00:00:12 -!- shikhin has quit (Ping timeout: 244 seconds). 00:02:29 I can see how a universal set would do that, but I thought there is no such things as a universal set. 00:02:47 A set cannot contain itself 00:03:12 the consequences is that universal set would not be, by definition, a set 00:03:53 http://en.wikipedia.org/wiki/Aczel%27s_anti-foundation_axiom 00:03:56 -!- dmstor has joined. 00:04:40 -!- dmstor has left. 00:06:06 to be fair, i've only seen an attempt at actually using it in a really shitty biology paper. 00:07:04 I don't deal with sets so much 00:07:16 But I can do Array = [], Array[Array] := Array 00:07:30 er, Array := [], Array[Array] := Array 00:07:34 missed a : on the first one 00:07:48 ( := is the assignment operator in the language I am using for this demonstration) 00:07:49 Command "=" not permitted. 00:17:33 GeekDude: Not in the usual set theories, but this was a set theory with such a universal set. It was probably NFU. 00:19:09 NFU? 00:19:44 Apparently this stands for “New Foundations with Urelements”. See https://en.wikipedia.org/wiki/New_Foundations . 00:20:46 -!- AndoDaan has quit (Ping timeout: 276 seconds). 00:21:00 My brain is frying here 00:21:14 I'm not too well versed in set theory 00:21:18 of any kind 00:23:13 I brought it up because there was already talking about the fact that if you take a different foundation, different axioms, you can get very different properties from what you’re used to. So I decided to give another example I’d heard of. 00:31:08 I need to write a simple PRNG, and am limited to only using integers between negative and positive 524287 00:32:13 A linear congruential generator sounds like it could work 00:32:29 with a small multiplier and increment 00:34:15 yeah, that should be easy. 00:34:59 724 for the modulus, that's roughly the square root of the max int limit 00:35:24 then perhaps something like 483 for the multiplier 00:35:35 and 238 for the increment 00:35:41 unless I'm completely misunderstanding the point 00:36:16 I could stuff a larger range of values in there I guess 00:36:33 hopefully you are not using this for crypto 00:36:37 nooooo way 00:36:47 who gives a damn then 00:36:58 I need values that look vaguely random 00:37:15 i had a school project last week where i got entropy by counting up until the user hit a button, #yolo 00:37:22 lol 00:37:42 no clock value you could mooch off? 00:39:15 that's basically what i was doing. didn't feel adventurous enough to figure out how to get at the oscillators directly 00:39:59 I've heard a last ditch effort for medium quality entropy is to ask the user to mash the keyboard 00:41:23 That's practically begging for bad entropy. 00:41:40 lol 00:41:46 you need high quality users 00:41:55 and maybe a blindfold 00:41:56 type without rhythm so as not to attract the worm. 00:42:01 Ah, "values that look vaguely random" though. 00:42:29 I hear the low bit of rdtsc is good enough. 00:43:42 rdtsc? 00:44:01 cycle counter on x86. 00:44:26 I'm not even close to using x86 though 00:44:44 Well then, that's a terrible idea. 00:44:49 I'm using BASIC on a 3ds 00:45:05 ... For a class project. 00:45:12 Is it SmileBASIC? 00:45:14 BASIC. On a 3DS. 00:45:16 Kay then. 00:45:26 zzo38: yes it is 00:45:29 the class project was me, geekdude might be doing some other shite 00:45:35 I know you can run SmileBASIC on 3DS, but it isn't really like BASIC; there are many differences too. 00:45:38 Bike: Oic 00:45:48 ¯\_(ツ)_/¯ all the variants have their giant quirks 00:45:52 me, i was doing a simon says game. security, of course, is paramount 00:46:20 Of course. 00:46:29 Someone might use it to choose passwords, or crypto keys. 00:46:46 I'm doing fractal trees 00:46:53 well, not really fractal 00:46:59 there was actually that pretty funny deniable encryption thing, but it used a game more like DDR 00:47:01 recursive trees 00:47:21 If it's self-similar it's fractal. 