 2012-10-08T12:28:00 how come there are 60 networks? 2012-10-08T12:28:08 because I live in toronto and that's normal 2012-10-08T12:28:13 oh 2012-10-08T12:28:14 each house basically has a network 2012-10-08T12:28:20 or more than one 2012-10-08T12:28:29 but wifi quickly dissipates 2012-10-08T12:28:34 and we pick up networks from our side of the street 2012-10-08T12:28:37 and across the street 2012-10-08T12:29:01 mcstar: it hasn't gotten the memo here 2012-10-08T12:29:22 lol 2012-10-08T12:29:35 these comsumer electronic networking devices mostly suck 2012-10-08T12:29:46 well I can't actually use more than 7 of the networks 2012-10-08T12:29:49 I just see the others 2012-10-08T12:29:52 I can't auth with them 2012-10-08T12:30:02 now thats more realistic 2012-10-08T12:30:13 you just sporadically get a beacon 2012-10-08T12:30:42 i should change my network to n 2012-10-08T12:30:42 at my mom's house there are a lot more that I can use 2012-10-08T12:30:58 because there's an apartment building across a walkway thing 2012-10-08T12:31:21 but about the same amount appear in a scan, I guess they flood each other out 2012-10-08T12:31:28 thestinger: hm, does that mean you have a house on your own? 2012-10-08T12:31:31 no 2012-10-08T12:31:36 means I live with my dad 2012-10-08T12:31:43 oh 2012-10-08T12:32:03 oops 2012-10-08T12:32:24 nothing 2012-10-08T12:32:38 i wonder if i can get my router to redirect all wifi traffic to my pc 2012-10-08T12:32:57 i dont see a reason why this wouldnt be possible 2012-10-08T12:33:38 why? 2012-10-08T12:33:51 mcstar: do you have openwrt on your router? 2012-10-08T12:34:04 i have no idea 2012-10-08T12:34:16 is that the firmware? 2012-10-08T12:34:17 no then 2012-10-08T12:34:42 well openwrt is replacement firmware 2012-10-08T12:34:47 like what cyanogenmod is for android phones 2012-10-08T12:35:05 mcstar: https://en.wikipedia.org/wiki/OpenWrt 2012-10-08T12:35:07 sure.... 2012-10-08T12:35:12 whats an android? 2012-10-08T12:35:14 like Data? 2012-10-08T12:35:18 XD 2012-10-08T12:36:03 yes 2012-10-08T12:36:03 except it fits in your palm 2012-10-08T12:36:39 the positronic pimp 2012-10-08T12:37:36 thestinger: what do you think og S7E2? 2012-10-08T12:37:49 i think it was better than the premier 2012-10-08T12:38:23 *** Simas_J has quit IRC (Ping timeout: 245 seconds) 2012-10-08T12:38:39 also, at one point deb and dex migh caugh fuck 2012-10-08T12:43:46 welcome to serial killer anonymous! 2012-10-08T12:49:38 *** InsaneMalkavian has quit IRC (Ping timeout: 240 seconds) 2012-10-08T12:50:09 *** foRei has joined #aichallenge 2012-10-08T13:06:52 *** replore_ has quit IRC (Remote host closed the connection) 2012-10-08T13:18:53 *** epicmonkey has quit IRC (Ping timeout: 246 seconds) 2012-10-08T13:21:43 *** replore_ has joined #aichallenge 2012-10-08T13:25:33 *** thestinger has quit IRC (Quit: WeeChat 0.3.9) 2012-10-08T13:26:07 *** replore_ has quit IRC (Remote host closed the connection) 2012-10-08T13:47:03 hey I can solve sudoku #2 2012-10-08T13:48:25 *** antimatroidl has joined #aichallenge 2012-10-08T13:52:32 *** dici has joined #aichallenge 2012-10-08T13:54:26 14% currently, pretty bad 2012-10-08T13:55:26 whats 14%? 2012-10-08T13:57:53 mleise: ^^ 2012-10-08T14:01:06 14% of the puzzles can be solved with my program 2012-10-08T14:01:37 one puzzle? 