00:06:07 <zzo38> I made a list of game rules in http://esolangs.org/wiki/User:Zzo38/Game_rules but does not include all of the games, yet.
00:13:59 <wryl> So I've been experimenting with an augmentation to FRACTRAN/Bag's rules.
00:15:31 <salpynx> korvo: it seems that BBB(n) winners must be non-terminating, and have no halt state (BBB(4) winner meets this) which effectively _adds_ another useful state to the machine. It's not going to be exactly equivalent to an n+1 machine... it's likely to be better, because a halting n+1 machine has at least one out of (n * symbol) combinations leading to Halt, whereas the BBB(n) can use the full (n * symbol) combinations to do what it needs to, so long as it
00:15:32 <salpynx> can create an arbitrary 'quasihalt' loop with an (n-1) non-terminating TM, form the quasihalt tape state.
00:15:52 <salpynx> ... so you're replacing a halt state with a n-1 state non-terminating TM which happens to overlap or co-exist with the useful machine you are designing
00:16:17 -!- craigo has quit (Ping timeout: 248 seconds).
00:16:36 <korvo> salpynx: Yeah, that seems to be the intuition that Ligocki is building. I admit that I don't have a good imagination for these.
00:17:31 <int-e> salpynx: eventually the oracle power will kick in and blow BB(n) out of the water
00:17:36 <zzo38> What augmentation to FRACTRAN/Bag's rules is it?
00:18:21 <wryl> One where instead of decrementing and incrementing by 1, you decrement and increment by the `min` of all of the registers mentioned on the LHS.
00:18:29 <int-e> like there'll be a /relatively/ small machine that simulates all TMs with 10^100 states in parallel and beeps whenever one of them halts, so BBB(<relatively small number>) > BB(10^100).
00:18:57 <wryl> And one where the LHS can't feature duplicates.
00:20:08 <wryl> It makes moving data around quite fast.
00:20:35 <b_jonas> salpynx: no, you don't make the halting state beep, you make the state that occurs the most frequently before the halt, and since there are very few states and BB is large, that's basically as many beeps as the BB. that proves BBB can't be significantly smaller than BB.
00:21:16 <b_jonas> that BBB is be much larger even for such low number of states isn't obvious to me
00:21:17 <int-e> b_jonas: it counts steps, not beeps
00:21:48 <int-e> I think the definition is carefully crafted to make BB(n) >= BBB(n) trivial.
00:22:11 <wryl> The caveat being that any catalyst-like rules that reduce the `min` to 1 reduce it to FRACTRAN/Bag semantics.
00:22:17 <b_jonas> int-e: probably doesn't matter much, but sure
00:23:32 <salpynx> b_jonas: as int-e said "it counts steps, not beeps", i was thinking it counted beeps first, but BB gets to output 1s on more that one state, BBB only gets 1 beep state, so by that measure BB has to get more 'outputs' in
00:23:47 <int-e> err that inequality is the wrong way around
00:24:32 <salpynx> ... then I got confused whether BB is 'producing the most output' / 'writing the most 1s to the tape' / 'running the longest' ... i think all of those end up be equivalent in the various formulation
00:25:02 <int-e> salpynx: all these notions exist but I think people focus on the number of steps until it halts
00:25:35 <b_jonas> though I wonder if BBB is much larger than BB, in the sense of whether there's no constant C such that BBB(n)<=BB(C*n) for large enough n. like, is BBB really more powerful, or does it just get some advantage for large n?
00:25:50 <b_jonas> s/advantage for large n/advantage for small n/
00:25:55 <int-e> b_jonas: did you see my comment about BB(10^100)?
00:26:49 <b_jonas> so that means BBB is much bigger
00:27:46 <salpynx> int-e: number of steps is the easiest to think about. I agree with your statement about the inequality crafted to make the inequality trivial. It does require understanding that it's a single beep marking the step index, not a beep count
00:28:43 <b_jonas> that shouldn't be different for large n though
00:29:56 <b_jonas> except of course if you want beep count then the operators will get annoyed by all the beeps and disconnect the alarm
00:31:09 <b_jonas> well, if you can still find operators after BBB days in first place
00:32:01 <korvo> Like, it's important to understand how this generalizes, or at least how it's made isomorphism-invariant. The machine only has to go through a "beep state" because we're assuming it's got an underlying FSM.
00:32:50 <korvo> If we forget that the machine has states, and just think of it in terms of observables, then the connection to halting is clear: we're waiting for the machine to satisfy a particular observational requirement.
00:33:24 <korvo> That observable might only manifest as a composite of machine actions or machine states. A classic example is old-school tape drives that walked across server rooms when given a sequence of head-moving operations.
00:34:04 <korvo> But it *must* generalize waiting for the machine to halt, because we can presumably observe whether the machine is halted yet.
00:34:38 <salpynx> this is really interesting, but i think it has more to do with halt definitions rather than IO, and something about it feels like state-count accounting tricks on n, like there will be an exact equivalence trade-off, with no rules broken or extra power conferred, but it certainly doesn't seem trivial to unpack.
00:35:48 <korvo> Yeah, it feels weird that something could still be hard even with an oracle for Halting, right? I struggled with this too.
