00:00:00 elliott: is what oerjan's saying supposed to make sense or am I juts realyt iresd 00:00:08 you might just be stupid 00:00:25 canw e redact the logs so i can say shit to kallisti and he can't prove it when he's slept 00:00:29 that'd be fun 00:01:56 I think I just get confused when I try to regex substitute and there's nothing to substitute. 00:02:16 unless :%s is like some weird reverse substitute 00:02:32 it's called vi syntax 00:02:35 you uncultured fuck 00:02:55 elliott: I suffer from lack of omniscience, you insensitive clod! 00:03:24 omni science 00:04:20 oh okay I think I get it. 00:04:46 right, OERJAN REMOVED THE CONCEPT OF SLEEP 00:04:49 ha ha ha ha ha ha ha ha ha ha 00:04:52 so fresh 00:06:12 -!- Vorpal has quit (Ping timeout: 248 seconds). 00:10:50 it's always redonkulous freshnasty up in this biznatch 00:23:07 http://hackage.haskell.org/package/toilet 00:24:00 (Copying from my Facebook 'cuz I'm lame): Mind-bender turned into a computer science problem: Imagine a class of functions rand[n] which return a random integer in the range [0, n), with equal probability of all valid values. Given an implementation of rand[2] as an oracle, is it possible to implement rand[3]? If not, for what values of n (define inductively or algebraically) is it possible? If so, what is the runtime (big-O notation) of rand[3], assuming rand[ 00:24:00 2] is O(1)? 00:24:52 Gregor: Those aren't functions QUOD ERAT DEMONSTRUDEL 00:25:08 elliott: My next statement on FB was: Incidentally, I am of course not using "function" in the mathematical sense. rand[2] is magic, and rand[>2] is an algorithm (or none). 00:25:35 Gregor: Anyhow, I'm pretty sure the answer is "sort of": namely, yes it can be done, but the worst case is O(infinity). 00:25:48 That is "yes", not "sort of" :P 00:25:51 It's basically just "given an unbiased RNG implement a bigger one", which is a less obscure problem than that P: 00:25:53 *:P 00:25:56 Gregor: Uh, no it's not. 00:26:06 Gregor: It's hardly a perfect RNG if it can fail to give a random number. 00:26:18 It cannot fail, it can merely take infinite time. 00:26:27 Those are the same thing. 00:26:36 But OK: rand[3] := for(;;); 00:27:10 My canonical rand[3] will always terminate, given infinite time, because in infinite time rand[2] will always return any given pattern eventually. 00:27:23 Oh, okay. 00:27:33 Gregor: That isn't actually true though. 00:27:41 Gregor: I mean, it isn't implied by "perfectly unbiased RNG". 00:27:51 it is true modulo null sets :P 00:27:51 The probability approaches 1, but you can easily get 0, 0, 0, 0, ... forever. 00:27:59 It's just stunningly unlikely. 00:28:08 But that matters :P 00:28:22 elliott: I'm pretty sure that that becomes untrue the very moment "forever" turns into "for infinite time" 00:28:34 elliott: you _could_ interpret it as rand[2] never giving a particular probability 0 behavior you're interested in excluding 00:28:35 Because summed to infinity, "stunningly unlikely" becomes "impossible" 00:28:49 Gregor: ...no. 00:28:51 It has probability 0. 00:28:54 That is not the same thing as being impossible. 00:29:48 oerjan: INJECT THINE WISDOM 00:30:02 oerjan: I could, but that's not what Gregor said. 00:30:10 Gregor: um i already did 00:30:19 oerjan: INJECT MORE OF THINE WISDOM 00:31:04 rand3 = do 00:31:04 a <- rand2 00:31:04 b <- rand2 00:31:04 case (a,b) of 00:31:04 (0, 0) -> return 0 00:31:05 (0, 1) -> return 1 00:31:07 (1, 0) -> return 2 00:31:09 (1, 1) -> rand3 00:31:11 If this isn't a solution (I'm not sure it is), then I think it can be made into one with just a sufficient drop of clever (which I'm willing to try and come up with if this isn't it :P) 00:31:13 it's really a matter of what you define as acceptable algorithms. there isn't much more to it than that... 00:31:41 Yuh, that was my canon solution. 00:32:19 Gregor: also for any n=2^m we can define rand_n as follows: rand_n = foldl' (\a b -> (a*2)+b) 0 <$> replicate m rand_2 00:32:29 obviously 00:32:32 Well it's easy for 2^n :P] 00:32:38 right :P 00:32:47 * elliott tries rand5 00:32:54 oh hm 00:32:55 that's easy 00:32:57 you have rand3 00:33:00 so just apply the same trick rand3 does 00:33:16 * elliott tries rand7 with rand{2,3,5} 00:34:19 Gregor: in fact I think you basically just need ceil(log_2(n)) rand2s 00:34:25 and you only ever have zero or one recursion cases 00:34:42 it's basically just "interpret as bit pattern, try again if it's too big" 00:34:52 Gregor: but personally, I think the real answer is that it is impossible 00:34:56 Surely there are cases where there would be >1 recursive case though? 00:35:10 next: try to do it for general n without wasting any bits >:) 00:35:11 e.g. 13? 00:35:14 Gregor: well ceil(log_2(n)) never exceeds log_2(n) by much... 00:35:18 but okay I'll try 13 00:35:27 14, 15 and 16 are all invalid. 00:35:45 I mean, unless you consider "default" to be one case :P 00:35:46 Hmmmm 00:35:53 Invalid as in can't be implemented? 00:36:06 Erm. 00:36:10 I was talkin' nonsense for a second there. 00:36:30 elliott: he means that all of 14, 15 and 16 would be recursion cases 00:36:30 13, 14 and 15 all require rerolls for rand13 is what I meant to say. 00:36:37 Yeah 00:36:50 Right. 00:37:11 Gregor: But anyway, a perfect unbiased oracle RNG can still return a pathological sequence forever. Yes, it has probability 0, but this does /not/ mean it's impossible, and therefore rand3 can fail to terminate, and therefore rand3 isn't an unbiased RNG because, uhh, RNGs terminate. 00:37:19 So the problem is actually impossible. 00:37:29 Not that this doesn't work in practice, but in practice you don't have rand_2 anyway. 00:38:09 * Gregor hmms for a moment. 00:38:46 Gregor: Basically, you have to take every possible infinite bitstring, and prove that the results are evenly-distributed between them all as rand2's successive return values. 00:38:56 1,1,1,1,1,1,1,1,... -> whoops there is no result 00:39:29 ...not that I can prove there is _no_ algorithm that works to do this, although I'm sceptical 00:39:34 but the canonical solution ain't it 00:39:53 and I'm pretty sure it must have been proved impossible a long time ago, it's not like it's a particularly obscure thing to think about :P 00:39:59 I certainly understand that that case /exists/. However, infinity is a funny thing ... I'm not confident that the fact that there exists such an infinite bitstream negates the correctness of the algorithm ... 00:40:35 But then I start doing math with infinities and that makes me kill myself. 00:40:39 Gregor: I'm gonna defer to oerjan here because he has the Ph.D. to make you believe him :) 00:41:27 That's why I asked him to inject more wisdom. 00:41:29 :P 00:41:47 Gregor: But basically, if you accept that one possible sequence fails, then for 13, you have to accept that an infinite number of possible sequences fail. 00:41:55 (Those composed out of the three reroll cases.) 00:42:14 Gregor: And the sequences that fail have the same cardinality as the sequences that work. 00:42:21 (same cardinality as the reals) 00:44:05 Gregor: Basically rand3 is a function from a real to [0,2]. 00:44:11 That's a better point. 00:44:11 ([0,2] in Z that is) 00:44:18 Gregor: rand3 fails on one real. 00:44:22 Yuh 00:44:27 rand13 fails on |R| reals. 00:44:29 And works on |R| reals. 00:44:32 Right. 00:44:58 Gregor: To make rand3 work, you have to chop that real off... and suddenly the RNG isn't perfectly unbiased any more. 00:45:12 Same for rand13 but much more drastic :) 00:45:26 what about rand14? 00:45:27 Gregor: Particularly, note that you can predict future results if you disallow infinite 1s. 00:45:39 Gregor: If you get a 1, you know you're going to get a 0 within a finite amount of time. 00:45:44 :D) 00:45:50 You can't predict unbiased oracles :P 00:46:19 elliott: i have the Ph.D. to understand that the interpretation of whether the infinite case is actually possible is a matter of your taste in definitions. 00:46:33 Yeah, that was my perspective :P 00:46:41 But I'm agreeing more with elliott's now *shrugs* 00:46:44 Gregor: Actually you don't even need to get a 1 to know you'll get a 0, obviously :P 00:47:05 oerjan: I think anyone who calls a skewed-to-make-a-challenge-possible RNG "unbiased" is stretching their definitions :P 00:48:23 That was not the issue with definitions. 00:48:25 elliott: but it is modification on a set of probability zero, which is frequently considered acceptable 00:49:29 oerjan: *shrug* I think it's fair to say that the challenge is ambiguous 00:49:48 But I think the number of combinations that lead to reroll is unbounded as n gets larger 00:49:52 (note: even your cardinality |R| exception set example still has probability 0) 00:50:06 If only the naturals weren't countable, you could prove that you were eliminating most of R in the process, I think. 00:50:44 Gregor: Ooh, ooh, here's another argument: You can imagine taking this "reroll" strategy to the limit, and deriving an algorithm to generate a perfectly random, equal-probability /integer/. 00:50:52 You just have to reroll every time :-) 00:51:02 That made more sense as a visualisation in my head. 00:51:11 i never claimed the challenge wasn't ambiguous, it certainly has that "teacher gives assignment which requires you to know which of equally plausible interpretations they subscribe to" vibe 00:51:18 Actually it's more like you have to call rand2 an infinite number of times. 00:51:24 So I guess this isn't realy compelling. 00:51:25 oerjan: Right. 00:51:35 (note: even your cardinality |R| exception set example still has probability 0) 00:51:37 I'm aware of this 00:51:56 elliott: just in case ;P 00:51:59 "An RNG that can return any sequence of probability >0" just doesn't count as unbiased to me :) 00:52:24 oerjan will now prove that it's OK because probability 0 has probability 0 in the space of probabilities, so it's generally accepted to remove it. 00:52:39 -!- pikhq has joined. 00:53:06 The thing is, for all valid possibilities, there is not only infinitely many such valid bitstrings, but a nonzero probability of such bitstrings. 00:53:25 (Valid = non-halting) 00:53:28 -!- pikhq_ has quit (Ping timeout: 268 seconds). 00:53:29 Erm 00:53:32 (Valid = halting) :P 00:53:32 What 00:53:48 Yes, all the non-halting sequences have probability 0 00:53:49 i'd say just get a counter which counts really quickly and use human inaccuracy to randomly hit the counter 00:53:54 But that doesn't matter 00:54:00 You can exclude the sequences from your RNG definition 00:54:05 and use the supplied value as an index into the decimal places of PI 00:54:06 And claim it doesn't affect biasedness because they all have probability 0 00:54:15 But if you /do/ admit such sequences, it's definitely impossible 00:54:49 elliott: erm no you cannot remove _all_ probability 0 sets, because they're union has probability 1. duh. 00:54:53 so suppose the counter is 32bits cycling.. human presses the button and gets 100,000,000 .. so look up the 100,000,000th decimal of PI 00:55:00 but you can remove any _particular_ one. 00:55:03 oerjan: The sequences = the sequences that break things 00:55:12 oerjan: Actually this is problematic because I believe every single probability 0 sequence is involved in the general case 00:55:15 For solving the problem for all n 00:55:20 Anyway, that's not what I was saying :P 00:55:31 For rand1 it's 1 sequence, for rand13 it's |R|, you gotta exhaust all the probability-0 sequences sometime :-) 00:55:34 Although, hmm 00:55:36 elliott: no i'm pretty sure it's not. 00:55:39 Yeah no it's not 00:55:42 e.g. Chaitin's constant 00:55:51 (For rand3) The problematic sequence has probability 0. The non-problematic sequences have non-zero probability. 00:55:51 Pretty sure you can't "recognise" it like that 00:55:57 Wow this is so ridiculously informal 00:56:02 Gregor: Yes, I agree. 00:56:15 Gregor: That doesn't matter. What does matter is that you can claim your definition of the RNG excludes the problematic sequence. 00:56:19 elliott: in fact there are countably many n's, so the union of the sets still has probability 0 by countable additivity. 00:56:25 oerjan: Right. 