The Nineteenth Byte

New Main Posts

00:42

0

(This challenge exists to extend sequence A267272 in the On-Line Encyclopedia of Integer Sequences, and perhaps create a new OEIS sequence1.) This is a code-challenge, which will have you write code to compute as many terms of this sequence as possible. Background A polytet is a kind of poly...

Redwolf Programs

03:18

Dang it, it's a week to the election and I forgot to finish political simulator :/

I guess I'll work extra fast and see if I can finish it in the next few days

Do any of you have way too many projects and can't keep track of them all? I feel like the guy in XKCD #1106

HyperNeutrino

03:34

i can relate to that xkcd especially the hover text lol

the one thing i usually get done the easiest is also the one thing i didn't actually really need to get done

5

Bubbler

03:53

And usually it is solving a code golf challenge

HyperNeutrino

for me it's actually cs assignments, which doesn't sound all that bad except it's usually ones that are due in like one or two weeks when i literally have an english assignment due on that day xD

Bubbler

Lol, same for me when I was in university

Redwolf Programs

For me it's writing challenges...I have probably 50 KoTH drafts that I've either never had time to finish, or didn't end up seeming fun

Razetime

04:30

Husk. Like if you agree

Bubbler

04:45

@Razetime -1 cuz obviously it must be APL

(jk)

Razetime

05:14

(nice)

Lyxal

09:57

CMQ: Have any of your programming languages (eso or practical) been mentioned in online articles/forums etc without you actually writing said inclusion?

I.e. You didn't post the article/have any involvement with it

Unrelated String

probably happens a fair deal to ais but i can't remember the last time i saw him around

John Dvorak

well... technically :D I have written an intcode implementation, and the language has been mentioned in several AOCs :D

Bubbler

Not mine, but Jelly (created by Dennis) and RAD (created by Zachary) have respective pages on APL wiki.

Lyxal

A whole bunch of esolangs mentioned in an article

John Dvorak

CMQ: How would you represent a pub in a 3x3 px pictogram?

@Lyxal what's up with that ginormous gray rectangle at the top?

Lyxal

10:08

@JohnDvorak I can only presume it's a jabberwocky

Not my article btw

John Dvorak

horizontal scroll - ewwwwww.

This is the first time in ever that I used the Firefox reader mode.

Lyxal

@JohnDvorak ^

Wow I thought that'd render bigger

I guess it is saved as a 3x3 pixel after all

It's supposed to be a glass with beer inside

John Dvorak

Got it, thanks

Lyxal

Do you need an enlarged version?

John Dvorak

Opening it in Gimp did the trick

Lyxal

10:20

Very cool. Do you think it resembles a pub?

Also, why 3x3? Pictograms can't exactly portray much in such a small space.

John Dvorak

I use it to distinguish objects that are 3x3 tiles in size on a map and I don't want to upscale.

Adám

@Lyxal My Extended Dyalog APL is mentioned repeatedly on APL Wiki.

John Dvorak

In completely unrelated news, would someone here be interested in a full map of Monster's Expedition?

Neil

@Sisyphus Bz0ṚFḄ also isn't a bijection, since it gives the same result of 7 for [2, 1, 1] and [2, 1, 2]

@Lyxal Ah nice, D♭ exists to complement C# (well, C♯) I guess?

caird coinheringaahing

11:24

> Brainfuck has inspired countless imitator languages, including: [...] JSFuck

I mean, in the same way that Java inspired Javascript

Sisyphus

@Neil You're right. I think it'd have to be something like BUz0FḄ.

Actually, that still doesn't work. Maybe tack another reverse in there like BUz0FṚḄ, but now it's getting much too long.

Neil

11:41

@Sisyphus yeah, I think another reverse works

Unrelated String

12:05

@Lyxal I find it strange that in the section on the palindrome challenge, the article neglected to mention that the solutions themselves have to be palindromes

and what's worse is that the author missed a space in his Java example

caird coinheringaahing

@UnrelatedString I thought it was even worse that they didn't credit half the language authors. Like Funciton was created by a PPCG user Timwi, it really wouldn't have been difficult to mention that

Redwolf Programs

12:19

@Lyxal The article calls APL an esolang...nobody tell Adám :p

#17 is "languages with 2-dimensional source code", so technically one of my languages falls under that category :p

Adám

@RedwolfPrograms Well, he questions that label: "(if you can call it that)".

