00:00:59 <FireFly> Hm, what was that anagolf task again...
00:01:51 <oerjan> > 1/(1/(1/0.30103-3)-3)
00:02:19 <oerjan> > 1/(1/(1/(1/0.30103-3)-3)-9)
00:02:29 <oerjan> > 1/(1/(1/(1/(1/0.30103-3)-3)-9)-2)
00:02:45 <oerjan> > 1/(1/(1/(1/(1/(1/(1/0.30103-3)-3)-9)-2)-2)-4)
00:03:30 <oerjan> > 1/(1/(1/(1/(1/(1/(1/0.30103-3)-3)-9)-2)-2)-4) :: Rational
00:04:38 <oerjan> > 1/(1/(1/(1/(1/(1/(1/0.30102-3)-3)-9)-2)-2)-4) :: Rational
00:06:03 <FireFly> oh, you changed a decimal digit
00:08:01 -!- Froox has changed nick to Frooxius.
00:08:18 <elliott> here I was hoping fizzie would take that bait too
00:08:27 <elliott> I guess I could, like, "measure" it or whatever
00:12:44 <shachaf> I think ret is generally easier to predict than an arbitrary indirect jump.
00:18:00 <int-e> shachaf: Right, but it wouldn't have to be that way. It's because compilers know that using 'ret' for returning to some other point than where the previous "call" was comes at a stiff price. I'm sure they would happily take advantage of a cheap instruction that just happens to jump to the address on top of the stack, popping it.
00:19:13 <int-e> The stiff price being that no CPU will ever predict the target correctly.
00:20:08 <int-e> and, of course, a cascade of similar failures further down the stack.
00:35:03 <FireFly> Why did anagolf accept a trailing space
00:35:22 <FireFly> I thought it was annoyingly anal about things like that
00:35:23 <ais523> it's not its job to correct your stupid mistakes
00:36:12 <HackEgo> [wiki] [[MNNBFSL]] http://esolangs.org/w/index.php?diff=41216&oldid=41202 * Oerjan * (+96) include original Japanese
00:37:16 <oerjan> FireFly: it is anal except at the very end of output
00:43:18 -!- boily has joined.
00:43:46 <HackEgo> [wiki] [[GridScript]] http://esolangs.org/w/index.php?diff=41217&oldid=41215 * Oerjan * (+48) fmt
00:54:20 -!- callforjudgement has joined.
00:54:55 -!- ais523 has quit (Read error: Connection reset by peer).
00:56:41 -!- callforjudgement has changed nick to ais523.
00:56:51 <HackEgo> [wiki] [[REGXY]] http://esolangs.org/w/index.php?diff=41218&oldid=7994 * Oerjan * (+91) The link on C2 Wiki was bitrotted so include the right one here
01:07:31 -!- Dulnes has joined.
01:07:38 <HackEgo> [wiki] [[Reg Xy]] N http://esolangs.org/w/index.php?oldid=41219 * Oerjan * (+65) Add redirect
01:11:33 <elliott> happy thank give to you too
01:12:29 <Dulnes> Im doing swell i guess
01:15:01 <elliott> nothing I want to go into in #esoteric :p
01:16:32 <Dulnes> These § why are they even added
01:16:47 <elliott> aw c'mon, I'd give that answer tons, it's just not a common question in here :p
01:17:45 <Bicyclidine> oh, right, it's section mark if you're not a nerd like i am
01:19:36 <HackEgo> [wiki] [[REBEL]] M http://esolangs.org/w/index.php?diff=41220&oldid=41103 * Kendfrey * (-11) Changed link to point to new site
01:22:15 <elliott> Taneb: pikhq_: how are *you* on this merry thank give
01:22:20 <elliott> yeah that was just an excuse to say thank give again
01:23:04 <Taneb> Well, I'm listening to I'm Sorry I Haven't A Clue
01:23:21 <pikhq_> I have no pants on and am playing Pokemon.
01:25:10 -!- mhi^ has quit (Quit: Lost terminal).
01:26:18 <ais523> I've been sort-of considering getting that myself
01:26:37 <Taneb> Not sure if I should go for Alpha Sapphire or Omega Ruby though
01:26:41 <ais523> but I'd need to buy a new console
01:26:58 <pikhq_> My impression is that it's where they actually put most of their work into, rather than XY.
01:27:02 * oerjan goes back to tatham's puzzles
01:27:25 <FireFly> Not sure if I can justify getting OR/AS given how much (or little, rather) time I spent on X/Y
01:27:58 <Dulnes> Taneb both are exactly the same and are relativaly short only difference is the main villain and the new graphics
01:28:12 <Taneb> Dulnes, hence the hard decisions
01:28:21 <ais523> pikhq_: well they put most of the work into ORAS back when they were making Ruby/Sapphire
01:28:25 <Dulnes> Team nerd or team bara
01:28:56 <Dulnes> Magma or what was the other one i only played ruby
01:28:58 <pikhq_> Hence why ORAS is where they threw a lot of gameplay tweaks.
01:29:25 <boily> quintopia: QUINTHELLOPIA! (should you ever answer while you're mexicaning.)
01:29:36 <FireFly> They had the new-rendering-engine thing for XY
01:29:40 <Dulnes> Yeah in the caves you cant traverse it by memory
01:29:50 <Dulnes> You need to defeat the gym leader
01:29:59 <FireFly> and new engine overall, I guess
01:30:21 * pikhq_ has greatly appreciated little details like "offering to teleport you instead of forced backtracking".
01:30:57 <pikhq_> Also, DexNav is glorious.
01:31:00 <Dulnes> My wife says Hi... i have to go eat family dinner
01:31:19 <pikhq_> It's like breeding Pokemon, only much less agonizing!
01:31:43 <ais523> what is DexNav, anyway?
01:31:58 <Taneb> pikhq_, if you've go Sapphire I shall get Omega Ruby!
01:32:18 <ais523> oh, it managed to get onto Bulbapedia before they locked the site down due to hacking
01:32:24 <elliott> sapphire was my first pokemon game
01:32:41 <Taneb> Emerald was mine, I'm young
01:32:48 <pikhq_> There's a menu in the PokeNav that lets you see Pokemon in the route you're in. You can tap on one of them to find more of 'em...
01:32:49 <elliott> aren't you like a lot older than me
01:32:53 <FireFly> Gold was mine, I'm slightly less young
01:33:00 <ais523> oh, it's basically Habitat List from BW2, but prettier
01:33:06 <elliott> I guess I'm actually just as old as old people now though
01:33:24 <Taneb> elliott, I'm like 9 months older
01:33:25 <Bicyclidine> but yeah i didn't get into pokemon like everyone i know did
01:33:33 <pikhq_> More you encounter the Pokemon, the better the Pokemon you find with DexNav are.
01:33:36 <ais523> Habitat List is great, anyway
01:34:00 <elliott> Taneb: what if you being born somehow metaphysically caused me to be conceived (per the hexham synchronicity)
01:34:04 <pikhq_> So you get decent odds of shiny Pokemon, of higher level, with egg moves, and 3+ maxed IVs.
01:34:45 <HackEgo> [wiki] [[GridScript]] http://esolangs.org/w/index.php?diff=41221&oldid=41217 * SuperJedi224 * (-63) /* 99 Bottles of Beer on the Wall */
01:35:14 <ais523> my theory is that the Pokémon people decided there was no way they could stop RNG abuse, and decided to make it unnecessary
01:35:20 <ais523> however, wild Pokémon with egg moves seems wrong to me
01:36:11 <pikhq_> It's a much smaller list of moves than the full list of possible egg moves.
01:36:15 <Taneb> elliott, somehow I doubt it
01:36:16 <pikhq_> Oh, also chance of hidden ability.
01:36:33 <HackEgo> Hexham es la ciudad mas importante de programación esotérico
01:36:47 <ais523> that said, I'm upset at the "no illegal egg move combinations"
01:37:08 <ais523> I'm /not/ upset at hidden abilities becoming commonplace, the old way of treating them like event Pokémon drove me out of competitive for a year
01:37:15 <ais523> (until BW2 mostly fixed it using Hidden Hollows)
01:37:30 <HackEgo> [wiki] [[GridScript]] http://esolangs.org/w/index.php?diff=41222&oldid=41221 * SuperJedi224 * (+0) /* 99 Bottles of Beer on the Wall */
01:37:37 <pikhq_> XY had the Friend Safari making them more common, and additionally made them breadable.
01:38:22 <pikhq_> Also fun, XY + ORAS has all the non-event legendaries.
01:38:31 <pikhq_> And one of the event legendaries.
01:38:31 <ais523> pikhq_: they always were breedable, from females
01:38:42 <ais523> getting the first female was very hard in BW and quite hard in B2W2, though
01:38:42 <pikhq_> Males have a low chance now.
01:38:54 <pikhq_> (Deoxys is just in ORAS' postgame)
01:38:57 <ais523> and some things they wanted to be rare were male only, e.g. eevee, starters
01:39:42 <pikhq_> Not to mention XY made IV breeding much, much simpler.
01:39:57 <ais523> I rather liked the old method where you screwed around with the RNG
01:40:16 <ais523> that said, I was recently doing random-IVs + flawless ditto to make the first round of parents
01:40:21 <ais523> one of them was quad-flawless (!)
01:40:26 <ais523> this was just for Subway, so I'm using it
01:40:31 <ais523> (quad in useful stats, as well)
01:40:46 <pikhq_> You have a flawless ditto? Wonderful.
01:41:07 <pikhq_> Congrats, it's possible to pass 5 of those to any Pokemon you want.
01:41:14 -!- Phantom_Hoover has quit (Quit: Leaving).
01:41:22 <pikhq_> Natures can also be inherited.
01:41:36 <ais523> pikhq_: natures have been inheritable for absolutely ages
01:41:40 <ais523> since the original R/S, I think
01:41:54 <ais523> also, destiny knot doesn't determine a specific parent to pass IVs
01:42:16 <ais523> flawless ditto was one of my earlier RNGs
01:42:35 <ais523> (the first was an event arceus, which was a bad place to start in retrospect, it took 11 hours and I still didn't get Adamant)
01:43:47 <Dulnes> In black and white on route 3 a shiny zorarork or however you spell it has a chance of just appearing
01:43:49 <pikhq_> Gen VI changes Destiny Knot to being 5, and the EV items force a specific IV as well.
01:44:00 <pikhq_> Still doesn't determine whose IVs you get, but hey.
01:44:07 <ais523> Dulnes: I don't believe you
01:44:09 <ais523> actually I can possibly prove it
01:44:25 <pikhq_> Yeah, wasn't Zoroark B/W shiny-locked?
01:44:27 <Dulnes> A shuckle can have the highest damage though
01:45:02 <ais523> pikhq_: more to the point, Zoroark and Zorua never appear wild in any version, barring Memory Link
01:45:09 <ais523> err, any gen 5 version
01:45:28 <ais523> I have a Zorua from Memory Link that can be bred into infinite zoroarks if necessary
01:45:30 <Dulnes> Mm then idk where this zorarok came from
01:45:39 <ais523> (together with straight-30 IVs, just like other Memory Link mons)
01:45:43 <Dulnes> Says i met it on route thtee
01:45:55 <pikhq_> AH, right. BW2 on, a parent holding an everstone will pass down its nature in particular.
01:45:55 -!- tlewkow has quit (Remote host closed the connection).
01:45:56 <FireFly> Hacked 'mon traded on GTS?
01:46:01 <pikhq_> That was the exact change.
01:46:15 <ais523> pikhq_: before that, there's a 50% chance
01:46:29 <pikhq_> I didn't pay much attention to breeding before XY.
01:47:04 <pikhq_> Hah, right. It's stupid, but one other thing: Pokeballs are also inherited in gen VI. :)
01:47:49 <ais523> pikhq_: I did know that
01:47:50 <FireFly> From which of the parents?
01:48:03 <ais523> Luxury Balls are thus what you should use for catches, if you can
01:48:18 <ais523> pikhq_: basically, AFAICT, Nintendo have decided to make legality checking as hard as possible
01:48:24 <ais523> to hide the fact that they're rubbish at it
01:48:47 <pikhq_> I'm kinda glad about how they implemented it though...
01:49:00 <pikhq_> This basically means that a cheater is not likely to have a huge advantage.
01:49:20 <Dulnes> I heard a lvl 100 shuckle if used correctly can deal the largest ammount of damage in the game
01:49:33 <ais523> Dulnes: "used correctly" = "the opponent cooperates"
01:49:54 <Dulnes> No i think.its the moves and items you use
01:50:47 <Dulnes> The opponent doesnt have to do much you just need really high defense to stand against its atk until you get your move set out
01:50:59 <ais523> Dulnes: is your opponent using Shedinja (maximum percent damage) / whatever it is for maximum raw damage?
01:51:15 <ais523> and no, the opponent also has to not switch while you're putting all the defence drops on them
01:51:23 <ais523> and not just OHKO the shuckle when you power trick
01:51:57 <Bicyclidine> me, i'm just excited to learn that metronome can do metronome
01:52:14 <Taneb> Bicyclidine, does it recurse?
01:52:45 <pikhq_> I like that Surfing on a Sharpedo is much faster now.
01:52:53 <FireFly> It'd be fun if it could be made to recurse multiple times in a row with RNG abuse
01:53:06 <FireFly> pikhq_: Sharpedo specifically?
01:53:22 <pikhq_> Surf in general was sped up, and Sharpedo in particular goes twice as fast.
01:53:29 <Taneb> With a malicious RNG it could recurse indefinitely
01:55:18 <Bicyclidine> i haven't played a pokemon in years, but i'm guessing yes because that sounds funny
01:56:23 -!- tlewkow has joined.
01:57:25 <pikhq_> Also, the encounter rate in water was reduced a lot; yay.
02:02:40 * pikhq_ has been getting a lot of use out of his 3DS.
02:03:48 <ais523> hmm, I spent some of last night reverse-engineering the formula for what happens when you use Rock Smash in HGSS
02:06:29 -!- hjulle has quit (Ping timeout: 264 seconds).
02:23:42 -!- J_Arcane_ has joined.
02:26:21 -!- J_Arcane has quit (Ping timeout: 265 seconds).
02:26:36 -!- J_Arcane_ has changed nick to J_Arcane.
02:32:08 -!- callforjudgement has joined.
02:32:22 -!- ais523 has quit (Read error: Connection reset by peer).
02:33:16 -!- AndoDaan has joined.
02:35:24 -!- boily has quit (Quit: APPLICATIVE CHICKEN).
02:35:25 -!- tlewkow has quit (Remote host closed the connection).
02:35:43 -!- tlewkow has joined.
02:46:09 <pikhq_> callforjudgement: Alas, if you want ORAS you should probably wait for the New 3DS to come out.
02:46:51 <callforjudgement> pikhq_: well, I'm a bit dubious about the whole 3DS thing in the first place, for the same reason I dislike the DSi
02:46:58 <pikhq_> Nintendo, in a typical show of being-Nintendo, managed to make a new DS model only come out along with a new Pokemon game in Japan.
02:47:05 <pikhq_> Dulnes: That's what it's called, the New 3DS.
02:47:19 <Dulnes> And does it do? new stuff
02:47:19 <pikhq_> callforjudgement: Unlike the DSi, it's a rather notable upgrade in hardware.
02:47:39 <pikhq_> callforjudgement: And there's, y'know, games worthwhile on it.
02:48:06 -!- callforjudgement has changed nick to ais523.
02:48:11 <pikhq_> Dulnes: Faster CPU, has a second analog stick, has face tracking so that the 3D sucks less.
02:48:18 <ais523> I'm more worried about things like streetpass
02:48:36 <pikhq_> Both are only ever used in fairly trivial ways by games.
02:48:36 <Dulnes> Ok whats the second analog stick do?
02:48:50 <madbr> is it equivalent to the "circle pad"?
02:48:57 <Dulnes> Im perfectly content with the DS i have
02:48:57 <pikhq_> madbr: It's more an analog nub.
02:49:16 <Dulnes> Why would they make a new one why not a new system in general
02:49:17 <pikhq_> It's presented as a second stick to games coded to read it, like Smash Bros.
02:49:49 <pikhq_> It also works for games that used the circle pad addon.
02:49:52 <Dulnes> Well the wii U is very slick
02:50:09 <Dulnes> Especially all these games they pumped out this yesr
02:50:25 <pikhq_> ais523: I think the most notable use of Streetpass in any 3DS game I'm aware of is the Streetpass Plaza games which come on the system.
02:50:39 <ais523> can I delete or otherwise get rid of them?
02:50:43 <pikhq_> It's otherwise a trivial, very minor feature that is 100% ignorable if you don't care about it.
02:50:48 <pikhq_> No, it's part of the firmware.
02:51:11 <Dulnes> You cant jailbreak it?
02:51:16 <pikhq_> Though, it won't streetpass *at all* unless you go in and set it to.
02:51:41 <pikhq_> Basically literally everything that does streetpass asks you if you want to enable streetpass functionality for the game.
02:51:59 <Dulnes> Majoras mask remake though
02:52:00 <pikhq_> (and you can go into the system settings later and undo it)
02:52:13 <madbr> wii u finally has its smash bros
02:52:56 <Dulnes> Nintendo is making every one hyped up for their latest installation to the LoZ series in 2015
02:52:58 <pikhq_> It also only streetpasses when the Wifi is on.
02:53:22 <Dulnes> If its as cartoony as the trailer i wont bother for it
02:53:30 <madbr> at least one thing nintendo understands: video games aren't movies
02:53:52 <pikhq_> The play coin thing isn't able to be turned off, but it also matters even less.
02:54:07 <pikhq_> ais523: Any other concerns?
02:54:23 <Dulnes> Whats the point of the puzzles
02:55:08 <ais523> pikhq_: the reason I didn't impulse-buy one a while back is that apparently they don't come with chargers in the EU
02:55:14 <pikhq_> That's just the 3DS XL.
02:55:22 <ais523> oh, the regular-size one does have a charger?
02:55:22 <pikhq_> THough I don't know *why*.
02:55:36 <Dulnes> Btw i bought my 3ds this June and still havent figured.it all out
02:56:29 <pikhq_> It uses the same charger as the DSi,FWIW.
02:56:29 <Dulnes> The XL has a charger or am i reading that wrong
02:56:44 <pikhq_> Dulnes: The US 3DS XL is the only one that comes with a charger.
02:57:08 <ais523> pikhq_: I don't have a DSi, though
02:57:11 <ais523> because the original is better
02:57:12 <pikhq_> Likewise the US New 3DS will be the only one that comes with a charger.
02:57:29 <pikhq_> I have a 3DS and an original DS. :)
02:57:43 <Dulnes> ? why does the Us only have chargers
02:57:48 <pikhq_> Dulnes: FUck if I know.
02:57:52 <madbr> so many game designers see what they're doing as building narratives and stories and whatever
02:57:56 <pikhq_> Admittedly the 3DS makes the DSi literally pointless.