00:47:30 ARM has a cycle counter register, looks like 00:47:37 i imagine something is accessible on the 3ds 00:47:55 I just about guarantee that SmileBASIC does not expose that though. 00:47:55 The branches resemble eachother in that they each have two sub branches (although at psuedo-random angles) 00:48:02 so that's uh, four bytes of entropy, hooray 00:48:03 pikhq: sadly not 00:48:18 do like old dos basic, just write instructions into the memory space 00:48:23 as previously mentioned, #yolo 00:48:38 Bike: If SmileBASIC works like that then some hackers are gonna be very happy. 00:48:57 oh, this is an official product. 00:49:01 'tis 00:49:02 Yes. 00:49:12 there's a new version coming out very soon to the US 00:49:14 should be loads better 00:49:17 i know some emulator people now so i just assume all game systems are cracked. 00:49:20 and it will support 3d output 00:49:25 It's, uh, the BASIC variant in the PetiteComputer thing? 00:49:33 yeah, Petit computer 00:49:44 Sadly not really. The DSi is only tenuously cracked, and the 3DS is utterly uncracked at this point. 00:49:50 "Also, you can utilize data in your program from the built-in accelerometer, gyro-sensor, microphone and the circle pad. Get the most out of potential of Nintendo 3DS!" 00:49:53 *And* it's got fucking region locking! 00:49:56 Sounds like entropy to meeeee 00:50:03 Bike: that's in the new version though 00:50:06 not released to the US yet 00:50:16 And the DSi and 3DS are region locked. 00:50:17 i live on the edge 00:50:18 I'm using the DSi version 00:50:21 Because Nintendo hates people. 00:50:25 ..........of the US and europe 00:50:41 I would think it would be better to do differently, if instead it has a Forth interpreter built-in (rather than being external software), with a virtual machine that you can access the full address space of (you don't need to access address space outside of the VM) 00:50:50 i wonder how good the accelerometer is. i tried my phone's and it was one hell of a hot mess 00:50:51 The original Nintendo DS is not region locked. 00:50:54 pikhq: I think this is more of a SmileBoom thing, they have to translate everything into english, and make sure it still works 00:51:17 GeekDude: The thing is, though, if the platform weren't region locked you could at least *run* foreign software. 00:51:19 then they plug their PRNG into a countdown timer 00:51:26 It might not be in a language you speak, but still. 00:51:40 I do have one foreign game card on my Nintendo DS 00:52:14 http://smilebasic.com/en/image/ss-story-help-4.png wow, it's like i'm really on my TI 00:52:37 I get "TIME$" and "DATE$" as my only two time sensitive inputs 00:52:46 Bike: it looks like a shitty dos game 00:52:58 that amazing font!!! 00:53:06 I still write games on DOS sometimes 00:53:18 I'm doing this http://i.imgur.com/ryhFSIA.jpg 00:53:28 but imagine green and brown, and more branches 00:53:42 oh, it's like an l system. 00:53:46 or... something. 00:53:47 it actually is 00:53:52 I've implemented an L system 00:54:05 #nailed it 00:54:19 though, two of the operations include "and add a random number" 00:54:31 so YMMV as to reproducability 00:54:31 nuthin wrong with a lil stochasticism. 00:55:29 RAND = (433*RAND + 238) % 724 00:55:33 woo, 'random' 00:55:46 nsa gonna hack ur l system 00:56:23 idgi. wouldn't that have a period of at most 724 00:56:37 lol 00:56:52 I said 'vaguely random' 00:57:05 no, that's true, you might as well use the largest modulus you can. 00:57:13 ¯\_(ツ)_/¯ 00:57:23 I'm not sure how stack overflow is handled 00:57:29 it says values larger than the max limit are "ignored" 00:57:32 stack?? 00:57:41 There is not stack 00:57:43 no* 00:57:52 er 00:57:54 you just said stack overflow, did you mean numeric overflow. 