00:00:17 <oerjan> shachaf: this may be the same problem with @let we discussed the other day
00:01:00 <oerjan> it passes things through Lang.Haskell.something without extensions set
00:02:07 <boily> quelqun_dautre: bonsoir. es-tu quelqu'un d'autre?
00:02:42 <oerjan> int-e said it might be easy to fix, but not before the weekend.
00:03:34 <oerjan> or wait, can it be that, doesn't that phase only do parsing
00:03:38 <shachaf> is it like a werewolf sort of bug
00:06:07 <oerjan> shachaf: it does two checks, the first only on the given declaration, then in L.hs context
00:12:49 -!- tromp has joined.
00:12:50 <douglass1> by "actual physics" I meant "run by the physics department." the chem department's version did indeed contain physics
00:14:42 -!- edwardk has joined.
00:22:01 <Bike> my chem manual last term had instructions on how to read a graph of a line, so i'm a bit skeptical of chemphysics now
00:24:10 -!- shikhout has joined.
00:27:29 -!- shikhin has quit (Ping timeout: 252 seconds).
00:32:41 -!- tromp has quit (Ping timeout: 264 seconds).
00:43:03 -!- hk3380 has quit (Ping timeout: 265 seconds).
00:50:38 -!- edwardk has quit (Ping timeout: 240 seconds).
01:01:24 -!- tromp has joined.
01:02:24 -!- edwardk has joined.
01:05:23 -!- edwardk has quit (Client Quit).
01:06:07 -!- edwardk has joined.
01:08:05 -!- tromp has quit (Ping timeout: 264 seconds).
01:20:12 <boily> AAAAAAAAAAAAAURGH! why. why is my neighbour vacuum cleaning for at least the third time tonight...
01:22:53 -!- edwardk has quit (Quit: Computer has gone to sleep.).
01:25:53 <oerjan> boily: it's the hygiene singularity approaching hth
01:28:20 -!- edwardk has joined.
01:29:08 <boily> I don't care about Approaching Singular Hygienes. I just want to be able to induce something that may resemble sleep on my person.
01:32:05 -!- Phantom_Hoover has quit (Quit: Leaving).
01:37:47 -!- hk3380 has joined.
01:39:04 <zzo38> The saving throws for healing spells in Dungeons&Dragons are not symmetric!
01:40:41 <zzo38> (Specifically, the saving throw is different for living and undead targets.)
01:41:11 <zzo38> More symmetric probably would be if the saving throw is the same except for the "(harmless)" marker.
01:47:45 -!- boily has quit (Quit: RECUUMED CHICKEN).
01:56:49 -!- edwardk has quit (Ping timeout: 258 seconds).
02:06:37 <zzo38> How well does keyless RC4 pass randomness tests?
02:15:49 -!- oerjan has quit (Quit: Nite).
02:16:38 <monotone> Do you mean RC4 without a stream key?
02:17:57 <monotone> The output would be deterministic at that point, no?
02:18:14 -!- mekeor` has joined.
02:18:45 <zzo38> Yes, although it could still be tested
02:19:03 -!- mekeor has quit (Ping timeout: 252 seconds).
02:19:17 <monotone> From a cryptographic standpoint, that wouldn't make much sense.
02:19:25 <zzo38> You can then add a key if you want it not to be the same each time, but make it the same for the same key.
02:20:01 <zzo38> SQLite uses RC4 with a 256-byte key for its random number generator.
02:20:14 <monotone> Yes, that's the idea of the key with all stream ciphers.
02:21:35 <zzo38> SQLite obtains the key from the VFS implementation, and is intended not to ever use the same key, although it is possible to do so.
02:21:43 <monotone> I don't think you can meaningfully define "random-looking" without a key, though.
02:22:12 <zzo38> Famicom Hangman does not use a key, but it generates a new random number several times per frame and normally ignores the result.
02:22:39 <zzo38> (It only uses the result when the space-bar is held down)
02:23:45 -!- Sorella has quit (Quit: It is tiem!).
02:24:21 <zzo38> How well do you expect that to work?
02:25:52 <monotone> If you just need something that looks kinda random then, hey, you can't do worse than Pokémon.
02:25:58 <zzo38> Well, it is a hangman game.
02:28:20 <monotone> RC4 is probably overkill for that unless your answers are TLS streams.
02:34:19 <zzo38> Another thing I wanted to use is to generate random booster packs (such as for Limited tournaments in Magic: the Gathering). In such a case there would be a key, but if one is not explicitly specified, it could be generated automatically using date/time and other things.
02:34:51 <zzo38> If you have 100 choices, you might use 7-bits and then if the result exceeds 99, you can try again until it doesn't exceed 99.
02:35:52 <zzo38> (At least, it is what I have implemented in any program which I implement a random number algorithm; I don't know how common it is.)
02:46:36 <kmc> RC4 is a shitty RNG
02:47:04 <kmc> it has hella biases
02:47:32 <kmc> http://en.wikipedia.org/wiki/RC4#Security
02:47:49 <kmc> most egregiously "the second output byte of the cipher was biased toward zero with probability 1/128 (instead of 1/256)"
02:48:02 <kmc> so no, it's not even suitable for non-cryptographic uses
02:48:26 <zzo38> It is commonly used though, it seems
02:49:39 <zzo38> Even without the modifications, although there are many modifications that are used too, such as skipping a lot of the start of the output, and/or modifying the key scheduling algorithm to run a large number of times, etc
02:50:54 <Bike> can't you just use an actual prng
02:55:34 <zzo38> It would seem clear now that RC4 is no good if used plainly, at least; you would have to use carefully. Actually any algorithm should be used carefully but RC4 has some specific weaknesses to consider.
