←2021-09-05 2021-09-06 2021-09-07→ ↑2021 ↑all
00:03:41 <esolangs> [[User:Xorol]] https://esolangs.org/w/index.php?diff=87923&oldid=86296 * Xorol * (+210) Added some info
01:26:43 <nakilon> maybe then he'e not automated but manned
02:46:57 <esolangs> [[Grue]] https://esolangs.org/w/index.php?diff=87924&oldid=87912 * PixelatedStarfish * (+28) /* Interpreter */
02:47:07 -!- Guest65 has joined.
02:47:13 -!- Guest65 has left.
02:48:48 <esolangs> [[Trivial brainfuck substitution]] https://esolangs.org/w/index.php?diff=87925&oldid=83915 * PixelatedStarfish * (-53) /* Syntax */
03:01:24 <nakilon> hah, few months ago I said "I should made a universal translator between all those brainfuck substitutions"
03:01:36 <nakilon> and someone already made it in that wiki page
03:02:26 <nakilon> oh wait, nope, it's just plain bf interpreter
03:02:55 <nakilon> I imagine a translator could go fetch the wiki pate,
03:03:13 <nakilon> *... page, find the instructions table and replace them automatically
04:10:05 -!- chiselfuse has quit (Remote host closed the connection).
04:10:23 -!- chiselfuse has joined.
04:11:01 -!- src has quit (Ping timeout: 252 seconds).
04:14:57 -!- APic has quit (Read error: Connection reset by peer).
04:15:23 -!- APic has joined.
04:16:07 -!- velik has quit (Ping timeout: 240 seconds).
04:54:57 -!- APic has quit (Read error: Connection reset by peer).
05:00:16 -!- APic has joined.
05:59:57 -!- APic has quit (Read error: Connection reset by peer).
06:00:14 -!- APic has joined.
06:04:45 -!- velik has joined.
06:09:57 -!- APic has quit (Read error: Connection reset by peer).
06:10:15 -!- APic has joined.
06:49:29 <b_jonas> nakilon: but nobody uses the substitutions other than brainfuck, Ook!, and the two or three that encode brainfuck with repetitions of one byte
06:52:39 <nakilon> I thought Chromium OS is something good, because there is no way to install anything in it, you only use docker, so it's so safe...
06:53:47 <nakilon> until I launched a container from another container mounting the docker socket in both, and server just stopped responding, disks were spinning like mad and even serial console didn't work
06:54:32 <nakilon> this stupid container overrated container thing
06:55:42 <nakilon> yeah I know that thing that I'm trying to build are not what people are supposed to try to build but I was told it's all so reliable...
06:55:52 <nakilon> *things
07:31:44 -!- ais523 has joined.
08:05:49 <nakilon> ok docker sockets were not a reason of hanging; it hanged again _-- maybe it's what happens to Chromium OS when it's out of RAM or idk, it just dies
08:06:37 -!- hendursa1 has joined.
08:09:27 -!- hendursaga has quit (Ping timeout: 276 seconds).
08:10:48 -!- ais523 has quit (Remote host closed the connection).
08:18:35 -!- velik has quit (Ping timeout: 252 seconds).
08:38:41 <nakilon> looks like the linux thing called "oom killer" kills a process but the system goes insane anyway and does not recover
08:39:13 -!- Sgeo has quit (Read error: Connection reset by peer).
08:43:27 -!- wib_jonas has joined.
08:44:10 -!- imode has quit (Ping timeout: 240 seconds).
08:44:12 <wib_jonas> does anyone happen to know where I can find human-readable info on how to interpret a robots.txt file?
08:49:00 <nakilon> IIRC it's barely standardized
08:50:07 <wib_jonas> yeah, https://www.w3.org/TR/html4/appendix/notes.html#h-B.4.1.1 has some info but not muc
08:56:39 -!- velik has joined.