He has redundant parens in his simple APL sample: +/(3+⍳4)

ais523

13:17

@UnrelatedString I look at the site and/or the TNB archives every now and then, but it's a pain to actually log in to the site and/or to chat (two different accounts with two different reputation levels)

although I posted an answer today, I golfed an old answer of mine down from 309 to 228 bytes when the next-best answer was 890 bytes

@Lyxal ugh, that article contains misinformation about INTERCAL copied from Wikipedia

2

caird coinheringaahing

@ais523 It is kinda weird when one of your answers shows up in the First Posts review queues, and I have to stop myself from my default "Welcome to the site, here are some helpful links" :P

ais523

heh, are they still going there? I think being locked at 11 rep breaks a number of site functions somewhat :-)

although somehow I dropped down to 9 rep, "user was removed", and had to find a post to edit

I checked all my edit suggestions and couldn't figure out what user had been removed to drop the rep, so maybe something weird's going on, but I've long since learned that questioning the Stack Exchange software is a path to sadness and insanity

caird coinheringaahing

I think it might actually be the late answers queue, because that IIRC works on rep, not account age

ais523

oh, "late answers" would be a perfect fit for that one

3 years after the previous answer, from a user with 11 rep, it almost certainly should go there

caird coinheringaahing

Yeah, I tend to lump them together cause posts go to both LA and FP queues all the time

@ais523 How does removing a user drop you 2 rep?

ais523

13:23

according to the documentation, it can't, the documentation says that removing a user undoes their votes but there's no way to get exactly +2 rep from a user by votes when all your posts are community wiki

the only rep on my account came from 5 suggested edits, and a suggested edit is +2, so presumably an edit suggestion rep boost got removed somehow

caird coinheringaahing

I think you lose edit rep if the edit is rolled back, but that doesn't involve removing a user :/

ais523

but all my edit suggestions were approved by users that still exist (or at least, had non-redacted usernames), and on posts by users that still exist, so ???

I hope this isn't fixed, at least, because it'll be a pain to get the account back to 11 rep if I end up losing that -2 at some point in the future

caird coinheringaahing

@ais523 Downvote a couple of LQ answers

ais523

can't

don't have the rep to even upvote

and downvoting needs a lot more

I am very happy that CWing posts needs less rep than voting on them, it helps avoid me voting on posts by mistake (this prevents accidentally upvoting the same post twice, as I have no way of remembering which posts I voted on using my previous account, and I don't think those votes were deleted with the user)

caird coinheringaahing

@ais523 IIRC, your votes don't get deleted if your account is older than a certain age, and I don't remember a bunch of people complaining about losing thousands of rep when you "left"

ais523

13:29

it's up to the moderators and/or admins whether they're deleted, I don't know for certain but it's very likely that they'd have chosen to keep them

Redwolf Programs

17:40

Posting Political Simulator in 5 hours

Actually I'm impatient. Posting now.

Adám

Volunteers wanted: We're getting ready for the 2021 APL Problem Solving Competition, and are seeking volunteers to review and solve the candidate problems, and test the website. Please ping me if you're interested ― you don't have to know APL. (Of course, this disqualifies you from winning.)

caird coinheringaahing

18:01

Browsing the Late Answers queue when suddenly

> ais523

And for some reason, Google tries to translate Jelly from Vietnamese

Redwolf Programs

I reviewed a suggested edit by them, pretty much an automatic "Approve" :p

New Main Posts

18:48

2

Q: KoTH: Political Simulator

It's election time, and your job is to beat your competitor in a head-on rivalry! You are both trying to win over a city of 256 people in a 16x16 grid. Right now, the city hasn't been divided into voting regions yet, but that's where your gerrymandering skills come in! You can also campaign in ar...

grid king-of-the-hill javascript

ais523

18:58

@RedwolfPrograms I found myself unexpectedly in need of 2 reputation so that I could CW the answer that I'd been thinking about on and off for the past 3 years

the funny thing is, I made the edit on a question that had been inactive for over 8 years, but it got approved by the original author rather than by voting

(it took a while to find a question that could do with a good edit, and I normally look at the least active questions for reputation edits as they're least likely to have had any problems with them fixed)

Neil

19:36

@Adám not knowing APL probably disqualifies me from winning even without testing the website

3

Adám

@Neil :-) Does that mean you're volunteering?

caird coinheringaahing

@Adám Does that mean we can we use languages other than APL to test the problems?