02:57:04 <zzo38> I would think any algorithm you are encrypting something with, you may want to compress it first, using a compression algorithm without headers and so on; you can even use especially what is known based on what kind of messages they are, for example if it is ASCII text one simple thing you can do is just ignore the bit7 of the results.
03:09:41 <zzo38> One thing you can do with any algorithm, assuming the messages do not get lost, is the message, once decrypted, contains a key for next message, or some part of the key.
03:10:40 <zzo38> You can also pad messages and add random junk inside of some parts and so on.
03:12:36 <zzo38> For example, an attacker who knows that the message contains "Meet Jane and me tomorrow at 3:30 pm" at a particular point can recover the keystream at that point from the ciphertext and plaintext. Then the attacker can replace the original content with any other content of exactly the same length, such as "3:30 meeting is cancelled, stay home" by encrypting it with the recovered keystream, without knowing the encryption key.
03:12:53 <zzo38> If you add random junk inside of the message then you can avoid such a thing.
03:27:30 <zzo38> I read about steganography (possibly in combination with encryption), and even about encoding the text inside of a diagram of a flower, or in music, or something. Once someone blink their eye by a Morse code. Once I tried to use a coughing code to give the answers to a test (simply to see if I could).
03:35:54 -!- Bike has quit (Ping timeout: 245 seconds).
03:40:40 -!- nucular has quit (Quit: Excess Food).
03:52:41 <zzo38> Another thing I thought of is to use a deck of cards, in combination with a book describing the rules for some card game (the game described having nothing to do with the encryption).
03:54:38 <zzo38> As well as sudoku (something described in Wikipedia), chess problems, astrology, and other things.
03:55:16 <zzo38> Or in a game played at a chess tournament...
03:55:25 -!- shikhout has quit (Ping timeout: 265 seconds).
03:56:56 <zzo38> Have these been done?
04:06:53 -!- hk3380 has quit (Ping timeout: 264 seconds).
04:10:10 <zzo38> I have used the following program to test the randomness of SQLite: create table d(v int); insert into d(v) with x(x) as (select 0 union all select x+1 from x where x<10000) select random('4d6dl') from x; select v, count() from d group by v order by v; What should be the proper probabilities of 4d6 drop-lowest having each total?
04:28:52 <zzo38> Maybe ten thousand samples would not be sufficient?
04:34:16 -!- Bike has joined.
04:44:28 <kmc> this is probably going to be one of those nights where I get really drunk and start writing weird / emo things on irc
04:44:32 <kmc> just letting you all know ahead of time
04:46:18 <kmc> fungot: are you ready
04:46:18 <fungot> kmc: hills; and even such whispers admitted that there was a long and close correspondence with carter had a still more mysterious realm where everything is white, and where the old chief fnord, they said, was the
04:50:05 -!- trout has changed nick to variable.
04:50:08 -!- variable has changed nick to constant.
04:50:11 -!- constant has changed nick to function.
04:58:38 <fowl> classic carter
04:59:28 -!- fowl has quit (Remote host closed the connection).
05:00:40 -!- fowl has joined.
05:00:42 <Sgeo> http://www.reddit.com/r/ideasfortheadmins/comments/25rcsz/allow_moderators_of_subreddits_to_see_ipuseragent/ am I mistaken in what I posted?
05:05:05 <Bike> "I assume that there would be fewer hashes than IP addresses" seems weird
05:09:38 <Bike> md5 hashes are 128 bits, so the same length as v6.
05:10:22 <Bike> "if the algorithm is public" etc. is a basic standard for crypto, all the good algorithms are open, you know that.
05:11:30 -!- function has changed nick to variable.
05:13:13 <Bike> ooh hey, SHA2 can do up to 512.
05:15:53 <Bike> anyway this system obviously wouldn't much help against a half-savvy spammer, they'd just restart DHCP.
05:17:41 <kmc> SHA-256, SHA-384, SHA-420
05:20:11 <newsham> birthday attacks around 2^1.5
05:20:29 <Bike> so like, three?
05:21:01 <Bike> You could just say three.
05:21:27 <newsham> dont accept every two bit hash you come across
05:22:53 <Bike> oh, you meant sha-3 as a three bit hash, not "the third SHA", which it atually is.
05:26:38 -!- nooodl has quit (Read error: Connection reset by peer).
05:41:24 -!- drdanmaku has quit (Quit: Connection closed for inactivity).
05:42:21 <Bike> yeah beyond me
06:00:03 -!- ^v has quit (Quit: http://i.imgur.com/Akc6r.gif).
06:39:30 -!- edwardk has joined.
06:56:33 -!- edwardk has quit (Quit: Leaving...).
06:57:08 -!- edwardk has joined.
07:08:34 -!- Bike has quit (Ping timeout: 240 seconds).
07:19:58 <HackEgo> ringcoin yabucoin retacoin dumbermiprelytinecoin duerecoin boulcoin madcoin frackledcoin cthcoin fettecoin brucoin quidcoin mkbcoin twingcoin magecoin mekacoin tunecoin ijulicoin reulcoin crainfusiccoin
07:22:51 -!- password2 has joined.
07:39:33 <Jafet> ringcoin http://www.youtube.com/watch?v=548qCuDHe_Q
08:15:31 <HackEgo> [wiki] [[Forobj]] http://esolangs.org/w/index.php?diff=39543&oldid=39505 * GermanyBoy * (+92)
08:16:18 <HackEgo> [wiki] [[User:GermanyBoy]] N http://esolangs.org/w/index.php?oldid=39544 * GermanyBoy * (+58) Created page with "Hello! == My languages == === Esoteric === * [[Forobj]]"
08:57:05 <kmc> http://regmedia.co.uk/2014/05/16/0955_peter_gutmann.pdf "It’s probably at least some sort of sign of the end times when your conference badge has a rootkit"
09:31:52 -!- conehead has quit (Quit: Computer has gone to sleep).