02:58:02 <madbr> games are not movies
02:58:05 <pikhq_> There is nothing the DSi does that the 3DS does not.
02:58:35 <Dulnes> LoZ is very good with its time travel sequences that if thought about to long hurt
02:58:35 <zzo38> Yes I agree games are not supposed to be movies; many games have too many cutscenes!
02:58:54 <pikhq_> Even though what the DSi does that the DS does not is, well, hardly anything. :)
02:59:10 <ais523> pikhq_: what the DSi does that the DS does not is why I don't like the DSi, though
02:59:12 <pikhq_> I think the most notable is that some games, like BW, can use WPA.
02:59:13 <Dulnes> Nintendo has short cutscenes that get straight to the point but still set a story
02:59:15 <ais523> things like the console having its own memory
02:59:21 <ais523> and accepting downloads
02:59:26 <ais523> this is not what I want from a games console at all
02:59:37 <ais523> Dulnes: I take it you've never played Pokémon Mystery Dungeon
02:59:43 <ais523> it has cutscenes long enough that they have multiple save points in
02:59:50 <pikhq_> ais523: The 3DS only has its own memory for DSi compat though; uses an SD card for storage otherwise.
02:59:50 <ais523> also, flashbacks to earlier in the same cutscene
03:00:04 <pikhq_> (and it works just fine without an SD card, though downloads won't work then)
03:00:06 <ais523> what's wrong with just storing things on the cartridge?
03:00:11 <Dulnes> ais523: that sounds awful
03:00:11 <zzo38> I happen to prefer just text windows that you can read at your own speed (or skip if you prefer) rather than having a lot of video cutscenes and stuff
03:00:19 <pikhq_> Note that 3DS carts *do* save on the cartridge.
03:00:55 <ais523> "Also since i cannot access the IRC because of some weird issue all i know is that im going to be a detrament to this community"
03:01:00 <Dulnes> Nintendo does put the skip option in for some games and all games have the manual text skip in a cutscene
03:01:10 <pikhq_> The SD card is used for downloaded games, pictures taken with the camera app, and game patches.
03:01:19 <ais523> XY needed several of those
03:01:32 <zzo38> I think save game should be stored in a memory card (such as CF cards), rather than in the console or in the cartridge (although it can be a cartridge that is also a memory card; of course a DVD or CD will be read-only though)
03:02:06 <ais523> Pokémon don't like people copying saves because you can clone Pokémon that way
03:02:17 <ais523> so presumably the saves are locked to the SD card somehow, for downloaded versions
03:02:28 <Dulnes> Do you have to buy the SD card in other countrys seperately from the console
03:02:48 <Dulnes> It comes with inside of it already
03:03:26 <Dulnes> I swear im having a brain fart
03:04:23 <pikhq_> Ah, so that's how they do it. The 3DS on-board storage is accessible to games...
03:04:34 <pikhq_> Pokemon stores a small key on the console.
03:04:44 -!- AndoDaan has quit (Read error: Connection reset by peer).
03:04:45 <pikhq_> Changing key each time it saves.
03:04:54 <ais523> for downloads, presumably
03:05:00 <ais523> I hope you can move the cart between consoles
03:05:10 -!- AndoDaan has joined.
03:05:48 <pikhq_> I suspect that they're using generic infrastructure for downloads there.
03:06:14 <madbr> In theory, the story of half life 2 is about aliens taking over. In practice, half life 2 is about the story of a mute repeatedly launching filing cabinets at people’s heads.
03:06:16 <pikhq_> (just like on the Wii, downloads are console-specific)
03:06:58 -!- AndoDaan has quit (Read error: Connection reset by peer).
03:07:03 <pikhq_> ais523: FWIW, I've even tested that.
03:07:20 <pikhq_> Just a matter of taking my AS cart and putting it in my girlfriend's 3DS.
03:07:21 -!- AndoDaan has joined.
03:08:17 <ais523> at this rate, we're going to have to prove the Pokémon video game TC so that it's ontopic
03:08:48 <pikhq_> Yeah, Gen I and II are TC.
03:08:49 <Bicyclidine> http://arxiv.org/abs/1203.1895 i find it hard to believe you haven't seen it, though, so maybe you mean something else?
03:09:08 <Bicyclidine> oh, wait, that's complexity not computation
03:09:24 <pikhq_> Gen I and II can be made to execute arbitrary code. :)
03:10:02 <ais523> Bicyclidine: oh, I have seen that, and realised what it was as soon as I saw the existence of the link
03:10:10 <ais523> (also, that's technically PSPACE-complete)
03:10:13 <ais523> pikhq_: Gen III's very close, too
03:10:21 <ais523> you should see what people do with the Pomeg Glitch
03:10:33 <pikhq_> Yeah, it's not actually arbitrary, but there's a decent bit of PEEK and POKE available with that one.
03:11:03 * pikhq_ is particularly fond of Pokemon glitched speedrunning
03:11:15 <ais523> http://bulbapedia.bulbagarden.net/wiki/Pomeg_glitch#Access_Pok.C3.A9mon_beyond_slot_6
03:11:26 <ais523> it's only a small step from there to completely screwing up memory
03:11:30 <ais523> as you already have a buffer overflow
03:11:45 <pikhq_> Yeah. It's not quite there but it's really close.
03:12:16 <pikhq_> And quite entertaining for speedrunning.
03:13:26 <pikhq_> Alas. I want to eat. I do not want to cook. I don't have leftovers. It is Thanksgiving.
03:14:03 <pikhq_> Probably keep watching MST3K.
03:15:43 <J_Arcane> I now have a customizable string split function. Go-go gadget easy parsing.
03:15:46 <J_Arcane> There's a slight bug though. When two delimiters get stuck next to each other you wind up with a little null string in the middle.
03:16:23 * pikhq_ has a gzip util that sucks less than standard, to go along with a zlib that sucks less than standard.
03:16:28 <ais523> that might be intentional, it's what you'd intuitively expect given what splitting does
03:17:18 <pikhq_> Pity the zlib API sucks.
03:17:36 <J_Arcane> ais523: Yeah. I think it's unavoidable given my algo, and it's easily filter'd for.
03:18:30 <pikhq_> elliott: Well, for starters it's significantly smaller.
03:18:55 <pikhq_> And it actually uses zlib instead of having its own code for it forked from an early version of infozip.
03:19:05 <elliott> haha, gzip doesn't use zlib?
03:20:06 <elliott> is this because GNU or something
03:20:11 <J_Arcane> I gotta say though, 4am is a really shitty time to finally figure out how to write your function ...
03:20:27 <pikhq_> I think it predates zlib actually.
03:21:00 <J_Arcane> elliott: Well, it's 5:19 here, but it was 4 when I woke with a horrible cramp and then couldn't sleep because my brain chose *that moment* to solve my funciton.
03:22:00 <pikhq_> The opposite-endianness of zlib vs. PNG makes me even more upset now.
03:22:06 <pikhq_> zlib was invented for the sole purpose of PNG.
03:23:42 <Taneb> It's impressive that GNU hello has 184 lines
03:24:00 <J_Arcane> It also took me longer because my comments didn't accurately describe what left$ and right$ do, and I'm considering changing what they do to be something more useful. It really seems like they should take the left or right half from a pivot index, rather than taking X from left or right. The latter seems less useful.
03:24:02 <madbr> pikhq : whatare their respective endianness?
03:24:10 <Bicyclidine> you know what's impressive? defibrillators
03:24:12 <ais523> Taneb: note that GNU hello exists as a sandbox for people to practice doing development correctly on
03:24:18 <Bicyclidine> just think about it. who came up with that? how did they /test/ it
03:24:26 <ais523> you should see the Debian packaging for GNU hello, it's just as overengineered for the same reason
03:24:32 <Taneb> ais523, oh really?
03:24:53 <Bicyclidine> "The primary purpose of GNU Hello is to demonstrate how to write other programs that do these things; it serves as a model for GNU coding standards and GNU maintainer practices."
03:25:16 <pikhq_> PNG uses big-endian for everything.
03:25:26 <pikhq_> zlib uses little-endian for its header.
03:25:43 <madbr> which one came first?
03:26:12 <int-e> Taneb: I find it more amazing that it still can't read mail.
03:26:21 <elliott> int-e: I thought it actually could
03:26:28 <pikhq_> zlib was the compression portion of the PNG reference implementation.
03:27:35 <pikhq_> They literally sat down and decided "it'll be big endian here and little endian here".
03:27:38 <zzo38> DVI is big-endian, Z-machine is almost always big-endian (no small-endian files exist), Knuth's MMIX instruction set is big-endian, etc. A lot of DOS programs use file formats with small-endian (since a PC uses small-endian natively).
03:28:08 <zzo38> Also programs that were originally designed for DOS tend to use small-endian even if they are now ported to other systems.
03:28:36 <madbr> don't most architectures that started 8 bits ended up little endian?
03:28:50 <madbr> like 6502 if I'm not mistaken
03:29:30 <madbr> trying to figure where the cultural divide comes from
03:29:34 <int-e> elliott: ah, yes it could. http://git.savannah.gnu.org/cgit/hello.git/commit/?id=1a962e1e873d382c376921ef41ad234bd1bf4339
03:30:11 <elliott> int-e: how dare they?! that feature is important to me!
03:31:00 <elliott> nice "portable" alloca implementation
03:31:45 <madbr> "Historically, byte order distinction was born out of the mainframe vs. microprocessor approach.[dubious – discuss] Until the 1970s virtually all processors were big-endian."
03:31:54 <pikhq_> elliott: What, in gnulib?
03:32:25 <pikhq_> Wait, no, that's not from gnulib.
03:33:00 <elliott> and doesn't actually allocate on the stack?
03:33:10 <elliott> it has linked lists. alloca using linked lists. why would you ever use this
03:33:36 <zzo38> Yes 6502 is also small-endian
03:33:41 <pikhq_> I'm pretty sure that's actually a property of Cray API.
03:33:58 <ais523> the basic difference between big-endian and little-endian is what happens if you cast a pointer from pointing at one size of integer, to pointing at a smaller size
03:34:08 <pikhq_> It seems as though the Cray *stack* formed a linked list.
03:34:12 <ais523> big-endian does an approximation scaled based on the full range; little-endian does modulo
03:34:13 <zzo38> ais523: Yes, and for that purpose small-endian works better
03:34:22 <ais523> and modulo is what's better if the value is actually within the range you want already
03:34:43 <madbr> little endian: x86, 6502, z80, dec alpha, atmel, vax
03:35:37 <madbr> big endian: 68k, superh, power, other atmel, system/360
03:36:37 <pikhq_> madbr: New POWER systems from IBM are little endian actually.
03:36:46 <zzo38> There are also systems with no endianness, although these are mostly VMs I suppose; in such a case the file format still has endianness although the runtime won't
03:37:04 <madbr> pikhq_ : it's like they changed the whole ABI?
03:37:11 <pikhq_> For reasons I'm not entirely sure of they decided to switch endianness and ABI.
03:37:28 <pikhq_> Note though that POWER has always been a bi-endian architecture.
03:37:43 <madbr> ARM is also bi-endian in theory
03:37:51 <madbr> in practice, it's little endian
03:37:56 <int-e> pikhq_: it's not uncommon for a stack to be organized in a linked list of stack frames; this facilitates unrolling the stack (for debuggers, or possibly for delivering exceptions)
03:38:12 <int-e> pikhq_: a tad old-fashioned, perhaps
03:38:15 <pikhq_> Yeah, POWER just had it be big endian in practice.
03:38:18 <zzo38> Some file formats can use both big-endian and small-endian, such as TIFF and Z-machine (although small-endian was never used, and was removed in EZIP). Some formats use the host's endianness (and int size), such as OASYS.
03:38:37 <pikhq_> Though run-time switchable.
03:38:54 <madbr> switchable endian formats sound like a bad idea to me
03:39:18 <elliott> 03:33:32 <pikhq_> It seems as though the Cray *stack* formed a linked list.
03:39:28 <madbr> do you want ppl to implement support for your format or do you want to drive them away?
03:39:39 <zzo38> madbr: Yes it does, but, TIFF has that feature. Z-machine had it but they removed that feature, I suppose they then knew how bad it is.
03:40:11 <pikhq_> IIRC TIFF did it as a weird compromise.
03:40:22 <pikhq_> TIFF is one of those "design by committee" standards.
03:41:01 <int-e> . o O ( People argue about little ends and big ends when we all know that eggs naturally come to rest on their sides... )
03:41:21 <madbr> also the reason why C/C++ still has digraphs after all these years
03:41:33 <pikhq_> That's even simpler actually.
03:41:39 <pikhq_> IBM *actually uses those*.
03:41:40 <Bicyclidine> the fact endianness is named after a joke argument and people still argue over it is fucking incredible, i love it
03:42:03 <pikhq_> Some of the EBCDIC code pages actually, really need it.
03:42:04 <zzo38> int-e: If you put into a flat surface it will probably fall down and end on its side, but, it seems that some people can manage to make it to stand up straight!
03:42:06 <Taneb> Middle-endian best endian
03:42:14 <pikhq_> And the jerks won't adopt UTF-EBCDIC.
03:42:18 <madbr> pikhq : ebcdic is also an abomination
03:42:28 <pikhq_> Or go out and shoot OS/360 in the head.
03:42:35 <zzo38> EBCDIC is much worse than ASCII, I agree that much.
03:42:48 <pikhq_> EBCDIC also was a mistake at the time.
03:44:00 <pikhq_> (no, literally: the IBM 360 mainframe was supposed to be ASCII or EBCDIC, with EBCDIC purely for compatibility with BCDIC data. But they couldn't fix bugs in the ASCII mode before they released.)
03:44:06 <elliott> int-e: side-endian would be, like, 32-bit values a, b, c and d would be laid out in memory as the bytes abcdabcdabcdabcd?
03:44:30 <int-e> Bicyclidine: to be fair, the joke argument is the basis of a whole book.
03:44:40 <pikhq_> int-e: NAh, it's the basis of a chapter.
03:45:15 <pikhq_> Lilliput is just a chapter in the Travels.
03:45:40 <pikhq_> Admittedly it's the first one.
03:46:12 <ais523> it's one chapter of four, IIRC
03:46:16 <ais523> so a quarter of a whole book?
03:46:40 <ais523> anyway, the chapter was about the way that ridiculous arguments can come about, so naming the computer science debate after the fictional one makes a ton of sense
03:46:48 <madbr> it's all that series of things that have no purpose other than breaking c++ code when you port it
03:47:21 <madbr> file system differences
03:47:32 <pikhq_> Writing C on a mainframe is pretty hilarious, too, cause you do it inside a Unix environment.
03:47:39 <pikhq_> A Unix environment... with EBCDIC.
03:48:03 <int-e> elliott: mixed endianness exists. ethernet encodes bytes in a little endian order, but the protocols above usually use big endian byte order.
03:48:29 <madbr> also: strange C99 features that aren't in C++
03:48:36 <zzo38> If you need to, don't write the program directly for the mainframe but for an emulator and run the emulator on the mainframe computer.
03:48:44 <int-e> so ... a 16 bit word would be transmitted as 89ABCDEF01234567
03:49:25 <elliott> int-e: yes; what I meant was more interleaving values with each other
03:49:25 <madbr> like variable size stack arrays
03:49:28 <int-e> (I'm numbering bits mathematically; 0 is the bit with value 2^0)
03:49:29 <zzo38> pikhq_: That is a possibility, but it could even be Inferno, or a PC emulator, or whatever
03:49:47 <int-e> elliott: yeah, but it's not interleaving that's happening here.
03:49:47 <pikhq_> Just naming Java because I know it's actually used on mainframes.
03:50:02 <elliott> int-e: but interleaving is closer to the egg on its side :)
03:50:02 <pikhq_> I suspect other emulators *might* actually need porting effort.
03:50:20 <madbr> C++ to C linking insanity
03:50:22 <pikhq_> zzo38: Of course a still simpler way of doing it is to just require a Linux VM on their hypervisor.
03:50:38 <zzo38> And then you can have your program to use ASCII even if it is a EBCDIC computer.
03:51:01 <pikhq_> At which point the only weird thing is that the system console is still one of IBM's really weird terminals rather than a VT100 alike.
03:51:03 <zzo38> pikhq_: How exactly does that work? I don't know a lot about that things
03:51:30 <ais523> there's probably some format out there that genuinely does use the forward slashes :-)
03:51:52 <pikhq_> (this is not observable from the OS point of view; the kernel maps between that terminal and a VT100)
03:52:03 <pikhq_> zzo38: Easily. Linux just has a mainframe port.
03:52:15 <elliott> int-e: nice, that 2^0 thing is a good argument for little endian actually
03:52:48 <ais523> elliott: AFAICT the only reason the argument came about in the first place is historical accident with the way people write numbes
03:52:48 <pikhq_> If you feel like it, and have oodles of money, you can just get an IBM mainframe and run nothing but Linux on it.
03:53:07 <pikhq_> A more *realistic* use of this is to get an IBM mainframe, run the IBM hypervisor, and run a *ton* of Linux on it.
03:53:13 <ais523> pikhq_: actually, ISTR the mainframes are cheaper if you only run Linux on them?
03:53:36 <ais523> like, they have some way of slightly damaging the hardware so that it won't run the old mainframe OSes, but that Linux can worka round
03:53:39 -!- shikhin has joined.
03:53:39 <ais523> presumably a reversible one
03:53:43 <elliott> what kind of specs do modern monolithic IBM mainframes have these days, anyway?
03:53:53 <ais523> to make sure that you aren't cheating and trying to run something other than Linux on a Linux-only mainframe
03:54:00 <pikhq_> The mainframe OS stuff is an excuse to leach money from big old companies anyways.
03:54:45 <madbr> old company cannot change due to inertia and people too close to retirement -> milk out $
03:54:52 <ais523> pikhq_: as I understand it, mainframes were once the only systems that could do what they do
03:54:54 <pikhq_> elliott: 5.5GHz hex core CPUs, up to 101 such CPUs.
03:55:04 <ais523> and nowadays, there's the reasonable alternative of "huge cluster of commodity hardware"
03:55:21 <pikhq_> ais523: Oh, certainly.
03:55:26 <elliott> pikhq_: and oodles of RAM?
03:55:52 <pikhq_> Way back in the day if you were a big enough company to really need computing power, a mainframe was the economical option.
03:56:28 -!- shikhout has quit (Ping timeout: 240 seconds).
03:56:45 <madbr> how much overlap does this have with cobol?
03:57:13 <pikhq_> IBM mainframe users tend to run COBOL extensively.
03:57:50 <zzo38> I did read somewhere that someone went to a store that sells computer, and asked for a mainframe computer because he/she wanted to learn to program in COBOL.
03:57:57 <pikhq_> elliott: They offer 3TB of RAM.
03:58:07 <zzo38> But now we have OpenCOBOL so you can run it on a PC and stuff too.
03:58:08 <pikhq_> And then 6.4TB of SSD for swap.
