←2014-11-05 2014-11-06 2014-11-07→ ↑2014 ↑all
00:00:16 -!- coppro has changed nick to NHcoppro.
00:00:56 -!- NHcoppro has changed nick to coppro.
00:04:18 -!- boily has joined.
00:10:23 -!- Sgeo has joined.
00:14:03 -!- Sgeo_ has quit (Ping timeout: 245 seconds).
00:14:17 -!- Sgeo_ has joined.
00:16:58 -!- Sgeo has quit (Ping timeout: 245 seconds).
00:19:19 -!- mihow has quit (Quit: mihow).
00:25:56 -!- Sgeo has joined.
00:27:15 -!- Sgeo_ has quit (Ping timeout: 244 seconds).
00:32:22 -!- Sgeo_ has joined.
00:36:08 -!- Sgeo has quit (Ping timeout: 245 seconds).
00:36:58 -!- Sgeo_ has quit (Ping timeout: 245 seconds).
01:07:54 -!- Bicyclidine has quit (Ping timeout: 256 seconds).
01:15:43 -!- Sgeo has joined.
01:18:41 <Sgeo> Let's see if I can stay connectedc
01:20:16 -!- nys has joined.
01:20:22 <HackEgo> [wiki] [[Subleq]] http://esolangs.org/w/index.php?diff=40786&oldid=40779 * Rdebath * (-115) You only need to say it once and English uses parentheses not square brackets.
01:28:29 -!- vanila has quit (Quit: Leaving).
01:36:52 <boily> Sgeo: so far so good.
01:44:21 -!- boily has quit (Quit: ABSTRACT CHICKEN).
01:45:05 <oerjan> darn and here i switched to the irc window just to see what concrete chicken message boily left with
01:45:07 <HackEgo> [wiki] [[Subleq]] http://esolangs.org/w/index.php?diff=40787&oldid=40786 * Rdebath * (+74) Move editing notes to a comment.
01:47:51 <Sgeo> How (im)practical is it to write code directly in LLVM IL assembly?
01:48:25 <Sgeo> *IR, which I was going to write but decided IL made more sense
01:48:57 <oerjan> just pletend you'le chinese
01:50:22 <Sgeo> I just want a good language to do another Braintrust implementation in which it's sufficiently comfortable to include a lightweight compiler for the language
01:50:36 <Sgeo> e.g. a statically-linked-in GHC in Haskell except that that's probably heavyweight
02:03:36 <Sgeo> Is rustc as a statically includable library plausible?
02:05:18 -!- AndoDaan has joined.
02:05:59 -!- AndoDaan has quit (Client Quit).
02:07:30 -!- AndoDaan_ has quit (Ping timeout: 272 seconds).
02:20:16 <oerjan> Sgeo: planning the Braint compiler?
02:20:33 <Sgeo> yes
02:21:17 <oerjan> if only kmc were around these days.
02:21:41 <oerjan> well he's on the network.
02:40:50 -!- adu has joined.
02:57:17 -!- paul2520 has quit (Quit: WeeChat 1.0).
03:05:16 -!- paul2520 has joined.
03:05:16 -!- paul2520 has quit (Changing host).
03:05:16 -!- paul2520 has joined.
03:15:47 -!- oerjan has quit (Quit: leaving).
03:23:16 -!- AndoDaan has joined.
03:29:13 -!- nys has quit (Quit: quit).
03:41:36 -!- ^v has quit (Quit: Leaving).
04:04:36 -!- Phantom_Hoover has quit (Ping timeout: 272 seconds).
04:06:15 -!- password2 has joined.
04:20:24 <tromp> what is braint?
04:26:15 -!- password2 has quit (Ping timeout: 256 seconds).
04:27:07 -!- lambdabot has quit (Ping timeout: 255 seconds).
04:32:54 -!- lambdabot has joined.
04:39:14 -!- password2 has joined.
05:01:28 -!- ^v has joined.
05:11:19 -!- adu has quit (Quit: adu).
05:12:47 <int-e> meh, "Encryption".
05:21:21 <Bike> more like shitcryption, imo.
05:22:45 <int-e> so now to do it in another language, hmm.
05:35:10 <HackEgo> [wiki] [[CA-1]] N http://esolangs.org/w/index.php?oldid=40788 * Notjohnconway * (+154) Created page with "Under construction n shit yo '''CA-1''' is the name of an esoteric programming language and a set of related programs, notably CA-1 Assembler. ==Example=="
05:35:50 <Sgeo> tromp: presumably accidental abbreviation of Braintrust
05:35:53 <Sgeo> That's what I assumed anyway
05:35:59 <lifthrasiir> Braintumor
05:39:38 -!- password2 has quit (Ping timeout: 244 seconds).
06:04:38 <Sgeo> minus = (x, y): x - y.
