00:06:33 -!- Phantom_Hoover has quit (Remote host closed the connection).
00:15:28 -!- uplime has joined.
00:25:20 <arseniiv> ha ha very funny, have they actually proven that that encoding is invertible?.. meh
00:26:07 <arseniiv> I almost want to write a comment there
00:39:48 <lambdabot> KOAK 082353Z 13005KT 10SM OVC033 14/11 A2994 RMK AO2 SLP140 60000 T01440111 10150 20128 56018
00:43:19 <int-e> arseniiv: there isn't much to prove?
00:43:59 <arseniiv> int-e: I think it’s outright invalid but I’m lazy to investigate
00:44:16 <arseniiv> the issue is with replacing each 0 with 69
00:45:07 <arseniiv> maybe two numbers could map to the same thing in the end
00:45:13 -!- sleepnap has left.
00:45:22 <int-e> Oh I trusted the description that "the 0" is only one.
00:46:25 <shachaf> https://www.ioccc.org/2018/mills/hint.html
00:47:02 <arseniiv> not quite—right before the example they write “All zeros in the code is now a 69.”
00:47:17 <HackEso> ioccclist is update notification for when a new year of the International Obfuscated C Code Contest is announced, or the winners for a year is announced, or the source codes of winners are released. http://www.ioccc.org/#news
00:47:17 <int-e> arseniiv: Since it's restricted in ASCII you won't find any collisions.
00:47:32 <arseniiv> not that it’s a bit inaccurate in its own right…
00:48:04 <int-e> provably since 0x69 is not divisible by 69 (decimal)
00:48:54 <arseniiv> well anyway I won’t be the only one to dislike languages that are print-only or, for the other cases, ill-described :P
00:49:57 <arseniiv> not that I dislike under-Turing-complete formalisms, but that these have been used in much more interesting esolangs than of this sort
00:51:11 <arseniiv> but I don’t know policies of the wiki and anyway I hope this fellow would think up something interesting after all
00:52:15 <arseniiv> hope is all we have when luck is proven nonexistent :D
00:52:32 <int-e> but you're right that as an encoding of natural numbers, this is not injective. For example, 1340 and 60 map to the same string: 69x1692C (omitting the /-/31169) (is this the smallest example?)
00:55:41 <int-e> err, 69x1692d -- I forgot the final shift of letters and the preference for lower case
01:13:31 -!- Sgeo_ has joined.
01:16:56 <arseniiv> seems indeed 60 ~ 1340 is the smallest pair
01:17:10 -!- Sgeo__ has quit (Ping timeout: 250 seconds).
01:18:26 -!- xkapastel has quit (Quit: Connection closed for inactivity).
01:19:16 <arseniiv> though it’s quite impolite to use just ASCII output these days, so even larger pairs should really matter :P
01:56:25 -!- danieljabailey has quit (Ping timeout: 268 seconds).
01:59:10 -!- imode has joined.
02:07:45 -!- Sgeo_ has quit (Read error: Connection reset by peer).
02:08:11 -!- Sgeo_ has joined.
03:22:06 -!- arseniiv has quit (Ping timeout: 260 seconds).
03:34:17 -!- Essadon has quit (Quit: Qutting).
03:53:23 -!- Lord_of_Life has quit (Ping timeout: 245 seconds).
03:55:51 -!- Lord_of_Life has joined.
03:59:45 -!- Sgeo_ has quit (Read error: Connection reset by peer).
04:00:09 -!- Sgeo_ has joined.
04:20:10 -!- oerjan has joined.
04:23:10 <oerjan> @tell ais523 You're right, with that assumption 1,0,0 (mod 4) works, and with a much simpler argument than I had for mod 5.
04:24:53 <oerjan> Googling today's Freefall claim, it looks to me like future Winston is passing on currently debunked science.
04:26:10 <oerjan> (Inert DNA simply does not last a million year unharmed, *and* the bacteria are suspiciously genetically close to modern ones.)
04:28:00 -!- FreeFull has quit.
04:53:24 -!- imode has quit (Ping timeout: 250 seconds).
04:54:49 -!- yaewa has joined.
05:02:09 -!- moei has quit (*.net *.split).
05:02:09 -!- zzo38 has quit (*.net *.split).
05:23:23 <esowiki> [[]] M https://esolangs.org/w/index.php?diff=59020&oldid=58913 * Salpynx * (+3) /* Examples */
05:31:01 -!- MDude has quit (Ping timeout: 268 seconds).
05:57:00 * oerjan eats the last nutella ball
06:25:38 -!- imode has joined.
06:32:41 -!- Sgeo_ has quit (Ping timeout: 268 seconds).
