←2014-04-16 2014-04-17 2014-04-18→ ↑2014 ↑all
00:00:38 <boily> cromulent.
00:04:37 <madbr> kinda wonder how hard it is to design a simple cpu, video rendering chip or sound playing chip
00:06:42 <boily> well, sound playing is the easiest. a few simple wave generators, subtractive synthesis, and you're golden.
00:06:58 <madbr> hm
00:07:08 <boily> in fact, you could design one with only basic discrete components.
00:07:15 <boily> (and a few logic gates if you wanna be fancy.)
00:07:28 <madbr> doesn't the filter require multipliers?
00:07:54 <madbr> in some feedback configuration
00:08:39 <boily> op amps. I always forget op amps.
00:08:47 <madbr> in verilog?
00:08:52 <boily> eeeeeeh...
00:09:02 <boily> I think you have them. I'd be very surprised if you didn't.
00:09:32 <boily> (but then, my understanding of verilog is quite limited and backwards, as I had urging deadlines and ethereal teammates to manage...)
00:10:27 * boily wonders if xcircuit is still maintained... “I have fond memories of that buggy and crashy piece of software...”
00:11:39 <madbr> hm, the one I've played around with is HADES
00:11:45 <madbr> and that's digital only
00:12:19 <boily> xcircuit is only for drawing. it was quite nice, when it was running.
00:13:05 <boily> I know I used something from xilinx a long time ago.
00:14:37 <madbr> mhmm
00:15:29 <madbr> I guess you could tape together 2 or 3 oscillators, pass through a textbook filter and multiply for volume yeah
00:15:40 <madbr> then replicate 6 times for a whole synth
00:17:03 <madbr> not very efficient but easy to design
00:18:56 <boily> I once combined a lo- and a hi-pass filter to make a no-pass filter. the end result was... unusual.
00:20:35 <madbr> filters were never too interesting to implement in sound LSIs
00:21:10 <madbr> until they became so large that they could just spend the extra multipliers on them
00:21:36 <madbr> (wavetable-based chips such as the soundblaster awe-32)
00:24:52 -!- ^v has joined.
00:39:21 -!- Sprocklem has joined.
00:39:21 -!- Sprocklem has quit (Changing host).
00:39:21 -!- Sprocklem has joined.
00:43:31 -!- tromp has joined.
00:48:27 -!- Phantom_Hoover has quit (Quit: Leaving).
00:49:32 <Bike> http://rxnm.files.wordpress.com/2012/08/image_thumb12.png?w=887&h=566 Meanwhile, science progresses
00:51:28 <oerjan> are you trying to point out the irony of the article vs. the number in the top bar
00:51:44 <Bike> yes
00:51:50 <oerjan> good, good
00:52:02 <kmc> his impact factor is over 9,000
00:55:40 <oerjan> Bike: perhaps you should stay away from inhabited planets twh
00:55:56 <kmc> every planet we reach is dead
00:56:11 <Bike> i don't actually know how impact factor is calculated. could you have 9000? is there like a singularity somewhere
00:56:30 <kmc> what's gauss's impact factor
00:56:30 <Bike> maybe over 1000 or so means everyone on earth is constantly thinking about your paper on cockroach neuroendocrine cells
00:56:39 <boily> what is an impact factor? against what is it impacted?
00:57:03 <Bike> It's a measure of your science penis
00:57:03 <oerjan> impalas
00:57:15 <oerjan> also imps
00:57:40 <oerjan> impomatic shouldn't feel too safe, either
00:57:51 <Bike> oh it's actually pretty simple: a journal's impact factor is number of citations of papers / number of papers written
00:58:19 <Bike> "The impact factor is based on the arithmetic mean number of citations per paper, yet citation counts follow a Bradford distribution (i.e., a power law distribution) and therefore the arithmetic mean is a statistically inappropriate measure." scientists are so awesomely bad at stats
00:59:41 <boily> oerjan: so, if I unabashedly mapole impomatic, I can augment my Factor?
01:00:36 <Bike> Annual Review of Immunology had the highest in 2006, apparently
01:00:41 <oerjan> `factor 9000
01:00:42 <HackEgo> 9000: 2 2 2 3 3 5 5 5
01:01:22 <oerjan> boily: there's also a danger of both your factor and impomatic being smashed to little primes
01:02:33 <kmc> http://mushroomobserver.org/image/show_image/386349?q=1uvb6&size=full_size
01:02:53 <oerjan> Bike: clearly they should simply define the new Power Impact which takes this into account
01:02:55 <boily> impomatic: do you accept being a victim^Wvaluable contributor to a scientific experience of utmost importance?
01:04:43 <oerjan> kmc: i take it the flavor just rolls off your tongue
01:04:59 -!- nisstyre has joined.
01:06:11 <oerjan> douglass_: btw the nick thiotimoline is vacant here hth
01:06:31 <kmc> you can register multiple nicks to the same nickserv account
01:06:49 <Bike> oerjan: there's an elaboration where you weigh cites based on the citing journal's impact factor, like PageRank. it's fucking crazy
01:09:45 <oerjan> kmc: yes but thiotimoline hasn't been registered yet. nor has douglass_ fwiw
01:09:58 <kmc> yeah
01:10:01 <kmc> that was to douglass_ not you
01:10:11 <oerjan> ok
01:11:02 <oerjan> there's some limit but i'm pretty sure it's higher than 3 (i just have the 2)
01:12:10 <oerjan> Bike: clearly they should use the improved bfjoust algorithm.
01:12:11 -!- augur has joined.
01:12:20 <Bike> yeah that would make sense
01:15:56 <kmc> elo ratings
01:16:30 <boily> ~duck elo
01:16:31 <metasepia> --- No relevant information
01:16:51 <oerjan> `hello elo
01:16:52 <HackEgo> Hello
01:16:53 <boily> stupide bot de mes deux de pas foutu de trouver de l'information quand qu'on lui demande non mais tsé quand même...
01:17:20 <oerjan> `hello hm
01:17:20 <HackEgo> Hello
01:17:27 <oerjan> `cat bin/hello
01:17:28 <HackEgo> echo Hello
01:17:28 <kmc> `thanks ants
01:17:29 <HackEgo> Thanks, ants. Thants.
01:17:35 <oerjan> `ello elo
01:17:36 <kmc> `thanks Hanks
01:17:36 <HackEgo> elello
01:17:36 <HackEgo> Thanks, Hanks. Thanks.
01:18:34 <Bike> https://www.youtube.com/watch?v=rBHjx4dYxjA&t=2m20s my daily commute
01:19:10 <kmc> haha
01:19:28 <kmc> i like the part where individuals from different generations walk side by side
01:19:54 <kmc> Bike: why do people throw boxes at you on your commute
01:20:15 <Bike> They think I'm ugly since I'm made of boxes
01:20:17 <Bike> racism is terrible
01:20:32 <kmc> they see me boxin', they hatin'
01:20:46 <Bike> the dinosaurs falling over rule though yes
01:20:57 -!- JZTech102 has joined.
01:20:58 <kmc> hey it could be worse https://www.youtube.com/watch?v=rBHjx4dYxjA&t=3m10s
01:22:07 -!- vravn has quit (Ping timeout: 240 seconds).
01:22:07 -!- JZTech101 has quit (Ping timeout: 240 seconds).
01:22:35 <Bike> yeah i laughed out loud at that
01:22:36 <Bike> like why
01:22:41 <kmc> me too
01:25:11 -!- vravn has joined.
01:26:35 <oerjan> et moi aussi
01:30:26 <oerjan> someone may have found a fly in the ointment of that inflation discovery http://physicsworld.com/cws/article/news/2014/apr/10/have-galactic-radio-loops-been-mistaken-for-b-mode-polarization
01:33:32 -!- JZTech102 has changed nick to JZTech101.
01:33:33 -!- JZTech101 has quit (Changing host).
01:33:33 -!- JZTech101 has joined.
01:34:04 * oerjan smells a bot
01:37:54 <oerjan> ( 10/85
01:37:54 <idris-ircslave> 0.11764705882352941 : Float
01:38:29 -!- oerjan has set topic: The channel with > 10% bots | PSA: fizzie is running the wiki now, contact him for any problems | https://dl.dropboxusercontent.com/u/2023808/wisdom.pdf http://codu.org/logs/_esoteric/ http://tunes.org/~nef/logs/esoteric/.
01:38:47 -!- nooodl has quit (Ping timeout: 265 seconds).
01:39:55 <douglass_> I'm using douglass_ instead of douglass because I registered douglass and then was too lazy to use it. I do like thiotimoline a bit better but it's less recognizable to people who know me in real life.
01:40:24 <Bike> ha ha, i knew i recognized the name from somewhere (no i didn't)
01:41:01 <boily> we have over 10% bots? wooooah...
01:41:05 <oerjan> douglass_: unless you have nick protection on, you don't _need_ to log in to use douglass
01:41:37 <douglass_> I don't even remember what I did when I set up this screen.