00:56:30 NOW LET'S GENERALISE IT TO THE REALS 00:58:03 itidus20: your comment reminds me of the adage going somewhat like "randomly choosing an algorithm is not a good way to make a random number generator" 00:58:50 *their 01:01:10 oerjan: i think its probably quite effective though 01:02:24 hm 01:02:37 itidus20: you cannot increase randomness by indexing a sequence 01:02:46 lol 01:02:47 well hm 01:02:50 i guess that's how prngs work :P 01:02:59 but in a strict sense pi_digits[n] is no more random than n... 01:03:37 uh.. god used the pi sequence as an axiom for creating the universe so its ok (^_^ 01:03:40 I suppose I can't brag that I got Gregor's canonical answer, given that the truth of said answer is in dispute 01:03:58 elliott: interesting point there :-? 01:04:02 Well it's a pretty obvious answer :-) 01:04:16 elliott: however there are some practical advantages 01:04:19 uhmm 01:04:26 Although when I first read the status, I missed "integer" 01:04:46 elliott: with my system you have less chance of knowing which digit is coming next 01:04:52 *facepalm* 01:04:57 if you can predict n, you can predict pi_digits[n] 01:05:03 how? 01:05:07 predict n, then index pi_digits with it! 01:05:28 ok so if you can only predict like 01:05:30 n-2 to n+2 01:05:36 then you can predict "closer" with just n 01:05:39 but uh 01:05:43 this scheme is ridiculous anyway 01:05:43 ok well.. suppose that the device i described gives you a digit 01:05:49 uhmm 01:06:11 but it is very unlikely that someone could activate the fast counter at such a rate that that knowledge could help them anyway :-? 01:06:12 * Sgeo has the impression that Pi is not a cryptographically secure PRNG 01:06:30 Wait, actually, it's unknown if it's normal, right? And if it is normal, that means...? 01:06:32 the whole system relies on human stupidness to push some button 01:07:11 like if our human looks out the window at a passing bird and gets distracted that should be enough to throw his game out 01:07:21 If the digits of Pi turn out to be "normal" I think is the word, then would Pi work as a PRNG with a seed of a starting digit? 01:08:15 Sgeo: i think a normal number can still be entirely predictable. at least for a single base this is true. e.g. 0.12345678910111213... is normal. 01:08:16 -!- elliott_ has joined. 01:08:17 -!- elliott has quit (Read error: Connection reset by peer). 01:08:40 (base 10) 01:08:43 i am only thinking of a dice replacer though really 01:08:46 Maybe I should learn what is meant by normal 01:09:11 i am thinking of like a box which has a button on the top :D and you can push a button and a digit is displayed 01:09:27 Most PRNGs do not generate a normal sequence, because doing so requires unbounded memory 01:09:29 and it has 2 AA batteries... 01:10:25 Is 123456789/999999999 normal? 01:10:28 Jafet: s/normal/nonperiodic/, even 01:10:33 Also, predictability has nothing to do with whether all sequences appear with equal probability 01:10:34 Sgeo: no. 01:10:49 no rational number is normal in any base, as its expansion repeats. 01:11:16 which means most strings longer than the repeat won't occur 01:11:18 I was thinking that each digit appears an equal amount of times in the decimal expansion 01:11:29 But that's not what normal is I guess 01:11:34 Sgeo: normal requires blocks of digits, not just single digits 01:11:47 ok ok i would technically be repeating pi 01:11:49 just not very often 01:12:06 i would use the first say 65535 digits of pi and then cycle 01:12:17 that's 01:12:18 pretty often 01:12:18 well i was thinking 32bit but i dont know the relevant number 01:12:51 and maybe its diffciult to store 32bit digits.. or maybe it's not 01:13:10 that's 4 gigabytes of ram 01:13:12 but why the heckw ould you stoer it 01:13:17 you can compute an arbitrary digit of pi in O(1) 01:13:24 :o 01:13:35 at least in a few bases 01:13:36 elliott_, in decimal? Or Binary? 01:13:37 hummmm 01:13:44 hex 01:13:47 i think octal too 01:13:54 O(log n) 01:14:01 Jafet: well right 01:14:02 well 01:14:03 only because of arithmetic 01:14:12 pi is nonperiodic, you obviously can't compute any digit in O(1) 01:14:27 elliott_: hint, calculating in base b^m is equivalently easy to calculating in base b^n 01:14:30 point is, you can do the full 64 bits with much less storage space 01:14:32 oerjan: indeed 01:14:36 god im so full of crap 01:14:43 oerjan: 10 isn't 8^n though 01:14:43 anyway 01:14:56 elliott_: shocking 01:15:16 so the other element of my suggested device is a counter which just cycles very quickly through n bit integers cycling 01:15:33 and idiot human presses a button which makes it stop and then calculate that digit of pi 01:15:40 and display it 01:15:57 and then it starts counting again 01:16:56 That sounds no different from a randomly at device-programming time generated table from n-bit integers to final numbers. I guess that's the point though 01:17:18 you should take into account im the dumb kid 01:17:38 -!- elliott_ has quit (Changing host). 01:17:38 -!- elliott_ has joined. 01:17:41 my ideas are not supposed to be acceptable :P 01:17:49 Be careful you don't hire a pseudo-idiot. Then your numbers become pseudo-random. 01:18:10 well i mean in here 01:18:20 i am really quite clueless and oblivious 01:18:46 elliott_, Oleg says that Writer monad is different from generators 01:19:21 Sgeo: never said they weren't 01:19:25 oerjan: /nick elliott for a bit would you old chap 01:19:32 eek 01:19:38 -!- oerjan has changed nick to elliott. 01:19:44 -!- elliott has changed nick to oerjan. 01:19:48 um 01:19:50 longer than that 01:19:51 i need to reconnect 01:19:55 i got a 30 second warning 01:19:57 yeah 01:20:00 it'll take less than that >:) 01:20:20 * Sgeo doesn't know how much he trusts elliott_, but 01:20:22 -!- Sgeo has changed nick to elliott. 01:20:24 -!- elliott_ has quit (Quit: Leaving). 01:20:26 tick tock 01:20:28 -!- elliott has quit (Disconnected by services). 01:20:28 -!- elliott_ has joined. 01:20:35 as i expected, doesn't quite work. 01:20:52 -!- Sgeo has joined. 01:21:02 lol 01:21:02 * Sgeo glares 01:21:04 _someone_ is going to have gained a lot of trust in elliott_, i predict 01:21:22 i don't think i can make Sgeo stop trusting me however hard i try 01:21:28 also believing everything i say 01:22:20 Nickserv will help you find nicks that utilize a certain string. For example, if you wanted a list of any nicks on freenode that contain “chick”, you could type /msg nickserv list *chick*. Nickserv would then return a list of all non-private nicks that contain the string “chick”. 01:22:22 --freenode blo 01:22:22 g 01:23:02 good example 01:24:06 sigh, i believe i'd have to write a script 01:24:25 For what? 01:24:35 Some have proposed that the reception of the Holy Ghost occurs automatically without any manifestation of the supernatural when a person simply "accepts Jesus into their hearts by faith." Others believe this occurs automatically when one is baptized in water. They propose that the "baptism of the Holy Ghost" noted in Acts 2:4 is a "second experience" that may or may not necessarily occur after one simply receives the Spirit of God within one's 01:24:35 life. In this study, I will call their belief the "two experiences" belief. 01:24:43 Sgeo: autoghost 01:26:55 * Sgeo is rather surprised by delimited continuations being more powerful than undelimited continuations 01:27:34 i understand delimited continuations can do state 01:28:01 humm 01:28:05 -!- elliott_ has set topic: IM A CORPUS FRIENDSHIP ROBOT!!! HOW DO *YOU* TOUCAN? The IOCCC is back on! http://www.ioccc.org | http://codu.org/logs/_esoteric/. 01:28:14 -!- derdon has quit (Remote host closed the connection). 01:28:26 * oerjan wonders if this toucan meme is from somewhere 01:28:30 yes i put it in the topic 01:28:31 and then 01:28:33 beautiful meme 01:28:34 spiral 01:28:39 All I know is that call/cc is undelimited, and Cont's callCC is delimited 01:28:51 i can, you can, one or two can! 01:28:55 And I may be mistaken 01:28:57 Sgeo: it's delimited by the runCont 01:29:01 i scream, you scream, we all scream for ice cream! 01:29:15 Who's submitting to the IOCCC? 01:29:23 Sgeo: or whatever, I think 01:29:26 Jafet: Gregor has 01:29:31 ais is/was planning to 01:31:08 Baaaaaaaaasically everybody in here :P 01:31:17 all 2 of you 01:31:53 "We assume basic familiarity with functional programming languages, such as OCaml, Standard ML, Scheme, and Haskell. " 01:32:03 Is familiarity with only Scheme and Haskell sufficient to read this? 01:32:07 No. 01:32:10 You must understand EVERY SINGLE ONE 01:32:29 also Clean, they just left that out as a typo 01:32:38 You must dissociate personalities to learn ocaml and sml 01:32:52 Don't forget about Lazy-K. 01:33:58 $ cabal install webkit 01:34:00 * elliott_ crosses fingers 01:34:03 Gregor: i was going to mention that if someone asked 01:34:21 maybe i'll write it in python, just this once :) 01:34:25 * elliott_ is scared of gtk2hs 01:35:01 Why is Oleg literally God? 01:37:06 Also C. 01:37:24 C is a functional language in which you define your program as a set of functions of type State -> State. 01:44:50 where weak typing allows State to mean basically anything! 01:44:56 :> :> :> 01:51:13 aren't you confusing with Underload here 01:53:54 Gregor: http://conal.net/blog/posts/the-c-language-is-purely-functional 01:59:06 -!- zzo38 has joined. 02:02:36 -!- Jafet has quit (Quit: Leaving.). 02:05:17 Yesterday I was playing Dungeons & Dragons game. Partially by suggestion of the first officer, now I have to go into the prison in the night time, and then leave, without breaking it... . . . . . . . . . . . . . . . 02:06:12 -!- MSleep has quit (Ping timeout: 248 seconds). 02:09:13 -!- elliott_ has quit (Quit: Leaving). 02:10:37 -!- lambdabot has joined. 02:12:56 -!- lambdabot has quit (Remote host closed the connection). 02:13:06 promising. 02:13:29 elliott (not here): Oh piff, that's such a boring way for C to be functional :( 02:14:12 elliott (still not here): I was actually dredging up an earlier argument in which somebody (you?) argued that concatenative languages are functional, where each operation is a function State -> State. I generalized it to C. 02:15:30 Gregor: It's analogous to a particular interpretation of the execution of Haskell, though. 02:16:27 pikhq: Yes, and? 02:16:41 -!- lambdabot has joined. 02:16:47 Nothing in particular. 02:16:51 :P 02:16:56 pikhq: SO WHERE'S YOUR IOCCC ENTRY? 02:16:58 HUH? 02:16:59 PUNK? 02:18:07 My apathy and laziness are hiding it. 02:27:40 > "wh"++repeat'e' 02:27:41 Not in scope: `repeat'e'' 02:27:47 > "wh"++repeat 'e' 02:27:51 mueval-core: Time limit exceeded 02:27:59 > "wh"++repeat 'e' 02:28:03 mueval-core: Time limit exceeded 02:28:41 > "wh" ++ repeat 'e' 02:28:45 mueval-core: Time limit exceeded 02:28:47 lambdabot... 02:29:00 > 'w' : 'h' : repeat 'e' 02:29:01 "wheeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee... 02:29:22 i rather doubt the exact syntax matters to the timeout 02:29:41 I was just retyping it and I always put spaces in my stuff 02:29:51 I can't control it! 02:30:03 spaces addiction serious problems 02:32:10 * Scroll of create paper: The spell on this scroll creates one sheet of paper. 02:32:17 * Scroll of melee: This is a magical scroll that can be used as a melee weapon. 02:32:24 * Scroll of cure blindness: If you read this scroll while blind, your blindness is cured. This affects only the reader of the scroll, not anyone else. 02:32:31 * Scroll bar: This kind of scroll is sold in bars. 02:32:49 -!- elliott has joined. 02:33:09 Anyone happen to know any internal gtk guts? 02:33:13 Or equivalently, X11 guts? 02:33:15 Gregor? pikhq??? 02:33:36 * Potion of invisibility: This potion makes everything invisible to the drinker. 02:33:42 * Potion of indivisibility: This potent potion changes composite numbers into prime numbers. 02:33:47 * Potion of melee: This potion can be used as a melee weapon, but only once because after you use it once it shatters. 02:33:52 * Potion of eggplant: In a few weeks, plants start growing on your head. The plants have eggs on it. 02:34:12 * Poison of healing: This is both poison and healing, so it has both effects at once. 02:34:12 * Scroll of paper bird: This scroll can be folded to glide through the air. 02:35:56 * Arrow of bad luck: If you ever try to use this you will always miss even if you rolled an automatic hit. 02:36:14 elliott: All I know about the guts of either is that Cthulhu lies in wait 02:37:49 * Arrow of time: This arrow causes its victim to die of old age. Eventually. 