06:04:42 <Sgeo> http://perl11.org/potion/pamphlet.html
06:04:58 <Sgeo> Sending... a block to a table to make an anonymous function?
06:08:29 <Sgeo> Or maybe calling the table as a function with the block?
06:17:08 -!- shikhin has joined.
07:32:27 -!- ^v has quit (Ping timeout: 240 seconds).
07:34:01 <Lymia> TIL the 3DS uses CTR mode for pretty much all its encryption.
07:34:29 <Lymia> Apparently Nintendo doesn't learn from its mistakes.
07:35:44 -!- ^v has joined.
07:42:04 <fizzie> What's wrong with (no-IV-reuse) CTR mode?
07:46:48 -!- shikhout has joined.
07:48:53 <int-e> fizzie: hmm. two ideas: 1. if you xor two encrypted messages, you get the xor of two plaintexts, possibly revealing more than intended. 2. if you don't bother to authenticate decrypted data, then the attacker can easily modify individual bytes of the plaintext.
07:49:24 <fizzie> int-e: I did say no IV reuse.
07:49:32 -!- shikhin has quit (Ping timeout: 245 seconds).
07:49:39 <fizzie> int-e: You don't get the xor of two plaintexts unless you do it stupidly.
07:49:42 <int-e> the first one should be avoided by the nonce, right. the second one is still a potential issue
07:50:08 <fizzie> I guess that's true; it's not an authenticated mode.
07:53:27 <fizzie> So maybe it's hard to recommend over, say, GCM. But it's not like it's pretending to be authenticated encryption, either.
07:56:34 <int-e> funny, this is the first time I run into a problem with the golf.shinh.org system being 32 bits.
07:57:08 <int-e> ("problem" meaning I have to spend extra characters on making the program work)
07:57:43 -!- yorick has quit (Read error: Connection reset by peer).
08:00:02 -!- yorick_ has joined.
08:02:07 <quintopia> fizzie: what algorithm do you use to generate cladograms?
08:08:15 <fizzie> I don't, really. I mean, that's more of a bioinformaticist thing. Though my last set of browser tabs (which I lost somehow) had a very nice-looking thing for that -- I just don't quite know how I could relocate it.
08:10:10 <elliott> int-e: needing a larger integer type?
08:23:01 <int-e> elliott: I'm using Haskell, so usually things are fine because stuff defaults to Integer.
08:25:08 <elliott> right
08:25:26 <elliott> dealing with gigabytes of data for golf? :p
08:25:45 -!- shikhout has changed nick to shikhin.
08:26:07 <int-e> not really.
08:27:20 <fizzie> quintopia: I think http://mrbayes.sourceforge.net/ was the page I had open, maybe.
08:32:09 <elliott> I admire the postscript golfers the most
08:38:09 <Sgeo> A clean Cy&H? What is the world coming to?
08:38:34 <elliott> jesus christ, you read Cy&H in 2014?
08:41:01 <Sgeo> At least I don't read that Station comic anymore
08:41:21 <Sgeo> Station V3
08:42:34 <Sgeo> ... I'm reading it again
08:43:23 <Sgeo> The list of storylines hasn't been updated since 2009... I'm assuming because there are no storylines
08:44:10 <Sgeo> There's a comment from two years ago complaining about it
08:44:34 <Sgeo> http://www.stationv3.com/d/20091213.html#comment-627106160
08:52:06 <Sgeo> http://www.stationv3.com/d/20050325.html
08:58:39 <Sgeo> http://www.stationv3.com/d/20050507.html
08:58:54 <Sgeo> 'business as usual' describes the comic, I think
09:02:08 <b_jonas> Sgeo: it had "storylines"? I thought it's just disconnected strips
09:04:45 -!- shikhin has quit (Remote host closed the connection).
09:06:37 -!- shikhin has joined.
09:08:52 -!- oerjan has joined.
09:09:55 <oerjan> <Sgeo> tromp: presumably accidental abbreviation of Braintrust <-- *swats Sgeo for missing the obvious pun -----###
09:10:11 <oerjan> also i cannot sleep :(
09:10:47 -!- AndoDaan has quit (Ping timeout: 265 seconds).
09:12:33 <oerjan> @ping
09:12:33 <lambdabot> pong
09:12:54 <oerjan> hm web seems molassey
09:16:27 <oerjan> looks like a router reboot helped
09:18:18 <oerjan> or not.
09:25:27 -!- MDude has quit (Ping timeout: 250 seconds).
09:28:08 <mroman> `js
09:28:10 <HackEgo> Invalid option "-e" \ Usage: java org.mozilla.javascript.tools.shell.Main [options...] [files] \ Valid options are: \ -?, -help Displays help messages. \ -w Enable warnings. \ -version 100|110|120|130|140|150|160|170 \ Set a specific language version. \ -opt [-1|0-9] Set optimizat
09:28:19 <mroman> `js "print(9)"
09:28:21 <HackEgo> No output.
