The Nineteenth Byte

ASCII-only

00:15

@neil wait a second, I just realized every multidirection is inconsistent

Neil

00:28

actually, it would make a difference in a reflect, but I don't know of a use case

Wît Wisarhd

03:09

ANyone would like to be playing strupremum with me? I have an itching for this game.

Pavel

dammit wheat

how long have you had this name

Wît Wisarhd

at most 9 days

But likely less

Oh cool today is my 888th consecutive day on the site

Pavel

nice

Quintec

05:01

Wow

Quintec

07:47

Is everyone asleep xd

Esolanging Fruit

Just finished my Mascarpone interpreter

Merry christmas to everyone in California!

New Meta Posts

08:50

0

Q: Change the use of the "Game" tag?

I was scrolling through the posts tagged with "Game" and noticed that it involved quite a few code golf competitions. Just for the sake of clarity and ease while searching, I believe that the definitions of some tags should be clarified to be as follows: Challenge: A play thing with an objectiv...

Neil

09:55

ugh, I forgot I told Windows to reboot as a Christmas present to me

Adám

10:12

@Neil I tried installing 1809. Had to revert to make Windows able to boot again…

New Main Posts

0

Q: Eight coins for the fair king

This is a "counterpart" of another puzzle, Eight coins for the fair king on Puzzling.SE. You can read the above puzzle for background. The details about this puzzle is as follows. Now that a set of 8 kinds of coins of different values are created, the king wants you to find out the maximum N su...

code-golf number-theory integer

ASCII-only

10:40

@Quintec ~~charcoal's wolfram language things~~

@dzaima yes. at the very least, don't make it so insanely verbose

@Anush 1. google sells domains? and 2. you use cname records

@Neil should I change all the multidirections so they're actually consistent sometime? If so, in which order should the directions be?

ASCII-only

11:18

Oh yeah. @Neil do you think it would be worth adding ternary (condition ? true : false form) and subscripting/slicing (a[b:c:d] python form) to verbose mode

RedClover

14:03

There was once a nice list of predefined colors in label editor on GitHub, why did they replace it with a button to get a random color from it? Is there any userscript to restore the previous color picker?

Neil

14:33

@ASCII-only I'm only ever going to use functional verbose mode, so do what you like

@ASCII-only I'm pretty sure I've got some answers that depend on the current order of multidirectionals as used when drawing polygons

flawr

15:04

Is anyone aware of answers that used a non-programmable calculator? I'm interested in the scoring method.

Erik the Outgolfer

@Neil um... man, you're in for serious trouble... Python beats JS

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

There are 1000 phones in a city. Add some switches so that we can freely decide linking pairs(Any amount of pairs). Two phones are connected iif some closed switches diretly or indirectly link them You need to answer with: A list of links, where 0~999 are the 1000 phones, and 1000~ are just po...

Neil

@EriktheOutgolfer let me know as soon as I can import itertools into JS to save some bytes

Erik the Outgolfer

:P

(actually, since you're only 26 bytes over Python, I'd say your answer handles the combinations with replacement part pretty well... if it does so)

also... the reduce(chain,...) part could've simply been sum(...,[]) if itertools.combinations_with_replacement returned a list instead of an itertools.combinations_with_replacement

waits for Arnauld

Adám

16:59

CMC: Total number of gifts during N days of Christmas.

Leaky Nun

17:11

@Adám Jelly, 4 bytes: RRSS

DJMcMayhem

@Adám Brain-flak, 38 bytes: {(({})<({({}[()])()}{})>[()])}{}({{}}), can probably be golfed better

Wît Wisarhd

17:26

@DJMcMayhem Correct

DJMcMayhem

Ah, of course

Dennis

@Adám Jelly, 4 bytes: +2c3

Wît Wisarhd

question: Is it possible to get a Brain-Flak program that grows better than quadratic using only one loop?

question: What is the fewest number of loops needed to create a function that is non-polynomial?

Ok you can do fibonacci in 1 loop. So the answers are yes and 1.

Adám

17:42

@Dennis What‽

FrownyFrog

the arguments go in arguments

Erik the Outgolfer

@Adám it means (n + 2)C3

or 3!+∘2 in APL

Dennis

@Adám 3!2+⊢ in APL.