09:36:29 -!- edwardk has quit (Quit: Computer has gone to sleep.).
09:37:17 -!- nooodl has joined.
09:42:23 -!- nooodl has quit (Ping timeout: 255 seconds).
09:43:52 -!- nooodl has joined.
09:47:14 -!- Slereah_ has joined.
09:49:51 -!- Slereah has quit (Ping timeout: 258 seconds).
09:56:33 -!- Phantom_Hoover has joined.
10:02:37 -!- MoALTz has joined.
10:30:54 -!- Frooxius has quit (Ping timeout: 245 seconds).
10:40:20 -!- edwardk has joined.
10:51:42 <mroman> "Exploit backwards-compatibility support in the CPU for bugs dating back to the 80286" sweet :)
10:52:22 <mroman> but what he's mostly saying is "physical access to a machine is enough"
10:57:01 <Slereah_> The NSA will look at it and go "Man fuck that"
11:02:43 -!- mhi^ has joined.
11:09:43 <mroman> "I love crypto, it tells me what part of the system not to bother attacking"
11:10:21 <Slereah_> Well some crypto is easy to attack
11:10:26 <nortti> http://regmedia.co.uk/2014/05/16/0955_peter_gutmann.pdf "It
11:11:14 <mroman> I had this idea for a secure key transmission protocol
11:11:42 <mroman> Alice want's to transmit m, so she chooses a and r
11:11:54 <mroman> she transmits m+a `mod` r
11:12:04 <mroman> bob receives it and chooses b and sends m+a+b `mod` r
11:12:09 <Slereah_> That's why physics is way better than cryptography
11:12:20 <Slereah_> In cryptography, Alice is always sending shit to bob
11:12:23 <mroman> alice receives it, subtracts her a and sends back m+a+b-a `mod` r
11:12:37 <mroman> bob subtracts his b again m+a+b-a-b `mod` r == m
11:13:08 <mroman> little that you know that you can just calculate b as (m+a+b)-(m+a)
11:26:35 -!- yorick has joined.
11:32:35 -!- hk3380 has joined.
11:32:39 -!- password2 has quit (Ping timeout: 265 seconds).
11:58:52 -!- oerjan has joined.
12:06:53 -!- MoALTz has quit (Quit: Leaving).
12:16:52 -!- nucular has joined.
12:18:17 <HackEgo> [wiki] [[Language list]] M http://esolangs.org/w/index.php?diff=39545&oldid=39533 * Oerjan * (+0) /* O */ rder
12:25:03 <Taneb> > maxBound :: Word
12:26:53 <HackEgo> [wiki] [[Special:Log/newusers]] create * Luiji * New user account
12:28:09 <lambdabot> MonadState s m => (s -> s) -> m ()
12:29:06 <HackEgo> [wiki] [[List of ideas]] http://esolangs.org/w/index.php?diff=39546&oldid=39457 * Luiji * (+198) /* Joke/Silly Ideas */ commented on the final fantasy idea
12:29:22 <HackEgo> [wiki] [[Owhelgossip]] http://esolangs.org/w/index.php?diff=39547&oldid=39532 * Oerjan * (+34) formatting
12:29:31 <lambdabot> MonadState s m => (s -> s) -> m ()
12:29:33 <int-e> shachaf: lambdabot has the monomorphism restriction turned off, but there was not {-# LANGUGAE NoMonomorphismRestriction #-} in L.hs.
12:34:10 <int-e> FreeFull: "dmr" = "dreaded monomorphism restriction"
12:34:30 <FreeFull> Too bad it's in the language spec
12:34:36 <int-e> (it's dreaded mainly because it leads to obscure type errors)
12:42:13 -!- shikhin has joined.
12:47:32 -!- nooodl has quit (Read error: Connection reset by peer).
12:48:09 -!- nooodl has joined.
12:48:53 -!- password2 has joined.
12:49:09 -!- MindlessDrone has joined.
12:57:58 -!- MoALTz has joined.
13:04:52 <HackEgo> [wiki] [[List of ideas]] http://esolangs.org/w/index.php?diff=39548&oldid=39546 * 188.120.201.244 * (+111)
13:22:54 -!- password2 has quit (Ping timeout: 240 seconds).
13:30:18 -!- idris-bot has quit (Ping timeout: 240 seconds).
13:31:12 -!- Melvar` has joined.
13:31:14 -!- Melvar has quit (Ping timeout: 240 seconds).
13:31:26 -!- Melvar` has changed nick to Melvar.
13:35:52 -!- password2 has joined.
13:36:32 -!- nucular_ has joined.
13:36:37 -!- password2 has quit (Max SendQ exceeded).
13:36:41 -!- nucular has quit (Ping timeout: 276 seconds).
13:37:03 -!- nucular_ has changed nick to nucular.
13:37:07 -!- nucular has quit (Changing host).
13:37:07 -!- nucular has joined.
13:37:11 -!- password2 has joined.
13:46:02 -!- boily has joined.
13:54:24 <Taneb> HELP I AM BOOKING A HOTEL FOR AN ANIME CONVENTION
13:54:34 <Taneb> I HAVE SIMULTANEOUSLY GAINED AND LOST CONTROL OF MY LIFE
13:55:00 <Taneb> That is, I'm going to attend an anime convention nearby
13:55:07 <HackEgo> Arrows are just strong monads in the category of profunctors.
13:56:14 <boily> Taneb: TANELLE. THIS IS GOOD FOR YOU. HAVE YOU EVER ATTENDED AN ANIME CONVENTION BEFOREHAND, AND/OR MITIGATED YOUR LIFE CONTROL?