09:28:24 <mroman> `js print(9)
09:28:26 <HackEgo> 9
09:28:44 <mroman> `js print(['10','10','10'].map(parseInt))
09:28:46 <HackEgo> 10,NaN,2
09:28:50 <mroman> lol wtf
09:31:09 <FireFly> Hehe
09:31:37 <FireFly> the explanation to that is that `map` passes in (value, index, array) and parseInt takes an optional second argument (the radix)
09:31:53 <FireFly> It's not a *good* reason for that, but it is a reason
09:32:12 <FireFly> `js print(['10','10','10'].map(Math.floor))
09:32:14 <HackEgo> 10,10,10
09:32:23 <FireFly> Always fun to abuse weak typing
09:32:58 <FireFly> `js print(parseInt('0x10', 0))
09:33:00 <HackEgo> 16
09:33:16 <FireFly> All right, so it treats a second argument of 0 the same as no second argument
09:34:25 <oerjan> `js print(parseInt('1111', 1))
09:34:27 <HackEgo> NaN
09:34:33 <oerjan> DISAPPOINTING
09:34:47 <oerjan> `js print(parseInt('0000', 1))
09:34:48 <HackEgo> NaN
09:36:10 -!- AndoDaan has joined.
10:15:03 <mroman> `js print(['10','10','10','10','10']).map(parseInt)
10:15:05 <HackEgo> 10,10,10,10,10 \ js: uncaught JavaScript runtime exception: TypeError: Cannot call method "map" of undefined
10:15:14 <mroman> `js print(['10','10','10','10','10'].map(parseInt))
10:15:16 <HackEgo> 10,NaN,2,3,4
10:15:19 -!- Phantom_Hoover has joined.
10:15:21 <mroman> although this is kinda funny
10:17:55 <oerjan> stackoverflow is no fun with molassey web :(
10:30:40 <elliott> oerjan: well 1 isn't a positional base
10:30:53 <elliott> okay I guess it is, just one that can only represent 0
10:31:40 <oerjan> thus my second test
10:36:44 -!- atehwa has joined.
10:40:43 -!- yorick_ has changed nick to yorick.
10:42:05 <mroman> `js [] + []
10:42:07 <HackEgo> No output.
10:42:11 <mroman> `js print([] + [])
10:42:12 <HackEgo> No output.
10:42:15 <mroman> `js print([] + []);
10:42:16 <HackEgo> No output.
10:42:21 <mroman> `js print({} + {});
10:42:23 <HackEgo> ​[object Object][object Object]
10:42:53 <mroman> I guess javascript can be a honorary esoteric language
10:54:00 <b_jonas> mroman: note that some of the apparent wierdness is that in some contexts javascript is like perl so {}+[] is parsed such that {} is an empty C-style statement block and then +[] is an expression with unary plus
10:54:22 <b_jonas> mroman: so sometimes you think you've printed the expression {}+[] but actually you print the expression +[]
10:58:06 <mroman> isn't {} an object?
10:58:10 <mroman> `js print({} + []);
10:58:11 <HackEgo> ​[object Object]
10:58:18 <mroman> `js print([] + {} );
10:58:20 <HackEgo> ​[object Object]
10:58:27 <mroman> `js print({'abc':9} + []);
10:58:29 <HackEgo> ​[object Object]
10:58:38 <mroman> `js print({'abc':9} + {});
10:58:40 <HackEgo> ​[object Object][object Object]
10:58:43 <b_jonas> mroman: it's an object if it occurs in expression context like in print({} + [])
10:58:57 <oerjan> `js print(({})+[])
10:58:59 <HackEgo> ​[object Object]
10:59:04 <b_jonas> mroman: but some people have tried to just type {} + [] in a firefox javascript console or something, in whcih case it was a statement
10:59:34 <b_jonas> `js print(eval("{}+[]"))
10:59:36 <HackEgo> 0
10:59:42 <b_jonas> `js print(eval("({}+[])"))
10:59:44 <HackEgo> ​[object Object]
10:59:47 <b_jonas> something like that
11:21:32 -!- MDude has joined.
11:25:08 <fizzie> "-- where score is a composite acoustic/language model score from the recognizer, on the bytelog scale. (A bytelog is a logarithm to base 1.0001, divided by 1024 and rounded to an integer.)"
11:30:57 -!- drdanmaku has quit (Quit: Connection closed for inactivity).
11:40:58 -!- boily has joined.
11:42:31 -!- hjulle has joined.
11:51:53 <elliott> fizzie: I...
11:52:58 <oerjan> > 1.0001**(1/1024)
11:53:00 <lambdabot> 1.0000000976513723
11:53:15 <oerjan> hm wrong way around?