01:41:47 <douglass_> Certainly too lazy to change anything until it dies.
01:41:47 <douglass_> .
01:41:49 <oerjan> OKAY
01:41:58 <Bike> "I inherited this dtach from my mother..."
01:42:51 <oerjan> boily: that also includes the logging bots.
01:43:55 <boily> and one unknown member, namely myndzi.
01:44:11 <Sgeo> I should play with Snap
01:44:15 * boily -> mattress
01:44:18 <oerjan> i didn't include myndzi in the 10, but e would certainly not hurt the number
01:44:36 <boily> go myndzi! embrace your siliconoïd side!
01:44:42 -!- boily has quit (Quit: ROBOT CHICKEN).
01:44:44 -!- metasepia has quit (Remote host closed the connection).
01:49:37 <Sgeo> http://www.christianet.com/creditcounseling/
01:49:43 <Sgeo> Christian... credit counseling
01:49:47 <Sgeo> I don't get it
01:52:25 <Sgeo> I tried to install Snap, it seems to be installing lens....
01:52:40 <Sgeo> Oh, right, Snap was the one that bothered me for having its own little weird lens library, iirc
01:52:43 <Sgeo> Has that changed?
01:53:50 <Bike> do you ever think about turbulence
01:56:35 -!- tromp has quit (Remote host closed the connection).
01:56:36 <madbr> isn't christianity generally against credit?
01:57:12 -!- tromp has joined.
01:57:26 <Bike> if you're like, living in 1200s Europe, maybe
01:59:54 <madbr> http://www.thebricktestament.com/the_law/lending_money/ex22_24.html
02:00:47 -!- xpte has quit (Ping timeout: 252 seconds).
02:01:13 <Bike> Yes as it turns out practice does not completely correspond to any legal code for more than about two seconds
02:01:43 -!- tromp has quit (Ping timeout: 276 seconds).
02:01:53 -!- ggherdov has quit (Ping timeout: 252 seconds).
02:02:07 <Bike> Insofar as legal codes even mean anything bla bla bla
02:02:50 -!- xpte has joined.
02:04:56 <oerjan> iiuc there was a time in europe when that was interpreted by the christians such that they could not take interest from anyone, but by the jews such that they couldn't take interest from other jews, but could do it from christians. and so jews became the major bankers.
02:05:48 <oerjan> (this iiuc comes with no guarantees of actually being correct)
02:15:51 -!- ggherdov has joined.
02:30:55 <Bike> That's what I was trying to get at with the 1200s comment, yeah. it was usury.
02:31:22 <HackEgo> [wiki] [[Boxy]] M http://esolangs.org/w/index.php?diff=39344&oldid=39341 * Zerk * (+224)
02:31:59 <Bike> the workarounds are pretty hilarious sometimes http://en.wikipedia.org/wiki/Islamic_banking#Bai.27_al_.27inah_.28sale_and_buy-back_agreement.29
02:50:09 <madbr> haha
02:55:10 <Bike> Hey non-Americans, question: in your early math classes what symbols were used to describe a basic linear equation
02:56:39 <shachaf> how early
02:57:13 <Bike> primary school? like "here's a line" early
02:57:25 <shachaf> not that i'm non-American, but i was in .il math classes until 4th grade or so
02:57:32 <shachaf> what do you mean by symbols
03:00:04 <madbr> y = ax + b
03:00:05 <madbr> ?
03:00:36 <Bike> something like that, yes.
03:00:50 <shachaf> i have no idea
03:01:01 <shachaf> did we even have those
03:01:11 <madbr> didn't see that until secondaire 1 tho (grade 7)
03:01:13 <Bike> i'm wondering because In American the usual is mx + b and the reason for the m is apparently mysterious
03:01:20 <Bike> In America rather
03:01:22 <madbr> oh yeah might have been
03:01:24 <madbr> mx + b
03:01:39 <Bike> quadratics are still ax² + bx + c though
03:01:43 <kmc> base and mantissa... wait no
03:01:54 <madbr> multiplier?
03:02:00 <Sgeo> I was at a friend's house, and some teacher(?) was quizzing me, and I knew things like derivatives but not what y=mx+b was about
03:02:03 <Bike> i literally found a citation for this saying "It just happened"
03:02:08 <Sgeo> This was many, many years ago
03:02:14 <kmc> multiplier? i hardly know 'er!
03:02:32 <Bike> and like apparently in swedeworld it's usually "kx + m", k for koefficient
03:02:48 <Sprocklem> m = slope; don't know why though
03:02:49 <shachaf> a⁢x² + m⁢x + b
03:02:53 <shachaf> just stick the m in there
03:02:57 <Bike> genius.
03:03:06 <shachaf> (i totally used U+2062 INVISIBLE TIMES, by the way)
03:03:10 <shachaf> (just sayin')
03:03:15 <Bike> whoa
03:03:23 <kmc> :O
03:03:24 * Sgeo verifies shachaf's clam.
03:03:31 <Sgeo> (Actually not fully verified)
03:03:39 <Sgeo> (And not a clam, but that was deliberate)
03:03:45 <Bike> wow, one of the citations is to an aol page
03:04:06 <shachaf> Bike: but aren't they both koëfficients :'(
03:04:12 <shachaf> wait maybe that's not valid swedish
03:04:20 <Bike> which is down, but thankfully i found a mirror. on tripod.
03:04:33 <Bike> shachaf: «where k may derive from "koefficient" in the Swedish word for slope, "riktningskoefficient."» is what it says
03:04:36 <Bike> but yes.
03:04:51 <Bike> rïktningsköefficient
03:05:11 <shachaf> Bïke
03:05:16 * oerjan swats Bike -----###
03:05:22 <Bike> :<
03:05:30 <oerjan> ö IS A SEPARATE LETTER
03:05:36 -!- password2 has joined.
03:05:37 <Bike> wat
03:05:46 <shachaf> oerjan: you should be swatting me, though
03:05:56 <shachaf> coefficient, more like scowefficient
03:06:25 <shachaf> (because it's such a scow)
03:07:48 <oerjan> incidentally sv:ko = cow but sv:kö = queue
03:08:02 <shachaf> what about no:kø
03:08:56 <oerjan> that's also queue
03:09:03 <oerjan> but cow is ku
03:09:15 -!- ^v has quit (Read error: Connection reset by peer).
03:09:26 <HackEgo> [wiki] [[Boxy]] M http://esolangs.org/w/index.php?diff=39345&oldid=39344 * Zerk * (+57) /* Toplevel */ fixed math by adding unwrapped 0/1 defs
03:09:43 -!- ^v has joined.
03:09:58 <Sgeo> :t maybe
03:09:59 <lambdabot> b -> (a -> b) -> Maybe a -> b
03:10:08 <oerjan> and ko is not a common word in norwegian.
03:10:14 <Sgeo> maybe 5 (error "Must be Nothing")
03:10:15 <Sgeo> vs
03:10:25 <Sgeo> maybe 5 (const $ error "Must be nothing)
03:10:32 <Sgeo> Any significant reason to use one vs the other?
03:10:43 <Sgeo> :t maybe 5 (error "Must be nothing")
03:10:45 <lambdabot> Num b => Maybe a -> b
03:10:46 <Bike> more dosh
03:10:50 <Sgeo> :t maybe 5 (const $ error "Must be nothing")
03:10:50 <lambdabot> Num b => Maybe a -> b
03:10:59 <Bike> the $ indicates xtreme dosh
03:11:02 <shachaf> @src maybe
03:11:03 <lambdabot> maybe n _ Nothing = n
03:11:03 <lambdabot> maybe _ f (Just x) = f x
03:11:16 <shachaf> they will compile to the same thing
03:11:50 <kmc> $wag overload
03:12:47 <oerjan> i don't think there's any difference in semantics
03:13:07 <kmc> benchmark it with criterion
03:13:22 <oerjan> given that maybe won't be applying any seq's to it.
03:14:20 <shachaf> that's actually what i meant
03:14:23 <madbr> kö sounds like a borrowing
03:14:28 <shachaf> i didn't mean that they would compile to the same thing
03:14:31 <oerjan> Sgeo: basically const $ error ... is equivalent to error ... unless you apply seq to it
03:14:39 <shachaf> i meant they have the same semantics and i was wondering whether they would also compile to the same thing
03:14:47 <shachaf> so i said the other one
03:14:47 <Sgeo> Even in terms of when the error message gets shown?
03:14:51 <shachaf> only if maybe gets inlined probably?
03:15:29 <Bike> i'm confused, this seems like the sort of thing haskell is pretty simple about
03:15:47 <Sgeo> Bike: undefined, and laziness in general, is a ... complication
03:15:59 <kmc> laziness: the ultimate side effect
03:16:05 <shachaf> it's not very complicated here
03:16:49 <Sgeo> I assume laziness can be described as a monad
03:17:27 <Sgeo> Although it kind of sucks to use lifted 'pure' functions in it, since they would be considered to be strict, I think
03:17:33 <kmc> Haskell didn't so much eliminate side effects as introduce a single pervasive side effect that doesn't play nice with others
03:17:40 * kmc checks if that's less than 140 characters
03:18:01 <Sgeo> Yeah but it allows for fancy tricks! At the expense of safety
03:18:06 <kmc> happiness is a warm monad
03:18:27 <kmc> the english word for cow is "cow"
03:19:15 <Bike> the english term for "cow" is "use-mention distinction"
03:20:50 <kmc> names with their own names!
