00:00:16 <Vorpal> Well there seem to be other ones
00:00:23 <Bike> yeah i don't think any of them are on sale or nuthin
00:00:28 <Bike> probably because noone cares
00:00:36 <elliott> this is where i link the datahand
00:00:38 <Koen_> it's open-source so it must be good
00:00:50 <Vorpal> Bike, http://www.amazon.com/The-PEREGRINE-Wearable-Interface-Medium/dp/B0035HABMM
00:00:50 <Bike> http://www.youtube.com/watch?v=Id3GQHXVjQY this is basically the opposite of futuristic
00:01:00 <Koen_> hey can you wire a keyglove to a pair of google glasses, and, you know, be *cool*
00:01:00 <Bike> uh don't talk back to me vorpal
00:01:02 <Vorpal> elliott, I already mentioned the datahand above
00:01:21 <Bike> the best thing about wearable computers is that people have been saying they look stupid since before the web
00:01:41 <Bike> Optimized for WoW, DotA, League of Legends, StarCraft 2
00:01:45 <Bike> i rest my case.
00:01:55 <Vorpal> Yes, and possibly sound stupid if you are using voice control
00:02:04 <Vorpal> "Google glass, take a picture"
00:02:15 <Vorpal> Saying that in public might not be such a smart idea
00:02:16 <Bike> Stupid and/or hallucinating.
00:02:23 <Bike> Apparently you can't use this thing as a keyboard.
00:02:36 <Vorpal> well okay, google search lied to me
00:02:41 <elliott> i think people are used to talking to nobody at this point
00:02:50 <elliott> given that handsfree stuff exists
00:03:01 <Bike> ##electronics is being exist again whyyyyy
00:03:14 <mnoqy> being exist: truly the worst curse
00:03:15 <elliott> sort of like everything else
00:03:18 <Vorpal> <Bike> http://www.youtube.com/watch?v=Id3GQHXVjQY this is basically the opposite of futuristic <-- yes, and it looks very hot too
00:03:21 <Koen_> hey, people talking in the phone with a handfree set already look stupid
00:03:25 <Bike> you know what i'm just not going to fix it
00:03:49 <Bike> Vorpal: "hot", I think you mean "optimized for arctic use". think like a marketer
00:03:50 <Vorpal> Koen_, even more so if bluetooth so there is no visible cable
00:04:08 <Vorpal> Bike, eh I have gloves like that. So not arctic. Just Nordic
00:04:17 <Koen_> and you know I keep thinking they're talking to me
00:04:20 <Bike> i'm pretty sure nordland is in the arctic.
00:04:42 <Vorpal> But not the south of Sweden where I live
00:04:51 <Vorpal> Not arctic climate here
00:05:00 <Koen_> oh hey I was at a bar earlier today
00:05:15 <mnoqy> did you just realize that Koen_
00:05:17 <Koen_> and I wanted a beer and the waiter told us what kind of beers he had
00:05:22 <Vorpal> Bike, it would be arctic climate if it wasn't for the gulf ocean stream, which carries hot water up here
00:05:32 <Koen_> and he had affligem, and when he said it it sounded like "african"
00:05:47 <Koen_> and you know, African beer? I had never tried that before so I asked for one
00:06:01 <Bike> i only drink asian beer
00:06:40 <Fiora> Bike: I know it's not quite what they were thinking of but "hand keyboard" makes me think of these things
00:06:43 <Fiora> http://ulva.com/images/maltron-right-hand785x581.jpg
00:06:48 <Koen_> mnoqy: I just wanted to share that african beer was a myth
00:07:21 <mnoqy> what a crappy myth
00:07:35 <Koen_> Fiora: I have no idea what position my hand is expected to be in to type on that
00:07:56 <Vorpal> elliott, hm.. "On a typical PC keyboard of today, the Caps Lock is pressed by the weakest finger pinky. The Ctrl key can be easily pressed with palm.", I just tried, on a laptop with ctrl outmost (nope), on my MS Natural (hell no) and on a standard straight PS/2 keyboard (not really without hitting other stuff, but least improbable of all the alternatives)
00:08:13 <elliott> well your hands are gigantic or whatever
00:08:43 <kmc> yeah ##electronics is not "kmc social values approved"
00:08:46 <mnoqy> who the heck presses ctrl with the palm
00:08:50 <Fiora> press the ctrl key with your palm? @_@
00:08:51 <kmc> perhaps i should have made that clear before
00:08:52 <elliott> though again I find hitting alt with thumb most comfortable
00:08:55 <mnoqy> ctrl is 100% a pinky key
00:09:01 <Vorpal> I can do ctrl-i with one hand on a MS Natural, awkward and not possible without turning myself yes, but that is the span
00:09:06 <Fiora> I have to move my entire hand over to do that
00:09:29 <Bike> kmc: well it was obvious within two minutes of going in. still quite irritating
00:09:49 <mnoqy> ##electronics sounds bad
00:09:52 <Bike> how do you press anything with your palm.
00:10:19 <Bike> the guest of christmas past has come for you, koen
00:10:31 -!- Guest21998 has quit (Changing host).
00:10:31 -!- Guest21998 has joined.
00:10:36 -!- Guest21998 has changed nick to Phantom_Hoover.
00:10:45 <Vorpal> Bike, I suspect you could reliably hit space with the lower part of your palm (though not on a laptop or ultraflat), but what would the point be
00:11:03 <Vorpal> Definitely possible on my MS Natural
00:11:25 <Koen_> I think there's some animal on the roof
00:11:42 <Vorpal> Btw, it was good getting a split keyboard, turned out i was hitting g with the wrong hand before.
00:12:03 <Vorpal> Phantom_Hoover, gnomes
00:12:37 <mnoqy> i dont worry about "correct hand".. i often hit things with the wrong hand if it's easier to hit them that way the way my hands are situated
00:12:48 <mnoqy> or i just always hit them with the wrong hand id
00:13:04 <Vorpal> elliott, with palm for ctrl I can't hit anything closer than the F keys with my fingers without them hurting
00:13:16 <elliott> anyway my point is not that ctrl is in a good position really
00:13:21 <mnoqy> i dont think it hurts my productivity or moral fibre
00:13:21 <elliott> just that caps lock isn't in a good position either
00:13:29 <Vorpal> well okay, that might be true
00:13:50 <Vorpal> elliott, where would you have it? I think I would actually suggest pedals for the modifiers
00:13:57 <Koen_> mnoqy: or i just always hit them with the wrong hand idk <<< wouldn't that make it the right hand then?
00:14:03 <elliott> well like I said the position of alt seems best for the main command modifier key
00:14:08 <mnoqy> Koen_: left, actually
00:14:15 <elliott> so maybe I will swap alt and ctrl next time I am on non-OS X
00:14:17 <Koen_> right as in notwrong
00:14:33 <Vorpal> elliott, say left foot down = ctrl, left foot forward = alt, right foot down = shift, right foot forward = altgr
00:14:48 <elliott> well I don't think foot pedals on a laptop is such a winning idea
00:14:57 <Vorpal> elliott, well yes, that would be an issue
00:15:07 <Koen_> Phantom_Hoover: can you hear them too?
00:15:26 -!- Phantom_Hoover has left ("Leaving").
00:15:29 <mnoqy> is this some sort of stupid joke or
00:15:45 <mnoqy> rest peacefully, ph
00:24:03 -!- Phantom_Hoover has joined.
00:32:28 <Bike> http://www.guardian.co.uk/uk-news/2013/jul/10/mod-trident-scotland-independence hahaha
00:33:06 <kmc> just let them have some nukes
00:33:08 <kmc> what's the big deal
00:33:22 <Bike> well they don't want them
00:33:33 <Bike> and if the UK doesn't have a nuclear submarine patrolling at all times France will invade
01:16:32 <HackEgo> johnny57: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.)
01:18:19 <Phantom_Hoover> for some reason when i saw that text i thought of fruit polos
01:21:54 <HackEgo> /home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: pblist: not found
01:21:55 <HackEgo> /home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: plist: not found
01:21:59 <HackEgo> /home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: blist: not found
01:22:00 <HackEgo> /home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: flist: not found
01:22:03 <HackEgo> pbflist: shachaf Sgeo quintopia
01:22:16 <elliott> Phantom_Hoover: that was sad
01:23:21 <Bike> do we have `alwllist
01:23:33 <Bike> `alillist actually
01:23:34 <HackEgo> /home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: alillist: not found
01:24:05 <Bike> i don't see any strips after honk, phantom.
01:24:22 <HackEgo> /home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: allelelist: not found
01:26:21 <Phantom_Hoover> (http://en.wikipedia.org/wiki/File:Hitchhiker's_thumbs.jpg omg there's a name for this?)
01:28:07 <Fiora> Huh, tongue rolling *isn't* inheritable?
01:29:51 <Fiora> "There is little laboratory evidence supporting the hypothesis that tongue rolling is inheritable and dominant. A 1975 twin study found that identical twins were no more likely than fraternal twins to both have the same phenotype for tongue rolling."
01:30:43 <Fiora> but earwax still is apparently?
01:31:01 * Fiora has the dominant gene yay
01:31:13 <kmc> what are the consequences of the dominant gene
01:31:30 <Fiora> the wet type is dominant and dry is recessive, it says?
01:31:41 <kmc> i don't know which type of earwax i have
01:31:45 <kmc> and maybe I would be happy not knowing
01:31:45 <Phantom_Hoover> it also says that dry is mostly found in asians and native americans
01:32:27 <Fiora> I guess I got that from my dad, though it just says "more likely", so
01:32:48 <Fiora> I guess tongue rolling is a lot easier to show around in a bio class xD
01:33:01 -!- Suchorski has quit.
01:33:14 <Fiora> oh! and there's the tasting "PTC" thing
01:34:40 <Phantom_Hoover> yeah, you can't exactly do much with ability to smell cyanide, or albinism
01:34:58 <Koen_> what's tongue rolling?
01:35:09 <Fiora> http://en.wikipedia.org/wiki/File:Rolled_tongue_flikr.jpg
01:37:00 <Koen_> I can roll my tongue the other way
01:37:01 <mnoqy> thats pretty gross ph
02:01:55 -!- Phantom_Hoover has quit (Read error: Connection reset by peer).
02:04:45 -!- Sgeo has joined.
02:19:51 <Sgeo> "any compiled language can have BOFs"
02:19:56 <Sgeo> [buffer overflows]
02:37:45 -!- atehwa_ has joined.
02:43:46 -!- Guest90495 has quit (*.net *.split).
02:43:47 -!- Koen_ has quit (*.net *.split).
02:43:47 -!- atehwa has quit (*.net *.split).
02:43:48 -!- SirCmpwn has quit (*.net *.split).
02:43:48 -!- Lymia has quit (*.net *.split).
02:45:06 <kmc> people are wrong on the internet
02:45:56 <Fiora> no sleeping tonight kmc
02:46:58 <Sgeo> kmc, I knew him IRL
02:47:00 <kmc> fortunately i have many techniques for not caring that people are wrong on the Internet
02:47:07 <kmc> Sgeo: oh, well, people are sometimes wrong IRL too
02:47:10 <Sgeo> He's a pen tester
02:47:12 <kmc> i am frequently wrong anywhere
02:47:20 <kmc> Sgeo: ballpoint or fountain?
02:47:34 <Sgeo> The best kind of pen
02:49:01 <kmc> pen 15 club
02:49:20 -!- Lymia has joined.
02:49:20 -!- Lymia has quit (Changing host).
02:49:21 -!- Lymia has joined.
02:49:37 <Bike> apparently hume did a pretty good criticism of gofai. i'm thrill'd
02:49:51 -!- SirCmpwn has joined.
02:50:07 <kmc> i saw a seagull basically hovering about 4 feet above the ground
02:50:10 <kmc> on an updraft
02:50:16 <kmc> they are so skilled
02:50:30 <Bike> birds own, imo
02:51:40 <kmc> seagulls will swoop down and take your food but they aim at a point away from you and change course at the last moment
02:51:45 <kmc> in order to avoid tipping off others
02:53:21 <kmc> there are a lot of mourning doves in my new neighborhood, which makes me happy
02:57:16 <Bike> is that german for sleeping
03:04:05 -!- nooodl has quit (Ping timeout: 246 seconds).
03:15:16 -!- Koen_ has joined.
03:15:41 <Koen_> Sgeo: I don't even know the difference between a compiled language and an interpreted one
03:16:31 <Bike> compiled languages are faster, duh
03:21:38 -!- sacje has joined.
03:25:16 <Sgeo> "compiled into machine code that is directly executed by the OS/Processor"
03:25:34 <Bike> no don't try a serious explanation
03:25:39 <Sgeo> I wasn't trying it
03:25:44 <Sgeo> That was his explanation
03:25:57 <Bike> ok well it's dumb still
03:26:25 <Sgeo> "trivial languages that you make up to prove your points are irelivant
03:27:14 <Bike> hey stop arguing with the wrong person
03:27:42 <Sgeo> About half an hour ago
03:27:48 <Sgeo> Because he had other things to do
03:52:58 -!- TeruFSX has quit (Ping timeout: 264 seconds).
03:59:17 -!- sprocklem has quit (Remote host closed the connection).
04:05:21 * Sgeo sticks out his tongue at those who say 8GB should be enough
04:05:28 <Sgeo> Almost reached that point earlier today
04:05:37 <Sgeo> Two users logged in, a lot of work stuff and some home stuff
04:06:03 -!- Koen_ has quit (Quit: Koen_).
04:09:08 -!- kallisti has joined.
04:19:45 -!- Bike has quit (Ping timeout: 245 seconds).
04:21:20 -!- Bike has joined.
04:23:21 -!- Koen_ has joined.
04:23:47 -!- Koen_ has quit (Read error: Connection reset by peer).
04:24:20 -!- Koen_ has joined.
04:39:43 -!- Bike has quit (Ping timeout: 245 seconds).
04:46:15 -!- Bike has joined.
04:53:17 -!- oerjan has joined.
04:57:37 -!- sacje has quit (Quit: sacje).
05:00:58 -!- conehead has quit (Quit: Computer has gone to sleep.).
05:03:08 <shachaf> Did Phantom_Hoover seriously `pbflist?
05:09:05 <shachaf> did you leave all the channels you used to be in!!
05:09:36 <mnoqy> especially #esoteric
05:09:46 <shachaf> #esoteric isn't a real channel
05:09:55 <shachaf> oonbotti........this was your cue........!!!
05:10:29 <Bike> oonbotti isn't alive it can't understand you
05:10:31 <Bike> also it's not here
05:10:45 <shachaf> bicycles aren't alive either
05:11:46 <Bike> uh dude i'm right here
05:14:20 -!- oonbotti2 has joined.
05:15:58 <nortti> it is a rewrite, doesn't have #esoteric but I can add it if you want
05:17:50 -!- oonbotti2 has quit (Remote host closed the connection).
05:18:07 -!- oonbotti2 has joined.
05:33:38 <oerjan> was that command ever intended to have a useful purpose?
05:34:01 -!- Sgeo_ has joined.
05:37:38 -!- Sgeo has quit (Ping timeout: 245 seconds).
05:40:56 <oerjan> huh Gregor is not here, or is he hiding under another nick again
05:41:35 <oerjan> i can see no sufficiently ridiculous ones :P
05:42:54 * oerjan checks a couple obscure ones just in case
05:43:20 -!- oerjan has set topic: Now playing Where's Gregor | <3 | http://underhanded.xcott.com/?page_id=5 | logs: http://codu.org/logs/_esoteric and http://tunes.org/~nef/logs/esoteric/?C=M;O=D.
05:44:10 <fizzie> 2013-07-11 00:17:35 (EEST) -!- Gregor!~Gregor@libdl.so has quit [*.net *.split]
05:51:28 <Sgeo_> Who got the infamous libc.so?