2012-10-08T14:01:46 but i cannot time it in release mode. i get odd numbers, like 140.735.607 ms 2012-10-08T14:02:19 no, 22 puzzles 2012-10-08T14:02:30 at least, i get a solution for them 2012-10-08T14:02:53 mleise: where did you get the puzzles from? 2012-10-08T14:03:18 i dont want to read norvig's site 2012-10-08T14:03:26 wait 2012-10-08T14:03:28 got it 2012-10-08T14:03:29 the links i posted yesterday. they are scattered in norvigs explanations 2012-10-08T14:03:30 skimmed it 2012-10-08T14:04:27 'easy' is in a block format 2012-10-08T14:04:39 my code already has more lines than some peoples code has characters ^^ 2012-10-08T14:04:46 lol 2012-10-08T14:05:08 yes, I wrote a parser that treats ascii as the beginning of a comment line 2012-10-08T14:05:20 i stop 'comment' mode when I see a \n or \r 2012-10-08T14:06:10 once there is a 0-9 or . it goes into 'grid' mode 2012-10-08T14:06:31 i just converted them to norvig's format 2012-10-08T14:06:37 e.g. comment mode is just to skip digits, that are contained in comments 2012-10-08T14:06:48 oh that's another solution 2012-10-08T14:06:53 ^^ 2012-10-08T14:08:21 *** UncleVasya has quit IRC (Read error: Connection reset by peer) 2012-10-08T14:10:59 lol 2012-10-08T14:11:03 this took some time 2012-10-08T14:11:24 emacs' query and replace highlighter had a hard time highlighting the 0s 2012-10-08T14:20:55 26% :) 2012-10-08T14:22:50 but now i need a better algorithm to push it further... or brute force. but 26% still feels too bad to switch to brute force 2012-10-08T14:23:49 *** alehorst has quit IRC (Ping timeout: 252 seconds) 2012-10-08T14:24:45 *** thestinger has joined #aichallenge 2012-10-08T14:25:26 mleise: http://sprunge.us/BAWR 2012-10-08T14:25:53 thats the board :) 2012-10-08T14:26:01 easy puzzle number 1 2012-10-08T14:26:50 map of int pairs to sets 2012-10-08T14:31:07 'easy'? if they were so easy my solve rate wasn't that low. 2012-10-08T14:31:42 I wouldn't be able to solve more without trial and error if I solved them myself 2012-10-08T14:37:15 mcstar: try "Grid 06" 2012-10-08T14:37:24 try what? 2012-10-08T14:37:27 i cant solve it 2012-10-08T14:37:33 me neither ^^ 2012-10-08T14:37:33 just setting up the necessary stuff 2012-10-08T14:50:07 mcstar: how can I define dependencies between slots? I have these cases where I can put '4' in one of two positions, but either way it would of course disable the option for '4' in other fields. 2012-10-08T14:50:58 '\o/' 2012-10-08T14:51:13 I believe that defining this kind of either/or would help solve some more of the puzzle 2012-10-08T14:52:16 is that a lol, or are you happy with something? 2012-10-08T14:54:38 mleise: http://sprunge.us/OLXC 2012-10-08T14:54:49 mleise: i spread my arms, dont know the answer 2012-10-08T14:55:22 the question is too general in fact 2012-10-08T14:55:31 hmm yeah. i guess I'l comtemplate that a bit and then switch to brute force :/ 2012-10-08T14:56:09 I basically want to forsee what brute force recursion would yield 2012-10-08T14:56:17 mleise: what do you think, how long is my code so far? 2012-10-08T14:56:32 i don't even know what it can do 2012-10-08T14:56:57 reads the puzzle into a Board 2012-10-08T14:57:03 and displays it like you saw 2012-10-08T14:57:40 35 lines ? 2012-10-08T14:58:02 I hope that was a good guess 2012-10-08T14:58:18 20 2012-10-08T14:58:22 so yeah 2012-10-08T14:58:38 20... I'm at 373 2012-10-08T14:59:47 *** Accoun has quit IRC () 2012-10-08T15:17:10 *** Accoun has joined #aichallenge 2012-10-08T15:22:14 i look away for a moment form #haskell, there are already 150 lines passed by 2012-10-08T15:24:11 *** cyphase has quit IRC (Quit: http://www.cyphase.com/) 2012-10-08T15:28:08 *** cyphase has joined #aichallenge 2012-10-08T15:28:59 mleise: heh, first puzzle done 2012-10-08T15:29:04 want to see my solution? 2012-10-08T15:29:14 cycle$cycle$cycle$cycle$cycle$cycle puzzle 2012-10-08T15:29:23 don't know ... what's that? 2012-10-08T15:29:31 thats my solution 2012-10-08T15:29:33 in short 2012-10-08T15:29:36 you just brute force it 2012-10-08T15:29:42 :) 2012-10-08T15:29:54 why not? for a first solution... 