03:58:15 <elliott> offer as in "up to" or as in what people actually get? nice, either way
03:58:28 <pikhq_> As in "if you wish to pay for that much, this is what you get".
03:58:28 <elliott> I guess that's "only" ~30 gigabytes per CPU if you get all of them
03:58:45 <elliott> I assume completely maxed-out systems aren't the norm, though...
03:59:29 <pikhq_> Actually, they probably are. These big suckers are leased by CPU usage, not by system specs.
04:00:15 <elliott> I think I knew that but forgot
04:00:18 <pikhq_> The smaller ones I think are sold.
04:01:49 <madbr> this is like... a collection of everything I want to stay as far as possible from :3
04:02:01 <madbr> buisiness computing etc
04:03:08 <pikhq_> A brand new one can cost millions.
04:03:20 <zzo38> Well, at least now you don't have to do all that worse stuff like IBM mainframes with EBCDIC and so on; now you can use OpenCOBOL for common business oriented programming instead.
04:03:47 <madbr> it's the kind of architecture that has decimal floating point too, right? :D
04:03:51 <pikhq_> The small one we have at work is more like $80k total.
04:04:06 <zzo38> (Of course you needn't even use COBOL if you want to use other programming languages instead.)
04:04:14 <pikhq_> (we're not a mainframe shop, but unfortunately we have clients that are, so we need to have one to test against)
04:04:54 -!- bb010g has joined.
04:07:00 -!- shikhin has quit (Ping timeout: 256 seconds).
04:07:32 <pikhq_> It is probably a good thing I find really weird computers entertaining.
04:07:54 <ais523> pikhq_: well, when C-INTERCAL was first announced, it was sent to a community for people entertained by old and weird computers
04:08:00 <ais523> because there weren't really esolang communities at the time
04:08:22 <pikhq_> It is also a *really* good thing that the mainframe stuff is kinda just a portion of my job rather than the whole thing.
04:08:34 <pikhq_> I imagine I'd go mad if I did that all day every day.
04:09:15 <madbr> dude, I had a java project once, and I was crawling up the curtains
04:09:53 <pikhq_> (for future reference: parsing COBOL with yacc is irritating, cause IBM doesn't believe in BNF)
04:10:25 <pikhq_> IBM believes in money.
04:10:34 <pikhq_> And the acquisition of wealth.
04:11:12 <madbr> ibm is the original out of touch technocratic company
04:11:19 <zzo38> Not that there is anything wrong with money, but, *love of* money is the root of all evil.......
04:11:34 <madbr> they're like microsoft squared in that respect
04:12:04 <zzo38> Parsing COBOL with yacc is irritating? Can you describe how exactly?
04:12:11 <elliott> @google ibm company songbook
04:12:12 <lambdabot> http://arstechnica.com/business/2014/08/tripping-through-ibms-astonishingly-insane-1937-corporate-songbook/
04:12:12 <lambdabot> Title: Tripping through IBM’s astonishingly insane 1937 corporate songbook | Ars T...
04:12:15 <J_Arcane> Even in the beginning, it was basically IBM, and everyone else. IBM does things the IBM way.
04:12:35 <pikhq_> zzo38: Well, see, the syntax is just *not very well specified* and only *barely* context-free.
04:13:02 <pikhq_> Also, tokenization is god damned insane.
04:13:25 <zzo38> pikhq_: It isn't specified? What? Can't you look at how OpenCOBOL does it?
04:13:44 <pikhq_> There's little details like "the first X and last Y columns are comments"...
04:14:00 <J_Arcane> pikhq_: Sounds like me trying to make sense of BASIC, *10.
04:14:16 <zzo38> Those things can be done in the tokenizer before you tell the parser what to do
04:14:33 <pikhq_> And "if there's character X in this column, you concatenate it with the previous line except when you're in a string literal, in which case you skip to the next instance of '"' and then concatenate the rest of the line".
04:14:42 <pikhq_> zzo38: Unfortunately not really.
04:14:56 <pikhq_> IBM COBOL is heavy on the non-standard extensions as well.
04:15:11 <zzo38> pikhq_: Maybe in yacc it is difficult, but did you try using Lemon? It is an alternative to yacc
04:15:14 <pikhq_> Which is all the more amusing because standard COBOL died.
04:15:20 <zzo38> And in my opinion, Lemon is much better
04:15:22 <pikhq_> I didn't know about Lemon.
04:15:30 <pikhq_> And I don't care now, because It Is Done.
04:16:13 <zzo38> Ah, OK, but maybe you can look at anyways in case it help later on for something.
04:16:40 <zzo38> Lemon uses no global variables so you can run multiple parsers at once, and the tokenizer calls the parser instead of the parser calling the tokenizer.
04:16:47 <zzo38> (Lemon is also in public domain.)
04:16:49 <elliott> pikhq_: how does that string literal thing work?
04:17:22 <pikhq_> elliott: You basically specify that a line is a continuation of the previous line, right?
04:17:55 <elliott> pikhq_: right, but where does the " come into it?
04:18:26 <pikhq_> elliott: It looks like this:
04:18:48 <pikhq_> abcde 01 LINE IS "This is a line. It is being continued.
04:19:01 <pikhq_> edfgh- "This is a continuation".
04:19:15 <pikhq_> LINE has "This is a line. It is being continued.This is a continuation"
04:19:23 <elliott> what about edfgh- qqqqq"test"
04:19:44 <elliott> imagine more spaces if that matters
04:19:53 <pikhq_> The contents between the continuation marker and the quote delimiter have to be whitespace.
04:20:33 <pikhq_> Newline will terminate the line, so that won't work.
04:21:02 <pikhq_> As will vertical tab I think?
04:21:07 <elliott> does it scan for the " just to let you indent continued strings?
04:22:10 <pikhq_> And the continuation line madness is really necessary. All text outside of columns 8 to 72 are comments.
04:22:18 <elliott> ...does LINE IS "test""foobar" even work?
04:22:28 <elliott> I guess it'd be "test"foobar" lol
04:24:27 <pikhq_> It's pretty impressive in its own way, with how it manages to be so utterly unlike anything you're used to.
04:29:28 -!- dts|feasting has changed nick to dts.
04:35:01 <Sgeo> I bought a book on COBOL because I confused it with CORBA
04:36:32 <madbr> kindof wondering what would be a good language for creating chiptune instruments
04:37:15 <zzo38> Depends what kind of chiptune instruments, I think
04:37:37 <madbr> also probably has to be as graphical as possible
04:37:59 <madbr> since sound designers rarely want to deal with text
04:38:26 <madbr> and text is a problem in most sound design/music production software anyways (DAWs)
04:38:33 <zzo38> Well, I really don't know; I know stuff such as Pure Data exists
04:38:50 <madbr> (the host program will capture your keyboard events, which is bad)
04:39:14 <madbr> and the typical user wants to be presented choices, not guess which choices are valid
04:43:39 -!- cluid has joined.
04:48:28 <FreeFull> Sgeo: I wonder if anyone ever bought a book on Pascal because they confused it with Haskell
04:51:10 <cluid> https://web.archive.org/web/20041103230415/http://geeksden.sourceforge.net/geekywiki/REGXY
04:51:16 <cluid> May I copy this verbatim onto esowiki?
04:51:20 <cluid> as "original specification"
04:52:30 <elliott> only public domain stuff can go on the wiki :/
04:52:33 <cluid> also I want to write a modern implementation, what langauge should I use?
04:53:55 <zzo38> cluid: I don't know? Possibly with Perl? It uses regular expressions.
04:54:13 <cluid> ill look into perl
04:54:29 <zzo38> I think Perl is a bit confusing though
04:54:34 -!- Oren has joined.
04:56:09 <Oren> no matter how many times I hear the word 'statistics' it still sounds like 'sadistic'
04:56:48 <int-e> sadistically speaking, 1 in seven people are afflicted by this condition?
05:03:38 <madbr> Oren : thos belies the fact that english is full of "T"s, "S"s, "D"s and "C"s :D
05:03:51 <madbr> especially t and s
05:05:21 <Oren> TS... no! bad oren's brain! do not make that stand for anything!
05:06:33 <cluid> I can translate regexy into perl easily
05:06:54 <cluid> http://lpaste.net/115247
05:07:03 <cluid> can wre write a regexy program to do this?
05:08:42 <cluid> a slight problem is that labels in regexy may not be valid labels in perl
05:09:12 <cluid> oh.. and lavbels in regex are comuputed from the regex.. so this translation isn't complete
05:10:31 <oerjan> cluid: btw i put the updated link to zzo38's REGXY page on the esowiki
05:10:40 <cluid> thanks a lot oerjan , i was going to do that
05:12:19 <cluid> http://zzo38computer.org/regxy/adbinery.txt
05:12:25 <cluid> actualy does this use computed labels or not? I think it does
05:12:34 <cluid> but i was mistaken
05:13:43 -!- tlewkow has quit (Remote host closed the connection).
05:13:51 <oerjan> Oren: tzatzikilly, this is a big problem hth
05:13:54 -!- callforjudgement has joined.
05:13:56 -!- ais523 has quit (Remote host closed the connection).
05:14:03 -!- callforjudgement has changed nick to ais523.
05:14:32 <zzo38> It doesn't looks computed label to me?
05:14:36 -!- Dulnes has quit (Quit: Connection closed for inactivity).
05:14:41 <cluid> thanks, I misread the code
05:14:49 <cluid> so RegXy doesn't support computed labels?
05:15:05 <cluid> it woudl be nice if it didnt
05:16:34 <cluid> I got an infinite loop trying to convert adbinery to perl
05:17:20 <cluid> http://lpaste.net/115248
05:18:04 <zzo38> Looks like not supporting computed labels from how I can see?
05:18:18 <cluid> i will have to use /s modifer in perl
05:18:24 <cluid> to treat multi line strings as a single string
05:18:31 <zzo38> I found a Visual Basic code and looks like not accepting computed labels.
05:18:32 <cluid> so that I can implement a regxy -> perl translation
05:19:59 -!- Dulnes has joined.
05:20:14 <Dulnes> What an awful way to end thanksgiving ;-;
05:20:25 <cluid> happy thanksgiving
05:20:48 <Dulnes> Merry Easter to you too
05:22:19 <Oren> whose birthday is it?
05:22:26 <Dulnes> Anyways my wife is pregnant and i hate children this is awful
05:24:19 <elliott> I'm not sure I believe that you have a wife
05:25:33 <cluid> why do you hate children
05:25:51 <Dulnes> They are loud sirens of pure malice
05:26:00 <elliott> don't you already have children
05:26:02 <cluid> isn't that a good thing?
05:26:12 <madbr> you were one at one time
05:26:13 <Dulnes> I dont want another one
05:26:27 <Dulnes> And my parents hated me
05:26:49 <Oren> So try to do better than them
05:27:11 <Dulnes> I have 3 kids i think thats enough for life
05:27:44 <ais523> is this the time when I give Dulnes a warning for being offtopic?
05:27:57 <ais523> offtopicness warnings in #esoteric is hard, because it's offtopic so often
05:28:12 <ais523> and yet some sorts of offtopicness are more obnoxious than others and it's hard to explain why
05:28:14 -!- oerjan has set topic: Child support channel | https://dl.dropboxusercontent.com/u/2023808/wisdom.pdf http://codu.org/logs/_esoteric/ http://tunes.org/~nef/logs/esoteric/.
05:28:27 <oerjan> there, no longer off topic hth
05:28:32 <cluid> I will finish the RegXy
05:28:40 <cluid> then we can be on topic
05:28:50 <Oren> i have 23 cousins
05:28:57 -!- ais523 has set topic: To kill a zombie, you must kill its parents | https://dl.dropboxusercontent.com/u/2023808/wisdom.pdf http://codu.org/logs/_esoteric/ http://tunes.org/~nef/logs/esoteric/.
05:29:04 <ais523> there, now it's on /both/ topics
05:29:22 -!- GeekDude has quit (Quit: {{{}}{{{}}{{}}}{{}}} (www.adiirc.com)).
05:29:22 <Dulnes> Whats that supposed to mean ais523
05:29:51 <Oren> i thought that rule was for vampires
05:30:04 <Oren> you kill dracula his minions also die
05:31:01 -!- Dulnes has set topic: There aren't any topics| https://dl.dropboxusercontent.com/u/2023808/wisdom.pdf http://codu.org/logs/_esoteric/ http://tunes.org/~nef/logs/esoteric/.
05:31:05 <elliott> btw if you didn't want another kid and you already have three, did you consider birth control
05:31:20 <Dulnes> Well idk maybe it didnt work
05:31:29 <ais523> Dulnes: you put a typo in the topic :-(
05:31:35 <Dulnes> We were using birth control
05:31:50 -!- ais523 has set topic: The international hub for esoteric programming language discussion, development and deployment | https://dl.dropboxusercontent.com/u/2023808/wisdom.pdf http://codu.org/logs/_esoteric/ http://tunes.org/~nef/logs/esoteric/.
05:31:55 <ais523> and you know what this means
05:32:01 <Oren> or you could try men instead that would prevent babies
05:32:22 <madbr> or have a sex change
05:32:43 <Dulnes> These are terrible ideas
05:32:54 <Dulnes> As they would break apart my family
05:32:56 <Oren> buy your wife a strapon
05:32:59 <elliott> let's not, re: all of this
05:33:18 <Oren> Ok i'll stop...
05:33:30 <madbr> are you religious?
05:34:17 <Dulnes> Sex changes cost money/ cheating on wife also causes divorce/money loss
05:34:38 <madbr> yeah I was just making a bad joke there
05:35:15 <Dulnes> Mmm whatever im going to drown myself in pie.
05:35:49 -!- ChanServ has set channel mode: +o ais523.
05:35:51 -!- ais523 has set channel mode: +m.
05:36:01 -!- ChanServ has set channel mode: +o elliott.
05:36:05 <ais523> in most channels, I have to get annoyed and forcibly change topic when something like this happens
05:36:11 <ais523> by coming up with an interesting tangents
05:36:14 <ais523> I'm all out of interesting tangents
05:36:18 <ais523> but I've realised I can do this
05:36:19 -!- elliott has set channel mode: -o ais523.
05:36:22 <elliott> don't worry the topic is now fish
05:36:23 -!- elliott has set channel mode: -m.
05:36:26 -!- elliott has set channel mode: -o elliott.
05:36:42 <elliott> -o'ing other ops is my favourite rogue op move
05:37:26 <cluid> I thought I could match anything that isn't a backslash
05:37:36 <cluid> but I actually need to include back slash escaped backslahes
05:37:37 <Dulnes> Maybe she was mistaken but whatever gnight
05:37:53 <elliott> ais523: https://en.wikipedia.org/wiki/Kickflip
05:37:55 <oerjan> cluid: you might take a look at my /// interpreter it has similar issues
05:38:02 <cluid> I could use this (([^\/]|\\\/)*)
05:38:04 <ais523> elliott: I mean in the context of IRC
05:38:08 <Dulnes> \/ thought that was a V
05:38:15 <cluid> oerjan, I invented a language similar to /// but less cool
05:38:23 <cluid> your /// stuff is very awesome
05:38:39 <ais523> I have a regex for matching slash-delimited backslash-escaping stuff in aiake
05:39:06 <madbr> still haven't figured out how I'll come up with my next horrible language
05:39:08 <ais523> push @aipath, $1 while $objvalue =~ m=((?:[^\\\/:]|\\.)*)( / | $ )=gsx;
05:39:24 <ais523> now I'm trying to figure out specifically what that regex matches
05:39:24 <cluid> backslashed backslashes
05:39:28 <cluid> I forgot about those
05:39:34 <cluid> (([^\/]|\\\/|\\\\)*)
05:39:36 <ais523> and looking at it, seems I missed a ?:
05:39:43 <cluid> this could be enough for me maybe I hope
05:39:46 <ais523> which will have efficiency effects
05:39:48 <madbr> is the mandelbrot set turing complete?
05:39:50 <ais523> cluid: best to do it correctly
05:39:54 <cluid> I have to use two loops in my code
05:39:59 <cluid> actuall y many, in two groups
05:40:05 <cluid> one for santizing labels into perl valid labels
05:40:13 <cluid> another for transforming RegXy rules into perl rules
05:40:20 <ais523> besides, \\. is easier than \\\/|\\\\ and more correct, at least in the case of ///
05:40:35 <cluid> ais523, oh nice idea ! thank you
05:40:51 <madbr> bicyclidine: yes but there's no upper bound on the number of iterations required to find if a number is in the set
05:41:13 <cluid> I wonder how regex $n works
05:41:14 <madbr> that seems a lot like the halting problem to me
05:41:16 <cluid> i will have to count
05:41:55 <Bicyclidine> well, finding the limit set of some mappings is turing equivalent.
05:42:07 <Bicyclidine> don't think that's the case for mandelbrot though. it's just squaring.
05:42:18 <ais523> incidentally, something else I'm annoyed at (and vaguely related to the LCRNG maths question I asked in here a few days ago)
05:42:33 <ais523> is when you have this nice elegant solution to something that isn't quite your problem
05:42:35 <madbr> it might be possible to build an accumulator out of the real and imaginary part
05:43:08 <cluid> $i =~ s/(^|\n)([^\/]*)\+(([^\/]|\\.)*)\/(([^\/]|\\.)*)\/(\n|$)/$1$2_plus$3$5$7/s;
05:43:33 <ais523> something that isn't quite my problem: you have sets of boolean variables (some variables appear in multiple sets); you want to find a mapping of variables to true/false so that there's an even number of true variables in each set
05:43:51 <newsham> howd the lcrng thing work out in the end?
05:43:51 <ais523> I found a nice algorithm to solve this in O(number of variables^2)
05:44:24 <ais523> newsham: I didn't solve it, even though it seems solvable, due to running out of ideas; however, I decided that I possibly had a "how can I use X to do Y?" problem
05:44:31 <Oren> brute force is 2^n
05:44:35 <ais523> and are looking at other potential ways to do Y, without doing X, even though X would have been cool
05:44:45 <ais523> also I wrote a brute forcer that works for n=32
05:44:53 <myname> ais523: easy: don't set any to true. O(1)
05:44:57 <ais523> n=64 would only take a few times longer than my expected lifetime
05:45:12 <ais523> myname: there's one specific variable that has to be set to true
05:45:52 <newsham> n=64 prob solvable on cloud clusters in your lifetime, or even this year
05:46:09 <ais523> right, that was on my laptop
05:46:11 <oerjan> ais523: it seems like you've reinvented XORSAT?
05:46:21 <ais523> oerjan: right, there was almost certainly a name for it
05:46:53 <cluid> can I write this with less brackets
05:47:13 <ais523> and, from a different point of view, reinvented linear programming
05:47:23 <Oren> yes the outer breackets are unecessay
05:47:31 -!- callforjudgement has joined.