09:01:01 <nakilon> lol I did "echo 2 > /proc/sys/vm/overcommit_memory" like this guy proposed https://serverfault.com/a/142003/67097 and now I can't spawn any process even while it's 500 mb of free ram
09:03:28 <nakilon> and now I can't even revert it because sudo sh -c "echo 1 >..." emits -bash: fork: Cannot allocate memory
09:04:10 <wib_jonas> nakilon: so use the kill builtin
09:04:17 <wib_jonas> that is why kill is a builtin command in bash
09:04:24 <nakilon> kill what?
09:05:15 <wib_jonas> yeah, that's the hard part
09:05:28 <nakilon> $ ps
09:05:28 <nakilon> -bash: fork: Cannot allocate memory
09:05:33 <wib_jonas> yeah
09:05:40 <wib_jonas> might be easier to reboot in practice
09:05:43 <nakilon> and this was just before that
09:05:44 <nakilon> $ free -h
09:05:44 <nakilon> total used free shared buff/cache available
09:05:44 <nakilon> Mem: 982Mi 290Mi 382Mi 1.0Mi 308Mi 544Mi
09:07:02 <nakilon> Chromium OS is just broken crap; I guess Google does not care about the scenarios when somethings runs out of memory -- they just stop the machine and spin up a bigger one, and probably expect me to do the same
09:07:11 <nakilon> on every single task
09:08:02 <nakilon> gotta switch to ubuntu
09:17:19 <fizzie> Does bash have a `readlink`-capable builtin? I imagine you could make a poor man's ps out of looking at /proc/pid/exe that way. Hmm, apparently not.
09:18:10 -!- velik has quit (Ping timeout: 240 seconds).
09:21:43 <fizzie> `` for p in /proc/[0-9]*/stat; do read -d'\t' pid prog rest <$p; echo $pid $prog; done
09:21:44 <HackEso> 10 (writeback) \ 11 (kcompactd0) \ 12 (crypto) \ 13 (kintegrityd) \ 14 (bioset) \ 15 (kblockd) \ 16 (kworker/0:1) \ 18 (kswapd0) \ 1 (init) \ 2 (kthreadd) \ 3 (ksoftirqd/0) \ 44 (bioset) \ 45 (kworker/0:1H) \ 48 (cat) \ 49 (`) \ 4 (kworker/0:0) \ 50 (`) \ 51 (`) \ 5 (kworker/0:0H) \ 6 (kworker/u2:0) \ 7 (lru-add-drain) \ 8 (kdevtmpfs) \ 9 (oom_reaper)
09:21:55 <fizzie> I feel like that probably maybe needs no processes.
09:23:22 -!- velik has joined.
09:24:52 -!- velik has quit (Remote host closed the connection).
09:26:31 <wib_jonas> fizzie: I think that will still fork, even if not exec
09:26:33 <wib_jonas> but I'm not sure
09:42:39 <fizzie> Hmm, well, from https://0x0.st/-wu8.txt I think not.
09:44:53 <fizzie> Interestingly, repeating `readlink /proc/self` on my work laptop increments the resulting number by pretty much exactly 13 for each iteration, which makes me think there's some sort of an extra per-command or per-prompt hook set up somewhere.
09:47:34 -!- ais523 has joined.
09:47:36 <wib_jonas> fizzie: cat /proc/*/forks gives the number of processes/threads created total on the system since boot, I think
09:47:44 <wib_jonas> or at least the user-space ones\
09:48:20 <wib_jonas> you can use taht to detect if anything has forked, which is useful to iterate through all processes eg. in /proc and be sure that you haven't missed any, eg. to kill all processes by a user
09:48:25 <ais523> hmm, does that count both fork and clone? there are two ways to create processes nowadays
09:48:31 <ais523> (I think Linux doesn't have a spawn in kernel-space)
09:48:32 <wib_jonas> ais523: yes, it counts clone
09:48:39 <wib_jonas> I checked that once in the kernel source code
09:48:45 <wib_jonas> of course things may have changed since\
09:48:51 <wib_jonas> let me look up the exact name of the file
09:51:20 <fizzie> There's a "processes N" field in /proc/stat at least.