05:58:18 <kmc> how goes it
05:58:57 <Sgeo_> http://whois.domaintools.com/libc.so
05:59:05 <Sgeo_> Tech Contact ID: mxm1148892775
05:59:13 <shachaf> kmc: turns out i'll be meeting the people near the powell st bart at 19:00
05:59:22 <kmc> that's a fine bart
06:00:19 <fizzie> Sgeo_: "Contact ID: mxm1148892775 -- Internationalized Name: Marcel Meyer -- Internationalized Organization: levelsystems -- Internationalized Street: Siedlungsstrasse 6c -- Internationalized City: Erding -- Internationalized Country: DE -- Email: marcel.meyer@levelsystems.de "
06:00:41 <kmc> germany is quite the internationalized country
06:00:45 <oerjan> no:bart = en:moustache, hth
06:00:46 <kmc> shachaf: what sort of people are they
06:00:49 <Gracenotes> shachaf: oh, incidentally, the sencha reminded me a bit of local anaesthetic
06:00:55 <fizzie> levelsystems: Wir arbeiten lieber für unsere Kunden als eine hübsche Webseite zu machen...
06:00:59 <Gracenotes> that was what I was thinking of earlier
06:01:25 <shachaf> Gracenotes: that was the green rubbery tea?
06:01:48 <HackEgo> slist: Taneb atriq Ngevd Fiora nortti Sgeo ThatOtherPerson alot
06:02:00 <Gracenotes> I should figure out at some point if all sencha is like that.
06:02:07 <shachaf> kmc: some of them irc people whom i've briefly met before
06:02:10 <Gracenotes> I probably should have stuck with jasmine, but I was curious
06:02:15 <Sgeo_> Ugh I am tired and hungry and I think sleep deprivation has been making me sick
06:02:30 <shachaf> Gracenotes: tea adventure is good hth
06:02:39 <kmc> shachaf: anyone i know?
06:04:07 <shachaf> Gracenotes and i had fancy tea today btw, it was p. good hth
06:04:18 <shachaf> should i become a tea snob
06:04:26 <kmc> sounds good
06:04:45 <shachaf> should i become a tea snob
06:05:17 -!- Koen_ has quit (Quit: Koen_).
06:05:36 <shachaf> misinformation on stackoverflow :'(
06:06:01 <Gracenotes> add a new answer, someone will surely upvote it
06:06:38 <shachaf> imo people much prefer answers like "ghc auto-memoized it" and "tail-call elimination" to answers like "GMP is just faster than Java's BigInteger"
06:07:02 <Gracenotes> yes, SO has many features that pinpoint it a lot more as a place-of-social-exchange rather than an authoritative-answer-on-unique-question.
06:07:25 <Bike> i'm... not sure how memoization would help with bignum arithmetic
06:07:42 <shachaf> well, it wouldn't help in this case, unless the function was called more than once
06:07:48 <Gracenotes> it's not really a fair marketplace of answers. it's an extremely unfair one that can still be effective, if you're lucky.
06:07:53 <Bike> i'm trying to imagine how it possibly could
06:08:15 <Bike> i guess you'd need a situation where the thing being memoized couldn't just be done by the ALU in negligible time anyway
06:08:18 <shachaf> Bike: well, if you're computing bignum fibonacci numbers or something naïvely it would probably be p. helpful
06:08:27 <Bike> ok but that's really stupid.
06:09:01 <Bike> like really dumb. might as well implement multiplication for fixnums yourself. in base 10.
06:09:06 <shachaf> in this case it's factorial which is even stupider since memoization doesn't even help (unless you use it more than once)
06:09:12 <Gracenotes> also, Haskell's Integer type isn't even GMP a lot of the time
06:09:32 <kmc> that's true, but maybe doesn't make a huge difference?
06:09:32 <Bike> computing factorials fast is a very interesting problem
06:09:33 <shachaf> also my favorite part was when someone said the haskell code was faster than the java code because ghc does strictness analysis
06:09:43 <kmc> I would expect that GMP is also pretty fast on machine-size integers
06:09:45 <kmc> shachaf: wow..........
06:10:00 <shachaf> the question at hand: http://stackoverflow.com/questions/17584630/why-is-factorial-calculation-must-faster-in-haskell-than-in-java
06:10:00 <elliott> java has a very basic strictness analyser
06:10:03 <Bike> what does that... but... aaaaaah
06:10:06 <elliott> takes up only 0 lines of code
06:11:28 <Bike> man java bignums are always so wonderfully verbose
06:11:45 <Bike> what the hchrist is this cache thing, help
06:12:05 <Bike> In Haskell, all functions are pure unless they're doing IO (see link)
06:12:17 <Gracenotes> strictly speaking Java uses an array of int to implement BigInteger, but it also stores more information (5 additional words, plus 'vtable').
06:12:22 <kmc> kind of true
06:12:23 <Bike> christ. christ help
06:12:33 <Bike> This is able to do tail-call optimization because it's just iteration. The same can be done in Java with a for loop as in your example, but it won't have the benefit of being functionally pure.
06:12:47 <elliott> true by way of being tautological
06:13:04 <elliott> though only because the condition is never satisfied
06:13:06 <Bike> ok so anyway. computing factorials is interesting.
06:13:11 <Gracenotes> Haskell, on the other hand, for quite big integers, stores the number of bytes used, fact that it's long, and the number itself.
06:13:11 <elliott> well I guess you could also say it has nothing to do with functions
06:13:11 <kmc> elliott: well if your flavour of Haskell has unsafePerformIO then it's true by being tautological, but not vacuous
06:13:21 <Bike> i'm not sure how practical it is to factorize and stuff
06:13:35 <Bike> since if you're doing 1000000! you probably don't need precision anyway?
06:13:35 <elliott> kmc: okay but, I don't think considering unsafePerformIO when talking about the semantics is worthwhile.
06:14:21 <shachaf> that's the real problem here
06:14:23 <elliott> who wants to read a pop sci article about homotopy type theory http://www.newscientist.com/article/dn23749-mathematicians-think-like-machines-for-perfect-proofs.html
06:14:32 <Bike> i get enough shit in pop sci about brains
06:14:33 <elliott> Bike: right from the title
06:14:38 <elliott> you know it's going to be amazing
06:14:38 <Bike> i... i can't take the math.
06:14:53 <mnoqy> im not up for anything that amazing imo
06:15:10 <Gracenotes> As well as know, types didn't exist before machines.
06:15:43 <lambdabot> 3.1415926535897932384626433832795028841972
06:16:02 <mnoqy> cant believe you dont know what cereal is
06:16:08 <Bike> > 2 ^ 7 -- i also forgot
06:16:33 <mnoqy> please tell me youve seen sourcereal.com bike
06:16:39 <Bike> i think i have actually
06:16:40 <lambdabot> "3.141592653589793238462643383279502884197169399375105820974944592307816406...
06:16:45 <Bike> > let stirling n = ((n/exp 1)^n)*(sqrt (2 * pi * n)) in stirling 1000000
06:16:46 <lambdabot> No instance for (GHC.Show.Show a0)
06:16:46 <lambdabot> arising from a use of `M582996387.sho...
06:16:52 <shachaf> sourcereal.com is v. important
06:17:07 <Bike> :t \n -> ((n/exp 1)^n)*(sqrt (2 * pi * n))
06:17:08 <lambdabot> (Floating b, Integral b) => b -> b
06:17:18 <Bike> :t \n -> ((n/exp 1)**n)*(sqrt (2 * pi * n))
06:17:25 <Bike> yeah everything's terrible imo
06:17:32 <Bike> > (\n -> ((n/exp 1)**n)*(sqrt (2 * pi * n))) 1000000
06:18:56 <kmc> > iterate log 10
06:18:57 <lambdabot> [10.0,2.302585092994046,0.834032445247956,-0.18148297420509205,NaN,NaN,NaN,...
06:19:07 <elliott> `addquote <Bike> what's cereal <mnoqy> cereal <mnoqy> cant believe you dont know what cereal is
06:19:07 <Bike> is there a log with base?
06:19:11 <HackEgo> 1071) <Bike> what's cereal <mnoqy> cereal <mnoqy> cant believe you dont know what cereal is
06:19:20 <Gracenotes> Anyway, the HoTT article has some nice quotes from Bauer
06:19:31 <Bike> > logBase 100 10
06:20:03 <Bike> > let shitfuckfuckfuck n = n * log n - n in shitfuckfuckfuck 1000000
06:20:16 <Bike> i am the worst
06:20:46 <kmc> > map (length . takeWhile (join (==)) . iterate log) $ iterate (*2) 1
06:20:47 <lambdabot> [3,3,4,4,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,...
06:21:01 <shachaf> mnoqy: you forgot to give a definition
06:21:18 <shachaf> Cereal can be defined as something that is made from cereal.
06:21:24 <shachaf> are we sure that's not cocereal
06:21:58 <Bike> oh, e7. i didn't fuck up too bad.
06:22:35 <Bike> > 1.2815510557964273e7 / log 10
06:22:41 <Gracenotes> > let x = map (length . takeWhile (join (==)) . iterate log) $ iterate (*2) 1 in findIndices not $ zipWith (==) x (tail x)
06:22:54 <Bike> so can haskell bignums handle five million digit numbers
06:23:05 <Gracenotes> > let x = take 1000 . map (length . takeWhile (join (==)) . iterate log) $ iterate (*2) 1 in findIndices not $ zipWith (==) x (tail x)
06:23:26 <Bike> yes well i don't know what gmp can handle
06:23:31 <Bike> i am the worst, as previously mentioned
06:23:31 <shachaf> Bike: that's a lot of digits
06:23:45 <shachaf> Bike: can we settle on two million digits
06:23:49 <Gracenotes> gmp can handle as many digits as can into... what's the order of magnitude... TB?
06:23:57 -!- Sgeo_ has quit (Read error: Connection reset by peer).
06:24:15 <Gracenotes> Well, modulo other constraints besides just pure size.
06:24:22 <Bike> why did i get the base 10 digits first hm
06:25:40 <Bike> re aczel, it's kind of a shame i'm too principled to pretend anything brain related depends on mathematical foundations, or else i'd totally write an awesme counterpart to shittyaczelpaper based on hott
06:26:37 <Gracenotes> brains? they're pretty great, but what about them
06:26:58 <elliott> ok so the abc conjecture proof includes a paper with the title "Inter-universal Teichmuller Theory I: Construction of Hodge Theaters"
06:27:02 <elliott> why did nobody tell me this before.
06:27:09 <Bike> elliott: dude those papers are great
06:27:15 <Bike> they make no sense
06:27:31 <Bike> Gracenotes: you're missing this whole "backstory" that i already explained to elliott
06:27:54 <Gracenotes> you should have been hanging out with me in Palo Alto
06:28:05 <Bike> that's like ten waleses away though.
06:28:10 <Gracenotes> rather than explaining things to elliott
06:28:14 <elliott> Bike: what aczel paper are we talking here
06:28:15 <mnoqy> thats a good paper title
06:28:22 <shachaf> higgledy piggledy / GMP Integer: / only constrained by the / size of your RAM. // sure, i'm neglecting the / addressability; / 32-bit are some / kind of a scam
06:28:35 <fizzie> Maximum number of limbs in a GMP number is whatever fits in an int on your platform; the limbs themselves are typically 32 or 64 bits.
06:28:57 <shachaf> these double dactyl things are hard ok
06:29:06 <Bike> elliott: the bikeologist one
06:29:15 -!- Taneb has joined.
06:29:31 <fizzie> So that's somewhere around 68719476736 or 137438953472 bits.
06:29:37 -!- FreeFull has quit (Quit: Gone).
06:29:57 <shachaf> write me a double dactyl and/or limerick plz thx hth
06:30:21 <Gracenotes> I will also accept poems in onegin stanza
06:30:31 <lambdabot> http://www.matisse.net/bitcalc/?input_amount=137438953472&input_units=bytes¬ation=legacy
06:30:31 <lambdabot> Title: Bit Calculator - Convert between bits/bytes/kilobits/kilobytes/megabits/megab...
06:30:41 <shachaf> @google 137438953472 bits in terabytes
06:30:43 <lambdabot> http://letconversion.com/data-storage-conversion/from-terabits/to-bytes
06:30:43 <lambdabot> Title: Terabits to Bytes Conversion Calculator - LetConversion
06:30:49 <Gracenotes> I will even take both tetrameter and pentameter
06:30:50 <shachaf> wow that's not a lot of TB
06:30:55 <shachaf> Gracenotes: sure, go for it
06:31:02 <Bike> 12:18 <Bike> a biologist mentioned it (it was a dumb mention) 12:18 <elliott> of course it was, a biologist made it
06:31:10 <Bike> that's "the backstory"
06:31:15 <shachaf> Bike: Wow, that's pretty rude.
06:31:37 <shachaf> hmm i was going to make a mean joke but i won't make it
06:32:12 <Taneb> There was a man on Wight/whose limericks were kinda shite/When asked for a ditty/his response was so shitty/His entire audience got a fright
06:32:16 <Bike> you know who aczel is right
06:32:20 <shachaf> imo onegin stanzas aren't so great usually? in english anyway? but maybe they are
06:32:41 <Bike> he did all that set theory stuff
06:32:50 <shachaf> Gracenotes: you could also just write me a plain old iambic pentameter shakespeare-style sonnet imo
06:32:59 <Bike> and a biologist was like "wow this is totally relevant to my field"
06:33:03 <Bike> even though it wasn't & that's dumb
06:33:09 <Bike> and wrote a paper about it.
06:33:10 <Gracenotes> shachaf: you still haven't given me something funny and specific you don't like
06:33:18 <Taneb> How was that limerick
06:33:37 <shachaf> Gracenotes: oh the limerick i was talking about before:
06:33:51 <Gracenotes> Taneb: good, maybe change last line to "His toupe perked up in fright"
06:34:02 <elliott> "Double dactyl verse form is, perhaps unsurprisingly, rare in popular music."
06:34:04 <Gracenotes> ...makes about as much sense? or less, which is better.
06:34:17 <shachaf> A Swede who was in #esoteric / Thought his rhymes were a little generic. / "I might use, in my prose, / ꙮs, / But my poetry's alphanumeric."
06:34:32 <elliott> nice fucking the unicode up!!
06:34:52 <shachaf> because ꙮ is alphanumeric?
06:35:41 <Taneb> Is it bad that I read that little square as "multi-ocular o"
06:35:55 <Taneb> When it didn't render at all
06:35:57 <Bike> the worst backstoyr
06:36:03 <elliott> my terminal is broken I guess!
06:36:12 <shachaf> http://codu.org/logs/log/_esoteric/s is rendering it "just fine"
06:36:17 <Taneb> And I am on a mobile device
06:36:34 <Taneb> Which may or may not support unicode
06:36:47 <Gracenotes> okay, I think I've reached a local maxima with mine
06:37:03 <Gracenotes> There was a young critic named shachaf / Who put #haskell flags on a mast-half / When asked why he said / "I don't think it's yet dead, / but I hope it'll get there quite fast-fast"
06:37:27 <Bike> who doens't like #h a s k el ldl
06:37:35 <shachaf> imo import more hebrew to rhyme with "shachaf"
06:38:08 <Gracenotes> yeah I know, I don't want/do code-switching
06:40:34 <elliott> having a hard time coming up with an anti-biologist limerick here
06:42:10 <Bike> how about make fun of vitalism
06:42:45 <Taneb> There were some poetic motes/Penned by a chap named Gracenotes/The meter was sluggish/The rhymes were rubbish/And it wasn't even the worst thing he/she wrote
06:43:56 <Taneb> I can't think of anything that rhymes with Taneb
06:44:25 <elliott> Bike: can you finish this anti-biologist limerick for me
06:44:28 <elliott> ps it's also an anti-bike limerick
06:47:54 <Bike> there once was a man in nantucket / he watched birds and the bees and a duck pit / til one day at noon / they drained the lagoon / and right then and there he said "fuck it"
06:47:57 <Bike> best i can do elliott
06:48:01 <Taneb> There once was a boy called elliott/Who thought biologists smell-iot/He wrote a short rhyme/In limerick time/And said "with the last line to hell-iot"
06:48:11 <elliott> yiou were meant to finish MINE
06:48:19 <Bike> wait where's yours
06:48:28 <elliott> ok here you go "there once was a biologist named Bike / who found his head stuck on a pike. / the previous day / channel regulars say /"
06:48:55 <Bike> does that first line even fit the meter...