13:56:27 <Taneb> I have attended one before (it was this one last year)
13:56:36 <Taneb> But that time I stayed at a friend's house
14:00:42 <impomatic> Is there anyone who's got a minute to read / comment on some text for a banner? If so, I'll /msg the link
14:01:32 <quintopia> a real banner or a web page advertisement?
14:02:26 -!- Slereah has joined.
14:02:36 <boily> quintopiaaaaaaaaaaaaaaaaaah!
14:03:06 <quintopia> boily: you never explained what exactly i would need to do to move (temporarily) to canada and work
14:03:13 <oerjan> exclamation points!!!!!!!!!!!!!!!!!!
14:05:05 -!- Slereah_ has quit (Ping timeout: 240 seconds).
14:05:44 <boily> quintopia: eh... as a whole, I think you need «une conjointe», some family, a referral, a job contract or something else to that same effect.
14:06:54 <boily> quintopia: fr:conjointe → en:de facto spouse.
14:07:28 <boily> our Fine Government wants either families or workers in a plug'n'play state.
14:10:47 <boily> quintopia: you could try workopolis → http://www.workopolis.com/rechercheemploi/trouver-emplois#pt=%257C2%257C&lg=FR&l=%257Cmontreal%252Cquebec%257C&pr=-1
14:10:57 <Taneb> boily, what if I have an Australian citizenship
14:11:13 <boily> Taneb: eh... is Australia part of the Commonwealth?
14:11:35 <boily> quintopia: that's where I found my current job. (previous one was through internships, so I can't really help you there.)
14:11:57 <boily> Taneb: besides, you're a UK resident, so you're already commonwealthian.
14:12:14 <Taneb> Australia sometimes has nicer treaties
14:12:24 <boily> I guess it'd be /slightly/ easier for Taneb to apply for a Canadian job.
14:13:10 <Taneb> I know it's easier to get a job in the US as an Australian than it is as a Brit
14:13:16 <boily> quintopia: Banque Nationale and Desjardins are booming right now. if you could land something with them, it'd be a nice cushiony job.
14:13:55 <boily> that's exactly what they need right fungotting now.
14:13:55 <fungot> boily: his daughter, the dawn is nigh, though only about two feet tall, paved with octagonal blocks and solidly vaulted overhead. there would he ever say he once dwelt as a prince, though here we knew him from his birth though he thought himself a king's son. he was pulled out of bed and into empty space. again the fnord of
14:15:01 <boily> fungot, the father of the Dwarven Octagonal Block Vaulting Champion.
14:15:01 <fungot> boily: tried to keep the guards in shape and size, and weathered and pitted to the very confines of infinity.
14:15:05 <Taneb> Not that I have any need or desire to work in Canada right now
14:15:14 <Taneb> What with being on the first year of a degree
14:17:54 <quintopia> i've got a chunk missing from my number 9 incisor that is what i need fixed right fungotting now
14:17:54 <fungot> quintopia: trapped though he was, however, are of the keenest interest and poignancy; suggesting with marvelous force the imminence of nameless terrors, and how taran-ish had died from fear and left a dead village whose corpses were horribly mangled, chewed, and clawed.
14:20:58 * boily twitches. “it's tooth number 21, not 9!”
14:25:48 <fungot> Available: agora alice c64 ct darwin discworld enron europarl ff7 fisher fungot homestuck ic irc iwcs jargon lovecraft* nethack oots pa qwantz sms speeches ss wp youtube
14:26:14 <fungot> Selected style: qwantz (Dinosaur Comics transcriptions 2003-2011)
14:26:27 <oerjan> fungot: you need to balance the darkness
14:26:28 <fungot> oerjan: t-rex, i have big news the other is a sad realization i always sort of a first world problem, yes, that language is good for domain-specific solutions, but it doesn't make me less of a man!
14:41:37 <quintopia> boily: do you guys number them differently?
14:46:07 -!- Frooxius has joined.
14:48:17 -!- hk3380 has quit (Ping timeout: 264 seconds).
14:52:17 -!- Patashu_ has quit (Ping timeout: 240 seconds).
14:54:41 <zzo38> I don't expect octagonal blocks to tile.
14:55:12 <boily> quintopia: we number them ISO style → http://en.wikipedia.org/wiki/Dental_notation
14:55:18 <boily> zzo38: you lack imagination.
14:55:38 <zzo38> (unless you put squares in between, too)
14:56:33 <quintopia> oh the iso system does seem fairly sane
14:56:45 <int-e> zzo38: how about non-convex octagons...
14:57:00 <quintopia> too bad my dentists would get confused if i said 21 (that's not even an incisor in the universal system)
14:57:07 <int-e> that or hyperbolic geometry
14:57:44 <zzo38> int-e: O, I forgot... I suppose especially if they are many different shapes of octagons; and, yes different kind of geometry too (at least, geometry is one thing I did think of)
14:58:38 <quintopia> zzo38: what about infinite regress octagonal fractal tiling
14:58:52 <zzo38> I don't know what that is.
14:59:17 <quintopia> you tile the plane with octagons. then you fill the spaces between them with octagons
14:59:19 <int-e> zzo38: but if you allow them to have different shapes then it becomes almost trivial.
14:59:31 <boily> quintopia: in our software, we have a tooth number class, with methods to get the international and American numbers.
14:59:39 <zzo38> int-e: Yes I know that
14:59:53 <int-e> heh. is a square removed from a square an octagon? it has 8 corners, 4 outside, 4 inside...
15:00:20 <zzo38> (and you can fill a shape with triangles of different shapes too)
15:03:07 <boily> int-e: ianam, but no.
15:03:40 <zzo38> int-e: I don't think so; I count nine sides. Still I don't really know.
