The Nineteenth Byte

4:00 PM

I like the completely useless findall there :-)

(that's the sort of thing that golfing languages can get away with, incidentally; having a command implicitly do something that's usually a no-op but, when it does something, will normally be useful)

user202729

But the formula is not there.

ais523

@Fatalize: incidentally, bagof is much more intuitive for me than findall; perhaps I should start using it by default

NieDzejkob

@user202729 well, crap

user202729

Just take a quick look over the paper, seems that I have to actually learn and not just "here is the formula, use it in the challenge"

Zgarb

@ais523 I don't know Prolog, so that's not meaningful to me. :P

user202729

4:05 PM

Anyone here know German?

Zgarb

Hmm, do you mean that findall produces the norm as the only possible result and wraps it into a list, which is then zipped with the input? If so, that does seem useless.

Mr. Xcoder

@user202729 I know a bit. Martin Ender and Dennis, on the other hand, speak it fluently

user202729

It's better not to annoy them. (interpret "annoy" here as "ask to do something not really necessary")

Fatalize

@Zgarb That's what it does yeah

So here we abuse the fact that there is only one possible result, which will then get properly zipped with the input

user202729

Google Translate is really useful for translating any language.

ais523

4:20 PM

@Zgarb: the main weirdness in Prolog is that you can use a value before you define it

findall/bagof differ in how they treat undefined values; findall will give you "fresh" undefined values if there's an undefined value in the the output of the thing you're finding all outputs from, bagof will attempt to share values that seem to be the same

but it's really unintuitive both ways so it's hard to describe exactly

@Fatalize: output that I discovered while doing experiments on undefined values, and I don't get it at all: 1000>ℕẹ outputs [0,1,1,0,1,0,0] Try it online!

Fatalize

that's.... interesting

Zgarb

@ais523 Does it have something to do with Brachylog's ≜? It seems to force a definite value on something based on the docs, but I don't know what or why.

Laikoni

@user202729 Yes, but no chemistry. But feel free to ask me anyway.

ais523

@Zgarb ≜ is a weirdness all of its own; not in what it does (which is pretty sane) but in why it's needed

basically it causes Brachylog to brute-force all values in its input that haven't received a definite value yet

to find a set of definite value that work

as opposed to the normal algorithm which carts unknowns around

Fatalize

@ais523 From the debugger, I think that's because ẹ implementation doesn't work properly for uninstantiated integers

because it uses findall which makes it find different possible answers which leads to this

ais523

4:32 PM

oh, ẹ having a broken flow pattern would explain a lot

I was actually expecting an infinite loop

which is the most common breakage in that case

but I guess it's doing something much funkier

Fatalize

you get the same result if you replace it with a Brachylog equivalent of how it's implemented in Prolog

ais523

oh, haha, it's literally just ∋ᶠ?

Fatalize

well, the Prolog equivalent yeah

ais523

I tried replacing it with ∋ᵇ and it failed, which is possibly an even more interesting result

Fatalize

It's explainable that way: a digit of a number between 0 and 999 can be: 0, or a non zero digit if it's 1-9, or a 1-9 digit then a 0-9 digit if it's 10-99, or 1-9 0-9 0-9

then it labelizes automatically at the end of the program leading to 0,1,1,0,1,0,0

Elementary really :p

ais523

4:42 PM

ooh, if you read it as "0, 1, 10, 100" it makes much more sense

four results that should be separate, but due to the findall, have somehow been flattened into one

Fatalize

much more than no sense at all, which isn't much :p

HyperNeutrino

hm I have 15 days left until Fanatic and 41 code-golf score left until code-golf tag badge. which do you think I'll get first? :P

Neil

15 days in which to get 41 upvotes? that's less than 3 a day on average

HyperNeutrino

I guess that depends on how much I post answers :P

the most recent time I've gotten any more than 3 score in a day is nov 16 :P

betseg

A number p is a prime number if it divides all the terms in the pth line of Pascal's triangle. (AKS test) CMC: given integer p, output the first term/terms p fails for if p is composite, output something else that can't be confused with an index otherwise.

Also, can it be a challenge in the site?