03:21:11 <Bike> so does anyone else think once you have a "series of monographs" in over 40 volumes you should consider calling it something else
03:22:48 <oerjan> i assume it's time to link https://en.wikipedia.org/wiki/Haddocks'_Eyes#Naming again
03:23:08 <oerjan> Bike: a polygraph hth
03:23:09 <kmc> Bike: a polygraph?
03:23:12 <kmc> efb
03:23:17 <Bike> wow how did i not think of that
03:23:23 <kmc> , he lied
03:23:42 <oerjan> efb?
03:23:54 <kmc> oerjan: yeah that's the only example I know, besides "tetragrammaton"
03:23:57 <kmc> edit: fuck, beaten
03:24:14 <Bike> logos
03:24:17 <oerjan> tdh
03:24:23 <kmc> hm?
03:25:07 <oerjan> "efb" tdh hth hand
03:25:14 <Bike> yeah
03:25:17 <kmc> what
03:25:23 <oerjan> `
03:25:24 <HackEgo> ​/home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: : not found
03:25:32 <kmc>
03:25:34 <shachaf> that did help hope that helped have a nice day
03:25:37 <oerjan> `
03:25:38 <HackEgo> ​/home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: : not found
03:25:44 <oerjan> oh
03:25:49 <oerjan> `? tdh
03:25:50 <HackEgo> tdh is the past tense of a successful hth. hth.
03:25:59 <kmc> tdnh
03:26:08 <shachaf> `àwesomé´
03:26:08 <HackEgo> ​/home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: àwesomé´: not found
03:27:09 <oerjan> `? tdnh
03:27:10 <HackEgo> tdnh does not help
03:28:09 <oerjan> `run echo $PATH
03:28:09 <HackEgo> ​/hackenv/bin:/opt/python27/bin:/opt/ghc/bin:/usr/bin:/bin
03:28:45 <oerjan> i guess there's no way to get an empty command name to work
03:29:05 <oerjan> inside the box
03:35:48 -!- Sorella has quit (Quit: It is tiem!).
03:44:25 -!- password2 has quit (Ping timeout: 276 seconds).
03:49:35 <HackEgo> [wiki] [[Boxy]] http://esolangs.org/w/index.php?diff=39346&oldid=39345 * Zerk * (-6) /* Toplevel */ properly* fixed math
03:49:41 -!- nisstyre has quit (Quit: WeeChat 0.4.3).
04:00:43 -!- kallisti_ has joined.
04:03:39 -!- kallisti has quit (Ping timeout: 250 seconds).
04:08:07 -!- tromp has joined.
04:23:44 -!- tromp has quit (Remote host closed the connection).
04:30:01 -!- oerjan has quit (Quit: leaving).
04:32:14 <quintopia> `? hand
04:32:15 <HackEgo> hand? ¯\(°​_o)/¯
04:46:04 -!- Sprocklem has quit (Quit: leaving).
04:47:13 -!- Bike has quit (Quit: leaving).
04:47:34 -!- Bike has joined.
04:52:53 -!- shikhout has joined.
04:54:44 <Sgeo> (About OverlappingInstances)
04:54:45 <Sgeo> "Can sometimes be simulated with the extra-method trick used in the Show class of the Prelude for showing lists of characters differently than lists of other things.
04:54:45 <Sgeo> "
04:54:48 <Sgeo> What trick is this?
04:55:47 -!- shikhin has quit (Ping timeout: 252 seconds).
04:55:49 -!- shikhout has changed nick to shikhin.
04:56:27 <Jafet> :t showList
04:56:29 <lambdabot> Show a => [a] -> ShowS
05:00:00 <Sgeo> Is it a bad idea to write code like
05:00:07 <Sgeo> blah >>= flip when $ do ...
05:00:18 <ion> yes
05:00:32 <shachaf> > "flip" `isInfixOf` "blah >>= flip when $ do ..."
05:00:32 <idris-ircslave> When elaborating an application of constructor __infer:
05:00:32 <idris-ircslave> No such variable isInfixOf
05:00:34 <lambdabot> True
05:00:42 <shachaf> idris-ircslave: go away or change your prefix
05:00:43 <idris-ircslave> When elaborating an application of constructor __infer:
05:00:43 <idris-ircslave> No such variable go
05:12:16 <fizzie> I understand these days you just bite the bullet and use @run.
05:13:02 <shachaf> lambdabot: @run away or change your prefix
05:13:04 <lambdabot> Not in scope: ‘away’Not in scope: ‘change’
05:13:04 <lambdabot> Perhaps you meant one of these:
05:13:04 <lambdabot> ‘range’ (imported from Data.Ix),
05:13:04 <lambdabot> ‘hang’ (imported from Text.PrettyPrint.HughesPJ)Not in scope: ‘your’Not in...
05:13:04 <lambdabot> Perhaps you meant ‘prefixed’ (imported from Data.List.Lens)
05:14:00 <fizzie> Idris also had something lambdabot did not respond to.
05:14:16 <shachaf> Yes, I used it a few lines above.
05:14:19 <shachaf> (Apparently.)
05:15:26 <fizzie> Something shorter than the full name, too.
05:16:18 <Jafet> ( 2 + 2
05:16:18 <idris-ircslave> 4 : Integer
05:16:25 <fizzie> That.
05:17:08 <shachaf> ( 2 + 2 )
05:17:08 <idris-ircslave> (input):1:7: error: expected: "$",
05:17:08 <idris-ircslave> "$>", "&&", "&&&", "*", "***",
05:17:08 <idris-ircslave> "+", "++", "-", "->", ".", "/",
05:17:08 <idris-ircslave> "/=", ":+", ":-", "::", ":::",
05:17:08 <idris-ircslave> ":=", "<", "<$", "<$>", "<*>",↵…
05:17:10 <shachaf> hmph
05:17:12 <fizzie> It's like, it acquired a non-conflicting prefix, but did not get rid of >.
05:18:06 <fizzie> Which I kind of hope it would, but haven't cared about enough to say anything.
05:19:09 <Jafet> Think about it, shachaf, this is helping to get rid of the ) backlog.
05:19:16 <Jafet> (()
05:19:46 <fizzie> (: just doing my part.
05:27:31 -!- kallisti_ has quit (Ping timeout: 258 seconds).
05:30:16 -!- kallisti has joined.
05:36:34 <Bike> http://freshbsd.org/search?project=openbsd&q=file.name:libssl they mad
05:37:33 -!- asie has joined.
05:37:34 <kmc> https://github.com/jmhodges/libssl/commits/master
05:37:41 <asie> hi! i used to be known as asiekierka here
05:37:43 <asie> a few years back
05:37:45 <asie> how are you?
05:41:00 <shachaf> kmc: hmm, you've written ~20% more words in #haskell than i have
05:42:32 <kmc> Bike: hm your link is better
05:42:53 <Bike> i am the superior being.
05:47:06 <kmc> hasn't jconn already ruined the paren-balance of this room?
05:47:36 -!- conehead_ has joined.
05:48:05 -!- conehead has quit (Ping timeout: 250 seconds).
05:50:07 <maurer> ( ()
05:50:07 <idris-ircslave> () : ()
05:50:23 <shachaf> ( () -- )
05:50:24 <idris-ircslave> () : ()
05:50:29 <shachaf> ( -- () -- )
05:50:29 <idris-ircslave> command not recognized or not supported
05:50:35 <kmc> is that a tie fighter
05:51:15 <kmc> rather a TIE Advanced x1
05:51:20 <shachaf> ( -- ( ⋈ ) -- )
05:51:20 <idris-ircslave> command not recognized or not supported
05:53:18 -!- MoALTz has quit (Quit: bbl).
05:55:46 <myname> https://www.dropbox.com/s/skmvujsn35s6q2x/IMG_20140417_075324.jpg wat
05:58:49 <oklopol> i don't think that's particularly weird
05:59:04 <oklopol> (except that i don't like any of those bands)
05:59:24 <myname> the second paragraph confuses me
05:59:31 <olsner> http://freshbsd.org/commit/openbsd/6fec7be36425adfd2b48f3695d24a03f6287d4fe wat
05:59:54 <myname> i don't get why he translated from german in a german university, either
06:00:00 <oklopol> the fucking?
06:00:14 <kmc> those are some good musicians
06:00:14 <myname> the fucking is one point, yes
06:00:16 <oklopol> maybe that's slightly weird
06:00:20 <olsner> myname: for the international students?