2012-10-08T15:29:55 well, I guess that works and is short ^^ 2012-10-08T15:30:05 not really 2012-10-08T15:30:12 i have 3 functions that i can make 1 2012-10-08T15:30:15 so you can already solve all the puzzles with it 2012-10-08T15:30:20 i dont think so 2012-10-08T15:30:28 well, maybe 2012-10-08T15:30:34 havent thought about it 2012-10-08T15:30:38 is it exhaustive or not? 2012-10-08T15:31:16 im just removing numbers that cant appear on a square 2012-10-08T15:31:39 um yes, that's how i solved #1, too. 2012-10-08T15:32:16 and when one square is down to 1 option I feed that square back into the algorithm 2012-10-08T15:35:03 what I need is a 'set solver' 2012-10-08T15:36:58 square_1 is one of [3,7,8] without [square_2, square_4, ...] 2012-10-08T15:42:30 *** pairofdice has quit IRC (Quit: In girum imus nocte et consumimur igni.) 2012-10-08T15:43:53 mleise: i factored it a bit http://hpaste.org/75990 2012-10-08T15:46:25 i dont understand it, but it looks concise 2012-10-08T15:50:26 'its easy, if you try...' 2012-10-08T15:55:16 *** bobby45 has joined #aichallenge 2012-10-08T15:56:48 mleise: there are easy puzzles in the 'easy' category 2012-10-08T15:56:49 *** wentam has quit IRC (Remote host closed the connection) 2012-10-08T15:57:02 and there are those, that cannot be solved with my method 2012-10-08T15:57:11 they reach a fixpoint 2012-10-08T15:58:10 you don't say ;) 2012-10-08T15:58:18 *** jacob_strauss has joined #aichallenge 2012-10-08T15:58:36 :) 2012-10-08T15:58:57 but i can solve quite a feq 2012-10-08T15:59:00 few 2012-10-08T15:59:34 hm 2012-10-08T15:59:41 not in the second half of the 50 :( 2012-10-08T16:00:06 oh, ok, doing good again 2012-10-08T16:01:33 mleise: for puzzle #50, theres only 1 square that i can solve this way 2012-10-08T16:01:41 lol :D 2012-10-08T16:01:54 let me try it, after all i have 80% of those 2012-10-08T16:02:37 no solution, but some parts solved 2012-10-08T16:03:08 I have 37/81 2012-10-08T16:04:01 mleise: there are 50 puzzles in this file 2012-10-08T16:04:25 i mean, in #50 i have 37 squares solved 2012-10-08T16:05:34 oh 2012-10-08T16:06:07 25 are given 2012-10-08T16:06:23 *** alehorst has joined #aichallenge 2012-10-08T16:21:08 let p=do{(x:xs)<-get;put xs;Node x<$>if all isDigit x then pure[]else replicateM 2p};r(Node x[])=x;r(Node x ys)="("++x++((' ':).r=< => 2012-10-08T16:21:17 "(+ (* 5 5) (- (+ 2 1) 42))" 2012-10-08T16:21:36 mleise: rpn to prefix converter ^^ just saw it on #haskell 2012-10-08T16:25:22 *** sigh has joined #aichallenge 2012-10-08T16:28:08 *** coeus has joined #aichallenge 2012-10-08T16:29:06 mleise: whats next? 2012-10-08T16:29:11 *** Cyndre_ has quit IRC (Quit: Leaving) 2012-10-08T16:29:17 should i just be doing a BFS? 2012-10-08T16:29:35 im not particularly enthusiastic about it 2012-10-08T16:36:41 Well, I just did what I always do when I play Sudoku 2012-10-08T16:37:45 How can I explain it... 2012-10-08T16:40:15 i played sudoku too 2012-10-08T16:40:30 sort of like, mine field 2012-10-08T16:40:39 mcstar: when you set a square to a certain value it not only removes that value from the row,column and block, but may also remove the last of 2 optional positions for another digit 2012-10-08T16:40:41 or whats the name of that ms game 2012-10-08T16:40:49 i know 2012-10-08T16:40:57 but that is a search 2012-10-08T16:41:24 I wouldn't call that search in my code 2012-10-08T16:41:26 fortunately, my version is immutable, i think i can do better with it than you :) 2012-10-08T16:42:09 i think i leave this part of the solver for another day 2012-10-08T16:42:14 I just remove options in loops from 1 to 9 mostly 2012-10-08T16:43:41 yes, i made a few mistakes while writing this. it isn't trivial. 