09:54:06 <wib_jonas> yes, the filename is /proc/stat
09:54:09 <fizzie> So I guess https://0x0.st/-wuT.txt is more direct evidence of not creating new processes.
09:54:13 <wib_jonas> you were faster than it took me to look it up
09:54:36 <wib_jonas> (it was among my emails, but was spoilered so a plain text search didn't find it easily)
09:54:48 <wib_jonas> it's not "forks"
09:55:02 <wib_jonas> yep
09:55:13 <wib_jonas> and it's a 64-bit counter, so it can't overflow either
09:56:28 <wib_jonas> mostly though I think if my computer hangs swapping and can't start a new process, I just reboot it
09:56:28 <fizzie> For some reason PIDs that are >= 2^16 just feel weird.
09:57:01 <wib_jonas> (possibly reboot with more and more rescue options if it happens right after boot with no chance to interrupt)
09:58:26 <wib_jonas> fizzie: it feels good, it means we finally managed to shed another piece of historical accident that held us back. there's still a control somewhere in /sys or /proc to limit pids if you don't like it though.
10:14:36 -!- wib_jonas has quit (Quit: Client closed).
10:20:56 -!- wib_jonas has joined.
10:41:24 -!- riv has joined.
10:45:01 -!- APic has quit (Read error: Connection reset by peer).
10:45:20 -!- APic has joined.
10:58:50 -!- tech_exorcist has joined.
11:15:37 -!- wib_jonas has quit (Quit: Client closed).
11:26:51 -!- dyeplexer has joined.
11:29:42 -!- tech_exorcist has quit (Remote host closed the connection).
11:32:01 -!- tech_exorcist has joined.
12:04:23 <nakilon> interesting, 2FA not via SMS but via phone call that you don't need to accept
12:04:33 <nakilon> you just type the last 4 digits of the number
12:04:53 <nakilon> so they own 10000 phone numbers just for that
12:09:37 -!- arseniiv has joined.
12:40:36 <ais523> maybe they're just spoofing the number?
12:42:35 -!- dyeplexer has quit (Ping timeout: 252 seconds).
12:44:57 -!- dyeplexer has joined.
13:41:59 -!- arseniiv has quit (Ping timeout: 252 seconds).
13:53:19 -!- arseniiv has joined.
14:00:04 -!- Everything has quit (Ping timeout: 252 seconds).
14:02:02 -!- Everything has joined.
14:40:30 <fizzie> "DPI SET to 67 x 34 (wanted 96 x 96)" is what I get when attaching to an xpra session, wonder what's up with that.
14:41:51 <fizzie> Tried to do `--dpi=96` on both `xpra start` on the remote side and `xpra attach` on the local side, but still.
14:55:09 -!- riv has quit (Quit: Leaving).
15:01:14 -!- wib_jonas has joined.
15:01:39 <wib_jonas> nakilon: well duh, it's easy to own a lot of phone numbers if you release a popular app that people will install on their smartphones
15:02:02 <wib_jonas> especially since many of those people will have their phones always on and will install just about anything
15:02:27 -!- riv has joined.
15:03:29 <wib_jonas> of course any other app on the same smartphone that has access to your call logs could spoof that part of the authentication as well
15:03:42 -!- src has joined.
15:04:05 -!- riv has quit (Client Quit).
15:05:15 <wib_jonas> it doesn't even have to be that same app, it could be another app that's a game with shiny colored balls and fewer ads than usual that they control that gives them the phone numbers
15:05:20 -!- riv has joined.
15:05:36 -!- riv has quit (Remote host closed the connection).
15:05:53 -!- riv has joined.
15:13:53 -!- imode has joined.
15:17:34 -!- hendursa1 has quit (Quit: hendursa1).
15:18:16 -!- hendursaga has joined.
15:26:34 -!- chiselfuse has quit (Remote host closed the connection).
15:27:40 -!- callforjudgement has joined.
15:27:40 -!- ais523 has quit (Ping timeout: 250 seconds).
