The Nineteenth Byte

Wheat Wizard

23:00

I like Adám's better though.

ais523

I'm not convinced, 999 bytes should be easily enough to produce a program whose behaviour mathematically can't be predicted (because if you could prove it didn't halt, it could prove your proof wrong)

there are programs that can't be proven either to halt, or to not halt, in ZFC…

Wheat Wizard

Depends on how many axioms you have.

You are not constrained to ZFC.

ais523

yes, but if you want to model the behaviour of the program correctly, you need to be sure that your axioms are actually correct

hmm… any proof system that you did use would have to be sufficiently complex that you couldn't represent it with a short Jelly program

otherwise, you could write a program that searched for a proof that it wouldn't halt, and halted if and only if it found one

Wheat Wizard

Yeah. We also I guess need a pretty formal definition of Jelly.

We have a python interpreter but that just kicks it down the road.

999 might be a bit big.

ais523

we've looked into busy beavers in #esoteric, and probably here on CGCC too

and the smallest programs that we don't know the behaviour of are pretty small, even in tarpits

Wheat Wizard

23:06

Yeah, I have spent a very long time with bbs for turing machines and brain-flak

My alma mater actually holds several bb records on turing machines.

"Print a number that Jelly code less than 66 bytes cannot output." Should slightly improve the issue.

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

Index Of A Color Of The Rainbow Challenge: The challenge is to output the index of a given color of the rainbow. For example, Red your program should output 0 (1 if you are using 1-indexed arrays). Rules: Input is case sensitive. Red should return 0 however red or RED should not. Invalid input c...

ais523

Jelly can implement a Turing-complete programming language in just 4 bytes (without using eval), so 66 is likely to still be quite a lot

Wheat Wizard

Ok "Print the 34th Jelly busy beaver."

Yeah, I don't know much about Jelly. I wonder how many bb are within somewhat easy reach.

ais523

actually, I'd love to see a busy beaver cops-and-robbers

Wheat Wizard

I've thought about it a lot.

ais523

23:09

cops write a program which the robbers can't predict the eventual behaviour of, shortest wins

but this would be different from a typical CnR because the cop doesn't have any particular solution in mind, they're just trying to create something as unsolvable as possible

Wheat Wizard

I think there is something like that.

ais523

actually, yes

7

A: Does the code terminate?

Brachylog (v2), 3 bytes in Brachylog's encoding ⟦cṗ Try it online! (will time out without doing anything visible, for obvious reasons) Full program; if run with no input, searches for the first Smarandache prime, and outputs true. if and when it finds one. It's an open question as to whether ...

there we go, a 3-byte Brachylog program for which nobody knows whether it will terminate or not (assuming infinite memory)

Wheat Wizard

I think the shortest Brain-Flak program we don't know the behavior of is around 38 bytes.

Hmm I think I might be double counting.

It is closer to 20.

We did everything under 18 when we proved the squaring program optimal. And I did a few after that.

Ok now I am going to have to give this another go.

ais523

now I'm wondering what busy beavers for JSFuck look like

it is one of the most inherently verbose languages out there, I think you need thousands of bytes to simply call a function

Wheat Wizard

Lenguage bbs.

You can probably calculate a great number of them.

ais523

23:18

they're like BF BBs, but exponentiated

although, Lenguage I/O is weird, it isn't a drop-in replacement for BF (as I noticed one day when actually trying to test a Lenguage program on the interpreter)

Wheat Wizard

I think you mean logarithmic?

There really needs to be a golfier length based language.

ais523

no, BF BBs are like the logarithm of Lenguage BBs

I've considered making a Jelly-based Lenguage

which requires the program to be given as a file, and simply looks at the file's size on disk without attempting to read it (like Lenguage does)

Wheat Wizard

When this brain-flak based golfing language is done it would be like 3 lines of code to make it also work based on length.

ais523

you can add random features to golfing languages, like length-based-ness and radiation hardening

sometimes even without damaging the original language much, if you look for unlikely statistical properties

like "the program contains 1000 or more consecutive copies of the same byte" (this is never useful when golfing normally, if you have an eval, because you could just generate the program and eval it using fewer bytes)

Wheat Wizard

Yeah, which is in my opiion why radiation weakening is much more interesting than radiation hardening.

ais523

23:23

what do you mean by radiation weakening? creating a program that breaks when any character is deleted?

I think it's pretty much the same in terms of language features required to make it easy; just look at A Pear Tree, for example

where if you modify a program, it can not only repair itself, but also figure out what was changed and how

Wheat Wizard

I mean programs that do particular things when modified in particular ways. It is a much more varied category of challenges.

ais523

right, it still benefits a lot from a parser specifically designed to handle modified programs though

Wheat Wizard

Of course.

caird coinheringaahing

@WheatWizard Like the "I ___ the source code, you ___ the input" challenges?