00:36:17 <b_jonas> hehe. for a turing machine you can observe haltin easily, but I think ais523 managed to define machines with much harder to observe halting conditions
00:38:43 <korvo> DNA computers are a fun example. AIUI we have autocatalytic DNA fragments, so we should be able to do universal computation somehow, but DNA computers don't ever halt, they just (possibly) reach equilibrium.
00:39:01 <salpynx> i know there have been esolang discussions where a tight loop was an uncontroversial alternative to a halt. This seems to be suggesting that unprovably non-halting loops can be used as halt too... which seems like a thing you can say, but, yeah, harder to work with in practice
00:39:31 <b_jonas> and of course you have to remember the context that BB(1000) is already larger than I can imagine
00:39:45 <korvo> Waterfall computers are another fun example. Technically, if you label every atom of water going over Niagara Falls in one second, you can compute by using the permutation of atoms as a reversible computer; but good luck actually tagging quadrillions of molecules.
00:41:00 <korvo> salpynx: Note that, for TMs, halting is definitely an observable, and that's what matters for TCness as it's originally defined. If you can emulate a TM in a system, and whether the emulator has halted is observable, then it doesn't matter if we can't measure the whole system at once.
00:41:14 <korvo> That might be super-obvious to you, sorry.
00:41:46 <salpynx> ugh, just had a though abotu machines that maximise the time for their first beep.. and then again for their second beep... isn't that just two BB's back to back, maybe sharing some states?
00:42:06 <b_jonas> korvo: by quadrillion you mean just 10**15 right? because 10**24 might be more than how many molecules in Niagara Falls
00:42:06 <int-e> Oh also instead of beeping whenever a TM halts, we can instead take the number of steps N so far, and enumerate N-state TMs to simulate (so now there's two separate batches of simulations, and the second one is reset whenever we find a larger N). So we can get BB(BB(10^100)) still with "relatively few" states. (I imagine tens of thousands if we do it with human-understandable means)
00:42:40 <korvo> Yep, you're on the same track as Aaronson and Ligocki (separately), who build nests of equations relating the two different degrees of computation. This sort of thing also shows up in algorithmic information theory.
00:42:40 <int-e> so as usual these constructions start stacking like crazy
00:43:06 <int-e> korvo: yeah I didn't think this would be a novel idea :)
00:43:10 <korvo> b_jonas: Yeah, the smaller amount.
00:44:44 <korvo> int-e: Hah, yeah. I've read the article by...you? Tromp? doing this with BLC. Same idea: the monoid of composition has recursive constraints on its presentation and we can use those to compute the cardinality of (slices of) the presentation.
00:45:54 <int-e> korvo: the term "oracle power" is even in the blog post... I'm just making this a bit more concrete for myself.
00:46:14 <b_jonas> ok never mind, there is 10**24 molecules too, I just can't count
00:47:14 <korvo> int-e: I keep thinking of the Erdös Book, or 0♯, or other indices of proofs. Stuff at the level of Halting is questions about maths. Stuff at the level of Beeping is questions about The Book.
00:47:40 <int-e> korvo: probably Tromp
00:49:09 -!- nitrix has changed nick to nitrix-or-treat.
00:49:37 <salpynx> i'm laughing just at the idea of not being able to count to 10**24 being a problem, that just struck me as very funny
00:50:28 * int-e just programs in BLC on occasion. Well, there is https://www.isa-afp.org/entries/Goodstein_Lambda.html which qualifies as academic and was motivated by BLC.
00:51:35 -!- amby has quit (Quit: so long suckers! i rev up my motorcylce and create a huge cloud of smoke. when the cloud dissipates im lying completely dead on the pavement).
00:54:13 <ais523> <b_jonas> hehe. for a turing machine you can observe haltin easily, but I think ais523 managed to define machines with much harder to observe halting conditions ← for the 2-state 3-symbol turing machine it's actually mostly the initial condition that's the problem, I think it's possible to design an initial condition so that the halt state becomes "goes to the left of the starting location" (although the halt state in my proof was more complicated)
00:55:53 <ais523> I think the attempt to calculate BB(3,3) has now encountered a machine that people are pretty sure halt, but are not yet sure how to simulate it all the way to its halt state
00:56:44 <esolangs> [[Translated !]] https://esolangs.org/w/index.php?diff=143891&oldid=143286 * None1 * (+1682)
00:59:00 <esolangs> [[Translated !/None1]] N https://esolangs.org/w/index.php?oldid=143892 * None1 * (+4208) Created page with "1. Take that sh*t <pre> C . , . 12299 ? . . , 10 . 2. . . ..."
00:59:23 <esolangs> [[Translated !/None1]] https://esolangs.org/w/index.php?diff=143893&oldid=143892 * None1 * (+39)
01:16:38 <salpynx> thinking in terms of emulators is interesing. It seems trivial to say any halting TM can be emulated by infinitely many non-halting TMs that do something else after the emulation has completed. What does that mean though? I don't think you can make claims about state counts, because it's also trivial to create examples where computations can be performed by fewer states (if you start with an inefficient state machine to begin with).