05:47:35 <oerjan> ais523: XORSAT is much simpler than linear programming
05:47:56 <oerjan> it's just solving a matrix equation (mod 2)
05:48:00 <callforjudgement> sadly, my /actual/ problem expressed in these terms are: you have a number of integer variables (mod n), and a number of linear polynomials of those variables
05:48:17 <callforjudgement> but the polynomials are inequalities rather than direct equalities
05:48:39 <oerjan> linear integer programming is NP-complete
05:48:57 <callforjudgement> oerjan: I was thinking of the subset of linear programming that's just solving matrix equations
05:49:28 <oerjan> well yes. but it doesn't work with inequalities.
05:49:46 <callforjudgement> nothing I can think of works with inequalities and it is annoying me
05:50:00 <madbr> the mandelbrot set is connected
05:50:08 <oerjan> oh hm you have (mod n). but i think linear integer programming is NP-complete even with just {0,1}
05:50:14 <madbr> I wonder if this means that it cannot be turing complete
05:50:24 <madbr> cluid: the wikipedia article says so
05:50:31 <oerjan> madbr: almost certainly those facts are unconnected
05:50:36 <cluid> its a joke beacuse it's hard to prove that
05:51:16 <madbr> oerjan : so in otherwose, almost certainly no? :D
05:51:40 -!- ais523 has quit (Ping timeout: 244 seconds).
05:51:42 -!- callforjudgement has changed nick to ais523.
05:52:33 <Oren> integer programming is NP-complete -- lecture today was on that fact
05:53:27 <Oren> in fact it is NP-hard
05:53:43 <oerjan> ais523: btw aaronson and friends think of XORSAT as "that problem which is so similar to SAT that many purported proofs of P != NP can be discarded by checking that they would prove XORSAT not in P because they fail to distinguish any real property of them"
05:53:50 <Bicyclidine> i uh, don't think it could be np-complete without being np-hard
05:53:56 <oerjan> (not a quote but a paraphrase)
05:54:02 <ais523> oerjan: oh, I think of 2SAT like that
05:54:18 <Oren> oh right. i'm still learning this stuff...
05:54:23 <ais523> xorsat might be better, though, I don't have much experience in finding fallacies in P≠NP proofs
05:54:39 <ais523> what's more common, incidentally, claims of P=NP proofs or claims of P≠NP proofs?
05:54:46 <Oren> how about a randomized algoritthm then?
05:54:54 <oerjan> ais523: that purported proof by deolalikar (sp?) a few years ago failed on that test
05:55:11 <ais523> I assume the latter, because unless a P=NP proof is existence rather than constructive, it's normally possible just to run it on a computer and see what happens
05:55:11 <oerjan> ais523: i don't remember
05:55:32 <oerjan> ais523: it might depend on the level of crankiness involved
05:56:15 <ais523> Bicyclidine: my favourite NP definition is "an problem is NP if it can be solved by a trustworthy P-time checker and untrustworthy TC oracle working together, such that the solution is always correct if the oracle happens to be trustworthy, and has no false positives regardless of what the oracle does"
05:56:36 <ais523> Bicyclidine: you need a lot of words, sadly
05:56:49 <cluid> How do I get the whole of STDIN as a string in perl?
05:57:21 <ais523> helping a student with this, we came up with the (possibly slightly mathematically unsound) "a problem is NP if a P-time algorithm with access to randomness has no false positives, and no false negatives with probability 1"
05:57:37 <ais523> cluid: {local $/; $string = <>}
05:58:00 <ais523> cluid: although that will read a file specified on the command line if there is one, in preference to stdin
05:58:05 <ais523> this is normally what you want but not always
05:58:10 <cluid> thats really nice thank you !
05:58:24 -!- ZombieAlive has quit (Remote host closed the connection).
05:59:35 <oerjan> madbr: i'm pretty sure you can easily construct a set that is connected and such that it's TC to check membership, yes: pick your favorite TC set of integers and connect with arcs or something.
06:02:21 <oerjan> <Oren> how about a randomized algoritthm then? <-- the current "most believed hypothesis" by experts is that (RNG) randomness doesn't give you anything beyond P
06:03:22 <Oren> does it at least make the constant teenier?
06:03:31 <ais523> oerjan: at least without some method of comparing the possible random results
06:03:43 <ais523> hmm, there are P-time PRNGs that are pretty good
06:03:58 <oerjan> <cluid> How do I get the whole of STDIN as a string in perl? <-- hey i _told_ you to look at my /// interpreter!
06:04:09 <ais523> it'd be something quite spectacular if there was an algo that worked with true-randomness, but not a CSPRNG
06:06:10 <oerjan> ais523: the hypothesis is based on the the theory that there are P-time PRNGs so good that they cannot be revealed as pseudo in P-time
06:06:42 <Oren> maybe true randomness doesn't exist?
06:06:43 <oerjan> although no one has proved they exist, as it's something stronger than P != NP
06:07:04 <ais523> right, P=NP would mean that such PRNGs could definitely be proven pseudo in P-time
06:07:14 <oerjan> Oren: quite possible but the mathematical concept of a true random algorithm still does.
06:10:06 <oerjan> iirc it follows from the also hypothesized existence of cryptographic hashes
06:11:48 <cluid> http://lpaste.net/115258
06:12:07 <cluid> do you think adbinery is wrong or my translation?
06:12:14 <cluid> the output seems wrong
06:12:56 <zzo38> Either could be wrong (possibly both)
06:14:12 -!- tlewkow has joined.
06:14:16 -!- madbr has quit (Quit: Pics or it didn't happen).
06:14:41 -!- oerjan has quit (Quit: Some other time).
06:16:49 <fizzie> cluid: One thing that at least is wrong is that in Perl's regexps, you use plain + instead of (sed, POSIX re) \+ for "one or more".
06:17:36 <fizzie> Though I guess that applies to whatever the upper block is too, now that I look at it.
06:17:41 <fizzie> (Sorry, just woke up.)
06:17:53 <Oren> he's trying to match + the operator plus
06:18:25 <cluid> I havea bug in my translation
06:19:09 -!- tlewkow has quit (Ping timeout: 258 seconds).
06:22:35 -!- AndoDaan has quit (Quit: Quit).
06:29:38 <cluid> I completed a self hosting regxy -> pl translator
06:29:49 <cluid> should I add this an ddetails to the wiki page?
06:42:48 <ais523> cluid: most commonly you'd post it somewhere else and add a link
06:43:22 <ais523> you could post it on the wiki if you had no better options, but normally there are better options; I know graue (who used to own esolangs.org) has volunteered to host almost arbitrary esolang compilers/interpreters/stuff on esoteric.voxelperfect.net/files
06:43:33 <ais523> but I'm not sure anyone's asked him to do that in ages
06:44:26 <cluid> i could put it on a paste bin but it might disappear
06:46:29 <ais523> some things do get posted to the wiki, and I'm not personally against that, but I know some other people are
06:47:51 <cluid> I wanted to make the RegXY article better
06:48:20 <elliott> some people use a user subpage and link it from mainspace
06:49:22 <ais523> I think everyone agrees that the impl should be posted somehow
06:49:25 <ais523> just there's debate as to how
06:49:37 <zzo38> I think if it isn't too long you can post directly on the wiki
06:50:11 <zzo38> Note that the wiki page normally isn't a valid program; however you can avoid this by using Perl's documentations features I suppose
06:50:49 <zzo38> As long as it is in the public domain too, it is OK then
06:51:19 <ais523> if it's only 10 lines, posting it directly's probably fine
06:51:24 <ais523> make sure to escape it properly, though
06:51:37 <cluid> is there a program to escape it?
06:51:39 <ais523> (wrapping in <pre><nowiki> to start and </nowiki></pre> works for most programs)
06:53:37 <zzo38> ais523: That works for display, but not for download unless you add # or Perl documentation commands or whatever.
06:54:01 <ais523> zzo38: it's not bad for download, you can copy-paste either from the displayed version or the wikitext
06:54:15 <zzo38> For example http://esolangs.org/wiki/Pure_BF/Implementation?action=raw&ctype=text/css is a valid Literate Haskell program.
06:54:27 <cluid> Can I put the reverse example code?
06:54:47 <cluid> from here https://web.archive.org/web/20041103230415/http://geeksden.sourceforge.net/geekywiki/REGXY
06:55:20 <zzo38> I don't know the license from that archived wiki, although the stuff in my directory is in public domain so you can use it
06:55:41 <zzo38> If the code is small enough it would probably be OK though
06:56:00 <ais523> cluid: the wiki is public domain things only, which is about the most restrictive thing possible in terms of what you can post to it
06:56:54 <zzo38> I don't know enough about Perl programming to know the best way to do a similar thing with Perl than what I have done with Haskell there.
06:58:42 <HackEgo> [wiki] [[REGXY]] http://esolangs.org/w/index.php?diff=41223&oldid=41218 * Cluid Zhasulelm * (+1917) Improved language definition and added examples along with implementation
07:01:32 -!- drdanmaku has quit (Quit: Connection closed for inactivity).
07:01:55 <HackEgo> [wiki] [[User:Cluid Zhasulelm]] M http://esolangs.org/w/index.php?diff=41224&oldid=41155 * Cluid Zhasulelm * (+75)
07:04:03 <HackEgo> [wiki] [[Talk:Nonsense Query List]] http://esolangs.org/w/index.php?diff=41225&oldid=31459 * Zzo38 * (+111)
07:08:55 <HackEgo> [wiki] [[REGXY]] M http://esolangs.org/w/index.php?diff=41226&oldid=41223 * Cluid Zhasulelm * (+467) 99 bottles of beer example
07:12:10 <HackEgo> [wiki] [[User:Cluid Zhasulelm]] M http://esolangs.org/w/index.php?diff=41227&oldid=41224 * Cluid Zhasulelm * (+227) wikis
07:24:00 -!- tlewkow has joined.
07:24:35 -!- Bicyclidine has quit (Ping timeout: 272 seconds).
07:28:55 -!- tlewkow has quit (Ping timeout: 258 seconds).
07:44:07 <cluid> what was the japanese blogger saying about MMMNBF?
07:48:30 <Oren> um, send me the page in question and I can try to translate.
07:57:07 -!- Patashu has joined.
07:57:25 <cluid> i dont know what page[s] it is
07:57:27 <Oren> if you mean copyright or whatever, the blog doesn't seem to have that info
07:57:52 <Oren> what information are you looking for?
07:57:54 <cluid> I was just curiouw what their views were
07:58:02 <cluid> if they had questions about it or who they were
07:58:34 <Oren> The blooger is the language designer. afaik.
07:59:09 <cluid> ah, i didnt know that
08:00:10 <Oren> I'll translate the whole page of the main post
08:00:18 <Oren> perhaps make it more clear
08:00:35 <Oren> title : まだ名前のないBrainfuck風スタック言語、略して MNNBFSL
08:01:17 <Oren> a not yet named brainfuck style stack language, for short mnnbfsl
08:01:42 <Oren> I have been thinking about it
08:01:57 <Oren> インタープリタの実装は yshl/MNNBFSL ? GitHub に。
08:02:41 <Oren> "an implementation of an interpreteris here, github"
08:02:57 <Oren> データスタックとリターンスタックの 2本のスタックがあります。
08:03:21 <Oren> "there are two stacks, a data stack and a return stack"
08:03:35 <lifthrasiir> the third-to-last paragraph: "It feels quite similar to Brainfuck, but [ and ] has been repurposed for program counter manipulation instead of jump. Forward jump can be done by getting PC via [, but backward jump would require getting PC and some adjustments to it and is slightly harder."
08:03:56 -!- callforjudgement has joined.
08:04:02 <Oren> スタック 2本あればチューリング完全らしいから
08:04:19 -!- ais523 has quit (Remote host closed the connection).
08:04:33 <Oren> Because if there are two stacks it is turing complete ish,
08:04:38 <lifthrasiir> the second-to-last and last paragraphs: "It keeps PC so I guess it is harder to optimize than Brainfuck when it comes to the transpiler to C." "At least it is better than Brainfuck that the behavior on EOF is defined..."
08:04:44 -!- callforjudgement has changed nick to ais523.
08:04:56 <Oren> チューリング完全なんじゃないかな。
08:05:16 <Oren> "I wonder if it is not turing complete"
08:09:09 <ais523> hmm, so despite the effort we're putting into translating this
08:09:13 <ais523> it is nonetheless just a BF derivative?
08:09:38 <cluid> its an intersting BF derivative
08:10:21 <Oren> looks like it is a bit different from BF
08:10:28 -!- qwertyo has joined.
08:10:34 <Oren> the problem is jusggleing the stacks
08:10:57 <Oren> the return address is not always there when you need it
08:11:36 <Oren> in any case from my reading, we have the spec right for all the instructions
08:12:50 -!- callforjudgement has joined.
08:14:06 <HackEgo> [wiki] [[MNNBFSL]] M http://esolangs.org/w/index.php?diff=41228&oldid=41216 * Orenwatson * (+1) count of instructions was wrong
08:15:07 -!- ais523 has quit (Ping timeout: 264 seconds).
08:16:10 <cluid> actually only superficially related to brainfuck
08:16:21 <cluid> its based on two stacsk not a tape
08:16:32 <cluid> and you have to compute jumps, not matching brackets
08:16:53 -!- callforjudgement has quit (Ping timeout: 240 seconds).
08:17:36 <b_jonas> cluid: um, two stacks is a tape.
08:18:27 <Oren> the blogger's profile says: "I am a biological human, unrelated to the Yokohama Shootout Hockey League"
08:19:12 <Oren> so his name is just: yshl
08:20:49 <Oren> he has bought a lot of manga, most of hisposts are listing the manga he bought that week.
08:24:07 <HackEgo> [wiki] [[MNNBFSL]] M http://esolangs.org/w/index.php?diff=41229&oldid=41228 * Orenwatson * (+84) added apparent author info
08:26:51 <fizzie> b_jonas: "Two stacks and a tape walk into a bar..."
08:27:09 <fizzie> (I don't know, it just sounded like a setup for a joke; no idea how it'd continue.)
08:30:25 <cluid> i can't translate brainfuck into MMBFSL
08:30:33 <cluid> a tape in bf is right infinite
08:30:41 <cluid> but the stacks start empty in MMBFSL
08:32:38 -!- tlewkow has joined.
08:39:08 <fizzie> cluid: Possibly you can use a tape layout like "1 a 1 b 1 c 0" for data "a b c", and make the analogue of > to check the 0/1 flag and extend the tape.
08:39:43 <fizzie> (Could be trickier than that in a sufficiently tarpitty language.)
08:39:54 <fizzie> (I haven't been following the discussion.)
08:40:04 <cluid> that's a good idea
08:40:32 <fizzie> elliott: I evaded the bait by cleverly being asleep, but OTOH I don't have an answer either, I probably would just have mumbled something incoherent about the branch predictor's hardware return stack.
08:41:35 -!- dts has quit (Read error: Connection reset by peer).
08:43:14 -!- dts has joined.
08:56:43 <Oren> I sent the blogger a (no doubt poorly written) email asking for more information.
08:57:02 <dts> i dont believe you
08:57:18 -!- tlewkow has quit (Remote host closed the connection).
08:57:56 <Oren> why not? I really did send him a n email. it's 6pm now so
08:58:05 <Oren> (in japan that is)
08:58:15 <dts> idk man. you just seem really shifty
08:58:18 <Oren> so he should be home from work soon
08:58:21 <cluid> cool Oren , kep me updated when you get a reply please?
08:58:24 <cluid> fizzie, can you ban dts?
09:05:58 -!- Patashu_ has joined.
09:05:58 -!- Patashu has quit (Disconnected by services).
09:24:36 -!- Dulnes has quit (Quit: Connection closed for inactivity).
09:27:30 <cluid> http://esolangs.org/wiki/Binary_combinatory_logic
09:27:36 <cluid> is this really string rewriting
09:27:52 <cluid> its term rewrting not string rewriting
09:27:58 <cluid> 11101xyz --> 11xz1yz
09:28:02 <cluid> its not characters x,y,z but terms
09:30:53 -!- qwertyo has quit (Quit: Leaving).
09:31:08 <Oren> mhm that's right
09:40:10 -!- ais523 has joined.
09:40:26 <cluid> http://esolangs.org/wiki/Infinity this page is really annoying
09:41:54 -!- sebbu has quit (Ping timeout: 264 seconds).
09:46:21 <cluid> http://esolangs.org/wiki/Cheese
09:47:06 <int-e> Does it really say "Infinity closely approximates the amount of nonsense on this page." or am I imagining things?
09:47:31 <dts> cheese++ is far better tbh
09:47:53 <cluid> dts, you should please design cheese+++ (and implement it in cheese++)
09:48:16 <dts> what would cheese+++ have in it?
09:49:02 <int-e> Oh I see, the BCL page is not even talking about string rewriting, it's just a category.
09:49:31 <cluid> it could use ideas from here http://en.wikipedia.org/wiki/List_of_cheeses
09:52:51 <int-e> it can be viewed as a sort of conditional string rewriting. "Kxy -> x if x and y are produced by the following CFG...", but I have yet to see such a formalism.
09:53:22 <int-e> (CFGs are, of course, string rewrite systems)
09:54:22 <cluid> Context-free grammar?
09:55:16 <int-e> consider the rules S -> 00; S -> 01; S -> 1SS, and look at which strings not containing S can be produced from S.
09:55:29 <cluid> that CFG describes the syntax well
09:55:43 <cluid> but the language is defined by term rewriting
09:56:06 <cluid> so it should be moved out of string rewrite category
09:56:43 <int-e> or perhaps not? is "term" a term or is it the set of strings defined under "syntax"?
09:57:57 <int-e> You can view it either way. Which means I don't feel strongly about the category.
09:58:40 -!- sebbu has joined.
09:58:53 <int-e> And since oerjan added the category I would leave the decision to him.
09:59:18 -!- sebbu has quit (Changing host).
09:59:18 -!- sebbu has joined.
10:00:27 <cluid> it can be a string rewriting system only if it has infinitely many rules
10:01:13 <cluid> so 1100xy --> x denotes all rules 11000000 --> 00, 11000001 --> 00, ...
10:01:32 <cluid> if you treat it as aterm rewrite system it is just 2 rules
10:01:51 <cluid> also it talks about "subterms of a given term" so it should be a term rewrite system
10:06:25 <int-e> cluid: I would normally agree, but there are notions of string rewriting where the rules may contain variables that represent arbtitrary substrings. (So instead of a term rewriting with unary function symbols, you'd have to do term rewriting with constants and one AC [associative, commutative] symbol.)
10:06:50 <cluid> ok i was just looking at the thue system thing on wikipedia
10:07:13 <cluid> on what you're saying it makes sense to call it string rewriting
10:07:19 <int-e> And in that contaxt 1100xy --> x fits perfectly, except that it would allow steps like 11001010 -> <empty>
10:08:51 <int-e> And the category is not overflowing with pages, so a broad definition of "string rewriting" seems appropriate.
10:09:02 <cluid> yeah, thanks for your input!