Adám

@cairdcoinheringaahing Yes, part of the testing is for clarity of spec and solvability of task.

ais523

19:57

Jelly would probably make a good test language, if you know that; it has a shared history with APL, so they're good at the same types of problems

caird coinheringaahing

@ais523 Yeah, I was considering doing a bit of volunteering with Jelly

@Adám I won't say no to reviewing a couple of the challenges

Adám

@ais523 Indeed. Last year, we had someone use J too.

Lyxal

20:42

@Adám if I'm able to use Python and other esolangs from around here, I'll be able to help out with some volunteering.

Adám

@Lyxal Absolutely.

@ais523 So, are you in?

Lyxal

@Adám Count me in then

Adám

OK, all that are in, email me: adam@ with the same domain as www.dyalog.com and we'll contact you with an NDA when we're ready for review.

Lyxal

@Adám what should the subject of the email be?

New Main Posts

0

Q: get length of dot from any point on a line

the goal is to make a function which takes in a list or vector of length 4, which represents a line in that the first item is the starting x position of the line, the second is the starting y position, the third is the ending x position, and the fourth is the ending y position, an x coordinate fo...

code-golf

Adám

20:56

@Lyxal "2021 contest volunteering"

caird coinheringaahing

21:09

@Adám Out of interest, why the whole "adam@ + same domain as ..." rather than just "email me at [email protected]?"

Adám

@cairdcoinheringaahing To avoid bots scraping email addresses for spam.

caird coinheringaahing

That makes sense

Also, sorry to whoever has [email protected] for the spam that might come your way

Lyxal

@Adám was my email received?

Adám

@Lyxal Yes, confirmed and added to internal list of beta testers. Thank you.

@cairdcoinheringaahing You're confirmed too.

caird coinheringaahing

@Adám :)

ais523

21:16

@Adám sorry, I can't guarantee being able to dedicate any time to something like this, so you'd probably be better off with someone else testing

Adám

No worries.

Lyxal

Out of curiosity has anyone ever asked a reverse 99 bottles of beer challenge here on cgcc

caird coinheringaahing

How do you reverse 99bob?

Adám

qoq66

Lyxal

You put the bottles back on the wall

ais523

21:20

it'd probably be too similar to a forwards bottles of beer challenge

Lyxal

And you just keep infinitely placing them

ais523

maybe slightly easier because the boundary conditions are easier

but not by enough to make it not a duplicate

Wheat Wizard

In my opinion we already have too many produce infinite output challenges.

We have two right now, and I think the third would be a duplicate of both of them. (I also think they are duplicates of each other but that's another thing.)

caird coinheringaahing

@WheatWizard Don;t we have like 4 challenges which are basically produce some form of infinite output?

Scream, lololol, produce infinite output and I think there's another one

Wheat Wizard

I can only think of 2. But I would not be surprised. There are a lot of bad challenges out there.

xnor

21:30

Can anyone help me understand what OP is saying in the last few comments here? codegolf.stackexchange.com/questions/214262/…

caird coinheringaahing

Personally I'm 100% for closing lololol as a dupe of scream but ¯\_(ツ)_/¯

Adám

@xnor Yeah, me too.

xnor

It looks like they want to put in a run-time limit, which might be good to make solutions less brute-forcey, but unfortunate to edit at this stage

Wheat Wizard

Oh no. Not "real number" input.

caird coinheringaahing

No idea what they're saying, but I reckon they should delete and post over to the sandbox

Wheat Wizard

21:32

The challenge seems very clear to me.

xnor

I probably should have directed them to the Sandbox in the first place rather than trying to quickly fix the challenge.