06:49:24 <shachaf> Taneb: can you please fix your meter thx
06:50:07 <Gracenotes> Some physicists sought to compete / With biologists who lived down the street / The biologists leisured / Since they couldn't be measured / Because all of their work was discrete.
06:50:28 <Bike> dude did you rhyme "leisured" with "measured"
06:50:30 <Bike> what kind of accent is taht
06:50:53 <Bike> oh right people say it like "ledgered"
06:51:11 <shachaf> Bike: did you rhyme "nantucket" with "duck pit"
06:51:36 <elliott> There once was a beautiful summer / fuck fuck fuck fuck fuck fuck fuck / fuck fuck fuck fuck / fuck fuck fuck fuck / fuck fuck fuck fuck fuck fuck fuck fuck
06:51:39 <Gracenotes> pretty good, eh. not exactly a searing punchline, but... it's basically true...
06:51:55 <Bike> shachaf: hey that works! almost
06:51:55 <mnoqy> id you rhyme summer with fuck
06:52:00 <Bike> the real flaw with mine is that it's terrible
06:52:03 -!- oerjan has set topic: Vogon poetry's got nothing on us | <3 | http://underhanded.xcott.com/?page_id=5 | logs: http://codu.org/logs/_esoteric and http://tunes.org/~nef/logs/esoteric/?C=M;O=D.
06:52:06 <Bike> just in general
06:52:44 <shachaf> oerjan: hey my original one was good
06:53:01 <shachaf> that was the best limerick of all
06:53:05 <Gracenotes> I'm not sure how to make one criticizing physicists without going for the obvious stuff (like cow spheres)
06:53:08 <mnoqy> maybe i should write a lickermich
06:53:12 <Taneb> A man once insulted my metrical technique/Despite my skills being at humanity's peak/So during the night/Purely out of spite/I killed him as he slept under a sheet
06:53:49 <Bike> Gracenotes: i'd be impressed to see a limerick criticizing renormalization
06:53:51 <elliott> it actually hurt to read that
06:54:20 <oerjan> There once was an expert on brainfuck / how the fuck do you rhyme with brainfuck
06:54:39 <Gracenotes> 'summer' rhymes with 'brianfuck' I think
06:55:33 <elliott> yo what's a bad thing you can do to someone that ends with -ition
06:55:55 <Bike> being petitioned sucks
06:56:06 <Gracenotes> (in Christian theology) A state of eternal punishment and damnation into which a sinful and unpenitent person passes after death.
06:56:08 <elliott> there once was a mathematician / fit only for eternal perdition / on type theory he said / "I'd rather be dead / than formulate proofs with precision"
06:56:24 <shachaf> stop it with the bad rhymes
06:56:32 <elliott> i bet mnoqy likes my limerick
06:56:35 <shachaf> better to not rhyme at all
06:56:43 <elliott> tell me how you really feel
06:56:45 <Bike> i think mathematician/perdition is ok
06:56:52 <Taneb> There once was an expert on brainfuck/On whose shoulder sat a plain duck/He looked at it and said/"Hop on to my head/"And this will be the best day of summer"
06:57:10 <oerjan> shachaf: i'm sorry, that _does_ rhyme. it just doesn't scan.
06:57:12 <Bike> Fuck the last line
06:57:25 <shachaf> oerjan: no, it doesn't do nothin'
06:57:27 <Bike> -ion is enough for it
06:57:35 <Bike> it seems like it scans to me...
06:57:45 <Bike> i guess 'eternal' is too long
06:57:58 <elliott> instead of "eternal perdition" i originally had "harsh extradition
06:57:59 <Bike> how about 'utter'
06:58:03 <mnoqy> (thinks) how do i make the rhyme (stops thinking)
06:58:24 <Taneb> A dangerous mission
06:58:26 <Bike> yeah, i think 'utter' works.
06:58:30 <Bike> you're welcome, elliott.
06:58:31 <shachaf> ok how about some drugz rhymes
06:58:33 <Gracenotes> A kitty did mew in the night / For her hunger did hit with such fright / She did look to her mother / As there was no other / With nipples
06:58:43 <Gracenotes> you had to expect an anti-limerick at some point, yes
06:58:55 <mnoqy> haven't half of these been anti limericks
06:59:31 <Gracenotes> I wanted to choose the most saccharine topic and the most predictable of subversions
06:59:50 <Bike> too many tildes help
07:00:11 <elliott> there once was a hoover ethereal / who despised derivative material / on the wiki he saw / some brainfucking lore / and gave brick-brain punishment imperial
07:00:33 -!- shachaf has left.
07:00:47 <mnoqy> there once was a dude called shachaf / now hes gone / rip
07:01:08 <Bike> elliott: you seem to like packing syllables into that second line
07:01:13 <Gracenotes> New rule: no limericks allowed for the next 2 months
07:01:15 <Bike> what did it ever do for you to deserve such bounty
07:01:22 <Gracenotes> all in favor yell angrily at your screens
07:01:55 <Bike> too tired for yelling
07:02:32 <elliott> there once was a lover of cereal / who sought to define the material / on sourcereal.com / they dropped the truth bomb / "it is something made from cereal"
07:02:59 <Taneb> There once was an onslaught of tilde s/By Which a poetry veteran was killed/Our painful rhymes/And meter less than sublime/Only stood to give him a pretty good thrill
07:03:09 <Bike> something that's made out of cereal
07:03:21 <elliott> Bike: http://sourcereal.com
07:03:28 <elliott> have to stay as true to the source material as possible
07:03:46 <mnoqy> Definition of Sour Cereal
07:03:46 <mnoqy> Cereal can be defined as something that is made from cereal. One can put it in a bowl, on a plate, or leave it in a box. Cereal is a nice thing to think about and sometimes eat. Sour cereal is an example of a type of cereal with spices.
07:04:08 <Taneb> Brb,.getting cereal
07:05:23 <Bike> <META NAME="description" CONTENT="This is a page about cereal, specifically sour cereal.">
07:05:24 -!- kallisti has quit (Ping timeout: 240 seconds).
07:09:16 -!- Taneb has quit (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com )).
07:24:05 -!- kallisti has joined.
07:24:05 -!- kallisti has quit (Changing host).
07:24:05 -!- kallisti has joined.
07:26:58 -!- Taneb has joined.
07:32:17 <Taneb> I enjoyed my cereal
07:48:29 <oerjan> i think badly scanning poetry is from now on a bannable offense hth
07:48:53 <oerjan> i'm sorry, elliott, that means you cannot write poetry any more
07:50:50 <elliott> oerjan: no, it just means I'll be banned from it. but happily I have the power to remove bans!
07:50:58 <elliott> or maybe not without +R. fuck &, also, shit
07:51:29 <oerjan> repeated offenses _may_ affect your ability to unban hth
07:51:59 <oerjan> it's just protecting sentient life in the universe, i hope you understand this is necessary
07:52:40 -!- MindlessDrone has joined.
07:54:49 <kmc> i missed a lot though
07:58:48 <kmc> hi elliott
07:58:51 <kmc> shachaf left :/
07:59:13 <kmc> i was about to nominate him as poet laureate
08:04:58 <fizzie> It's special-characters hour here at the #esoteric.
08:06:04 <fizzie> ıf you w®ıþ€ wıþh əlþ-ŋ® ðowñ oñ þhıš ĸ€yßoə®ð ləyouþ, þh€ €ñ𠮀šulþ šþıll ĸıñð of looĸš lıĸ€ ®€ŋulə® þ€×þ.
08:06:29 <mnoqy> imo thats kind of like badly scanning poetry
08:07:02 <fizzie> The use of altgr-b for the not-beta-but-German-ss ß is kind of rude. "Looks like B" indeed.
08:12:00 <kmc> it's Compose S S
08:13:15 <kmc> ſ is Compose F S which seems... all right?
08:13:57 <oerjan> ("if you write with alt-gr down on this keyboard, the result is rather disappointing.)
08:25:23 -!- kiview has joined.
08:25:49 -!- kiview has left.
08:27:30 -!- carado has joined.
08:30:38 -!- Taneb has quit (Ping timeout: 240 seconds).
08:53:56 -!- epicmonkey has joined.
09:00:51 -!- Phantom_Hoover has joined.
09:01:28 <oerjan> oh, this simon tatham puzzle was made by ben olmstead
09:03:29 <mnoqy> is simon tatham a brand
09:04:37 <oerjan> he made putty and this puzzle collection, both of which i use
09:23:23 <elliott> oerjan: you won't be using one of those for long, hth
09:23:56 <fizzie> That sounds like a threat hth
09:24:22 <elliott> yes. I'm going to threaten him with a linux installation CD
09:45:23 <fizzie> LIBDBUSMENU-GLIB-WARNING **: Trying to remove a child that doesn't believe we're it's parent.
09:50:29 <elliott> that happened to mnoqy too
10:01:10 -!- kallisti has quit (Ping timeout: 248 seconds).
10:05:00 -!- epicmonkey has quit (Read error: Operation timed out).
10:13:37 -!- nooodl has joined.
10:27:41 <Deewiant> Bike: ghc -e 'print (product [1..905381] :: Integer)' | wc -L ==> 5000004
10:49:21 -!- oerjan has quit (Quit: LAter).
10:50:57 <elliott> coppro: hi I have clang questions
10:51:20 <elliott> coppro: 1. is clang's output without any -O at all as completely-terrible as gcc's is i.e. do I always want at least -O or -Os or whatever
10:51:29 <elliott> coppro: 2. does clang give worse warnings without -O like gcc does
10:52:38 <Deewiant> 1. Yes to the part before the i.e., I think, but definitely no to the part after it
10:53:10 <elliott> have you seen gcc's -O0 output
10:53:13 <Deewiant> 2. I'm pretty sure that the answer is no
10:53:47 <Deewiant> I've probably seen some of it at some point
10:54:38 <elliott> well I don't see how you could stand ever using it :P
10:54:52 <elliott> it's like the output of a toy C compiler you would write in a weekend for fun
10:55:01 <Deewiant> Do you compile stuff in order to look at the generated assembly?
10:55:12 <Deewiant> I think clang only uses FastISel at -O0
10:55:18 <elliott> no but I don't want to feel like I'm running a Python program :(
10:55:23 <Jafet> Do you ever want to read -O0 assembly
10:55:34 <elliott> it does shave 15 seconds off the build time though
10:55:39 <Deewiant> Which means it should be significantly (once your program is big enough) faster to compile stuff at -O0
10:55:42 <elliott> which is nice because I'm really lazy
10:55:54 <elliott> well, compared to -O2, a bit less compared to -Os
10:56:06 <Deewiant> Isn't -O equal to -O2 in gcc and hence clang
10:56:30 <elliott> pretty sure -O1 is a thing
10:56:34 <Jafet> elliott: python cannot run, it slithers
10:56:39 <elliott> or was -O made to alias to -O2 at some point
10:56:46 <fizzie> -O is equal to -O1, I believe.
10:57:06 <elliott> also I think clang doesn't support gcc's fancy -ggdb3 stuff
10:57:10 <Deewiant> I thought it was -O2 ever since the -O[123] were introduced
10:57:21 <elliott> and apparently it doesn't do any more than line numbers if you use anything other than -O0???
10:57:21 <Deewiant> But I'm probably thinking of a different compiler
10:57:28 <elliott> can I have a compiler refund
10:57:37 <Deewiant> Any more than line numbers in what?
10:57:51 <elliott> -g Generate debug information. Note that Clang debug information
10:57:51 <elliott> works best at -O0. At higher optimization levels, only line number
10:57:51 <elliott> information is currently available.
10:57:56 <elliott> admittedly, this is more reason to use -O0
10:57:57 <Jafet> The return shipping for your gnu is one gnu.
10:58:14 <Deewiant> I don't think I've ever used debug information for anything other than line numbers
10:58:26 <elliott> well I just like knowing there's something fancy involved you know
10:58:39 <elliott> I think -ggdb3 made something perceptibly nicer when I switched to it with gcc
10:58:46 <Deewiant> Which is actually a bit surprising now that I think about it, but oh well
10:59:17 -!- rom__ has joined.
11:01:08 -!- rom__ has quit (Client Quit).
11:02:04 <Jafet> Is there any other debugging information than line numbers and symtab
11:02:06 <elliott> oh good that made the build slower
11:02:21 <Jafet> Perhaps your ramdisk is swapping
11:08:29 -!- epicmonkey has joined.
11:12:01 -!- shachaf has joined.
11:34:49 -!- ais523 has joined.
11:35:01 <ais523> so, I came across a fun pathological function in my research yesterday
11:35:27 <ais523> that I'm having trouble expressing in any programming language (other than languages as low-level as, say, VHDL)
11:35:47 <ais523> when you call it, it returns two functions/closures, let's call them f() and g()
11:36:20 <ais523> err, there are side effects involved here, so let me try to be clear about the types
11:36:36 <ais523> my function is (IO ()) -> (IO (), IO ())
11:36:44 <ais523> not very good at Haskell
11:37:00 <elliott> maybe it should be -> IO (IO (), IO ())?
11:37:26 <ais523> no, I'm reasonably sure it isn't, in this case; there are no side effects involved until you actually execute the return values
11:37:57 <elliott> so it's just two IO () -> IO () functions
11:38:20 <ais523> yep, it's equivalent to that, although conceptually weird thinking about it
11:38:35 <ais523> maybe I do need an extra IO in order to avoid having to compare functions
11:38:53 <elliott> um, that sounds implausible but ok
11:38:55 <ais523> I guess it is IO () -> IO (IO (), IO ()), just so that the two return values can be aware of each other
11:39:18 <ais523> anyway, if you call either of the return values, then the original argument starts executing
11:39:39 <ais523> and the return values stop executing when either the original argument stops executing, or the other return value is called
11:39:52 <ais523> as in, while the original argument is executing, there has to be at least one of the return values executing
11:39:54 <ais523> but they can take it in turns
11:40:11 <Deewiant> Stop as in they can continue or as in they die
11:40:28 <ais523> so, e.g., I call pathological(delay 1000), that gives me two functions f and g
11:40:40 <ais523> if I call f, then it won't return for 1000 time or until I call g
11:40:52 <ais523> if I call g, then f returns, now g won't return until the original 1000 time is up or I call f again
11:40:54 <elliott> conflation of functions and programs upsets :(
11:41:20 <ais523> actually in Haskell this doesn't need two functions, because you could call the same function simultaneously from different threads
11:41:50 <ais523> if you spend a few days staring at linear logic, you start ceasing to be able to comprehend the idea of copying data
11:42:20 <ais523> so I could just make it (IO ()) -> IO (IO ())
11:42:27 <shachaf> So this is like a weird version of spawn?
11:42:49 <ais523> it feels like a concurrency primitive, but it doesn't match any of the concurrency primitives I've ever seen
11:43:05 <ais523> it was a counterexample to something I was trying to prove
11:43:19 <shachaf> Such that if you run the action from one thread, it returns in any other thread that's running it?