15:05:02 <int-e> boily: I agree, but I found the thought amusing.
15:05:44 <boily> it's like a car has eight wheels. two in front, two in the back, two left and two right :D
15:07:16 <zzo38> I also count twelve corners in the shape described.
15:07:47 <zzo38> I am probably both wrong, though.
15:08:12 <int-e> zzo38: http://int-e.eu/~bf3/tmp/tile.png is what I had in mind when I mentioned non-convex octagons
15:10:14 <quintopia> boily: oh you work with dental. i forgot
15:11:29 <ion> http://i.imgur.com/jmgzg0w.png
15:11:35 <int-e> I guess the square with a square hole is, as a polygon, a degenererated decagon (with two coincident edges connecting an outside corner to an inside corner)
15:11:55 <Taneb> ion, I am pretty sure that a number of those characters are not ASCII
15:12:26 -!- alone has changed nick to syndrome.
15:12:27 <Jafet> It seems to be possible to get six, eight, nine or twelve sides, but no others.
15:13:08 <boily> ion: oh. ooooooooh...
15:13:45 <int-e> Jafet: sides for what?
15:17:22 <Jafet> A polygon obtainable by removing a square from a square.
15:18:02 <Jafet> Actually, only six, eight or nine.
15:20:28 <int-e> ok, I see what you mean
15:22:02 <int-e> when you had 12 you should also get 10 (with two corners of the inner square touching the sides)
15:24:22 <Jafet> One of the sides is not necessary, so it should not be counted.
15:35:02 -!- shikhout has joined.
15:35:40 -!- boily has quit (Quit: BROTHABLE CHICKEN).
15:38:01 -!- shikhin has quit (Ping timeout: 252 seconds).
15:39:50 -!- Melvar has quit (Ping timeout: 258 seconds).
15:41:54 -!- Melvar has joined.
15:49:41 -!- hk3380 has joined.
15:50:41 -!- password2 has quit (Ping timeout: 240 seconds).
15:57:58 -!- Bike has joined.
16:00:47 <ion> Phantasy Star Online (Japanese: ファンタシースターオンライン Hepburn: Fantashī sutā Onrain
16:03:38 -!- oerjan has quit (Quit: Marufelusu).
16:09:11 -!- password2 has joined.
16:09:52 -!- password2 has quit (Max SendQ exceeded).
16:10:10 -!- MoALTz has quit (Quit: Leaving).
16:10:36 -!- password2 has joined.
16:11:30 -!- password2 has quit (Max SendQ exceeded).
16:12:11 -!- password2 has joined.
16:14:09 -!- password2 has quit (Max SendQ exceeded).
16:15:16 -!- shikhout has changed nick to shikhin.
16:17:42 -!- password2 has joined.
16:18:33 -!- password2 has quit (Remote host closed the connection).
16:20:03 <Sgeo> https://www.youtube.com/watch?v=SOD7Ni_3NIc
16:20:31 <Sgeo> So, they're doubling this guy's health repeatedly, but in the middle someone adds a +3 to that characters health, which totally ruins the pretty numbers imo :(
16:21:55 <Sgeo> Actually, I think that +3 didn't register, hmm
16:27:51 -!- Sorella has joined.
16:32:51 <int-e> Nice 32 bit integer overflow though.
16:39:25 -!- MoALTz has joined.
16:53:11 <HackEgo> [wiki] [[PHL 1.0]] N http://esolangs.org/w/index.php?oldid=39549 * GermanyBoy * (+2094) Created page with "'''Proceed High Language 1.0''' (also known as '''PROCEED''') is an esoteric programming language created by [[user:GermanyBoy]] in 2012. == Syntax == PHL 1.0 is based on su..."
17:00:22 <HackEgo> [wiki] [[PHL 1.0]] http://esolangs.org/w/index.php?diff=39550&oldid=39549 * GermanyBoy * (+533)
17:02:16 -!- edwardk has quit (Quit: Computer has gone to sleep.).
17:03:58 <HackEgo> [wiki] [[User:GermanyBoy]] http://esolangs.org/w/index.php?diff=39551&oldid=39544 * GermanyBoy * (+14)
17:04:42 <HackEgo> [wiki] [[Language list]] http://esolangs.org/w/index.php?diff=39552&oldid=39545 * GermanyBoy * (+14) /* P */
17:07:36 <ion> LibreSSL – An OpenSSL replacement. The first 30 days, and where we go from here. http://www.openbsd.org/papers/bsdcan14-libressl/mgp00001.html
17:10:46 <Bike> ebcdic support, lol
17:11:52 -!- boily has joined.
17:19:47 <HackEgo> [wiki] [[PHL 1.0]] http://esolangs.org/w/index.php?diff=39553&oldid=39550 * GermanyBoy * (+951)
17:20:17 <ion> http://image.gamespotcdn.net/gamespot/images/2002/vgnews/052802/gc_keyboard_790screen001.jpg
17:23:48 <HackEgo> [wiki] [[PHL 1.0]] http://esolangs.org/w/index.php?diff=39554&oldid=39553 * GermanyBoy * (+64)
17:29:56 -!- password2 has joined.
17:31:04 -!- password2 has quit (Client Quit).
17:36:53 -!- mhi^ has quit (Ping timeout: 264 seconds).
17:37:25 -!- mhi^ has joined.
17:39:23 -!- nooodl has quit (Ping timeout: 252 seconds).
17:41:38 -!- mhi^ has quit (Ping timeout: 240 seconds).
17:42:39 -!- mhi^ has joined.
17:44:49 -!- mhi^ has quit (Client Quit).
17:47:53 -!- KingOfKarlsruhe has joined.
17:48:46 <zzo38> Which PC motherboards includes a Forth system (or BASIC or assembler) in BIOS so that it can boot without an operating system?