10:18:37 -!- nooga has joined.
10:29:18 -!- ais523 has quit (Ping timeout: 264 seconds).
10:32:45 <cluid> so is it easy to recognize CFGs inside a string?
10:33:04 <cluid> maybe this is similar to regex
10:35:37 <cluid> idea for esolang: Write a CFG, then write CFG substitution rules to define a program
10:35:49 <cluid> then you can implement BCL in it for example, quite easily
10:53:30 -!- ais523 has joined.
10:54:59 -!- Phantom_Hoover has joined.
11:03:24 -!- dts has changed nick to dts|Zzzz.
11:14:25 -!- callforjudgement has joined.
11:14:47 -!- ais523 has quit (Disconnected by services).
11:14:49 -!- callforjudgement has changed nick to ais523.
11:18:22 -!- boily has joined.
11:18:29 -!- dts|Zzzz has quit (Ping timeout: 240 seconds).
11:20:12 -!- HackEgo has quit (Ping timeout: 272 seconds).
11:39:12 <cluid> hm the esolang wiki is down
11:39:20 <cluid> but I was reading the BCL page and this is interesting,
11:40:03 <boily> Gregor: AAAAAAAAAAAAAAAAAAAH! http://www.downforeveryoneorjustme.com/esolangs.org
11:40:05 <cluid> you can define it in two steps, first syntax: S --> 00 | 01 | 1 S S
11:41:00 <fizzie> boily: I'm provisionally going to suggest it's probably that shifty VPS provider at work again.
11:41:02 <cluid> and then rewrite laws: 100 S:x S:y --> x, 101 S:x S:y S:z --> 11 x z 1 y z
11:41:19 <fizzie> boily: (Doesn't answer to SSH, so I can't do much.)
11:41:21 <cluid> this implements BCL
11:41:33 <cluid> you could also implement other programs in this way
11:43:34 <boily> fizzie: I wouldn't be surprised at all. :/
11:43:49 <boily> cluid: this sounds an awful lot like an L-system...
11:44:00 <cluid> its much more powerful than L system
11:44:28 <cluid> you can match an arbitrary CFG in the string to provide a replacement, rather than just a character
11:44:48 <cluid> and your rules can shrink the string as well as grow it, so you can compute more things
11:47:09 <fizzie> I don't think the term "L-system" is technically restricted to having only context-free rules, even if that's the most common case.
11:47:43 <fizzie> http://en.wikipedia.org/wiki/L-system#Context_sensitive_grammars is counted as a "variation". Not that there are probably any "official" definitions.
11:50:39 <cluid> I should use different symbos for the CFG production, and the rewrite rules
11:50:47 <cluid> ==> maybe for rewrite
11:51:20 <cluid> it's cool that you can implement BCL in 2 lines, maybe there are other program you can write it in
11:58:14 -!- tlewkow has joined.
12:00:13 -!- Patashu_ has quit (Ping timeout: 240 seconds).
12:02:37 -!- tlewkow has quit (Ping timeout: 258 seconds).
12:06:42 -!- boily has quit (Quit: MANIFOLD CHICKEN).
12:20:24 -!- ais523 has quit.
12:20:58 -!- nooga has quit (Ping timeout: 255 seconds).
12:23:09 -!- Bicyclidine has joined.
12:28:15 -!- idris-bot has quit (Quit: Terminated).
12:28:54 -!- idris-bot has joined.
12:34:49 -!- sebbu2 has joined.
12:35:25 -!- sebbu2 has quit (Changing host).
12:35:26 -!- sebbu2 has joined.
12:37:12 -!- sebbu has quit (Ping timeout: 250 seconds).
12:44:21 -!- ZombieAlive has joined.
12:58:05 -!- GeekDude has joined.
14:13:01 -!- cluid has quit (Quit: Leaving).
14:29:34 <lambdabot> oerjan said 14h 55m 31s ago: <mroman> which is usually suffixed with (embed) <-- if you want people to follow conventions you should actually state them on the website hth
14:29:49 <mroman> @message oerjan It's not my website
14:29:50 <lambdabot> Maybe you meant: messages messages-loud messages?
14:30:04 <mroman> @bell oerjan It's not my website
14:31:26 <fizzie> So it's not your (singular) website, but arguably it's then your (plural) website.
14:32:23 <fizzie> (Unless it was the kind of passive-voice-style "we".)
14:34:11 <fizzie> Gregor: I guess we could nickping you to maybe have a poke and/or a prod at the management console (if they have one) of the esolangs VPS, to see if it says anything informative. (It's down w.r.t. HTTP, SSH at least at the moment.)
14:46:05 -!- Oren has quit (Ping timeout: 246 seconds).
14:51:36 -!- copumpkin has joined.
14:52:16 -!- contrapumpkin has joined.
14:56:13 -!- copumpkin has quit (Ping timeout: 255 seconds).
14:58:06 -!- contrapumpkin has changed nick to copumpkin.
14:58:52 -!- HackEgo has joined.
15:03:23 -!- contrapumpkin has joined.
15:06:39 -!- copumpkin has quit (Ping timeout: 265 seconds).
15:09:21 -!- Oren has joined.
15:15:11 -!- mig22_ has joined.
15:15:48 -!- mig22_ has changed nick to mig22.
15:18:17 -!- mig22 has quit (Remote host closed the connection).
15:19:13 -!- mig22 has joined.
15:19:20 -!- copumpkin has joined.
15:21:20 -!- drdanmaku has joined.
15:22:51 -!- contrapumpkin has quit (Ping timeout: 255 seconds).
15:34:58 -!- mig22 has quit (Quit: mig22).
15:36:11 -!- contrapumpkin has joined.
15:38:59 -!- copumpkin has quit (Ping timeout: 240 seconds).
15:42:12 -!- copumpkin has joined.
15:46:06 -!- contrapumpkin has quit (Ping timeout: 264 seconds).
15:48:01 -!- contrapumpkin has joined.
15:51:33 -!- propumpkin has joined.
15:51:39 -!- copumpkin has quit (Ping timeout: 255 seconds).
15:52:55 -!- GeekDude has quit (Quit: {{{}}{{{}}{{}}}{{}}} (www.adiirc.com)).
15:55:42 -!- copumpkin has joined.
15:55:42 -!- contrapumpkin has quit (Ping timeout: 272 seconds).
15:59:18 -!- propumpkin has quit (Ping timeout: 264 seconds).
16:03:47 -!- Oren has quit (Ping timeout: 246 seconds).
16:07:54 -!- Oren has joined.
16:08:01 -!- contrapumpkin has joined.
16:11:01 -!- propumpkin has joined.
16:11:27 -!- copumpkin has quit (Ping timeout: 255 seconds).
16:14:39 -!- contrapumpkin has quit (Ping timeout: 245 seconds).
16:18:12 -!- copumpkin has joined.
16:21:29 -!- propumpkin has quit (Ping timeout: 240 seconds).
16:26:11 -!- contrapumpkin has joined.
16:30:08 -!- copumpkin has quit (Ping timeout: 264 seconds).
16:38:11 -!- copumpkin has joined.
16:38:33 -!- blsqbot has quit (Ping timeout: 240 seconds).
16:39:34 -!- mroman has quit (Ping timeout: 244 seconds).
16:40:41 -!- propumpkin has joined.
16:42:08 -!- contrapumpkin has quit (Ping timeout: 264 seconds).
16:43:49 -!- copumpkin has quit (Ping timeout: 245 seconds).
16:43:52 -!- contrapumpkin has joined.
16:46:21 -!- copumpkin has joined.
16:46:59 -!- propumpkin has quit (Ping timeout: 240 seconds).
16:49:42 -!- contrapumpkin has quit (Ping timeout: 255 seconds).
16:54:48 <Oren> what kind of bot logs in and logs out every 3 minutes?
16:55:20 <Oren> what's going on then?
16:55:37 <Oren> bad connection?
16:59:53 <Oren> hmm, seems my scrip7 interpreter is still buggy
17:00:12 <Oren> the loops are not working
17:00:22 -!- contrapumpkin has joined.
17:01:38 <Oren> this program crashes my scrip7 interpreter: $ { # _ p 64 } .
17:01:52 <Oren> trying to fixt that
17:02:07 <elliott> contrapumpkin: fix your connection :p
17:03:44 <Oren> the problem is with the null variable _
17:03:49 -!- copumpkin has quit (Ping timeout: 245 seconds).
17:04:04 <Oren> which isn'tin the spec becuase it keeps screing up
17:04:19 <Oren> s/screing/screwing/
17:06:11 -!- copumpkin has joined.
17:07:29 -!- ChanServ has set channel mode: +o elliott.
17:07:36 -!- elliott has set channel mode: +b $a:copumpkin$##fixyourconnection.
17:07:45 <elliott> someone poke me in an hour
17:07:58 <FireFly> Oh, A057755 is about to run out
17:08:32 <fizzie> Still two days for Dominosa and it's so far seen very little action.
17:08:38 <fizzie> And most of what it's seen is #esoteric's fault.
17:09:57 -!- contrapumpkin has quit (Ping timeout: 255 seconds).
17:10:18 <elliott> is it legit or just a really bad cheat?
17:10:59 <elliott> how many bits do you get from the $$ setter? I guess a bit less than 16?
17:11:38 <elliott> b_jonas: but some other processes are already on the system, right?
17:11:44 <elliott> you can't collide with those
17:11:44 -!- copumpkin has quit (Ping timeout: 245 seconds).
17:11:54 <b_jonas> elliott: a bit less than 15 bits really
17:11:54 <elliott> oh, I guess it's actually faked
17:12:02 <elliott> as in your pid isn't really that
17:12:14 <elliott> are you sure? anagol fakes many syscalls
17:12:57 <fizzie> elliott: I've been assuming it's legit, because even just uncompressed embedding is just 240B + whatever language-specific overhead there is.
17:13:07 -!- tlewkow has joined.
17:13:10 <fizzie> Or was it even less than that.
17:13:26 <elliott> fizzie: it has a shebang, documentation, and obeys PEP-8
17:13:57 <fizzie> I was supposed to check what that was all about, but forgot.
17:14:15 <elliott> http://golf.shinh.org/reveal.rb?Wow/PatchiKnowsWhatsUp_1415208522&py
17:15:27 <FireFly> Especially the space after print, as someone pointed out
17:15:46 <int-e> I would've done it again, too.
17:16:09 <elliott> int-e: oh, is that solution yours? or what do you mean
17:16:24 <int-e> No, I liked the space in particular.
17:18:17 <FireFly> Heh, someone did A057755 in jq
17:27:09 <int-e> perl... $$.2, yuck.
17:28:03 <Oren> append 2 to variable $$
17:28:10 <Oren> is how I read that
17:28:27 <Oren> but perl may not read it the same way
17:28:35 <int-e> yes. $$ is the process id, it'll be 315652
17:28:44 <int-e> after appending the 1
17:32:24 <FireFly> So have we decided that Dominosa is a compression task now?
17:32:45 <int-e> so what kind of literals are those C solutions using?
17:32:51 -!- GeekDude has joined.
17:34:07 <b_jonas> http://golf.shinh.org/p.rb?Make+24 is public now
17:34:30 <b_jonas> so now you can see the craziest regex I've ever written:
17:35:39 <b_jonas> this is the solution: http://golf.shinh.org/reveal.rb?Make+24/b_jonas_1414698860&pl
17:36:42 <fizzie> FireFly: It's not a compression task, but I'm sure the best solutions are embeds.
17:37:37 <fizzie> FireFly: Out of my burlesque ones, the un-suffixed one is legit, the "cheat" makes one unwarranted assumption that happens to cause no problems for the three test cases, and the "supercheat" just embeds the outputs.
17:37:51 <elliott> int-e: idea: programming language that somehow uses $$ as part of the control flow/code...
17:38:12 <elliott> save a whole two bytes of information!!
17:39:43 <int-e> of course Dominosa is a data compression task, so where are the Perl solutions :/
17:41:54 <elliott> anagol should just add zlib as a language
17:42:07 <elliott> oh I guess zlib can't condition on input at all though :p
17:42:24 <elliott> hmm, could you easily add input to the "languages" of decompressors?
17:43:31 -!- sebbu2 has changed nick to sebbu.
17:50:39 <fizzie> elliott: I was looking for a feature like that in the end-user gzip/zcat/etc. tools, but couldn't find anything.
17:50:48 <fizzie> (For that one bash+zcat thing.)
17:51:32 <elliott> fizzie: what I mean is considering the decompressors as interpreters for languages like "emit these bytes" and "reference N bytes" ago... I know zip is enough to do a quine, so maybe adding input somehow would be enough to get it to do golf cheats all by itself?
17:51:36 <elliott> maybe you knew I meant that though
17:53:06 <fizzie> Oh, no; I assumed you meant what I was looking for, which was to compress input Y as if it was preceded by input X to something short called Z, and then being able to decompress Z back to Y if given X.
17:54:28 <fizzie> It sounds somewhat reasonable that many compressors could get a shorter Z (compared to just compressing Y alone to Z') if X and Y are similar, but they don't seem to offer that as an option.
17:54:32 <elliott> hand-programming of zip files is where it's at
17:55:51 <fizzie> I think the zlib library has something a bit like that.
17:56:09 <fizzie> See http://www.zlib.net/manual.html deflateSetDictionary + inflateSetDictionary.
17:56:19 <fizzie> "Initializes the compression dictionary from the given byte sequence without producing any compressed output. This function must be called immediately after deflateInit, deflateInit2 or deflateReset, before any call of deflate. The compressor and decompressor must use exactly the same dictionary (see inflateSetDictionary). without producing any compressed output."
17:56:29 <fizzie> "The dictionary should consist of strings (byte sequences) that are likely to be encountered later in the data to be compressed, with the most commonly used strings preferably put towards the end of the dictionary. Using a dictionary is most useful when the data to be compressed is short and can be predicted with good accuracy; the data can then be compressed better than with the default empty ...
17:56:41 <fizzie> But as far as I could tell, the command-line tools didn't provide for that.
17:58:25 <fizzie> (Even though they could very easily have a "do {deflate,inflate}SetDictionary with the contents of file X before {,de}compression" option, and it might even conceivably be useful, to someone, somewhere.)
17:59:26 <int-e> anyway, extending zlib seems wrong, because the trick is usually to come up with a special purpose decompressor that exploits a lot of regular structure in the output that is know a priori.
18:01:52 <int-e> (For example, no matter how much you change its data portion, my Make24 entry will never print "Hello, world!")
18:02:24 <elliott> int-e: yes, I was mostly joking on the zlib solutions that float around.
18:02:56 -!- Oren has quit (Quit: Page closed).
18:03:11 <int-e> right, but usually they don't come out on top overall
18:04:53 <elliott> well, neither does Ada and it's still there :p
18:07:18 <elliott> int-e: actually, you know what might be interesting? an arithmetic coder language
18:07:42 <elliott> I think it's some PAQ variant that represents the compressed data as "prediction model algorithm" + "arithmetic coder output for that data with the prediction model"
18:08:07 <elliott> hard to see how to extend that to multiple outputs, sadly
18:08:16 -!- tromp has quit (Read error: Connection reset by peer).
18:08:42 <fizzie> I will try to use speech recognition participate in the discussion though
18:08:50 -!- tromp has joined.
18:09:11 <elliott> fizzie: lovely to see the spectacular fruits of your research
18:09:32 <fizzie> this is nothing really sucks this is the Google speech recognizer
18:09:42 * elliott pats the fizziebot's metallic head
18:09:43 <fizzie> not my research I mean
18:09:49 <b_jonas> int-e: hehe, mine won't either,
18:09:52 <elliott> fizzie: ah, well, you'd better go improve that, right?
18:09:58 <elliott> not that that would ever happen.
18:10:10 <b_jonas> `perl -e $_="Hello, world!"; s/\w\K\B/+/g;s/\w/hex$&/ge;s/(.*)=/($1)\xc3\x97/g;print
18:10:28 <fizzie> louhi accent is not entirely comfortable with it
18:10:33 <elliott> you're so much cuter in lowercase
18:10:45 <fizzie> I don't know how to speak in uppercase
18:11:24 <fizzie> my wife is finding this amusing
18:11:36 <elliott> is using this to talk part of the corporate culture there
18:12:12 <int-e> b_jonas: In my case it's a close call. you can make it print things like Hel+lo,+world+!
18:12:32 <fizzie> Autolab speech recognizer out it would be in my eye me listing
18:12:58 <int-e> b_jonas: because the numbers are actually taken from stdin.
18:14:20 <fizzie> I tried to make something based on the input numbers but forgot I had would have had to rearrange them
18:14:40 <elliott> what does your wife think of fungot
18:14:40 <fungot> elliott: looks more to me like a bush. actually, now i'm all about " do unto you!
18:15:21 <b_jonas> I don't use the input numbers at all
18:15:52 <b_jonas> in one version I read the first line to determine which of the three testcases I have, but $$ turns out to be shorter
18:16:27 <int-e> b_jonas: I was so happy when I found the -a options of perl.
18:17:00 <fizzie> it's happened to me when I was demonstrating this to some students
18:17:40 <fizzie> tried to say: how old is she
18:17:55 <fizzie> what came out was something something is shaped
18:18:13 <int-e> b_jonas: yes. it just so happened that I was already inserting the operators by indexing into an array, and this gave me another array to do exactly the same for the numbers, virtually for free.
18:19:12 -!- nooga has joined.
18:19:56 <elliott> fizzie: at least it's not as bad as dear aunt, let's set so double the killer delete select all
18:20:41 <b_jonas> dear fungot, let's set so double the killer delete select all
18:20:42 <fungot> b_jonas: that as well, i see!
18:20:49 <fungot> b_jonas: i look forward to. every. single. time.
18:21:16 <b_jonas> fungot, what came out was something something is shaped?
18:21:16 <fungot> b_jonas: that, but also a lot!
18:21:55 <b_jonas> fungot, how do you define a function template that can return void or non-void? can you really only do it with two separate overloads?
18:21:55 <fungot> b_jonas: to the last, i will grapple with thee! from hell's heart, i stab at thee! from hell's heart, i stab at thee! from hell's heart, i stab at thee! from hell's heart, i stab at thee! from hell's heart, i stab at thee! from hell's heart, i stab at thee! from hell's heart, i stab at thee! from hell's heart, i stab at thee! from hell's heart, i stab at thee! from hell's heart, i stab at thee! from hell's heart, i stab at thee
18:22:25 <fungot> elliott: but a good one! it's funny this one time in high school, a friend, but i have a problem, t-rex?
18:22:39 -!- elliott has set channel mode: -b $a:copumpkin$##fixyourconnection.
18:23:40 -!- elliott has set channel mode: -o elliott.
18:25:09 <elliott> "Benedict Cumberbatch Can Charm Humans, but Can He Fool a Computer?" this is the best headline I've ever read
18:25:25 <elliott> "no, he can't fool computers into thinking he's human"
18:25:32 <elliott> by headline I mean wolfram blog post title
18:25:58 <b_jonas> fungot, what's the best headline you've ever read?