15:27:57 -!- callforjudgement has quit (Client Quit).
15:28:24 -!- chiselfuse has joined.
15:56:44 -!- wib_jonas has quit (Quit: Client closed).
16:17:40 -!- chiselfuse has quit (Remote host closed the connection).
16:17:51 -!- chiselfu1e has joined.
16:31:39 <esolangs> [[School]] https://esolangs.org/w/index.php?diff=87926&oldid=85635 * AceKiron * (+82)
16:42:47 <esolangs> [[User talk:Heavpoot]] https://esolangs.org/w/index.php?diff=87927&oldid=85250 * Martsadas * (+111)
17:04:29 <arseniiv> fungot being a pita is no news
17:04:30 <fungot> arseniiv: and china could not copy usa technology and flaunt it just keeps on growing. am a bit puzzled about why you asked
17:05:04 <arseniiv> fungot: why, I didn’t ask anything (yet?)
17:05:04 <fungot> arseniiv: what is remarkable, i think. mostly common lisp and scheme implementations ( tm))
17:05:38 <arseniiv> I need to write some lispy language sometime, yeah
17:05:55 <int-e> that sounded like pure propaganda
17:06:11 <int-e> (fungot, I mean)
17:06:11 <fungot> int-e: you're suggesting cluttering the namespace with names like fnord and fnord
17:06:31 <int-e> fungot: okay, no need to get insulting
17:06:31 <fungot> int-e: just seems silly to construct a list; it does whatever the code surrounding the ( x y) x y
17:07:09 <arseniiv> something far from CL and Scheme, just using S-expressions for its syntax and maybe using symbols, though good symbols might ask their full CL treatment which isn’t easy for a language developer
17:07:32 <arseniiv> you're suggesting cluttering the namespace with names like fnord and fnord => lol
17:09:31 <arseniiv> I agree it’s silly to construct a list; either it’s already there in platonic sense, or it’s nowhere at all, in which case trying to construct it would be in vain
17:11:31 <arseniiv> if it’s already there, one needs to just fetch it with (get-list-from-outer-planes <name of the list>)
17:12:15 <arseniiv> if it doesn’t have any names, then I guess we’re broke
17:13:03 <arseniiv> either it’s already there in platonic sense, or it’s nowhere at all => though I’m unnecessarily classical here. Maybe there is a third option
17:19:04 <int-e> arseniiv: "what is a list" has the potential to unfold into all philosophy ever conceived and some new one
17:19:49 <arseniiv> int-e: it better do!
17:20:14 <arseniiv> or did
17:20:22 <int-e> at some point you'll have to make a list of possible interpretations
17:20:28 <int-e> just to keep track of it all
17:21:08 <int-e> (not circular; you can make a list without fully understanding what it is)
17:21:39 <Corbin> Lists usually require infinity; they don't trivially always exist. https://ncatlab.org/nlab/show/free%20monoid explains the details.
17:22:52 <shachaf> Saying "lists require infinity" sounds like saying "natural numbers require infinity".
17:23:09 <shachaf> Where I think it's maybe more like, the style you use to talk about natural numbers requires infinity, or something.
17:24:01 <shachaf> In Haskell lists are certainly not free monoids (either in the standard sense or in the Haskell sense, where those are also distinct).
17:24:07 <Corbin> Maybe there's a more grammatical way to put it. FinSet is a topos, so it's not possible to constructively tell whether or not infinite objects really exist; we have to assume them.
17:24:17 <int-e> shachaf: fortunately we can justify this claim on a ridiculously informal level: Have you ever looked at the number 8 from the side? ;-)
17:26:05 <shachaf> It's well-established that the Peano naturals stop at 88.
17:26:17 <arseniiv> I think something like shachaf too. For example, though I don’t know ultrafinitist constructions, we could imagine lists of length more than N fail to exist, so cons or ++ are partial but for some reason we can’t stumble on that fact
17:26:24 <arseniiv> oh, 88
17:27:13 <shachaf> Man, "what if sufficiently big numbers don't exist? we'd never know it" is a funny take on ultrafinitism.