Adám

Yeah, to me too, except the 1 comment.

xnor

@WheatWizard I rewrote it, but this as always has the risk that OP meant something different or pipes in with new requirements

I do think it would be cool if solutions used actual math rather than testing the distance to every point on the segment up to some precision

caird coinheringaahing

That challenge has a bunch of hidden problems that they won't have anticipated (floating point values, needs test cases, whatever they're trying to say in the comments etc.) that I reckon it should at least move to the Sandbox for a bit of cleaning, then edited and undeleted once its fully clear

Then again, I am on the more cautious side when it comes to posting challenges

xnor

Agreed. Can mods do that?

Wheat Wizard

21:35

Hmmm. I wonder.

caird coinheringaahing

Seems like that might come off as a bit mod abuse-y, especially if the user doesn't want it

Wheat Wizard

I can close it. I don't know about moving it to the sandbox.

Adám

I think it is close enough to OK that we can leave it. No answers yet either.

caird coinheringaahing

I'm not opposed to having it hammered, Sandboxed then reopened, but it is definitely close enough to be easily fixed with some involvement by the OP

Wheat Wizard

In my opinion it is better to close things when there are no answers. Closure prevents answers from coming in when things really get hairy.

xnor

21:37

Yeah, especially since solutions might become invalid with run-time or precision requirements

Wheat Wizard

Closure is also not a huge deal. Very easy to get reversed.

caird coinheringaahing

Especially in this case

Adám

Then we should at least notify the OP about it not being a rejection of their challenge, but just a constructive mechanism working as planned.

caird coinheringaahing

Unrelated: I like finding inactive challenges to compete in, but there are two major pet peeves I have that a lot of old challenges have/use: banning builtins and time restrictions

Adám

@cairdcoinheringaahing Not to mention awkward I/O.

caird coinheringaahing

21:39

Example. Interesting challenge right up until "no builtin functions for prime testing, prime generation or prime gaps"

Wheat Wizard

> Hello, and welcome to the site. I think this challenge is interesting. I am going to close it to prevent any answers while the details from above are ironed out. This is by no means permenent. You can still hammer it out here or in the sandbox. When these have been fixed users can vote to reopen or you can flag it for moderator attention and we can have a check to open it back up

caird coinheringaahing

@Adám I can handle that, because I usually use the current site standards as my main, then include a longer solution that complies with "space separated STDIN input"

@WheatWizard Sounds good IMO

Adám

@WheatWizard Go for it.

ais523

21:50

@cairdcoinheringaahing this sort of thing gets hard to define objectively, e.g. does Brachylog's "decompose into prime factors" builtin count? what about unmultiplication with an unknown?

Adám

Heh, and in some APLs, if you apply product -1 times then you get the factors. Built-in?

caird coinheringaahing

@ais523 Yeah. That's one of my two main complaints against banning builtins (the other being "banning builtins doesn't make your boring challenge interesting, but it does make your interesting challenge boring")

But, it's an old question, so it's not that surprising

ais523

to be fair, that one is pretty trivial with builtins

caird coinheringaahing

There's nothing wrong with trivial questions IMO

ais523

although, "find the indexes of the largest elements so far in this sequence" (which is what it boils down to in Jelly after builtin usage) is probably an interesting building block to write in its own right

actually, those are some of the most interesting things to microgolf, in golfing languages

caird coinheringaahing

21:54

Sure Jelly, 05AB1E or Brachylog might be able to do it in <5 bytes, but trivial questions usually lead to some of the most interesting answers in some languages

ais523

the sorts of things that seem like they should be builtins but are just slightly too complex to have been added to the language (and not just because they're a trivial combination of two builtins and giving them a name wouldn't be any shorter than just writing the original commands)

5 seems low for that, actually

caird coinheringaahing

TBH, I didn't actually give it a go after seeing the ban, so I have very little frame of reference :P

5 was just a low number I chose :P

Just gave it a shot (using prime builtins) and it took me 16 bytes so ¯\_(ツ)_/¯

Half of that was the infinite loop + printing tho

ais523

my quick (and very inefficient) Jelly attempt was 17: ÆRIUMP=Ẓ¶Ṅ1Ç?¬µ1#

actually I can shave off 2 bytes by writing the loop manually I think

yep: ÆRIUMP=Ẓ¶Ṅ¹Ç?‘ß

caird coinheringaahing

My 16 bytes

ais523

22:10

you're supposed to only output the new record largest prime gaps

caird coinheringaahing

Oh right, I misread the challenge

In which case ¯\_(ツ)_/¯ meh

ais523

"You can also show off versions with builtin functions if they are interesting. Be creative."