11:43:58 <ais523> or well, the idea is that once you start it running, it finishes, but threads can pretty much change /which/ thread is running it arbitrarily
11:44:42 <ais523> the analogy I use as a mental model of it is, you start some long-running task that needs someone to constantly supervise it
11:44:43 <shachaf> Instead of being an action can this be a sort of lockish thing?
11:44:50 <ais523> but you can change who's supervising it
11:45:09 <shachaf> Such that you "wait" on it and then someone else can wait on it instead.
11:45:21 <ais523> shachaf: I tried to implement it with locks, but failed, and it's quite easy to prove (using the same research that discovered it in the same place) that that's impossible with a static concurrency model
11:45:47 <ais523> (e.g. one where each fork() has to have a matching wait() called from the same thread)
11:47:41 <ais523> I guess you could handle it with an integer semaphore and a thread that's happy to clean up for itself and doesn't need to report back to a parent
11:48:58 <quintopia> where is that asshole Phantom_Hoover
11:49:00 <ais523> hmm… now I'm tempted to add this as a concurrency primitive to INTERCAL
11:49:17 <ais523> it fits the action-at-a-distance flavour of INTERCAL quite well
11:49:28 <quintopia> he triggered the pbflist and there is no new pbf. false positives are a waste of my time!
11:49:40 <shachaf> quintopia: That's what I said!
11:50:07 <shachaf> Phantom_Hoover: Oh, you're here now. Don't do the thing you did.
11:50:37 <HackEgo> pbflist: shachaf Sgeo quintopia
11:50:54 <shachaf> You're just being annoying, you know.
11:51:45 <Phantom_Hoover> i think on some level you're to blame for taking the `list system seriously
11:52:59 <quintopia> any system for systematically pinging people should be taken seriously
11:53:12 <Vorpal> ais523, what use is this new concurrency primitive?
11:53:14 <quintopia> because pinging people for no reason is annoying. it's also called "spam"
11:53:22 <ais523> Vorpal: I don't know that it's useful
11:53:25 <quintopia> the people who do it are called "spammers"
11:54:06 <Vorpal> ais523, Are you sure you couldn't implement it with other primitives anyway? I believe I can think of a way, though it isn't wait-free.
11:54:10 <elliott> i suggest a collective dropping of the topic
11:54:39 <ais523> Vorpal: as I said, it's provable that you can't do it with a static concurrency model, but if you have more general concurrency models it's implementable
11:54:55 -!- FreeFull has joined.
11:55:44 <Vorpal> ais523, I admit I'm not an expert on this area, and I'm not quite sure what a static concurrency model is, but I'm thinking in terms of pthreads with semaphores message queues.
11:57:08 <Vorpal> Or maybe not. I'll have to work out the details
11:58:25 <ais523> I think pthreads doesn't have any particular "parent" / "child" division, or at least lets children outlive their parents
11:58:28 <ais523> so it's probably workable
11:58:36 <ais523> it's surprisingly nontrivial even if your language is expressive enough, though
11:58:39 <Vorpal> Well yes, is that a requirement for static concurrency?
11:58:45 <Vorpal> I suspect you could do it in terms of a couple of semaphores and mutexes in fact.
11:59:25 <ais523> Vorpal: that's pretty much what static concurrency is
12:00:37 <Vorpal> Here is how I would do it: Transform the task into three functions: f,g and h, have h execute the actual task, in a separate thread, have f and g wait on a signaling semaphore, these can either be signaled by h completing, or the other function of f or g being called. You will need to refine it a bit to handle two functions calling f, or f/g being called after h completes but the idea should work I beli
12:01:16 <Vorpal> ais523, unless I misunderstood the problem that seems like a pretty straight forward implementation in a pthreads style environment
12:01:20 <ais523> Vorpal: yeah, I think that works
12:01:58 <Vorpal> I believe it is an utterly useless thing though. :)
12:02:47 <elliott> kmc: do you have any particular recommendations wrt what ThinkPads are good to buy (I realise this is hopelessly vague)
12:02:47 <Vorpal> out of interest ais523, what else defines a static concurrency model?
12:03:07 <shachaf> elliott: kmc has a thinkpad x1 carbon and it looks p. nice
12:03:24 <FreeFull> I love STM, but people say there might be something better possible
12:03:28 <ais523> Vorpal: well it basically means that your only thread-creation primitive is a "run X and Y in parallel" function, or can be implemented in terms of that
12:03:33 <ais523> which doesn't return until both X and Y return
12:04:16 <Vorpal> ais523, But you are free to do whatever in terms or semaphores, message queues, critical sections and so on?
12:04:30 <shachaf> Oh well, he helped me with a thing in /msg the other day, he can keep the extra karma.
12:05:12 <elliott> shachaf: /set window_history on
12:05:15 <ais523> Vorpal: yes, there's no restriction on synchronization primitives
12:05:23 <Vorpal> Also btw, I never understood the difference between a critical section and a mutex. They just seems like two ways to express the same thing?
12:05:36 <elliott> Vorpal: a mutex is one way to implement a critical section
12:05:49 <elliott> a critical section is a semantic property rather than some specific implementation
12:06:08 <shachaf> elliott: Why do I want that?
12:06:17 <elliott> shachaf: it makes up-arrow work better
12:06:26 <elliott> also re: X1 Carbon, that is one of the models I am looking at
12:06:44 <shachaf> I frequently e.g. /msg things to lambdabot and later up-arrow in another window.
12:06:53 <shachaf> Well, I suppose I could up-arrow in the lambdabot window.
12:07:01 <shachaf> I'll keep it like this for now.
12:07:12 <shachaf> (The reason for the mis-up-arrow is that my palm touched my touchpad.)
12:07:13 <ais523> huh, I didn't realise there were clients that didn't have per-window history
12:07:14 <elliott> the resolution on them is kind of bad though it seems :(
12:07:22 <shachaf> (Which scroled the scroll-wheel, etc.)
12:07:27 -!- oonbotti2 has quit (Ping timeout: 240 seconds).
12:08:10 <Vorpal> ais523, you could emulate non-static concurrency in a static concurrency using a concept of worker thread pools? Just start x number of threads and make them never complete. Now you can use message queues to submit jobs, and then you can implement the weird primitive above
12:08:15 <Vorpal> Or am I missing something?
12:08:51 <elliott> shachaf: also they're p. expensive
12:09:02 <Vorpal> You can just emulate a non-static concurrency on top using semaphores to signal "completion" of your threads and so on
12:09:18 <ais523> Vorpal: oh, in my case that didn't work because we can't send anything more complex than an int
12:09:27 <ais523> the perils of hardware compilation
12:09:41 <ais523> I guess you could translate everything into bytecode and then interpret it at the other end :)
12:09:49 <shachaf> elliott: you should get the clevo W740SU so you can tell me whether it's any good hth
12:09:59 <elliott> ideally I would like something with an ultrabook form factor and battery life and workstation-replacement specs and netbook price........
12:10:06 <Vorpal> ais523, anyway, the concept of static concurrency seems a bit ill-defined if you can just emulate non-static concurrency on top of it.
12:10:19 <ais523> Vorpal: it's ill-defined if the rest of your system is also ill-defined
12:10:42 <ais523> if you have a sufficiently well-defined model of what you can and can't do, it's quite easy to define static concurrency too
12:10:46 <quintopia> elliott: i want all that too! let me know when those are invented.
12:10:49 <ais523> `quote removing concurrency
12:10:51 <HackEgo> 236) <ais523> gah, who'd have thought removing concurrency from algol could be so difficult
12:10:59 <Vorpal> ais523, well okay, is having full blown semaphores and message queues ill-defined?
12:11:35 <ais523> Vorpal: if you even have to ask the question, rather than having an answer ready to hand, it means that your system as a whole is ill-defined
12:11:53 <ais523> like, does this system have higher-order functions? does it have closures?
12:12:05 <ais523> in particular, can we send a closure through a message queue and get some sort of useful result?
12:12:22 <ais523> (these problems are nontrivial, e.g. compare "can I send a filehandle over a network connection?")
12:12:35 <ais523> (in most programming languages, the answer to that is "no"; in CLC-INTERCAL, it's "yes")
12:12:37 <Vorpal> In unix you kind of can
12:12:55 <Vorpal> Erlang allows that fully across all nodes of course
12:13:12 <Vorpal> ais523, you wouldn't need to send closures though, as you said above yourself, bytecode.
12:13:42 <ais523> Vorpal: well, now you reach the question of "is this language capable of decompiling arbitrary functions that are passed in?"
12:13:58 <ais523> again, in some languages (e.g. Java), the answer is "yes", but it's typically "no"
12:15:53 <Vorpal> ais523, anyway, you could even in C just implement an interpreter for your own byte code. Or you might not even need that. Something like an enum defining predefined tasks and some arguments to it could be enough to implement a limited number of tasks if that is all you need.
12:16:14 <elliott> shachaf: also they only have 8 gigs of RAM :'(
12:16:37 <shachaf> elliott: Which one, the X1?
12:17:06 <shachaf> But back in the day you couldn't even get i7+8GB at once. You had to choose 4GB or i5.
12:17:26 <shachaf> Dark old days when kmc was looking at it.
12:17:35 <ais523> "back in the day" "4 GB"
12:17:54 <ais523> I never really used computers in the "640KB is enough for everyone" era
12:18:08 <shachaf> elliott: imo what's not to like about that clevo thing, 4-core haswell cpu, 1920×1080, <2kg, supports 16GB ram, no nvidia optimus
12:18:21 <ais523> but I did use them in the era when 640KiB was the base memory and you needed to use unusual APIs to access the other 2MiB, with many programs stuck in the bottom 640
12:18:28 <elliott> which one is the clevo again
12:18:29 <shachaf> except i'm not sure if it's actually good. this is where you come in
12:18:42 <shachaf> the one system76.com is selling among other places
12:18:49 <Vorpal> elliott, if you want a powerhouse computer, the business Dells laptops are quite good. Not very portable, and pretty bad battery life. Matte screen and trackpoint though.
12:19:45 <elliott> Vorpal: not really what I'm looking for, I want the laptop part more for the portability indoors than for it being a desktop you can lug around elsewhere
12:19:52 <elliott> shachaf: rightb ut I can't find a clevo on system76.com
12:19:52 <fizzie> Hey, I have a Clevo-based laptop.
12:19:56 <Vorpal> ais523, I used my dads Mac back then. I never had to deal with that weirdness
12:20:12 <shachaf> elliott: https://www.system76.com/laptops/model/galu1 hth
12:20:19 <Vorpal> elliott, yeah I think it was months since I took out my work laptop of the docking station XD
12:20:19 <fizzie> It's from the 2011s, though; and big-and-heavy.
12:20:28 <ais523> Vorpal: well by then, packagers had got used to the weirdness
12:20:33 <shachaf> elliott: clevo is the manufacturer's name hth
12:20:36 <ais523> and gave simple instructions about what to put in config.sys
12:20:56 <Vorpal> ais523, but wouldn't it possibly break between different programs you wanted to run?
12:21:05 <elliott> wonder if the touchpad is any good
12:21:39 <ais523> Vorpal: not really, by that stage it was mostly a case of increasing minimum settings
12:21:48 <ais523> if you set the settings too high for a program, it'd still run just waste memory
12:22:01 <shachaf> elliott: see, that's the kind of question i need you for answering
12:22:06 <Vorpal> ais523, waste memory that could have been used for what?
12:22:06 <ais523> and that was OK because the programs that could make do with lower settings normally were written for older computers anyway which had less memory
12:22:21 <ais523> putting the settings up higher meant the kernel had to use them
12:22:29 <ais523> err, the kernel had to use more memory
12:22:32 <ais523> leaving less for userspace
12:22:47 <Vorpal> ais523, didn't some DOS programs put the computer in a 32-bit mode while running?
12:22:59 <fizzie> http://www.avadirect.com/gaming-laptop-configurator.asp?PRID=19611 <- I have this thing | it is so big.
12:23:08 <elliott> shachaf: are there any system76 display models in stores anywhere
12:23:36 <ais523> Vorpal: there were "DOS extenders" that put the system into protected mode, then turned it back into realmode whenever something tried to make a system call, so that DOS would work correctly
12:23:38 <Vorpal> fizzie, 17" heh, yeah those are big
12:23:40 <elliott> how lenient is their return policy
12:23:47 <ais523> and which handled some 32-bit system calls themselves
12:23:56 <ais523> NetHack for DOS works that way, for instance
12:24:16 <shachaf> elliott: Here's Clevo's own web page on what I'm told is the same thing: http://www.clevo.com.tw/en/products/prodinfo.asp?productid=472
12:24:25 <fizzie> (I think. At least a lot of games come with DOS4GW.)
12:24:27 <ais523> fizzie: I had to mess around with DOS extenders for making the NetHack TAS
12:24:41 <ais523> including fixing the emulator so that it could handle at least one of htem
12:25:25 <shachaf> almost no one has haswell so far
12:25:33 <Vorpal> what is new in haswell? Even stupider instruction names?
12:25:49 <shachaf> i hear you should wait until september or so if you want reasonable haswell options
12:25:55 <fizzie> Aren't there some Apples with Haswells?
12:26:00 <elliott> the macbook airs have haswell
12:26:12 <elliott> unfortunately the pros don't, I was hoping to see what the retina would be like with haswell
12:26:21 <elliott> since that could be a pretty compelling option for me
12:26:31 <elliott> wow this system76 clevo is cheap
12:26:47 <shachaf> elliott "the x1 carbon is p. expensive" elliott
12:26:57 <fizzie> Vorpal: Are you saying VPUNPCKHQDQ is a stupid name!
12:27:12 <elliott> it only has 2.0 ghz so it's going to feel very slow no matter how fast it is :'(
12:27:21 <Vorpal> btw, what does that one stand for?
12:27:34 <elliott> help, which kind of SSD is the best
12:27:44 <shachaf> the most expensive one hth
12:27:52 <Vorpal> elliott, I like the Intel 520 SSD at work. 250 GB
12:28:23 <shachaf> Vorpal: vex parallel unpack something something hth
12:28:56 <Vorpal> fizzie, does sandy bridge have that one? I want to test it out
12:28:56 <elliott> 16 gigabytes of RAM and 256 gigabyte SSD and 1 TB HD for $1432...... that's pretty good
12:29:07 <Vorpal> fizzie, to see if it is the right choice for me
12:29:14 <fizzie> Vorpal: V for "vector" (probably), P possibly as a prefix for "packed", then UNPCK for unpack, and HQDQ for "High Quadwords to Double-Quadword".
12:29:26 <fizzie> Vorpal: And no, it was an AVX2 instruction.
12:29:38 <shachaf> fizzie: Doesn't V stand for "three-operand"?
12:29:40 <fizzie> Vorpal: You probably do have the plain PUNPCKHQDQ though.
12:29:59 <fizzie> shachaf: Oh, sure, "thVree-operand", of course.
12:30:13 <shachaf> https://en.wikipedia.org/wiki/VEX_prefix
12:30:15 <Vorpal> elliott, both SSD and HDD? Must be a pretty heavy laptop
12:30:46 <fizzie> shachaf: I still guesstimate that the 'V' in the name originally derives from the word vector.
12:30:50 <shachaf> fizzie: Anyway if PFOO exists in SSE and you have plain old AVX isn't that enough to be able to use VPFOO with SSE registers?
12:30:55 <elliott> Vorpal: 1.72 kg by default apparently
12:30:59 <elliott> Vorpal: default = HD and no SSD
12:31:16 <elliott> doubt the SSD is heavy enough to make the total not light
12:31:31 <fizzie> shachaf: Only the floating-point stuff is in plain old AVX, I think.