18:25:58 <fungot> b_jonas: don't i know you from somewhere? but, i mean, a male, i can be one of those people are going to think you're a pedophile, and he's on a friggin' universe. and then there'll be a day shortly afterwards when we can simulate universe on our cell, because there'd have been no controversy.
18:26:04 <elliott> "Turing machines were one of the focal points of the movie, and we launched a prize in 2007 to determine whether the 2,3 Turing machine was universal." come oooon
18:28:24 <elliott> http://blog.wolfram.com/data/uploads/2014/11/confident-results.png benedict cumberbatch revealed to be actually the same person as alan rickman by stephen wolfram
18:38:43 <fizzie> elliott kalanwww.fronter finnish speech recognizer ovat
18:39:47 <fizzie> langat funka puunkaato unikot kankaat langat kankaat
18:39:57 <fizzie> it's no use I can't say it
18:40:28 <elliott> I hope your research does a lot better than this
18:40:53 <fizzie> that last one was my wife crying
18:41:30 <fizzie> I will try to improve the phone with recognition I believe this office recognize out of
18:41:33 <elliott> can your marriage survive speech recognition
18:41:49 <Bicyclidine> i don't know if i can survive speed recognition
18:41:59 <fungot> Available: agora alice c64 ct darwin discworld enron europarl ff7 fisher fungot homestuck ic irc iwcs jargon lovecraft nethack oots pa qwantz* sms speeches ss wp youtube
18:42:09 <fizzie> high-speed speech recognition
18:43:01 <Bicyclidine> now i want to try taking a speech recognition thing and hold it up to a radio announcer rattling off caveats for an insurance policy
18:43:09 <fungot> fizzie: are you you're going trick-or-treating this year will be better in the future. a year, a balloon! the balloon goes up some of the way, we can consider the real question, which is a good thing! stupid problems
18:44:03 <fizzie> I had to speak fun and got separately
18:44:19 <elliott> did you find the west gutter yet
18:45:10 <elliott> is this on a computer or a phone, anyway?
18:45:25 <fizzie> I'm doing it on my android tablet
18:45:57 <fizzie> so everything I say it sent to Google satellite
18:46:33 <fizzie> can't reach Google at the moment it says
18:47:12 <fizzie> I think this works better for native English speakers
18:47:42 <fizzie> also I have it set to recognize both in Spanish and English
18:48:12 <fizzie> voisi puhua myös suomea
18:49:47 <elliott> you still seem so tiny in lowercase.
18:49:55 <elliott> ...do you even speak spanish?
18:53:32 <fizzie> thanks to machine translation I want need to
18:55:23 <elliott> can it automagically do speech recognition -> translation -> IRC?
18:57:05 <fizzie> its can do a speech recognition to translation to speech synthesis
18:57:17 <elliott> just hook that up to speech recognition again, then
18:57:47 <fizzie> sounds like the best idea
18:58:03 <fizzie> I don't have a second device though
18:58:25 <elliott> they don't shower you with devices when they initiate you?
18:58:40 <elliott> maybe you could use some of Your Research on your computer as the other speech recogniser
18:58:56 <fizzie> I haven't actually started yet
18:59:28 <fizzie> so it's mostly about to finish
19:01:02 -!- AndoDaan has joined.
19:07:27 <fizzie> I tried to say in Microsoft you about finish
19:08:05 -!- ion has quit (Ping timeout: 250 seconds).
19:08:07 <fizzie> my research ace about the Finnish language
19:08:39 <elliott> okay, so translate english -> finnish
19:09:10 <elliott> don't worry, I can translate what you say back to english
19:09:31 <fizzie> I don't have a lot of things up hahaha system have at home
19:09:36 <HackEgo> [wiki] [[GridScript]] http://esolangs.org/w/index.php?diff=41230&oldid=41222 * 71.184.241.244 * (+3)
19:09:58 <Bicyclidine> that person still hasn't seen my talk page note, have they
19:12:27 -!- Oren has joined.
19:13:02 -!- Lorenzo64 has joined.
19:15:57 <elliott> fizzie: so is google speech recognition only going to work in finnish soon
19:20:09 <elliott> Busy testing speech recognition I physeal I'm testing speech recognition to
19:20:15 -!- ion has joined.
19:20:32 <elliott> this is probably even worse
19:20:48 <elliott> okay it seems to be better if I don't claim username
19:21:37 <elliott> and so glad when this is speech recognition
19:23:52 -!- AndoDaan_ has joined.
19:24:13 <fizzie> metsästysseura human computer interaction
19:24:38 <fizzie> it's the phoyoutubefuture
19:24:58 <fizzie> having more than one language enabled is probably not a good idea
19:25:41 <fizzie> it was correct for a moment. Before the trip lasted with the finish world
19:25:54 <elliott> Angela venture furniture S you N G oh T
19:26:12 -!- AndoDaan has quit (Ping timeout: 265 seconds).
19:26:32 <elliott> this is a applesauce action
19:27:10 <fizzie> raision sähköpalvelu grinders
19:27:10 <HackEgo> 722) <Taneb> hang on I have bright idea <Taneb> navajo to f me 1 in 3 people
19:27:20 <elliott> this is apples fault Picture children
19:27:39 <fizzie> so are you using an Apple device
19:28:20 <elliott> of course it gets that I Word right
19:28:36 <fizzie> it has its priorities in autumn
19:28:47 <elliott> it doesn't work very well with the terminal on my accent all my accent all my accent all my accent or my accent
19:29:59 <fizzie> at first you don't succeed
19:30:08 <fizzie> latvia sveitsi live stream
19:30:22 <fizzie> ok I'm switching this to English only
19:30:31 <elliott> it's weird it's takes out of her it types out there wrong words but then backspaces and cracked them and corrects them in a split Second
19:30:46 <fizzie> this thing does that too
19:31:13 <elliott> how come it has given you any Spanish not knot yep yeah yet
19:31:27 <elliott> how come did not give you any Spanish
19:32:17 <elliott> cannot you join into can you cannot and Again they can invent Dawn Nathan Doctor doom
19:32:24 <fizzie> the Spanish was a mistake
19:32:35 <fizzie> no I have only English subject did
19:32:36 <elliott> can then Tanya can then oh Tanya was close
19:32:58 <elliott> Tanya you should join into in to oh
19:33:21 <Taneb> I did this ages ago and it didn't end well
19:33:22 <fizzie> oh you meant Ahmed lemon finance 10m kinetic Thunderhead Gangnam
19:33:26 <elliott> you are typing now your tendon you can you you Tanya
19:33:54 <fizzie> about the software has improved drastically
19:34:17 <elliott> what was metal and finance 10 m meant to be
19:34:42 <elliott> Connecticut thunderhead getting them
19:34:47 <elliott> I'm so glad it knows gang them
19:35:16 <fizzie> I'm sure that train this with us out of sorts queries
19:36:23 <fizzie> it inserts on that last E
19:37:07 <fizzie> tanee be Tanabe I am dB
19:38:31 <fizzie> I don't think I can get this right
19:38:51 -!- Sprocklem has joined.
19:39:13 <elliott> was philosophy osteopathic
19:39:13 <fizzie> le up Colin it doesn't know you are an A
19:39:27 <elliott> that's just call him coughing
19:39:36 <elliott> let's just call him: Colin
19:39:55 <fizzie> I feel bad for the laundry.
19:40:11 <elliott> fizzy isn't so great either
19:40:29 <elliott> your real name would be worse L Shell so so
19:40:30 <fizzie> it's quite close at least
19:41:07 <fizzie> I think it knows my name
19:41:17 <fizzie> it's also knows where I live
19:41:31 <ion> are you chatting with speech recognition I want to try to this is fun
19:41:36 <fizzie> it didn't know my birthday when I asked
19:41:45 <elliott> honey doesn't look as my recognise my name at all
19:42:11 <ion> Hello Elliot and fizzy
19:42:27 <fizzie> hey i on what device are you using Autobots truck driver
19:42:49 <ion> Google voice input on a samsung galaxy phone
19:43:01 <fizzie> this is Google Voice invite on the Nexus tablet
19:43:04 <elliott> Busy MacBook is this actually getting married
19:43:24 <fizzie> honeymoon after getting married
19:43:35 <J_Arcane> this entire channel has become a recursive markov chain ...
19:43:36 <elliott> it's too complicated can you walk me to it
19:44:05 <fungot> fizzie: this i do believe so! excuse to stomp! little does the dromiceiomimus know that i intend to do today! we can't hang out in paradise, alone, to raise the friggin'. fragile, species can survive their electrical onslaught
19:44:21 <ion> Testing. Explicit punctuation seems to work.
19:44:26 <fizzie> I have to stop talking in the middle of fun and got
19:44:40 <fizzie> I am calling I tried suicide calling but it didn't recognize
19:45:17 <elliott> peace and police officer please don't call yourself busy
19:45:29 <fizzie> ion:and get drunk show a shin is sometimes but not often
19:46:01 <ion> I have trouble importing fungal
19:46:20 <fizzie> I don't remember what I was trying to say when it recognize it by side
19:46:35 <elliott> I have trouble importing fungal to
19:46:46 <J_Arcane> omg https://github.com/jhallen/joes-sandbox/tree/master/exorsim/mpl
19:46:53 <fizzie> try saying fun than typing to post tense tense I got
19:46:56 <elliott> are you still tan sexy fizzy
19:47:10 <fizzie> yes I insist a lot of sexy
19:48:46 <elliott> I'll be getting married fizzy
19:49:04 <fizzie> who are you are getting married to
19:50:16 -!- hjulle has joined.
19:50:38 <fizzie> and don't think you can get somebody to talk concept
19:51:25 <Oren> what was wrong with using a keyboard?
19:51:27 <elliott> I can talk and settlement all I want
19:51:35 <elliott> it's not the future on Oren
19:52:07 <fizzie> it's so much more convenient this way
19:52:09 <Oren> what about when there are 25 people in the room with you
19:52:21 -!- AndoDaan_ has quit (Ping timeout: 264 seconds).
19:52:30 <Oren> all talking into their computers
19:52:57 <zzo38> To me keyboard is better and I can type fast too
19:52:58 <elliott> is he loves problems like that fizzy
19:53:07 <zzo38> Speaking into the computer is good if you want a sound recording though
19:53:19 <zzo38> But I don't need a sound recording
19:53:22 <elliott> if I want to send according I can progress into speech since that
19:53:31 <elliott> if I want a sound recording I can progress into synthesiser
19:54:31 <fizzie> but he's the best apps Mario's full on the phone with no good kid lock
19:54:36 -!- AndoDaan has joined.
19:54:53 <fizzie> invite in public is probably Mario s photo on phone with no good keyboard
19:55:05 <fizzie> voice input models for
19:57:24 <fizzie> checklist google image search
19:57:50 -!- AnotherTest has joined.
20:17:31 -!- AndoDaan has quit (Ping timeout: 258 seconds).
20:23:44 <FireFly> You should try speech-recognizer | bash | speech-synthesizer
20:24:29 <HackEgo> [wiki] [[GridScript]] http://esolangs.org/w/index.php?diff=41231&oldid=41230 * 71.184.241.244 * (+543)
20:25:38 <fizzie> actually my colleague my speech recognizer into xD O'Toole
20:25:48 <fizzie> so that it would type into any window that has focus
20:26:01 <fizzie> I tried to write one paragraph of my teeth is using it
20:26:04 <b_jonas> well, I'm not an engineer, I don't know.
20:28:22 <fizzie> it was kind of dangerous
20:29:09 <Vorpal> fizzie, hm? xdotool? Isn't that for moving windows around and such?
20:29:17 <FireFly> xD O'Toole is a good name for software
20:31:10 <Vorpal> Right, a different one according to the logs. Speaking of which, why didn't znc replay the log like it is configured to when I reconnected?
20:31:48 <b_jonas> fungot, why didn't znc replay the log like it is configured to?
20:31:49 <fungot> b_jonas: dromiceiomimus, i don't!
20:33:51 -!- AndoDaan has joined.
20:35:21 -!- tlewkow has quit (Remote host closed the connection).
20:42:50 <fizzie> yes you can use it to type 2
20:43:12 <fizzie> in addition to moving windows and stuff
20:43:33 <fizzie> it pretends to be a kid lot
20:43:42 <fizzie> send stupid a key events
20:44:45 <FireFly> translating speech recognition output → english is quite tricky at times
20:47:48 -!- Sprocklem has quit (Ping timeout: 258 seconds).
20:48:14 <Vorpal> Is there a way to say to cabal "fix mixed versions"? I.e rebuild any packages depending on different versions of a package so that everything just works, and I don't get weird errors.
20:48:24 -!- ion has quit (Ping timeout: 255 seconds).
20:53:45 <elliott> using cabal sandbox is probably the best option but I'm a year+ out of date with haskell
20:54:05 <Vorpal> elliott, I mean, gentoo managed to make this shit work, so why can't haskell
20:54:23 -!- zzo38 has quit (Remote host closed the connection).
20:54:33 <elliott> well, the problem is inherently hard. cabal has some questionable behaviour in response to it, though
20:55:23 <Vorpal> I'm fine with it telling me that "these packages will need to be rebuilt if you update this package, do you want me to do that, recursively?"
20:55:42 -!- AndoDaan has quit (Quit: Quit).
20:55:49 <Vorpal> Or of course "this will break, and the version restrictions in the dependencies make it unsolvable"
20:56:14 <Vorpal> It should never let me end up in a state where I have a pretty much broken environment
20:56:30 <Vorpal> So yes it would need rollback support
20:56:50 <elliott> does gentoo actually allow coinstallation of multiple versions?
20:57:59 <Vorpal> elliott, not in general, but for some packages that marks it as ok (for example, kernel, major gcc versions, ...)
20:58:49 <Vorpal> It does have infrastructure for rebuilding broken packages though
20:59:21 <elliott> well, cabal allows multiple versions of the same package. so the problem is fundamentally different
21:00:11 <Vorpal> True, but I'd in general prefer to not have multiple versions because right now I'm dealing with two different packages that want different versions of a common dependency
21:00:28 <Vorpal> So I have to rebuild stuff, but I can't find any automated way to do it
21:00:33 <elliott> Vorpal: are you sure you'd prefer that? what about when half of the library ecosystem depends on QuickCheck 2 and the other half QuickCheck 3?
21:01:06 <elliott> Gentoo has control over all its packages
21:01:12 <elliott> so it can avoid things like that
21:01:15 <elliott> if you want something like that, check out stackage.
21:01:27 <Vorpal> elliott, however, what am I supposed to do when I have B needs A1 and C needs A1 and I need to exchange A-data between B and C
21:01:37 -!- ion has joined.
21:01:44 <Vorpal> Which is my current issue
21:01:55 <Vorpal> elliott, I will check that out, thanks
21:01:56 <elliott> I think you typo'd somewhere there
21:02:05 <Vorpal> elliott, err yea, A1 and A2
21:02:41 <elliott> well, nothing on earth can solve that problem. it is true that it is all too easy to get multiple versions of a library linked within a single *program*, which is only okay if those libraries are implementation details for other libraries you use, basically.
21:02:55 <elliott> (you *can* work with multiple ones explicitly with PackageImports I think, but please god don't)
21:03:11 <elliott> cabal sandbox would help avoid that because it'd install your dependencies "all at once"
21:03:25 <Vorpal> elliott, It isn't like B or C actually need different versions of A, it is just they ended up being built against different versions
21:03:38 <elliott> yes, hence my cabal sandbox remark
21:04:29 <Vorpal> elliott, I added C to the project much later, I guess in that case I would have to rebuild the sandbox from scratch? I will have to look into how that works. And stackage. And find out which option is best for me.
21:04:46 <elliott> I forget if it'd just work or not.
21:04:56 <elliott> stackage is something you use *with* the sandbox, generally, I think
21:05:04 <elliott> the nice thing about the sandbox is that your projects are fully isolated from each other
21:05:13 <elliott> also, I think by now "rebuild" may not actually be rebuilding, because cabal caches builds properly
21:05:32 <Vorpal> My main point here is that there doesn't seem to be much help for you once you run into an issue like that. On gentoo at least there is tools to clean up the mess.
21:06:39 <elliott> it's easy to clean up with cabal sandbox.
21:06:42 <elliott> just blow away the sandbox.
21:06:53 <elliott> if you don't isolate, then the analogy is rm -rf ~/.cabal
21:07:07 <Vorpal> I didn't isolate, because I didn't know I could do that.
21:07:28 <elliott> I agree the haskell library experience is hell.
21:07:42 <HackEgo> [wiki] [[Scrip7]] http://esolangs.org/w/index.php?diff=41232&oldid=41167 * Orenwatson * (+423) better interpreter version, brainfuck interpreter
21:08:19 <Vorpal> Right. Also that will take a while to rebuild, because one of the libraries in there has like 100 of deps... (not related to the issue though).
21:08:38 -!- tlewkow has joined.
21:08:39 <newsham> its so simple.. just create a different user for each project!
21:10:42 <Vorpal> Yeah I definitely think cabal could do a much better job of preventing situations like this and helping you fix them.
21:11:11 <Vorpal> I should look into how NixOS does this. Since it allows multiple versions to co-exist
21:11:44 <Vorpal> elliott, do you happen to know how nixos deals with this?
21:11:55 <Vorpal> That should be a much better analogy than gentoo in fact
21:13:19 <Oren> Does the requirement that a pointer to memory be a fixed size in memory make a language turing-incomplete?
21:13:44 <Bicyclidine> having a finite number of possible states rules out turing completeness, yes
21:13:56 <fizzie> the Mondeo system also I love smoking bowl out of snow
21:14:19 <fizzie> http://modules.sourceforge.net/
21:14:32 <fizzie> it's very simple though
21:14:59 <fizzie> cheated by copying pasting the f you out of hell
21:15:19 <Vorpal> "<fizzie> the Mondeo system also I love smoking bowl out of snow" <-- what
21:15:38 <fizzie> the muscular system of a lot of installing multiple lation
21:15:51 <fizzie> the more the old system the money he owes
21:15:51 <Vorpal> fizzie, are you drunk?
21:16:17 <fizzie> there we go I was pronouncing it wrong
21:16:34 <fizzie> so the module system allows installing multiple version
21:17:01 <Vorpal> fizzie, why are you using voice recognition?
21:17:33 -!- Oren_ has joined.
21:17:52 -!- Oren has quit (Ping timeout: 246 seconds).