11:53:20 <oerjan> > 1.0001**(1024)
11:53:22 <lambdabot> 1.107820842039981
11:56:12 <fizzie> > round . (*1024) . logBase 10 <$> [0.00001, 0.001, 0.1, 1, 10, 100, 100000, 1000000000]
11:56:14 <lambdabot> [-5120,-3072,-1024,0,1024,2048,5120,9216]
11:56:31 <fizzie> I think they must have meant "multiplied by 1024", otherwise that doesn't make all that much sense.
11:56:43 <fizzie> Oh, I forgot the 1.0001.
11:56:51 <fizzie> > round . (*1) . logBase 1.0001 <$> [0.00001, 0.001, 0.1, 1, 10, 100, 100000, 1000000000]
11:56:52 <lambdabot> [-115135,-69081,-23027,0,23027,46054,115135,207243]
11:56:58 <fizzie> > round . (/1024) . logBase 1.0001 <$> [0.00001, 0.001, 0.1, 1, 10, 100, 100000, 1000000000]
11:57:00 <lambdabot> [-112,-67,-22,0,22,45,112,202]
11:57:07 <fizzie> Maybe it is that, indeed.
11:57:44 <oerjan> that 1.107820842039981 is the scale factor which increases it by 1
12:03:57 <J_Arcane> This is an utterly amazing piece, about computer history, human networking, and reliving the past. https://medium.com/message/networks-without-networks-7644933a3100
12:04:46 -!- AndoDaan has quit (Ping timeout: 264 seconds).
12:07:14 -!- Frooxius has quit (Quit: *bubbles away*).
12:21:28 -!- boily has quit (Quit: DICHOTOMIC CHICKEN).
12:47:52 -!- Phantom_Hoover has quit (Ping timeout: 245 seconds).
13:33:19 * oerjan concludes that the `civd subset of unlambda is not TC, assuming he's calculated correctly
13:47:02 -!- shikhout has joined.
13:50:26 -!- shikhin has quit (Ping timeout: 260 seconds).
14:03:23 -!- atriq has joined.
14:05:23 -!- drdanmaku has joined.
14:05:41 <atriq> Today I finally learned officially about group homomorphisms
14:07:42 <atriq> Is the one-element group a terminal object in that category?
14:09:06 <atriq> (the category of group homomorphisms)
14:11:57 -!- MDream has joined.
14:12:36 -!- atriq has changed nick to Baneet.
14:12:41 -!- Baneet has changed nick to Banet.
14:12:55 -!- Banet has changed nick to atriq.
14:14:59 -!- MDude has quit (Ping timeout: 264 seconds).
14:24:29 <oerjan> atriq: yep
14:24:40 <atriq> Yay
14:25:03 <oerjan> also initial hth
14:25:13 <atriq> Oh?
14:25:29 <atriq> Oh yeah, that makes sense
14:25:58 <atriq> Because it's a subgroup of any other group, the only homomorphism 1 -> G is inclusion?
14:26:06 -!- Phantom_Hoover has joined.
14:26:10 <oerjan> right
14:26:14 <oerjan> well
14:26:34 <atriq> Isomorphic to a subgroup
14:26:38 <oerjan> right
14:30:51 <Phantom_Hoover> hey oerjan what's the fundamental group of the configuration space of 3 distinct unlabelled points on a circle
14:31:09 <Phantom_Hoover> i'm fairly sure it's Z but i can't get a grip on proving it properly
14:31:12 <oerjan> i know most of those words.
14:31:28 <oerjan> but not "configuration space".
14:32:26 <Phantom_Hoover> as in, the space of arrangements of 3 points on a circle
14:32:34 <oerjan> oh hm
14:32:41 <oerjan> so 3-element subsets essentially
14:33:04 <oerjan> ok you can rotate around the circle and get back
14:33:17 <Phantom_Hoover> it's T^3 less the planes where two coördinates are the same and quotiented by permutations of the coördinates
14:34:10 <Phantom_Hoover> well basically my argument is that given some basepoint configuration, the points can't pass so they're always in the same order on the circle
14:34:21 <oerjan> ok i have an argument.
14:34:31 <Phantom_Hoover> so up to homotopy all you can do is move each point to the next one clockwise or anticlockwise
14:34:33 <oerjan> or wait
14:34:49 <oerjan> yes i do.
14:35:47 <oerjan> assume a closed path in this space. pick one of the arbitrary points, since they never collide you can follow it.
14:36:32 <oerjan> now adjust the path so that that point is kept still.
14:36:43 <oerjan> then at the end, do a rotation to the right place.
14:36:52 <oerjan> (if it doesn't end up as itself.)
14:37:08 <oerjan> *adjust the path by rotation
14:38:19 <Phantom_Hoover> so you're ending up with Z as well?
14:38:37 <oerjan> i think so.
14:39:27 -!- Sgeo has quit (Read error: Connection reset by peer).
14:39:43 -!- Sgeo has joined.
14:40:04 <oerjan> you can deform your path into two parts, in one one of the point is kept still while the others do their complicated dance
14:40:13 <oerjan> in the other all points are simply rotated.