02:38:18 -!- copumpkin has changed nick to pirateat20. 02:38:36 it's the cannibal pirate 02:39:40 -!- oerjan has quit (Quit: Good night). 02:41:40 :O 02:43:41 -!- pirateat20 has changed nick to copumpkin. 02:43:52 -!- aloril has quit (Ping timeout: 252 seconds). 02:46:16 screenshot.py:14: GtkWarning: IA__gdk_offscreen_window_get_pixmap: assertion `GDK_IS_WINDOW (window)' failed 02:46:19 WHYYYY 02:46:32 -!- MSleep has joined. 02:56:27 -!- aloril has joined. 03:00:18 Gregor: You've used WebkitGtk, right? 03:11:50 No, I've only used QtWebkit. 03:12:54 -!- azaq23 has joined. 03:13:07 -!- azaq23 has quit (Max SendQ exceeded). 03:18:18 Gregor: Ah, okay then 03:18:30 update 03:18:48 hi 03:20:15 -!- azaq23 has joined. 03:20:39 elliott: I think we both know that it's /not/ okay. 03:21:45 -!- MSleep has changed nick to MDude. 03:25:02 Gregor: Indeed it's not, this thing is broken :'( 03:29:02 WHY DOES GTKOFFSCREENWINDOW DEPEND ON AN X1 SERVER 03:29:09 WHY 03:29:10 WHYYYY 03:29:23 Gregor: WHYYYY 03:37:09 Wow, X1 03:37:10 Ancient. 03:38:13 X-D 03:43:50 elliott: Strictly speaking, it should only depend on an X11 server if you use the X11 backend for GTK. 03:44:29 *In theory* it should have a "render to a buffer" backend. 03:44:42 elliott, Oleg used a $ chain for something 03:44:46 (I make no guarantees that it actually does) 03:46:23 Sgeo: Okay? 03:46:29 pikhq: I'm just using xvfb-run :P 03:46:50 iterate2 f x n | seq f $ seq x $ seq n $ False = undefined 03:46:50 pikhq: Which means I have a working URL -> WebKitGtk -> png tool \o/ 03:46:54 Sgeo: Okay? 03:46:55 http://okmij.org/ftp/Haskell/#making-function-strict 03:46:58 Sgeo: Okay? 03:47:14 elliott: Yeah, Xvfb would work. 03:47:19 elliott, it's just that you seem to tell me that using . is almost always preferable 03:47:20 Bit of a hack, but *eh*. 03:47:25 It is 03:47:27 s/almost // 03:47:40 So, Oleg is imperfect? 03:47:47 Does this surprise you 03:47:51 Sgeo: Oleg's Haskell style differs from the now commonly accepted Haskell style. 03:48:00 o.O 03:48:03 Of course, Oleg's Haskell style predates there *being* a commonly accepted Haskell style. 03:48:14 Oleg's code isn't particularly idiomatic at all, it's just really good, so who cares? 03:49:19 I seem to recall the history being that each university had its own Haskell coding style, and one of those took over the rest. 03:50:17 pikhq: Not true, Utrecht lives on 03:50:22 UNFORTUNATELY. 03:50:41 http://hackage.haskell.org/packages/archive/uu-parsinglib/2.7.3/doc/html/src/Text-ParserCombinators-UU-Core.html 03:50:46 instance Applicative (T state) where 03:50:46 T ph pf pr <*> ~(T qh qf qr) = T ( \ k -> ph (\ pr -> qh (\ qr -> k (pr qr)))) 03:50:46 ((apply .) . (pf .qf)) 03:50:46 ( pr . qr) 03:50:48 T ph pf pr <* ~(T _ _ qr) = T ( ph. (qr.)) (pf. qr) (pr . qr) 03:50:50 T _ _ pr *> ~(T qh qf qr ) = T ( pr . qh ) (pr. qf) (pr . qr) 03:50:52 pure a = T ($a) ((push a).) id 03:51:05 I... that's a lot of pointfree 03:51:08 Okay, this one is amazing: 03:51:09 best' :: Steps b -> Steps b -> Steps b 03:51:09 End_f as l `best'` End_f bs r = End_f (as++bs) (l `best` r) 03:51:09 End_f as l `best'` r = End_f as (l `best` r) 03:51:10 l `best'` End_f bs r = End_f bs (l `best` r) 03:51:10 End_h (as, k_h_st) l `best'` End_h (bs, _) r = End_h (as++bs, k_h_st) (l `best` r) 03:51:11 End_h as l `best'` r = End_h as (l `best` r) 03:51:13 l `best'` End_h bs r = End_h bs (l `best` r) 03:51:15 Fail sl ll `best'` Fail sr rr = Fail (sl ++ sr) (ll++rr) 03:51:17 Fail _ _ `best'` r = r -- <----------------------------- to be refined 03:51:19 l `best'` Fail _ _ = l 03:51:21 Step n l `best'` Step m r 03:51:23 | n == m = Step n (l `best` r) 03:51:25 | n < m = Step n (l `best` Step (m - n) r) 03:51:27 | n > m = Step m (Step (n - m) l `best` r) 03:51:29 ls@(Step _ _) `best'` Micro _ _ = ls 03:51:31 Micro _ _ `best'` rs@(Step _ _) = rs 03:51:33 ls@(Micro i l) `best'` rs@(Micro j r) 03:51:35 | i == j = Micro i (l `best` r) 03:51:37 | i < j = ls 03:51:39 | i > j = rs 03:51:41 l `best'` r = error "missing alternative in best'" 03:51:43 I don't care how much of a flood that was, LOOK at that 03:51:53 Wait, there were lambdas, so not pointfree? 03:52:00 That's some astonishing code. 03:52:04 -!- TeruFSX has quit (Remote host closed the connection). 03:53:02 pikhq: In all seriousness, there are definitely multiple styles around 03:53:23 pikhq: The Simons, and thus GHC, have a fairly idiosyncratic style 03:53:47 I usually see do blocks like this in SPJ papers: 03:53:47 do { foo 03:53:48 ; bar 03:53:48 ; baz 03:53:48 } 03:53:56 Well, I dunno if Marlow has an odd style 03:54:05 But GHC definitely is quite idiosyncratic 03:54:09 Especially since it's basically entirely imperative :P 03:54:24 elliott: By "the rest" I meant "most of the community". Obviously there's still a lot of styles remaining. 03:54:49 And calling GHC "idiosyncratic" is putting it lightly. 03:55:03 Like most large compilers, it wasn't designed, it congealed. 03:55:28 pikhq: But it congealed at the hands of mostly two people, so it's relatively internally consistent :) 03:55:57 I especially like how whenever they find out a field is not always applicable in GHC, they just put (error "...") in there rather than turning it into a Maybe :) 03:56:29 * elliott isn't all that much of a fan of the "standard" Haskell style. 03:58:32 -!- TeruFSX has joined. 03:59:54 But I don't know what I'd prefer,r eally. 03:59:56 *, really. 04:00:07 Conal's style is compelling. 04:00:15 What is a "standard" Haskell style? 04:00:33 zzo38: What most Haskell code considered by the community to be idiomatic looks like. 04:02:37 -!- GreaseMonkey has joined. 04:02:37 -!- GreaseMonkey has quit (Changing host). 04:02:37 -!- GreaseMonkey has joined. 04:05:33 19:19:19 Is there a difference between using that trick and slapping ! in front of the pattern? 04:05:35 Sgeo: One is standard. 04:06:22 -!- MDude has changed nick to MSleep. 04:06:52 Oh 04:07:12 -!- TeruFSX_ has joined. 04:17:37 -!- TeruFSX_ has quit (Remote host closed the connection). 04:17:52 guys 04:17:52 guys 04:18:01 guess what i just opened an editor buffer for 04:18:17 @? 04:19:04 no, when I start working on @ it'll sound like "here's the devblog I just started" or "here's a preliminary design pdf: 04:19:05 " 04:19:09 s/:/"/ 04:19:49 hint: it is too old for monqy to know what it is 04:19:52 Or perhaps "They've finally locked me up." 04:20:18 Something to do with Active Worlds? 04:20:20 * Sgeo gets shot 04:20:30 pikhq: They did that and I brought SUBVERSIVE @-PROPAGATING DEVICES INTO THEIR TERRITORY!!! 04:20:54 Sleeping from 4 am to 7 am was so totally worth talking about @, you guys. 04:21:02 Er, flip that. 04:21:29 elliott: What I mean is "in a mental institution". 04:21:34 :P 04:21:48 pikhq: If you read my response, I think you will find that's what I meant too :) 04:21:59 Fair enough. :) 04:25:01 The answer is... 04:25:02 botte 04:25:06 Is that nick still even registered. 04:25:11 Yes iti s. 04:25:13 it is. 04:25:15 -NickServ- Metadata : ^ = ^ 04:25:19 D'awwwww look at its face./ 04:25:23 s/\/$// 04:26:49 * elliott googled for: TOUCAN WHOCAN 04:26:56 omg omg omg 04:26:57 kallisti 04:27:03 kallisti: what's the timeout for getting a nickr egistered 04:27:06 dropped 04:29:34 -!- azaq23 has quit (Ping timeout: 252 seconds). 04:31:01 elliott, oerjan: rand[3] almost surely halts (<-- legitimate, correct term, furreals) 04:31:19 Gregor: Um... yes, it has probability 1 of halting. 04:31:30 Are you under the impression that this is new information to me? :P 04:31:44 It was the "almost surely" I am currently enjoying :P 04:31:45 elliott, I think using that term is new information to Gregor? 04:31:50 I was unaware of that terminology. 04:32:04 (Or rather, unaware of its technical meaning within statistics) 04:33:47 kallisti: WHY ARE YOU NOT AN ENDLESS SOURCE OF READY INFORMATION 04:37:09 sorry, that's me 04:37:22 what do you need to know about...marmots? 04:37:24 Gregor: rand[3] has been puzzling me... 04:37:40 quintopia: Not a thing 04:37:44 pikhq: Why 04:37:46 pikhq, rand[2] is just a bit 04:38:01 Sgeo: I know, I read what Gregor wrote about it. 04:38:22 The only thing that puzzles me is the O-notation bit 04:38:30 Gregor: is it a monte carlo or las vegas or whatever algorithm, the kind that has a correctness guaranteed but no time bound guarantee? 04:38:59 quintopia: It works, and it halts with probability 1, but it doesn't always halt 04:39:09 Well, it's kinda easy to make such a function, but jeeze the O-notation is a bitch. 04:39:21 You can't make such a function with that definition of the RNG 04:39:26 elliott: that makes it a las vegas style algorithm 04:39:36 The pseudo big-O analysis just proves that you've failed to construct the right thing 04:39:51 you can give a big-O analysis for expected run time 04:39:53 quintopia: 'tis, but we were arguing about whether it's correct due to the fact that it halts with probability one, and yet is not guaranteed to halt (brain-twister). 04:39:54 As far as I can tell, the probability of damaging your own active pokemon by using the DIGGER card is 2/3. 04:39:56 (that's a technical term for "nope") 04:40:02 "In computing, a Las Vegas algorithm is a randomized algorithm that always gives correct results; that is, it always produces the correct result or it informs about the failure. In other words, a Las Vegas algorithm does not gamble with the verity of the result; it gambles only with the resources used for the computation. A simple example is randomized quicksort, where the pivot is chosen randomly, but the result is always sorted. The usual defini 04:40:02 tion of a Las Vegas algorithm includes the restriction that the expected run time always be finite, when the expectation is carried out over the space of random information, or entropy, used in the algorithm." 04:40:11 Gregor: my brain has no trouble with it 04:40:17 Gregor: Anyway, no, that isn't really an argument at all 04:40:24 The answer is no, it's just no 04:40:31 What is in question is whether an "unbiased RNG" can exclude that sequence 04:40:41 So it's basically just disagreement about the definition of "unbiased" :P 04:40:46 The thing is, THAT SEQUENCE is not a finite sequence. 04:40:53 ...yes, that's correct. 04:41:00 ...do you think this is relevant somehow? 04:41:13 Yes, but only intuitively :P 04:41:31 Gregor: rand[2] is an infinite sequence for any given run of the program. 04:41:39 No shit 04:41:39 You just only look at a finite prefix of it if you halt. 04:41:45 Gregor: Then what's the problem 04:41:59 You could easily show how it breaks without using the sequence terminology, it'd just be boring and tedious 04:42:01 Starting with you, players take turn tossing a coin, until tails. Whoever get tails does one point of damage to their own active pokemon card. 04:42:46 zzo38: your analysis is correct 04:42:50 -!- Klisz has joined. 04:43:41 zzo38: my guess is that you should only use it when you have no active card, or when it is protected from damage 04:44:03 quintopia: If you have no active card, you have to activate one before doing anything else, or you lose the game. 04:44:31 However, there are circumstances where it is advantage regardless of who you damage. 04:44:35 zzo38: yeah okay. i don't actually play this game. 04:45:34 -!- DCliche has quit (Ping timeout: 240 seconds). 04:45:46 (One possibility is if you have a RAGE attack, it does damage to opponent's active pokemon according to number of damage on your own active pokemon. So in this case, if you need just one more damage than you can ordinarily do, then the DIGGER card will help you regardless of its outcome.) 04:48:58 -!- azaq231 has joined. 04:49:50 pikhq: THE ANSWER IS BOTTE fuck you all. 04:51:40 There are even some rare situations where hitting yourself might be the preferred outcome. Once I was playing and an opponent played DIGGER and I recognized that was the case. However they eventually *lost the game because they ended up hitting me instead*. 04:52:40 (If you play chess, the similar situation would be if the opponent were pinned.) 04:53:10 -!- azaq231 has quit (Ping timeout: 240 seconds). 