06:52:38 <HackEso> Your omnidryad saddle principal ideal golfing toe-obsessed "Darth Ook" oerjan the shifty evil grinch is a punctual expert in minor compaction. Also a Groadep who minces Roald Dahl. He could never remember the word "amortized" so he put it here for convenience. His ark-nemesis is Noah. He twice punned without noticing it.
06:52:53 <shachaf> Hmm, it already says you're an evil grinch, I guess there's no upgrade from that.
06:56:07 -!- uplime has quit (Quit: WeeChat 2.2).
07:11:17 -!- Sgeo has joined.
07:33:13 <esowiki> [[User talk:Graue]] https://esolangs.org/w/index.php?diff=59021&oldid=30275 * Qpliu * (+713)
08:10:50 -!- AnotherTest has joined.
08:22:07 -!- xkapastel has joined.
09:04:50 -!- nfd has joined.
09:06:31 <b_jonas> `bobadventureslist http://bobadventures.comicgenesis.com/d/20190108.html
09:06:32 <HackEso> bobadventureslist http://bobadventures.comicgenesis.com/d/20190108.html: b_jonas
09:07:10 -!- nfd9001 has quit (Ping timeout: 244 seconds).
09:07:24 -!- imode has quit (Ping timeout: 272 seconds).
09:09:05 <oerjan> b_jonas: it looks clear that we can keep the cells non-negative in ais523's simplified two-loop BF construction
09:15:49 <b_jonas> oerjan: and, just to be clear, it still uses only a bounded number of cells, right?
09:17:05 <oerjan> Sure, the size is a function of the number of waterclocks converted
09:18:11 <oerjan> although i don't know the minimum number of that
09:26:36 -!- oerjan has quit (Quit: Nite).
09:57:18 -!- b_jonas has quit (Quit: leaving).
10:04:04 -!- Sgeo has quit (Read error: Connection reset by peer).
10:04:31 -!- Sgeo has joined.
10:14:33 <esowiki> [[]] https://esolangs.org/w/index.php?diff=59022&oldid=59020 * Salpynx * (+69) /* External resources */ WIP interpreter
11:00:48 -!- Sgeo has quit (Read error: Connection reset by peer).
11:01:17 -!- Sgeo has joined.
12:54:50 -!- Essadon has joined.
13:06:53 -!- MDude has joined.
13:15:31 <esowiki> [[Brainfuck]] https://esolangs.org/w/index.php?diff=59023&oldid=58997 * B jonas * (+63) /* Computational class */
13:16:07 <esowiki> [[Brainfuck]] https://esolangs.org/w/index.php?diff=59024&oldid=59023 * B jonas * (-63) rv self
13:39:25 -!- arseniiv has joined.
14:01:28 -!- danieljabailey has joined.
14:16:42 -!- AnotherTest has quit (Ping timeout: 252 seconds).
14:51:44 -!- arseniiv has quit (Ping timeout: 246 seconds).
14:59:45 -!- arseniiv has joined.
15:02:40 -!- Sgeo_ has joined.
15:05:52 -!- Sgeo has quit (Ping timeout: 246 seconds).
15:08:29 -!- uplime has joined.
15:22:41 -!- AnotherTest has joined.
15:27:06 -!- AnotherTest has quit (Ping timeout: 252 seconds).
15:29:42 -!- yaewa has quit (Quit: Leaving...).
15:37:36 -!- LKoen has joined.
15:44:17 -!- Sgeo has joined.
15:46:02 -!- Sgeo_ has quit (Ping timeout: 244 seconds).
15:46:51 -!- Sgeo_ has joined.
15:47:55 -!- uplime has quit (Quit: brb).
15:48:44 -!- wob_jonas has joined.
15:48:53 <HackEso> olist 1151: shachaf oerjan Sgeo FireFly boily nortti b_jonas
15:48:57 -!- wob_jonas has quit (Client Quit).
15:50:00 -!- uplime has joined.
15:50:41 -!- Sgeo has quit (Ping timeout: 268 seconds).
15:51:51 -!- Lord_of_Life_ has joined.
15:55:05 -!- Lord_of_Life has quit (Ping timeout: 246 seconds).
15:55:06 -!- Lord_of_Life_ has changed nick to Lord_of_Life.
16:13:06 -!- arseniiv has quit (Ping timeout: 252 seconds).
16:22:12 -!- arseniiv has joined.
16:26:07 -!- AnotherTest has joined.
16:37:08 -!- AnotherTest has quit (Ping timeout: 250 seconds).
16:59:21 -!- AnotherTest has joined.
17:05:34 -!- AnotherTest has quit (Ping timeout: 272 seconds).
17:18:30 -!- arseniiv_ has joined.
17:20:23 -!- imode has joined.
17:20:47 -!- arseniiv has quit (Ping timeout: 252 seconds).
17:57:14 -!- AnotherTest has joined.