00:57:55 integer overflow* 00:58:12 right, well, just find a prime around 500000 and call it good. 00:58:17 My head is on slightly crooked 00:58:29 "Ok google, closest prime to five hundred thousand" 00:58:41 er... well i don't remember the requirements... they're on wikipedia. 00:58:58 http://www.wolframalpha.com/input/?i=prime+near+50000 the future is great 00:59:49 * GeekDude needs to test integer overflow first 01:00:09 ah 01:00:17 seems to just print "overflow" and ignore the command 01:00:41 I need to do two commands then I think 01:00:49 one to multiply and the other to do other things 01:02:02 RAND = 483 * RAND:RAND = (RAND + 238) % (LIMIT - 238) 01:02:07 that would work I think 01:02:15 : is the command delimiter thing 01:02:18 like ; in C++ 01:02:47 what's LIMIT? 01:02:58 the maximum integer size 01:03:06 In BASIC a colon can be used for multiple commands in one line; otherwise line breaks delimit multiple commands. 01:03:07 in this case, roundabout 524287 01:03:25 hey zzo38, I'm starting up a SmileBASIC channel if you wanna join 01:03:28 It is different than in C and C++, where a semicolon marks the end of a statement, regardless of line breaks 01:03:37 well, a new SmileBASIC channel 01:03:43 the other one was on EFNet and only had 2 people 01:03:45 GeekDude: I saw your invitation, and no I don't have any Nintendo 3DS 01:03:47 me and 01:03:49 aww 01:04:00 Where'd you learn about SmileBASIC? 01:04:05 or was that just google talking 01:05:10 i think every random number you generate will be 1 (mod 3) 01:05:27 I like to program Famicom though 01:05:52 -!- AndoDaan has joined. 01:05:53 I did learn about SmileBASIC from a message on NESdev I think. 01:06:00 I do not remember. 01:06:22 because 3 | 524287 - 238, and 3 | 483, so your formula mod 3 becomes just (0r + 1) = 1 01:07:15 I'm not sure I'm following 01:09:24 256 * 1024 is the max integer limit 01:09:28 er 01:09:31 512 * 1024 01:09:51 so if I use a multiplier just below 512, I can use a max of just around 1024 01:11:32 -!- copumpkin has quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…). 01:13:12 IF RAND<1024 THEN RAND=RAND*503 01:13:12 RAND=(RAND+9203)%1024 01:13:22 That should be overflow free, and users primes for both M and C 01:13:29 er, a and c 01:13:34 m is 1024 01:14:04 multiples of two for m is standard 01:18:19 this seems to be working fairly okay 01:18:58 There is an actual PRNG on here, but I can't seed it 01:19:01 Famicom even has two game pads, microphone, keyboard, drawing tablet, light gun, tape recording, stereovision, dance mat, and more other things. 01:20:11 What's up with the connector port on the bottom? 01:20:26 at least, on the USA NES 01:20:36 The NES connector port was going to be used for the NES disk drive, but they never released that. 01:20:48 It connects also to the cartridge port and a few other things; it isn't very useful. 01:21:17 And not all versions of NES even have such a port anyways. 01:21:33 Perhaps the most useful thing to do with it would be to jumper the cartridge port lines and the sound lines, and then mod a Famicom adapter to put the sound lines on those pins. 01:22:11 (the Famicom had a sound line on the cartridge port; on the NES, this was moved to the expansion port. Unfortunately, Famicom games *actually used* the sound line.) 01:22:49 Sound input, output, or both? 01:23:32 Output. 01:23:53 and the famicom games used this for... ? 01:23:54 Actually, the Famicom has both. 01:23:57 I can't think of a goo use case 01:24:01 good* 01:24:10 Better synthesizers on the cartridge. 01:24:20 The FDS itself used this feature. 01:24:26 If you decide the output, can't you just use local feedback? 01:24:26 The audio goes through the cartridge; usually it is just a wire or mixed with extra audio on the cartridge, and another pin goes that to TV set. 01:25:02 zzo38: Ah. 01:25:09 I thhought it was mixed in the console. 01:25:59 -!- tromp_ has joined. 