06:00:22 <kmc> perhaps i should find this person and engage in some of these activities
06:00:40 <myname> kmc: go for it
06:01:00 <asie> )
06:02:52 <myname> kmc: in return, name a cool esolang i could write a compiler for, that is not too hard to make
06:05:46 * Jafet memorizes that list of musicians for later hipster usage
06:06:08 <myname> lol
06:06:32 <kmc> yeah radiohead is a pretty obscure hipster band
06:07:18 <Jafet> It looks like they forgot to translate Strawinsky
06:07:32 <myname> translate from what
06:07:34 <Bike> tschaikowsky
06:08:48 -!- ^v has quit (Read error: Connection reset by peer).
06:08:50 <kmc> wait are all those gerunds meant to apply to the noun "music"
06:09:06 <myname> kmc: i am not sure
06:09:16 <oklopol> that was my interpretation
06:09:16 -!- ^v has joined.
06:09:21 <myname> especially how "fucking music" would look like
06:09:23 <kmc> it makes... more and less sense simultaneously
06:09:26 <kmc> yeah
06:09:51 <oklopol> i thought the joke was that it sounds like he is proposing sex, but then it turns out he is still talking about music
06:09:58 <kmc> heh
06:10:07 <myname> but what the hell is fucking muic?
06:10:09 <myname> *s
06:10:48 <oklopol> (and that maybe his original wording in german made that slightly clearer, although i guess that's a bit unlikely)
06:11:04 <olsner> myname: that guy is, and the friends who respond?
06:11:26 <oklopol> he says something like in increasing order of emotional whatever
06:12:22 <kmc> perhaps he refers to products such as http://www.ohmibod.com/ [nsfw i guess]
06:12:49 <kmc> music controlled vibrators
06:13:18 <myname> sounds awesome
06:13:18 <oklopol> perhaps the meaning of (emotionally ordered) was clearer originally
06:13:27 <oklopol> and yeah those sound awesome
06:14:07 <kmc> eh better to build a USB control kit for the Magic Wand, i think
06:14:13 <Jafet> http://www.gamegirladvance.com/2002/10/sex-in-games-rezvibrator.html
06:14:23 <myname> kmc: go and sell the result
06:14:29 <myname> you will be rich in no time
06:14:41 <kmc> maybe...
06:14:44 <kmc> selling stuff is hard
06:14:47 <kmc> then it has to be, like, good
06:14:53 <kmc> i've never made a physical thing that was good
06:14:54 <kmc> it's hard
06:15:23 <kmc> the Hitachi Magic Wand contains a 120 V DC motor
06:15:31 <kmc> the two speed settings correspond to half-wave and full-wave rectifiers
06:15:53 <kmc> ask me how i know this
06:16:03 <myname> how do you know this?
06:16:08 <kmc> by taking one apart
06:16:17 <myname> that poor thing
06:16:21 <kmc> wasn't mine
06:16:27 <kmc> and we put it back together
06:16:29 <myname> i could think of better use cases
06:16:33 <myname> ah, okay
06:17:34 -!- conehead_ has quit (Quit: Computer has gone to sleep).
06:22:39 -!- ^v has quit (Quit: http://i.imgur.com/DrFFzea.png).
06:23:55 <kmc> at one of the sex toy shops in SF they have a museum with a bunch of old vibrators
06:25:40 <kmc> the vibrator was the fifth ever electric domestic appliance
06:25:45 -!- madbr has quit (Quit: Rouringu de hajikunda!).
06:26:00 <myname> at which point is the washing mashine?
06:28:15 <kmc> much later I think
06:28:46 <kmc> complicated and needs lots of power
06:29:00 <kmc> early appliances just connected in place of a light bulb
06:29:27 <shachaf> at what point was mr burbujas
06:29:37 <kmc> mr burbujas exists outside of time and space
06:29:48 <kmc> our entire universe is but one bubble in his frothy beard
06:30:13 <kmc> for the un-enlightened: http://www.sanfranciscodays.com/photos/large/mr-burbujas.jpg
06:30:23 <myname> kmc: the resistance?
06:33:53 * kmc -> bed (?)
06:34:50 <shachaf> bͭ̒̎ȅ̛̽d̯̤̈́?
07:32:44 -!- shikhin has quit (Ping timeout: 252 seconds).
07:49:12 -!- Tritonio has joined.
08:10:44 -!- Frooxius has quit (Quit: *bubbles away*).
08:10:55 -!- Frooxius has joined.
08:14:05 -!- Patashu has joined.
08:14:50 -!- Patashu[Zzz] has joined.
08:14:50 -!- Patashu has quit (Disconnected by services).
08:39:18 -!- Sellyme has quit (Excess Flood).
08:40:24 -!- Sellyme has joined.
08:41:22 -!- xpte has quit (Quit: Connection closed for inactivity).
08:50:28 -!- asie has quit (Ping timeout: 240 seconds).
08:52:17 -!- asie has joined.
09:02:53 <impomatic> boily, oerjan: anything in the name of science...
09:05:25 <fizzie> Bike: It's quite commonly "kx + ???" in Finland, and the "k" can be argued to stand for "kulmakerroin". I don't remember what the ??? usually is.
09:05:58 <Bike> mmhm mmhm
09:06:20 <fizzie> I think maybe "kx + b", though there is no Finnish explanation for b.
09:07:02 <fizzie> http://opinnot.net/kokonaisuudet/index.php?id_kokon=164&kieli_id=1&taso_kokon=0&oppiaine_kokon=2 and so on.
09:09:07 <fizzie> Searching for "y=kx" in Finnish has "b" for 9 out of 10 hits on the first page. (Last one just has "y = kx".)
09:10:11 <fizzie> ("kerroin" = coefficient, "kulma" = angle.)
09:10:30 <fizzie> ("kerroin" = "I told".)
09:13:39 <fizzie> It's the noun for coefficient, the first-person singular indicative past form of the verb "kertoa" 'to tell, to multiply (transitive)', and (which I didn't even think about at all before looking at Wiktionary) the instructive case plural form of the noun "kerta" (approx. 'occasion'). Possibly because the instructive case is such nonsense.
09:13:48 <fizzie> http://en.wikipedia.org/wiki/Instructive_case
09:13:54 <fizzie> What's that good for, anyway?
09:15:41 <fizzie> I guess it's used in something like "kaksin kerroin".
09:20:37 -!- yorick has joined.
09:22:26 -!- AnotherTest has joined.
09:44:09 -!- Tritonio1 has joined.
09:45:22 <fizzie> Fun fact: a = 0; printf("%d %d %d %d", a, a++, a, a++); produces "2 1 2 0" on many GCC versions on x86 and x86-64, "2 0 2 1" on GCC 4.7 on ARM and ICC 12.1.0 on x86-64, "0 0 1 1" on some versions of clang on x86-64, and "2 1 2 1" on the PGI C compiler on x86-64.
09:45:44 <fizzie> That's probably not an exhaustive list.
09:45:59 -!- Tritonio has quit (Ping timeout: 250 seconds).
09:47:26 <fizzie> The one with no zeroes in it is quite nice.
09:57:36 <b_jonas> fizzie: on x86, doesn't it also depend on -mcpu/-mtune?
09:57:54 <b_jonas> and maybe on other optimization switches
09:58:01 <fizzie> I couldn't coax other outputs with flags, but I'm sure it could be possible.
09:59:09 <lifthrasiir> we should make a C compiler that launches /usr/games/hack on compiling such program
10:16:31 -!- boily has joined.
10:29:22 -!- jconn has quit (Ping timeout: 276 seconds).
10:29:56 -!- asie has quit (Quit: I'll probably come back in either 20 minutes or 8 hours.).
10:32:47 <Jafet> You hacker
11:01:03 -!- Phantom_Hoover has joined.
11:05:06 -!- boily has quit (Quit: Ping: Read: Error: Timeout: Remote: Chicken).
11:07:37 -!- nucular has joined.
11:07:37 -!- nucular has quit (Changing host).
11:07:37 -!- nucular has joined.
11:17:54 -!- jconn has joined.
11:19:08 -!- atslash has joined.
11:25:49 -!- atslash has quit (Remote host closed the connection).
11:26:19 -!- atslash has joined.
11:27:47 -!- password2 has joined.
11:37:49 -!- jconn has quit (Ping timeout: 258 seconds).
11:38:17 -!- slereah has joined.
11:40:17 -!- jconn has joined.
12:10:43 -!- MindlessDrone has joined.
12:11:31 -!- Sorella has joined.
12:18:00 -!- Sgeo has quit (Read error: Connection reset by peer).
12:44:05 <slereah> Hm
12:44:23 <slereah> What activation function would be faster for a backpropagation algorithm
12:44:46 <slereah> The basic sigmoid has an exponential in it, but on the other hand the derivative can be expressed simply by the function
12:45:06 <int-e> atan has 1/(1+x^2) as its derivative
12:45:23 <slereah> And the one that is 1/(1 + |x|) is simple to calculate but the derivative has to be calculated independantly
12:45:29 <slereah> Well probably anywa
12:46:37 <Jafet> Changing the activation also changes the learning rate, doesn't it
12:46:42 -!- Sellyme has quit (Excess Flood).