Pavel

5:00 PM

@HyperNeutrino Do votes on challenges count for tag badges?

HyperNeutrino

no

ais523

hmm, reading about that: what do you even call an algorithm with O(n¹²) running time?

(no wonder people said that AKS was slow…)

Mr. Xcoder

@HyperNeutrino Oh lol JS ties Jelly?!

(I assume you know what I am talking about)

HyperNeutrino

string challenges :P

Mr. Xcoder

@HyperNeutrino definitely code golf

Pavel

5:11 PM

I'm honestly not sure if I even have 41 code-golf score

Most of my rep is from challenges

Mr. Xcoder

@Pavel you have 142.

Pavel

Huh

Mr. Xcoder

And bronze code-golf...

Pavel

That is surprisingly high

TIL I have answers with more than 3 votes

When did that happen

Mr. Xcoder

A lot of which comes from this

Pavel

5:16 PM

@Mr.Xcoder That is a very crappy answer

ais523

now I'm wondering where Pavel's 4 downvotes came from

Pavel

@ais523 How do you see how many downvotes I have?

ais523

@Pavel I'm not surprised, bad answers are normally worth more than good ones

@Pavel I don't, but your reputation in the tag ends with 2, and upvotes give you rep in chunks of 10

so the 2 must be from an upvote minus four downvotes

Pavel

Or 9 downvotes. Or 14 downvotes.

ais523

incidentally, that question is a bad question because it's way too broad

Pavel

5:19 PM

The minecraft one has -2

ais523

huh, that question's especially weird because the accepted answer doesn't sort first

has something changed recently?

Mr. Xcoder

@Pavel People vote in a strange manner

If they wouldn’t, I am pretty sure I would have half of my current reputation

Pavel

@ais523 here's another -1: codegolf.stackexchange.com/questions/111380/…

ais523

basically, the SE voting system is optimized to reward answers that are amusing, probably not a great fit for the question, and require little thought

@Pavel I guess that was from someone who disagreed that you were using a programming language

although I don't see why it's any less valid than vim is

Pavel

@ais523 Technically it isn't, it's just unusable for almost all challenges.

ais523

5:25 PM

right, I was disagreeing with the person who disagreed with you

whoever it was

Mr. Xcoder

@ais523 ... the only special case is when the answer outgolfs another answer by a lot shortly after it has been posted (basically FGITW-outgolf)

ais523

it needs to be very shortly

Pavel

Hmmm....
I wonder if I can claim Visual Studio + C# as my programming language in order to get the import statements it inserts into C# classes by default for free.

ais523

you can comprehensively outgolf an answer and still get many fewer votes

maybe it depends on how many answers are posted in between

Mr. Xcoder

That too, but it’s weird anyway

ais523

5:26 PM

@Pavel there was a debate about that a while ago when it comes to BASIC; my guess is no, unless it also does it when reading from a file

I guess you could potentially score in keystrokes, but that sounds like a nightmare

Mr. Xcoder

And to be honest, does this deserve 8 upvotes? I just chained two built-ins in Python

Pavel

@ais523 Scoring in keystrokes is my thoughts exactly. The shortcuts to creating a new project are less than typing using System.Linq;using System.Collections.Generic;...

@Mr.Xcoder I can downvote it for you if you want :P

ais523

huh, I just realised that not only does "add a language to a polyglot" have 158 answers, all of them are at at least +3

this is quite possibly unique on the Stack Exchange network

Mr. Xcoder

@Pavel I don’t like downvoting for no reason either, but I just al unsure of how some members of the community vote

Fatalize

Tip: stop worrying about whether things are deserved (valid for PPCG and life in general)

Mr. Xcoder

5:29 PM

Why?

ais523

@Fatalize the reason I'm concerned is not due to the undeserved rep, it's because I want high-quality answers but the site tries to persuade people to post low-quality answers instead

if the incentives were different, the questions and answers would be more useful

Fatalize

I would bet they wouldn't be

there would be as many good answers, just fewer trivial ones

Pavel

Applies to challenges too. Like how codegolf.stackexchange.com/questions/123685/covfefify-a-string has several hundred votes.

ais523

@Mr.Xcoder that's the first time I've seen a nontrivial challenge where the Python and Jelly are literal translations of each other