18:01:17 <imode> what memory topologies lead to automata that aren't TC?
18:01:54 -!- AnotherTest has quit (Ping timeout: 252 seconds).
18:02:45 <imode> a single one-way tape is one, off the top of my head.
18:03:21 -!- LKoen has quit (Remote host closed the connection).
18:14:09 <imode> for cellular automaton, any graph that isn't infinite.
18:18:56 -!- LKoen has joined.
18:45:14 <imode> . o O (though, I guess anything that isn't infinite can't be TC...)
18:47:27 <Taneb> imode: a single stack?
18:48:35 <imode> yup, that'd be one.
19:13:26 <imode> what'd be nice is to find the "general rule", so that you could recognize, for example, patterns of access over a particular graph and say "yup, that can't possibly be turing complete".
19:16:21 <imode> "what topologies can't be used as memory spaces for automata to be turing complete".
19:17:07 <imode> another trivial one is the graph with finite vertices but infinite edges (unless you can change the edge labels).
19:17:58 -!- LKoen has quit (Remote host closed the connection).
19:30:31 <rain1> I don't understand
19:30:58 <rain1> is this like how regular languages are the finite automata?
19:46:18 <imode> it's related to the memory spaces for various automata. turing machines, for example, use an unbounded tape that's able to contain symbols. a cellular automaton like the game of life, for example, relies on the idea of an unbounded 2D grid.
19:47:32 -!- AnotherTest has joined.
19:50:13 -!- Phantom_Hoover has joined.
19:50:13 -!- Phantom_Hoover has quit (Changing host).
19:50:13 -!- Phantom_Hoover has joined.
19:51:22 <imode> these are examples of spaces. they have some topology. their access patterns are "walks" along that topology. their topology is usually discrete/digital.
19:52:08 <imode> my question is, what spaces can't be used to simulate a turing machine.
19:53:06 <imode> it kind of boils down to "what spaces can you not embed a tape".
19:58:12 -!- ais523 has joined.
19:59:43 -!- AnotherTest has quit (Ping timeout: 244 seconds).
20:00:44 <ais523> imode: it seems hard to characterise, e.g. imagine a directed graph formed out of connected loops, O→O→O→O…, that's TC if and only if there's no maximum limit on how large the largest loop can be
20:00:45 -!- ais523 has quit (Client Quit).
20:00:57 -!- ais523 has joined.
20:01:09 <imode> are all the loops the same size?
20:01:28 <ais523> not necessarily; if they are it's sub-TC
20:01:37 <imode> that's what I figured.
20:02:11 <ais523> hmm, even more confusingly: suppose you have a root node that connects to a loop in one direction, and a copy of the same graph but with bigger loops in the other direction
20:02:27 <imode> could you visualize that?
20:02:38 <ais523> this is /also/ sub-TC even though, for any halting computation, you can find a tape on which it could be run
20:02:49 <ais523> because you don't have enough memory to reliably find the tape in question
20:04:36 <ais523> something like ⋌⋌⋌⋌⋌⋌⋌⋌⋌⋌ except imagine that everything is directed (to the right), you kern it more tightly, and there's an increasingly large loop connected to each of the top connections
20:04:51 <ais523> sorry, my extended-ASCII art isn't very good due to a lack of appropriate characters in Unicode
20:05:20 <imode> is it possible to even construct an algorithm to recognize graphs that we can't embed a tape in?
20:06:03 <ais523> given that we're talking about infinite graphs, just defining the I/O format for that would be hard
20:06:19 <imode> I'd be content with just LBAs.
20:06:37 <imode> "infinite, but to a certain point".
20:06:53 <imode> that's suceptible to induction.
20:07:05 <ais523> oh, I think you can run an LBA on that graph (assuming the loop sizes increase faster than linear) so long as you get to read the input twice
20:07:44 <ais523> it can be done with linearly-increasing loop sizes too if the constant factor is sufficiently large
20:08:05 <ais523> or, actually, no, even if the constant factor is small, you just need to increase the number of states
20:09:30 <ais523> now I'm thinking of another interesting graph: it's a directed graph with infinitely many connections, and each vertex has a (directed) path to every other, but the connections are chosen randomly
20:09:51 <imode> subset of the infinite complete directed graph?
20:10:05 <ais523> yes, but not just any subset
20:10:22 <ais523> needs to be a single connected component and infinitely large
20:10:30 <imode> you could actually simulate this with an infinite bitvector.
20:10:34 <ais523> I have a suspicion that this is TC with probability 1, but not absolutely TC
20:10:47 <imode> starting with the complete graph, enumerate all edges.
20:11:00 <imode> if there exists an edge, put a 0. if there doesn't, put a 1.
20:11:12 <imode> you can "carve" the particular graph out of the complete graph.