2012-10-08T16:44:09 i had it invalidate an already solved square for example 2012-10-08T16:44:35 turned out i was using the wrong 'row' variable 2012-10-08T16:59:24 *** jacob_strauss has quit IRC (Remote host closed the connection) 2012-10-08T17:19:41 *** Scooper has quit IRC (Quit: Leaving) 2012-10-08T17:21:45 *** antimatroidl has quit IRC (Ping timeout: 248 seconds) 2012-10-08T17:33:01 *** Kineman has joined #aichallenge 2012-10-08T17:35:07 *** Kineman has left #aichallenge 2012-10-08T17:43:10 *** antimatroidl has joined #aichallenge 2012-10-08T17:51:56 *** wentam has joined #aichallenge 2012-10-08T17:53:24 mleise: hm, im actually debugging something in haskell 2012-10-08T17:53:31 weird moment 2012-10-08T17:55:09 i think i know the reason 2012-10-08T17:58:03 *** bobby45 has quit IRC (Quit: Page closed) 2012-10-08T17:58:38 im tired 2012-10-08T18:00:42 *** antimatroidl has quit IRC (Ping timeout: 246 seconds) 2012-10-08T18:06:14 *** thestinger has quit IRC (Quit: WeeChat 0.3.9) 2012-10-08T18:07:45 why did I notice the highlight just now? 2012-10-08T18:07:58 *** Kakwa has quit IRC (*.net *.split) 2012-10-08T18:08:10 it must have blinked since 5 minutes O.o 2012-10-08T18:08:13 your reflexes are slow 2012-10-08T18:08:40 I'm tired too, but I will quickly hack together the brute force recursion 2012-10-08T18:08:45 im trying to index a block by normal row/col indices... how stupid of me 2012-10-08T18:08:49 I don't need new code for that at least 2012-10-08T18:09:15 yes, it ended up being a mess of % 3 / 3 * 3 stuff 2012-10-08T18:10:07 ok segmentation fault :-/ 2012-10-08T18:11:02 and my solver object has to be copied for every recursion level. it is 1.5KB. 2012-10-08T18:14:33 *** Accoun has quit IRC (Read error: Connection reset by peer) 2012-10-08T18:14:50 *** Accoun has joined #aichallenge 2012-10-08T18:16:58 thats not big 2012-10-08T18:17:08 haha 2012-10-08T18:17:14 mleise has a little stack 2012-10-08T18:17:15 mleise has a little stack 2012-10-08T18:19:30 actually that includes my solver stack, yes. I should try to avoid copying that 2012-10-08T18:19:30 fuck 2012-10-08T18:19:48 why do i have an out of bounds like error??? 2012-10-08T18:20:56 me too, just now that i refactored 2012-10-08T18:21:43 *** Kakwa has joined #aichallenge 2012-10-08T18:22:05 oh 2012-10-08T18:22:08 i know why 2012-10-08T18:22:34 *** foRei has quit IRC (Read error: Connection reset by peer) 2012-10-08T18:23:19 its working 2012-10-08T18:23:28 disappointing 2012-10-08T18:23:49 mleise: can solve all of them 2012-10-08T18:24:16 good 2012-10-08T18:25:45 http://sprunge.us/UFOY?haskell 2012-10-08T18:25:52 2 plus functions 2012-10-08T18:26:08 the intput has all the 50 puzzles 2012-10-08T18:26:46 come on 67 loc ? 2012-10-08T18:27:13 comments, the useless instance declaration 2012-10-08T18:27:24 all the input and conversion functions 2012-10-08T18:27:30 yeah, its a lot 2012-10-08T18:27:49 (well, not useless, it shows me the board nicely) 2012-10-08T18:29:20 *** dici has quit IRC (Quit: ( www.nnscript.com :: NoNameScript 4.22 :: www.esnation.com )) 2012-10-08T18:30:08 mleise: puzzle #36 takes very long 2012-10-08T18:30:35 it solves it, but more than an order of magnitude slower 2012-10-08T18:31:26 I can't really tell. The total was so fast until now, that I need recursion 2012-10-08T18:31:33 and get these out of bounds errors 2012-10-08T18:32:15 mleise: nice 2012-10-08T18:32:22 i put my cycle as a first pass 2012-10-08T18:32:27 it is the same speed 2012-10-08T18:32:49 mleise: im working in the interpreter 2012-10-08T18:32:57 cant really tell the 'real' speed 2012-10-08T18:33:01 this is btecode 2012-10-08T18:33:49 let me compile it 2012-10-08T18:35:36 mleise: 0.59s for