Wheat Wizard

What a clear rule.

caird coinheringaahing

That's even worse than banning them

"This version uses builtins, but its the shortest, so it's interesting"

Wheat Wizard

Yeah. Just out of curiosity ... what challenge is this?

caird coinheringaahing

22:14

Linked it above, but this

WW quietly modhammers it

Wheat Wizard

Ah yes 2014. That figures.

caird coinheringaahing

Yeah, it's a very old challenge

Redwolf Programs

You can tell because GolfScript is still in the lead :p

caird coinheringaahing

<2015: Golfscript, 2015-2017: CJam, >2017: Jelly/05AB1E

Wheat Wizard

Wth did someone just upvote it? It is a really bad challenge by today's standards.

Redwolf Programs

22:17

Oh, that was a misclick on my part. Glad you noticed that.

Wheat Wizard

Oh it has a bonus too.

Redwolf Programs

Kinda reminds me of this :p

H.PWiz

@ais523 It seems that the answers to the challenge are also printing the primes responsible for the gaps. But if we just want the gap sizes, Husk has 6 bytes: ü≤Ẋ-İp

ais523

oh right, Husk would naturally be really good at this

caird coinheringaahing

@WheatWizard As if it couldn't get worse :P

ais523

22:21

it's the go-to golfing language for infinite sequences

caird coinheringaahing

Yeah, Husk is irritatingly good at sequence based challenges

Wheat Wizard

@H.PWiz You can do either, you get a +10 penalty for just the gap zizes.

caird coinheringaahing

@RedwolfPrograms Oh please tell me that's a photoshop

Redwolf Programs

It is, yes :p

H.PWiz

Yeah. Also to prove ais's earlier point ü≤ is a neat way to get the record breakers

Adám

22:22

-4 on a comment‽

Redwolf Programs

It's part of the joke, I edited it

caird coinheringaahing

> you may take a string containing the english words for the index of an item in the list

But of course, using sWAPPED tITLE cASE, obviously

H.PWiz

The obvious way would be uGY

caird coinheringaahing

@RedwolfPrograms It keeps getting better the more I read it :P

Redwolf Programs

I read the entire list of things not to do and did all of them

Adám

22:25

@RedwolfPrograms Doubles? Too easy. Why not singles?

caird coinheringaahing

What about triples?

ais523

the requirement to violate exactly three standard loopholes is interesting

caird coinheringaahing

> you may not use [...] any C variant (although C itself is fine)

Idk why, but that got me

Wheat Wizard

I do wonder which 3 loopholes are easiest to violate simultaneously.

This challenge is so complex, "that interpret the challenge literally" may not be super helpful.

caird coinheringaahing

Homoglyphs would work well with any two

ais523

22:27

whitespace to hide things offscreen is free, you can mix that with pretty much anything (or refuse to use it)

Redwolf Programs

Aw, triple precision floating point doesn't actually exist. Apparently octuple precision does, though...

ais523

especially as the challenge is underhanded

Adám

@WheatWizard Hm, that could make an interesting challenge; a trivial task (e.g. cat) where you must use 3 loopholes.

Wheat Wizard

It is a bit of a meta challenge.

Redwolf Programs

Wait, it actually would be fun. Is there any way it would be allowed?

ais523

22:28

one thing that's got me thinking is simply trying to find a language that has pointers, and yet can distinguish a double from a float (the former normally implies statically typed and unboxed, the latter dynamically typed and boxed)

caird coinheringaahing

Using a made up language which uses a Cyrillic a instead of an a would violate two

Adám

@RedwolfPrograms Why not, if well-specified?