12:31:40 <fizzie> shachaf: AVX2 extends to integer datatypes.
12:32:09 <elliott> my current laptop is like uh
12:32:12 <shachaf> fizzie: Yes, but if you just need SSE instructions, i.e. on xmm registers, I think you can use the VEX prefix.
12:32:18 <elliott> so I bet this one would feel super heavy :(
12:33:16 <Vorpal> elliott, I think mine is closer to 2.5 kg or something
12:33:46 <fizzie> shachaf: Look, all I know is that VPUNPCKHQDQ is in http://software.intel.com/en-us/blogs/2011/06/13/haswell-new-instruction-descriptions-now-available
12:34:28 <shachaf> fizzie: Maybe they mean that it works on YMM registers now?
12:34:42 <elliott> shachaf: what is your current laptop
12:35:19 <elliott> i hope the blurriness of the screen in https://c12278716.ssl.cf2.rackcdn.com/galu1-e791e840dd9d7623d0228ca15867ae74.jpg is just a bad photo
12:35:35 <elliott> oh it's not even blurry there
12:35:40 <elliott> it's just downsized on the web page so it looks blurry?
12:37:22 <shachaf> fizzie: Anyway I could've used AVX2 back when I was writing some SSE code. :-(
12:37:51 <fizzie> "Haswell, will travel."
12:38:40 <shachaf> elliott: anyway tell me if you find a better computer than that one along the relevant axes
12:42:39 <fizzie> ais523: Are there any INTERCAL implementations that use the Haswell PDEP/PEXT instructions for select/mingle?
12:45:30 <ais523> fizzie: not as far as I know
12:45:56 <ais523> perhaps it could be added as an alternative in C-INTERCAL's libick
12:46:10 <ais523> it /would/ be the sort of thing I'd like doing, but I wouldn't be able to test it
12:46:32 <elliott> do you need someone to donate you a haswell computer
12:46:58 <ais523> elliott: if they just donate me a patch, and convincingly claim it works
12:47:11 <elliott> ais523: there are lazy people with money
12:48:50 <ais523> how many of them care about haswell-accelerated INTERCAL, though?
12:49:15 <Vorpal> elliott, seems haswell is less energy efficient than ivy bridge hm
12:49:22 <Vorpal> https://en.wikipedia.org/wiki/Haswell_(microarchitecture)#Performance
12:49:31 <Vorpal> And harder to overclock
12:49:32 -!- johnny57 has left.
12:50:07 <elliott> Vorpal: isn't it meant to be much more energy efficient
12:50:36 <Vorpal> Well I don't know, I just pointed you to the source I saw it on. It could be that it is more energy efficient for the given performance
12:51:18 <elliott> not like anyone wants to overclock a laptop anyway
12:51:28 <Vorpal> Yay wikipedia is quoting two sources that are both the same text???
12:51:35 <elliott> The Haswell architecture is specifically designed[5] to optimize the power savings and performance benefits from the move to FinFET transistors on the improved 22 nm process node.[6]
12:52:21 <ais523> Vorpal: probably an attempt by one of them to drive traffic to their site
12:52:22 <elliott> https://en.wikipedia.org/wiki/File:Haswell_Chip.jpg what causes the repeated patterns on chips like these
12:52:28 <elliott> like I assume they're a bunch of duplicates of the same thing?
12:52:49 <Vorpal> elliott, isn't that a waffer with multiple CPUs?
12:52:51 <ais523> elliott: most commonly it's RAM (e.g. L2 cache)
12:52:59 <ais523> if it's a large repeating area
12:53:15 <elliott> unfortunately we will never know because ais523 won't click it :P
12:53:17 <Vorpal> elliott, but inside each CPU there are repeating patterns as well yes
12:53:32 <ais523> although that particular image may just be multiple chips
12:53:37 <elliott> I have a hard time believing the entire CPU is much smaller than a pin
12:53:42 <Vorpal> elliott, well it says waffer in the description below
12:54:16 <ais523> elliott: basically chips have to be that small, a single impurity makes the entire chip unusable
12:54:18 <Vorpal> elliott, I would guess the CPU is about as bit as the pin, that is, the largest scale repeated structure is one CPU each.
12:54:23 <ais523> so they need to make them small to fit between impurities on the wafer
12:55:00 <elliott> Vorpal: well it would be awfully thin if so
12:55:12 <Vorpal> ais523, isn't that the point of three core CPUs? One of the cores had an impurity in it?
12:55:23 <Vorpal> elliott, not sure that is unlikely though
12:55:38 <fizzie> Haswell GT2's die size is listed as 177 square millimetres; that's a 13 mm x 13 mm square, if square.
12:55:50 <ais523> they're failed attempts to make quad core CPUs
12:56:09 <Vorpal> ais523, presumably some of the dual core CPUs are quad cores with two failed cores as well?
12:56:23 <elliott> AMD still does 6-core CPUs right? what's up with that?
12:56:27 <ais523> Vorpal: that's less likely, that'd be rare enough that it wouldn't be worth their time to separate them out and market them
12:56:28 <elliott> 8-core CPUs where two failed??
12:56:30 <ais523> they'd probably be thrown away
12:56:34 <ais523> elliott: they might just be made as 6
12:56:55 <elliott> I guess you don't really get 8 cores on a single CPU much
12:58:02 <Vorpal> elliott, anyway yeah, I suspect within each CPU, there is L2 cache in one area of the chip and then 4 cores, then some shared stuff at the top. That would fit with other (annotated) images I have seen of dies.
12:58:54 <fizzie> Vorpal: The GPU is a big block there too, probably.
12:59:00 <Vorpal> elliott, the bottom part might be cache, then each of the 4 repeated bits could be a core, then at the "top" there is some common bus, power supply, clocks or similar. That is a rough guess.
12:59:08 <Vorpal> fizzie, yeah I suggested that above.
12:59:23 <fizzie> Vorpal: The mostly dark-red one, I think, based on the shapes of http://forums.anandtech.com/showthread.php?t=2234017
12:59:48 <Vorpal> elliott, anyway most of a physical CPU's area is just to allow for all the pins at a scale where the pins won't break if you touch them.
13:00:10 <Vorpal> fizzie, oh okay, so reverse of what I was thinking then
13:00:56 <Vorpal> wait, you said GPU not CPU sorry
13:01:01 <Vorpal> the words, they look similar
13:02:24 -!- Koen_ has joined.
13:04:05 <Vorpal> elliott, that is one reason why the companies try to put more stuff on the CPU (such as GPUs), less pins to go to external locations. (The other reason I can think of is less latency)
13:04:35 <Vorpal> Well that is an effect of less external pins
13:05:42 <Vorpal> SOCs (System-On-a-Chip), such as found in phones, takes that even further, often integrating sound, wifi, bluetooth, DSPs and so on as well.
13:06:35 <Vorpal> The PC is in a bit of a weird spot there, since adding the GPU onto the chip doesn't actually reduce number of pins currently. You still need that PCI express bus anyway.
13:06:53 <Vorpal> So latency would be the primary reason there I assume
13:07:22 <Vorpal> And reduced costs, not having to make separate integrated GPU chipsets (at least for Intel)
13:09:51 <Vorpal> ais523, What is the next step after 22nm?
13:10:15 <Vorpal> Also why did they go for 22 nm rather than 21 or 23? Or maybe 22.5?
13:12:00 <fizzie> 14 nm is the next one.
13:12:59 <fizzie> Then 10, 7 and 5, but that's a bit speculative?
13:13:02 <ais523> Vorpal: I don't really care, I just write compilers that design chips, I don't actually make them
13:13:33 <Vorpal> And still they only recently started taking advantage of 3D structures I believe? Previously it was mostly layered 2D right?
13:14:21 <Vorpal> Wouldn't it be more efficient to design a cubic CPU? That would make the signal paths shorter it seems to me? A bit more complicated cooling though I guess
13:15:10 <fizzie> A "bit" more complicated to manufacture using current lithography techniques, most likely.
13:15:59 <Vorpal> fizzie, how do you even do more than one layer currently?
13:16:28 <Vorpal> Deposit a new layer of silicon on top?
13:17:15 <fizzie> I don't know; another question along the same lines (where the answer is most likely "by being clever") is "how do they draw 20nm lines using light with wavelengths in the hundreds of nm?"
13:17:37 <Vorpal> They use electron beams nowdays iirc?
13:18:17 <fizzie> Well, it *exists*; I don't know what e.g. Intel uses at the moment.
13:18:25 <ais523> well even the really low tech photolithography that schoolchildren use to make PCBs uses ultraviolet
13:18:36 <ais523> that cuts your nanometres down a bit
13:18:38 <fizzie> "Intel has already outlined a path to use 193 nm immersion lithography down to 11 nm node" that's still light.
13:19:11 <fizzie> ("Immersion lithography is a photolithography resolution enhancement technique for manufacturing integrated circuits (ICs) that replaces the usual air gap between the final lens and the wafer surface with a liquid medium that has a refractive index greater than one. The resolution is increased by a factor equal to the refractive index of the liquid.")
13:19:15 <Vorpal> fizzie, I know they use double masking. Where they first use one mask then switch to another offset mask.
13:20:52 <fizzie> "Extreme ultraviolet lithography" has the greatest name.
13:20:54 <Vorpal> fizzie, hm: https://en.wikipedia.org/wiki/Three-dimensional_integrated_circuit
13:21:08 <Vorpal> Seems like it is in the works
13:21:34 <fizzie> Well, they have to do something.
13:22:10 <Vorpal> More cores? More optimised instructions? There are lots of way to continue increasing performance in the future
13:22:46 <Vorpal> Maybe an FPGA section of the chip that could be optimised for each application
13:24:16 <fizzie> Apparently Intel has done some forays to the EUL direction.
13:24:38 <Vorpal> oh, Extreme ultraviolet lithography?
13:25:08 <Vorpal> Isn't that into xrays already?
13:25:25 <Vorpal> "Extreme ultraviolet lithography (also known as EUV or EUVL) is a next-generation lithography technology using an extreme ultraviolet (EUV) wavelength, currently expected to be 13.5 nm."
13:25:45 <fizzie> Apparently it's still ultraviolet down to 10 nm.
13:25:54 <fizzie> Then x-rays from 10 to 0.1.
13:26:30 <fizzie> "The Intel Lithography Roadmap" requires a login. I wonder if I have an Intel account.
13:27:06 <ais523> why would a lithography roadmap require a login?
13:27:50 <fizzie> ais523: http://noggin.intel.com/content/the-intel-lithography-roadmap -> click link -> "To access this additional content, you must first log in. If you haven't yet registered, please sign up. It's quick, it's free, it's easy!"
13:27:59 <fizzie> I guess that's not an answer to the "why".
13:28:10 <ais523> yeah, that's merely a statement that the roadmap requires a login
13:28:21 <fizzie> "Lithography technologies, such as 193nm, 157nm, and EUV lithography, which have benefited from Intel investment, have gained industry acceptance, while competing technologies, such as x-ray lithography, are no longer being pursued."
13:28:26 <fizzie> I like the way they put it.
13:28:40 <elliott> fizzie: how good is the clevo's touchpad :P
13:28:46 <fizzie> elliott: I think it's p. standard.
13:29:44 <shachaf> ive never met a touchpad i didnt dislike.....
13:30:45 <Vorpal> Same, go for trackpoints.
13:31:13 <ais523> fizzie: that's a pretty evasive way to praise themeselves
13:31:43 <ais523> like, I'd be surprised if there were any ASIC manufacturing technologies that Intel hadn't put at least a bit of money into determining how well they worked
13:34:39 <fizzie> I think it sounds more like a threat. "Nice lithography technology you've got here. Would be shame if something happened to it, and it was no longer pursued."
13:35:09 <fizzie> Except they're the ones giving the money instead of getting it.
13:36:14 <Vorpal> Is there actually any sharp difference between 11 nm UV and 10 nm x-ray?
13:36:27 <Vorpal> It sounds more like an arbitrary line
13:36:44 <fizzie> There's a one nanometer difference hth
13:36:50 <fizzie> "That's pretty sharp."
13:37:04 <ais523> I think most of the distinctions between different sorts of light are pretty arbitrary
13:37:13 <ais523> even visible, because not everyone's eyes are identical
13:41:59 <fizzie> "Note that there are no precisely defined boundaries between the bands of the electromagnetic spectrum; rather they fade into each other like the bands in a rainbow (which is the sub-spectrum of visible light)." Some Wikipedia editor's being very poetic there.
13:42:07 <fizzie> (I was looking for standard definitions.)
13:43:59 <fizzie> ITU apparently has standard names, numbers and frequency boundaries for things that can be called radio, from EHF to ELF, but don't extend outside those bounds.
13:48:06 <fizzie> Oh, the low-frequency end only starts from ULF, officially. (It goes ULF, VLF, LF, MF, HF, VHF, UHF, SHF and EHF.)
13:48:45 <fizzie> Wait, no; there is ELF, but it's four bands wide, and in a separate table; and SLF is missing. Way to be logical, guys.
13:49:08 <ais523> I thought ELF was an executable format
13:50:26 <shachaf> Ultra-Linkable Format, Very Linkable Format, Linkable Format, ...
13:50:47 <fizzie> The Super-Linkable format is missing.
13:51:00 <fizzie> Fortunately, the Extremely Linkable Format is objectively better.
13:51:01 <ion> High Frequency, Very High Frequency, Unusually High Frequency, Surprisingly High Frequency, Excessively High Frequency
13:52:47 <fizzie> ion: Apparently some authors[1] use THF for "Tremendously High Frequency".
13:52:56 <fizzie> Your names fit that pretty well.
13:53:10 <fizzie> [1] Tanenbaum, Andrew (2002). Computer Networks. page 101: Prentice Hall. p. 912. ISBN 978-0-13-066102-9.
13:54:00 -!- FireFly has quit (Ping timeout: 256 seconds).
13:55:23 -!- FireFly has joined.
13:58:47 -!- jsvine has joined.
14:03:28 <Vorpal> Someone should figure out a way to take photos in radio wavelengths. I want to look what it looks like when I stand in front of a wifi router.
14:05:32 <ais523> Vorpal: I imagine it'd look much the same as taking a photo of a lightbulb
14:06:03 <ais523> routers don't send signals that turn corners for the same reason that lasers don't shoot out light sideways in real life, like they do in films
14:06:11 <Vorpal> Hm, surely I wouldn't be completely invisible?
14:06:25 <ais523> oh, I thought you were taking a photo of it
14:06:33 <ais523> rather than someone taking a photo of you next to it
14:06:38 <ais523> yeah, you'd probably be at least slightly visible
14:06:40 <Vorpal> ais523, me standing in between camera and the router
14:06:44 <ais523> although human bodies don't block radio waves very well
14:07:03 <ais523> (this is part of the reason radio is used for communication)
14:07:31 <Vorpal> Most of the world would be transparent. That would be pretty cool I think
14:07:51 <Vorpal> Well not fully transparent
14:08:31 <ais523> hmm, I imagine the main technical obstacle to building a camera like that is that it would have to be tens of metres across, or perhaps more depending on the wavelength
14:11:22 <Vorpal> Hm, reading up on MIMO (curse you wikipedia, I started out reading about lithography and ended up at MIMO...). That is pretty clever.
14:15:15 <Vorpal> If antennas length need to match the wavelength, why does normal FM radio work? You can just turn a dial or press a few buttons, and it doesn't extend or retract the antenna afaik.
14:19:22 <ais523> Vorpal: it doesn't need to match exactly; it works better the closer the match is
14:19:33 <ais523> but it'll still work at non-100% efficiency if the match isn't perfect
14:19:41 <ais523> you get some of the signal reflecting rather than being transmitted
14:30:09 -!- copumpkin has quit (Ping timeout: 240 seconds).