14:40:24 <oerjan> equally.
14:40:35 <oerjan> *in one part, one
14:40:43 <oerjan> *points
14:41:28 <oerjan> or even better
14:41:56 <oerjan> deform it into three parts, such that the rotation is in the middle, and the points are distributed equilaterally during it
14:42:13 -!- Sprocklem has quit (Ping timeout: 256 seconds).
14:44:29 <oerjan> has to be Z
14:46:58 <oerjan> although hm
14:48:08 <oerjan> i'm thinking there might be a more elegant method, which maps configurations to equilateral ones
14:48:26 <oerjan> as a retraction
14:49:44 <oerjan> right, if the whole configuration space retracts to the equilateral ones, it would be very pretty
14:52:27 <Phantom_Hoover> the elegant method probably involves covering maps or something
14:52:48 <oerjan> well sure
14:53:18 <oerjan> in that case, probably the covering map from the _labeled_ version?
14:54:14 <oerjan> if you know how to solve that. i'm definitely rusty.
14:55:11 <oerjan> hm the first question is whether you can get a retract at all, deformation or not
14:56:17 <oerjan> what happens if you map to the configuration given by the _midpoints_ between the original ones, then iterate and take the limit?
14:56:29 <oerjan> does that tend toward equilateral
14:56:53 <oerjan> also you can obviously do each step as a deformation, i think
14:57:44 -!- atriq has quit (Quit: leaving).
14:57:48 <oerjan> ok the new arc lengths are the averages of the originals
14:59:11 <oerjan> > iterate (\(x,y,z) -> ((y+z)/(x+y+z),(x+z)/(x+y+z),(x+y)/(x+y+z))) (0.1,0.9,0.8) !! 10
14:59:12 <lambdabot> (0.6661241319444444,0.6669921875,0.6668836805555556)
14:59:17 <oerjan> looks good
14:59:31 <oerjan> i think you can get a deformation retract that way
15:00:32 <oerjan> (i divided by x+y+z rather than 2 to keep the sum stable)
15:01:02 <oerjan> oh hm
15:01:30 <oerjan> it may not always be obvious which way to turn
15:01:43 <oerjan> but perhaps that will fix itself after the first few steps
15:02:02 <oerjan> or wait, duh
15:02:20 <oerjan> you have to turn the way that doesn't cross the other points
15:02:55 <oerjan> this method is clearly symmetric.
15:03:46 <oerjan> there should be a simple formula for the point you get in the limit, then you can just move straight to it
15:04:37 <oerjan> oh there might be problems with continuity
15:06:20 -!- `^_^v has joined.
15:07:20 <oerjan> oh hm
15:07:52 <oerjan> it's just the square root of the product in the complex unit circle
15:08:16 <oerjan> although you need to choose the right one...
15:09:52 <oerjan> sqrt(yz), sqrt(xz), sqrt(xy), then you iterate to x^(1/2)y^(1/4)z^(1/4) etc.
15:11:27 <oerjan> hm that's obviously going to go like x^m y^n z^n where m+2n=1 and m and n get closer
15:11:49 <oerjan> so it's in some sense (xyz)^1/3
15:12:47 <oerjan> but with the question, how do you choose the _right_ third root to retract to
15:13:02 <oerjan> or well
15:13:13 <oerjan> they're all the same configuration, aren't they
15:13:49 <oerjan> but you still want to do it as a deformation
15:14:40 <oerjan> x^(1-2*t/3)y^(t/3)z^(t/3)
15:15:26 <oerjan> oh hm
15:16:05 <oerjan> if you unroll the circle this is just finding the average of a point with the two on each side
15:16:27 <oerjan> (taking logarithm)
15:20:46 <oerjan> Phantom_Hoover: ok let's say the three points are e^(ix), e^(iy), e^(iz), with x < y < z < x+2pi. then map e^(iy) to e^(i(2y+x+z)/3*t). there's your retract as it works on that point.
15:20:55 <oerjan> *deformation retract
15:21:09 <oerjan> er
15:21:17 <oerjan> not the right formula
15:22:09 <oerjan> * e^i(y + (x+z-2y)*t/3) is better
15:23:36 <oerjan> so basically, the retract works by averaging angles with the neighboring points
15:24:49 <oerjan> where some of the intuition here comes from thinking of the points as lying on R and repeating every 2pi
15:25:19 <oerjan> which is of course an isomorphic representation
15:28:14 <oerjan> Phantom_Hoover: well i think this is pretty elegant, anyway
15:28:31 <oerjan> and geometric
15:28:44 <Phantom_Hoover> so... what, you've found a retraction onto a point?
15:29:00 <oerjan> no, onto the subspace of equilateral configurations
15:29:22 <oerjan> which is obviously homeomorphic to the circle
15:31:42 -!- Sprocklem has joined.