Erik the Outgolfer

@ais523 that was an issue, (/ on empty list is error) before a new rule (no empty input) was added

Wheat Wizard

5:36 PM

@ais523 It's very likely that there are 3 users that have upvoted every answer. I know that I pretty much upvote every answer I see there because they do take quite a bit of effort to make.

ais523

well, yes

but I can't deny that this is very different from the normal SE voting pattern

Wheat Wizard

That's certainly true

ais523

@EriktheOutgolfer I wonder if that got added in response to me trying to add the opposite rule?

incidentally, in the new golfing language I'm working on, almost every binary operator comes with a default output for if it's reduced over an empty list

Wheat Wizard

Especially since the challenge is nearly a year old

ais523

to avoid this sort of problem

Erik the Outgolfer

5:38 PM

@ais523 btw you can't just suggest an edit like that

that's a very old issue with how SE works, and a reason why you may hate it

ais523

hmm, now I'm wondering which question on all of SE has lead to the highest total reputation added to the system

Erik the Outgolfer

that's possibly why it got rejected

ais523

@EriktheOutgolfer you can technically suggest pretty much any edit you want, especially if your rep is low enough that the OP sees it first

Erik the Outgolfer

huh? how can you be sure OP sees it first?

the issue is that you need 50 rep to comment, which is, in my opinion, a bit wrong

ais523

when my rep got high enough to edit anything, I ended up having to use comments to ask the OP if I should edit, and rather slowing things down

Wheat Wizard

5:39 PM

@ais523 Might be show case of languages. Sounds like its a good time to make a SEDE query

ais523

@EriktheOutgolfer you can't for certain, but the OP gets pinged and nobody else does

so if the OP's online they're likely to be the first person to see it

Erik the Outgolfer

"if the OP's online" that

ais523

and if the OP sees the edit, they have final say regardless of what the review queue says

Erik the Outgolfer

basically, I'm not really sure of a reason to limit comments to ≥50 rep

@ais523 also there's a possibility your IP gets banned from editing if you get too many rejects, so it doesn't entirely solve the problem

ais523

@EriktheOutgolfer I think it's because people kept on using comments in ways that SE didn't like

Wheat Wizard

5:42 PM

It's very strange that while I can make any edit with impunity I can't accept any edit with impunity. I mean the system already trusts me to make edits that are helpful why does it not trust me to choose which edits are good? The whole edit system is a little screwy.

ais523

this backfired, because it caused people to start using answers in ways that SE didn't like because they couldn't post comments, and people still don't use comments in the intended way pretty much ever

Erik the Outgolfer

well, at least answers can be converted to comments by mods

AdmBorkBork

@ais523 The recent "why does man print gimme gimme gimme" has had a gigantic influx of rep. I'm sure there have been days of repcap from that.

Wheat Wizard

The obvious solution is to have a comment review queue. /s

ais523

@AdmBorkBork that's why I said "added the most rep to the system", very popular answers like that just give you a few days of repcap and so are only worth the equivalent of around +60 or +80

whereas with slow-burning questions, basically every upvote counts

Erik the Outgolfer

5:45 PM

@WheatWizard that's what I thought of too, but pretty much an /s because SE won't really work with that, and we should try not to add even more to SO

@AdmBorkBork what about this?

or maybe this

AdmBorkBork

@EriktheOutgolfer No, you have to think bigger. Like this one.

ais523

@EriktheOutgolfer I know, but as far as I can tell, suggested edit results are indistinguishable from randomness anyway

Erik the Outgolfer

@AdmBorkBork ninja

ais523

(nearly every time I suggest an edit it goes 2:1 rather than 2:0)

Erik the Outgolfer

@ais523 your IP remains the same except if you use a VPN or something?

Wheat Wizard

5:47 PM

Are we counting bounties as adding rep to the system? If not does a halved bounty count as a loss?

ais523

@EriktheOutgolfer no, it's very dynamic

I've never been edit banned, I just happen to know the ban exists

AdmBorkBork

@EriktheOutgolfer by a split second :p

ais523

@WheatWizard a halved bounty is a loss, yes

actually that's one of the very few ways to remove rep from the system; the others are downvotes and account deletions

Erik the Outgolfer

@WheatWizard well, bounties actually go to the answers though

@ais523 and red deletions

like, a post being deleted as spam or offensive

ais523

@EriktheOutgolfer how often does that happen to a user with 2+ rep anyway?