14:30:45 -!- copumpkin has joined.
14:34:06 -!- mnoqy has quit (Quit: hello).
14:45:31 <nortti> @tell oerjan < oerjan> was that command ever intended to have a useful purpose? <-- no
14:54:44 -!- oerjan has joined.
14:55:48 <HackEgo> olist (899): shachaf oerjan Sgeo FireFly
14:55:57 <shachaf> Oh, I bet oerjan joined to tell me.
14:56:27 <shachaf> hmm i bet 900 will be big and important??
14:56:58 <oerjan> um were previous round numbers important?
14:57:35 <shachaf> http://www.giantitp.com/comics/oots0300.html looks kind of important i guess
14:58:19 <ais523> well, given what happened in 898
14:58:23 <ais523> it's probably the end of the story arc
14:58:47 <shachaf> Do you want to be added to the list?
14:59:00 <ais523> shachaf: nah, I just check manually from time to time
14:59:10 <ais523> I have no particular reason to have up-to-the-minute updates
14:59:14 <ais523> given that it updates kind-of slowly anyway
14:59:19 <ais523> also that I'm often not in the channel
14:59:46 <ais523> and I'm often offline even when I'm not in the channel
15:00:01 <shachaf> You were gone for so long that the channel sprouted some new ops.
15:01:08 <ais523> I'm probably going to leave again tbf, except perhaps when something new happens that I care about talking about
15:02:07 <shachaf> Well, it's nice having you here, when you're here.
15:27:35 <fizzie> Re "photos in radio wavelengths", there's a couple of cameras that take "photos" of sound -- http://phys.org/news/2013-05-world-handheld-camera-ready.html and the like.
15:30:55 <oerjan> if an NFS mount gets so broken that you cannot even kill -9 processes trying to access it, must a linux machine then be rebooted?
15:31:16 -!- AnotherTest has joined.
15:31:21 <fizzie> You should fix the NFS server and hope it recovers hth
15:31:47 <oerjan> well the thing is the _server_ is fine, it's the client which has problems :P
15:33:02 <oerjan> also it's not actually me who is administrating this, i'm just wondering if i can expect to get kicked off when they fix it.
15:34:15 <oerjan> by the server is fine i mean that i can access my mail inbox fine from a third machine.
15:36:23 <fizzie> That doesn't really guaratee the health of the NFS service operating on it, though. (I think in most cases rebootless fixes should be possible.)
15:37:15 <oerjan> (the third machine also accesses via NFS, of course)
15:37:59 <fizzie> (Or "non-fixes" of the kind of "let the process stay dead, lazy-unmount the broken NFS mount, and wait for the next scheduled reboot for a more permanent solution".)
15:39:21 <oerjan> as long as they can remount a second time...
15:39:39 <oerjan> otherwise no one can access mail from this machine.
15:43:15 <fizzie> A lazy umount just removes the mount from the filesystem tree; it shouldn't be a problem mounting something working in its place, unless of course the NFS client code makes that a problem.
15:45:25 <fizzie> The NFS hangups I've seen at work lately have ended at I/O errors in the process (in a reasonable time; not days) -- possibly corresponding to people doing something to the server. But I don't admin those either.
15:48:08 <oerjan> i suppose there may not be any admins active at the moment. yesterday there was a different but probably related problem and i quickly got an email back.
15:48:43 <kmc> elliott: yes, the X1 Carbon is nice, I can talk about its strengths and weaknesses if you like
15:51:10 <elliott> kmc: sounds good! I am really tired but if you have a monologue I will read it and I might have more questions tomorrow
15:58:15 -!- conehead has joined.
16:00:52 <oerjan> ais523: i haven't read all the logs yet, but it looks to me like your concurrency primitive should be trivial to implement in haskell
16:01:22 <ais523> oerjan: I'd like to see the impl; also I don't know how concurrency works in Haskell, which is part of the reason I wouldn't try to work it out myself
16:03:28 <oerjan> oh except for one thing, it definitely cannot be pure, you'd want to allocate an MVar to communicate between invocations.
16:04:18 <elliott> kmc: alt. we can do it tomorrow
16:07:44 <shachaf> --------------------------------------------------------------------------------
16:17:46 <Vorpal> Urgh, lame doesn't take flac as input. Now what? (No I can't use ogg here, the device I need to put this on only supports mp3)
16:18:34 <Jafet> Insert lame pun here
16:20:07 <elliott> Vorpal: flac --decode --stdout foo.flac | lame --preset standard /dev/stdin ... or whatever
16:20:10 <elliott> maybe you need a temp file
16:20:34 <Vorpal> What does flac --decode output? PCM?
16:21:43 <Deewiant> ffmpeg -i foo.flac -o foo.mp3 or whatever
16:22:33 <Vorpal> elliott, lame --help says - will work for stdin so that should be good
16:22:44 <Vorpal> Deewiant, Is that better than lame?
16:24:31 <Vorpal> for i in *.flac; do flac --decode --stdout "$i" | lame --preset extreme - "${i/.flac/}.mp3"; done <-- this should work right? Not sure about that /.flac/ syntax, but I think it is correct?
16:25:12 <Vorpal> except this takes ages, how to parallelize? I usually use xargs -P
16:25:33 <elliott> there's that GNU parallel thing.
16:25:50 <elliott> btw, --preset extreme is overkill.
16:25:53 <Vorpal> Hm, yeah doing this with xargs would be complicated, maybe through a sh call
16:26:06 <Vorpal> elliott, anything between standard and extreme?
16:26:13 -!- jsvine has quit (Quit: Leaving.).
16:26:40 <fizzie> Deewiant: But ffmpeg is deprecated!1
16:26:54 <elliott> uh, I don't think so. but I really doubt standard fails to be transparent on non-pathological examples without equally pathological listening conditions
16:27:18 <fizzie> Vorpal: The command-line tool, I mean; tou're supposed to use "avconv" these days.
16:27:30 <elliott> it's like 190 kbps and LAME is p. good at encoding
16:27:45 <elliott> fizzie: did the original ffmpeg project give up or is it still a hostile fork situation?
16:28:01 <fizzie> See: http://sprunge.us/TbBh
16:28:19 <fizzie> And I think avconv's part of the official ffmpeg now, or something. I haven't really been following.
16:29:24 <fizzie> I guess it's a "libav" thing, mhmm.
16:29:28 <Deewiant> ffmpeg 0.10's changelog from 2012 says "
16:29:29 <Deewiant> all features from avconv merged into ffmpeg"
16:31:40 <fizzie> Okay, I guess you shouldn't be using avconv if your libavcodec and friends comes from FFmpeg and not libav.
16:31:47 <fizzie> I'm glad it's not confusing or anything.
16:33:50 <elliott> http://blog.pkh.me/p/13-the-ffmpeg-libav-situation.html
16:34:40 <kmc> elliott: so the main good things are, it's thin, light, and has a screen with a not embarassingly low resolution
16:34:57 <kmc> (1600x900 while the other ultraportable thinkpads are 1366x768)
16:35:33 <kmc> battery life is fine (I get about 5 hours in a moderately tuned Linux system) but the battery charges really quickly, which is nice
16:35:56 <kmc> like it will charge from 20% to 80% in 30 min or so
16:35:58 <elliott> (includes stuff about the deprecation warning)
16:36:04 <fizzie> "-- a FFmpeg/Libav war child --"
16:36:12 <fizzie> A tragedy. I shed a tear.
16:36:16 <kmc> I hear there are Ivy Bridge power management improvements in Linux 3.10 and I should try that
16:36:44 <kmc> also it suspends and resumes from RAM *really* fast
16:36:50 <kmc> like resume takes 1 second before the machine is completely usable again
16:37:04 <kmc> there's no reason why suspend to RAM should be slow, but it's always been a bit slower on other thinkpads i've owned
16:38:08 <Vorpal> WARNING: gnome-keyring:: couldn't connect to: /home/arvid/.cache/keyring-cU64W7/pkcs11: No such file or directory
16:38:21 <Vorpal> Why is ffmpeg doing that?!
16:39:40 <elliott> kmc: do you have any idea if it might be getting Haswell any time soon
16:40:03 <kmc> i'm no expert on lenovo kremlinology
16:40:12 <kmc> I couldn't even get a straight answer as to when the i7/8GB model would be available
16:40:34 <Vorpal> kmc, is it not possible to add a few extra RAM modules yourself?
16:40:45 <elliott> kmc: I think you can get i7 + 8GB now
16:40:45 <kmc> it's an "ultrabook", RAM is soldered on
16:40:58 <Vorpal> kmc, yet another reason to go for normal laptops
16:41:13 <kmc> itt Vorpal and I have different utility functions
16:41:15 <elliott> have to admit I fail to see the reasoning here
16:41:16 <kmc> elliott: yeah, also there's a touch version now
16:41:35 <kmc> so yeah, you can't upgrade the RAM
16:41:40 <elliott> "you can't get 8 gigs of RAM so you should get a machine that is worse at doing what it's for in every respect other than it happens to have 8 gigs of RAM"
16:41:48 <kmc> except I have 8 gigs of RAM!!
16:41:52 <kmc> but I wish I had 16 :(
16:41:55 <kmc> you maybe can't remove the SSD either; I need to try that
16:42:07 <kmc> in the past the ability to quickly remove the hdd from thinkpads has been very useful in disaster type situations
16:42:22 <kmc> I've heard the SSD is a mSATA card but I've also heard otherwise
16:42:24 <Vorpal> elliott, I like having the ability to extend the hardware easily as my needs changes.
16:42:54 <elliott> kmc: yes my main misgivings look like (a) no quad core (solved by haswell??) (b) intel graphics (made less bad by haswell??) (c) 8 gigs of ram limit
16:43:02 <kmc> seems reasonable
16:43:02 <elliott> (d) the screen could do with being even higher-resolution...... :(
16:43:13 <kmc> yeah there will probably be a 'retina' version someday
16:43:21 <kmc> you could get a chromebook pixel instead ;P
16:43:27 <kmc> or I think samsung has a high dpi laptop too now
16:43:40 <Vorpal> kmc, on my thinkpad I only need to remove a single screw to change the HDD. I need to remove 4 to reach the RAM modules.
16:43:44 <elliott> I am also looking at https://www.system76.com/laptops/model/galu1 because of shachaf
16:43:53 <kmc> Vorpal: yes I have owned thinkpads before as well
16:44:02 <Vorpal> kmc, Anyway should my ram break I can replace it. Couldn't if it was soldered on.
16:44:10 <kmc> have you ever had RAM break
16:44:17 <elliott> for which my main misgiving is whether the touchpad is any good
16:44:21 <fizzie> Fiora: I interpreted that as "I have 32 Thinkpads".
16:44:30 <elliott> (and whether iris pro is actually decent enough)
16:44:32 <Vorpal> kmc, Not for ages. (Early 2000s iirc)
16:44:34 <kmc> other issues: Linux hardware support isn't *quite* there yet. it's mostly fine, but occasionally sound freaks out, this can usually be fixed by removing and reinstalling the kernel module, but not always
16:44:46 <kmc> I've occasionally had similar issues with wireless too
16:45:00 <kmc> neither is frequent enough to really be annoying
16:45:22 <kmc> I wonder if I can keep listing drawbacks of this machine without Vorpal assuming that I'm a dumbass who didn't know these things when I bought it
16:45:38 <kmc> and listing irrelevant facts at me
16:45:39 <elliott> Fiora: how much have you actually managed to use
16:45:44 <Vorpal> kmc, I'm not assuming you are a dumbass :P
16:45:44 <elliott> I mean ignoring disk caches and stuff
16:46:14 <Vorpal> And finding out linux hw support in advance tends to be rather hit or miss I found.
16:46:34 <elliott> kmc: don't forget to mention the keyboard
16:46:36 <kmc> elliott: also it takes a different power adapter plug than other thinkpads :( (the standard round barrel is too thick for the case)
16:46:52 <elliott> well I don't have any thinkpads and never have so that's not really a concern at least :P
16:47:00 <Fiora> elliott: um... I think I probably use like up to 2 or 3 GB for a game? plus a few gigs for other things so probably like 8-10
16:47:02 <kmc> I had quite an investment in old-style power adapters
16:47:10 <Fiora> the disk cache is nice though! albeit not quite as useful with an SSD
16:47:28 <Fiora> oh right! when I was playing around with ECM factoring I got to use 23 GB and actually ran out? xD
16:47:38 <FreeFull> kmc: Do you do any programming?
16:47:42 <kmc> also there's no onboard Ethernet; it ships with a USB Ethernet adapter that is only supported by recent-ish Linux (see http://mainisusuallyafunction.blogspot.com/2012/12/hex-editing-linux-kernel-modules-to.html)
16:47:52 <elliott> I could probably use 32 GB just by compiling enough Haskell packages in parallel I guess
16:47:58 <kmc> FreeFull: what, do you mean "do i do any work instead of hanging out on IRC all day"
16:48:15 <kmc> I have a meeting in 10 min
16:48:23 <kmc> but yes I did write code yesterday
16:48:28 <elliott> "how is your IRC performance"
16:48:29 <fizzie> Incidentally, do you have to give Kensington® some money if you put a Kensington® hole® in your thing?
16:48:35 <elliott> "have you sold any ThinkPads lately"
16:48:39 <elliott> "are they believing your Mozilla cover story"
16:48:49 <FreeFull> kmc: You could take a look at the driver code and fix it once and for all
16:48:50 <Vorpal> <Fiora> elliott: um... I think I probably use like up to 2 or 3 GB for a game? plus a few gigs for other things so probably like 8-10 <-- I mainly found that I use my 16 GB of RAM for running hundreds of chrome tabs.
16:49:01 <kmc> elliott: also the only video port is Mini DisplayPort; if you want VGA or DVI or HDMI you need an adapter as well, and that one is not included
16:49:08 <Vorpal> Also the occasional game, minecraft with mods especially.
16:49:27 <Vorpal> And of course panorama stitching, but that is not a very common activity.
16:49:35 <fizzie> I mainly use the 16 gigs at work to hold large matrices.
16:49:44 <FreeFull> Vorpal: I have hundreds of tabs open regularly in Firefox, and it doesn't take anywhere near that much RAM
16:49:53 <kmc> FreeFull: oh also my code was building, that's my other excuse
16:49:59 <fizzie> Just Gaussian random data if I can't figure out what else to put there! (Not really.)
16:50:04 <elliott> kmc: have you added any code yet
16:50:08 <Vorpal> FreeFull, well of course chrome doesn't take it all. Usually about 1 GB total as far as I can tell.
16:50:13 <FreeFull> fizzie: I hope your matrix multiplication algorithm has good cache locality
16:50:16 <Vorpal> FreeFull, but it gives room for still doing other stuff
16:50:44 <kmc> elliott: https://github.com/mozilla/servo/commit/9f6de7e5e30ff84bbe04d22c04a22fa33e4aa8e7 https://github.com/mozilla/servo/commit/cfffd0542404b60923f3f524f5144693d9b89f00 v. important changes
16:50:47 <fizzie> FreeFull: It's MATLAB's matrix multiplication algorithm, I'm sure it has all the good things and none of the bad things; I mean, MATLAB. It's got MAT right in the name.
16:50:53 <FreeFull> The only benefit I would have from 16GB would be more filesystem cache
16:51:14 <kmc> it's amazing how quickly n+1 browser tabs can eat up 8 GB
16:51:15 <FreeFull> I wonder what you need matrices that big for
16:51:31 <elliott> kmc: those both look more like removing code to me
16:51:31 <kmc> Rodina MAT
16:51:36 <kmc> elliott: yes isn't it great
16:51:40 -!- conehead has quit (Ping timeout: 276 seconds).