15:33:10 <Phantom_Hoover> ah
15:54:13 -!- nys has joined.
16:07:54 -!- Phantom_Hoover has quit (Ping timeout: 265 seconds).
16:28:42 <Taneb> Man, I wish I thought to ask oerjan for help with my homework.
16:34:12 <oerjan> :D
17:00:51 -!- MDream has changed nick to MDude.
17:06:44 -!- Sprocklem has quit (Ping timeout: 255 seconds).
17:11:09 <mroman> `learn_append oerjan he also does your homework .
17:12:47 -!- oerjan has quit (Quit: UNRELIABLY).
17:18:54 -!- MDream has joined.
17:21:14 -!- MDude has quit (Ping timeout: 258 seconds).
17:27:25 -!- MoALTz has joined.
17:29:04 -!- shikhout has changed nick to shikhin.
17:40:52 <fizzie> Every home should have an oerjan, to take care of those household math problems that arise.
17:44:53 <quintopia> fizzie: but you made cladograms for bfjoust programs once
17:48:38 <fizzie> If you want to call them that, though from my point of view they were just dendrograms about a hierarchical clustering directly in a simple feature space. A "cladogram" sounds much fancier. Maybe that's a pretty subtle distinction.
17:50:15 <mroman> what's a dendrogram?
17:50:43 <fizzie> mroman: http://zem.fi/egostats/plot_cluster.png that kind of thing
17:51:02 <fizzie> At any rate, those (IIRC) were just your regular bottom-up greedy hierarchical clustering, with a Manhattan distance metric between points and the average-of-all-pairwise-distances metric for clusters.
17:51:54 <fizzie> mroman: The position of the vertical "bar" denotes the distance between the things it connects.
17:57:22 -!- MDream has changed nick to MDude.
17:59:47 -!- Sprocklem has joined.
18:03:34 <mroman> distance of what?
18:03:38 <mroman> points? ranking?
18:12:11 <mroman> http://codepad.org/YPYwJ9lh
18:12:21 <mroman> interesting sketch i got laying around here
18:13:01 <mroman> except that int doesn't form a multiplicative group
18:26:03 <HackEgo> [wiki] [[Special:Log/newusers]] create * Icedvariables * New user account
18:26:49 <Taneb> My plan to kick ass at BF Joust has come to a problem
18:26:55 <Taneb> It's a big search space to brute-force
18:30:08 -!- AnotherTest has joined.
18:30:53 <Taneb> As n -> + inf, number-of-BF-Joust-programs-with-length-n approaches k^n where 6 < k < 7, I think
18:33:58 <Taneb> number-of-BF-Joust-programs-with-length 10 is ~1.5 million
18:34:23 <Taneb> s/10/8/
18:34:38 <int-e> that's without syntactic sugar?
18:34:46 <Taneb> Yeah
18:35:03 <Taneb> And I don't think a length 8 BF Joust program is going to do much
18:39:04 <Taneb> ...I'm listening to one of my lecturers using the State monad as an analogy
18:39:45 <Taneb> I'm not sure what for
18:39:52 <Taneb> A minute a go he was talking about hanggliders
18:40:01 <int-e> hmm, 919139 is not 1.5 million.
18:41:29 <Taneb> I definitely get 1542195
18:41:29 <int-e> > let f n | n<0 = 0 | otherwise = f' !! n; f' = 1 : [6 * f (n-1) + sum [f k * f (n-2-k) | k <- [0..n-2]] | n <- [1..]] in f 8
18:41:31 <lambdabot> 3172478
18:41:35 <int-e> err
18:41:45 <int-e> > let f n | n<0 = 0 | otherwise = f' !! n; f' = 1 : [5 * f (n-1) + sum [f k * f (n-2-k) | k <- [0..n-2]] | n <- [1..]] in f 8
18:41:47 <lambdabot> 919139
18:42:12 <int-e> > let f n | n<0 = 0 | otherwise = f' !! n; f' = 1 : [5 * f (n-1) + sum [f k * f (n-2-k) | k <- [0..n-2]] | n <- [1..]] in take f' -- 0..
18:42:14 <lambdabot> Couldn't match expected type ‘GHC.Types.Int’
18:42:14 <lambdabot> with actual type ‘[a0]’
18:42:20 <int-e> > let f n | n<0 = 0 | otherwise = f' !! n; f' = 1 : [5 * f (n-1) + sum [f k * f (n-2-k) | k <- [0..n-2]] | n <- [1..]] in f' -- 0..
18:42:22 <lambdabot> [1,5,26,140,777,4425,25755,152675,919139,5606255,34578292,215322310,13519788...
18:42:45 <int-e> so how many do you get for length 4?