the 50 puzzles 2012-10-08T18:35:46 im running at 1GH 2012-10-08T18:35:48 z 2012-10-08T18:37:34 you have special 1 Ghz. it's almost as fast as 2 Ghz. I'll take that into account ^^ 2012-10-08T18:39:33 yay! solved #50 in 250 ms 2012-10-08T18:39:33 mleise: what? 2012-10-08T18:39:42 :) 2012-10-08T18:39:55 300ms vs 250ms 2012-10-08T18:40:07 i dont mind the difference 2012-10-08T18:40:15 also, this is a bad haskell code 2012-10-08T18:40:21 idk about your D 2012-10-08T18:40:44 mleise: wait 2012-10-08T18:40:52 ok 9077 µs for the first 50. Now I can optimize... 2012-10-08T18:40:53 you solved 50 puzzles 2012-10-08T18:40:57 or the 50th? 2012-10-08T18:41:04 the 50th in debug mode 2012-10-08T18:41:23 mleise: heh, you tricked me on purpose? 2012-10-08T18:41:29 console output isn't particularly fast in D 2012-10-08T18:41:39 mcstar: a bit 2012-10-08T18:41:50 didn't notice the ambiguity at first 2012-10-08T18:42:11 i dont have a debug mode 2012-10-08T18:42:38 well, I mean with debug messages, assertions, unoptimized code... that stuff 2012-10-08T18:42:39 mleise: the only error i made in the code, is that i first omitted this line: (br,bc) = ((r-1)`div`3+1, (c-1)`div`3+1) 2012-10-08T18:43:35 Well I think I stumbled over some code generator bugs in D once again. I couldn't cleanly copy the Solver structure for some reason. Parts of it were uninitialized. 2012-10-08T18:44:29 i sure your algorithm is better on a theoretical level 2012-10-08T18:46:04 I have that additional step in there, yes. 2012-10-08T18:46:57 if you didn't improve on what you had (1 additional solved square for #50), that is 2012-10-08T18:47:41 i mean generally, my algorithm is not too good 2012-10-08T18:47:49 im doing too much work 2012-10-08T18:48:00 but hey, didnt spend too much time on it either 2012-10-08T18:48:23 I hate how slow the top95 puzzles are 2012-10-08T18:48:34 600 ms for all 2012-10-08T18:50:20 im not brave enough to try those yet 2012-10-08T18:51:06 it will work, your algorithm can do it 2012-10-08T18:51:15 yes we can! 2012-10-08T18:58:38 *** replore_ has joined #aichallenge 2012-10-08T19:10:49 *** sigh has quit IRC (Remote host closed the connection) 2012-10-08T19:14:20 *** sigh has joined #aichallenge 2012-10-08T19:18:05 mleise: well, 185ms 2012-10-08T19:18:21 converted to 2ghz to look better 2012-10-08T19:19:05 i changed the priority, i pick those squares and try those first, that have only 2 possible values 2012-10-08T19:21:15 wow, your recursion sure kicks ass 2012-10-08T19:22:24 :( 2012-10-08T19:22:27 mock me 2012-10-08T19:23:23 no, I mean it. I'm impressed. You do realize you are 3 times faster than my D program ? 2012-10-08T19:24:07 what? 2012-10-08T19:24:13 no, i didnt realize 2012-10-08T19:24:18 i still dont 2012-10-08T19:24:30 you quoted 9077 µs for the first 50. 2012-10-08T19:25:02 i see 2012-10-08T19:25:05 I thought you tried the top95 puzzles ? 2012-10-08T19:25:08 i didnt solve the top95 2012-10-08T19:25:14 no way 2012-10-08T19:25:23 i can try... 2012-10-08T19:25:50 *** replore_ has quit IRC (Remote host closed the connection) 2012-10-08T19:26:18 well if I run the test 10 times, the first 50 come down to ~3900 µs 2012-10-08T19:26:29 on average 2012-10-08T19:28:03 wtf 2012-10-08T19:28:11 0.73s??? 2012-10-08T19:28:32 great 2012-10-08T19:28:38 it has unsolved puzzles 2012-10-08T19:28:45 clever optimization 2012-10-08T19:28:48 *** replore has joined #aichallenge 2012-10-08T19:28:51 :p 2012-10-08T19:29:06 i never checked my solutions btw. 2012-10-08T19:29:32 mleise: not even for consistency? 2012-10-08T19:29:42 or just neve by inspection? 