17:27:36 -!- Koen_ has joined.
17:28:28 <int-e> you'll need imaginary numbers then
17:28:38 <int-e> that you imagine to exist
17:29:05 <shachaf> The thing that makes it funny is that it still takes some platonic notion of existence, and imagines that it applies to some numbers, and not others. I hadn't really thought of that perspective.
17:30:10 <Corbin> I'm not even being finitist here. It's a variation on Skolem's paradox. In order to talk about existence of objects, we need a formal mathematical logic for those objects. For questions of size/cardinality/measurement, set theory is traditional. But constructive set theory is confined within the rules of topoi, so if a topos lacks some classical property, then the property didn't exist universally in the first place.
17:30:40 <Corbin> Just like with Skolem's paradox, we might have taken some property of Set (say, that there's natural numbers or free monoids) and imagined that it must be universal, even though some topoi like FinSet refute it.
17:30:50 -!- src has quit (Quit: Leaving).
17:30:58 <arseniiv> <int-e> that you imagine to exist => hm then we may end up with an existence hierarchy
17:31:04 <shachaf> I was talking about arseniiv's thing.
17:31:16 <shachaf> But, hmm, I can talk about the existence of objects without a formal mathematical logic for those objects.
17:31:56 <Corbin> I guess. Maths just makes it easy: Every describable object exists, except for those which would contradict themselves if they existed.
17:32:21 <shachaf> Would you say that existence is a property that some things have and some things don't, or that everything exists?
17:32:55 <Corbin> I'd say that "existence" is a linguistic meme, a language-game which has a faulty assumption about the nature of reality.
17:33:51 <arseniiv> there were some logics with two sets of quantifiers, the usual ∀ and ∃ and then “outer” ones which have strange semantics
17:34:20 <shachaf> Hmm, Internal Set Theory?
17:34:55 <arseniiv> a quantified variable in those runs over inexisting things. Don’t remember if it’s connected to IST
17:35:07 <arseniiv> ow, not inexisting things, but all things
17:35:34 <arseniiv> I think that was for modal logics or free logics, hm
17:36:38 -!- riv has quit (Quit: Leaving).
17:37:25 <int-e> "real thought" is a nice oxymoron
17:37:25 <arseniiv> I guess I saw that here: https://plato.stanford.edu/entries/logic-free/#Meinong
17:37:37 -!- riv has joined.
17:39:44 <arseniiv> once I thought some kind of free logic would be what I wanted an ideal classical logic to be, yet it failed when I read this article: all kinds seem to have their flaws
17:40:52 <arseniiv> hopefully a type-theoretical framework is good. At least I haven’t seen someone discussing flaws like those for it to have
17:41:54 <Koen_> it's one of those "there are 15 types of logic, but they all have their flaws. i'm going to develop a new logic that unifies them all" -> "there are 16 types of logic, but they all have their flaws"
17:42:44 <int-e> But it makes for a great story. https://en.wikipedia.org/wiki/Homo_narrans
17:43:20 <Corbin> arseniiv: The "field with one element" doesn't exist, because of the definition of fields. Nonetheless it has pages on WP and nLab, because there is a field-with-one-element-shaped-hole in projective geometry.
17:43:38 <b_jonas> "It's well-established that the Peano naturals stop at 88." => I don't think that can satisfy the axioms
17:44:09 <Corbin> When combined with the fact that the category of fields is disconnected (fields of different characteristic aren't related to each other), it suggests that fields are an incomplete notion; we have the wrong definition!
17:44:17 <int-e> b_jonas: that just means that the theory is flawed model of reality
17:44:23 <arseniiv> Corbin: yeah I agree it should exist in some fashion
17:44:33 <arseniiv> the field F₁, I mean
17:45:37 <b_jonas> "a quantified variable in those runs over inexisting things" => is that second order logic?