12:46:55 <slereah> Probably
12:47:01 <slereah> Dont know which one is better for it
12:47:53 -!- Sellyme has joined.
12:52:39 <fizzie> The Elliott activation function has the best name, obviously.
12:53:20 <slereah> http://www.dontveter.com/bpr/activate.html
12:53:25 <slereah> BACKPROAGATOR REVIEW
12:53:26 <slereah> Ahah
12:54:03 -!- Phantom_Hoover has quit (Ping timeout: 250 seconds).
12:55:46 <FreeFull> tanh is a sigmoid too
12:56:17 <FreeFull> Not sure if it would be suitable though
12:56:29 <slereah> "Both these sigmoid approach their extremes more slowly. This means that if you are trying to output numerical values it will take more iterations to reach your target value. But if you're doing a classification problem you really only care to get the correct output value greater than the other outputs and here these functions will save on CPU time without influencing the number of iterations required by very much. "
12:56:30 <slereah> Hm
12:57:26 <slereah> I guess I'll worry about that later and try several activation functions
12:57:37 <fizzie> "However, while simulation is faster with elliotsig, training is not guaranteed to be faster, due to the different shapes of the two transfer functions. Here, 10 networks are each trained for tansig and elliotsig, but training times vary significantly even on the same problem with the same network." (MATLAB Neural Networking toolbox, "Optimize Neural Network Training Speed and Memory" article.)
13:01:37 -!- conehead has joined.
13:13:20 <elliott> does it really misspell his name?
13:14:30 <slereah> elliott, tell us about your activation function
13:14:45 <elliott> that might be hard
13:14:49 <fizzie> elliott: It does; it also does not credit him in any way.
13:15:53 <fizzie> Not that I think MATLAB documentation usually does, but still.
13:17:33 <fizzie> (It also doesn't mention what the actual function is, anywhere; I had to use "type" to check that it actually is the same as that David Elliott one.)
13:17:50 <elliott> it's a lot more common as a surname than a first name, it seems
13:20:50 -!- douglass_ has quit (Ping timeout: 255 seconds).
13:22:23 -!- douglass_ has joined.
13:41:46 -!- Patashu[Zzz] has quit (Ping timeout: 276 seconds).
14:06:40 -!- Phantom_Hoover has joined.
14:31:02 <slereah> Man my perceptron converges values all to 0.5 :(
14:31:10 <slereah> What went wrooong
14:34:00 <int-e> it doesn't care
14:35:54 <Jafet> It learned the truth
14:37:44 <slereah> The problem isn't from the learning rate and inertia apparently
14:38:18 <slereah> For some reason all my weights converge to 0
14:39:18 <slereah> It's that Stephen King book all over again
14:48:31 <Phantom_Hoover> slereah, i thought you were shachaf for quite a long time
14:48:38 <Phantom_Hoover> plz. recapitalise your nick
14:49:02 <slereah> It is.
14:56:48 <Bike> phantom's right
14:57:56 <slereah> slereah may not be upper cased
14:57:58 <slereah> But Slereah_ is
14:58:01 -!- `^_^v has left ("Leaving").
14:58:05 -!- `^_^v has joined.
15:02:04 -!- Tritonio1 has quit (Ping timeout: 240 seconds).
15:03:02 -!- Tritonio has joined.
15:04:26 <Bike> if all weights are converging to zero does that just mean it's learning a constant function
15:04:41 <slereah> I'm trying to make an autoencoder
15:04:52 <slereah> So the training set is just the identity function
15:05:03 <slereah> Currently the input is just (1 2 3 4 5)
15:05:27 <slereah> And I'm trying to get the same as output
15:05:50 <slereah> But for some reason weights all decrease with each update until they're all 0
15:06:08 <Bike> "Reading an embargoed paper that cites 71 papers in the first two sentences."
15:06:17 <Bike> well, that sounds like a bug probably /helpful
15:07:06 <slereah> You don't sa
15:07:06 <slereah> y
15:10:31 <fizzie> Are you doing it all DIY, or with some sort of a thing?
15:12:28 <slereah> A Thing indeed
15:12:32 <slereah> internship thing
15:12:39 <slereah> Oh wait
15:12:45 <slereah> Do you mean coded from scratch
15:12:48 <slereah> Then yes
15:13:02 <slereah> Though I did look at a few libraries of it before
15:17:53 <fizzie> Right, right, that just makes bug-in-the-implementation the likelier reason than doing-it-wrong.
15:18:11 <slereah> Well doing it wrong is still on the table
15:18:34 <slereah> I guess maybe I should do a very simple one
15:18:44 <slereah> 2 in, 2 out, 1 hidden neurone
15:22:18 <Bike> can you do an identity map if everything has to go through one neuron...?
15:22:55 <slereah> Dunno
15:23:09 <slereah> Hm
15:23:29 <slereah> It's unfortunate that there are no step by step examples of the backpropagation algorithm
15:23:34 <slereah> I could check where it went wrong
15:25:46 -!- Tritonio has quit (Ping timeout: 276 seconds).
15:31:47 <fizzie> Bike: It should at least do a credible approximation, anyway, as long as things work out.
15:33:15 <fizzie> Autoencoder features are quite "hip" at the moment in the speech recognition field.
15:33:25 <Bike> how would it distinguish, like, 01 from 10
15:33:38 <slereah> Bike : different weights?
15:34:17 <Bike> The hidden neuron outputs one bit, doesn't it? How do you encode two bits into one bit
15:35:00 <fizzie> Bike: It's not usually binary; but in any case, by having first output equal to the hidden bit, and the second its complement.
15:35:07 <fizzie> Bike: That way 0 -> 01 and 1 -> 10.
15:35:21 <Bike> but then you couldn't also do 00 and 11.
15:35:28 <slereah> The network has real values and real weights
15:35:31 <Jafet> Neural networks can output any real number
15:35:52 <fizzie> Bike: Sure, sure; but it's supposed to learn the "best" low-dimensional representation it can.
15:36:14 <Jafet> You interpret the number if you want a classifier
15:36:50 <Bike> okay i guess i thought after the sigmoid you had a binary threshhold. oopsie.
15:37:26 <Jafet> The sigmoid is used to introduce nonlinearity (and to compress outputs between layers)
15:38:52 <fizzie> http://jmlr.org/papers/volume11/vincent10a/vincent10a.pdf is the kind of thing people do on speech too.
15:39:37 <fizzie> (Have to go do some things.)
15:40:00 <slereah> l8er
15:44:37 -!- password2 has quit (Ping timeout: 265 seconds).
15:52:05 -!- tertu has joined.
15:55:00 -!- mekeor has joined.
15:55:54 -!- ^v has joined.
15:59:21 -!- tertu has quit (Ping timeout: 252 seconds).
16:00:13 -!- slereah has quit (Quit: Leaving).
16:01:49 <Taneb> Today I got a free hat
16:08:34 -!- Sprocklem has joined.
16:09:08 -!- doesthiswork has joined.
16:27:26 -!- ^v has quit (Read error: Connection reset by peer).
16:27:54 -!- ^v has joined.
16:28:41 -!- shikhin has joined.
16:36:34 <doesthiswork> so I've got a fun little thing. variables can only hold random bits. You can set a variable to the result of xoring or averageing other variables together. Print prints out the name of the variable that most closely matches the chunk of bits coming after it.
16:37:03 <doesthiswork> the control structures are "label" and "call" but procedures do not accept arguments
16:38:06 <Jafet> So, just xor all variables with themselves
16:38:14 <doesthiswork> call calls the chunk of code labeled with the most similar chunk of bits to the one its called with
16:38:24 <doesthiswork> yes 0 is easy to get
16:38:49 <doesthiswork> but 0 doesn't do anything special
16:39:31 <Jafet> Ok, so what is special
16:40:15 <doesthiswork> you can holographically store values in other values
16:40:36 <doesthiswork> using xor to make a pair and averaging to combine several pairs
16:44:05 -!- Slereah_ has quit (Ping timeout: 252 seconds).
16:45:33 <doesthiswork> I'm pretty sure it isn't turing complete because you can only store a finite amount of information in an anonymous value
16:45:47 -!- Slereah_ has joined.
17:12:57 <FreeFull> What sort of ternary operations would be useful for a ternary computer?
17:13:18 <FreeFull> I was thinking tritwise max would be a rough equivalent of OR
17:13:19 <b_jonas> FreeFull: all the trintercal ones?
17:13:31 <b_jonas> wait... useful?
17:13:38 <FreeFull> =P
17:14:40 <b_jonas> dunno, I'd rather stick to binary (with two's complement or one's complement, and fibonacci and negafibonacci)
17:16:07 <FreeFull> But ternary is the easiest way to render a sierpinski carpet to the screen
17:17:46 <doesthiswork> there are 3^9 different two argument ternary operators
17:18:13 <doesthiswork> but yeah, max and min seem useful
17:18:30 <int-e> so that's one operation then, the sierpinsky operation: 222;202;222. For the other, addition, perhaps?