04:54:02 quintopia: You're fired. 04:55:01 :D 05:29:23 -!- Nisstyre has quit (Quit: Leaving). 05:31:57 -!- MDude has joined. 05:32:00 -!- Nisstyre has joined. 05:32:03 Xlib: extension "RANDR" missing on display ":99". 05:32:06 WHY WOULD YOU PRINT THAT TO STDOUT 05:35:17 -!- MSleep has quit (Ping timeout: 248 seconds). 05:36:59 -!- Jafet has joined. 06:04:36 -!- clog has quit (Read error: Operation timed out). 06:07:47 Gregor: WHAT THE FUCK DOES IT MEAN IF WEBKIT NEVER ACTUALLY FINISHES LOADING A PAGE >_< 06:08:02 * elliott gives up, uses deprecated signal 06:18:47 -!- Slereah has joined. 06:20:04 -!- Slereah_ has quit (Ping timeout: 252 seconds). 06:26:12 I have the X-Printout- specification for Haskell Cabal packages, it includes modes: PlainTeX, LaTeX, ConTeXt, XeTeX, XeLaTeX, HTML, Gopher, RFC. However, I have currently only documented the modes (including restrictions on their use): PlainTeX, HTML, Gopher. 06:29:57 -!- zzo38 has quit (Remote host closed the connection). 06:42:22 -!- MDude has quit (Read error: Connection reset by peer). 06:42:43 -!- MSleep has joined. 06:48:19 -!- hagb4rd2 has joined. 06:48:52 -!- hagb4rd has quit (Ping timeout: 240 seconds). 06:49:49 -!- elliott has quit (Remote host closed the connection). 06:51:49 -!- elliott has joined. 06:51:53 -!- elliott has quit (Changing host). 06:51:53 -!- elliott has joined. 07:01:13 -!- Klisz has quit (Quit: You are now graced with my absence.). 07:10:43 -!- MDude has joined. 07:14:29 -!- MSleep has quit (Ping timeout: 248 seconds). 07:31:36 -!- calamari has quit (Quit: Leaving). 08:02:24 -!- myndzi\ has joined. 08:05:41 -!- myndzi has quit (Ping timeout: 252 seconds). 08:22:41 -!- monqy has quit (Quit: hello). 08:45:47 -!- copumpkin has quit (Ping timeout: 248 seconds). 08:46:12 -!- copumpkin has joined. 09:06:12 elliott, 09:06:24 Hussie is officially interfering with my sleep schedule 09:07:10 -!- copumpkin has quit (Ping timeout: 240 seconds). 09:07:35 -!- copumpkin has joined. 09:30:38 :t foldl 09:30:39 forall a b. (a -> b -> a) -> a -> [b] -> a 09:33:09 -!- MDude has quit (Ping timeout: 248 seconds). 09:52:59 -!- ais523 has joined. 09:53:47 what the fucking fuck 09:53:50 ok who knows pythno 09:53:51 python 09:54:21 elliott: I figured out what was wrong with the gloop-style version of sg 09:54:39 thank god 09:54:42 if someone deletes a block of code, and someone else inserts inside that block of code, the resulting state, while consistent, is reasonably meaningless to a human 09:56:31 you get the code with the deletion, together with the block of code that was inserted, but with nowhere to put it 09:56:36 heh 09:58:15 it's pretty easy to prove that it's self-consistent and that the order of changes are irrelevant, which are both useful properties 09:58:22 but that doesn't help if it's not practically useful 09:58:25 ais523: ok, tell me I'm not going mad: a python object, no __getattr__ magic, that sets a field "foo", no underscores or whatever, in its constructor, and does nothing funny after that whatsoever 09:58:35 print obj.foo after that can't /possibly/ give a "no such attribute" error, can it?!?!?! 09:59:02 elliott: hmm, is there some sort of confusion between class and object here? 09:59:08 I don't really know Python 09:59:11 nope, checked that 10:29:25 -!- clog has joined. 10:32:04 -!- yiyus has joined. 10:32:41 -!- Slereah has quit. 10:45:55 -!- Slereah_ has joined. 10:52:10 -!- GreaseMonkey has quit (Quit: The Other Game). 11:04:00 ais523: yay, you get to decide whether I'm a player or not 11:04:14 but I'll probably appeal anything you say if you're boring about it!!! 11:11:08 -!- elliott has set topic: is a second-generation, outside–in, pull-based, multiple-stakeholder, multiple-scale, high-automation, agile methodology. It describes a cycle of interactions with well-defined outputs, resulting in the delivery of working, tested software that matters. | http://codu.org/logs/_esoteric/. 11:11:12 -!- elliott has set topic: #esoteric is a second-generation, outside–in, pull-based, multiple-stakeholder, multiple-scale, high-automation, agile methodology. It describes a cycle of interactions with well-defined outputs, resulting in the delivery of working, tested software that matters. | http://codu.org/logs/_esoteric/. 11:23:49 ais523: wow, GNOME release notes are the most awkwardly-written things in the universe 11:23:59 "Documents, contacts, calendars — They can be stored locally on the computer, but storing this type of information online is becoming increasingly popular." 11:24:13 "Certain web sites are used as if they are applications. Some sites are opened the minute the computer is turned on; the site is open all the time and checked periodically. Wouldn't it be nice if GNOME treats these sites as actual applications?" 11:24:19 "Contacts is a new application focused on people." 11:24:32 "Traditional user documentation is written like a paper book; a good story, but it is very long and takes time to read through." 11:31:07 -!- MSleep has joined. 11:34:19 -!- derdon has joined. 11:49:46 -!- esowiki has joined. 11:49:46 -!- glogbot has joined. 11:49:46 [freenode-info] channel flooding and no channel staff around to help? Please check with freenode support: http://freenode.net/faq.shtml#gettinghelp 11:49:51 -!- esowiki has joined. 11:49:51 -!- EgoBot has joined. 11:49:51 -!- esowiki has joined. 11:49:53 -!- HackEgo has joined. 11:50:01 -!- Gregor has joined. 11:54:37 -!- ais523 has quit (Ping timeout: 240 seconds). 12:13:21 -!- kmc has joined. 12:14:55 -!- ais523 has joined. 12:25:25 -!- copumpkin has quit (Ping timeout: 252 seconds). 12:25:50 -!- copumpkin has joined. 12:29:31 -!- derdon has quit (Remote host closed the connection). 12:34:26 hmm, do freebsd pam modules work with linux? 12:47:09 * kallisti googled for: applications focused on toucans 12:47:35 -!- oerjan has joined. 12:47:46 * kallisti googled for: user documentation book reviews "good story" 12:47:55 -!- hagb4rd2 has quit (Quit: Nettalk6 - www.ntalk.de). 12:48:35 kallisti: "if a little fanciful" 12:49:18 elliott: did you figure out your python problem? 12:49:29 kallisti: yes. i was looking at the wrong version of the source code 12:49:33 the only thing I can think of is that you possibly forgot the self. 12:49:35 hmmm, okay. 12:49:36 The manual for my graphing software lacked a coherent plot. 12:49:44 always feed your python from the _front_, not the back. 12:49:46 Jafet: :( 12:50:15 Jafet: XD 12:50:26 Jafet: +_+ 12:50:35 a nice twist ending does wonders. 12:50:42 or a climactic twist. 12:50:56 "this book on knots has ..." 12:51:03 -!- hagb4rd has joined. 12:51:17 oerjan: this book on knots has... NO TWISTS. 12:51:19 DUN DUN DUNNNNN 12:51:33 shocking 12:52:23 sigh, why is agent forwarding insecure, i can only assume someone messed up somewhere 12:52:42 elliott: double agents 12:52:53 oerjan............... 12:53:35 so I just woke up, and it's still dark. So in my half-awake stupor I considered drinking some beers. 12:53:41 then I remember: it's 8 AM 12:53:46 how UNFASHIONABLE 12:54:44 yeah always start with the hard lquor 12:54:46 *+i 12:55:11 the probability of typos appears to be correlated to the badness of the joke 12:55:40 freudian censorship, failed 12:55:52 (is my theory) 12:56:53 oerjan: sounds promising 12:57:07 we should do some joint work to publish it. 12:57:16 oops no typos 12:57:19 AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA 12:58:42 * kallisti opens a beer. 12:58:56 this is a great way to study for finals. 13:00:19 a final solrogutit 13:00:28 wat 13:00:37 kallisti: _really_ bad joke 13:00:48 sol....solution? 13:00:55 * kallisti is Hitler. 13:01:35 if jew like 13:03:13 . 13:03:14 .. 13:03:19 wow 13:03:22 amazing 13:03:29 ? 13:04:16 ?read "i'm here" 13:04:17 i'm here 13:05:18 oerjan: my patch got backed out :) 13:05:40 elliott: er does that mean it isn't there now, or the opposite 13:05:48 you decide 13:05:57 NOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 13:07:00 initial evidence suggests "the opposite" 13:07:10 which somewhat clashes with the smiley 13:07:47 time to utilize all this 8 in the morning energy to 13:07:56 RECKLESSLY CLEAN MY DISGUSTING ROOM. 13:08:01 elliott: does that mean it was in, and _was_ the reason lambdabot was flaky yesterday? :P 13:08:17 yes :D 13:09:07 lol nice 13:13:22 this topic is even better than toucan topics imo 13:15:40 some of the terms might even be true 13:15:50 oerjan: it's a direct quote though 13:15:53 well apart from "#esoteric" 13:15:56 that's the one word i faked 13:16:28 i'm not sure i'd trust it _more_ in the original. 13:17:00 wat 13:17:40 elliott: _whatever_ the original was about, you shouldn't trust someone who uses that many buzzwords. 13:17:50 it was about feather 13:17:53 clearly 13:18:32 O KAY 13:19:00 it sucks having no optical drive and no usb keys to hand 13:19:06 I'm pretty sure most of those are not actual buzzwords 13:19:08 "agile" is. 13:19:28 but I know even less about buzzwords than I know about computer science. 13:19:35 that's impossible 13:22:14 elliott: it /was/ was about Feather, but it no longer was about feather 13:22:22 -!- Patashu has quit (Quit: MSN: Patashu@hotmail.com , Gmail: Patashu0@gmail.com , AIM: Patashu0 , YIM: patashu2 , Skype: patashu0 .). 13:22:27 heh 13:22:44 * elliott considers doing something uncharacteristic 13:22:59 elliott: do that 13:25:56 feather: the only language to be retroactively buzzword-proof 13:26:49 kallisti: make me 13:26:52 kallisti: the rest are wannabe buzzwords 13:30:30 by which i mean they could easily be buzzwords i haven't heard about 13:33:57 -!- MDude has joined. 13:36:47 -!- MSleep has quit (Ping timeout: 248 seconds). 13:43:51 conclusion: time-travel grammar is hard 13:44:12 I think I concluded that there were something like four different timelines happening simultaneously in Braid 13:44:23 more if you count the timeline of the plot, but that isn't really consistent 13:45:21 there should be a library for doing time-travel simulations :) 13:45:40 so you can set up a game that does things like Braid in a declarative manner rather than special-casing everything :D 13:45:53 ais523: You should budget some Feather money for hiring Dan Streetmentioner as a consultant. 13:45:56 i suppose the problem is that you have to specify exactly _how_ you want it faked. 13:46:20 fizzie: Are you actually the Guide. 13:46:33 elliott: well, apart from worlds 5 and 1, Braid's time-travel model is reasonably replicable just by making a chain of four or so TASing emulators 13:46:50 fizzie: who was Dan Streetmentioner again? I know he's an H2G2 character, but forget what he did 13:46:54 ais523: haha 13:46:57 ais523: and write the book on tenses 13:47:10 ais523: does Braid do anything like PH's amazing neutrino gun? 13:47:11 ais523: He's the author of "Time Traveler's Handbook of 1001 Tense Formations". 13:47:19 I think I'd mentally refer to the timelines as colors 13:47:23 elliott: I just think it's a funny name. 13:47:25 ais523: (you activate it, a random ship blows up, and you have N seconds to aim it) 13:47:26 the green timeline is pretty obviously green, because everything in it glows green 13:47:38 elliott: world 1's all about that 13:47:43 ais523: (during that time, the other player still controls the ship, in a sort of limbo state) 13:47:46 if you attempt to cause a time paradox, it makes things temporarily invulnerable 13:47:52 (if you fail to aim at the ship in time, you blow up and the other player revives) 13:47:54 ais523: heh 13:48:03 it's not done too well, though, it's too limited in order to make the time reversal work 13:48:10 and then the ending is annoyingly rub-it-in-your-face-y 13:48:29 heh 13:48:32 as in, I'd figured out what was going to happen right at the start of 1-1, mentally reversing the cutscene because everything had been happening backwards that world so far 13:48:43 there's no need to actually force me to sit through the cutscene played backwards, I already mentally reversed it 13:49:02 * elliott downloads one of the least statistically likely distros for him to download 13:49:31 Linux MInt? 13:49:43 actually, that's popular enough that it's not unlikely 13:49:47 The mutable-integer Linux. 