17:49:20 <nortti> the original PC, at least
17:50:16 <zzo38> Better is if, such a thing also includes command for CMOS setting and for internet connection and serial port, so it is still possible to use it as a simple telnet client or terminal to directly connected equipment, and using for downloading an operating system using internet.
17:50:17 <nortti> if there ever were PCs with open firmware those'd have forth
17:50:34 <zzo38> Is Open Firmware compatible with the PC BIOS, though?
17:50:55 <zzo38> Then it isn't much PC.
17:50:56 <nortti> but one could've simulated it, like EFIs do
17:53:05 * boily lightly mapoles nortti for having mentioned EFI
17:53:07 <Bike> uefi boards often have anetwork stack
17:53:20 <boily> I hate EFI, UEFI, and all that weird crap.
18:20:26 -!- KingOfKarlsruhe has quit (Read error: Connection timed out).
18:20:34 <boily> (suh-weeet! “You finish putting on the +3 ring mail "Shoim" {+Inv rF+ rN+ Str+3 SInv}.”)
18:21:26 <zzo38> Now will they let you remove it too?
18:22:11 <HackEgo> [wiki] [[Gentzen]] http://esolangs.org/w/index.php?diff=39555&oldid=39542 * Zzo38 * (+132)
18:23:09 <boily> it's nice! the centaur in me is happy ^^
18:23:25 <zzo38> There is a centaur in you?
18:24:38 <boily> elliott: then answer it.
18:25:04 <HackEgo> [wiki] [[Gentzen]] http://esolangs.org/w/index.php?diff=39556&oldid=39555 * Zzo38 * (+28)
18:27:00 <zzo38> What is your opinion of the http://esolangs.org/wiki/Gentzen so far?
18:31:11 <boily> at a glance, it feels like an uncompressed ursala mixed in with abbreviated ruby-ish perl.
18:31:22 <boily> but it has “iosys” in it, which in my book is a big plus.
18:32:09 <boily> (holy fungot. two scrolls of acquirement on Lair:1. the RNG will hit me hard soon...)
18:32:09 <fungot> boily: so, hey, it was our table! if he's interested, he'll push it back to your side. the salt shaker over to their side, this means that that one incident to say that sentence:
18:33:03 <zzo38> Uncompressed ursala? I don't know of that. And how is it like mixed with abbreviated ruby-ish perl.
18:33:25 <zzo38> elliott: But if such an armour is cursed, then it would make it difficult to acquire a better one, I think?
18:33:53 <elliott> remove curse scrolls are incredibly common in crawl :p
18:34:08 <boily> elliott: that's exactly what I did. “the +2 pair of gloves of Sakiem (worn) {Str+2}”
18:34:45 <nortti> zzo38: there actually are some HP laptops with firmware (I think) linux system that is very limited and shitty
18:35:20 <zzo38> Ah, OK then use remove curse scroll if you find a better armor and then find that the one you just put on cannot come off.
18:36:39 <zzo38> nortti: How limited is it? I would think is better having a BIOS with Forth included, but doesn't include a whole bunch of stuff like application programs and so on.
18:37:11 <nortti> just a out of date firefox and a shitty filesystem browser, iirc
18:37:58 <zzo38> That's really silly
18:39:31 <zzo38> Better would be to have a Forth system with direct hardware access, and perhaps a terminal emulator
18:41:15 <zzo38> boily: I don't know about Ursala (I cannot find the information), and I don't know what "abbreviated ruby-ish perl" quite means. Can you explain it better perhaps?
18:44:02 <boily> zzo38: cpressey sez: “I also wanted to implement it in URSALA, but its official website seems to have vanished. (Well, I didn't stand a chance, anyway, really.)” on his user page.
18:44:35 <boily> I think https://github.com/gueststar/Ursala is pretty legit about ursala.
18:45:06 <boily> ruby-ish perl is just that nice visual over-abundance of punctuation symbols seasoning a program listing.
18:50:54 -!- shikhin has quit (Ping timeout: 240 seconds).
19:14:39 -!- edwardk has joined.
19:20:14 -!- augur has quit (Remote host closed the connection).
19:29:45 -!- conehead has joined.
19:34:22 -!- MindlessDrone has quit (Quit: MindlessDrone).
19:41:22 -!- AnotherTest has joined.
19:49:02 <zzo38> It doesn't seems to resemble URSALA to me
19:50:27 -!- boily has quit (Quit: ELVISH CHICKEN).
20:06:55 -!- nooodl has joined.
20:08:03 <coppro> marking first-year logic is fun
20:09:21 -!- edwardk has quit (Quit: Computer has gone to sleep.).
20:15:20 -!- edwardk has joined.
20:25:40 -!- edwardk has quit (Quit: Computer has gone to sleep.).
20:26:42 <zzo38> Do you prefer to stay in a hotel where the floor labels aren't numbers but instead are colors that indicate what color the pokemons are that correspond to the room numbers on that floor?
20:29:23 <Taneb> Is there such a hotel?
20:36:20 -!- augur has joined.
20:40:05 <coppro> that sounds like an amazing hotel
20:40:10 <coppro> we should all pool our money and build it
20:51:02 <int-e> this sort looks affordable. http://technabob.com/blog/2010/02/08/five-stars-cardboard-hotel/
20:52:04 <coppro> I should give one of my students the 'most likely to be a doctor' award
21:02:56 -!- edwardk has joined.
21:10:05 -!- nortti has changed nick to lawspeaker.
21:10:39 -!- lawspeaker has changed nick to nortti.
21:11:41 -!- AnotherTest has quit (Ping timeout: 252 seconds).
21:14:42 -!- edwardk has quit (Quit: Computer has gone to sleep.).