17:19:29 <fizzie> Not going the balanced ternary route? (In which case multiplication is I guess quite a natural one.)
17:19:53 -!- nisstyre has joined.
17:20:20 <doesthiswork> you could try creating random ternary circuts out of a selection of ternary gates and see which set seems to produce the remainder of the gates most fairly
17:25:54 -!- Phantom_Hoover has quit (Ping timeout: 240 seconds).
17:33:59 <FreeFull> A single binary operation has many ternary equivalents
17:34:18 <FreeFull> Tritwise multiplication and tritwise min are both equivalents of AND
17:42:15 <kmc> what's a minimal basis for all ternary operations?
17:45:12 <fizzie> "A single two-input ternary operator and a single one-input ternary operator are proposed, which together with the logic values 1 and 2 form a functionally-complete set of ternary operators."
17:46:08 <fizzie> Tokmen, V. H. "A functionally-complete ternary system", Electronic Letters, Feb 1978, 14(3), pp. 69-71, http://dx.doi.org/10.1049/el:19780048
17:46:39 -!- Phantom_Hoover has joined.
17:47:48 <fizzie> The two-input operator it mentions is the {max(a, b) if a != b; 0 if a = b} one, and the unary operator is the {2 if a = 0; 0 if a != 0} one.
17:48:13 <Bike> no zero? nice
17:48:34 <fizzie> That's just the unary operator applied to 1 or 2, after all.
17:48:50 <Jafet> You don't need 0 for binary logic, either
17:49:17 <Bike> yeah, i still like it
17:49:41 <Jafet> 0 is nothing compared to 1
17:49:41 <fizzie> "The particular merit of the above operators is their ease of realisation in bipolar or m.o.s., technology."
17:49:59 -!- shikhin has changed nick to draumr.
17:50:04 <fizzie> The two-input operator is called the TOR, short for "Ternary exclusive-OR".
17:51:08 <fizzie> Also the notation is quite funky. TOR is denoted by a τ b, and the unary one by ⁰a⁰ of all things.
17:51:18 <fizzie> Sorry, that's ⁰(a)⁰.
17:51:31 <Jafet> The teddy ears operator
17:51:44 <Jafet> I'm sure it'll catch on
17:53:09 <Bike> too cute not to
17:53:47 -!- draumr has changed nick to shikhin.
17:54:20 <Jafet> I wonder if they ever realised it in bipolar or m.o.s. technology.
17:54:38 <fizzie> Apparently it's a friend of ¹(a)¹ and ²(a)² -- the number denotes what a must be to get the value 2.
17:54:52 -!- nortti has changed nick to hvidie.
17:56:00 -!- hvidie has changed nick to nortti.
17:59:33 -!- atslash has quit (Ping timeout: 250 seconds).
18:02:17 <fizzie> Whose idea it was anyway to make it unclear whether calling something a "binary operator" (resp. a ternary operator) means that it takes two (resp. three) inputs or that it has two (resp. three) states for those inputs?
18:05:24 <Bike> huh. i don't think i've heard the latter. until um, just now.
18:05:45 <kmc> whose idea was it that the term "special case" should have two opposite meanings
18:07:06 <Jafet> kmc: http://nlab.mathforge.org/nlab/show/red%20herring%20principle
18:07:17 <Jafet> "The mathematical red herring principle is the principle that in mathematics, a “red herring” need not, in general, be either red or a herring."
18:08:02 <fizzie> The "two-input ternary operator" was an instance of the latter use.
18:08:20 <fizzie> And this other "Ternary Exclusive Or" is an instance of the former.
18:08:59 <kmc> Taneb: what kind of hat did you get for free; also how?
18:09:03 <fizzie> (It's just a boolean logic connective that's true iff exactly one of its three inputs is true.)
18:09:05 <kmc> `coins
18:09:07 <HackEgo> mopocoin relyknitiationcoin hairequitcoin pingbraicoin liheidifycoin rpecoin bdallcoin duplycoin rubeofthecoin pcoin buggonaltcoin peracoin claimcoin aftgreurochesecoin festcoin selectpcoin hercoin wildcoin obfurlocoin mckecoin
18:09:19 <fizzie> fi:mopo == en:moped.
18:09:31 <kmc> rubeofthecoin
18:09:53 <fizzie> duplycoin doesn't even try to prevent double-spending.
18:10:49 <Taneb> kmc, some form of knit cap
18:12:04 <fizzie> Huh, I didn't even know the next Ubuntu was going to be the "Trusty Tahr" and now it's released.
18:12:50 <Jafet> I hope ubuntu dies after Zebra
18:12:58 <Jafet> If for no other reason
18:13:30 <fizzie> The development code name page already goes up to 20.04 that will start with F; they just loop around.
18:13:42 <fizzie> The one where they list suggestions, that is.
18:14:39 <fizzie> The U names sound p. stupid.
18:52:44 -!- mekeor has quit (Remote host closed the connection).
18:54:16 -!- MindlessDrone has quit (Quit: MindlessDrone).
18:55:37 <b_jonas> fizzie: no, duplycoin is like the treasure in Bellatrix Lestrange's vault in Gringotts: it's enchanted so that when you touch it it produces duplicates that look similar but are worthless, and also glow hot
18:56:48 <kmc> i've said it before but they should loop around with fungi next
18:57:21 <kmc> Ascendent Amanita, Beautiful Bolete, Charming Chanterelle
18:58:34 -!- vravn has quit (Ping timeout: 240 seconds).
19:00:03 -!- xpte has joined.
19:02:06 <b_jonas> kmc: good idea. then bacteria, then plants after that.
19:04:20 -!- vravn has joined.
19:07:55 -!- nisstyre has quit (Quit: WeeChat 0.4.3).
19:15:10 <kmc> delicious dekkera, esoteric ergot, faithful flammulina, glistening ganoderma, happy hypholoma, intriguing inocybe, jovial jelly, keen kuehneromyces
19:22:33 <kmc> leafy laetiporus, magnificent morel, noble neolentinus, omniscient omphalotus, pleasing pleurotus, quixotic queletia, rustic russula
19:22:51 -!- AnotherTest has quit (Ping timeout: 252 seconds).
19:23:13 -!- Phantom_Hoover has quit (Ping timeout: 252 seconds).
19:28:18 <shachaf> `olist (949)
19:28:18 <HackEgo> olist (949): shachaf oerjan Sgeo FireFly boily nortti
19:30:22 <kmc> superb saccharomyces, tasty truffle, unctuous urnula, vernal velvet-foot, wonderful wine-cap, xenodochial xylaria
19:30:30 <kmc> i'll get back to you on y and z
19:30:46 -!- MoALTz has joined.
19:31:26 <kmc> i have no idea if velvet foot mushrooms actually grow in the spring; also it's the same critter as flammulina
19:33:28 -!- jix has quit (Remote host closed the connection).
19:34:26 -!- jix has joined.
19:41:27 -!- Sprocklem has quit (Ping timeout: 265 seconds).
19:45:01 -!- Sprocklem has joined.
19:45:01 -!- Sprocklem has quit (Changing host).
19:45:01 -!- Sprocklem has joined.
19:46:03 -!- Phantom_Hoover has joined.
19:51:42 <kmc> loveable lactobacillus
19:56:25 -!- shikhout has joined.
19:58:21 -!- shikhin has quit (Disconnected by services).
19:58:25 -!- shikhout has changed nick to shikhin.
20:17:31 <kmc> http://livegrep.com/search/linux?q=CHICKEN
20:25:36 <HackEgo> [wiki] [[Hollang]] N http://esolangs.org/w/index.php?oldid=39347 * Doesthiswork * (+2349) new holographic language
20:26:22 <HackEgo> [wiki] [[Hollang]] M http://esolangs.org/w/index.php?diff=39348&oldid=39347 * Doesthiswork * (+2)
20:31:09 -!- ^v has quit (Quit: http://i.imgur.com/DrFFzea.png).
20:34:19 -!- nisstyre has joined.
20:39:23 -!- doesthiswork has quit (Quit: Page closed).
20:39:39 -!- nooodl has joined.
20:41:14 <b_jonas> `8-ball
20:41:15 <HackEgo> Concentrate and ask again.
20:41:19 <b_jonas> 'k
20:43:26 <Bike> http://opensslrampage.org/post/83019384273/ok-there-was-a-need-for-openssl-cleanse-instead all your favorite drama, now in tumblr form
20:43:40 <myname> what the hell @ hollang
20:59:38 -!- adu has joined.
21:09:33 <Bike> ok i'm loving this, i looked up what OPENSSL_malloc is
21:09:37 <Bike> OPENSSL_malloc is a macro that expands to a call to CRYPTO_malloc, which calls a function pointer. the default value of that function pointer is a function that calls another function pointer. the default value of /that/ function pointer is, finally, malloc.
21:10:34 -!- nucular has quit (Quit: Excess Food).
21:11:19 <elliott> Bike: are you sure there's not one too many levels of funptr indirection there...?