01:17:27 <salpynx> for BBB, why can't there be more future beeps, if you can define an observable halting condition like: consider the emulation halted if there are no beeps for m steps, or when there are two consecutive beeps, or some other fixed or relative period of silence from some pre-beeped point
01:23:15 <salpynx> What I joined earlier to say was about bfjoust, but got swiped by bf, IO, and BBB
01:23:26 <salpynx> I have another optimised bfjoust warrior, descended from the same parent as Lymia.nyuroki2 (#2 on the hill). At 2k, it's 1/4 the size of Lymia.nyuroki2, and 1/3 the size of their common ancestor, nyuroki-esoteric. It gets 1st place with 22.21 points, and a Markov score of 59.82.
01:24:02 <salpynx> It's ridiculously manually optimised (maybe fragile?), but the basic algorithm is that of nyuroki-esoteric. nyuroki2 has a slightly more complex version, but it's the same idea as Lymia's original of 8x rushes, and some fiddly magic numbers here and there.
01:24:31 <salpynx> I'm not submitting it to the hill because I need to figure out how to apply the MIT license of the original written in a custom JoustExt script / scala / bfjoust, and mine which is bash scripted bfjoust. I'm treating warriors more like discovering integers or algorithms with interesting properties. Maybe I'm missing something about the required creativity for algorithms made from 8 symbols, or etiquette of re-using code? I do want to attribute prior
01:26:34 <salpynx> I'm using the same basic algorithm of nyuroki, and tweaked some of the parameters (setup minimal defense, then 8 potential rushes, one with params for short tapes), but have significantly increased the Kolmogorov complexity of the rush. Common ancestor nyuroki has only 1 rush type, current hill nyuroki2 has 3 kinds, my versions uses 2 distinct rushes.
01:34:05 -!- salpynx has quit (Remote host closed the connection).
02:05:04 -!- op_4 has quit (Remote host closed the connection).
02:05:34 -!- op_4 has joined.
05:26:28 -!- ais523 has quit (Quit: quit).
05:56:20 <esolangs> [[Deadfish/Implementations (nonalphabetic and A-L)]] https://esolangs.org/w/index.php?diff=143894&oldid=143849 * PrySigneToFry * (+728)
06:43:03 <Lymia> There's still life in BFJoust.
06:57:31 -!- tromp has joined.
07:13:31 <Lymia> (will aim this at salpynx later, but dropping this here before I forget)
07:15:51 <Lymia> So... nyuroki2 was kinda always a somewhat poorly tuned version of an at-the-time-new (?) strategy group that plays unfair in a lot of small ways. Break the rule of 9 severely, intentionally use decoy setup clashes, straight-up skip a flag without confirming for sure it's not the flag, etc.
07:21:36 <Lymia> Its last closest relative (strategy-wise) was margins3, which tries to win very short tapes and tie longer ones. Nyuroki overall tries to take a more extreme version of this strategy, winning very short and long tapes. Against older bots coded before it, it usually ends up outright losing (generally) tapes from about 11-14 and winning everything else, leading to a ridiculously winning overall score against the entire hill.
07:23:15 <Lymia> It's still got the best W/L record overall (score ladder), and... its dominance there actually kinda distorts the markov ladder. Since, defeating Nyuroki at the cost of several of the weaker bots that Nyuroki still wins you the ladder overall.
07:26:31 <Lymia> Feel free to post variants of it, just... I think it'd end up amounting to hill tuning against to beat the new bots. It'll probably take a few bots in the general strategy group tuned differently for the ladder to "stabilize" again and so that new bots need to beat the whole general strategy "solidly" instead of just taking advantage of flaws/bugs in Nyuroki to get high on the ladder. I'd encourge new similar bots, but also encourge them to be
07:26:31 <Lymia> completely new code based on the same principles.
07:32:35 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…).
07:37:25 -!- Lord_of_Life has quit (Ping timeout: 248 seconds).
07:43:22 -!- Lord_of_Life has joined.
08:13:00 -!- tromp has joined.
08:20:27 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143895&oldid=143880 * Ractangle * (+22) /* Syntax */
08:22:39 <b_jonas> Lymia: you may want to copy some of that to https://esolangs.org/wiki/BF_Joust_champions#2017
08:28:21 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143896&oldid=143895 * Ractangle * (+154) /* MarkupL syntax */
08:29:34 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143897&oldid=143896 * Ractangle * (+20) /* Hello, world */
08:31:50 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143898&oldid=143897 * Ractangle * (+30) /* Hello, world */
08:32:22 <Lymia> It's kinda hard to describe succinctly because it does a lot of small things rather than any one big thing.
08:32:27 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143899&oldid=143898 * Ractangle * (+0) /* Hello, world */
08:32:45 -!- Sgeo has quit (Read error: Connection reset by peer).
08:33:56 <b_jonas> I don't think it needs to be more succint than what you said above. I think the information on how it affected the ladder can fit there.
08:34:53 <wWwwW> i had an esolang idea
08:35:01 <wWwwW> and i want feedback
08:35:16 <b_jonas> or maybe ask ais523, since I'm not really a bfjoust player
08:36:10 <wWwwW> = which is a daller step
08:36:13 <Lymia> Yeah, I can update the champion thing.