21:18:09 <Oren_> Because C requires ability to acesss pointers to arrays of pointers, so pointers have to have a fixed size
21:18:17 <fizzie> I will start working for Google Nexus. Or so I thought I'd try out what about system its like
21:18:57 <Oren_> which means the set of languages in the real world that are turing complete is rather small
21:19:08 <fizzie> and not the Nexus next year
21:20:05 <Vorpal> Oren_, unless you do some tricky stuff with IO to add extra storage. For a non-seekable file (i.e. pipe), there is no size limit as far as I know
21:20:58 <Oren_> oh! that could work
21:21:18 <fizzie> it's kind of external the language love
21:22:09 <Vorpal> Yes use a pipe to self as an infinite queue, probably needs POSIX on top of C though, don't think the pipe functions are in standard C
21:22:28 <fizzie> prostate spike buck Futter sought a very limited in size
21:22:54 <elliott> prostate spike buck Futter
21:23:10 <fizzie> this doesn't work but it well otto technical topics
21:23:26 <fizzie> I'll cheat with a keyboard for a moment.
21:23:28 <Oren_> he says pipe buffers are small
21:23:35 <fizzie> You can't write an infinite amount into a POSIX pipe.
21:23:36 <Vorpal> Oren_, how did you decode that
21:23:48 <elliott> if you speak it out loud and imagine it's muffled and look at context that can help
21:23:55 <Vorpal> fizzie, well that is a technical limitation of the real world, ulimit and so on
21:24:01 <Oren_> I did that in my head yeah
21:24:01 <elliott> "posix pipe buffers are very limited in size"
21:24:15 <FireFly> Couldn't you address a potentially infinite amount of files in C?
21:24:33 <FireFly> I guess there is technically a limit on the length of the filename
21:24:41 <Vorpal> Yes in the real world they are very limited, and so is RAM. but I don't know that it is specified to not be potentially infinite
21:24:42 <FireFly> since the filename has to be in RAM
21:24:49 <elliott> Oren_: did you, like, logread old #esoteric arguments about C turing completeness
21:25:23 <Oren_> i was wondering if my language scrip7 could possibly be turing complete.
21:25:29 <Vorpal> FireFly, yes, it becomes number of possible files * sizeof(fpos_t) or something like that
21:25:38 <Vorpal> Which is HUGE, but not infinite
21:25:53 <Oren_> it seems probably not
21:25:57 -!- Frooxius has quit (Quit: *bubbles away*).
21:26:18 <elliott> unless your language is defined in terms of C
21:26:19 <Oren_> a language has to be quite abstract to be turing complete it seems
21:26:39 <elliott> nah, C is just unusually airtight in that respect, IMO
21:26:40 <Oren_> my language inherits many features from C but has a terse syntax
21:26:55 <fizzie> Vorpal: The fact that PIPE_BUF is defined kind of implies there must be a limit.
21:26:58 <Oren_> features like pointer arthmetic
21:27:14 <elliott> looking at https://esolangs.org/wiki/Scrip7, does anything in the language prevent you from making pointers bignums?
21:27:15 <fizzie> But it's possible it's not quite entirely explicitly required.
21:27:41 <fizzie> (The references to PIPE_BUF are mostly about atomicity and behaviour w.r.t. blocking.)
21:27:56 -!- Frooxius has joined.
21:28:14 <Vorpal> fizzie, where is that defined? I don't have the PDF on this computer, so I'm trying to find the man page
21:28:16 <Oren_> well, you are supposed to be able to go o>i to access the ith element of a list of pointers at o
21:28:17 <fizzie> And it's indeed defined as the "maximum number of bytes that is guaranteed to be atomic", so.
21:28:42 <elliott> Oren_: just imagine an abstract model of the language where pointers are represented by arbitrary natural numbers
21:28:51 <elliott> and memory is an infinite list of those
21:29:01 -!- Patashu has joined.
21:29:15 <elliott> I guess if you can access things as chars then you can "see" the representation of bignums which might be awkward
21:29:15 <newsham> [11:25] < Oren_> a language has to be quite abstract to be turing complete it seems
21:29:27 <fizzie> Vorpal: The constant's defined in the POSIX spec of <limits.h>. But you're right that maybe it's not strictly speaking a limit for the maximum amount of pending data. It's of course not defined by C.
21:29:28 <newsham> C integers can be unbounded size, no?
21:29:44 <newsham> does the standard mandate a finite size?
21:29:52 <Oren_> no there is intmax.h or something
21:29:58 <Vorpal> fizzie, well yeah I don't think pure C can be made TC
21:30:05 <fizzie> newsham: At least you can make a strong argument that the value of sizeof must be a finite number.
21:30:34 <newsham> so obviously someone needs to make a C variant!
21:30:52 <Oren_> I kind of made a c variant
21:31:03 <Vorpal> Taneb, only at compile time, and with infinite template evaluation depth iirc
21:31:04 <FireFly> Taneb: does C++ include the C++ preprocessor?
21:31:10 <Oren_> it has different syntax but very similar semantics
21:31:19 <Taneb> FireFly, I don't think so :(
21:31:40 <Vorpal> FireFly, the preprocessor of C and C++ is limited, maximum recursion depth and so on
21:31:48 <Oren_> or at least C as it is commonly implemented on x86
21:31:59 <Taneb> I did make a factorial function using GHC's type system
21:32:05 <Vorpal> Think it is specified as a "at least" though
21:32:19 <FireFly> Well, 'at least' is fine :P
21:32:21 <Taneb> It runs surprisingly quickly, but if you ask for the factorial of some big numbers things go wrong
21:32:26 <FireFly> That's a bound in the "right" direction
21:32:37 <Vorpal> FireFly, not quite sure. It is something like "at least x recursive include", not sure if has to limit it at all
21:32:46 <Taneb> (I think when log_2 (n!) > 200 or so)
21:32:54 -!- Sprocklem has joined.
21:33:11 <Vorpal> FireFly, I recommend checking the spec, which I don't have downloaded on this computer
21:33:42 <Vorpal> FireFly, similar for the recusion of templates in C++, "at least x iterations"
21:34:02 <Vorpal> Pretty sure GCC limits both though
21:34:08 <Vorpal> And neither is at runtime of course
21:34:17 <Oren_> GCC isn;t the spec tho
21:34:52 <Oren_> some madman might have made a wildly screwed up compiler that breaks conventions
21:35:07 <fizzie> I'm pretty sure the C preprocessor is generally considered not TC, but I've forgotten the argument against why recursive include doesn't make it so.
21:35:10 <Vorpal> Well obviously, but it is a common implementation. Just saying it isn't viable to use it practically
21:35:15 <fizzie> The macro expansion has explicit no-recursion rules.
21:35:16 <Oren_> and allows as many templates as there is memory for
21:35:40 <Vorpal> fizzie, C++ templates are TC I'm fairly certain though
21:35:45 <fizzie> The templates are, yes.
21:36:01 <Vorpal> Again with infinite recursion
21:36:03 <fizzie> And if you allow repeated executions of the C preprocessor, with the output fed in as the new input, it is too, I believe.
21:36:35 <Oren_> cpp can't generate newline
21:36:55 <Oren_> so the number of lines decreases each iteration of cpp
21:37:03 <Vorpal> Or it could stay equal
21:37:12 <Vorpal> There is no need for it to be multiple lines either
21:37:18 <Vorpal> What matters is statements
21:37:24 <fizzie> Well, I mean. You know, this thing: http://www.ioccc.org/2001/herrmann1.hint
21:37:31 <Oren_> #define isn't a statement
21:37:43 <Oren_> it ends at newline not ;
21:38:02 <Vorpal> True, but you can't generate directives anyway
21:38:05 <elliott> fizzie: I don't know if it's TC or not. it can implement a functional language
21:38:09 <elliott> it does seem to be weirdly limited somehow though
21:38:15 <elliott> I can't figure out how to even do the things the wizards do with it
21:38:28 <Vorpal> Oren_, also look at the Boost Preprocessor library, it is quite interesting how much you actually CAN do with it
21:38:44 <Oren_> isn't boost C++ tho
21:38:48 <fizzie> chaos-pp/order-pp are IMO the most interesting preprocessor abuses.
21:39:02 <Vorpal> Oren_, Sure but the preprocessor bit is basically the same for both
21:39:08 <fizzie> But they too have some non-TC upper bounds.
21:39:25 <Vorpal> fizzie, what is chaos-pp/order-pp?
21:39:27 <elliott> ah, what is order-pp's upper bound?
21:39:29 <elliott> that was the language i was talking about
21:39:33 <fizzie> https://github.com/rofl0r/order-pp/blob/master/example/fibonacci.c
21:39:46 <b_jonas> Vorpal: that, and also the P99 library
21:39:47 <FireFly> https://github.com/rofl0r/chaos-pp/blob/master/chaos/preprocessor/algorithm/merge_sort.h ...
21:39:52 <fizzie> "Unless you are already convinced, you should check, by preprocessing this example, that the actual parameter to `printf' in the above code is just a single string that contains the 500th Fibonacci number in base 10."
21:40:05 <fizzie> elliott: IIRC, it's configurable. But it must be an integer.
21:40:21 <elliott> right, but I mean, what's it an upper bound *on*?
21:40:57 <elliott> I would really like cpp to be TC...
21:41:22 <b_jonas> Vorpal: http://p99.gforge.inria.fr/
21:41:24 -!- dts has joined.
21:42:29 <b_jonas> but I think the boost preprocessor library has more powerful stuff
21:42:36 <fizzie> elliott: Something like that. It's part of chaos, and that's even more confusing than order. Something something exponential.
21:43:16 <Vorpal> b_jonas, Well I seen some simple Boost PP code evaluating lists of stuff and such mostly
21:43:33 <elliott> Vorpal: should just click fizie's fibonacci link :p
21:44:02 <Vorpal> And no I can't read it :P
21:44:38 <Vorpal> I think those lists are limited to 255 elements in boost preprocessor though
21:44:41 <fizzie> The thing with all the 8s seems to be characteristic of preprocessor things that "go too far".
21:44:55 <fizzie> (Sometimes it's some other number.)
21:45:55 <Vorpal> Does it need to be a number?
21:45:55 <fizzie> I had some sort of intuition why it's required, but I've completely blacked that out.
21:46:31 <elliott> it stops it being interpreted as a preprocessor token thingy, I think
21:47:12 -!- dts has quit (Quit: Leaving).
21:47:32 -!- dts has joined.
21:47:58 <Oren_> oh yah, did i mention the bf interpreter for scrip7 expands the memory dynamically?
21:48:02 <Oren_> Q=(a>1S=PS+NS~ON*2bNNO=PN/2O+NaS0N*2#H+3G=H`)
21:48:24 <b_jonas> um, is that a chemical formula?
21:48:33 <Oren_> its a function in scrip7
21:48:52 <fizzie> "8foo" is a single pp-number in the translation phase that decomposes the source to preprocessing tokens.
21:48:55 <Oren_> it conditionally expands the bf memory if it needs to
21:49:00 <Vorpal> Hm, Boost PP should be usable in C I assume
21:49:07 <b_jonas> does it deliberately try to look like a chemical formula?
21:49:24 <Oren_> but i see what you mean
21:49:39 <Oren_> what with NN0 and such
21:49:58 <fizzie> (The "pp-number" syntax includes just about everything that starts with a digit, it can be followed by arbitrary identifier characters.
21:50:25 <fizzie> But it's still a pp-number and doesn't participate in macro expansion and such in the same way as identifiers; it probably has something to do with that.)
21:50:30 <Oren_> N is register 5 accessed as a 64bit int
21:50:55 <Oren_> O is register 0 acessed as an adress
21:50:58 <fizzie> Also I didn't realize order-pp grew out of the Boost preprocessor library: https://github.com/rofl0r/order-pp/blob/master/README.md
21:51:38 <Oren_> because someone went completely bonkers?
21:52:33 <Bicyclidine> hm, smiles uses @@ for stereochemistry. bizarre
21:52:38 <FireFly> Apparently order-pp has (.)
21:52:50 <FireFly> "8compose(f, g) : (a -> b) -> (c -> a) -> (c -> b)"
21:54:15 <Oren_> ~ is equality test becuase = was taken by assignment
21:55:16 <Vorpal> FireFly, that looks haskellish
21:55:19 <Oren_> what kind of chemical formula uses *2 tho?
21:55:38 <FireFly> Vorpal: looks like the type signatures in https://github.com/rofl0r/order-pp/blob/master/doc/notes.txt are haskell-inspired
21:55:40 <elliott> fizzie: rather, chaos-pp did
21:55:54 <elliott> that repo is chaos-pp with all the parts that aren't needed for order-pp removed, I guess
21:56:31 <FireFly> Oren_: is : used for anything?
21:56:53 <Vorpal> elliott, okay, so which one should you use? chaos or order?
21:56:53 <Oren_> : is used for "set and move forward"
21:57:07 <elliott> chaos-pp is a cpp library.
21:57:11 <elliott> order-pp is a language implemented in cpp.
21:57:16 <elliott> that uses chaos-pp, I think.
21:57:25 <Oren_> like if you have (struct){int i; long j; char *s}
21:57:53 <FireFly> K has : for assignment and = for equality, at any rate
21:57:55 <Oren_> then you do i:4 I:2 o:"foobar" to set each variable in turn
21:58:31 <Oren_> this grew out of a config file format
21:59:13 <FireFly> A config file that is essentially compressed C?
21:59:18 <Oren_> yes now config files can modify arbitrary memory inside the game
21:59:28 <Oren_> but it is interpreted
22:00:03 <Oren_> but it uses ragisters, not variables
22:00:04 <elliott> 8paste(l, r) : a -> b -> c -where 8isnt_edible(r)
22:00:06 <FireFly> Well, I guess it's not too different from emacs' config files
22:00:20 <Oren_> which are lisp iirc
22:03:18 <Oren_> if my game was any good I'd postit
22:05:04 <b_jonas> Oren_: well, that looks like it's easily turing complete to me. you can allocate memory, point into it with several pointers, move those pointers, read and write through them, you have loops and conditionals (even if I'm not sure how those conditionals and loops work).
22:06:05 <Oren_> the conditionals are skips: i~0 skdkasj # will skip to the # if the 0 register points to a 32 bit value of zero
22:06:27 <b_jonas> Oren_: so they skip to the next # commadn?
22:06:43 <Oren_> and $ is unconditional jump
22:06:50 <b_jonas> and the # command itself is a no-op?
22:07:18 <Oren_> (){}[] are all unconditional loops inside unconditional jumps
22:07:40 <b_jonas> and you can exit from them with a skip instruction
22:07:47 <Oren_> liek goto b a: goto a b:
22:08:05 <b_jonas> not that it's really needed, but is there an easy way to have a nested if?
22:08:21 <elliott> b_jonas: C has all those properties too
22:08:43 <Oren_> reverse condition and use i!0{# dfsdsdf}
22:09:16 <b_jonas> elliott: meh, it's turing-complete enough for my purposes.
22:09:22 <Oren_> scrip7 should be about as TC as C
22:10:04 <elliott> b_jonas: but the topic was whether scrip7 was actually TC, with the (correct) assumption that C isn't.
22:10:20 <elliott> (C is not turing complete enough for esowiki categorisation, say, which is relevant here)
22:10:32 <b_jonas> Oren_: what I don't understand is, why are the registers of different types aliased? wouldn't it be better if they were separate, and there was a command for assigning the pointer from one to another?
22:11:01 <Oren_> so that you can easily set a bunch of members of a struct in order
22:11:12 <b_jonas> Oren_: in fact, are there commands to read and write the address where a register is pointing to to the memory?
22:11:50 <b_jonas> oh, so that's what the addr type does!
22:12:05 <Oren_> it gets the value of the pointer itself
22:12:30 <b_jonas> that also explains how the G register works
22:12:45 <Oren_> yeah that's how you make subroutines
22:13:05 <Oren_> by convention i've been using H to store the return address
22:13:34 <Oren_> if you look at the bf interpreter the > and < commands use subroutines
22:13:49 <Oren_> it's really more of a compiler...
22:14:11 <Oren_> it turns bf into scrip7 and then jumps directly into that code
22:16:11 <b_jonas> Oren_: scrip7 also looks like it's almost compilable, but not quite, because writes to the G register could be a bit difficult to compile
22:16:46 <b_jonas> you could probably still compile it, you'd just have to try to figure out some of the easier cases of writing to G, and if any remains, handle it with a big indirect jump table
22:17:09 <Oren_> yeah. even if you don't support jumping to data space You would need to ensure that each instruction's compiled form is the same size
22:17:18 -!- Sprocklem has quit (Ping timeout: 264 seconds).
22:17:50 <b_jonas> Oren_: you don't really have to if there's a table mapping the scrip7 addresses to compiled addresses
22:18:01 <b_jonas> though there's the problem that the scrip7 can have self-modifying code
22:18:11 <b_jonas> if that's allowed, you can't really compile
22:18:21 <b_jonas> because basically any instruction could write there
22:18:32 <Oren_> right.... but semantics of that are just as undefined as writing to a function pointer in C
22:18:56 <b_jonas> but if you forbid that and also forbid wild indirect jumps, you could compile it pretty efficiently
22:19:37 <Oren_> whoops need a cast through (void*)
22:20:08 <fizzie> Not really through void *.
22:20:16 <fizzie> Casting it directly to char * is no better or worse.
22:20:42 <fizzie> (But it does need a cast.)
22:21:05 <Oren_> using (void*) is instinct
22:21:23 <fizzie> It wouldn't work the other way around, anyway.
22:21:45 -!- AnotherTest has quit (Ping timeout: 264 seconds).
22:22:11 <Oren_> ((void (*)(char*))"Hello, world")("Foobar");
22:22:22 <fizzie> char *data; int (*f)(int) = (int (*)(int))data; needs a cast to the function pointer type, and going to a void * won't help.
22:23:39 <Oren_> line that i wrote above compiles with GCC
22:23:57 <fizzie> Yes, there's nothing "wrong" with it.
22:24:30 <Oren_> everything about it is evil, but you could theoretically use that for inline machine code
22:24:41 <b_jonas> Oren_: but what I don't see is how you're using this as a configuration format?
22:24:56 <fizzie> Not on very many systems.
22:25:00 <Oren_> Becasue the interpreter gets a pointer to the data space
22:25:20 <fizzie> I think we had some machine code oneliners here, but they're complicated by an ugly required mprotect call.
22:25:21 <Oren_> void scrip7(char *code,void *data);
22:25:44 <Oren_> so if you pass in a struct for the data, your code can alter the contents of that
22:26:49 <Oren_> and the main in scrip7.c is a weak symbol so the one in my game overrides it
22:27:13 <Oren_> while compiling it alone gives you an interpreter
22:27:45 <b_jonas> using such a low-level language as configuration file language sounds a bit unsafe, but whatever suits you
22:28:22 <fizzie> `run echo 'const char f[] = "1\xc0""1\xff""1\xd2\xff\xc0\xff\xc7H\x8d""5\n\0\0\0\xb2\x06\x0f\x05\xb0<1\xff\x0f\x05hello\n"; int main(void) { mprotect((void *)((unsigned long)f&~0xffful), 0x1000, 7); ((void (*)())f)(); }' | gcc -x c - -o /tmp/x && /tmp/x # with some implicit function declaration action
22:28:23 <Oren_> the whole point was for the configuration to be able to allocate memory for wtrings and such
22:29:20 <fizzie> There's `runc and `! c, but I always forget their peculiarities.