Wheat Wizard

That would be a bit of a subset yeah.

ais523

23:26

that's more source-layout than radiation-hardening, although it's a continuum

Wheat Wizard

The Geiger counter is another one which doesn't quite fall into that template but is radiation weakening.

ais523

the program's often changed quite dramatically

Wheat Wizard

I very much like source-layout too.

ais523

although you could merge them as "I delete the nth character of the source code, you delete the nth character of the input"

(which may have already been done, and seems almost impossible in most languages, and fairly easy in Befunge and A Pear Tree)

Wheat Wizard

I think that challenge exists.

ais523

23:27

actually, Befunge may be nontrivial

Wheat Wizard

I think it could be done in Klein.

ais523

because the standard radiation hardening trick there involves runs of duplicate characters, which make the challenge unsolvable because you can't figure out which element of the run got deleted

Wheat Wizard

Oh yeah, that's the issue.

There goes any hope of a lost solution.

Well actually lost's inability to do input is probably a more immediate problem.

I should publish my revised version of lost that can do input.

ais523

@WheatWizard I checked, I don't think it exists (at least, I skimmed everything tagged source-layout)

Wheat Wizard

28

Q: Radiation Detector!

A radiation-hardened program is a program where, if any character of the code is removed, the program will still function the same. For this question, we'll be writing a program that detects when it is irradiated. Write a program or function that, when any single byte is removed, the resulting p...

Maybe this is what I was thinking of. A little similar

But I thought that exact challenge existed.

Aha!

29

Q: Pass on your radiation

The task here is pretty simple. You should write a program that takes a string as input and outputs it (that is a cat program). Additionally when the \$n\$th byte of your program is removed (without replacement) the augmented program should take a string as input and remove the \$n\$th characte...

code-challenge source-layout radiation-hardening self-referential

This is what I was thinking of. It is not the same.

But you get a perfect score if you can do that task.

It was solved perfectly in ><>

It does not have a pear tree answer yet.

ais523

23:35

I can't beat the perfect 0, might be able to beat it on byte count though

Wheat Wizard

I will edit in a tie breaker if you would like.

ais523

maybe I'll have a go later

do we have to follow proper quine rules? that actually makes quite a difference here

Wheat Wizard

I am not sure.

I'm not the biggest fan of proper quine rules since I think they can be unclear, but it looks like it might be the default for the tag.

ais523

the challenge doesn't have the tag though

Wheat Wizard

self-referential seems to have the same restriction by default.

ais523

23:38

aha

probably more interesting that way anyway

Bubbler

IIRC quine enforces the "code and data" stuff, self-referential doesn't

Wheat Wizard

The tag wiki is a little vague, but says you must not read your source.

I think the code and data stuff also doesn't make sense for a lot of self-referential questions.

I've wanted to experiment with pear tree. However I have very little interest in writing perl. I should maybe try making a version for a language I dislike less.

Bubbler

I mean, code and data stuff != not reading source code

Wheat Wizard

Yeah of course.

ais523

@WheatWizard well, it's designed for polyglots, so just write it in some other language that happens to also be valid Perl

it should probably have a successor language at some point, though, and that doesn't necessarily need to be Perl-based

Wheat Wizard

23:43

I see some joker put a link to the self-referential tag wiki in the description of the self-referential tag wiki.

ais523

so far, though, all I really have is the name, "Another Pear Tree"

Wheat Wizard

Names can be hard. I have some language I have not published because I can't come up with a name I like.

So if you have a name you are happy with you may be well on your way.

ais523

the original name was just a throwaway joke for the "add a language to a polyglot" challenge, but I take esolanging seriously enough that I try to create a language that's seriously usable for something even when it's just a throwaway joke

Wheat Wizard

I would say the joke works a little more generally for polyglotting

ais523

the whole language was designed around the joke, including the polyglot builtins and the radiation hardening

because I needed to make sure that almost any program that was submitted to the "add a language to a polyglot" challenge would error out (and thus print a partridge)

so the question was, how do you design a useful language where most programs error, and what sort of purpose for a language would make it behave like that?

Wheat Wizard

23:48

Ah I see what you mean.

ais523

and the answer was, "most of this source file isn't your program, have some way to mark your bit of the program so that the rest of it can do its own thing"

pick a marker that gets disturbed by radiation, and suddenly you have a language that's great for radiation hardening because you can put multiple markers in it and only the undisturbed bit runs

the main upshot of this conversation is that I should make throwaway jokes more often :-D

Wheat Wizard

Yeah it's a neat idea.

Transcript for