17:46:08 <int-e> arseniiv: oh, third possibility: does Schrödinger's cat exist?
17:46:47 <int-e> Hypothetically speaking, of course ;)
17:46:51 <arseniiv> b_jonas: no, that one is one-sorted
17:47:53 <b_jonas> int-e: unlikely. Schrödinger died in 1961, his cat would have died and decomposed by now unless he got it mummified or put it on a relativistic speed rocket or something
17:48:18 <int-e> b_jonas: seems fair
17:48:48 <arseniiv> int-e: it, I mean the paradox itself, does in global culture, but as of the cat itself, I’d think decoherence theory gives us a very tight estimate on the time the cat’s still in superposition, like attoseconds I think
17:49:11 <arseniiv> oh my poor grammar
17:51:54 <b_jonas> what I wodner is, if the poison gas is released in the box to kill Schrödinger's cat, wouldn't it also kill whoever opens the box when they open it?
17:51:57 -!- delta23 has joined.
17:52:39 <b_jonas> I guess they could be careful and use a chemistry lab smoke hood or something
17:52:46 -!- Sgeo has joined.
17:53:30 <arseniiv> hm what if we call a Schrödinger’s cat any cat which lives in Schrödinger’s children’s homes as an unbreaking line, or a tree. Then this kind of a collective cat is probably still alive
17:54:29 <arseniiv> or if his children all disliked cats, then not
18:17:19 -!- dyeplexer has quit (Remote host closed the connection).
18:18:58 -!- Everything has quit (Quit: leaving).
18:42:28 <b_jonas> I'm not sure if any cat would be willing to own the family that deliberately killed with 50% probability a previous cat just for a tought experiment
18:42:50 <b_jonas> they would certainly just pick a better choice of servent
18:49:42 -!- tech_exorcist_ has joined.
18:51:05 -!- tech_exorcist has quit (Ping timeout: 252 seconds).
19:33:18 -!- Koen_ has quit (Quit: Leaving...).
20:16:16 <arseniiv> are there good aperiodic 1d tilesets?
20:17:27 <arseniiv> if so, either the matching rules should be quirky, or tiles should at least be nonconnected
20:18:14 <arseniiv> I feel it might end up not that hard to enforce for example 0 → 01, 1 → 10 blow-up rules
20:30:26 <b_jonas> arseniiv: I don't think aperiodic 1d tilesets are possible
20:30:55 <b_jonas> 1d tiles just behave like regular languages
20:31:27 <riv> yes
20:31:31 <b_jonas> unless perhaps it's some weird non-measurable ones, but even so I doubt they can force aperiodic
20:31:52 <riv> you can do an aperiodic 1d tiling, but there's no tile set that forces aperiodicity
20:32:07 <arseniiv> :(
20:32:44 <int-e> +finite
20:32:46 <arseniiv> how to make a connection with regular languages clearer?
20:33:01 <riv> each tile is a transition from one state to another
20:33:03 <arseniiv> finite, yeah, I wouldn’t want an infinite tileset
20:33:03 <int-e> arseniiv: it's quite literally an NFA
20:33:15 <arseniiv> hmmmhmhmhm
20:34:42 <arseniiv> I see, even if a tile has many holes and connects with many previous and following tiles, it still would be a finite number of neighborhoods allowed for all of them
20:34:54 <arseniiv> that’s sad
20:36:08 <riv> it's interesting
20:36:25 <riv> so for wang tiles to work, the pattern needs height
20:36:32 <riv> that increases without bound
20:36:44 <riv> any finite height strip would just be equivalent to a 1d tileset
20:37:05 <b_jonas> um wait
20:37:26 <arseniiv> ah at least this 1d thing gives an insight (if it does)
20:38:32 <b_jonas> what if you have two tiles, one is the set of rationals between 0 and sqrt(3), the other the set of all irrationals between 0 and 1?
20:38:42 <b_jonas> one is the set of rationals between 0 and sqrt(3), the other the set of all rationals between 0 and 1?