08:36:14 <wWwwW> there is a pointer
08:36:28 <wWwwW> it climbs up the nearest ladder(right prorioty)
08:36:28 <Lymia> ... if it's alive, may actually implement my latest idea for an evolver.
08:36:35 <Lymia> And hope it works this time.
08:36:45 <wWwwW> if its at bottom of ladder goes
08:36:50 <wWwwW> up and other way around
08:36:58 <wWwwW> if its somewhere in the middle in between
08:37:25 <wWwwW> goes oppistive of the last ladder direction
08:39:20 <b_jonas> Lymia: fizzie revived zemhill, so it should be alive
08:40:21 <b_jonas> perlbot call ztest bad (>)*9((+.)*180>(+)*100(+.+..)*30>)*20
08:40:24 <perlbot> b_jonas: {"result":"ok","message":"web.b_jonas-bad: points -28.95, score 4.74, rank 47/47\n"}
08:42:35 <b_jonas> a few people were submitting programs too
08:47:01 <b_jonas> perlbot has been here for a year or two, and it's actually rather old outside this channel, it's a new instance of some parts of buubot3, where buubot3 is a very old bot that I contributed into and is a complete rewrite of buubot2.
08:47:02 <perlbot> b_jonas: Stored has been here for a year or two, and it's actually rather old outside this channel, it's a new instance of some parts of buubot3, where buubot3 is a very old bot that I contributed into and is a complete rewrite of buubot2.
08:47:30 <b_jonas> perlbot search has been there
08:47:30 <perlbot> b_jonas: [<*:##NULL> macro true facts is `eval use Acme::ConspiracyTheory::Random -all; bad_punctuation theory!] [<*:##NULL> has a lot of the same features, but pork has been around for, what, 20 years?] [<*:##NULL> has been here for a year or two, and it's actually rather old outside this channel, it's a new instance of some parts of buubot3, where buubot3... [Output truncated. http://perl.bot/p/6vzai7 ]
08:47:37 <b_jonas> perlbot search has been there for a year or two
08:47:38 <perlbot> b_jonas: [<*:##NULL> has been here for a year or two, and it's actually rather old outside this channel, it's a new instance of some parts of buubot3, where buubot3 is a very old bot that I contributed into and is a complete rewrite of buubot2. ] [<*:##NULL> macro true facts is `eval use Acme::ConspiracyTheory::Random -all; bad_punctuation theory!] [<*:##NU... [Output truncated. http://perl.bot/p/pclksz ]
08:47:55 <b_jonas> drat, how do I make it tell what the copula in that is
08:48:08 <b_jonas> perlbot fact has been here for a year or two, and it's actually rather old outside this channel, it's a new instance of some parts of buubot3, where buubot3
08:48:08 <perlbot> b_jonas: a very old bot that I contributed into and is a complete rewrite of buubot2.
08:48:14 <b_jonas> perlbot forget has been here for a year or two, and it's actually rather old outside this channel, it's a new instance of some parts of buubot3, where buubot3
08:48:14 <perlbot> b_jonas: Forgot has been here for a year or two, and it's actually rather old outside this channel, it's a new instance of some parts of buubot3, where buubot3
08:48:29 <b_jonas> perlbot search has been there for a year or two
08:48:29 <perlbot> b_jonas: [<*:##NULL> macro true facts is `eval use Acme::ConspiracyTheory::Random -all; bad_punctuation theory!] [<*:##NULL> itself is nominally opensource, but it's got a history that would make any lawyers upset if you tried to get a proper license for it (3 main authors, me buu and a user by the name of b_jonas). buu disappeared for a while but is now b... [Output truncated. http://perl.bot/p/3vcap4 ]
08:48:44 <b_jonas> ah, that other one looks accidental too
08:48:48 <perlbot> b_jonas: nominally opensource, but it's got a history that would make any lawyers upset if you tried to get a proper license for it (3 main authors, me buu and a user by the name of b_jonas). buu disappeared for a while but is now back, b_jonas has been gone for years and none of us have ever decided to give it a proper license. the rest is all stuff i've spun off that's got proper
08:51:14 <b_jonas> Lymia: anyway, perlbot can send HTTP GET queries, though it has some annoying limitation in how it uses that, it always tries to parse the reply as html or xml, only gives you the parse tree rather than the original source, and sometimes the parsing fails outright, so you can't just download arbitrary content that's not encapsulated in html or xml, though plain text files will often work because they'll
08:51:47 <Lymia> Oh, is the bot not alive?
08:51:50 <b_jonas> Lymia: and also zemhill happens to accept bfjoust submissions of in a GET query, which is a bug, but I decided I'd use it
08:52:15 <zemhill> b_jonas: "!ztest progname code". See http://zem.fi/bfjoust/ for documentation.
08:52:25 <b_jonas> but I think using perlbot is funnier
08:52:32 <b_jonas> !ztest bad (>)*9([(+.)*150]>)
08:52:33 <zemhill> b_jonas.bad: points -32.86, score 3.02, rank 47/47
08:53:42 <b_jonas> the whole infrastructure, zemhill's hill and website and bot, are rather old and fizzie basically made it mostly work using duct tape
08:53:55 <b_jonas> (of course perlbot is also very old)
08:56:38 <b_jonas> though on the plus side, zemhill was originally written in ruby 1.9, while jevalbot was originally written in ruby 1.8, so zemhill needed fewer layers of duct tape
09:06:36 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143900&oldid=143899 * Ractangle * (+2) /* MarkupL add-ons */
09:10:17 <Lymia> (My last attempt at an evolver was... in retrospect, overcomplicated.)