01:26:12 But yeah. The Famicom sound hardware was somewhat limited, and having an extra synth on the cartridge bought you a lot of extra power. 01:26:50 Least it's not as limited as the poor 2600 was. 01:28:02 I still think super famicom was where it was at 01:28:23 -!- tromp__ has quit (Ping timeout: 240 seconds). 01:28:27 It was overall a pretty awesome system. Crazy to emulate though. 01:29:11 Cartridges on there can bus master. 01:30:06 Emulating heterogenous multiprocessing like that is such a pain. 01:30:33 opencl to the rescue 01:31:02 I don't like Super Famicom; I like Famicom. 01:31:14 Bike: Doesn't help that much here. 01:31:36 It's shared memory heterogenous multiprocessing. 01:32:04 This is about where most people start sobbing. 01:32:20 Famicom is faster than Commodore 64, too. 01:33:18 By about 700kHz, sure. 01:33:52 I am even writing a Z-machine interpreter on Famicom. I designed a custom mapper for this purpose; I have the description of it if you want it. 01:34:02 Spiffy. 01:34:22 That's almost as awesome as the SCUMM interpreter on Famicom. 01:34:24 Infocom never did, but they did it on Commodore 64. However, looking at their codes they did many things slow. 01:36:30 The mapper I use is pretty unusual (although it can be made using a few 74xx series): For one thing, the story file is stored on a separate ROM chip, which is bankswitched two bytes at a time, and the 64K RAM in the cartridge is bankswitched one byte at a time. For another, the registers overlap mirrors of the Famicom's internal RAM. 01:37:01 pikhq: famicom SCUMM interpreter? 01:37:06 how's the audio? 01:38:08 Only thing that uses it is the official port of Maniac Mansion, so a bit hard to judge. :) 01:38:27 (I don't know any other mapper with such small bank sizes, or with registers overlapping the mirrors of internal RAM.) 01:50:08 hmm 01:50:18 My random number generator, expectedly, is quite bad 01:50:26 a small seed starts off with a small input 01:50:32 er, small starting random values 02:14:01 -!- Sorella has quit (Ping timeout: 244 seconds). 02:30:25 Another possible random number generator is ARCFOUR, which is what SQLite uses. 02:32:50 Finally found some good values for my random 02:34:09 IF RAND < 515 THEN RAND = RAND * 1018 02:34:10 RAND = (RAND + 1018) % 515 02:34:27 gives pretty good results 02:34:40 though... 02:34:46 now that I think about it it's always less than 515 02:35:19 actually, I think I have an overflow opportunity here too 02:35:25 I might try choosing a slightly lower coprime pair 02:35:32 or a slightly lower multiplier 02:36:10 -!- tromp__ has joined. 02:37:53 -!- SvenGek has quit (Quit: [Quit: [Quit: SvenGek] [Quit: SvenGek]]). 02:38:27 -!- tromp_ has quit (Ping timeout: 258 seconds). 02:40:15 ah 02:40:18 nope, I'm good 02:41:03 1018+514*1018 == 524270, which is less than 524287 02:44:23 What's the feature known to be introduced in new NetHack/ 02:46:14 -!- boily has joined. 03:02:07 -!- AndoDaan has left. 03:03:48 Sgeo, you can offer the amulet to moloch now or something 03:12:28 That can't be a win, can it 03:12:42 What happens? 03:30:12 -!- coppro has quit (Ping timeout: 258 seconds). 03:30:55 -!- coppro has joined. 03:32:17 How can you win an infinite collection of envelopes (the first containing $1, the second having $2, the third having $3, the fourth having $4, etc) from "Let's Make a Deal"? 03:34:15 you can't, there isn't an infinite number of envelopes to win 03:34:26 or would that be an arent 03:34:28 anyways, https://db.tt/NPeEG7ml 03:34:56 what's an arent? 03:34:57 And the output, http://i.imgur.com/unpvLzV.jpg 03:35:53 night all 03:36:09 GoodnightkDude! 03:36:11 Where would you keep all of them? In a black hole? Possibly it won't fit even in a black hole. Also, they would add up to a negative number. 03:36:47 wasn't there a "proof" out there about adding up all the positive integers equals -1/12? 