18:43:13 <Taneb> 931
18:44:22 <Taneb> > [p|p<-replicateM 4 syms,q 0 p] where q n ""=n==0;q n ('[':xs)=q(n+1)xs;q n (']':xs)=q(n-1)xs;q n (_:xs)=q n xs
18:44:24 <lambdabot> <hint>:1:32: parse error on input ‘where’
18:44:47 <Taneb> > let q n ""=n==0;q n ('[':xs)=q(n+1)xs;q n (']':xs)=q(n-1)xs;q n (_:xs)=q n xs in [p|p<-replicateM 4 syms,q 0 p]
18:44:48 <lambdabot> Not in scope: ‘syms’
18:44:56 <Taneb> > let q n ""=n==0;q n ('[':xs)=q(n+1)xs;q n (']':xs)=q(n-1)xs;q n (_:xs)=q n xs in [p|p<-replicateM 4 "+-<>.[]",q 0 p]
18:44:57 <lambdabot> ["++++","+++-","+++<","+++>","+++.","++-+","++--","++-<","++->","++-.","++<+...
18:45:04 <Taneb> > let q n ""=n==0;q n ('[':xs)=q(n+1)xs;q n (']':xs)=q(n-1)xs;q n (_:xs)=q n xs in length[p|p<-replicateM 4 "+-<>.[]",q 0 p]
18:45:05 <lambdabot> 931
18:45:11 <Taneb> > let q n ""=n==0;q n ('[':xs)=q(n+1)xs;q n (']':xs)=q(n-1)xs;q n (_:xs)=q n xs in length[p|p<-replicateM 8 "+-<>.[]",q 0 p]
18:45:15 <lambdabot> mueval-core: Time limit exceeded
18:45:29 <int-e> Taneb: you include things like ][
18:45:38 <Taneb> Fuck, I do
18:45:52 <tromp> you can also exclude +- and -+
18:46:08 <tromp> and <> and ><
18:46:08 <Taneb> tromp, I'm doing this stupidly naively
18:46:24 <Taneb> I can exclude anything starting with <, too
18:46:29 <int-e> tromp: those are valid though, ][ is invalid syntax
18:47:17 <Taneb> > let q n _|n<0=False;q n ""=n==0;q n ('[':xs)=q(n+1)xs;q n (']':xs)=q(n-1)xs;q n (_:xs)=q n xs in length[p|p<-replicateM 8 "+-<>.[]",q 0 p]
18:47:21 <lambdabot> mueval-core: Time limit exceeded
18:47:22 <Taneb> > let q n _|n<0=False;q n ""=n==0;q n ('[':xs)=q(n+1)xs;q n (']':xs)=q(n-1)xs;q n (_:xs)=q n xs in length[p|p<-replicateM 4 "+-<>.[]",q 0 p]
18:47:23 <lambdabot> 777
18:47:28 <Taneb> int-e, that the number you get?
18:47:37 <int-e> Taneb: yes, see above.
18:47:46 <tromp> jackpot!
18:50:50 <Taneb> OK, this makes it SLIGHTLY smaller search space
18:50:58 <Taneb> If I make tromp's improvements, less further
18:51:19 <Taneb> tromp, I think in the wrong place +- or >< could have an effect
18:51:22 <Taneb> in BF Joust
18:51:53 <tromp> sorry; i dont know BF Joust
18:52:21 <tromp> but i can see how they affect multithreaded BF :)
18:52:36 <int-e> Taneb: actually my code isn't so hard, it's based on the following observation: A program of length n is either one of +-<>. followed by a program of length n-1, or it's [ followed by a program of some length k, followed by ], followed by a program of length n-2-k. Hence a_n = 5*a_(n-1) + sum[k=0 to n-2, a_k * a_(n-2-k)].
18:53:18 <HackEgo> [wiki] [[Replace]] N http://esolangs.org/w/index.php?oldid=40789 * Icedvariables * (+874) A esoteric language based around regular expressions and self-modifying code
18:53:18 <int-e> (and a_0 = 1)
18:53:43 -!- password2 has joined.
18:57:26 <HackEgo> [wiki] [[Replace]] http://esolangs.org/w/index.php?diff=40790&oldid=40789 * Icedvariables * (+65) /* External resources */
19:01:41 <HackEgo> [wiki] [[Language list]] http://esolangs.org/w/index.php?diff=40791&oldid=40777 * Icedvariables * (+14) Added Replace
19:05:21 <fizzie> mroman: Manhattan distance between the (-1, 0, 1)-valued entire results against all N-2 opponents the two compared programs have in common, in that case.
19:06:02 <fizzie> mroman: There's also another similar plot using Euclidean distance for the concatenation of the tape heat-maps, possibly averaged somehow.
19:12:07 <int-e> fizzie: asumptotically the number should be O(7^n / n^1.5). (There are Catalan numbers in there, that's where the n^1.5 comes from. I say "should be" because I've only convinced myself, not proved it formally.)
19:12:16 <int-e> *asymptotically
19:13:36 <fizzie> Isn't that more to Taneb than me?