Wheat Wizard

It might be more interesting if the challenge is clearly impossible.

caird coinheringaahing

@RedwolfPrograms It's allowed, it just might not be well-received

Adám

@RedwolfPrograms Sure they do!

ais523

22:28

the "violate n loopholes" challenge could be solved pretty easily by inventing a language for the purpose in which the zero-byte program solves the challenge

Redwolf Programs

The only problem I can think of is the requirements changing when new loopholes are added or voted on

ais523

that's one loophole, and you can easily tune how many others it violates by changing the definition of the language

caird coinheringaahing

Ok, I think I've got it: "The challenge doesn't say anywhere that I can't make up my own language where <insert homoglyph here> prints 'Hello World', therefore, I win"

Redwolf Programs

What would the best scoring criterion be?

caird coinheringaahing

popularity-contest

Duh

Wheat Wizard

22:30

Pop-con obviously. Got to go with the worst possible.

caird coinheringaahing

Ironically, it's probably one of the very few challenges where pop-con would be the best scoring criterion

ais523

popcon is at least objective, and actually works well for some challenges

Redwolf Programs

Yeah, true

ais523

you can get the disadvantages without the advantages as "my favourite program wins, code-challenge"

Wheat Wizard

Throw in a bonus.

ais523

22:32

(except that this isn't as FGITW-heavy as a popcon)

caird coinheringaahing

You could mandate that no two answers can use the same three loopholes

Redwolf Programs

"code golf with a -Infinity bonus for most votes"

ais523

"code golf/popcon hybrid: -1 byte bonus for every upvote"

Wheat Wizard

"my favourite program wins, +10 bytes to answers with vowels in the source code"

caird coinheringaahing

code-golf - pop-con

ais523

22:32

(actually I think someone tried something like that once)

Redwolf Programs

Actually I like the no-reusing-loopholes thing, but maybe allow each loophole to be used 3 times

caird coinheringaahing

programming-puzzle "Whichever answer guessing what programming language I'm thinking of wins"

ais523

the "changing your username for a username-dependent puzzle" would be fun, as you could create a username-dependent language to solve the problem

Wheat Wizard

Give an impossible task. You can use 1 loop hole per answer. Each player gets 1 point per shortest answer they have using a certain loophole.

Redwolf Programs

That seems really fun too

ais523

22:34

maybe a rosetta-stone wincon, but with loopholes rather than languages?

caird coinheringaahing

The task would have to be impossible, but would still have to force a few different things (e.g. usernames) to cover all loopholes

ais523

but only a few of the loopholes can be used to solve impossible tasks, and they do so via not actually solving them

it's probably better to have a task that is clearly too difficult, rather than impossible

e.g. kolmogorov-complexity for the entirety of Wikipedia has several ways to loophole it

and nobody is insane enough to do it legitimately

caird coinheringaahing

There are a few you can't cover all at the same time (namely the c-n-r related ones, or the koth ones)

Wheat Wizard

It would be decently hard to do the invented language for that one.

For one your interpreter would be quite large.

But kc with a several gb text might be a bit of a toned down version.

ais523

it doesn't seem that ridiculous, just download a Wikipedia database dump while compiling the interprer

caird coinheringaahing

22:37

TBH a very basic task might do well, as it could lead to some very stupidly funny answers

Redwolf Programs

Someone suggested cat, I think that's a good one

caird coinheringaahing

Also, dibs on exactly duplicating an answer

ais523

oh, it's probably worth mentioning that kolmogorov-complexity for the first 100 MB of a Wikipedia database dump is a long-running golfing contest that you can get cash prizes for

Redwolf Programs

Maybe reverse the input, since some langs' empty programs are cat

Wheat Wizard

The funny though would be the legitimate golf task to download all of wikipedia.

ais523

22:38

as long as you beat the existing records

Wheat Wizard

Is that in a particular language?

ais523

err, first gigabyte, I think

prize.hutter1.net/hrules.htm

and yes, x86 or x86_64 Windows or Linux machine code

although, the expected size of the program is so large that bundling an interpreter with the program isn't likely to cost all that much

Redwolf Programs

Ooh, like a jelly interpreter with pre-loaded jelly code?

caird coinheringaahing

Given the diversity of the loopholes, maybe an answer chaining could work, where the next answer has to follow a one sentence task specified in the current answer?