16:51:51 -!- ais523 has quit (Ping timeout: 240 seconds).
16:51:51 <Vorpal> fizzie, is that the "brand" of the holes?
16:51:54 <elliott> go write some hacks to handle awful CSS edge-cases
16:52:11 <fizzie> Vorpal: Officially it's a "Kensington Security Slot", I think.
16:52:11 <Vorpal> fizzie, I never seen a Kensington lock, I only ever seen other brands that fit the same hole.
16:52:34 <elliott> "Address pcwalton's comments" "Address pcwalton's nits" wow pcwalton, give em a break
16:52:46 <fizzie> I've only seen the holes. Well, I guess I've seen a couple of locks at stores and so on, don't know which brand.
16:53:05 <Vorpal> fizzie, we have locks at work.
16:54:18 <Vorpal> Some no-name brand lock I believe. Not that we need it. Just a general company laptop security policy.
16:54:49 <fizzie> FreeFull: Well, you know... so many things are en-squared. If you have a 30k-dimensional vectors and ask for something that wants to make a covariance matrix, that's 30000 x 30000 = 900 million 8-byte floats = 6.7 gigs right there.
16:56:21 <FreeFull> Now I'm wondering where you need 30k-dimensional vectors
16:56:44 <elliott> finns live in a 30k-dimensional universe
16:56:54 <elliott> this is how fizzie plans his commute
16:58:01 <FreeFull> fizzie: At least you don't have to calculate a 30k x 30k x 30k tensor, then you'd never have enough RAM =P
16:58:42 <fizzie> FreeFull: Activation vectors for a 30k-element dictionary of atoms hth
17:00:57 <fizzie> FreeFull: Also sometimes you have 30k 39-dimensional vectors and accidentally give it to the function transposed, because half the functions expect observations in column vectors and other half in row vectors. (The truth comes out.)
17:01:28 <FreeFull> fizzie: Oh, you're doing a physics simulation?
17:01:44 <elliott> fizzie: it's ok to admit finns live in a bizarre and disorienting world, incomprehensible to all outsiders.
17:01:47 <elliott> I mean, we already know that
17:02:03 <fizzie> You know all those dimensions from string theory? Yeah, they've all curled up in Finland.
17:02:49 <fizzie> (Also I don't really do physics simulations, I just like to refer to audio frames as "observations".)
17:08:45 <FreeFull> Wait, 30000 x 30000 matrices for audio? Just what are you doing?
17:20:14 <fizzie> As I mentioned, activations for a 30k element dictionary of "exemplars".
17:20:36 <fizzie> It's a source separation thingie based on a NMF decomposition.
17:30:39 -!- jsvine has joined.
17:32:26 -!- tswett has quit (Changing host).
17:32:27 -!- tswett has joined.
17:32:52 -!- carado has quit (Ping timeout: 260 seconds).
17:33:21 -!- conehead has joined.
17:34:13 -!- TeruFSX has joined.
17:42:56 -!- epicmonkey has quit (Read error: Operation timed out).
18:03:05 <Gracenotes> hm. I went to bed at a reasonable time, but only properly woke up just now
18:03:31 <kmc> did you improperly wake up before that?
18:03:49 <Fiora> oddly enough elliott just went to sleep
18:04:36 <Gracenotes> then I had increasingly weird REM half-sleeps
18:04:51 <Gracenotes> so I suppose it's not like I was getting actual rest out of that
18:11:44 <fizzie> Fiora: Have you installed a small spy camera in elliott's room? (Clearly the most likely conclusion.)
18:14:30 <Deewiant> Occam's razor suggests that assuming it's small is superfluous
18:14:36 <Fiora> I was talking with him -_-
18:15:08 <fizzie> I think that's just a cover story.
18:15:51 <Fiora> he's a cutiepants okay I can't help it
18:19:01 -!- nooodl has quit (Quit: Leaving).
18:20:11 -!- nooodl has joined.
18:26:56 <kmc> tricked again by /bin/sh -> /bin/dash
18:32:10 <kmc> static THROBBER: [char, ..8] = [ '⣾', '⣽', '⣻', '⢿', '⡿', '⣟', '⣯', '⣷' ];
18:34:22 <kmc> I didn't realize where those characters are from but it should have been obvious really
18:41:52 <ion> I bet blind people get totally confused when reading the output from that software.
18:43:39 <fizzie> I bet it looks fancy as anything, though.
18:43:50 <Gracenotes> there are fancy devices that do dynamic embossing, but that seems no harder to do with plain letters.
18:44:58 <Gracenotes> people who use Braille likely know ASCII better than any of us not from the punchcard era
18:45:00 <fizzie> Also, what *is* that from? Isn't Braille just three rows?
18:45:22 <nooodl> super cool extended braille isn't
18:51:25 <nooodl> > text [chr $ ord '⣟' Data.Bits..&. ord '⣾']
18:52:19 <kmc> the Debian installer has lots of support for braille ttys, also for speech synthesis of the menus
18:52:22 <kmc> pretty cool
18:54:17 -!- epicmonkey has joined.
18:59:01 <fizzie> > text [chr $ ord '⣟' `xor` ord '⣾'] -- less cute
18:59:17 <fizzie> They should've put that stuff right at the beginning of Unicode, for cuteness.
19:00:33 <Bike> blind superiority
19:00:43 <fizzie> > text [chr $ ord '⠀' + (ord '⣟' `xor` ord '⣾')] -- magic space fix
19:03:16 -!- mnoqy has joined.
19:03:50 <fizzie> lambdabot: You could've done a little more than that.
19:03:56 <kmc> clearly we should use these as a text transport for binary data
19:04:04 <lambdabot> "\10240\10241\10242\10243\10244\10245\10246\10247\10248\10249\10250\10251\1...
19:04:12 <kmc> 8 bits each right?
19:04:14 <fizzie> (The multiline query version looked nice.)
19:04:22 <kmc> are some combos verboten
19:04:33 <kmc> btw today is free slurpee day at 7-Eleven
19:04:36 <kmc> maybe only in the USA
19:04:50 <fizzie> AIUI, there's quite a few Braille variants, esp. concerning the 8-cell versions.
19:04:58 <fizzie> But Unicode has the whole set of symbols.
19:05:02 <kmc> the reason: date +%m-%d
19:06:30 <fizzie> There are no 7-Elevens in Finland.
19:07:02 <fizzie> (There are some in Denmark, Norway and Sweden, which I don't think is entirely fair.)
19:08:02 -!- copumpkin has quit (Ping timeout: 246 seconds).
19:08:38 -!- copumpkin has joined.
19:09:56 <fizzie> Ohh, the Steam Summer Sale. So that's why everyone's[weasel words][dubious - discuss] talking about Steam, and it's down.
19:10:55 <fizzie> That would be Tal... right
19:11:27 <fizzie> I'm not even mentioned at Talk:Bath bomb, and Talk:Bath fizzie is a redlink. :/
19:11:49 <fizzie> [[ "Wave 4" was released in August 2005, and it included Bertie, Purkle, Spangle, Dilly, Skinker, and Fizzie. ]]
19:11:57 <fizzie> In good company, I'd say?
19:12:18 <fizzie> But "Bath fizzie" isn't a redirect.
19:12:30 <Gracenotes> or there was some mess with exchanging articles, redirects, and disambiguation pages, and the talk pages went along for the ride
19:12:38 <fizzie> It's a really kinda weird article instead, and completely different than the Bath bomb one.
19:12:57 <fizzie> I think perhaps written by a chemist, or something.
19:13:02 -!- Koen_ has quit (Quit: Koen_).
19:13:30 <fizzie> "-- designed to effervesce in personal bath water -- amorphous grains of homogeneous mixture -- solid boluses of homogeneous or inhomogeneous mixture -- ingredients must include one or more acid(s) and one or more water-soluble bicarbonate, sesquicarbonate, and/or carbonate -- other water-soluble, water-dispersible, and/or volatile ingredients --"
19:13:33 <Gracenotes> oh. slap a merge tag on it, or something.
19:13:50 <fizzie> What, me, edit Wikipedia? I'm not one of those guys.
19:14:24 <fizzie> Did you know that: In principle a fizzie could release phosphoric anhydride gas, but the release of gas from phosphate salts is so slow that any phosphates present in either beverage or bath fizzies are for other purposes.
19:14:55 <fizzie> In practice, I only release phosphoric anhydride gas after very specific meals.
19:15:32 <fizzie> (There's also gratuitous use of bold.)
19:15:46 <Gracenotes> Is it you who releases it or intestinal bacteria?
19:16:57 -!- TeruFSX2 has joined.
19:22:25 -!- mtve- has joined.
19:30:58 -!- TeruFSX has quit (*.net *.split).
19:31:01 -!- mtve has quit (*.net *.split).
19:31:01 -!- yiyus has quit (*.net *.split).
19:31:01 -!- Fiora has quit (*.net *.split).
19:37:38 -!- Koen_ has joined.
19:39:09 -!- Koen__ has joined.
19:39:10 -!- Koen_ has quit (Read error: Connection reset by peer).
19:41:04 <Vorpal> Oh god, a full screen installer with a blue gradient background. This is old!
19:41:12 <Vorpal> Why did they ever do that btw?
19:41:26 -!- yiyus has joined.
19:41:26 -!- Fiora has joined.
19:43:33 -!- AnotherTest has quit (Quit: Leaving.).
19:44:57 <ion> https://bugs.kde.org/show_bug.cgi?id=322255
19:46:11 -!- tswett_ has joined.
19:47:53 -!- quintopi1 has joined.
19:48:51 -!- tswett_ has quit (Changing host).
19:48:51 -!- tswett_ has joined.
19:52:40 -!- upgrayeddd has quit (Ping timeout: 246 seconds).
19:52:42 -!- quintopia has quit (Ping timeout: 246 seconds).
19:52:44 -!- tswett has quit (Ping timeout: 246 seconds).
19:53:51 <ion> http://www.imdb.com/media/rm465020160/tt2724064?ref_=butt
19:58:18 -!- tswett_ has changed nick to tswett.
20:12:31 <tswett> Okay, I wonder if all you wonderful people can help me with a problem I'm facing.
20:12:41 -!- variable has changed nick to function.
20:12:44 <tswett> So I want to store a directed acyclic graph in a database.
20:13:21 <tswett> Whenever I try to add a new edge to the graph, I first look up all paths that would create a cycle with that edge.
20:13:30 <tswett> Meaning that I have a table containing all paths.
20:13:56 <tswett> The table has three columns: "head", which is the first edge in the graph; "last", which is the last; and "tail", which is a reference to the tail of the path.
20:14:46 <tswett> So the path table forms a tree.
20:16:01 <tswett> Whenever I add a new edge, I need to add new paths containing that edge. But I can't figure out a good way to do that.
20:16:08 -!- oerjan_ has joined.
20:18:16 <Fiora> all paths? isn't that an exponential number?
20:18:43 <Fiora> eesh, that sounds painful
20:18:46 <tswett> I guess the algorithm is reasonably straightforward, now that I think of it.
20:18:52 -!- oerjan has quit (Disconnected by services).
20:19:03 <Fiora> maybe it would be better to have some storage method that doesn't involve exponential space <.<
20:19:06 <tswett> Start with a list of all relevant paths already in the database, then just recursively extend them backwards.
20:19:11 -!- oerjan_ has changed nick to oerjan.
20:19:15 <tswett> If performance gets to be an issue, I'll rethink stuff.
20:21:10 <oerjan> if anyone spoke to me in privmsg in about the last 2 hours, my nick has been a ghost
20:21:32 <kmc> tswett: you look for paths that would create a cycle just to throw an error if you find one, or for some other reason?
20:21:44 <kmc> if you just want to know "would this new edge form a cycle" then there are simpler data structures you can use
20:22:13 <tswett> kmc: for some other reason. I need to do something with every edge in the cycle.
20:22:29 <tswett> And then delete one of the edges.
20:24:04 <oerjan> isn't hpaste.org alive any more?
20:24:42 <Fiora> might it not be a little less exponential to like, store the results of an all-pairs shortest path thing?
20:24:50 <Fiora> I think that's N^2 if you do it the dynamic programmy way
20:25:10 -!- quintopi1 has changed nick to quintopia.
20:25:17 -!- quintopia has quit (Changing host).
20:25:18 -!- quintopia has joined.
20:25:57 <tswett> Fiora: well, adding a single edge could create exponentially many cycles, and I could potentially want to operate on all of them.
20:26:41 <tswett> So whatever method I use needs to be able to tell me about all of those cycles.
20:27:14 <Fiora> yeah, but like, you can have exponentially many *paths* without having exponentially many cycles
20:27:58 <Fiora> how do you distinguish different cycles? like what counts as two different cycles
20:28:32 <Bike> made up of different vertices?
20:28:38 <Fiora> like if A->B, B->C, B->D, C->A, and D->A, is A->B->C->A and A->B->D->A two different cycles?
20:28:57 <Bike> i would think so...
20:29:34 <Bike> why wouldn't they be?
20:31:05 <Fiora> actually um. that's a question if you have a directed acyclic graph with V vertices and E edges, what's the maximum number of cycles adding an edge could create...
20:32:21 <quintopia> Fiora: it should be maximized by adding the final mossing edge to make a complete graph
20:34:30 -!- MindlessDrone has quit (Quit: MindlessDrone).
20:36:04 <Fiora> could such a graph have no cycles?
20:36:32 <Fiora> http://i.imgur.com/BbPmpQi.png here's an example I was thinking of... I guess this example does create an exponential number of cycles
20:36:40 <Fiora> (red is the new edge)
20:36:57 <Fiora> as the number of vertex in the diamond gets larger, I think the number of cycles approaches 2^V?
20:37:06 <Fiora> since like, there's a branching factor of 2
20:38:20 <quintopia> i know some people who were working on problems of graphs like this
20:39:45 <quintopia> the branching factor is 2 for the bottom half, but maybe no be on the top half. the actual number of paths is given by a nice recurrence relation i think
20:40:42 <kmc> https://github.com/mozilla/rust/issues/7048
20:40:52 <fizzie> I think the recurrence relation will involve a two times something n-1.
20:43:25 <fizzie> (Intuitively speaking, the paths after going left first + the paths after going right first, which look like they'd be smaller semi-diamond-like graphs.)
20:45:02 -!- oerjan has quit (Quit: Sleep soon).
20:45:10 <kmc> hpaste is down :/
20:45:14 <kmc> need to link some amusing things from there
20:46:36 <fizzie> (#paths(n) = f(n,n) where f(n,m) = f(n-1,m) + f(m,n-1); add some base cases and solve; and the n in #paths(n) is sqrt(V). Or some-such.)
20:47:03 <fizzie> FreeFull: Does lpaste have hpaste pastes?
20:47:49 <FreeFull> http://lpaste.net/1 dated 2008
20:48:26 <FreeFull> fizzie: I think it actually is hpaste
20:48:32 <FreeFull> Just with a different name and domain
20:49:54 -!- nooodl has quit (Ping timeout: 264 seconds).
20:51:10 -!- nooodl has joined.
20:51:55 <fizzie> http://oeis.org/A000984 is the number of paths for a diamond with edge size (number of vertices) of n, I think. (I might have gotten that all wrong.)
20:53:11 <fizzie> The comments seem to bear it out. E.g. "Number of possible values of a 2*n bit binary number for which half the bits are on" aka "number of ways you can take left and right paths to end up at the top of the diamond".
20:54:28 <fizzie> ("The number of lattice paths from (0,0) to (n,n) using steps (1,0) and (0,1)" is even closer.)
20:57:24 <Bike> isn't that exactly the same
20:57:45 <fizzie> They're comments for the same sequence, of course they're exactly the same.
20:58:08 <fizzie> I just meant "closer to Fiora's Diamond™".