21:19:14 -!- oerjan has joined.
21:28:39 -!- nucular has quit (Quit: Excess Food).
21:30:27 <zzo38> I read about SWIG but it doesn't have some, such as SQL, Haskell, Forth, FORTRAN, AWK, and a few others.
21:35:17 -!- drdanmaku has joined.
21:43:18 -!- edwardk has joined.
21:44:26 <Bike> https://twitter.com/mc_hankins/status/467766548162412545/photo/1/large Statistics.
21:54:06 -!- edwardk has quit (Quit: Computer has gone to sleep.).
22:01:31 <zzo38> Is it OK to make a random selection from a list you don't know how long, if you first have a 1/1 chance to pick the first one, a 1/2 chance to discard it and put the second one instead, a 1/3 chance to discard it and put the third one instead, etc?
22:03:15 <Phantom_Hoover> Bike, twitter replies point out people might just be writing p = 0.05 rather than 0.050
22:04:11 <Bike> Phantom_Hoover: in the replies he mentions he just cut out .05 entirely since he couldn't text search it. the point is the asymmetry.
22:05:07 <int-e> Phantom_Hoover: a truly significant finding
22:05:35 <int-e> (there must be a better pun in it but I didn't fine it)
22:05:51 -!- augur has quit (Remote host closed the connection).
22:06:03 <int-e> https://xkcd.com/882/ is relevant
22:06:56 <Bike> he probably should have indicated this on the graph, but oh well
22:07:20 <Bike> why is he wearing goggles
22:08:17 <int-e> I believe there are 4 people involved in that story
22:08:55 <int-e> he, she, scientist A and scientist B.
22:09:21 <int-e> I base this on the first panel, which has voices coming from the right.
22:09:36 -!- edwardk has joined.
22:09:39 <Bike> how many are wearing goggles
22:09:57 <int-e> one, scientist ... mumble ... A, I guess.
22:11:37 <Phantom_Hoover> i note that what-if is now using xkcd comics as a reference
22:12:02 <int-e> has what-if become interesting again?
22:13:14 <Phantom_Hoover> where 'interestng' = 'describes apocalyptically ridiculous events qualitatively, rather than just working out how much milk it'd take to drown all the rabbits'
22:22:27 -!- edwardk has quit (Quit: Computer has gone to sleep.).
22:22:34 <int-e> oh, this *is* cute, "Have you ever tried to make a tower of peanut butter?"
22:31:32 -!- atehwa has quit (Ping timeout: 258 seconds).
22:33:29 <zzo38> Do you know an answer of my question about a random selection?
22:33:40 -!- edwardk has joined.
22:33:40 -!- atehwa has joined.
22:34:12 <Bike> yes. the answer is yes.
22:35:47 <zzo38> If the items have different weights can I do for example, 5/5, 4/9, 3/12, 2/14, 1/15, etc?
22:36:28 <Bike> You can do whatever you dream.
22:36:57 <zzo38> I know I can obviously program that in, but I meant, would it work OK?
22:37:29 <oerjan> zzo38: i'm pretty sure i've seen your 1/n algorithm before.
22:37:30 <Bike> i don't know what your criteria for "OK" are, so probably.
22:37:46 <zzo38> Where have you seen it?
22:39:13 <oerjan> for one thing, it seems implied in the fisher-yates shuffle.
22:39:37 <oerjan> (if you use it only for finding the first element, say)
22:40:34 <zzo38> I am only searching for a single element of the list, although they may have different weights (possibly even zero) and the weights may vary.
22:41:35 <oerjan> i'm pretty sure you can adapt it to use weights.
22:42:15 <oerjan> basically, switch in a number with the probability of it + the remaining numbers.
22:42:36 <int-e> right. just calculate the probability of the current item w.r.t. the items seen so far, and replace the item you have by the current item with that probability.
22:42:51 <zzo38> Would the way I specified work?
22:43:15 <oerjan> no, not the probability of the current item, the probability of the current item _or any later one_
22:43:40 <fizzie> I'm pretty sure I've seen the (unweighted) 1/n approach, too.
22:43:53 <int-e> > let xs = [5%5,4%9,3%12,2%14,1%15] in zipWith (/) xs (scanl1 (+) xs)
22:43:55 <lambdabot> [1 % 1,4 % 13,9 % 61,36 % 463,84 % 2399]
22:44:35 <int-e> those are your probabilities for discarding the previous item and replacing it by the current one.
22:44:45 <int-e> > let xs = [1%1,1,1,1,1,1,1] in zipWith (/) xs (scanl1 (+) xs)
22:44:46 <lambdabot> [1 % 1,1 % 2,1 % 3,1 % 4,1 % 5,1 % 6,1 % 7]
22:45:17 <zzo38> No, I mean the probability to discard and replace with the current one are 5/5, 4/9, 3/12, 2/14, 1/15.
22:45:29 <fizzie> http://en.wikipedia.org/wiki/Reservoir_sampling
22:45:35 <oerjan> argh brain closing down
22:46:26 <zzo38> (If the weight is 5,4,3,2,1)
22:46:35 <zzo38> Note, the weights will not be known ahead of time
22:46:47 <int-e> > let ys = zipWith (/) [5%5,4%9,3%12,2%14,1%15] (scanl (+) 1 ys) in ys
22:46:49 <lambdabot> [1 % 1,2 % 9,9 % 80,720 % 11767,564816 % 20298727]
22:47:14 <oerjan> hm i think int-e is about right, anyway
22:48:20 <oerjan> although there might be an off by one error or not in there
22:48:54 <zzo38> Maybe I have made several kind of mistake, I don't know
22:48:55 <oerjan> you don't want scanl (+) 1 ys, just scanl1 (+) ys
22:49:19 <oerjan> i think the first was right.