03:36:50 i'm not sure the world financial system is down with ramanujan summation. 03:36:57 Yes, I have seen that 03:37:03 That it add up to -1/12 03:37:25 Bike: But the world financial system is not infinite. 03:37:28 ramanujan and... i think cesaro? give you -1/12 03:37:57 oh, no it's not cesaro summable. 03:38:07 Isn't that the one where you take averages? 03:38:16 Yes. 03:38:27 -!- GeekDude has quit (Quit: {{{}}{{{}}{{}}}{{}}} (www.adiirc.com)). 03:38:33 I say all positive numbers are, on average, about five. 03:38:45 (trust me, I'm an engineer.) 03:38:48 I don't think so. 03:39:18 you know what's weird? the erdos-kac theorem 03:39:54 -!- mauris has quit (Ping timeout: 246 seconds). 03:40:08 wikipedia even says "randomly chosen large integer". 03:41:54 that is strangely fascinating. 03:42:17 eh, any integer is gonna be about five anyway. 03:43:36 that's what I said, but zzo38 is being mean. *pfbfbfbfbt* 03:43:53 c'mon zzo, they're an engineer! they know all about numbers. and five. 03:45:30 me, i've never even seen a number in person, to be perfectly honest 03:45:45 Being an engineer is not quite good enough 03:46:08 It also mean, you need to be mathematician, and you also need to prove it, too. 03:46:29 IANAMAA. 03:47:13 http://existentialcomics.com/comic/36 03:47:40 -!- AndoDaan has joined. 03:48:39 Bike: :D 03:51:30 'night all! 03:51:45 -!- boily has quit (Quit: CHILLING CHICKEN). 04:01:25 "P.S. I can't end the post without sharing my favorite self-reproducing program: the one-line shell script #!/bin/cat." 04:02:21 -!- copumpkin has joined. 04:03:31 Sgeo: Sadly, not a shell script and won't work on all POSIX systems. 04:03:45 However, this one will: cat "$0" 04:07:42 Can't tell anything about a .zip file until the entire .zip file is loaded, but can quickly make changes 04:08:15 Although, the latter seems irrelevent... a central header should be sufficient these days, right? Since central footer mostly makes using floppies easier? 04:10:21 Puting the list at the end may also make it easier to write through a pipe in some cases (although might be more difficult to read), and possibly for cheating at Day of Zeux. 04:11:50 The footer of the .zip file is strictly necessary. 04:12:19 That's the only part of the file that most anything looks at first. 04:12:42 Day of Zeux? 04:13:04 Wait, .zip the Z-Machine format? 04:13:08 -!- Sorella has joined. 04:13:13 I meant sufficient in the sense of, if .zip were redesigned with modern requirements in mind, would we still want to place the directory data in the footer 04:13:46 The extension .zip (sometimes also .dat) is also used for Z-machine files, although this is not recommended due to confusion with ZIP archive files. 04:13:47 Oh. Yes. We would not do so at all. 04:13:58 -!- Sorella has quit (Changing host). 04:13:58 -!- Sorella has joined. 04:14:14 As evidenced by how ZIP is nearly unique in that property. :) 04:15:15 * Sgeo goes to read the DEFLATE spec 04:15:45 DEFLATE is a moronic and insane file format with equal parts weird microoptimization and utterly pointless redundancy. 04:16:40 Frankly, I'm pretty sure Phil Katz was drunk when he made it. 04:17:11 Is ZPAQ better then? 04:18:18 From what I know, it's a bit absurd in the sheer flexibility of the format, but it sure seems better designed. 04:19:05 o.O 04:19:28 (for those who don't know, ZPAQ is literally a compression format that specifies a VM for decompression algorithms.) 04:19:43 Ooh that sounds fun 04:20:07 I'm going to go read that instead even if no one uses it 04:20:46 Yep. It's not a common format, but it's certainly a unique design for a format. 04:21:35 I don't entirely like all the way that the VM in ZPAQ works; it requires more than one VM as well as a few other things. What I think would be good way, is the archive file is a memory image for a "Decompression Computer"; the decompression program then loads te memory image and emulates the Decompression Computer. The data on the output port is then treated as a tape archive. 