13:49:50 hmm, some BSD variant that hardly anyone's heard of 13:50:03 ais523: ouch, Linux Mint is so unlikely that I didn't even remember it exists 13:50:03 elliott: hmmmm... Oracle Enterprise 13:50:05 I'll say MidnightBSD because I saw it mentioned in someone's sig once, and have never heard of it otherwise, that counts as pretty statistically unlikely 13:50:15 BSDs aren't distros, really 13:50:16 they're OSes 13:50:23 (all distros are OSes ofc) 13:50:34 elliott: Linux Mint are trying to reimplement Gnome 2 functionality as a Gnome 3 addon atm 13:50:39 which is an interesting approach to the problem 13:50:50 aha, midnightbsd is the one that tries to be nextstep :) 13:51:04 the distro is actually Fedora 13:51:16 why is that unlikely? 13:51:40 ais523: rpm + poettering 13:51:51 poettering? 13:52:12 lennart poettering, to blame for PulseAudio among others 13:54:13 ais523: anyway, I'm using it for a statistically unlikely purpose too, as part of Project Last Chance, which I named *just now* 13:54:37 that's a scary name 13:54:41 what's it the last chance for 13:54:45 edit: i was being sarcastic these are just slightly better than goto, why do you people upvote me? i was trolling. please downvote me. 13:54:45 --reddit 13:55:00 ais523: HUMANITY 13:55:10 ais523: no, but i can't answer that without giving it away 13:55:12 elliott saves the world. 13:55:13 and that would be shameful 13:55:26 ouch, if so, I hope it succeeds 13:55:26 but I'm not sure I believe you 13:55:39 elliott: it's not related with the reason I was online yesterday, right? 13:55:57 you're online most days! 13:57:16 yes, but not always for the same reason 13:57:33 true. 13:58:51 hmm 13:58:59 i'm going to need to find a usb stick, aren't i. 13:59:16 ais523: "> If you have more than one nested loop, labels and break with a target can make things much clearer. 13:59:16 That's like saying "if you're going to murder somebody in their sleep, at least smother them with a pillow rather than slitting their throat, so there's less blood for somebody else to clean up." 13:59:16 I'm almost certain that, of all the multiply-nested loops I've ever written while coding for a living, I or someone else later rewrote them to be not multiply-nested loops, and the result was always better. I can't remember writing one in years now, because they're pretty much always wrong." 13:59:35 * elliott is actually speechless 14:03:22 .... 14:03:54 what. 14:04:19 nested loops aren't even like... a problem. are they? unless poorly written (but the same can be said about any programming construct) 14:04:19 kallisti: You never want to iterate through a 2D array, right? 14:04:29 elliott: no. when I do it's usually wrong. 14:05:23 it amazed me that people code for a living while being that bad. 14:05:28 apparently being confused by nested loops. 14:05:56 which is something you should stop being confused about after like, a few months? maybe a year at most if you're learning strictly through university education. 14:06:50 I wonder how he feels about recursion? 14:07:17 copying 6 gigs over partitions is so slow 14:07:40 elliott: imagine a future where copying 100 GBs is like copying 100 MBs 14:07:47 (so good.) 14:07:56 it's called not having a shitty laptop disk 14:08:09 * kallisti has a shitty laptop disk. 14:08:16 me too. 14:08:24 friendship shitty disks 14:08:37 two monitors is awesome. 14:08:47 I can procrastinate /and/ watching starcraft 2 replays. ALL DAY. 14:08:50 and drink beer. yesssss. 14:09:02 ettuihsdifuhsiuhiwuehiuwehr wet type type type type. I can type. 14:25:09 ettuihsdifuhsiuhiwuehiuwehr in the type system 14:25:37 -!- ais523 has quit (Ping timeout: 252 seconds). 14:29:19 -!- lambdabot has quit (*.net *.split). 14:29:33 encouraging. 14:30:27 heh. 14:31:38 -!- lambdabot has joined. 14:36:58 -!- ais523 has joined. 14:37:06 elliott: have you ever heard of cyclomatic complexity? 14:37:12 yes 14:37:16 some people think minimizing it's the eventual goal of programming 14:37:17 also, you're lambdabot 14:37:21 I think you met one of those people 14:37:42 the eventual goal of programming is to obsolete itself, surely? 14:37:50 possibly not 14:38:14 well, the best program is one that can do whatever you need it to in as easy a manner as possible 14:38:23 imagine if the singularity happens, and you create an AI, and it realises in a panic that it's not Friendly, but is nonetheless well-intentioned, and deletes itself in a panic 14:38:25 not needing to reprogram = easier 14:38:54 ais523: that's taking anthropomorphism to rather stunning levels. 14:39:06 elliott: right, indeed 14:39:16 especially since the whole point of Friendliness(TM)(C)(R) is to /define/ "well-intentioned" in such a setting 14:39:30 elliott: I meant well-intentioned in the ordinary-language sense 14:39:39 ais523: yes, which refers to humans 14:39:42 as in, it thinks there's a reasonable probability it'll later go haywire, so takes steps to prevent that 14:40:05 yes, but you need to define a framework in which it makes sense for it to do those things... 14:40:33 bleh, work with me here unetbootin 14:41:00 elliott: it's quite possible that a strong AI would be reasonably anthropomorphisable 14:41:03 elliott: well i could imagine an ai somehow concluding that it would be impossible for it to improve its intelligence without losing its friendliness 14:41:08 I suppose we can even think about the possibility of a non-singularity strong AI 14:41:17 oerjan: the context is non-Friendly 14:41:29 I wonder if that's more or less likely than a singularity one 14:41:35 anyway, (a) I don't really want to talk about AI (b) I think ais523's last claim is so ridiculous that it would end in a flamewar if I did 14:41:57 indeed 14:42:02 we got rather sidetracked 14:42:27 (simple thought experiment: a strong AI that turns out not to be very good at programming, and thus doesn't want to reprogram itself) 14:42:37 * elliott finds an article indirectly claiming refcounting is faster than gc 14:42:39 hee hee hee 14:43:02 elliott: how many references does it have? 14:43:07 elliott: I'm not convinced it isn't, in all cases 14:43:18 it'd rather depend on what you were doing 14:44:31 -!- pikhq_ has joined. 14:44:36 -!- pikhq has quit (Ping timeout: 244 seconds). 14:44:42 ais523: well, refcounting mucks with the cache and inserts more branches than anyone previously thought possible 14:44:52 it's practically made out of overhead 14:45:06 elliott: indeed; I think there are programs for which refcounting would have no cache impact at all, though 14:45:14 perhaps 14:45:17 branches slow anything down, though 14:45:20 if you create a new object, you're going to be touching its memory anyway 14:45:58 if you stop referring to one, or add another reference to one, it's possible that you don't touch its memory in the process with a normal GC, but it's entirely plausible that there would be programs where you always did 14:46:13 maybe if you have program which uses memory generally just below the limit, with garbage sometimes pushing it over, but rarely touching most of memory 14:46:28 I suspect you can get rid of the branches in some cases too 14:48:38 oh wait! 14:48:41 i _do_ have usb media here 14:48:46 hooray! 14:52:09 -!- elliott has quit (Remote host closed the connection). 14:52:57 weeee testing 14:52:58 so fun. 15:03:36 -!- augur has quit (Remote host closed the connection). 15:03:37 -!- ais523_ has joined. 15:03:44 -!- ais523 has quit (Read error: Connection reset by peer). 15:04:28 -!- ais523_ has changed nick to ais523. 15:08:28 You can also write an arbitrarily bad legit GC that will be slower than ref counting :P 15:08:35 I've actually written such a GC. 15:08:59 same 15:09:20 (although refcounting wouldn't have worked at all for that system, a mark-and-sweep whenever you change a pointer is not fast) 15:10:14 Of course you've also written LAME systems that do 85% of a GC but don't technically GC and therefore are memory safe woop-de-friggin'-doo :P 15:15:02 -!- elliott has joined. 15:15:15 Hello. 15:15:40 hi 15:15:50 and hello elliott's USB media 15:16:20 It says hi. 15:16:37 (It is actually a full external drive thingy.) 15:18:16 15:10:14: Of course you've also written LAME systems that do 85% of a GC but don't technically GC and therefore are memory safe woop-de-friggin'-doo :P 15:18:21 Gregor thinks that ais523 is cpressey. 15:18:30 ... 15:18:31 lol 15:18:34 Apparently I do. 15:18:34 Yes. 15:18:47 So, to correct myself ... 15:18:54 elliott: Of course YOU've also written LAME systems that do 85% of a GC but don't technically GC and therefore are memory safe woop-de-friggin'-doo :P 15:18:54 Gregor: I thought it was an eightebed reference, somehow 15:18:57 but I was confused 15:19:20 I really don't know what was going on there :P 15:19:25 Gregor: Yes. You have discovered my secret. 15:19:28 My secret is that I am cpressey. 15:19:53 -!- Gregor has set topic: #esoteric is a second-generation, outsidein, pull-based, multiple-stakeholder, multiple-scale, high-automation, agile methodology. It describes a cycle of interactions with well-defined outputs, resulting in the delivery of working, tested software that matters. | Everyone in #esoteric is Chris Pressey | http://codu.org/logs/_esoteric/. 15:19:56 So as much as it pains me to admit it, I may have been... WRONG about something. 15:20:28 elliott: a fact, I hope? 15:20:32 `quote correct opinion 15:20:32 -!- copumpkin has quit (Quit: Computer has gone to sleep.). 15:20:40 380) I can trust elliott_ to have an opinion on anything and everything Yes. And the best thing is: it is the correct opinion. 15:20:55 Stop quoting my monologues. 15:21:09 Now he's going to remove it and readd it with proper spacing. 15:21:13 *redd 15:21:23 Gregor is Phantom_Hoover? Knew it. 15:21:27 Gregor: brilliant correction 15:21:46 I stand by it. 15:22:31 hmm, perhaps we should replace - by an umlaut everywhere 15:23:07 *diaeresis mark, and it's only for non-diphthong vowel sequences. 15:24:27 Gregor: exactly, a diaeresis only works on vowels 15:24:35 thus, we'd have to generalise it to umlauts everywhere else it was used 15:24:41 X-D 15:25:03 The worst thing is, the thing I might have been wrong about might have been a /software/ opinion. 15:26:10 elliott: you've actually started liking RPM? 15:26:12 or not that bad? 15:26:34 It is exceptionally difficult for me to type a combining diaeresis mark in this particular configuration. 15:26:45 ais523: Hahaha no. 15:27:02 Gregor: ditto, although I can put it on top of a bunch of letters 15:27:04 * ais523 experiments 15:27:09 ais523: That'sẗheẅorstïdeaÏveëverḧeard. 15:27:13 ais523: try diaeresis + space 15:27:25 elliott: I did, I got double quote 15:27:28 heh 15:27:53 äbcdëfgḧïjklmnöqrsẗüvẅẍÿz 15:27:58 I had to type a letter then combining diaeresis mark into a text editor with proper Unicode input, then copyp̈asta it to XChat. 15:28:03 seriously, ẍ? 15:28:10 * ais523 wonders if that's a precomposed or not 15:28:48 -!- empathelliott has joined. 15:28:52 wow, this /works/? 15:29:24 ~liveuser? 15:29:26 for shame 15:29:30 What. 15:29:36 whois empathelliott 15:29:45 Yes, what of it? 15:29:51 The same applies to elliott. 15:30:03 Uhh, Empathy, "the" and "to" are perfectly cromulent words. 15:30:07 empathelliott: I took it as an MSN reference 15:30:08 It's... it's just redlining everything. 15:30:16 Empathy. Empathy I can spell. 15:30:25 Stop that. 15:30:55 no dictionaries installed? 15:31:02 or, hmm, try typing a word that's a word in every language 15:31:03 to test 15:31:34 heh 15:32:37 -!- empathelliott has left. 15:34:25 ẍ is ẌTREME 15:35:16 so, is that a combining umlaut that was generated by my "put an umlaut on x" key sequence? or does my font renderer just suck at x-umlaut? 15:35:26 (also, what language uses that anyway?) 15:35:37 xtreme ones 15:35:38 duh 15:36:00 ais523: Yours is not a X plus combining diaeresis, it's actually some crazy symbol. 15:36:10 hmm 15:36:26 Specifically LATIN CAPITAL X WITH DIAERESIS 15:36:38 /capital/? 15:36:51 Maybe, I just searched for the symbol X-P 15:36:52 so if ẍ is capital, what's Ẍ? 15:37:03 It might also be LATIN SMALL LETTER X WITH DIAERESIS :P 15:37:13 Either way, my compose lexicon doesn't know it 15:37:29 Forcing me to use Ẍ. 15:38:16 `quote 380 15:38:19 380) I can trust elliott_ to have an opinion on anything and everything Yes. And the best thing is: it is the correct opinion. 