I forgot that rule because I can't understand why it exists

Erik the Outgolfer

5:49 PM

to discourage you from posting such stuff?

ais523

actually, even regular deletions refund votes, which can remove rep from the system too, but it's just rep that was added from that post

I'd have thought the threat of a ban would be enough reason to not post spam or offensive content

(technically, either a suspension or an account destruction/deletion)

Wheat Wizard

I feel like its more to remove privileges than to deter behavior.

Erik the Outgolfer

what prevents you from using a VPN with an entirely new account just to post spam again?

@WheatWizard posting bans do already exist, but I don't think for a single spam post

Uriel

@AdmBorkBork I thought that votes in the magnitude of thousands were only in the backyard of the pre-2012 SO answers like "how do I assign a variable in Python". turns out they're still achievable

ais523

@EriktheOutgolfer this is a problem that a great many websites are struggling with!

Erik the Outgolfer

5:52 PM

well, there's no solution to that unfortunately

ais523

the most frequently seen solutions that are at least semi-working involve a human manually adding anti-abuse rules to stop specific abusers

combined with something like a CAPTCHA to stop drive-by spam that isn't targeted at you specifically

Erik the Outgolfer

although SE does have local network bans, VPNs can get through that

@ais523 unfortunately that could block good posts too

ais523

OK, I looked up the rules for edit bans: you get edit-banned for 5 rejects, 3 accepts cancel one reject, edits more than a week old don't count

@EriktheOutgolfer I tend to be very conservative when adding anti-abuse rules myself

Erik the Outgolfer

to what site(s)?

ais523

the answer didn't say, so I assume it's a network-wide rule but counts per-site

Erik the Outgolfer

5:56 PM

I mean where do you add rules?

not talking about SE

ais523

oh, Esolang typically

I have the power to do it on NetHackWiki too but don't think I've ever needed to

IP bans and per-page edit restrictions have been enough there when I need to fight abusive people

Erik the Outgolfer

well, bigger sites tend to need more than that

ais523

well, I used to be an admin on Wikipedia – can't get much bigger than that! – but didn't get involved in that side of things

DJMcMayhem

@ais523 But if you think about it, it makes sense that it's difficult to remove rep from the system. Rep is just to encourage behavior, not to form an economy of sorts, so as the number of users and the amount of behavior goes up, the total amount of rep in the system should too

(Even though you hate the behavior they're trying to encourage)

ais523

yes, I don't think rep inflation is necessarily bad in its own right

there are lots of problems with the rep system but that isn't really one of them

it's just an interesting measure to look at

Erik the Outgolfer

6:08 PM

"rep inflation" isn't a thing, since rep isn't a currency :p

for example Jon Skeet has over 995M rep over SO, that hasn't caused any inflation at all

ais523

you mean putting bounties to buy answers to really specific questions isn't intended?

Erik the Outgolfer

well, sure

ais523

also, it's generally believed that Jon Skeet repcaps entirely off old answers, so the only benefit to him posting new answers from a rep point of view is to gain bounties

Erik the Outgolfer

but it's not intended to be used as a currency

DJMcMayhem

@ais523 And accepts (+15, not subject to the daily cap)

Fatalize

6:10 PM

@ais523 and validated answers

Erik the Outgolfer

that's why you first have to wait for 48 hours before putting a bounty

Fatalize

......

ais523

oh right, accepts

DJMcMayhem

CMC: what was the most recent day in which Jon Skeet earned less then 200 rep? (Not including giving away bounties)

ais523

I hate those things :-(

DJMcMayhem

6:11 PM

@ais523 For a competitive coding site, I agree. But it makes perfect sense for a Q&A site

ais523

not always; it only makes sense for questions which want a "good enough" answer

it doesn't work for questions which are looking for the best answer

DJMcMayhem

How so?