04:22:56 But yeah. ZIP was devised by an alcoholic and it shows. :( 04:24:34 O, and I would then have Huffman-encoded instructions, and many of the instructions implement important parts of decompression algorithms and other high-level stuff too; it is a very CISC computer. 04:26:33 Also the output has to be made, it be set to go to a buffer or to stdout. (It could also have stdin, which is going to be used for encryption.) 04:48:02 -!- weissschloss has quit (Remote host closed the connection). 04:49:13 pikhq: Looks like adding data can get corrupted if there's a hash collision in SHA1? 04:50:49 ^^probably not relevent 04:52:10 -!- weissschloss has joined. 04:52:25 Although SHA1 is disliked these days, isn't it? 04:53:24 Yes, there's theoretical exploits against it that weaken its collision resistance. 04:54:11 GNU C is getting a bit slow even when I didn't turn on the optimization 04:54:18 Though strictly speaking it's not attackable in practice yet to our knowledge, it's to the point that you should really move away from it *now* so you don't get caught with your pants down. 04:54:52 Is it necessary to tell it to turn off the optimizations? 04:55:12 pikhq: is ZPAQ updatable to make dedup reliant on SHA-256 instead? 04:55:17 I do not know. 04:56:07 :( at C++ API... so many lesser known programming languages can talk to C but not C++ 04:57:27 Can you use it in C++ if you use a extern "C" block to load them? 04:58:44 "time gcc -g -O0 -o nanozil nanozil.c" says it is 15 seconds 05:00:00 pikhq: so... this model stuff... would it make sense to, say, make a specialized Brainfuck model to compress more easily brainfuck code? 05:00:18 Which presumably could then be decompressed with any compliant ZPAQ decompressor 05:00:50 Sgeo: Things like these are some of things I have wanted to do, too 05:04:09 "To mark the end of the data, each decoded byte 05:04:10 is preceded by a EOS (end of segment) bit, which is 1 after the last byte, coded with p1 very near 0. " 05:04:31 So, is it possible that a model that makes a strong prediction but gets it wrong confuses the decompressor? 05:08:25 Clearly, the optimization I have implemented is broken. 05:11:03 This is the input file: http://sprunge.us/JIdV This is output with optimization turned off: http://sprunge.us/TPAM This is output with optimization turned on: http://sprunge.us/SXPa 05:13:07 -!- Guest15560 has quit (Ping timeout: 245 seconds). 05:17:05 I suppose it is to be expected; I didn't test it until now. 05:17:49 Can you help me to fix it? 05:23:39 I added #ifndef OMIT_OPTIMIZATIONS and -DOMIT_OPTIMIZATIONS for now; it causes GCC to run much faster now. 05:26:13 -!- AndoDaan_ has joined. 05:29:09 -!- AndoDaan has quit (Ping timeout: 260 seconds). 05:30:15 -!- conehead has joined. 05:39:45 -!- conehead has quit (Quit: Computer has gone to sleep). 05:51:58 -!- augur has quit (Quit: Leaving...). 05:55:15 -!- conehead has joined. 06:04:19 I should stop being surprised when this Celtic music sometimes sounds Enya-esque 06:07:15 -!- shikhin has joined. 06:07:21 -!- shikhin has quit (Client Quit). 06:07:39 -!- shikhin has joined. 06:15:50 -!- augur has joined. 06:23:14 -!- password2 has joined. 06:27:33 -!- Patashu_ has joined. 06:27:34 -!- Patashu has quit (Disconnected by services). 06:27:39 -!- Patashu_ has changed nick to Patashu. 07:04:42 -!- Lymia has quit (Remote host closed the connection). 07:04:48 web.test: points 2.17, score 22.16, rank 14/47 07:05:52 is that a way to submit programs via the web 07:06:09 Yeah, http://zem.fi/bfjoust/ at the bottom 07:06:21 I just made a mistake submitting it for real. 07:07:10 web.test: points -46.00, score 0.00, rank 47/47 (-33) 07:08:16 -!- password2 has quit (Quit: Leaving). 