22:29:35 <fizzie> One needs all \s doubled because it allows \n so that you can use preprocessor directives.
22:29:39 <HackEgo> #!/bin/bash \ t=`tempfile` \ echo -e "$@" | gcc -trigraphs -o $t -x c - 2>/dev/null && $t \ rm $t
22:29:55 <elliott> is that because there's a trigraph for newlines or something
22:30:00 <HackEgo> #!/bin/sh \ . lib/interp \ interp_file "./interps/gcccomp/gcccomp c"
22:30:00 <fizzie> Possibly both of them need \\s, actually.
22:30:11 <elliott> `cat interps/gcccomp/gcccomp
22:30:12 <HackEgo> #!/bin/bash \ LANG="$1" \ echo >>"$2" \ \ case "$LANG" in \ c) \ HEAD='#include <stdio.h>\n#include <stdlib.h>\n#include <string.h>\n#include <sys/types.h>\n#include <unistd.h>\nint main(int argc, char **argv) {' \ TAIL='; return 0; }' \ EXT='c' \ GCC='gcc' \ FLAGS='-std=gnu99' \ ;; \ \ c++)
22:30:21 <elliott> `url interps/gcccomp/gcccomp
22:30:22 <HackEgo> http://codu.org/projects/hackbot/fshg/index.cgi/file/tip/interps/gcccomp/gcccomp
22:30:26 <fizzie> That one has the "try it with a main function" thing.
22:30:30 <b_jonas> isn't there one that's like geordi, translating backslashes outside strings to newlines?
22:30:42 <fizzie> Like candide, you mean. :p
22:30:49 <fizzie> (The ##c answer to ##c++'s geordi.)
22:30:56 <fizzie> And no, we don't have anything fancy.
22:31:18 <fizzie> The \! c is closest in that it tries compiling with a provided main wrapper first.
22:31:19 <HackEgo> ./interps/gcccomp/gcccomp: line 53: 307 Segmentation fault /tmp/compiled.$$
22:31:22 <b_jonas> mind you, it would have to tokenize C for that, which isn't trivial with comments and raw strings and trigraphs and suchlikes
22:31:41 <elliott> b_jonas: you could maybe do something with libclang.
22:32:52 <fizzie> `! c int main(void) { puts("foo"); } /* there's also this problem */
22:33:21 <fizzie> That's because it turns into int main(void) { int main(void) { puts("foo"); } return 0; } which compiles and runs just fine.
22:33:34 <elliott> is that actually legal C99?
22:33:38 <fizzie> No, but it's legal GCC.
22:33:40 <elliott> wait, nested functions aren't C99
22:33:49 <elliott> isn't main reserved or something? can you really shadow functions??
22:33:54 <Oren_> nested functions are GCC11 or something
22:34:05 <fizzie> You can do int f(void) { int f = 42; return f; } so why not?
22:34:14 <fizzie> There was an attempt to add some flags that disable nested functions, but that ran into some other issue that I forget.
22:34:22 <Oren_> because gcc puts everything into the default version
22:34:38 <elliott> fizzie: probably it broke doing int f() { return 42; } printf("%d\n", f()); or such
22:34:56 <b_jonas> couldn't it just try to link without the wrapper first, then recompile and link with the wrapper if it fails to link?
22:34:57 <fizzie> `! c int main(void) { puts("you can do this but it's real silly"); } main();
22:34:59 <HackEgo> you can do this but it's real silly
22:35:29 <elliott> `! c void main() { puts("see, this is the right signature after all"); } main();
22:35:31 <HackEgo> see, this is the right signature after all
22:37:54 <elliott> not unless you want me to ban ou
22:39:01 -!- dts has changed nick to Sargon_.
22:39:24 -!- Sargon_ has changed nick to dts.
22:39:38 <fizzie> `run sed -i -e '47s/SOURCE/2/;48s/2/SOURCE/' interps/gcccomp/gcccomp # now that you mention it, I see no immediate reason why the order could not just be swapped
22:39:58 <fizzie> `! c int main(void) { puts("foo"); }
22:40:46 <Oren_> `! c puts("\e[32mdoes this work");
22:41:17 <Oren_> `! c puts("\33[32mdoes this work");
22:41:45 <elliott> fizzie: what was the other old unix box you have, an SGI Indy and a...?
22:41:50 <b_jonas> or even better, use the convention geordi does: if the code starts with an open bracket, then find the block of code (again not so easy), move it to the end, and make it the main functino
22:41:52 <fizzie> elliott: SparcStation 5.
22:42:24 <elliott> fizzie: right. I wonder if you can run Windows NT 4.0 on the SGI Indy.
22:42:27 <elliott> that sure would be something.
22:42:40 <Oren_> my dad has a AIX box still running
22:42:55 <fizzie> I have absolutely no idea what sort of machines the MIPS port of Windows NT ran on.
22:42:55 <elliott> b_jonas: windows used to support alpha, mips and powerpc, didn't you know? :)
22:43:18 <fizzie> Somehow I don't think it's SGI machines though.
22:43:24 <elliott> of course you can't find much in the way of third-party binaries, but they aren't nonexistent, either.
22:43:29 <Oren_> makes Mac moving to intel not a big deal
22:43:39 <elliott> well, I mean, nobody actually /used/ this
22:43:51 <elliott> except I think the Alpha port got some use because the chip could emulate x86 pretty well or something
22:44:08 <elliott> Windows NT was deliberately designed to be very portable, though
22:44:21 <fizzie> Apparently ARC-compliant MIPS systems.
22:44:22 <elliott> they designed it on a random Intel RISC chip that they never actually shipped it on to ensure that
22:44:29 <b_jonas> elliott: really? I thought it supported only one architecture besides x86, possibly if you don't count intel64 which was later
22:44:31 <fizzie> So it's even borderline possible, because Indy is on that list.
22:44:53 <fizzie> "Windows NT 3.1 was released for Intel x86 PC compatible, DEC Alpha, and ARC-compliant MIPS platforms. Windows NT 3.51 added support for the PowerPC processor in 1995, specifically PReP-compliant systems such as the IBM Power Series desktops/laptops and Motorola PowerStack series; --"
22:45:16 <elliott> fizzie: it would reveal how much of a dork I am if I said having a real live non-x86 machine running Windows has been a pipe dream of mine for years now, right?
22:45:27 <elliott> I mean, I'd get absolutely no use out of it.
22:45:32 <b_jonas> elliott: you mean non-emulated, right?
22:45:34 <elliott> but what a beautiful object it'd be.
22:45:38 <b_jonas> because emulated it'd be easy
22:45:40 <elliott> b_jonas: I said real live :p
22:45:50 <elliott> I've done it emulated, I think, maybe. it's not as easy as it should be.
22:45:56 <b_jonas> elliott: no, I mean the box would be real live, but it ran an x86 emulator with windows in it
22:46:06 <b_jonas> that's the config that would be easy
22:46:08 <Oren_> i have a NES, I was born in 1993 when the SNES was getting old
22:46:16 <fizzie> elliott: "Intergraph Corporation ported Windows NT to its Clipper architecture and later announced intention to port Windows NT 3.51 to Sun Microsystems' SPARC architecture,[41] but neither version was sold to the public as a retail product." You should try to get hold of that.
22:46:43 * elliott has bricked two PPC macs trying to get unix onto them
22:46:51 <fizzie> Though "announced intention" is particularly vague.
22:47:19 <elliott> fizzie: I should track down Dave Cutler and pry the Intel i860 version of NT from his cold, dead hands.
22:47:30 <b_jonas> (I run an emulated windows 3.11, but that's on an x86_64 machine)
22:47:52 <elliott> fizzie: I wonder how hard it'd be to get NT 4 running on a PPC mac
22:47:55 <Oren_> I have like 300 DOS games
22:47:55 <elliott> I mean, very hard, I'm sure.
22:48:04 <elliott> but it sounds kind of like a small matter of (lots and lots of) programming.
22:48:35 <fizzie> I was browsing through my closets the other day (looking for a UK-compatible power cable I knew I had somewhere), and came across the 13W3-to-4xBNC monitor cable for the Sparc/Indy (they share the same Sun 13W3 display connector), and a 25-pin-serial to 9-pin-serial cable I think I used as a serial console for one of them.
22:48:39 <Oren_> like a moon landing or something
22:49:28 <fizzie> What I'm slightly worried about is that I didn't locate the 13W3-to-13W3 cable that used to connect the Indy to its own SGI-issue monitor. But maybe it's in the basement wrapped around the monitor or something.
22:50:01 <elliott> are you sure the computers are still there? how do you know they haven't fleed?
22:50:10 <Oren_> if your house is like mine then they are still there
22:50:21 <elliott> whoa, I just realised NTFS = NT FS
22:50:39 <elliott> I think that's the second most obvious thing I've ever missed
22:50:41 <Oren_> what does NT stand for?
22:50:59 <Vorpal> * elliott has bricked two PPC macs trying to get unix onto them <-- what? how?
22:51:02 <fizzie> Incidentally, 13W3 is the silliest connector: it's like a regular "D-style" connector, with 10 regular pins, and then 3 tiny tiny coax-style pins.
22:51:04 <elliott> because the i680 was codenamed N10.
22:51:05 <fizzie> http://en.wikipedia.org/wiki/DB13W3#mediaviewer/File:13W3_Stecker.jpg
22:51:08 <elliott> but now it's "New Technology".
22:51:21 <FireFly> elliott: ...what is the most obvious thing you've missed?
22:51:39 <elliott> FireFly: honestly, I've forgotten. but I remember the groans of others when I realised it.
22:51:45 <Vorpal> elliott, surly you could just put in a OS CD yeah?
22:51:55 <Vorpal> Technically bricked I guess
22:52:10 <elliott> maybe I only bricked one of them
22:52:25 <elliott> I managed to get linux-or-was-it-BSD to start to boot on one of them, but then it panicked, and that's it.
22:52:36 <elliott> Windows would be way cooler an achievement, though.
22:52:43 <Vorpal> I booted a live cd on my ibook once
22:52:55 <elliott> Vorpal: they were an old PowerMac and a G3 iMac, respectively
22:53:02 <elliott> the PowerMac was more of a fuss
22:53:11 <fizzie> I put a Debian on my Performa I-can't-remember-the-number.
22:53:13 <elliott> the G3, I just couldn't convince Open Firmware to boot from the CD for whatever reason.
22:53:20 <fizzie> It went in via the MkLinux route.
22:53:31 <Vorpal> elliott, First model ibook (G3) is the one I booted from a live cd
22:53:36 <elliott> fizzie: I think the one I had was a Performa or a Quadra, maybe.
22:53:42 <elliott> right, I tried MkLinux I think.
22:53:43 <Vorpal> Didn't ever go further, due to the 3.2 GB HDD
22:54:00 <Oren_> how many years from now will x86 lose its dominance in PCmarket?
22:54:07 <fizzie> MkLinux is that thing where you have a Mach microkernel, and run a Linux kernel as one of its tasks.
22:54:26 <fizzie> The kind of thing that causes people to make a "kernel on your kernel" jokes.
22:54:28 <b_jonas> Oren_: not very soon, I expect
22:54:54 <elliott> Oren_: more like how many years from now will the PC market dissolve
22:54:58 <Vorpal> FireFly, and 32 MB RAM
22:55:01 <elliott> ARM is the king of the mobile market
22:55:49 <Oren_> but the mobiles are getting bigger and you stick a keyboard onto the side and you have a laptop thingy
22:55:55 <fizzie> Possibly it was just the MkLinux bootloader or something.
22:56:06 <Oren_> I consider that to be a laptop
22:56:28 <elliott> you consider an iPad to be a laptop? ok.
22:56:31 <b_jonas> elliott: no way. there'll always be a PC market.
22:56:49 <b_jonas> if you came with C not being turing complete
22:56:51 <FireFly> I wish I still had the Compaq Armada
22:56:54 <elliott> b_jonas: sure, but it might be a mostly-irrelevant one.
22:56:57 <b_jonas> but still, it is here to stay for a long time
22:56:58 <Oren_> it is a mini laptop with no keyborad
22:57:02 <FireFly> I think I have an old Thinkpad somewhere around here
22:57:05 <elliott> it's already becoming that way for many people.
22:57:10 <Oren_> my laptop has a touchscreen
22:57:15 <Vorpal> FireFly, anyway I didn't want to remove MacOS, I wanted to dual boot
22:57:18 <elliott> our perspectives on this as nerds is hopelessly skewed and it's hopeless to generalise from that.
22:57:27 <Vorpal> FireFly, as in, classic Mac OS 9
22:57:30 <b_jonas> elliott: well, it might become smaller, but still
22:57:50 <FireFly> Oren_: my phone almost has a keyboard
22:57:51 <Vorpal> FireFly, now 3.2 GB is not very much any longer :P
22:58:42 <Oren_> my desktop machine has 3.2 gb of cache I think... something like that
23:00:03 <Oren_> nope off by an order of magnitude, butonly one
23:00:05 <Vorpal> I doubt that, that seems absurd
23:01:00 <Vorpal> Oren_, 3.2 MB seems reasonable, 32 MB seems very large too. But possible I guess. 320 MB I don't believe in
23:01:16 <Vorpal> What sort of beast computer is it
23:01:35 <b_jonas> Vorpal: it depends on which level of cache
23:01:46 <Vorpal> Oren_, so 3 orders of magnitude then?
23:01:57 <elliott> some core 2s had 32 mb I think
23:01:59 <fizzie> Vorpal: Depends on your base. :p
23:02:00 <Oren_> dammit i am bad at math
23:02:21 <Oren_> that's why i have computers to do it for me
23:02:21 <Vorpal> elliott, really? I know core 2 had large caches, but THAT large?
23:03:08 <Vorpal> model name: Intel(R) Core(TM)2 Duo CPU P8400 @ 2.26GHz
23:03:35 <fizzie> Vorpal: There's a thing on Haswell that there's a bit of DRAM that works as L4 cache both for the CPU and the GPU.
23:03:39 <Oren_> best recent processor from intel is Intel® Core™ i7-5960X Processor Extreme Edition (20M Cache, up to 3.50 GHz)
23:03:58 <fizzie> Vorpal: "GT3e [Iris Pro] version with 40 EUs and on-package 128 MB of embedded DRAM (eDRAM), called Crystalwell, is available only in mobile H-SKUs and desktop (BGA-only) R-SKUs. Effectively, this eDRAM is a Level 4 cache; it is shared dynamically between the on-die GPU and CPU, and serving as a victim cache to the CPU's Level 3 cache."
23:04:32 <elliott> okay, maybe 12 was the most. still.
23:04:41 <fizzie> It's kind of debatable if that counts.
23:04:57 <Vorpal> I have seen POWER CPUs though with 512 MB cache iirc
23:05:10 <fizzie> This regular desktop thing reports a cache size of 6 MB.
23:05:47 <fizzie> elliott: I think I mentioned that there's no Linux driver for the Indy's graphics option (XZ graphics)? IRIX would be the low-effort option. (If you can call it that.)
23:06:07 <Vorpal> model name: Intel(R) Core(TM) i5-2500 CPU @ 3.30GHz
23:06:17 <Vorpal> That is a Sandy Bridge
23:06:41 <Vorpal> And this might be off since it is in a Xen domU:
23:06:45 <Vorpal> model name: Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz
23:07:32 <fizzie> elliott: The SS5 has a Sun CG3 graphics card in it, I think, that's very supported. And something fancy in the networking department, I forget exactly what.
23:07:54 <elliott> fizzie: does NetBSD have a driver for the Indy's graphics? :p
23:08:21 <elliott> (does NetBSD have a driver for my toaster?)
23:08:28 <fizzie> elliott: I doubt that. It works with a serial console just fine, but that's perhaps kind of missing the point of a SGI box.
23:09:29 <fizzie> I wonder if I got somehow someone to send me one of Sun's quad Ethernet dealies.
23:09:55 <fizzie> Vorpal: "model name: Intel(R) Xeon(R) CPU E5-2695 v2 @ 2.40GHz" "cache size: 30720 KB"
23:10:24 <fizzie> (The front-end node of our cluster.)
23:10:46 <fizzie> It also has "address sizes: 46 bits physical, 48 bits virtual", which is the most physical bits I've seen so far, at least on Intels.
23:11:10 <Vorpal> fizzie, same on the domU I checked on
23:11:34 <fizzie> Well, same family, presumably, based on the model numbers. (E5-2xxxx)
23:12:30 <Vorpal> address sizes: 36 bits physical, 48 bits virtual is the core i5 above
23:12:38 <elliott> fizzie: 32 TB of RAM sounds pretty doable
23:12:42 <elliott> I guess you'd be limited by slots.
23:12:54 <elliott> do servers with hundreds of gigs of RAM just have ten billion RAM slots?
23:13:17 <Vorpal> elliott, you can get 8 GB modules easily these days, maybe larger
23:13:20 <fizzie> elliott: The cluster's got some (2?) nodes with 1 TB of RAM, but that's it.
23:13:31 <Vorpal> You can get 8 GB laptop modules even
23:13:52 <fizzie> 1024/8 is still 128, though.
23:14:17 -!- Oren_ has quit (Ping timeout: 246 seconds).
23:14:26 <fizzie> I think the particular server model goes up to 2TB in the configu-o-tron.
23:15:00 <fizzie> http://www8.hp.com/us/en/products/proliant-servers/product-detail.html?oid=4142916
23:15:19 <Vorpal> So presumably there are more than 8 GB modules?
23:15:19 <fizzie> "Memory, maximum: 2TB" "Memory slots: 64 DIMM slots"
23:15:28 <Vorpal> So 16 GB per module then
23:15:52 <fizzie> (For the 2 TB option; they might well be using those even if you order it only half-full.)
23:16:36 <fizzie> 32 GB DIMM sounds somehow familiar.
23:22:41 <fizzie> The front-end node also seems to have a load average of 17.06, with one MATLAB running at 100% CPU, and exactly 16 cryptically named processes running at 25% CPU. I think someone's running their actual jobs accidentally on the front-end again.
23:22:57 <fizzie> "We ask you to refrain from running multi-GB, many-core applications of the frontend."
23:29:58 -!- zzo38 has joined.
23:31:50 -!- Dulnes has joined.
23:33:16 <Dulnes> Wat oh ive set it to disconnect
23:42:09 -!- dts has quit (Ping timeout: 264 seconds).
23:44:18 -!- Frooxius has quit (Quit: *bubbles away*).
23:46:45 -!- dts has joined.
23:48:29 -!- copumpkin has joined.
23:49:49 <copumpkin> hi, sorry about yesterday, I won't stay long
23:50:31 -!- Frooxius has joined.
23:51:21 -!- nooga has quit (Ping timeout: 256 seconds).
23:58:34 -!- copumpkin has quit (Ping timeout: 240 seconds).