20:59:00 <Bike> i meant the number of lattice paths is the same as paths in the diamond
20:59:39 <fizzie> I mean, it was supposed to be, too.
21:06:20 <Fiora> I guess you could just chop off the top half of the diamond and have everything connect to one top point
21:09:22 -!- Bike_ has joined.
21:09:58 -!- Bike has quit (Ping timeout: 248 seconds).
21:14:55 -!- Bike_ has changed nick to Bike.
21:16:05 <fizzie> That'd be, what, 2^n paths, with V = n*(1+n)/2?
21:18:08 <fizzie> (I guess +1 for the star at the top.)
21:20:30 <shachaf> kmc: hpaste is up, it's just the domain.
21:20:32 <lambdabot> http://paste.tryhaskell.org/new/haskell
21:21:04 <shachaf> Oh, lpaste is the same thing now.
21:25:12 <quintopia> i'm glad to see fizzie figured out the same recurrence i did while i was in the shower :D
21:26:41 <shachaf> did elliott decide on a laptop
21:26:52 <Fiora> he's asleep right now, I think
21:26:53 -!- Bike has quit (Ping timeout: 245 seconds).
21:29:04 -!- Bike has joined.
21:29:44 <quintopia> The number of direct routes from my home to Granny's when
21:29:44 <quintopia> Granny lives n blocks south and n blocks east of my home
21:29:44 <quintopia> in Grid City. To obtain a direct route, from the 2n
21:29:44 <quintopia> blocks, choose n blocks on which one travels south. For
21:29:45 <fizzie> Maybe that's just what he wants us to think?
21:29:46 <quintopia> example, a(2)=6 because there are 6 direct routes
21:29:56 <quintopia> i wish i had thought of that solution
21:30:06 <quintopia> geez it's so obvious in retrospect
21:32:51 -!- TeruFSX2 has quit (Read error: Connection reset by peer).
21:33:37 <fizzie> Finding Granny: so easy, a pushdown automaton could do it.
21:34:32 <FreeFull> I wonder what sort of space you get if you add diagonal paths
21:35:35 <FreeFull> Since it wouldn't be taxicab space anymore
21:36:08 <fizzie> For the southeast? Isn't that just +f(n-1,m-1)?
21:38:53 <fizzie> http://oeis.org/A001850 and so on.
21:51:24 -!- copumpkin has quit (Ping timeout: 260 seconds).
21:52:00 -!- copumpkin has joined.
21:55:33 -!- Lymia has quit (Ping timeout: 245 seconds).
21:55:49 -!- upgrayeddd has joined.
21:56:12 <fizzie> (The general taxicab-with-diagonals distance is called Chebyshev distance, or chessboard distance (king moves), or maximum metric (from D(a,b) = max_i |a_i - b_i|), or probably by some further names.)
21:59:19 -!- function has changed nick to trout.
22:02:02 -!- jsvine has quit (Quit: Leaving.).
22:02:12 <FreeFull> I remember looking at it earlier but I forgot about it
22:03:42 <FreeFull> Seems Chebyshev and Taxicab metrics are duals of each other or something
22:04:22 -!- fungot has quit (Ping timeout: 248 seconds).
22:06:24 -!- fungot has joined.
22:09:21 <kmc> shachaf: one cool thing in Rust is that you can put debug print statements all over your code, and then enable them at runtime with an env var
22:09:34 <kmc> and you can do this on a per-module basis, without needing to reiterate the name of the module on each print statement
22:09:49 <kmc> I feel like a lot of languages should have this, but I don't think I've seen it before
22:10:28 <FreeFull> kmc: So it's like assert in C but for printing?
22:11:00 <Bike> it's a debug print statement, haven't you seen those before
22:11:33 <FreeFull> Bike: kmc's point is that they are always in the code, but only turn on when compiling with debugging
22:11:35 <kmc> not ones where you leave them in and then you say "enable debug printing for module foo:bar" at runtime
22:11:41 <kmc> FreeFull: no do pay attention please
22:11:47 <kmc> > enable them at runtime with an env var
22:11:48 <lambdabot> Not in scope: `enable'Not in scope: `them'Not in scope: `runtime'Not in sco...
22:11:58 <Bike> that's what you get for 4chanquoting.
22:12:13 <FreeFull> Bike: That quote style is much older than 4chan
22:12:13 <Bike> These pens are not for everyone. They are pens that are for people that appreciate quality and are willing to spend a little extra on having something that is fairly unique. EiMIM stands for Everything in Moderation, Including Moderation. Most of the time I try to keep things as simple as possible, but every once in a while I need to embrace indulgence. Will a $2 pen write? Absolutely. The thing is, there is no soul in that pen. It was ...
22:12:20 <Bike> ... produced with a million other pens exactly like it. These pens are different.
22:12:29 <Fiora> Bike: yks is amazing xD
22:12:39 <kmc> I've seen lots of logging systems but usually you're required to name some "logger" on each print statement and that's what you can enable/disable
22:12:50 <kmc> rather than it being implicitly controlled by the current module
22:13:00 <fizzie> kmc: Some of fancy C 'dprintf' macros (or whatnot) do enablation based on __FILE__ or top-of-source-file "MODULE" macro.
22:13:01 <FreeFull> Bike: I'd say to that person, "If you like pens so much, make them"
22:13:09 <shachaf> kmc: #define log(s) real_log(__FILE__, s)?
22:13:09 <kmc> fizzie: that's cool
22:14:09 <fizzie> Also __VA_ARGS__ so you can be all format string.
22:16:58 -!- Bike has quit (Ping timeout: 264 seconds).
22:17:59 -!- Bike has joined.
22:18:49 <Bike> https://s3.amazonaws.com/ksr/assets/000/728/934/bf0f707a2fbf5f442603d2a948ac8761_large.png?1373294172
22:20:06 <Bike> FreeFull: oh they are making them. they are in-deed.
22:20:35 <shachaf> Bike: Is Amazon Aws where they serve all the kitten pictures from?
22:22:50 <Bike> Phantom_Hoover: uh did you see the table i think it's pretty clear.
22:25:48 <Bike> http://www.kickstarter.com/projects/eimim/eimim-x-y-and-z-pens1 here
22:26:03 -!- epicmonkey has quit (Ping timeout: 245 seconds).
22:26:41 <Bike> yes and that thing is not enough craftsmanship
22:26:58 <ion> http://www.guardian.co.uk/world/video/2013/jul/08/mos-def-force-fed-guantanamo-bay-video
22:28:21 <Bike> oh, http://www.kickstarter.com/projects/eimim/eimim-x-y-and-z-pens
22:28:45 <shachaf> Bike: There's a better chart at http://www.sourcereal.com/
22:29:03 <Bike> not in png form though!!
22:29:26 <shachaf> Bike: Well, print it out and photograph it if you want a png.
22:29:58 -!- fungot has quit (Ping timeout: 248 seconds).
22:29:59 <Bike> why would you use a png for a photograph
22:30:13 <kmc> finally, pens for insufferable people
22:30:34 <shachaf> Bike: because it's a photograph of text
22:31:36 <Phantom_Hoover> i like how the 'seamless design' is accomplished by cutting grooves all the way up the pen to make it hard to see the seam with the cap
22:31:55 <Phantom_Hoover> "I use the strongest permanent neodymium magnets available. They provide a unique and innovative method of attaching the cap to the back of the pen during use."
22:32:17 <Phantom_Hoover> you are, er, calling the use of magnets to stick things together innovative
22:32:29 <kmc> "yeah bitch! magnets!"
22:32:38 <FreeFull> Strongest possible neodymium magnets
22:32:42 <Bike> Phantom_Hoover: to be fair i don't think i've ever seen that in pens.
22:32:47 <FreeFull> That cap isn't coming back off
22:33:20 <Phantom_Hoover> Bike, yes, probably because for, say, any kind of practical pen design it wouldn't really work
22:33:28 <Bike> don't be a h8r.
22:36:22 <Bike> hahaha this is so dorky.
22:36:36 <Bike> patented magnetic engineering physics.
22:37:21 <Bike> is this seriously $30 for two pens.
22:37:54 <Fiora> Bike: http://yourkickstartersucks.tumblr.com/post/46417947427/hey-congrats-on-inventing-the-bowl-the-cookie
22:38:04 <Fiora> High milk displacement dunk technology.
22:38:57 <Bike> nope, $7,734 was donated (though it wasn't funded, so)
22:39:11 <Bike> well i mean, they're probably writing it in a jokey tone
22:39:25 <Phantom_Hoover> http://www.kickstarter.com/projects/1339254269/ron-paul-road-to-revolution
22:39:25 <Bike> as a smokescreen over the fact that this is bullshit
22:39:32 <Bike> ron paul is a classic.
22:39:47 <Bike> Phantom_Hoover: that's the game, right? i don't think it's been made yet
22:40:12 <Fiora> wasn't that the thing where like the guy ended up admitting he had no idea how to make a game?
22:40:28 <Bike> so it just kinda died
22:40:36 <Bike> Basically, here's the problem: that finger extending, fist jamming, glass tilting lean to stretch for a dunk in those last few ounces of milk. It's an awful scenario. Best case? A partial dunk. Worst case? Milk spillage or you drop the cookie in and fish it out with a fork.
22:40:58 <Bike> https://s3.amazonaws.com/ksr/assets/000/438/812/28ef07a4bfb36e7852b766519f2f734e_large.png?1363050910
22:42:05 <Bike> Is this part of the new face of American Manufacturing?
22:42:08 <Bike> Yes! This is about how a small business, with a 3d printer in the back room (we don't even have a garage!), can iterate and develop an idea and then use Kickstarter to raise funds for production tooling. We work with local companies because ethically we think it's the right thing to do and it makes the most sense for a small company to manage the logistics. Our injection molder, packaging printer, assembly facility, warehouse, and order fulfi
22:42:30 <Phantom_Hoover> also that thing reminds me of those glasses that are supposedly modelled on a breast or something
22:43:12 -!- zzo38 has joined.
22:44:49 <kmc> yourkickstartersucks is the greatest
22:46:39 <Phantom_Hoover> one of the potential refills listed are the ones for fisher space pens
22:47:07 <Phantom_Hoover> thing is, fisher space pen refills are a complete ink + nib assembly
22:47:29 <Phantom_Hoover> so he's not just charging $55 for a ballpoint pen, he's charging $55 for a ballpoint pen casing
22:48:55 <kmc> look if you buy expensive things then you're just flaunting wealth but if you buy expensive "minimalist" things then you can somehow pretend that it's morally justified and makes you better on an aesthetic plane as well
22:49:23 <kmc> anyway as usual: a pen that does what I want is minimal, a pen that does what you want is bloated
22:50:00 <Phantom_Hoover> (are you... insinuating that buying expensive things isn't morally justified)
22:50:23 <zzo38> I think I read somewhere that all Disney movies have Mickey Mouse in them, but does that include Pixar?
22:50:53 <kmc> http://www.w3.org/TR/CSS21/aural.html
22:50:58 <Koen__> you shouldn't believe all that you read!
22:51:11 <Bike> Oh, you mean hidden mickeys?
22:51:59 <zzo38> I have sometimes found Mickey Mouse in some Disney movies I have seen, without being told about it before, but mostly I haven't done so.
22:52:29 <zzo38> Phantom_Hoover: I don't know. I haven't seen it in any Pixar movies.
22:52:37 <zzo38> (But maybe I just didn't notice it)
22:53:17 <Phantom_Hoover> i mean i don't know if it's a joke but it's brilliant either way
22:54:30 <kmc> http://static.rust-lang.org/doc/std/local_data.html weird module imo
22:56:48 <Bike> http://www.warriorforum.com/warrior-special-offers-forum/682220-memejacker-limited-reopening-popular-demand-rave-reviews-fastest-results-i-have-seen.html finally, i can make my own website.
22:58:00 <kmc> don't jack my memes
23:00:32 <ion> It took me a while to begin realizing that might not actually be a joke.
23:05:04 <Phantom_Hoover> http://en.wikipedia.org/wiki/List_of_countries_by_income_equality
23:05:42 <Phantom_Hoover> i mean according to these statistics the poorest ten percent of the bolivian population earn more than the richest 10%, afaict
23:06:58 <Fiora> it says 93.9, so that's 93.9 times more for the richest 10% than the poorest 10%
23:07:25 <Bike> that's still positive.
23:07:46 <Bike> the average income of a member of the top 10% is 93.9 or 157.3 times the average income of a member of the bottom 10%.
23:08:01 <Bike> oops, over one i meant, yeah, not positive.
23:08:44 <Bike> unless you think the top 10% of the UK earn a tenth of what the bottom 10% do ;)
23:09:11 <Bike> gini is a percentage, though.
23:09:27 <Bike> (not that it's all that great as a metric but that's another matter)
23:13:36 <Bike> yeah having % in the column title desn't help
23:13:57 <Phantom_Hoover> http://yourkickstartersucks.tumblr.com/post/47199723497/a-set-of-wooden-apple-keyboard-replacement-keys
23:14:47 <FreeFull> Hmm, what would tactical keys be like?
23:15:55 <Bike> "...I could totally see one of them under my iMac screen, sat on my chipped old marble-topped desk" -WIRED
23:16:35 <Bike> ugh this thing is so ugly
23:16:45 <FreeFull> Aren't those basically just pieces of wood, made sticky on one side, and cut to match the keys?
23:17:18 <Bike> it's a tactile experience
23:18:04 <Bike> Every set of Engrain Keys has it's own inherent beauty and natural tactility, turning your keyboard into an individualized work of art that will compliment your desktop, and engage your senses every time you type.
23:18:38 <Bike> Note: Engrain Keys are not recommended for installation on laptops, as the clearance is inadequate when closed and could result in scratching the screen.
23:19:33 <Bike> "The idea for the Engrain Tactile Keys originally spawned from my graduate thesis work " must not stereotype must not stereotype
23:19:52 <Bike> https://s3.amazonaws.com/ksr/assets/000/378/842/d09309f192b8847a72846f77ef5d0cf8_large.jpg?1360358317 aghaghagh
23:20:38 <Bike> it's like typing on lays
23:20:50 <Phantom_Hoover> i wonder if they're actually cut to the grain of the wood or if they just cut some template onto it
23:21:21 <Bike> looks like they just cut holes in a plank
23:21:26 <Bike> you get the rest of the plank with the keys
23:22:09 <Phantom_Hoover> in fact i imagine it's very hard to get any other kind of plank
23:22:33 <Bike> if it's not flat it's not a plank it's just part of a tree
23:26:25 -!- Sgeo has joined.
23:27:47 <kmc> Bike: when I went camping someone else had this amazingly comfortable foam mat made of fancy foam with circular holes cut in it... they told me that there's a company selling earplugs to the military and they sell the leftover sheets of foam as camping mats
23:27:52 <kmc> i was super impressed
23:28:46 <Phantom_Hoover> i'm sure i've heard other stories about industrial processes like that
23:45:32 <shachaf> https://en.wikipedia.org/wiki/Wikipedia:Terminal_Event_Management_Policy
23:46:03 <Fiora> Bike: http://jeb.biologists.org/content/early/2013/06/27/jeb.087809
23:46:20 <Bike> Fiora: yeah that's been going around
23:46:27 <Bike> fucking crazy far as i'm concerned
23:46:52 <Bike> i guess i should actually read the paper, the news i've seen is a bit ambiguous
23:47:01 <Bike> something about needing a refresher training session after the decaptation
23:47:21 <Bike> (man jeb has the coolest papers)
23:49:29 <Bike> also good: i've been to campus twice, haven't paid tuition yet, aand have still gotten dozens of papers from campus access.
23:55:59 <Bike> Fiora: ...the paper has line numbers. i have never seen this before.