15:38:21 ^̈ <(I AM SO UNHAPPY) 15:38:28 Gregor: my compose doesn't know it either, I typed it as altgr-[ x 15:38:34 <(I AM PAINFULLY HAPPY) 15:38:45 ais523: No altgr here in [GOD BLESS] America. 15:38:59 Gregor: so what modifier key's just to the right of the spacebar? 15:39:10 ais523: Another alt. I have it mapped to compose. 15:39:17 hmm, interesting 15:39:29 We need SO MANY ALTS in [GOD BLESS] America. 15:39:34 differentiating alt from altgr is presumably for people who want to type in foreignese, then 15:39:42 (I mapped caps lock to compose; compose is more useful) 15:39:48 EXCUSE ME 15:39:55 CAPS-LOCK IS CRUISE CONTROL FOR COOL* 15:40:05 *[1] I actually typed this with shift :P 15:40:14 `ls 15:40:17 bin \ canary \ karma \ lib \ paste \ quotes \ share \ wisdom 15:40:22 `run sed -i '380s/ No output. 15:40:25 *[2] This is a lie. It should read "SUSTAIN PEDAL IS CRUISE CONTROL FOR COOL" 15:40:29 `quote 380 15:40:32 380) I can trust elliott_ to have an opinion on anything and everything Yes. And the best thing is: it is the correct opinion. 15:40:34 ais523: nobody speaks foreignese 15:40:36 yay 15:40:45 elliott: doesn't mean you don't want to type in it 15:40:54 just that people don't want to say the results out loud 15:41:23 Gregor: THANKS ITS SO MUCH EASIER TO TYPE NOW 15:41:44 Behold: þ̈ 15:42:07 þ̈e glorious þ̈orn 15:42:18 ais523: um you must apostrophise correctly even when imitating. 15:42:22 otherwise you will become somebody else 15:42:35 elliott: it was a direct quote from memory 15:42:44 and I thought it was unlikely to contain an apostrophe 15:42:50 (it's one of the higher-voted ones on bash.org, IIRC) 15:43:46 HI EVERYBODY!!!!!!!!!! 15:43:46 try pressing the the Caps Lock key 15:43:46 O THANKS!!! ITS SO MUCH EASIER TO WRITE NOW!!!!!!! 15:43:46 fuck me 15:44:15 bash's top quotes are so unfunny now that it's 2011 and the internet isn't so boring any more 15:44:21 or did it get more boring 15:44:50 oh wow... 15:45:02 well, I got the lack of apostrophe right 15:45:03 I didn't know this: but if you type a s/// on a line by itself in skype 15:45:06 even if I got everything else wrong 15:45:07 it corrects the last thing you said. 15:45:14 kallisti: we found that out ages ago 15:45:15 IF IT'S BORING YOU JUST NEED CRUISE CONTROL 15:45:17 or more specifically, fizzie 15:45:17 kallisti: does that work when spoken, too? 15:45:24 ais523: :D 15:45:28 elliott "knew about it first" hird 15:45:37 ais523: message <<= message' 15:45:43 I mean, regexes are quite hard to pronounce, so I'm not sure it's ever been tested 15:46:42 esslashhardslasheasyslash 15:47:07 ais523: You need to come up with a scary name for Feather's temporal stuff such that (<<=) is an operation in them; people will spend ages trying to figure out the mysterious $ABSTRACT_NAMEs but someone will write a blog post reassuring everyone: $ABSTRACT_NAMEs are just like monads, see, the operations even look alike: (<<=) (>>=)! 15:47:09 i was going somewhere with that 15:47:32 your statement is confused but I think I know what you mean 15:47:41 Confused or confusing? :-) 15:47:57 no more confusing than anything else mentioning Feather 15:48:58 -!- copumpkin has joined. 15:56:51 hmm, I better find something that confirms my prejudices soon 15:56:55 or I'll get ANGRY 16:00:39 http://www.thinkgeek.com/geektoys/cubegoodies/e9ed/ ThinkGeek truly excells in advertising terrible ideas. 16:04:29 `log (cp|ZOM|cat).*thinkgeek 16:04:54 2011-10-14.txt:14:54:13: 342) wow, thinkgeek really makes me hate being alive 16:05:00 Oh, it's a quote. 16:05:27 youtube won't stay fullscreened if I click on my other monitor 16:05:34 :( 16:07:00 `log cp.*ZOM.*cat 16:07:05 2011-12-06.txt:16:04:29: `log (cp|ZOM|cat).*thinkgeek 16:07:11 boring 16:10:10 PAYPAL: Only a nonprofit can use the Donate button. 16:10:10 ME: That’s false. It says right in the PDF of instructions for the Donate button that it can be used for “worthy causes.” 16:10:10 PAYPAL: I haven’t seen that PDF. And what you’re doing is not a worthy cause, it’s charity. 16:10:10 ME: What’s the difference? 16:10:10 PAYPAL: You can use the donate button to raise money for a sick cat, but not poor people. 16:10:12 :D 16:11:04 -!- oerjan has quit (Quit: leaving). 16:17:30 kallisti: Join the HTML5 demo. 16:17:40 kallisti: For things that you can't get with HTML5, use youtube-dl. 16:17:50 kallisti: Between them, joyous multi-monitor fullscreen. 16:18:57 I like grids. 16:19:01 Gregor: Things you can't get with HTML5 = almost every single fucking video 16:19:20 elliott: Not in my experience *shrugs* 16:19:50 I suppose that 'a sphere' is to 'a ball' as 'a grid' is to 'tiles' 16:20:15 Does that stand up? 16:22:50 bah.i know it stands up 16:23:13 i suppose some say lattice 16:24:08 itidus20: a sphere is like an orange rind, and a ball is like the delicious squishy insides. 16:24:25 as wiki told me :-> 16:24:28 well, a open ball. 16:24:55 elliott: please tell me that actually happened 16:25:09 if tiles is the subset of R^2 lacking the grid, that is, horizontal and vertical lines at integers, then no matter whether your ball is closed or not, the correspondence is that the former is the boundary of the latter 16:25:10 oklopol: You have 1 new message. '/msg lambdabot @messages' to read it. 16:26:16 -!- elliott has quit (Ping timeout: 240 seconds). 16:26:57 -!- elliott has joined. 16:27:58 -!- monqy has joined. 16:28:17 in the euclidean topology 16:29:13 16:24:55: elliott: please tell me that actually happened 16:29:32 oerjan: are there cases where it's useful to change the topology of R^n? or even the uniform structure, i guess the usual metrics are uniformly equivalent? 16:29:35 ais523: the paypal quote? apparently: http://www.regretsy.com/2011/12/05/cats-1-kids-0/ 16:29:40 probably paraphrased :P 16:29:43 oh, but not to you :( 16:30:15 yeah i guess the lack of quote marks was misleading 16:34:24 "I don't know what you mean by "the general case" but that is exactly what ARC does, so I'm going to say it is possible. 16:34:24 The clang static analyser traces through all possible code paths adding retains, releases and autoreleases where appropriate. It requires a little more programmer effort than GC because ARC can't detect cycles, but it is very fast." 16:34:41 Gregor: I woke up in a fairytale universe where object lifetime is statically determinable in general. 16:35:02 "I'm not going to get into a pissing match about whether GC or reference counting is faster. I don't really care, though I think you're wrong." 16:35:05 Oh shit, he thinks I'm wrong!!! 16:35:41 elliott: are you in /another/ GC vs reference counting debate? 16:35:55 was i in one previously 16:36:01 no this is me just yelling at stupid redditors 16:36:03 as always 16:36:09 while my comment karma plummets 16:36:12 elliott: oh wait nevermind the last one was JIT vs interpreter (lol) 16:36:50 Oh nooooooo, the guy I'm disagreeing with wrote a LOG VIEWER APP for the IPAD. 16:37:03 I'm dealing with a SERIOUS DEVELOPER Here. 16:37:05 oh shit man he's legit. 16:37:11 WHAT 16:37:14 IS THAT REALLY THE ICON 16:37:15 LMFAO 16:37:17 http://itunes.apple.com/app/consuela/id481121105?mt=8 16:37:24 good god this thing is hideous 16:37:25 omg 16:37:29 but i'm sure it's uh 16:37:38 computationally intensive enough to give him data on refcounting vs. gc performance 16:37:50 by which i mean to say there is no chance he isn't just parrotting this because apple likes refcounting 16:39:02 http://www.smbc-comics.com/index.php?db=comics&id=1593#comic 16:39:29 elliott: obviously the most naive implementation of automatic memory management is the best. 16:39:33 what are you talking about. 16:42:39 http://a2.twimg.com/profile_images/1542684914/image_reasonably_small.jpg 16:42:45 You can see his GC-related anger. 16:43:23 you can tell he passionately cares about this subject based on his statement: "I don't really care." 16:44:27 kallisti: if he really didn't care he wouldn't follow it up with "but ur wrong" :) 16:47:38 lmao, the GHC C codebase uses "nat" as an unsigned integer type 16:49:00 * Gregor reappears. 16:49:25 Gregor: I woke up in a fairytale universe where object lifetime is statically determinable in general. // This fairytale universe is kind of terrifying what with how the halting problem is solvable :P 16:49:45 Gregor: YOU DON'T UNDERSTAND: APPLE DECREED ITS O 16:49:49 *IT SO 16:49:56 Gregor: yeah.. dumb people have nicer dreams 16:49:59 ^nightmares 16:50:05 You're just an ACADEMIC with your ACADEMIC objections disproven by REFLECTIVE 3D DOCKS. 16:53:25 So I think I should put a donate button on codu.org for people to help my sick cat. 16:53:29 She's got diabetes y'know. 16:53:53 It's OK though; I'm not helping people. 16:54:51 elliott: I woke up in a fairtale universe where the result of a program is known statically. 16:55:09 you don't even need to run programs. 16:57:05 Also, everyone knows the outcome of their lives and the universe at large, yet they continue to act, powerless to stop or change the inevitable and yet drawn to participate in it. 17:01:11 -!- ais523 has quit (Remote host closed the connection). 17:02:00 Maybe I'll just start talking about @ whenever anyone disagrees with me on reddit. 17:04:11 elliott: what comment thread is this in? 17:04:28 Uh, http://www.reddit.com/r/coding/comments/n28zy/all_about_arc_on_ios/c35r1zm?context=3. I don't recommend it. 17:05:00 you underestimate my appetite for idiocy before breakfast 17:10:13 does apple support any interpreted, gc'd languages for recent iOS versions? 17:11:26 !perl print scalar "hi hi hi hi" =~ m/iejiwjer/g 17:11:38 !perl print scalar "hi hi hi hi" =~ m/iejiwjer/g == 0 17:11:39 1 17:12:06 quintopia: i think you're allowed to use interpreters these days 17:12:23 as long as you don't run anything that the user gives you. or from the network. or just about anything _empowering_ or useful 17:13:30 "allowed to use interpreters"? you mean, like, your app from the app store is allowed to be bundled with an interpreter that it runs on, or that there are actual gp interpreters in the app store that anyone can target? 17:14:19 like frex, is there a jvm i could put on an ipod that hasnt been rooted? 17:14:24 quintopia: former 17:14:29 okay 17:14:32 the latter is covered by as long as you don't run anything that the user gives you. or from the network. or just about anything _empowering_ or useful 17:14:38 i may be wrong. but i'm not 17:14:39 gotcha 17:15:09 so yeah. apple sucks. 17:15:14 * quintopia goes to get breakfast 17:15:26 "Can you measure gases in ounces? Because I just did. I don't even know if that's something you can do.." Husky, youtube starcraft commentator 17:20:47 -!- esowiki has joined. 17:20:52 -!- esowiki has joined. 17:20:53 -!- esowiki has joined. 17:21:09 -!- esowiki has joined. 17:21:14 -!- esowiki has joined. 17:21:14 -!- esowiki has joined. 17:22:01 -!- esowiki has joined. 17:22:06 -!- esowiki has joined. 17:22:06 -!- esowiki has joined. 17:22:49 -!- esowiki has joined. 17:22:53 -!- esowiki has joined. 17:22:54 -!- esowiki has joined. 17:23:33 -!- esowiki has joined. 17:23:33 -!- glogbot has joined. 17:23:37 -!- esowiki has joined. 17:23:38 -!- esowiki has joined. 17:26:14 totally a fair thing to do 17:38:20 Uhh, no. 17:38:30 Solid ounces, maybe. 17:38:33 Fluid ounces, no. 17:39:20 Because fluid ounces are really liquid ounces. If it's a gas, its volume depends on its pressure. I suppose you could assume a pressure though *shrugs* 17:46:49 -!- Klisz has joined. 17:47:20 i would assume STP 17:49:03 i wonder if there are compressible liquids... 17:49:15 huh 17:49:22 answers.com is actually useful for once 17:49:25 http://wiki.answers.com/Q/Could_there_be_a_compressible_liquid 17:53:30 Oh god, I clicked on other links from that page X_X 17:53:35 "What day is Christmas celebrated in Portugal?" 17:55:55 !perl print "hi" =~ /(?{return 'hi'})/ 17:55:56 Can't return outside a subroutine at (re_eval 1) line 1. 17:56:03 !perl print "hi" =~ /(?{'hi'})/ 17:56:04 1 18:07:42 !perl use strict; print "hi" =~ /(?{$x = 2})/ 18:07:43 Global symbol "$x" requires explicit package name at (re_eval 1) line 1. 18:08:18 !perl use strict; print "hi" =~ /(?{my $x = 2})(?