20:38:46 <b_jonas> no
20:38:49 <b_jonas> argh
20:38:55 <b_jonas> one is the set of rationals between 0 and sqrt(3), the other the set of all irrationals between 0 and 1?
20:39:26 <b_jonas> then you'll have to overlay the two tiles such that the first kind repeats with period sqrt(3), the second with period 1
20:39:30 <b_jonas> of course it's cheating
20:39:45 <b_jonas> but you should try to explain why that problem can't happen without cheating
20:39:56 <b_jonas> wait that doesn't work
20:40:02 <b_jonas> you have to be more evil than that
20:41:08 <b_jonas> it has to be the set of rational linear combinations of 1 and sqrt(3) between 0 and sqrt(3) half-inclusive, and the set of numbers between 0 and 1 that are linear independent from {1,sqrt(3)}
20:41:34 -!- Corbin has quit (Ping timeout: 252 seconds).
20:41:41 <riv> well there is the projection trick
20:42:20 <riv> http://sig3.ecanews.org/isac2010/lectures/06a_fujita_tilings_a.pdf slide 13 14
20:42:50 <riv> you start with a slope of irrational gradiant and project it down
20:43:58 <b_jonas> riv: sure, but in 1D that'll just lead to an aperiodic tiling
20:44:06 <b_jonas> not a tileset that forces it
20:55:44 -!- Lord_of_Life_ has joined.
20:55:52 -!- Lord_of_Life has quit (Ping timeout: 252 seconds).
20:58:21 -!- Lord_of_Life_ has changed nick to Lord_of_Life.
21:00:09 <riv> can a PDA output an aperiodic string?
21:01:22 <keegan> PDA doesn't output a string, it only accepts or rejects a string
21:01:38 <riv> i seee...
21:01:41 <keegan> but I suppose you could easily define an output version (pushdown transducer?)
21:02:06 <riv> maybe it has to be context sensitive
21:02:49 <int-e> b_jonas: You can stick to natural numbers or integers; for example, if you let X = { n | n >= 0 such that n can be written in base 4 using only digits 0 and 1}, then {X u X+2} = 4X+{0,1,2,3}, so you can tile N, and {X u X-2} = 4X+{-2,-1,0,1} so you can tile Z. Obviously those are aperiodic
21:03:00 <keegan> anyway i think it could output a nonperiodic string
21:03:19 <keegan> you can use the stack to count so you could do like ABBAAABBBBAAAAABBBBBB
21:03:40 <keegan> and there is no bound to the stack depth
21:04:34 <int-e> (so infinite tiles are a problem)
21:05:29 <riv> keegan, oh good point! i was overlooking that completely
21:05:55 <keegan> i'd have to formalize it to be sure though
21:06:04 <keegan> maybe it's not so easy to count because you lose the count as you consume it
21:07:38 <int-e> the pumping lemma for PDAs means you can always have long stretches of periodic substrings...
21:07:42 <keegan> remember also there is a difference in power between deterministic and nondeterministic PDA (unlike for finite automata or for turing machines)
21:07:55 <int-e> (assuming it accepts strings of arbitrary length)
21:08:23 <arseniiv> int-e: infinite tiles are a problem for sure, I think using disconnected tiles we could force using them like keegan’s ABBAAABBBB: 1 2 3 4 etc., something like making them each 0, 1, 2, 3, … holes and carefully assigning distances
21:08:55 <keegan> (fwiw i wasn't paying attention to the earlier tiling discussion)
21:09:17 <arseniiv> it was about if I can make a finite 1d nonperiodic tileset
21:09:24 <int-e> arseniiv: what I have is similar in spirit, really; I just find it cute to make it work with a single tile (though that means you have to keep track of the positioning of the tile for "aperiodicity")
21:10:17 <int-e> (but to my mind that makes sense: there's no placement of tiles that is repeated periodically)
21:10:17 <arseniiv> int-e: I think I saw that tile somewhere, even, if I infer its construction right (I skimmed what you wrote)
21:10:35 <int-e> arseniiv: it's reminiscent of the cantor set
21:13:03 <arseniiv> int-e: something like p-adically? Oh I guess I should just read in
21:15:39 <int-e> > fix (\x -> 'X':tail (x >>= \c -> [c,c,'.','.']))