ais523

because I can't tell, when an answer is posted, whether it's the best answer or not

so don't know whether I should accept or not

Erik the Outgolfer

@DJMcMayhem nov 19

DJMcMayhem

@ais523 So then don't accept. Wait a couple days until activity dies down, and then accept

ais523

6:14 PM

answers on Q&A sites also have a tendency to get out of date, that includes accepted answers too

DJMcMayhem

Even if you make a mistake, you can unaccept and reaccept a new better answer

ais523

I've only asked one question on SO (it was on a work account; don't go looking for it, you won't find it)

Erik the Outgolfer

well, accepting an answer means it helped you at that time, not necessarily helping everyone

ais523

I didn't get any correct answers, but it's possible that one will be found

I would have given rep to the comment saying it was probably impossible, but you can't give rep to comments

DJMcMayhem

@ais523 Honestly, is there anything about SE that you don't hate?

Simply Beautiful Art

6:19 PM

x.x so my class is trying to design high school senior t-shirts

Everything on google is trash

x.x

ais523

@DJMcMayhem the community

that's about it though

I held off on joining PPCG for years because I hated SE so much, but there are some really amazing and creative users here, and some really good questions that are worth looking at

the whole "quest for Tetris" thing was both a) a huge effort that was really worth doing, and b) an utter abuse of the way the site's designed to work

(the question should, based on any sensible reading of the rules, have been downvoted and closed, but I'm glad it wasn't)

DJMcMayhem

@ais523 I don't understand why you think that

ais523

@DJMcMayhem the specific close reason is "too broad": "there are either too many possible answers, or a correct answer would be too long for this format"

it's clear that there's a large range of reasonable answers, and when an answer actually came, it was too long to fit in a single post

so it's a perfect fit for the close reason

Erik the Outgolfer

actually it's an on-topic challenge, since it asks for a solution which can encode tetris in GoL

ais523

(that said, someone did try to close the Sandbox once, and on a pure rules basis, they were probably right…)

quartata

6:25 PM

The latter part doesn't really apply to us at all. People make answers in Lenguage or Unary all the time that are too long to fit in a single post

ais523

@EriktheOutgolfer being offtopic is only one close reason, though

quartata

but they're still valid answers

Erik the Outgolfer

@ais523 that was last year Winter Bash, and it was intended for a hat

@ais523 still, it doesn't seem too broad

ais523

@quartata right, we routinely ignore the SE rules here, and it's the only reason the site functions at all

Erik the Outgolfer

also what quartata said, since SE isn't the most perfect platform for a code golf community

ais523

6:26 PM

@EriktheOutgolfer oh, I assumed it was just a joke, and a pretty good one at that

i found it amusing, anyway

what sort of hat persuades people to go round closing highly popular Meta posts, anyway?

Erik the Outgolfer

none

it's in WB's rules that normal site rules still apply

Simply Beautiful Art

@quartata O.o

Erik the Outgolfer

so abusing privs isn't a thing you can do even during WB

ais523

also, you can fit Unary into a post just fine using run-length encoding

(Lenguage is potentially a different matter, but not normally)

Simply Beautiful Art

OH RIGHT. WB is coming woo!

ais523

6:27 PM

@EriktheOutgolfer you can, just shouldn't

I expect a lot more rules abuse goes on during WB than other times

Erik the Outgolfer

@ais523 last year Dennis unfortunately had to pin a message on the starboard here saying that whoever abuses their privs after that incident will instead get a nice suspension until WB ends

DJMcMayhem

@ais523 I don't think anything here can be closed because there's a large range or reasonable answers, otherwise nearly every question here could be closed for that reason since A) We allow duplicate languages, even if the code ends being very similar and B) There are thousands of possible languages you could answer it in. Too Broad means (to a Q&A site) "Your question is not narrowed down to a single answerable problem". For example,

"I'm trying to learn Python but it's confusing. Please help" would be a too broad question. Here on our challenge site, a too broad challenge would be a challenge that isn't narrowed down enough into a single task. For example, something like "Print a string", or "Demonstrate something cool that your language does." QFT is definitely not too broad.

ais523

well, we do interpret most of the close meanings as meaning something other than what they literally say

DJMcMayhem

Additionally, there are answers all around the network that don't fit into a single post. QFT is a bit of an exception given how long it is (probably the longest one I know of), but that doesn't make it off-topic

ais523

clearly you need to golf it down a bit :-D

Erik the Outgolfer

6:30 PM

@ais523 I'm not sure if we can edit the text

Anonymous

@ais523 We follow the "golf your code, not youe explanation" philosophy. The QFT explanation needed to be long.