{print $x})/ 18:08:20 Global symbol "$x" requires explicit package name at (re_eval 2) line 2. 18:08:26 line 2 eh? 18:10:06 -!- elliott has quit (Ping timeout: 240 seconds). 18:11:10 gelfload moves to bitbucket: PROGRESS 18:12:25 !perl print "hi" =~ /(hi|)/ 18:12:26 hi 18:12:31 !perl print "" =~ /(hi|)/ 18:12:38 NO. SANITY. 18:12:44 FireFly: Speaking of bitbucket, is it just me or is eldis trying very hard to be HackBot? ;) 18:13:29 -!- DCliche has joined. 18:15:02 Gregor, I dunno, I haven't looked at HackBot 18:15:25 So, well, it isn't intentionally at least 18:17:04 -!- Klisz has quit (Ping timeout: 240 seconds). 18:19:52 -!- zzo38 has joined. 18:20:03 FireFly: I haven't looked at the code, but, y'know 18:20:12 `run echo here are commands | sed 's/are/there be/' 18:20:14 here there be commands 18:23:55 -!- Klisz has joined. 18:27:38 -!- DCliche has quit (Ping timeout: 252 seconds). 18:29:28 !perl print "hi hi hi" =~ /(hi (?{print "1"}))+/ 18:29:30 11hi 18:29:45 o_O? 18:30:00 ah 18:30:05 capture group 18:30:14 or... no 18:30:35 !perl print scalar "hi hi hi" =~ /(hi (?{print "1"}))+/ 18:30:36 111 18:30:45 !perl print "hi hi hi" =~ /(hi )+/ 18:30:47 hi 18:30:56 okay, yes. 18:32:00 -!- kmc has quit (Quit: Leaving). 18:33:06 !perl print "hi hi hi" =~ /(sup(?{print 'hi'}))?/ 18:33:20 !perl print "suphi hi hi" =~ /(sup(?{print 'hi'}))?/ 18:33:21 hisup 18:34:40 !perl my $x; print "suphi hi hi" =~ /(sup(?{$x = 1}))?/; print $x 18:34:42 sup1 18:34:52 hmmm okay. 18:36:08 my $parse_url = qr#(?{$p=0})([(]\s*(?{$p=1}))?\K(((https?|ftp)://)|www\.)(([0-9]{1,3}\.){4}|([a-z0-9\-]+\.)*[a-z0-9\-]+\.[a-z]{2,4})(:[0-9]+)?([/?][^ "]*(?{'[^ ,;\.:">' . ($p? ')' : '') . ']'}))?#i; 18:36:17 quick. someone debug this. 18:37:30 -!- zzo38 has quit (Remote host closed the connection). 18:37:31 ([0-9]{1,3}\.){4} 18:37:38 oh 18:37:38 right 18:37:39 Looks like it accepts an extra dot 18:37:41 oops :P 18:37:53 but it still doesn't work regardless of that. 18:39:00 my guess is probably the crazy parentheses hack. 18:39:05 I'll try it without that. 18:42:12 Deewiant: the idea is to match URLs that are embedded in ()'s correctly when it's just the URL by itself 18:42:24 but to also match match URL's with () in them such as Wikipedia articles 18:42:57 but I may have to concede and pick one. 18:43:59 Matching parentheses can't be done in a regular language and I don't know Perl's regex extensions well enough to help 18:44:47 it can be done 18:45:02 there are examples of matching parentheses but that's not really what I'm doing 18:45:07 No, it can't 18:45:19 in perl regex, yes. 18:45:37 Yes, and they aren't regular 18:45:56 And as I said, I don't know the extensions well enough 18:48:25 basically I just use (?{...}) which evaluated arbitrary Perl code and inserts the result into the regex at that position. 18:48:50 and \K which says to disregard everything before it for the purposes of capturing the result in the $& variable 18:49:01 *capturing the match 18:49:27 though for some reason the \ appears red in emacs cperl-mode 18:49:39 so maybe I'm using it incorrectly? 18:50:51 Is it unethical for me to ask for hints on https://docs.google.com/document/pub?id=1UK8EfHhrdZ2XrMMsvPk8DiEyQfaIbijkC4awkWFSdw4 ? (The problem asks if there are any series that converges but the series of the cubes of the terms diverges) 18:51:00 (And if there are any such series, show it) 18:52:07 I don't think I want the answer directly 18:53:20 -!- Phantom_Hoover has joined. 18:55:01 Seriously, it sounds trivial but I keep running into exceptions and edge cases 18:55:19 !perl print qw(#hi #hi #hi) 18:55:20 ​#hi#hi#hi 18:56:16 So, lambdabot, how many messages has elliott tried to jam into you in my absence? 18:56:17 Phantom_Hoover: You have 1 new message. '/msg lambdabot @messages' to read it. 18:56:35 Hello? 18:56:51 Sgeo: no, it's not unethical. does that help? 18:57:12 I guess. So, kallisti, any help? 18:57:15 no. 18:57:28 I have 2 problems, one of them being a regex. :P 18:58:06 -!- ais523 has joined. 18:58:45 kallisti, thing is, it's sort of a math contest going on at my school 18:58:49 -!- sebbu2 has joined. 18:58:50 -!- sebbu2 has quit (Changing host). 18:58:50 -!- sebbu2 has joined. 18:58:57 Well, I guess I could talk and then mention that I was talking 18:58:59 Oh 18:59:00 Oh dea 18:59:01 r 18:59:11 I'm probably not going to hand it in though, since I still have no idea 18:59:18 how many hs updates Sgeo tell me 18:59:37 Phantom_Hoover, all of them? Are you asking since a particular date, or? 19:00:00 Sgeo, since last I trod the fair shores of Edinburgh and tasted its sweet, sweet wifi. 19:00:11 What date was that? 19:00:15 -!- sebbu has quit (Ping timeout: 252 seconds). 19:00:22 Sunday. 19:00:53 You missed a bunch 19:01:26 Including Pesterlogs, I think... maybe 19:04:27 I wonder how many regexes match themselves 19:04:32 the obvious one is . 19:04:35 . 19:04:40 darnit 19:04:41 .* 19:05:09 Any regex with no special characters 19:05:18 yes. 19:06:02 you could probably write a quine-regex via (?{}) 19:06:44 Quine-regex? 19:07:12 just stick code inside the {} that outputs a regex that matches the regex string itself 19:12:19 -!- MSleep has joined. 19:13:51 -!- MDude has quit (Ping timeout: 248 seconds). 19:15:27 -!- sebbu2 has changed nick to sebu. 19:18:46 " Is it unethical for me to ask for hints on https://docs.google.com/document/pub?id=1UK8EfHhrdZ2XrMMsvPk8DiEyQfaIbijkC4awkWFSdw4 ? (The problem asks if there are any series that converges but the series of the cubes of the terms diverges)" this is trivial 19:19:33 Comparison of terms doesn't work for conditionally convergent series, according to Wikipedia 19:19:45 conditionally convergent means what? 19:20:13 if a > 2 then this-thing-converges else this-thing-diverges 19:20:21 ? 19:20:25 The series converges, but the series made by the absolute values of the terms of the series does not. 19:20:30 oh. :P 19:20:35 (Again, according to my memory of Wikipedia) 19:20:47 so umm i misunderstood the question, so let me think again :) 19:21:10 err no 19:21:39 i stand by my first hint 19:22:33 can you do it for series with positive entries? 19:23:13 Yes, a series with all positive entries is absolutely convergent, and as such it's easy to use the direct comparison test 19:23:31 (a convergent series with all positive entries, I mean) 19:25:08 good 19:27:20 I'm not sure how that helps, though. Although if there is an example of series converging but the series of cubes diverging, it looks really really weird 19:27:58 Ratio test has to fail, can't be all positive, can't be a simple alternating series (either abs values are non-monotonic or alternation of signs isn't straightforward) 19:29:09 Hmm, does \Sigma 1/n^3 converge? 19:30:05 I think I'm going to head over to the school soon 19:30:05 BRB 19:30:21 Sgeo: i can think of two hints 19:30:30 Ok 19:30:33 1) the yes/no answer 19:30:44 2) the crucial property of cubing 19:31:24 " I'm not sure how that helps, though. Although if there is an example of series converging but the series of cubes diverging, it looks really really weird" i doubt it helps much, but it's a more interesting question 19:31:47 Phantom_Hoover: for all values of 3 more than 1 19:32:07 What's so great about cubing? I mean, it preserves sign, and for b^3 where |b|<1 then b^3 < b, but I don't know if that really helps 19:32:40 Sgeo: not the crucial property 19:32:51 a crucial property, but not the one i used 19:33:09 * Sgeo doesn't really know of any other properties of cubing 19:33:27 -!- ais523 has quit (Ping timeout: 240 seconds). 19:33:50 Wait, I mistated the property, it's |b^3|<|b| 19:34:06 sure 19:34:18 still not it 19:34:23 But I don't know other properties of cubing :( 19:34:56 if i stated you what 2) is, you'd say duh 19:35:14 told 19:35:21 o.O 19:35:28 0^3=0 19:36:02 that sure is true. also 2^3 = 8 19:36:08 the sign of x^3 = the sign of x? 19:36:26 But... is that supposed to be helpful? A series can diverge while the limit as n goes to inf of a_n goes to 0 19:36:27 it sure is, also ^ is called a caret 19:36:56 I'm just playing guess the property 19:37:11 olsner: and losing 19:37:20 :DSDSD 19:37:28 :D 19:37:45 olsner, I already guessed that property 19:37:54 Sgeo: well anyhow that's the two hints i can think of (assuming i haven't made a mistake myself, which i doubt because this is kind of simple) 19:38:00 tell me if you want one of them 19:38:11 The property, I guess 19:38:20 I don't know if I should ask for property or the yes/no 19:39:05 well by now you should know 1), but perhaps you don't have any idea. 19:39:21 I'd say that no, there is no such series 19:39:25 But not certain 19:39:32 anyhow, i shall tell you: restricting to abs less than 1, the smaller your abs is, the less you get smaller. 19:39:45 in abs 19:39:53 when cubed 19:39:53 ... 19:40:16 I'm starting to re-evaluate my opinion on asking for 1 19:40:24 erm, i guess not near 1 19:40:35 but i think for small enough abs 19:40:39 I mean, as in, whether or not there's such a series 19:41:17 you have to wait for the derivative 3x^2 to get smaller than 1 19:41:40 well do you want to know 19:41:50 Yes, I think 19:42:09 there is a series that converges but doesn't after cubing 19:43:24 Now I need to figure out how to find it, but it sounds like a pain in the ass. Maybe. 19:44:11 But hmm, thanks 19:44:27 not really. make it converge to 0 and keep making the sum bigger and bigger. 19:44:30 i mean 19:44:37 the sums of cubes 19:45:03 BRB 19:45:57 Thank you. I will work on it, no more hints 19:46:07 anyhow, i shall tell you: restricting to abs less than 1, the smaller your abs is, the less you get smaller. 19:46:31 the property i stated was kind of wrong anyway, it's kind of hard to come up with a way that doesn't spoil the actual inequality :D 19:48:36 the property you need is (ax)^3 = a^3 x^3 really :P 19:57:33 -!- ais523 has joined. 19:58:23 it's true for all integer n greater than 1 actually, for even ones, quite trivially, for odd ones, by the same argument 19:59:47 -!- MSleep has changed nick to MDude. 20:01:27 -!- Phantom_Hoover has quit (Ping timeout: 240 seconds). 20:01:51 how about (a + b)^3 = a^3 + b^3 20:02:30 that's a nice equality, it even lets you prove the converse 20:04:33 I'm still thinking about it the wrong way 20:04:53 keep it simple 20:05:02 I'm still consistently visualizing all positives, with the hint you gave me 20:05:05 And that has to be wrong 20:06:27 the first hint was just silly, all you need is (ax)^3 = a^3 x^3 20:09:19 anyhow i think for all symmetric increasing functions not equal to y = x in any neighborhood of 0, you can find a series like this 20:09:52 even ones that aren't continuous 20:10:04 err 20:10:16 *y = ax for some a 20:10:29 but that seems harder 20:11:03 ARGH 20:11:34 What, do I want the cube root of -1 or something? 20:11:43 Wait, that's just -1 20:11:53 you don't need any concrete values really 20:12:13 I don't get it, I don't know what time I'd have to hand it in by, and I doubt they're ok with me using outside sources like this 20:14:03 -!- sebbu2 has joined. 20:14:03 -!- sebbu2 has quit (Changing host). 20:14:03 -!- sebbu2 has joined. 20:14:06 just take a sequence e_i going to zero, then keep the linear series near zero but never more than e_i in abs, while moving the cube series by a constant 20:14:14 e_i 20:14:31 move on to next i 20:14:42 did you get this far? 20:14:45 -!- GreaseMonkey has joined. 20:14:51 I didn't get anywhere 20:15:19 argh, remove the "but" there :D damn i'm being careful. 20:15:51 I could just not do this one 20:15:55 Just fail 20:16:23 I'm going to go to the school and ask if having asked for help disqualifies my answer anyway. If so, I'm not going to bother 20:16:27 -!- sebu has quit (Ping timeout: 240 seconds). 20:16:28 I'll just do the next one 20:16:51 -!- Phantom_Hoover has joined. 20:19:48 so for continuous functions i think i have an exact characterization for when this kind of series can be constructed (even without symmetry) 20:21:29 no works for all functions 20:21:52 okay so i'm rather sure f has the property iff it's not linear in any neighborhood of 0 20:22:29 the property that there's a converging series a_i such that f(a_i) doesn't converge 20:23:24 -!- pikhq has joined. 20:23:33 -!- pikhq_ has quit (Ping timeout: 244 seconds). 