Redwolf Programs

I really like that idea, actually

caird coinheringaahing

22:40

The lack of specificity could lead to loopholes being used more creatively

Wheat Wizard

That seems quite insane

I don't know how you could have any reasonable scoring criterion

caird coinheringaahing

Pop-con

Or most answers

Redwolf Programs

TBF a challenge based on exploiting loopholes should be quite insane

Wheat Wizard

Ah yes pop-con. The scoring criterion for when you can't think of good one.

caird coinheringaahing

11 mins ago, by Wheat Wizard

Pop-con obviously. Got to go with the worst possible.

ais523

22:42

ah, I didn't misremember the size of the Wikipedia prefix to compress; it was increased recently

Wheat Wizard

I'm kind of doing a similar thing with Brain-flak.

I'm trying to make a language that is just really compressed brain-flak.

I'm taking the grammar for brain-flak programs and pruning branches that produce redundant or stupid programs, then using the contents of a file to create a path through the grammar.

caird coinheringaahing

@RedwolfPrograms If you are going to do a loophole-based challenge, you'll have to specify which loopholes are fair game. A lot of them are <5 and twice, or are deleted

Wheat Wizard

But there is also some statistical analysis to make certain operations easier to reach in fewer bits.

Redwolf Programs

Isn't there a requirement that valid loopholes have at least +5 score and twice as many upvotes?

Wheat Wizard

Yeah something like that.

caird coinheringaahing

22:46

@RedwolfPrograms Yeah, but in a challenge built around abusing loopholes, abusing that rule wouldn't be out of character

Wheat Wizard

There was at one point a problem of a certain loophole which was right on the edge of that and kept flipping back and forth.

Redwolf Programs

@cairdcoinheringaahing There's a difference between insanity and anarchy :p

caird coinheringaahing

"What loophole are you using? Well, I looked on Meta, but decided 'screw the loopholes', and decided to just post this Sandbox proposal as an answer instead"

Wheat Wizard

You might be able to actually construct the liar's paradox with loophole based challenges.

Particularly using the "interpret the challenge too literally" loophole.

caird coinheringaahing

@cairdcoinheringaahing Or, to frame it as a loophole, "nowhere in the challenge did it say I wasn't allowed to just post anything I wanted to"

Wheat Wizard

22:49

For example if the challenge were: Print "Hello world!" using exactly one loophole.

Does: print "\"Hello world!\" using exactly one loophole" count as interpreting the challenge to literally?

Adám

Which is the simplest code golf challenge that cannot be fully specified in 100 characters or less?

caird coinheringaahing

Yeah, looks like it to me

@Adám You know how much we like our objectivity, define "simplest"

Adám

@cairdcoinheringaahing Shortest possible spec.

ais523

surely it'll be a tie between all the challenges which require 101 characters to fully specify

:55968141 Jelly, 1 byte: ß

(note: this works as a specification of the challenge, in addition to being a solution to it)

Adám

@ais523 No.

caird coinheringaahing

22:53

@ais523 That wouldn't actually loop forever tho, due to recursion limits no?

ais523

I think it's tail-recursive, not sure though

Wheat Wizard

What is the shortest challenge spec in which no one can come up with a valid answer shorter than the spec?

Redwolf Programs

"factor RSA prime shortest wins"

Adám

@WheatWizard π

Wheat Wizard

I don't think that is a clear challenge.

caird coinheringaahing

22:55

"last digit of π, first wins"

Adám

Print π backwards.

Wheat Wizard

Ok limited to code-golf, and challenges that can be solved.

Redwolf Programs

Or, an actually solvable one: "output wikipedia. shortest wins"

ais523

I think in order to make that work, the challenge needs to talk about some concept that's common knowledge but not a builtin in anything

Wheat Wizard

VTC as unclear. Wikipedia is constantly changes.

It's not too hard.

caird coinheringaahing

22:57

"ζ(s)=0, Re(s)=1/2. Get s or 0 if no s"

Wheat Wizard

"Output a number which cannot be output by a Jelly program less than 999 bytes."

ais523

I'm not convinced that's solvable with certainty

Wheat Wizard

You can prove busy beaver numbers correct.

ais523

you can solve it probabilistically if you have a sufficiently high-quality source of randomness

Adám

"Solve any task requiring code longer than this text."

ais523

22:58

but Jelly is Turing-complete so you can't compute busy beaver numbers that are too big for it to produce

caird coinheringaahing

Even with the 999byte constraint?

Wheat Wizard

You can prove a that a particular number is the 1000th jelly bb.

It is hard. Really hard, but can be done.

Transcript for