09:11:08 <wWwwW> would waterfall be TC if you could just reset the value
09:11:16 <Lymia> (An approach came to mind for a BF Joust evolver that follows through how I tend to make bots in practice. Have a concept of "behaviors" which are nested "monotonically".)
09:11:32 <Lymia> (Not sure what the proper term is, but. An ordered list of "behaviors" which can jump to other behaviors that follow them.)
09:11:43 <Lymia> (So complicated strategies can evolve that still reuses code for different "sections".)
09:11:48 <Lymia> (Might still run into fitness landscape issues, but.)
09:12:14 <Lymia> (Basic instructions for constructs like an offset clear or clears in general should help.)
09:20:01 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143901&oldid=143900 * Ractangle * (+608) /* Examples */
09:21:41 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143902&oldid=143901 * Ractangle * (-2) /* MarkupL add-ons */
09:33:23 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…).
09:45:22 -!- chiselfuse has quit (Remote host closed the connection).
09:45:41 -!- chiselfuse has joined.
09:50:48 -!- tromp has joined.
09:53:11 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143903&oldid=143902 * Ractangle * (+376) /* Cat program */
10:27:01 -!- iddi01 has joined.
10:27:41 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143904&oldid=143903 * Ractangle * (+7) /* Cat program */
10:28:37 <iddi01> !zjoust another_rush_program ++>>>(+)*5<(-)*5<(+)*33(>)*4(>[([+[+[--[-[-[(-[{(+)*23(+[{(+)*95(+.)*23(>(+)*117(+.)*23)*-1}])%22}])%20]]]]]]>)*25]++>-)*15
10:28:37 <zemhill> iddi01.another_rush_program: points 0.98, score 21.57, rank 20/47
10:30:46 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143905&oldid=143904 * Ractangle * (+22) /* Cat program */
10:49:19 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143906&oldid=143905 * Ractangle * (+464) /* MarkupL syntax */
10:49:29 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143907&oldid=143906 * Ractangle * (+7) /* MarkupL syntax */
10:54:01 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143908&oldid=143907 * Ractangle * (+66) /* MarkupL syntax */
10:56:20 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143909&oldid=143908 * Ractangle * (+32) /* Cat program */
10:56:55 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143910&oldid=143909 * Ractangle * (-11) /* Cat program */
11:09:00 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…).
11:15:31 <Lymia> !zjoust worst_program (><)*-1
11:15:32 <zemhill> Lymia.worst_program: points -33.05, score 3.06, rank 47/47
11:19:14 <esolangs> [[Comment]] https://esolangs.org/w/index.php?diff=143911&oldid=143557 * Ractangle * (+17) /* 4ME */
11:26:30 -!- __monty__ has joined.
11:26:45 <esolangs> [[User talk:PythonshellDebugwindow]] https://esolangs.org/w/index.php?diff=143912&oldid=143422 * Cycwin * (+119)
11:28:17 <esolangs> [[4ME]] https://esolangs.org/w/index.php?diff=143913&oldid=139947 * Ractangle * (-83)
11:32:54 <b_jonas> question. do bfjoust abbreviations let you compactly describe a quasi-infinite string P where P = [QP[RP]] where Q and R are given programs with brackets balanced, and the closing square brackets are never reached?
11:33:23 <b_jonas> if it's just P = [QP] then you can write that as ([Q{}])*-1
11:52:38 -!- X-Scale has joined.
11:58:17 -!- X-Scale has quit (Ping timeout: 256 seconds).
12:03:46 <b_jonas> wait no, that's the wrong question
12:04:12 <b_jonas> I'll have to figure out the real question later
12:14:08 <esolangs> [[User:ATProtogen/Sandbox]] M https://esolangs.org/w/index.php?diff=143914&oldid=143886 * ATProtogen * (-54)
12:21:08 -!- amby has joined.
12:29:58 <esolangs> [[MarkupL]] https://esolangs.org/w/index.php?diff=143915&oldid=143910 * Ractangle * (-21) /* MarkupL syntax */
12:41:35 -!- tromp has joined.
12:59:30 <esolangs> [[User:PkmnQ/Alternate Universe Underload]] https://esolangs.org/w/index.php?diff=143916&oldid=143855 * PkmnQ * (+2283) More minimization
13:00:58 -!- Everything has joined.
13:08:32 <esolangs> [[User:ATProtogen/Sandbox]] M https://esolangs.org/w/index.php?diff=143917&oldid=143914 * ATProtogen * (+921)
13:15:11 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…).
13:16:32 -!- tromp has joined.
13:17:07 -!- wWwwW has quit (Quit: Client closed).
13:22:26 <esolangs> [[User:ATProtogen/Sandbox]] M https://esolangs.org/w/index.php?diff=143918&oldid=143917 * ATProtogen * (+6)
13:43:45 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…).
13:47:12 -!- tromp has joined.
13:47:13 -!- supercode has joined.