21:11:25 <elliott> I'd... be surprised but not surprised enough
21:11:48 <Bike> yes. the outer function pointer is "malloc_ex_func" which is like malloc but takes __LINE__ and __FILE__
21:12:03 <elliott> nice.
21:12:40 <Bike> see, the whole purpose for this is that if nothing's been allocated yet, you (as in you, outside the library) can call a function to switch all these pointers to something else
21:13:01 <elliott> that's actually useful sometimes, for what it's worth
21:13:03 <elliott> gmp offers it too
21:13:20 <Jafet> BECAUSE LIBC MALLOC IS TOO SLOW
21:14:47 <Bike> hm does mercurial's web interface let you grep code or no
21:17:22 <Bike> p[-1] = (0xdeadbeef << 31) + 0xdeafdeed; <-- nifty (in gmp)
21:17:53 -!- adu has quit (Quit: adu).
21:19:36 <Jafet> Why would they shift a 32-bit number by 31 bites
21:19:47 <Jafet> I must be hungry
21:19:54 <Jafet> s/es/s
21:21:27 <int-e> it could be 64 bits
21:21:35 <Bike> it's aprt of a debug section
21:21:58 <int-e> or 36
21:22:10 <int-e> (to quote another popular value ;-) )
21:22:17 <Jafet> 0xdeadbeef is a 32-bit constant
21:22:37 -!- Sprocklem has quit (Ping timeout: 276 seconds).
21:23:33 <int-e> it's an int
21:23:57 <int-e> so you don't know how many bits it has from the compiler's perspective
21:24:52 <fizzie> libpng has a compile-time option (PNG_USER_MEM_SUPPORTED) that lets you switch around the memory allocation functions.
21:25:03 <Jafet> Unless that's in the cray mpn code or something, it's going to be a 32-bit unsigned int
21:27:08 <fizzie> And GMP has at least reasonable chances of running on such a thing, compared to many other pieces of code.
21:27:44 <Jafet> Yes, I know there is cray mpn code (and probably people still using it)
21:31:53 <Bike> https://gmplib.org/repo/gmp/file/55ff6b8d9a92/memory.c#l130 i come bearing context
21:32:36 <fizzie> Well, that's curious.
21:32:39 <Jafet> I guess it's probably what it is for no particular reason
21:33:04 <fizzie> Since mp_ptr is an mp_limb_t *, and quite often mp_limb_t is unsigned long instead.
21:33:16 <Jafet> It's repeated a dozen times, though, and you'd think it'd be a macro at that point
21:34:08 <fizzie> Would presumably make sense to convert the constants to mp_limb_t.
21:35:45 <Jafet> Yes, it looks like a "bug"
21:38:18 -!- nisstyre has quit (Quit: WeeChat 0.4.3).
21:39:44 <fizzie> Also I believe it's possibly undefined behaviour on a system where int is larger than 32 bits but smaller than 64.
21:40:44 <Jafet> Is it? The shift is valid as long as unsigned int is 32-bit or larger.
21:40:59 <fizzie> But the type of an unsuffixed hex constant 0xdeadbeef is signed int if it can represent the value.
21:41:19 <Jafet> Huh, evil.
21:41:22 <fizzie> And E1 << E2, when E1 has a signed type, is undefined if E1 * 2^E2 is not representable in the result type.
21:45:30 <fizzie> It's also kind of nasty, the way the types of integer literals changed from C89 to C99.
21:46:19 <fizzie> I mean, I'm sure they had to, thanks to the introduction of long long. But there used to be a "fallback" option for also unsuffixed decimal constants to end up as "unsigned long" if they were too long for a long.
21:48:02 <fizzie> http://sprunge.us/SCIF and so on.
21:48:28 <Jafet> This is why I always cast numeric values before doing anything weird with them
21:48:29 <fizzie> Also the silliest messages, __int128 is not unsigned.
21:53:55 <fizzie> Fun fact: C89 (or at least the "ansi.c.txt" draft that keeps circulating around) does not seem to really define E1 << E2 at all if E1 is signed.
21:54:25 <fizzie> "The result of E1 << E2 is E1 left-shifted E2 bit positions; vacated bits are filled with zeros. If E1 has an unsigned type, the value of the result is E1 multiplied by the quantity, 2 raised to the power E2, reduced modulo ULONG_MAX+1 if E1 has type unsigned long, UINT_MAX+1 otherwise. (The constants ULONG_MAX and UINT_MAX are defined in the header <limits.h> .)"
21:54:34 <fizzie> That's all it says on the topic of left shifts.
21:55:52 <fizzie> (C99 adds the "if signed type and nonnegative value and E1*2^E2 is representable" sentence.)
21:56:43 <fizzie> http://www.open-std.org/jtc1/sc22/wg14/www/docs/dr_081.html apparently it was entirely implementation-defined for C89.
21:56:44 -!- oerjan has joined.
22:02:30 <fizzie> "There is an ongoing discussion about what privacy means and if it is still needed. Some tracks differ from the versions on the albums listed. It was a tremendous success and marked the beginning of Italian operatic dominance north of the Alps."
22:02:49 <fizzie> Nice spam-filter avoidance header, I even thought it was something sane at first.
22:04:11 -!- idris-ircslave has quit (Ping timeout: 252 seconds).
22:05:31 -!- Melvar has quit (Ping timeout: 276 seconds).
22:06:22 <Jafet> fungot, do you write spam subject lines
22:06:23 <fungot> Jafet: the honourable member will be aware that all eurocontrol does is coordinate flights a priori. we are also acknowledging that this parliament sends the message that i would not be accepted by the council last week, when we discuss matters such as pay and conditions in a coastal area into account when drawing up the olive sector, the communal services, the commission generally agrees with.
22:07:37 <oerjan> i can believe that would avoid some filters.
22:07:49 <Jafet> The concerning thing, mr president, is that we are at more-or-less the point where fungot can speak more coherently than some healthy adults
22:07:49 <fungot> Jafet: mr president, i agree with you in mind. amendments nos 84 and 85 on reviewing the priority list do not allow the commission to take due account of the particular problem arising from the strategies will be adopted today in the role of the court of justice in the fnord session of the parliament, to support the airlines in their country.
22:08:28 <oerjan> Jafet: this can be fixed by redefining healthy again hth
22:08:35 -!- Bicyclidine has joined.
22:08:48 <Jafet> If you hit them in the head, they are no longer healthy
22:09:38 <oerjan> i imagine the fnord session of the parliament is where they discuss the secret illuminati stuff.
22:10:39 <Bicyclidine> ActiveX components enable you to display ActiveX controls in your GUI. They are available only on the Microsoft® Windows® platform.
22:11:29 <fizzie> Are you doing MATLAB®™ there?
22:11:53 <Bicyclidine> yes
22:12:02 <Bicyclidine> wondering if i should bother figuring out how to grey out a textbox
22:12:13 <oerjan> hm there are symbols for trademarks and copyrights, is there one for patents
22:12:38 <fizzie> oerjan: U+1F4A9 hth
22:12:55 <oerjan> `unicode U+1F4A9
22:12:56 <HackEgo> Unknown character.
22:13:08 <oerjan> NOPE
22:13:23 <oerjan> `cat bin/unicode
22:13:24 <HackEgo> ​#!/usr/bin/env python \ # -*- encoding: utf-8 -*- \ import re \ import sys \ import unicodedata \ def l(c): m = re.match('(?:U+)?([0-9a-f]{1,5})$', c, re.I); return unicodedata.lookup(c) if m is None else unichr(int(m.group(1),16)) \ try: \ print u''.join(map(l, sys.argv[1:])).encode('utf-8') \ except KeyError: \ print u'Unknown characte
22:13:45 <fizzie> Oh.
22:13:48 <fizzie> Unescaped +
22:13:54 <fizzie> Shows how well I test these things.
22:14:10 <Jafet> Not just the standard unicode executable?
22:14:24 <oerjan> `run sed -i 's/U[+]/U[+]/' bin/unicode
22:14:26 <HackEgo> No output.
22:14:31 <oerjan> `unicode U+1F4A9
22:14:32 <HackEgo> ​💩
22:14:33 <Bicyclidine> uh this is #esoteric, _sir_
22:14:42 <Bicyclidine> wow owned. good one
22:15:07 <oerjan> `unidecode 💩
22:15:08 <HackEgo> ​[U+1F4A9 DUNNO]
22:15:08 <Jafet> python isn't very esoteric (below 100 lines, at least)
22:15:29 <Jafet> Incidentally, that character describes these programs also
22:15:40 <Jafet> `cat bin/unidecode
22:15:41 <HackEgo> ​#!/usr/bin/env python \ import sys \ import unicodedata \ print u" ".join("[U+{0:04X} {1}]".format(ord(c), unicodedata.name(c, "DUNNO")) for c in " ".join(sys.argv[1:]).decode("utf-8")).encode("utf-8")
22:15:58 <fizzie> It's all up to the version of Python unicodedata that's installed there.