Adám

@EriktheOutgolfer Ah, so 3!+∘2

Erik the Outgolfer

double ninja lol

Dennis

17:45

Or that.

Adám

Double ninja'd.

Triple.

trichoplax

CodeReview.SE gets a custom behaviour - maybe there's hope for our customisation requests...?

Erik the Outgolfer

lol Shog be like "I just sat down and wrote a new check for the purpose" after >2 years

New Main Posts

18:14

0

Q: Operation Zalgo

This is a free style question, where your program must write itself into another file with additions that must be valid. t͍̦̩̱̖̻̻h̡͕̤̹͂͐̓ͩi͈͕̓̔̀̔̔̍̒͝šͣ̑̓̉̄ ̯̺̳͕̞̠͎ͧͥ̃ͨͦ͛i̻̭͙͙̱̍̋̃̂ͅs̸̬̼͚̣̫̟̝̓̎ͫ̎ͣ ͤ͝z̼̬ͭͭ̎̌͑ͦ́͝ͅá̶̦̜͔̚l͕͔͈͆̓ͯ̆ͯ̚g͓̲̳̞̎o̘̞̯̝̟ͨͣ̈́̽̀ The Task: Your code must write itself with z...

self-referential zalgo

Mr. Xcoder

19:12

@Mr.Xcoder sure thing...

Neil

19:23

@EriktheOutgolfer actually I'm down to only 2 bytes over Python now

Erik the Outgolfer

yeah saw that a few hours ago :P

nice work :D

dzaima

20:58

@ASCII-only yeah i wanted to make it less verbose, but beyond replacing t.objectWrap.classProp with "class" and having sane unique name generation there doesn't seem to be much to do, beyond replacing all news with methods

ASCII-only

22:06

@dzaima ... you shouldn't need to type new CArrow(blankThis.castAs(t.objectWrap.type()), t.objectWrap.classProp) to generate ((Obj) (this))->class

@NewMainPosts :|

dzaima

@ASCII-only I want to allow easily globally changing names of things, e.g. if at some point I wanted to streamline the Obj type. same with the class property. and mixing raw code & specific properties feels strange

ASCII-only

@dzaima raw code and specific properties?

dzaima

@ASCII-only class & Obj = specific property, ((...) ...)-> = raw code

ASCII-only

oh also ew new CNumExpr(1)

also. CALLOC.call? 1. why is CALLOC in ALL CAPS 2. call? does that mean there are other useful methods on whatever that is

dzaima

@ASCII-only what would you suggest in place of that? just 1 would make me have to have the methods 2nd parameter be Object so in place of it could also be an expression which is very very very horrible

ASCII-only

22:13

@dzaima ... Java's overloads are terrible then???

dzaima

@ASCII-only what would you want a methods signature to be for it to be able to receive both a number and a specific object‽

@ASCII-only all caps because constant, and it's an object so i have no option but to have a pointless .method()

ASCII-only

@dzaima specific object?

@dzaima ew. use a lambda at least. also i don't really count complex objects as constant, especially if they represent functions, but ok

dzaima

@ASCII-only what where would a lambda even be usable there

ASCII-only

@dzaima make CALLOC a lambda that does what CALLOC.call currently does maybe?

Java lambdas are only syntax sugar for class instances after all...

dzaima

@ASCII-only to call a lambda you still have to do lambdaFun.itsonlymethod(...), and yes that is annoying

@ASCII-only e.g. there's a method call. what would its definition look like for both call(1) and call(new CExpr()) both to work and not be ugly as hell

ASCII-only

22:18

@dzaima ew :/

dzaima

@ASCII-only does c# provide something better?

ASCII-only

@dzaima call(int) and call(CExpr)...

@dzaima yes. actual, first class (?) lambdas...

dzaima

@ASCII-only so have 2 different methods‽

ASCII-only

heck, you can even make methods look like lambdas

@dzaima yes. overloads are a thing, you know

dzaima

@ASCII-only oh yeah that answers my question

@ASCII-only but that's a very very very very very pointless overload

ASCII-only

22:20

@dzaima saves having to type new CNumExpr(1) every. single. time.

dzaima

and internally that overload would still have to do the new CNumExpr()

ASCII-only

ofc

but less clutter/boilerplate = more readability

and that's what's important here imo

like... you've already shortened the type names to the point of unreadability

dzaima

should i also have an overload for two operands to addition passed as an arg to that? how about multiplication? how about both raw & expr variants for args of those?