13:51:23 <esolangs> [[Talk:lang]] https://esolangs.org/w/index.php?diff=143919&oldid=140439 * Yes * (+87)
13:58:00 <supercode> Is there a proof available for Brainfuck being Turing Complete?
13:59:40 <esolangs> [[User:PkmnQ]] https://esolangs.org/w/index.php?diff=143920&oldid=142715 * PkmnQ * (+311)
14:08:59 -!- iddi01 has quit (Quit: Client closed).
14:16:48 <esolangs> [[User:ATProtogen/Sandbox]] M https://esolangs.org/w/index.php?diff=143921&oldid=143918 * ATProtogen * (-201)
14:21:02 -!- Everything has quit (Ping timeout: 272 seconds).
14:22:35 -!- Everything has joined.
14:23:17 <esolangs> [[User:ATProtogen/Sandbox]] M https://esolangs.org/w/index.php?diff=143922&oldid=143921 * ATProtogen * (+12)
14:28:33 <esolangs> [[BubbleLang/Error]] https://esolangs.org/w/index.php?diff=143923&oldid=139584 * PrySigneToFry * (+339)
14:31:09 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…).
14:32:35 -!- amby has quit (Remote host closed the connection).
14:57:36 -!- tromp has joined.
15:02:01 -!- amby has joined.
15:02:02 <esolangs> [[Snakel]] https://esolangs.org/w/index.php?diff=143924&oldid=143859 * Ractangle * (+42)
15:06:40 <esolangs> [[Snakel/Errors]] N https://esolangs.org/w/index.php?oldid=143925 * Ractangle * (+210) Created page with "{{WIP}}If you make an error in the code (for example putting an undefined variable into the code) in a Snakel interpreter. Then the terminal will emit an error and will explain and show where you made a mistake"
15:07:00 <esolangs> [[Snakel/Errors]] https://esolangs.org/w/index.php?diff=143926&oldid=143925 * Ractangle * (+1)
15:07:28 <esolangs> [[SKR]] N https://esolangs.org/w/index.php?oldid=143927 * Hakerh400 * (+7075) +[[SKR]]
15:08:06 <esolangs> [[Language list]] https://esolangs.org/w/index.php?diff=143928&oldid=143848 * Hakerh400 * (+10) +[[SKR]]
15:08:18 <esolangs> [[User:Hakerh400]] https://esolangs.org/w/index.php?diff=143929&oldid=141802 * Hakerh400 * (+10) +[[SKR]]
15:09:23 <esolangs> [[SKR]] M https://esolangs.org/w/index.php?diff=143930&oldid=143927 * Hakerh400 * (+2)
15:59:57 -!- Everything has quit (Quit: leaving).
16:11:22 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…).
16:34:07 -!- tromp has joined.
17:06:50 <esolangs> [[Snakel/Errors]] https://esolangs.org/w/index.php?diff=143931&oldid=143926 * Ractangle * (+14)
17:35:45 <esolangs> [[Hostile snakes]] N https://esolangs.org/w/index.php?oldid=143932 * Yayimhere * (+801) Created page with "{{Lowercase}} '''hostile snakes''' is an esolang created by [[User:Yayimhere]]. hostile snakes is a 2d self modifying esolang with ladders. dont worry about the name, its just an overcomplicated pun == how it works == the pointer starts at lower right corner, an
17:42:17 <esolangs> [[Snakel/Errors]] M https://esolangs.org/w/index.php?diff=143933&oldid=143931 * PythonshellDebugwindow * (+17) Back
17:43:29 <esolangs> [[Hostile snakes]] M https://esolangs.org/w/index.php?diff=143934&oldid=143932 * PythonshellDebugwindow * (+81) Categories
17:45:06 -!- fowl has quit (Ping timeout: 246 seconds).
17:50:13 <esolangs> [[Snakel/Errors]] https://esolangs.org/w/index.php?diff=143935&oldid=143933 * Ractangle * (+25) /* TypeError */
17:50:59 <esolangs> [[Snakel/Compatibility methods]] https://esolangs.org/w/index.php?diff=143936&oldid=143725 * Ractangle * (+17)
17:52:46 <esolangs> [[Snakel]] https://esolangs.org/w/index.php?diff=143937&oldid=143924 * Ractangle * (+33) /* Truth-machine */
17:55:41 -!- craigo has joined.
17:55:41 -!- craigo has quit (Client Quit).
18:01:57 <int-e> korvo: https://paste.debian.net/1332858/
18:04:09 -!- craigo has joined.
18:16:22 <korvo> int-e: Wow, fast. I'll reproduce in a bit.
18:17:37 <esolangs> [[Snakel/Errors]] https://esolangs.org/w/index.php?diff=143938&oldid=143935 * Ractangle * (+198) /* Using a wrong type */
18:18:06 <esolangs> [[Snakel/Errors]] https://esolangs.org/w/index.php?diff=143939&oldid=143938 * Ractangle * (+53)
18:19:00 <esolangs> [[Snakel/Errors]] https://esolangs.org/w/index.php?diff=143940&oldid=143939 * Ractangle * (+41) /* Using a wrong type in a variable */
18:20:39 -!- X-Scale has joined.