DJMcMayhem

Yeah, otherwise we would

Erik the Outgolfer

@Mego exactly that; this is code golf, not explanation golf

Anonymous

@SimplyBeautifulArt Crap. My first winterbash as a TNB room owner...

Anonymous

I need more whiskey

DJMcMayhem

6:32 PM

@Mego Irish coffee.

Anonymous

@DJMcMayhem Why did you repeat what I just said? :D

DJMcMayhem

Just a recommendation :P

Erik the Outgolfer

@DJMcMayhem you're not exactly helping here :-D

Anonymous

The joke I was making is, whiskey is Irish coffee.

DJMcMayhem

No, Irish coffee is coffee with Whiskey in it I though

Erik the Outgolfer

6:33 PM

well, Irish coffee is actually whiskey with a WB2016 reference in it

Anonymous

Bah. The coffee is ancillary

Anonymous

@DJMcMayhem It's actually usually coffee + baileys + jameson

DJMcMayhem

@Mego I have good news and I have bad news. The good news is that SE finally changed the stupid review queue icon! The bad news, is that it's now an even more confusing icon that looks too much like the stupid site switcher (a square with three lines on it)

Anonymous

@DJMcMayhem Wow, SE is really dedicated to continuing to downgrade the top bar

Erik the Outgolfer

@DJMcMayhem slow cache perhaps? :p

DJMcMayhem

6:35 PM

@Mego They should add some cute animal icons to it. Maybe a pony and some puppies. I'm not sure what it should do, but people like puppies, right?

Anonymous

@DJMcMayhem Penguin icon for the review queue

Anonymous

I'd actually be happy with that

AdmBorkBork

@DJMcMayhem I think it's an improvement over the I'm-writing-a-journal icon that was there previously.

ais523

OK, so Stack Overflow has no questions with 100+ answers which are all upvoted (SEDE link)

@Fatalize (and anyone else interested): I was interviewed about esolangs recently, here's the interview; it contains a brief section on why I like golfing languages (using Brachylog as an example) although it's mostly about tarpits

Fatalize

thanks for the link

aditsu

6:54 PM

hey folks, my Plain GUI "Hello, World!" challenge has been sitting in the sandbox for some time, any more feedback before I post it?

ais523

I wouldn't be surprised if it turned out to be a duplicate, although being sandboxed for a while without finding one is encouraging

20

Q: Hello World window

Create a window that says hello world inside and nothing else. You can not have any ok or cancel buttons inside. It must be just a plain window that says hello world. I do not care about the contents of the window frame or how the window looks, but inside it must say hello world only. Example: ...

that looks suspiciously like a duplicate to me, unfortunately

aditsu

wat

ais523

there are a few solutions to that one which wouldn't be valid in yours, but most of the solutions are valid to both problems

and the Quomplex solution is unlikely to be beatable without a hello world builtin (what were the odds of there being a language that displays anything it doesn't understand in a GUI window by default?…)

aditsu

yeah I guess it's a duplicate.. I wonder how it escaped detection before...

ais523

finding duplicates on PPCG is really hard

I basically guessed at what the most likely search term was that you were unlikely to have tried, "window" on its own, and that was one of the top answers

but there's no guarantee that that sort of approach will work

aditsu

7:02 PM

I probably should have searched for [hello-world] [graphical-output] which finds it too

Conor O'Brien

7:19 PM

CMC: Interpret deadfish. You may output in numbers or characters and may disregard the spec detailing that a newline should be outputted for unrecognized characters.

ais523

25

Q: Write an interactive Deadfish interpreter

Deadfish is a joke "programming language" with four commands. Because the Esolang page is a bit contradictory and the interpreters on that page don't all work exactly the same, you should implement the following variation: Specification There is an accumulator which is at least 16 bits in si...