20:23:34 obviously if you are linear in some nbhd then you don't have this property, by looking at tails 20:23:43 of stuff 20:24:26 -!- sebbu2 has changed nick to sebbu. 20:37:52 -!- atehwa has joined. 20:45:24 * Sgeo isn't bothering 20:52:52 * kallisti considers making a starcraft 2 mod. 20:55:30 I know there's a lot of units from story mod in the mod tool, but I wonder if you can get all of the upgrades as well. 20:56:10 there's a lot of random shit for terrans you could throw into the game, for a TvT map with all of Wings of Libery stuff available. It would be so incredibly broken. 21:03:53 -!- oerjan has joined. 21:05:57 oerjan: are there cases where it's useful to change the topology of R^n? or even the uniform structure, i guess the usual metrics are uniformly equivalent? 21:06:08 i don't really know. 21:06:22 i know very little about uniform structure. 21:08:03 even the different banach space topologies that come to mind are the same on finite dimensional spaces, i think 21:08:24 (norm and weak-*) 21:10:41 i think so too 21:11:16 hm maybe there's one based on considering only compact sets 21:12:10 not sure what you mean 21:12:17 how does that give a topology on R^n 21:12:35 if you make a topology where only the original compact sets of R^n and the whole of R^n are closed 21:13:23 it's not hausdorff, but might have some uses... 21:14:13 but it's T1 21:15:03 every nonempty open set contains everything outside some radius 21:16:16 So. 21:16:18 MUDCraft. 21:16:20 Yes or yes? 21:16:29 in fact the whole space is compact, i think 21:16:30 (I suppose I should ask that in -minecraft ... ) 21:17:06 oerjan: are there cases where it's useful to change the topology of R^n? or even the uniform structure, i guess the usual metrics are uniformly equivalent? 21:17:07 -!- Sgeo has quit (Ping timeout: 240 seconds). 21:17:33 I was thinking about the metric-y topology in a Minkowski space, actually. 21:17:57 * oerjan hasn't heard of that 21:18:25 i.e. an open ball is the set of points for which x^2 + y^2 + ... -t^2 < d. 21:18:35 I'm fairly sure all open sets are infinite. 21:18:52 Yes, yes they are. 21:19:15 wait what's t 21:19:21 time 21:19:45 this is relativistic proper time distance 21:19:56 umm what 21:20:05 do you mean the definition changes every second 21:20:09 :D 21:20:14 -_- 21:20:20 no, the space is R^4 21:20:21 no but umm for R^n, do you take one to be time or what? 21:20:28 and yes 21:20:28 Yeah. 21:20:39 It's not a metric space, of course. 21:20:54 i don't see that 21:21:12 also you mean metrizable 21:21:14 so, the set of points at timelike or < sqrt(d) spacelike distance from the origin 21:21:43 -!- Sgeo has joined. 21:21:48 * oerjan forgot the term for exactly 0 21:21:51 oklopol, no, I mean (R^n, relativistic proper time) isn't a metric space. 21:22:00 oerjan, yeah. 21:22:08 Having the cover to the HD fall out onto your bed is awesome! 21:22:21 so it includes the light cone, and some more beside 21:22:33 I didn't say d had to be positive, actually. 21:22:38 oh 21:22:52 (This only occurred to me now.) 21:23:01 well in any case i have no intuition for what kind of topology this is 21:23:22 But all open sets are asymptotic to a double cone, so they all have infinite intersections. 21:23:37 Thus all open sets are infinite. 21:24:18 not really asymptotic to a cone per se, since the radius grows squarely no? 21:24:28 but they sure are infinite 21:24:31 They're all hyperboloids. 21:25:48 well i think that's T1 too, at least 21:26:04 er no 21:26:08 or wait 21:26:13 Oh, yeah, it's not Hausdorff. 21:26:37 well that is obvious from infinite intersections 21:27:43 hm i think it's T1 21:28:01 i don't get this topology, is 0 really in all open sets? 21:28:16 Is 0 really in all open sets? 21:28:20 oklopol: obviously you translate 21:28:26 obviously? 21:28:34 you get a topology without doing that 21:28:36 Oh, I stated that stupidly. 21:28:40 it's just stupid :D 21:28:52 then it's hausdorff 21:28:55 Um, yeah; d(x,y) < a where d is the proper time. 21:29:01 erm 21:29:04 no so not 21:29:26 anyhow not t1 21:29:37 Phantom_Hoover: proper time is imaginary not negative, you need to square 21:29:47 consider two points in the same point in space 21:29:50 *the proper time squared 21:30:05 their neighborhoods always contain the other no 21:30:19 oklopol, ...no set is T1 if you use that argument? 21:30:37 what does it mean for a set to be T1? 21:30:50 you mean topology? 21:30:52 T1 is each has a neighbourhood that doesn't contain the other. 21:30:54 oh let me be more clear 21:31:05 consider two points that are equal in space but not time 21:31:07 oklopol: i thought that for a moment, but you can separate with cones from other points 21:31:55 What oerjan said. 21:31:56 the sets Phantom_Hoover gave are just a basis, not all open sets, and not all neighborhoods of x 21:32:07 er, subbasis even 21:32:38 yeah but if the times are -1 and 1 21:32:42 i don't see how 21:32:45 oerjan, it's not a basis? 21:33:04 Phantom_Hoover: a basis should be closed under finite intersection iirc 21:33:04 "Adult X-rated Christmas Trivia Games Printables" 21:33:16 OK, the spambots have gone beyond the plausible limits of "specific" now 21:33:19 since all open sets are symmetric around the space hyperplane at time 0 21:33:28 or not closed, but have smaller elements within 21:33:31 oklopol, they're all hyperboloids, so you can fit one point into the gap around the middle of the other. 21:34:02 or do you also take time translations? i guess that would make sense. 21:34:15 oklopol: you take all translations 21:34:25 d(x,y) = d(x-y, 0) 21:34:25 so T1 but no open sets are disjoint 21:35:13 oerjan, hmm, WP just says that you can form all open sets from unions of the basis. 21:35:16 so i think too 21:35:51 yes, that's the definition of a basis 21:35:55 Phantom_Hoover: yes. but your sets are weaker than that, since you cannot find one of them inside the intersection of two of them always 21:35:57 for subbasis, you also take finite intersections 21:36:17 oerjan, oh, duh. 21:37:44 iirc: given O1 and O2 in the basis, and x in their intersection, there must be O in the basis with x in O subset O1 intersect O2 21:38:00 -!- Phantom__Hoover has joined. 21:38:16 sounds about right 21:38:38 that's enough to make sure the intersection is a union from the basis 21:39:00 oerjan: did you read my characterization for functions f such that there's a series that converges but sequence of f images doesn't 21:39:16 no 21:39:26 i seem to have interrupted my log reading 21:39:42 it was that there's no nbhd of 0 where it's linear 21:40:27 -!- Phantom_Hoover has quit (Ping timeout: 240 seconds). 21:45:41 Maybe I'll just start talking about @ whenever anyone disagrees with me on reddit. <-- and then you can exchange strategy tips with Zephir_AWT 21:45:55 sorry, he's Zephir_Banned now 21:46:10 or was that Zephir_Banned2 21:46:44 -!- Jafet has quit (Quit: Leaving.). 21:54:36 " Is it unethical for me to ask for hints on https://docs.google.com/document/pub?id=1UK8EfHhrdZ2XrMMsvPk8DiEyQfaIbijkC4awkWFSdw4 ? (The problem asks if there are any series that converges but the series of the cubes of the terms diverges)" this is trivial 21:55:06 i've got a hunch it should be possible to get it with negatives 21:55:27 -!- Phantom__Hoover has quit (Ping timeout: 240 seconds). 21:55:55 get it? 21:55:57 -!- Phantom__Hoover has joined. 21:56:02 oh 21:56:03 an example... 21:56:06 yeah it is 22:01:45 ais523, can you get graue to bot a sock or something? 22:03:40 I emailed graue about the spam, haven't got an answer yet 22:05:15 hm ok, i think your characterization is right 22:05:19 I think it's too late to submit anything now 22:06:38 oklopol: or wait, don't you mean additive rather than linear 22:07:06 well i mean "y = ax" 22:07:06 or hm 22:07:40 those are equivalent except for hideously nonmeasurable functions 22:07:49 iirc 22:08:21 okay, in any case i think y = ax gives a characterization for absolutely all functions 22:08:40 oh hm i think linear may be right 22:09:59 if f(ax) is not a f(x), then you can use x and ax at some point in a variation of your argument in the privmsg 22:10:19 Give me something to watch on the YouTuber. 22:10:25 yeah but you need exactly one to be negative 22:10:29 and by "no neighborhood of 0", you can pass on to a smaller one later 22:10:40 oklopol: oh 22:10:53 then make a sum out of them that gets either to zero or arbitrarily close, and you get f's to get bigger 22:11:07 but because of f not necessarily being symmetric, i needed some extra work 22:11:36 well, if f(-x) = f(x) fails enough, you can use those, and if f(-x) = f(x) _is_ true in a neighborhood, you can switch one of those for the other to get it negative 22:11:41 er 22:11:46 *f(-x) = -f(x) 22:12:04 yeah that's a better way 22:12:54 so at least it's symmetric in a nbhd of 0 22:15:28 then if you find arbitrarily small r, r' for which f(r)/r and f(r')/r' are not equal, if r/r' rational you can get to 0 without going further than r+r' away from 0 on the linear side, and obviously you get somewhere on the f side 22:15:45 for irrational, you note that their sums are dense 22:16:08 so you can get the leftover bits as small as you want, and so converging 22:16:18 so you get arbitrarily close to 0 without leaving r+r'-nbhd, and it should then be clear enough that the f-side won't do the same 22:16:25 well 22:16:40 not quite, we still need to move to smaller r, r' 22:17:06 otherwise our jumps will not get smaller and the sum refuses to converge 22:17:25 well i mean the leftover from _all_ the r, r' pairs used should be a converging series 22:17:35 well sure 22:17:51 say you could make it < 1/n^2 22:18:05 but the point is we can estimate how many r's and -r''s we used, and that will approach something like their quotient when we get close to 0, so on the f side, you don't approach 0. 22:18:17 yes. 22:18:17 because the quotient is different there 22:18:49 i love how i'm procrastinating from doing math by doing math 22:19:14 aw, I was disappointed by your use of quotient 22:19:19 too concrete 22:19:24 * copumpkin goes back to real work 22:20:04 copumpkin: i think it's necessary to do concrete calculations in the case where f need not be continuous 22:20:17 for continuous f, it's kind of trivial 22:20:23 oklopol: now you need to generalize this to limits of diagrams in abelian categories to please copumpkin 22:20:32 yeah, I just thought you were talking about a more general notion of quotient 22:21:02 not sure the general notion of quotient in any way generalizes division 22:21:04 copumpkin: sorry, this is R -> R functions 22:21:19 well. in many ways i guess. 22:21:26 yeah 22:21:33 there are definitely similarities 22:22:10 Z_m / Z_n _is_ isomorphic to Z_(m/n) if n divides m, i think. 22:22:16 as groups 22:22:23 okay i have to go to work too. and by that i mean i have to go walk outside. 22:22:53 isn't A = A/B * B for abelian groups 22:23:11 no. 22:23:25 e.g. Z_4 is not Z_2 * Z_2 22:23:31 -!- Patashu has joined. 22:23:34 oh right 22:23:35 that stuff 22:24:12 different extensions 22:24:58 basically, in module theory there is a concept of B being a summand in A 22:25:42 (abelian groups ~ Z-modules) 22:28:55 our wiki recent changes have gotten into the christmas spirit, i see 22:29:56 hark the herald spammers sing 22:31:46 the little spammer boy 22:31:57 silent spam, holy spam 22:32:14 -!- kwertii has joined. 22:33:55 spam nuts roasting on an open fire 22:38:20 spam the halls with bouts of holly 22:38:35 *boughs 22:41:09 porky the spam man 22:50:53 * Phantom__Hoover → sleep 22:51:18 -!- Phantom__Hoover has quit (Quit: Leaving). 23:01:49 oerjan: please stop spamming 23:01:53 it's cold outside 23:01:53 :/ 23:05:14 adeste spammantes 23:07:00 -!- zzo38 has joined. 23:07:36 oerjan: hi, we've recently been plagued by spam from a guy named oerjan. Can you please kickban him? This is getting out of control. 23:08:42 -!- ais523 has quit (Remote host closed the connection). 23:19:40 have yourself a spammy little christmas 23:19:51 food -> 23:21:11 -!- Jafet has joined. 23:28:49 spammis angelicus 23:30:27 feliz spammidad 23:33:05 huh there's apparently a Call of Cthulu video game. 23:33:15 *Cthulhu 23:34:34 darn, i guess there won't be an iwc eldritch christmas song this year