ASCII-only

@dzaima no need. just basic literals

dzaima

@ASCII-only why are they an exception

ASCII-only

22:22

because you can type literals in Java

but... I guess you can't really overload every single function :/

dzaima

should there also be an overload for receiving a string there? it's syntactically valid, though doesn't make sense for calloc

ASCII-only

I'd say no, but... brb trying something

dzaima

ok though i guess specializing builtins would make sense, as sizeof is already specializing as its parameter has to be a type not an expression which is very weird

ASCII-only

java doesn't have this right

dzaima

@ASCII-only yeah, but imo that's ugly and as it itself states, very implicit

also what

imo implicit things don't fit in OOP

ASCII-only

22:32

@dzaima uhm... of course. i haven't defined ToString for Foo...

dzaima

java is that idea taken too far

ASCII-only

@dzaima exactly

OOP isn't always the solution :/

no need to pointlessly restrict people to a single paradigm

dzaima

@ASCII-only but imo in the case of a wrapper around the 1 for the parameter, it is

ASCII-only

@dzaima ?

dzaima

@ASCII-only imo the new CNumExpr(1) makes perfect sense

@ASCII-only but the fact that new Foo(3) did 2 completely different things as an operand to +

ASCII-only

22:35

@dzaima because the first one casts to int maybe

dzaima

that numbers do that already is very strange, extending that to everything is even stranger

ASCII-only

@dzaima nothing does that unless you tell it to though?

dzaima

@ASCII-only allowing everything to do that if it wants to* is even stranger

ASCII-only

suuuuure. but i sure as hell am not going to do "foo " + a.toString() + " bar " + b.toString() every single time

dzaima

@ASCII-only yeah, i agree, that's why + should cast to strings in the event of confusion

ASCII-only

22:38

@dzaima confusion?

dzaima

@ASCII-only confusion = both args arent numbers

ASCII-only

@dzaima that... that is even worse

what about.... "foo " + aNumber.toString()...

~~"foo " + 1 == "foo!"~~

dzaima

really i'd prefer if the char for addition and concatenation wouldn't be the same, but that's not really an option

ASCII-only

@dzaima not really an option?

dzaima

@ASCII-only what's your proposal that wouldn't make 50% of people hate you?

@ASCII-only what about that? that'd be identical to "foo " + aNumber

ASCII-only

22:41

you do know plenty of languages have user-defined operators right

> both args arent numbers

dzaima

@ASCII-only oh i meant !(both args are numbers) :|

@ASCII-only yeah, i've been thinking about how to do that in a way that isn't very very horrible and just can't

ASCII-only

@dzaima those languages are widely used though

link

dzaima

@ASCII-only ok ill be fair i haven't used operator overloading in pretty much any language

ASCII-only

not a lot of languages, but most are pretty popular

Veskah

You haven't lived until you overload all your arithmetic operators to do utter nonsense

ASCII-only

22:47

@Veskah only arithmetic?

Veskah

@ASCII-only Only arithmetic

ASCII-only

:(

hmm. anyone know how to golf turing machines?

dzaima

23:06

is the implicit casting what makes this work?

ASCII-only

@dzaima either that or operator overloading, yes

dzaima

so a class Foo can choose to overload int + Foo and Foo + int? what happens if both Foo and Bar want control of Foo + Bar?

Adám

@dzaima We'll have a system that dictates whether the left or right argument is stronger.

dzaima

oh so it seems to error only if there's a Foo + Bar in the code

@Adám that'd be a dictation per operator or per implementer?

Adám

@dzaima Per language ;-)

dzaima

23:20

@Adám and you've got a whole system for that? as far as i know, there's only one Dyalog APL :p

Adám

@dzaima Just a simple rule. E.g. where Foo defines Foo + y and Bar defines x + Bar, Foo's code will be run.

That being said. The project currently only has "proposal" status, so things may change. Let me have a look if it speaks about that in detail. Hang on…

dzaima

@Adám ah, so saying whether the left or right arguments type is stronger would've been clearer

ASCII-only

@dzaima nah, depends which one defines an overload, i think?

dzaima

@Adám would that system allow specifically overloading Foo + Bar?