code-golf interpreter deadfish

not sure if it's worth having a CMC that duplicates a main site question

although that one requests printing the prompts, like the official interpreter does

Conor O'Brien

my cmc is simpler

and non-interactive

ais523

yep, looking at the question now, it is a bit simpler

Conor O'Brien

besides it's a CMC, I don't think they're subject to "duplication". don't do it you don't want to, just ignore it :P

ais523

let's try it in a language that wasn't there for the original

Mr. Xcoder

7:24 PM

Deadfish, 0 bytes. Put the input in the "code" field

Pavel

@Mr.Xcoder 0/10 input cannot be hardcoded

Conor O'Brien

I do this "cmc" to learn languages, so I have a few on my hard drive somewhere

ais523

or, ugh, was going to do it in Jelly but you can't put increment and decrement instructions in a string in Jelly

and I'm not really in the mood for putting a lot of effort into a CMC to, e.g., encrypt the string somehow to get round the restriction

DJMcMayhem

@ConorO'Brien You could transpile that to brain-flak classic really easily

ais523

hmm, so I was reading about the AKS algorithm, and the first step is checking if a number is a perfect power

Pavel

7:30 PM

#!/usr/bin/env ruby -n
$a=0unless$a
case$_
when?i
$a+=1
when?d
$a-=1
when?s
$a**=2
when?o
p$a
end

¯\_(ツ)_/¯

Anonymous

@ConorO'Brien Actually, 28 bytes: 0@⌠"?uD²■"@"idso"íu@E⌡MΣ'Xoƒ. It transpiles deadfish to Actually and runs it.

ais523

I was curious about how fast that step is, and the algorithm is crazy in a "how dare you write that sort of nonsense in a mathematical paper sort of way

among other things, it uses floating point, which is something I never expected to see in a number-theory algorithm

Mr. Xcoder

Try it online! (Can golf more, but I am lazy)

aditsu

@ConorO'Brien what about setting the value to 0 when it's out of range (>256)?

Conor O'Brien

@aditsu that is necessary, and in the described fashion (only wrapping on -1 and 256)

Pavel

7:37 PM

Ah

Conor O'Brien

4, 598 bytes

aditsu

so.. -2 and 257 are acceptable?

ais523

you can't create -2 anyway, but 257 is acceptable

Conor O'Brien

assuming one instruction resulted in -2 and 257. It can at no point make -2 since you'd have to step through -1

aditsu

you can, via overflow, in some languages

ais523

7:39 PM

the fact that you can "square to above 256" is half the fun of deadfish

also, Deadfish as originally implemented is bignum, I think?

Conor O'Brien

it was originally implemented in ruby, yeah, bignum

number sizes don't matter tho for the cmc

aditsu

anyway, it's already too complicated for a CMC

Pavel

#!/usr/bin/env crystal
$a=0u8unless$a
case gets
when?i
$a+=1
when?d
$a-=1
when?s
$a**=2
when?o
p$a
end

ais523

where's the test for 256 in that?

oh, you're using a u8, that misses the whole point

Conor O'Brien

@aditsu you'd think that, but people have asked much larger tasks as CMCs :P

ais523

7:42 PM

which is that, say, 25² is 625, and definitely not 625%256

aditsu

CMC: beat AlphaGo

Conor O'Brien

Sep 29 at 16:31, by AdmBorkBork

@JohnDvorak CMC: Find a prime larger than 2^74,207,281 - 1

ETHproductions

CMC: Implement Java from scratch

ais523

@ETHproductions I was actually considering setting a subset of the JVM as a main site question

there are a lot of hard things there, but a lot of easy things too

if you get rid of classes, objects and methods, there's still enough there for Turing-completeness and yet the space of what you have to do is fairly small

Anonymous

@ConorO'Brien 2^2^74207281 + 1 (conjecture)

Pavel

7:47 PM

@ETHproductions Bash, 8 bytes: javac $1

Mr. Xcoder

@ETHproductions ಠ_ಠ

ais523

@Mego isn't it widely believed that all Fermat primes have already been discovered?

although not proven

ETHproductions

CMC: Implement Java in Scratch