07:08:51 !bfjoust BeatYouMate >(-)*6>(+)*7>-(+)*17>(-)*12>(+)*8>(-)*7>(+)*8>(+)*3>[(-)*7[+]]>>[(+)*7[-]]>>([(+)*14[-]]>)*3([(-)*14[+]]>)*3[(-)*7[+]]>>[(+)*8[-]]>>([(+)*14[-]]>)*3[(-)*14[+]]>[(+)*14[-]]>[(-)*16[+]]>[(-)*7[+]] 07:08:51 AndoDaan_.BeatYouMate: points 2.38, score 22.40, rank 13/47 07:09:03 -!- password2 has joined. 07:11:46 !bfjoust BeatYouMate <>< 07:11:46 AndoDaan_.BeatYouMate: points -46.00, score 0.00, rank 47/47 (-34) 07:11:55 -!- AndoDaan_ has changed nick to AndoDaan. 07:11:59 !bfjoust BeatYouMate >(-)*6>(+)*7>-(+)*17>(-)*12>(+)*8>(-)*7>(+)*8>(+)*3>[(-)*7[+]]>>[(+)*7[-]]>>([(+)*14[-]]>)*3([(-)*14[+]]>)*3[(-)*7[+]]>>[(+)*8[-]]>>([(+)*14[-]]>)*3[(-)*14[+]]>[(+)*14[-]]>[(-)*16[+]]>[(-)*7[+]] 07:11:59 AndoDaan.BeatYouMate: points 2.95, score 22.58, rank 13/47 (+26) 07:13:24 web.DeadFish: points -46.00, score 0.00, rank 47/47 07:14:03 oh jousting 07:14:12 Indeed. 07:17:09 penile jousting 07:25:11 -!- Tod-Autojoined has changed nick to TodPunk. 07:30:59 -!- oerjan has quit (Quit: More like piscile). 07:48:50 -!- conehead has quit (Quit: Computer has gone to sleep). 07:54:53 "Every compressor that can compress any input must also expand some of its input. However, the expansion never needs to be more than one symbol. Any compression algorithm can be modified by adding one bit to indicate that the rest of the data is stored uncompressed." 07:55:01 I wonder how many people fail to consider that last part 08:29:21 [wiki] [[Special:Log/newusers]] create * MaurineWooten * New user account 08:41:38 -!- drdanmaku has quit (Quit: Connection closed for inactivity). 08:48:36 -!- shikhin has quit (Quit: Lost terminal). 08:49:07 -!- shikhin has joined. 09:04:48 Sgeo: i would imagine that any compression algorithm worth its salt would include that logic, or something equivalent to it? 09:05:54 quintopia: most compression algorithms include similar logic, but not exactly that strict, so they might expand strings by slightly more than one symbol 09:08:38 b_jonas: depends on how you define symbol :P but yeah, I'd expect them to at least expand no input by more than a constant 09:09:09 quintopia: they don't usually do that, because they want a streamable compression (with very large buffer), 09:09:37 so eg. if the first gigabyte of the input is uncompressible, they write that first gigabyte uncompressed with a length marker before reading the second gigabyte 09:09:44 so you get expansion by a small constant factor 09:10:06 they don't want to have to read the whole input to decide whether to use a constant long marker marking that all the data is uncompressible 09:10:29 b_jonas: not all compressed data needs to be streamed. for streamable data, i expect each encoded unit to expand by no more than a constant 09:11:19 sure, not all, but all the compression methods for very long input are. there's methods for limited sized inputs, where obviously the input extends only by a constant for any encoder with finite state. 09:13:12 in practice, the important part is that the compression doesn't expand the input by too much 09:14:36 also note that some compressions, such as gzip (zlib with the gzip header), are defined as concatenatable: if you concatenate two compressed outputs, you can decompress them as the catenation of the two inputs. that already precludes the compressor being able to compress with just a constant marker at the end, because after such marker it couldn't find where you concatenate the next gzip file. 09:14:48 presumably this was useful for casette tapes. 09:20:21 it doesn't preclude that each of the two parts that were concatenated has only a constant tag 09:26:16 quintopia: it does 09:26:39 quintopia: if there's a constant tag that quotes all the rest of the file, then you can't recognize where that file ends and the next compressed one starts 09:27:15 quintopia: this works even if the form of the constant length quote tag varies as long as you can quote any input with a constant length prefix tag 09:41:33 -!