19:13:45 <int-e> ... yes. Sorry.
19:13:55 <fizzie> Serves me right for opening my virtu-mouth.
19:14:17 <int-e> I have this bug where I reply to the last non-int-e person who spoke.
19:16:57 -!- Sprocklem has quit (Quit: [).
19:18:34 -!- FreeFull has quit (Ping timeout: 244 seconds).
19:21:54 -!- FreeFull has joined.
19:22:11 -!- nys has quit (Ping timeout: 255 seconds).
19:34:14 -!- nys has joined.
19:47:55 -!- shikhout has joined.
19:51:06 -!- shikhin has quit (Ping timeout: 250 seconds).
19:59:03 <int-e> wee I've been running lambdabot for almost a year now.
20:01:13 <fizzie> Do you get some kind of a sticker or a commemorative plaque or something when you have?
20:04:06 <int-e> No, I get an invoice from the hosting provider.
20:04:39 <fizzie> I guess you can frame that, too.
20:04:48 <int-e> hehe
20:04:52 -!- ais523 has joined.
20:05:00 -!- ais523 has quit (Changing host).
20:05:01 -!- ais523 has joined.
20:05:25 <int-e> My fault, I *could* easily run the thing on a university server instead and nobody would mind.
20:07:01 -!- ais523 has left.
20:12:33 <int-e> Bah, web designers keep finding new ways of hiding the "logout" feature.
20:13:06 <int-e> This time, I had to click on "Hello, $name" which popped up a menu with a logout item.
20:22:25 -!- vyv has quit (Ping timeout: 245 seconds).
20:44:50 -!- password2 has quit (Ping timeout: 250 seconds).
20:55:33 <HackEgo> [wiki] [[CA-1]] http://esolangs.org/w/index.php?diff=40792&oldid=40788 * 50.115.194.66 * (+1263)
20:57:01 -!- password2 has joined.
21:01:13 <fizzie> int-e: They count on you being conditioned to do that by the "shutdown from the start button" thing.
21:01:35 <fizzie> Though maybe in this era that's obsolete knowledge too.
21:12:45 <int-e> fizzie: it doesn't look like a button at all. http://int-e.eu/~bf3/tmp/start.png
21:13:22 -!- AnotherTest has quit (Remote host closed the connection).
21:13:59 -!- AnneFrank has joined.
21:14:51 <AnneFrank> Kewww
21:15:13 <int-e> `relcome AnneFrank
21:15:14 <HackEgo> AnneFrank: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: <http://esolangs.org/>. (For the other kind of esoterica, try #esoteric on irc.dal.net.)
21:15:44 <AnneFrank> Oh god rainbow txt
21:16:44 <AnneFrank> ~hello
21:22:10 <int-e> `welcome AnneFrank
21:22:11 <HackEgo> AnneFrank: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: <http://esolangs.org/>. (For the other kind of esoterica, try #esoteric on irc.dal.net.)
21:22:20 <int-e> there, we have a plain version, too.
21:23:09 <AnneFrank> No i saw
21:23:18 <AnneFrank> It just surprised me thats all
21:24:51 <AnneFrank> Ugh hold on my name always ends up like this
21:24:54 -!- AnneFrank has quit (Quit: Page closed).
21:25:23 -!- CakeMeat has joined.
21:25:30 <CakeMeat> There
21:25:54 -!- password2 has quit (Ping timeout: 244 seconds).
21:28:53 <CakeMeat> Gah This thing is annoying
21:30:34 <CakeMeat> Reads through old logs
21:31:19 -!- `^_^v has changed nick to nycs.
21:33:05 <CakeMeat> Sooo Whats up
21:34:37 -!- shikhout has changed nick to shikhin.
21:39:13 -!- password2 has joined.
21:39:21 <CakeMeat> gnight
21:39:24 -!- CakeMeat has quit (Quit: Page closed).
21:51:59 <Gregor> AnneCakeMeatFrank certainly didn't give us much time.
21:52:43 <int-e> I would've been in time to answer "the sky" but decided not to.
22:13:52 <HackEgo> [wiki] [[StackStacks]] http://esolangs.org/w/index.php?diff=40793&oldid=40768 * Oj742 * (+570) /* Examples */ Added Quine
23:14:32 -!- oerjan has joined.
23:26:15 -!- Phantom_Hoover has joined.
23:28:36 -!- zzo38 has joined.
23:31:51 -!- nycs has quit (Ping timeout: 272 seconds).
23:32:46 -!- adu has joined.
23:51:20 -!- vanila has joined.
23:51:20 <vanila> hi
23:57:12 <elliott> hi
23:58:26 <oerjan>
23:59:06 <vanila> I want to learn how to write compilers
23:59:14 <vanila> do you have some advice
23:59:37 <b_jonas> vanila: read a good book on them
←2014-11-05 2014-11-06 2014-11-07→ ↑2014 ↑all