Adám

@dzaima It says right side wins.

dzaima

23:30

@Adám oh hmm, personally i'd prefer an error over an arbitrary choice

Adám

@dzaima Isn't a rule akin to APL's one-rule-for-all-functions order-of-evaluation over an arbitrary hierarchy?

@dzaima No, it doesn't seem so from the current description. Rather, Foo would say :Implements ⍶+⍹ to define + for two magic arrays (obviously, one has to be a Foo).

dzaima

@Adám oh, can it not implement Foo + normalArray?

Adám

@dzaima Sure, but then it has to have :Implements ⍶+⍵

dzaima

@Adám ah, that's why the underbars

@Adám ah if the operator implementing is that simple, i guess that makes sense

ASCII-only

:/

328 — ASCII-only 1 min ago

i abused the interpreter so much

Adám

23:36

@dzaima Yup, and even more magically (!) what's on the right of :Implements can be any expression, e.g. ⍺((⍺⍺¨⊂)⍤1)⍹

ASCII-only

@Adám :O nice

dzaima

@Adám oh what

Adám

Well, there are few restrictions, currently:
∘ It is a single derived function applied to one or two arguments.
∘ The arguments can be any of `⍺` `⍶` `⍵` `⍹`, with underlines for magic arrays, or constant values. At least one argument must be underlined.
∘ No argument token can be used twice, including with an underline.
∘ `⍵` must be used. If `⍵⍵` is used, `⍺⍺` must be as well.

dzaima

if i did :Implements -×⍹, would -(×obj) call that?

Adám

@dzaima ^^ -×⍹ isn't valid, but (-×)⍹ is.

dzaima

23:40

@Adám ⍵ must be used? don't you mean ⍹?

Adám

@dzaima Oh, it probably means one of ⍵ or ⍹.

Obviously, you can't have both or both ⍺ and ⍶ either.

dzaima

@Adám what would a :Implements look like with no magic arrays?

Adám

@dzaima That wouldn't be meaningful, as the code would never be called. Clearly, if there is no ⍶ there must be ⍹, and if there is ⍵, there must be ⍶.

dzaima

also why the restriction of only one of ⍵&⍹ and ⍺&⍶?

Adám

@dzaima :Implements has a phrase containing one (derived) function applied to one or two arguments. A function can only have one left argument and one right argument. Any argument may be magic or not, but it can't be both magic and not magic at the same time.

Here's an example of what it could look like:
https://tio.run/##dVdPTxtHFL/zKaa3HbHGOOrJSVWhikSWYhKFSDkQDst6sIfuzm521sEIuJCKghNHTauUXipVUQ5WFCmHhKa11Iv5JvtF6HtvZv8YA4fIO/Pe7/37vTcvXhzUOnteEHUvL5ttryv9lSTx9th67CVaLDCWjf5kdM48unDUEm@yFRYKv@cpqUO2tcd2e9LvMU@xlYf3WZxE3cQLQ5EwH46iOGVpZJHEIBaJFMoXTEehYNE2S3uCxZ5U@HtLSNXNcWQYByIUKo2SpQXUf/229SA7/mXZYq34vtCa@ZFKkyj4Hk6bd6UIOmy958Wi/GypjgRJo9QCODKZeM9FwBwdJanouKyv5LO@4KXWqrGtSeuHKEmEjiNAAv8AQRrMUvyuDAKGfyjurCo/iLTocCYMDGRIKLSbCCY1U4QgBhTWyc9sTew624hAOea3Q0//uIBoTRvkw/5WIH1z1MrzosEvpdOk70OK6I4ihxRloy@ERIfoGpwhPn0jOHy7eJCd/lUK2kSR/j9u4UOeCTjHs7qba4DnG…

ASCII-only

23:57

@Downgoat why not use fn in VSL HW (actually, are you ever even going to remove fn)

dzaima

:Implements ⌽(⍤⍺)⍹?

Downgoat

@ASCII-only 1. VSL HW? 2. I think I'll leave both so people can choose which one they like. Might throw warning if people mix different fn/func/function

Transcript for