22:51:11 <int-e> > let ys = scanl (*) 5 (map (\p -> 1/(1-p)) $ tail [5%5,4%9,3%12,2%14,1%15]) in zipWith (-) ys (0:ys)
22:51:19 <oerjan> i sense a small complication if an initial segment has weights 0, though.
22:51:36 <int-e> (that was a bit tricky to get right)
22:52:01 <oerjan> oh you were calculating in reverse
22:52:01 <zzo38> Yes clearly such thing is then wrong, but based on the algorithm actually in use it probably isn't a problem if the weight is 0.
22:52:26 <oerjan> zzo38: well you just need to handle division by zero correctly.
22:52:48 <zzo38> I am not actually doing any division to generate the random numbers, though.
22:52:56 <oerjan> or just skipping the initial elements, though.
22:52:59 <int-e> oerjan: I was verifying zzo38's calculation, yes. (and cheating a bit, too, see the hard-coded 5)
22:53:10 <oerjan> zzo38: well you need division if you are starting from a list of weights
22:53:25 <oerjan> to get the probabilities
22:53:29 <zzo38> The list of weights is not accessed ahead of time, though.
22:53:41 <zzo38> I have to do it without accessing the list ahead of time.
22:54:15 <oerjan> right, so does this, i was wrong about the idea of looking at later weights, you need the sum of the previous ones like int-e's scanl1 (+)
22:54:34 <oerjan> which does not look ahead.
22:54:46 <zzo38> Sum of previous weights (including the current weight) is what I did.
22:56:30 <oerjan> and i guess you don't need division if you are generating a random number in [0..sum of weights-1]
22:56:49 <zzo38> Yes, that is what I am doing.
22:56:50 <oerjan> although you still need to handle the sum being 0
22:57:29 <oerjan> well then, this seems to work
22:58:21 <zzo38> The algorithm I use is to generate however many bits of random is needed, and then if the result is too high, try again until it is in range. For example to make 0 to 99, make 7-bits number, and if it exceed 99 then try again.
22:58:30 <int-e> and as an exercise you can process the list in chunks
22:59:00 <int-e> zzo38: ah you shouldn't discard all of the number
23:00:03 <int-e> zzo38: instead, note that if a random 7 bit number n exceeds 99, then n-100 is uniformely distributed between 0 and 27 (inclusively).
23:00:31 <zzo38> sqlite3_result_text(cxt,sqlite3_db_filename(sqlite3_context_db_handle(cxt),sqlite3_value_text(*argv)),-1,SQLITE_TRANSIENT);
23:00:32 <int-e> adding another two bits to that gives you something uniform in 0..103.
23:01:07 <zzo38> Wouldn't doing alll that stuff just make it more confusing?
23:01:09 <int-e> It's just an optimisation, of course, but it's pretty cheap to implement.
23:01:53 -!- impomatic has quit (Ping timeout: 240 seconds).
23:03:49 <int-e> http://sprunge.us/ZEhj?c
23:04:37 <int-e> while (u < n) -> for ( ; u < n; u <<= 1)
23:06:14 <int-e> And actually it's safe to replace that u < n by u <= n. (The invariant that the code maintains is that v is uniform in 0..u-1)
23:06:58 -!- tertu has joined.
23:09:46 <int-e> so trying again: http://sprunge.us/iWjS?c
23:14:24 <zzo38> Prove it better than the one I did.
23:19:51 <int-e> for 0..99, yours needs 8.96 random bits on average (7 * 128/100), mine needs 7.48(36) (7 + 28/128 * (2 + 4/104 * 5) * 1/(1-28/128 * 4/104))
23:24:19 <int-e> And if you keep the (v,r) state around rather than clearing it every time, you can get to the optimal log_2(100) random bits average.
23:24:22 <zzo38> I don't quite see how it would work though, but I can see how those how many it need in average
23:24:58 <zzo38> Also, the program I am using can only generate eight bits at a time, and I have it generating sixty-four at a time (due to some other reasons)
23:28:03 <int-e> Actually even that way it still waste many bits of randomness. (you're using that number 0..99 to make a binary choice, say between 0..41 and 42..99. You can actually add the remaining uniform 0..41 or 0..57 (99-42) back into the (v,r) state. These ideas come from arithmetic coding in data compression.
23:28:16 -!- edwardk has quit (Quit: Computer has gone to sleep.).
23:29:55 -!- edwardk has joined.
23:30:08 <int-e> To add a random number x in [0..n-1] to (v,r), the calculation is r = r*n (we get n times as many distinct values) and v = v*n + x. One has to be a bit careful about overflows when doing that though.
23:32:00 <int-e> Sorry I made a mistage above; we won't get to log_2(100) average this way.
23:32:09 <int-e> I also cannot type.
23:32:15 -!- Patashu has joined.
23:36:12 <zzo38> I can see what you mean now about binary choice
23:38:40 <zzo38> But I don't think the random number generator is quite the slowest thing in the program, and adding too many things might confuse it (and, sometimes may slow it down due to many instructions, regardless what the O(...) is), rather than making a simple and clearly programming
23:39:17 <zzo38> But is good you make such idea anyways, to study how much it could be done, in case it can make some of these improvements later if it seem it would be help much.
23:42:38 -!- nooodl has quit (Ping timeout: 240 seconds).
23:53:00 <Taneb> Fiora hasn't been here for a while, did something happen?
23:53:20 -!- mekeor` has quit (Ping timeout: 255 seconds).
23:54:46 <kmc> she left because of some homophobic comments
23:56:13 <kmc> if you particularly want to know, read the logs
23:57:38 <shachaf> sounds like a good reason to leave a channel
23:59:31 -!- yorick has quit (Remote host closed the connection).