2012-10-08T19:29:44 no :p 2012-10-08T19:29:48 lol 2012-10-08T19:29:49 nothing 2012-10-08T19:29:53 great 2012-10-08T19:29:56 just return the memory 2012-10-08T19:30:04 and, do you want to? 2012-10-08T19:30:15 cause the comparisons are meaningless otherwise 2012-10-08T19:30:38 *** sigh has quit IRC (Remote host closed the connection) 2012-10-08T19:30:39 well, I will check in debug mode 2012-10-08T19:31:09 mleise: just copy my prettyprinter 2012-10-08T19:31:24 This is how we work in the D community. We use asserts which get removed in a release compile. 2012-10-08T19:31:29 *** replore has quit IRC (Remote host closed the connection) 2012-10-08T19:31:55 that doesnt matter 2012-10-08T19:32:03 a pretty printer is useful 2012-10-08T19:32:11 you can have both at once 2012-10-08T19:32:37 mleise: you are probably not aware of it, but haskell has the best unit test framework 2012-10-08T19:33:19 I have a pretty print function, but how can it possibly be a validator ? 2012-10-08T19:33:41 you look at the board 2012-10-08T19:33:49 and you can immediately see if it is filled 2012-10-08T19:33:53 and if it is solved 2012-10-08T19:34:44 this is weird 2012-10-08T19:35:02 according to me test functions, that non-filled board is not bad and solved 2012-10-08T19:35:49 lol 2012-10-08T19:35:51 ok 2012-10-08T19:36:01 my solved condition is wrong 2012-10-08T19:36:13 fortunately i dont use it in the solution process 2012-10-08T19:36:18 just for hand-testing 2012-10-08T19:39:31 anyway, the recursion that im doing is far from ideal 2012-10-08T19:39:48 now is the time for me to stop though 2012-10-08T19:40:02 but id really like to know if you solutions are indeed solutions 2012-10-08T19:41:25 mleise: great, i can solve those unsolved ones, by running my 'solve' function on them again 2012-10-08T19:41:45 idk why do some of the partially filled boards escape my recursion 2012-10-08T19:42:32 lol 2012-10-08T19:42:55 omg! "Invalid result" errors 2012-10-08T19:43:02 what's happening 2012-10-08T19:43:05 oh 2012-10-08T19:43:35 mleise: even in the 50 easy ones? 2012-10-08T19:44:42 the very first one fails, yes 2012-10-08T19:45:13 lol 2012-10-08T19:45:38 mleise: i think the 4th of top95 takes me into an infinite loop 2012-10-08T19:46:39 maybe my test is wrong 2012-10-08T19:48:30 shit, no it is correct. I have a block with two '2's. I'm doomed. 2012-10-08T19:48:46 great 2012-10-08T19:48:55 i think we should go to bed 2012-10-08T19:49:02 mleise: good night 2012-10-08T19:49:03 oh wait... i looked at the wrong row, should add separators 2012-10-08T19:49:05 night 2012-10-08T19:49:11 wait 2012-10-08T19:49:17 is the solutino correct after all? 2012-10-08T19:49:37 i haven't hand validated all blocks yet, but it may be, yes 2012-10-08T19:49:52 see you tomorrow 2012-10-08T19:49:54 *** mcstar has quit IRC (Quit: mcstar) 2012-10-08T20:13:30 *** thestinger has joined #aichallenge 2012-10-08T20:38:51 *** mleise has quit IRC (Ping timeout: 246 seconds) 2012-10-08T20:43:50 *** mattswe has joined #aichallenge 2012-10-08T20:49:21 *** mleise has joined #aichallenge 2012-10-08T20:57:17 *** amstan has joined #aichallenge 2012-10-08T20:57:17 *** ChanServ sets mode: +o amstan 2012-10-08T21:02:09 *** Apophis_ has joined #aichallenge 2012-10-08T21:02:59 *** Apophis has quit IRC (Read error: Connection reset by peer) 2012-10-08T21:04:30 *** lurtz has joined #aichallenge 2012-10-08T22:12:57 *** thestinger has quit IRC (Quit: sleep) 2012-10-08T22:45:43 *** Cyndre has joined #aichallenge 2012-10-08T22:50:29 *** replore has joined #aichallenge 2012-10-08T23:06:41 *** replore has quit IRC (Remote host closed the connection) 2012-10-08T23:11:14 *** lurtz has quit IRC (Ping timeout: 240 seconds) 2012-10-08T23:19:26 *** Accoun has quit IRC (Ping timeout: 256 seconds) 2012-10-08T23:22:58 *** Accoun has joined #aichallenge