22:16:10 <Jafet> `python --version
22:16:10 <HackEgo> Python 2.7.3
22:16:29 <fizzie> At least we've got it out of UCS-2 nowadays; it was stuck in that for ages.
22:16:58 <Jafet> `run python -c $'import unicodedata\nprint unicodedata.unidata_version'
22:16:59 <HackEgo> 5.2.0
22:17:47 <fizzie> So close.
22:18:07 -!- Melvar has joined.
22:18:10 <Bicyclidine> "To add an image to a push button or toggle button, assign the button's CData property an m-by-n-by-3 array of RGB values that defines RGB (Truecolor) Images."
22:18:22 <fizzie> oerjan: In case you don't feel like looking it up, it was just PILE OF POO.
22:19:16 <oerjan> well that was obvious once i could see it properly; putty only displayed the left part
22:20:20 <fizzie> Also, stalker mode has stopped eworking for me.
22:20:47 <Bicyclidine> "You can extend the file types that the open command recognizes to include any file having a three-character extension. Do this by creating a MATLAB code file with the name openxyz.m. xyz is the file extension for the type of files to be handled. Do not, however, take this approach for opening FIG-files, because openfig.m is a MATLAB function which is needed to open GUIs."
22:20:51 <fizzie> I thought it was just something with the Android tablet's browser, but it doesn't seem better on this desktop either.
22:20:52 <Bicyclidine> good system.
22:21:10 <fizzie> It keeps on scrolling and blinking.
22:24:57 -!- adu has joined.
22:27:21 * oerjan doesn't use stalker mode anyway.
22:28:43 -!- tromp has joined.
22:30:10 <oerjan> <Sgeo> blah >>= flip when $ do ... <-- the precedences are wrong for that to work.
22:30:31 <oerjan> @tell Sgeo <Sgeo> blah >>= flip when $ do ... <-- the precedences are wrong for that to work.
22:30:31 <lambdabot> Consider it noted.
22:30:46 <oerjan> wtf is Sgeo not here
22:31:17 <shachaf> :t return True >>= do flip when $ Just ()
22:31:18 <lambdabot> Maybe ()
22:31:21 <shachaf> hth
22:31:43 <oerjan> um
22:32:05 -!- Sprocklem has joined.
22:32:43 <oerjan> fancy use of do
22:32:54 <oerjan> :t do flip when $ Just ()
22:32:55 <lambdabot> Bool -> Maybe ()
22:33:02 <shachaf> > Just $ 5 :: Int
22:33:04 <lambdabot> Couldn't match expected type ‘GHC.Types.Int’
22:33:04 <lambdabot> with actual type ‘Data.Maybe.Maybe s0’
22:33:05 <shachaf> > Just $ do 5 :: Int
22:33:07 <lambdabot> Just 5
22:33:17 <Bicyclidine> just do int
22:33:32 <oerjan> hm i'm not sure if the last one is portable
22:33:55 <Jafet> Portable to what
22:34:00 <oerjan> is do with just one expression guaranteed to work if the expression doesn't have monadic type
22:34:01 <Bicyclidine> hugs obviously
22:34:08 <oerjan> also uhc
22:34:08 <elliott> oerjan: yes IIRC
22:34:22 <oerjan> elliott: in the report?
22:34:42 <elliott> yes IIRC
22:34:48 <elliott> the desugaring doesn't talk about types I think
22:34:53 <elliott> only that it uses Prelude return/(>>=)
22:34:58 <oerjan> i guess if so then you actually _can_ use do as a general-purpose precedence fixer, huh
22:36:35 <Jafet> `file /usr/share/unicode/unicodeData.txt
22:36:35 <HackEgo> ​/usr/share/unicode/unicodeData.txt: ERROR: cannot open `/usr/share/unicode/unicodeData.txt' (No such file or directory)
22:36:38 <Jafet> `file /usr/share/unicode/UnicodeData.txt
22:36:39 <HackEgo> ​/usr/share/unicode/UnicodeData.txt: ERROR: cannot open `/usr/share/unicode/UnicodeData.txt' (No such file or directory)
22:36:46 <Jafet> `lsb_release -a
22:36:47 <HackEgo> No LSB modules are available. \ Distributor ID:Debian \ Description:Debian GNU/Linux \ Release:n/a \ Codename:n/a
22:38:07 <Bicyclidine> kind of a boring codename
22:38:11 <oerjan> @run do let x = 5; x+2
22:38:13 <lambdabot> <hint>:1:18:
22:38:13 <lambdabot> parse error (possibly incorrect indentation or mismatched brackets)
22:38:23 <oerjan> @run do let {x = 5}; x+2
22:38:25 <lambdabot> 7
22:39:25 -!- Patashu has joined.
22:39:39 <oerjan> sadly the brackets are needed there
22:39:56 <Jafet> n/a is a rock that andy picked up on the way home. He hasn't named it yet.
22:40:17 <Bicyclidine> haskell needs more guillemets
22:49:29 -!- tromp has quit (Remote host closed the connection).
22:50:49 <Jafet> https://packages.debian.org/sid/unicode-screensaver
22:51:03 <Jafet> I wonder if it does the emoji block
22:53:59 <Jafet> `fetch http://sources.debian.net/data/main/u/unicode/0.9.7/unicode
22:54:03 <HackEgo> 2014-04-17 22:53:59 URL:http://sources.debian.net/data/main/u/unicode/0.9.7/unicode [26623/26623] -> "unicode" [1]
22:56:39 <Jafet> `fetch http://ftp.unicode.org/Public/UNIDATA/UnicodeData.txt
22:56:41 <HackEgo> 2014-04-17 22:56:37 URL:http://ftp.unicode.org/Public/UNIDATA/UnicodeData.txt [1367023/1367023] -> "UnicodeData.txt" [1]
22:56:49 <Jafet> Wonder how long this will ta.. nevermind.
22:57:32 <Jafet> `run mv bin/unicode{,.old} && mv unicode UnicodeData.txt bin && chmod +x bin/unicode
22:57:34 <HackEgo> No output.
22:57:47 <Jafet> `unicode U+1F4A9
22:57:47 <HackEgo> U+1F4A9 - No such unicode character name in database \ UTF-8: f0 9f 92 a9 UTF-16BE: d83ddca9 Decimal: &#128169; \ 💩 (💩) \ Uppercase: U+1F4A9 \ Category: Cn (Other, Not Assigned)
22:57:56 <Jafet> o kay
23:00:29 <Jafet> `run sed -i 's,\./UnicodeData\.txt,/hackenv/UnicodeData.txt,' && unicode U+1F4A9
23:00:30 <HackEgo> sed: no input files
23:00:38 <Jafet> `run sed -i 's,\./UnicodeData\.txt,/hackenv/UnicodeData.txt,' bin/unicode && unicode U+1F4A9
23:00:40 <HackEgo> U+1F4A9 - No such unicode character name in database \ UTF-8: f0 9f 92 a9 UTF-16BE: d83ddca9 Decimal: &#128169; \ 💩 (💩) \ Uppercase: U+1F4A9 \ Category: Cn (Other, Not Assigned)
23:01:22 <Jafet> `unicode U+1F4A9
23:01:23 <HackEgo> U+1F4A9 PILE OF POO \ UTF-8: f0 9f 92 a9 UTF-16BE: d83ddca9 Decimal: &#128169; \ 💩 \ Category: So (Symbol, Other) \ Bidi: ON (Other Neutrals)
23:01:31 <Jafet> Maybe that needs to be trimmed.
23:02:29 <oerjan> `unicode PILE OF POO
23:02:30 <HackEgo> U+1F4A9 PILE OF POO \ UTF-8: f0 9f 92 a9 UTF-16BE: d83ddca9 Decimal: &#128169; \ 💩 \ Category: So (Symbol, Other) \ Bidi: ON (Other Neutrals)
23:03:04 <oerjan> oh well i guess it contains all the information needed
23:03:53 <oerjan> `file bin/unicode
23:03:54 <HackEgo> bin/unicode: Python script, ASCII text executable
23:16:24 -!- Patashu[Zzz] has joined.
23:16:24 -!- Patashu has quit (Disconnected by services).
23:16:33 -!- tromp has joined.
23:18:43 -!- adu has quit (Quit: adu).
23:21:47 <HackEgo> [wiki] [[Talk:Hollang]] N http://esolangs.org/w/index.php?oldid=39349 * Zerk * (+225) Created page with "Could you clarify "holographically embed them into variables"? How does averaging two bitstrings work, some operation between an ''and'' and an ''or''?--~~~~"
23:22:39 -!- yorick has quit (Remote host closed the connection).
23:35:31 <kmc> "EXIT() is really exit(), a gentle surprise but… OPENSSL_EXIT() is really just return()"
23:39:37 -!- ^v has joined.
23:40:08 -!- adu has joined.
23:49:52 -!- adu has quit (Quit: adu).
23:52:55 -!- adu has joined.
←2014-04-16 2014-04-17 2014-04-18→ ↑2014 ↑all