20:11:23 <ais523> right, you can define formats that describe a specific graph of this nature
20:11:45 <ais523> programming in this would be weird, the main issue would be trying to find the tape again whenever you expanded memory
20:12:15 <ais523> (assuming a deterministic or probabilistic language, a true-nondeterministic language would have no trouble)
20:12:30 <imode> the reason I ask is that I'm probing the possible space of physically realizable computations. universes with different topologies admit different tape embeddings of TMs.
20:12:54 <imode> under different physical laws, how would computation be affected.
20:13:15 <imode> and what's a good general "model".
20:13:23 <imode> something that can be ported from space to space.
20:17:03 <ais523> anything that's usable like a TM tape needs to contain a sufficiently large loop somewhere
20:17:10 <ais523> (necessary condition, not sufficient)
20:17:40 <ais523> otherwise you only have finite memory
20:17:42 <imode> TM tapes aren't looped.
20:17:53 <ais523> they are, the loop goes out to the right then back to the left
20:18:00 <ais523> I'm assuming you can visit the same cell multiple times
20:18:26 <imode> the traditional definition of a TM, unless I'm mistaken, involves a tape that extends infinitely to the right.
20:18:41 -!- AnotherTest has joined.
20:18:45 <ais523> right, but the reason it's TC is that you can send the tape head to the right, then back to the left to reread data it's previously written
20:18:47 <Taneb> imode: the cells are connected in a directed graph which has cycles
20:18:52 <imode> ohhh right right right.
20:18:56 <imode> sorry, got mixed up.
20:19:25 <ais523> actually I don't think my O→O→O→O thing is IO-complete unless it can read data infinitely many times
20:19:25 <imode> yeah, so something like "tileable loops of sufficient size" would probably be a good generalization of the topology of a tape.
20:19:53 <ais523> as if you discover you don't have enough memory, you have to follow a → and lose everything but the TM's current state (which has finitely many possibilities)
20:20:06 <ais523> if you don't take input it's fine, just restart from scratch, if you do you have nowhere to store it though
20:21:27 <imode> that's actually interesting. "bubbles" of state.
20:21:56 -!- uplime has quit (Quit: brb - please send all requests, no matter how silly they are, to blkshp asap).
20:22:22 <esowiki> [[Nellephant]] https://esolangs.org/w/index.php?diff=59025&oldid=58994 * Ais523 * (+359) /* The preprocessor */ macro arguments
20:24:07 <ais523> imode: something that I've noticed in several languages is connections that the TM can only cross in one state
20:24:40 <ais523> like, you have a loop of tape, but one particular point on the tape causes the TM to forget what it's doing as it crosses it
20:25:37 <ais523> An Odd Rewriting System is like that, the "tape head" forgets what it's doing when it goes from the rightmost end of the tape back to the leftmost end, it made the TCness proof somewhat difficult (although AORS lets you add extra elements "inside" the loop to get its infinite memory, and that operation seems to help in other ways too)
20:27:02 -!- uplime has joined.
20:30:21 <imode> interesting! I'll brb.
20:34:36 -!- AnotherTest has quit (Ping timeout: 260 seconds).
20:41:06 -!- imode has quit (Ping timeout: 250 seconds).
22:08:30 -!- LKoen has joined.
22:12:10 -!- b_jonas has joined.
22:13:31 <b_jonas> ais523: "suppose you have a root node that connects to a loop in one direction, and a copy of the same graph but with bigger loops in the other direction" =>
22:13:51 <b_jonas> I don't see why that wouldn't be Turing-complete. can't you just copy the tape to the next tape when you run out of it?
22:14:03 <ais523> no, this is unidirectional
22:14:10 <ais523> so once you've gone to a loop you can't go back
22:14:48 <b_jonas> you can only move in one direction
22:18:16 <b_jonas> unrelated: anyone wants to recommend speedruns on AGDQ2019 that are worth to watch?
22:18:39 <b_jonas> Oh! There's an "Another world" run. I must watch that.
22:19:03 <rain1> the sonic run was fun
22:21:19 <b_jonas> which sonic? I'm guessing there's more than one sonic run on a GDQ
22:21:27 <pikhq> There was a whole block.
22:39:01 -!- Sgeo_ has quit (Read error: Connection reset by peer).
22:39:47 <b_jonas> there's an _original_ sonic game? wow
22:39:55 -!- imode has joined.
22:40:32 -!- imode has quit (Client Quit).
22:48:21 <rain1> I don't see why you're insulting me like that
22:54:11 -!- imode has joined.
23:16:22 -!- Sgeo has joined.
23:19:43 -!- b_jonas has quit (Quit: leaving).
23:48:49 <esowiki> [[NEGATOR]] https://esolangs.org/w/index.php?diff=59026&oldid=59016 * Areallycoolusername * (+639)