18:25:13 <esolangs> [[Snakel/Errors]] https://esolangs.org/w/index.php?diff=143941&oldid=143940 * Ractangle * (+256) /* Using a wrong type in a variable */
18:33:24 -!- craigo has quit (Quit: Leaving).
18:56:28 -!- supercode has quit (Quit: Client closed).
18:56:50 <esolangs> [[User:ATProtogen/Sandbox]] M https://esolangs.org/w/index.php?diff=143942&oldid=143922 * ATProtogen * (-16)
19:05:17 -!- mniip has joined.
19:07:18 <esolangs> [[4ME]] https://esolangs.org/w/index.php?diff=143943&oldid=143913 * Ractangle * (+98) /* Rewritten 4ME */
19:08:46 -!- sam__ has joined.
19:09:11 -!- sam__ has changed nick to Guest1779.
19:11:58 <esolangs> [[4ME]] https://esolangs.org/w/index.php?diff=143944&oldid=143943 * Ractangle * (+56) /* Commands */
19:12:30 -!- Guest1779 has quit (Client Quit).
19:12:56 <esolangs> [[User:ATProtogen/Sandbox]] M https://esolangs.org/w/index.php?diff=143945&oldid=143942 * ATProtogen * (-93)
19:13:02 -!- sam__ has joined.
19:13:04 -!- sam__ has quit (Client Quit).
19:16:32 <korvo> int-e: Yep, your example works. Very nice! Do you prefer this pastebin for permanent links, or can I save a copy to the bb-gauge repo?
19:18:49 -!- Guest22 has joined.
19:19:11 -!- mniip has quit (Quit: Client closed).
19:19:14 -!- Guest22 has quit (Client Quit).
19:19:36 <int-e> I prefer the pastebin for the time being
19:21:08 -!- Guest22 has joined.
19:21:28 -!- ais523 has joined.
19:21:41 -!- Guest22 has quit (Client Quit).
19:22:37 <ais523> Lymia: I wanted to tell you that nyuroki2 is more impressive than you probably think it is – the way you check for clashing decoy setups is innovative and it's actually a surprisingly hard program to beat with something other than a slightly tweaked copy of itself
19:23:05 <ais523> even the slightly broken antishudder has been causing me huge problems, it is harder to exploit than either a fully working antishudder or an absent one
19:23:14 <korvo> Oh, I'm going to have to remember how to XML-RPC in order to get a raw copy of the paste, aren't I... How would I do that from Nix? This is suddenly an engineering problem.
19:28:38 <int-e> korvo: hmm there's no visible link for that but https://paste.debian.net/plain/1332858/ works
19:29:01 <korvo> int-e: TIL! I had no idea.
19:29:15 -!- sam__ has joined.
19:29:30 <int-e> err, actually there *is* a link at the very bottom
19:29:38 -!- sam__ has changed nick to Guest4255.
19:30:12 <int-e> ah, no, that's just for the creator of the paste.
19:31:45 -!- Guest4255 has quit (Client Quit).
19:34:51 <korvo> int-e: Well, I'm happy to record this as an "Interp(BCL)" weighing in at 2087 characters for now. I'll push that in a bit.
19:36:09 -!- X-Scale has quit (Ping timeout: 256 seconds).
19:37:11 <int-e> korvo: It's a bit less if you drop the first loop (which serves as a comment). But also, the restriction to terms is noteworthy.
19:38:21 <int-e> (full BCL does the same thing as BLC where a self-delimited program is parsed and the rest of the input bits are passed to the program as an argument)
19:38:49 <ais523> now I am wondering how BCL compares to Iota/Jot
19:39:08 <ais523> they are conceptually very similar, just different primitives
19:39:24 <ais523> and directly compiling in either direction would bloat the size of the program…
19:39:31 <korvo> Hm, I should double-check that my optimizer is removing the first loop. I'm canonicalizing all of the BF programs so that we're on a level playing ground; the gauge would not be very useful if it were calibrated by the length of English comments rather than code.
19:41:16 <b_jonas> ais523: is "BCL" a typo for BLC?
19:41:42 <int-e> BCL is the spiritual precursor of BLC
19:41:43 <b_jonas> you mean BCL as in an encoding for S-K
19:42:05 <b_jonas> I should just call that one BBS (binary bird sociology)
19:48:58 <int-e> korvo: the fact that there are no links at the bottom is weird. compare https://paste.debian.net/1332870
19:49:19 <korvo> int-e: Ugh, confirmed that the first loop isn't being dropped. I guess I know what I'm doing after lunch. Thanks for catching that!
19:49:53 <korvo> ...Oh, that's creepy. Wonder why it's different? Is there a premium tier or something?
19:50:14 <int-e> I declared it as Brainfuck so maybe something is wrong with that.