- impomatic_ has joined. 09:56:55 -!- Phantom_Hoover has joined. 10:16:48 -!- Lymia has joined. 10:16:50 Opps! 10:29:46 IDSPISPOPD 11:06:56 -!- Sorella has quit (Ping timeout: 260 seconds). 11:21:55 -!- evalj has joined. 11:56:55 -!- shikhout has joined. 12:00:13 -!- shikhin has quit (Ping timeout: 260 seconds). 12:14:22 -!- password2 has quit (Ping timeout: 240 seconds). 12:20:47 -!- GeekDude has joined. 12:46:12 b_jonas: not if the constant tag encodes the length of the file 12:46:27 -!- AndoDaan_ has joined. 12:47:14 quintopia: sure, but then it's no longer constant 12:47:27 b_jonas: with the qualification of course that "any length" is upper bounded by a constant 12:48:21 -!- AndoDaan has quit (Ping timeout: 246 seconds). 12:48:23 yeah 12:51:28 -!- AndoDaan_ has quit (Read error: Connection reset by peer). 12:51:47 -!- AndoDaan has joined. 12:53:50 -!- boily has joined. 13:00:17 -!- Patashu has quit (Ping timeout: 245 seconds). 13:02:28 Best webdesign: phone number form can't be tabbed out or moved within with cursor keys (part of input validation, apparently) and the submit button fires a jQuery $.ajax request with only a "success" handler, which proceeds to fail (20-second timeout) and leave the page to a "please wait" + spinning thing mode. 13:04:08 fizzie: yeah, typical 13:04:31 -!- FreeFull has quit (Quit: BBL). 13:07:37 -!- Sorella has joined. 13:08:29 -!- Sorella has quit (Changing host). 13:08:29 -!- Sorella has joined. 13:09:07 lol 13:09:22 Why would you need jquery to submit a form? 13:10:45 [wiki] [[Special:Log/newusers]] create * MarylinYagan * New user account 13:11:18 HackEgo: help 13:11:24 pfft 13:14:46 It's one of those multi-step "wizard"-style forms, except it's just the one page and they move from step to step with scripts that do ajax requests and populate the later steps based on the results, then animate fancy side-swipe things to move to the next step. 13:15:01 T_T 13:15:56 * GeekDude wonders if you could do it with a complicated css script that relies on the status of check boxes to hide or show other elements 13:20:37 fizzie: stupid 13:20:50 GeekDude, my guess? "Yes, but don't" 13:20:57 lol 13:21:02 I want to do it for the lulz 13:21:05 Which, this being #esoteric and me being me, I'll give it a go this evening probably 13:21:13 I don't know enough CSS though 13:22:36 GeekDude, can you send me a @tell to remind me? I'm heading out now 13:22:41 uhh 13:22:44 I can send youa mmemo 13:22:46 memo* 13:22:49 with memoserv 13:22:57 Or a tell with lambdabot 13:23:03 @tell GeekDude like this 13:23:04 Consider it noted. 13:23:05 I don't know these bots 13:23:17 ah 13:23:18 okay 13:23:55 @tell Taneb GeekDude wonders if you could do it with a complicated css script that relies on the status of check boxes to hide or show other elements 13:23:55 Consider it noted. 13:46:52 -!- Sorella has quit (Ping timeout: 245 seconds). 13:47:06 -!- AndoDaan_ has joined. 13:47:54 -!- AndoDaan has quit (Ping timeout: 244 seconds). 13:48:52 -!- AndoDaan_ has changed nick to AndoDaan. 13:50:19 GeekDude: the bots are our leaders. 13:50:23 @metar CYUL 13:50:24 CYUL 281300Z 25010KT 15SM FEW240 18/15 A3013 RMK CI1 CI TR SLP202 DENSITY ALT 200FT 13:51:00 I lead a guy on for a minute or two by copy and pasting into my IRC client. He though he had shell access on my bot 13:51:26 -!- paul2520 has joined. 13:52:47 -!- password2 has joined. 13:54:39 hunter2? 13:56:37 -!- GeekDude has quit (Quit: {{{}}{{{}}{{}}}{{}}} (www.adiirc.com)). 13:58:41 -!- GeekDude has joined. 14:05:22 [wiki] [[Object oriented thue]] http://esolangs.org/w/index.php?diff=40548&oldid=40515 * 4D enthusiast * (+139) /* stdio */ Simplified escpe codes, and made the description mroe explicit 14:09:29 -!- GeekDude has quit (Quit: {{{}}{{{}}{{}}}{{}}} (www.adiirc.com)). 14:11:12