21:15:41 <lambdabot> "XX..XX..........XX..XX..........................................XX..XX........
21:16:17 <arseniiv> hm wait why use 4 but not 3?
21:18:01 <int-e> with digits 0 and 1? that would leave a gap of size 1 near the beginning: XX.XX.... etc.
21:18:06 <int-e> and nothing to fill it
21:18:30 <arseniiv> ah, I see, we can’t use just 1
21:18:46 <int-e> and using digits 0 and 2 confuses me, but I imagine it still doesn't work
21:20:19 <int-e> I mean we can take X.X...X.X.........X.X...X.X[...] and shift it by one, and that will result in XXXX..XXXX........XXXX..XXXX[...] and now we have chunks of length 4 and some gaps of length 2
21:20:31 <int-e> in contrast, base 4 just works
21:21:05 <arseniiv> -X- -X- --- --- -X- --- --- --- --- --- --- --- --- -X- --- --- --- --- --- --- --- --- --- --- --- --- --- <- with 1 in ternary
21:21:34 <arseniiv> yeah 4 with 0 and 1 looks more comfortable
21:21:39 <int-e> basically by virtue of {0,1,2,3} = {0,1}+{0,2}, where {0,1} are used as digits for the tiles and {0,2} as offsets for shifting the tiles
21:22:31 <int-e> Or, for covering the integers, {-2,-1,0,1} = {0,1}+{-2,0}
21:23:33 <arseniiv> how I like simple solutions
21:27:36 <int-e> Hah, how about this abstraction as an exercise: "Find a subset X of N={0,1,...} such that the map (x,y) |-> x+2y is a bijection between XxX and N."
21:29:45 <int-e> (The X from above is the only solution.)
21:32:07 <riv> would that just be the odd numbers?
21:32:29 -!- olsner has joined.
21:32:58 <b_jonas> int-e: sure, with infinite tiles it's easy. you can make one huge infinite tile with tiny holes far from each other, where the holes are randomly 2 or 3 long, plus small 2 and 3 long tiles to fill one of those holes, and that's aperiodic too.
21:33:16 <int-e> riv: no, then you don't get any even numbers as results... but also plenty of duplicates
21:33:49 <b_jonas> you can make that work with just two tiles in fact, no need for different sizes of holes, just place the holes randomly
21:33:53 <int-e> b_jonas: your tiles are effectively infinite too, by virtue of incommunserability
21:34:42 <int-e> b_jonas: right. but as I said, I found the task of making a single tile work interesting
21:34:59 -!- APic has quit (Read error: Connection reset by peer).
21:35:00 <riv> difficult problem :O
21:35:41 -!- APic has joined.
21:36:45 <int-e> riv: well, 0 needs to be in X (otherwise you don't get 0 as a sum)... 1 needs to be in X (otherwise you don't get 1). 2 can't be in the set (because otherwise we'd get 2+2*0 = 0+2*1... and it basically continues like that.
21:37:26 <int-e> riv: at some point you'll switch to binary representation.... and look at every second digit... and that's how you get to base 4.
21:37:49 <riv> that is so cool, it's like a sudoku or something
21:38:17 <int-e> I think it's doable, but it does require some technique and persistence.
21:45:56 -!- Corbin has joined.
22:02:48 -!- tech_exorcist_ has quit (Quit: see you tomorrow).
22:24:53 -!- APic has quit (Read error: Connection reset by peer).
22:25:32 -!- arseniiv has quit (Ping timeout: 245 seconds).
22:42:13 -!- APic has joined.
23:08:55 -!- sprock has quit (Read error: Connection reset by peer).
23:09:28 -!- sprock has joined.
←2021-09-05 2021-09-06 2021-09-07→ ↑2021 ↑all