19:51:49 <int-e> Hmm, no, it's not that. It chokes on the unicode ⇒
19:52:08 <korvo> Curious. I notice that it's Perl, so maybe they're using something in the Perl ecosystem instead of Pygments. For contrast, here's a Python version, expy in 24hrs: https://bpa.st/M3CAHS76F4E7R6HK3W3WZRBQC4
19:56:53 <int-e> half of that issue was reported here (2 1/2 years ago) https://github.com/formorer/paste.pl/issues/17
19:57:56 <korvo> Heh, cleaning out my gists, found something tangent to combinators: https://gist.github.com/MostAwesomeDude/3c2b2a37f88e750f386fd9053f481351
19:58:08 <int-e> hmmmmm https://github.com/formorer/paste.pl/blob/master/lib/Paste/Template/Plugin/Highlight.pm#L494
19:58:48 <korvo> Those names are from Metamath's intuitionistic logic database, but I did it like Agda with lambdas. I do *not* remember why I was doing this; maybe I was proving that the concept worked before writing a compiler from Metamath.
20:01:06 <korvo> Oh wait, I remember. https://github.com/sorear/metamath-turing-machines did the same thing, but to NQL, for making small TMs. So this must have been part of a proof-search project.
20:02:01 <int-e> (the reason I used the unicode arrow is that => interferes with Brainfuck code. And there's no nice ASCII substitute. =} looks bad.)
20:03:40 <b_jonas> int-e: kvikkalkul uses (: for the <= arrow
20:04:32 <esolangs> [[4ME]] https://esolangs.org/w/index.php?diff=143946&oldid=143944 * Ractangle * (+6) /* Commands */
20:04:48 <int-e> Hmm. I already have machinery for skipping over BCL (sub-)terms. I could probably do "full non-interactive BCL" without too much trouble.
20:05:58 <int-e> (It becomes interactive with lazy evaluation and on-demand input; that doesn't fit well with the program as written.)
20:06:42 -!- timothytomato has joined.
20:07:10 <int-e> oh and lazy output too
20:11:53 -!- timothytomato has quit (Quit: Client closed).
20:35:05 <esolangs> [[4ME]] https://esolangs.org/w/index.php?diff=143947&oldid=143946 * Ractangle * (-258)
20:47:03 <esolangs> [[Snakel/Errors]] https://esolangs.org/w/index.php?diff=143948&oldid=143941 * Ractangle * (+4) /* Using a wrong type in a variable */
21:01:54 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…).
21:35:46 <esolangs> [[Translated ORK/PSTF Again16]] https://esolangs.org/w/index.php?diff=143949&oldid=143692 * MihaiEso * (+42)
21:43:39 <Lymia> ais523: what was the "slightly broken" thing you found there anyways?
21:43:42 <Lymia> Haven't looked at it in a while.
21:49:12 <esolangs> [[Translated ORK/Mihai Again13]] N https://esolangs.org/w/index.php?oldid=143950 * MihaiEso * (+2154) Created page with "[[Translated ORK/PSTF Again16|<span style="font-family:Unifont;">Warning: You have a Samsung Galaxy Tab 2 that is old and usedV46L}4=0 `lA`}y;}b@jSHOmZif_&E,WY>OKy2C+Z#7F.ca=Onk{G bT`j;O1-~Z!]</span>]] 1. Drag out that uNJw..."
21:52:24 <esolangs> [[User talk:MihaiEso]] https://esolangs.org/w/index.php?diff=143951&oldid=143836 * MihaiEso * (+98) /* Make Translated ORK/PSTF Again16 it scarier ! */
21:52:50 <esolangs> [[User talk:PrySigneToFry]] https://esolangs.org/w/index.php?diff=143952&oldid=143594 * MihaiEso * (+335) /* Make Translated ORK/Mihai Again10 Mor ary! */
22:09:13 <ais523> Lymia: on some timings it moves on from a shuddered cell without checking it twice
22:10:04 <ais523> which meant that ash couldn't lock it in its antishudder – it *could* trick it off the tape on some timings but that seems to be probabilistic and it zeroes the flag first on most timing combinations
22:10:42 <esolangs> [[Overload/Turing-completeness proof]] https://esolangs.org/w/index.php?diff=143953&oldid=143861 * Fffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff * (-4) /* Translation */
22:24:58 -!- __monty__ has quit (Quit: leaving).
22:45:38 <Lymia> Yeah, that's intentional. If it sees the cell as 0 twice (total, not twice in a row), it assumes it's not the flag.
22:52:04 <esolangs> [[And]] M https://esolangs.org/w/index.php?diff=143954&oldid=71054 * Timothytomato * (-30) fixed codeblock formatting
22:58:40 <Lymia> (Reasoning there is that the anti-shudder clear (with fairly high chance (?), haven't done the math) should have already set the cell to zero two times in a row once that condition's reached, even if the code's not directly observed it.)
23:00:22 <esolangs> [[Spain without the S]] N https://esolangs.org/w/index.php?oldid=143955 * Timothytomato * (+7302) created pain
23:02:39 <ais523> empirically it seems to be well above 50%
23:06:21 <esolangs> [[User:Timothytomato]] https://esolangs.org/w/index.php?diff=143956&oldid=132870 * Timothytomato * (+26)
23:09:57 <esolangs> [[Spain without the S]] https://esolangs.org/w/index.php?diff=143957&oldid=143955 * Timothytomato * (+12) honestly idk the computational class
23:12:44 <esolangs> [[Brainword]] https://esolangs.org/w/index.php?diff=143958&oldid=132730 * Timothytomato * (+29) nondeterministic
23:42:56 -!- supercode has joined.